View - Kowalewski, M. - Virginia Tech

THE PALEONTOLOGICAL SOCIETYPAPERSVolume 8 October 2002THE FOSSIL RECORD OF PREDATIONMichal Kowalewski and Patricia H. Kelley, EditorsSeries Editor: Russell D. WhiteA Publication of The Paleontological Society

THE FOSSIL RECORD OF PREDATIONMIchal Kowalewski and Patricia H. Kelley, EditorsPresented as a Paleontological Society Short Courseat the Geological Society of America, Denver, CO,October 26, 2002.Convened byMichal KowalewskiVirginia Polytechnic Institure and State UniversityPatricia H. KelleyUniversity of North Carolina, WilmingtonSpecial Publication EditorRussell D. WhitePeabody Museum of Natural HistoryYale University

The Paleontological Society Special PublicationsA Publication of The Paleontological SocietySeries EditorRussell D. (“Tim”) WhiteDivision of Invertebrate PaleontologyPeabody Museum of Natural HistoryYale University170 Whitney Ave., P.O. Box 208118New Haven, CT 06520Copyright © 2002 by The Paleontological SocietyAll rights reserved. This entire publication may not be reproduced, stored in a retrievalsystem, or transmitted in any form or by any means, electronic, mechanical,photocopying, recording, or otherwise, without permission in written form from theSecretary of the Society. Diagrams, figures, tables, illustrations, and graphs may bereproduced by photocopying and distributing free of charge for educational purposes, ifproper credit is given to the author(s) and the Society.Printed by Yale University Reprographics & Imaging Services, New Haven, CT.

OFFICERS OF THE PALEONTOLOGICAL SOCIETYFor the year ending December, 2002PresidentPatricia (“Trisha”) H. Kelley, Wilmington,NCPresident-ElectWilliam I. Ausich, Columbus, OHPast-PresidentPeter R. Crane FRS, Richmond, Surrey,UKSecretaryCarl W. Stock, Tuscaloosa, ALTreasurerThomas W. Kammer, Morgantown, WVProgram CoordinatorMark A. Wilson, Wooster, OHCouncilorsChristopher G. Maples, Bloomington, INStephen M. Holland, Athens, GAJournal of Paleontology Co-EditorsAnn F. (“Nancy”) Budd, Iowa City, IABrian J. Witzke, Iowa City, IAJulia Golden, Iowa City, IAJonathan M. Adrain, Iowa City, IAPaleobiology Co-EditorsWilliam DiMichele, Washington, DCJohn Pandolfi, Washington, DCSpecial Publications EditorRussell D. (“Tim”) White, New Haven, CTPriscum EditorPeter J. Harries, Tampa, FLEducation CoordinatorDale A. Springer, Bloomsburg, PAStudent RepresentativeGregory S. Herbert, Davis, CASECTION CHAIRSCordilleran SectionJeff Myers, Monmouth, ORNorth-Central Section Co-ChairsSteve Lo Duca, Ypsilanti, MINortheastern SectionC. Frederick Lohrengel, II, Cedar City, UTRocky Mountain SectionRobert C. Thomas, Dillon, MTSouth-Central SectionDavid M. Rohr, Alpine, TXSoutheastern SectionMichael Savarese, Ft. Meyers, FL

AuthorsRICHARD K. BAMBACHBotanical MuseumHarvard University26 Oxford StreetCambridge, MA 02138TOMASZ K. BAUMILLERDepartment of Geological Sciences andMuseum of PaleontologyUniversity of Michigan1109 Geddes RoadAnn Arbor, MI 48109-1079STEFAN BENGTSONDepartment of PalaeozoologySwedish Museum of Natural HistoryBox 50007SE-104 05 StockholmSwedenCARLTON E. BRETTDepartment of GeologyUniversity of Cincinnati500 Geology/Physics BuildingCincinnati, OH 45221-0013KAREN CHINCU Museum/Geological SciencesUniversity of Colorado, BoulderUCB 265Boulder, CO 80309< Karen.Chin@colorado.edu>STEPHEN J. CULVERDepartment of GeologyEast Carolina UniversityGreenville, NC 27858GREGORY P. DIETLDepartment of ZoologyNorth Carolina State UniversityRaleigh, NC 27695JAMES O. FARLOWDepartment of GeosciencesIndiana-Purdue University at Fort Wayne2101 Coliseum Boulevard EastFort Wayne, IN 46805FOREST J. GAHNDepartment of Geological Sciences andMuseum of PaleontologyUniversity of Michigan1109 Geddes RoadAnn Arbor, MI 48109-1079GARY HAYNESAnthropology DepartmentUniversity of Nevada, RenoReno, NV 89557THOMAS R. HOLTZDepartment of GeologyUniversity of MarylandCollege Park, MD 20742

IAN JENKINSDepartment of Earth SciencesUniversity of BristolWills Memorial Building, Queen’s RoadBristol, BS8 1RJ, EnglandPATRICIA H. KELLEYDepartment of Earth SciencesUniversity of North Carolina at Wilmington601 South College RoadWilmington, NC 28403-5944GEERAT J. VERMEIJDepartment of GeologyUniversity of California at DavisDavis, CA 95616SALLY E. WALKERDepartment of GeologyUniversity of GeorgiaAthens, GA 30602MICHAL KOWALEWSKIDepartment of Geological SciencesVirginia Polytechnic Institute andState UniversityBlacksburg, VA 24061CONRAD C. LABANDEIRASmithsonian InstitutionNational Museum of Natural HistoryDepartment of PaleobiologyWashington, DC 20560JERE H. LIPPSDepartment of Integrative BiologyMuseum of PaleontologyUniversity of California at BerkeleyBerkeley, CA 94720BLAIRE VAN VALKENBURGHDepartment of Organismic Biology,Ecology, and Evolution621 Young Drive, SouthUniversity of California, Los AngelesLos Angeles, CA 90095

THE FOSSIL RECORD OF PREDATIONMichal Kowalewski and Patricia H. Kelley, editorsContents PageIntroduction ................................................................................................................. 1Michal Kowalewski and Patricia H. KelleySection I: MethodsThe Fossil Record of Predation: An Overview of Analytical Methods ....................... 3Michal KowalewskiAnalyses of Coprolites Produced by Carnivorous Vertebrates .................................. 43Karen ChinArchaeological Methods for Reconstructing Human Predation onTerrestrial Vertebrates ................................................................................................ 51Gary HaynesSection II: PatternsThe Trophic Role of Marine Microorganisms Through Time ................................... 69Jere H. Lipps and Stephen J. CulverPredators and Predation in Paleozoic Marine Environments .................................... 93Carlton E. Brett and Sally E. WalkerPost-Paleozoic Patterns in Marine Predation: Was there a Mesozoicand Cenozoic Marine Predatory Revolution?...........................................................119Sally E. Walker and Carlton E. BrettFossil Record of Parasitism on Marine Invertebrates with Special Emphasison the Platyceratid-Crinoid Interaction ................................................................... 195Tomasz K. Baumiller and Forest J. GahnPaleobiology of Predators, Parasitoids, and Parsites: Death and Accomodationin the Fossil Record of Continental Invertebrates ....................................................211Conrad C. LabandeiraThe Fossil Record of Predation in Dinosaurs .......................................................... 251James O. Farlow and Thomas R. Holtz, Jr.

Section III: ProcessesEvolutionary Patterns in the History of Permo-Triassic and CenozoicSynapsid Predators .................................................................................................. 267Blaire Van Valkenburgh and Ian JenkinsOrigins and Early Evolution of Predation ............................................................... 289Stefan BengtsonSupporting Predators: Changes in the Global Ecosystem Inferred fromChanges in Predator Diversity ................................................................................. 319Richard K. BambachThe Fossil Record of Predator-Prey Arms Races: Coevolution andEscalation Hypotheses ............................................................................................. 353Gregory P. Dietl and Patricia H. KelleyEvolution in the Consumer Age: Predators and the History of Life........................ 375Geerat J. VermeijClosing RemarksThe Fossil Record of Predation: Methods, Patterns, and Processes ........................ 395

INTRODUCTIONTHE FOSSIL RECORD OF PREDATIONAn IntroductionBiologists and paleontologists agree that direct interactions among organisms are importantecological mechanisms that may play a key role in evolution. Among various biotic interactions, predationhas often been recognized as a particularly significant ecological force. Its evolutionary importance ismuch more controversial, however, and researchers still debate the role predation has played in shapingthe history of life. The fossil record is our primary source of the data needed to address this issue. Inrecent years, paleontologists have provided critical documentation of prey-predator interactions overevolutionary timescales, and from there have generated fruitful hypotheses regarding the history of lifeand the role of predation in evolution.The goal of this short course is to provide a comprehensive and up-to-date overview of the currentknowledge and understanding of the fossil record of predation—from direct indicators provided bytrace fossils and coprolites to more indirect proxies provided by taphonomic data, functional morphology,and phylogenetic relationships. The short course includes presentations by experts in the areas ofmicropaleontology, invertebrate paleontology, paleoentomology, vertebrate paleontology, andanthropology. The fossil record left by parasites is also included because predators and parasites representendmembers of the same ecological adaptation. This short course is organized into three distinct sections.The first, methodological, section (Methods) provides an overview of collecting methods, analyticaltechniques, and statistical strategies that are applied in the study of the fossil record of predation.Reviews from different fields, from invertebrate paleontology to physical anthropology, bring togethervarious methodological perspectives and offer guidelines from collecting strategies to analyticalapproaches.The second, descriptive, section of the short course (Patterns) provides an up-to-date overview ofcurrent knowledge on the fossil record of predation. Various lines of evidence, from trace fossils tofunctional morphology, are reviewed for microfossils, marine invertebrates, terrestrial invertebrates,and vertebrates. The descriptive section illustrates the wealth of data already amassed by researchers,but also points to various temporal and taxonomic gaps that call for future research.The descriptive section sets the empirical stage for the interpretive, process-oriented section of theshort course (Processes). This final section focuses on higher-order interpretations of the fossil recordof predation, including models derived from or tested against it. This section of the short course presentsvarious views regarding the role of predation in shaping the history of life on our planet.To our knowledge, this short course volume represents the first book that focuses solely on thepaleobiology of predation. The volume provides a comprehensive synthesis of our knowledge of thefossil record of predatory behavior, demonstrates the amazing wealth of data on predator-prey interactionsthat can be extracted from the fossil record, and shows how these data are instrumental in developingnew interpretations and hypotheses regarding the evolutionary history of ecological interactions. Wehope that this short course will stimulate further research on predation and aid future investigators inidentifying unexplored and fertile areas of study.We thank the reviewers R. Alexander (Rider University), R. Bambach (Harvard University), A.Behrensmeyer (Smithsonian Institution), J. Bernhard (University of South Carolina), K. Chin (Universityof Colorado), G. Dietl (North Carolina State University and the University of North Carolina atWilmington), J. Farlow (Indiana-Purdue University), D. Fisher (University of Michigan), T. Hansen(Western Washington University), S. Hasiotis (University of Kansas), G. Haynes (University of Nevada),A. Hoffmeister (Virginia Tech), L. Leighton (San Diego State University), D. Meyer (University of1

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Cincinnati), A. Miller (University of Cincinnati), W. Miller (Humboldt University), R. Molnar (Museumof Northern Arizona), J. Rice (North Carolina State University), D. Rodland (Virginia Tech), R. Rogers(Macalester College), D. Smith (University of Colorado), S. Snyder (East Carolina University), B. VanValkenburgh (University of California, Los Angeles), G. Vermeij (University of California, Davis), S.Walker (University of Georgia), and M. Wilson (College of Wooster). The expert help of the reviewersgreatly improved the quality of this volume.Our special thanks go to our assistant editor, Alan Hoffmeister, for his scrupulous pursuit of errors:from typos, citation errors, format inconsistencies, and stylistic deficiencies to more substantive errorsof facts and interpretations. Also, we thank the Short Course Series Editor, Tim White, for his patienceand extensive help, especially during the final stages of preparation of this volume. Last but not least,we thank the contributors for providing high-quality manuscripts and for responding quickly and patientlyto our sometimes unreasonable demands.Michal KowalewskiDepartment of Geological SciencesVirginia Polytechnic Institute and State UniversityBlacksburg, VAPatricia H. KelleyDepartment of Earth SciencesUniversity of North CarolinaWilmington, NC2

KOWALEWSKI—ANALYTICAL METHODSTHE FOSSIL RECORD OF PREDATION:AN OVERVIEW OF ANALYTICAL METHODSMICHAL KOWALEWSKIDepartment of Geological Sciences, Virginia Polytechnic Institute and State University,Blacksburg, Virginia 24060 USAABSTRACT—Paleontological research on predation has been expanding rapidly in scope, methods, and goals. Thegrowing assortment of research strategies and goals has led to increasing differences in sampling strategies,types of data collected, definition of variables, and even reporting style. This methodological overview serves asa starting point for erecting some general methodological guidelines for studying the fossil record of predation.I focus here on trace fossils left by predators in the skeleton of their prey, arguably one of the most powerfulsources of direct data on predator-prey interactions available in the fossil record. A critical survey of samplingprotocols (data collecting strategy, sieve size, and sample size) and analytical approaches (predation intensitymetrics, strategies for evaluating behavioral selectivity of predators, and taphonomic tests) reveals that variousapproaches can be fruitful depending on logistic circumstances and scientific goals of paleoecological projects.Despite numerous caveats and uncertainties, trace fossils left by predators on skeletons of their prey remain oneof the most promising directions of research in paleoecology and evolutionary paleobiology.INTRODUCTIONIN RECENT YEARS, paleontological researchon predation has become increasingly sophisticatedin terms of complexity of tested hypotheses,intricacy of sampling designs, and quality ofanalytical methods. Moreover, its thematic scope hasexpanded abruptly as we now collect much morediverse data for a much broader spectrum oforganisms over a much wider range of observationalscales, from individual interactions to global-scalesecular trends. Unfortunately, albeit perhapsinevitably, our data are collected in various, oftendisparate ways, so our research efforts arecontaminated with methodologically undesirableidiosyncrasies. The irreconcilable differences insampling strategies, types of collected information,definition of variables, and even reporting stylemake it difficult to compare directly manyotherwise valuable data sets, and hamper metaanalyticalattempts to explore hoards of dataamassed in the rapidly growing literature on thefossil record of predation.The methodological overview presented in thischapter and the two subsequent contributionsincluded in this volume (Chin, 2002; Haynes, 2002)bring together a diversity of methods used forstudying the fossil record of predation. Thesereviews should help us in collecting and reportingfuture data in a more congruent manner so as toavoid the confusion that we often encounter whencommunicating our research.This chapter focuses primarily on trace fossilsfound on skeletons of marine invertebrate prey. Suchfossilized traces of predation provide arguably therichest source of quantifiable data on prey-predatorinteractions available in the fossil record (seeespecially Kitchell, 1986) and have been widely usedin paleontological research to date. Other importantlines of evidence for studying predator-preyinteractions are discussed here only briefly. Thesubsequent methodological contributions includedin this volume review the methods employed toinvestigate coprolites (Chin, 2002) and the distinctstrategies used to study hominids and othervertebrate predators and prey (Haynes, 2002).Following Bambach (2002), predators aredefined here as organisms that “…hunt or trap,3

KOWALEWSKI—ANALYTICAL METHODSKelley and Hansen, 1993; Alexander and Dietl,2001; Dietl and Kelley, 2002).2. Coprolites and stomach contents withidentifiable prey remains consumed by predators area direct indicator (e.g., Bishop, 1977; Sohn andChatterjee, 1979; Stewart and Carpenter, 1990,1999; Coy, 1995; Becker et al., 1999; Richter andBaszio, 2001; Carrion et al., 2001) that is particularlywidely used in studying the fossil record of terrestrialvertebrates (e.g., Chin et al., 1998; Andrews andFernandez Jalvo, 1998). Coprolite-based researchis discussed later in this book (Chin, 2002). A sourceof direct evidence, similar to coprolites, is providedby stomach contents. Although rarely preserved inthe fossil record, instances of prey remains found inthe digestive system of a predator are known for awide range of predator-prey systems (e.g., Spencerand Wright, 1966; Moy-Thomas and Miles, 1971:Alpert and Moore, 1975).3. Exceptional Preservational Events (EPE),in which two or more individuals are preservedtogether while interacting (e.g., Baumiller, 1990;Carpenter, 2000), also represent a direct indicatorof biotic interactions. Trace fossils left byinteracting organisms can also be included here (e.g.,trails left by predators chasing their prey; Lockleyand Madsen, 1992; Pickerill and Blissett, 1999).Although such indicators can be an insightful sourceof information on biotic interactions, they are veryrare and thus of limited use in quantitative analysesor any large-scale studies that require multiplerecords through time or space.4. Taphonomic patterns offer a wide range ofindirect evidence such as the degree andcharacteristics of shell and bone fragmentation,preservation of predators and prey in close spatialassociation, midden deposits, or some other distinctbiostratinomic characteristics in the arrangement ofprey skeleton fragments (e.g., Wilson, 1967; Cadée,1968, 1994, 2000; Mayhew, 1977; Stallibrass, 1984,1990; Wilson, 1987; Todd and Rapson, 1988; VanValkenburgh and Hertel, 1993; Cate and Evans, 1994;Lyman, 1994; Brandt et al., 1995; Chin, 1997; Stewartet al., 1999; Merle, 2000). However, these indicatorsare often limited to unique taphonomic settings andtheir interpretation tends to be ambiguous. Forexample, fragmentation may occur due to variouscauses other than direct biotic interactions (althoughbreakage of biotic origin may be admittedly adominant factor; e.g., Cadée, 1968; Cate and Evans,1994; Oji et al., 2001). Other taphonomic lines ofevidence are also debatable. The close spatialassociation of presumed predators and prey mayreflect unique preservational circumstances (e.g.,taphonomic traps such as tar pits) and accumulationsof abiotic origin may be so similar to middens thattheir differentiation requires a careful statisticalanalysis (e.g., Henderson et al., 2002).5. Indirect evidence is also provided byinferring predatory or defensive behaviors fromfunctional morphology of fossils and phylogeneticaffinities of studied groups, including ecology andethology of their nearest living relatives. There areobvious dangers of interpretations based onfunctional morphology and phylogeny. Organismsmay change their behavior, but, due to exaptationsor various constructional constraints, may still retainmorphological characters reflecting their previousecology and behavior. Also, a morphology that canbe interpreted as serving a particular function (e.g.,prey defensive traits) may have evolved due toabiotic factors, and distinguishing between the twocauses may be difficult (e.g., Wood, in press).Phylogenetic affinities are also a dangerous toolgiven the arguably high evolutionary plasticity ofecology and behavior of organisms. Moreover,feeding ecologies may be non-randomly distributedwithin and across metazoan clades, as is suggestedby the derived nature of herbivory observed at manyscales of phylogenetic analysis (Vermeij andLindberg, 2000). Even more discouragingly, indirectstrategies based on phylogeny or functionalmorphology typically provide information aboutonly one component of the biotic interaction (e.g., aparticular predator, a given prey clade, etc.), but tellus next to nothing about types of organisms withwhich the studied group may have interacted.Moreover, such indicators offer no quantifiablepaleoecological data on the frequency of predatoryattacks, prey selectivity, size refuge, and otheraspects of predator-prey interactions (see alsoLeighton, 2002). Thus, they offer limited interpretive5

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002power relative to trace fossils or coprolites,especially for detailed paleoecological analyses.Despite those caveats, these indirect indicatorsrepresent a valuable source of data becausefunctional morphology can be applied to nearly allfossil specimens, and interpretations based onphylogenetic affinities can be postulated forvirtually any lineage. Not surprisingly, theapproach has proven a powerful tool in numerousstudies, especially in large-scale studies of temporaland spatial gradients (e.g., Vermeij, 1977, 1987,2002; Signor and Brett, 1984; Leighton, 1999;McRoberts, 2001; Dietl and Kelley, 2001; VanValkenburgh and Jenkins, 2002; Bambach, 2002;and references therein).VIRTUES OF TRACE FOSSILSAS PREDATION INDICATORSI focus here on the most common quantifiableindicator of predation in the fossil record: trace fossils.These direct indicators have numerous virtues:1. Traces of predation are common acrossvarious depositional environments. For example,drill holes and repair scars of predatory origin occurnot only in a wide spectrum of marine environments(see Vermeij, 1987, for numerous references) butalso can be found on skeletons of terrestrial prey(e.g., Vermeij, 1987, 2002; Ørstan, 1999;Gittenberger, 1999). Similarly, vertebrate tooth andgnawing marks can be found in both terrestrial(Lyman, 1994; Haynes, 2002) and marine prey(Kauffman and Kesling, 1960; Stewart andCarpenter, 1999; Tsujita and Westermann, 2001; butsee Kase et al., 1998).2. Traces of predation are found on skeletons ofa wide spectrum of prey—from protists (e.g., Sliter,1971, 1975; Lipps, 1988; Culver and Lipps, in press;Lipps and Culver, 2002; Hageman and Kaessler,2002) through virtually all groups of metazoans withbiomineralized skeletons, including marineinvertebrates (see especially Vermeij, 1987; Brett andWalker, 2002; Walker and Brett, 2002; Brett, in press;Alexander and Dietl, in press; Kelley and Hansen, inpress), terrestrial invertebrates (LaBandeira, 2002),and terrestrial vertebrates, including hominids (e.g.,Lyman, 1994; Haynes, 2002).3. Traces of predation are ubiquitous in thegeological record and span virtually the entire fossilrecord of metazoan organisms with biomineralizedskeletons. Starting with enigmatic tubes of the LatePrecambrian Cloudina (Bengtson and Yue, 1992;Bengtson, 2002) and followed by diverse tracefossils found in Cambrian prey (e.g., Alpert andMoore, 1975; Miller and Sundberg, 1984; Jensen,1990; Conway Morris and Bengtson, 1994; Nedin,1999), drill holes, punch holes, repair scars, andother traces left by predators litter the fossil recordof marine invertebrates (see reviews and datacompilations by Vermeij, 1983, 1987; Signor andBrett, 1984; Alexander, 1986b; Kabat, 1990;Kowalewski et al., 1998; Brett, in press; Brett andWalker, 2002; Walker and Brett, 2002). Tracefossils left in terrestrial invertebrate (Labandeira,2002, and references therein) and especiallyvertebrate prey (e.g., Jacobsen, 1997, 1998; Farlowand Holtz, 2002; Haynes, 2002; and referencestherein) are also well documented.4. Trace fossils left by predators are made inbiomineralized skeletons. Consequently, such traceshave as good, or almost as good (see below), potentialfor preservation as the skeletons of their prey.5. Finally, and perhaps most importantly, tracesof predation provide direct evidence of bioticinteraction and thus offer a rich array of quantifiabledata on predator-prey interaction. Drill holes offera particularly spectacular example of the incrediblewealth of data that can be retrieved from traces leftby predators in the skeletons of their victims (Fig. 1).CRITERIA FOR IDENTIFYINGAND INTERPRETINGPREDATION TRACESTrace fossils that may record predatory activityare often controversial in terms of their origin andneed to be assessed rigorously. I advocate here athree-phased evaluation approach.First, the biotic nature of the traces needs tobe demonstrated. Chemical and physical agents canalso create marks on shells and bones. For example,Lescinsky and Benninger (1994) documented a6

KOWALEWSKI—ANALYTICAL METHODSFIGURE 1—The wealth of ecological and behavioral information that may be obtained by analysis of justone type of trace fossils left by predators: drill holes (after Kitchell, 1986). Whereas many lines of evidenceincluded on this diagram involve debatable assumptions, the figure illustrates the interpretive potential ofthe paleontological record provided by trace fossils left by predators on skeletons of their prey.number of diagenetic alterations in marineinvertebrate shells that could be potentiallymisidentified as traces of biological activity. Variousabiotic processes, ranging from impacts of wavebornestones to compaction, can result in fractures(or even repair scars in case of pre-mortem damage)and be mistaken for records of biotic interactions.Second, demonstrating biotic origin is notenough. Not all biotic traces represent interactionsbetween two living organisms. For example,substrate borings or substrate attachment scars canbe postulated for structures interpreted as traces ofpredation or parasitism (see discussions in Carrikerand Yochelson, 1968; Richards and Shabica, 1969;Kase et al., 1998; Kaplan and Baumiller, 2000,2001; Wilson and Palmer, 2001; Tsujita andWestermann, 2001) and self-inflicted damage ofburrowers can be misinterpreted as records of anencounter with a predator (Checa, 1993; Cadée etal., 1997). Thus, before attempting any analysis, itis necessary to demonstrate that the studied tracesrecord contacts between living organisms (i.e., liveliverather than live-dead interactions).Third, the specific ecological nature of theinteraction needs to be identified. For example,many traces are ambiguous in that they may havebeen formed by predators, scavengers, parasites,amensals, or commensals. There are also caseswhen trace fossils represent self-inflicted damagesuffered by a predator during an attack on its preyas in the case of predatory attacks by Busycon(Dietl, pers. comm., 2001). Thus, it is even possible7

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002incorrectly to attribute self-inflicted damage to apredator as a record of an attack by another predator(although this error can be often avoided by carefulexamination of the damage; Dietl, pers. comm.,2002). Finally, it is also useful to try to differentiatetraces that represent sublethal damage (e.g., repairscars, healed drill holes, etc.) from lethal traces(e.g., extensive breakage, complete drill holes). Itis noteworthy that sublethal damage may recordtwo different types of events. First, it may representfailed lethal predation; for example, unsuccessfulattacks of crabs that failed to kill and eat theirmollusk prey. Second, sublethal damage mayrecord successful partial predation by carnivoresthat only partially consume their prey. Such victimsoften recover, even though the attempt wassuccessful from the perspective of the attacker.Examples of prey that are frequently subjected topartial predation in present-day ecosystemsinclude, among others, echinoderms (e.g.,Kowalewski and Nebelsick, in press), corals (e.g.,Wood, in press), and bivalves (e.g., Vermeij, 2002).Numerous lines of evidence can be used torecognize trace fossils produced by bioticinteractions (see also Carriker and Yochelson, 1968;Bishop, 1975; Chatterton and Whitehead, 1987;Rohr, 1991; Lyman, 1994; Bromley, 1996;Baumiller et al., 1999; Alexander and Dietl, in press):1. Traces have distinct geometric shape (e.g., drillholes, double punctures, peeling, tooth marks). Thiscriterion helps to rule out abiotic origin of traces.2. Traces show a relatively narrow size range.This pattern also suggests biotic origin of traces,as abiotic traces tend to be more variable in size.3. The nature of traces suggests that they weremade to gain access to the inside of the protectivearmor of the prey or host. This type of evidenceincludes, for example, holes and punctures thatpenetrate external skeletons from outside and donot go through the opposite side of those skeletons(traces that go through are likely to representsubstrate borers; see Richards and Shabica, 1969;Kaplan and Baumiller, 2000).4. Traces are distributed non-randomly acrosstaxa. Such species-selectivity is strong evidencethat traces are biotic in origin, especially if taxaare comparable in mineralogy, microstructure, andphysical durability of skeletons, as it is hard toimagine that destructive abiotic (physical andchemical) processes would be highly selectiveamong taphonomically comparable fossils.5. Biotic traces made by predators or parasitesare often non-random in their distribution on preyskeletons. Drill holes made by snails mayconcentrate in a particular area of the shell (e.g.,Reyment, 1971; Negus, 1975; Berg, 1976; Kelley,1988; Leighton, 2001) and vertebrate predatorssuch as owls may preferentially break or otherwisedamage only certain types of bones (Dodson andWexlar, 1979; Kusmer, 1990; Lyman, 1994). Thistype of site-selectivity can not only help us todemonstrate the biotic origin of traces but may alsoallow us to postulate the specific type of behaviorrecorded by these traces. For example, traces maybe distributed to give optimal access to muscletissues, suggesting predatory rather than parasiticbehavior (e.g., Hoffmeister et al., submitted).6. Traces may occur preferentially in aparticular size class of fossils. This pattern suggestssize-selective behavior and thus points to bioticinteractions. For example, Hoffmeister et al.(submitted) demonstrated that drill holes inPennsylvanian brachiopods are restricted to anarrow size range of prey specimens.7. Traces made by predators tend to be singularwhereas parasitic traces are often multiple. Forexample, echinoid tests drilled by cassid snailstypically bear one hole only (e.g., Nebelsick andKowalewski, 1999) whereas those drilled byparasitic eulimid snails often contains several holes(e.g., Wáren et al., 1994; Kowalewski andNebelsick, in press). However, there areexceptions. For example, predatory octopods tendto drill two or more holes to inject the venom moreeffectively (Bromley, 1993); and even textbookpredators such as naticids and muricids are knownto drill multiple successful holes in some prey (e.g.,Dietl and Alexander, 2000; Dietl, 2000).8. The presence of complementary scars onopposite sides of the skeleton suggests that these traceswere made by a scissor-like weapon such as a crabclaw or bird beak (e.g., Kowalewski et al., 1997).8

KOWALEWSKI—ANALYTICAL METHODS9. The correlation between size of traces andsize of fossils that contain them can also supportpredatory/parasitic origin of traces (but seediscussion of taphonomic biases below).10. Attachment scars can sometimes beobserved in association with trace fossils,suggesting that the trace maker was attached to itsvictim for a prolonged period of time. Suchattachment scars are typically interpreted asevidence for parasitic origin of traces (e.g.,Matsukuma, 1978; Baumiller, 1990).Typically, because of the nature of availabledata, only some of the above criteria are applicablein any given case study. Many of those criteria areinsufficient when applied alone and, ideally,multiple lines of evidence should be applied to makea convincing case. The criteria can be supplementedwith indirect lines of evidence such as the repeatedco-occurrence of possible trace makers with theirvictims in many fossil assemblages through timeand space, or such spatio-temporal changes infrequency of traces that are more likely to reflectchanges in intensity of biotic interactions rather thanchanges in the intensity of taphonomic and otherabiotic processes.It is dangerous to assume, however, that alltraces were made by the same type of organisms(e.g., naticid snails), or even represent a singlebehavior (e.g., predation) (see Gibson and Watson,1989 for a convincing example). In fact, given awide variety of origins that can be postulated forany given type of trace fossil (see below), it seemslikely that trace fossil assemblages often contain amix of records representing a whole spectrum ofbehaviors, including even abiotic traces. Ausichand Gurrola (1979) made a case for simultaneouspresence of drill holes of parasitic and predatoryorigin in the same fossil assemblage, whereas alively discussion between Kaplan and Baumiller(2000, 2001) and Wilson and Palmer (2001) offersa good example of methodological and practicaldifficulties in dealing with this issue. Ultimately,Bayesian statistical approaches may be needed todeal with those issues in a formalized way, butmethodological strategies for dealing with thisproblem are still in their infancy.CAN WE IDENTIFYTHE CULPRIT?Once the behavioral origin of the traces isdemonstrated, one may attempt to pinpoint thebiological identity of the culprit. However, as withany trace fossil (see Bromley, 1996), theidentification of a predator from traces is a riskybusiness (see especially Bromley, 1993). Differentclades of predators often produce similar traces.For example, even the most morphologicallydistinct and informative traces such as drill holescan be made by a whole spectrum of organismsincluding as many as 14 different groups ofpredatory or parasitic invertebrates (Table 2; seealso Vermeij, 1987; Kabat, 1990; Kowalewski,1993; Brett and Walker, 2002; and Walker andBrett, 2002). Moreover, the same species ofpredator can produce traces that vary notably inmorphology (e.g., Wodinsky, 1969; Bromley,1970). In fact, even the same individual preyingon a single prey type can make traces that varygreatly in shape and size, as demonstrated for singlespecimens of Octopus vulgaris preying on strombidgastropods (Arnold and Arnold, 1969).In addition to these problems, the morphologyof traces is not just a function of the anatomy andbehavior of predators but may also vary greatlydepending on prey morphology and many otherfactors. For example, at least seven additional factorsare known to affect the geometry of a drill hole(Table 3). Because drill holes are widely consideredto be one of the most unambiguous sources ofinformation on predator-prey interactions in the fossilrecord (e.g., Kitchell, 1986), Table 3 is likely to bethe best-case scenario. Arguably, other less distinctand inherently more variable traces such as repairscars, fractures, chewing marks and so on are aneven more capricious source of information on thepredator’s identity, although exceptions certainlyexist. For example, distinct double punctures madeby stomatopods may be possible to identify in thefossil record (Geary et al., 1991; Bauk and Radwaski,1996; but see Alexander and Dietl, in press).In sum, different organisms can make similartraces, the same organisms can make different9

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 2—Drill hole producers reported in the biological and paleontological literature (expanded andupdated after Kowalewski, 1993).Drilling OrganismSelected ReferencesNEMATODS Sliter 1971, 1975; Arnold et al., 1985FLATWORMS Yonge, 1964; Woelke, 1957GASTROPODSNudibranchs Zilch, 1959; Young, 1969; Taylor et al., 1983Pulmonates Wächtler, 1927; Degner, 1928; Carriker and Yochelson, 1968;Mordan, 1977Platyceratids Baumiller, 1990, 1993, 1996, 2002; Baumiller and Macurda, 1995;Baumiller et al., 1999Naticids Fischer, 1922-1966; Ziegelmeier, 1954; Carriker, 1961, 1981;Sohl 1969, Reyment, 1963-1967; Berg and Nishenko, 1975; Berg, 1976;Wiltse, 1980; Taylor et al., 1980; Savazzi and Reyment, 1989Muricids Fischer, 1922-1966; Carriker, 1943-1981; Reyment, 1963-1967;Sohl, 1969; Carriker and Van Zandt, 1972a,b; Matsukuma, 1977;Taylor et al., 1980, 1983Cassids Hirsch, 1915; Roughley, 1925; Day, 1969; Sohl. 1969;Hughes and Hughes, 1971, 1981; Nebelsick and Kowalewski, 1999;Kowalewski and Nebelsick, in pressEulimids Warén, 1980, 1981; Warén and Crossland, 1991; Crossland et al., 1991;Rinaldi, 1994; Warén et al., 1994; Kowalewski and Nebelsick, in pressCapulids Orr, 1962; Kosuge and Hayashi, 1967; Matsukuma, 1978; Bromley 1981Nassarids Fisher, 1922, 1962a; Reyment, 1967; Morton and Chan, 1997Marginellids Ponder and Taylor, 1992Buccinids Peterson and Black, 1995; Harper et al., 1998OCTOPODS Fuijta, 1916; Arnold and Arnold 1969; Nixon, 1979; Ambrose, 1986;Guerra and Nixon, 1987traces, and traces may also vary in morphology forreasons unrelated to the identity of a trace maker.Thus, with few exceptions, trying to identify thespecific organism responsible for traces found inthe fossil record is difficult. In fact, suchidentification efforts can only distract us from thereal strength of trace fossils: their informative valueas records of interactions that affected populationsof an identifiable prey.DATA COLLECTIONAt least six strategies have been used to acquiredata on traces of predation:1. Outcrop Surveys rely on visual screening ofoutcrops for fossils with traces of predation,including accidental discoveries of such specimensby researchers collecting fossils for other reasons(e.g., Kowalewski and Flessa, 1994). Outcrop10

KOWALEWSKI—ANALYTICAL METHODSTABLE 3—Confounding factors that may hamper reliable identification of the taxonomic identity of adrill hole maker; based on review by Kowalewski (1993).ConfoundingFactorsSite of the drill holeon the prey shellThickness of thebored shellStructure ofbored materialHardness ofbored materialOrnamentation ofprey shellGeometry ofprey shellTaphonomicalterationMorphological ConsequencesHoles with an imperfectly developed form are producedwhen drilling is localized at the edge of the shell.The vertical shape of the hole is imperfectly developedif shells are too thin.The stepped appearance of drill holes can result fromthe effect of the presence of the hard conchiolin layerin the shell of the prey (e.g., corbulid bivalves).Changes in size and shape of drill holes may dependon the hardness of drilled skeleton, even when allholes were made by the same drilling predator.Drill hole geometry also can be controlled byornamentation. For example, the drill hole morphotype"F" (sensu Arua and Hoque, 1989) always occurs betweenthe ribs of ribbed shells (see Table 5 in Arua and Hoque,1989).The unusual oval shape of the drill holes bored inscaphopod shells is an effect of the cylindricalgeometry of the shell.Taphonomic processes can affect drill hole morphologyfrom subtle alteration of its outline to major modificationsthat completely change the shape of the drill hole. Whereasdetailed studies exploring this issue are lacking, marginalremarks can be found in many previous works.ReferencesVermeij, 1980Ziegelmeier, 1954;Kitchell et al., 1981;Taylor et al., 1983;Yochelson et al., 1983Ziegelmeier, 1954;Fischer, 1963; Taylor etal., 1983; Cauwer, 1985Nixon, 1979Arua and Hoque, 1989Yochelson et al., 1983Vermeij andDudley, 1982;Taylor et al., 1983;Allmon et al., 1990surveys represent a highly uncontrolled samplingstrategy that has limited use in quantitativeanalyses. However, they are an effective methodfor maximizing the chances of finding evidence ofpredation when such evidence is not expected tobe common (e.g., in pilot projects that focus on timeintervals, depositional environments, or prey typesknown to be a poor source of data on predation).2. Direct Bulk Sampling provides quantitativedata of highest quality by offering a full control onsampling design and data acquisition strategies.This method has been widely used to studypredation (e.g., Hoffman and Martinell, 1984;Kelley and Hansen, 1993, 1996; Hagadorn andBoyajian, 1997; Stewart et al., 1999; Dietl et al.,2000). Although arguably superior to othersampling strategies, bulk methods suffer somedrawbacks. First, bulk samples are limited spatiotemporallyand, consequently, may offer a nonrepresentativeestimate of the sampled fossilassemblage. Second, bulk samples are limitedvolumetrically and very large specimens may be11

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002severely under-sampled (see also Dietl et al., 2000).Finally, a high time demand is involved in bulksampling. Thus, typically, such studies are limitedto few sites with a narrow stratigraphic andgeographic coverage (e.g., Colbath, 1985;Kowalewski, 1990). However, large scale researchprograms that apply direct bulk sampling strategiesare feasible and can yield some of the mostcomprehensive insights into the history of predation,as demonstrated by efforts of Kelley, Hansen, andtheir students and collaborators (e.g., Kelley andHansen, 1993, 1996, in press; Hansen and Kelley,1995; Kelley et al., 2001).3. Indirect Bulk Sampling is based on reusingpreviously collected, either processed or—betteryet—unprocessed, bulk materials. For example,despite highly constrained spatial, temporal, andenvironmental scopes of their study (north-centralEurope, early middle Miocene, marine clasticepicontinental facies), Kowalewski et al. (2002)obtained numerous, high-quality bulk samplesfrom museums, university collections, individualresearchers, and private collectors. Hoffmeister andKowalewski (2001) used the resulting dataset in adetailed quantitative analysis of drilling predationpatterns and produced data comparable to thoseobtained when direct bulk sampling strategies areemployed. Although the indirect bulk samplingmethod does not offer full control on selection ofsampling sites and sample processing procedures,the resulting data can be of comparable quality andtheir acquisition can be much less time consuming,thus permitting studies with much broader spatialand/or temporal scopes.4. Museum Surveys are a widely used methodbased on revisiting existing collections of fossilsstored in museums and research collections (e.g.,Hoffman et al., 1974; Allmon et al., 1990;Baumiller, 1993, 1996; Harper et al., 1998; Harperand Wharton, 2000; Hoffmeister, 2002;Hoffmeister et al., submitted). These types ofprojects can be particularly effective when tracesof predation are rare and large collections need tobe examined to achieve quantitative data (e.g.,Hoffmeister, 2002). Museum collections may bebiased by removal of “imperfect” specimens byoverzealous curators (Walker, 1989) or bypaleoecologists interested in studying specimenswith trace fossils (Baumiller, pers. comm., 1999).Kowalewski (1990) showed that bulk samples ofMiocene mollusks from the Korytnica Clays(central Poland) yielded similar estimates ofdrilling predation intensity for most (but not all)taxa that were analyzed previously by Hoffman etal. (1974; see also Zlotnik, 2001). However, somemuseum collections can offer materials that arecomparable to bulk materials for individual taxa(Harper et al., 1998) or even for entire assemblages(Hoffmeister, 2002), providing suitable materialsfor quantitative studies.5. Monograph Surveys represent a lamentablyunderutilized research strategy. Monographsprovide thousands of photographs and thus allowus to examine quickly large numbers of specimensalready identified and documented in terms ofsampling site and stratigraphy. Such surveys are agreat tool for pilot studies in exploring understudiedgroups of prey or geological time intervals knownto have limited records of predation. Moreover,despite many obvious drawbacks, monographs canyield data that provide useful information onpredator-prey interactions, and can go beyond merepilot studies. Kowalewski et al. (2000) examinedthe series of monographs of Cooper and Grant(1972-1976) and were able not only to show thatdrilling predation (or parasitism) was continuouslypresent in the Permian, but also to explore somequantitative patterns regarding behavioral stereotypyof the drillers. Even the quantitative estimatesobtained by Kowalewski et al. (2000) were notunreasonable, as demonstrated in a subsequent study(Hoffmeister, 2002) of a brachiopod collectionhoused in the Smithsonian Institution (thecollection is a nearly complete representation ofbulk materials processed by Cooper and Grantduring the preparation of their monographs).6. Meta-Analytical Literature Compilationscombine data assembled in previous qualitative andquantitative studies and provide a powerful tool forlarge-scale analyses of global secular trends inpredation (e.g., Vermeij, 1987; Kowalewski et al.,1998). Although such studies are admittedly12

KOWALEWSKI—ANALYTICAL METHODShampered by methodological differences among casestudies, they can provide first-order approximationsfor long-term trends that are otherwise difficult toaccess (but see Harper et al., 1999).The six categories represent end members in aspectrum of possible approaches. Also, the dataobtained with different approaches can becombined to broaden the scope of the study (e.g.,Allmon et al., 1990; Hagadorn and Boyajian, 1997;Dietl et al., 2000). The numerous case studies citedfor each strategy illustrate that all of the aboveapproaches can be fruitful depending on logisticcircumstances and research goals.DATA PROCESSINGNumerous decisions are made when designingany research project. These decisions, often forcedby pragmatic aspects of a particular investigation,can influence the quality and informative value ofthe data. Sieve size, sample size requirements,tallying strategies, and styles of data reporting areparticularly important.Sieve Size.—In many cases, paleoecologicalsamples are processed with sieves to separatefossils from the enclosing sediment or rock. Sievesizes can vary greatly among case studies, evenfor projects that target the same type of fossils. Itis intuitively obvious that the choice of the meshsize used can greatly affect any quantitativepaleoecological estimates derived from theanalyzed residue. Because predators may be sizeselective (either directly by selecting certain sizeclasses of prey or indirectly by selectingpreferentially species from certain size classes), thesieve effect can be severe. To illustrate thisproblem, I re-analyze here a large dataset ofmollusks from the Miocene of Europe (Hoffmeisterand Kowalewski, 2001). The data include over3500 specimens that were measured in terms ofsize and analyzed for presence of drill holes. Aseries of computer simulations was used torandomly sub-sample the database whilemimicking the sieve effect (Fig. 2). The resultsindicate that, in this particular case, the drillingintensity rises as the mesh size is incrementallyincreased from 1mm to 10mm (the frequency ofholes increases roughly by half: from ca. 15% for1mm mesh to 23% for 10mm mesh). This simpleexample shows that the mesh size can influencethe estimates of predation. Because smallspecimens can be excluded analytically, thecompatibility of a study with previous studies isincreased when the mesh size is small: the finer isthe sieve, the more comparable will be the resultingdata in future meta-analytical studies.The above exercise shows how the exclusionof small specimens may affect the analysis. Theexclusion or under-sampling of large specimens,which may be associated with bulk sampling (seeabove), may introduce similar types of biases intothe analysis.Sample Size Requirements.—Because traces ofpredation do not occur in all collected specimensand some specimens may bear more than one tracefossil, the sample size can be computed in threeways as (1) number of specimens; (2) number oftrace fossils; and (3) number of specimens withtraces fossils. Depending on the target of ouranalysis and frequency of traces, different samplingrequirements may apply. For example, if theintensity of predation is the primary parameter ofinterest and traces are common, individual samplesof 30 to 50 specimens may be sufficient to evaluatethe analyzed patterns in a meaningful way. Notethat maximizing the number of samples byreducing their size is a statistically advantageoustradeoff in quantitative paleoecological analyses(see Bennington and Rutheford, 1999). On theother hand, if the spatial distribution of traces is ofprimary interest and traces are rare, severalthousand specimens may be required to obtain datathat are statistically meaningful (e.g., Hoffmeister,2002). Finally, the sample size also may be predeterminedby demands of statistical tests (e.g.,contingency tables [goodness of fit tests] require acertain number of observations per cell), althoughthis issue can be partly alleviated by applyingcomputer-intensive methods.Tallying Strategies.—When processingsamples, specimens and trace fossils can be talliedin several ways. First, data entry may be limited to13

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 2—An empirical example of the effect of sieve size on the estimates of drilling intensity. Alarge dataset of Miocene mollusks (> 3500 gastropod and bivalve shells) was obtained by processingbulk samples using fine sieves with mesh below 1mm (Hoffmeister and Kowalewski, 2001; Kowalewskiet al., 2002). An effect of sieve size was then mimicked in a computer simulation by removing from thedatabase all specimens below a given mesh size. For each sieve size, 1000 subsamples of 100specimens were randomly selected and assemblage frequencies (AF; see Equation 1) were computed,including mean drilling frequencies (large solid points) and 95% confidence intervals around them(small solid points). The confidence intervals are based on 2.5 and 97.5 percentile of 1000 AF valuesobtained for each mesh size.only specimens with traces. This is usually notdesirable because such data do not allow us tocompute frequency of traces in bulk samples, andmany other types of analyses (see below) cannotbe conducted. However, this method may beeffective in extracting some quantitative data, whenvery large collections are screened for traces ofpredation. Preferably, a data entry table shouldinclude at least one row per specimen (multiplerows per entry are recommended if multiple tracesare found on the same specimens).Data Reporting.—With increasing use ofonline repository systems, many journals allow forelectronic publication of raw data tables. This isthe most desirable way of reporting data that givesother researchers full access to all informationcollected in a project. Because the posting ofrepository materials may be delayed (authors mayrightly feel it premature to disclose their data), itis also important to report clearly the results of thestudy, so the reader is able to distinguish, forexample, counts of trace fossils from counts offossils with traces and counts of valves of bivalvesfrom counts of shells of bivalves.ANALYSES OF PREDATIONINTENSITY (FREQUENCY)The frequency of traces is arguably the mostimportant and widely used metric in quantitativeanalyses of the fossil record of predation thatestimates the frequency of predator-prey interactionsand may serve as a proxy for predation intensity(but see below). Four different, albeit related,methods have been used for estimating the frequencyof predator-prey interactions in the fossil record.14

KOWALEWSKI—ANALYTICAL METHODS1. Lower Taxon Frequency (LTF) aims toestimate the frequency of interactions for a givenlower taxon of prey (typically species, genus, orfamily). Examples of LTF estimates include thefrequency of drill holes in turittellid gastropods(e.g., Allmon et al., 1990) or the frequency offractures in hominid bone assemblages (e.g., Villaand Mahieu, 1991). LTF is computed as follows:Equation 1: LTF = D K/N K,where K is a lower taxon target in the analysis, D Kis the number of specimens of that taxon thatcontain at least one successful predation trace andN Kis the total number of specimens of that taxonin the sample. Similarly, LTF can be used tocompute the frequency of failed attacks (e.g.,frequency of repair scars in gryphaeid oysters),although the interpretation of such estimates tendsto be more complicated (e.g., Dietl et al., 2000).2. Assemblage Frequency (AF) aims toestimate the overall frequency of predator-preyinteractions recorded by a fossil assemblage. Inpractice, this metric typically estimates frequencyof traces in a higher taxon targeted by the bulksampling protocol (e.g., all mollusks or allbrachiopods). Examples include frequency ofdrilling in all mollusks found in bulk samples (e.g.,Robba and Ostinelli, 1975; Hoffman and Martinell,1984; Colbath, 1985; Kowalewski, 1990; Kelleyand Hansen, 1993) or frequency of cut marks foundin a survey of bones for all types of small or largebovids (e.g., Bunn and Kroll, 1986). AF iscomputed as follows:Equation 2: AF = ΣD i/ΣN i,where D iis the number of specimens of i-th specieswith at least one predation trace and N iis the totalnumber of specimens of i-th species in the sample.3. Highest Lower Taxon Frequency (LTF MAX)aims to estimate the highest frequency of tracesobserved among lower taxa. LTF MAXis computedas follows:Equation 3: LTF MAX= D MAX/N MAX,where MAX is a lower taxon with the highestfrequency of traces in the assemblage, D MAXis thenumber of specimens of that species that containat least one successful predation trace, and N MAXisthe total number of specimens of that species inthe sample. This metric can be derived only frombulk materials when the data on frequency of tracesfor the entire sampled assemblage are available.4. Assemblage Taxon Frequency (ATF),proposed by Vermeij (1987; see also Hansen andKelley, 1995), is comparable to AF in that it alsoaims to estimate the overall predation intensity ina fossil assemblage. However, unlike the AF metric,the ATF metric uses a proportion of lower taxarather than the proportion of specimens to derivean estimate of the overall predation intensity. ATFis computed as follows:Equation 4: ATF = D T/N T,where D Tis a number of common taxa thatfrequently bear traces of predation and N Tis thetotal number of taxa in the sample. The terms“common” and “frequently” are defined a priorinumerically. Vermeij (1987, p. 308), whenanalyzing drilling intensity for bivalved organismsthrough the Phanerozoic, defined “commonspecies” as those represented by at least 20 valvesand “frequently attacked” as those with LTF > 10%.However, if traces are rare we may decide to defineas “frequently attacked” lower taxa with as few asone specimen with traces of predation (i.e., LTF >0) (e.g., Hoffmeister, 2002).Except for ATF, all of the metrics listed aboveuse some estimate of the number of specimens withtraces of predation versus some estimate of thenumber of specimens. It is important to stress twocaveats here. First, the number of specimens withtraces of predation [D K] is not synonymous withthe total number of traces found in those specimensunless all specimens bear singular traces (i.e.,multiple traces are completely absent). Whencomputing predation intensity we should alwaysuse the number of prey specimens attacked (i.e.,the number of specimens with traces) and not thenumber of attacks (i.e., the number of traces).Second, the strategies for computing the totalnumber of specimens may vary greatly dependingon the number of unique elements, degree of15

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002disarticulation of those elements, and statistical andtaphonomic assumptions made in the course of astudy (see especially Gilinsky and Bennington,1994 and references therein). These equations arethus directly applicable only to specimens withsingle-element skeletons and may requiremodifications when applied to multi-elementskeletons (see also below).Discussion of Metrics.—LTF provides a metricsystem that offers the best chance for a biologicallymeaningful analysis of predator-prey interactionin the fossil record (see also Leighton, 2002;Vermeij, 2002). This is because lower taxa are morelikely to represent a single behavioral andecological type of prey, which may interact with asimilar suite of predators through time and space.Also, potential taphonomic biases associated withdifferential preservation of taxa are not as severewhen the analysis is restricted to a single genus orfamily (see also Leighton, 2002).The fact that theestimate is restricted to one prey type also decreasesthe chances for variable behavior of the tracemaker, which may change its behavior dependingon prey type (e.g., drilling predation may beobligatory for bivalve prey that are able to shuttheir valves tightly and facultative for prey withvalves that allow a predator to insert its probosciswithout drilling; e.g., Frey et al., 1986).In contrast, the AF metric, by combining preywith a wide range of morphological and ecologicalcharacteristics, is less reliable both in terms ofbiologic interpretations and potential taphonomicbiases. However, AF offers a significant pragmaticadvantage: it can be computed for any fossil sampleand thus provides a metric that is comparableanalytically (if not biologically) throughout thefossil record. In contrast, few lower taxa arecontinuously abundant through long intervals ofgeological time and virtually none can be used tostudy very long secular trends: an LTF studyencompassing the entire Phanerozoic cannot bedone, except perhaps for such extremelyconservative, long-lived lower taxa like Lingulidae.Thus, in the case of comprehensive long-termstudies, AF is the lesser of two evils: it makes moresense to compare overall assemblage frequenciesbetween the Paleozoic and the Cenozoic than tocompare a specific family of Permian bivalves witha different family of Cenozoic bivalves. AF can bea useful indicator as long as we recognize that it isnot likely to provide estimates for specific predatorpreyinteractions but rather represents a proxy forthe overall predation pressure in the ecosystem.Whereas the credibility of AF has been recentlydebated (e.g., Leighton, 2002; Vermeij, 2002), itis worth pointing out here that assemblage-levelmetrics also provide an important baseline for theoverall intensity of a particular behavior (e.g.,drilling) through time. This baseline can providean important reference standard against whichspecific lineages can be compared. Also, as shownbelow, the metrics tend to correlate highly (theyare obviously dependent) so it may actually notmatter that much which one is used. Thus, althoughAF may be a misleading metric when applied totest a specific model such as the Hypothesis ofEscalation (this is yet to be demonstratedempirically), we should not discard it entirely.LTF MAXresembles AF in that it does not focuson specific interactions of a particular lineage ofprey, but rather tries to estimate the overall intensityin the assemblage by targeting the most frequentlydrilled taxon. AF should tend to be more reliablebecause a maximum is a highly volatile parameterboth in a statistical as well as biological sense.Although ATF is akin conceptually to AF, it doesdiffer fundamentally from AF in that it provides anestimate for how widespread predation was acrossprey taxa rather than across prey specimens. ATFmay also provide an indirect proxy for behavioral,ecological, and maybe even taxonomic diversity ofpredators. Thus, in the best-case scenario, andnotwithstanding all caveats listed above and below,AF may tell us how intense was the overall predationpressure ecologically (what proportion of biota wasbeing killed by predators), and ATF can tell us howintense was predation pressure macro-evolutionarily(what proportion of phylogenetic lineages wasaffected by predators).All four estimates are expected to show somecorrelation with one another: as frequency of tracesincreases the metrics all should go up. This is16

KOWALEWSKI—ANALYTICAL METHODSespecially so in the case of AF, LTF, and LTF max,which can become nearly synonymous when theanalyzed fossil assemblages are close tomonospecific, which happens occasionallyespecially in Paleozoic marine fossil assemblages(e.g., Chatterton and Whitehead, 1987). However,the extent to which the metrics approximate eachother may vary notably, as exemplified in Figure 3for these three metrics (AF, LTF, ATF). In thePaleozoic brachiopod assemblages from Texas, allmetrics correlate tightly; whereas in the Miocenemollusk assemblages from Europe, the correlationis much poorer.I do not advocate any of these metrics asnecessarily superior to the others, as they all haveadvantages and drawbacks and all may beapplicable depending on logistic circumstances andresearch goals. However, four recommendationsseem appropriate here. First, researchers shouldmake sure that they explicitly and precisely definethe intensity metric up front. Second, wheneverpossible, data should be collected to make itpossible to compute all metrics. Third, as suggestedrecently by Leighton (2002), the intensity metricscan be enhanced by combining these variables withpopulation and size-class data, both within andacross species. Finally, when the required data areavailable, multiple metrics should be computed sowe can evaluate their relative volatility andinterpret them jointly. As of now, only a few studieshave used more than one type of metric whenstudying the intensity of predation (e.g., Hansenand Kelley, 1995; Hoffmeister, 2002).Complicating Factors.—Regardless of themethods used, there are general complicatingfactors that need to be considered. Traces ofpredation represent a record of specific predatorpreyinteraction and there are many predators inecosystems that kill without leaving any evidencebecause they feed by whole-animal ingestion oraccess soft tissue without damaging the preyskeleton. There are also many predators thatdestroy prey skeletons entirely. For all thesereasons the assemblage-level estimates such as AFand ATF are likely to underestimate the overallpredation pressures. The interpretation becomeseven more ambiguous for specific predator-preysystems because frequencies of traces recorded inlower taxa (especially the LTF metric) can bothunderestimate and overestimate the intensity ofpredation. For example, a given frequency of testsof echinoids drilled by cassid snails mayunderestimate the importance of cassid-echinoidinteractions because cassids are facultative drillersthat occasionally access the soft tissues of their preyvia peristomal or periproctal membranes (Hughesand Hughes, 1981), and because drillings of cassidsmay be mistakenly attributed to parasitic eulimidgastropods (see Kowalewski and Nebelsick, inpress). On the other hand, the drilling frequencymay overestimate the importance of cassidsbecause many other predators of echinoids tend todestroy prey tests (e.g., Nebelsick, 1999;Kowalewski and Nebelsick, in press), resulting intoo high a percentage of preserved tests killed bydrilling predators. The importance of cassids mayalso be overestimated because eulimid drillings canbe mistakenly attributed to cassids (e.g.,Kowalewski and Nebelsick, in press).A unique host of problems affects repair scars,healed drill holes, and other traces that recordunsuccessful predation events. Such traces areinherently difficult to interpret (see also Schoener,1979; Schindel et al., 1982; Vermeij, 1983; Walkerand Voight, 1994; Cadée et al., 1997; Kowalewskiet al., 1997; Leighton, 2002) because they cannotbe used directly to estimate predation intensity: aprey population with a repair scar frequency of 20%may be preyed upon at much higher rates by anefficient predator or at much lower rates by aclumsy predator. In fact, some predators are knownto repeatedly attack unsuitable prey (e.g., Vermeij,1982), and thus it is feasible that a “prey” withfrequent repair scars is never subjected tosuccessful predation. Also, if a predator is (at leastoccasionally) successful, the repair scars representonly a subsample of all attacks. Unless theunsuccessful and successful attacks are statisticallyindistinguishable in their ecological and behavioralaspects (e.g., prey size, site of attack), a quantitativeanalysis of repair scars may provide misleadinginsights into predation.17

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 3—Comparison of three different metrics to estimate intensity of predation for the same sets ofsamples. Each point represents one sample. Symbols: + – highest per-taxon estimates plotted againstassemblage estimates, o – taxon estimates plotted against assemblage estimates, r – Spearman rankcorrelation coefficient (reported only if p < 0.05). A. Drilling predation estimates for 21 museum samplesof Permian brachiopods of West Texas (data from Hoffmeister, 2002). B. Drilling predation estimates for15 bulk samples of Miocene mollusks of central Europe (data from Hoffmeister and Kowalewski, 2001).18

KOWALEWSKI—ANALYTICAL METHODSFinally, traces of predation may both decreaseand increase the preservational potential of the preyskeleton, biasing quantitative estimates based onfrequency of specimens with traces (e.g., Roy et al.,1994; Hagstrom, 1999; Zuschin and Stanton, 2001;see also below for a discussion of taphonomic biases).How to Correct Predation Frequency Estimatesfor Disarticulated Elements.—The equationsdiscussed in the above section can be applied directlyto taxa with single-element skeletons (gastropodshells, foraminiferan tests, etc.). However, manyprey animals possess skeletons that consist of twoor more elements that tend to disarticulate afterdeath. If a predation event is recorded on one ofthose elements only, then the probability of findingevidence of predation (preserved by only onedisarticulated element) is smaller than the probabilityof finding prey (represented by any of its elements).Consider, for example, a bivalve mollusk killed bya predator that drilled a hole in one of its valves.Assuming that both valves have the samepreservational potential, the probability of findingone of the two valves of the prey is two times morelikely than finding specifically the valve that wasdrilled. Thus, a correction by a factor of 2 is required.It is worth stressing here that, regardless of whetherthe sampling domain is infinite and all sampledvalves are unique or the sampling domain is finiteand some valves come from the same individuals(see Gilinsky and Bennington, 1994), this correctionis required (see also Bambach and Kowalewski,2000; Hoffmeister and Kowalewski, 2001).The issue of correction may appear trivial but itturns out that there are two ways of making thiscorrection and both of them are used in the literature.Equation 5: f d= d/0.5nEquation 6: f d= 2d/n,where f drepresents the estimate of drilling frequency,d is the number of valves in the sample that containat least one successful drill hole, and n is the totalnumber of valves in the sample. These two equationsmay appear synonymous but, from a statisticalperspective, they are not. Equation 6 produces anestimate with a sample size that is two times higherthan an estimate produced by Equation 5.Consequently, Equation 6 offers much more powerthan Equation 5. Table 4 shows a hypotheticalexample of two samples of bivalves. If Equation 6is employed all statistical tests used indicate thatthe two samples differ significantly in drillingfrequency. If Equation 5 is used none of the testsrejects the null hypothesis that the two samplescame from a single underlying population. Whichequation is correct?The answer to this question is not intuitivelyobvious. Whereas Equation 6 doubles theTABLE 4—A hypothetical example illustrating differences in statistical power of the two equations usedto correct frequency estimates for drill holes in bivalved fossils. Symbols: N – total number of valves,D – number of drilled valves, R – drilling frequency, P chi, P G, and P Fisher– The probability estimates (Chisquare,Log-likelihood G, and Fisher’s Exact tests, respectively) for the null hypothesis that the twosamples came from a population with the same drilling frequency. All tests are significant at alpha=0.05level for Equation 6, but none is significant for the more conservative Equation 5. Computer simulations(Fig. 4) show that Equation 5 yields correct estimates of Type I Error.Sample 1 Sample 2Drilling IntensityEquation 5 Equation 6N 1 =30 N 2 =40 N 1 =15, N 2 =20; D 1 =11, D 2 =5 N 1 =30, N 2 =40; D 1 =22, D 2 =10D 1 =11 D 2 =5 R 1 =73.3%, R 2 =25% R 1 =73.3%, R 2 =25%R 1 =36.7% R 2 =12.5% P Chi =0.09, P G =0.08, P Fish =0.13 P Chi =0.02, P G =0.01, P Fish =0.0219

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002observations in the numerator, increasing the powerof the test, Equation 5 halves the number ofobservations in the denominator, decreasing thepower of the test. A simple computer simulation canresolve this issue. If we draw random samples froma known underlying distribution and use a=5% toreject the null hypothesis, which we know to becorrect in this case, we should reject incorrectly 5%of the tests. Results show that, regardless of samplesize, when Equation 5 is used ca. 5% of tests arerejected and if Equation 6 is used over 11% of testsare rejected (Fig. 4). This simulation indicates thatEquation 5 performs correctly and should beemployed in future studies whereas Equation 6clearly is too powerful and should not be used tocorrect for disarticulated elements.Note here that the example considered aboveassumes the following: (1) the two opposite valvesare equally likely to be preserved; (2) the predatoralways produces a trace in one valve only; (3) thetrace does not weaken the skeleton; and (4) thepredator does not show valve selectivity. All theseassumptions are questionable, and more complexcorrective strategies (most likely, based on theBayesian approach) should be developed in the future.Escalation parameters.—Escalation parametersare estimates that provide some measure of thepredator’s failure. A relative frequency of failedattacks (often referred to as “prey effectiveness”;e.g., Vermeij, 1987; Alexander and Dietl, in press;Kelley et al., 2001), as recorded by tracesdocumenting unsuccessful attacks (e.g., repair scarFIGURE 4—A series of computer simulations testing the statistical power of Equations 5 and 6. In thesimulation, samples of specimens are drawn randomly from an infinite population of disarticulated valvesof bivalve mollusks with a predefined drilling frequency of 50%. The correct null hypothesis (drillingfrequency = 50%) is then tested for each random sample using Fisher’s Exact Test and alpha=0.05.When Equation 5 is used the Type Error I (the erroneous rejection of the correct null hypothesis) variesaround 5% (mean=5.32) indicating that this test performs correctly. When Equation 6 is used over 11% oftests are significant indicating that Equation 6 offers over two times more statistical power than it should.20

KOWALEWSKI—ANALYTICAL METHODSor healed drill hole), is the most frequent metric ofescalation. It is typically computed as follows:Equation 7: P E= T F/T T,where P Edenotes prey effectiveness computed asT F, the number of a particular trace fossil ofpredatory origin that records failed attacks (e.g.,the number of incomplete and healed drill holes),divided by T T, the total number of these tracefossils (e.g., the total number of drill holes). Notethat, unlike for the intensity metrics above, thenumbers are computed using the number of tracesand not the number of specimens with traces (if apredator left two repair scars on a single prey, theprey survived twice, not once, and the predatorfailed twice, not once).As intensity metrics, escalation parameters andescalation tests are not without problems. Forexample, in a case of repair scars, numerousconfounding factors need to be considered (basedpartly on Alexander and Dietl, in press):1. No single method for quantifying shellbreakage can be applied when prey include a widerange of taxonomy and morphology (e.g., highlyornamented forms provide many more indicationsof repair than do smooth forms) (see also Schindelet al., 1982; Cadée et al., 1997).2. Repairs may accumulate on skeletal partsthat are less readily preserved (e.g., opercula butnot the shell of a gastropod; see Alexander andDietl, in press).3. Lethal shell damage is often unrecognizableand thus the denominator of Equation 7 isunderestimated and the prey effectiveness isoverestimated.4. Predators may mistakenly attack a skeletonof a dead prey. Consequently, post-mortem attackscan be confused with successful attacks, thedenominator of Equation 7 is overestimated, andthe prey effectiveness is underestimated (seeespecially Walker and Yamada, 1993).5. Prey skeletons are often completely destroyedby predators (e.g., Alexander and Dietl, in press).6. Frequency of attacks can be severelyunderestimated in the case of prey that experienceecdysis. For example, if a trilobite molted 5 timesand survived one unsuccessful attack, the frequencyof repairs will be underestimated five-fold.7. Different morphs of species that displaysexual dimorphism or developmentalpolymorphism may be preyed upon with differentintensity and/or different predatory success.8. Spatial variability in escalation patterns mayobscure temporal trends (e.g., Hoffmeister andKowalewski, 2001).9. Ambiguities in distinguishing failed andsuccessful attacks can confuse the computation ofthe prey effectiveness (see above and Alexanderand Dietl, in press).10. Disarticulation may complicate computingthe prey effectiveness (e.g., Alexander and Dietl,2001).Despite this depressingly long list of problems,there are exceptions when the complicating factorscan be partly or entirely eliminated (e.g., Kohn andArua, 1999; Alexander and Dietl, in press).It is noteworthy that, very much as was the casefor the intensity metrics above, Equation 7 can becomputed at various scales—from specific lowertaxa affected by a single type of trace fossil up toentire assemblages of prey including various highertaxa and a wide range of trace fossil types. All ofthe above factors complicating the prey effectivenessanalysis are increasingly likely to mask or distortthe patterns when the taxonomic resolution isdecreased. Thus, the prey effectiveness is bestapplied to specific lineages (see also Vermeij, 2002).However, as in the case of intensity metrics and forsimilar reasons, we should not discard assemblageestimates completely. In this case again they mayserve as an important reference baseline and overallproxy of failure rates. The assemblage approach isagain more debatable when used to test specificcausative hypotheses (although, again, it is yet tobe demonstrated beyond reasonable doubt that theproblems above render these assemblage testscompletely invalid).ANALYSES OF SELECTIVITYThe second major analytical focus of researchon predation traces deals with selectivity patternsrevealed by non-random distribution of traces (1)across prey taxa (taxon selectivity), (2) on prey21

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002skeletons (site selectivity), and (3) among prey withdifferent sizes (size selectivity). This type of analysisis routinely included in detailed quantitative studiesof predation traces and can provide many importantinsights into the nature of predator-prey interactions.Taxon Selectivity.—Predators are often highlyselective in choosing the species (or lower taxon)of their prey. In the most general terms, taxonselectivity simply means that a given taxon isattacked more frequently than is expected by chance.Based on the work of Botton (1984), Alexander andDietl (in press) suggested that this type ofevaluation can be performed using the StraussIndex, which can be defined as follows:Equation 8: L I= R I—P I,where R Iis the percentage of specimens with tracesbelonging to taxon i computed relative to allspecimens with traces found in the assemblage, andP Iis the percentage of taxon i in the assemblage.This index can help us to detect prey taxa withunusually high or low proportion of traces.Another possible strategy is to apply computersimulations to evaluate how likely it is to obtainthe observed distribution of traces across lower taxafor a given sampling design. In an example shownin Fig. 5, all common genera of silicifiedbrachiopods from the Permian of West Texas (4452specimens from 37 genera; data from Hoffmeister,2002) are plotted and the frequency of traces ofpredation (drill holes) is marked by black parts ofthe bars. Notice that holes are generally rare(overall predation rate AF = 1.1%) so many generado not include any specimens with traces (theFIGURE 5—Evaluation of genus selectivity in drilling on Permian brachiopods from West Texas (datafrom Hoffmeister, 2002 and in prep.). See text for details. The SAS/IML code for the Monte Carlo modelshown in the inset plot is provided in Appendix 1.22

KOWALEWSKI—ANALYTICAL METHODSproportion of drilled brachiopod genera ATF =43.6%). It is difficult to assess visually if thedistribution of traces is random or not. A MonteCarlo model (see Appendix 1 for SAS/IML code)was therefore used to draw random samples of4452 specimens assigned to 37 genera ofbrachiopods, mimicking the actual sample sizes forthose genera. The simulated specimens were then“drilled” by the computer with an a priori assignedprobability of 1.1%. The inset plot shows the resultof the simulation. Only one time in 999 iterationswas the simulated ATF value lower than or equalto 43.6%, demonstrating that drill holes aredistributed non-randomly (i.e., if drilling wasrandom, significantly more genera should containholes than is observed in the data). The reportedp=0.002 includes 999 random values and the actualsample (see Manly, 1995).One should be careful in interpreting taxonselectivity, especially in the case of paleontologicaldata. Because of time-averaging, the assemblagemay contain prey taxa that never encountered thepredator even though they are preserved in the sameassemblage; so the “selectivity” may have nothingto do with active selection by predators but maysimply reflect the fact that predators never had achance to encounter some of their “contemporary”prey. Also, and this problem applies toneontological studies as well, the lack of traces maymean that the predators failed in their attacks, andnot that they did not try.Leighton (2002) points to another seriousproblem of selectivity analyses related to thesequential nature of many predator-preyinteractions. Because marine benthic predators tendto encounter one prey at a time, the frequency ofattacks may reflect the frequency of preyencounters, and not a preferential selection by anoptimally foraging predator. If, as Leighton (2002)argues, the “zero-one rule” is in effect (i.e., anyprey type that is ever taken will always be taken),then differences in frequency of drill holes or repairscars observed across prey taxa may reflect therelative abundance and/or accessibility of thoseprey taxa. Although the zero-one rule may bequestioned—many models postulate that thebehavior of individual organisms varies throughtime so any prey type that is ever taken does nothave to be always taken (e.g., Evolutionarily StableStrategy; see Dawkins, 1976 for an excellentreview)—the sequential nature of encountersmakes selectivity analyses based on relativefrequencies of predation traces questionable (seeLeighton, 2002 for more details).Site Selectivity.—The location of trace fossils leftby a predator on skeletons of its prey may provideuseful information about the behavior of that predatorand its interaction with its prey. Many predators,parasites, or amensal organisms are behaviorallystereotyped in showing a preference for a particularlocation for their attack or attachment site.Consequently, biotic traces often display non-randompatterns in terms of their spatial distribution on preyor host skeletons. The evaluation of such siteselectivity is useful for several reasons. First, thepresence of site selectivity provides strong evidencefor the biotic origin of traces (see section above oncriteria for identifying traces of predation). Second,the specific nature of site selectivity may provideclues as to the nature of biotic interactions recordedby traces (e.g., drill holes located around areas thatgive direct access to muscle tissues of victims aremore likely to represent predation than parasitism).Third, changes in site selectivity through time mayoffer a good tool for evaluating various evolutionarymodels (e.g., Hagadorn and Boyajian, 1997; Dietland Alexander, 2000). In exceptional circumstances,multimodal patterns in distribution of trace fossilsalong the growth axis of its prey may be used todetect seasonality of predation and growth curvesof the prey (e.g., Kowalewski and Flessa, 2000).Although the evaluation of site selectivity canbe applied to various types of traces, includingfractures, repair scars, cut marks, and tooth marks(e.g., Babcock and Robinson, 1989; Lyman, 1994;Kowalewski et al., 1997; Kowalewski and Flessa,2000), most of the studies in the marine fossil recordhave focused on drill holes in invertebrate shells(e.g., Reyment, 1971; Negus, 1975; Kelley, 1988;Kowalewski, 1990; Anderson, 1992; Roopnarineand Beussink, 1999; Dietl and Alexander, 2000;Hoffmeister and Kowalewski, 2001; and many23

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 624

KOWALEWSKI—ANALYTICAL METHODSreferences therein). This is not surprisingconsidering that drill holes are highly localizedtraces (i.e., attack sites can be unambiguouslydetermined) that often represent a uniform recordof a single behavior and are frequent enough tomake statistical analyses possible.The strategies used to study site selectivity aredependent on prey type and scale of analysis. Theyinclude two general approaches: (1) among-elementselectivity and (2) within-element selectivity.Among-element tests for selectivity are basedon evaluating distribution of traces across differentskeletal elements or symmetry axes. Naticids mayprefer to drill a left rather than a right valve of abivalve mollusk prey, or a carnivore maypreferentially gnaw and chew particular types ofbones (e.g., Dodson and Wexlar, 1979; Kusmer,1990; Lyman, 1994). The evaluation methods arestraightforward analytically (although biologicalinterpretations may not be—see especially Kaplanand Baumiller, 2000; Wilson and Palmer, 2001;Kaplan and Baumiller, 2001) and typically involvea comparison of expected frequency of traces(given the relative frequency of elements) againsttheir observed frequencies. The Fisher’s exact testcan be applied for cases when two elements areinvolved, and Homogeneity (G-test, Chi-square)and/or Kolmogorov-Smirnov tests can be appliedfor multiple elements. If elements vary in theirpreservational potential, a more complicatedanalysis can be considered: Kaplan and Baumiller(2001) combined trace fossil distribution patternswith taphonomic information in a way somewhatakin to Bayesian statistical methods.Within-element methods test for nonrandomnessin the spatial distribution of tracefossils on a particular type of skeletal element: forexample, distribution of repair scars on brachiopodvalves along the growth axis (e.g., Kowalewski etal., 1997), angular distribution of drill holes on snailshells (e.g., Dietl and Alexander, 2000), or spatialdistribution of attack sites across the symmetry axisof a bilateral organism (e.g., Babcock, 1993).Numerous analytical strategies have beendeveloped over the years. The five main types canbe distinguished here:1. Qualitative Approach is based onsuperimposing all traces on a single “standard”element (Fig. 6a). This method is highly imprecise(because of the high potential operator error andmorphological and allometric variability amongspecimens) and produces data that cannot beconverted to data usable in other approaches listedbelow. On the other hand, the approach allows usto define sectors that are biologically meaningful,and it can be used to test very specific hypothesesabout the nature of traces (see especially Leighton,2001; Zlotnik, 2001).2. Sector Approach is based on partitioning theprey skeleton into sectors and tallying thefrequency of traces in each sector. The resultingdistribution can then be evaluated statistically usinghomogeneity tests (e.g., Reyment, 1971; Kowalewskiet al., 1997), the Shannon-Weaver Evenness Index(Dietl et al., 2001), or computer-intensive methods(e.g., Kowalewski et al., 1997). The sector-basedapproaches include two variants. Uneven-sectorapproach is a widely used strategy (e.g., Kelley,1988; Kowalewski, 1990; Anderson, 1992;Hoffmeister and Kowalewski, 2001) based on←FIGURE 6—Examples of strategies for evaluating site selectivity of predation. All examples are based ondrill holes in marine invertebrate prey. A. Qualitative approach based on plotting drill holes on a standardizedprey skeleton. The diagram shows distribution of large (>1mm) (triangles) and small (

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002partitioning a skeleton into a small number(typically

KOWALEWSKI—ANALYTICAL METHODSwhich the size of a trace is evaluated against the sizeof a prey specimen that contains that trace (e.g.,Kitchell et al., 1981; Kowalewski, 1990; Anderson,1992; Harper et al., 1998). The standard correlationtests can be applied here. However, size data tend tobe non-normally distributed and rarely representcontinuous variables (e.g., drill holes may vary byas little as 1 mm so even if the measurementprecision is 0.1 mm the resulting variable is discretein practice: we group holes in 10 or so sizecategories). Consequently, Spearman rankcorrelation rather than Pearson correlation shouldbe employed for analysis. The size estimate istypically based on the maximum dimensions of askeleton. However, Centroid Size (CS) is preferredhere because this landmark-based measure tends tobe isometric (e.g., Dryden and Mardia, 1998). CS isespecially convenient to use when geometricmorphometric methods are applied to quantify preyshape and trace fossil position (see above).In the case of drill holes, several studiesintegrated size selectivity with cost-benefit analysis(e.g., Kitchell et al., 1981; Kelley, 1988; Andersonet al., 1991; see recent review by Kelley andHansen, in press, for more examples). In suchanalyses the cost-benefit analysis is used toestimate the expected size class of prey that shouldbe preferred by the predator, and the resultingprediction is contrasted against the observed patternestimated from the distribution of traces across allsize classes of prey (but see Leighton, 2002).Size analysis is prone to various biases. In thecase of repair scars, the larger specimens tend tohave more scars for two reasons that have nothingto do with size selectivity. First, the larger preyare often more likely to survive predatory attacks;and second, the larger prey lived longer andtherefore had a higher chance of encounteringpredators (e.g., Vermeij and Dudley, 1982;Vermeij, 1987; Kowalewski et al., 1997;Alexander and Dietl, in press). Size analysis mayalso be particularly sensitive to taphonomicproblems. A large trace made in a small skeletonmay substantially weaken that skeleton and makeits preservation less likely. Notice that typicallya positive correlation is expected intuitively(larger predators eat larger prey), but suchcorrelation can be enhanced because the skeletonof a small prey attacked by a large predator is lesslikely to be preserved than a skeleton of a largeprey attacked by a large predator.TESTING FORTAPHONOMIC BIASESTraces of predation may weaken prey skeletonsor affect their hydrodynamic properties. Even iftraces do not affect the skeleton notably, postmortemprocesses may obscure or remove tracesleft by predators. Consequently, taphonomic filtersranging from pre-burial processes to compactionand diagenesis may bias quantitative patternspreserved in the geological record. For example,in the case of drill holes, Lever et al. (1961)demonstrated that drilled specimens weretransported farther up the beach than undrilledvalves (note that most fossil samples used inpaleoecological studies are from deposits thataccumulated in more offshore settings, for whichthe results of the Lever et al. (1961) study may notapply directly). Such post-mortem sorting maydistort quantitative estimates of drilling intensity(see also Kornicker et al., 1961). Roy et al. (1994)used an experimental approach to show that drilledvalves of the mactrid bivalve Mulinia were weakerunder point-load compression than were undrilledvalves. On the other hand, sediment compactionexperiments indicate that drilled specimens of thebivalve Anadara break less frequently thancomparable undrilled valves from the same genus(Zuschin and Stanton, 2001; see also Kaplan andBaumiller, 2000); and Hagstrom (1996) showedthat the point-load weakening observed by Roy etal. (1994) may be a serious problem only in highenergyenvironments.Many other taphonomic processes may distortthe data. Post-mortem encrusters may veneer overa repair scar left by a predator, and incomplete drillholes can become “complete” through subsequentremoval of their thin bottoms (e.g., the translucent,ultra-thin flooring of incomplete drill holesobserved commonly in Spisula solidissima [R.R.27

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Alexander, pers. comm., 2002] can easily beremoved by taphonomic processes).The evaluation of paleoecological data onpredation can involve both experimental andindirect taphonomic approaches. An experimentalapproach is exemplified by the work of Kaplan andBaumiller (2000), who performed a series ofexperiments with casts mimicking the morphologyof the studied prey (the brachiopod Onniella) andused the results of their experiment to evaluate fortaphonomic biases.Taphonomic data can also be used indirectlyto evaluate biases. Roy et al. (1994) suggested aset of simple questions to assess for taphonomicbias in drill holes. Do fragments of prey skeletonfrequently show partial traces? Frequent partialtraces indicate that most of the fractures passthrough the trace fossils and may have beeninitiated by those traces. This pattern suggests thata bias exists. Are fragments generally common inthe fossil assemblage? If they are rare, thefragmentation bias cannot be severe.Nebelsick and Kowalewski (1999), in a studyof drilling predation on echinoids, proposed a simpletaphonomic model to test for bias. They argued that,if drill holes have no taphonomic effect, theirdistribution should be independent of thetaphonomic alteration of drilled tests (i.e., uniformacross taphonomic grades); if drill holes affect thepreservation potential of echinoids by substantiallyweakening their tests, the proportion of drilled testsshould decrease with the increase in taphonomicalteration; and, finally, if drill holes are oftaphonomic (post-mortem) origin (i.e., they arepseudo-drillings), a proportion of drilled tests shouldincrease with the increase in taphonomic alteration.Nebelsick and Kowalewski (1999) then showed thatdrilled specimens are as common among testsseverely altered by taphonomic processes as amongtests that are still pristine; that is, the proportion ofdrilled tests does not decrease or increase with theincrease in taphonomic alteration of the tests (Fig. 7).They concluded that drill holes are unlikely to havea serious taphonomic effect even for the small, thintests of the clypeasteroid echinoids used in theirstudy. They noted, however, that the neontologicalmaterial they studied, unlike most of the fossilizedtests, was unaffected by compaction, during whichpreferential breakage of drilled tests would be morelikely to occur. Finally, in a daring study, Kaplanand Baumiller (2000) proposed the use oftaphonomic data to correct for biases in aquantitative way. This approach, specificallydesigned for bivalved organisms, estimates themagnitude of the differential bias in the preservationof opposite valves of an organism and uses theresulting estimate to correct the frequency data oftrace fossils found in those valves.METHODOLOGICALRECOMMENDATIONSIt is clear from the above review that there isa multitude of approaches for collecting traces ofpredation, processing and tabulating the resultingmaterials, analyzing the resulting data, andinterpreting the analytical outputs. It is also clearthat interpretations are rarely unambiguousbecause of the complexity of ecologicalinteractions, the confounding effects of abioticfactors, and the obscuring and biasing effects oftaphonomic processes.It would be foolish to suggest at this point (orperhaps at any point in the development of a scientificdiscipline) that we should erect strict guidelinesregarding how to collect, analyze, or interpret the data(see also Feyerabend, 1978). Consequently, whereasI do propose here some general methodologicalrecommendations, they are primarily geared towardmaking our data more compatible and readable (i.e.,more useful to other researchers).1. Given a wide range of data collecting andprocessing strategies, a method that maximizes thecompatibility of resulting data with futurecomparative analyses should be preferred. Forexample, if possible, fine mesh size should be usedin sieving the samples because that way data canbe compared (by eliminating analytically smallspecimens) to other datasets that were processedwith coarser meshes.2. Results should be reported in a clear mannerso that future researchers can combine the reported28

KOWALEWSKI—ANALYTICAL METHODSFIGURE 7—Taphonomic comparison of drilled and undrilled tests of clypeasteroid echinoids from theNorthern Bay of Safaga, Red Sea. A. Fibularia ovulum. B. Echinocyamus crispus. Symbols: n – samplesize. Modified after Kowalewski and Nebelsick (in press). Taphonomic Grade is a semi-quantitative rankvariable that varies from 1 for the least altered echinoid tests to 10 for the most heavily altered tests.data with other datasets or compute differentindices than those computed in an original study.Also, one should not restrict her/his reports toprocessed/corrected results from which raw datacannot be recomputed. We should followStuckenrath’s (1977, p. 187) plea not to correct rawdata because “…eventually these corrected [data]will have to be uncorrected in order to berecorrected in order to be correct…”.3. When faced with different tabulating andanalytical strategies, a method that makes theresulting data usable in the widest possible rangeof analytical approaches should be preferred. Forexample, if possible, data should be collected in away that allows one to compute various measuresof predation intensity rather than only one metric.A particular metric may indeed be the most usefuland appropriate for a specific case study, but othersmay want to use the data for other reasons (eventhe authors may, perhaps, agree retrospectively thattheir preferred metric does not work for these newresearch goals).4. When faced with a multitude of analyticalchoices, one should keep in mind that the choiceof the method may not always be important. Ifpossible, a comparison of different metrics mayhelp us to resolve the issue (if all correlate highlythe approach may not matter, but if they vary a lot29

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002one needs to make an explicit claim as to why oneof them is selected for the analysis).5. Various protocols and analytical methods canbe fruitful depending on the logistic circumstancesand scientific goals of a paleoecological project.6. The most fruitful analyses are based onrelative comparative approaches. Latitudinalchanges in predation intensity or temporal shifts inbehavioral stereotypy are more insightful than a givenabsolute frequency of predation or a specific degreeof stereotypy. Comparative analyses can also helpavoid various biases. For example, taphonomic biasor spatial/environmental overprint can be minimizedif comparisons are done for samples from similartaphonomic settings and comparable depositionalenvironments (the “isotaphonomic approach” ofBehrensmeyer and Hook, 1992).7. Finally, regardless of the above points, itseems particularly useful to provide raw data (eitheras repository data or appendices) so that futureresearchers can re-analyze these data in new ways.Consider all the data on predation traces that havebeen collected over the last 40 years and cannot beaccessed. At best, a few succinct tables and graphsare all that remain. This is the one mistake we neednot repeat in our future efforts.CLOSING REMARKSThe methodological dimension of research onpredation traces is a rapidly growing field of study.Based on current activities, the futuremethodological themes that are likely to benefitour discipline include (1) laboratory experimentsthat should help us in dealing with varioustaphonomic biases, (2) neontological analyses thatprovide reference baselines and should further helpus to understand various confounding factors thatneed to be accounted for before proposing anyinterpretation (e.g., spatial gradients in predation),and (3) numerical modeling that should continueto improve our arsenal of statistical tools andanalytical strategies.Despite all caveats and problems, distincttraces of predation such as drill holes offer one ofthe best sources of quantitative data inpaleoecology. Such traces provide unusuallyfavorable research conditions for testing newmethodologies and for pushing our interpretivepowers to the highest possible limits. Research onpredation traces can thus be viewed as one of theforemost areas for testing the scientific limits ofour discipline—by examining traces of predationwe can examine the limits, strengths, methods, andassumptions of paleoecology.ACKNOWLEDGMENTSThis project was supported by the NSF grantEAR-9909225. I thank Richard Alexander, GregDietl, Thor Hansen, Alan Hoffmeister, PatriciaKelley, and Lindsey Leighton for numerouscomments and suggestions that improved thismanuscript considerably.REFERENCESALEXANDER, R. R. 1981. Predation scars preserved in Chesterian brachiopods: probable culprits and evolutionaryconsequences for the articulates. Journal of Paleontology, 55:192–203.ALEXANDER, R. R. 1986a. Resistance to repair of shell breakage induced by durophages in Late Ordovicianbrachiopods. Journal of Paleontology, 60:273–285.ALEXANDER, R. R. 1986b. Frequency of sublethal shell-breakage in articulate brachiopod assemblages throughgeologic time. In P. R. Racheboeuf and C. C. Emig (eds.), Les Brachiopodes Fossiles et Actuels, FirstInternational Brachiopod Congress Proceedings, Biostratigraphie du Paleozoique, 4:159–166.ALEXANDER, R. R., AND G. P. DIETL. 2001. Shell repair frequencies in New Jersey bivalves: a recent baseline fortests of escalation with Tertiary, Mid-Atlantic congeners. Palaios, 16:354–371.30

KOWALEWSKI—ANALYTICAL METHODSALEXANDER, R. R., AND G. P. DIETL. In press. The Phanerozoic history of shell-breaking predation on marinebivalves and gastropods. In P. H. Kelley, M. Kowalewski, and T. A. Hansen (eds.), Predator-Prey Interactionsin the Fossil Record. Topics in Geobiology Series, Plenum Press/Kluwer, New York.ALLMON, W. D., J. C. NIEH, AND R. D. NORRIS. 1990. Drilling and peeling of turritelline gastropods since the LateCretaceous. Palaeontology, 33:595–611.ALPERT, S. P., AND J. N. MOORE. 1975. Lower Cambrian trace fossil evidence for predation on trilobites. Lethaia,8:223–230.AMBROSE, R. F. 1986. Effects of Octopus predation on motile invertebrates in a rocky subtidal community. MarineEcology Progress Series, 30:261–273.ANDERSON, L. C. 1992. Naticid gastropod predation on corbulid bivalves: Effects of physical factors, morphologicalfeatures, and statistical artifacts. Palaios, 7:602–620.ANDERSON, L. C., D. H. GEARY, R. H. NEHM, AND W. D. ALLMON. 1991. A comparative study of naticid gastropodpredation on Varicorbula caloosae and Chione cancellata, Plio-Pleistocene of Florida, U.S.A.Palaeogeography, Palaeoclimatology, Palaeoecology, 85:283–290.ANDREWS, P., AND Y. FERNANDEZ JALVO. 1998. 101 uses for fossilized faeces. Nature, 393:629–630.ARNOLD, A. J., F. D’ESCRIVAN, AND W. C. PARKER. 1985. Predation and avoidance responses in the Foraminifera ofthe Galapagos hydrothermal mounds. Journal of Foraminiferal Research, 15:38–42.ARNOLD, J. M., AND K. O. ARNOLD. 1969. Some aspects of hole-boring predation by Octopus vulgaris. AmericanZoologist, 9:991–996.ARUA, I., AND M. HOQUE. 1989. Fossil predaceous gastropod borings from Nigeria. Palaeogeography,Palaeoclimatology, Palaeoecology, 73:175–183.AUSICH, W. I., AND R. A. GURROLA. 1979. Two boring organisms in a Lower Mississippian community of southernIndiana. Journal of Paleontology, 53:335–344.BABCOCK, L. E. 1993. Trilobite malformations and the fossil record of behavioral asymmetry. Journal ofPaleontology, 67:217–229.BABCOCK, L. E., AND R. A. ROBINSON. 1989. Preferences of Palaeozoic predators. Nature, 337:695–696.BAUK, W., AND A. RADWASKI. 1996. Stomatopod predation upon gastropods from the Korytnica Basin, and fromother classical Miocene localities in Europe. Acta Geologica Polonica, 46:279–304.BAMBACH, R. K. 2002. Supporting predators: Changes in the global ecosystem inferred from changes in predatordiversity. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record of Predation. Paleontological SocietySpecial Papers, 8 (this volume).BAMBACH, R. K., AND M. KOWALEWSKI. 2000. How to count fossils. Geological Society of America Abstracts withPrograms, 32:A-95.BAUMILLER, T. K. 1990. Non-predatory drilling of Mississippian crinoids by platyceratid gastropods. Palaeontology,33:743–748.BAUMILLER, T. K. 1993. Boreholes in Devonian blastoids and their implications for boring by platyceratids. Lethaia,26:41–47.BAUMILLER, T. K. 1996. Boreholes in the Middle Devonian blastoid Heteroschisma and their implications forgastropod drilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 123:343–351.BAUMILLER, T. K. 2002. Multi-snail infestation of Devonian crinoids and the nature of platyceratid-crinoidinteractions. Acta Palaeontologica Polonica, 47:133–139.BAUMILLER, T. K., AND F. GAHN. 2002. Fossil record of parasitism on marine invertebrates with special emphasison the platyceratid-crinoid interactions. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record ofPredation. Paleontological Society Special Papers, 8 (this volume).BAUMILLER, T. K., AND D. B. MACURDA, JR. 1995. Borings in Devonian and Mississippian blastoids (Echinodermata).Journal of Paleontology, 69:1084–1089.BAUMILLER, T. K., L. R. LEIGHTON, AND D. THOMPSON. 1999. Boreholes in brachiopods of the Fort Payne Formation(Lower Mississippian, central USA). Palaeogeography, Palaeoclimatology, Palaeoecology, 147:283–289.BECKER, M. A., J. MEIER, AND W. SLATTERY. 1999. Spiral coprolites from the Upper Cretaceous Wenonah-Mt.Laurel and Navesink formations in the northern coastal plain of New Jersey. Northeastern Geology andEnvironmental Sciences, 21:181–187.31

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BEHRENSMEYER, A. K., AND R. W. HOOK. 1992, Paleoenvironmental contexts and taphonomic modes, p. 15–136. InA. K. Behrensmeyer, J. D. Damuth, W. A. DiMichele, R. Potts, H. D. Sues, and S. L. Wing (eds.), TerrestrialEcosystems through Time. The University of Chicago Press, Chicago.BENGTSON, S. 2002. Origins and early evolution of predation. In M. Kowalewski and P. H. Kelley (eds.), The FossilRecord of Predation. Paleontological Society Special Papers, 8 (this volume).BENGTSON, S., AND Z. YUE. 1992. Predatorial borings in late Precambrian mineralized exoskeletons. Science,257:367–369.BENNINGTON, J. B., AND S. D. RUTHEFORD. 1999. Precision and reliability in paleocommunity comparisons based oncluster-confidence intervals; how to get more statistical bang for your sampling buck. Palaios, 14:506–515.BERG, J. C. 1976. Ontogeny of predatory behavior in snails (Prosobranchia: Naticidae). Nautilus, 90:1–4.BERG, J. C., AND S. NISHENKO. 1975. Stereotypy of predatory boring behavior of Pleistocene naticid gastropods.Paleobiology, 1:258–260.BISHOP, G. A. 1975. Traces of predation, p. 261–281. In R. W. Frey (ed.), The Study of Trace Fossils. Springer-Verlag, New York.BISHOP, G. A. 1977. Pierre feces: A scatological study of the Dakoticancer assemblage, Pierre Shale (UpperCretaceous) of South Dakota. Journal of Sedimentary Petrology, 47:129–139.BOOKSTEIN, F. L. 1991. Morphometric Tools for Landmark Data. Cambridge University Press, New York, 435 p.BOTTON, M. L. 1984. Diet and food preference of the adult horseshoe crab Limulus polyphemus in Delaware Bay,New Jersey, USA. Marine Biology, 81:199–207.BOUCOT, A. J. 1990. Evolutionary Paleobiology of Behavior and Coevolution. Elsevier, Amsterdam, 725 p.BRANDT, D. S., D. L. MEYER, AND P. B. LASK. 1995. Isotelus (Trilobita) “hunting burrow” from Upper Ordovicianstrata, Ohio. Journal of Paleontology, 69:1079–1083.BRETT, C. E. In press. Predation in Paleozoic marine communities. In P. H. Kelley, M. Kowalewski, and T. A.Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topics in Geobiology Series, Plenum Press/Kluwer, New York.BRETT, C. E., AND S. E. WALKER. 2002. Predators and predation in Paleozoic marine environments. In M.Kowalewski and P. H. Kelley (eds.), The Fossil Record of Predation. Paleontological Society Special Papers,8 (this volume).BROMLEY, R. G. 1970. Borings as trace fossils and Entobia cretacea Portlock, as an example. Geological JournalSpecial Issue, 3:49–90.BROMLEY, R. G. 1981. Concepts in ichnotaxonomy illustrated by small round holes in shells. Acta GeologicaHispanica, 16:55–64.BROMLEY, R. G. 1993. Predation habits of octopus past and present and a new ichnospecies, Oichnus ovalis.Geological Society of Denmark Bulletin, 40:167–173.BROMLEY, R. G. 1996. Trace Fossils; Biology, Taphonomy and Applications. Chapman and Hall, 361 p.BRUNTON, H. 1966. Predation and shell damage in a Visean brachiopod fauna. Palaeontology, 9:355–359.BUNN, H. T., AND E. M. KROLL. 1986. Systematic butchery by Plio/Pleistocene hominids at Olduvai George,Tanzania. Current Anthropology, 27:431–452.CADÉE, G. C. 1968. Molluscan biocoenoses and thanatocoenoses in the Ria de Arosa, Galicia, Spain. ZoologischeVerhandelingen, Leiden, 95:1–121.CADÉE, G. C. 1994. Eider, shelduck, and other predators, the main producers of shell fragments in the WaddenSea; palaeoecological implications. Palaeontology, 37:181–202.CADÉE, G. C. 2000. Herring gulls feeding on a Recent invader in the Wadden Sea, Ensis directus. In TheEvolutionary Biology of the Bivalvia. Geological Society Special Publications, 177:459–464.CADÉE, G. C., S. E. WALKER, AND K. W. FLESSA. 1997. Gastropod shell repair in the intertidal of Bahia la Choya(N.Gulf of California). Palaeogeography, Palaeoclimatology, Palaeoecology, 136:67–78.CARPENTER, K. 2000. Evidence of predatory behavior by carnivorous dinosaurs. Gaia, 15:135–144.CARRIKER, M. R. 1943. On the structure and function of the proboscis in the common oyster drill, Urosalpinxcinerea Say. Journal of Morphology, 73:441–506.CARRIKER, M. R. 1957. Preliminary study of behavior of newly hatched oyster drills Urosalpinx cinerea (Say).Journal of the Elisha Mitchell Scientific Society, 73:328–351.32

KOWALEWSKI—ANALYTICAL METHODSCARRIKER, M. R. 1961. Comparative functional morphology of the boring mechanisms in gastropods. AmericanZoologist, 1:263–266.CARRIKER, M. R. 1981. Shell penetration and feeding by naticacean and muricacean predatory gastropods: asynthesis. Malacologia, 20:403–422.CARRIKER, M. R., AND D. VAN ZANDT. 1972a. Predatory behavior of a shell-boring muricid gastropod, p. 157–244.In H. E. Winn and B. L. Olla (eds.), Behavior of Marine Animals: Current Perspectives in Research, Vol. 1Invertebrates. Plenum Press, New York.CARRIKER, M. R., AND D. VAN ZANDT. 1972b. Regeneration of the accessory boring organ of muricid gastropodsafter excision. Transactions of American Microscope Society, 91:455–466.CARRIKER, M. R., AND E. L. YOCHELSON. 1968. Recent gastropod boreholes and Ordovician cylindrical borings.U.S. Geological Survey Professional Paper, 593-B:1–26.CARRION, J. S., J. A. RIQUELME, C. NAVARRO, AND M. MUNUERA. 2001. Pollen in hyaena coprolites reflects lateglacial landscape in southern Spain. Palaeogeography, Palaeoclimatology, Palaeoecology, 176:193–205.CATE, A. S., AND I. EVANS. 1994. Taphonomic significance of the biomechanical fragmentation of live molluscanshell material by a bottom-feeding fish (Pogonias cromis) in Texas coastal bays. Palaios, 9:254–274.DE CAUWER, G. 1985. Gastropod predation on corbulid bivalves; paleoecology or taphonomy? Annales du SociétéZoologique de Belgique, 115:183–196.CHATTERTON, B. D. E., AND H. L. WHITEHEAD. 1987. Predatory borings in inarticulate brachiopod Artiotreta fromthe Silurian of Oklahoma. Lethaia, 20:67–74.CHECA, A. 1993. Non-predatory shell damage in Recent deep-endobenthic bivalves from Spain. Palaeogeography,Palaeoclimatology, Palaeoecology, 100:309–331.CHIN, K. 1997. What did dinosaurs eat? Coprolites and other direct evidence of dinosaur diet, p. 371–382. In J. O.Farlow and M. K. Brett-Surman (eds.), The Complete Dinosaur. Indiana University Press, Bloomington.CHIN, K. 2002. Analyses of coprolites produced by carnivorous vertebrates. In M. Kowalewski and P. H. Kelley(eds.), The Fossil Record of Predation. Paleontological Society Special Papers, 8 (this volume).CHIN, K., T. T. TOKARYK, G. M. ERICKSON, AND L. CALK. 1998. A king-sized theropod coprolite. Nature, 393:680–682.COLBATH, S. L. 1985. Gastropod predation and depositional environment of two molluscan communities from theMiocene Astoria Formation at Beverly Beach State Park, Oregon. Journal of Paleontology, 59:849–869.CONWAY MORRIS, S., AND S. BENGTSON. 1994. Cambrian predators: Possible evidence from boreholes. Journal ofPaleontology, 68:1–23.COOPER, G. A., AND R. E. GRANT. 1972–1976. Permian brachiopods of West Texas, I–V. Smithsonian Contributions toPaleobiology, 14:1–231; 15:233–793; 19:795–1298; 19:1300–1921; 21:1923–2285; 21:2288–2607; 24:2609–3159.COY, C. E. 1995. The first record of spiral coprolites from the Dinosaur Park Formation (Judith River Group,Upper Cretaceous) southern Alberta, Canada. Journal of Paleontology 69:1191–1194.CROSSLAND, M. R., R. A. ALFORD, AND J. D. COLLINS. 1991. Population dynamics of an ectoparasitic gastropod,Hypermastus sp. (Eulimidae), on the sand dollar, Arachnoides placenta (Echinoidea). Australian Journal ofMarine and Freshwater Research, 42:69–76.CULVER, S. J., AND J. H. LIPPS. In press. Predation on and by foraminifera. In P. H. Kelley, M. Kowalewski, and T.A. Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topics in Geobiology Series, PlenumPress/Kluwer, New York, in press.DAWKINS, R. 1976. The selfish gene. Oxford University Press, New York, 224 p.DAY, J. A. 1969. Feeding of the cymatiid gastropod, Argobuccinum argus, in relation to the structure of theproboscis and secretions of the proboscis gland. American Zoologist, 9:909–919.DEGNER, E. 1928. Über das Fleisch und Kalkbeddurfnis von Cepaea nemoralis. Archiv für Molluskenk., 60:209–213.DIETL, G. P. 2000. Successful and unsuccessful predation of the gastropod Nucella lapillus (Muricidae) on themussel Mytilus edulis from Maine. Veliger, 43:319–329.DIETL, G. P., AND R. R. ALEXANDER. 1995. Borehole site and prey size stereotypy in naticid predation on Euspira(Lunatia) heros Say and Neverita (Polinices) duplicata Say from the southern New Jersey coast. Journal ofShellfish Research, 14:307–314.DIETL, G. P., AND R. R. ALEXANDER. 2000. Post-Miocene shift in stereotypic naticid predation on confamilial preyfrom the Mid-Atlantic shelf: Coevolution with dangerous prey. Palaios, 15:414–429.33

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002DIETL, G. P., AND P. H. KELLEY. 2001. Mid-Paleozoic latitudinal predation gradient: Distribution of brachiopodornamentation reflects shifting Carboniferous climate. Geology, 29:111–114.DIETL, G. P., AND P. H. KELLEY. 2002. The fossil record of predator-prey arms races: Coevolution and escalationhypotheses. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record of Predation. PaleontologicalSociety Special Papers, 8 (this volume).DIETL, G. P., R. R. ALEXANDER, AND W. F. BIEN. 2000. Escalation in Late Cretaceous–early Paleocene oysters(Gryphaeidae) from the Atlantic Coastal Plain. Paleobiology, 26:215–237.DIETL. G., R. R. ALEXANDER, P. H. KELLEY, AND T. A. HANSEN. 2001. Stereotypy of naticid predation on bivalves since theCretaceous: trends, controlling factors, and implications for escalation. PaleoBios 21(supplement to number 2):78.DODSON, P., AND D. WEXLAR. 1979. Taphonomic investigations of owl pellets. Paleobiology, 5:279–284.DRYDEN I. L., AND K. V. MARDIA. 1998. Statistical Shape Analysis. Wiley, New York, 347 p.FARLOW, J. O., AND T. R. HOLTZ. 2002. The fossil record of predation in dinosaurs. In M. Kowalewski and P. H.Kelley (eds.), The Fossil Record of Predation, Paleontological Society Special Papers, 8 (this volume).FEYERABEND, P. K. 1978. Against method: Outline of an Anarchistic Theory of Knowledge. Verso, London, 339 p.FISCHER, P. H. 1922. Sur les gastéropodes perceurs. Journal de Conchyliologie, 67:1–56.FISCHER, P. H. 1962a. Perforations de fossiles Pré-Tertiaries attribuées a des gastéropodes prédateurs. Journal deConchyliologie, 102:68–78.FISCHER, P. H. 1962b. Action des gastéropodes perceurs sur des Mesalia de L’Étage Lutétien. Journal deConchyliologie, 102:95–97.FISCHER, P. H. 1963. Corbules fossiles perforées par des gastéropodes prédateurs. Journal de Conchyliologie, 103:29–31.FISCHER, P. H. 1966. Perforations de fossiles Tertiaires par des gastéropodes prédateurs. Journal de Conchyliologie,106:66–96.FREY, R. W., J. D. HOWARD, AND J. S. HONG. 1986. Naticid gastropods may kill solenid bivalves without boring:Ichnologic and taphonomic consequences. Palaios, 1:610–612.FUJITA, S. 1916. On the boring of pearl oyster by Octopus (Polypus) vulgaris Lamarck. Dobytsugaku Zasshi, 28:250–257.GEARY, D. H., W. D. ALLMON, AND M. J. REAKA-KUDLA. 1991. Stomatopod predation on fossil gastropods from thePlio-Pleistocene of Florida. Journal of Paleontology, 65:355–360.GILINSKY, N. L., AND B. BENNINGTON. 1994. Estimating numbers of whole individuals from collections of bodyparts: A taphonomic limitation of the paleontological record. Paleobiology, 20:245–258.GITTENBERGER, E. 1999. Predatory bore-holes in shells of terrestrial snails: Roth has priority. Basteria, 63:164.GIBSON, M. A., AND J. B. WATSON. 1989. Predatory and non-predatory borings in echinoids from the Upper Ocala Formation(Eocene), north-central Florida, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 71:309–321.GUERRA, A., AND M. NIXON. 1987. Crab and mollusc shell drilling by Octopus vulgaris (Mollusca: Cephalopoda)in the Ria de Vigo (north-west Spain). Journal of Zoology, London, 211:515–523.HAGADORN, J. W., AND G. E. BOYAJIAN. 1997. Subtle changes in mature predator-prey systems; an example fromNeogene Turritella (Gastropoda). Palaios, 12:372–379.HAGEMAN, S. A., AND R. L. KAESSLER. 2002. Fusulinids: Predation damage and repair of tests from the UpperPennsylvanian of Kansas. Journal of Paleontology, 76:181–184.HAGSTROM, K. M. 1996. Effects of compaction and wave-induced forces on the preservation and macroevolutionaryperception of naticid predator-prey interactions. Unpublished Master’s Thesis, 63 p.HANSEN, T. A., AND P. H. KELLEY. 1995. Spatial variation in naticid gastropod predation in the Eocene of NorthAmerica. Palaios, 10:268–278.HARPER, E. M., G. T. W. FORSYTHE, AND T. PALMER. 1998. Taphonomy and the Mesozoic marine revolution:Preservation state masks the importance of boring predators. Palaios, 13:352–360.HARPER, E. M., G. T. W. FORSYTHE, AND T. PALMER. 1999. A fossil record full of holes: The Phanerozoic history ofdrilling predation: Comment. Geology, 27:959.HARPER, E. M., AND D. S. WHARTON. 2000. Boring predation and Mesozoic articulate brachiopods. Palaeogeography,Palaeoclimatology, Palaeoecology, 158:15–24.HAYNES, G. 2002. Archeological methods for reconstructing human predation on terrestrial vertebrates. In M. Kowalewskiand P. H. Kelley (eds.), The Fossil Record of Predation. Paleontological Society Special Papers, 8 (this volume).HENDERSON, W. G., L. C. ANDERSON, AND C. R. MCGIMSEY. 2002. Stratigraphic, taxonomic, and taphonomic criteriato distinguish shell-rich chenier and archaeological deposits of the Louisiana chenier plain. Palaios, in press.34

KOWALEWSKI—ANALYTICAL METHODSHIRSCH, G. C. 1915. Die Ernährungsbiologie fleischfressender Gastropoden (Murex, Natica, Pterotrachea,Pleurobranchea, Tritonium). I Teil. Makroskopischer Bau, Nahrungsaufnahme, Verdauung, Sekretion.Zoologische Jahrbuch, 35:357–504.HIRSCH, K. F., A. J. KIHM, AND D. K. ZELENITSKY. 1997. New eggshell of ratite morphotype with predation marksfrom the Eocene of Colorado. Journal of Vertebrate Paleontology, 17:360–369.HOFFMAN, A., AND J. MARTINELL. 1984. Prey selection by naticid gastropods in the Pliocene of Emporda (NortheastSpain). Neues Jahrbuch für Geologie und Paläontologie, Monatshaft, 1984:393–399.HOFFMAN, A., A. PISERA, AND M. RYSZKIEWICZ. 1974. Predation by muricid and naticid gastropods on the LowerTortonian mollusks from the Korytnica Clays. Acta Geologica Polonica, 24:249–260.HOFFMEISTER, A. P. 2002. Quantitative analysis of drilling predation patterns in the fossil record: Ecological and evolutionaryimplications. Unpublished Ph.D. dissertation, Virginia Polytechnic Institute and State University, Blacksburg, 293 p.HOFFMEISTER, A.P., AND M. KOWALEWSKI. 2001. Spatial and environmental variation in the fossil record of drillingpredation: A case study from the Miocene of Central Europe. Palaios, 16:566–579.HOFFMEISTER, A. P., M. KOWALEWSKI, R. K. BAMBACH, AND T. K. BAUMILLER. Submitted. Intense drilling predationon the brachiopod Cardiarina cordata Cooper 1956 from the Pennsylvanian of New Mexico. Lethaia.HUGHES, R. N., AND P. I. HUGHES. 1971. A study on the gastropod Cassis tuberosa (L.) preying upon sea urchins.Journal of Experimental Marine Biology and Ecology, 7:305–314.HUGHES, R. N., AND H. P. I. HUGHES. 1981. Morphological and behavioural aspects of feeding in the Cassidea(Tonnacea, Mesogastropoda). Malacologia, 20:385–402.JACOBSEN, A. R. 1997. Tooth marks, p. 738–739. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs.Academic Press, San Diego.JACOBSEN, A. R. 1998. Feeding behavior of carnivorous dinosaurs as determined by tooth marks on dinosaurbones. Historical Biology, 13:17–26.JENSEN, S. 1990. Predation by Early Cambrian trilobites on infaunal worms; evidence from the Swedish MickwitziaSandstone. Lethaia, 23:29–42.KABAT, A. R. 1990. Predatory ecology of naticid gastropods with a review of shell boring predation. Malacologia,32:155–193.KAPLAN, P., AND T. K. BAUMILLER. 2000. Taphonomic inferences on boring habit in the Richmondian Onniellameeki epibole. Palaios, 15:499–510.KAPLAN, P., AND T. K. BAUMILLER. 2001. A misuse of Occam’s Razor that trims more than just the fat. Palaios,16:525–528.KASE, T., P. A. JOHNSTON, A. SEILACHER, AND J. B. BOYCE. 1998. Alleged mosasaur bite marks on Late Cretaceousammonites are limpet (patellogastropod) home scars. Geology, 26:947–950.KAUFFMAN, E. G., AND R. V. KESLING. 1960. An Upper Cretaceous ammonite bitten by a mosasaur (South Dakota).Contributions from the Museum of Paleontology, University of Michigan, 15:193–248.KELLEY, P. H. 1988. Predation by Miocene gastropods of the Chesapeake Group: Stereotyped and predictable.Palaios, 3:436–448.KELLEY, P. H., AND T. A. HANSEN. 1993. Evolution of the naticid gastropod predator-prey system: An evaluation ofthe hypothesis of escalation. Palaios, 8:358–375.KELLEY, P. H., AND T. A. HANSEN. 1996. Naticid gastropod prey selectivity through time and the hypothesis ofescalation. Palaios, 11:437–445.KELLEY, P. H., AND T. A. HANSEN. In press. The fossil record of drilling predation on bivalves and gastropods. In P.H. Kelley, M. Kowalewski, and T. A. Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topicsin Geobiology Series, Plenum Press/Kluwer, New York.KELLEY, P. H., T. A. HANSEN, S. E. GRAHAM, AND A. G. HUNTOON. 2001. Temporal patterns in the efficiency ofnaticid gastropod predators during the Cretaceous and Cenozoic of the United States Coastal Plain.Palaeogeography, Palaeoclimatology, Palaeoecology, 166:165–176.KITCHELL, J. A. 1986. The Evolution of predator-prey behavior: naticid gastropods and their molluscan prey, p.88–110. In M. H. Nitecki and J. A. Kitchell (eds.), Evolution of Animal Behavior: Paleontological and FieldApproaches. Oxford University Press, New York.35

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002KITCHELL, J. A., C. H. BOGGS, J. F. KITCHELL, AND J. A. RICE. 1981. Prey selection by naticid gastropods: Experimentaltests and application to the fossil record. Paleobiology, 7:533–552.KOHN, A. J., AND I. ARUA. 1999. An Early Pleistocene molluscan assemblage from Fiji: Gastropod faunal composition,paleoecology and biogeography. Palaeogeography, Palaeoclimatology, Palaeoecology, 146:99–145.KORNICKER, L. S., C. D. WISE, AND J. M. WISE. 1961. Factors affecting the distribution of opposing molluskvalves. Journal of Sedimentary Petrology, 33:703–712.KOSUGE, S., AND S. HAYASHI. 1967. Notes on the feeding habits of Capulus dilatatus A. Adams, 1860 (Gastropoda).Scientific Reports of Yokosuka City Museum, 13:45–54.KOWALEWSKI, M. 1990. A hermeneutic analysis of the shell-drilling gastropod predation on mollusks in the KorytnicaClays (Middle Miocene; Holy Cross Mountains; Central Poland). Acta Geologica Polonica, 40:183–213.KOWALEWSKI, M. 1993. Morphometric analysis of predatory drillholes. Palaeogeography, Palaeoclimatology,Palaeoecology, 102:69–88.KOWALEWSKI, M., AND K. W. FLESSA. 1994. A predatory drillhole in Glottidia palmeri Dall (Brachiopoda; Lingulidae)from Recent tidal flats of northeastern Baja California, Mexico. Journal of Paleontology, 68:1403–1405.KOWALEWSKI, M., AND K. W. FLESSA. 2000. Seasonal predation by migratory shorebirds recorded in shells oflingulid brachiopods from Baja California, Mexico. Bulletin of Marine Science, 66:405–416.KOWALEWSKI, M., AND J. H. NEBELSICK. In press. Predation and parasitism on recent and fossil echinoids. In P. H.Kelley, M. Kowalewski, and T. A. Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topics inGeobiology Series, Plenum Press/Kluwer, New York.KOWALEWSKI, M., K. W. FLESSA, AND J. D. MARCOT. 1997. Predatory scars in the shells of a Recent lingulidbrachiopod: Paleontological and ecological implications. Acta Palaeontologica Polonica, 42:497–532.KOWALEWSKI, M., A. DULAI, AND F. T. FÜRSICH. 1998. A fossil record full of holes: The Phanerozoic history ofdrilling predation. Geology, 26:1091–1094.KOWALEWSKI, M., M. G. SIMÕES, F. F. TORELLO, L. H. C. MELLO, AND R. P. GHILARDI. 2000. Drill holes in shells ofPermian benthic invertebrates. Journal of Paleontology, 74:532–543.KOWALEWSKI, M., K. GÜRS, J. H. NEBELSICK, W. OSCHMANN, W. PILLER, AND A. P. HOFFMEISTER. 2002. Multivariatehierarchical analyses of Miocene mollusks of Europe: Paleogeographic, biostratigraphic, and paleoecologicalimplications. Geological Society of America Bulletin, 114:239–256.KUSMER, K. D. 1990. Taphonomy of owl pellet deposition. Journal of Paleontology, 64:629–637.LABANDEIRA, K. 2002. Paleobiology of predators, parasitoids, and parasites: Death and accommodation in thefossil record of continental invertebrates. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record ofPredation. Paleontological Society Special Papers, 8 (this volume).LAWRENCE, J. M., AND J. VASQUEZ. 1996. The effects of sublethal predation on the biology of echinoderms.Oceanologica Acta, 19:431–440.LEIGHTON, L. R. 1999. Possible latitudinal predation gradient in middle Paleozoic oceans. Geology, 27:47–50.LEIGHTON, L. R. 2001. New example of Devonian predatory boreholes and the influence of brachiopod spines onpredator success. Palaeogeography, Palaeoclimatology, Palaeoecology, 165:53–69.LEIGHTON, L. R. 2002. Inferring predation intensity in the marine fossil record. Paleobiology, 28: in press.LESCINSKY, H. L., AND L. BENNINGER. 1994. Pseudo-borings and predator traces; artifacts of pressure-dissolution infossiliferous shales. Palaios, 9:599–604.LEVER, J., A. KESSLER, P. VAN OVERBEEKE, AND R. THIJSSEN. 1961. Quantitative beach research, II, The ‘hole effect’: Asecond mode of sorting lamellibranch valves on sandy beaches. Netherlands Journal of Sea Research, 1:339–358.LIPPS, J. H. 1988. Predation on foraminifera by coral reef fish: Taphonomy and evolutionary implications. Palaios,3:315–326.LIPPS, J. H., AND S. J. CULVER, 2002. The trophic role of marine microorganisms through geologic time. In M. Kowalewskiand P. H. Kelley (eds.), The Fossil Record of Predation, Paleontological Society Special Papers, 8 (this volume).LOCKLEY, M. G., AND J. MADSEN, JR. 1992. Permian vertebrate trackways from the Cedar Mesa Sandstone ofEastern Utah: Evidence for predator-prey interaction. Ichnos, 2:147–153.LYMAN, R. L. 1994. Vertebrate Taphonomy. Cambridge University Press, Cambridge, 524 p.MARCUS, L. F., M. CORTI, A. LOY, G. J. P. NAYLOR, AND D. E. SLICE (eds.). 1996. Advances in Morphometrics.NATO ASI Series A: Life Sciences, 284, Plenum, New York, 587 p.36

KOWALEWSKI—ANALYTICAL METHODSMARTILL, D. M. 1990. Predation on Kosmoceras by semionotid fish in the Middle Jurassic lower Oxford Clay ofEngland. Palaeontology, 33:739–742.MATSUKUMA, A. 1977. Notes on Genkaimurex varicosa (Kuroda 1953) (Prosobranchia: Neogastropoda). Venus,36:81–88.MATSUKUMA, A. 1978. Fossil boreholes made by shell-boring predators or commensals, Part I: Boreholes ofcapulid gastropods. Venus, 37:29–45.MAYHEW, D. F. 1977. Avian predators as accumulators of fossil mammal material. Boreas, 6:25–31.MCROBERTS, C. A. 2001. Triassic bivalves and the initial marine Mesozoic revolution; a role for predators? Geology,29:359–362.MERLE, D. 2000. Premiere étude taphonomique de la predation affectant de grands mollusques benthiques dansl’Eocene de Gan (Pyrenées-Atlantiques, France). Comptes Rendus de l’Academie des Sciences, Serie II.Sciences de la Terre et des Planetes, 330:217–220.MILLER, R. H., AND F. A. SUNDBERG. 1984. Boring Late Cambrian organisms. Lethaia, 17:185–190.MORDAN, P. B. 1977. Factors affecting the distribution and abundance of Aegopinella and Nesovitrea (Pulmonata: Zonitidae)at Monks Wood National Nature Reserve, Huntingdonshire. Biological Journal of the Linnean Society, 9:59–72.MORTON, B., AND K. CHAN. 1997. The first report of shell-boring predation by a representative of the Nassariidae(Gastropoda). Journal of Molluscan Studies, 63:476–478.MOY-THOMAS, J. A., AND R. S. MILES. 1977. Palaeozoic Fishes, 2 nd ed. Saunders, Philadelphia.NEBELSICK, J. H. 1999. Taphonomic legacy of predation on echinoids, p. 347–352. In M. D. Candia Carnevali andF. Bonasoro (eds.), Echinoderm Research 1998. Balkema, Rotterdam.NEBELSICK, J. H., AND S. KAMPFER. 1994. Taphonomy of Clypeaster humilis and Echinodiscus auritus from the Red Sea, p.803–808. In B. A. David, A. Guille, J. P. Féral, and M. Roux (eds.), Echinoderms through Time. Balkema, Rotterdam.NEBELSICK, J. H., AND M. KOWALEWSKI. 1999. Drilling predation on Recent clypeasteroid echinoids from the RedSea. Palaios, 14:127–144.NEDIN, C. 1999. Anomalocaris predation on nonmineralized and mineralized trilobites. Geology, 27:987–990.NEGUS, M. 1975. An analysis of boreholes drilled by Natica catena (Da Costa) in the valves of Donax vittatus (DaCosta). Proceedings of the Malacological Society of London, 41:353–356.NEUMANN, C. 2000. Evidence of predation on Cretaceous sea stars from northwest Germany. Lethaia, 33:65–70.NIXON, M. 1979. Hole boring in shells by Octopus vulgaris in the Mediterranean. Malacologia, 18:431–444.OJI, T., C. OGAYA, AND T. SATO. 2001. Shell fragments in fossil shell beds as a result of durophagous predation: Resultsfrom field observation and tumbling experiments. Geological Society of America Abstracts with Programs, 33:A307.Orr, V. 1962. The drilling habit of Capulus danieli (Crosse) (Mollusca Gastropoda). Veliger, 5:63–67.ØRSTAN, A. 1999. Drill holes in land snail shells from western Turkey. Schriften zur Malakozoologie aus demHaus der Natur, 13:31–36.PETERSON, C. H., AND R. BLACK. 1995. Drilling by buccinid gastropods of the genus Cominella in Australia.Veliger, 38:37–42.PICKERILL, R. K., AND D. BLISSETT. 1999. A predatory Rusophycus burrow from the Cambrian of southern NewBrunswick, eastern Canada. Atlantic Geology, 35:179–183.PONDER, W. F., AND J. D. TAYLOR. 1992. Predatory shell drilling by two species of Austroginella (Gastropoda;Marginellidae). Journal of Zoology, 228:317–328.PRATT, B. R. 1998. Probable predation on Upper Cambrian trilobites and its relevance for the extinction of softbodiedBurgess Shale-type animals. Lethaia, 31:73–88.PYKE, G. H. 1984. Optimal Foraging Theory: A critical review. Annual Reviews of Ecology and Systematics,15:523–575.REYMENT, R. A. 1963. Bohrlöcher bei Ostrakoden (sowie einige paläoethologische Bemerkungen). PaläontologischeZeitschrift, 37:283–291.REYMENT, R. A. 1966a. Preliminary observation on gastropod predation in the western Niger Delta.Palaeogeography, Palaeoclimatology, Palaeoecology, 2:81–102.REYMENT, R. A. 1966b. Studies on Nigerian Upper Cretaceous and Lower Tertiary Ostracoda, Part 3: Stratigraphical,paleoecological and biometrical conclusions. Stockholm Contributions in Geology, 14:1–151.37

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002REYMENT, R. A. 1967. Paleoethology and fossil drilling gastropods. Transactions of the Kansas Academy ofScience, 70:33–50.REYMENT, R. A. 1971. Introduction to Quantitative Paleoecology. Elsevier, Amsterdam, 226 p.RICHARDS, R. P., AND C. W. SHABICA. 1969. Cylindrical living burrows in Ordovician dalmanellid brachiopodbeds. Journal of Paleontology, 43:838–841.RICHTER, G., AND S. BASZIO. 2001. Traces of a limnic food web in the Eocene Lake Messel; a preliminary reportbased on fish coprolite analyses. Palaeogeography, Palaeoclimatology, Palaeoecology, 166:345–368.RINALDI, A. C. 1994. Frequency and distribution of Vitreolina philippi (De Rayneval and Ponzi, 1854)(Prosobranchia, Eulimidae) on two regular echinoid species found along the southern coast of Sardinia.Bolletino Malacologia, 30:29–32.ROBBA, E., AND F. OSTINELLI. 1975. Studi paleoecologici sul Pliocene Ligure, I: Testimoniaze di predazione suimolluschi Pliocenici di Albenga. Rivista Italiana Paleontologia Stratigraphia, 81:309–372.ROHR, D. M. 1976. Silurian predator borings in the brachiopod Dicaelosia from the Canadian Arctic. Journal ofPaleontology, 50:1175–1179.ROHR, D. M. 1991. Borings in the shell of an Ordovician (Whiterockian) gastropod. Journal of Paleontology,65:687–688.ROOPNARINE, P. D., AND A. BEUSSINK.1999. Extinction and naticid predation of the bivalve Chione von Mühlfeld inthe Late Neogene of Florida. Palaeontologia Electronica, 2(1), 24, 719 KB.ROUGHLEY, T. 1925. The perils of the oyster. Australian Museum Magazine, 2:277–284.ROY, K., D. J. MILLER, AND M. LABARBERA. 1994. Taphonomic bias in analyses of drilling predation: Effects ofgastropod drill holes on bivalve shell strength. Palaios, 9:413–421.SAVAZZI, E., AND R. A. REYMENT. 1989. Subaerial hunting behavior in Natica gualteriana (naticid gastropod).Palaeogeography, Palaeoclimatology, Palaeoecology, 74:355–364.SCHINDEL, D. E., G. E. VERMEIJ, AND E. ZIPSER. 1982. Frequencies of repaired shell fractures among the Pennsylvaniangastropods of North-Central Texas. Journal of Paleontology, 56:729–740.SCHOENER, T. W. 1979. Inferring properties of predation and other injury-producing agents from injury frequencies.Ecology, 60:1110–1115.SHEEHAN, P. M., AND P. LESPÉRANCE. 1978. Effects of predation on the population dynamics of a Devonianbrachiopod. Journal of Paleontology, 52:812–817.SIGNOR, P. W., III, AND C. E. BRETT. 1984. The mid-Paleozoic precursor to the Mesozoic marine revolution.Paleobiology, 10:229–245.SLITER, W. V. 1971. Predation on benthic foraminifers. Journal of Foraminiferal Research, 1:20–29.SLITER, W. V. 1975. Foraminiferal life and residue assemblages from Cretaceous slope deposits. Geological Societyof America Bulletin, 86:897–906.SMITH, S. A., C. W. THAYER, AND C. E. BRETT. 1985. Predation in the Paleozoic: Gastropod-like drillholes inDevonian brachiopods. Science, 230:1033–1035.SPENCER, W. K., AND C. W. WRIGHT. 1966. Asterozoans , p. U4–U107. In R. Moore (ed.), Treatise on Invertebrate Paleontology,Pt. U, Echinodermata 3, Vol. 1. Geological Society of America and University of Kansas Press, New York.SOHL, N. F. 1969. The fossil record of shell boring by snails. American Zoologist, 9:725–734.SOHN, E. G., AND S. CHATTERJEE, 1979. Freshwater ostracodes from Late Triassic coprolites in Central India.Journal of Paleontology, 53:578–586.STALLIBRASS, S. 1984. The distinction between the effects of small carnivores and humans on post-glacial faunalassemblages, p. 259–269. In C. Grigson and J. Clutton-Brock (eds.), Animals and Archaeology: 4. Husbandryin Europe. British Archaeological Reports International Series, 227.STALLIBRASS, S. 1990. Canid damage to animal bones: Two current lines of evidence, p. 151–165. In D. E. Robinson(ed.), Experimentation and Reconstruction in Environmental Archaeology. Oxbow Books, Oxford.STEWART, J. D., AND K. CARPENTER. 1990. Examples of vertebrate predation on cephalopods in the Late Cretaceousof the Western Interior, p. 203–207. In A. J. Boucot (ed.), Evolutionary Paleobiology of Behavior andCoevolution. Elsevier, Amsterdam.STEWART, J. D., AND K. CARPENTER. 1999. Examples of vertebrate predation on cephalopods in the Late Cretaceousof the Western Interior. Bulletin of the Southern California Paleontological Society, 31:66–73.38

KOWALEWSKI—ANALYTICAL METHODSSTEWART, K. M., L. LEBLANC, D. P. MATTHIESEN, AND J. WEST. 1999. Microfaunal remains from a modern EastAfrican raptor roost; patterning and implications for fossil bone scatters. Paleobiology, 25:483–503.STUCKENRATH, R. 1977. Radiocarbon: Some notes from Merlin’s diary. Annals of the New York Academy ofSciences, 288:181–188.TAYLOR, J. D., N. J. MORRIS, AND C. N. TAYLOR. 1980. Food specialization and the evolution of predatory prosobranchgastropods. Palaeontology, 23:375–409.TAYLOR, J. D., R. J. CLEEVELY, AND N. J. MORRIS. 1983. Predatory gastropods and their activities in the BlackdownGreensand (Albian) of England. Palaeontology, 26:521–553.THOMAS, R. D. K. 1976. Gastropod predation on sympatric Neogene species of Glycymeris (Bivalvia) from theeastern United States. Journal of Paleontology, 50:488–499.TODD, L. C., AND D. J. RAPSON. 1988. Long bone fragmentation and interpretation of faunal assemblages; approachesto comparative analysis. Journal of Archaeological Science, 15:307–325.TSUJITA, C. J., AND G. E. G. WESTERMANN. 2001. Were limpets or mosasaurs responsible for the perforations in theammonite Placenticeras? Palaeogeography, Palaeoclimatology, Palaeoecology, 169:245–270.VAN VALKENBURGH, B., AND F. HERTEL. 1993. Tough times at La Brea: Tooth breakage in large carnivores of theLate Pleistocene. Science, 261:456–459.VAN VALKENBURGH, B., AND I. JENKINS. 2002. Evolutionary patterns in the history of Permo-Triassic and Cenozoicsynapsid predators. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record of Predation. PaleontologicalSociety Special Papers, 8 (this volume).VERMEIJ, G. J. 1977. The Mesozoic marine revolution: Evidence from snails, predators, and grazers. Paleobiology,3:245–258.VERMEIJ, G. J. 1980. Drilling predation of bivalves in Guam: Some paleoecological implications. Malacologia,19:329–334.VERMEIJ, G. J. 1982. Gastropod shell form, breakage, and repair in relation to predation by the crab Calappa.Malacologia, 23:1–12.VERMEIJ, G. J. 1983. Traces and trends of predation, with special reference to bivalved animals. Palaeontology,26:455–465.VERMEIJ, G. J. 1987. Evolution and Escalation. Princeton University Press, Princeton, New Jersey, 527 p.VERMEIJ, G. J. 2002. Evolution in the consumer age: Predators and the history of life. In M. Kowalewski and P. H.Kelley (eds.), The Fossil Record of Predation. Paleontological Society Special Papers, 8 (this volume).VERMEIJ, G. J., AND E. C. DUDLEY. 1982. Shell repair and drilling in some gastropods from the Ripley Formation(Upper Cretaceous) of the Southeastern U.S.A. Cretaceous Research, 3:397–403.VERMEIJ, G. J., AND D. R. LINDBERG. 2000. Delayed herbivory and the assembly of marine ecosystems. Paleobiology,26:419–430.VERMEIJ, G. J., E. ZIPSER, AND E. C. DUDLEY. 1980. Predation in time and space: Peeling and drilling in terebridgastropods. Paleobiology, 6:352–364.VERMEIJ, G. J., D. E. SCHINDEL, AND E. ZIPSER. 1981, Predation through geological time: Evidence from gastropodshell repair. Science, 214:1024–1026.VILLA, P., AND E. MAHIEU. 1991. Breakage patterns of human long bones. Journal of Human Evolution, 21:27–48.WÄCHTLER, V. W. 1927. Zür biologie der Raublungenschnecke Poiretia (Glandina). Algira Brug. ZoologischenAnzeiger, 72:191–197.WALKER, S. E. 1989. Hermit crabs as taphonomic agents. Palaios, 4:439–452.WALKER, S. E., AND C. E. BRETT. 2002. Post-Paleozoic patterns in marine predation: Was there a Mesozoic andCenozoic marine predatory revolution? In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record ofPredation. Paleontological Society Special Papers, 8 (this volume).WALKER, S E., AND S. B. YAMADA. 1993. Implications for the gastropod fossil record of mistaken crab predationon empty mollusc shells. Palaeontology, 36:735–741.WALKER, S. E., AND J. R. VOIGT. 1994. Paleoecologic and taphonomic potential of deepsea gastropods. Palaios, 9:48–59.WARÉN, A. 1980. Revision of the genera Thyca, Stilifer, Scalenostoma, Mucronalia and Echineulima (Mollusca,Prosobranchia, Eulimidae). Zoologica Scripta, 9:187–210.39

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002WARÉN, A. 1981. Eulimid gastropods parasitic on echinoderms in the New Zealand region. New Zealand Journalof Zoology, 8:313–324.WARÉN, A., D. R. NORRIS, AND J. T. TEMPELADO. 1994. Description of four new eulimid gastropods parasitic onirregular sea urchins. Veliger, 37:141–154.WARÉN, A., AND M. R. CROSSLAND. 1991. Revision of Hypermastus Pilsbry, 1899 and Turveria Berry, 1956 (Gastropoda:Prosobranchia: Eulimidae), two genera parasitic on sand dollars. Records of Australian Museum, 43:85–112.WILTSE, W. I. 1980. Predation by juvenile Polinices duplicatus (Say) on Gemma gemma (Totton). Journal ofExperimental Marine Biology and Ecology, 42:187–199.WILSON, J. B. 1967. Palaeoecological studies on shell-beds and associated sediments in the Solway Firth. ScottishJournal of Geology, 3:329–371.WILSON, M. V. H. 1987. Predation as a source of fish fossils in Eocene lake sediments. Palaios, 2:497–504.WILSON, M. A., AND T. J. PALMER. 2001. Domiciles, not predatory borings: A simpler explanation of the holes inOrdovician shells analyzed by Kaplan and Baumiller, 2000. Palaios, 16:524–525.WOELKE, C. E. 1957. The flatworm Pseudostylochus ostreophagus Hyman, a predator of oysters. Proceedings ofthe National Shellfisheries Association, 47:62–67.WOOD, R. In press. Predation in ancient reef-builders. In P. H. Kelley, M. Kowalewski, and T. A. Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topics in Geobiology Series, Plenum Press/Kluwer, New York.WODINSKY, J. 1969. Penetration of the shell and feeding on gastropods by Octopus. American Zoologist, 9:997–1010.YOCHELSON, E. L., D. DOCKERY, AND H. WOLF. 1983. Predation on sub-Holocene scaphopod mollusks from southernLouisiana. United States Geological Survey Professional Paper, 1282:1–13.YONGE, C. M. 1964. Rock borers. Sea Frontiers, 10:106–116.YOUNG, D. K. 1969. Okadaia elegans, a tube-boring nudibranch mollusc from the central and west Pacific.American Zoologist, 9:903–907.ZAR, J. H. 1999. Biostatistical analysis, 4 th ed. Prentice Hall, Upper Saddle River, 663 p.ZIEGELMEIER, E. 1954. Beobachtungen über den Nahrungserwerb bei der Naticide Lunatia nitida Donovan(Gastropoda Prosobranchia). Helgoländer wissentschaftliche Meeresuntersuchungen, 5:1–33.ZILCH, A. 1959. Euthyneura, p. 1–701. In O. H. Schindewolf (ed.), Handbuch der Paläozoologie, 6. Berlin.ZOTNIK, M. 2001. Size-related changes in predatory behaviour of naticid gastropods from the middle MioceneKorytnica Clays, Poland. Acta Palaeontologica Polonica, 46:87–97.ZUSCHIN, M., AND R. J. STANTON, JR. 2001. Experimental measurement of shell strength and its taphonomicinterpretation. Palaios, 16:161–170.40

KOWALEWSKI—ANALYTICAL METHODSAPPENDIX 1—A SAS/IML and SAS/STAT programto test for non-random distribution of traces acrosstaxa. The program generates four outputs: (1)frequency of traces for each taxon and eachiteration; (2) Null proportion of taxa expected tocontain traces sorted by iteration; (3) Null highestper-taxon frequency of traces sorted by iteration;and (4) Null highest per-taxon frequency of tracesby iteration and taxon. Written by M. Kowalewski.________________________________________________%let prob=0.011; * - assemblage-level drilling frequency (AF);%let times=999; * - number of iterations;PROC IML;X={292,280,280,280,220,200,200,200,140,140,140,120,120,100,100,100,100,100,100,100,80,80,80,80,80,80,60,60,60,60,60,60,60,60,60,60,60};* - enter the specimen numbers for all taxa here;START drill(X,fin);Z=X; a=1-&prob; b=&prob; k=nrow(Z);DO i=1 to k;new=shape(i,Z[i],1);size=shape(Z[i],Z[i],1);newsize=new||size;new2=new2//newsize;END;out1=new2; k2=nrow(out1);DO j=1 to k2;c=rantbl(0,a,b); c1=c-1;c2=c2//c1;END;out2=c2; out3=out1||out2;fin=out3;FINISH drill;START simul(X,out);do i=1 to ×run drill(X,fin);fin2=fin||shape(i,nrow(fin),1);fin3=fin3//fin2;end;out=fin3;FINISH simul;RUN simul(X,out);CREATE new from out;APPEND from out;CLOSE new;RUN;DATA output;set new; taxon=col1; nspec=col2; drill=col3; iter=col4;PROC sort; by iter taxon nspec;PROC univariate noprint;by iter taxon nspec; var drill;output out=final n=n mean=mean sum=sum;title1*Simulation Output for &times iterat & drilling prob &prob*;title2’Total output’;PROC print;var iter taxon nspec mean sum;DATA new2;set final;if sum>0 then sum=1;PROC univariate noprint;by iter; var sum;output out=final2 mean=mean;title2’Proportion of taxa drilled by iteration’;PROC print;PROC univariate data=final noprint;by iter; var mean;output out=final3 max=max;title2’The highest per-taxon drilling frequency by iteration’;PROC print;PROC sort data=final; by iter nspec;PROC univariate noprint;by iter nspec; var mean;output out=final4 max=max;DATA new3;set final4; keep iter nspec max;title2’The highest drilling frequency by taxon sample size byiteration’;PROC print;QUIT;_______________________________________________________APPENDIX 2—A SAS/IML and SAS/STAT programfor a two-sample bootstrap test for difference inmeans for circular (azimuth) data. The algorithmfor the Watson-Williams two-sample test is basedon equations from Zar (1999, p. 626, Example27.8). F statistic is not corrected (i.e., K=1; seeZar, 1999, Equation 27.11). The correction is notneeded here because the probability densityfunction for F is established empirically throughbootstrap simulation. Bootstrap probability [p] iscalculated as follows: p=s+1/i+1, where s is thenumber of bootstrap values larger than or equal toactual F and i is the number of iterations (thisequation includes the actual samples in computingp; see Manly, 1995). Written by M. Kowalewski.Note: This code was used recently by Dietl andAlexander (2000) to analyze changes in siteselectivity of drilling on Cenozoic gastropods._______________________________________%let TIMES=999; *—number of times to randomize;%let DATA1=’ang1.dat’; *—file containing first variable;%let DATA2=’ang2.dat’; *—file containing second variable;Title1’2-sample bootstrap for circular data (Watson-Williamstest)’;Title2’written by Michal Kowalewski, October 11, 1999';DATA data1; infile &DATA1; input var1;DATA data2; infile &DATA2; input var2; RUN;PROC IML; %let pi=3.1415926535;USE data1; READ all var{var1} into X1;USE data2; READ all var{var2} into X2;START RANVEC(in,v_out); k=nrow(in); v_index=in;DO i=1 to k; rand=floor((k-i+1)*ranuni(0) + 1);v_ran=v_ran||v_index[rand]; v_index=remove(v_index,rand); END;v_out=v_ran; FINISH RANVEC;START MIXUP(X,times,template);n=nrow(X); template=t(1:n)*j(1,times,1);DO i=1 to times; run ranvec(template[,i],out);template[,i]=t(out); END; DO i=1 to n; runranvec(t(template[i,]),out); template[i,]=out; END; FINISH MIXUP;START WATSON(D1,D2,F);Y1=D1/(180/&pi); Y2=D2/(180/&pi); Y=Y1//Y2;C1=sum(COS(Y1))/nrow(Y1); S1=sum(SIN(Y1))/nrow(Y1);R1=sqrt(C1**2+S1**2)*nrow(Y1); a1=arcos(C1/(R1/nrow(Y1)))*(180/&pi);C2=sum(COS(Y2))/nrow(Y2); S2=sum(SIN(Y2))/nrow(Y2);R2=sqrt(C2**2+S2**2)*nrow(Y2); a2=arcos(C2/(R2/41

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002nrow(Y2)))*(180/&pi);C=sum(COS(Y))/nrow(Y); S=sum(SIN(Y))/nrow(Y);R=sqrt(C**2+S**2)*nrow(Y); mean_a=arcos(C/(R/nrow(Y)))*(180/&pi);Fstat=((nrow(Y)-2)*(R1+R2-R))/(nrow(Y)-R1-R2);F=a1||a2||mean_a||Fstat; FINISH WATSON;START BOOT(X1,X2,times,dist);RUN watson(X1,X2,aF); X=X1//X2; k=nrow(X1);j=nrow(X2);RUN mixup(X,times,template);Do i=1 to times; Y1=X[template[1:k,i]];Y2=X[template[(k+1):(k+j),i]];RUN watson(Y1,Y2,F); rF=rF//(i||F); END; rand=rF;act=shape(aF,nrow(rand),ncol(aF)); dist=rand||act; FINISH BOOT;RUN BOOT(X1,X2,&times,dist);CREATE OUT from DIST [colname={‘i’‘r1’‘r2’‘mean-r’‘rF’‘a1’ ‘a2’ ‘mean_a’ ‘aF’}];APPEND from DIST; CLOSE OUT;DATA report; set OUT; if i=1; keep a1 a2 mean_a aF;DATA count; set OUT; if rF>=aF then p=1; else p=0;PROC univariate noprint; var p; output out=last sum=s N=n;DATA prob; set last; n=n+1; p=(s+1)/n; keep p n;DATA final; merge prob report;PROC print data=final noobs split=’*’;label a1=’mean angle for the first sample’; label a2=’meanangle for the second sample’; label mean_a=’mean angle for pooleddata’; label aF=’Waston-Williams Stat. without K-factor correction’;label n=’number of random samples (# iterations + 1)’; labelp=’probabil. that 2 samples have the same mean angle’; RUN; QUIT42

CHIN—ANALYSES OF COPROLITES PRODUCED BY VERTEBRATESANALYSES OF COPROLITESPRODUCED BY CARNIVOROUS VERTEBRATESKAREN CHINMuseum of Natural History/Department of Geological Sciences, University of Colorado at Boulder,UCB 265, Boulder, Colorado 80309 USAABSTRACT—The fossil record contains far more coprolites produced by carnivorous animals than by herbivores.This inequity reflects the fact that feces generated by diets of flesh and bone (and other skeletal materials)contain chemical constituents that may precipitate out under certain conditions as permineralizing phosphates.Thus, although coprolites are usually less common than fossil bones, they provide a significant source of informationabout ancient patterns of predation. The identity of a coprolite producer often remains unresolved, but fossilfeces can provide new perspectives on prey selection patterns, digestive efficiency, and the occurrence of previouslyunknown taxa in a paleoecosystem. Dietary residues are often embedded in the interior of coprolites, but muchcan be learned from analyses of intact specimens. When ample material is available, however, destructive analysessuch as petrography or coprolite dissolution may be used to extract additional paleobiological information.INTRODUCTIONWHEN PREDATOR-PREY interactions cannotbe observed directly, fecal analysis provides the nextbest source of information about carnivore feedingactivity because refractory dietary residues oftenreveal what an animal has eaten. This approach isvery effective in studying extant wildlife, and it canalso be used to glean clues about ancient trophicinteractions. Yet, although fossilized carnivore fecesare often present in fossiliferous sediments, theirantiquity makes analysis more difficult. Diageneticalteration of specimens obscures the originaldigestive residues, and it is often impossible toascertain which animal produced the coprolite.Fossil feces also show significantly more variationin morphology and composition than skeletal fossils:animal diets are often highly variable, and soft fecalmaterial can assume a range of shapes—especiallyafter post-depositional deformation. In spite of suchcomplexities, coprolites provide importantinformation about ancient predator-prey interactionsthat is not available from body fossils. Coproliteanalyses, however, require different approachesfrom those used to study most skeletal fossils.Furthermore, different types of coprolites mayprovide different types of information.Diet and depositional environment largelydetermine which animal feces may be fossilized andthe quality of preservation of a lithified specimen.Significant concentrations of calcium and phosphorusin bone and flesh often favor the preservation ofcarnivore feces by providing autochthonous sourcesof constituents that can form permineralizing calciumphosphates (Bradley, 1946). Thus, althoughherbivores outnumber carnivores in terrestrialecosystems, carnivore feces are more likely to bepreserved. Preservation is also facilitated by rapidburial, so coprolites from aquatic taxa usually faroutnumber those from terrestrial animals. Thesetaphonomic biases explain why coprolites arerelatively rare in most terrestrial deposits, while fishcoprolites can be quite common.WHAT CARNIVORE COPROLITESTELL US ABOUT PREDATOR-PREY INTERACTIONSThe identity of the animal that produced acoprolite is very difficult, if not impossible, topinpoint because our knowledge of ancient faunasis incomplete and because fecal material is so43

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002variable. Although we know that spiral coprolites(or intestinal casts; see Williams, 1972) wereproduced by one of the groups of fish with spiralintestinal valves (e.g., sharks, lungfish, or gars;Gilmore, 1992), morphology usually provides littleinformation about the animal of origin becausemany animal droppings produced by different taxaare quite similar. Coprolite contents, composition,size, and stratigraphic placement can, however,constrain the number of likely perpetrators.Carnivore coprolites are usually easy todifferentiate from herbivore coprolites becausethey are typically phosphatic and often containskeletal inclusions. Coprolite size is also veryinformative, since fecal volumes generally scalewith animal size. While fecal amounts are variable,it’s clear that small animals cannot produce largeindividual fecal deposits. Field guides to modernscat provide analogs that can be used to roughlyapproximate the size range of animals that producefeces of a given mass. Thus, even when the taxa offecal producers are unknown, carnivore coprolitesprovide evidence that predators of an approximatesize range frequented a given habitat. This indicatesa trophic niche that may be further defined bydietary residues that reveal prey selection patterns.In many cases, inclusions within a coproliteprovide more information about prey animals thanabout the coprolite producer itself. The integrityof included digestive residues depends on theircomposition and the extent of their exposure todigestive and diagenetic processes. Somecoprolites contain no recognizable inclusions, butrefractory skeletal constituents have been found innumerous carnivore coprolites (e.g., Hantzschel etal., 1968). When dietary residues are incompletelydigested, the morphology of elements such asmollusk shells, ganoid scales, and small bones mayallow identification of prey. Specific taxa ofmollusks (e.g., Speden, 1969; Stewart andCarpenter, 1990), crustaceans (Bishop, 1977; Sohnand Chatterjee, 1979), fish (e.g., Zangerl andRichardson, 1963; Waldman, 1970; Coy, 1995),reptiles (e.g., Parris and Holman, 1978), andmammals (e.g., Martin, 1981; Meng et al., 1998)have been recognized in coprolites. Furthermore,fragments of larger bones may be ascribed tohigher-level taxonomic groups on the basis ofhistological analysis (e.g., Chin et al., 1998).Coprolites may also show signs of ingested softtissues as organic residues or in the form of threedimensionalimpressions. Some exceptionalspecimens have revealed evidence of feathers(Wetmore, 1943), fur (Meng and Wyss, 1997),insect exoskeletons (Northwood, 1997), andmuscle tissues (Chin et al., 1999). Such remarkablepreservation requires depositional conditions thatminimize diagenetic recrystallization.These studies show that coprolites may containdietary residues that provide concrete evidence ofancient carnivory. It is clear, though, that thechallenges of coprolite analysis stem from thedifficulties involved in determining the animal oforigin and in identifying residual dietarycomponents. This analytical complexity reflects thefact that coprolites are relatively anonymouspackages that represent varying diets, digestiveprocesses, and diagenetic alteration. Such markedvariability necessitates that care be exercised indrawing conclusions from a limited number ofsamples. Even so, coprolites provide cumulativeclues that help flesh out our understanding ofpatterns of predation in ancient environments byidentifying prey species within a givenpaleoecosystem, and by indicating general size and/or age classes of prey animals (e.g., Zidek, 1980;Martin, 1981). The composition and integrity ofthese inclusions may also provide information aboutthe diet and digestive processes of the predator itself.COPROLITE ANALYSISDocumentation.—Both destructive and nondestructivetechniques may be used to analyzecoprolites, and the choice of analytical methoddepends on the questions addressed by the researchproject. Regardless of experimental approach, eachstudy of coprolites must include carefuldocumentation of provenance and of the physicalcharacteristics of each specimen.The collection of coprolites is similar to thecollection of vertebrate fossils where documentation44

CHIN—ANALYSES OF COPROLITES PRODUCED BY VERTEBRATESof locality and stratigraphic information is of criticalimportance. In many cases, coprolites have erodedout of their encasing sediments and are collected asfloat. Although such specimens contribute notablebaseline information about a given formation, theinformative value of coprolites is greatly enhancedwhen they are collected in situ and can be correctlyplaced within a detailed stratigraphic column.Mapping specimens in place will also indicatecoprolite orientation and density. Such taphonomicinformation contributes important information aboutthe environment of deposition.Photographic documentation of specimens isas useful for coprolite analysis as it is for researchon other fossils. Unfortunately, most coprolites arenot as strikingly photogenic as many skeletalfossils! Nevertheless, images of seeminglyamorphous coprolitic masses are quite usefulbecause they document the range of coprolite sizeand morphology, and help create search images forpaleontologists likely to encounter fossilized fecesin the field (Fig. 1). Such records become evenmore important when destructive analyses alter theoriginal form of a specimen. Photos should alwaysinclude a scale.If a coprolite specimen is very important orhas an unusually distinctive morphology, standardpaleontological molding and casting methods (seeGoodwin and Chaney, 1994) can be used toreplicate the external form. Care should be taken,however, if this technique is applied to fragilespecimens; some coprolites are composed of softmaterials that can be easily scratched or gouged(possibly obscuring paleobiologically informativeimpressions), or may absorb compounds appliedas separators or mold release agents. This techniqueshould not be used on highly fractured specimens.Non-destructive Analyses.—Intact coprolitespecimens can be characterized by morphology, size,and surface features. Although coprolites from manydifferent taxa can have similar traits, documentationof the physical characteristics of coprolite specimensis important because recurring features may revealdistinct morphological categories. Such forms maybe designated as coprolite morphotypes. One earlyclassification scheme differentiated spiralcoprolites into “heterpolar” or “amphipolar” types,depending on the spacing of coils along the longaxis of the specimen (Neumayer, 1904). Thissystem has been applied to other spiral coprolites,though there is debate as to whether thesemorphotypes have any taxonomic significance (e.g.Price, 1927; Zidek, 1980).Linear dimensions of coprolites (e.g., diameterand length) provide rough approximations of fecalsize, but volumetric measurements give much moreinformative assessments. Volume can be measuredin several ways. The volume of small, densespecimens can be determined by measuring waterdisplaced by submerged samples; porousspecimens should be allowed to absorb waterbefore displacement is measured. This approachcan also be applied to large, fractured specimensby using water displacement to calculate thedensity of small fragments; volume can then bedetermined by extrapolation after weighing theentire specimen (e.g., Chin et al., 1998). In a fewcases a coprolite may be so large and fractured thatit remains in the plaster jacket in which it wascollected (e.g., Fig. 1a), and cannot be accuratelyweighed. The volume of such specimens can beapproximated by visualizing the mass as beingcomposed of one or more geometric shapes whosevolumes can be calculated.When a large number of coprolites from agiven locality represents an unbiased sample,recurring size classes and other physicalcharacteristics may indicate specimens producedby a small number of taxa and/or age groups. Suchgroupings may be subtle, however. Edwards andYatkola (1974) analyzed 106 coprolites from theWhite River Formation and found no distinct sizeclasses within a continuum of coprolite diametersranging from 15 to 36 mm. But when the data wasre-analyzed using the mean diameter of in situcoprolites that occurred in small “clusters”(suggesting individual fecal deposits), they foundthree distinct size groups that may reflect the sizesof the carnivores that produced them. Correlatingcoprolite size with other physical attributes willrefine interpretations, though some features (suchas color and degree of flattening) probably provide45

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 1—Coprolites can have many morphologies, from easily recognizable forms to nondescriptmasses. (a) Very large Cretaceous coprolite from Alberta (Royal Tyrrell Museum, TMP 98.102.7). Thisspecimen was so fractured from weathering that it was removed in a plaster jacket. (b) Probable fishcoprolite from the Pennsylvanian black shale of Indiana (Field Museum of Natural History; FMNHPF2210; see Zangerl and Richardson, 1963). Photograph shows lateral view of sliced specimen; thespecimen originally appeared as a lump in the sediments, but was sliced through the center to revealthe coprolitic material sandwiched between the black shale. (c) Small coprolite with more typical,‘sausage’ fecal shape (Royal Tyrrell Museum, TMP 80.16.1098).little taxonomic information (Sawyer, 1981).Non-destructive analyses of coprolites shouldalso include careful examination of specimensurfaces for distinctive impressions or inclusions.These surfaces can be scanned with a stereomicroscope; if warranted, higher magnifications canbe obtained using scanning electron microscopy.Scrutiny of the outer surface of a coprolite mayreveal dietary inclusions protruding from thecoprolitic ground mass. Mineralized skeletalelements are commonly observed in carnivorecoprolites, and inclusions on outer surfaces may beaccentuated by weathering processes. Coproliteexteriors may also show three-dimensionalimpressions of the vegetation or detritus on whichthe feces were deposited. Such impressions are ofinterest, but should be distinguished from those thatrepresent dietary components.46

CHIN—ANALYSES OF COPROLITES PRODUCED BY VERTEBRATESBroken coprolites allow scrutiny of fracturedinternal surfaces. In most cases, recognizablefeatures evident in the interior of a specimen canbe confidently attributed to diet. This includesthree-dimensional impressions that may indicateundigested soft tissues.Destructive Analyses.—As a general rule,damage of fossil specimens should be studiouslyavoided. At the present state of our technology,however, destructive analyses appear to providesome of the most effective means to extractpaleobiological information from coprolites.Techniques such as petrographic analysis or aciddissolution may reveal dietary components thataren’t evident on coprolite surfaces. Because suchanalyses destroy the morphological integrity of aspecimen, several factors should be consideredbefore a sample is altered. If numerous comparablecoprolites are available or if a specimen is verylarge and/or fragmented, the information obtainedfrom destructive tests is likely to compensate forthe loss of some coprolitic material. But thedecision of whether to perform destructive analysesbecomes more difficult if the tests will damage aunique specimen. When destructive analyses areplanned, they should be preceded by carefulmeasurements, photo-documentation, and scrutinyof accessible surfaces (see above).Thin sections of coprolites provideexceptionally informative views of specimencontents because they permit analysis withcompound microscopes. Such analyses may revealdietary inclusions with considerable histologicaldetail. They also shed light on patterns of diageneticmineralization. The jumbled nature of fecalcontents makes thin section sampling ratherunpredictable, however, because identification ofdietary components depends on fortuitous slicesthrough recognizable structures. Fortunately, somefeatures diagnostic of certain taxonomic groups(such as patterns of bone vascularization) may beevident on small fragments.Thin sections can be made from relatively smallpieces of coprolite, and careful scrutiny of coprolitefragments or intact specimens will help identifyoptimal sampling sites. When possible, thin sectionscan be taken from the end of a specimen in order topreserve more of the original morphology. Althoughtechniques for preparing coprolite thin sections aresimilar to those for preparing standard petrographicsections of rock, more efforts are made to minimizedamage to and loss of coprolitic material (seeWilson, 1994 for a useful discussion of methods forpreparing fossil thin sections).Saws with diamond-embedded blades are usedto reduce large samples to sizes that can be affixedto glass slides. A diamond saw is also necessary toshave off the thin sample slices that are mountedon slides (the sample can be sliced thin before orafter it is mounted on the slide). The use of aprecision saw with a thin diamond wafering bladewill facilitate more accurate cuts and help reduceloss of coprolite material during the cuttingoperation. In a few cases, indurate coprolites canbe cut with a precision saw without embedding,but fragile or fractured specimens should beembedded in or impregnated with an epoxy orpolyester resin before being cut.The cut surface of a specimen must be groundsmooth before it is affixed to a microscope slidewith a strong epoxy bonding agent. Standardpetrographic slides are 27 × 46 mm, but specimenscan also be mounted on larger slides (or glass plates)as well. Grinding/polishing machines or lappingwheels are used to grind and polish the sample toan appropriate thickness (around 30–40 µm,depending on the nature of the sample). Slides canthen be cover-slipped for examination with a lightmicroscope, or finely polished for chemical analyseswith a microprobe or scanning electron microscope.In a few studies, coprolites have beenmechanically disaggregated in order to releasedietary residues from the ground mass. Thistechnique will be most effective when the prey havebeen poorly digested, and it may reveal thepresence of small prey taxa that are not otherwiserepresented in a faunal assemblage. Aciddissolution may also facilitate chemical andmorphological studies of amorphous organicresidues (e.g., Hollocher et al., 2001). Some of thetechniques used in the acid preparation ofvertebrate fossils may be modified for use on47

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002coprolites (see Rutzky et al., 1994). In suchprocedures, the exposed portions of skeletalelements that are encased in well-lithifiedsediments are carefully coated with a hardener forprotection before the specimens are immersed inweak acid solutions (usually phosphate-bufferedacetic or formic acids). As this process is repeated,embedded skeletal elements are gradually releasedfrom the sediments.Because the ground mass of most coprolites isphosphatic, acid solutions prepared for dissolvingthis material should not have a phosphate buffer(unless it is very weak). The acid dissolutionprocedure will be easier when the ground mass of acoprolite is more susceptible to acid than the dietaryinclusions. The operation will be more challenging,however, when both included skeletal elements andthe coprolitic ground mass show similarsusceptibility to the acid solution. Because thephosphatic ground masses of different coprolites canhave widely varying compositions, it is clear thatexperimentation will be necessary to identify themost effective methods for releasing inclusions froma given type of coprolite. Some authors (Sohn andChatterjee, 1979) have used formic acid to releaseostracods from coprolites. Others (Burmeister et al.,1999) report that using ultrasonication with weakacetic acid is effective in freeing inclusions fromcoprolites that have a significant carbonatecomponent. Mechanical disaggregation alone mayalso be used on coprolites that have a softer groundmass (e.g., Parris and Holman, 1978).CONCLUSIONSCoprolite analysis is quite different from thestudy of fossil skeletal elements. Becausemorphology is usually not diagnostic, the chemicaland physical composition of a coprolite assumesgreater importance and may provide as much (ormore) paleobiological information than size andshape. The ambiguity of coprolite morphology alsomakes it difficult to unequivocally associate acoprolite with the animal that produced it.Despite these challenges, some carnivorecoprolites provide unique perspectives on ancientpredator-prey activities. Although they may notprovide complete information about the carnivoroushabits of specific animals, coprolites can supplyimportant fossil evidence that reveals prey selectionpatterns, digestive efficiency, and the occurrence ofsmaller fauna in a given paleoenvironment.REFERENCESBISHOP, G. A. 1977. Pierre feces: a scatological study of the Dakoticancer assemblage, Pierre Shale (UpperCretaceous) of South Dakota. Journal of Sedimentary Petrology, 47(1):129–136.BRADLEY, W. H. 1946. Coprolites from the Bridger Formation of Wyoming: their composition and microorganisms.American Journal of Science, 244:215–239.BURMEISTER, K. C., J. J. FLYNN, J. M. PARRISH, AND A. R. WYSS. 1999. New fossil vertebrates from the northernMorondava Basin, Madagascar, and the recovery of microvertebrates from coprolites. PaleoBios, 1999California Paleontology Conference Abstracts, 19(supplement to Number 1):3.CHIN, K., D. A. EBERTH, AND W. J. SLOBODA. 1999. Exceptional soft-tissue preservation in a theropod coprolitefrom the Upper Cretaceous Dinosaur Park Formation of Alberta. Journal of Vertebrate Paleontology, Abstractsof Papers, 19(Supplement to 3):37–38.CHIN, K., T. T. TOKARYK, G. M. ERICKSON, AND L. C. CALK. 1998. A king-sized theropod coprolite. Nature, 393: 680–682.COY, C. E. 1995. The first record of spiral coprolites from the Dinosaur Park Formation (Judith River Group,Upper Cretaceous) southern Alberta, Canada. Journal of Paleontology, 69(6):1191–1194.EDWARDS, P. D., AND D. YATKOLA. 1974. Coprolites of White River (Oligocene) carnivorous mammals: origin andpaleoecological significance. Contributions to Geology, University of Wyoming, 13:67–73.GILMORE, B. G. 1992. Scroll coprolites from the Silurian of Ireland and the feeding of early vertebrates.Palaeontology, 35(2):319–333.48

CHIN—ANALYSES OF COPROLITES PRODUCED BY VERTEBRATESGOODWIN, M. B., AND D. S. CHANEY. 1994. Molding, casting, and painting, p. 235–284. In P. Leiggi and P. May(eds.), Vertebrate Paleontological Techniques, Volume One. Cambridge University Press, Cambridge.HANTZSCHEL, W., F. EL-BAZ, AND G. C. AMSTUTZ. 1968. Coprolites: an annotated bibliography. Geological Societyof America Memoir, 108:1–132.HOLLOCHER, T. C., K. CHIN, K. T. HOLLOCHER, AND M. A. KRUGE. 2001. Bacterial residues in herbivorous dinosaurcoprolites and a role for these bacteria in its mineralization. Palaios, 16:547–565.MARTIN, J. E. 1981. Contents of coprolites from Hemphillian sediments in northern Oregon, and their significancein paleoecological interpretations. Proceedings of the South Dakota Academy of Science, 60:105–115.MENG, J., AND A. R. WYSS. 1997. Multituberculate and other mammal hair recovered from Palaeogene excreta.Nature, 385: 712–714.MENG, J., R. ZHAI, AND A. R. WYSS. 1998. The late Paleocene Bayan Ulan fauna of Inner Mongolia, China, p.148–185. In K. C. Beard and M. R. Dawson (eds.), Dawn of the Age of Mammals in Asia. Bulletin of theCarnegie Museum of Natural History, No. 34, Pittsburgh.NEUMAYER, L. 1904. Die Koprolithen des Perms von Texas. Palaeontographica, 51:121–128.NORTHWOOD, C. 1997. Coprolites from the early Triassic Arcadia Formation, Queensland, Australia. Journal ofVertebrate Paleontology, Abstracts of Papers, 17(Supplement to 3):67A.PARRIS, D. C., AND J. A. HOLMAN. 1978. An Oligocene snake from a coprolite. Herpetologica, 34(3):258–264.PRICE, P. 1927. The coprolite limestone horizon of the Conemaugh Series in and around Morgantown, WestVirginia. Annals of the Carnegie Museum, 17:211–254RUTZKY, I. S., W. B. ELVERS, J. G. MAISEY, AND A. W. A. KELLNER. 1994. Chemical preparation techniques, p. 155–186. In P. Leiggi and P. May (eds.), Vertebrate Paleontological Techniques, Volume One. Cambridge UniversityPress, Cambridge.SAWYER, G. T. 1981. A study of crocodilian coprolites from Wannagan Creek Quarry (Paleocene—North Dakota).Scientific Publications of the Science Museum of Minnesota, New Series, 5(2):1–29.SOHN, E. G., AND S. CHATTERJEE. 1979. Freshwater ostracodes from Late Triassic coprolites in Central India.Journal of Paleontology, 53(3):578–586.SPEDEN, I. G. 1969. Predation on New Zealand Cretaceous species of Inoceramus (Bivalvia). New Zealand Journalof Geology and Geophysics, 14(1):56–70.STEWART, J. D., AND K. CARPENTER. 1990. Examples of vertebrate predation on cephalopods in the Late Cretaceous ofthe Western Interior, p. 205–207. In A. J. Boucot (ed.), Evolutionary Paleobiology of Behavior and Coevolution.Elsevier, Amsterdam, 750 p.WALDMAN, M. 1970. Comments on a Cretaceous coprolite from Alberta, Canada. Canadian Journal of EarthSciences, 7:1008–1012.WETMORE, A. 1943. The occurrence of feather impressions in the Miocene deposits of Maryland. The Auk, 60:440–441.WILLIAMS, M. E. 1972. The origin of “spiral coprolites.” The University of Kansas Paleontological Contributions,59:1–19.WILSON, J. W. 1994. Histological techniques, p. 205–234. In P. Leiggi and P. May (eds.), Vertebrate PaleontologicalTechniques, Volume One. Cambridge University Press, Cambridge.ZIDEK, J. 1980. Acanthodes lundi, new species (Acanthodii) and associated coprolites from uppermost MississippianHeath Formation of central Montana. Annals of the Carnegie Museum, 49:49–78.ZANGERL, R., AND E. S. RICHARDSON, JR. 1963. The Paleoecological History of Two Pennsylvanian Black Shales.Fieldiana: Geological Memoirs, Volume 4. Chicago Natural History Museum, Chicago, 352 p.49

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 200250

HAYNES—RECONSTRUCTING HUMAN PREDATIONARCHEOLOGICAL METHODS FOR RECONSTRUCTINGHUMAN PREDATION ON TERRESTRIAL VERTEBRATESGARY HAYNESAnthropology Department, University of Nevada, Reno, Nevada 89557 USAABSTRACT—Archeological interest in predation ranges from studies of the earliest evidence for human meateating,to attempts to understand the fossil record’s ambiguity about the meaning of associated animal bones andhuman artifacts. A number of quantitative measures are used to find the meaningful patterns in archeologicalassemblages, and taphonomic research has also provided analogs and comparative standards for interpreting humanbehavior based on the evidence for predation. The most important methods, approaches, and interests are discussedhere, using case studies to illustrate the way archeologists have thought about the record of humans as predators.INTRODUCTIONTHIS PAPER IS A SURVEY of methodsarcheologists use to recognize and analyze predationby humans. The literature is large and the issues ofinterpretation are complex; a quick topical scan likethis is not meant to seem authoritative or complete.Rather it is an entrée into a varied and fascinatingsubfield of scholarship that has produced reviewsof the evidence (e.g., Behrensmeyer, 1987), casestudies, and reconsidered interpretations (Stanfordand Bunn, 2001).Archeologists have two main interests inpredation. One is in distinguishing active huntingfrom the more passive process of scavenging, onegoal being to find the beginning of predation, andanother goal being to understand foraging patternsthroughout all of prehistory. The two different foodprocurement tactics are widely thought todistinguish pre-human from modern humanbehavior over the long course of human evolution(see Klein, 2000; Stiner, 1990; also see thereferences in Stiner, 1991). Another main interestis in modeling human economic activities thatinfluence and are affected by predation, such asfood-getting technology or social organization.Prehistoric human groups colonized and lived inmany different habitats, and the technologicaldemands of predation in each habitat wereextremely diverse. Details about prehistoric groupfissioning and fusing, the scheduling of economicactivities, and the methods used to exchangeforaging information can be partly recoveredthrough careful analyses of archeological evidencefrom predation patterns.Other kinds of interest in predation aresometimes seen in the archeological literature aswell—such as the effort to understand whethercannibalism ever provided a non-negligible amountof calories to the human diet, and the desire torecognize murder and mass killings as special casesof predation. There is also an interest in trying todefine the extent of carnivore predation onhominids throughout prehistory.Laughlin (1968, p. 304) proposed that huntingis “the master behavior pattern of the humanspecies.” But hunting is a risky and oftenunsuccessful activity, and human hunter-gatherersin recent times rarely get more than 30–50% of theircalories from animal tissues (Lee, 1968; see thereferences in Lee and DeVore, 1968). However, atdifferent times in prehistory, pre-industrial foragersapparently acquired far more. Richards et al. (2000)provide a late Pleistocene example in which isotopicstudies of human bones from Britain demonstratedmuch more meat-eating than seen in ethnographicallydocumented (“recent”) non-industrial foragers.Stable-isotope studies of human bones can revealthe extent of meat-eating in prehistoric human diets,but unfortunately the huge majority of prehistoric51

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002archeological sites contain no human bones, or thebones’ chemistry has been completely altered bydiagenesis. Thus isotopic studies are impossible inmost cases. For that reason, I do not discuss isotopicgeochemistry in this paper (see DeNiro, 1987 foran introduction to the literature).Here I describe the methods used to addressquestions about humans as predators, with greaterattention given to the research relating to humanpredation on terrestrial animals.A TRICKY ISSUE:WHAT IS “PREDATION”?When humans prey on animals, they oftenbehave extraordinarily. Not only do humansactively pursue and kill animals for food, they alsokill animals for skins, fur, and other by-products,even when no meat is to be eaten. These practicesare unique forms of predation. A modern exampleof predation for purposes other than human feedingis the case of the subadult caribou killed by Alaska’snative Nunamiut Eskimo people in late summerand fall (Binford, 1978, p. 86) to acquire the softskin for clothing; no meat is taken except fromheads, and the Nunamiut people’s dogs are allowedto eat the rest of the carcasses.A prehistoric example of procurement forproducts other than food may be found in EurasianUpper Paleolithic assemblages containing bonesfrom fur-bearing animals (Soffer, 1985, p. 310-327),usually interpreted as the result of hunting for furs.The common fur-bearers—fox, wolf, bear, andwolverine—do not provide much meat and are notabundant compared to herbivores. Most commonlyrepresented at such sites are the bones from the furbearers’lower limbs, presumably still attached tothe furs after skinning. Further support for the furhuntinghypothesis may also come from seasonalitydata—the optimal season for best fur yield wouldbe winter—and bone cutmark distributions, whichshould show a pattern resulting from carcassskinningrather than meat-stripping or body-partsectioning; this topic will be discussed below.Another prehistoric example can be seen inUpper Paleolithic sites in central Europe whichcontained huge concentrations of woolly mammothbones (see Soffer,1985 for site descriptions andoriginal references). The tons of bones are thoughtto have functioned as walls and supports for dwellingstructures. The conventional archeologicalinterpretation is that some or most of the mammothsdied natural deaths and their bones were gatheredfor buildings and fire-fuel because little wood wasavailable regionally (Soffer, 1985; also see Haynes,1989). Support for scavenging rather than huntingis found in the different degrees of weathering onthe bones (suggesting the mammoths died over anextended period of time), plus the general scarcityof butchering marks on bone surfaces and the sheervolume of bones from dozens of animals, whichmany archeologists do not see as a rational foragingchoice for spear-wielding humans.THE EVOLUTION OF HUMANPREDATION: SCAVENGING ANDHUNTING IN PREHISTORYThe prehistoric record of human meat-eatingis far from clear about the prevalence of scavengingin the earliest human economies. One faction ofprehistorians (including Binford, 1981;Blumenschine, 1995; Blumenschine and Cavallo,1992; Selvaggio, 1994; and Selvaggio and Wilder,2001) argue that Plio-Pleistocene hominids atOlduvai Gorge, Tanzania, were scavengers whosought defleshed carcasses abandoned by leopards,hyenas, or other carnivores. However, anotherfaction (including Bunn and Kroll, 1986; Bunn andEzzo, 1993; Bunn 2001) disagrees and proposesthat the hominids at Olduvai Gorge actively huntedor aggressively chased competing four-leggedcarnivores away from recent kills in order toprocure meat. The disagreement is not resolved,but methods for making the distinctions are stillbeing developed.By the end of the Lower Paleolithic (latemiddle Pleistocene times, around 200,000-400,000years ago), hominid behavior is thought to haveincluded relatively less passive scavenging andmore active hunting (Klein, 1999), althoughscavenging was occasionally done (Grayson and52

HAYNES—RECONSTRUCTING HUMAN PREDATIONDelpech, 1994). Archeological evidence suggeststhat the active hunting often targeted large gamespecies; for example, several sharply pointedwooden spears were found in Schöningen, Germany,along with numerous bones of Pleistocene horsesdated around 400,000 years old (Thieme, 1997). Thespears are interpreted as killing tools, and the horsesare thought to be the prey of Homo heidelbergensisor archaic H. sapiens.By the Middle Paleolithic, beginning around250,000 years ago, archeological sites containpossible evidence of even larger and moredangerous prey being hunted as well as scavenged(Patou-Mathis, 2000). The site of La Cotte de St.Brelade in the Channel Islands between France andEngland contained very incomplete skeletons ofseveral woolly mammoths and woolly rhinocerosesalong with stone artifacts, although no hearths orburnt bones were discovered. The assemblage hasbeen interpreted as what was left in a processingsite where certain meat-bearing body parts weretransported for butchering after mass kills (Scott,1980, 1986).DISTINGUISHING HUNTINGFROM SCAVENGINGEven fully modern foragers can be satisfied withpassive scavenging of already dead animals when itis less costly than the active pursuit and killing oflive prey. But humans are never more than part-timeor opportunistic scavengers and no specializedscavenging human groups are known from theethnographic record. The opportunistic scavengingof animals involves cultural practices that seemalmost identical to those used in active hunting.Modern small-scale foraging groups such as theHadza in Tanzania (O’Connell et al., 1988a, 1988b;Woodburn, 1968) make and use the same tools tokill animals—spears, clubs, or bows-and-arrows—as they do to scare away lions and hyenas fromanimal carcasses. The Hadza (and any other humanforagers) also use the same kinds of butchering toolsto extract food from scavenged carcasses as fromkills they make themselves (Bunn et al., 1988). Thismeans that the artifacts and technology found inprehistoric archeological sites do not by theirpresence distinguish between scavenging andkilling. Other kinds of evidence must be used tomake the distinction. The methods available to makethe distinction between hunting and scavenging mustbe applicable in all contexts, from the very earliestsites to the most recent ones. The immediate costsand benefits of hunting versus scavenging must beunderstandable in order to reconstruct situationallyspecific choices made in the past.One possible if indirect measure of how muchpredation rather than scavenging was done byhumans in prehistory may come frommusculoskeletal analysis of prehistoric humanbones (Peterson, 1998). In cases where there areno animal bones associated with implements madeby humans, the human skeleton itself might be ableto provide indicators (specifically stress markers)that predation-related activities were frequent, suchas habitual overhand throwing, thus supporting thehypothesis that spears were delivered by armmotion or through the use of atlatls, specializedthrowing boards or sticks. Frequent spear-throwinglogically implies frequent predation.Unfortunately, extremely few prehistoricarcheological sites contain human skeletal material,and thus other methods are needed to find theevidence for predation in the fossil record.FINE-SCALE METHODSFOR ANALYZING PREDATIONThe goal of a step-by-step predation analysis isto identify the costs and benefits of humaninvolvement in the killing or utilization of an animalcarcass. Before any economic interpretations can bemade, a first step is to establish the contemporaneityand behavioral association between humans and theanimal remains in question. Bone decomposition,also known as weathering, by itself is an impreciseindicator of the contemporaneity of different skeletalelements (Behrensmeyer, 1978; Lyman and Fox,1989). Weathering varies within similarenvironments due to element differences (thickcompact bone weathers differently than thinnerwalledcancellous bone) or microhabitat differences53

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002(bones shaded in thick vegetation weather moreslowly than bones in direct sunlight).Microstratigraphic studies of buried in situspecimens may indicate if there was an actualtemporal association of human tools and animalbones. Radiometric dating of bone specimens,depositional strata, and organic artifacts also canserve to establish contemporaneity and association.But “association” or “contemporaneity” are not thesame as predation (Haynes and Stanford, 1984);humans may make use of animal bones as buildingmaterials or raw materials for artwork even when theyhave not killed the animals. To establish whetheractual hunting was done, the next necessary step afterestablishing contemporaneity and association is arelatively fine-scale taphonomic analysis.Here I list several direct methods of analyzingthe extent of human predation on animal remains.These methods were developed mainly throughexperimental and observational research. Studies ofthe processes undergone by recent animal carcassesin natural and cultural contexts are termed“neotaphonomy” and experimental archeology,respectively. Blumenschine (1986), Selvaggio(1994), Shipman (1981), and Haynes (1981, 1991),among many others, have brought archeologists’concerns to the study of animal bones. They havecounted modern carcasses in different habitats ofthe world, conscientiously recorded the actions ofscavengers and predators, examined carnivoretoothmarks on bone surfaces, modeled carcassfeedingby carnivores under a variety of differentconditions, experimentally butchered dead animalsto examine toolmarks or observed foragersprocessing animal carcasses, and measured potentialyields from butchered and scavenged animal parts.In an ideal analysis, archeologists would firstvisually or microscopically examine every animalbone fragment (all of which are saved fromexcavations, no matter how scrappy) for thepresence of marks made by butchering tools,carnivore teeth, hammerstones, or other objects.Also, the element name and body side of allspecimens, whole or fragmentary, would beidentified so a minimum number of individualanimals could be established. Much of this workmust be by comparison to reference collections andrequires training and practice (Lyman, 1994; seeespecially Blumenschine et al., 1996). The goalsof such labor-intensive analyses are very specific;I list them here in no special order:1. Spatial associations and congruent absolutedating of animal bones and human-made implements(“artifacts”) may suggest predation or scavenging.I list this method first but temper its apparentusefulness with the observation that even thetightest spatial/stratigraphic and temporalassociation of skeletal elements with human-madeartifacts such as killing implements or butcheringtools may be the most overvalued potentialindicator of predation. Artifact associations can beover-interpreted to indicate behavioral actions thatnever happened, such as the killing and butcheringof animals. Artifacts discarded at the same placeswhere animals died natural deaths do not alwaysreflect human involvement with the deaths.Careful consideration must be given to thelandscape settings of associated artifacts and animalbones, because animal remains may be naturallyabundant in certain localities such as around watersources or in caves and rockshelters where bothhumans and carnivores provisioned their young orbrought prey body parts to eat. In these locationsanimal bones and human tools or bones often cooccurbut may reflect neither human scavenging norpredation, and instead result from a generalterrestrial mammalian inclination to find shelter orresources in similar places. Thus the co-occurrenceof artifacts and animal bones does not clearly implyhuman predation in many settings, as demonstratedfor such sites as the famous South African Plio-Pleistocene hominid-bearing cave deposits (Brain,1981, 1993; de Ruiter and Berger, 2000).An interesting example of how a uniquelandscape setting actually conditioned humans toscavenge large terrestrial vertebrates is theShestakova site in western Siberia, dated to justbefore the Last Glacial Maximum (18-25,000radiocarbon years BP). Derevianko et al. (2000)proposed that woolly mammoths preferentiallycongregated at a salt-rich rock exposure in a hollowon a riverbank, where at least 18 animals died of54

HAYNES—RECONSTRUCTING HUMAN PREDATIONnatural causes at various times. This featureattracted human foragers to the same place, and thepeople made use of some bones, transporting themaway from the deathsites. In this case, the scavengingwas not for food but for building material or perhapsfuel for heating or cooking fires.Human foragers transport animal body partsto homesites in caves, rockshelters, or open-airsettings regularly for processing and cooking, aswell as for tool-making using the bones as rawmaterial. A zooarcheological analyst is thereforecommonly faced with the need to distinguishpatterning in bone accumulations that resulted fromhuman versus non-human processes. Thearcheological record shows that ambiguity extendsfar back into prehistory. For example, Gargett(1996) demonstrated that Ursus spelaeus (cavebear) behavior in the middle Pleistocene creatednon-random animal-bone assemblage patterns—in spatial distributions and skeletal partfrequencies—that could be mistaken for culturalpatterning, thus introducing potentially seriouserror into interpretations of Pleistocene humanbehavior. Several methods to tease apart humanfrom nonhuman agencies have been developed byarcheologists. Well illustrated guides show thespecific traces that animal-scavenging creates inbone assemblages. For example, Haynes (1982)detailed the sequence and timing of Canis lupus(wolf) feeding on Bison bison (bison) carcasses;Selvaggio and Wilder (2001) described felid andhyenid damage on African Plio-Pleistocene animalbones; Sampson (2000) found that South Africanraptor-accumulated tortoise bones could be easilydistinguished from human-accumulated tortoiseassemblages by bone condition and elementrepresentation, such as the presence of abundantneck and head bones if raptors accumulated bones,or relatively more plastrons and carapaces in ahuman accumulation.It has become very clear throughneotaphonomic studies (such as Haynes, 1991) thatspatial and stratigraphic associations are notadequate or unassailable evidence of killing. Eventhe discovery of human-made artifacts directlybedded with animal bones does not prove primafacie that the animals were killed and butcheredby humans using those tools. The tools inassociation with animal bones ideally must showmicroscopic evidence of use either as projectilesor butchering implements (see Keeley, 1980 foran introduction to microwear analysis).Unless microwear studies are carried out, thespatial association of stone tools and animal bonescan be seriously misinterpreted. For example,microwear analyses of sharp stone-tool edges fromAfrican Plio-Pleistocene cultural assemblagesindicate that some of the ancient stone flakesclosely associated with fossil animal bones were notused to butcher carcasses, but rather to scrape woodor cut siliceous plants such as grasses and sedges(Blumenschine and Selvaggio, 1988). Thus thepresence of such tools does not always reflecthominid predation or scavenging. Necessary furthersupport for the interpretation of killing would bethe presence of toolmarks or hammerstone impactmarks preserved directly on the animal bone surfaces(reflecting butchering of the carcass when fresh),the association of bones with actual projectilepoints, or the presence of burnt bones withincultural features such as hearths or pits, alsoimplying the processing of fresh carcasses.Modern sites of African elephant bones (Fig. 1)resulting both from human kills and from nonculturaldeaths—such as drought-caused die-offsFIGURE 1—Mass death site of Loxodonta africana(African elephant) in Zimbabwe. Climate-causedstarvation killed over 70 animals around this naturalwater point in 1994 and 1995.55

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002ABFIGURE 2—A, Cutmarks on the distal end of a metapodial of Alces alces (moose). The marks werecreated by a steel knife used to skin the animal. B, Cutmarks on fragment of diaphysis of Equus caballus(domestic horse) experimentally butchered with stone tools. The marks were made during meat-stripping,and preceded the breaking of the element with a hammerstone. The signs of skinning versus meat-strippingare distinguished not by the morphology of each mark, but by the placement and intensity of the marks.(Haynes, 1995, 1999)—are sources of multivariatemeasures that might one day provide criteria fordistinguishing predation from scavenging ofprehistoric proboscideans, still an unsettled issue forsome time periods and world regions (Leshchinskiy,2001; Niven, 2001; Péan, 2001; Vasil’ev, 2001). Thepotential variables in ongoing neotaphonomic fieldstudies, such as whether the elephant bone sitescontain transported or in situ bones, whether theywere created serially or en masse, or whether theywere culturally or nonculturally produced, are stillbeing identified.2. Cutmark analysis may provide evidence thatdistinguishes hunting from scavenging.When tools are used to cut away meat fromthe skeleton, distinctive marks are often made atspecific attachment points on bone surfaces.Likewise, the removal of skin or the separation ofbody parts at joints leaves different marks ondifferent parts of the skeleton, which thus identifythe particular processes the carcasses underwent—skinning versus meat-stripping versus body-partsectioning (Fig. 2). Most carnivore toothmarks maybe distinguished from toolmarks by either visualor microscopic examinations (Fig. 3) (Binford,FIGURE 3—Scanning electron microscope (SEM)photomicrograph of cutmarks created by a stone toolused to strip meat from the femur of an experimentallybutchered Equus caballus (domestic horse).56

HAYNES—RECONSTRUCTING HUMAN PREDATIONFIGURE 4—Toothmarks of Canis lupus (wolf) ona femur shaft fragment from Bison bison (bison).The marks were created by the carnassial teethstripping meat, and are very similar to marksproduced by humanly made butchering tools.1980; Blumenschine et al., 1996; Haynes, 1981;Shipman and Rose, 1983), although some marksare not easily distinguished (Fig. 4). The presenceof cutmarks on the meat-bearing portions of bonesis an indication that meat was still present andcarnivores had not fed on or “ravaged” the remainsfirst. The cuts imply that relatively complete andfresh carcasses were processed, and suggest thathumans either made the kills or acquired thecarcasses immediately after the kills were madeby other carnivores.Very well preserved killsites need not alwayscontain complete toolmarked skeletons becausehumans often transport body parts to homesites orconsumption areas removed from the kill/processingstations. “Utility” indices have been devised(Binford, 1978) to predict how the various parts ofanimal carcasses would be valued for meat,mesentery and muscle fat, and marrow, and thusimply the likelihood of specific bones or body partsbeing transported away from kills (Lupo, 2001; seealso Grayson,1989, and Lyman, 1985 for otherconsiderations). These indices differ by animal taxon(see, for example, Outram and Rowley-Conwy, 1998on horse; Binford, 1978 on domestic sheep andcaribou; Metcalfe and Jones, 1988 on caribou;Lyman, 1994 lists indices for several other taxa).These quantitative measures help explain humandecisions to remove or not to remove anatomicalparts from killsites; plus, they are also potentiallyuseful for suggesting whether humans actuallykilled the animals or found them dead aftercarnivores had ravaged the remains. Bones foundin situ with very few butchering marks andconsisting mainly of low-utility elements such aslower legs would not have provided meat or muchuseful skin. If the animal remains had beenencountered by human foragers only aftercarnivore feeding (in other words, if the humansscavenged them), relatively few butchering markswould have been made when the last scraps ofusable tendon or meat were cut off, or when limbelements were broken for salvageable marrow. Thelocation and intensity of butcher-marking plus thelow utility of the butcher-marked body parts wouldsuggest that the animal carcass was found aftercarnivore feeding.Carcasses butchered when dried, frozen, or stiffwith rigor mortis may be butcher-marked byhumans in ways different from fresh carcasses.Lupo (1994) has identified a category of stiffcarcasscutmarks seen on bones butchered whencarcasses were not fresh. Relatively more forcefulcutting or chopping may leave unique marks whenfrozen or dry tissue is difficult to remove; certainbones may be chopped through rather thansectioned at joints (ethnographic examples are inBinford, 1978, 1984, 1988; and Lupo, 1994;Saunders and Daeschler, 1994 illustrate what theyconsider to be an archeological example). Hunterswho use poison-tipped projectiles may not be ableto butcher animals immediately after they die ifthe prey must be tracked for a long distance (Lupo,1994); this technology could be associated withstiff-carcass buchering relatively often. Binford(1978, p. 480) also noted different placement andintensity of marks on bones when carcasses werebutchered to feed dogs rather than people.Another possible cause of variability inbutchering is the quantity of animal prey killed atone time. In the large prehistoric mass kills of bisonon the North American High Plains, such as Olsen-57

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Chubbock, Colorado (Wheat, 1972), manycarcasses were fully sectioned or meat-stripped butothers were not, probably because human butchersreached the limit of how much meat they couldprocess, store, or carry. This site and others like itare examples of “surplus killing” by humanpredators. Surplus killing has been recognized inmany other predators, such as grey wolf(DelGuidice 1998).3. Bone-breakage analysis may also provideevidence about hunting or scavenging.Humans often break animal bones to get at themarrow or oil-rich cancellous interiors (Fig. 5). Inaddition, bone fragments can be boiled to make asoup from the fat in them. But carnivores also breakhigh-utility bones, and animal trampling canfragment elements (Fig. 6). The agencies thatcaused the bones to break must be accuratelyidentified, and a few guides are available to helpidentify whether carnivores, animal trampling,human processing, or other agencies areresponsible (Binford, 1981; Blumenschine et al.,1996; Haynes, 1981, 1983, 1991; Hesse andWapnish, 1985; Shipman, 1981; Shipman andRose, 1983). The utility indices mentioned aboveFIGURE 5—Fragments of a femur diaphysis fromBos taurus (domestic cow). The bone was brokenby a hammerstone to extract marrow. Note thenotched edges on the specimen to the right; thesenotches mark the spots where hammerstoneimpacts occurred.FIGURE 6—Fragment of a femur shaft fromSyncerus caffer (Cape buffalo). The bone wasfragmented by a Crocuta crocuta (spotted hyena)feeding on the carcass. Note that the notchededges, created by the hyena’s teeth, are very similarto hammerstone-impact notches (see Figure 5).are often used to determine if bones were brokendeliberately by humans following a rational agenda(for example, bones with plenty of marrow shouldbe frequently broken, while bones lacking marrowor fat should not be).The morphology of fractures must be carefullyevaluated to determine if the breaks were madewhen bones were fresh, and if human percussionaccounts for the fragmentation. Blumenschine andSelvaggio (1988) published descriptions ofdistinctive hammerstone marks in order todifferentiate them from animal toothmarks or othernon-cultural damage to bone surfaces.4. Mortality-profile analysis may indicatewhether animals were hunted or scavenged.Klein (1994, for example) and others (Stiner,1990, 1991) have reconstructed hominid huntingmethods based on age profiles of animal boneassemblages. Catastrophic age profiles may implynonselective hunting (the killing of animals ofdifferent age and sex classes in the sameproportion as in the population), such as throughherd surrounds (see Saunders, 1980) or drivesover high cliffs (see Reher and Frison, 1980;Frison, 1970); while non-catastrophic profilesmay indicate deliberate selection of morevulnerable animals, such as the very young or veryold (see Klein, 1994 for a comparison of huntingtactics reconstructed for southern Africa’s Middle58

HAYNES—RECONSTRUCTING HUMAN PREDATIONand Later Stone Ages). However, nonselective andselective age profiles may both result underdifferent conditions, depending upon how theanimals die, so other kinds of evidence beyondmere age-profiling must be available to supportan interpretation of scavenging or hunting.Plio-Pleistocene animal bone sites aresometimes interpreted as the result of humanpredation, but absolute proof is often lacking. Siteswith multiple individuals of large mammalian taxabut without clearly associated artifacts may havemortality profiles interpreted to mean that humanskilled the animals. An example is the Nanjing Manmiddle Pleistocene site in China, which yieldedfragmented bones and teeth of several extinct rhino(Dicerorhinus mercki) individuals. No cutmarks orstone tools were associated with the bones (Tong,2001). Tong reasoned that even though the rhinoswere mostly subadults they were so large that noother predator except hominids could haveaccounted for the deaths. Tong also argued that therhinos must have been actively hunted because thebones were not located in any kind of natural trapyet were fragmented and incomplete; thus theymust have been transported from killsites by theagent that had killed them. Unfortunately, theNanjing assemblage was not analyzed in terms ofsuch simple measures of bone survivability as bonedensity—which could explain why certainelements of the entire skeleton were notpreserved—nor in terms of transport utility (fordiscussions of the complex interplay between bonetransport data and differential destruction due tobone density, see Grayson, 1989; Lyman, 1985,1994). Without a fair appraisal of all the potentialattritional processes affecting the skeletons, andall the preserved elements’ utility, the interpretationcannot be confidently accepted that humans aloneaccount for this entire assemblage. An alternativehypothesis is that the bones represent a residualnon-cultural accumulation such as a palimpsestcarnivore densite containing scavenged remainsfrom carnivore kills.5. Economic-return models provide evidenceabout the costs and benefits of predation.A set of analytical measures is available toprovide indirect evidence about the costs andbenefits of predation. These measures are data- andtheory-based—that is, they have been derived fromboth empirical observations and deductions abouthuman needs and abilities. Zooarcheologicalstudies (for discussion of the basic methods, seeHesse and Wapnish, 1985; Klein and Cruz-Uribe,1984; see also Grayson, 1984, and Lyman, 1994for additional discussion) begin with best estimatesof the minimum number and relative frequenciesof bone elements (whole bones with individualnames such as humerus, rib, femur) from animalsof each taxon represented in a cultural assemblage,usually calculated as both NISP (number ofidentified specimens) and MNE (minimum numberof elements). These are counts of bones andfragments that do not have overlapping orduplicated morphological features, and ideally takeinto consideration variations due to age and sex.All whole and partial specimens from each taxonare identified to the level of element name, side ofthe body, and segment of element (proximal, distal,diaphyseal). Using these numbers, an estimate thencan be made of the minimum numbers ofindividuals (MNI of each taxon) represented byall the specimens. An alternative measurementcalled minimum animal unit (MAU) is calculatedby counting the different elements in assemblages(such as, for example, femora, both rights andlefts), then dividing by the number of such elementsin a skeleton (two, in the case of femora—thus 24left femora would be considered to represent only12 animals or animal “units,” thereby providing ameasure of actual yield rather than number ofindividuals killed). Lyman (1994, p. 104–110)discussed the potential differences in terms ofinterpretative implications.Like the utility indices, these measures can beused to determine if human foragers transportedbody parts away from killsites or into homesites,or if the bones of whole animal skeletons weredifferentially “subtracted” through weathering,carnivore ravaging, fluvial action, etc. (seeGrayson, 1984, 1989). These measures have alsobeen used by archeologists to estimate meat yieldsand degrees of carcass utilization by humans, both59

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002from transported and residual assemblages, butoften without an awareness of the possibleproblems discussed by Grayson (1984, 1989) orLyman (1994 and references therein). Nevertheless,modeling predation outcomes based on thefundamental zooarcheological quantifications(such as MNI or MAU) often inspires testable ideasabout human hunting strategies, especially whensupported by animal behavior models, foragingtheory, or other kinds of evidence.Economic models that refer to these measuresand to possible yields of meat and carcass byproducts(skins, horns, antlers, bones) usually citetheoretical concepts or possibilities whenreconstructing predation practices. For example,Todd (1991) analyzed Paleoindian boneassemblages from mass kills in the American West(see also Davis and Wilson, 1978; Frison, 1978,1998). Based on bison tooth eruption he determinedthat the killing was done in the late fall or earlywinter. Through analysis of cutmarks and body-partfrequencies, he also found that the carcass butcheringwas a “bulk-processing” pattern that optimized thestorage potential of the lean meat rather than therecovery of bone fats. A human diet consistingexclusively of lean meat is not feasible (Speth, 1990;Speth and Speilmann, 1983), so Todd proposed thatextremely mobile bison hunters were able to survivea meat-heavy diet low in carbohydrates and fatsby consuming the carbohydrates out of butcheredrumens. Thus, the hunters did not need to spendtime processing low-yield plant foods for neededcarbohydrates or high-cost bones for the marrow/fat, and were seasonally specialized predators. Thispattern was not in place in later prehistory, whenmuch more effort went into bone processing andthe storage of fats and carbohydrates.The cost-benefit analysis of classic foragingtheory (Stephens and Krebs, 1986; for summariesand archeological uses, see Kelly, 1995; Bettinger1991) may suggest how human foragers rankeddifferent prey species, and how they may havepursued and processed a range of prey taxa. Butdiet breadth models require quantified data suchas search and handling times and energy yieldsfrom foods, which cannot be measured directlyfrom archeological assemblages. Nonetheless,archeologists are willing to make even roughestimates of these variables based on meat-yieldassessments if they aid in understanding pasthuman predation. The estimates may be based onlayered assumptions about prehistoric animaldistributions, dangers of the hunt, and otherunproven variables, but they are intended to betested against archeological data and often arefertile sources of new ideas.6. Replicative and analogical studies mayprovide evidence about the feasibility (andplausibility) of predation.Another line of evidence that may be helpfulin understanding human predation comes from“replicative” studies where ancient behavior iscopied. Such activities are termed experimentalarcheology. One example would be the manufactureand use of replica prehistoric tools to process animalcarcasses in order to record the suitability of ancienttechnology (see Huckell, 1979, for an example),and to document the places on bones where theFIGURE 7—Opportunistic replicative butchering ofLoxodonta africana (African elephant) carcasses inZimbabwe using steel and stone cutting implements.All animals were killed by government-sanctionedwildlife rangers for management purposes. After thebutchering was complete, bones were carefullyexamined for cutmarks resulting from meatstripping,skin removal, and body sectioning.60

HAYNES—RECONSTRUCTING HUMAN PREDATIONbutcher marks would be found when differentcarcass processing strategies were used (Fig. 7).Future replicative study may incorporate thetesting of copied prehistoric weaponry for accuracyand dropping power, thus supporting interpretationsof predation on certain species. Other future studiesmay try out various methods of hunting such asambushing or driving different taxa. These twolatter kinds of replicative experiments have notbeen attempted or reported, for understandablereasons, but would be very useful. While suchstudies cannot provide direct evidence abouthuman predation in the past, they do offer a degreeof plausibility or feasibility that is worth havingwhen reconstructing variable hunting and carcassprocessingpractices. By understanding weaponuses, archeologists may then also be able toreconstruct different predation tactics such assustained pursuit or mass driving, partly reflectedin the discarded weaponry at archeological sites.For example, clubs and simple thrusting spears maybe numerous where drivelines and corrals arefound, whereas ambush sites may contain killingimplements delivered from a distance, such asthrowing spears or arrows.IMPLICATIONS OFANALYTICAL RESULTSThe main goal of all the archeological methodssurveyed above is not just to list the animal specieseaten by human foraging groups in prehistory orthe numbers of animals killed in any unit of time.The bigger goal is to understand how predationaffects the ways that human cultures organizethemselves in all aspects—including socialorganization and kinship systems, settlementmobility and scheduling, technology and landscapeuse, and so forth. Humans that prey on migratorylarge mammals may or may not be migratorythemselves, and the way carcasses are treated infact will reflect this important characteristic.Foraging humans often share meat and carcassparts from kills, and the network of people whowill be given pieces of animal carcasses may belarge and complex. Obviously predation practicesand body-part distributions will differ when meatsharingis regularly done, as compared to thepractices of groups that do not cement social orkinship bonds with regular meat re-distribution.The last topic surveyed in this paper ispredation on humans, by both other humans andquadrupedal carnivores.ANALYZING PREDATIONBY HOMINIDS ON HOMINIDSCannibalism.—Cannibalism is the eating ofhuman bodies by other humans. Ritual cannibalismhas been practiced at many times and in many partsof the world; culinary or dietary cannibalism areterms used when the practice is frequent andprovides a substantial part of the human diet(Keeley, 1996, p. 103–106). Late in the 20 th century,anthropologists generally declared humansfeeding-on-humansa fiction or at most only asymbolic action (Arens, 1979; Peter-Röcher, 1994;but see Forsyth, 1983, 1985). However, thearcheologist Lawrence Keeley (1996) haspersuasively argued that at least some historic andprehistoric societies derived a non-negligibleamount of calories from cannibalism.Middle Paleolithic (Neandertal) cannibalismin Europe has been hypothesized based on markedskeletal elements, but re-analysis indicates thatsome instances may have been mis-interpreted.One ambiguous example is the Krapina (Croatia)Neandertal bones, which were affected by nonculturalpostmortem events such as carnivoregnawing or trampling (Trinkaus, 1985), but also mayhave been cut by Neandertal stone tools to preparea fresh body for secondary burial (Russell, 1986).By contrast, the 100,000-year-old bones of sixNeandertal individuals from the Moula-Guercy cavesite in France do seem to have been cut and brokenfor “culinary” cannibalism (Defleur et al., 1999).Paleoanthropologists and archeologists believethey can identify when either the “ritual” or the“culinary” form of the practice occurred (seeespecially Turner and Turner, 1999; see also Villa,1992; White 1992). The methods used to identifyinstances of culinary cannibalism are similar to61

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002those employed to analyze predation in general.The meat available on a typical human bodyis distributed similarly to meat on quadrupeds;therefore, cutmarking during butchering, bodysectioning, or skinning by humans using stone ormetal should also be similar. Bone breakage formarrow extraction will also be identifiable, asopposed to bone breakage in order to section thebody for re-burial or trophy distribution. In otherwords, culinary cannibalism will leavebutchermarks or damage to human bones patternedlike the marking and damage to animal-prey bones.Thus, for an interpretation of cannibalism towithstand scrutiny, it is necessary that a site’shuman and animal bones be similarly butchermarkedand fragmented (Turner and Turner, 1999;White, 1992). Otherwise the human skeletalsectioning represented by cutmarks, chop damage,and bone breakage may reflect not butchering afterpredation but preparation for re-burial or ceremonialtreatment, which many ethnographicallydocumented human groups are known to have done.Murder, violent killings of other humans.—Urbanized state societies see plenty of violentkillings, such as in organized warfare, gang wars,serial or mass murder, and individual murder, butthese are not forms of “predation” as defined byecologists. Keeley (1996, p. 32–33) surveyed theliterature, and estimated that non-state societieswere at war with other social groups even morefrequently than were state societies: 70–90% ofnon-state societies seem to have been at war onceevery couple of years; and even the simplest,lowest-density forager groups (the Kalahari Sanor “Bushman” groups in southern Africa, forexample) have had murder rates as high or higherthan among modern urban populations.Murder as a form of predation is abioarcheological and forensic issue, and theanalytical and interpretive literature to helpunderstand human actions based on archeologicalbones is slowly growing. Walker (2001) providesa good introduction to the topic; Haglund and Sorg(1997) provide a wide survey of case studies andemerging principles.When prehistoric human bodies are so wellpreserved that soft tissue survives, archeologists canuse standard forensic analysis to determine if thecause of death was murder. Some of the “bog bodies”from Iron Age Europe were probably killed as eithersacrifices or executions; a few have nooses aroundtheir necks, or show severe soft-tissue trauma suchas slashed throats (see Brothwell, 1986; Glob, 1971).In the absence of signs of violent damage to softtissue and bone, archeologists also may submit tissuesamples in search of poison or the presence of toxicchemicals in human tissues.In prehistoric cases where no soft tissue ispresent, a careful examination of human skeletonsfrom ancient graveyards or single-burial sites mayreveal projectile points embedded in bones. Forexample, the bones of 8,400-radiocarbon-year-old“Kennewick Man” from Washington state had ahealing wound in his pelvis where a stone-tippedspear had been thrust or thrown (Chatters, 2001).The famous 5,300-radiocarbon-year-old “Iceman”discovered in the Italian/Austrian alps (Spindler,1995) had died from an arrow shot into his back(Glausiusz, 2002). There are other known instancesof prehistoric human skeletons—in Africa, NorthAmerica, Europe—discovered with probable arrowor spear points bedded with them in graves orcemetery pits, strongly suggesting death at thehands of other humans.Killing for trophies such as enemy heads,other body parts, or scalps (Keeley, 1996, p. 99–103; Milner et al., 1991, and references therein)can be discerned in the prehistoric records frommany parts of the world. Like murder or ritualexecutions, these examples are generally notconsidered clear cases of predation, because thebodies usually were not butchered to be eaten;but they are instances where a battery of similaranalytical methods can discern the surprisingextent of humans killing humans in thearcheological record (Keeley, 1996).Carnivore predation on humans.—Occasionally an ancient human skeleton is recoveredthat appears to have been ravaged by carnivores.Most were probably individuals who died from avariety of non-predation causes such as old age ordisease, and whose bodies were scavenged by62

HAYNES—RECONSTRUCTING HUMAN PREDATIONcarnivores soon after death. A careful examinationof the toothmarks on the bones, the skeletalsectioning, and the range of bone damage mayindicate the timing of the scavenging and the taxaresponsible for it. Examples of useful guides are thepublished reports such as Carson et al. (2000), whichdescribes the effects of bear scavenging on humanskeletons; or Haglund (1991), who described coyotegnaw-damage and feeding sequences as well as otherprocesses affecting modern human bodies.Brain (1981) hypothesized that at least oneAustralopithecus individual represented in the Plio-Pleistocene cave deposits of South Africa was thevictim of leopard predation, based on toothmarkingand breakage of bones. Gargett (1989, 1999) arguedthat many Middle Paleolithic hominid skeletonsin caves were not purposeful burials as ofteninterpreted, but were merely fortuitously wellpreserveddeaths. Some may have been the remainsof humans that were scavenged or even preyed onby carnivores such as cave lions or hyenas. Theseand other archeological examples perhaps faintlysuggest that predation on humans may have beenan ever-present danger in prehistory.CONCLUSIONThere is so much to talk about when referringto human predation that it is an almost overwhelmingtopic to summarize. I have narrowed the themes tojust a few—such as how archeologists distinguishactive procurement (killing) from passiveprocurement (scavenging) of terrestrial mammals,or how the location of cutmarks on animal bonesreveals whether humans were skinning, meatstripping,or sectioning animal carcasses. Manymore themes could have been explored, such asanalytical approaches to interpreting the gearingupactivities and toolkit maintenance of humanpredators, or the archeological methods used tounderstand seasonal shifts in predation patterns.Not discussed here are other predation practicessuch as fishing and whaling, which were importantin human prehistory, too. Shell middens and fishbones in Pleistocene sites indicate that even premodernpredatory humans had an impact on marineecosystems as well as terrestrial ones.As is the case with other topics in archeologicalresearch, the future research agenda includes acontinuing urge to expand and improve the methodscurrently in use in order to eliminate the possibilityof interpretive error caused by equifinality orambiguity. Unfortunately, neotaphonomic andexperimental field studies are now being carried outless often than in the 1970s and 1980s whentaphonomic research burgeoned in archeology. Thereduction is mainly due to funding shortages, butthere is also a sort of complacency that has settledover many archeologists who think that all thenecessary work has been done, and that recipe booksare now available to allow unerring interpretationsof all possible animal bone modifications.Meanwhile, although the methods areimperfect, they do allow archeologists to makewarranted and supportable interpretations of humanpredation even in the most distant past, and theyhelp us to better understand human economies andsubsistence strategies.REFERENCESARENS, W. 1979. The Man-Eating Myth. Oxford University Press, Oxford (U.K.).BEHRENSMEYER, A. K. 1978. Taphonomic and ecologic information from bone weathering. Paleobiology, 4:150–162.BEHRENSMEYER, A. K. 1987. Taphonomy and hunting, p. 423–450. In M. H. Nitecki and D. V. Nitecki (eds.), TheEvolution of Human Hunting. Plenum Publishing Corporation, New York.BETTINGER, R. L. 1991. Hunter-Gatherers: Archaeological and Evolutionary Theory. Plenum Press, New York.BINFORD, L. R. 1978. Nunamiut Ethnoarchaeology. Academic Press, New York, 509 p.BINFORD, L. R. 1980. Willow smoke and dogs’ tails: Hunter-gatherer settlement systems and archaeological siteformation. American Antiquity, 45:4–20.BINFORD, L. R. 1981. Bones: Ancient Men and Modern Myths. Academic Press, New York, 320 p.63

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BINFORD, L. R. 1984. Faunal Remains From Klasies River Mouth. Academic Press, New York, 283 p.BINFORD, L. R. 1988. Fact and fiction about the Zinjanthropus floor. Current Anthropology, 29:123–149.BLUMENSCHINE, R. J. 1986. Carcass consumption sequences and the archaeological distinction of scavenging andhunting. Journal of Human Evolution, 15:639–659.BLUMENSCHINE, R. J. 1995. Percussion marks, tooth marks, and experimental determinations of the timing ofhominid and carnivore access to long bones at FLK Zinjanthropus, Olduvai Gorge, Tanzania. Journal ofHuman Evolution, 29:21–51.BLUMENSCHINE, R. J., AND J. A. CAVALLO. 1992. Scavenging and human evolution. Scientific American, 1992:90–96.BLUMENSCHINE, R. J., C. W. MAREAN, AND S. D. CAPALDO. 1996. Blind tests of inter-analyst correspondence andaccuracy in the identification of cut marks, percussion marks, and carnivore tooth marks on bone surfaces.Journal of Archaeological Science, 23:493–507.BLUMENSCHINE, R. J., AND M. M. SELVAGGIO. 1988. Percussion marks on bone surfaces as a new diagnostic ofhominid behavior. Nature, 333:763–765.BRAIN, C. K. 1981. The Hunters or the Hunted? University of Chicago Press, Chicago, 365 p.BRAIN, C. K. (ed.). 1993. Swartkrans, A Cave’s Chronicle of Early Man. Transvaal Museum Monographs, 8.BROTHWELL, D. 1986. The Bog Man and the Archaeology of People. British Museum Publications, London, 128 p.BUNN, H. T. 2001. Hunting, power scavenging, and butchering by Hadza foragers and by Plio-Pleistocene Homo,p. 199–218. In C. B. Stanford and H. T. Bunn (eds.), Meat-Eating and Human Evolution. New York, OxfordUniversity Press, 370 p.BUNN, H. T., L. BARTRAM, AND E. M. KROLL. 1988. Variability in bone assemblage formation from Hadza hunting,scavenging, and carcass processing. Journal of Anthropological Archaeology, 7:412–457.BUNN, H. T., AND J. A. EZZO. 1993. Hunting and scavenging by Plio-Pleistocene hominids: Nutritional constraints,archaeological patterns, and behavioural implications. Journal of Archaeological Science, 20:365–398.BUNN, H. T., AND E. M. KROLL. 1986. Systematic butchering by Plio/Pleistocene hominids at Olduvai Gorge,Tanzania. Current Anthropology, 5:431–452.CARSON, E. A., V. H. STEFAN, AND J. F. POWELL. 2000. Skeletal manifestations of bear scavenging. Journal ofForensic Sciences, 45(3):515–526.CHATTERS, J. 2001. Ancient Encounters: Kennewick Man and the First Americans. Simon and Schuster, NewYork, 303 p.DAVIS, L. B., AND M. WILSON. 1978. Bison Procurement and Utilization: A Symposium. Plains AnthropologistMemoir, 14:1–361.DEFLEUR, A., T. WHITE, P. VALENSI, L. SLIMAK, AND É. CRÉGUT-BONNOURE. 1999. Neanderthal cannibalism at Moula-Guercy, Ardèche, France. Science, 286:128–131.DELGIUDICE, G. D. 1998. Surplus killing of white-tailed deer by wolves in northcentral Minnesota. Journal ofMammalogy, 79(1):227–35.DENIRO, M. J. 1987. Stable isotopy and archaeology. American Scientist, 75(2):182–191.DEREVIANKO, A. P., V. N. ZENIN, S. V. LESHCHINSKIY, AND E. N. MASHCHENKO. 2000. Peculiarities of mammoth accumulationat Shestakovo site in west Siberia. Archaeology, Ethnology & Anthropology of Eurasia, 3(3):42–55.FORSYTH, D. W. 1983. The beginnings of Brazilian anthropology: Jesuits and Tupinamba cannibalism. Journal ofAnthropological Research, 39(1):147–178.FORSYTH, D. W. 1985. Three cheers for Hans Staden: The case for Brazilian cannibalism. Ethnohistory, 32(1):17–36.FRISON, G. C. 1970. The Glenrock Buffalo Jump, 48CO304. Plains Anthropologist Memoir, 7:1–66.FRISON, G. C. 1978. Prehistoric Hunters of the High Plains. Academic Press, New York, 457 p.FRISON, G. C. 1998. Paleoindian large-mammal hunters on the plains of North America. Proceedings of theNational Academy of Sciences of the USA, 95:14576–14583.GARGETT, R. H. 1989. Grave shortcomings: The evidence for Neandertal burial. Current Anthropology, 30:157–190.GARGETT, R. H. 1996. Cave Bears and Modern Human Origins: The Spatial Taphonomy of Pod Hradem Cave,Czech Republic. University Press of America, Lanham, 265 p.GARGETT, R. H. 1999. Middle Palaeolithic burial is not a dead issue: The view from Qafzeh, Saint-Césaire,Kebara, Amud, and Dederiyeh. Journal of Human Evolution, 37:27–90.GLAUSIUSZ, J. 2002. A stone-age murder mystery. Discover, 23(1):63.64

HAYNES—RECONSTRUCTING HUMAN PREDATIONGLOB, P. V. 1971 (orig. 1965). The Bog People: Iron-Age Man Preserved. Ballantine Books, New York, 142 p.GRAYSON, D. K. 1984. Quantitative Zooarchaeology. Academic Press, Orlando, 202 p.GRAYSON, D. K. 1989. Bone transport, bone destruction, and reverse utility curves. Journal of ArchaeologicalScience, 16:643–652.GRAYSON, D. K., AND F. DELPECH. 1994. The evidence for middle Palaeolithic scavenging from Couche VIII,Grotte Vaufrey (Dordogne, France). Journal of Archaeological Science, 21:359–375.HAGLUND, W. D. 1991. Applications of Taphonomic Models to Forensic Investigations. Unpublished Ph.D.dissertation in Anthropology, University of Washington, 252 p.HAGLUND, W. D., and M. H. Sorg. 1997. Forensic Taphonomy: The Postmortem Fate of Human Remains. CRCPress, Boca Raton, 636 p.HAYNES, G. 1981. Bone Modifications and Skeletal Disturbances by Natural Agencies: Studies in North America.Unpublished Ph.D. dissertation in Anthropology, Catholic University of America, 527 p.HAYNES, G. 1982. Utilization and skeletal disturbances of North American prey carcasses. Arctic, 35(2):266–281.HAYNES, G. 1983. Frequencies of spiral and green-bone fractures on ungulate limb bones in modern surfaceassemblages. American Antiquity, 48(1):102–114.HAYNES, G. 1989. Late Pleistocene mammoth utilization in northern Eurasia and North America. Archaeozoologia,3(1–2):81–108.HAYNES, G. 1991. Mammoths, Mastodonts, and Elephants: Biology, Behavior, and the Fossil Record. CambridgeUniversity Press, Cambridge (U.K.), 413 p.HAYNES, G. 1995. The role of mammoths in rapid Clovis colonization. Paper presented at 1 st International MammothSymposium, November, 1995, St. Petersburg, Russia.HAYNES, G. 1999. The role of mammoths in rapid Clovis dispersal, p. 9–38. In G. Haynes, J. Klimowicz, and J. W.F. Reumer (eds.), Mammoths and the Mammoth Fauna: Studies of an Extinct Ecosystem. Deinsea (Annualof the Natural History Museum Rotterdam), 6.HAYNES, G., AND D. STANFORD. 1984. On the possible utilization of Camelops by early man in America. QuaternaryResearch, 22:216–230.HESSE, B., AND P. WAPNISH. 1985. Animal Bone Archeology: From Objectives to Analysis. Manuals on Archeology,5. Taraxacum, Washington, D.C., 132 p.HUCKELL, B. B. 1979. Of chipped stone tools, elephants, and the Clovis hunters: An experiment. PlainsAnthropologist, 24:117–189.KEELEY, L. H. 1980. Experimental Determination of Stone Tool Uses: A Microwear Analysis. University ofChicago Press, Chicago, 212 p.KEELEY, L. H. 1996. War Before Civilization. Oxford University Press, New York, 245 p.KELLY, R. J. 1995. The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Smithsonian Institution Press,Washington, D.C., 446 p.KLEIN, R. G. 1994. Southern Africa before the Iron Age, p. 471–519. In R. S. Corruccini and R. L. Ciochon (eds.),Integrative Paths to the Past: Paleoanthropological Advances in Honor of F. Clark Howell. Prentice-Hall,Englewood Cliffs, 716 p.KLEIN, R. G. 1999. The Human Career. University of Chicago Press, Chicago, 810 p.KLEIN, R. G. 2000. Archeology and the evolution of human behavior. Evolutionary Anthropology, 9(1):17–36.KLEIN, R. G., AND K. CRUZ-URIBE. 1984. The Analysis of Animal Bones From Archaeological Sites. University ofChicago Press, Chicago, 266 p.LAUGHLIN, W. S. 1968. Hunting: An integrating biobehavior system and its evolutionary importance, p. 304–320.In R. B. Lee and I. DeVore (eds.), Man the Hunter. Aldine Publishing Company, Chicago, 415 p.LEE, R. B. 1968. What hunters do for a living, or, How to make out on scarce resources, p. 30–48. In R. B. Lee andI. DeVore (eds.), Man the Hunter. Aldine Publishing Company, Chicago, 415 p.LEE, R. B., AND I. DEVORE. 1968. Man the Hunter. Aldine Publishing Company, Chicago, 415 p.LESHCHINSKIY, S. V. 2001. The late Pleistocene beast solonetz of western Siberia: “Mineral oases” in mammothmigration paths, foci of the Palaeolithic man’s activity, p. 293–298. In G. Cavarretta, P. Gioia, M. Mussi, andM. R. Palombo (eds.), The World of Elephants: Proceedings of the 1 st International Congress. ConsiglioNazionale delle Richerche—Roma, Rome (Italy), 739 p.65

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002LUPO, K. D. 1994. Butchering marks and carcass acquisition strategies: Distinguishing hunting from scavengingin archaeological contexts. Journal of Archaeological Science, 21:827–837.LUPO, K. D. 2001. Archaeological skeletal part profiles and differential transport: An ethnoarchaeological examplefrom Hadza bone assemblages. Journal of Anthropological Archaeology, 20:361–378.LYMAN, R. L. 1985. Bone frequencies, differential transport, in situ destruction and the MGUI. Journal ofArchaeological Science, 12:221–236.LYMAN, R. L. 1994. Vertebrate Taphonomy. Cambridge University Press, Cambridge (U.K.), 524 p.LYMAN, R. L., AND G. L. FOX. 1989. A critical evaluation of bone weathering as an indication of bone assemblageformation. Journal of Archaeological Science, 16:293–317.METCALFE, D., AND K. T. JONES. 1988. A reconsideration of animal body-part utility indices. American Antiquity,53:486–504.MILNER, G., E. ANDERSON, AND V. SMITH. 1991. Warfare in Late Prehistoric west-central Illinois. American Antiquity,56:581–603.NIVEN, L. 2001. The role of mammoths in Upper Paleolithic economies of southern Germany, p. 323–327. In G.Cavarretta, P. Gioia, M. Mussi, and M. R. Palombo (eds.), The World of Elephants: Proceedings of the 1 stInternational Congress. Consiglio Nazionale delle Richerche—Roma, Rome (Italy), 739 p.O’CONNELL, J. F., K. HAWKES, AND N. BLURTON-JONES. 1988a. Hadza hunting, butchering, and bone transport andtheir archaeological implications. Journal of Anthropological Research, 44:113–161.O’CONNELL, J. F., K. HAWKES, AND N. BLURTON-JONES. 1988b. Hadza scavenging: Implications for Plio-Pleistocenehominid subsistence. Current Anthropology, 29:356–353.OUTRAM, A., AND P. ROWLEY-CONWY. 1998. Meat and marrow utility indices for horse (Equus). Journal ofArchaeological Science, 25:839–849.PATOU-MATHIS, M. 2000. Neanderthal subsistence behaviours in Europe. International Journal of Osteoarchaeology,10:379–395.PÉAN, S. 2001. Mammoth and subsistence practices during the Mid Upper Palaeolithic of Central Europe (Moravia, CzechRepublic), p. 331–336. In G. Cavarretta, P. Gioia, M. Mussi, and M. R. Palombo (eds.), The World of Elephants:Proceedings of the 1 st International Congress. Consiglio Nazionale delle Richerche—Roma, Rome (Italy), 739 p.PETER-RÖCHER, H. 1994. Kannibalismus in der prähistorischen Forschung. Studien zu einer paradigmatischenDeutung und ihren Grundlagen. Universitätsforschung Prähistorisch Archäologie, 20:1–265.PETERSON, J. 1998. The Natufian hunting conundrum: spears, atlatls, or bows? Musculoskeletal and armatureevidence. International Journal of Osteoarchaeology, 8:378–389.REHER, C. A., AND G. C. FRISON. 1980. The Vore Site, 48CK302, a Stratified Buffalo Jump in the Wyoming BlackHills. Plains Anthropologist Memoir, 16:1–190.RICHARDS, M. P., R. E. M. HEDGES, R. JACOBI, A. CURRANT, AND C. STRINGER. 2000. FOCUS: Gough’s Cave andSun Hole Cave human stable isotopes indicate a high animal protein diet in the British Upper Palaeolithic.Journal of Archaeological Science, 27:1–3.DE RUITER, D. J., AND L. R. BERGER. 2000. Leopards as taphonomic agents in dolomitic caves—implications for boneaccumulations in the hominid-bearing deposits of South Africa. Journal of Archaeological Science, 27:665–684.RUSSELL, M. 1986. Mortuary practices at the Krapina Neandertal site. American Journal of Physical Anthropology,72:381–397.SAMPSON, C. G. 2000. Taphonomy of tortoises deposited by birds and Bushmen. Quaternary Research, 27:779–788.SAUNDERS, J. J. 1980. A model for man-mammoth relationships in late Pleistocene North America, p. 87–98. In N.W. Rutter and C. E. Schweger (eds.), The Ice-Free Corridor and the Peopling of the New World. CanadianJournal of Anthropology 1, 139 p.SAUNDERS, J. J., AND E. B. DAESCHLER. 1994. Descriptive analyses and taphonomical observations of culturallymodifiedmammoths excavated at “The Gravel Pit,” near Clovis, New Mexico in 1936. Proceedings of theAcademy of Natural Sciences of Philadelphia, 145:1–28.SCOTT, K. 1980. Two hunting episodes of Middle Paleolithic age at La Cotte de Sainte-Brelade, Jersey (ChannelIslands). World Archaeology, 12(2):137–152.SCOTT, K. 1986. The bone assemblages of layers 3 and 6, p. 159–183. In P. Callow and J. M. Cornford (eds.), LaCotte de St. Brelade, 1961–1978: Excavations by C. B. M. McBurney. Geo Books, Norwich (U.K.), 433 p.66

HAYNES—RECONSTRUCTING HUMAN PREDATIONSELVAGGIO, M. M. 1994. Evidence From Carnivore Tooth Marks and Stone-Tool-Butchery Marks For Scavengingby Hominids at FLK Zinjanthropus Olduvai Gorge, Tanzania. Unpublished Ph.D. dissertation in Anthropology,Rutgers, The State University of New Jersey, 388 p.SELVAGGIO, M. M., AND J. WILDER. 2001. Identifying the involvement of multiple carnivore taxa with archaeologicalbone assemblages. Journal of Archaeological Science, 28:465–470.SHIPMAN, P. 1981. Life History of a Fossil: An Introduction to Taphonomy and Paleoecology. Harvard UniversityPress, Cambridge (MA), 222 p.SHIPMAN, P., AND J. ROSE. 1983. Early hominid hunting, butchering, and carcass-processing behavior: Approachesto the fossil record. Journal of Anthropological Archaeology, 2:57–98.SOFFER, O. A. 1985. The Upper Paleolithic of the Central Russian Plain. Academic Press, New York, 539 p.SPETH, J. D. 1990. Seasonality, resource stress, and food sharing in so-called “egalitarian” foraging societies.Journal of Anthropological Archaeology, 9:148–188.SPETH, J. D., AND K. A. SPIELMANN. 1983. Energy source, protein metabolism, and hunter-gatherer subsistencestrategies. Journal of Anthropological Archaeology, 2:1–31.SPINDLER, K. 1995. The Man in the Ice. Orion Paperbacks, London, 305 p.Stanford, C. B., and H. T. Bunn (eds.). 2001. Meat-Eating and Human Evolution. Oxford University Press,Oxford (U.K.), 432 p.STEPHENS, D. W., AND J. R. KREBS. 1986. Foraging Theory. Princeton University Press, Princeton, 247 p.STINER, M. C. 1990. The use of mortality patterns in archaeological studies of hominid predatory adaptations.Journal of Anthropological Archaeology, 9:305–351.STINER, M. C. (ed.). 1991. Human Predators and Prey Mortality. Westview Press, Boulder, 276 p.THIEME, H. 1997. Lower Paleolithic hunting spears from Germany. Nature, 385:807–810.TODD, L. C. 1991. Seasonality studies and Paleoindian subsistence strategies, p. 217–238. In M. C. Stiner (ed.),Human Predators and Prey Mortality. Westview Press, Boulder, 276 p.TONG, H. 2001. Age profiles of rhino fauna from the middle Pleistocene Nanjing Man site, south China—explainedby the rhino specimens of living species. International Journal of Osteoarchaeology, 11:231–237.TRINKAUS, E. 1985. Cannibalism and burial at Krapina. Journal of Human Evolution, 14:203–216.TURNER, C., AND J. TURNER. 1999. Man Corn: Cannibalism and Violence in the Prehistoric American Southwest.University of Utah Press, Salt Lake City, 547 p.VASIL’EV, S. A. 2001. Man and mammoth in Pleistocene Siberia, p. 363–366. In G. Cavarretta, P. Gioia, M.Mussi, and M. R. Palombo (eds.), The World of Elephants: Proceedings of the 1 st International Congress.Consiglio Nazionale delle Richerche—Roma, Rome (Italy), 739 p.VILLA, P. 1992. Cannibalism in prehistoric Europe. Evolutionary Anthropology, 1(3):93–104.WALKER, P. L. 2001. A Bioarchaeological Perspective on the History of Violence. Annual Review of Anthropology,30:573–596.WHEAT, J. B. 1972. The Olsen-Chubbuck Site: A Paleo-Indian Bison Kill. Society for American ArchaeologyMemoir, 26:1–180.WHITE, T. D. 1992. Prehistoric Cannibalism at Mancos 5MTUMR-2346. Princeton University Press, Princeton, 492 p.WOODBURN, J. 1968. An introduction to Hadza ecology, p. 49–55. In R. B. Lee and I. DeVore (eds.), Man theHunter. Aldine Publishing Company, Chicago, 415 p.67


LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSTHE TROPHIC ROLE OF MARINE MICROORGANISMSTHROUGH TIMEJERE H. LIPPS 1 AND STEPHEN J. CULVER 21Department of Integrative Biology and Museum of Paleontology, University of California,Berkeley, California 94720-3140 USA2Department of Geology, East Carolina University, Greenville, North Carolina 27858-4353 USAABSTRACT—Microorganisms (prokaryotes and protists) seldom fossilize, but they form much of the trophicstructure in marine pelagic and benthic environments, chiefly as primary producers and secondary consumers.The fossil record of unskeletonized groups is meager or non-existent. Skeletonized groups have excellent recordsbut represent a small portion of the total microbial diversity.The evolution of trophic structures and roles of microorganisms can be reconstructed broadly for most ofgeologic history. When life first evolved, it had a trophic structure. The first microbial fossils appear to bebenthic mats; these are abundant in the Precambrian but sparse later; body fossils are very rare. The Archeansaw pelagic and benthic prokaryotes and possibly protists later on. Proterozoic trophic structures becameincreasingly complex as protists entered pelagic environments. Benthic assemblages likewise became complex,as prokaryotes and protists formed mats and stromatolites in many environments. At the end of the eon, animalsappeared; microbial primary producers and predation on microorganisms and among animals fueled theseassemblages. The fundamental trophic structures that developed then persisted with modification into moderntimes. Phanerozoic ecosystems became very complex as skeletonized animals and protists evolved. Among theimportant trophic developments in the Phanerozoic history of microorganisms were the early diversification ofphytoplankton and siliceous micro-zooplankton (Cambrian), algal endosymbiosis with benthic metazoans(Cambrian to Recent) and rock-forming foraminifera (late Paleozoic to Recent), the radiation of pelagic skeletalprimary producers and micro-zooplankton (mid-Mesozoic), and radiations in the deep sea, reefs, and shallowareas (Mesozoic and Cenozoic). Each evolutionary change increased trophic complexity by adding more speciesat each level, while episodic mass extinctions decreased species diversity and trophic complexity.Marine trophic structures evolved over immense intervals of geologic time, growing complex and then sufferingdestruction at major extinction events. The effects of human impact on these structures should be examined, forwithout them, Earth may change dramatically.INTRODUCTIONORGANISMS NEED ENERGY to grow andreproduce. Thus, a fundamental property of anyecosystem is its trophic structure—the way in whichenergy, as nutrients and food, is used and distributedthroughout it. The first living organisms, no matterhow simple, had a trophic structure. Microorganismshave always dominated marine trophic structures;indeed for the first 3 billion years of the geologicrecord, ecosystems consisted wholly of them, andlater when larger animals and plants appeared,microorganisms still provided the bulk of primaryproduction and consumption. For the vast majorityof microorganisms, estimated at millions ofunknown or poorly known species of unskeletonizedprokaryotes (Torsvik et al., 2002) and protists(Patterson, 1999), no fossil record exists at all. Yetmicrofossils and microbial traces are the only directevidence in the record of the most important aspectsof trophic structures—primary production and itsprocessing up to larger and more complex organismsthrough predation (defined broadly as theconsumption of any other organism).Microfossils represent a vast array oforganisms, from prokaryotes to parts of largermetazoa and metaphyta (Table 1). They are unitedas “microfossils” chiefly because of their small size69

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002and geological utility, but they have little biologicalunity, either ecologically, evolutionarily, ortrophically. In essence, microfossils includerepresentatives of every known group of organismsfrom all environments. Thus, we restrict this paperto marine single-celled eukaryote (referred to hereas protists) and prokaryote (Bacteria and Archaea)fossils or their traces. Their systematic range stillencompasses all domains of life, and all areimportant in marine trophic structures.Not only are microorganisms diverse in form,but also for much of geologic time they probablyoccupied every habitat in the seas from the deepestoceans to the most variable marginal environment,and all levels in the water column. The range is soenormous even for single groups that a simplesummary of all trophic relationships is impossible.The living prokaryotes and protists with fossilrecords exhibit a multitude of complex trophic styles(Lipps and Valentine, 1970; Azam, 1998); hencetheir roles in ancient ecosystems were likelysimilarly varied (Fig. 1). Some inferences arepossible based on the biological properties of livinggroups of related organisms and the relationship ofthe microfossils to other fossils in co-occurringassemblages. Little direct evidence, however, existsfor the trophic relationships of these tiny fossils. Verylittle about predation either on these microorganismsor by them on other organisms is reliably knownfor long intervals of geologic time. This is true evenfor well-fossilized groups like foraminifera (Lipps,1983; Culver and Lipps, in press).Prokaryote fossils include rare preserved cellsthemselves, and indirect evidence of them from thefossil microbial mats and stromatolites they built.Fossil pelagic prokaryotes are rare. Protists includebenthic and planktic primary producers; microherbivoresthat fed on the producers; and microcarnivoresthat ate the herbivores, other microcarnivores,and smaller metazoans (Table 1). Theirindividual roles also may have varied opportunisticallyeven during their life spans (Lipps, 1983).Through geological time, the relative importanceof these groups undoubtedly changed as marineecosystems evolved. The chief protists involvedinclude the enigmatic but certainly primaryproducing,early-appearing acritarchs (Vidal andMoczydlowska-Vidal, 1997), the primary producersmostly of the last half of the Phanerozoic (thecalcareous nannoplankton, dinoflagellates,diatoms, and silicoflagellates), and the skeletonizedheterotrophic protists (thecamoebians, radiolarians,tintinnids, and benthic and planktic foraminifera).As in today’s seas, prokaryotes, naked protists, andunfossilizable larger zooplankton were likelypresent (see Fig. 1) and important, but largelyunrecorded in the fossil record.The purpose of this paper is to provide ageneral review of the evolution of trophicrelationships of microorganisms through time. Oursummary of current data is designed for universitypaleontology and micropaleontology courses.Additional information is in the references cited.MODERN MARINE TROPHICRELATIONSHIPSMarine trophic relationships are oftendiagrammed as food chains, but these are simplifiedand incorrect for the most part, as the differentorganisms involved are commonly varied andopportunistic in their trophic roles. Theserelationships are shown best as a web of species orassociated groups rather than as a chain (Fig.1).Marine food webs are based mostly on singlecelledprimary producers, many of which rarelyfossilize. Further, these photosynthesizingmicrobes are still incompletely known in modernoceans (Azam, 1998); hence complete inferencesfor ancient oceans cannot be made either. Today,about half of Earth’s total net primary productiontakes place in oceanic environments (Azam, 1998).This value must have been higher before theevolution (early Paleozoic) of metaphytes or evenparticular plant groups, like angiosperms(Cretaceous), on land. Thus, the trophic relationshipswe deal with here were probably even moresignificant in the past than now. Although differentspecies were involved at different times in the distantpast, the fundamental transfer of energy throughprimary producers to pelagic and benthic herbivoresthen on to carnivores, omnivores, and scavengers70

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSTABLE 1—The chief fossilizable prokaryotes and protists and their trophic functions and geologic age.Modified from Lipps (1993).Bacteria or Protists Habitat Function Age RangeHeterotrophic bacteria All Consumers Archean–RecentAutotrophic bacteria,especially cyanobacteriaAll Primary producers Paleoproterozoic–RecentAcritarchsChiefly onshore/offshore pelagicPrimary ProducersMesoproterozoic–RecentDinoflagellatesOceanic, nearshore,lacustrinePrimary ProducersPaleoproterozoic?–RecentCoccolithophorids andrelated calcareous algaeOceanic, eutrophicto oligotrophicPrimary ProducersTriassic–RecentSilicoflagellates Oceanic Primary Producers;secondary consumers?Cretaceous–RecentDiatomsOceanic, chieflyeutrophic; lacustrinePrimary ProducersCretaceous–RecentForaminiferaBenthicPelagicHerbivorous,omnivorous,carnivorousCambrian–Recent.Jurassic–Recent forpelagic forms.Radiolarians Oceanic Herbivorous, carnivorous Cambrian–RecentTintinnids Neritic, oceanic Herbivorous, carnivorous Ordovician–Recentmust have been somewhat similar in function.Pelagic Food Webs.—By most definitions, thepelagic environment comprises the water columnof the open oceans seaward of the continental shelf.Organisms present there may be planktonic (passivedrifters that do not swim very much, at leasthorizontally) or nektonic (swimming forms). Theseorganisms range in size from the tiniest picoplanktonthrough nano- and microplankton to the relativelylarge zooplankton on which huge vertebrates graze(Table 2). At all size ranges, the herbivores,carnivores, and other consumers are supportedultimately by mostly tiny primary producers.Prokaryotes, commonly constituting the largestbiomass in the water column (Christaki et al.,2001), are supplied with food and nutrients by largeamounts of dissolved organic matter (DOM) andparticulate organic matter (POM) derived fromnumerous sources, as well as viruses, protists, andother living organisms. They are heterogeneouslydistributed even within a single water sample, withsome habitats within the water particularly rich inthem; for example, marine snow, detritus, andphytoplankton find abundant bacteria attackingthem. Bacteria constitute the major pathway foroceanic primary production (about 50%), whichtakes the form of the so-called bacterial loop(Fig.1) (Pomeroy , 1974; Azam, 1998). This loopis variable and dynamic in its functions (Falkowskiet al., 1998; Riemann et al., 2000; Shinada et al.,2001; and it is largely self-sustaining andindependent of larger organisms.Two other pathways exist in pelagic systems:the “grazing food web” and the “sinking flux”71

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 1—Generalized pelagic and benthic Recent marine trophic structures (Modified from Lippsand Valentine, 1970, and Azam, 1988). DOM and POM are dissolved and particulate organic matter.72

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMS(Fig. 1). In the grazing component, larger primaryproducers (microphytoplankton) are taken byherbivorous zooplankton, which are in turn eatenby larger carnivores ranging from small crustaceansand other invertebrates to fish and whales. Thisstandard view of trophic resources in the sea ismisleading and does not apply to most of Earth’shistory. Another part of the food web sinks to thesea floor where it becomes available for benthicorganisms. Fecal pellets, aggregations of otherorganisms into marine snow, and the dead and evenlive bodies of organisms comprise the sinking flux.This flux removes a variable amount of carbonfrom the water column.The bacterial loop dominates over the grazingfood web in the oligotrophic parts of the oceans.The grazing web prevails in eutrophic situations,geographically and seasonally (Kiorboe, 1993;Shinada et al., 2001), but even here the bacterialloop is significant. In eutrophic areas, phytoplanktonare larger, resulting in a web with fewer levels orlinks. Because smaller phytoplankton are stillpresent, more complex webs coexist.Benthic Food Webs.—The benthic environmentsof the world’s oceans are far more diverseand heterogeneous than pelagic ones. This isreflected in the species diversity of groups withboth benthic and pelagic members. For example,among Recent foraminifera, well over 4000 benthicspecies are known but only 42 or so species areplanktic. In prokaryotes, where species have noteven been described, pristine marine sedimentcontains far more genomic diversity than the watercolumn (Torsvik et al., 2002). The benthic habitatsfor microorganisms include all known marineenvironments (see Nybakken, 2001, for summary).Perhaps the most significant habitats through timeand even today are microbial mats (Figs. 2–4, 6).While these mats have changed over the past 3.5billion years, they still appear very similar acrossthe ages. The other major environments of theworld’s oceans have undergone significant andconstant oceanographic modifications through timeas continental positions and climates have changed.At any period in the history of the oceans,thousands of distinctive habitats and individualtrophic structures existed. Today, because of theprominent thermal gradients from north to south(Buzas et al., 2002; Crame, 2002) and shallow todeep, the latitudinal array of the continents, andthe evolutionary complexity and diversity oforganisms, habitat heterogeneity is perhaps at leastas great as at any other time in Earth history.MICROORGANISMS AND THEEVOLUTION OF MARINETROPHIC STRUCTURESNothing, of course, is known of the earliesttimes on Earth from its formation approximately4.55 Ga to about 3.8 Ga, other than what can beinferred from extraterrestrial bodies that fell toTABLE 2—Marine pelagic organisms. Primary producers indicated with an asterisk.Groups Size (microns) Main Taxa Age RangeNekton 2×10 3 –3.3×10 7 Invertebrates, fish, reptiles, mammals Cambrian–RecentZooplankton 2×10 3 –2×10 3 Invertebrates, vertebrate larvae Cambrian–RecentMicroplankton 20–2×10 2 Acritarchs*, dinoflagellates*, diatoms*,silicoflagellates*, radiolarians,foraminifera, tintinnidsProterozoic; chieflyMesozoic–RecentNanoplankton 2–20 Coccolithophorids* Mesozoic–RecentPicoplankton 0.2–2 Prokaryotes*, cyanobacteria* Archean–Recent73

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Earth, and from a few crystals incorporated intolater-formed rocks. Based on crater counts onnearby bodies in the Solar System, Earth wassubjected to intense bombardment by bolides inits early history. The impacts were sufficient toimpede the accumulation of water or oceans andthe sustained development of life (Nisbet andSleep, 2001). After half a billion years or so, thebombardment slowed and decreased in intensityas the bolides became smaller and fewer. Thesecomets, asteroids, and meteorites likely brought tothe Earth’s surface considerable amounts of organicmaterial—sufficient for life to start (Anders, 1989;Chyba et al., 1990). The organic molecules servedas the foundation of life; and once life started, thosecompounds, and the continued synthesis of carbonmolecules in the waters of Earth, served as nutrientsfor the newly evolving prokaryotes. Crystals inyounger rocks provide evidence for surface waterat least by 4.3 Ga (Mojzsis et al., 2001). Thus, lifehad what was needed to form and survive very earlyin Earth history: the proper chemicals, a source ofenergy, liquid water, and a trophic resource. Alongwith the first life came the first microbial trophicinteractions.ARCHEAN: 4.0 TO 2.5 GAArchean metasedimentary rocks make up asmall percentage of exposures on the continentstoday, and the oceans of most of this eon areessentially unknown (Lowe, 1992). Life, too, ispoorly known. Thick exposures at Isua, Greenland,dated at about 3.8 Ga, contain carbon isotopicsignatures interpreted as having been fractionatedby biologic activity (Rosing, 1999), although atleast some of that signal may be of metasomaticorigin (Fedo and Whitehouse, 2002; see alsoSchidlowski, 2001). No definitive fossils areassociated with these deposits. Prokaryotic fossilswere described in rocks from about 3.5 Ga (Schopfand Packer, 1987; Schopf, 1993), although thesehave been questioned recently (Brasier et al., 2002).In spite of these fossils’ status, life’s presence isindicated by isotopic evidence (Rosing, 1999) andby other fossil occurrences at about this same timeor slightly later (see Schopf, 1992a for summary).Organisms present in the Archean werebacteria, probably Archaea, and possibly earlyeukaryotes (Brock et al., 1999; Summons et al.,1999), likely both benthic and planktic (Table 3,4).No evidence directly links protists to these earlysystems, but they could have been present—wesimply have no way to tell, other than biomarkersof uncertain reliability. Some molecularphylogenies suggest an early divergence ofeukaryotes (Gu, 1997; Hedges et al., 2001), butthis is an indirect conclusion subject to alternativeinterpretations (Katz, 1999).Benthic trophic structures were likely confinedto prokaryote aggregations, films and mats, someof which made stromatolites and laminatedsedimentary rocks (Simonson and Carney, 1999;Nisbet and Sleep, 2001). A prokaryotic trophicstructure also probably operated in open waters.Both benthic and pelagic trophic structures wouldhave been much simpler than later structures (Table3, 4), as they were surely dominated by prokaryoteswithout larger or more complex organisms. Theorigins of these prokaryotic trophic structures mayhave been benthic, associated with hydrothermalvents, because molecular phylogenies indicate thatthermophilic bacteria and archaeans are basal toother extant groups (Reysenbach and Shock, 2002).This scenario has been seriously questioned onseveral grounds (Bada and Lazcano, 2002; Brocierand Philippe, 2002), and origins and evolution incooler waters were also possible. The basal positionof thermophilic bacteria in the current phylogeniesdoes not necessarily indicate origins, but, if correct,indicates only that the last common ancestor ofliving bacterial lineages was thermophilic. Suchan interpretation excludes all fossil and livinglineages yet to be discovered or studied—far morework needs to be done on living and fossilindicators of prokaryotic and protistan presencebefore confident reconstructions can be made. Yetthe earliest ecosystems probably did not haveparticularly unusual kinds of trophic interactions,although they likely were limited and lackedoxygenic photoautotrophs (Nisbet and Sleep,2001). These benthic and pelagic ecosystems were74

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSalso probably widely distributed in Earth’s seas.Pelagic realms of the Archean must have ratherquickly evolved a prokaryotic biota (Table 4).Prokaryotic trophic interactions would have beenless complex than modern ones, chiefly becauselarger plankton and sources of organic debris wereabsent, and diversity was probably lower.Nevertheless, this pelagic system would have beenfueled by organic matter from the prokaryotes,from other ecosystems elsewhere on Earth, fromviruses, from co-occurring species, and from thecontinued infall of meteorites. The system wouldhave operated quite independently of the benthos,although the export of parts of it to the benthosmust have taken place and DOM and POM fromthe benthos may have been mixed into the openwaters (Fig. 1).After their initial evolution, prokaryotes likelyradiated into most benthic marine habitats. Therecord reveals only sparse prokaryotic fossils andfilms, mats, and stromatolites, known from severallocations, which represent shallow-watercommunities. These, however, cannot be fullyrepresentative of the range of microbial habitats inthe Archean oceans because they represent so fewof the possible Archean environments and the knownhabitats that microbes occupied later in their history.Evidence of predation in these communities is absenttoo, although surely the producers were harvestedby other prokaryotes and possibly protists, eitherdirectly or as decomposed matter.PROTEROZOIC: 2.5 TO 0.545 GADuring the two billion years of the Proterozoic,trophic structures evolved from simple typesdominated by prokaryotes to those that were theimmediate precursors of modern marine types(Table 3, 4), which included large metazoans andabundant larger primary producers in the benthosand a poorly preserved but likely complex groupof plankton. Trophic relationships changed throughtime as newly evolved kinds of organismsinterposed themselves into previously existingcommunities. The fossil record throughout thePaleoproterozoic and Mesoproterozoic is stillrelatively unknown, although there is enough fossilevidence to indicate that protists had becomeimportant parts of the benthic and pelagicecosystems (Vidal and Moczydlowska-Vidal, 1997;Javaux et al., 2001). In the Neoproterozoic, fossilspermit reasonable reconstructions of trophicinteractions.Global events in the Proterozoic may have hada significant impact on marine ecosystems. Theseevents include a poorly known but extensive glacialperiod about 2.4 Ga, the so-called Snowball Earth I(Kirschvink et al., 2000). Oxygen became plentifulin the atmosphere and oceans between 2.2 and1.8Ga, and this may have spurred protists todiversify and occupy more habitats. Fossil protistsare present in the record starting about 1.8 Ga (Vidaland Moczydlowska-Vidal, 1997) and indicaterelatively complicated ecologic relationships. Thefirst protistan groups known later in the record alsoappear during this eon (Javaux et al., 2001). Mostsignificantly, a second set of Snowball Earthglaciations took place about 750 to 600 Ma.Snowball Earth II ended just before the appearanceof metazoans in the most complicated ecosystem ofthe first seven-eighths of Earth history.Paleoproterozoic: 2.5 to 1.6 Ga.—During theearly Paleoproterozoic, cyanobacteria likelyappeared and produced free oxygen by way ofphotosynthesis. Thus, for the first time in Earthhistory, at about 2.2 Ga, the hydrosphere andatmosphere became significantly oxygenated.Acritarchs, the first protist fossils, occur in rocksas old as 1.8 Ga (Vidal and Moczydlowska-Vidal,1997). Most acritarchs were cysts of planktic algaealthough some may have been prokaryotes orbenthic protists (Butterfield, 1997). Clearly thepelagic ecosystem changed drastically from onemade up almost solely of prokaryotes to one thatincluded oxygenic-photosynthesizing nano- tomicrophytoplankton (Table 4). Although no fossilsare known, microzooplankton was likely presentas well, as unskeletonized protists that preyed onboth prokaryotes and phytoplankton (Christaki etal., 2001). Ciliates, flagellates, and sarcodines ortheir ancestors are possible candidates for this role.Benthic communities of prokaryotes werecommon in shallow waters and probably elsewhere75

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 3—Increasing complexity in benthic trophic structures from the Archean to Recent.ArcheanPaleoproterozoic;MesoproterozoicNeoproterozoic Paleozoic Mesozoic;CenozoicCarnivoresCarnivoresCarnivoresPhotoendosymbiosisPhotoendosymbiosisHerbivores Herbivores HerbivoresProtist herbivores Detritovores Detritovores DetritovoresFleshy algae Fleshy algae Large algae Large algaeAlgal protists in matsAlgal protistsin matsAlgal protists onsubstrataAlgal protists(attached diatoms)AnoxygenicprokaryotematsProkaryotes includingcyanobacteria matsProkaryotesincludingcyanobacteriaProkaryotesincludingcyanobacteriaProkaryotesincludingcyanobacteriaTABLE 4—Increasing complexity of pelagic trophic structures from the Archean to Recent.ArcheanPaleoproterozoic;MesoproterozoicNeoproterozoic Paleozoic Mesozoic;CenozoicMega-nektonNektonNektonZooplankton Zooplankton ZooplanktonProtist herbivores Microherbivores Microherbivores MicroherbivoresProtist phytoplankton Protist phytoplankton ProtistphytoplanktonProtistphytoplanktonProkaryotesProkaryotes includingcyanobacteriaProkaryotes includingcyanobacteriaProkaryotesincludingcyanobacteriaProkaryotesincludingcyanobacteria(Schopf, 1992b). They built laminated mats, whichstabilized the substrate, and stromatolites domingupwards. Some of these contain body fossils ofbacteria, including some interpreted ascyanobacteria. The trophic structure of such matsand stromatolites might well have resembled matsand stromatolites existing today in hypersalineenvironments, but without advanced eukaryotessuch as diatoms and metazoans (Figs. 2, 3).Mesoproterozoic: 1.6 to 1.0 Ga.—Pelagictrophic structures seem closely similar to those ofthe Paleoproterozoic as far as fossils indicate, but76

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSboth protist and prokaryote diversity likelyincreased without leaving a record. The trophicroles of these new additions were probably similarto the preexisting types. Benthic ecosystems alsodiversified and became more complex, withdiversity and abundance greater in shallow- thanin deep-water settings (Javaux et al., 2001).Microbial diversity continued to increase acrossthe boundary into the Neoproterozoic (Sereev, etal., 1997). A single trace fossil occurrence mayrepresent metazoan-like organisms moving on mudmore than 1.2 Ga (Rasmussen et al., 2002).Neoproterozoic: 1.0 to 0.545 Ga.—During thefirst half of the Neoproterozoic, benthic and pelagictrophic structures were much like those of theearlier Proterozoic. Although protists continued todiversify and build complex ecosystems in bothbenthic and pelagic environments, the totaldiversity remained low compared to later times(Knoll, 1994). Small, early multicellular stemmetazoans may have been widespread, butevidence is lacking (Lipps et al., 1992). EarlyNeoproterozoic pelagic trophic structures(Tables3, 4) surely included a bacterial loop, whichby then encompassed phytoplankton, their organicproducts, possible herbivorous protists, carnivores,and consumers. In the benthos, mats andstromatolites were abundant.Snowball Earth II, proposed to account forglacial evidence in low latitudes between 750 and600Ma (Harland, 1964; Kirschvink, 1992;Hoffman et al., 1998), comprised several glacialperiods, each of which would have had significantimpact on pelagic and benthic ecosystems and theirtrophic structures, depending on the geographicextent of the glaciations. Snowball Earthpossibilities include complete, global glaciationsand paving of the oceans with ice, partialglaciations reaching low latitudes but with an openequatorial ocean, and glaciations restricted tomedium-to-high latitudes. The fossil record ofplankton continues through this time interval withvariations in diversity but evidently with littleextinction (Vidal and Moczydlowska-Vidal, 1997).FIGURE 2—Modern stromatolites, Cargla Point, Shark Bay, Australia. The stromatolites (50–60 cmhigh here) grow in hypersaline waters that exclude herbivorous invertebrates and fish from the area.Other algal-rich forms are intermingled with the stromatolites including microbial mats (foreground),oolites, and floating clumps of bacteria. (Image by J. H. Lipps and S. Culver, Feburary, 2002.)77

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002ABFIGURE 3—Modern microbial mats. A, A temperatemat flooring an ephemeral lake surrounded bySpartina and Juncus vegetation, Pea Island, NorthCarolina, USA. The mat is generally continuouswhen submerged, but becomes shrunken, furledand sometimes overturned upon drying. (Photo byS. J. Culver.) B, A tropical mat consisting ofcyanobacteria, diatoms, and other microbes builton a tidal flat at Tamae, Moorea, French Polynesia.The flat is in a tidal inlet with salinity usually rangingfrom 38–68 o / oo. These mats are thin and variable insurface texture and topography. Some are furled(as in the image), rolled-up, flat, or disrupted. (Photoby Shelene Poetker.) Measuring tape in cm.With the oceans covered with ice, primaryproduction would have been severely restricted,although phytoplankton would not necessarily havebecome extinct. The continuous record ofacritarchs can be interpreted under the ice-coveredoceans scenario as the survival of these forms inareas where the ice was cracked and fissured bytidal and current action, in melt water pools on ornear the surface of the ice, near and below holes inthe ice cover, and in regions where ice could notform because of volcanic activity. In either of theother two scenarios, primary production might havebeen constrained to lower latitudes, but it couldstill have been substantial and supportive ofcomplex trophic interactions. Benthic ecosystems,of course, would also have been affected differentlyunder each scenario because most deeper watersystems are dependent on surface productivity. Butthere is currently no record of these biotas.Shallow-water mats and stromatolites are presentin many places around the globe, presumably fromshallow areas unaffected by ice. As the ice agesended, thick deposits of laminated, algal-dominatedcarbonates were deposited. None of these algalassociations appear to have been grazed.Trophic relationships changed fundamentallyjust after 600 Ma when large animals, the Ediacarabiota, first appeared (Narbonne, 1998). This biotahas been found on all continents except Antarcticaand includes over 250 described species of softbodiedand trace fossils (Runnegar, 1992a; 1992b).Sometimes referred to as “an experiment,” theseorganisms were hardly a trial or test, for they livedfor over 25 million years, were fully functional,were quite diverse, comprised precursors to thelater metazoan record, and would be consideredvery successful had they lived later (Runnegar,1992c). Originally they were described mainly asanimals (Glaessner, 1984), but were later consideredto be other kinds of organisms based on alternativemorphologic interpretations and the perceivedtaphonomic difficulties of preserving soft bodies(Seilacher, 1989; Buss and Seilacher, 1994).Suggested alternatives were xenophyophorans, anagglutinated group of protists (Zhuravlev, 1993),lichens (Retallack, 1994), or extinct sister groups78

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSFIGURE 4—Microbial mat texture (“elephant skin”) preserved on the surface of a bedding plane. Vendian,White Sea, Russia. (Photo by UC Photo Services of a UCMP specimen.) Scale in cm.of the metazoans (Buss and Seilacher, 1994). Thesehypotheses have been refuted (Fedonkin, 1992;Jenkins, 1992; Runnegar, 1992c; Waggoner, 1995;Fedonkin and Waggoner, 1997; Knoll and Carroll,1999; Collins et al., 2000; Martin et al., 2000;Ivantsov and Fedonkin, 2001) and the fossils arenow considered to be mostly metazoan, either asmembers of a few extant lineages or as stem groups(Dewel et al., 2001; Valentine, 2002).More important to the trophic analysis here,the Ediacara biota has been called a “Garden ofEdiacara” (McMenamin, 1986; Seilacher, 1999) or“Peaceful Kingdom” (Monastersky, 1998) becausepredation was thought to be absent. Instead, the biotarelied on the uptake of dissolved organic matter(DOM) or photosymbiosis. While the idea may havemerit (Runnegar, 1992c), no direct evidence existsfor it. If this hypothesis were correct, thenmicroorganisms would have played little or no roleas a food resource in the rise and maintenance ofthese early animal trophic structures. However, thefossils possess tentacles, colonial polyps, digestivechannels or caecae, guts, and body volumesindicative of feeding on other organisms (Glaessner,1984; Fedonkin, 1992; Jenkins, 1992; Fedonkin,1994; Fedonkin and Waggoner, 1997; Dzik andIvantsov, 2002); and they produced traces offeeding activities (Fedonkin, 1992; Seilacher,1999). The Ediacaran animals likely consumed awide variety of food. Phytoplankton (acritarchs atleast) (Vidal and Moczydlowska-Vidal, 1997),benthic prokaryote-eukaryote mats (Fig. 4), anddetritus (Gehling, 1999; Seilacher, 1999) wereabundant and probably supported the biota at thebase of the food web, and micro- and macroorganismslikely consumed each other.Many trophic roles existed in the Ediacaranbiota. Nemiana (Fedonkin and Runnegar, 1992)and Eoporpita, for example, represent benthictentaculate feeders that probably caught prey andingested it whole. The toothed fossil Redkinia(Fedonkin, 1992) may represent holding orcatching teeth, filtering devices, sieves, or supportsfor a food collecting apparatus. Cloudina, askeletonized tube, was bored by some animal,protist, or perhaps cyanobacterium (Bengtson and79

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Zhao, 1992). The benthic biota was also tiered(Clapton and Narbonne, 2002) with frond animalsstanding above the others. In analogous modernsoft-bodied communities (Fig. 5), predation is acommon strategy (Madin, 1988; Carre et al., 1989;Mills, 1993, 1995). The Ediacaran biota includedsimilar cnidarian- and ctenophore-like animals(Gehling, 1988; Fedonkin, 1992; Jenkins, 1992;Wood et al., 2002). Ediacara was certainly nopeaceful garden!Trails and scratch marks indicate substratesearch and feeding strategies. Large Dickinsonia andYorgia were found with trails behind them (Ivantsovand Fedonkin, 2001); the presence of digestiveorgans on their undersides (Dzik and Ivantsov,2002), marks on the underlying sediment, and thesmoothing of the surface under the fossils indicatethat they groomed the substrate and ingested food.Trace fossils representing locomotion and feeding,chiefly at the sediment surface (Crimes, 1992) orwithin the mats (Seilacher, 1999), are common inEdiacaran biotas, and fecal pellets occur on bedsurfaces (Fedonkin and Runnegar, 1992), showingdigestion and excretion. Scratches on the sedimentin front of the anterior end of Kimberella wereinterpreted as feeding marks (Seilacher, 1999).The Ediacaran biota was the earliest fullyfunctioning animal community with herbivores,detritovores, and predators on both benthic andplanktic resources. They were not especiallyunusual although some major feeding types hadnot yet evolved by the Neoproterozoic, such asinfaunal bioturbation. Nevertheless, a variety offeeding strategies was present, all based onmicroorganism primary producers. These trophicstructures presaged the more complex ones inPhanerozoic and modern marine communities(Table 3, 4).What initiated the evolution of the Ediacaranbiota? The trophic developments suggest thatresources not previously available either in typeor abundance played a role, although animal andmicrobial ancestors had to have been present andresponsive to selection. The waning Snowball IIFIGURE 5—A modern soft-bodied community in Jellyfish Lake, Palau, where predation is rampanteven though some organisms may contain photoendosymbionts. Here anemones (Entacmaeamedusivora) attached to a log capture the jellyfish Mastigias (arrows) as they float into contact with thetentacles (Fautin and Fitt, 1991). No evidence of the consumed jellyfish remains as they are eatenwhole. Mastigias possesses dinoflagellate endosymbionts, but it also captures its own food. The jellyfishon the right is about 15 cm long. (Photo by J. H. Lipps.)80

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSglaciations would have altered the surface currentsof the world’s oceans, as the temperature gradientschanged and ice disappeared. Increased nutrientsupply would likely result from oceanographicreorganization, and this might have profoundlyaffected the pelagic primary-producing microorganismsand the nearshore benthic associations,just as eutrophication of oceanic waters does today.Increased productivity could account for theabundant and widespread mat development at thistime, an inferred increased phytoplanktonabundance, and changing diversity of the acritarchsas well. Energy flow, and hence trophic structures,is the only direct link between the many elementsof the Neoproterozoic biota.PHANEROZOIC: 0.545 TO 0.0 GACompared to most of the previous history oflife, Phanerozoic trophic structures were relativelyconstant throughout the entire eon (Table 3, 4),chiefly because all the phyla of animals, most ofthe other eukaryotes, and many photosynthesizingorganisms were present soon after the start of theeon (Valentine et al., 1991). Although trophicstructures varied, especially as complexityincreased through the addition of new species,skeletonization, and body plans in the Cambrian;and especially following mass extinctions, thefundamental energy flows did not change much.With regard to microfossils, six major evolutionarydevelopments of trophic importance occurred:1)evolution of benthic foraminifera and siliceousplankton in the Cambrian; 2) appearance of larger,symbiont-bearing foraminifera in late Paleozoicshallow carbonate shelf environments; 3) inferredsymbiosis between reef organisms and algalsymbionts; 4) emergence and diversification ofmajor groups of organic-walled, siliceous orcalcareous phyto- and zooplankton in the mid-Mesozoic; 5) Mesozoic radiation of deep-water, reef,and nearshore foraminifera; and 6) mass extinctionand rediversification of many species several times.In prokaryotes, mats and stromatolites weresubjected to increased grazing when bioturbatingand grazing metazoans evolved, first in the laterProterozoic and then in the early Phanerozoic(Awramik, 1971). Mats and even stromatolites haveremained abundant but are more restricted inmodern environments.Paleozoic: 545 to 248 Ma.—Cambrianecosystems differed from previous ones largelybecause new functional groups of animalsevolved. Microorganisms also developed some ofthe same features as the animals. These involvedfive important evolutionary events that relate totrophic structures: 1) radiation of skeletonizedorganisms at the start of the period; 2) appearanceof new body plans; 3) further diversification ofmicroplankton and suspension-feeding benthicanimals; 4) radiation of infaunal bioturbatinganimals; and 5) development of archaeocyathidreefs possibly facilitated by symbioses withphototrophs. While these developments were new,the fundamental trophic structures established inthe late Neoproterozoic remained intact.At the base of the Cambrian, skeletonization andornamentation evolved among protists (Culver,1994; Knoll, 1994; Lipps and Rozanov, 1996), largeralgae (Knoll and Lipps, 1993), and animals(Valentine et al., 1991). Animals radiated by way ofskeletonization, body plan proliferation throughdevelopmental gene duplications and rearrangements,and behavior (Crimes, 1992; Valentine et al.,1999). These radiations may have been in responseto changes in the marine environment—especiallyincreased primary productivity; acritarchs radiatedin concert with animals (Vidal and Moczydlowska-Vidal, 1997; Butterfield, 2001). Since phytoplanktonrespond chiefly to oceanic conditions, especiallywater mixing and nutrient supply, a change inprimary production may have initiated a cascade ofselection among other organisms in all environments,resulting in a rapid diversification. Today,benthic animals and protists respond quickly tochanges in water column primary production withchanges in metabolism, life histories, reproduction,and behavior (Erskian and Lipps, 1987; Graf, 1989;Anderson, 1993; Linke and Lutze, 1993; Altenbachet al., 1999; Schmiedl et al., 2000). The producerstoo were adapting to the selective pressures exertedon them by microzooplankton (protists, small81

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002animals, and larvae) herbivory, so that they, too,developed new strategies and morphologies. TheCambrian radiation was likely due to a number ofinteracting oceanographic, sedimentologic, andbiologic factors; but microbes were definitely andfundamentally involved.The pelagic biome continued to diversifythrough the Cambrian. In the middle and laterCambrian, acritarchs (Vidal and Moczydlowska-Vidal, 1997) and radiolarians (Dong et al., 1997;Won and Below, 1999) diversified. The radiation ofthese primary producers and herbivorous/carnivorous microplankton, as well as the presenceof graptolites and conodonts, indicates that the openocean was far more complex biotically andtrophically; and very little is known about the nonfossilizablecomponents. With a large diversity ofbenthic animals, planktic larvae were likely far morecommon in the Cambrian than in earlier times.Archaeocyathids evolved shortly after thebeginning of the Cambrian and formed reefs andbioherms in low latitudes (Rowland and Shapiro,2002). The individual specimens and reefs havecharacteristics (Cowen, 1983) indicative of thepresence of microbial photoendosymbionts. Todaysponges, a group related to archaeocyathans, harborarchaeans, bacteria (including cyanobacteria), andprotist symbionts (Osinga et al., 2001), so thatmodern analogy supports the hypothesis ofphotoendosymbiosis. Dinoflagellates, a commonphotoendosymbiont group, or their ancestors havea biomarker record extending at least to theCambrian (Moldowan and Talyzina, 1998).Archaeocyathans may have been the first animalsthat hosted such symbionts but this strategy ofphotoendosymbiosis becomes quite commonthrough the subsequent geologic record.The radiations in the Cambrian made trophicstructures far more complex. Now there wereprimary producers, nano- and microherbivores,both tiny and larger zooplanktic carnivores, andmost likely meroplankton, as indicated by anincrease in diversity and turnover rate in protists(Knoll, 1994) as well as animals. A bacterial loopwas probably well-developed at this time too. Onthe sea floor, benthic prokaryotes, protists, andanimals made trophic structures even morecomplex in terms of diversity of the various trophiccomponents. Bioturbating and surface detritalfeeders, suspension feeders, herbivores, and majorand minor predators were all present. Trophicstructures in benthic and pelagic environmentsassumed a fundamentally modern aspect, althoughfurther modification by the evolution and extinctionof taxa continued.The Ordovician was a time when protists andanimals diversified even more (Knoll, 1994). Thediversification of each group tracks that of theother, indicating that evolutionary processestransgressed organizational levels, ecologicrestrictions, biogeography, and behavior. After theCambrian, prokaryotes became less obvious in thefossil record (because grazers restricted theirdistribution), but without doubt they existed in bothpelagic and benthic environments. Among protists,benthic agglutinated foraminifera diversifiedgreatly and moved into a variety of habitats. They,like their modern counterparts, were most likelygenerally omnivorous with some taxa takingdetritus, prokaryotes, and other protists, andperhaps being fed upon intentionally by juvenileand specialized invertebrates (such as scaphopodsthat first appeared in the Ordovician) andaccidentally by non-selective detritovores andgrazers (Lipps, 1983; Culver and Lipps, in press).Evidence for these kinds of trophic interactions isdifficult to assess because most of them result incomplete destruction of prey. Heterotrophic protistscontinued to move into the plankton, withradiolarians diversifying (Casey, 1993) andtintinnids first appearing (Tappan, 1993). In the lateOrdovician, a major extinction eliminated 85% ofthe marine biota; it was followed by a radiation oftaxa into similar habitats during the early Silurian(Sheehan, 2001). Protists tracked the invertebrates(Knoll and Lipps, 1993).Later in the Paleozoic, foraminifera evolvedcalcareous tests. The fusulinids originated in theSilurian, then developed heavily calcified, largetests in tropical carbonate environments in theCarboniferous and Permian (Culver, 1993). Theselikely were the first coevolved photosymbiont–82

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSforaminiferal systems. The association wasimmensely successful, and lasted some 100 millionyears. The strategy of symbioses between algalprotists and other protists and animals became morewidespread, as various animals throughoutcarbonate shelves adopted it as well.During the Paleozoic, continental configurationchanged from fairly well-dispersed Cambrian landmasses to all the Earth’s plates assembled into asingle supercontinent. These positional changeswould have caused much change in the distributionand ecology of microorganisms. The trophicstructures over the era must have been different indetail, but no interpretations for microfossils areavailable. As plates moved across latitudes andoceanography changed, different regions of theoceans were subjected to different nutrient supplies,temperature regimes, and hydroclimates that alteredtrophic structures and biogeographies (Ross, 1970).The Paleozoic ended with the largest extinction ever.This event terminated multiple animal (Erwin, 1993)and protist lineages, such as benthic foraminifera(especially fusulinids) and at least four lineages ofpelagic radiolarians (Casey, 1993; Culver, 1993).Mesozoic: 250 to 66 Ma.—The rediversificationof benthic organisms, including foraminifera, andthe radiation of the oceanic groups of skeletalmicroplankton and nekton like the ammonites, fish,and reptiles were the major events of the Mesozoic(Gale, 2000). Marine diversity remained low untilthe mid-Triassic. Microbial mats seem to haveflourished in the earliest Triassic (Fig. 6A), whenmetazoan diversity was very low. Diversity soonincreased to that of the Paleozoic, and marinecommunities developed a normal appearance(Erwin, 1996). Scleractinian corals built reefs andmade use of photoendosymbionts, most surelydinoflagellates. Benthic foraminifera diversified aswell, but their trophic functions can only beinferred. They ate small primary producers,secondary consumers of all kinds, and probablyeven larger motile invertebrates caught using theirpseudopodia (Culver and Lipps, in press). Manyspecies of symbiont-bearing foraminifera existedthroughout the Mesozoic, the fusiformalveolinellids (Fig. 7) in particular. In some placesABFIGURE 6—Phanerozoic microbial mats. A, LowerTriassic domed mats, Kockatea Shale, Blue Hills,Western Australia. B, Jurassic mats in theGeological Natural Reserve of Haute-Provence atEntrages, France. Both mats cover a wide areawith smaller domes interspersed on them. Thedomes may be true stromatolites built by microbesor they may have formed by gas accumulationsunder the mats. (Photos by J. H. Lipps.)83

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 7—A living larger foraminiferan,Alveolinella quoyi, containing diatomendosymbionts. While foraminifera host a varietyof algal symbionts, most are dinoflagellates ordiatoms. The darker areas on the specimen aregreen or brown in life. Specimen is one cm long.(Photo by J. H. Lipps at the Motupore IslandResearch Station, Papua New Guinea.)where conditions may have been unfavorable tometazoans microbial mats grew (Fig. 6B). Otherbenthic prokaryotes and protists left few or noremains and nothing can be inferred about them.They must have been present in great diversity andconsiderable abundance.Mesozoic skeletonized nano- andmicroplankton radiated starting in the Triassic andcontinuing through the Cretaceous. Pelagiccoccolithophorids, dinoflagellates, diatoms,silicoflagellates, radiolaria, and foraminifera reachedtheir highest diversities during this era. They becameprovincial in their distributions from north to south,their taxonomic distributions generally reflecting thewater masses and sea surface circulation patterns.The continents were arrayed mostly across latitudes,forcing boundary currents along their margins anda gyre system in the mid-latitude centers of theoceans. These currents distributed fairly warm andhomogeneous waters from low to high latitudes. Theopening of the Atlantic Ocean, and thus thedecreasing width of the proto-Pacific, caused theseoceanographic patterns to be restricted. Thediversification of pelagic microorganismsthroughout the Mesozoic relates to the intensifiedand restricted surface circulation (Lipps, 1985). Asthese currents encountered continental margins,particularly on the eastern sides of ocean basins,upwelling—bringing nutrients into the surficialwaters—occurred more readily than before. Thus,the oceanic environments became more heterogeneousand separated, allowing diversificationthrough the creation of new pelagic environmentsand isolation. While this general model may accountfor the evolutionary patterns in some respects, moreinformation is required for a complete interpretation.Since much of the Mesozoic deep sea floor has beensubducted or translated to regions far removed fromits origin, collecting this information is an oneroustask. However, the nano- and microplankton are wellknown from sediments preserved on land or in afew places in the deep sea. These microfossils showthat a major shift in the sites of carbonate depositiontook place between the Paleozoic and the Mesozoic.Most Paleozoic carbonate was deposited in shallowseas by larger invertebrates, calcareous algae, andbenthic foraminifera; but in the mid- to lateMesozoic the deposition of carbonate was chieflycarried out by calcareous phytoplankton andforaminifera. Siliceous deposition in the deep seaalso became prevalent as radiolaria, diatoms, andsilicoflagellates incorporated vast quantities of silicathat was then deposited onto the sea floor.Trophic structures in these pelagic systemsbecame more complex than ever before (Table3,4).Not only were different ecosystems now in place indifferent parts of the oceans, but they also had moreelements in them. A bacterial loop must have beenpresent, primary producers included prokaryotes,and algal protists included skeletonized forms.Microcarnivores likewise had unskeletonized andskeletonized elements. The skeletonized formswere dependent on particular areas in the oceanwhere their skeletal materials were available—carbonate in widespread areas of the oceans andsilica mostly in upwelling areas in eastern boundarycurrents and along the equator. Food webs inupwelling systems were probably bimodal, aslarger phytoplankton grew rapidly and supporteddirect feeding by larger herbivores and carnivores.Smaller nano- and microplankton were presentwith their own set of herbivores and carnivores.The advent of larger plankton was exploited bypelagic fish and reptiles.84

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSSymbiosis with algal protists apparently wasimportant in the carbonate, tropical zones. Notonly were larger foraminifera present, butscleractinian corals and rudist bivalves developedextensive reef systems (Kauffman and Johnson,1988; Stanley, 1988). All of these organisms mayhave harbored photoendosymbionts. This type ofsymbiosis increases the ability of the host tosecrete large amounts of carbonate, and hence tobuild a massive skeleton (Cowen, 1983).The Mesozoic ended with a mass extinction thateliminated the symbiont-bearing but not otherbenthic foraminifera (Culver, in press), and reducedeach of the planktic microfossil groups to less than10% of their diversity. Likewise the reef builders ofthe Mesozoic perished completely in the extinctionevent, which seems to have affected shallower-,warmer-water organisms more than others.Cenozoic: 66 to 0 Ma.—The end-Mesozoicextinction left only a few members of eachmicrofossil group in pelagic and carbonateenvironments. From these few species offoraminifera, radiolaria, diatoms, coccolithophorids,and others, the Cenozoic biota (Pickering, 2000) wasestablished in the Paleocene. The species were quitedifferent in morphology, but their trophic roles wererestored. The trophic structures of the Cenozoicwere, as far as the fossil record indicates, more orless identical to those of the Mesozoic. After aboutthree million years, symbiont-bearing taxaappeared again in both benthic and plankticforaminifera and probably radiolaria; diatoms werethe dominant larger primary producer in eutrophicregions, while coccolithophorids dominated in theoligotrophic central gyres. Foraminifera againbecame important contributors to reefs. Deep-seabiotas were trophically and taxonomically similarto those of the Cretaceous but experienced a laterextinction as deep-water circulation changed at theend of the Paleocene. Although the biota changed,the foraminifera probably continued to cycledetrital material and prokaryotes to largerinvertebrates that preyed on them. At the end ofthe Eocene, shallow-water and pelagic biotassuffered another extinction that resulted in a sharpdecrease in diversity (Lipps, 1985). The microfossildiversity slowly recovered through the Oligocene,and by the Miocene the trophic structures seem ascomplex in terms of species involved as before.From the mid-Miocene on, glaciations in highlatitudes became increasingly intense until highlatitudeparts of continents were covered with largeice sheets. This refrigeration of Earth had littleeffect on the trophic structures in either benthic orpelagic microbiotas other than to shift the entirebiotas biogeographically toward lower latitudes andback again during warmer periods. Upwellingintensified along the equator, along the eastern sidesof continents, and around Antarctica. This allowedincreased primary production in these areas and aconcentration of silica deposition by the diatomsand radiolarians. Pleistocene glacial-interglacialshifts in upwelling intensity and bottom waterscaused restructuring of deep-water assemblages.EFFECTS OF EXTINCTIONSON FOOD WEBSExtinctions during the Phanerozoic changedthe microbial trophic structures enormously. Eachextinction, in general, appears to have had asimilar impact—the diversity of most elementsin shallow and pelagic ecosystems was reduceddrastically (Lipps, 1985; Culver, in press), and itmost commonly occurred rapidly (Table 5).Whether or not this happened because a particularpart of the food web was destroyed with ancillaryeffects cannot be determined. Species diversityof primary producers declined as much as that ofthe consumers and predators, and at the same time.Overall primary production may have declinedor the ecosystem changed so as to cause a decreasein carbon isotopic values (d’Hondt, 1998).Evidence from oxygen isotopic values of plankticforaminifera in the Mesozoic and Cenozoicindicates that the thermocline was commonlydegraded at mass extinction events, and this wouldcertainly have had a major effect on theavailability of nutrients for the photosyntheticorganisms. If they declined in productivity, thenthe entire trophic structure of much of the oceansmay well have collapsed too. Further studies of85

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 5—Trophic structure changes across the Cretaceous-Tertiary boundary to the Eocene in pelagicand benthic ecosystems.Pelagic EcosystemsLatest Cretaceous Earliest Paleocene EoceneLarge reptilesFish, cephalopodsZooplankton (diverseforaminifera, radiolaria,tintinnids, crustaceans,mollusks, etc.)CetaceansFish, cephalopods? Zooplankton (diverseforaminifera, radiolaria,tintinnids, crustaceans,mollusks, etc.)Microherbivores (diverseforaminifera, radiolaria,tintinnids, etc.)Protist phytoplankton(diverse coccolithophorids,diatoms, silicoflagellates, etc.)Prokaryotes includingcyanobacteriaMicroherbivores (few speciesof foraminifera, etc.)Protist phytoplankton (fewspecies ofcoccolithophorids, etc.)Prokaryotes includingcyanobacteriaMicroherbivores (diverseforaminifera, radiolaria,tintinnids, etc.)Protist phytoplankton (diversecoccolithophorids, diatoms,silicoflagellates, etc.)Prokaryotes includingcyanobacteriaBenthic EcosystemsLatest Cretaceous Earliest Paleocene EoceneMacrocarnivores (fish, mollusks,echinoderms, etc.)Microcarnivores (diverseforaminifera, crustaceans,mollusks, etc.)Microherbivores (diverseforaminifera, tinymollusks, etc.)Algal-foraminiferal andinvertebrate endosymbiosesProtist primary producers(diatoms, flagellates, etc.)Prokaryotes includingcyanobacteriaMicrocarnivoresMicroherbivores (foraminifera,mollusks, echinoderms, etc.)Protist primary producers(naked forms, etc.)Prokaryotes includingcyanobacteriaMacrocarnivores (fish,mollusks, echinoderms, etc.)Microcarnivores (diverseforaminifera, crustaceans,mollusks, etc.)Microherbivores (diverseforaminifera, mollusks,echinoderms, etc.)Algal-foraminiferal andinvertebrate endosymbioses.Protist phytoplankton (diversediatoms, flagellates, etc.)Prokaryotes includingcyanobacteria86

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSthe trophic structures in mass-extinctiontransitions are required, with special emphasis onthe preserved primary producers, including theacritarchs, calcareous nannoplankton, anddiatoms. The microherbivore and carnivorerecords should be examined as indicators of thenature of primary production. Whether or notdeclining primary production was intimatelyinvolved in the overall extinction dynamics, thephysical oceanographic mechanisms influencingprimary productivity may have differed in eachextinction event. Some of these mechanisms mayhave included asteroid impacts, changingoceanographic circulation, and changing degreeof oceanic heterogeneity as controlled by densitystratification. The patterns in Phanerozoicextinctions seem sufficiently consistent thatregardless of physical causes, trophic dynamicswere affected severely.IMPLICATIONS FORCONSERVATION BIOLOGYImportant for the conservation of marinebiodiversity, but commonly overlooked, is themaintenance of the trophic mechanisms andstructures without which these diverse assemblagescould not function. These structures took four billionyears to evolve, and climatic and oceanographicchanges—or a few human misjudgments—coulddamage them considerably. The fossil record andmodern marine ecology demonstrate the importanceof prokaryotes and protists in the trophic interactionsthat energize all marine ecosystems. Microfossiltrophic analyses have much to contribute to ourunderstanding of trophic dynamics and where thesesystems are most vulnerable. Modern ecologicalstudies are severely constrained by a lack ofhistorical perspective; experiments cannot bedesigned long enough to record the consequencesof events occurring on ecologic scales.Paleontologists must provide that perspective withdetailed studies of their own to complement andstrengthen ecologic work.ACKNOWLEDGMENTSJHL thanks J. Valentine, C. Blank, B.Waggoner, M. Fedonkin, A. Rozanov, A. Ivantsov,and J. Gehling for discussion of the Precambrianand Cambrian bacterial and metazoan biotas, andW. Hamner and L. Gershwin for information aboutmodern cnidarians. He also thanks the UC GumpResearch Station on Moorea and its ResearchDirector Dr. Neil Davis for the facilities to makethe study of tropical microbial mats possible; andthe Geological Natural Reserve of Haute-Provenceand M. Guiomar for permission to examine fossilsthere. SJC thanks D. Merritt, M. Buzas, F. Banner,J. Wallace, D. Stewart, and C. Smith for assistanceand support. JHL and SJC are indebted to D. Haigfor organizing an amazing week-long field tripduring Forams2002 to study microbial mats andstromatolites in Western Australia. Researchunderpinning much of this summary was supportedby grants NSF EAR 85-09301 to JHL, NSF EAR9317247 and NSF EAR 9814845 to J. Valentineand JHL. SJC thanks the USGS for funding theNorth Carolina Coastal Geology Cooperativeresearch program at ECU. Permission to undertakeresearch on the Outer Banks’ Pea Island NationalWildlife Refuge is gratefully acknowledged. Thisis UC Museum of Paleontology Contribution 1787and UC Gump Research Station Contribution 96.REFERENCESALTENBACH, A. V., U. PFLAUMANN, R. SCHIEBEL, A. THIES, S. TIMM, AND M. TRAUTH. 1999. Scaling percentages and distributionalpatterns of benthic foraminifera with flux rates of organic carbon. Journal of Foraminiferal Research, 29:173–185.ANDERS, E. 1989. Prebiotic organic matter from comets and asteroids. Nature, 342:255–257.ANDERSON, O. R. 1993. The trophic role of planktonic foraminifera and radiolaria. Marine Microbial Food Webs, 7:31–51.AWRAMIK, S. M. 1971. Precambrian columnar stromatolite diversity: reflection of metazoan appearance. Science,174:825–827.87

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002AZAM, F. 1998. Microbial control of oceanic carbon flux: the plot thickens. Science, 280:694–696.BADA, J. L., AND A. LAZCANO. 2002. Some like it hot, but not the first biomolecules. Science, 296:1982–1983.BENGTSON, S., AND Y. ZHAO. 1992. Predatorial borings in Late Precambrian mineralized exoskeletons. Science, 257:367–369.BRASIER, M. D., O. R. GREEN, A. P. JEPHCOAT, A. K. KLEPPE, M. J. VAN KRANENDONK, J. F. LINDSAY, A. STEELE, ANDN. V. GRASSINEAU. 2002. Questioning the evidence for Earth’s oldest fossils. Nature, 416:76–81.BROCIER, C., AND H. PHILIPPE. 2002. A non-hyperthermophilic ancestor for Bacteria. Nature, 417:244.BROCK, J. J., G. A. LOGAN, R. BUICK, AND R. E. SUMMONS. 1999. Archean molecular fossils and the early rise ofeukaryotes. Science, 285:1033–1036.BUSS, L. W., AND A. SEILACHER. 1994. The phylum Vendobionta: a sister group of the Eumetazoa? Paleobiology, 20:1–4.BUTTERFIELD, N. J. 1997. Plankton ecology and the Proterozoic-Phanerozoic transition. Paleobiology, 23:247–262.BUTTERFIELD, N. J. 2001. Ecology and evolution of Cambrian plankton, p. 200–216. In A. Y. Zhuravlev and R.Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.BUZAS, M. A., L. S. COLLINS, AND S. J. CULVER. 2002. Latitudinal difference in biodiversity caused by highertropical rate of increase. Proceedings of National Academy of Sciences, USA, 99: 7841–7843.CARRE, D., C. CARRE, AND C. E. MILLS. 1989. Novel cnidocysts of narcomedusae and a medusivorous ctenophoreand confirmation of kleptocnidism. Tissue and Cell, 21:723–734.CASEY, R. E. 1993. Radiolaria, p. 249–284. In J. H. Lipps (ed.), Fossil Prokaryotes and Protists. Blackwell ScientificPublications, Boston.CHRISTAKI, U., A. GIANNAKOUROU, F. VAN WAMBEKE, AND G. GREGORI. 2001. Nanoflagellate predation on auto- andheterotrophic picoplankton in the oligotrophic Mediterranean Sea. Journal of Plankton Research, 23:1297–1310.CHYBA, C. F., P. J. THOMAS, L. BROOKSHAW, AND C. SAGAN. 1990. Cometary delivery of organic molecules to theearly Earth. Science, 249:366–373.CLAPTON, M. E., AND G. M. NARBONNE. 2002. Ediacaran epifaunal tiering. Geology, 30:627–630.COLLINS, A. G., J. H. LIPPS, AND J. W. VALENTINE. 2000. Modern mucociliary creeping trails and the body plans ofNeoproterozoic trace-makers. Paleobiology, 26:47–55.COWEN, R. 1983. Algal symbiosis and its recognition in he fossil record, p. 431–479. In M. J. S. Tevesz and P. L.McCall (eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum Press, New York.CRAME, J. A. 2002. Evolution of taxonomic diversity gradients in the marine realm: a comparison of Late Jurassicand Recent bivalve faunas. Paleobiology, 28:184–208.CRIMES, T. P. 1992. The record of trace fossils across the Proterozoic-Cambrian boundary, p. 177–202. In J. H.Lipps and P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum Publishers, New York.CULVER, S. J. 1993. Foraminifera, p. 203–247. In J. H. Lipps (ed.), Fossil Prokaryotes and Protists. BlackwellScientific Publications, Boston.CULVER, S. J. 1994. Early Cambrian foraminifera from the southwestern Taoudeni Basin, West Africa. Journal ofForaminiferal Research, 24:191–202.CULVER, S. J. In press. Benthic foraminifera across the Cretaceous-Tertiary (K-T) boundary—a review. MarineMicropaleontology.CULVER, S. J., AND J. H. LIPPS. In press. Predation on and by foraminifera. In M. Kowalewski and P. H. Kelley(eds.), Predation in the Fossil Record. Kluwer Publishing Co.DEWEL, R. A., W. C. DEWEL, AND F. K. MCKINNEY. 2001. Diversification of the Metazoa: ediacarans, colonies, andthe origin of eumetazoan complexity by nested modularity. Historical Biology, 15:93–118.D’HONDT, S. 1998. Isotopic proxies for ecological collapse and recovery from mass extinctions. The PaleontologicalSociety Papers, 4:179–211.DONG, X. P., A. H. KNOLL, AND J. H. LIPPS. 1997. Late Cambrian Radiolaria from Hunan, China. Journal ofPaleontology, 71:753–758.DZIK, J., AND A. Y. IVANTSOV. 2002. Internal anatomy of a new Precambrian dickinsoniid dipleurozoan fromnorthern Russia. Neues Jahrbuch fur Palaontologie Abhandlungen. In press.ERWIN, D. H. 1993. The Great Paleozoic Crisis: Life and Death in the Permian. Columbia University Press, New York, 327 p.ERWIN, D. H. 1996. Understanding biotic recoveries: Extinction, survival, and preservation during the end-Permianmass extinction, p. 398–418. In D. Jablonski, D. H. Erwin, and J. H. Lipps (eds.), Evolutionary Paleobiology.University of Chicago Press, Chicago.88

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSERSKIAN, M. G., AND J. H. LIPPS. 1987. Population dynamics of the foraminiferan Glabratella ornatissima (Cushman)in northern California (USA). Journal of Foraminiferal Research, 17:240–256.FALKOWSKI, P. G., R. T. BARBER, AND V. SMETACEK. 1998. Biogeochemical controls and feedbacks on ocean primaryproduction. Science, 281:200–206.FAUTIN, D. G., AND W. K. FITT. 1991. A jellyfish-eating sea anemone (Cnidaria, Actinaria) from Palau: Entacmaeamedusivora sp. nov. Hydrobiologia, 216/217:453–461.FEDO, C. M., AND M. J. WHITEHOUSE. 2002. Metasomatic origin of quartz-pyroxene rock, Akilia, Greenland, andimplications for Earth’s earliest life. Science, 296:1448–1452.FEDONKIN, M. A. 1992. Vendian faunas and the early evolution of Metazoa, p. 87–129. In J. H. Lipps and P. W.Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum Publishers, New York.FEDONKIN, M. A. 1994. Vendian body fossils and trace fossils, p. 370–388. In S. Bengtson (ed.), Early Life onEarth; 84th Nobel Symposium, Karlskoga, Sweden, May 16, 1992. Columbia University Press, New York.FEDONKIN, M. A., AND B. N. RUNNEGAR. 1992. Proterozoic metazoan trace fossils, p. 389–395. In J. W. Schopf andC. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.FEDONKIN, M. A., AND B. M. WAGGONER. 1997. The Late Precambrian fossil Kimberella is a mollusc-like bilaterianorganism. Nature, 388:868–871.GALE, A. S. 2000. The Cretaceous world, p. 4–19. In S. J. Culver and P. F. Rawson (eds.), Biotic Response toGlobal Change: The Last 145 Million Years. Cambridge University Press, Cambridge.GEHLING, J. G. 1988. A cnidarian of actinian-grade from the Ediacaran Pound Subgroup, South Australia. Alcheringa,12:299–314.GEHLING, J. G. 1999. Microbial mats in terminal Proterozoic siliciclastics: Ediacaran death masks. Palaios, 14:40–57.GLAESSNER, M. F. 1984. The Dawn of Animal Life: A Biohistorical Study. Cambridge University Press, Cambridge, 244 p.GRAF, G. 1989. Benthic-pelagic coupling in a deep-sea benthic community. Nature, 341:437–439.GU, X. 1997. The age of the common ancestor of eukaryotes and prokaryotes: statistical inferences. MolecularBiology and Evolution, 14:861–866.HARLAND, W. B. 1964. Critical evidence for a great infra-Cambrian glaciation. Geologische Rundschau, 54:45–61.HEDGES, B., H. CHEN, S. KUMAR, D. Y-C. WANG, A. S. THOMPSON, AND H. WATANABE. 2001. A genomic timescale forthe origin of eukaryotes. BMC Evolutionary Biology 2001, 1:4. (www.biomedcentral.com/1471-2148/1/4)HOFFMAN, P. F., A. J. KAUFMAN, G. P. HALVERSON, AND D. P. SCHRAG. 1998. A Neoproterozoic snowball earth.Science, 281:1342–1346.IVANTSOV, A. Y., AND M. A. FEDONKIN. 2001. Trails of active locomotion: final proof of animal nature of the Ediacaraorganisms, p. 133–137. In V. M. Podobina (ed.), Evolution of Life on Earth. NTL, Tomsk [in Russian].JAVAUX, E. J., A. H. KNOLL, AND M. R. WALTER. 2001. Morphological and ecological complexity in early eukaryoticecosystems. Nature, 412:66–69.JENKINS, R. J. F. 1992. Functional and ecological aspects of Ediacaran assemblages, p. 131–176. In J. H. Lippsand P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum Press, New York.KATZ, L. A. 1999. The tangled web: gene genealogies and the origin of eukaryotes. American Naturalist,154(Supplement):S137–S145.KAUFFMAN, E. G., AND C. C. JOHNSON. 1988. The morphological and ecological evolution of Middle and UpperCretaceous reef-building rudistids. Palaios, 3:194–216.KIORBOE, T. 1993. Turbulence, phytoplankton cell size and the structure of pelagic food webs. Advances in MarineBiology, 29:1–72.KIRSCHVINK, J. L. 1992. Late Proterozoic low-latitude global glaciation: The Snowball Earth, p. 51–52. In J. W.Schopf and C. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.KIRSCHVINK, J. L., E. J. GAIDOS, L. E. BERTANI, N. J. BEUKES, J. GUTZMER, L. N. MAEPA, AND R. E. STEINBERGER. 2000.Paleoproterozoic snowball Earth: Extreme climatic and geochemical global change and its biological consequences.Proceedings of the National Academy of Sciences, USA, 97:1400–1405.KNOLL, A. H. 1994. Proterozoic and Early Cambrian protists: Evidence for accelerating evolutionary tempo.Proceedings of the National Academy of Sciences, USA, 91:6743–6750.KNOLL, A. H., AND S. B. CARROLL. 1999. Early animal evolution: emerging views from comparative biology andgeology. Science, 284:2129–2137.89

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002KNOLL, A. H., AND J. H. LIPPS. 1993. Evolutionary history of prokaryotes and protists, p. 19–29. In J. H. Lipps(ed.), Fossil prokaryotes and protists. Blackwell Scientific Publications, Boston.LINKE, P., AND G. F. LUTZE. 1993. Microhabitat preferences of benthic foraminifera—a static concept or a dynamicadaptation to optimize food acquisition? Marine Micropaleontology, 20:215–234.LIPPS, J. H. 1981. What, if anything, is micropaleontology? Paleobiology, 7:167–199.LIPPS, J. H. 1983. Biotic interactions in benthic foraminifera, p. 331–376. In M. J. S. Tevesz and P. L. McCall(eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum Press, New York.LIPPS, J. H. 1985. Extinction dynamics in pelagic ecosystems, p. 87–104. In D. K. Elliott (ed.), Dynamics ofExtinction. John Wiley & Sons, New York.LIPPS, J. H. 1993. Introduction to fossil prokaryotes and protists, p. 1–10. In J. H. Lipps (ed.), Fossil Prokaryotesand Protists. Blackwell Scientific Publishers, Boston.LIPPS, J. H., AND A. YU. ROZANOV. 1996. The Late Precambrian–Cambrian agglutinated fossil Platysolenites.Paleontological Journal, 30:679–687.LIPPS, J. H., AND J. W. VALENTINE. 1970. The role of foraminifera in the trophic structure of marine communities.Lethaia, 3:279–286.LIPPS, J. H., S. BENGTSON, AND J. D. FARMER. 1992. The Precambrian-Cambrian transition, p. 453–457. In J. W.Schopf and C. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.LOWE, D. R. 1992. Major events in the geological development of the Precambrian Earth, p. 67–75. In J. W.Schopf and C. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.MADIN, L. P. 1988. Feeding behavior of tentaculate predators: in situ observations and a conceptual model. Bulletinof Marine Science, 43:413–429.MARTIN, M. W., D. V. GRAZHDANKIN, S. A. BOWRING, D. A. D. EVANS, M. A. FEDONKIN, AND J. L. KIRSCHVINK. 2000.Age of Neoproterozoic bilatarian body and trace fossils, White Sea, Russia: implications for metazoanevolution. Science, 288:841–845.MCMENAMIN, M. A. S. 1986. The garden of Ediacara. Palaios, 1:178–182.MILLS, C. E. 1993. Natural mortality in NE Pacific coastal hydromedusae: grazing predation, wound healing andsenescence. Bulletin of Marine Science, 53:194–203.MILLS, C. E. 1995. Medusae, siphonophores, and ctenophores as planktivorous predators in changing globalecosystems. ICES Journal of Marine Science, 52:575–581.MOJZSIS, S. J., T. M. HARRISON, AND R. T. PIDGEON. 2001. Oxygen-isotopic evidence from ancient zircons forliquid water at the Earth’s surface 4,300 million years ago. Nature, 409:178–181.MOLDOWAN, J. M., AND N. M. TALYZINA. 1998. Biogeochemical evidence for dinoflagellate ancestors in the EarlyCambrian. Science, 281:1168–1170.MONASTERSKY, R. 1998. The rise of life on Earth: Life grows up. National Geographic Magazine, 193:100–115.NARBONNE, G. M. 1998. The Ediacara biota: A terminal Neoproterozoic experiment in the evolution of life. GSAToday, 8:1–6.NISBET, E. G., AND N. H. SLEEP. 2001. The habitat and nature of early life. Nature, 409:1083–1091.NYBAKKEN, J. W. 2001. Marine Biology, 5 th ed. Benjamin Cummings, San Francisco, 516 p.OSINGA, R., E. ARMSTRONG, J. G. BURGESS, F. HOFFMANN, J. REITNER, AND G. SCHUMANN-KINDEL. 2001. Spongemicrobeassociations and their importance for sponge bioprocess engineering. Hydrobiologia 1, 461:55–62.PATTERSON, D. J. 1999. The diversity of eukaryotes. American Naturalist, 154:S96–S124.PICKERING, K. T. 2000. The Cenozoic world, p. 20–34. In S. J. Culver and P. F. Rawson (eds.), Biotic Response toGlobal Change: The Last 145 Million Years. Cambridge University Press, Cambridge.POMEROY, L. R. 1974. The ocean’s food web: A changing paradigm. Bioscience, 24:499–504.RASMUSSEN, B., S. BENGTSON, I. R. FLETCHER, AND N. J. MCNAUGHTON. 2002. Discoidal impressions and trace-likefossils more than 1200 million years old. Science, 296:1112–1115.RETALLACK, G. J. 1994. Were the Ediacaran fossils lichens? Paleobiology, 20:523–544.REYSENBACH, A.-L., AND E. SHOCK. 2002. Merging genomes with geochemistry in hydrothermal ecosystems. Science,296:1077–1082.RIEMANN, L., G. F. STEWARD, AND F. AZAM. 2000. Dynamics of bacterial community composition and activityduring a mesocosm diatom bloom. Applied and Environmental Microbiology, 66:578–587.90

LIPPS AND CULVER—TROPHIC ROLE OF MARINE MICROORGANISMSROSING, M. T. 1999. 13 C-depleted carbon microparticles in >3700-Ma sea-floor sedimentary rocks from westGreenland. Science, 283:674–676.ROSS, C. A. 1970. Development of fusulinid (Foraminiferida) faunal realms. Journal of Paleontology, 41:1341–1354.ROWLAND, S. M., AND R. S. SHAPIRO. 2002. Reef patterns and environmental influences in the Cambrian andearliest Ordovician. SEPM Special Publication, 72: 95–129.RUNNEGAR, B. N. 1992a. Proterozoic fossils of soft-bodied metazoans (Ediacara faunas), p. 999–1007. In J. W.Schopf and C. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.RUNNEGAR, B. N. 1992b. Proterozoic metazoan trace fossils, p. 1009–1015. In J. W. Schopf and C. Klein (eds.),The Proterozoic Biosphere. Cambridge University Press, Cambridge.RUNNEGAR, B. N. 1992c. Evolution of the earliest animals, p. 65–93. In J. W. Schopf (ed.), Major Events in theHistory of Life. Jones and Bartlett Publishers, Boston.SCHIDLOWSKI, M. 2001. Carbon isotopes as biogeochemical recorders of life over 3.8 Ga of Earth history: evolutionof a concept. Precambrian Research, 106:117–134.SCHMIEDL, G., F. DE BOVEE, R. BUSCAIL, B. CHARRIERE, C. HEMLEBEN, L. MEDERNACH, AND P. PICON. 2000. Trophiccontrol of benthic foraminiferal abundance and microhabitat in the bathyal Gulf of Lions, westernMediterranean Sea. Marine Micropaleontology, 40:167–188.SCHOPF, J. W. 1992a. Paleobiology of the Archean, p. 25–39. In J. W. Schopf and C. Klein (eds.), The ProterozoicBiosphere. Cambridge University Press, Cambridge.SCHOPF, J. W. 1992b. Proterozoic prokaryotes: Affinities, geologic distribution, and evolutionary trends, p. 195–218. In J. W. Schopf and C. Klein (eds.), The Proterozoic Biosphere. Cambridge University Press, Cambridge.SCHOPF, J. W. 1993. Microfossils of the early Archean Apex Chert—new evidence of the antiquity of life. Science,260:640–646.SCHOPF, J. W., AND B. M. PACKER. 1987. Early Archean (3.3 billion to 3.5 billion-year-old) microfossils fromWarrawoona Group, Australia. Science, 237:70–73.SEILACHER, A. 1989. Vendozoa: Organismic construction in the Proterozoic biosphere. Lethaia, 22:229–240.SEILACHER, A. 1999. Biomat-related lifestyles in the Precambrian. Palaios, 14:86–93.SERGEEV, V. N., A. H. KNOLL, AND P. YU. PETROV. 1997. Paleobiology of the Mesoproterozoic–Neoproterozoic transition:the Sukhaya Tunguska Formation, Turukhansk Uplift, Siberia. Precambrian Research, 85:201–239.SHEEHAN, P. M. 2001. The late Ordovician mass extinction. Annual Reviews of Earth and Planetary Sciences,29:331–364.SHINADA, A., T. IKEDA, S. BAN, AND A. TSUDA. 2001. Seasonal dynamics of planktonic food chain in the Oyashioregion, western subarctic Pacific. Journal of Plankton Research, 23:1237–1247.SIMONSON, B. M., AND K. E. CARNEY. 1999. Roll-up structures: evidence of in-situ microbial mats in late Archeandeep shelf environments. Palaios, 14:13–24.STANLEY, G. D., JR. 1988. The history of early Mesozoic reef communities: a three-step process. Palaios, 3:170–183.SUMMONS, R. E., L. L. JAHNKE, J. M. HOPE, AND G. A. LOGAN. 1999. 2-methylhopanoids as biomarkers forcyanobacteria oxygenic photosynthesis. Nature, 400:554–557.TAPPAN, H. 1993. Tintinnids, p. 285–303. In J. H. Lipps (ed.), Fossil Prokaryotes and Protists. Blackwell ScientificPublications, Boston.TORSVIK, V., L. OVREAS, AND T. F. THINGSTAD. 2002. Prokaryotic diversity-magnitude, dynamics, and controllingfactors. Science, 296:1064–1066.VALENTINE, J. W. 2002. Prelude to the Cambrian explosion. Annual Review of Earth and Planetary Sciences,30:285–306.VALENTINE, J. W., D. JABLONSKI, AND D. H. ERWIN. 1999. Fossils, molecules and embryos: new perspectives on theCambrian explosion. Development, 126:851–859.VALENTINE, J. W., S. M. AWRAMIK, P. W. SIGNOR, AND P. M. SADLER. 1991. The biological explosion at thePrecambrian-Cambrian boundary. Evolutionary Biology, 25:279–355.VIDAL, G., AND M. MOCZYDLOWSKA-VIDAL. 1997. Biodiversity, speciation, and extinction trends of Proterozoicand Cambrian phytoplankton. Paleobiology, 23:230–246.WAGGONER, B. M. 1995. Ediacaran lichens—a critique. Paleobiology, 21:393–397.91

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002WON, M.-Z., AND R. BELOW. 1999. Cambrian Radiolaria from the Georgina Basin, Queensland, Australia.Micropaleontology, 45:325–363.WOOD, R. A., J. P. GROTZINGER, AND J. A. D. DICKSON. 2002. Proterozoic modular biomineralized metazoan fromthe Nama Group, Namibia. Science, 296:2383–2386.ZHURAVLEV, A. Y. 1993. Were Ediacaran Vendobionta multicellulars? Neues Jahrbuch fur PalaontologieAbhandlungen, 190:299–314.92

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSPREDATORS AND PREDATION INPALEOZOIC MARINE ENVIRONMENTSCARLTON E. BRETT 1 AND SALLY E. WALKER 21Department of Geology, University of Cincinnati, Cincinnati, Ohio 45221-0013 USA2Department of Geology, University of Georgia, Athens, Georgia 30602-2501 USAABSTRACT—The Paleozoic body fossil record of potential benthic predators includes nautiloid and ammonoidcephalopods, phyllocarids, decapods, and several lineages of gnathostomes. The latter group, in particular, radiatedrapidly during the Devonian. In the pelagic realm, predator-prey interactions involving cephalopods and somenektonic arthropods probably appeared in the Ordovician. Again, evidence indicates intensification of pelagicpredation, much of it by arthrodires and sharks on other fishes, during the Devonian radiation of gnathostomes.Trace fossils provide direct evidence of predatory attack from the Ediacarian and Early Cambrian onward,but with a substantial increase in the Siluro-Devonian. Brachiopod and molluscan shells and trilobite exoskeletonsshow evidence of healed bite marks and peeling from the Cambrian onward, but with an increased frequency inthe Devonian. Predatory drill holes with stereotypical position and prey-species preference are found inbrachiopods (Cambrian onward) and mollusks (Ordovician onward); boreholes also show increased frequencyin the middle Paleozoic. Certain of these boreholes are tentatively attributable to platyceratid gastropods.Hard-shelled benthic organisms with thicker, more spinose skeletons may have had a selective advantage asdurophagous predators increased. Brachiopods, gastropods, trilobites, and crinoids show an abrupt increase inspinosity beginning in the Siluro-Devonian. But spinosity decreases after the early Carboniferous. Late Paleozoicbenthos may have taken refuge in smaller size and resistant, thick-walled skeletons, as well as endobenthic andcementing modes of life. Conversely, in the pelagic realm, external armor was reduced, while more efficient, fastswimmingmodes of life (e.g., in sharks) increased in the post-Devonian.INTRODUCTIONPREDATION, THE KILLING and ingestionof one animal by another carnivorous organism,has undoubtedly been an important interaction inmarine environments throughout Phanerozoichistory (see Connell, 1970; Paine, 1974; Vermeij,1977, 1987; Signor and Brett, 1984; Brett, 1992,in press; Bambach, 1993, for reviews). Arguably,predation is a key driving force in evolution.However, documentation of ancient predation isdifficult. Not surprisingly, despite compilations(e.g., Vermeij, 1987), many questions regarding thepattern of evolution of predator-prey interactionsremain unanswered. How rapidly did predationdevelop, and through what stages? Was the rise ofpredators gradual and steady, or episodic? Is thereevidence for replacement in particular predatoryguilds following mass extinctions?In this paper, the varied types of Paleozoicmarine predators are reviewed in chronologicalorder. Both pelagic and benthic ecosystems areconsidered. The latter are more thoroughlydocumented, and thus they are afforded morediscussion. Basic lines of evidence for ancientpredator-prey interactions include: a) evidence forpredatory adaptation, and b) evidence of predation.A key line of evidence for predation is the body fossilrecord of organisms in which morphology (e.g.,claws, jaws, teeth) or phylogenetic relationshipindicates a durophagous carnivorous habit (Signorand Brett, 1984). Inference of predatory behavior,obviously based on analogy with living organisms,becomes more tenuous in ancient, extinct fossilorganisms. Direct evidence of predation, asdocumented in the fossil record, includes thoserarely preserved body fossils showing predator-preyinteractions as well as other direct evidence in the93

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002form of trace fossils (e.g., shell repair, drilling, andcoprolites; see Figs. 4, 5).This paper also reviews evidence for responseof potential prey organisms to various phases inthe evolution of predators in marine ecosystemsthrough the Phanerozoic. We consider the predictedevolutionary consequences of intensified attack onhard-shelled prey, and compare these to the actualrecord of changes in skeletal morphology. Finally,we discuss several ecologically significantcorrelations of change in marine ecosystems thatmay reflect predator-prey co-evolution.CAMBRIAN RISE OF PREDATORSRecord of Marine Predators.—There isevidence of marine predation as early as the latestProterozoic (Conway Morris and Jenkins, 1985;FIGURE 1—Ranges of various taxa of Paleozoic durophagous (hard shell feeding) predators. Thinlines: present, but minor; thick lines: abundant; broken lines: possibly present but rare. Carb:Carboniferous; Miss: Mississippian: Penn: Pennsylvanian.94

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSBengtson, 1994). Certainly, by the Early Cambrian,predators were impacting the marine shelly benthos(Fig. 1) (Babcock and Robison, 1989; Babcock,1993; Conway Morris and Bengtson, 1994),although predation styles were less sophisticatedthan during the rest of the Paleozoic (ConwayMorris, 2001). Nevertheless, for the CambrianBurgess Shale communities at least, the predatoryguild was fully functioning (Conway Morris, 2001).Anomalocaridids.—Among the oldest large(up to 1 m) predators were the anomalocaridids,an enigmatic but widely distributed Cambriantaxon, with a circular slicing oral ring (Fig. 2). Bitemarks on Cambrian trilobites have been attributedto these large predators (Conway Morris andJenkins, 1985). Nedin (1999) postulated thattrilobites were captured by the large anteriorappendage of the anomalocaridids and then forcedinto the mouth, where the victim was repeatedlyflexed to crack its exoskeleton. Consequently,among the earliest lines of irrefutable evidence forpredation are trilobites that show healed divots orscalloped areas removed from portions of the dorsalexoskeleton (Fig. 3.1). Some of these bite marks inCambrian trilobites have been attributed toanomalocaridids (Nedin, 1999). Many of these bitemarks occur on the posterior right pleural lobes oftrilobites (Babcock and Robison, 1989; Babcock,1993). The consistent location of bite marks implieseither that much predation occurred from the rear,or that anterior attacks were more commonly fatal.It further suggests left-right asymmetries(lateralization) in mode of attack by visual predators,or in behavioral response of the attacked trilobites,or both (Babcock, 1993).Trilobites.—Trilobites themselves (Fig. 3.1)have been cited as primitive predators on softbodiedorganisms. Trace fossil assemblages fromthe Cambrian show numerous instances of trilobiteproducedRusophycus and Cruziana interceptingPlanolites or Teichichnus traces attributable toinfaunal worms. These interception traces havebeen interpreted as evidence for foraging andhunting behavior in trilobites (Bergström, 1973;Fortey and Owens, 1999).Other Predators.—Trilobite sclerites,ostracodes, and hyolithids have also been found ingut traces of other large Cambrian arthropods,including Sidneyia (Briggs et al., 1994) andUtahcaris (Conway Morris and Robison, 1988);the enigmatic arthropods Yohoia and Branchiocarisalso may have been durophagous predatorsFIGURE 2—Anomalocaris; reconstruction based on material from Middle Cambrian, Burgess Shale,×0.5. Drawing by Marianne Collins, from Gould (1989); reprinted by permission.95

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 3—Examples of Paleozoic marine predators. 1—Reconstruction of trilobite Elrathia kingi withappendages, showing a divot in the lower right pleuron probably from an anomalocaridid; ×1. 2—TheRecent phyllocarid Nebalia. 3—Large predaceous eurypterid Pterygotus; ×0.25. 4—Enlargement ofchelicera of Pterygotus showing serrated cutting edges; ×0.5. 5, 6—Platyceratid gastropods: 5—Platyceras; Devonian; 6—Naticonema; Silurian-Devonian; ×1. 7—Nautiloid cephalopod, reconstructedwith outlines of dental arcade of the shark Petalodus, based on specimen with rows of punctures fromPennsylvanian of Kentucky. 1, based on specimen illustrated by Babcock (1993); 2, from Clarkson(1996); 5, 6, from Tasch (1980); 7, from Mapes and Hansen (1984).96

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTS(Vermeij, 1987). In addition, specimens ofpriapulid worms from the Burgess Shale have beenfound with hyolithids in their gut trace (ConwayMorris, 1977; Briggs et al., 1994, their fig. 73).The earliest report of cannibalism is also knownfrom the Burgess Shale. One specimen of Ottoia,a priapulid, had a proboscis of another Ottoiapreserved in its gut (Briggs et al., 1994). Modernpriapulids are also known to be cannibals. Thus,priapulid feeding behavior has remainedremarkably similar for 530 million years (Fig. 1).Nautiloid cephalopods appear late in theCambrian, but they only attain abundance and largesize during the ensuing Ordovician Period. Allknown living cephalopods are carnivorous, butearly forms were small, poor swimmers that mayhave been bottom-feeding scavengers (Bandel,1985; Lehmann, 1988), and only later didnautiloids develop as major predators.Trace Fossils.—Circular borings made byputative predators that are 0.1 mm to 4 mm indiameter occur on shells from the Early Cambrianonward (Bengtson and Zhao, 1992; Conway-Morris and Bengtson, 1984; Miller and Sundberg,1984). Minute pits are reported in the enigmaticphosphatic shell of the Early Cambrian Moburgella(Bengtson and Zhao, 1992; Conway Morris andBengtson, 1994). These borings were evidentlyproduced by an organism capable of drillingphosphatic shells. This borer may have persistedinto the middle Paleozoic. Chatterton andWhitehead (1987) reported similar cylindrical drillholes on about 10% of the valves of a lingulatebrachiopod from the Silurian of Oklahoma.Putative predator borings are also known tooccur in the exoskeletons of agnostoid trilobites(Babcock, 1993). Some of these tiny pits have pearllikeplugs, evidently secreted by the trilobites inresponse to the predatory action of the predator (orparasite). These ancient pits resemble borings madeby modern nematodes (Sliter, 1971) and providecircumstantial evidence for the existence of boringnematodes as far back as the Early Cambrian.Häntschel et al.’s (1968) compendium oncoprolites cites only 25 reports of pre-Devoniancoprolites; most of these are small and phosphatic.Subsequently, there have been several reports ofCambrian coprolites containing trilobite sclerites,echinoderm ossicles, and fragments of inarticulatebrachiopods (Sprinkle, 1973; Conway Morris andRobison, 1986, 1989; Babcock, 1993; Nedin, 1999).Pelagic Predators?—Seemingly, there was littleto no development of a pelagic predator-prey systemduring the Cambrian, as there are relatively fewdefinite pelagic forms. Cephalopods remained smallbenthic forms, and open swimming vertebrates,other than possible conodont animals, had yet toappear. It is possible that large nektonic arthropods,such as Sidneyia, may have preyed upon each otheror on conodont animals.CAMBRIAN RESPONSE:EARLY PALEOZOIC MARINEPREDATOR REVOLUTIONIn many ways, the Cambrian revolution ofpredators was the first major episode of escalationin marine ecosystems, although the effect of newlyevolved groups of biting and drilling predators isso pervasive that it might be overlooked. With theexception of tiny boreholes in some of the earliestsmall calcareous shelly organisms, Cloudina(Bengtson and Zhao, 1992), there is, as yet, noevidence of predators in the latest Proterozoic(Vendian). This observation led to the scenario ofa nearly predator-free, early “Garden of Ediacara”phase in Earth’s history (McMenamin, 1986;McMenamin and McMenamin, 1990).The first wave of predation may have instigatedthe acquisition of hard skeletons by numerous taxaduring the Cambrian explosion (Bengtson, 1994;Conway Morris, 2001). The apparently “explosive”development of phosphatic and calcitic sclerites,valves, and armor in the Early Cambrian may wellhave been driven by biting organisms. The early andevolutionarily critical rise in Cambrian predators isreviewed by Babcock (in press), who proposes tocall this the “early Paleozoic marine revolution.”The appearance of skeletons was geologicallyrapid, probably encompassing no more than tenmillion years, and was one of the most dramaticepisodes of convergent evolution in the history of97

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002metazoans. Some nine phyla of animals, as wellas algae (receptaculitids), almost simultaneouslyacquired skeletal coverings of varied physiologicalorigins and compositions, including protein, chitin,silica, apatite, calcite, and aragonite.Babcock (1993) notes that the frequency ofhealed bite marks in trilobites actually declines inthe Late Cambrian and Early Ordovician;moreover, the proportion of right posterior bitemarks declines. This trend apparently coincideswith the disappearance of anomalocaridids.However, a host of new organisms appearing inthe middle Paleozoic ushered in a new wave ofpredation: the “Middle Paleozoic Revolution”.MIDDLE TO LATE PALEOZOICMARINE PREDATORSThe Ordovician brought on a further phase ofpredator escalation, but one that remainedsomewhat subdued until the middle Paleozoic.While large predators were present from Cambriantimes onward, a number of new marine predatorsappeared by the Middle Ordovician, includingasteroids, varied arthropods, larger cephalopods, andprobably drilling gastropods. Subsequently,durophagous predators showed an abrupt increasein the Devonian (Fig. 1). These included crustaceansand, most notably, several fish groups.Arthropods and their Traces.—Arthropods nodoubt continued to occupy predator guilds in themiddle to late Paleozoic. Further examples oftrilobite “hunting burrows” have been described (e.g.Brandt et al., 1995), and the morphology of enditesin larger trilobites, such as Isotelus and phacopids,suggests that these appendages served to grasp andperhaps masticate weakly skeletonized organisms.Durophagous arthropods of the early to middlePaleozoic include eurypterids and phyllocaridcrustaceans (Figs. 3.2–3.4). Eurypterids appear inthe Ordovician in marine environments and, duringthe Silurian, included some of the largest arthropodsthat have ever lived. Pterygotids with estimatedlengths in excess of four meters were also equippedwith formidable chelate chelicerae (Selden, 1984,1992) (Fig. 3.4). There seems little doubt that theseclaws were used in seizing and slicing prey.However, it is unlikely that these organisms dwelledin open marine environments. Indeed, theseeurypterids are most commonly associated withbrackish estuarine facies, suggesting that theyinhabited marginal marine environments (Selden,1984). Fossil associations suggest that othereurypterids, and perhaps non-marine vertebrates,may have formed a part of their diet.Phyllocarid crustaceans (Fig. 3.2) appeared inthe Cambrian, but diversified in the Devonian(Signor and Brett, 1984). Stout, molariform, andcalcified gastric teeth may have been utilized in preymastication. Unlike eurypterids, the phyllocaridswere relatively common in open marineenvironments, and may have preyed upon shellfish.Decapod crustaceans with claws for crushingprey appeared in the Devonian and diversified inthe later Paleozoic, but they were mainly small anduncommon. The ancestors of stomatopods probablydiverged from the rest of the malacostracans in theDevonian (Schram, 1982, 1984; Hof, 1998);paleostomatopods occupied nearshore habitats andare known from North America and Europe(Schram, 1977; Jenner et al., 1998). Primitivestomatopods that do not smash their prey, such ashemisquillids, can eat solitary corals, crabs, bivalves,and fish (Basch and Engle, 1989).Scalloped fractures of the outer lips of Paleozoicgastropod shells (Fig. 4.1) resemble marks made bymodern predatory crustaceans that “peel” gastropodshells to reach the body of the snail (Vermeij et al.,1981; Schindel et al., 1982; Ebbestad and Peel, 1997;Ebbestad, 1998). However, these peeled shells arenot attributed to durophagous crabs (which do notappear until mid Mesozoic times; see Walker andBrett, this volume) and, at present, the predatorremains unknown. This type of probable arthropodpeeling trace is known in shells from the MiddleOrdovician (Peel, 1984) onward, but is rare—generally < 7% of shells—in the early to middlePaleozoic (Schindel et al., 1982; Peel, 1984).The middle Paleozoic appears to have been atime of intensification of this type of interaction.Devonian and Carboniferous gastropod andammonoid shells show increased frequencies of98

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSshell repair, though slightly lower frequencies thanthose recorded for snails of comparable size in thelater Mesozoic (Vermeij et al., 1981; Schindel etal. 1982; Brett and Cottrell, 1982; Bond andSaunders, 1989).Cephalopods.—Large nautiloids, endoceratoids,and actinoceratoids, some exceeding five meters inlength, were abundant in marine benthic assemblagesfrom the Early Ordovician onward (Fig. 3.7). Allknown cephalopods are carnivorous (Nixon, 1988).By analogy with modern Nautilus, these nautiloidsprobably possessed chitinous beaks capable ofshearing skeletons (Alexander, 1986a; Saunders andWard, 1987). The oldest known cephalopod jawsare Carboniferous in age and are similar to Recentcoleoid jaws (Lehmann, 1988). Radulae ofcephalopods date back to the Silurian (Mehl, 1984),and are also similar to Recent coleoid radulae,suggesting stasis in feeding morphology from themiddle Paleozoic to the Recent for these groups.Alexander (1986a) attributed divots andcrescentic healed breakages in Ordovicanbrachiopod shells, especially strophomenides, tonautiloids; and Rudkin (1985) described aspecimen of the Late Ordovician trilobitePseudogygites with crescentic bite marks, whichhe attributed to an endoceratoid. Brunton (1966)and Elliot and Brew (1988) also noted predatoryfractures preserved on Carboniferous brachiopodsthat they attributed to nautiloid predation.Possible crop residues from large nautiloids inthe Ordovician contain abundant trilobite fragments(Brett, unpublished data). Kloc (1987) described apyritized coprolite from the Late Devonian that heattributed to a nautiloid, and Zangerl et al. (1969)reported possible nautiloid coprolites.In addition to nautiloids and ammonoids,coleoids first appeared in the Early Devonian(Lehmann, 1976), and recently a Carboniferous“octopod” has been reported from the Mazon Creekfossil Lagerstätte (Kluessendorf and Doyle, 2000).Because of their soft-bodied construction thesecephalopods have a very poor fossil record andtheir impact as predators is not known.Gastropods and Drilling Predation.—Moderngastropods of several families are voraciouspredators that use a combination of chemical andmechanical radular drilling to penetrate the shellsof their prey (Carriker and Yochelson, 1968;Carriker, 1969, 1981; Kabat, 1990), formingdistinctive bore holes termed Oichnus by Bromley(1981). For many years it was assumed that thistype of drilling was confined to meso- andcenogastropods, and Oichnus in Paleozoic shellswas ascribed to another type of unknown predator(Carriker and Yochelson, 1968; Smith et al., 1985).However, recent discoveries suggest thatplatyceratid archaeogastropods were also predatorydrillers. Baumiller (1990) and Baumiller et al.(1999) documented gastropod-like drill holesbeneath the shells of attached platyceratidgastropods on a crinoid and on a brachiopod shell.These intriguing cases, although possibly recordingparasitism, prove the capacity for radular drillingamong platyceratids.The family Platyceratidae spans the periodfrom the Middle Ordovician to the Late Permian(Bowsher, 1955); highly modified genera, such asPlatyceras itself (Fig. 3.5), were clearly commensal/parasitic on pelmatozoan echinoderms. However,others, notably Cyclonema (Ordovician-Silurian)and Naticonema (Ordovician-Devonian) (Fig. 3.6),retained unspecialized shells and may have beenfacultatively free-living scavengers and predators.Predatory drill holes provide direct evidencefor carnivory. It is important to use specific criteriato recognize drill holes in the fossil record, sincesubstrate borers or pressure dissolution can makeholes similar to drillings (Richards and Shabica,1969; Lescinsky and Benninger, 1994) (Figs. 4, 5).Kowalewski et al. (1998) used specific criteria torecognize predatory borings. First, completed drillholes are generally single and unhealed; second,drill hole position is consistently located over afood-rich area of the prey; third, there should beno attachment scars—such scars would indicatethat the drill holes were made by parasitic, ratherthan carnivorous, organisms. Kowalewski et al.(1998) also suggested that the ratio of inner to outerdiameter of successful beveled borings shouldexceed 0.5, as in most modern predatory drill holes.However, this criterion does not apply to cylindrical99

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002borings of the type made by muricid gastropods.Drill holes of at least two distinctive typesoccur in middle to late Paleozoic shells (Ausichand Gurrola, 1979). These were termed Type A andType B boreholes and they are morphologicallysimilar to the later muricid and naticid gastropoddrill holes, respectively (see Figs. 4.3, 4.4). TypeA drill holes (Figs. 5.1, 5.2) are smaller, cylindrical,and may penetrate shells from below; they mayFIGURE 4—Traces of predation, in fossil and Recentshells. 1—Shell of Devonian gastropodPalaeozygopleura with sublethal healed fracture ofthe outer lip. 2—Permian bivalve shell with healedcrescentic fractures along valve margin; probably theresult of attempted predation by a fish. 3—Incompletebore hole of Polinices duplicata; note raised boss atcenter. 4—Complete drill hole of Natica severa; notebeveled outer margin. Redrawn from photographsin the following sources: 1, Brett and Cottrell (1982),2, Boyd and Newell (1972); 3, 4, Carriker andYochelson (1968). Figure modified from Brett (1992).record attacks by parasitic organisms (Ausich andGurrola, 1979). However, Type A drill holes occuras a single drill hole per shell and display a nonrandomstereotyped pattern characteristic ofcarnivores (Fig. 5.6) (Smith et al., 1985; Leighton,2001a, b). Cylindrical holes are known from LateOrdovician (Cincinnatian) brachiopods (Bucher,1938; Cameron, 1967). Some of these have provento be domichnial borings (Trypanites) made in deadshells (Carriker and Yochelson, 1969; Richards andShabica, 1969). But Kaplan and Baumiller (2000)argued recently that at least some of these holesshow non-random positioning, and hence wereprobably produced by predatory organisms.Rohr (1976) observed prey and site selectivityof small boreholes in Silurian orthid brachiopods.Liljedal (1985) also noted Type A borings insilicified Silurian bivalves. Similarly, Type Aborings occur in about 11% of the Early Devonianbrachiopod Discomyorthis and show evidence ofsize and site selectivity on the prey shells (Sheehanand Lespérance, 1978). Buehler (1969) reported alow frequency (2.25%) of cylindrical borings inMiddle Devonian shells, as did Rodriguez andGutschick (1970). However, the jury is still out onthe issue of whether these were predatory or merelyparasitic in nature (Leighton, 2001a, b).Type B boreholes are parabolic, 1–3 mm indiameter, and display a chamfer or bevel;incomplete boreholes possess a central raised knobor boss (see Figs. 4.3–5.5). These most closelyresemble drillings of modern naticid gastropods.These boreholes first become common in Devonianbrachiopods (Fenton and Fenton, 1931, 1932;Smith et al., 1985; Kowalewski et al., 1998); earlierpossible examples are known from Ordovicianbrachiopods (S. Felton, pers. comm.) but have notbeen documented in the literature. Brunton (1966)reports frequencies of up to 30% of brachiopodsdrilled with this type of hole in assemblages of lateCarboniferous age. Relatively few typical Type Bborings are reported from the Upper Carboniferousto Permian (see Kowalewski et al., 2000).However, a series of papers document small (< 2mm) boreholes with chamfering, which shouldperhaps be assigned to a third category; these occur100

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSFIGURE 5—Gastropod-like boreholes in brachiopods from the Middle Devonian Hamilton Group ofNew York State. Note scale bars. 1, 2—Type A borehole on the brachiopod Rhipidomella; 1 showsexterior view, 2, with negative reversed for comparison, shows blister of healed shell on interior of theshell. 3—Incomplete hole in Rhipidomella showing central boss. 4—Two overlapping sediment-filledboreholes on Parazyga; upper hole is complete; note chamfer (bevel) well displayed on lower hole.5—Incomplete borehole in Douvillina showing central raised boss; note truncated pseudopunctae ofshell. 6—positions of drill holes on the brachiopod Rhipidomella showing stereotypy of positioning overmain visceral mass of brachiopod. Modified from Smith et al. (1985).primarily in diminutive brachiopods belonging tothe family Cardiarinidae (Cooper, 1956; Bassett andBryant, 1993; Grant 1988; Morris, 1994). Hoffmeisteret al. (2001a) report drilling frequencies of up to33% in Cardiarina. The boreholes show stereotypywith respect to valves and preferred site on shells.Until recently, it was quite unclear whatorganisms were responsible for Type B boreholes,but the discovery of platyceratid gastropods in directassociation with this type of drill hole on LowerCarboniferous crinoids suggests that these snailswere among the culprits (Baumiller, 1990, 1996;Baumiller et al., 1999). Several studies have shownthat Type B hole-drillers display a distinct preferencefor particular prey taxa, notably athyrid and certainstrophomenid brachiopods. They also show101

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002stereotypical positioning on valves (see Fig. 4.5),as is typical for predatory gastropods (Ausich andGurrola, 1979; Smith et al., 1985; Leighton, 2001a).As with durophagous predation, there is evidencefor intensification of shell drilling in the Devonian(Kowalewski et al., 1998), although the fossil recordof platyceratids shows relatively little increaseduring this time. Initially, it was thought that thefrequency of drilling declined in the late Paleozoic(Kowalewski et al. 1998), and this seems to besupported in some cases (Kowalewski et al., 2000;Hoffmeister et al., 2002). However, for individualspecies data, drilling frequencies can be similar tothose reported for the Late Cretaceous to Cenozoic—which can exceed 30% per species examined (Ausichand Gurrola 1979; Hoffmeister et al. 2001a, b).Asteroids.—Asteroids evidently developed theirnotoriously predaceous habits early in their history(Blake and Guensburg, 1992). Middle Ordovicianstarfish have been found with gastropod shells intheir gut cavities (Spencer and Wright, 1966). Stillolder possible examples of starfish predation areknown from as early as the Whiterockian (MiddleOrdovician; G. C. Baird, pers. comm.).There is controversy as to whether or not starfishdeveloped extraoral digestion in the Paleozoic (seeDonovan and Gale, 1990). However, Blake andGuensburg (1994) describe an OrdovicianPromopalaeaster in apparent feeding position on abivalve, a characteristic behavior related to extraoraldigestion. Similarly, Clarke (1921) illustratedprobable examples of starfish predation from theDevonian of New York, where specimens ofDevonaster apparently were overwhelmed bysediment while in feeding position on bivalves.Gnathostomes.—The earliest well-knownpredatory gnathostome fishes are Silurianacanthodians, although possible acanthodian spinesand chondrichthyan (shark) denticles are known fromthe Middle Ordovician (Benton, 1997). These fishesand their later Paleozoic descendants possessed sharpteeth with cutting plates adapted for predation on softto chitinous invertebrates and other fishes.The earliest major radiation of durophagous(shell-crushing) fishes undoubtedly occurred in theEarly to Middle Devonian. Varied placoderms,including rhenanids with blunt crushing plates andray-like benthic adaptation, and ptyctodonts withhypermineralized tritors, also evolved during theDevonian. The ptyctodonts and rhenanids mayhave been important crushers of hard-shelled prey(Figs. 5.1, 5.2), although their remains areuncommon in most marine invertebrate-richassemblages. Nonetheless, there are reports ofptyctodonts in normal marine shell beds (Moy-Thomas and Miles, 1971). Ptyctodonts, in fact, aremost commonly associated with fragmentaryremains of arthrodires. Their blunt, crushing teethmay have been adapted for cracking the armor ofarthrodires during scavenging.Placoderms became extinct by the end of theDevonian (Moy-Thomas and Miles, 1971), but werereplaced by varied sharks (Fig. 6). Especially duringthe Carboniferous and Permian, many types of sharksevolved, including the symmoriaformes, hybodontids,and ctenacanthoids, some of which developedbroadened teeth and were durophagous (Moy-Thomas and Miles, 1971; Mapes and Benstock, 1988)(Fig. 6). For example, Boyd and Newell (1972) reporta high percentage of Permian bivalves with divots inthe shells probably produced by sharks (see Fig. 4.2).Chimaeras or holocephalans (e.g., helodontoids,cochliodontoids, and petalodontids) possessedautostylic (fused) skulls and hypermineralized,crushing dentition analogous to that of earlierptyctodonts (Fig. 6.4). Certain Carboniferouschimaeras, such as Helodus, have been implicated asproducers of distinct crush marks in Carboniferousand Permian brachiopod and bivalve shells (Brunton,1966; Boyd and Newell, 1969; Alexander, 1981).Hansen and Mapes (1990) also reported crush marksin Upper Carboniferous nautiloids that theyattributed to the shark Petalodus (Fig. 3.7).Chimaeroids underwent a five-fold increase intaxonomic richness in the Carboniferous relative tothe Devonian (Mapes and Benstock, 1988). However,durophagous holocephalans also underwent a majordecline in the Upper Carboniferous and Permian(Mapes and Benstock, 1988).In addition, during the Carboniferous, deepbodiedchondrostean fishes of the Doryopteridaedeveloped well-defined tooth plates for crushing102

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSFIGURE 6—Middle Paleozoic predatory fishes. 1—Rhenanid, ray-like placoderm Gemuendina, EarlyDevonian. 2—Ptyctodont placoderm Ctenurella, Devonian. 3—Upper and lower dentition of holocephalanDeltoptychius, showing hypermineralized tritor grinding gnathal plate. 4—Holocephalan Helodus,Pennsylvanian. 5—Stethacanthus; note toothed brush structure, Mississippian. 6—Eugeneodontiformeshark Sarcoprion; tooth whorl in lower jaw opposes pavement in rostrum, Pennsylvanian. 7—Hybodus,hybodont shark. 1, from Moy Thomas and Miles, 1971; 2, 3, from Stensiö, 1969; 4, from Patterson, 1965;5, 6, from Moy-Thomas and Miles, 1971; 7, from Zangerl, 1981.hard-shelled prey. These reef-dwelling fishes showmany similarities to specialized reefal teleost fishesof the Cenozoic (Moy-Thomas and Miles, 1971;Benton, 1997).There is limited information concerning thecoprolites or gut contents of the shark group.However, the few trace fossils available reveal thatbenthic organisms formed a food source for someof these predatory sharks. For example, coprolitesand gut contents of Carboniferous-Permianholocephalans, sharks, and other fish containfragmented brachiopods and crinoid ossicles(Zangerl and Richardson, 1963; Malzahn, 1968;Moy-Thomas and Miles, 1971).103

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Pelagic Predators.—Although pelagicpredation by swimming arthropods or cephalopodsmay have occurred as early as the Cambrian, thefirst direct evidence for predator-prey relationshipsamong nektonic organisms is from the Devonian(Fig. 7). The rise of goniatite ammonoids in theDevonian may have impacted the pelagic ecosystem.These presumably predatory cephalopods arecommomly found in settings in which there is littleor no benthic fauna (e.g., black shales recordinganoxic sea floors). Hence, these organisms may havefed on pelagic organisms, such as swimmingcrustaceans, other cephalopods, conodont animals,and perhaps small fish (Lehmann, 1976, 1988).Many of the arthrodires, with sharp shearinggnathal plates, were undoubtedly piscivorous. Hlavin(1973, 1990) reports on an articulated specimen ofthe arthrodire Holdenius, from the Upper DevonianCleveland Shale, preserved adjacent to the remainsof its prey: a ctenacanth shark, which had been bittenin half (Fig. 8). This is direct evidence forpredation—although failed predation in this case:an anterior dorsal spine from the ctenacanth wasfound lodged in the palate and extending into thebraincase of the Holdenius. The arthrodire wasprobably killed instantly when it was impaled onthe spine of its prey (a lose-lose situation!).Other groups, such as cladodont sharks, withsharp, cusped teeth, clearly had an impact on fishand on certain probably pelagic invertebrate prey.Williams (1990) provides an excellent summaryof evidence for cladodont predation from fecalFIGURE 7—Time-line of appearance of major groups of Paleozoic pelagic predators and prey.104

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSFIGURE 8—Reconstruction of the arthrodireHoldenius attacking a ctenacanth shark; basedon specimen from Upper Devonian ClevelandShale, northern Ohio. Drawing by J. P. Lufkin;from Hlavin (1990).masses and gut residues in the body cavities ofmore than 50 well-preserved cladoselachid sharksfrom the Upper Devonian Cleveland Shale in Ohio.The most commonly preserved ingesta are scalesand bones of palaeoniscoid fishes (present in 64%of shark specimens)—with a few showing bothhead-first and tail-first swallowing orientations—followed by remains of the crustaceanConcavicaris (found in 28%). One unidentifiedcladoselachid had two ctenacanth shark spinesembedded in its jaw and at least two Cladoselachehave smaller specimens of Cladoselache in the gutcavity, indicating cannibalistic behavior in thispelagic predator. About 5% of the Cladoselachecontained conodonts, and all of the conodontbearingsharks (including one in the body cavityof a larger shark!) are small individuals, whichfurther suggests size partitioning of food resources.Ctenacanth sharks have been found with smallarthrodires in the gut cavity. In turn, the largerpalaeoniscoid osteichthyan fishes of the ClevelandShale also show evidence of pelagic predatorybehavior. Ironically, these osteichthyans have smallsharks and arthrodires in their gut cavities.Trace fossil evidence of attacks by the sharkSymmorium is also known from shells of UpperCarboniferous coiled nautiloids (Mapes and Hansen,1984; Hansen and Mapes, 1990). Shells of thenautiloid Domatoceras show punctures that matchthe spacing of tooth files in the associated shark(Fig. 9). Zangerl and Richardson (1963) and Zangerlet al. (1969) also report abundant evidence fromcoprolites, regurgitates, and gut contents for sharkpredation on other fishes preserved in the UpperCarboniferous Mecca Quarry Shale in Illinois.The Cleveland Shale and Upper Carboniferousshark-bearing shales generally lack benthic bodyor trace fossil assemblages, and were evidentlydeposited in anoxic bottom waters. Hence, thesecomplex food webs involved an entirely pelagiccommunity. Many of these early shark,osteichthyan, and arthrodiran predators may havehad little impact on marine benthic communities.DEVONIAN-PERMIAN:MIDDLE PALEOZOIC MARINEREVOLUTIONSignor and Brett (1984) explored severalPaleozoic adaptive trends that served to strengtheninvertebrate skeletons or make them more difficultto attack. They inferred that these trends were, atleast in part, a response to increased predationintensity during the middle Paleozoic “precursor”to the Mesozoic marine revolution. This term isperhaps inappropriate as it implies a preliminarybuild-up to the later revolution. In fact, we arguethat the two actually involved separate radiationsof predators and were separated by a majorreorganization of predator-prey interactions and105

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 9—Shark predation on Pennsylvanian nautiloids. Bottom—reconstruction of shark Symmoriumshown attacking a small coiled nautiloid. Top—sequence of events involved in producing punctures ina nautiloid shell rotated over tooth row; adapted from Mapes and Hansen (1990).other aspects of ecology as a consequence of thePermo-Triassic extinction. Hence, we herein referto the purported middle Paleozoic escalation phasesimply as the “Middle Paleozoic Revolution.”Possible responses (aptations sensu Gould andVrba, 1982; Vermeij, 1987; Gould, 2002) to predationpressure may follow two patterns: a) changes inbehavior and mode of life, and b) changes inmorphology. Both types of aptations can be inferredfor middle to late Paleozoic organisms. Theseresponses may represent direct adaptations toincreased predation pressure, such as thepreferential survival of spiny organisms, as wellas exaptions that spring from pre-existing skeletalfeatures that can be co-opted for a different function(Gould and Vrba, 1982; Gould, 2002). In a majorityof cases it is not possible to distinguish betweenthese modes of origin.Micro-architecture.—One such exaptation thatmay impede predation is shell micro-architecture. Forexample, micro-architecture such as pseudopunctaein brachiopod shells may help to prevent propagationof shell fractures (Alexander, 1986a, 2001). Perhapsthe advantage of this micro-architecture providesone of the reasons that pseudopunctate brachiopods,especially productids and chonetids, becomedominant in the late Paleozoic.Shell Architecture.—In some groups ofgastropods, the presence of an open umbilicusweakens the shells and makes them more easilycrushed (Vermeij, 1983, 1987). Therefore, one mightpredict a decline in umbilicate forms in the face ofincreasing predation pressure. In a sample of some60 genera of bellerophontids, Signor and Brett(1984) found a substantial decline in umbilicateforms, beginning in the Silurian Period (Fig. 10).Ribbing and fluting also render shells moreresistant to crushing; again Signor and Brett foundan increased incidence of sculpture in post-Silurian nautiloids. Alexander (1986b) observeda parallel decrease in the incidence of smoothimplicate shells and increase of coarse ribs in post-Silurian brachiopods. Similarly, he observed adeclining proportion of rectimarginate shells infavor of stronger uniplicate and strongly ribbedshells. Such morphological features could aid inthe resistance to shell breaking and crushingpredators. Strongly plicate brachiopods, such as106

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSrhynchonellides, also appear to have been avoidedby shell boring organisms, at least in the Devonian(Bordeaux and Brett, 1990).Fluted margins may also give shells tighterclosure against the prying action of starfish, as doesinterlocking hinge dentition, and central placementof adductor muscles (Alexander, 2001). LaterPaleozoic brachiopods and bivalves show increasedfrequency of fluting, interlocked hinge teeth, andcentralization of adductor muscle scars.Greater shell thickness could also have beenadvantageous during a time of increaseddurophagous predation, drilling predation, orFIGURE 10—Morphological trends in bellerophontidmollusks. Upper curve shows total number ofgenera; area under lower curve representsproportion of total genera that show indicatedfeature: (a) presence of sculpture; (b) anomphalous(lacking umbilicus); (c) disjunct coiling. Note loss ofgenera with disjunct coiling and increase inproportion of genera with sculpture and lacking anumbilicus. Redrawn from Signor and Brett (1984).parasitism. Leighton (2001b) also notes a tendencyfor brachiopods to develop thickened muscleplatforms in the most drill-prone centrally locatedshell areas. A trend toward increased plate thicknessis evident in late Paleozoic crinoids. CertainPermian taxa are extraordinarily thickly plated(Signor and Brett, 1984).Spinosity.—Spinose skeletons may deter bothdurophagous and drilling predators. An increasein the frequency of taxa with skeletal spines duringthe Paleozoic is documented by Signor and Brett(1984); spines may also increase in length andsharpness. Articulate brachiopods show a strongincrease in the presence of spines on both thepedicle and brachial valves, reflected in the rise todominance of the productides, in the later Paleozoic(Signor and Brett, 1984). Although the spines onthe deeply convex pedicle valve of productides mayhave served as “rooting” spines for these semiendofaunalbrachiopods (Grant, 1966; Rudwick,1970) (Fig. 10), they may also have been functionalin preventing predatory attack, particularly frombelow by infaunal predators. Leighton (2001a)showed that among Late Devonian brachiopods thespinose Devonoproductus had a much lowerfrequency of completed boreholes than eithercontemporaneous atrypids or Douvillina.Among gastropods, relatively few Paleozoicgenera (~5%) show spines; however, here theexceptions may prove the rule. No spinose generaare known from the lower Paleozoic and spinoseforms first appear in the Silurian. Moreover, all ofthe spinose gastropods are inferred to have beenrelatively sedentary. Notably, several species ofspinose platyceratids appear in the Devonian. Apermanently sessile commensalistic/parasiticlifestyle (Bowsher, 1955; Rollins and Brezinski,1988; Baumiller, 1990; Boucot, 1990) may haverendered these gastropods particularly vulnerableto predatory attack, and conferred a selectiveadvantage to species that evolved spines.Trilobites also show an abrupt, but short-lived,burst of spinosity during the Devonian. The wellknown and highly diverse trilobites from theEmsian-Eifelian of Morocco and North Americashow a high frequency of spinose genera in several107

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002lineages (Kloc, 1992). Kloc (1992, 1993) has alsodocumented the occurrence of possible camouflagestrategies in the Early Devonian selenopeltid trilobiteDicranurus. The elongate cephalic spines aretypically heavily bored and encrusted. Kloc suggeststhat these encrusted spines served to obscure theimage of the trilobite from visual predators in astrategy analogous to that of decorator crabs.Long spines on the calyces and tegmens ofcrinoids are reasonably interpreted as a deterrentto would-be predators. Therefore, it is significantthat no crinoids display spinose calyces prior tothe Wenlock (Silurian), when Calliocrinus displayslarge tegminal spines (Signor and Brett, 1984).Both camerate and cladid crinoids in severalfamilies show a substantial increase in theproportion of spinose genera commencing in EarlyDevonian time (Fig. 11). The proportion of spinosegenera increases to a maximum in Visean time andthen declines in the late Paleozoic in concert withthe decline of camerate crinoids during theChesterian crisis identified by Ausich et al. (1994).Other crinoids, primarily Devonian-LowerCarboniferous camerates, but also a few latePaleozoic cladids, developed elongate spines onthe calyx (Fig. 12). A few genera developed spinoseFIGURE 11—Spinosity in brachiopods. 1—Reconstruction of the productid brachiopod Waagenoconchafrom the Permian of Russia; note juveniles attached to algae by “clasping spines,” and quasi-infaunal modeof life, with “rooting” spines in adults, ×1; from Grant (1966). 2—Brachiopod genera, primarily productides,with spines on the pedicle or both valves. Both show consistent trends; from Signor and Brett (1984).108

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSplates on the axillaries of the arms. Arthroacantha,a very common and widespread Devoniancamerate, possessed articulated spines on the calyx,as well as spines on the arms (Fig. 12.1) (Keslingand Chilman, 1975). Within this genus there is alsoa trend of increasing spine length into the LateDevonian (G.C. McIntosh, pers. comm., 2001).Aronson (1991) argued that if predationpressure were a significant factor in crinoidcommunities a major decline in crinoid thicketswould be expected between pre-Devonian andCarboniferous benthic assemblages. Thisprediction was based in part on evidence thatstalked crinoids migrated offshore in the face ofthe Mesozoic marine revolution of predators(Meyer, 1985). Aronson made corrections fordifferences in rock volume of various ages andpredicted the frequency of dense crinoidassemblages for each age. He found that densecrinoid thickets did not, in fact, show a declineduring this interval. This provides negativeevidence for the escalation hypothesis and mightsuggest that predation pressure was not, in fact, amajor factor in controlling crinoid density.Alternatively, Aronson suggested that the generallack of reefs in the Lower Carboniferous caused aFIGURE 12—Spinosity in Paleozoic crinoids. 1—Reconstruction of Devonian crinoid Arthroacanthawith attached (coprophagous) Platyceras gastropod, ×2; note jointed spines on calyx and spines onaxillaries of arms. 2—Percentages of spinose genera in three subclasses of crinoids through thePaleozoic Era. 1, Modified from Kesling and Chilman (1975); 2, from Signor and Brett (1984).109

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002decline in specialized reef-dwelling fish predators.This is questionable since most known Devoniangnathostome fish fossils are not associated withreefs, but occur rather in open marine settings. Onemight, alternatively, suggest that crinoids were ableto adjust, up to a point, to the increased pressureof grazing by sharks, holocephalans, and otherfishes. The rise in spinosity and plate thickness mayhave been effective temporarily in preventingdecimation by predators.Surprisingly, all crinoid subclasses exhibitdecreased spinosity in the late Paleozoic, followinga Lower Carboniferous high (Signor and Brett,1984). Waters and Maples (1991) suggest thatpredators were able to “keep up” with the armamentsof their prey and that spinose plates becameineffective as a defensive strategy; smaller size andcompactness of the calyx may then have been moreeffective strategies. This trend toward smaller sizesmay have other meanings, such as declining foodresources, although no correlations are obvious.A majority of the common Devonianplatyceratid host crinoids were spiny, and nearlyall spiny crinoids were at least occasional hosts ofplatyceratids. In contrast, none of the commonOrdovician or Silurian crinoid or cystoid hosts werespiny. Obviously, the spines were not a deterrentto platyceratids. Arthroacantha, the most widelycited host genus (with populations showing up to70% individual infestation by Platyceras)possessed both movable spines on the calyx andaxillary spines on the arms. Intriguingly, Platycerasdumosum, one of the common symbionts, alsopossessed long spines. Brett (in press) suggests thatthe development of spines in crinoids was an antipredatoryadaptation mediated by the presence ofgastropods. Even if crinoids were not tasty prey(as has been suggested by some modern studies:Meyer and Ausich, 1983), gastropods may havebeen. If gastropod-bearing crinoids were frequently“targeted” by durophagous predators, they mayhave experienced a higher selection pressure toevolve spines (as did the gastropods themselves)than did non-host crinoids.Life Habit Changes.—Vertebrates typicallyshow a pattern of decreasing skeletal armor duringthe late Paleozoic. Early agnathan “ostracoderms”and placoderms were heavily encased in dermalbone. Dermal bone may have served non-defensivefunctions, such as areas for muscle attachment andphosphate sinks. However, it is also probable thatthis armor deflected predatory attack, especiallyfrom contemporary invertebrate predators.Ironically, it may be the preference of these earlyvertebrates for marginal marine environments thatfostered escalation, as these environments werealso home to large predaceous eurypterids.Subsequently, the rise of gnathostome fishesmust have placed additional predation pressure onother vertebrates. It is perhaps surprising that someof the largest predators of the Devonian—thearthrodires—had heavily armored heads. This mayreflect the evolution of still more effective, fasterswimmingsharks, or it may merely reflect anothertype of adaptation—possibly for phosphateexchange—unrelated to predation. In any case,heavy dermal armor was largely lost with theextinction of placoderms. The successful predatorsof the later Paleozoic probably reduced armor asan adaptation for increased maneuverability andrapid swimming. This apparently was a highlysuccessful tradeoff. The appearance of varied finspines in sharks and peculiar spine and brush“headgear” in the stethacanthids (Fig. 6.5) mayrepresent anti-predatory or sexually selectiveadaptations (Zangerl, 1981).Life habit changes among Paleozoic organismsinclude the development of endobenthic andcemented modes of life. Semi-endofaunal(frequently termed quasi-infaunal in earlierliterature) habits were adopted by many orthide andstrophomenide brachiopods as early as EarlyOrdovician time, but the proportion of semiendobenthicbrachiopods increased in the latePaleozoic with the rise to dominance of productidesand chonetid brachiopods (Thayer, 1983) (Figs. 11,13). This change coincides with the middle to latePaleozoic revolution of predators.Bottjer (1985) related increasing intensity ofpredation to the progressive occupation by bivalvesof successively deeper endobenthic tiers (Ausichand Bottjer, 1982). Endobyssate and shallow-110

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSburrowing bivalves occupied an upper endofaunaltier, 0 to -6 cm (i.e., at and up to 6 cm below thesediment–water interface), from Cambrian timesonward. Invasion of an intermediate (-6 to -12 cm)tier by venerid and pholadomyid bivalves occurredlater during the Devonian; pholadomyids pusheddownward into the deep, -12 to -100 cm, tierslightly later in the Lower Carboniferous. Bottjer(1985) attributes this increased tiering to the mid-Paleozoic escalation of predators. Deependobenthic modes of life were limited during thePaleozoic by the absence of mantle fusion and lackof true siphons in most bivalve clades (Stanley,1970, 1977) (Fig. 13).Cementation of shells to hard substrates alsomakes them harder to dislodge by durophagouspredators (Harper, 1991). Alexander (2001) notesthat two major groups of brachiopods (productidesand orthotetaceans) show an increased frequencyof cemented forms during the later Paleozoic (Fig.13). During the Carboniferous, pseudomonotidbivalves also adopted a cemented mode of life andevolved shell spines, presumably in response toincreased predation pressure.SUMMARYThere is a growing body of evidence thatpredation on hard-shelled marine organismsintensified during the middle Paleozoic (Fig. 13).The direct fossil record of potential predators showsa substantial increase in durophagous shellcrushingpredators, as well as pelagic pursuit andambush piscivorous predators. Trace fossilevidence provides a strong case for the existenceof predatory attack on shelled organisms as earlyas the Cambrian.Predation in marine environments evolvedthrough several phases of intensification withminor setbacks following mass extinctions(Fig. 13). The first phase might be termed theCambrian Revolution. In this phase, largepredators, such as anomalocaridids, otherenigmatic arthropods, and perhaps trilobites, firsthad an impact on marine communities.A middle Paleozoic phase of predationintensification, emphasized by Signor and Brett(1984), involved the rise of nautiloid andammonoid cephalopods, phyllocarids, asteroids,and several lineages of gnathostome fishes. Thelatter group in particular radiated rapidly duringthe Devonian to produce diverse durophagous andpiscivorous placoderms and sharks. Major LateDevonian extinctions terminated the placoderms,but their guilds were rapidly replaced by evolvingsharks, holocephalans, and bony fishes.Brachiopod and molluscan shells and trilobiteexoskeletons show evidence of healed bite marksand peeling from the Cambrian onward, but with amarked increase in frequency in the later Paleozoic.Predatory drill holes with stereotypical position andprey-species preference are found in brachiopods(Cambrian onward) and mollusks (Ordovicianonward), but boreholes also show increasedfrequency in the middle Paleozoic.The Permo-Triassic extinction crisisconstituted a major setback for all marinecommunities. This certainly included manypredatory taxa (e.g., many ammonoids, nautiloids,phyllocarids, predatory archeogastropods).However, it is likely that certain active predatoryorganisms (e.g., fishes) were not as stronglyaffected as others. Studies by Knoll et al. (1996)note that many sedentary benthic organisms (e.g.,brachiopods, echinoderms) have a lower capacityfor controlling CO 2concentrations than do some“high energy” organisms, including activepredatory arthropods and vertebrates. Theseauthors postulate preferential extinction of manyfilter-feeding invertebrates during an interval ofhypercapnial stress.Drilling predation appears to have beencommon in the Paleozoic (Kowalewski et al.,1998). The evolution of Paleozoic drilling actuallyoccurs in two phases: a Precambrian to Silurianstage, and a Silurian to Carboniferous phase(Kowalewski et al., 1998, their figure 3). Drillingwas globally widespread in the Permian, but thefrequency of drilling bivalve or brachiopod preywas relatively low (Kowalewski et al., 2000). Asmore data are added to our knowledge of drillingpredation, it appears that there is not an everincreasingescalatory trajectory through the111


BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSPhanerozoic of drilling predation (as depicted inVermeij, 1987); rather, it appears in phases relatedto the predators that evolved during that time.Predator-prey interactions were probably offundamental importance in shaping and directinglong-term trends by evolutionary adaptation andcooptation (Vermeij, 1977, 1987). The response ofbenthic organisms to the Cambrian rise of predatorsmay be one of the most significant events in thehistory of life: the nearly synchronous evolutionof sclerotized and biomineralized armor and theappearance of an abundant skeletal fossil record.Brachiopods, gastropods, trilobites, andcrinoids, among others, show an abrupt increasein spinosity in the Middle Devonian and LowerCarboniferous. There are also possible patterns ofincreased spinosity along latitudinal gradients inthe Carboniferous. But spinosity decreases afterthe early Carboniferous. Late Paleozoic forms mayhave taken refuge in smaller size and resistant,thick-walled skeletons.Hard-shelled organisms may have respondedto crushing and drilling predation by evolving avariety of thicker, more spinose skeletons.Although escalation is sometimes cast as anongoing “arms race,” in actuality, escalation ofpredator-prey relationships may have developed ina series of incremental steps during episodes ofabrupt biotic reorganization punctuating longerinterludes of relative stability.ACKNOWLEDGMENTSWe dedicate this paper to Richard Bambach, agourmand of life’s history. We greatly appreciatethe efforts of Michal Kowalewski and Tricia Kelleyin putting together this volume and allowing us toproduce this synthesis. Glenn Storrs, WilliamMiller, III, Alan Hoffmeister, and an anonymousreviewer provided useful reviews that improvedthis paper. Finally, we thank Ruth Mawson andPeter Cockle, Macquarie University, Australia, forfacilitating the international editing of this paper.REFERENCESALEXANDER, R. R. 1981. Predation scars preserved in Chesterian brachiopods: Probable culprits and evolutionaryconsequences for the articulates. Journal of Paleontology, 55:192–203.ALEXANDER, R. R. 1986a. Resistance to and repair of shell breakage induced by durophages in Late Ordovicianbrachiopods. Journal of Paleontology, 60:273–285.ALEXANDER, R. R. 1986b. Frequency of sublethal shell-breakage in articulates through geologic time, p. 159–166.In P. R. Racheboeuf and C. Emig (eds.), Les Brachiopodes Fossiles et Actuels. Université de BretagneOccidentale. Biostratigraphie du Paleozoique.ALEXANDER, R. R. 2001. Functional morphology and biomechanics of articulate brachiopod shells, p. 145–170. InS. J. Carlson and M. Sandy (eds.), Brachiopods Ancient and Modern: A Tribute to G. Arthur Cooper.Paleontological Society Papers, 7.ARONSON, R. B. 1991. Escalating predation on crinoids in the Devonian: Negative community-level evidence.Lethaia, 24:123–128.AUSICH, W. I., AND D. J. BOTTJER. 1982. Tiering in suspension-feeding communities on soft substrata throughoutthe Phanerozoic. Science, 216:173–174.←FIGURE 13—Ranges of possible anti-predatory traits in various benthic marine invertebrates. Thin lines:present, but minor; thick lines: abundant; broken lines: possibly present, but rare. Abbreviations: Burrow:burrowing; Carb: Carboniferous; Miss: Mississippian; Penn: Pennsylvanian; Semi: semi-endobenthic.113

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002AUSICH, W. I., AND R. A. GURROLA. 1979. Two boring organisms in Lower Mississippian community of southernIndiana. Journal of Paleontology, 53:335–344.AUSICH, W. I., T. W. KAMMER, AND T. K. BAUMILLER. 1994. Demise of the middle Paleozoic crinoid fauna: A singleextinction event or rapid faunal turnover. Paleobiology, 20:345–361.BABCOCK, L. E. 1993. Trilobite malformations and the fossil record of behavioral symmetry. Journal of Paleontology,67:217–229.BABCOCK, L. E. In press. Trilobites in Paleozoic predator-prey systems and their role in reorganization of earlyPaleozoic ecosystems. In P. H. Kelley, M. Kowalewski, and T. Hansen (eds.), Predator-Prey Interactions inthe Fossil Record. Topics in Geobiology, Kluwer Academic/Plenum Publishers, New York.BABCOCK, L. E., AND R. A. ROBISON. 1989. Preferences of Paleozoic predators. Nature, 337:695–696.BAMBACH, R. K. 1993. Seafood through time: Changes in biomass, energetics, and productivity in the marineecosystem. Paleobiology, 19:372–397.BANDEL, K. 1985. Cephalopod morphology and function, p.190. In D. J. Bottjer, C. S. Hickman, and P. D. Ward(eds.), Mollusks: Notes for a Short Course. University of Tennessee Department of Geological Sciences,Studies in Geology 13.BASCH, L. V., AND J. M. ENGLE. 1989. Aspects of the ecology and behaviour of the stomatopod Hemisquillaensigera californiensis (Gonodactyloidea: Hemisquillidae), p. 199–212. In E. Ferrero (ed.), Biology ofStomatopods. Selected Symposia and Monographs, U. Z. I., volume 3.BASSETT, M. G., AND C. BRYANT. 1993. The micromorphic rhynchonelloidean brachiopod Lambdarina from thetype Dinantian. Journal of Paleontology, 67:518–527.BAUMILLER, T. K. 1990. Non-predatory drilling of Mississippian crinoids by platyceratid gastropods. Palaeontology,33:743–748.BAUMILLER, T. K. 1996. Boreholes in Devonian blastoids and their implications for boring platyceratids. Lethaia, 26:41–47.BAUMILLER, T. K., L. R. LEIGHTON, AND D. L. THOMPSON. 1999. Boreholes in Mississippian spiriferide brachiopods and theirimplications for Paleozoic gastropod drilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 147:283–289.BENGTSON, S. 1994. The advent of animal skeletons, p. 412–425. In S. Bengtson (ed.), Early Life on Earth.Columbia University Press, New York.BENGTSON, S., AND Y. ZHAO. 1992. Predatorial borings in late Precambrian mineralized exoskeletons. Science,257:367–369.BENTON, M. J. 1997. Vertebrate Palaeontology, 2nd edition. Chapman and Hall, London, New York, 452 p.BERGSTRÖM, J. 1973. Organisation, life and systematics of trilobites. Fossils and Strata, 21:1–69.BISHOP, G. A. 1975. Traces of predation, p. 267–281. In R. Frey (ed.), The Study of Trace Fossils. Springer-Verlag.BLAKE, D. B., AND T. E. GUENSBURG. 1992. Predatory asteroids and the fate of articulate brachiopods. Lethaia, 23:429–430.BLAKE, D. B., AND T. E. GUENSBURG. 1994. Predation by the Ordovician asteroid Promopalaeaster on a pelecypod.Lethaia, 27:235–238.BOND, P. N., AND W. B. SAUNDERS. 1989. Sublethal shell injury and shell repair in Upper Mississippian ammonoids.Paleobiology, 15:414–428.BORDEAUX, Y. L., AND C. E. BRETT. 1990. Substrate specific associations on Middle Devonian brachiopods:Implications for paleobiology. Historical Biology, 4:203–220.BOTTJER, D. J. 1985. Bivalve paleoecology, p. 112–137. In D. J. Bottjer, C. S. Hickman, and P. D. Ward (eds.), Mollusks:Notes for a Short Course. University of Tennessee Department of Geological Sciences, Studies in Geology 13.BOUCOT, A. J. 1981. Principles of Benthic Marine Paleoecology. Academic Press, NewYork, 463 p.BOUCOT, A. J. 1990. Evolutionary Paleobiology of Behavior and Coevolution. Elsevier, Amsterdam, 725 p.BOWSHER, A. L. 1955. Origin and adaptation of platyceratid gastropods. University of Kansas PaleontologicalContributions, Mollusca, 5:1–11.BOYD, D. W., AND N. D. NEWELL. 1972. Taphonomy and diagenesis of a Permian fossil assemblage from Wyoming.Journal of Paleontology, 46:1–17.BRANDT, D. S., D. L. MEYER, AND P. B. LASK. 1995. Isotelus (Trilobita) “hunting burrow” from Upper Ordovicianstrata, Ohio. Journal of Paleontology, 69:1079–1083.BRETT, C. E. 1992. Predation, 4.13.1, Marine, p. 368–372. In D. E. G. Briggs and P. W. Crowther (eds.),Palaeobiology: A Synthesis. Blackwell Scientific, Oxford.114

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSBRETT, C. E. In press. Durophagous predation on Paleozoic marine benthic assemblages. In P. H. Kelley, M.Kowalewski, and T. Hansen (eds.), Predator-Prey Interactions in the Fossil Record. Topics in Geobiology,Kluwer Academic/Plenum Publishers, New York.BRETT, C. E., AND J. F. COTTRELL. 1982. Substrate specificity of the Devonian tabulate coral Pleurodictyum.Lethaia, 15:248–263.BRIGGS, D. E. G., D. H. ERWIN, AND F. J. COLLIER. 1994. The Fossils of the Burgess Shale. Smithsonian InstitutionPress, Washington, D.C. and London.BROMLEY, R. G. 1981. Concepts in ichnotaxonomy illustrated by small, round holes in shells. Acta GeologicaHispanica, 16:55–64.BRUNTON, H. 1966. Predation and shell damage in a Visean brachiopod fauna. Palaeontology, 9:355–359.BRUNTON, C. H. C., AND C. CHAMPION. 1974. A Lower Carboniferous brachiopod fauna from the Manifold Valley,Staffordshire. Palaeontology, 17:107–111.BUCHER, W. H. 1938. A shell-boring gastropod in a Dalmanella bed of upper Cincinnatian age. American Journalof Science, 36:1–7.BUEHLER, E. J. 1969. Cylindrical borings in Devonian shell. Journal of Paleontology, 43:1291.CAMERON, B. 1967. Oldest carnivorous gastropod borings found in Trentonian (Middle Ordovician) brachiopods.Journal of Paleontology, 41:147–150.CARRIKER, M. 1969. Penetration of calcium carbonate substrates by lower plants and invertebrates. AmericanZoologist, 9:629–1020.CARRIKER, M. R. 1981. Shell penetration and feeding by naticacean and muricacean predatory gastropods: asynthesis. Malacologia, 20:403–422.CARRIKER, M. J., AND E. L. YOCHELSON. 1968. Recent gastropod boreholes and Ordovician cylindrical borings. USGeological Survey Professional Paper, B593:1–26.CHATTERTON, B., AND H. L. WHITEHEAD. 1987. Predatory borings in inarticulate brachiopods Artiotrema from theSilurian of Oklahoma. Lethaia, 20:67–74.CLARKE, J. M. 1921. Organic dependence and disease: Their origin and significance. New York State MuseumBulletin 221/222. 113 p.CLARKSON, E. N. K. 1996. Invertebrate Palaeontology and Evolution, 3rd edition. Chapman and Hall, London, 434 p.CONNELL, J. H. 1970. A predator-prey system in the marine intertidal region: I Balanus glandula and severalpredatory species of Thais. Ecological Monographs, 40:49–78.CONWAY MORRIS, S. 1977. Fossil priapulid worms. Special Papers in Palaeontology, 20:1–95.CONWAY MORRIS, S. 2001. Significance of early shells, p. 31–40. In D. E. G. Briggs and P. R. Crowther (eds.),Palaeobiology II. Blackwell Science, Oxford.CONWAY MORRIS, S., AND S. BENGTSON. 1994. Cambrian predators: Possible evidence from boreholes. Journal ofPaleontology, 68:1–23.CONWAY MORRIS, S., AND R. J. F. JENKINS. 1985. Healed injuries in Early Cambrian trilobites from South Australia.Alcheringa, 9:167–177.CONWAY MORRIS, S., AND R. A. ROBISON. 1986. Middle Cambrian priapulids and other soft-bodied fossils fromUtah and Spain. University of Kansas Paleontological Contributions, Paper 117:1–17.CONWAY MORRIS, S., AND R. A. ROBISON. 1988. More soft-bodied animals and algae from the Middle Cambrian ofUtah and British Columbia. University of Kansas Paleontological Contributions, Paper 122:1–48.COOPER, G. A. 1956. New Pennsylvanian brachiopods. Journal of Paleontology, 30:521–530.DONOVAN, S. T., AND A. S. GALE. 1990. Predatory asteroids and the decline of the articulate brachiopods. Lethaia, 23:77–86.EBBESTAD, J. O. R. 1998. Multiple attempted predation in the Middle Ordovician gastropod Bucania gracillima.Geologiska Förenings I Stockholm Forhandlingar, 120:27–33.EBBESTAD, J. O. R., AND J. S. PEEL. 1997. Attempted predation and shell repair in Middle and Upper Ordoviciangastropods from Sweden. Journal of Paleontology, 7:1007–1019.ELLIOT, D. K., AND D. C. BREW. 1988. Cephalopod predation on a Desmoinesian brachiopod from the NacoFormation, central Arizona. Journal of Paleontology, 62:145–147.FENTON, C. L., AND M. A. FENTON. 1931. Some borings of Paleozoic age. American Midland Naturalist, 12:522–528.FENTON, C. L., AND M. A. FENTON. 1932. Orientation and injury in the Devonian genus Atrypa. American MidlandNaturalist, 13:63–70.115

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FORTEY, R. A., AND R. M. OWENS. 1999. Feeding habits in trilobites. Palaeontology, 42:429–465.GOULD, S. J. 1989. Wonderful Life: The Burgess Shale and the Nature of History. Norton, New York, London, 347 p.GOULD, S. J. 2002. The Structure of Evolutionary Theory. Belknap-Harvard Press, 1433 p.GOULD, S. J., AND E. S. VRBA. 1982. Exaptation—a missing term in the science of form. Paleobiology, 8:4–15.GRANT, R. E. 1966. Spine arrangement and life habits of the productoid brachiopod Waagenoconcha. Journal ofPaleontology, 40:1063–1069.GRANT, R. E. 1988. The Family Cardiarinidae (Late Paleozoic rhynchonellid Brachiopoda). SenckenbergianaLethaia, 69:121–135.HANSEN, M. C., AND R. H. MAPES. 1990. A predator-prey relationship between sharks and cephalopods in the late Paleozoic,p. 189–192. In A. J. Boucot, Evolutionary Paleobiology of Behavior and Coevolution. Elsevier, Amsterdam.HÄNTZSHEL, W., F. EL-BAZ, AND G. C. AMSTUTZ. 1968. Coprolites: An annotated bibliography. Memoir of theGeological Society of America, 108:1–132.HARPER, E. M. 1991. The role of predation in the evolution of cementation in bivalves. Palaeontology, 34:455–460.HLAVIN, W. J. 1973. A review of the vertebrate fauna of the Upper Devonian (Famennian) Cleveland Shale:Arthrodira. Unpublished MA Thesis, Boston University, Boston, Mass.HLAVIN, W. J. 1990. Arthrodire-ctenacanth shark, p. 192–195. In A. J. Boucot, Evolutionary Paleobiology ofBehavior and Coevolution. Elsevier, Amsterdam.HOF, C. H. J. 1998. Fossil Stomatopods (Crustacea: Malacostraca) and their phylogenetic impact. Journal ofNatural History, 32:1567–1576.HOFFMEISTER, A. P., M. KOWALEWSKI, R. K. BAMBACH, AND T. K. BAUMILLER. 2001a. Intense drilling predation onthe brachiopod Cardiarina cordata (Cooper, 1956) from the Pennsylvanian of New Mexico. GeologicalSociety of America Abstracts with Programs, 33:A248.HOFFMEISTER, A. P., M. KOWALEWSKI, R. K. BAMBACH, AND T. K. BAUMILLER. 2001b. Evidence for predatory drillingin late Paleozoic brachiopods and bivalve mollusks from West Texas. North American PaleontologicalConvention, Paleobios, 21 (Supplement 2):66–67.HOFFMEISTER, A. P., M. KOWALEWSKI, R. K. BAMBACH, AND T. K. BAUMILLER. 2002. A boring history of drilling predationon thePaleozoic brachiopod Composita. Geological Society of America Abstracts with Programs, 34:A116.JENNER, R. A., C. H. J. HOF, AND F. R. SCHRAM. 1998. Palaeo- and archaeostomatopods (Hoplocarida: Crustacea) from theBear Gulch Limestone, Mississippian (Namurian), of Central Montana. Contributions to Zoology, 67:155–185.KABAT, A. R. 1990. Predatory ecology of naticid gastropods with a review of shell boring predation. Malacologia,32:155–103.KAPLAN, P., AND T. K. BAUMILLER. 2000. Taphonomic inferences on boring habit in the Richmondian Onniellameeki Epibole. Palaios, 15:499–510.KESLING, R.V., AND R. B. CHILMAN. 1975. Strata and Megafossils of the Middle Devonian Silica Formation.University of Michigan Museum of Paleontology, Papers on Paleontology 8, 408 p.KLOC, G. J. 1987. Coprolites containing ammonoids from the Devonian of New York. Geological Society ofAmerica, Abstracts with Programs, 19:23.KLOC, G. J. 1992. Spine function in the odontopleurid trilobites Leonaspis and Dicranurus from the Devonian ofOklahoma. North American Paleontological Convention Abstracts and Program. Paleontological SocietySpecial Publication 6:167.KLUESSENDORF, J., AND P. DOYLE. 2000. Pohlsepia mazonensis, an early ‘octopus’ from the Carboniferous of Illinois,USA. Palaeontology, 43:919–926.KNOLL, A. S., R. K. BAMBACH, J. P. GROTZINGER, AND D. CANFIELD. 1996. Comparative Earth history and LatePermian mass extinction. Science, 273:452–457.KOWALEWSKI, M., A. DULAI, AND F. T. FÜRSICH. 1998. A fossil record full of holes: The Phanerozoic history ofdrilling predation. Geology, 26:1091–1094.KOWALEWSKI, M., M. W. SIMÕES, F. F. TORELLO, L. H. C. MELLO, AND R. P. GILARDI. 2000. Drill holes in shells ofPermian benthic invertebrates. Journal of Paleontology, 74:532–543.LEHMANN, U. 1976. The Ammonites: Their Life and Their World. Cambridge University Press, Cambridge, 245 p.LEHMANN, U. 1988. On the dietary habits and locomotion of fossil cephalopods, p. 633–640. In J. Wiedmann andJ. Kullmann (eds.), Cephalopods Present and Past. 2 nd International Cephalopod Symposium, E.116

BRETT AND WALKER—PREDATION IN PALEOZOIC MARINE ENVIRONMENTSSchweizerbart’sche Verlagsbuchhandlung, Stuttgart, Germany.LEIGHTON, L. R. 2001a. New examples of Devonian predatory boreholes and the influence of brachiopod spineson predator success. Palaeogeography, Palaeoclimatology, Palaeoecology, 165:71–91.LEIGHTON, L. R. 2001b. New directions in the paleoecology of Paleozoic brachiopods, p. 185–206. In S. J. Carlson and M.Sandy (eds.), Brachiopods Ancient and Modern: A Tribute to G. Arthur Cooper, Paleontological Society Papers 7.LESCINSKY, H. L., AND L. BENNINGER. 1994. Pseudoborings and predatory traces: Artifacts of pressure dissolutionin fossiliferous shales. Palaios, 9:599–604.LILJEDAL, L. 1985. Ecological aspects of a silicified bivalve fauna from the Silurian of Gotland. Lethaia, 18:53–56.MALZAHN, E. 1968. Uber neue Funde von Janassa bituminosa (Schloth.) im neiderrheinisschen Zechstein.Geologisch Jarhbuch, 85:67–96.MAPES, R. H., AND E. J. BENSTOCK. 1988. Color pattern on the Carboniferous bivalve Streblochondria? Newell.Journal of Paleontology, 62:439–441.MAPES, R. H., AND M. C. HANSEN. 1984. Pennsylvanian shark-cephalopod predation: A case study. Lethaia, 17:175–183.MCMENAMIN, M. 1986. The Garden of Ediacara. Palaios, 1:178–182.MCMENAMIN, M., AND D. L. S. MCMENAMIN. 1990. The Emergence of Animals: The Cambrian Breakthrough.Columbia University Press, New York, 217 p.MEHL, J. 1984. Radula und Fangarme bei Michelinoceras sp. asu dem Silur von Bolivien. PaläeontologicaZeitschrift, 58:211–229.MEYER, D. L. 1985. Evolutionary implications of predation on Recent comatulid crinoids. Paleobiology, 11:154–164.MEYER, D. L., AND W. I. AUSICH. 1983. Biotic interactions among Recent crinoids and among fossil crinoids, p.377–427. In M. J. S. Tevesz and P. L. McCall (eds.), Biotic Interactions in Recent and Fossil BenthicCommunities. Plenum Press, New York.MILLER, R. H., AND F. A. SUNDBERG. 1984. Boring Late Cambrian organisms. Lethaia, 17:185–190.MORRIS, P. G. 1994. Some new Carboniferous cardiniarid brachiopods from the Middle Limestones (Visean, Chadian)of North Staffordshire, England. Neues Jahrbuch fur Geologie und Palaontologie Monatshefte, 5:267–276.MOY-THOMAS, J. A., AND R. S. MILES. 1971. Palaeozoic Fishes. W. B.Saunders Co., Philadelphia, 259 p.NEDIN, C. 1999. Anomalocaris predation on non-mineralized and mineralized trilobites. Geology, 27:987–990.NIXON, M. 1988. The feeding mechanisms and diets of cephalopods—Living and fossil, p. 642–652. In J. Wiedmannand J. Kullmann (eds.), Cephalopods Present and Past. 2nd International Cephalopod Symposium, E.Schweizerbart’sche Verlagsbuchhandlung, Stuttgart, Germany.PAINE, R. T. 1974. Intertidal community structure: Experimental studies on the relationship between dominatcompetitor and its principal predator. Oecologia, 15:93–120.PATTERSON, C. 1965. The phylogeny of the chimaeroids. Philosophical Transactions of the Royal Society, SeriesB, 249:101–209.PEEL, J. S. 1984. Attempted predation and shell repair in Euomphalopterus (Gastropoda) from the Silurian ofGotland. Lethaia, 32:163–168.RICHARDS, R. P., AND C. W. SHABICA. 1969. Cylindrical living burrows in Ordovician dalmanellid brachiopodshells. Journal of Paleontology, 43:838–841.RODRIGUEZ, R. P., AND R. C. GUTSCHICK. 1970. Late Devonian-Early Mississippian ichnofossils from westernMontana and northern Utah, p. 407–438. In J. P. Crimes and L. C. Harper (eds.), Trace Fossils. GeologicalJournal, Special Issue 3. Seel House, Liverpool.ROHR, D. M. 1976. Silurian predator borings in the brachiopod Dicaelosia from the Canadian Arctic. Journal ofPaleontology, 65:687–688.ROLLINS, H. B., AND D. K. BREZINSKI. 1988. Reinterpretation of crinoid-platyceratid interaction. Lethaia, 21:189–292.RUDKIN, D. M. 1985. Exoskeleton abnormalities in four trilobites. Canadian Journal of Earth Sciences, 22:479–483.RUDWICK, M. J. S. 1970. Living and Fossil Brachiopods. Hutchinson University Library, London, 199 p.SAUNDERS, W. B., AND P. D. WARD. 1987. Ecology, distribution, and population characteristics of Nautilus, p.137–162. In W. B. Saunders and N. H. Landman (eds.), Nautilus : The Biology and Paleobiology of a LivingFossil. Plenum Press, New York.SCHAEFER, B., AND M. WILLIAMS. 1977. Relationships of fossil and living elasmobranchs. American Zoologist,17:293–302.117

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002SCHINDEL, D. E., G. J. VERMEIJ, AND E. ZIPSER. 1982. Frequencies of repaired shell fractures among the Pennsylvaniangastropods of north-central Texas. Journal of Paleontology, 56:729–740.SCHRAM, F. R. 1977. Palaeozoogeography of Late Paleozoic and Triassic Malacostraca. Systematic Zoology,26:367–379.SCHRAM, F. R. 1982. The fossil record and evolution of the Crustacea, p. 93–147. In L. G. Abele (ed.), The Biologyof the Crustacea: Systematics, the Fossil Record, and Biogeography. Academic Press, Inc., New York.SCHRAM, F. R. 1984. Upper Pennsylvanian arthropods from black shales of Iowa and Nebraska. Journal ofPaleontology, 58:197–209.SELDEN, P. A. 1984. Autecology of Silurian eurypterids. In M. G. Bassett and J. D. Lawson (eds.), Autecology ofSilurian Organisms. Special Papers in Palaeontology, 32:39–54SELDEN, P. A. 1992. Biomechanics, p. 322–326. In D. E. G. Briggs and P. R. Crowther (eds.), Palaeobiology: ASynthesis. Blackwell Scientific Publications, Oxford.SHEEHAN, P. M., AND P. J. LESPERANCE. 1978. Effect of predation on population dynamics of a Devonian brachiopod.Journal of Paleontology, 52:812–817.SIGNOR, P. W., III, AND C. E. BRETT. 1984. The mid-Paleozoic precursor to the Mesozoic marine revolution.Paleobiology, 10:229–245.SLITER, W. V. 1971. Predation on benthic foraminifers. Journal of Foraminferal Research, 1:20–29.SMITH, S. A., C. W. THAYER, AND C. E. BRETT. 1985. Predation in the Paleozoic: gastropod-like drill holes inDevonian brachiopods. Science, 230:1033–1037.SPENCER, W. K., AND C. W. WRIGHT. 1966. Asterozoans, p. U1–U157. In R. C. Moore (ed.), Treatise on InvertebratePaleontology, Pt. U. Geological Society of America and University of Kansas Press.SPRINKLE, J. 1973. Morphology and Evolution of blastozoan echinoderms. Museum of Comparative Zoology,Harvard University Publication, 284.STANLEY, S. M. 1970. Relation of shell form to life habits of the Bivalvia (Mollusca). Memoir of the GeologicalSociety of America, 125:1–296.STANLEY, S. M. 1977. Rates, trends, and patterns of evolution in the Bivalvia, p. 209–250. In A. Hallam (ed.),Patterns of Evolution as Illustrated by the Fossil Record. Elsevier, Amsterdam.STENSIÖ, E. 1969. Les cyclostomes fossiles ou ostracodermes, p. 71–692. In J. Piveteau (ed.), Traité desPaléontologie 4. Masson, Paris.TASCH, P. 1980. Paleobiology of the Invertebrates: Data Retrieval from the Fossil Record. John Wiley, New York, 975 p.THAYER, C. W. 1983. Sediment-mediated biological disturbance and the evolution of the marine benthos, p. 479–595. In M. J. S. Tevesz and P. J. McCall (eds.), Biotic Interactions in Recent and Fossil Benthic Communities.Plenum Press, New York.VERMEIJ, G. J. 1977. The Mesozoic marine revolution: Evidence from snails, predators, and grazers. Paleobiology,3:245–258.VERMEIJ, G. J. 1983. Shell breaking predation through time, p. 649–669. In M. J. S. Tevesz and P. L. McCall(eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum Press, New York.VERMEIJ, G. J. 1987. Evolution and Escalation. Princeton University Press, Princeton, NJ, 527 p.VERMEIJ, G. J. 1995. Economics, volcanoes, and Phanerozoic revolutions. Paleobiology, 21:125–152.VERMEIJ, G. J., D. E. SCHINDEL, AND E. ZIPSER. 1981. Predation through geological time: Evidence from gastropodshell repair. Science, 214(27):1024–1026.WATERS, J. W., AND C. MAPLES. 1991. Mississippian pelmatozoan community reorganization: A predation-mediatedfaunal change. Paleobiology, 17:400–410.WILLIAMS, M. E. 1990. Feeding behavior in Cleveland Shale fishes, p. 273–287. In A. J. Boucot, EvolutionaryPaleobiology of Behavior and Coevolution. Elsevier, Amsterdam.ZANGERL, R. 1981. Chondrichthyes, 1: Paleozoic elasmobranches. Handbook of Paleoichthyology, 3A:1–115.ZANGERL, R., AND E. RICHARDSON. 1963. Paleoecological history of two Pennsylvanian black shales. FieldianaGeological Memoir, 4, 352 p.ZANGERL, R., B. G. WOODLAND, E. S. RICHARDSON, AND D. L. ZACHRY. 1969. Early diagenetic phenomena in theFayetteville Black Shale (Mississippian) in Arkansas. Sedimentary Geology, 3:87–120.118

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONPOST-PALEOZOIC PATTERNS IN MARINE PREDATION:WAS THERE A MESOZOIC AND CENOZOIC MARINEPREDATORY REVOLUTION?SALLY E. WALKER 1 AND CARLTON E. BRETT 21Department of Geology, University of Georgia, Athens, Georgia 30602 USA2Department of Geology, University of Cincinnati, Cincinnati, Ohio 45221-0013 USAABSTRACT—Mesozoic and Cenozoic evolution of predators involved a series of episodes. Predators reboundedrather rapidly after the Permo-Triassic extinction and by the Middle Triassic a variety of new predator guildshad appeared, including decapod crustaceans with crushing claws, shell-crushing sharks and bony fish, as wellas marine reptiles adapted for crushing, smashing, and piercing shells. While several groups (e.g., placodonts,nothosaurs) became extinct in the Late Triassic crises, others (e.g., ichthyosaurs) survived; and the Jurassic toEarly Cretaceous saw the rise of malacostracan crustaceans with crushing chelae and predatory vertebrates—inparticular, the marine crocodilians, ichthyosaurs, and plesiosaurs. The late Cretaceous saw unprecedented levels ofdiversity of marine predaceous vertebrates including pliosaurids, plesiosaurs, and mosasaurs. The great Cretaceous-Tertiary extinction decimated marine reptiles. However, most invertebrate and fish predatory groups survived; andduring the Paleogene, predatory benthic invertebrates showed a spurt of evolution with neogastropods and newgroups of decapods, while the teleosts and neoselachian sharks both underwent parallel rapid evolutionary radiations;these were joined by new predatory guilds of sea birds and marine mammals. Thus, although escalation is sometimescast as an ongoing “arms race,” in actuality the predatory record shows long interludes of relative stabilitypuncturated by episodes of abrupt biotic reorganization during and after mass extinctions. This pattern suggestsepisodic, but generally increasing, predation pressure on marine organisms through the Mesozoic-Cenozoicinterval. However, review of the Cenozoic record of predation suggests that there are not unambiguous escalatorytrends in regard to antipredatory shell architecture, such as conchiolin and spines; nor do shell drilling and shellrepair data show a major increase from the Late Mesozoic through the Cenozoic. Most durophagous groups aregeneralists, and thus it may be that they had a diffuse effect on their invertebrate prey.“As evolutionists, we are charged, almost bydefinition, to regard historical pathways as theessence of our subject. We cannot be indifferent tothe fact that similar results can arise by differenthistorical routes.” —Gould and Vrba, 1982“This is not to say that selection is notimportant, but that its invocation is not justifieduntil the role of chance in the operation of abasically stochastic universe is ruled out.”—Schram, 1986“A science grows only as it is willing toquestion its assumptions and expand itsapproaches.” —Hickman, 1980INTRODUCTIONTHE CONCEPT OF predator-prey escalationis, in large measure, an outgrowth of the extensivestudies of Vermeij (1977, 1987) on the so-called“Mesozoic Marine Revolution.” This term mightseem to imply that a dramatic development ofmarine predators was initiated at the Triassic; acontinuous intensification of predator-preyrelationships has been envisaged. In actuality, theMesozoic and Cenozoic evolution of predatorsinvolved a series of episodes. In this paper wedocument the diverse predatory guilds of theMesozoic and Cenozoic, especially vertebrates thatputatively devoured invertebrate prey, withcomments on their modes of feeding and possible119

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002impact on potential prey, focusing on benthicinvertebrates. Predator guilds (e.g., marine reptiles)during the Mesozoic are surprisingly similar tothose in the Cenozoic (e.g., marine mammals),except that the players have changed. Despite thesetbacks of mass extinctions, the diversity ofpredators remained at a nearly constant proportionfrom the Late Triassic to mid-Cretaceous(Bambach and Kowalewski, 1999). However,during the last 110 million years (Late Cretaceous–Neogene), predators diversified faster than the restof the fauna (Bambach and Kowalewski, 1999).We also examine the patterns of predation inpost-Paleozoic shell structure (e.g., conchiolin andspines), shell repair, and shell drilling. However, datafrom drilling and shell repair thus far do not showunambiguous escalatory trends. For the sake ofargument, escalation is not a continual trend fromthe Mesozoic to the Cenozoic; rather, within eachera it is dependent on the suite of predators and prey.In the final section of the paper we reconsiderthe Mesozoic Marine Revolution hypothesisproposed by Vermeij (1977, 1978, 1987), and askquestions that may guide future research: Is thepattern of putative armor in invertebrates strictlyrelated to predation, or might there be otherhypotheses that could explain armor in organisms?Is there evidence that most predators are specialistson particular prey and thus might cause extremeselection in invertebrates who then respond withvarious escalating strategies (e.g., spines,chonchiolin) to mitigate the increased predationpressure? Does diffuse selection from generalistpredators cause antipredatory armor to arise in anumber of groups of invertebrates? If mostpredators are generalists, as appears to be the casebased on the evidence amassed herein, then perhapsthere was not a sufficiently intense selective forceto produce a major “sea change” in antipredatoryarmor in any one group of marine invertebrates,especially in post-Paleozoic organisms. Thus, forexample, durophagy may not necessarily mean thata predator ate molluscs; durophagous dentitioncould also indicate the eating of crustaceans, otherhard-shelled prey, or even soft prey (e.g., Plotkinet al., 1993; Wilga and Motta, 2000). Perhaps also,as Gould and Vrba (1982) have recognized, thereare a number of historical and non-adaptive routesby which specific aptations may ultimately arisein organisms, and such may also be the case withcertain antipredatory strategies.Despite its length, this paper is not an exhaustivereview. However, we did strive both to provide abroad overview of Mesozoic and Cenozoic predatorsand their potential prey, and, perhaps moreimportantly, to demonstrate that there are alternativeways to think about these predatory patterns.TRIASSIC PREDATORSAND PREDATIONAll marine benthic ecosystems were profoundlyaltered by the Permo-Triassic extinction (Fig. 1).Many Paleozoic predators were eliminated, includingmost phyllocarids, platyceratid gastropods, goniatiteammonoids, and many primitive lineages of sharks.Other active predatory groups preferentially madeit through this bottleneck, including the hybodontidsharks and the root-stocks of Mesozoic crustaceansand ammonoids (Knoll et al., 1996).Gastropods and Bivalves.—Varied archaeo- andmesogastropod taxa rediversified in the Triassic.Nonetheless, records of gastropod drilling predationare surprisingly rare in this period (Kowalewski etal., 1998). However, in a few instances, drillers seemto have had a significant impact (Fürsich andJablonski, 1984; Kowalewski et al., 1998). The firstnaticid-like mesogastropods (Ampullina) areknown from this time (see Fig. 7) (Fürsich andJablonski, 1984; Newton, 1983; Kowalewski et al.,1998). Given the rarity of naticid-like boreholesfrom the Late Triassic to the mid-Cretaceous(Albian), it has been suggested that predatorydrilling was relatively ineffective and largely lostduring the Triassic, only to be evolved again,successfully, during the Cretaceous (Kowalewski etal., 1998). Predatory septibranch bivalves alsooriginated at this time (Skelton et al., 1990).Ammonoids.—Ammonoids were nearlyextinguished by the Permo-Triassic crises.However, the ceratitic ammonoids staged a rapidrediversificiation in the Triassic. Like other120

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 1—Ranges of various taxa of Mesozoic and Cenozoic durophagous (hard-shell crushing)predators. Thin lines: present, but of limited abundance; thick lines: abundant; broken lines: possiblypresent but rare as fossils.ammonoids, the ceratites are assumed to have beenpredaceous, although data are very sparse. Stomachor crop contents of ammonoids are very rare, butwhen they are found provide important evidencefor trophic relationships. One specimen of an EarlyTriassic ammonoid (Svalbardiceras) hadostracodes and foraminiferans among its gastriccontents and may have been a predatory nektoniccarnivore (Westermann, 1996, p. 675).Crustacea and Ostracodes.—The mostimportant post-Paleozoic groups of decapod,isopod, and amphipod crustaceans appeared in theLate Paleozoic, but they did not diversifysignificantly until the Jurassic (Briggs and Clarkson,1990). Four out of 27 Paleozoic families survivedinto the Mesozoic, and only a few groups are knownfrom the Triassic (Briggs and Clarkson, 1990).Various lobster groups evolved in the Triassic(Table 1). Their appendages indicate that they weredurophagous, but modern lobsters feed on a widevariety of prey and are not specialists on molluscanprey. Ostracodes are known to be predators on121

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 1—Geographic and temporal ranges and feeding preferences of crustacean predators.TaxonGeologic RangeGeographic rangeand depth range(shallow, < 70 m;deep, > 70 m)Specialist feeders on specific Geologic range of durophagousinvertebrate prey? anatomical features (molariformclaws, mouth parts, etc.)Infraorder BrachyuraFamily PortunoideaCarcininaePortuninaeCretaceous(Maastrichtian)–RecentBoreal to temperate Generalists(bivalves, like oysters and mussels;scavengers; fish, plants, benthicinvertebtrates, jellyfish, epibionts onZostera, tunicates)UnknownFamily XanthidaeCretaceous–Recent(Schram, 1986)Temperate to tropical Generalists(mangrove detritus, other plants anddetritus, small crustaceans, oysters,barnacles; M. mercenaria eats molluscs,like oysters, polychaetes, barnacles andother crustaceans)Menippe mercenaria is the mostfamous shell-crusher of this group; thegenus Menippe originated in the Middleto Upper Eocene (Glaessner, 1969), andM. mercenaria is known only from thePleistocene (Rathbun, 1935)Family CalappiideaEocene–Oligocene to Miocene;Recent (Glaessner, 1969);Cretaceous–Recent (Schram,1986)South Pacific, Caribbean,North America, EuropeGeneralists The shell-peeling Calappa flammea isknown from the Oligocene (Rathbun,1930; Ross et al., 1964)Decapod crustaceans(Order Decapoda)(earliest decapod isPalaeopalaemon, Schram, 1986)Generalists UnknownLate Devonian to Recent(Schram, 1986); Permo-Triassicto Recent (Williams, 1996)Stomatopods(Order Stomatopoda)Extant: 350 spp, 66 genera,12familiesDevonian to RecentChiefly tropical shallowto deep water; rare intemperate zones, nonein polar regionsSpecialists Raptorial mouthparts, Devonian–Recent; folding raptorial thoracopods, inpaleostomatopods, Carboniferous;gonodactyloid shell smashers,?Cretaceous, Upper Miocene to Recent;Squillids, ?Cretaceous to Recent122

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONMolariform teeth and shell-crushingclaws (Miocene to Recent)Gore and Soto (1979, p. 67) state thatparthenopids are omnivores anddetritivores; Debelius (1999, p. 259)figures Parthenope horrida eating apufferfish; Vermeij 1978, p. 40, statesthat Parthenope (Daldorfia ) horrida isa shell-crusher in the labMolariform teeth on major crusher clawas in Homarus (Cretaceous to Recent;Glaessner, 1969)Chiefly boreal andtemperate; shallow anddeep waterGeneralists (sea urchins, polychaetes,scavengers, molluscs)Temperate to tropical Generalists, omnivores, detritivores; onespecies known as a shell-crusher(Parthenope horrida )GeneralistsBoreal, temperate totropicalGeneralists(sea urchins)Temperate to tropical;deep and shallow waterGeneralists(mollusks, sea cucumbers; Williams, 1984)Temperate zones, NorthAtlantic, Europe; shallowand deep water (manymigrate)Generalists(molluscs, hydroids, crustaceans,lobsters, polychaetes, brittle stars,mussels, limpets, lobster molts, chitons,bryozoans, scallops, oysters, sea urchins,seaweeds; Williams, 1984; Lawton andKavalli, 1995)Temperate to tropical Generalists?(?scyphozoan medusae; Williams, 1984)(Table 1, cont.)Family CancridaeEocene to Recent(Schram, 1986)Family ParthenopidaeEocene to Recent(Schram, 1986)Family MajidaeEocene to Recent(Schram, 1986)Family GrapsidaeEocene to Recent(Schram, 1986)Infraorder PalinuroideaFamily Palinuridae(Lobsters)?Lower Triassic, Mid-Triassic toRecent (Williams, 1996)Superfamily NephropidaeFamily Nephropidae(Clawed marine lobsters)Permo-Triassic to Recent(Williams, 1996)Family Scyllaridae(Spanish Lobsters)?Lower Cretaceous; ?lowerEocene to Recent(Williams, 1996)123

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002polychaete annelids, and scavengers on deadpolychaetes, fish, and squid (Vannier et al., 1998).They have serrated appendages that act like knives(or sandpaper) to abrade their food. Based on theirfeeding appendages, the Early Triassic cypridinids(and possibly the Late Ordovician myodocopids)may have been predators or scavengers oncephalopod carcasses (Vannier et al., 1998).Chondricthyes.—The long-lived hybodontidsflourished in the Triassic and became the dominantJurassic sharks (Maisey, 1982). Hybodontspossessed varied dentition, ranging from highcuspedimpaling teeth to low-crowned crushers,indicating rather generalized predatory diets(Maisey, 1982); they gave rise to swimming,piscivorous sharks, as well as pavement-toothedforms. The hybodont sharks may have arisen inthe Devonian, but they underwent strong adaptiveradiation during the Triassic (Maisey, 1982).Sauropterygian clade.—The sauropterygianclade (Figs. 2, 6; Table 2) contains the Triassic stemgroups such as placodonts, pachypleurosaurs,nothosaurs, and pistosaurs, and the Jurassic-Cretaceous crown groups known as plesiosaurs,pliosaurs, and elasmosaurs (Rieppel, 1999). Verylittle is known about the feeding mechanics ofTriassic stem-group sauropterygians, whichsecondarily became aquatic from their terrestrialancestors (see Rieppel, 2002). Feeding underwater,as the stem-group sauropterygians did, required asuite of anatomical and behavioral adjustments thathad to allow for their adaptive radiation into earlyMesozoic seas (Rieppel, 2002). Suction feedingappears to be the most efficient hydrodyamic wayto solve the underwater feeding dilemma (Lauder,1985); however, “quick snapping bites” at the airwaterinterface (or underwater) are also used,especially by crocodilians (Rieppel, 2002). Triassicsauropterygians covered all styles of feeding, andthus have little overlap in hypothesized feedingstrategies. The varied nearshore habitats in theMiddle Triassic, with lagoonal basins interspersedamong reef habitats, may have accounted for thetrophic-functional diversity of stem-groupsauropterygians (Rieppel, 2002).Placodonts.—During the Middle Triassic, theplacodonts (Figs. 1, 2.1–2.3) evolved fromunknown diapsid reptilian ancestors (Benton, 1993,1997). The Triassic placodonts, sister group to allother Sauropterygia, have members that areinterpreted to have been benthic predators on hardshelledinvertebrate prey. Placodus, for example,had pachyostosis (complete covering of the cheekby dermal bone), which added weight to the jawand thus may have functioned as an adaptation fordurophagy (Rieppel, 2002). Additionally, theprocumbent and chisel-shaped premaxillary anddentary teeth may have functioned to pick offinvertebrates from their substrate, which were thencrushed with the posterior tooth plates before theywere swallowed (Westphal, 1988). Biomechanically,the tooth plates of Placodus were positioned in sucha way as to enhance crushing, but not increase loadto the jaw (Rieppel, 2002). The basal stock ofPlacodus already had large crushing tooth platesand procumbent premaxillary teeth (Rieppel, 2002),indicating that durophagy was an ancestral conditionin this group. The few durophagous taxa ofplacodonts may have had an impact on therediversifying molluscan communities of theTriassic, but they became extinct in the major crisestoward the end of that period.Not all placodonts had dentition indicating thatthey ate benthic hard-shelled prey. More derivedcyamodontids (Placochelys and Psephoderma) lackpremaxillary and anterior dentary teeth, and mayhave picked up benthic soft-substrate invertebrates(like crustaceans) through suction action (Rieppel,2002). Another basal cyamodontoid, Henodus, hasmuch reduced crushing dentition, and may have hadbaleen that was used in sieving benthic invertebrates.Henodus is thus interpreted to have been a bottomfeeder—perhapsan herbivore or omnivore—but itwas not durophagous (Rieppel, 2002).Pachypleurosauria.—Pachypleurosaurs(Fig. 2.7) were swimming reptiles with long headsand interlocking lower and upper sharp teethpresumably for the capture of fish (Benton, 1997).Pachypleurosauria are considered to be the sistergroup of the Nothosauroidea, or the sister taxon toall other Eusauropterygia (composed of nothosaursand plesiosaurs; Rieppel, 2002). Pachypleurosauria124

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 2—Triassic predatory reptiles. 1–3, Placodont reptile Placodus. 2, 3, Lateral and palatal viewsof skull; note spatulate incisors and broad “pavement” teeth in maxilla and palate. 4, IchthyosaurMixosaurus. 5, 6, Ichthyosaur Grippia; lateral and dorsal views of skull. 7, Nothosaur Pachypleurosaurus.Adapted from figures in Benton (1997).are among the smallest of the sauropterygians: mostattain a length of 50 cm; few attain lengths of 120cm (Carroll and Gaskill, 1985; Rieppel, 2002).Based on a relatively large tympanic membrane andlimited bone ballast, these marine reptiles may haveinhabited shallow, coastal, and estuarine waters(Taylor, 2000). Pachypleurosauria (e.g.,Neusticosaurus) had delicate jaws with homodontdentition; loading conditions of the jaw indicate thatthey were not efficient in subduing vigorous prey(Rieppel, 2002). Pachypleurosaurs probably werepelagic predators that used suction and rapid closureof the jaws to subdue soft-bodied cephalopods andsmall fish (Sander, 1989; Rieppel, 2002).Nothosaurs.—The Middle TriassicNothosauroidea (up to 4 m in length) are a majorclade of the Eusauropterygia, members of whichmay have eaten fish, other sauropterygians, and125

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 2—Cenozoic marine vertebrate predator guilds.Marine Reptile Geologic Range;BiogeographyFeeding Realm Feeding Type Putative Prey Gastric Evidence orother evidenceReferencePlacodontia Middle Triassic(chiefly); late Lower toUpper Triassic ofEurope, North Africa,Middle East; distributionrestricted to westernperiphery of TriassicTethys (Lucas, 1997)Benthic Durophagous (basal andderived cyamodontoids) tosuction and sieving(Henodus ); crush guild ofMassare, 1997Hard-shelled sessile to nonsessileinvertebrates(cyamodonts); soft shelledinvertebrates to plants(Henodus )Pachypleurosaurs Middle Triassic Pelagic Suction feeding Soft-shelled invertebrates,cephalopodsNothosaurs Middle Triassic(chiefly); late EarlyTriassic to UpperTriassic; endemic toEuropean/Asiancontinent (Lucas, 1997)Pelagic Suction feeding to fish-trapdentitionSimosaurus : hard-shelledinvertebrates; all others fish, softshelledinvertebratesJuventile placodont inLariosaurus stomachRiepple, 2002;Massare, 1997Riepple, 2002Riepple, 2002;Sanders, 1989;Tschanz, 1989Pistosauroidea Middle Triassic Pelagic Puncturing dentition Soft-shelled invertebrates; fish Riepple, 2002Plesiosaurs Late Triassic (Rhaetian)to Late Cretaceous;global distribution in theJurassic andCretaceous, but had highendemism (Lucas, 1997)Pelagic Needle-shaped teeth; pierceI and II guilds of Massare,1997Soft-shelled invertebrates orfleshy prey; some may havestrained small prey from water(Cryptoclidus )Jurassic forms withcephalopod hooklets;gastroliths; ?regurgitatesthought to be fromplesiosaurs with ammonoidlarvae and shells ofBaculitesRiepple, 2002;Massare, 1997;Pollard, 1968;Martill et al.,1994; Wetzel,1960Pliosaurs Cretaceous Pelagic Pursuit predators; Cut andPierce II guild of Massare,1997; shake feedingLarge, fleshy prey: other reptiles,fish, and cephalopodsCephalopod hooklets inPeloneustesRiepple, 2002;Massare, 1997;Martill et al.,1994Elasmosaurs Cretaceous Pelagic Robust teeth, pierce I and IIguild of Massare, 1997Riepple, 2002;Massare, 1997Ichthyosaurs Early Triassic to lateCenomanian (LateCretaceous); achievedglobal distribution by theMiddle Triassic (Lucas,1997)Nearshore (Triassic)to Pelagic (Jurassicto Cretaceous); somewere deep divers to adepth of 600 m to themesopelagic layer ofthe ocean(Ophthalmosaurus );Shonisaurus (UpperTriassic, Nevada)was outer shelf orbasinal in distribution(Hogler, 1992)Cut, pierce, smash, crunch,and crush guilds of Massare(1987, 1997)Triassic icthyosaurs hadheterodont dentitionsuggesting ambush,generalist predators innearshore habitats; inJurassic, mostly homodontdentition suggesting pelagicpursuit predators thatspecialize in a certain typeof prey (Massare andCallaway, 1990)Large, fleshy prey; soft prey; softprey with internal hardparts; preywith bony scales or hard, thinexoskeleton; prey with a veryhard exteriorLarge, rear teeth of MiddleTriassic icthyosaurs(Phalarodon, Omphalosaurus)are suggestive of mollusccrushing(Massare and Callaway,1990), however this typedisappeared by the Late Triassic;the Lower Triassic Grippiamay also have been durophagous(Lingham-Soliar, 1999)Gastric mass withcephalopod hooklets, mostlikely from belemnites orother types, in manyichthyosaur skeletons;coprolites indicate fishremains such as the Liassicnectonic or necto-benthicfish, Pholidophorus;?Jurassic cuttle-fish;?marine reptiles; wood;none found with ammonoidor belemnite shells ingastric contentsMassare, 1987,1997; Pollard,1968; Keller,1976; Motani etal., 1999126

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATION(Table 2, cont.)Mosasaurs Late Cretaceous(Cenomanian to endCretaceous); most wereendemic to one region(Lucas, 1997)Crocodiles Early Jurassicto RecentPelagic, ?Benthic Ambush predators; mostmosasaurs occupied theCut guild; Crush andPierce II guilds alsooccurred (Massare, 1997)Opportunistic generalists(Massare, 1997)Benthic Ambush predators;generalists, occupying theCut, Pierce I and II, Crunch,and Crush guilds ofMassare, 1997; Globidensonly crush guild marinereptile since Triassic(Massare, 1997)Large, fleshy prey, prey with hardexterior; Globidens wascosmopolitan in distribution and isthought to be durophagous(Lingham-Soliar, 1999)Teleosaurids (teleosaurs): fish,turtles, and ammonites;Metriorynchids (geosaurs):ammonites, belemnites,pterosaurs, and giant fishLeedsichthysSea Turtles Jurassic to Recent Benthic to pelagic Generalists to specialists One form today feeds onmolluscs (Caretta ); one form inthe Late Cretaceous may alsohave fed on molluscs (Hirayama,1997); Others feed on sea grass,jellyfish, and crustaceansHybodont sharks Benthic;necto-benthicscavengers to predators;Asteracanthus haddurophagous dentition(Jurassic)Predators on surface-livingammonoidsStomach contents indicateammonoids, birds, fish,smaller mosasaurs;Clidastes, had a marineshark and a diving marinebird Hesperornis in gastriccontents (Martin and Bjork,1987)Crocodile tooth associatedwith turtle scutes in ateleosaurid;Metriorhynchus withabundant cephalopodhooklets; associated scalesof Lepidotes associatedwith skeletal elements ofSteneosaurus;Metriorhynchus toothembedded in giant fishLeedsichthysKauffman andKesling, 1960;Martin and Bjork,1987; Massare,1997Martill, 1985,1986; Hua andBuffetaut, 1997;Martell et al.,1994No fossil evidence Hirayama, 1997Belemnites; one specimen(Hybodus ) with over 250rostra in gastric contentsPollard, 1968Chimaeras Jurassic Benthic Martill et al.,1994NeoselachiansharksJurassic Benthic some with durophagousdentitionMartill et al.,1994127

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002hard- and soft-shelled invertebrates in the pelagicrealm (Rieppel, 1998). Teeth of some nothosaurs,such as Simosaurus, have a bulbous shape whichmay have had a somewhat durophagous function(Rieppel, 2002). Anatomical evidence indicatesthat Simosaurus had strong neck muscles and wascapable of rapid jaw opening; suction also mayhave been used to round up shelled ammonoids orfish (Rieppel, 2002). Nothosaurus mirabilis hadspecialized jaw adductor muscles, heterodontdentition with procumbent fangs, and a very narrowand elongate skull (Rieppel, 2002). The heterodontdentition suggests most nothosaurs ate fish, althoughthe gastric contents of one nothosaurid (Lariosaurus)contained placodonts and small pachypleurosaurs(Sander, 1989; Tschanz, 1989; Rieppel, 2002). Theymay also have eaten soft-bodied invertebrates, suchas cephalopods. Simosaurus may have eaten hardshelledprey (Rieppel, 2002). Some nothosaurs,because of their large size, may have been at the topof the food chain.Pistosauroidea.—The Triassic Pistosaruoideagave rise to the plesiosaurs that were common inthe Jurassic and Cretaceous seas. Some pistosaurianshad jaws similar to those of the putatively fish-eatingnothosaurids (e.g., Nothosaurus), and others, suchas Pistosaurus, had narrow and elongated pincertypejaws, that had less numerous and widely-spacedheterodont dentition with maxillary fangs (Rieppel,2002). Puncturing prey, rather than suction feeding,may have been the modus operandi of thesecreatures, and they may have fed on soft-shelledpelagic invertebrates and fish (Rieppel, 2002).Ichthyosauria.—Ichthyosaurs (Order Ichthyosauria)(Figs. 2.4–2.6; 4) are known from the LowerTriassic to Cenomanian (Bardet, 1994), but theyhave only recently been studied in detail (Callaway,1997a). Triassic ichthyosaurs were nearly as diverseand widespread as Jurassic ichthyosaurs, but arenotoriously affected by preservational bias(Callaway, 1997b; Sander, 1997). Because of thispreservational bias, little is known about theevolution of dentition in Triassic ichthyosaurs. Mostof them likely had heterodont dentition, indicatingthat they were generalist feeders in nearshore waters(Massare and Calloway, 1990).Some Middle Triassic ichthyosaurs with largerear teeth may have been molluscivorous (Massareand Callaway, 1990). In the Jurassic and Cretaceous,the dentition became homodont, indicating that theymay have become specialized on pelagic prey(Massare and Callaway, 1990). Perhaps they werespecialists on fish and/or soft-bodied cephalopods(Sander, 1997; Massare and Callaway, 1990).Gastric contents indicate that they may have fed onbelemnite cephalopods, although no belemnite orammonoid shells have ever been found inichthyosaurian stomach contents. Based on bodyform, by the end of the Triassic, ichthyosaurs werehydrodynamically advanced and were very fastswimminganimals (Lingham-Soliar, 2001).Pterosaurs.—The appearance of pterosaurs inthe middle of the Triassic Period (Benton, 1993) mayhave increased predation pressure on near-surfacenektonic organisms, including fish and cephalopods.The long jaws and impaling spike-like teeth ofrhamphorhynchids and many pterosauroids suggestsa piscivorous diet in these flying reptiles.TRIASSIC BENTHICORGANISMS:ANTIPREDATORY RESPONSES?Varied morphological and behavioral featuresof benthic invertebrates have been interpreted asantipredatory adaptations (Fig. 3), although manyof these features may be merely exaptations(sensu Gould and Vrba, 1982). Triassic benthicfaunas are decidedly “no frills” relative to thoseof the late Paleozoic (Valentine, 1973). Shells arerelatively thin and mainly lacking in spines. Inaddition, several groups of cemented bivalves—the ostreids, gryphaeids, plicatulids, andterquemids—first became abundant on hardsubstrates in the Triassic. Harper (1991) hasdemonstrated experimentally that predators avoidcemented bivalves when given a choice.The evolutionary breakthrough of mantlefusion in bivalves led to the rapid development ofinfaunal clades in the early Mesozoic (Stanley,1977). Mud- and rock-boring bivalves also firstbecame common during this time (Seilacher, 1985;128

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 3—Temporal ranges of various potential anti-predatory adaptive or exaptative strategies invarious invertebrate groups; semi: semi-endobenthic (quasi-infaunal); burrow: burrowing endobenthicorganisms. Thin lines: present, but of limited abundance; thick lines: abundant; broken lines: possiblypresent but rare as fossils.129

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Bottjer, 1985). Deep burrows and endolithic cryptsmay have been particularly effective protectionfrom grazing predators, such as placodonts. Thenear-synchronous development of this strategy inat least three independent lineages of venerid(Skelton et al., 1990) and myoid bivalves, as wellas the great increase in endobenthic and endolithicanomalodesmatans—all during the Early Triassic—suggests intensification of some selective pressure(Fig. 3). Bottjer (1985) and Skelton et al. (1990)drew attention to the coincidence of this downwardpush with the Mesozoic Marine Revolution.However, McRoberts (2001) has recently argued thatdurophagous predators may not have beensufficiently abundant or widespread during theTriassic to account for the early radiation ofendobenthic strategies. It is possible that theantipredatory advantages of living cryptically weremerely a side-benefit of adaptation driven by otherpressures, such as competition (McRoberts, 2001).The overall frequency of shell repair, due eitherto predators or to physical factors in theenvironment (see Cadée et al., 1997; Cadée, 1999;Ramsay et al., 2001) is also low during this timeinterval, with repair frequencies evidently evenlower than those of the late Paleozoic, as recordedby Vermeij et al. (1982) (Table 3).Ammonoids.—Vermeij (1987) drew attentionto the fact that surveys of ammonoid shellarchitecture and traces of predation on cephalopodsare critically needed for the whole Phanerozoic. Ageneral view of ammonoids suggests that theirmorphology is related to their pelagic, demersal,or planktic lifestyle (Westermann, 1996, p. 689),rather than to antipredatory features.Early Triassic ceratitic ammonoids fromplatform environments are considered to have beenchiefly nektonic in habit, although some planktonicand demersal forms occurred (Westermann, 1996).Offshore bituminous limestones of the Middle andUpper Triassic in Europe, North America, and Chinaalso contained coarsely costate to smooth ammonoidmorphotypes, all of which were interpreted to bepelagic (including some with planktonic lifestyles).In the Late Triassic (Norian), however, most highlysculpted evolute ammonoid morphotypesdisappeared, whereas smooth involute formssurvived, and the first heteromorphs appeared.Deep outer-shelf and upper-slope environmentsfrom the Early Triassic of China contained bothsmooth and costate ammonoids; deep basinammonoids were smooth-shelled, some with finesculpture, many of which are interpreted to havebeen pelagic (Westermann, 1996). Coarse sculpture,however, is thought to be commonly associated withbasin-slope habitat.Large pelagic predators, such as ichthyosaurs,plesiosaurs, placodonts, and turtles had evolved bythe Late Triassic, and many are thought to haveeaten ammonoids; ceratitic ammonoids do notseem to show classic antipredatory defenses.However, the temporal trends, if any, of ceratiteshell injuries remain to be studied.Echinoderms.—Echinoderms went through anevolutionary bottleneck after the Permianextinction, with at least five classes surviving intothe Early Triassic (Simms, 1990). From lowdiversity in the Triassic, echinoids and crinoidsdiversified in the Middle Triassic, but some cladeswent extinct during the mid-Carnian. A seconddiversification occurred in the Norian and the EarlyJurassic for both groups. Triassic crinoid forms reevolved“passive” filtration systems like theirPaleozoic forebearers. Most post-Paleozoic crinoidsare thought to be anatomically similar to theirPaleozoic ancestors; however, Donovan (1993) andOji (2001) provide evidence that the Mesozoiccrinoids (especially the Jurassic forms) were agileand could actively relocate—this may have provideda selective advantage as predation pressure increased(Meyer, 1985). Pseudoplanktic pentacrinitids,paracomatulids, and the true comatulids evolved inthe Late Triassic and occupied niches altogetherdifferent from their Paleozoic counterparts (Simms,1990). These new modes of life probably do notreflect escalation; overall, predation on theseechinoderms is deemed to have been relatively lowduring the Triassic (Schneider, 1988).The two main clades of echinoids, theDiadematacea and Echinacea, diversified in theLate Triassic and Early Jurassic, but still retainedtheir mid-Paleozoic diversity levels (Simms, 1990).130

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 4—First appearances of major groups of Mesozoic and Cenozoic pelagic predators andpotential prey. Arrows point to approximate time of first appearance of taxa.The echinaceans developed carnivorous andherbivorous habits, and others started to bore intorock substrates. As noted, this latter could reflectan antipredation strategy, but there is littleinformation available on predation-related injuriesin these organisms during the Triassic.JURASSIC PREDATORSMuch more is known about Jurassic predatorschiefly because of the greater extent ofepicontinental sea deposits compared to the Triassic.There are also many significant Lagerstätten fromthe Jurassic. Still, information concerning predatorpreyrelationships for the Jurassic is limited. Muchof the evidence for vertebrate predation on Jurassicprey is circumstantial, based on overlapping faunalcompositions of predator and prey, interpretation oftooth form, and attempts to match dental form withputative bite marks. Nevertheless, there is tantalizingevidence of predation. Fish (including sharks),ichthyosaurs and plesiosaurs are considered thedominant vertebrate predators (Figs. 1, 4;Tables 1, 2). These organisms could function asboth pelagic and benthic predators, so their predatoryactivities cannot be exclusively tied to either of theserealms (see Martill, 1990). Alternatively, someinterpretations suggest that the dominant marinereptiles at this time were all pelagic predators131

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 3—Shell repair in bivalves and gastropods from Mesozoic and Cenozoic Localties. Only papers that cite their raw data, or semblanceof such, are included in this table. According to Vermeij (1982, p. 233), the higher the frequency of scars, the greater is the role of the shellin protecting the gastropods against lethal breakage; if selection for armor is weak, then the frequency of repair should be low. The exceptionalpreservation required to detect shell repair is rare, especially in Paleozoic and Mesozoic gastropods (Vermeij et al., 1981).Taxonomic Group;Age/Locality;ReferenceEvidence of ShellRepairCalculation of ShellRepair FrequencyRange/Mean/median frequencyGastropodsUpper Triassic; Northern Italy, St.Cassian Group, dolomites;species describedVermeij et al.(1982)Broken edge is repaired;tiny chips ignored; somehad only body whorlsexposed, so not all shellareas were examined;small shells examined(

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONRepair scars that divert orcut across the normalconcentric growth lamellae(including: scallops,meandering clefts, divotedrepairs, and moreextensive repairs includingirregular fractures)Total number of repairscars on left valvesdivided by the totalsample size of repairedand uninjured valvesNo median recorded.Exogyra: 0.00–1.11*/0.40; Pycnodonte: 0.12–0.38/0.23Combined (Exogyra and Pycnodonte ) mean frequency: 0.34*Because more than one repair per shell was used, thefrequency can exceed 1.0.x, Campanian: 0.11; x, Maastrichtian: 0.44n =?528 or ?525 Exogrya spp., ?460 or ?450 Pycnodontespp. for Cretaceous localities, total ~?988; they state n = ">1600 indivs," 7 spp. (Paleocene + Cretaceous)As for bivalves As for bivalves 0.07–0.09/0.08; for all size classes, mean = 0.10 (from theirtable 5)n = ?77, 1 sp. Pycnodonta dissmilarisJagged repair scars onshellsNumber of individualswith one scar/totalindividuals of thatspeciesExcluding samples with 1 individual:0.03–0.59/0.29/no median (their table 10, p. 349–350)n = 3090 indiv.; 21 spp.Jagged repair scars onshellsAs for gastropods 0.07–0.35/0.21n=896, 2 spp. (corbulids had the highest shell repairfrequency at 0.35)Repairs were recordedonly if they extended 20%or more of the whole in aspiral direction or if theyinvolved subjectivelysubstantial breakage;excludes minor lip breaksNumber of repairedinjuries divided by thetotal number of shells inthe sample; frequencyof repair is looselycorrelated with thenumber of species ofshell-peeling calappidsMedian frequency reported (their table 1)*: Recent: n= 5735,0.54; Pleistocene: n =110, 0.55; Pliocene, n = 314, 0.54;Miocene, n = 549, 0.57; Paleogene, n = 136, 0.47* n must equal number of samples, but it is not clear; statedthat samples with ten or more individuals were used;frequencies of repair have remained unchanged from theEocene to PresentNo information; appears tobe large, jagged repairscars (his fig. 1)Frequency of snails withdamaged shellsestimated for eachpopulationMany localities, here totaled together: 0.00–0.48/0.11n = 4593(Table 3, cont.)Bivlaves: Osteids, Exogyra,PycnodonteLate Cretaceous (North AtlanticCoastal Plain, New Jersey)Dietl et al. (2000)Bivalves: Ostreids, Pycnodontesearly PaleoceneDietl et al. (2001)Gastropodslower Pliocene; Albenga, ItalyRobba and Ostinelli (1975)Bivalves: Anadara and Corbulaonlylower Pliocene; Albenga, ItalyRobba and Ostinelli (1975)Gastropods: TerebridsEocene–Recent; tropical tosubtropicalVermeij et al. (1981)Gastropods: LittorinidsRecent; cold temperateRafaelli (1978)133

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002(Massare, 1987). Importantly, the range in tooth formand function in Jurassic (and Cretaceous) marinereptiles was at least as great as that of modern marinemammals (Massare, 1987).Gastropods.—Although naticid mesogastropods(Fig. 6.4) existed in the Jurassic, boreholes are quiterare. However, recent discovery of drilled shellsproves that the capacity for drilling predation didexist (Kowalewski et al., 1998).Nautiloids.—Nautiloids were very diverse inthe Paleozoic, but there were few nautiloids in theMesozoic (House and Senior, 1981). Nautiloids arethought to have continued with their Paleozoicpredatory mode of life, perhaps scavenging orpreying on crustaceans (Fig. 1). A nautiloid with acomplete jaw apparatus (rhyncolites) is knownfrom lithographic limestone, Upper Jurassic ofsouthwestern Germany (Dietl and Schweigert,1999). Modern nautiloids can repair their shells(Meenakshi et al., 1974), although little is knownabout shell repair in Mesozoic nautiloids.Ammonoids.—Shell shape in ammonoids issometimes used to infer directly whether or notthey were predatory. For example, largemacroconchs of oxyconic forms are interpreted tobe mobile predators (Westermann, 1996). Shellshape, sculpture (Fig. 6.1), and size (especially formacro- and microconchs) can also be explainedby sexual dimorphism (Westermann, 1996). Interms of direct evidence, there is only one EarlyJurassic example of an ammonoid (Hildoceras)with aptychi of juvenile ammonids within its bodychambers (Westermann, 1996).Middle Jurassic ammonoids appeared tooccupy a number of trophic functional groups, fromplanktonic to demersal forms that presumably fedon ostracodes and microgastropods in algal mats(Westermann, 1996), although there is no data ongastric contents to confirm this. The lower ToarcianPosidonia shale (northwestern Europe) is knownto have clusters of fragmented harpoceratineammonoids, presumably from cephalopodpredation (Lehmann, 1975). In turn, the stomachcontents from a harpoceratine indicate that it preyedon small or juvenile ammonoids (Lehmann, 1975).Finally, Late Jurassic ammonoids hadtrophically complex functional groups similar tothose in the Middle Jurassic. Some ammonites mayhave fed on both the plankton and the benthos,depending on food availability and benthic anoxia.Ammonoid forms at this time had costae or nodosemacroconchs, and microconchs with horns on somespecies; smooth shelled ammonoids were alsocommon. Numerous records of ammonoid aptychiare reported from the body chambers of haploceratidammonites, indicating predation; and specimens ofthe Late Jurassic ammonoid, Neochetoceras, haveaptychi of conspecific juveniles within their bodychambers, indicating cannibalism (Westermann,1996, p. 676). A rare find of a Saccocoma crinoidamong the stomach contents of Physodoceras isknown from the Solnhofen Limestone (Milson,1994). Saccocoma is variously interpreted as eitherplanktic or benthic in habit (Milson, 1994), anddepending on the interpretation of the life mode forSaccocoma, the ammonoid is interpreted as either aplanktic or a benthic feeder (the latter interpretationis favored by Westermann, 1996).Echinoderm Predators.—Living families ofasteroids (e.g., Forcipulatida and Notomyotida)have their roots in the Early Jurassic (Hettangian)of Germany and Switzerland (Blake, 1993).Complete asteroids are exquisitely preserved inpelletoidal calcarenite from this time period.Modern forcipulatids are known to prey on otherechinoderms, molluscs, barnacles, and many othertypes of invertebrates. The presence of many armsin asteroids (e.g., solasteroids) suggests that theywere predators of active prey, such as otherasteroids. Predation on active prey by solasteroidsmost likely evolved in the Jurassic (Blake, 1993).Asteriids, in contrast, continued to feed onmolluscs and other benthic prey as their Paleozoicancestors did. During the Jurassic, asteriids hadprominent adambulacral spines that their moderndescendants no longer have; it is thought that thesespines functioned to trap prey (Blake, 1993).Decapods.—Despite the common assumptionthat shell-crushing crabs evolved during the Jurassic,in reality, only one group of lobsters (theNephropidae) is known to have evolved during thistime. All other groups evolved during either the134

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONPaleozoic, Triassic, or Cenozoic (Table 1). Hermitcrabs evolved in the Jurassic (Glaessner, 1969), andwhile they may crush mollusc shells (Vermeij,1987), it is difficult to assess their overall importanceas predators on molluscan groups. Hermit crabs canbe scavengers, carnivores, filter feeders, ordetritivores (Schram, 1986).Chondricthyes.—The rapid radiation of sharksand marine reptiles (Figs. 1, 4) in the middleMesozoic may have been triggered by the rise ofvast numbers of squids and actinopterygian fishes,including semionitids and basal teleosts (Theis andReif, 1985). The advances of increased swimmingefficiency and maneuverability, and sensory abilityenabled the neoselachians to pursue fast-swimmingthin-scaled fishes and squids in nearshoreenvironments (Packard, 1972).Hybodont sharks of the Triassic (Fig. 5.4) werelargely supplanted by the expanding neoselachiansharks during the Late Jurassic. The evolution ofhighly flexible, hyostylic jaws clearly marked anew level of sophistication in shark predation(Maisey, 1996). In hyostylic suspension the upperjaw is loosely articulated to the braincase and canby swung downward and forward on thehyomandibular bone. This enables sharks to thrustthe jaws forward and gouge out large chunks offlesh from prey. This adaptive breakthroughfomented an adaptive radiation of sharks, whichcontinued through the present day. Modern sharksshow varied feeding modes, including grasping andswallowing, suction feeding, cutting, gouging, andcrushing (Moss, 1977). One strongly modifiedclade from within the neoselachian shark lineageis the highly successful Batomorpha: rays andskates. These first appeared in the Late Jurassicbut diversified in the Cretaceous. The dental platesof rays and chimaeroids of this type may be usedfor digging up shelled invertebrate prey, and thencrushing them, leaving only fragments.Osteichthyes.—Among the Jurassic bonyfishes there is evidence for common piscivoroushabits; for example, the famed Upper JurassicSolnhofen Limestone provides many instances ofpredator-prey interactions (Voihl, 1990). Most“fossilized interactions” involve fish carcassescontaining partially ingested smaller fish. Jurassicpycnodont reef fish developed deep-bodiedmorphologies. For example, Daepedium (Fig. 5)was a deep-bodied Jurassic marine fish with heavyganoid scales, but with pebble-like teeth forcrushing. Jurassic pycnodonts evolved batteries ofrounded, shell-crushing teeth, plus specializednipping teeth. A few pycnodontids even developedstout pavement teeth possibly for crunching corals;rare specimens have been found with coralfragments in the gut (Viohl, 1990). The generalmorphology of these fishes overlaps with that ofdeep-bodied platysomids of the Paleozoic andmany reef-dwelling Cenozoic teleosts.Fish with durophagous dentition, such asSemionotidae (Lepidotes, Heterostrophus),Pycnodontidae (Mesturus), as well as hybodontsharks (Asteracanthusare) and chimaeroids(Brachymulus, Pachymylus, Ischyodus), arethought to have been predators of ammonoids fromthe Middle Jurassic of the Lower Oxford Clay ofEngland (Martill, 1990). Many well-preservedammonoid fragments are thought to be the resultof fish predation rather than physical factors(Martill, 1990). One ammonite specimen, aKosmoceras, was found to have bite marks thatwere similar to the dental battery of the semionotidfish, Lepidotes macrocheirus (Martill, 1990).Sea turtles.—Turtles are the only livingreptiles that are fully adapted to a marine existence(except for egg laying). Many fossil sea turtlesare only known from their plastron and carapace(Nicholls, 1997). The earliest sea turtles are thePlesiochelyidae, possible predators that lived inshallow, coastal waters.Sauropterygians: Plesiosaurs and pliosaurs.—The plesiosaurs are thought to have diversified intotwo major grades during the Jurassic (Fig. 5;Table2): the short-necked forms as fast-swimmingpursuit predators (pliosaurs), and the long-neckedforms as lurking ambush predators (plesiosauroidsand elasmosaurids). O’Keefe (2002), however, hascalled this an oversimplified view of their actualmorphological diversity.A cladistic analysis revealed that plesiosaurspresent a spectrum of body forms, and do not135

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 5—Mesozoic predatory marine vertebrates. 1, Cretaceous ichthyodectid teleost Xiphactinus. 2,Skeleton of the pycnodontid fish Proscinetes, Jurassic. 3, Skull of ichthyodectid teleost fish Cladocyclus,Cretaceous; Late Cretacous. 4, Hybodid shark, Hybodus. 5, Geologic distribution of marine reptiles, fromleft to right: ichthyosaurs, plesiosauroids, pliosauroids, teleosaurs, metriorhynchids, and mosasaurs.6–9, Skeletons of Mesozoic marine reptiles: 6, Mosasaur Plotosaurus, Late Cretaceous. 7, PlesiosauroidMuraenosaurus; Jurassic. 8, Pliosauroid Peloneustus; Late Jurassic. 9, Ichthyosaur Ophthalmosaurus;Early Jurassic. 10, Cretaceous foot-propelled diving bird Hesperornis. 11, Skeleton of Creaceous marinebird Ichthyornithyes. Figures 1–4, 10, 11 adapted from Benton (1997); Figures 5–9 from Massare (1987).136

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONdiscretely fall into two basic shapes: from the longnecked,small-headed elasmosaurs to the shortnecked,large-headed pliosauromorphs (O’Keefe,2002). By the Late Cretaceous, these pelagicmarine reptiles were globally distributed (Rieppel,1997). But the taxonomy of this group is still poorlyknown because of the inadequacy of type material,and preservational problems such as skull-lessskeletons (Carpenter, 1997).The plesiosaurs (clade Plesiosauria) wereamong the most diverse, geologically long-lived,and widespread of the Jurassic to Cretaceousmarine reptiles, with a fossil record extending fromthe Triassic-Jurassic boundary to the LateCretaceous (Figs. 5.5, 5.7) (Carroll and Gaskill,1985; Rieppel, 1997). These large reptiles (up to15 m long) had long paddle-shaped limbs(considered hydrofoils), short tails, long necks,needle-shaped conical teeth, and may have swamlike modern sea lions (Godfrey, 1984; Carroll andGaskill, 1985). Plesiosaurs are the only marineanimals in which both forelimbs and hindlimbsperformed as lift-based appendicular locomotion(i.e., as hydrofoils; Storrs, 1993). Pliosaurs had largeskulls (up to 3 m) and jaws with large fang-like teeth(Taylor, 1992), and were capable of dismemberingtheir prey (Taylor and Cruickshank, 1993).The small relative skull size and neck lengthof plesiosaurs, in addition to their dentition,suggests that many of them may have fed on smallfish and soft-bodied cephalopods; some may havealso strained the water for prey (Massare, 1987;Rieppel, 1997). Their evolution may have beenstimulated by the new abundance of largeractinopterygian fishes and sharks in offshoremarine environments. Plesiosaurs from the MiddleJurassic of the Oxford Clay also are thought to havebeen specialists on soft-bodied cephalopods andfish (Martill, 1990). The gastric contents of one LateJurassic plesiosaur, Pliosaurus brachyspondylus,included cephalopod hooklets (Tarlo, 1959). Wetzel(1960) has reported small ammonites in coprolitesattributed to plesiosaurs.Case studies from the Middle Jurassic OxfordClay, United Kingdom, provide a window into themarine trophic relationships of this time period. Thecarnivorous plesiosaurs (Liopleurodon, Pliosaurus)were considered to be at the top of the MiddleJurassic food chain, presumably feeding on fish and“naked” (without a shell) cephalopods (Martill,1990). The ichthyosaur Ophthamosaurus wasthought to be a specialist on naked cephalopods,while marine crocodilians (Metriorhynchus,Steneosaurus) presumably fed on fish and nakedcephalopods (Martill, 1986a, 1986b, 1990). Massare(1987) examined the conical pointed teeth form ofsome Jurassic ichthyosaurs and plesiosaurs, andconcluded that the teeth functioned to pierce softprey. Fish from these deposits were either planktonfeeders or fed on smaller fish, indicating that theMiddle Jurassic had a highly complex marine foodweb (Martill, 1990).Ichthyosaurs.—Lower Jurassic localities fromEurope (e.g., Lyme Regis, England; Holzmaden,Germany) indicate that marine reptile guilds at thistime were dominated by a diverse array ofichthyosaurs (Figs. 5.5, 5.9) (Massare, 1987). Gastriccontents from ichthyosaurs are known from LowerJurassic localities in Europe (Pollard, 1968; Keller,1976; Massare, 1987). The majority of preservedfood remains were cephalopod hooklets (Massare,1987, her table 1, p. 128). For example, preservedcephalopod hooklets (interpreted to be frombelemnites), fish remains, and wood were presentin the gastric contents from the small (< 3 m) LowerJurassic ichthyosaur Stenopterygius (Keller, 1976).Putative phragmoteuthid cephalopods also werepreserved in the stomach contents of the smallLower Jurassic icthyosaur, Ichthyosaurus (Pollard,1968). No belemnite hardparts (besides hooklets)have been reliably found in ichthyosaur gutcontents (Massare, 1987; but see Pollard, 1968).In contrast to the Lower Jurassic, Middle to LateJurassic assemblages indicate a number of changesin the vertebrate predatory ensemble (Massare,1987). Although the same functional feeding types(based on tooth form and wear) were present, thereptile groups were different, with pliosauroids andcrocodiles dominating the assemblages, and withreduced ichthyosaur diversity (Massare, 1987). TheMiddle Jurassic cephalopod-eating ichthyosaur,Ophthalamosaurus (Fig. 5.9) is inferred to have137

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002dived to depths of 600 m, based on an analysis ofits eyes and bone condition (Motani et al., 1999).Ichthyosaurs are also thought to have regurgitatedhardparts of indigestible food. In the Peterboroughquarry in England, Peter Doyle (unpublished,2002) discovered “ichthyosaur regurgitates” of 160million-year-old acid-etched juvenile belemnites.The acid-etched fossils indicate that they were oncewithin the ichthyosaur stomach.Marine Crocodiles.—Little is known about thefossil history of marine crocodiles (SuborderMesosuchia) compared to other marine reptiles (Huaand Buffetaut, 1997). The earliest crocodiles(Teleosauridae) are known from the Early Jurassic.This group shows little adaptation for marine life,and it is only because they are found in shallowmarine deposits that they are inferred to have beenmarine crocodilians (Hua and Buffetaut, 1997).Later in the Jurassic, these forms showed anatomicalfeatures that were more characteristic of life inmarine conditions (e.g., streamlined skull, reductionin bony armor, and reduction of the forelimb). Someforms (Steneosaurus) had long, slender teeth andmay have been piscivorous. Other teleosaurids hadblunt teeth, and more robust jaws, and are thoughtto have been durophagous predators on ammonoidsor sea turtles (Hua and Buffetaut, 1997).The Early Jurassic to Early CretaceousMetriorhynchidae include crocodilians with bothlong (longirostrine) and short (brevirostrine) snoutsthat may reflect dietary differences (Hua andBuffetaut, 1997). This group, because of its morestreamlined body form and a skull similar tomosasaurs, is thought to have been pelagic. Thestomach contents of a brevirostrine form(Metriorhynchus) contained ammonites, belemnites,pterosaurs (Rhamphorhynchus), and the large fishLeedsichthys (Martill, 1986b). Metriorhynchus andtheir ilk were probably lunging ambush predatorsthat captured their prey by sudden bursts ofswimming (Massare, 1987).Two other groups of marine crocodiles, thePholidosauridae (Lower-Upper Cretaceousboundary) and Dryosauridae (Upper Creatceous tolate Eocene) had fresh- and salt-water members (Huaand Buffetaut, 1997). The marine species ofPholidosauridae (Teleorhinus) are thought to havebeen piscivorous. Two groups of marine dryosaurids(Phosphatosaurinae and Hyposaurinae) are known:the phosphatosaurins had blunt teeth and robustjaws, and are thought to have preyed upon turtlesand nautiloids; the hyposaurins, most common inthe Paleogene, had long slender jaws and pointedteeth and were probably piscivorous (Hua andBuffetaut, 1997). Crocodilians are known to undergorapid changes in dental morphology in response toenvironmental change related to dietarymodification. It is thought that the piscivorous modeof life became more common after the Cretaceousmass extinction, when ammonoids and hardshelledmarine reptiles were not as common(Denton et al., 1997). However, the extinction ofthe dryosaurids in the Eocene is thought to haveresulted from the expansion of whales, which mayhave competed with them for food (Hua andBuffetaut, 1997). Crocodiles also regurgitate theirprey and such remains have been reported fromthe Paleocene of Wyoming (Fisher, 1981a, 1981b)but not from the Cretaceous.CRETACEOUS PREDATORSThe Early Cretaceous marked the beginningsof a major reorganization of marine predators,including the rise of neogastropods, numerouscephalopod predators, and several new vertebratepredatory guilds (Figs. 1, 4–6). The Early Cretaceoussaw the radiation of large teleost fish (> 3 m inlength) and sharks, and the non-dominance of marinereptiles (Massare, 1987). Massive shell-crushingmosasaurs (e.g., Globidens) did not evolve untilthe Late Cretaceous. This major specializedfunctional feeding type had been essentially absentthroughout most of the Mesozoic, since theextinction of Triassic placodonts (Massare, 1997).Late Cretaceous marine reptiles were dominatedby ambush predators, such as mosasaurs; marinefish (including sharks) were much more commonat this time and became more dominantcomponents of the predator functional feedingguild than ever before in the Mesozoic. Marinereptiles such as plesiosaurs and ichthyosaurs wereminor components of the Cretaceous predatory138

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 6—Representative Mesozoic marine invertebrate predators. 1, Ammonoid; note flutedornamentation. 2, Homarid lobster. 3, Belemnite; Belemnitella. 4, Naticid gastropod. 5, Brachyurancrab. 1, 3 from Tasch (1980); 2, 4, 5 from Robison and Kaesler (1987).food webs. In fact, ichthyosaurs and thepholidosaur marine crocodiles became extinct inthe Early Cretaceous, and pliosaurids were rare(Massare, 1997). Ichthyosaur extinction may havebeen associated with the Cenomanian-Turonianboundary events, following a severe depletion intheir putative belemnite prey (Bardet, 1992).During the Cretaceous and Tertiary, the offshoremovement of fast-moving fishes and coleoids mayhave stimulated evolution of offshore huntingamong the neoselachian sharks (Benton, 1997).Gastropods.—The Cretaceous marks animportant time of evolution in the predaceous shelldrillinggastropods. Several groups appeared and/or diversified during the Late Cretaceous and theirdistinctive drilling traces (Oichnus) becomecommon at this time (Kowalewski et al., 1998).Naticids (Figs. 6.4, 7, 8.3) become abundant in theLate Cretaceous as do their diagnostic boreholes(Fig. 8.4) (see reviews by Kabat, 1990; Kowalewski,139

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 20021993). The drilling frequencies in some Cretaceoussamples equal or exceed those observed in earlyCenozoic samples from the Gulf Coastal Plains(Kelley and Hanson, 1993, 2001; Kelley et al., 2001;see discussion below). These studies are possiblebecause naticids leave a unique type of countersunkdrillhole in scaphopod, bivalve, gastropod, andconspecific gastropod prey, as well as otherorganisms (Carriker and Yochelson, 1968; Sohl,1969; reviewed by Kabat, 1990; Kowalewski, 1993).Muricids (Neogastropoda) also evolved in theLate Cretaceous; predaceous muricids producecharacteristic cylindrical, non-chamfered boreholes(Fig. 8.3). Muricids form an ecletic gustatorygroup, ranging from herbivores to carrion feeders;however, most are shell drillers (Kabat, 1990).Shell drilling is most likely a pleisomorphicbehavioral trait within the Muricidae, although notall muricid genera bore through hard exoskeletons(Vermeij and Carlson, 2000). In contrast to naticidFIGURE 7—Diversification patterns of shell drilling through time: upper figure shows diversity of drillinggastropod clades through the Mesozoic and Cenozoic eras; lower figure shows frequency of drilledprey per million years through the Phanerozoic. Adapted from Sohl (1969) and Kowalewski et al. (1998).140

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONboreholes, holes drilled by muricids are considerablyless frequent in the Cretaceous than in most Eoceneand younger samples (Vermeij, 1987).Cephalopods.—As in the Jurassic, there werea host of belemnoids, ammonoids and nautiloidspresent in the Cretaceous (Fig. 6). All of these wereprobably nektonic predators, though their food mayhave ranged from zooplankton and larvae to othercephalopods (Packard, 1972). Cretaceous octopiare also known. While modern nautiloids appearto be extensively drilled by octopods (Saunders etal., 1991), no data exists for drilling predation onMesozoic nautiloids. As in the Jurassic, Cretaceousammonoids are thought to have been carnivorous.However, many species are thought to have eatenzooplankton (Ward, 1986).Stomatopods.—Stomatopods are known tohave extreme specialization in their limbs that isrelated to their predatory activities; no other majorextant malacostracan group has such specialization(Kunze, 1983). All stomatopods are obligatecarnivores (Table 1)—they eat only live prey—anduse their large raptorial second maxillipeds for preycapture (Kunze, 1983). These folding raptorialthoracopods can be used in two ways: as eithersmashing or spearing appendages. Folding raptorialthoracopods are known from the Carboniferouspalaeostomatopods (Schram, 1969), and within theMesozoic forms. The extant superfamilies ofstomatopods are thought to have originated in theCretaceous (approximately 100 Ma; Ahyong andHarling, 2000); however, the true fossil record ofthis group begins in the Cenozoic, and will bediscussed in that section.Based on fossil mouthparts, specialization forthe stomatopod’s zealous carnivorous life styleevolved very early, by the Late Devonian or EarlyCarboniferous, and the trend continued into theMesozoic (Schram, 1979). Mouthparts shred theprey, and food is stuffed into the mouth, not unlikethe way an energetic, hungry teenager feeds.Undigestible shell and cuticular material isregurgitated. The regurgitated remains have notbeen examined from a taphonomic perspective.Decapods.—In contrast, to stomatopods,decapods are not obligate carnivorous predators;most are scavengers (Schram, 1986). The majorityof the durophagous forms evolved in the Cenozoic,with just a few forms evolving in the Cretaceous(Table 1). The portunids and xanthids evolved inthe Cretaceous, and today are generalist andopportunistic feeders, occasionally eating hardshelledprey like molluscs. The slipper lobsters mayhave evolved in the Late Cretaceous, and they arethought to feed on scyphozoans (Table 1).Chondricthyes.—The neoselachian sharksradiated during the Cretaceous. Cartilaginous sharkskeletons do not fossilize well, and consequently,their teeth are used to infer their feeding behavior(Shimada, 1997). Despite popular accounts thatCretaceous sharks were some of most voraciousof all predators, it is still not clear whether theirattacks were on live or scavenged organisms.Healed injuries are usually taken to be attacks onlive prey, but these are rare in the fossil record.Necrosis around bite marks is also used to inferpredatory shark attacks (Schwimmer et al., 1997).Additionally, animals associated with sharkremains are usually interpreted as the shark’s lastmeal or as associative potential prey. For instance,in the Late Cretaceous Niobrara Chalk, a lamniformshark (Cretoxyrhina mantelli) is accompanied bywell-preserved cartilagenous skeletal elementspresumably from its last meal, the fish Xiphactinusaudax (Shimada, 1997).Late Cretaceous lamniform sharks(Cretoxyrhina) up to 6 m in length attacked orscavenged mosasaurs and perhaps plesiosaurs, and,in turn, were themselves possibly attacked orscavenged by other sharks (anacoracids; Shimada,1997). Dental arcades of Cretoxyrhina are similarto those of modern predatory mako sharks, and,not surprisingly, they belong to the FamilyLamnidae that includes the mako (Isurus), greatwhite (Carcharodon), and salmon shark (Lamna)(Shimada, 1997). Although shark taxa are differentthrough geologic history, Late Cretaceous sharks’functional feeding capabilities in ecosystems showparallels to modern sharks (Shimada, 1997).Direct evidence of shark predation on mosasaursis very rare. Shimada (1997) discusses severalreports of putative shark attacks on mosasaurs, either141

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 8—Traces of predation in fossil and Recent shells. 1, Shell of Cretaceous ammonitePlacenticeras with rows of punctures, probably made by a mosasaur. 2, Recent gastropod shell(Fasciolaria) exhibiting peeling damage inflicted by a callapid crab, ×1. 3, 4, Profile views of incompleteand complete gastropod drill holes in cross sections of bivalve shell: 3, cylindrical muricid drill holes;4, typical naticid holes; note parabolic cross section and central boss in incomplete borehole. 5, Locationsof most frequent drill holes in Recent bivalves from the Niger Delta. Redrawn from photographs in thefollowing sources: 1, Kauffman and Kesling (1960); 2, Bishop (1975); 3, Reyment (1971); 4, Sliter(1971). Figure modified from Brett (1990).with shark teeth embedded in bone, tooth marksslashed into the bone, or evidence of gastric-acidetching on putative prey items. Rothchild and Martin(1990) report on a shark tooth embedded in mosasaur(Clidastes) bone, which subsquently was repairedand ultimately caused spondylitus. Bite marks onmosasaurs without evidence of healing are alsoreported (Hawkins, 1990). Shark bite marks are alsoknown from elasmosaurid plesiosaur bones(Williston and Moodie, 1917; Welles, 1943).Late Cretaceous galeomorph selachian sharks(Squalicorax) are thought to have been scavengers142

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONpar excellence in the eastern Gulf Coastal Plainand Western Interior of the United States(Schwimmer et al., 1997). All living neoselachiansharks are carnivores, while galeomorph sharksmay prey on mollusks, crustaceans, andvertebrates; Squalicorax is also thought to scavengemany types of prey (see table 1 of Schwimmer, etal., 1997). Evidence of scavenging may includeembedded teeth which do not show evidence ofwound healing or tissue necrosis, and shark teethassociated with putatively scavenged remains(Schwimmer et al., 1997). Squalicorax, amoderately sized shark at 3.5 m, had serrateddentition of the “cutting type,” which may indicaterelatively diverse feeding strategies. IsolatedSqualicorax teeth, sometimes with associated bitemarks, were reported embedded in a decayedmosasaur vertebra and a juvenile hadrosaurmetatarsal (Schwimmer et al., 1997). Putative gutcontents from Squalicorax include mosasaur limbbones (Druckenmiller et al., 1993).Even fewer examples are known of sharkattacks on benthic invertebrates. Molluscan shells,such as those of inoceramid bivalves, are known tohave marginal edge fragmentation, and frequentlyare preserved as fragmented remains sometimesassociated with the putative shell-crushing shark,Ptychodus (Kauffman, 1972). One inoceramidspecimen, Inoceramus tenuis, is described as havingshell injuries on its left valve perhaps directlystemming from Ptychodus predation (Kauffman,1972). The right valve is uninjured, and Kauffman(1972) explains that this lack of injury is compatiblewith the life habits of the inoceramid, as the leftvalve was more exposed. Speden (1971) interpretsaggregations of fragmented inoceramid shells fromLate Cretaceous sites in New Zealand as evidenceof regurgitated or fecal material from vertebratepredators. Inoceramids are usually found in quiet,deep-water settings, either in chalks or black shales(Kauffman, 1972)—thus, any information on theirpotential predators would illuminate the littleknownpaleoecology of deep-water fauna in theLate Cretaceous.The batoid rays and skates first appeared in theEarly Jurassic and diversified during the LateJurassic and Cretaceous (Benton 1997; see alsoVermeij, 1987). These specialized elasmobranchswere adapted in large part for durophagous benthicpredation. Rays evolved stout hypermineralizedpavement plates for crushing hard-shelled prey, suchas crustaceans and molluscs (Fig. 9). Ray dentitionis thus similar to the pavement teeth of Devonianptyctodonts and rhenanids, and late Paleozoicholocephalans, Triassic–Jurassic semionotid fish,and Triassic placodonts. In all cases, crushing ofhard-shelled prey is inferred, but of these groupscertainly the batoid rays have been most successful.Many rays, exemplified by the cow nose rays, arecapable of excavating shallow pits in sandysubstrates in pursuit of infaunal bivalve, gastropod,polychaete, and other prey (see Fig. 12). Possibleancient ray pits have been reported from depositsas old as Late Cretaceous (Howard et al., 1977).Osteichthyans.—The neoselachian radiationsaw its counterpart in the Cretaceous osteichthyanteleost fishes (Fig. 5.1–5.3). The achievement ofimproved buoyancy via swim bladders,development of deep bodies, and anterior placementof pectoral and pelvic fins, represent coordinatedadaptations for increased swimming efficiency andmaneuverability during the Jurassic Period. Duringthe Late Cretaceous, the development of hingedmaxillae-premaxillae and highly protrusible mouthsfurther gave rise to a new mode of suctorial predatoryfeeding. These adaptations in turn fostered a majoradaptive radiation of neoteleost predators in the seaand in fresh water.In the Cretaceous, large basal teleosts clearlydominated in the intermediate- to large-sized fisheatingpredator guild. Many specimens of the largeXiphactinus (Fig. 5.1) from the Cretaceous of NorthAmerica have been found with ingested fish in thebody cavity. These include specimens from Kansaswith as many as ten fish in the stomach and a 4.5-meter specimen with a 1.6-meter relatedichthyodectid fish inside (Benton, 1997)! Specimensof the pavement-toothed Tribodus from theCretaceous Santana Formation of Brazil hadstomach contents that included shrimp andfragmentary molluscan shells (Maisey, 1996).Advanced acanthomorph teleosts evolved143

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 9—Cladogram of chondrichthyes (sharks) showing the repeated evolution of durophagy(indicated by “CRUNCH”; after Wilga and Motta, 2000).further defense in the Late Cretaceous–Tertiary,without substantial loss of mobility, in the form oferectile fin spines. These adaptations may indeedhave made the swallowing of whole prey sufficientlydifficult that individuals possessing longer, sharperfin spines were frequently spared and/or avoidedby experienced predators, thus driving adaptivetrends in neoteleosts (Patterson, 1994).Sea turtles.—Although modern turtles are allmorphologically similar, Mesozoic sea turtles werefar more disparate (Hirayama, 1997). There wereup to three separate radiations of sea turtles in theMesozoic (Nicholls, 1997). The Plesiochelyidaeevolved in the Jurassic. The second group(Pelomedusidae) is presently restricted to freshwater, but in the Late Cretaceous and Early Tertiary,members of this group were present in shallowmarine environments (Nicholls, 1997). The thirdgroup (Chelonioidea) first appeared in the late EarlyCretaceous and includes the Dermochelyidae,Cheloniidea, and the Protostegidae (Hirayama,1997). Of these, the jellyfish-eating stock(Dermochelyidae) arose in the Santonian and is stillextant, and the other omnivorous and herbivorousgroups (Cheloniidae) arose in the Aptian and arestill extant, having reached a diversity peak duringthe Late Cretaceous (Hirayama, 1997). TheProtostegidae were restricted to the LateCretaceous. The Chelonidae and Dermochelyidaesurvived the mass extinction at the end of theCretaceous, while most other marine reptiles, withthe exception of the crocodiles, went extinct. Theskull of Late Cretaceous Protostegidae turtles issimilar to that of the modern freshwatermolluscivorous turtle (Malayemys). Based on thissimilarity, it may have fed on pelagic ammonoids(Hirayama, 1997). The Protostegidae were thelargest sea turtles known, characterized by massiveheads, like that of the late Campanian Archelon. Thisgigantism was short-lived, as the Protostegidae wentextinct before the end of the Cretaceous (Hirayama,1997). The skulls of the Dermochelyidae areimperfectly known; however, it appears that thenarrow lower jaw and other skeletal features suggestthat the jellyfish-eating mode was developed duringthe Cenozoic (Hirayama, 1997).144

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONPlesiosaurs.—Plesiosaurs are thought to havebeen top predators of Mesozoic seas, but there hasbeen little evidence to support this claim (Sato andTanabe, 1998). The oldest firm evidence ofpredator-prey associations between ammonoidsand plesiosaurs is from a Late Cretaceous (UpperCenomanian) outcrop from Hokkaido, Japan. Fromthis locality, Sato and Tanabe (1998) describegastroliths, 30 isolated and disarticulatedammonoid jaws, a shark tooth, and molluscanshells from the putative gastric contents of apolycotylid plesiosaur. While the head of theplesiosaur was missing, comparable teeth in otherpolycotylids suggest that they were poorly adaptedto crush ammonite shells, and may have swallowedtheir prey whole. Plesiosaur gastric contents fromthe Early Cretaceous are known to includecephalopod jaws in association with gastroliths(Sato and Tanabe, 1998). Gastric residue from otherLate Cretaceous plesiosaurs, however, had fishvertebrae, pterodactyl bones, and thin-shelledammonites (Massare, 1987, her table 1, p. 128).Mosasaurs.—Mosasaurs originated anddiversified worldwide in less than 25 million yearsduring the Late Cretaceous (Fig. 10), but met theiruntimely demise during the end-Cretaceousextinction event (Lingham-Soliar, 1999). By the timeof their origin, the ichthyosaurs had gone extinct,and only a few plesiosaur families were still extant(Lingham-Soliar, 1999). Not since the Triassicplacodonts, had a reptile group so dominated thedurophagous functional lifestyle (Massare, 1987).Mosasaurs, with elongated snouts and elongated,fusiform bodies, include the largest marine reptilesever known (e.g., Mosasaurus hoffmanni at over17 m in length; Lingham-Soliar, 1998a). Becausethe Late Cretaceous sea levels were the highestrecorded during the Mesozoic, these giant reptileswere more likely to be preserved than were otherMesozoic marine reptiles, and so we have a betterunderstanding of their habits.Bone microstructure and bone density are usedto infer the ecological distribution of mosasaurs inthe water column (Sheldon, 1997). Reduced bonedensity of two common mosasaurs (Clidastes andTylosaurus) indicates that they lived at great depthFIGURE 10—Temporal distribution of severalgenera of mosasaurs (indicated by letters and theirapparent ammonite prey, based on bite/crushmarks from Cretaceous deposits of the WesternInterior Seaway. From Kauffman (1990).(Sheldon, 1997). Thus, even deep-water LateCretaceous ammonoids that were thought to usedepth as a refuge against predation (Westermann,1996) may not have been immune to their attacks,which may have fragmented the shells completely.Evidence of deep diving in mosasaurs comes fromavascular necrosis of their bones, indicating the“bends”—decompression syndome (Martin andRothschild, 1989; Taylor, 1994). Some mosasaurs,however, had pachyostosis (bone thickening), whichrequired that they increase lung volume to remainneutrally bouyant; in turn, increased lung volumemeans a larger rib cage, and thus more drag on theanimal, making it a slow swimmer (Sheldon, 1997).Mosasaurs with pachyostosis (e.g., Platecarpus)usually lived in shallow waters, but even these forms145

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002could dive to deeper depths as suggested byavascular necrosis in their bones. Similarly, skeletalelements of modern Belgula whales havepachyostosis, and consequently these whales spendmost of their time in shallow water. However,Belgula whales also are susceptible to the bends ifthey dive too deeply, and this is recorded in theirbones (Sheldon, 1997). Most deep-diving animalsusually do not suffer the bends, although someturtles may show skeletal evidence of having haddecompression syndrome (Motani et al., 1999).Consequently, considering evidence fromdentition, body form, thickness of skeletalelements, and avascular necrosis of the bones,mosasaurs are interpreted to have been top ambushpredators that once foraged in lagoonal to openoceanenvironments. Prey were swallowed whole,crushed, pierced, rammed, and shredded to namebut just a few means of prey demise (Lingham-Soliar, 1998a, 1998b, 1999). The dentition ofmosasaurs was so diverse that dentition patternsfit all five of Massare’s (1987) functional predatorygroups for Mesozoic marine reptiles (i.e., cut,pierce, smash, crunch, and crush guilds), afunctional feat accomplished in just a shortevolutionary time period (Lingham-Soliar, 1999).Because of these varied feeding modes, mosasaursmost likely fed on an array of benthic and pelagicorganisms (Lingham-Soliar, 1999).The West African Pluridens walkeri, forexample, had broad-based and short tooth crownsthat are speculated to be powerful enough to havecrushed thin-shelled invertebrate exoskeletons(Lingham-Soliar, 1998b). Globidens, from theUpper Cretaceous of Belgium, had rounded anddeeply wrinkled mushroom-shaped teeth that arethought to be specialized for crushing thick-shelledmolluscs (Lingham-Soliar, 1999). In fact, Globidens,along with another coeval mosasaur, Carinodens,shows the most remarkable durophagous crushingdentition since the demise of the placodonts(Lingham-Soliar, 1999). While the biomechanicalimportance of such dentition was discussed(Lingham-Soliar, 1999), it still remains to beindependently verified using experiments whetherthe many varieties of mosasaur teeth were capableof crushing ammonite shells or other putative prey,as well as how they crushed the shells.Such biomechanical studies would bebeneficial at least in coming to some conclusionas to how some ammonites received large holes intheir shells; and indeed, Kase et al. (1998)performed such tests using modern Nautilus shellsand a “mosasaur robot” modeled after a putativemosasaur predator of ammonites (Prognathodonovertoni). Seilacher (1998) also performedbiomechanical tests using steel pliers and nautiloidshells. While it can be argued that Nautilus shellsare not analagous to ammonite shells with respectto biomechanical loading (Tsujita and Westermann,2001), it is still important to experimentally testthe predatory hypothesis.Numerous specimens of the ammonite,Placenticeras, from the Late Cretaceous PierreShale and Bearpaw Formation of the westerninterior of North America, show putative mosasaurtooth marks (Fig. 8.1) (Kauffman and Kesling,1960; Kauffman, 1990; Hewitt and Westermann,1990) that have been reinterpreted to be limpethoming scars that were enhanced by diagenesis(Kase et al., 1998). Kase et al.’s biomechanical testsindicated that robot bite marks on live Nautilustypically had jagged edges that did not show theconcentric cracks characteristic of putative bitemarks in Placenticeras. The innermost nacreouslayer was shattered in the experiment, whereasinternal shell layers under the putative mosasaurbite marks on Placenticeras were not (Kase et al.,1998). They also found few examples ofPlacenticeras with holes corresponding tomosasaur jaw shape. Consequently, they concludedthat the holes in ammonites were limpet homescars, and not mosasaur predatory bite marks.Tsujita and Westermann (2001) rejected thefindings of Kase et al. (1998) and provided furtherobservations in support of the mosasaurian originof the holes. In fact, they pointed out that some ofthe experimental robot-induced holes that Kase etal. (1998) figured (e.g., their fig. 3b, p. 948)resembled those of putative mosasaurian bite markson Placenticeras meeki (Tsujita and Westermann,2001, fig. 3a,b, p. 251). Further, they argued that146

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONthe biomechanical loading was oversimplified, andneeds to be reanalyzed using various ammonitemodels in addition to testing various loadingfunctions attributed to the jaws of mosasaurs. Thelack of exact matching between jaw shape andputative bite marks is explained by the fact thatlike most marine reptiles, mosasaur jaws were notperfectly occluded; the lower jaw was loose enoughto pivot laterally. Further, while well-preservedlimpet fossils are found in the Pierre Shale, theyhave yet to be found in the Bear Paw Formation ofAlberta. Thus, Tsujita and Westermann concludethat putative predatory holes on ammonites maybe only rarely associated with limpet home scars,and the vast majority are from mosasaur predation.No one has done a quantitative comparision of thesize of the bite marks, the size of the limpet homescars, the diameter of the preserved limpets, andrelated it to the range of tooth sizes found incontemporaneous mosasaurs.Gastric contents from mosasaurs includecephalopod hooklets, fish, belemnites, turtle bones,and birds (Massare, 1987, her table 1, p. 128). Forexample, gastric contents from a single specimenof the mosasaur Clidastes included a marine sharkand a diving marine bird, Hesperornis (Martin andBjork, 1987). At least one squid gladius from thePierre Shale exhibits bite marks attributable to amosasaur (Stewart and Carpenter, 1990). Dollo(1913) reported a broken test of the echinoid,Hemipneustes, between the teeth of the mosasaurCarinodens; and numerous ammonites have toothmarks, presumably from mosasaur predation(Kauffman and Kesling, 1960; Kauffman, 1990).Some of these tooth marks, however, may also belimpet homing scars on some specimens (Kase etal., 1998). To date, no ammonites are known frommosasaur gastric contents (Martin and Bjork, 1987;Massare, 1987), and this may be due to taphonomicbias against the preservation of aragonititc ammonitesin gastric contents (Tsujita and Westermann, 2001).Sea and Shore Birds.—Finally, in the LateCretaceous, two orders of marine diving birdsevolved: the flightless, foot-propelledHesperornithiformes and the swimming-wingedIchthyornithiformes. Both taxa had elongate beakswith rows of sharply pointed teeth, presumably forfish capture. Fish remains have been found incoprolites associated with Hesperonis (Benton,1997, p. 273). Presumably, these taxa filled theguild presently occupied by diving sea birds,although the Cretaceous orders were evolutionarydead ends. One mosasaur specimen also containsingested Hesperonis skeletal elementsJURASSIC–CRETACEOUSBENTHIC PREY AND THEIRPOSSIBLE ANTIPREDATORYRESPONSESPossible Behavioral Responses ofInvertebrates.—During the Jurassic and Cretaceous,a host of organisms from sponges to worms,barnacles, and bivalves independently evolved anability to bore into stiff mud, rock, carbonate,hardgrounds, shell substrates, and wood (Palmer,1982; Seilacher, 1985; Wilson and Palmer, 1990,1992). Submarine crypts and caverns seeminglyprovided a refuge for certain primitive groups, suchas sclerosponges, many bryozoans, sedentary tubedwellingpolychaetes, and pediculate brachiopods(Palmer, 1982; Wilson and Palmer, 1990). Anumber of sedentary invertebrate groups persistedon exposed hard substrata during the Mesozoic.But these, in particular, show allegedly strongantipredatory skeletal adaptations (Fig. 3): they arestrongly cemented (oysters, corals, barnacles), havethick, heavy shells (e.g., rudists, oysters),camouflage, and spines/spicules or toxins.A major decline in free-resting benthicinvertebrates occurred in the Mesozoic (relative tothe Paleozoic). Quasi-infaunal forms, such asgrypheid and exogyrid oysters, remained commonon Mesozoic soft substrates, but these organismswere partially hidden and evolved thick, robustshells (Fig. 3). Exposed epifaunal brachiopods,corals, and crinoids were greatly reduced or absentfrom shallow marine soft-substrate settings duringthe Mesozoic (Thayer, 1983; Vermeij, 1987).Thayer (1983) argued that this decline in epifaunalsuspension feeding may have been fostered by therise of deeply burrowing infaunal “bulldozers,”147

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002especially siphonate bivalves, during the Mesozoic.Moreover, predator grazing may be a further causefor the decline of epifaunal suspension feeders(Stanley, 1977).Vermeij (1977) hypothesized that shellbreakingpredation became a more important causeof mortality and was a driving force in evolutionfrom the Mesozoic to Cenozoic. He termed thispostulated major escalation of predator-preyinteractions the “Mesozoic Marine Revolution”(MMR); in fact, most of the trends he described inadaptive morphology continued from the Mesozoicto the Recent, so the present discussion combinesevidence from both eras.Gastropods.—Gastropods show a number oftrends, probably in response to shell crushing/drilling predation pressure (Vermeij, 1977, 1983,1987). These include a further decrease in umbilicateand loosely coiled shells. The few remaining looselycoiled gastropods lived within sponges (sillquarids),or were cemented (vermetids), the latter sometimesforming large aggregates, or “reefs” (Vermeij, 1983).Other trends among Mesozoic gastropods includeincreased proportions of thick-lipped shell apertures,slit-like apertures, varices, and spines or knobs. Thesimultaneous increase in these features suggests acommon evolutionary pressure, presumably theincrease in durophagous predation (Vermeij, 1983,1987). However, as this review shows, most of thesedurophagous predators are generalist feeders,feeding on a variety of hard-shelled prey, and notjust molluscs. More work must be done to determinewhether shell ornamentation really reducessuccessful predation. Shell sculpture and variceshave been shown to decrease predation by sea starsand durophagous crustaceans in laboratoryexperiments (Donovan et al., 1999). Fish predationis also deterred by shell sculpture (e.g., stout spines)on gastropods (Palmer, 1979). However, fieldexperiments demonstrate that shell sculpture maynot always be a deterrant to predation for somegastropods (Ray and Stoner, 1995). For queen conch(Strombus gigas), living in aggregations andattaining large overall size was found to be moreimportant in deterring predation than was shellsculpture (Ray and Stoner, 1995). Longer spines andheavier shells do not necessarily reduce predationmortality in queen conch, especially when predatorsattack through the aperture, as do crustaceans andpredatory molluscs (Ray and Stoner, 1995).Bivalves.—Much of the literature concerningbivalve shell ornamentation in relation to predationhas been based on largely circumstantial evidence(Fig. 3) (Harper and Skelton, 1993). Spondylidbivalves provide an interesting example. These spinyepifaunal bivalves appear in the Middle Jurassic andshow increasingly spinose shells up to the presentday (Harper and Skelton, 1993). However, thesespines apparently do not increase shell strength(Stone, 1998; Carlson, pers. comm., 2000), but doincrease effective size and make shells more difficultto attack. Spines are also commonly worn off, andit is not known how this affects the survival of thesecemented groups (see Logan, 1974).Shell microstructures and the development ofspines in some groups of bivalves may haveoriginated in their Paleozoic ancestors (Table 5).Additionally, changes in thickness and arrangementof shell microstructure may also be primarilycontrolled by water chemistry and temperature,rather than by predation. Some microstructures,however, may secondarily function to reduce crackpropagation, such as cross-lamellar structures, andincrease abrasion resistance (Currey and Kohn,1976). However, there are many ways to build crosslamellarstructures (Schneider and Carter, 2001).Shell microstructure such as spines, thickness ofparticular shell layers, and types of shell layers mayreflect a phylogenetic constraint. Thickening of shellmargins through extensive inductural deposits maybe related to photosymbiosis, and not directly topredation (Schneider and Carter, 2001).Cardiid (Jurassic to Recent) bivalve shellmicrostructure exhibits several evolutionary trendsthat may not be related to predation (Table 5). SomeCretaceous cardiids evolved stronger reflection ofthe shell margins, and increased thickness orsecondary loss of the ancestral prismatic outer shelllayers. However, these changes appear to be relatedto water chemistry and temperature. For example,microstructural convergences may be directly orindirectly tied to ocean chemistry and temperature:148

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONcool climate may lead to thicker prismatic outer shelllayers in some cardiid bivalves, similar to those ofvenerid bivalves (Schneider and Carter, 2001).Cardiid bivalves are known for their spines.Cardiid spines can be formed in different ways: 1)by a mantle that is strongly reflected over exteriorshell surfaces; 2) by extensions of the normal outer,or outer and middle shell layers; or 3) by theperiostracum (formed on the undersurface of theperiostracum, and cemented to the shell exterior).Cemented periostracal granules or spines inCarboniferous astartids, and in three subfamiliesof cardiids (i.e., colpomyid and mytilidpteriomorhians and trigonioid palaeoheterodonts),suggest that periostracal mineralizing isplesiomorphic for the bivalvia, and is merelyretained by many anomalodesmatans (Schneiderand Carter, 2001). Thus, some spine forms in theseMesozoic and Cenozoic groups may be partly orlargely the result of phylogenetic and physicalenvironmental contraints.Ammonoids as Prey.—Ammonoids are knownto have sublethal injuries from the Mesozoic thatmay not have affected their bouyancy as much assublethal injuries in nautiloids (Kröger, 2002).Unfortunately, little quantitative data exists forshell repair in ammonoids during this time.Westermann (1996) suggests that ammonoids livedin deeper-water areas to avoid predators, especiallyin the Cretaceous. However, there is now extensiveevidence that marine reptiles were able to dive todeep depths during this time.Vermeij suggests that shell repair increases inammonoids during the Mesozoic, although he makesa plea that more data be accumulated in order toreally assess this claim (Vermeij, 1987, p. 283–284).To date, little if any data exist to analyzeantipredatory features and predation on Mesozoicammonoids. Because shallow-water and deep-waterforms were abundant, and because ammonoidsoccupied many different habitats within thosesettings during the Jurassic and Cretaceous, theywould be ideal organisms by which examineenvironmental records of predation.Ward (1986, p. 818) states that there is“abundant evidence…suggesting that predation byshell-breaking predators commonly occurred, forbreak marks are common in Jurassic and Cretaceousammonites,” but he does not provide data to supportthis statement. Data are needed on the number ofammonoid shells with evidence of healed injuries,and on whether this varies by environment ofdeposition, and on shell ornamentation through theMesozoic. Equally important would be acomparative examination of healed scars onmicroconchs versus macroconchs, and on demersalversus more planktonic forms of ammonoids.A few direct records of predation on ammonoidshave been reported. Several examples of ammonoidswith smaller ammonoid shells in their bodychambers are cited above. Ammonite shell fragmentsare known from fish feces from the SolnhofenLimestone in Germany (Schindewolf, 1958). Anunknown marine reptile apparently left twentypossible bite marks on a specimen of the MiddleJurassic ammonoid Kosmoceras gulielmi from theMiddle Oxford Clay, England (Ward andHollingworth, 1990). The bite marks are surroundedby an inclined ring of fractured shell, and becausethere was no sign of healing, the bites are consideredto have been fatal to the ammonoid (Ward andHollingworth, 1990). It is also thought, because ofthe diversity of predatory marine reptiles, fish, andbelemnites, that ammonoids may have lived indeeper, slightly more oxygen-deficient waters at thistime (Westermann, 1996). Vermeij (1987, p. 283)reviewed the limited anecdotal informationconcerning shell repair on ammonoids and suggestedthat the incidence of shell repair was low in Earlyand Middle Jurassic ammonoids.If benthic durophagous predators were preyingon ammonoids, the ammonoid prey should show atrend in antipredatory ornamentation and shellrepair through the Mesozoic in accordance withthe Mesozoic Marine Revolution theory of Vermeij(1977, 1987). As is the case for the Triassic, littleis known about antipredatory effects of ammonoidshell shape and sculpture, although shell crushingmarine reptiles, fish, and other cephalopods werequite diverse in the Jurassic and Cretaceous.Costae and spines in ammonoids have beenconsidered to be antipredatory (Logan, 1974;149

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Westermann, 1996). Costae presumablystrenthened the shell against predators (Checa andWestermann, 1989; Westermann, 1990), yet thereare numerous examples of smooth-shelledammonoid groups living contemporaneously insimilar habitats. Spines and spine-like antipredatoryfeatures of adult ammonoids include spines onancyclocone ammonoids, and protuberances suchas lappets, rostra, and horns on microconchs(Westermann, 1996).Ward (1981) argued, based on figures in theTreatise of Invertebrate Paleontology, that manyof the highly sculptured Cretaceous ammonoidsevolved primarily as a defense against shellcrushing predators. Ammonites showed limitedtrends in shell ornamentation during the Mesozoicrelative to their Paleozoic counterparts (Ward,1981). Ward documented trends toward increasedfluting and ribbing in ammonite shells during theJurassic and Cretaceous.There is only a slight increase in fine-tomoderateshell ornament in Middle to Late Jurassicammonoids compared to Early Jurassic ammonoids(Ward, 1981, his fig. 2, p. 98). Additionally, thereis no difference in moderately coarse to very coarseshell ornamentation between Early Jurassic andMiddle to Late Jurassic ammonoids (Ward, 1981).The proportion of ammonoids with very coarseornamentation stays the same through the Jurassic;moderate to strong ornamentation does increase byabout 10%, but then changes little throughout therest of the Mesozoic. Vermeij (1987) surmised fromthis data that armor in ammonoids was no longer asuccessful anti-predation strategy after the Turonian.Some long-lived groups of ammonoids, such as theLytoceratidae and Phylloceratidae, remainmorphologically similar through their geologicrange, while other long-lived ammonoid familiesare morphologically diverse (Ward and Signor,1983). It is not known what causes morphologicalstasis in some forms but not in others.Ward (1981) found that 40% of LowerCretaceous, and approximately 42% of UpperCretaceous ammonoids had moderate to strongribbing on their shells (his fig. 2, p. 98); all othershell surface types (i.e., no ornamentation, fine tomoderate ribbing, and very strong ribbing)appeared to be similar for both time periods.Essentially, there appears to be no difference inornamentation between the Lower and UpperCretaceous ammonoids, despite the origin andevolution of durophagous mosasaurs in the UpperCretaceous. Ward did not differentiate betweenbenthic, planktonic, and pelagic ammonoids.Ammonoid shell shape was also examined fromthe Berriasian to Maastrichtian, and little change wasnoted for coarsely ornamented ammonoids (Ward,1986, his fig. 3, p. 9). Non-streamlined (nonornamented)forms stay roughly the same throughtime, with slightly more in the Berriasian. Thus, itappears that shell ornamentation in ammonoids isnot a direct result of predation.A great deal of work remains to be done ontesting various ammonoid shell forms in relationto predation. For example, it would still bebeneficial to examine benthic versus pelagicammonoids to determine if there is a differencein shell ornamentation between these two types.A shift to more offshore ammonoid faunas inthe Late Cretaceous, and the extinction of nearshoreNorth Pacific forms prior to the Maastrichtian(Ward, 1986), may have resulted from increasedcompetition and/or predation. However, there werenumerous offshore, deep-diving predators in theLate Cretaceous (e.g., globally distributedmosasaurs, sharks, and other fish) that may havepreyed on pelagic ammonoids and other pelagicinvertebrate fauna.The last ammonoids of the Late Cretaceous arebest known from continental slope deposits, andinclude nektonic and planktonic forms (Ward,1987); curiously, it is the durophagous nautiloidsthat survived the Cretaceous-Tertiary extinctionevent, and not the pelagic, perhaps chieflyplanktivorous ammonoids of that time. Thisextinction may not be directly related to theirplanktic habit, but rather to the fact that thatammonoids had a planktic part of their early lifecycle, whereas nautiloids had a benthic stage(Ward, 1986). However, nautiloids are dependenton other invertebrates for food, including crustacea,which have a planktonic period in their life cycle150

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONColeoids.—Other nektonic organisms gave uparmor in exchange for more efficient swimmingduring the Mesozoic and Cenozoic (Vermeij, 1983).Soft-bodied coleoid cephalopods traded off externalarmor for increased speed and mobility and evasivedefenses, such as sepia ink for camouflage (Packard,1972; Lehmann, 1975). The majority of gastriccontents preserved in marine reptiles, however,consist of hooklets from belemnites, and few fromnaked cephalopods. While the belemnites did notsurvive the Cretaceous mass extinction, the softbodiedcoleoids did. They faced a renewed groupof predators in the Cenozoic—the marine mammals.Despite two revolutions in their predators, theirmorphology has remained remarkably static.Decapods and Ostracodes.—Little is knownabout predation on decapods during the Cretaceous.Evidence of drilling predation in Cretaceous toRecent ostracode assemblages from Texas includesdrillholes from juvenile naticid gastropods(Maddocks, 1988). It is not known if ostracodesdeveloped antipredatory armor, although theincidence of drilling appears to increase from theCretaceous into the Tertiary, but then declines inthe Holocene (Maddocks, 1988). It is thought thatsmooth shells may be preferentially drilled, or atleast may make drillholes more discernable to thepaleontologist (Maddocks, 1988). Because of theirabundance and worldwide distribution in a varietyof environments, ostracodes would provide animportant database from which to test the variousmarine revolutions; but they remain little studied.Echinoderms.—Regional large deposits ofcrinoid grainstones and packstones (encrinites) arenot present after the Jurassic (Ausich, 1997). Thepresence of regional encrinites since the Ordovicianillustrated the domination of many shallow-shelfenvironments by crinoids and other stalkedechinoderms perhaps for millions of years up tothe Jurassic (Ausich, 1997). Comatulid crinoidsevolved rapidly from the stalked forms during theLate Triassic–Early Jurassic (Meyer and Macurda,1977), but their diversity remained fairly low (fivespecies) during the Jurassic. Modern comatulidsare known to be preyed upon by reef fishes (Meyerand Ausich, 1983; Meyer, 1985). There is alsolimited information concerning predation oncrinoids during the Jurassic (Schneider, 1988). Theoffshore retreat of “primitive” groups, such asstalked crinoids, has been suggested to be a generaltrend that might be related to increased predationpressure (Jablonski et al., 1983; Vermeij, 1987;Bottjer and Jablonski, 1988; Jablonski and Bottjer,1990). For example, Meyer and Macurda (1977)documented an offshore migration of stalkedcrinoids during the Jurassic. This onshore-offshorepattern in crinoid distribution needs to be reexaminedin light of new data.Most isocrinids (except for Pentacrinitidae)lived in shallow waters until the Mid-Cretaceous,whereas in the Cenozoic these forms inhabiteddeeper water (Bottjer and Jablonski, 1988).In the Early Jurassic, the biggest evolutionaryinnovation in echinoderms appeared with the adventof irregular echinoids (Simms, 1990). The flattenedtests of these creatures are thought to have been anaptation that provided greater stability within thesubstrate. At the same time, the periproct movedaway from the apex of the test, in accord with theirsediment-eating habits. By the Middle Jurassic,endobenthic irregular echinoids had evolved andrapidly diversified. Today, their descendantscomprise nearly half of all extant echinoids (Simms,1990). The aboral spines and the anal sulcus of thesecreatures (e.g., Galeropygidae) were consistent withtheir endobenthic lifestyle (Simms, 1990). Theevolution of pencillate tube feet in these groupsallowed them to pick up finer sedimentary particlesvia mucous adhesion (Simms, 1990). A peri-oraltube foot also allowed them to expand into newtrophic realms. Was this endobenthic lifestyleprovoked by predation, or merely by the opportunityfor better feeding conditions? It should be noted thatepibenthic echinoids were also diversifying at thistime, with the Cassiduloids and their offshoots.Fish are the dominant predators of modernophiuroids (Aronson, 1988). Little is known aboutpredation on Mesozoic ophiuroids, although therate of arm regeneration appears to be low forJurassic compared to Recent ophiuroids (Aronson,1987, 1991). However, there is no clear evidencethat ophiuroids developed antipredatory armor, as151

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002their morphology has remained relatively the samesince their origin in the Ordovician (Aronson,1991). It is possible that they developed betterautotomization of their arms, like some crinoids;or maybe they developed into distasteful prey(Aronson, 1991). The evolutionary radiation of thespatangoids (and holasteroids) in the Cretaceousmay have been the result of the Jurassic innovationof pencillate tube feet, a feature shared with noother echinoid group (Simms, 1990). EarlyCretaceous spatangoids (Hemiaster eleganswashitae) are reported to have drillholes fromparasitic gastropods (Kier, 1981). Parasitic drillingsare commonly associated with deformation of theechinoid ossicles where the parasite housed itself.Fish bite marks are well preserved on completeasterioids and asteroid ossicles from the LateCretaceous White Chalk of northwestern Germany(Neumann, 2000). On some specimens, serratedtooth marks may be related to galeoid sharkpredation. Regurgitate pellets are also common inthe White Chalk, and indicate that predation bythese durophagous fish may have been sizeselective (Neumann, 2000).Vertebrates.—As with late Paleozoic fish,armor does not appear to have been a significantpart of the response to escalation among Mesozoicmarine vertebrates. With the exception of relativelyslow-moving placodonts and marine turtles, noneof the marine vertebrates developed any unusualarmor during the Mesozoic. Indeed, withinactinopterygian fish there is a distinct trend towardreduction of ganoid scales in favor of lighter andless protective cycloid and ctenoid types (Patterson,1994; Benton, 1997). Presumably, this scalereduction reflects the ineffectiveness of dermalarmor against large predators, which demonstrablyswallowed prey whole (Voihl, 1990). This furtherreduction in armament is clearly coordinated withthe development of improved swimming speed,buoyancy control, and maneuverability in theCretaceous teleosts. In turn, this increased mobilitymay well have triggered adaptations for improvedspeed, maneuverability, and/or stealth among largerpredators, such as the neoselachian sharks,plesiosaurs, ichthyosaurs, and mosasaurs.CENOZOIC PREDATORSThe Cretaceous-Tertiary extinction had adevasting impact on pelagic ecosystems. Ammonitesand belemnoids, as well as large vertebrate predators,were decimated by this event. All of the marinereptilian predator guilds, except sea snakes and seaturtles, became extinct during this crisis—includingmosasaurs, plesiosaurs, and ichthyosaurs, in additionto the flying pterosaurs. This left only the highlysuccessful neoselachian sharks and teleost fishes inthe vertebrate predator realm. Marine mammalsemerged in the Eocene to essentially take over theecological void left by Mesozoic marine reptiles(Table 4). In fact, tooth dentition in marine mammalsclosely parallels that of the Mesozoic marine reptiles(Massare, 1987, 1997).Conversely, many benthic invertebratepredators, such as naticid and muricid gastropodsand various decapod crustaceans, were seeminglylittle affected by the terminal Cretaceous extinctions.Several groups of shell-drilling predators evolvedor diversified within the Cenozoic (for review, seeVermeij, 1987); some groups, such as theneogastropods, evolved in the Late Cretaceous.Prosobranch gastropod predators, the dominantdrillers, were much more common in the Cenozoicthan at any other time, though the Mesozoic recordneeds to be more throughly examined (Kowalewskiet al., 1998). The record of octopod shell drilling ischiefly Cenozoic, with the soft-bodied octopodfossil record primarily within the Cretaceous toPaleogene (Engeser 1988; Harper, 2002).Several major groups of vertebrate shellcrushingand shucking predators that may haveseriously impacted benthic and pelagic marinebiotas evolved or diversified during the Cenozoic:shell-crushing sea turtles (i.e., the single genusCaretta), the coral reef teleost fishes and otherteleosts, rays and skates, diving marine and shorebirds, pinnipeds, sea otters, gray whales, andhumans. Among mammals, the origination ofpinnipeds (seals and walruses), the cetaceans(especially the gray whale) in the Eocene, and seaotters (Carnivora; Family Mustelidae) in the lateMiocene also potentially impacted Cenozoic152

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONTABLE 4—Cenozoic marine vertebrates and their functional feeding groups.TaxonomicGroupFunctionalFeedingGroupPreyForensic Evidence(potentially traceable in thefossil record)Mustelidae(Sea otters)GeneralistcarnivoresAbalone (Haliotis ); seaurchins (Strongylocentrotusfranciscanus, S. purpuratus );kelp crabs (Pugettia ); rockcrabs (Cancer ); turban snails(Tegula ); octopus (Octopus );bivalves (Tivela, Saxidomus,Tressus ); sea stars (Pisaster )Cetaceans Generalists Krill, whales, dolphin, squid,callianassids, small bivalves;no specialistsPinnipeds Generalist Two genera are specialists onmolluscs and crustaceaSea Turtles Generalist One genus (Caretta ) specialiston molluscsShell damage consisted of:fractured middle sections ofshells as a result of being hit withstones by sea otter; larger shellsmay have fractured middlesections; edge damage may bedue to otter gnawing on theedges of the shell or chippingwith a stoneGray whales suck pits in thebenthos to gather food; pits maybe ephemeral, but may bepreserved; no other informationavailableWalruses may leave marks onshells, but no record of this as yetCaretta may leave marks onmolluscan prey, but no record ofthis as yetSea Snakes Generalists Crush prey, but no record of theirpredatory prowess as yetDiving Marine Birds Generalists Eat crustacea, molluscs, fish;one genus appears tospecialize on molluscsMarine Crocodiles Generalist Birds, fish, turtles, humans,golf balls, etc.Extensive literature on birds andhow they forensically alter preyNo forensic information availablebenthic invertebrate prey. Vermeij (1987) reviewedthe molluscivorous habits of some of these groups,and here we discuss their more generalist feedingbehavior, add or update several other groups, andsuggest possible alternative scenarios to hisescalatory hypothesis.Stomatopods.—Stomatopod crustaceans areobligate carnivores and vicious predators.Stomatopods that crush the shells of their prey bypounding them with blunt expanded segments oftheir maxillipeds (e.g., Burrows, 1969) did notevolve until the Cenozoic (Hof and Briggs, 1997;Hof, 1998). Two major groups of stomatopods existtoday: the squilloids and the gonodactyloids, whichhave very different means of feeding. The squilloidseither attack prey with their dactylar spines, or graspprey between the toothed margins of their propodusand dactylus (Kunze, 1983). Squilloids typicallyprey on fish, polychaetes, and very small crustaceans(Schram, 1986). In gonodactyloids, the propodus isswung from an anterioventral position, and prey is“smashed” on the lower part of the dactylus (Kunze,1983). Gonodactyloids feed typically on hardshelledprey like molluscs and large crustaceans(Schram, 1986). Both types, however, can also scoopup prey from the benthos with their maxillipeds.153

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002The gonodactyloids can smash small to largeholes in molluscan prey (Hof, 1998; Ahyong andHarling, 2000). Stomatopods can also sheargastropod shells in half and break the outer lip(Geary et al., 1991). Additionally, small puncturewounds in molluscan prey, called ballistic traces(e.g., the trace fossil Belichnus), are attributed tostomatopods (Pether, 1995). Most of the extantgroups have an actual fossil record extending backonly to the Eocene, with the shell-smashinggonodactylid group originating in the Miocene(Hof and Briggs, 1997; Hof, 1998). Approximately400 living stomatopod species are recognized(Manning, 1995).Despite their long history, only a few examplesof gonodactylid shell-breaking predation areknown from the Neogene fossil record. Geary etal. (1991) described a few cases of putativestomatopod shell damage from Pliocene localitiesin Florida. Baluk and Radwanski (1996) alsodocumented stomatopod damage on diversegastropods from Miocene localities in Europe.Stomatopod predatory damage should be easilyrecognized, and documentation of this damage inmore assemblages would enhance thepaleoecological picture of these creatures.Brachyuran crabs and lobsters.—The secondmajor wave of crustacean adaptive radiationoccurred in the Paleogene. Brachyuran crabs hadappeared in the Mesozoic, but new families of crabssuch as the Portunidae, Cancridae, Grapsidae, andOcypodidae arose in the Eocene (Table 1). Mostbrachyuran crabs are generalist and opportunisticfeeders, and few are durophagous (Table 1).Brachyurans with heavily toothed claws apparentlyevolved in the Paleocene (Vermeij, 1983), but thishas not been studied in detail. Crushing claws area formidable tool for peeling and crushingmolluscan shells and also for crushing othercrustacea or hard-shelled prey. The distinctivepeeling of calappid crabs has been documented infossil and recent shells (Bishop, 1975; Vermeij,1982, 1987). The parthenopid crabs, whichoriginated in the Late Cretaceous, are known toeat molluscs only in the laboratory (Vermeij, 1978),and the few reports available show them eatingpuffer fish or non-molluscan invertebrates(Table1). Clearly more work needs to be done onthe parthenopid crabs. Durophagous cancrid crabs(Cancer spp.) eat a diversity of prey, such aspolychaetes, squid, crustaceans, fish, andechinoderms, following the dominant macroinvertebratesin the habitat; whereas Ovalipid crabsmay predominantly eat molluscs (Stehlik, 1993).Lobsters also crush shells, but usually onlyfragments are left (Cox et al., 1997). In modern seas,rock lobsters are known to prey extensively onmolluscs, such as abalones and turban snails in somelocalities (Van Zyl et al., 1998; Branch, 2000), andechinoderms in others (Mayfield et al., 2001). Verylittle is known about lobster foraging and how it wouldaffect the fossil record of invertebrate hard-shelledprey (reviewed in part by Walker et al., 2002).Gastropods.—Gastropod predators that chipand wedge open molluscan prey (e.g., Buccinidae,Fasciolaridae, and Melongenidae) originated in theLate Cretaceous, but diversified in the Cenozoic;however, the shell-chipping record in prey shellsis known only from the Pliocene (Vermeij, 1987).Buccinid gastropods chip their outer lips in theprocess of preying on other mollusks, and thensubsequently repair their self-inflicted breakage(Carriker, 1951; Nielsen, 1975). Dietl and Alexander(1998) noted that this type of lip damage occurs inbuccinids as old as Miocene. Older buccinids, whichrange back to the Late Cretaceous, have not yieldedevidence of this distinctive lip damage. Hence, theshell-prying habit of buccinids may have evolvedwithin the Neogene.The best evidence for predation in the Cenozoicfossil record comes from predatory drillholespreserved in prey ranging from protists, such asforaminifera, to many phyla of invertebrates, suchas bryozoans, molluscs, brachiopods, andechinoderms (Carriker and Yochelson, 1968; Sohl,1969; Taylor, 1970; Sliter, 1971; Bishop, 1975;Boucot, 1981, 1990; Bromley, 1981; Vermeij, 1987;Kabat, 1990; Kowalewski, 1993; Kowalewski andFlessa, 1997). Prosobranch gastropods are theprimary shell drillers in marine environments,although nudibranchs (Vayssiereidae), flatworms,nematodes, and the protist foraminiferans have also154

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONevolved drilling apparatuses (Woelke, 1957; Sliter,1971; Carriker, 1981; Hallock and Talge, 1994).Eight families or superfamilies of molluscshave evolved drilling behavior. Naticids are the beststudied, but the other groups of shell drillers werereviewed by Kabat (1990). Gastropods, inparticular, have evolved a variety of means to preyupon hard-shelled prey, and one major family(Cassidae) and two little known groups(Marginellidae and Nassariidae) of shell drillersoriginated in the Cenozoic; the naticids andmuricids persisted from the Mesozoic into theCenozoic (Fig. 7). In drilling gastropods (e.g.,muricids, naticids), the small rasping organ (theradula) and the accessory boring organ are used todrill holes in prey shells (Carriker, 1969). Predatorycephalopods, such as the octopus, also use theradula to bore into mollusc shells.Capulids (Capulidae, Mesogastropoda) arespecialized ectoparasites on molluscs andechinoderms (Kabat, 1990). Capulids drill their preyto extract nutrients from the host’s feeding currents.They drill sharp-sided cylindrical holes and leavean attachment scar on their host’s shell (Matsukama,1978; Kabat, 1990). Capulid-host associations dateback to the Late Cretaceous, where capulids areknown to associate with inoceramid bivalves(Hayami and Kanie, 1980). Drilled inoceramids,however, are not reported for this assemblage.Capulid attachment scars and shell morphology thatconforms to their host are reported in modern andmiddle Pleistocene assemblages (Orr, 1962; Grant-Mackie and Chapman-Smith, 1971). Suchassociations should be highly reliable, and couldpotentially be found in the fossil record (Boucot,1990). Actual evidence of capulid drilling, however,is known only from one report from the latePleistocene of Japan (Matsukama, 1978). Thus, verylittle is known about this intriguing parasitic drillingbehavior in the fossil record.Cassid (Tonnoidea, Mesogastropoda) predatoryholes (not true drillholes, but rather rasped areas) inechinoderms date back to the Early Cretaceous, buthave been little studied despite their ubiquity inCenozoic and modern echinoids (Hughes andHughes, 1981; Nebelsick and Kowalewski, 1999).Cassids use sulfuric acid from their proboscis glandand the radula to cut out (rather than drill) anirregular hole in echinoderm tests (Kabat, 1990);however, most workers use the term “drillhole” fortheir predatory traces. Although the earliest drilledechinoderm dates back to the Early Cambrian, mostdrilling predation on echinoids is known only fromthe Cretaceous and Cenozoic (Sohl, 1969; Beu etal., 1972; Nebelsick and Kowalewski, 1999). Theearliest drillholes attributed to cassids weredescribed from the Early Cretaceous (Albian) ofTexas, but the cassid drilling record is much moreextensive in the Cenozoic, especially from theEocene to present (Hughes and Hughes, 1981;McClintock and Marion, 1993; Nebelsick andKowalewski, 1999).Muricids (Neogastropoda) diversified greatlyin the Paleogene, occurring primarily in tropicalto subtropical waters, although they are found intemperate and cooler regions as well (Vokes, 1971,1990; Vermeij, 1996; Vermeij and Carlson, 2000).During times of reduced food, muricids may drillconspecifics (Spanier, 1986, 1987). An increase insuch cannibalistic boring in muricids has beenassociated with sea level changes in the Red Sea(Spanier, 1987). Rarely, some muricids drill theirown opercula or bore into dead empty shells(Prezant, 1983). While increasing drillhole size canbe correlated with increasing size of muricidpredator for some species, this does not hold forothers (Urrutia and Navarro, 2001). Muricids mayalso change their drilling behavior and preferreddrilling location on the prey with ontogeny (Urrutiaand Navarro, 2001). Drillholes in inarticulatebrachiopods are rare but reported in Recentcommunities, and may be due to muricid predation(Paine, 1963; Kowalewski and Flessa, 1997).Similar drillholes in fossil inarticulate brachiopodsreported from the Tertiary of Seymour Island,Antarctica, and the eastern United States (Cooper,1988; Wiedman et al., 1988; Bitner, 1996) may beattributed to a muricid predator.The record of naticid predatory drillholes hasbeen used extensively to examine the evolution ofpredatory behavior, escalation hypotheses, andcost-benefit analyses in modern and fossil155

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002assemblages (Taylor et al., 1980; Kitchell et al.,1981; Kitchell, 1986; Kelley, 1988; Kabat, 1990;Anderson et al., 1991; Kelley and Hansen, 2001;see below). The naticid subfamily Polinicinaediversified greatly in the Cenozoic, and thepolinicid body fossil and predatory trace fossilrecord is extensive, especially after the Oligocene(Sohl, 1969; Taylor et al., 1983).The Nassariids (Neogastropods) are carnivoresor scavengers, and until recently, their predatorydrilling habits were in question (Kabat, 1990;Kowalewski, 1993). Kabat (1990), in fact,suggested that drilling did not occur in this group,although the possibility of nassariid drilling wasmentioned by Fischer (1963). Recently, Mortonand Chan (1997) have shown unequivocally thatsome nassarids can drill prey. A few (8 of 30individuals) laboratory-reared juveniles ofNassarius festivus were found with stereotypicallysitedboreholes on the ventral surface of their mainbody whorl (Morton and Chan, 1997, their fig. 1).The boreholes varied in morphology, fromelongate, irregular drillholes to sphericalcountersunk borings that were clearly rasped withthe radula and aided by chemical dissolution. It isthought that drilling may be a juvenile behaviorthat is lost in the adults, as no adult nassariids haveunequivocally been found to drill prey.To date, two species (Austroginella johnsoniand Austroginella muscaria) of marginellidgastropods from southeastern Australia are knownto drill into molluscan prey (Ponder and Taylor,1992). Parabolic in sectional shape and circular tosubcircular in outline, the studied drillholes rangein length from 1.13 mm to 3.1 mm. Additionally,marginellid drillholes are countersunk with a verysmall inner diameter relative to the outer diameter.This inner opening may have an irregular shapethat can be used to distinguish these borings fromthose of other predatory gastropods such asnaticids. Naticids make larger drillholes (seeKowalewski, 1993). However, marginelliddrillholes are similar to octopus drillings (Ponderand Taylor, 1992), and thus may be difficult todistinguish in the fossil record. Like octopods,marginellids may only use the drillhole for injectingtoxins to relax the prey, rather than feeding throughthe hole. Under SEM, the calcareous microstructureis seen to be greatly etched, suggesting a dominantsolutional mechanism for drilling.Cephalopods.—Shell-crushing and crustaceancrushingnautiloids diversified after theCretaceous-Tertiary extinction, and remainedrelatively abundant into the Miocene, whennautiloids were quite diverse and abundant incontinental shelf habitats across the globe (Ward,1987). The earliest Nautilus is known from theEocene–early Oligocene, but no fossils are knownfrom the upper Oligocene to Pleistocene (Teichertand Matsumoto, 1987). In modern seas, nautiloidsextend from Fiji in the east to the Indian Ocean inthe west, and from New Caledonia to Japan (Ward,1987). Natiloids forage for prey or crustacean moltsacross great depth ranges. There may be up to sevenspecies of nautiloids in modern oceans, but severalof the species designations are debated (Saundersand Ward, 1987; Ward, 1987).The prey of Nautilus is seemingly quite differentfrom that of the Mesozoic ammonoids.Unfortunately, the feeding ecology of Nautilus ispoorly known, but it is thought to be both a predatorand a scavenger (Ward, 1987). While directobservations of predation are lacking, evidence fromcrop dissections suggests that nautiloids eatcrustaceans, especially crabs (Saunders and Ward,1987; Ward, 1987; Nixon, 1988). The crop ofNautilus macromphalus, for example, has often beenfound to contain many hermit crabs of one species(Ward, 1987). However, this dietary finding maybe biased in that nearly all Nautilus studied arecaught in traps, which also attract crustaceans.Additionally, nautiloids have been directly observedby divers to eat molts from lobsters and slipperlobsters. With their large, chitinous jaws tipped withcalcium carbonate, nautiloids shred their prey orscavenged items into very small pieces of about 5mm 3 (Nixon, 1988). Predators of Nautilus includesharks, triggerfish, humans, octopods, and perhapsother nautiloids (Ward, 1987).Ammonites and nearly all belemnoids becameextinct during the Cretaceous-Tertiary crisis.However, other coleoid cephalopods, such as the156

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONcuttlefish Sepia, the squids (Loligo), and Nautilus,are common in the Cenozoic (Nixon, 1988). TheSepiida, in particular, diversified greatly in theCenozoic. Benthic sepiids and cuttlefish mainly feedon small crustacea, such as prawns (Nixon, 1988).Drillholes in molluscan prey from Recentoctopods are well known (Fugita, 1916; Pilson andTaylor, 1961; Wodinsky, 1969; Nixon, 1980;Bromley, 1981; Kowalewski, 1993). Octopuses usesecretions and abrasion from an accessory salivarypapilla in the drilling process (Nixon, 1979, 1980).Their drillholes are distinctly irregular or oval, witha very small inner borehole diameter (Kabat 1990).Despite their ubiquity in modern habitats, theirability to select particular prey, and their shellcollectinghabits, few of their borings (trace fossilsof Oichnus spp.) have been reported in the fossilrecord (Bromley, 1993; Harper, 2002). Robba andOstinelli (1975) first reported octopod drillingsfrom the Pliocene of Italy. Bromley (1993) reportsoctopus drillings from the Pliocene of Greece.Walker (1991, 2001) reports octopod drillings fromthe late Pleistocene of the Galápagos Islands, andfor the late Pliocene of Ecuador. Harper (2002)records octopus drillings from the Plio-Pleistoceneof Florida. Octopods may also drill nautiloids,many shells of which have multiple drillholes(Saunders et al., 1991). Octopods also make“kitchen middens” of their favored prey type whichcan be found outside their den; the shells arecommonly drilled (Walker, 1990).Echinoderms.—Evidence for the rise of asteroidpredation in the Cenozoic is reviewed by Vermeij(1987). Gastropods were found within the oral discof the sea star, Ctenophoraster in Eocene–Oligocene deposits from Antarctica (Blake andZinsmeister, 1979). This type of in situ predationhas a long fossil history dating back to thePaleozoic, but is rarely reported from Cenozoiclocalities. It is important to note that manyCenozoic predators that ingest their prey whole,such as sea stars, don’t leave an imprint on theirprey (see Vermeij, 1987). These predators are stillabundant in the Cenozoic, and some have evolvedto prey on reef corals (such as Acanthaster plancii).Regular echinoids have emerged as a majorpredatory group. Using the jaws of their Aristotle’slanterns, echinoids are able to graze corals and evennibble on the tests of distantly related clypeasteroid“sand dollars” (Kier, 1977). An unusual modernpredatory interaction between deep-water cidaroidsand crinoids was documented by Baumiller et al.(1999): the cidaroid, Calcocidaris micans, devoursthe stalked isocrinoid, Endoxocrinus parrae.Another cidaroid, Histocidaris nuttingi, alsocontained crinoids in its gut (Baumiller et al., 1999).Chondrichthyes.—Most sharks areopportunistic predators, with limited exceptionssuch as the planktivorous whale sharks (Cortés,1999). Intriguingly, both sharks and bony fisheshave evolved similar suites of prey capturestrategies, including suction, grasping, biting,gouging, and filter feeding (Motta et al., 2002).Inertial suction feeding is thought to be ancestralin bony fishes, while the ancestral condition ofsharks most likely involved grasping the prey anddismembering it with little upper jaw protrusion(Lauder, 1985; Motta et al., 2002). Some sharks,especially durophagous forms, use an inertialsuction prey capture method similar to the bonyfish. Suction feeding has arisen many times withinthe shark group, chiefly in relation feeding onbenthic prey (Motta et al., 2002). Specializationsfor suction feeding include rapid jaw opening, around terminal mouth, reduced dentition, and theability to produce large suction pressures (Mottaet al., 2002). Whale sharks (Rhincodon) possessthese features, but are planktivorous. Thus,durophagy may be an exaptation from a primarilyadaptive form of suction feeding in sharks.Few shark groups are known to have evolveddurophagous members, and thus durophagy isconsidered a rare form of feeding. Seven speciesof chimaerids (Holocephali), one species of hornshark (Heterodontidae), one species of nurse shark(Orectolobiformes), two species of the classicallypredatory Carcharhiniformes, and seven species ofrays (Rhinobatoidea, Rajoidea, Myliobatioidea) areknown to be durophagous. Stout, flattened teethand robust jaws are the hallmarks of durophagy.Durophagy in sharks, however, does notnecessarily mean that they eat molluscs; many157

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Figure 11158

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONdurophagous sharks eat crustaceans and/or fish(Wilga and Motta, 2000). For example, thechimaeroids have pavement tooth structures likethe myliobatids, and feed primarily on molluscsand crabs (Di Gaincomo and Perier, 1996). Hornsharks feed primarily on limpets, bivalves, and bluecrabs (Smith, 1942; Wilga and Motta, 2000). Thenurse shark, Ginglymostoma cirratum, eats fish andcrustaceans (Motta et al., 2002). The bonnethead(Carcharhiniformes) shark, Sphyrna tiburo, hasmolariform teeth and modified jaw structures thatfunction in crushing hard-shelled crustaceans, butit can also eat fish (Wilga and Motta, 2000).Another member of Carcharhiniformes, thesmoothhound or dogfish, Mustelus, has cuspedteeth and also feeds primarily on crustaceans(Russo, 1975; Yamaguchi and Taniuchi, 2000).Some skates (Rajoidea), guitarfish (Rhinoboidei),and rays (Myliobatoidea) have crushing andgrinding dentition for crunching crustacean ormolluscan prey but they can also feed on polychates(Gregory et al., 1979; Wilga and Motta, 2000).Neoselachian sharks became the top predatorsin the Cenozoic seas. During the Paleogene Periodthe successful galeomorphs (Galea) radiated intoseveral clades, including the dogfish and graysharks that feed on crustaceans and molluscs, thebasking and whale sharks that strain krill from seawater, and the white sharks that eat fish, seals, andcetaceans (Benton, 1997). Members of the lattergroup attained large size, culminating withCarcharodon megalodon (Fig. 11.2–11.4) in theMiocene and Pliocene, a 10–20-meter-long sharkwith teeth up to 17 cm long. This giant shark mayhave been specialized for feeding upon whales(Gottfried et al., 1996). Shark teeth were foundembedded in a whale jaw preserved in Pliocenesediments, suggesting a potential shark attack onthe whale (Demere and Cerutti, 1982).The radiation of deep-water Neoselachian-Squaliformes sharks, to which most of the modernforms belong (e.g., Somniosinae, Centrophorinae,most Etmopterinae, Oxynotinae), began in deepwaters with demersal forms originating after theCenomanian-Turonian anoxic event; the secondradiation of Squaliformes sharks (most of theDalatiidae) began in the Early Tertiary after theCretaceous mass extinction, and these epipelagicsharks radiated into shallow waters (Fig. 9; see alsoAdnet and Cappetta, 2001, their fig. 4, p. 241). Thedentition of most of these groups is quite varied,but most are heterodont. Most squaliformes dineon fish and cephalopods (Cortés, 1999).All major living families of durophagous rayswere established by the middle Eocene (Vermeij,1987 after Maisey, 1982), but their effect on theresultant fossil record of molluscs and other preyis not known. It is clear that despite their pavementtype dentition, rays eat a wide variety of food thatis not necessarily hard-shelled prey. Some preyitems ingested, however, may be incidental to theirforaging for larger prey items. Modern bat rays,such as Myliobatis californica, feed on bivalves,crustaceans, and polychaetes; bivalves are thedominant food item for most size classes, exceptfor the adults (Gray et al., 1997). Prey items within←FIGURE 11—Cenozoic marine vertebrate predators. 1, Teleost paracanthopterygian fish, Mcconichythes.2, Silhouettes of modern great white shark Carcharodon carcharias and Neogene C. megalodon. 3,Outline and skeleton of C. megalodon. 4, Carcharodon megalodon; reconstructed jaws of C. megalodon(perhaps overestimated). 5, Batoid sting ray Raja. 6, Reconstruction of flightless marine, wing-propelledswimming birds drawn at same scale in swimming posture: lower figure is modern Emperor penguin;upper shows reconstruction of extinct pelecaniforme plotopterid. 7, 8, Early whale with limbs,Ambulocetus in two postures; Middle Eocene. 9, Reconstruction of oldest known whale Pakicetus(Early Eocene). 10, Early large whale Basilosaurus; note tiny head with distinctive, multi-cusped teeth;Late Eocene. 11, Early (Oligocene–Miocene) pinniped Enalioarctos. 12, Desmatophocid sealAllodesmus, Miocene. Figures adapted from Benton (1997).159

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Myliobatis stomach contents vary ontogeneticallyand with the sex of the ray (Gray et al., 1997). Inthe study by Gray et al. (1997), juvenile bat raysfed on small clams (< 5 mm clam siphon diameter),benthic shrimp, and polychaetes. Adult rayspredominately fed on polychaetes, large clams(>5mm clam siphon diameter), and Cancer crabs.The largest rays preferred large clams and Cancercrabs. Large clams and crabs were eaten by adultfemale rays, whereas subadult and adult male raysfed primarily on polychaetes and burrowing shrimp.In other studies, large females also predominantlyfed on echiuran worms (Karl and Obrebski, 1976).Rays and skates excavate shallow pits in softsediments in search of prey (Fig. 12) (Gregory etal., 1979). The cownose ray (Rhinoptera bonasus)and possibly other batoids repeatedly inhalesediments and water through the mouth and ventit out the gill slits; the pectoral fins act to move theFIGURE 12—Benthic feeding by batoid rays.Upper figure, diagram of eagle ray Myliobatis jettingwater through gill slits to excavate circular feedingdepression. Lower figure, drawing of ray in feedingposition on excavated pit. From Gregory et al.(1979) and Howard et al. (1977).sediment away and to enlarge the burrow (Gray etal., 1997). Similar shallow pits are present inPleistocene localities, and are indicative of rayfeeding activities (Fig.12) (Howard et al., 1977).These pits could be correlated with associatedfragmented mollusc deposits, but, to date, this hasnot been examinedPods of deep-water gastropods attributed toeither fecal masses or regurgitated remains fromshell-eating sharks or other predators, weredescribed from bathyal Pliocene deposits fromEcuador (Hasson and Fischer, 1986, p. 35).However, a recent analysis of these shell “nests”revealed that they are not related to predation(Walker, 2001).Osteicthyes.—The diversification of teleosts(Fig. 11.1) in the Cenozoic is unprecedented amongvertebrates: presently some 23,670 species areassigned to 38 orders and 425 families (Patterson,1994). This is largely the result of development oftwo clades during the Cenozoic, the Ostariophysiin fresh water and the very successfulAcanthomorpha (over 21,000 extant species) in allenvironments (Maisey, 1996). Teleosts, rangingfrom tarpons to tunas, became the most commonpiscivorous open-water predators during theCenozoic. Certain fast-swimming large predatoryteleosts, such as swordfish, seemingly filled a partof the fast-swimming piscivorous predator guildheld by ichthyosaurs and primitive teleosts (e.g.Xiphactinus) during much of the Mesozoic.Ray-finned teleosts with molluscivorous habitsoriginated and diversified in the Eocene, a few othergroups in the Oligocene and Miocene (Vermeij,1987). Additionally, a major evolutionary radiationoccurred in the tropical reef fish fauna of the Eocene.Most of the fossil record of reef fish comes fromLate Cretaceous to Miocene Tethyan reef depositsof southern Europe (Rosen, 1988; Choate andBellwood, 1991). The best reef fish fossils, however,are from the Eocene of Monte Bolca, Italy (Blotte,1980; Choate and Bellwood, 1991; Bellwood, 1996).These fossils are excellently preserved—some retainpigmentation—and represent mass mortality events,probably related to poisonous algal blooms (Choateand Bellwood, 1991).160

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 13—Recent shells dredged from Ria de Arosa, Galicia, Spain, showing healed (2, 23, 24) andunhealed fracturing attributable to crustaceans. 1, 2, Fragmented bivalve Chlamys. 3–14, Fragmentsof bivalve Venus. 15–23, Gastropod Nassarius. 24–28, Turritella. From Cadée (1968).This extraordinary group of reef teleostsevolved rapidly, coinciding with the evolution ofthe coral taxa that dominate reefs today (Rosen,1988; Choate and Bellwood, 1991). Within a 20-million-year period, most teleost families that occurin modern reefs had appeared, with the exceptionof the durophagous Sparidae, which evolved in theMiocene (Choate and Bellwood, 1991). Thus, withthe evolution of the scleractinian coral species inthe Eocene (Acropora, Porites, and Pocillopora),the reef fishes evolved as well. Since that time,reef fish morphology has remained relatively stable161

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002through the duration of the Cenozoic.Today, predatory and grazing reef fish in theIndo-Pacific alone comprise over 4000 species offish, representing 18% of all living fishes (Choateand Bellwood, 1991). The triggerfish(Tetraodontidae, Balistidae) are known to crushprey in their jaws. Triggerfish durophagousspecializations include the loss of jaw protrusion,enlarged jaw adductor muscles, and stout teeth.Sparidae (parrot fish) also crush corals and otherhard prey.Fish other than tropical reef fish can fragmentshell material (Cadée, 1968; Cate and Evans, 1994;Norton, 1995); although they may be responsiblefor a majority of fragmented shelly remains in thefossil record, direct evidence linking them to thescene of the crime is lacking. Alternatively, fishthat puncture shelly hardparts are known (Norton,1988), and it would be possible to trace this specifictype of shell damage in the fossil record, althoughthis has not been attempted.Sea snakes.—Sea snakes evolved from varanidancestors (as did mosasaurs) in the mid-to-LateCretaceous, and diversified greatly in the Cenozoic(Caldwell and Lee, 1997; Lee and Caldwell, 2000).From the Late Cretaceous to Eocene, there wereseveral genera of marine snakes representative ofthe booidean family Palaeophidae (Hecht et al.,1974; Heatwole, 1987). Early Tertiary fossils ofsea snakes are very abundant and globallywidespread. However, the Palaeophidae are not thedirect ancestors of modern sea snakes; rather, theFamily Elapideae (terrestrial, venomous snakes ofthe cobra family) is thought to have given rise tothe extant sea snake fauna between 35 and 25million years ago, in the Oligocene to Miocene.However, the modern genera are not well knownas fossils (Heatwole, 1987). The modern fauna ofsea snakes, Laticaudinae and Hydrophiinae, arecomprised of 12 genera and approximately 48species (Hecht et al., 1974) distributed chiefly insubtropical to tropical oceans. A few saltmarsh andestuarine snakes also occur in temperate NorthAmerica. Most sea snakes are nearshore creatures(within the 100 m isobath; Hecht et al., 1974).Although they not particularly well studied, it isthought that most sea snakes are piscivorous, with afew species that are “generalists”—that is, that feedon both fish and invertebrates such as crustaceansand molluscs (McCosker, 1975; Glodek and Voris,1982; Voris and Voris, 1983; Heatwole, 1987).Saltmarsh snakes (natricines) eat small fish andfiddler crabs; the granulated file snake (Acrochordusgranulatus) eats fish, crustaceans, and snails(Heatwole, 1987). Sea snakes swallow their preywhole, but it would be very useful to know what thetaphonomic quality of the invertebrates are once theypass through the gut of the sea snake. That is, whatsize fragments? Is there any indication of gastricacids on the fragments? How much do they eat ofvarying prey items?Sea turtles.—Cenozoic fossil turtles are knownfrom a number of localities dating from thePaleocene (Weems, 1988). The Chelonoidea firstappeared in the late Early Cretaceous and theCenozoic fauna includes the survivors of theCretaceous-Tertiary mass extinction: Dermochelyidaeand Cheloniidea (Hirayama, 1997). Ofthese, the Dermochelyidae, with their thin shellsand fontanellization, are poorly preserved;chelonids are better preserved, and are slightly betterknown (Weems, 1988). Despite this taphonomicproblem, no catastrophic terminal Cretaceous eventis evident in the record of sea turtles (Weems, 1988).Sea turtles had declined in diversity by the lateCampanian, and were low in diversity during theMaastrichtian and Danian, but recovered in theThanetian and Ypresian stages of the early Cenozoic(Weems, 1988). The cheloniids underwent a majordiversification in the late Paleocene (Weems, 1988).This pattern of diversity matches the global patternof oceanic cooling and warming in Late Cretaceousto early Tertiary time (Weems, 1988, p. 143, hisfig. 27). The later Tertiary sea turtles are too poorlyknown to allow us to extrapolate diversity at thistime, but in general, diversity declined from theEocene to the five cosmopolitan species remainigtoday (Weems, 1988).Although modern turtles are morphologicallysimilar, their feeding preferences differ: theCheloniidae are omnivorous, and herbivorous(Hirayama, 1997) adult green turtles (Chelonia162

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONmydas) eat sea grasses and algae; the hawksbills(Eretmochelys imbricata) eat sponges, and Kemp’sridley (Lepidochelys kempii) eats crustaceans. TheDermochelyidae are rather fond of jellyfish.Caretta has massive jaws with a trituratingsurface; this genus is most common in temperatewaters, and the present distribution of this sea turtledates back to the early Pliocene (Dodd and Morgan,1992; Parris et al., 2000). Caretta eats sea pens,crustaceans, and molluscs among many other preyitems (Hendrickson, 1980; Plotkin et al., 1993;Nicholls, 1997). One species, Caretta caretta, theloggerhead turtle, is known to eat a variety of prey(Mortimer, 1982; Bjorndal, 1985). Caretta off theTexas shelf feed primarily on sea pens in the spring,and benthic crabs in the summer and fall (Plotkinet al., 1993). They can also eat molluscs,anthropogenic debris (e.g., fishing line, plastic trashbags), Diopatra tube worms, barnacles, fish,seaweed, whip coral, sea pansies, sea anemones,stomatopods, shrimp, and jellyfish (Plotkin et al.,1993). The small bivalve molluscs present in theirstomachs may come from the digested tubes ofDiopatra, and perhaps may not have been directlyfed upon. Scavenging gastropods that feed on deadfish or crabs, such as Nassarius acutus, may havebeen eaten accidently as the turtle went afterdecaying fish (Plotkin et al., 1993).Caretta populations from different geographicareas feed on different types of prey. In the westernMediterranean, Caretta caretta eat fish and tunicatesalps, although they can also eat benthiccrustaceans and molluscs (Tomas et al., 2001).However, the fish may be from scavenged by-catchthat is thrown overboard by fisherman. Caretta canalso forage on jellyfish at the ocean surface (Plotkinet al., 1993). The variety of prey that theseloggerheads eat is impressive, denoting a generalist(Plotkin et al., 1993). Of all the sea turtles that existtoday, Caretta is the only generalist.Sea and Shore Birds.—The extinction ofpterosaurs and early toothed diving birds in the LateCretaceous left open another important niche formarine piscivorous predators. It seems that thisvoid was filled rapidly by the evolution ofneognathan sea birds (Fig. 11.6). Aquatic birds areincluded among the oldest fossils of the DivisionNeognathae, with Late Cretaceous records for thetransitional shore birds (Feduccia, 1995). Severalorders of marine birds have fossil records extendingat least to the early Paleogene; these includeAnseriformes (ducks), Gaviiformes (loons), andCharadriformes (shore birds).Foot-propelled loons (O. Gaviiformes; LateCretaceous(?) to Recent) and grebes (O.Podicepiformes; Miocene to Recent) appear highlyconvergent on the Cretaceous ichthyornithines andhesperornithines, but are not closely related(Chiappe, 1995). Contrastingly, gliding albatrosses(O. Procellariformes; Eocene to Recent), some withwingspans exceeding 3.5 m, gulls (O.Charadriformes; Eocene to Recent), and pelicansand cormorants (O. Pelicaniformes; Eocene toRecent) seemingly fill a guild similar to theMesozoic sea-going, piscivorous pterosaurs.Finally, the penguins (O. Sphenisciformes; Eoceneto Recent) (Fig. 11.6), including some 25 genera,have an excellent fossil record, primarily in theSouthern Hemisphere (Simpson, 1975); diversefossils are especially common in the Eocene toMiocene of New Zealand and Seymour Island.These amphibious birds have become specializedfor rapid underwater flight even as they have lostaerial flying ability. Interestingly, an extinct cladeof flightless pelicaniform birds, the Plotopteridae(Eocene to Miocene), convergently evolvedelongate paddle-like wings for underwater flight.Some of the Pacific plotopterids attained lengthsof 2 m (Olson and Hasegawa, 1979). All of thesebirds are primarily piscivores and their abundanceattests to the proliferation of small teleost fish innear-surface seawater.The diversification of diving and other coastalmarine birds also may have greatly impacted thefossil record of crustaceans and molluscs (Vermeij,1977, 1987). Although several groups originated inthe Late Cretaceous, the diversification of shorebirds and diving marine birds took place chiefly inthe Paleogene (Vermeij, 1987). Oyster catchers,however, originated in the Neogene (Olson andSteadman, 1978). Diving marine birds catch fishthat had previously preyed on molluscs; the163

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002molluscs are then transported to nests often farfrom their original habitat (Teichert and Serventy,1947; Smith, 1952; Lindberg and Carlton, 1969;Lindberg and Kellogg, 1982). Coastal birds suchas the oyster catchers and eiderducks can preydirectly on large numbers of molluscs and otherinvertebrates (Schäfer, 1972; Cadée, 1989,1995).Oyster catchers penetrate molluscan prey bystabbing slightly gaping valves with their beaks,resulting in broken mollusc shells (Drinnan, 1957;Carter, 1968; Cadée, 1995). Oyster catcherpredation often produces distinctive shell damage,with one valve fragmented and the otheruntouched (Drinnan, 1957). However, at least halfof the prey shells may be left intact as these birdsalso insert their beak between the valves,supposedly without harming the shells.The diet of herring gulls may consist of up to70% marine molluscs. Prey are directly scoopedup by the gull from the marine environment andtransported to the shore where the shells aredropped over tidal flats or other hard surfaces,which fragments the shells so that the meat can beextracted (Cadée, 1995). Often, birds selectparticular sizes of prey, which can affect theresultant fossil record of coastal molluscanassemblages (Cadée, 1989). Cadée (1995) hasestimated for the Dutch Wadden Sea that shellcrushingshore birds may fragment up to 35% ofthe annual shell carbonate production. Because theDutch Wadden Sea benthos has up to 75% shellfragments, many of these fragments may be ofbiological rather than physical origin (Cadée,1995). Thus, in shallow Cenozoic seas, coastal andsea birds would have been important agents of shellfragmentaion (Cadée, 1995). These fragments mustoccur in the fossil record (Trewin and Welsh, 1972),but it is generally impossible to pinpoint exactlywho fragmented the shells.Shore birds may also leave benthic feedingtraces in soft sediment. For example, gulls such asLarus ridibundus may make troughs up to 3 m long,15 cm wide, and 3 cm deep in soft shore sedimentsas they forage for food (Cadée, 1990). Shelducksmake smaller pits (60 cm in diameter and 10 cmdeep). These feeding traces are similar to thoseproduced by foraging rays and flatfish (Cadée,1990). However, rays excavate sand around thecircumference of the foraging pits, whereasshelducks excavate to only one side (Cadée, 1990).These distinguishing characteristics could beobliterated by the tides, and therefore it may bedifficult to distinguish bird foraging pits from thoseof benthic fish predators. Foraging pits of aquaticbirds are known only from the Holocene.Pinniped Mammals.—Pinnipeds—seals, sealions, and walruses (Fig.11.11)—in modern seashave a global distribution, occur in enormousnumbers in some regions of the world, and are ableto dive to great depths in the ocean in search offood (Table 4). Therefore, some forms must havehad a significant impact on hardshelledcommunities, although many eat fish. Some of thepinnipeds evolved molar crowns with hypermineralizedcutting edges for crushing and piercingthe hard exoskeletons of crustaceans and molluscs(Haley, 1986). Pinnipeds evolved in the Eocene, andthus have had over 40 million years to affect theevolutionary history of molluscs and arthropods;however, there is no indication that they did so.Walruses (Family Odobenidae) feed mainly onbenthic invertebrates, and have a peculiar feedingstyle: they suck out the siphon or foot of bivalvesusing their piston-like tongue while their mouthworks as a vacuum pump (Muizon, 1993).Walruses have large and deep palates, a wide, bluntsnout with strong muscular insertions, and areduction of maxillary dentition (Muizon, 1993).The tusks are thought to have a primarily social,rather than foraging function. Walruses also leavelong, narrow feeding tracks or small excavated pitsthat can be seen in side-scan sonar (Oliver et al.,1983; Nerini, 1984, her fig. 3).All known odobenine odobenids (walruses) arebottom feeders and are first known from the Miocene(Repenning, 1976). Several extinct species from thePliocene are known to have been molluscivorous(Repenning, 1976) and were widespread at that timein the northern hemisphere (Muizon, 1993).However, the modern walrus (Odobenus) has a fossilrecord only from the Pleistocene (Repenning, 1976).There is no direct fossil record of pinniped predatory164

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONbehavior, with the possible exception of putativecoprolites packed with crab parts, attributed to seals(Glaessner, 1960; Boucot, 1990).Sea otters.—Sea otters (Table 4) evolvedduring the cooling period of the late Miocene, andare restricted to temperate regions (Van Blaricomand Estes, 1988). Otters occur in shallow coastalareas, where they eat a variety of invertebrate prey(Haley, 1986; Van Blaricom and Estes, 1988). Insoft-sediment habitats, they are known to prey uponendobenthic bivalves (Kvitek et al., 1992, 1993).Sea otters ingest copious amounts of echinodermand molluscan prey—taking in up to 35% of theirbody mass in invertebrate prey every day (Hinesand Pearse, 1982)—but their predatory effects inthe fossil record remain unknown. Hines andPearse (1982) used the size, structure, and breakagecharacteristics of empty abalone shells to documentthe selectivity of the predator and the source ofabalone mortality in a rocky subtidal habitat off ofcentral California. Gormand sea otters preferabalones in California, and can consume about tenabalones a day (Costa, 1978). Similarly, crackedshells were used to infer otter predation on bivalveprey in southeastern Alaska (Kvitek et al., 1992).In this area, sea otters substantially impacted thepopulation of endobenthic bivalves and epibenthicurchins (Kvitek et al., 1992, 1993). Additionally,foraging pits dug by otters attracted predatory seastars, which then ate any exposed molluscs. Ottersdigging for clams also exhumed buried shells(Kvitek et al., 1992), suggesting that biologicalremainie is common in these areas. The reworkedshells then become settling sites for epibenthicinvertebrates. Curiously, sea otter populations maybe controlled by paralytic shellfish poisoning inthese areas (Kvitek et al., 1993).The predatory record of these creatures shouldbe discernible because sea otters have peculiarcarnassial teeth that are flat and rounded forcrushing prey, and their lower incisors are used toscoop meat out of shells. One thing is certain,however: where sea otters occur, their effects onpopulations of their favored food items should begreat. Sea otter predation on sea urchins has aconsiderable effect on nearshore communitystructure (Estes and Palmisano, 1974; Estes et al.,1982). It is also known that where humans havepreyed on sea otters for their furry pelts, theresultant fossil record is skewed towardherbivorous limpets and sea urchins; where seaotters are not preyed upon, the stratigraphic recordshows abundant kelp beds and fish populations(Simenstad et al., 1978).Cetaceans.—Cetaceans (whales and dolphins)originated from land-dwelling artiodactyls in theearly to middle Eocene (Gingerich et al., 2001;Thewissen et al., 2001). Early forms such asAmbulocetus (Fig. 11.7–11.9) probably wereamphibious and may have behaved like seals(Thewissen et al., 2001). It is not known what theseancient toothed whales fed on. By the late Eocenethe gigantic (20 m) and fully marine Basilosaurusseems to have occupied the guild of large Mesozoicmarine reptiles, such as mosasaurs (Fig. 11.10). Itssharp, multiply cusped, undifferentiated teeth wereapparently adapted to fish capture, although arelatively small head limited prey size (Benton,1997). The toothed whales (Suborder Odontoceti)diverged in the Oligocene and radiated during theMiocene into a large number of smaller, dolphinlikelineages (Barnes, 1984). These whales evolvedhighly sensitive echolocation and fast-swimmingbehaviors. They are well adapted for chasing downand capturing fish, sharks, and, in some cases, otherwhales. Apparently, these odontocetes re-evolvedmany of the adaptations of Mesozoic pursuitpredators, specifically the ichthyosaurs (seeMassare, 1987, 1997). The largest toothed whales,sperm whales, are of uncertain origin, but molecularstudies of Milinkovitch (1995) suggest that they mayactually have been derived, in the Oligocene Epoch,from the baleen whales rather than the odontocetewhales. Sperm whales are well adapted for deepdiving in pursuit of squid prey and perhaps occupythe guild of some Cretaceous mosasaurs.An unprecedented find of a walrus-like whaleskull from the Pliocene of Peru indicates that onerare form of whale may have been durophagouson molluscs and/or crustaceans (Muizon, 1993).Odobenocetops peruvianus did not have anelongated rostrum, but had large ventrally directed165

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002premaxillary tusks, a deep-vaulted palette withoutteeth, and strong muscle scars on the premaxillae,which indicate durophagy (Muizon, 1993). Itsmorphology is similar to the Beluga and narwhalwhales (Monodontidae).The Suborder Mysticeti (baleen whales)originated in the Oligocene (Whitmore and Sanders,1976) and developed sheets of horn- or hair-likebaleen for sieving water to collect pelagic organisms,especially krill—a form of predation previouslyevolved by certain bony fish (e.g., Mesozoicpachycormids, and whale sharks) and perhaps by aTriassic marine reptile, the placodont Henodus. Allof these organisms attained large size, and mysticetewhales include the largest known organisms.While a number of cetaceans may eat somebenthic fauna, it is only the gray whale (Mysteceti,Eschrichtiidae, Eschrichtius robustus) thatconsistently raids the benthos in search ofinvertebrates (e.g., tubiculous amphipods andcallianassid shrimp) to complement its fare ofpelagic prey such as squids, mysid shrimp, and fish(Norris et al., 1983; Nerini, 1984). Gray whales arealso known to skim eelgrass mats for bothcrustaceans and sea grass/algae, and sandy muddyhabitats for gastropods, bivalves, and tube-buildingpolychaetes (e.g., Diopatra and Onuphis; Nerini,1984, her table 2). Buccinids, neptunids, thaids, andnaticids are just a few of the gastropods that havebeen found among gray whale stomach contents;Macoma, Mya, and Mytilus are some of the ingestedbivalves. The gray whale is able to sieve sedimentsthrough its thick baleen plates, which have coarserhairs than other baleen whales (Nemoto, 1970).Gray whales leave very large feeding pits inshallow, nearshore to intertidal mudflats that areoften the only record of their feeding behavior(Nerini, 1984, her fig. 1). On one benthic foragingdive, it is possible for one whale to make a series ofshallow pits that are usually arrayed in a slight curveand range from 1 to 3 m long and from 0.5 to 1.5 mwide. Gray whales are known to commonly feedin Baja California lagoons, along their migratoryrange from the Bering Sea to Baja California (a6000-km range), and in the northern Bering,Chukchi, and Beaufort Seas (Nerini, 1984). An entirepopulation of gray whales (estimated in 1984 at15,500 whales) could turn over 3,565 km 2 /yr of seabottom while feeding, considerably impacting thebenthic communities where they feed (Nerini, 1984).Gray whale fossils, however, are only knownfrom the late Pleistocene, although several closelyrelated groups are known from the Miocene of NorthAmerica (Barnes and McLeod, 1984). The obligatebarnacle parasite of gray whales, Cryptolepas, is alsoonly known from the late Pleistocene (Barnes andMcLeod, 1984). It is known that there were twoallopatric populations of the gray whale in the earlyHolocene, one in the North Pacific and one in theNorth Atlantic, which is now extinct.Order Sirenia (sea cows).—Sea cows date fromthe Eocene, and are a very small group of mammalsthat feed chiefly on sea grasses, algae, or waterhyacinths (Domning, 1976; Savage, 1976). Oneparticular fossil Sirenian, however, may have fedon benthic molluscs. Miosiren from the lateMiocene of Belgium displays thickened toothenamel and cusp modifications, which indicate thatit may have fed on molluscs (Savage, 1976).Other mammals that forage for marineinvertebrates.—Raccoons (Procyon) forage forcrustaceans in temperate to subtropical tidepoolsand salt marshes (Ricketts et al., 1985; Walker, pers.obs., 1997). The first known Procyon is from theupper Pliocene; there are several Pleistocene fossilspecies as well (Arata and Hutchison, 1964).Fossils of Procyon are known from all over thecontinental United States, as well as Baja Californiaand Canada (Arata and Hutchison, 1964). Coyotesand other mammals also can feed in the intertidalzone of temperate regions (Ricketts et al., 1985).Rats, in particular, can prey on over 40 differenttypes of intertidal organisms, especially key holelimpets, porcellanid crabs, and cancrid crabs(Navarrete and Castilla, 1993).Humans.—Lastly, the origination of humansin the late Pleistocene added to the potential forcoastal foraging and selection of particularinvertebrate food items as evidenced by abundantkitchen midden sites around the world, as well astools embedded in late Pleistocene coral reefs (seeWalter et al., 2000). Humans have been using sea166

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONcreatures for food and ornamentation for manythousands of years based on archaeological shellmiddens (e.g., Speed, 1969; Avery and Siegfried,1980; Jerardino et al., 1992). For example, in Chile,the rocky coast has been exploited by humans forfood for at least 8,500 years (Moreno, 2001). Thisforaging was tied closely to settlement of thePacific region of South America, and has onlyrecently been recognized as a force that affects theresultant ecological community structure of an area(Moreno et al., 1984). Ecological shifts in seafoodbiota directly or indirectly caused by humans areknown from the present day (Castilla and Duran,1985; Castilla, 1999) and from the stratigraphicrecord (Simenstad et al., 1978; Kirch, 1983).EVOLUTIONARY VIGNETTES:SELECTED PATTERNSOF PREDATIONFROM THE CENOZOICThe Mesozoic Marine Revolution hypothesis(Vermeij, 1977, 1987) has been subjected to manytests, from several sources of evidence, chiefly todetermine: (1) if shell armor increases through time;(2) if shell predators increase through time; and (3)if lethal shell injuries increase through time (see alsoVermeij, 1983). In the following sections we reviewand critique some of the primary lines of argument.Most of the putative durophagous functional groupsre-evolved in the Cenozoic, and, one could argue,became more common during this time than in theMesozoic. However, some of this apparent increasemay represent biases such as the Raupian “pull ofthe Recent” and the better record of well-preservedfossils. It is also well known that aragoniticMesozoic invertebrates, especially molluscs, are notas well preserved as calcitic forms (except forammonoids in black shales); whereas in theCenozoic more aragonitic forms are preserved,giving us a more detailed picture of the potentialpredatory panorama. Shell repair, drilling, and otherfeatures can be distinguished on Cenozoic hardpartsmuch more easily than on older ones. This is not agloom-and-doom scenario, just a realistic one.Examples of prey in coprolites or regurgitatedremains, predation preserved in situ, and preyorganisms in stomach contents are rare inCenozoic deposits just as they are rare inMesozoic and Paleozoic assemblages (Häntzschelet al., 1968; Boucot, 1990; Brett, 1990). Whilethere is an extensive literature on coprolites, moststudies focus on terrestrial and vertebrate remains;few if any coprolites in marine environments canbe tied with reliability to a specific predator(Bishop, 1975; Boucot, 1990).Echinoderms.—In many localities, not least inthe Danish basin, the Cretaceous-Tertiary extinctiongreatly affected the invertebrate biota. However,several echinoderm groups do not appear to havebeen greatly affected by this extinction event, andshow an increase in diversity directly above theboundary in the Danish basin (Kjaer and Thomsen,1999). There are several examples of shallow-waterstalked crinoids from the early Cenozoic (Oji, 1996);and further movement offshore of isocrinid (stalked)crinoids occurred in the Miocene in the Caribbeanregion (Bottjer and Jablonski, 1988; Donovan,2001). Deeper-water crinoids have a relativelyconstant generic composition from the Miocene tothe Recent; the Plio-Pleistocene regional extinctionhad little effect on this group (Donovan, 2001).Shallow water areas remain populated by stem-lesscomatulid crinoids (Donovan, 2001). This suggeststhat mobility and cryptic habitats may have enabledthis group to survive in the face of high predation inshallow water.Arm autotomy is common in stalkless crinoids,but has not been well documented in stalkedcrinoids (Oji, 1986). The ability to autotomizecrinoid arms dates back at least to the Triassic (Ojiand Okamoto, 1994). It is thought that autotomyacts as a “lizard-tail” defense (after Baumiller etal., 1999): arms can be dropped quickly into themouths of predators, while the main body of thecrinoid is left to regenerate new arms. It is possiblethat isocrinids exploited this ability and that this iswhat allowed them to survive the putative increasedfish predation in the late Mesozoic (Oji andOkamoto, 1994). Modern crinoids from bathyaldepths have more regenerated arms than crinoidsfrom deeper depths (Oji, 1996).167

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Stalk shedding is also a common occurrencein isocrinoids, and may be a deterrent to predators.Baumiller et al. (1999) hypothesize that crinoidshave evolved various antipredatory strategies sincethe Devonian: a planktonic (e.g., Uintacrinus) orpseudoplanktonic (e.g., Seirocrinus) lifestyle, stalksheddingabilities (e.g., in isocrinids, comatulids),short-bursts of swimming (e.g., comatulids), andlife in cryptic habitats (e.g., comatulids).Kier (1977) plotted global diversity ofechinoids through the Cenozoic, and showedlimited diversity in the Paleocene and OligoceneEpochs, with peaks in echinoid diversity in theEocene and Miocene–Pliocene Epochs; the recordof regular echinoids was not as good as that ofirregular echinoids. Regular echinoids arecommonly fragmented, and their fragments usuallyare not studied by taxonomists (Greenstein, 1993)or are not collected (Oyen and Portell, 2001).Clypeasteroids evolved in the Paleocene anddiversified rapidly, aided by the evolutionaryinnovation of numerous small tube feet and spinefreebranching food grooves. Flattening of the testmeant that only the top fraction of the sedimentcould be sieved for food particles (Kier, 1982).Records of predation on Cenozoic echinoidsare rare, even though in modern seas predation onechinoids is well documented (Nebelsick, 1995,1999). Drilled echinoid tests are known from theEocene Upper Ocala Formation in North CentralFlorida (Gibson and Watson, 1989). Some of thesedrillholes were predatory; others were parasitic.Parasitic eulimids are known to drill the aboralsides of echinoids; commonly an echinoid displaysmultiple drillholes made by parasitic gastropods(Berry, 1956). Cassid drillings on irregularechinoids are known from the Eocene of theAtlantic Coastal Plain (Woodcock and Kelley,2001) and elsewhere (see Cassid review, thispaper). Sand dollars (Parascutella hobarthi) fromthe lower Miocene of the Austrian Molasse Zonedisplayed repaired scallop-shaped areas on theirtests resulting from predation, possibly by regularechinoids (Nebelsick, 1999). Lethal predation wasindicated by large round holes cutting through theechinoid test or by bite marks penetrating the oralsurface (Nebelsick, 1999). Fish bite marks onclypeasteroid echinoids are also reported from theupper Miocene of Argentina (Zinsmeister, 1980).In the modern Atlantic and Gulf region, thereare 95 asteroid species in 56 genera, with adepauperate (because of lack of work) record inthe Cenozoic Caribbean region dating to the earlyPaleocene (Donovan, 2001). Asteroids are knownfrom the Eocene to Pleistocene in Florida, and insome horizons their fragments are very abundant(Oyen and Portell, 2001). Amazing preservation ofcomplete specimens of Heliaster microbranchius isknown from the Pliocene of Florida (Oyen andPortell, 2001).Ophiuroids are one of the most diverse extantechinoderm groups in the Caribbean region, buthave a “poor” fossil record because of their easilydisarticulated skeletons and a lack of work on thesecreatures (Donovan, 2001; Oyen and Portell, 2001).Nonetheless, a number of dense stalked crinoidophiuroidassociations are known from before theJurassic; a near absence of these dense assemblagesafter the Jurassic was postulated to be due topredation pressure (Aronson, 1987, 1991).Intriguingly, however, the Tertiary La MesetaFormation, Antarctic Peninsula, contains localizeddense assemblages of autochthonous ophiuroidsand crinoids representing shallow-water facies(Aronson et al., 1997). The incidence of sublethalarm injuries was low in this assemblage, suggestingthat predation was rare; possibly in high latitudecool-water areas predation is suppressed.Molluscs.—Molluscs provide the mostimportant Cenozoic database for examiningevolutionary questions regarding the fossil recordof predation because they are globally widespread,very abundant, well preserved, and present in manydifferent facies. Therefore, most studies havefocused primarily on escalation in marine molluscs.Shell repair and shell drilling in molluscs haveprovided the database by which to examinePhanerozoic predatory trends. Shell repair data hasnot been applied with as much success as drillingpredation, most likely because shell repair can bea consequence of a variety of physical andbiological destructive factors. Shell repair may168

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONshow an increase in the Cretaceous and Cenozoic,or it may not, and a closer examination of shellrepair during this time is warranted (Table 3). Littlework has been done for Cenozoic localities toexamine shell repair with respect to habitat, species,and stratigraphic interval.Do lethal shell injuries (or shell repair)increase through late Mesozoic-Cenozoic time?—Traces of non-fatal peeling in molluscs are evidentas scars on the shell (Figs. 8.2, 13; Table 4) thatresult from repair of the outer shell lip by the mantleedge (Robba and Ostinelli, 1975; Raffaelli, 1978;Elner and Raffaelli, 1980; Vermeij et al., 1982;Vermeij, 1982; Allmon et al., 1990; Cadée et al.,1997). Frequency of shell repair is often cited inorder to compare temperate with tropical and deepseawith shallow-sea habitats, as well as to examinewithin- and between-habitat predation, and thetemporal dynamics of shell repair. It appears thatshell repair may increase through the Phanerozoic,with higher incidence of shell repair in theCenozoic—indicating that durophagous predatorsbecome more of a threat to molluscan prey (e.g.,Vermeij et al., 1980, 1982; Vermeij, 1983, 1987;Dietl et al., 2000). But analysis of the data on shellrepair (Table 3) illustrates that there are no realdifferences in shell repair frequency between theMesozoic and Cenozoic, despite the better recordof marine durophagous predators at this time.Shell repair must be interpreted with caution,as researchers use different methods andinterpretations in analyses of shell repair data. Twomethods are used to estimate shell repairfrequencies. First, shell repair frequencies can beestimated by dividing the number of shells withone major repair (jagged scar) by the total numberof shells in the sample (after Robba and Ostinelli,1975; Raffaelli, 1978; Elner and Raffaelli, 1980;Geller, 1983; Vale and Rex, 1988, 1989; Cadée etal., 1997; Walker, 2001). This is the moreconservative estimate for shell repair, as snails cansurvive injury more than once. If the snail is older,it may display more instances of shell repair. Second,shell repair frequency has also been calculated asthe total number of scars in all shells divided by thetotal number of shells in the sample (Table 3)(Vermeij et al., 1980, 1982; Vermeij, 1982). Thismethod does not take into account the fact that oldershells may have more shell repair than youngershells, and thus can result in an overestimate of shellrepair for an assemblage (although Vermeij hasrecognized this problem). Further, more instancesof shell repair than actual sample size are commonlyreported which makes the data difficult to interpret.Therefore, it is important to determine which methodis most useful in examining the fossil record of shellrepair and to be consistent with that method.Comparing papers that use different methods isdifficult and tenuous at best.Interpretations of shell repair must be carefullyevaluated especially in regard to equating frequencyof shell repair with intensity of predation (Cadée etal., 1997; Cadée, 1999). There are several factorsthat complicate the interpretation of shell repair.First, it is difficult to distinguish repair that mayhave been provoked by physical factors, such asburial or crushing between stones (e.g., Raffaelli,1978; Cadée, 1999). Self-inflicted wounds resultingfrom the process of predation that are thensubsequently repaired can also inflate estimates ofshell repair. For example, buccinid gastropods chiptheir outer lips in the process of preying on othermolluscs and then repair their self-inflicted breakage(Nielsen, 1975). Second, shell repair frequencies donot directly correlate with the intensity of predation,as a total absence of scars may mean either thatpredation did not occur or that predators were 100%efficient (Schoener, 1979). Third, the incidence ofrepair on a shell needs to be tied to the age of theorganism, as older snails may exhibit more shellrepair than younger ones. This may be especiallytrue for deep-sea snails that may exhibit slowergrowth rates and increased longevity with depth(Vale and Rex, 1988). Fourth, certain life historytraits (slow growing vs. fast growing, particularbehavior) and feeding mode may affect whether andwhen a shell is exposed to predation. Fifth, somespecies may be more prone to predation than othersin an assemblage (Hoffmeister and Kowalewski,2001; Kelley and Hansen, 2001; Walker, 2001); and,using the metric of only one species’ repairfrequencies can bias the results for an entire169

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002assemblage. Lastly, for a time-averaged assemblage,shell repair frequencies might be higher than whatwould be found in the living population at any onetime because of the patchiness of predation (andassociated physical factors).The consensus is, however, that conspicuousshell repair (i.e., conspicuously peeled shells withsubsequent repair) is most likely the result ofpredation. That is, only deeply peeled injuries thatare subsequently repaired can reliably be used inthe analysis of shell repair, whereas repaired nipsor edge chippings may not be indicative ofpredation (Walker and Voight, 1994; Walker, 2001).Consequently, although shell repair is not a goodindicator of predation intensity, it is instrumentalin providing a record of predators within a habitatwhen body fossils of the predators are missing.Shell Drilling through Time.—Shell drillingfrequency is less ambiguous in interpretation: acompleted drillhole signifies prey mortality. Also,particular borehole morphologies may beassociated with specific gastropod or octopodpredators (Carriker and Yochelson, 1968; Kabat,1990; Kowalewski, 1993; Kowalewski et al.,1998). Nonetheless, certain caveats also apply tothe study of drilling predation.Escalation studies of drilling predators andtheir prey have not generally taken into accountthe particular facies and associated biota ofanalyzed assemblages (with the exception ofHoffmeister and Kowalewski, 2001). Essentially,all assemblages are treated as if they were the samefacies (e.g., onshore and offshore assemblages aregrouped). Environmental differences betweenassemblages, however, can affect the morphologyof the taxa—some species are larger in nearshoreenvironments than they are in offshore environments(or vice versa). This gradient in morphology maynot be related to predation.Sedimentary facies could also havetaphonomic effects. For example, assemblagesdeposited above storm wave base may sort drilledand undrilled shells differently compared tooffshore assemblages. Drilled and undrilled shellscan be differentially transported in nearshoresettings and thus there may be a bias toward anoverabundance of drilled shells in some localities.Additionally, drilled shells are more prone totaphonomic breakage than undrilled shells, andsuch breakage may be more common in somelocalities than others (Roy et al., 1994). Left vs.right valves of bivalves and pedical vs. brachialvalves of brachiopods are also differentiallytransported and/or preserved (Brett and Allison,1998). Thus, it would be important to know thevalve frequencies of an assemblage, and whetherthey are biased. It would also be important to knowif drilling predators were actually found in the sameassemblage as the drillholes (e.g., Hansen andKelley, 1995), but not all papers that examinedrilling predation discuss this issue.It is also important to examine more than onelocality within a time period, as the record ofpredation is strongly controlled by habitat (Vermeijet al., 1981; Geller, 1983; Hansen and Kelley, 1995;Cadée et al., 1997; Hoffmeister and Kowalewski,2001). Location within a sequence may also affectthe density of drilled shells, as transgressive lagdeposits formed after a major sequence boundary(e.g., extinction?) commonly contain more bioticinformation as a result of longer time averaging(Brett, 1995; Holland, 2000). Therefore, one mustbe careful in interpreting the pattern and processof drilling through the Phanerozoic, as it is not assimple as merely counting drilled taxa per temporalstratigraphic sequence. As Boucot said, “Naturedoes not take place within an ecological vacuum”;nor should evolutionary interpretations using thefossil record be decoupled from facies studies.Given these caveats, based on an analysis ofover 150,000 gastropod and bivalve shells fromthe Gulf and Atlantic Coastal Plain (GACP), Kelleyand Hansen (2001) suggested that the interactionbetween naticid drilling predators and their preydoes not necessarily show escalation from theCretaceous to Oligocene. After examination of anumber of localities, they found that there is anepisodic pattern to drilling frequency, with massextinctions resetting the “arms race” for faunas.Drilling within the most of their Cretaceouslocalities was greater than several of their lateEocene localities and similar to early Oligocene170

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONlocalities (their table 8.1, p. 153). This is asignificant finding given that previously Vermiej(1987) had used naticid drilling as one line ofevidence illustrating escalation in Cretaceous toEocene faunas. In the Cretaceous Vermeij (1987)found limited drilling, but by the Eocene, drillinghad reached modern levels.Kelley and Hansen (1993), in contrast, did notfind an ever-increasing trend in naticid drillingfrequencies from the Cretaceous to Eocene forGACP molluscs. Escalation could also mean that apredator gets better at selecting prey; however,Kelley and Hansen (1993) did not find any temporaltrends toward increased drillhole site stereotypy innaticids. Molluscan prey were found to have moreincomplete drillholes and multiply drilled shells,indicating that prey effectiveness may haveescalated, but Kelley and Hansen’s (2001) data didnot show a trend for most of the periods examined.Kelley and Hansen (2001) also examineddifferences in morphology within molluscan generathat may be related to escalation. Although manygenera were examined, four particularly longrangingMiocene genera from the GACP—twogastropod predators (Euspira heros and Neveritaduplicata) and two frequently drilled naticid prey(Bicorbula idonea and Stewartia anodonta)—wereanalyzed for different morphological charateristics(their table 8.3, p. 159). In this case, however, thegastropod predators are also cannibalistic. Resultsshowed that shell size (height) did not change foreither E. heros, B. idonea, or S. anodonta (no dataare reported for N. duplicata), indicating that theseprey species found no size refuge from predationover time. Shell thickness (which would make a preyitem more difficult to drill) did not change for E.heros, decreased for N. duplicata, slightly increasedfor B. idonea and increased greatly for S. anodonta.Internal volume (an indicator of the amount of fooda predator can take in) did not change within theMiocene. Thus, it appears that most prey charactersdeemed to be directly related to predatory escalationdid not demonstrably change within the Miocene(except for shell thickness in S. anodonta). It wouldbe interesting to know whether drilling frequencyincreased or stayed relatively the same across thevarious assemblages examined.Escalated species are thought to be moresensitive to changes in primary productivity becausemaintaining heavy armor or high speeds to avoidpredators requires high metabolic rates and thus anuninterrupted food source (Vermeij, 1987).Therefore, Hansen et al. (1999) tested whetherpurported escalated species (those withantipredatory adaptations such as heavy armor) weremore vulnerable to extinctions caused by climatechange and associated environmental changes. Tenshell characters deemed important for predatorresistance were evaluated for GAPC molluscs acrossvarious mass extinction events associated withclimatic cooling and/or a decline in primaryproductivity (e.g., Cretaceous-Paleocene; Eocene-Oligocene; middle Miocene; Pliocene-Pleistocene).Importantly, all these assemblages were depositedin relatively shallow shelf environments withroughly similar grain sizes; all but one assemblagewas a bulk collection. Hansen et al. (1999) foundthat escalated species, overall, were not morevulnerable to climate-related mass extinction. Onlyornamented Pliocene gastropod species were moresusceptible to extinction than their weaklyornamented counterparts. In another study, Kelleyet al. (2001) found that recovery faunas after a massextinction event were not more vulnerable toenhanced drilling pressure, contrary to hypothesizedpredictions. Additionally, no overall trend inunsuccessful drilling was seen from the lateCretaceous to Pleistocene.Spatial trends in drilling predation vary byenvironment in fossil studies. Hansen and Kelley(1995) used 27,554 specimens of GACP molluscsfrom the Eocene and found a statistically significantdifference in drilling frequency between the innertomiddle- shelf Moodys Branch Formation andthe outer-shelf Yazoo Formation, the deeper sitehaving a higher frequency of drilling predation.However, for the five other assemblages examinedfrom the Moody’s Branch, there was no significantbathymetric trend. Drilling frequency was alsohighly correlated with the percentage of naticidsand their preferred prey within each assemblage.Hoffmeister and Kowalewski (2001) examined171

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002spatial and environmental variation in drillingpredation in the middle Miocene of Central Europe.The sampling methodology allowed forcomparisons within provinces, between provinces(Boreal vs. Paratethys), and between facies (finegrainedvs. coarse-grained siliciclastics). Theyfound that unsuccessful and multiple drillholesoccurred more frequently in the Boreal provincethan the Paratethys province; the same facies alsoincluded molluscs with different drillingfrequencies—with as much as a three-folddifference between samples collected in adjacentsites from the same facies! They concludedunequivocally that spatial variation should beevaluated independently before any large-scaletemporal trends are inferred for predation. Clearlymultiple collections with emphasis on facies needto be included in the temporal analysis of escalatorypredation hypotheses.Shell ornamentation (spines).—Spinedevelopment requires extra amounts of calciumcarbonate in seawater, and shallow tropical marinewaters meet this requirement (Nichol, 1965;Stanley, 1970). Spines have inspired variedhypotheses concerning their function asantipredatory architecture (Table 5). Few studies,however, have focused on alternative hypothesessuch as whether spines are phylogenetic legaciesof shell building (as in cardiid bivalves), non-aptiveconstructional artifacts, exaptations, or adaptations.Few workers have endeavored to apply suchphilosophical rigor, and many have created adaptivescenarios. Another important possible function ofspines could be to increase surface area for thesettlement of epibionts, and for trapping debris thatcamouflages the shells (Vance, 1978; Feifarek, 1987;Stone, 1998). Stone (1998) has shown that spineson epifaunal bivalves deter the attack of muricidshell-drilling predators, but muricids can still borein areas where spines are absent. In the same study,spines were found not to deter predatory attacks bysea stars that engulf their prey. The rise of spinoseornamentation in bivalves predates the radiation ofthe predatory Muricidae in the Albian, and actuallyextends back to the late Paleozoic in thesuperfamily Pectinoidea (Stone, 1998).Molluscan conchiolin layers: Are theyantipredatory?—Conchiolin is the organiccomponent of molluscan shells composed ofproteins, polysaccharids, and glycosaminoglycans(Table 6) (Wilber and Simkiss, 1968; Gregoire,1972; Wainwright et al., 1982). The periostracumand non-calcareous operculae are composedchiefly of conchiolin, while the nacreous layersand other shell microstructures contain variousquantities of conchiolin. Thus, conchiolin has a veryold history, putatively stemming from the oldestshelled mollusc in the Cambrian Period. What ispuzzling, however, is that only a few groups ofmolluscs—chiefly the freshwater bivalves (e.g.,Margaritiferidae, Unionidae, Mutelidae), estuarineto marine bivalves (Corbiculidae and Solenidae),and a few marine species—have conchiolinrepresented as separate sheets within their shells(Taylor et al., 1969; Anderson, 1992; Harper, 1994).Conchiolin, as a protein, is thought to form at a highmetabolic cost to the organism—and, perhapsbecause of this, there appears to be an evolutionarytendency to lose conchiolin layers (references inKardon, 1998). Thus, there must be someevolutionary reason for maintaining conchiolin inmolluscan shells despite its high metabolic cost ofproduction (Table 6). It has long been hypothesizedthat the conchiolin sheets deterred predation bydrilling molluscan predators, such as naticid (Lewyand Samtleben, 1979) or muricid gastropods (Taylor,1970,1981), and most of the work done to test thishypothesis has focused on the corbulid bivalves(Fischer, 1963; Lewy and Samtleben, 1979; DeCauwer, 1985; Anderson et al., 1991; Anderson,1992; Harper, 1994). The corbulids (FamilyCorbulidae) are small, inequivalved bivalves with aglobose shell form, a single byssus thread, andshallow burrowing habits (e.g., Stanley, 1970). Theyfirst appeared in the Middle Jurassic, with thegreatest diversification taking place in theCretaceous and Eocene (e.g., Hallam, 1976).There appear to be three contrasting temporal“trends” related to whether conchiolin reducespredation. The first is that conchiolin doeseffectively reduce predation on corbulids throughtheir evolutionary history. Fischer (1963), for172

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONTABLE 5—Hypotheses for the origin of skeletal spines in marine invertebrates.Hypothesis Evidence Examples ReferenceSpines develop in calciumcarbonate supersaturated seawaters; most common in tropicsLess energetically costly to makespines in tropical watersSpondylus americanus Stanley, 1970Spines are antipredatoryPrimary spines that projectoutward may protect mantle edgeSpondylus americanus Logan, 1974Hollow spines and keels are forpelagic/planktonic existencesEconomy of mass Ammonoids Birkelund,1981Spines have no function Constructional artifact? Alternative hypothesisfor any invertebrateSpines function in filter-feeding Spines cover opening to animal PoricthophenidbrachiopodsCarter, 1967;Kauffman,Rudwick, 1970Spines are an ancestralcondition; phylogeneticconstraintSpines form in various ways, evenwithin closely related familiesCardiid bivalves;anomalodesmatansSchneider andCarter, 2001Spines vary withenvironmental conditionsof the substrateVarious spine types depending onsubstrate the larvae attach toSpondylus americanus(Jurassic–Recent)Logan, 1974Attachment to substrateSpines discourage epibiontsSpines acts as supports forsensory mantle tissue; "mantleoutposts" to give early warningsignals of dangerSpines serve a camouflagicfunction, breaking updistinctive outline of shellSpines stabilize the shell on ashifting substrateSpines act as attachmentmechanismsSpines and pedicellaria in someechinoderms discourage biontsettlement; perhaps barbedsecondary spines of Spondylusamericanus reduce biontsettlementCemented bivalves likeSpondylus whose rightvalve is attached tosubstrate; thishypothesis does notfunction for the left valveSea urchins; the bivalveSpondylus americanus-- Brachiopods (JurassicAcanthothirus )Hair-like barbed spines typical ofthe neanic stage of the left valve ofSpondylus which get covered withalgae and sediment; spines arethickly encrusted with epibointsLate Paleozoicproductoid brachiopod,Waagenoconcha;Spondylus americanusThe CretaceousSpondylus spinosusLogan, 1974Logan, 1974Rudwick, 1965;Logan, 1974Grant, 1966;Logan, 1974Logan, 1974;Carter, 1972Spines keep the feedingmargins of the shell abovethe substrate-- -- Logan, 1974173

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 6—Evolutionary findings concerning whether conchiolin serves an anti-predatory function againstdrilling predators for corbulid bivalves. Alternative hypotheses are discussed in text.EvolutionaryconclusionAdaptationExaptation*EvidenceConchiolin arises at the same time asdrilling predators in the CretaceousConchiolin arises in the Middle Jurassic,well before origin of drilling naticids;*however, no temporal trend in drilling predationReferenceHarper (1994)Kardon (1998)Not Anti-predatory No temporal trend in drilling predation Anderson et al. (1991);Anderson (1992)PhylogeneticConstructional ArtifactConclusions drawn from synthesis ofthe literatureThis paperexample, suggested that Recent corbulid specieswere less likely to be completely drilled than fossilspecies. Kardon (1998, p. 73) also suggests that intemporally and spatially separated fossil samplesof corbulid bivalves, conchiolin layers are effectivedeterrents of naticid predation. The second is thatdrilling predation actually increased in corbulidsover their evolutionary history. For example,although using a limited data set, Taylor et al.(1983) suggested that corbulids showed enhancedpredation from the Late Cretaceous to Eocene. Andlastly, others suggest that there is no spatial ortemporal trend in drilling predation in corbulids.For example, in Late Cretaceous to Pliocenecorbulid fossils from Europe and North America,De Cauwer (1985) found no trend toward increasedcomplete drilling was found. Similarly, forMiocene to Pleistocene fossil corbulids from theDominican Republic and Florida, there appears tobe no spatial or temporal pattern in complete orincomplete drilling, strongly indicating thatconchiolin layers are not effective deterrents tonaticid predation (Anderson, 1992). Likewise,Harper (1994) reported that there was nosignificiant difference in drilling frequency amongalmost all geological samples examined from theCretaceous to Plio-Pleisotocene. Further, Harper(1994) found that there was no significantdifference in possession of conchiolin sheetsbetween temperate and tropical localities. Giventhese contrasting findings, it is important toexamine some of the salient evolutionaryhypotheses regarding conchiolin as anantipredatory deterrent, such as cost-benefitanalyses and whether conchiolin is an adaptationor exaptation (or neither) against predation.Corbulids have small size and effective valvearmor (i.e., relatively thick valves with conchiolinsheaths), and thus, according to Kitchell et al.’s(1981) cost-benefit model, would represent a highdrilling investment with low benefits (De Cauwer,1985). For example, Kelly (1988) found thatpredation on corbulids was lower than would bepredicted by the cost-benefit model (but seeAnderson, 1992). Yet corbulids are heavily drilledin many localities, and this may be a result of theirtendency to cluster, their shallow burrowing depths,and their sluggishness (De Cauwer, 1985). Perhapsdrilling predators may mistakenly drill empty shellsin the presence of chemical attractants in theexhalant water of the corbulid associations(Carriker, 1981; De Cauwer, 1985); or there maybe hydrodynamic and taphonomic reasons for thepreponderence of drilled corbulids in some174

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONlocalities (De Cauwer, 1985).Anderson et al. (1991) tested the cost-benefitmodel of Kitchell et al. (1981), and showed that acorbulid bivalve (Varicorbula caloosae) was nomore likely to be drilled by a naticid predator thanby a venerid bivalve (Chione cancellata), amongPleistocene fossils from Florida. Anderson (1992)examined many species of corbulids from theMiocene and Pliocene of the Dominican Republicand from the Pliocene and Pleistocene of Floridaand found that the incidence of drilled,incompletely drilled, and multiply drilled valveswas highly variable in space and time. This resultwas similar to other studies on drilled bivalves, andtherefore indicates that conchiolin was generallynot part of the antipredatory arsenal. Rather,alternative evolutionary hypotheses, such asconchiolin as a retardant of shell dissolution or adeterrent to crab-crushing predation (Anderson,1992; Kardon, 1998), for example, need to beadvanced and tested. The main evolutionaryquestion, as Harper (1994) pointed out(paraphrasing Gould and Vrba, 1982), is whetherconchiolin layers are a beneficial trait that isenhanced by natural selection (adaptation), orwhether conchiolin layers are an exaptation, abeneficial trait that is secondarily co-opted foranother function. Experimental testing is requiredto determine if a trait is truly beneficial; and thereshould also be a temporal correspondence betweenthe evolution of the trait and the proposed selectiveagent (Harper, 1994).Accordingly, Kardon (1998) tested threehypotheses concerning the evolutionary importanceof conchiolin: 1) it retards shell dissolution; 2) itincreases shell strength and thus deters crushingpredation; and 3) it inhibites shell drilling by naticidgastropods. She also examined the fossil record ofnaticid drilling predators, and compared it to thatof conchiolin-bearing corbulids (which hail fromthe Middle Jurassic; but see Harper, 1994) toexamine the evolution of the trait in associationwith its putative selective agent (the naticids). Herexperimental results show that conchiolin didretard shell dissolution, although—as she clearlypointed out—the majority of corbulids live incalcium carbonate–saturated regions, and havedone so for most of their geologic history.The most promising line of research concerningconchiolin, however, stems from the finding ofmechanical tests that the conchiolin in corbulids mayfunction to inhibit crack propagation, which in turnmay be a deterrent to shell-crushing predation(Kardon, 1998). It remains to be tested whetherconchiolin layers do inhibit shell-crushing predators.Biomechanical tests using corbulid bivalves, inaddition to feeding experiments with livedurophagous crustaceans, are needed to address thishypothesis. An historical analysis of shell repair incorbulids through time is warranted.Lastly, Kardon (1988) found that naticiddrilling rates were not significantly slowed byconchiolin layers. Further, although Kardon (1998)states that conchiolin has acted as an effectivedeterrent against drilling predation in the corbulidfossil record (p. 73, but see her p. 76), her data donot support this claim (p. 75, her table 2). Her results,in fact, support the findings of Anderson et al. (1991)and Anderson (1992) that there is temporal variationin drilling through time in corbulids, with noapparent trend. It would also be important to knowfrom which facies these corbulids came, and whethertaphonomic (hydrodynamic or biotic) conditionsaffected their preservation.Although Kardon (1998) suggests thatconchiolin is an exaptation, and Harper (1994)suggests that conchiolin is an adaptation, a reviewof the data to date indicates that conchiolin maybe an artifact of construction. Of course, thisstatement needs to be refuted by scientific tests.That is, without further tests with shell-crushingpredators, we cannot know if conchiolin is indeeda beneficial trait, either co-opted or evolved by theorganism against predation. Other hypotheses werediscussed by Harper (1994), such as protectionagainst nonpredatory borers or assistance withhermetic sealing, and these could be rigorouslytested as well (see Table 6). The oldest corbulids(Jurassic Corbulomima) were marine organisms,and had conchiolin before the evolution of naticiddrilling during the Early Cretaceous (Kardon,1998), further suggesting that conchiolin was not175

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002evolved as a deterrent to drilling predation.However, it would be very important to know theenvironmental conditions of the origin anddiversification of corbulids; also, for thosecorbulids with more than one conchiolin layer,whether they are from “physiologically” morestressful environments, such as brackish water oranoxic environments. It would also be importantto know their cladistic relationships with respectto their environment and conchiolin form.SUMMARY DISCUSSION:AN EPISODIC HISTORYOF PREDATIONPredation in marine communities evolvedthrough several phases of intensification withminor setbacks following mass extinctions(Fig.14). The Permo-Triassic extinction crisisformed a major setback for all marine communities.This certainly included many predatory taxa (e.g.,many ammonoids, nautiloids, phyllocarids,predatory archeogastropods). However, certainmarine predators, notably bony fishes and sharks,seem to have been less strongly affected by thismajor extinction than were many benthicinvertebrates (Knoll et al., 1996). Thus, predatorsseem to have rebounded rather rapidly and by theMiddle Triassic a variety of new predator guildshad appeared, including decapod crustaceans withcrushing claws, and shell-crushing sharks and bonyfish. However, data from the Triassic regardingshell repair and drilling predation are almost nonexistent.New groups of carnivorous marine reptilesalso appeared in the Triassic, includingdurophagous placodonts, and piscivorous andperhaps cephalopod-eating pachypleurosaurs,nothosaurs, ichthyosaurs, and the first plesiosaurs.Ceratite ammonoids and some marine reptiles(e.g., placodonts, nothosaurs) became extinctduring Late Triassic crises. However, other lineages(e.g., ammonites, ichthyosaurs, plesiosaurs)survived to form the stem groups for new Jurassicradiations. The Jurassic to Early Cretaceous sawthe rise of malacostracan crustaceans with crushingchelae and predatory vertebrates—in particular, themarine crocodilians, ichthyosaurs, and plesiosaurs.Following a setback in the Late Triassic, predatorsmade a major re-advance in the mid-Mesozoic withthe evolution of new groups of decapods,ammonites, neogastropods, and teleost fishes, aswell as neoselachian sharks and marine reptiles.Some of these groups are thought to have beendurophagous, but that does not mean they ateexclusively molluscan prey. Limited data from thistime indicates that drilling predation existed, butoccurred at low very frequencies.The Late Cretaceous saw unprecedented levelsof diversity of marine predaceous vertebratesincluding pliosaurs, plesiosaurs, and mosasaurs.The great Cretaceous-Tertiary extinction decimatedmarine reptiles. Drilling and shell peelingfrequencies pick up in the Late Cretaceouscorresponding to the evolution of new durophagousand shell-drilling groups. The drilling frequenciesfrom this time are no different from those reportedfrom Cenozoic localities; indeed, drilling and shellrepair data from the later Cretaceous and Cenozoicshow no apparent trends.The Cretaceous-Tertiary mass extinctioneliminated all large marine predators, including themosasaurs, plesiosaurs, and many sharks and fish.Additionally, pterosaurs and early marine birdswere eliminated. However, many benthicinvertebrate and fish predatory groups survived;and during the Paleogene, predatory benthicinvertebrates showed a spurt of evolution withneogastropoda and new groups of decapods, whilethe teleosts and neoselachian sharks both underwentparallel rapid evolutionary radiations; these werejoined by new predatory guilds of sea birds andmarine mamals. Ultimately, many of the largevertebrate predator guilds were refilled by newlyevolved groups of marine mammals (cetaceans,pinnipeds) and birds (gulls, albatrosses, penguins).Despite the fact that a new suite of predatorsevolved in the Cenozoic, there are no apparentescalatory trends in durophagous predation.All of this would seem to suggest episodic, butgenerally increasing predation pressure on marineorganisms through the Mesozoic–Cenozoicinterval. Theoretically, there should have been a176

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONFIGURE 14—Summary diagram, showing phases of escalation in marine predator-prey systems andmajor extinction events.177

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Paleogene marine revolution in the molluscanrealm, because of the increased abundance ofdrilling neogastropods, the first real records ofdurophagous stomatopods and decapod crustaceans,and the evolution of specialized bird and mammalianpredators. However, most of these durophagousgroups are generalists, and it may be that they hada diffuse effect on their invertebrate prey.Finally, several new groups of carnivorousmarine mammals and birds originated in theMiocene. Walruses, gray whales, and humans arosein the Neogene and affected the coastal hard-shelledbiota in the areas where they foraged or settled. Thus,a Neogene phase of further predator intensificationis also suggested. However, there is no directevidence that prey were selectively affected (exceptfor the widespread decimation of species by humansand their alteration of marine habitats).The Cenozoic record seems to provide anexcellent window into predation and its effects, butfew have examined the temporal trends in predationduring this time (except for naticid molluscandrillers; e.g., Hansen et al., 1999; Kelley et al., 2001).Given that many predators leave their signature onshells and other prey, it is just a matter of reexaminingthe fossil record with the specific intentto look for predation. More work needs to be donein this area, especially on drilling records fromother gastropod groups, and on putative shell repairrecords that allow a comparison of Paleogene withNeogene localities. Additionally, in the Cenozoic,vast deep-sea (bathyal and deeper) fossil depositsof molluscs are well preserved in uplifted terracesin tectonically active regions of the world, allowingfor comparisons of predation (shell repair, shelldrilling) between shallow benthic and deep seafossil assemblages (Walker and Voight, 1994;Walker et al., 2002).Although escalation is sometimes cast as anongoing “arms race,” in actuality the predatoryrecord shows episodes of abrupt bioticreorganization during and after mass extinctions,punctuating longer interludes of relative stability(Brett et al., 1996). Some clades may retain thehistorical legacy of the Paleozoic predatoryrevolutions, as could be argued for the stalkedcrinoids in modern oceans; other clades maycontinuously evolve new predators, as Vermeij(1987) has argued based on the gastropod fossilrecord. Schneider and Carter (2001) show thatcardiid spine forms in Mesozoic and Cenozoicgroups appear to be a Paleozoic ancestral condition,and appear not to be related to the putativeMesozoic Marine Revolution. Thus, a clade-bycladeanalysis of predation would be most useful,as the different groups each have their ownevolutionary histories and ecological constraints.This review shows that not all morphology inbenthic organisms need be directly related topredation. Additionally, most durophagous predatorsdo not prey specifically on molluscs. They also preyon hard-shelled crustaceans, a major group oforganisms deemed to have caused selective pressuretoward escalated armor in gastropods (Vermeij,1987). We also must strive to examine predation inassemblages spatially across different environments,mindful of taphonomic bias, if we are to deriveevolutionarily and paleoecologically meaningfulinterpretations. The Phanerozoic record ofpredation is there, but it has not been fully explored;it is especially important to consider multipleworking hypotheses about Phanerozoic predationas we seek to interpret this record.Vermeij (1987) reviewed the record ofmolluscivorous predators, and their multifariousmethods of predation in the Phanerozoic. He madea plea for more data on the responses of prey speciesin Mesozoic and Cenozoic assemblages (Vermeij,1987, p. 239). Fifteen years later, his plea still stands.ACKNOWLEDGMENTSWe dedicate this second entrée to RichardBambach for his pioneering work concerningseafood through time. We greatly appreciateM.Kowalewski and P . Kelley for commissioningus to produce this synthesis. We thank L. Anderson,R. Feldmann, C. Hickman, A. Hoffmeister,G.Storrs, H. Greene, O. Rieppel, J. S. Pearse,R.Portell, and J. Voight for their thoughtfulcomments concerning numerous predation-relatedqueries. A. Hoffmeister, W. Miller III, and twoanonymous reviewers greatly improved the first178

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONdraft of this paper. E. Reitz graciously allowedaccess to her zooarchaeology collections at theGeorgia Museum of Natural History, Athens,Georgia. We also wish to thank Ruth Mawson andPeter Cockle, of Macquarie University, Australia,for facilitating the international editing of thismanuscript. This is not an exhaustive treatment,and any predatory comments regarding this paperneed to be addressed to the first author who willbristle with delight!REFERENCESADNET, S., AND H. CAPPETTA. 2001. A paleontological and phylogenetical analysis of squaliform sharks(Chondrichthyes: Squaliformes) based on dental characters. Lethaia, 34:234–248.AHYONG, S. T., AND C. HARLING. 2000. The phylogeny of the stomatopod Crustacea. Australian Journal of Zoology,48:607–642.ALLMON, W. D., J. C. NIEH, AND R. D. NORRIS. 1990. Drilling and peeling of turritelline gastropods since the LateCretaceous. Palaeontology, 33:595–611.ANDERSON, L. 1992. Naticid gastropod predation on corbulid bivalves: effects of physical factors, morphologicalfeatures, and statistical artifacts. Palaios, 7:602–620.ANDERSON, L. C., D. H. GEARY, R. H. NEHM, AND W. D. ALLMON. 1991. A comparative study of naticid gastropodpredation on Variborbula caloosae and Chione cancellata, Plio-Pleistocene of Florida, USA. Palaeogeography,Palaeoclimatology, Palaeoecology, 85:29–46.ARATA, A. A., AND J. H. HUTCHINSON. 1964. The Raccoon (Procyon) in the Pleistocene of North America. TulaneStudies in Geology, 2:21–27.ARONSON, R. B. 1987. Predation on fossil and Recent ophiuroids. Paleobiology, 13:187–192.ARONSON, R. B. 1988. Palatability of five Caribbean ophiuroids. Bulletin of Marine Science, 43:93–97.ARONSON, R. B. 1991. Predation, physical disturbance, and sublethal arm damage in ophiuroids: a Jurassic-Recentcomparison. Marine Ecology Progress Series, 74:91–97.ARONSON, R. B., D. B. BLAKE, AND T. OJI. 1997. Retrograde community structure in the late Eocene of Antarctica.Geology, 25:903–906.AUSICH, W. I. 1997. Regional encrinites; a vanished lithofacies, p. 509–519. In C. E. Brett and G. C. Baird, PaleontologicalEvents: Stratigraphic, Ecological and Evolutionary Implications. Columbia University Press, New York.AVERY, G., AND W. R. SIEGFRIED. 1980. Food gatherers along South Africa’s seashores. Oceans, 13:33–37.BALUK, W., AND A. RADWANSKI. 1996. Stomatopod predation upon gastropods from the Korytnica Basin, and fromother classical Miocene localities in Europe. Acta Geologica Polonica, 46:279–304BAMBACH, R. K., AND M. KOWALEWSKI. 1999. Diversity of predators compared to the records of prey-predator escalation:Two tales of the history of predation. Geological Society of America Abstracts with Programs, 31(7):A336.BARDET, N. 1992. Stratigraphic evidence for the extinction of ichthyosaurs. Terra Nova, 4:649–656.BARDET, N. 1994. Extinction events among Mesozoic marine mammals. Historical Biology, 7:313–324.BARNES, L. G. 1984. Whales, dolphins and porpoises: evolution of the Cetacea, p. 139–154. In P. D. Gingerichand C. E. Badgely (eds.), Mammals: Notes for a Short Course. University of Tennessee Department ofGeological Sciences, Studies in Geology, 8.BARNES, L. G., AND S. A. MCLEOD. 1984. The fossil record and phyletic relationships of gray whales, p. 3–32. In M. L.Jones, S. L. Swartz, and S. Leatherwood (eds.), The Gray Whale Escrichtius robustus. Academic Press, Orlando, FL.BAUMILLER, T. K., R. MOOI, AND C. G. MESSING. 1999. Predator-prey interactions between cidaroids and crinoids and theirecological and evolutionary implications. Geological Society of America Abstracts with Programs, 31:172–173.BELLWOOD, D. R. 1996. The Eocene fishes of Monte Bolca: the earliest coral reef assemblage. Coral Reefs, 15:11–19.BENTON, M. J. (ed.) 1993. The Fossil Record 2. Chapman and Hall, London, 845 p.BENTON, M. J. 1997. Vertebrate Palaeontology, 2 nd edition. Chapman and Hall, London, New York, 452 p.BERRY, S. S. 1956. A new west Mexican prosobranch mollusk parasitic on echinoids. American Midland Naturalist,56:355–357.179

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BEU, A. G., R. A. HENDERSON, AND C. S. NELSON. 1972. Notes on the taphonomy and paleoecology of NewZealand. New Zealand Journal of Geology and Geophysics, 15:275–286.BISHOP, G. A. 1975. Traces of predation, p. 2671–281. In R. W. Frey (ed.), The Study of Trace Fossils. Springer-Verlag, New York, Heidelberg, Berlin, 562p.BJORNDAL, K. A. 1985. Nutritional ecology of sea turtles. Copeia, 1985:736–751.BITNER, M. A. 1996. Brachiopods from the Eocene La Meseta Formation of Seymour Island, Antarctic Peninsula, p. 65–100. In A. Gazdzicki (ed.), Palaeontological results of the Polish Antarctic expeditions. Palaeontologia Polonica, 55.BLAKE, D. B. 1993. A new asteroid genus from the Jurassic of England and its functional significance. Palaeontology,36:147–154.BLAKE, D. B., AND W. J. ZINSMEISTER. 1979. Two early Cenozoic sea stars (Class Astedroidea) from SeymourIsland, Antarctic Peninsula. Journal of Paleontology, 53:1145–1154.BLOTTE, J. 1980. La faune ichthyologique des gisements due Monte Bolca (Province de Vérone, Italie). Cataloguesystématique présentant l’état actuel des recherches concernant cette faune. Bulletin Muséum Nationald'Histoires Naturelles, Paris (Ser. 4), 2C, 4:339–3.BOTTJER, D. J. 1985. Bivalve paleoecology, p. 112–137. In D. J. Bottjer, C. S. Hickman, and P. D. Ward (eds.), Mollusks:Notes for a Short Course. University of Tennessee Department of Geological Sciences, Studies in Geology, 13.BOTTJER, D., AND D. JABLONSKI. 1988. Paleonevironmental patterns in the evolution of post-Paleozoic marineinvertebrates. Palaios, 3:540–560.BOUCOT, A. J. 1981. Principles of Benthic Marine Paleoecology. Academic Press, NewYork, 463 p.BOUCOT, A. J. 1990. Evolutionary Paleobiology of Behavior and Coevolution. Elsevier, Amsterdam, 725 p.BRANCH, G. M. 2000. Interrelations among rock lobsters, sea urchins, and juvenile abalone: implications forcommunity management. Canadian Journal of Fisheries and Aquatic Science, 57:2175–2185.BRETT, C. E. 1990. Predation, 4.13.1, p. 368–372. In D. E. G. Briggs and P. W. Crowther (eds.), Palaeobiology: ASynthesis. Blackwell Scientific, Oxford.BRETT, C. E. 1995. Sequence stratigraphy, biostratigraphy, and taphonomy in shallow marine environments.Palaios, 10:597–616.BRETT, C. E., AND P. A. ALLISON. 1998. Paleontological approaches to environmental interpretation of marinemudrocks, p. 301–349. In Schieber, J., W. Zimmerle, and P. S. Sethi (eds.), Shales and Mudstones, Vol. 1:Basin Studies, Sedimentology, and Paleontology. E. Schweizerbart’sche, Stuttgart.BRETT, C. E., L. C. IVANY, AND K. M. SCHOPF. 1996. Coordinated stasis: An overview. Palaeogeography,Palaeoclimatology, Palaeoecology, 127:1–20.BRIGGS, D. E. G., AND E. N. K. CLARKSON. 1990. The late Palaeozoic radiation of malacostracan crustaceans, p.165–186. In P. D. Taylor and G. P. Larwood (eds.), Major Evolutionary Radiations. Clarendon Press, Oxford.BROMLEY, R. G. 1981. Concepts in ichnotaxonomy illustrated by small, round holes in shells. Acta GeologicaHispanica, 16:55–64.BROMLEY, R. G. 1993. Predation habits of octopus past and present and a new ichnospecies, Oichnus ovalis.bulletin of the Geological Society of Denmark, 40:167–173.BURROWS, M. 1969. The mechanics and neural control of the prey capture strike in the mantid shrimps Squilla andPseudosquilla. Zeitschrift der Vergleichen Physiologie, 62:361–381.CADÉE, G. C. 1968. Molluscan biocoenoses and thanatocoenoses in the Ria de Arosa, Galicia, Spain. ZoologischeVerhandelingen, 95:1–121.CADÉE, G. C. 1989. Size-selective transport of shells by birds and its palaeoecological implications. Palaeontology,32:429–437.CADÉE, G. C. 1990. Feeding traces and bioturbation by birds on a tidal flat, Dutch Wadden Sea. Ichnos, 1:23–30.CADÉE, G. C. 1995. Birds as producers of shell fragments in the Wadden Sea, in particular the role of the Herringgull. Geobios, 18:77–85.CADÉE, G. C. 1999. Shell damage and shell repair in the Antarctic limpet Nacella concinna from King GeorgeIsland. Journal of Sea Research, 41:149–162.CADÉE, G. C., S. E. WALKER, AND K. FLESSA. 1997. Gastropod shell repair in the intertidal of Bahia la Choya (N.Gulf of California). Palaeogeography, Palaeoclimatology, Palaeoecology, 136:678–78.180

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONCALDWELL, M. W., AND M. S. Y. LEE. 1997. A snake with legs from the marine Cretaceous of the Middle East.Nature, 386:705–652.CALLAWAY, J. 1997a. Introduction, Part I: Ichthyosauria, p. 3–16. In J. Callaway and E. L. Nicholls (eds.), AncientMarine Reptiles. Academic Press, San Diego.CALLAWAY, J. 1997b. A new look at Mixosaurus, Introduction, p. 45–59. In J. Callaway and E. L. Nicholls (eds.),Ancient Marine Reptiles Academic Press, San Diego.CARPENTER, K. 1997. Comparative cranial anatomy of two North American Cretaceous plesiosaurs. In J. Callawayand E. L. Nicholls (eds.), Ancient Marine Reptiles. Academic Press, San Diego.CARROLL, R. L., AND P. GASKILL. 1985. The nothosaur Pachypleurosaurus and the origin of plesiosaurs. PhilosophicalTransactions of the Royal Society of London, B, 309:343–393.CARRIKER, M. 1951. Observations on the penetration of tightly closing bivalves by Busycon and other predators.Ecology, 32:73–83.CARRIKER, M. 1969. Penetration of calcium carbonate substrates by lower plants and invertebrates. AmericanZoologist, 9:629–1020.CARRIKER, M. R. 1981. Shell penetration and feeding by naticacean and muricacean predatory gastropods: asynthesis. Malacologia, 20:403–422.CARRIKER, M. J., AND E. L. YOCHELSON. 1968. Recent gastropod boreholes and Ordovician cylindrical borings. USGeological Survey Professional Paper, B593:1–26.CARTER, R. M. 1967. The shell ornament of Hysteroconcha and Hecuba (Bivalvia): a test case for inferentialfunctional morphology. Veliger, 10:59–71.CARTER, R. M. 1968. On the biology and palaeontology of some predators of bivalved mollusks. Palaeogeography,Palaeoclimatology, Palaeoecology, 4:29–65.CARTER, R. M. 1972. Adaptations of British Chalk Bivalvia. Journal of Paleontology, 46:325–340.CASTILLA, J. C. 1999. Coastal marine communities: trends and perspectives from human-exclusion experiments.Trends in Ecology and Evolution, 14:280–283.CASTILLA, J. C., AND L. R. DURAN. 1985. Human exclusion from the rocky intertidal zone of central Chile: theeffects on Concholepas concholepas (Gastropoda). Oikos, 45:391–399.CATE, A. S., AND I. EVANS. 1994. Taphonomic significance of the biomechanical fragmentation of live molluscanshell material by a bottom-feeding fish (Pogonias cromis) in Texas coastal bays. Palaios, 9:254–274.CHECA, A., AND G. E. G. WESTERMANN. 1989. Segmental growth in planulate ammonites: inferences on costalfunction. Lethaia, 22:95–100.CHIAPPE, L. M. 1995. The first 85 million years of avian evolution. Nature, 378:349–355.CHOATE, J. H., AND D. R. BELLWOOD. 1991. Reef fishes: their history and evolution, p. 39–66. In P. Sales (ed.), TheEcology of Fishes on Coral Reefs. Academic Press, San Diego.COOPER, G. A. 1988. Some Tertiary brachiopods of the East Coast of the United States. Smithsonian Contributionsto Paleobiology, 64:1–54.CORTÉS, E. 1999. Standardized diet compositions and trophic levels of sharks. ICES Journal of Marine Science,56:707–717.COSTA, D. 1978. The ecological energetics, water, and electrictrolyte balance of the California sea otter, Enhydralutris. Dissertation, University of California, Santa Cruz, CA.COX, C., J. H. HUNT, W. G. LYONS, AND G. E. DAVIS. 1997. Nocturnal foraging of the Caribbean spiny lobster(Panulirus argus) on offshore reefs of Florida. Marine and Freshwater Research, 48:671–679.CURREY, J. D., AND A. J. KOHN. 1976. Fracture in the crossed-lamellar structure of Conus shells. Journal of MaterialScience, 11:1615–1623.DEBELIUS, H. 1999. Crustacea Guide of the World. IKAN-Unterwasserarchiv Press, Frankfurt, 321 p.DE CAUWER, G. 1985. Gastropod predation on corbulid bivalves: palaeoecology or taphonomy? Annales de laSociété Royale Zoologique de Belgique, 115:183–196.DEMERE, T. A., AND R. A. CERUTTI. 1982. A Pliocene shark attack on a cethotheriid whale. Journal of Paleontology,56:1480–1482.DENTON, R. K., JR., J. L. DOBIE, AND D. C. PARRIS. 1997. The Marine Crocodilian Hyposaurus in North America,p. 375–397. In J. Callaway and E. L. Nicholls (eds.), Ancient Marine Reptiles. Academic Press, San Diego.181

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002DIETL, G., AND R. R. ALEXANDER. 1998. Shell repair frequencies in whelks and moon snails from Delaware andsouthern New Jersey. Malacologia, 39:151–165.DIETL, G., AND G. SCHWEIGERT. 1999. A nautiloid with a complete jaw apparatus still in situ from the LithographicLimestone of Nusplingen (Upper Jurassic, SW Germany). Neues Jahrbuch für Geologie und Palaeontologie,Abhandlungen, 211:75–87.DIETL, G., R. R. ALEXANDER, AND W. F. BIEN. 2000. Escalation in Late Cretaceous-early Paleocene oysters(Gryphaeidae) from the Atlantic Coastal Plain. Paleobiology, 26:215–237.DI GIANCOMO, E. E. AND M. R. PERIER. 1996. Feeding habits of cockfish, Callorhinchus callorhyncus (Holocephali:Callorhynchidae) in Patagonian waters (Argentina). Marine and Freshwater Research, 47:801–808.DODD, C. K., JR., AND G. S. MORGAN. 1992. Fossil sea turtles from the Early Pliocene Bone Valley Formation,central Florida. Journal of Herpetology, 26:1–8.DOLLO, L. 1913. Globidens fraasi, mosasaurien mylodonte nouveau du Maestrichtien (Crétacé supérieur) duLimbourg, et l’ethologie de la nutrition chez les mosasuriens. Archiv Biol, 28:609–626.DOMNING, D. P. 1976. An ecological model for late Tertiary sirenian evolution in the North Pacific Ocean. SystematicZoology, 25:352–362.DONOVAN, S. K. 1993. Contractile tissues in the cirri of ancient crinoids: criteria for recognition. Lethaia, 26:163–169.DONOVAN, S. K. 2001. Evolution of Caribbean echinoderms during the Cenozoic: moving towards a completepicture using all of the fossils. Palaeogeography, Palaeoclimatology, Palaeoecology, 166:177–192.DONOVAN, D. A., J. P. DANKO, AND T. H. CAREFOOT. 1999. Functional significance of shell sculpture in gastropodmolluscs: test of a predator-deterrent hypothesis in Ceratostoma foliatum (Gmelin). Journal of ExperimentalMarine Biology and Ecology, 236:235–251.DRINNAN, R. E. 1957. The winter feeding of the oystercatcher (Haematopus ostralegus) on the edible mussel(Mytilus edulis) in the Conway estuary, North Wales. Fishery Investigations, Series 2, 22:1–15.DRUCKENMILLER, P. S., A. J. DAUN, J. L. SKULAN, AND J. C. PLADZIEWICZ. 1993. Stomach contents in the UpperCretaceous shark Squalicorax falcatus. Journal of Vertebrate Paleontology, 13(Suppl. 3):33A–34A.ELNER, R. W., AND D. G. RAFFAELLI. 1980. Interactions between two marine snails, Littorina rudis Maton andLittorina nigrolineata Gray, a predator, Carcinus maenas (L.), and a parasite, Microphallus similis Jägerskiold.Journal of Experimental Marine Biology and Ecology, 43:151–160.ENGESER, T. 1988. Fossil “octopods”—a critical review, p. 81–87. In M. R. Clarke and E. R. Trueman (eds.), TheMollusca, Volume 12, Paleontology and Neontology of Cephalopods. Academic Press, San Diego.ESTES, J. A., AND J. F. PALMISANO. 1974. Sea otters: their role in structuring nearshore communities. Science,185:1058–1060.ESTES, J. A., R. J. JAMESON, AND E. B. RHODE. 1982. Activity and prey selection in the sea otter: influence ofpopulation status on community structure. The American Naturalist, 120:242–258.FEDUCCIA, A. 1995. Explosive evolution in Tertiary birds and mammals. Science, 267:637–638.FEIFAREK, B. P. 1987. Spines and epibionts as antipredator defenses in the thorny oyster Spondylus americanusHermann. Journal of Experimental Marine Biology and Ecology, 105:39–56.FISCHER, P. H. 1963. Corbules fossiles perforées par des gastéropodes prédateurs. Journal de ConchyliologieParis, 103:29–31.FISHER, D. C. 1981a. Crocodilian scatology, mirovertebrate concentrations, and enamel-less teeth. Paleobiology,7:262–275.FISHER, D. C. 1981b. Mode of preservation of the Shotgun local fauna (Paleocene, Wyoming) and its implicationfor the taphonomy of microvertebrate concentration. University of Michigan Contributions to the Museumof Paleontology, 25:247–257.FUGITA, S. 1916. On the boring of pearl oysters by Octopus (Polypus) vulgaris Lamarck. Dobytsugaki Zasshi,28:250–257.FÜRSICH, F. T., AND D. JABLONSKI. 1984. Late Triassic naticid drill holes: carnivorous gastropods gain a majoradaptation but fail to radiate. Science, 224:78–80.GEARY, D. H., W. D. ALLMON, AND M. J. REAKA-KUDLA. 1991. Stomatopod predation on fossil gastropods from thePlio-Pleistocene of Florida. Journal of Paleontology, 65:355–360.182

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONGELLER, J. B. 1983. Shell repair frequencies of two intertidal gastropods from northern California: Microhabitatdifferences. The Veliger, 26:113–115.GIBSON, M. A., AND J. B. WATSON. 1989. Predatory and non-predatory borings in echinoids from the upper Ocala formation(Eocene), North Central Florida, U.S.A. Palaeogeography, Palaeoclimatology, Palaeoecology, 71:309–321.GINGERICH, P. D., M. L. HAG, I. S. ZALMOUT, I. HUSSAIN KHAN, AND M. SADIQ MALKANI. 2001. Origin of whalesfrom early Artiodactyls: hands and feet of Eocene Protocetidae from Pakistan. Science, 293:2239–2242.GLAESSNER, M. F. 1960. The fossil decapod Crustacea of New Zealand and the evolution of the Order Decapoda.New Zealand Geological Survey Paleontological Bulletin, 31:1–63.GLAESSNER, M. F. 1969. Decapoda, p. R399–R533, R626–R628. In R. C. Moore (ed.), Treatise on InvertebratePaleontology, part R, Arthropoda. University of Kansas and Geological Society of America, Lawrence, KS.GLODEK, G. S., AND H. K. VORIS. 1982. Marine snake diets: Prey consumption, diversity, and overlap. Copeia, 3:661–666.GODFREY, S. 1984. Plesiosaur subaqueous locomotion: a reappraisal. Neues Jahrbuch für Geologie undPaläontologie, Monatshefte, 11:661–672.GORE, R. H., AND L. E. SCOTTO. 1979. Crabs of the Family Parthenopidae (Crustacea, Brachyura: Oxyrhyncha)with notes on specimens from the Indian River region of Florida. Memoirs of the Hourglass Cruises, FloridaMarine Research Institution, 3:1–98.GOTTFRIED, M. D., L. J. V. COMPAGNO, AND S. C. BOWMAN. 1996. Size and skeletal anatomy of the giant “megatooth”shark Carcharodon megalodon, p. 55–66. In A. P. Klimley (ed.), Great White Sharks: The Biology ofCarcharodon carcharia. Academic Press, San Francisco.GOULD, S. J., AND E. VRBA. 1982. Exaptation—A missing term in the science of form. Paleobiology, 8:4–15.GRANT, R. E. 1966. Spine arrangement and life habits of the productoid brachiopod Waagenoconcha. Journal ofPaleontology, 40:1063–1069.GRANT-MACKIE, J. A., AND M. CHAPMAN-SMITH. 1971. Paleontological notes on the Castlecliffian Te Piki Bed,with descriptions of a new molluscan taxa. New Zealand Journal of Geology and Geophysics, 14:665–704.GREENSTEIN, B. J. 1993. Is the fossil record of regular echinoids really so poor?—A comparison of living andsubfossil assemblages. Palaios, 8:587–540.GRAY, A. E., F. J. MULLIGAN, AND R. W. HANNAH. 1997. Food habits, occurrence, and population structure of thebat ray, Myliobatis californica, in Humbolt Bay, California. Environmental Biology of Fishes, 49:227–238.GREGOIRE, C. 1972. Structure of the molluscan shell, p. 45–102. In M. Florikin and B. T. Scheer (eds.), ChemicalZoology, Volume 7, Mollusca. Academic Press, New York.GREGORY, M. P., P. F. BALANCE, G. W. GIBSON, AND A. M. AYLING. 1979. On how some rays (Elasmobranchia)excavate feeding depressions by jetting water. Journal of Sedimentary Petrology, 49:1125–1130.HALEY, D. 1986. Marine Mammals. Pacific Search Press, Seattle, Washington, 296 p.HALLAM, A. 1976. Stratigraphic distribution and ecology of European Jurassic bivalves. Lethaia, 9:245–259.HALLOCK, P., AND H. K. TALGE. 1994. A predatory foraminifer, Floresina amphiphaga n. sp., from the FloridaKeys. Journal of Foraminiferal Research, 24:210–213.HANSEN, T. A., AND P. H. KELLEY. 1995. Spatial variation of naticid gastropod predation in the Eocene of NorthAmerica. Palaios, 10:268–278.HANSEN, T. A., P. H. KELLEY, V. D. MELLAND, AND S. E. GRAHAM. 1999. Effect of climate-related mass extinctionson escalation in mollusks. Geology, 27:1139–1142.HÄNTZSHEL, W., F. EL-BAZ, AND G. C. AMSTUTZ. 1968. Coprolites: an annotated bibliography. Memoir of theGeological Society of America, 108:1–132.HARPER, E. M. 1991. The role of predation in the evolution of cementation in bivalves. Palaeontology, 34:455–460.HARPER, E. M. 1994. Are conchiolin sheets in corbulid bivalves primarily defensive? Palaeontology, 37:551–578.HARPER, E. M. 2002. Plio-Pleistocene octopod drillling behavior in scallops from Florida. Palaios, 17:292–295.HARPER, E. M., AND P. SKELTON. 1993. The Mesozoic marine revolution and epifaunal bivalves. Scripta Geologica,Special Issue, 2:127–153.HASSON, P. F., AND A. G. FISCHER. 1986. Observations on the Neogene of northwestern Ecuador. Micropaleontology,32:32–42.HAWKINS, W. B., JR. 1990. Taphonomy of an Upper Cretaceous (Maastrichtian) mosasaur, Braggs, Alabama.Journal of Vertebrate Paleontology, 10(Suppl. 3):26A.183

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002HAYAMI, I., AND Y. KANIE. 1980. Mode of life of a giant capulid gastropod from the Upper Cretaceous of Saghalienand Japan. Palaeontology, 23:689–698.HEATWOLE, H. 1987. Sea Snakes. The New South Wales University Press, Australia, 85 p.HECHT, M. K., C. KROPACH, AND B. M. HECHT. 1974. Distribution of the yellowbellied sea snake, Pelamis platurus,and its significance in relation to the fossil record. Herpetologica, 30:387–396.HENDRICKSON, J. R. 1980. The ecological strategies of sea turtles. American Zoologist, 20:597–609.HEWITT, R. A., AND G. E. G. WESTERMANN. 1990. Mosasaur tooth marks on the ammonite Placenticeras from theUpper Cretaceoous BearPaw Formation of Alberta. Canadian Journal of Earth Science, 27:469–472.HICKMAN, C. S. 1980. Gastropod radulae and the assessment of form in evolutionary paleontology. Paleobiology,6:276–294.HINES, A. H., AND J. S. PEARSE. 1982. Abalones, shells, and sea otters: dynamics of prey populations in centralCalifornia. Ecology, 63:1547–1560.HIRAYAMA, R. 1997. Distribution and diversity of Cretaceous cehloniods, p. 225–241. In J. Callaway and E. L.Nicholls (eds.), Ancient Marine Reptiles. Academic Press, San Diego.HOF, C. H. J. 1998. Fossil Stomatopods (Crustacea: Malacostraca) and their phylogenetic impact. Journal ofNatural History, 32:1567–1576.HOF, C. H. J., AND D. E. G. BRIGGS. 1997. Decay and mineralization of mantis shrimps (Stomatopoda: Crustacea)—a key to their fossil record. Palaios, 12:420–438.HOFFMEISTER, A. P., AND M. KOWALEWSKI. 2001. Spatial and environmental variation in the fossil record of drillingpredation: a case study from the Miocene of Central Europe. Palaios, 16:566–579.HOGLER, J. A. 1992. Taphonomy and paleoecology of Shonisaurus popularis (Reptilia: Ichthyosauria). Palaios,7:108–117.HOLLAND, S. M. 2000. The quality of the fossil record: a sequence stratigraphic perspective, p. 148–168. In D. H.Erwin and S. L. Wing (eds.), Deep Time: Paleobiology’s Perspective. The Paleontological Society.HOUSE, M. R., AND J. R. SENIOR (eds.). 1981. The Ammonoidea. Academic Press, London.HOWARD, J. D., T. V. MAYOU, AND R. W. HEARD. 1977. Biogenic sedimentary structures formed by rays. Journal ofSedimentary Petrology, 47:339–346.HUA, S., AND E. BUFFETAUT. 1997. Introduction, Part V: Crocodylia, p. 357–374. In J. Callaway and E. L. Nicholls(eds.), Ancient Marine Reptiles. Academic Press, San Diego.HUGHES, R. N., AND P. I. HUGHES. 1981. Morphological and behavioral aspects of feeding in the Cassidae (Tonnacea,Mesogastropoda). Malacologia, 20:385–402.JABLONSKI, D., AND D. J. BOTTJER. 1990. Onshore-offshore trends in marine invertebrate evolution, p. 21–75. In W. Allmonand R. D. Ross (eds.), Causes of Evolution: A Paleontological Perspective. University of Chicago Press, Chicago.JABLONSKI, D., J. J. SEPKOSKI, JR., D. J. BOTTJER, AND P. M. SHEEHAN. 1983. Onshore-offshore patterns in theevolution of Phanerozoic shelf communities. Science, 222:1123–1125.JERARDINO, A., J. C. CASTILLA, J. M. RAMIREZ, AND N. HERMOSILLA. 1992. Early coastal subsistence patterns in Central Chile: asystematic study of the marine-invertebrate fauna from the site of Curaumilla-1. Latin American Antiquity, 3:43–62.KABAT, A. R. 1990. Predatory ecology of naticid gastropods with a review of shell boring predation. Malacologia,32:155–203.KARDON, G. 1998. Evidence from the fossil record of an antipredatory exaptation: conchiolin layers in corbulidbivalves. Evolution, 52:68–79.KARL, S., AND S. OBREBSKI. 1976. The feeding biology of the bat ray, Myliobatis californica in Tomales Bay, California,p. 181–186. In C. A. Simenstad and S. J. Lipovski (eds.), Fish Food Habit Studies. Washington Sea Grant, Seattle.KASE, T., P. JOHNSON, A. SEILACHER, AND J. BOYCE. 1998. Alleged mosasaur bite marks on Late Cretaceous ammonitesare limpet (patellogastropod) home scars. Geology, 26:947–950.KAUFFMAN, E. G. 1972. Ptychodus predation in a Cretaceous Inoceramus. Palaeontology, 15:439–444.KAUFFMAN, E. G. 1990. Mosasaur predation on ammonites during the Cretaceous—an evolutionary history, p. 184–189. In A. J. Boucot (ed.), Evolutionary Paleobiology of Behaviour and Coevolution. Elsevier, Amsterdam.KAUFFMAN, E. G., AND R. V. KESLING. 1960. An Upper Cretaceous ammonite bitten by a mosasaur. Contributionsof the Museum of Paleontology, University of Michigan, 15:193–248.184

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONKELLER, T. 1976. Magen- und Darminhalte von Ichthyosaurien des süddeutschen Posidonienschiefers. NeusJahrbuch fur Geologie und Paläontologie Monatshefte, 5:266–283.KELLEY, P. H. 1988. Predation by Miocene gastropods of the Chesapeake Group: stereotyped and predictable.Palaios, 3:436–448.KELLEY, P. H., AND T. A. HANSEN. 1993. Evolution of the naticid gastropod predator-prey system: an evaluation ofthe hypothesis of escalation. Palaios, 8:358–375.KELLEY, P. H., AND T. A. HANSEN. 2001. The role of ecological interactions in the evolution of naticid gastropodsand their molluscan prey, p. 149–170. In W. D. Allmon and D. J. Bottjer (eds.), Evolutionary Paleoecology.Columbia University Press, New York.KELLEY, P. H., T. A HANSEN, S. E. GRAHAM, AND A. G. HUNTOON. 2001. Temporal patterns in the efficiency ofnaticid gastropod predators during the Cretaceous and Cenozoic of the United States Coastal Plain.Palaeogeography, Palaeoclimatology, Palaeoecology, 166:165–176.KIER, P. M. 1977. The poor fossil record of the regular echinoid. Paleobiology, 3:168–174.KIER, P. M. 1981. A bored Cretaceous echinoid. Journal of Paleontology, 55:656–659.KIER, P. M. 1982. Rapid evolution in echinoids. Palaeontology, 25:1–9.KIRCH, P. V. 1983. Man’s role in modifying tropical and subtropical polynesian ecosystems. Archaeology inOceania, 18:26–31.KITCHELL, J. A. 1986. the evolution of predatory-prey behavior: naticid gastropods and their molluscan prey, p.88–110. In M. Nitecki and J. A. Kitchell (eds.), Evolution of Animal Behavior: Paleontological and FieldApproaches. Oxford University Press, Oxford,KITCHELL, J. A., C. H. BOGGS, J. F. KITCHELL, AND J. A. RICE. 1981. Prey selection by naticid gastropods: experimentaltests and application to the fossil record. Paleobiology, 7:533–552.KJAER, C. R., AND E. THOMSEN. 1999. Heterochrony in bourgueticrinid sea-lilies at the Cretaceous-Tertiary boundary.Paleobiology, 25:29–40.KNOLL, A. S., R. K. BAMBACH, J. P. GROTZINGER, AND D. CANFIELD. 1996. Comparative Earth history and LatePermian mass extinction. Science, 273:452–457.KOWALEWSKI, M. 1993. Morphometric analysis of predatory drillholes. Palaeogeography, Palaeoclimatology,Palaeocology, 102:69–88.KOWALEWSKI, M., AND K. FLESSA. 1997. Predatory scars in the shells of a Recent lingulid brachiopod: paleontologicaland ecological implications. Acta Palaeontologica Polonica, 42:497–532.KOWALEWSKI, M., A. DULAI, AND F. T. FÜRSICH. 1998. A fossil record full of holes: the Phanerozoic history ofdrilling predation. Geology, 26(12):1091–1094.KRÖGER, B. 2002. On the efficiency of the buoyancy apparatus in ammonoids: evidences from sublethal shellinjuries. Lethaia, 35:61–70.KUNZE, J. C. 1983. Stomatopoda and the evolution of the Hoplocarida, p. 165–188. In F. R. Schram, CrustaceanPhylogeny. A. A. Balkema, Rotterdam.KVITEK, R. G., J. S. OLIVER, A. R. DEGANGE, AND B. S. ANDERSON. 1992. Changes in Alaskan soft-bottom preycommunities along a gradient in sea otter predation. Ecology, 73:413–428.KVITEK, R. G., C. E. BOWLBY, AND M. STAEDLER. 1993. Diet and foraging behavior of sea otters in southeastAlaska. Marine Mammal Science, 9:168–181.LAUDER, G. V. 1985. Aquatic feeding in lower vertebrates, p. 210–229. In M. Hildebrand, D. M. Bramble, K. F.Liem, and D. B. Wake (eds.), Functional Vertebrate Morphology. Belknap Press, Cambridge.LEE, M. S. Y., AND M. W. CALDWELL. 2000. Adriosaurus and the affinities of Mosasaurs, Dolichosaurs, and Snakes.Journal of Paleontology, 74:915–937.LEHMANN, U. 1975. Über Nahrung und Ernahrungsweise der Ammoniten. Paläontologische Zeitschrift, 49:187–195.LEWY, Z., AND C. SAMTLEBEN. 1979. Functional morphology and palaeontological significance of the conchiolinlayers in corbulid pelecypods. Lethaia, 12:341–351.LINDBERG, D., AND J. CARLTON. 1969. Intertidal marine mollusca of Southeast Farallon Island, San Francisco,California. Annual Report of the Western Society of Malacologists, 11:7.LINDBERG, D., AND M. KELLOGG. 1982. Bathymetric anomalies in the Neogene fossil record: the role of divingmarine birds. Paleobiology, 8:402–407.185

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002LINGHAM-SOLIAR, T. 1998a. Unusual death of a Cretaceous giant. Lethaia, 31:308–310.LINGHAM-SOLIAR, T. 1998b. A new mosasaur Pluridens walkeri from the Upper Cretaceous, Maastrichtian of theIullemmeden Basin, southwest Niger. Journal of Vertebrate Paleontology, 18:709–717.LINGHAM-SOLIAR, T. 1999. The durophagous Mosasaurs (Lepidosauromorpha, Squamata) Globidens and Carinodensfrom the Upper Cretaceous of Belgium and The Netherlands. Palaeontological Journal, 33:638–647.LINGHAM-SOLIAR, T. 2001. The ichthyosaur integument: skin fibers, a means for a strong, flexible and smoothskin. Lethaia, 34:287–302.LOGAN, A. 1974. Morphology and life habits of the Recent cementing bivalve Spondylus americanus Hermannfrom the Bermuda platform. Bulletin of Marine Science, 24:473–492.LUCAS, S. 1997. Marine reptiles and Mesozic biochronology, p. 423–434. In J. Callaway and E. L. Nicholls(eds.), Ancient Marine Reptiles. Academic Press, San Diego.MADDOCKS, R. E. 1988. One hundred million years of predation on ostracods: The fossil record in Texas, p. 637–657. In T. Hanai, N. Ikeya, and K. Ishizaki (eds.), Evolutionary Biology of Ostracoda. Elsevier, Amsterdam.MAISEY, J. G. 1982. The anatomy and relationships of Mesozoic hybodont sharks. American Museum Novitates, 2724:1–48.MAISEY, J. G. 1996. Discovering Fossil Fishes. Henry Holt, New York, 223 p.MANNING, R. B. 1995. Stomatopod crustacea of Vietnam: The legacy of Raoul Serène. Crustacean Research,Special Number 4 (Tokyo: The Carcinological Society of Japan), 339 p.MARTILL, D. 1986a. The preservation of marine vertebrates in the Lower Oxford Clay (Jurassic) of central England.Philosophical Transactions of the Royal Society of London, B, 311:155–165.MARTILL, D. 1986b. The diet of Metriorhynchus, a Mesozoic marine crocodile. Neues Jahrbuch für Geologie undPalaeontologie, Monatshefte, 1986(10):621–625.MARTILL, D. 1990. Predation on Kosmoceras by semionotid fish in the Middle Jurassic Lower Oxford Clay ofEngland. Palaeontology, 33:739–742.MARTILL, D., M. A. TAYLOR, K. L. DUFF, J. B. RIDING, AND P. R. BOWN. 1994. The trophic structure of the biota ofthe Peterborough Member, Oxford Clay Formation (Jurassic), UK. Journal of the Geological Society, London,151:173–194.MARTIN, J. E., AND P. R. BJORK. 1987. Gastric resudues associated with the mosasaur Tylosaurus from the LateCretaceous (Campanian) Pierre Shale, South Dakota. Dakoterra, 3:68–70.MARTIN, L. D., AND B. ROTHSCHILD. 1989. Paleopathology and diving mosasaurs. American Scientist, 77:460–467.MASSARE, J. A. 1987. Tooth morphology and prey preference of Mesozoic marine reptiles. Journal of VertebratePaleontology, 7:121–137.MASSARE, J. A. 1997. Introduction, Part VI: Faunas, Behavior, and Evolution, p. 401–421. In J. Callaway andE. L. Nicholls (eds.), Ancient Marine Reptiles. Academic Press, San Diego.MASSARE, J. A. AND J. M. CALLAWAY. 1990. The affinities and ecology of Triassic ichthyosaurs. Geological Societyof America Bulletin, 102:409–416.MATSUKAMA, A. 1978. Fossil boreholes by shell-boring predators or commensals, I: Boreholes of capulid gastropods.Venus (Japan Journal of Malacology), 376:29–45.MAYFIELD, S., E. DE BEER, AND G. M. BRANCH. 2001. Prey preference and the consumption of sea urchins andjuvenile abalone by captive rock lobsters (Jasus lalandii). Marine and Freshwater Research, 52:773–780.MCCLINTOCK, J. B., AND K. R. MARION. 1993. Predation by the King Helmet (Cassis tuberosa) on six-holed sanddollars (Leodia sexiesperforata) at San Salvador, Bahamas. Bulletin of Marine Science, 52:1013–1017.MCCOSKER, J. E. 1975. Feeding behavior of Indo-Australian Hydrophiidae, p. 217–232. In W. A. Dunson (ed.),The Biology of Sea Snakes. University Park Press, Baltimore.MCROBERTS, C. A. 2001. Triassic bivalves and the initial Mesozoic marine revolution: a role for predators. Geology,29:359–362.MEENAKSHI, V. R., A. W. MARTIN, AND K. M. WILBUR. 1974. Shell repair in Nautilus macromphalus. MarineBiology, 27:27–35.MEYER, D. L. 1985. Evolutionary implications of predation on Recent comatulid crinoids. Paleobiology, 11:154–164.MEYER, D. L., AND W. I. AUSICH. 1983. Biotic interactions among Recent crinoids and among fossil crinoids, p.377–427. In M. J. S. Tevesz and P. L. McCall (eds.), Biotic Interactions in Recent and Fossil BenthicCommunities. Plenum Press, New York.186

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONMEYER, D. L., AND D. B. MACURDA, JR. 1977. Adaptive radiation of the comatulid crinoids. Paleobiology, 3:74–82.MILINKOVITCH, M. C. 1995. Molecular phylogeny of cetaceans prompts revision of morphological transformations.Trends in Ecology and Evolution, 10:328–334.MILSON, C. V. 1994. Saccocoma, a benthic crinoid from the Jurassic Solnhofen Limestone, Germany. Palaeontology,37:121–130.MORENO, C. A. 2001. Community patterns generated by human harvesting on Chilean shores, a review: AquaticConservation. Marine and Freshwater Ecosystems, 11:19–30.MORENO, C. A., J. P. SUTHERLAND, AND H. F. JARA. 1984. Man as a predator in the intertidal zone of southern Chile.Oikos, 42:155–160.MORTIMER, J. A. 1982. Feeding ecology of sea turtles. In K. A. Bjorndal (ed.), Biology and Conservation of SeaTurtles. Smithsonian Institution Press, Washington, D.C.MORTON, B., AND K. CHAN. 1997. First report of shell boring predation by a member of the Nassariidae (Gastropoda).Journal of Molluscan Studies, 63:476–78.MOSS, S. A. 1977. Feeding mechanisms in sharks. American Zoologist, 17 (2):355–364.MOTANI, R., B. M. ROTHSCHILD, AND W. WAHL. 1999. Large eyeballs in diving ichthyosaurs. Nature, 402:747.MOTTA, P. J., R. E. HUETER, T. C. TRICAS, AND A. P. SUMMERS. 2002. Kinematic analysis of suction feeding in thenurse shark, Ginglymostoma cirratum (Orectolobiformes, Ginglymostomatidae). Copeia, 1:24–38.DE MUIZON, C. 1993. Walrus-like feeding adaptation in a new cetacean from the Pliocene of Peru. Nature, 365:745–748.NAVARETTE, S. A., AND J. C. CASTILLA. 1993. Predation by Norway rats in the intertidal zone of Central Chile.Marine Ecology Progress Series, 92:187–199.NEBELSICK, J. H. 1995. Comparative taphonomy of Clypeasteroids. Ecologae Geologicae Helvetiae, 88:685–693.NEBELSICK, J. H. 1999. Taphonomic comparison between Recent and fossil sand dollars. Palaeogeography,Palaeoclimatology, Palaeoecology, 149:349–358.NEBELSICK, J. H., AND M. KOWALEWSKI. 1999. Drilling predation on Recent clypeasteroid echinoids from the RedSea. Palaios, 14:127–144.NEMOTO, T. 1970. Feeding pattern of baleen whales in the ocean, p. 241–252. In J. H. Steele (ed.), Marine FoodChains. Oliver & Boyd, Edinburgh.NERINI, M. 1984. A review of gray whale feeding ecology, p. 423–450. In M. L. Jones, S. L. Swartz, and S.Leatherwood (eds.), The Gray Whale Escrichtius robustus. Academic Press, Orlando.NEUMANN, C. 2000. Evidence of predation on Cretaceous sea stars from northwest Germany. Lethaia, 33:65–70.NEWTON, C. R. 1983. Triassic origin of shell-boring gastropods. Geological Society of America Abstracts withPrograms, 15:652–653.NICHOL, D. 1965. Ecologic implications of living pelecypods with calcareous spines. Nautilus, 78:109–115.NICHOLLS, E. L. 1997. Introduction (to Part III: Testudines), p. 219–223. In J. Callaway and E. L. Nicholls (eds.),Ancient marine reptiles. Academic Press, San Diego.NIELSON, C. 1975. Observations on Buccinum undatum L. attacking bivalves and on prey responses with a shortreview of attack methods of other prosobranchs. Ophelia, 13:87–108.NIXON, M. 1980. The salivary papilla of Octopus as an accessory radula for drilling shells. Journal of Zoology, 190:53–57.NIXON, M. 1988. The feeding mechanisms and diets of cephalopods—Living and fossil, p. 642–652. In J. Wiedmannand J. Kullmann (eds.), Cephalopods Present and Past. 2 nd International Cephalopod Symposium, E.Schweizerbart’sche Verlagsbuchhandlung, Stuttgart, Germany.NORRIS, K. S., B. VILLA-RAMIREZ, G. NICHOLS, B. WURSIG, AND K. MILLER. 1983. Lagoon entrance and otheraggregations of gray whales, Eschrictius robustus, p. 259–293. In R. Payne (ed.), Behavior and Communicationof Whales. Westview Press, Boulder, Colorado.NORTON, S. F. 1988. Role of the gastropod shell and operculum in inhibiting predation by fishes. Science, 241:92–94.NORTON, S. F. 1995. A functional approach to ecomorphological patterns of feeding in cottid fishes. EnvironmentalBiology of Fishes, 44:61–78.OJI, T. 1986. Skeletal variation related to arm regeneration in Metacrinus and Saracrinus, Recent stalked crinoids.Lethaia, 19:355–360.OJI, T. 1996. Is predation intensity reduced with increasing depth?—Evidence from the west Atlantic stalked crinoidEndoxocrinus parrae (Gervais) and implications for the Mesozoic Marine Revolution. Paleobiology, 22:339–351.187

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002OJI, T. 2001. Fossil record of echinoid regeneration. Microscopy, Research and Technique, 55(6):397–402.OJI, T., AND T. OKAMOTO. 1994. Arm autotomy and arm branching pattern as anti-predatory adaptations in stalkedand stalkless crinoids. Paleobiology, 20:27–39.O’KEEFE, F. R. 2002. The evolution of plesiosaur and pliosaur morphotypes in the Plesiosauria (Reptilia:Sauropterygia). Paleobiology, 28:101–112.OLIVER, J. S., P. N. SLATTERY, E. F. O’CONNER, AND I. F. LOWRY. 1983. Walrus feeding in the Bering Sea; a benthicperspective. Fisheries Bulletin, 81:501–512.OLSON, S. L., AND Y. HASEGAWA. 1979. Fossil counterparts of giant penguins from the north Pacific. Science,206:688–689.OLSON, S. L., AND D. W. STEADMAN. 1978. The fossil record of the Glareolidae and Haematopodidae (Aves:Charadriformes). Proceedings of the Biological Society of Washington, 91:972–981.ORR, V. 1962. The drilling habit of Capulus danieli (Crosse) (Mollusca: Gastropoda). Veliger, 5:63–67.OYEN, C. W., AND R. W. PORTELL. 2001. Diversity patterns and biostratigraphy of Cenozoic echinoderms fromFlorida. Palaeogeography, Palaeoclimatology, and Palaeoecology, 166:193–218.PACKARD, A. 1972. Cephalopods and fish: The limits of convergence. Biological Reviews, 47:241–307.PAINE, R. T. 1963. Ecology of the brachiopod Glottidia pyramidata. Ecological Monographs, 33:187–213.PALMER, A. R. 1979. Fish predation and the evolution of gastropod shell sculpture: experimental and geographicevidence. Evolution, 33:697–713.PALMER, T. J. 1982. Cambrian to Cretaceous changes in hardground communities. Lethaia, 15:309–323.PARRIS, D. C., J. L. DOBIE, AND A. B. MCPHERSON. 2000. A fossil marine turtle from east central Louisiana. TexasJournal of Science, 52:230–234.PATTERSON, C. 1994. Bony fishes, p. 57–84. In D. R. Prothero and R. M. Schoch (eds.), Major Features of VertebrateEvolution. Short Courses in Paleontology, 7. Paleontological Society, Knoxville, Tennessee.PETHER, J. 1995. Belichnus new ichnogenus, a ballistic trace on mollusc shells from the Holocene of the Benguelaregion, South Africa. Journal of Paleontology, 69:171–181.PILSON, M. E., AND P. B. TAYLOR. 1961. Hole drilling by octopus. Science, 134:1366–1368.PLOTKIN, P. T., M. K. WICKSTEN, AND A. F. AMOS. 1993. Feeding ecology of the loggerhead sea turtle Carettacaretta in the Northwestern Gulf of Mexico. Marine Biology, 115:1–15.POLLARD, J. E. 1968. The gastric contents of an ichthyosaur from the Lower Lias of Lyme Regis, Dorset.Palaeontology, 11:376–388.PONDER, W. F., AND J. D. TAYLOR. 1992. Predatory shell drilling by two species of Austroginella (Gastropoda:Marginellidae). Journal of Zoology, London, 228:317–328.PREZANT, R. S. 1983. Auto-drilling in the oyster drill Thais haemastoma (Muricidae). The Nautilus, 97:81–84.RAFFAELLI, D. G. 1978. The relationship between shell injuries, shell thickness, and habitat characteristics of theintertidal snail Littorina rudis Maton. Journal of Molluscan Studies, 44:166–170.RAMSAY, K., C. A. RICHARDSON, AND M. J. KAISER. 2001. Causes of shell scarring in dog cockles, Glycymerisglycymeris L. Journal of Sea Research, 45:131–139.RATHBUN, R. 1930. Fossil decapod crustaceans from Mexico. Proceedings of the United States National Museum,78:1–10.RATHBUN, R. 1935. Fossil crustacea of the Atlantic and Gulf Coastal Plain. Geological Society of America SpecialPaper, 2:1–160.RAY, M., AND A. W. STONER. 1995. Predation on a tropical spinose gastropod: the role of shell morphology. Journalof Experimental Marine Biology and Ecology, 187:207–222.REPENNING, C. A. 1976. Adaptive evolution of sea lions and walruses. Systematic Zoology, 25:375–390.REYMENT, R. A. 1971. Introduction to Quantitative Palaeoecology. Elsevier, Amsterdam.RICKETTS, E. F., J. CALVIN, AND J. W. HEDGEPETH. 1985. Between Pacific Tides. Stanford University Press, Stanford, 652 p.RIEPPEL, O. 1997. Introduction (to Part II: Sauropterygia), p. 107–119. In J. Callaway and E. L. Nicholls (eds.),Ancient marine reptiles. Academic Press, San Diego.RIEPPEL, O. 1998. Corosaurus alcovensis Case, and the phylogenetic interrelationships of Triassic stem-groupSauropterygia. Zoological Journal of the Linnean Society, 124:1–41.188

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONRIEPPEL, O. 1999. Phylogeny and paleobiogeography of Triassic Sauropterygia: problems solved and unresolved.Palaeogeography, Palaeoclimatology, and Palaeoecology, 153:1–15.RIEPPEL, O. 2002. Feeding mechanics in Triassic stem-group sauropterygians: the anatomy of a successful invasionof Mesozoic seas. Zoological Journal of the Linnean Society, 135:33–63.ROBBA, E., AND F. OSTINELLI. 1975. Studi paleontologici sul Pliocene Ligure, I: Testimonmianze di predazione suimolluschi Pliocenici di Albenga. Rivista Italiana Paleontologia, 81:309–318.ROBISON, R. A., AND R. L. KAESLER. 1987. Phylum Arthropoda, p. 205–267. In R. S. Boardman, A. H. Cheetham,and A. J. Rowell (eds.), Fossil Invertebrates. Blackwell Scientific Publications, Palo Alto, Oxford, London.ROSEN, B. R. 1988. Progress, problems, and patterns in the biogeography of reef corals and other tropical marineorganisms. Helgoland Wissenschaftlichen Meeresuntersuchen, 42:269–301.ROSS, A., J. E. LEWIS, AND R. J. SCOLARO. 1964. New Eocene decapods from Florida. Quarterly Journal of theFlorida Academy of Sciences, 27:187–196.ROTHSCHILD, B. M., AND L. D. MARTIN. 1990. Shark-induced infectious spondylitis; evidence in the Cretaceousrecord. Abstracts of Papers; 5th annual meeting, Society of Vertebrate Paleontology, 10 (3):39A.ROY, K., D. J. MILLER, AND M. LABARBERA. 1994. Taphonomic bias in analyses of drilling predation: effects ofgastropod drill holes on bivalve shell strength. Palaios, 9:413–421.RUDWICK, M. J. S. 1965. Sensory spines in the Jurassic brachiopod Acanthothirus. Palaeontology, 8:604–617.RUDWICK, M. J. S. 1970. Living and Fossil Brachiopods. Hutchinson University Library, London, 199 p.RUSSO, R. A. 1975. Observations on the food habits of leopard sharks (Triakis semifasciata) and brownsmoothhounds (Mustelus henlei). California Department of Fish and Game, 61:95–103.SANDER, P. M. 1989. The pachypleurosaurids (Reptilia: Nothosauria) from the Middle Triassic of Monte SanGiorgio (Switzerland), with the description of a new species. Philosophical Transactions of the Royal Societyof London, B, 325:561–670.SANDER, P. M. 1997. The paleobiogeography of Shastasaurus, p. 17–43. In J. Callaway and E. L. Nicholls (eds.),Ancient Marine Reptiles. Academic Press, San Diego.SATO, T., AND K. TANABE. 1998. Cretaceous plesiosaurs ate ammonites. Nature, 394:629–630.SAUNDERS, W. B., AND P. D. WARD. 1987. Ecology, distribution, and population characteristics of Nautilus, p.137–162. In W. B. Saunders and N. H. Landman (eds.), Nautilus: The Biology and Paleobiology of a LivingFossil. Plenum Press, New York.SAUNDERS, W. B., R. L. KNIGHT, AND P. N. BOND. 1991. Octopus predation on Nautilus: Evidence from Papua NewGuinea. Bulletin of Marine Science, 49:280–287.SAVAGE, R. J. 1976. Review of the early Sirenia. Systematic Zoology, 25:344–351.SCHÄFER, W. 1972. Ecology and Paleoecology of Marine Environments. University of Chicago Press, Chicago, 568 p.SCHINDEL, D. E., G. J. VERMEIJ, AND E. ZIPSER. 1982. Frequencies of repaired shell fractures among the Pennsylvaniangastropods of north-central Texas. Journal of Paleontology, 56:729–740.SCHINDEWOLF, O. H. 1958. Über Aptychen (Ammonoidea). Palaeontolographica, A, 111:1–46.SCHNEIDER, J. A. 1988. Frequency of arm regeneration of comatulids in relation to life habit, p. 97–102. In R. D.Burke, P. V. Mladenov, P. Lambert, and R. L. Parsley (eds.), Proceedings of the 6 th international EchinodermConference, Victoria, British Columbia.SCHNEIDER, J. A., AND J. G. CARTER. 2001. Evolution and phylogenetic significance of Cardioidean shellmicrostructure (Mollusca: Bivalvia). Journal of Paleontology, 75:607–643.SCHOENER, T. W. 1979. Inferring properties of predation and other injury-producing agents from injury frequencies.Ecology, 60:1110–1115.SCHRAM, F. R.1969. Polyphyly in Eumalacostraca? Crustaceana, 16:243–350.SCHRAM, F. R. 1979. The genus Archaeocaris, and a general review of the Palaeostomatopoda (Hoplocarida:Malacostraca). Transactions of the San Diego Society of Natural History, 19:57–66.SCHRAM, F. R. 1986. Crustacea. Oxford University Press, New York, 606 p.SCHWIMMER, D. R., J. D. STEWART, AND G. D. WILLIAMS. 1997. Scavenging by sharks of the genus Squalicorax inthe Late Cretaceous of North America. Palaios, 12:71–83.SEILACHER, A. 1985. Bivalve morphology and function, p. 88–101. In D. J. Bottjer, C. S. Hickman, and P. D. Ward (eds.),Mollusks: Notes for a Short Course. University of Tennessee Department of Geological Sciences Studies in Geology 13.189

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002SEILACHER, A. 1998. Mosasaur, limpets or diagenesis; how Placenticeras shells got punctured. Mitteilungen ausdem Museum für Naturkunde in Berlin, Geowissenschaftliche Reiche, 1:93–102.SHELDON, A. 1997. Ecological implications of mosasaur bone microstructure, p. 333–354. In J. Callaway and E.L. Nicholls (eds.), Ancient Marine Reptiles. Academic Press, San Diego.SHIMADA, K. 1997. Paleoecological relationships of the Late Cretaceous lamniform shark, Cretoxyrhina mantelli(Agassiz). Journal of Paleontology, 71 (5):926–933.SHOUP, J. B. 1968. Shell opening by crabs of the genus Calappa. Science, 160:887–888.SIMENSTAD, C. A., J. A. ESTES, AND K. L. KENYEN. 1978. Aleuts, sea otters, and alternative stable state communities.Science, 200:403–411.SIMMS, M. J. 1990. The radiation of post-Paleozoic echinoderms, p. 287–304. In P. D. Taylor and G. P. Larwood(eds.), Major Evolutionary Radiations. Clarendon Press, Oxford.SIMPSON, G. G. 1975. Fossil penguins, p. 19–41. In B. Stonehouse (ed.), The Biology of Penguins. MacMillan, London.SKELTON, P. W., J. A. CRAME, N. J. MORRIS, AND E. M. HARPER. 1990. Adaptive divergence and taxonomic radiationin post-Palaeozoic bivalves, p. 91–117. In P. D. Taylor and G. P. Larwood (eds.), Major EvolutionaryRadiations. The Systematics Association, Special Volume 42, Clarendon Press, Oxford.SLITER, W. V. 1971. Predation on benthic foraminifers. Journal of Foraminferal Research, 1:20–29.SMITH, A. 1952. Shells from the bird guano of southeast Farallon Island, California, with description of a newspecis of Liotia. Proceedings of the California Academy of Sciences, 27:383–387.SMITH, B. G. 1942. The heterodontid sharks: their natural history and the external development of Heterodontusjaponicaus based on notes and drawings by Bashford Dean. The Bashford Dean Memorial Volume: ArchaicFishes. American Museum of Natural History, 8:649–770.SOHL, N. F. 1969. The fossil record of shell boring by snails, p. 725–734. In M. R. Carriker, et al. (eds.), Penetrationof calcium carbonate substrates by lower plants and invertebrates. American Zoologist, 9.SPANIER, E. 1986. Cannibalism in muricid snails as a possible explanation for archaeological findings. Journal ofArchaeological Science, 13:463–468.SPANIER, E. 1987. A fossil record of shell boring: possible evidence for sea level changes in the Red Sea. Estuarine,Coastal and Shelf Science, 24:873–879.SPEDEN, I. A. 1971. Notes on New Zealand fossil mollusca, 2: Predation on New Zealand Cretaceous species ofInoceramus (Bivalvia). New Zealand Journal of Geology and Geophysics, 14:56–60.SPEED, E. 1969. Prehistoric shell collectors. South African Archaeological Bulletin, 24:193–196.STANLEY, S. M. 1970. Relation of shell form to life habits of the Bivalvia (Mollusca). Memoir of the GeologicalSociety of America, 125:1–296.STANLEY, S. M. 1977. Rates, trends, and patterns of evolution in the Bivalvia, p. 209–250. In A. Hallam (ed.),Patterns of Evolution as Illustrated by the Fossil Record. Elsevier, Amsterdam.STEHLIK, L. L. 1993. Diets of the brachyuran crabs Cancer irroratus, C. borealis, and Ovalipes ocellatus in theNew York Bight. Journal of Crustacean Biology, 13:723–735.STEWART, J. D., AND K. CARPENTER. 1990. Examples of vertebrate predation on cephalopods in the Late Cretaceousof the Western Interior, p. 203–207. In A. J. Boucot (ed.), Evolutionary Paleobiology of Behavior andCoevolution. Elsevier, Amsterdam.STONE, H. M. I. 1998. On predator deterrence by pronounced shell ornament in epifaunal bivalves. Palaeontology,41:1051–1068.STORRS, G. W. 1993. Function and phylogeny in sauropterygian (Diapsida) evolution. American Journal of Science,293A:63–90.TARLO, L. B. 1959. Pliosaurus brachyspondylus from the Kimmeridge Clay. Palaeontology, 1:283–291.TASCH, P. 1980. Paleobiology of the Invertebrates: Data Retrieval from the Fossil Record. John Wiley, New York, 975 p.TAYLOR, J. D. 1970. Feeding habits of predatory gastropods in a Tertiary (Eocene) molluscan assemblage fromthe Paris Basin. Palaeontology, 13:254–260.TAYLOR, J. D. 1981. The evolution of predators in the Late Cretaceous and their ecological significance, p. 229–240. InP. L. Forey (ed.), The Evolving Biosphere. British Museum (Natural History) and Cambridge University Press.TAYLOR, J. D., R. D. CLEEVELY, AND C. N. TAYLOR. 1980. Food specialization and the evolution of predatoryprosobranch gastropods. Palaeontology, 23:375–409.190

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONTAYLOR, J. D., R. D. CLEEVELY, AND N. J. MORRIS. 1983. Predatory gastropods and their activities in the BlackdownGreensand (Albian) of England. Palaeontology, 26:521–553.TAYLOR, J. D., W. J. KENNEDY, AND A. HALL. 1969. The shell structure and mineralogy of the bivalvia—Introduction,Nuculacea-Trigonacea. Bulletin of the British Museum of Natural History, Zoology, 3:1–125.TAYLOR, M. A. 1992. Functional anatomy of the head of the large aquatic predator Rhomaleosaurus zetlandicus(Plesiosauria, Reptilia) form the Toarcian (Lower Jurassic) of Yorkshire, England. Philosophical Transactionsof the Royal Society of London, B, 335:247–280.TAYLOR, M. A. 1994. Stone, bone, or blubber? Buoyancy control strategies in aquatic tetrapods, p. 151–161. In L.Maddock, Q. Bone, and J. M. V. Rayner (eds.), Mechanics and Physiology of Animal Swimming. CambridgeUniversity Press, Cambridge, England.TAYLOR, M. A. 2000. Functional significance of bone ballast in the evolution of buoyancy control strategies byaquatic tetrapods. Historical Biology, 14:15–31.TAYLOR, M. A., AND A. R. I. CRUICKSHANK. 1993. Cranial anatomy and functional morphology of Pliosaurusbrachyspondylus (Reptilia: Plesiosauria) from the Upper Jurassic of Westbury, Wiltshire. PhilosophicalTransactions of the Royal Society of London, B, 341:399–418.TEICHERT, C., AND D. L. SERVENTY. 1947. Deposits of shells transported by birds. American Journal of Science,245:322–328.TEICHERT, C., AND T. MATSUMOTO. 1987. The ancestry of the genus Nautilus, p. 25–32. In W. B. Saunders and N. H.Landman (eds.), Nautilus: The Biology and Paleobiology of a Living Fossil. Plenum Press, New York.THAYER, C. W. 1983. Sediment-mediated biological disturbance and the evolution of the marine benthos, p. 479–595. In M. J. S. Tevesz and P. J. McCall (eds.), Biotic Interactions in Recent and Fossil Benthic Communities.Plenum Press, New York.THEIS, D., AND W. E. REIF. 1985. Phylogeny and evolutionary ecology of Mesozoic Neoselachii. Neues Jahrbuchfur Geologie und Palaeontologie Abhandlungen, 169:331–369.THEWISSEN, J. G. M., E. M. WILLIAMS, L. J. ROE, AND S. T. HUSSAIN. 2001. Skeletons of terrestrial cetaceans and therelationship of whales to artiodactyls. Nature, 413:277–281.TOMAS, J. F., J. AZNAR, AND J. A. RAGA. 2001. Feeding ecology of the loggerhead turtle Caretta caretta in thewestern Mediterranean. Journal of Zoology (London), 255(4):525–532.TREWIN, N. H., AND W. WELSH. 1972. Transport, breakage, and sorting of the bivalve Mactra corallina on AberdeenBeach, Scotland. Palaeogeography, Palaeoclimatology, Palaeoecology, 12:193–204.TSCHANZ, K. 1989. Lariosaurus buzzii n. sp. from the Middle Jurassic of Monte San Giorgio (Switzerland), withcomments on the classification of nothosaurs. Palaeontographica, A208:153–179.TSUJITA, C. J., AND G. E. G. WESTERMANN. 2001. Were limpets or mosasaurs responsible for the perforations in theammonite Placenticeras? Palaeogeography, Palaeoclimatology, Palaeoecology, 169:245–270.URRUTIA, G. X., AND J. M. NAVARRO. 2001. Patterns of shell penetration by Chorus giganteus juveniles (Gastropoda:Muricidae) on the mussel Semimytilus algosus. Journal of Experimental Marine Biology and Ecology, 258:141–153.VALENTINE, J. W. 1973. Evolutionary Paleoecology of the Marine Biosphere. Prentice-Hall, Englewood Cliffs,New Jersey, 511 p.VALE, F. K., AND M. A. REX. 1988. Repaired shell damage in deep-sea prosobranch gastropods from the westernNorth Atlantic. Malacologia, 28:65–79.VALE, F. K., AND M. A. REX. 1989. Repaired shell damage in a complex of rissoid gastropods from the uppercontinental slope south of New England. The Nautilus, 103:105–108.VANCE, R. R. 1978. A mutualistic interaction between a sessile marine clam and its epibionts. Ecology, 59:679–685.VANNIER, J., K. ABE, AND K. IKUTA. 1998. Feeding in myodocopid ostracods: functional morphology and laboratoryobservations from videos. Marine Biology, 132:391–408.VAN BLARICOM, G. R., AND J. A. ESTES. 1988. The Community Ecology of Sea Otters. Ecological Studies, 65,Springer-Verlag, Berlin, 247 p.VAN ZYL, R. F., S. MAYFIELD, A. PULFRICH, AND C. L. GRIFFITHS. 1998. Predation by West Coast rock lobsters (Jasus lalandii)on two species of winkle (Oxystele sinensis and Turbo cidaris). South African Journal of Zoology, 33:203–209.VERMEIJ, G. J. 1977. The Mesozoic marine revolution: Evidence from snails, predators, and grazers. Paleobiology,3:245–258.191

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002VERMEIJ, G. J. 1978. Biogeography and Adaptation, Patterns in Marine Life. Harvard University Press, Cambridge, 332 p.VERMEIJ, G. J. 1982. Gastropod shell form, breakage, and repair in relation to predation by the crab Calappa.Malacologica, 23:1–12.VERMEIJ, G. J. 1983. Shell breaking predation through time, p. 649–669. In M. J. S. Tevesz and P. L. McCall(eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum Press, New York.VERMEIJ, G. J. 1987. Evolution and Escalation. Princeton University Press, Princeton, NJ, 527 p.VERMEIJ, G. J. 1996. Marine biological diversity: muricid gastropods as a case study, p. 355–375. In D. I. Jablonski,D. H. Erwin, and J. H. Lipps (eds.), Evolutionary Paleobiology. Chicago University Press, Chicago.VERMEIJ, G. J., AND S. CARLSON. 2000. The muricid gastropod subfamily Rapaninae: phylogeny and ecologicalhistory. Paleobiology, 26:19–46.VERMEIJ, G. J., D. SCHINDEL, AND E. ZIPSER. 1981. Predation through geological time: evidence from gastropodshell repair. Science, 214:1024–1026.VERMEIJ, G. J., E. ZIPSER, AND E. C. DUDLEY. 1980. Predation in time and space: Peeling and drilling in terebridgastropods. Paleobiology, 6:352–364.VERMEIJ, G. J., E. ZIPSER, AND R. ZARDINI. 1982. Breakage-induced shell repair in some gastropods from the UpperTriassic of Italy. Journal of Paleontology, 56:233–235.VOIHL, G. 1990. Piscivorous fishes of the Solnhofen Limestone, p. 287–303. In A. J. Boucot ( ed.), EvolutionaryPaleobiology of Behavior and Coevolution. Elsevier, Amsterdam.VOKES, E. H. 1971. The geologic history of the Muricinae and the Ocenebrinae. Echo, 4:37–54.VOKES, E. H. 1990. Cenozoic Muricidae of the western Atlantic region, Part VIII—Murex s.s., Haustellum,Chicoreus, and Hexaplex; additions and corrections. Tulane Studies in Geology and Paleontology, 23:1–96.VORIS, H. K., AND H. H. VORIS. 1983. Feeding strategies in marine snakes: an analysis of evolutionary,morphological, behavioral and ecological relationships. American Zoologist, 23:411–425.WAINWRIGHT, S. A., W. D. BIGGS, J. D. CURREY, AND J. M. GOSLINE. 1982. Mechanical design in organisms. PrincetonUniversity Press, Princeton, New Jersey, 423 p.WALKER, S. E. 1990. Biological taphonomy and gastropod temporal dynamics, p. 391–421. In W. Miller III (ed.),Paleocommunity Temporal Dynamics: The Long-term Development of Multispecies Assemblies. ThePaleontological Society Special Publication, 5.WALKER, S. E. 1991. Taphonomy and paleoecology of Villamil fossil megagastropods of Isla Isabella, p. 423–437. In M. J. James (ed.), Galápagos Marine Invertebrates. Plenum Press, New York.WALKER, S. E. 2001. Palaeoecology of gastropods preserved in turbiditic slope deposits from the Upper Plioceneof Ecuador. Palaeogeography, Palaeoclimatology, Palaeoecology, 166:141–163.WALKER, S. E., K. PARSONS-HUBBARD, E. POWELL, AND C. E. BRETT. 2002. Predation on experimentally deployedmolluscan shells from shelf to slope depths in a tropical carbonate environment. Palaios, 17:147–170.WALKER, S. E., AND J. VOIGHT. 1994. Paleoecologic and taphonomic potential of deep sea gastropods. Palaios, 9:48–59.WALTER, R. C., R. T. BUFFLER, J. H. BRUGGEMANN, M. M. GUILLAUME, S. M. BERHE, B. NEGASSI, Y. LIBSEKAL, H.CHENG, R. L. EDWARDS, R. VON CASEL, D. NERAUDEAU, AND M. GAGNON. 2000. Early human occupation of theRed Sea coast of Eritrea during the last interglacial. Nature, 405:65–69.WARD, D. J., AND N. T. J. HOLLINGWORTH. 1990. The first record of a bitten ammonite from the Middle OxfordClay (Callovian, Middle Jurassic) of Bletchley, Buckinghamshire. Mesozoic Research, 2:153–161.WARD, P. D. 1981. Shell sculpture as a defensive adaptation in ammonoids. Paleobiology, 7:96–100.WARD, P. D. 1986. Cretaceous ammonite shell shapes. Malacologia, 27:3–28.WARD, P. D.1987. The Natural History of Nautilus. Allen & Unwin, Boston, 267 p.WARD, P. D., AND P. W. SIGNOR III. 1983. Evolutionary tempo in Jurassic and Cretaceous ammonites. Paleobiology,9:183–198.WEEMS, R. E. 1988. Paleocene turtles from the Aquia and Brightseat formations, with a discussion of their bearingon sea turtle evolution and phylogeny. Proceedings of the Biological Society of Washington, 101(1):109–145.WELLES, S. P. 1943. Elasmosaurid plesiosaurs with description of new material from California and Colorado.Memoirs of the University of California, 13:125–254.WESTERMANN, G. E. G. 1996. Ammonoid life and habitat. In N. H. Landmam, K. Tanabe, and R. A. Davis (eds.),Ammonoid Paleobiology. Plenum Press, New York.192

WALKER AND BRETT—POST-PALEOZOIC PATTERNS IN MARINE PREDATIONWESTPHAL, F. 1988. Pflasterzahnsaurier (Placodonten) aus dem süddeutschen Muschelkalk (Mitteltrias), p. 151–165. In H. Hagdorn (ed.), Neue Forschungen zur Erdgeschichte von Crailsheim. Goldschneck, Korb (Stuttgart).WETZEL, W. 1960. Nachtrag zum Fossilarchiv der Fuiriquina-Schichten. Neues Jahrbuch für Geologie undPaläontologie, Monatshefte, 10:439–446.WHITMORE, F. C., JR., AND A. E. SANDERS. 1976. Review of the Oligocene cetacea. Systematic Zoology, 25:304–320.WIEDMAN, L. A., R. L. FELDMANN, D. E. LEE, AND W. J. ZINSMEISTER. 1988. Brachiopoda from the La Meseta Formation(Eocene), Seymour Island, Antarctica, p. 531–539. In R. M. Feldmann and M. O. Woodburn (eds.), Geologyand Paleontology of the Seymour Island, Antarctic Penninsula. Geological Society of America Memoir 169.WILBER, K. M., AND K. SIMKISS. 1968. Calcified shells, p. 229–295. In M. Florkin and E. H. Stotz (eds.), ComprehensiveBiochemistry, Volume 26, Part A, Extracellular and Supporting Structures. Elsevier, Amsterdam.WILGA, C. D., AND P. J. MOTTA. 2000. Durophagy in sharks: feeding mechanics of the Hammerhead Sphyra tiburo.The Journal of Experimental Biology, 203:2781–2796.WILLIAMS, A. 1996. Taxonomy and evolution, p. 13–21. In J. R. Factor (ed.), Biology of the Lobster, Homerusamericanus. Academic Press, San Diego.WILLISTON, S. W., AND R. L. MOODIE. 1917. Ogmodirus martinii, a new plesiosaur from the Cretaceous of Kansas.The Kansas University Science Bulletin, 10:61–73.WILSON, M. A., AND T. J. PALMER. 1990. A review of evolutionary trends in carbonate hardground communities.The Paleontological Society Special Publication, 5:137–152.WILSON, M. A., AND T. J. PALMER. 1992. Hardgrounds and Hardground Faunas. University of Wales, AberystwythInstitute of Earth Studies Publication, 9:1–131.WODINSKY, J. 1969. Penetration of the shell and feeding on gastropods, p. 997–1010. In M. R. Carriker et al.(eds.), Penetration of calcium carbonate substrates by lower plants and invertebrates. American Zoologist, 9.WOELKE, C. E. 1957. The flatworm Pseudostylochus ostreophagus Hyman, a predator of oysters. National ShellfishAssociation Proceedings, 47:62–67.WOODCOCK, T. C., AND P. H. KELLEY. 2001. Predation on irregular echinoids (Eocene) from the Castle HayneLimestone, Atlantic Coastal Plain (southeastern North Carolina). Geological Society of America, SoutheasternSection, Abstracts with Programs, 33:15–16.YAMAGUCHI, A., AND T. TANIUCHI. 2000. Food variations and ontogenetic dietary shift of starspotted-dogfish Mustelusmanazo at five locations in Japan and Taiwan. Fisheries Science, 66:1039–1048.ZINSMEISTER, W. J. 1980. Observations on the predation of the clypeasteroid echinoid, Monophoraster darwin,from the Upper Miocene Enterrios Formation, Patagonia, Argentina. Journal of Paleontology, 54:910–912.193


BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESFOSSIL RECORD OF PARASITISM ON MARINEINVERTEBRATES WITH SPECIAL EMPHASIS ON THEPLATYCERATID-CRINOID INTERACTIONTOMASZ K. BAUMILLER AND FOREST J. GAHNMuseum of Paleontology, University of Michigan, Ann Arbor, Michigan 48109-1079 USAABSTRACT—The paleontological literature on marine invertebrates is rich in supposed examples of parasitismand our tabulation shows a nearly even distribution of reported cases through the post-Cambrian Phanerozoic.Slightly lower frequencies characterize the Triassic and Jurassic and higher frequencies the Cretaceous and Tertiary,and the pattern roughly mirrors Sepkoski’s (1984) marine diversity curve. The total number of parasitic associationsfor any geologic period rarely exceeds a dozen, yet few of the reported examples provide explicit criteria distinguishingparasitism from predation, commensalism, or mutualism. We evaluated the published examples using the followingcriteria: (1) evidence of a long-term relationship between two organisms, (2) benefit of interaction to supposedparasite, and (3) detriment of interaction to the host. We found that only in exceptional cases were these criteriafulfilled. One example that provides much information on parasitic interactions involves platyceratids and crinoidsand we summarize the evidence for the parasitic interaction between these two groups of organisms.INTRODUCTIONBIOTIC INTERACTIONS can be representedin terms of the ecological and/or evolutionaryconsequences that they have on the interactingorganisms. From an ecological perspective,interactions may be considered to be positive (+),neutral (0), or negative (-). In the case of twointeracting organisms, there exist six possiblecombinations: amensalism (0,-), competition (-,-),mutualism (+,+), commensalism (+,0), exploitation(+,-), and toleration (0,0) (Clarke, 1954). Two typesof exploitative behavior are recognized: predationand parasitism. Both are interesting ecologically andevolutionarily because they may lead to reciprocalevolution, or an “arms race,” between predator andprey, or host and parasite. However, the differencebetween parasitism and predation is one of degreerather than kind. One difference is that whilepredators typically kill their prey, parasites mightkill their hosts, but not without first making use oftheir living victims for an extended period. As afirst step in recognizing parasitism it is thereforenecessary, though not sufficient, to establish that anassociation involved a long-term relationshipbetween two organisms. It must further bedemonstrated that the interaction benefited theinfester and was detrimental to the host. It isunderstood that we must know the identity of thehost, and, if we are to explicitly demonstrate thebenefit of the interaction to the parasite, we shouldknow its identity as well. Not surprisingly, forpaleontologists, “…recognition of a fossilassociation as being parasitic may not be easy”(Conway Morris, 1981, p. 491).FOSSIL RECORD OF PARASITESON MARINE INVERTEBRATES:A SURVEYMethods.—We reviewed the literature toexplore the fossil record of parasites on marineinvertebrates. We were interested not only indiscerning temporal patterns of inferred parasitism,but in the type and quality of data that were used toestablish the nature of the relationship. Our list maynot be exhaustive, but it should not be biased in anysystematic way. In collecting the data we relied195

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 1—Fossil record of parasites on marine invertebrates.Host ParasiteCategoryParasitebenefitHostdetrimentEvidence for durationof relationshipFirstoccurrenceLastoccurrenceReferencesarthropod fungi 3 nutrients ? fungal mass in coprolites Silurian Tertiary Taylor & Osborn, 1996belemnite acrothoracid barnacle 3 ? ? teardrop-shaped boring Jurassic E. Cretaceous Seilacher, 1968bivalve capulid gastropod 1 ? ? attachment, scars, borings L. Cretaceous Tertiary Hayami & Kanie, 1980bivalve cestode or trematode 3 ? ? pearls Triassic Tertiary Herdman, 1906bivalve Clionid sponge 3 ? ? borings Cretaceous Tertiary Barry et al., 1972bivalve polychaete or sipunculid 2 ? ? tube-like borings Tertiary Tertiary Savazzi, 1995bivalve trematode 3 ? ? pits TertiaryTertiary Boucot, 1990(venerid)(Miocene)blastoid platyceratid gastropod 1 nutrients nutrients attached, scars, borings Devonian Mississippian Baumiller, 1996brachiopod acrothoracid barnacle 2 ? ? elongate borings L. Devonian Pennsylvanian Rodriguez &Gutschick, 1977brachiopod polychaete orgastropod3 ? ? borings M. Ordovician Tertiary Baumiller et al., 1999;Leighton, 2001brachiopod uncertain 3 ? ? tube secretions Devonian Devonian MacKinnon & Biernat,1970; Chatterton, 1975brachiopod uncertain 3 ? ? calcareous partitions L. Silurian L. Silurian Boucot &McCutcheon, 1986bryozoan algae and fungi 3 ? ? microstructure Ordovician Tertiary Elias, 1966bryozoan hydroid/3 ? ? embedment structures LateLatePalmer & Wilson, 1988colonian ascidianOrdovician Ordovicianbryozoan serpulid polychaete 3 ? ? tube secretions UpperCretaceousU. Cretaceous Ehrhard Voight, 1955chitinozoa bacteria and fungi 3 ? ? borings Ordovician Mississippian Grahn, 1981;Sutherland, 1994conularid uncertain 3 ? ? pearls Pennsylvanian Pennsylvanian Babcock, 1990crinoid myzostomid annelid 2 ? ? galls, pits Ordovician Tertiary Welch, 1976;Brett, 1978crinoid platyceratid gastropod 1 nutrients nutrients attached, scars, borings M. Ordovician Permian Rollins &Brezinski, 1988crinoid uncertain 3 ? ? circular pits M. Ordovician Permian Brett, 1985cystoid platyceratid gastropod 1 nutrients nutrients attached, scars, borings M. Silurian M. Silurian Boucot, 1990decapod bopyrid isopod 2 nutrients soft tissues carapace deformation L. Jurassic Tertiary Glaessner, 1969decapod rhizocephalan 2 nutrients sex organs claw feminization/castration L. Jurassic? Tertiary Boucot, 1990echinoid barnacle 2 ? ? elongate borings L. Cretaceous Tertiary Madsen & Wolff, 1965;Marouf, 1999echinoid copepod 2 ? ? internal galls Jurassic Tertiary Solovjev, 1961; Mercier,1937; Marouf, 1999echinoid eumelid gastropod 2 nutrients soft tissues borings Cretaceous Tertiary Kier, 1981; Alekseev &Endelman, 1989echinoid gastropod 1 ? ? deformed spines Cretaceous Tertiary Tasnádi-Kubacska, 1962echinoid myzostomid annelid 2 ? ? galls Tertiary (Miocene) Tertiary Roman, 1952echinoid prosobranch gastropod 2 ? ? skeletal malformations E. Cretaceous Tertiary Boucot, 1990gastropod acrothoracid barnacle 2 ? ? elongate borings E. Devonian Pennsylvanian Baird et al, 1990(platyceratid)graptolite uncertain 3 ? ? closed blisters E. Ordovician L. Silurian Bates and Loydell, 2000graptolite uncertain 3 ? ? open-ended tubes E. Ordovician L. Silurian Bates and Loydell, 2000graptolite uncertain 3 ? ? tubothecae E. Ordovician E. Ordovician Conway Morris, 1981196

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESheavily on two important reviews of parasitism:Conway Morris’ (1981) “Parasites and the fossilrecord,” and Boucot’s (1990) “Evolutionarypaleobiology of behavior and coevolution.”We evaluated the published examples in termsof several criteria (Table 1). In considering theidentity of the parasite, if the inferred parasite wasdirectly observable in association with the host, itwas assigned to “category 1.” If the identity of theparasite was inferred from a well-constrainedproxy, namely the trace fossil that it produced, weassigned it to “category 2.” In all other instances,the identity of the parasite was considered morespeculative and assigned to “category 3.”We also considered the evidence presented forthe benefit that was gained by the parasite and forthe detrimental effects on the host. Finally, dataused to infer a long-term duration of the associationwere noted.Results.—Our search produced a substantialnumber of inferred cases of parasitism (Table 1),and, as Figure 1 illustrates, reports of parasiticassociation are quite evenly distributed through thepost-Cambrian Phanerozoic. Although slightlylower frequencies characterize the Triassic andJurassic and higher frequencies the Cretaceous andTertiary, the coarse temporal and taxonomicresolution prevents us from assigning muchsignificance to these differences, though it is worthnoting that the pattern roughly mirrors Sepkoski’s(1984) marine diversity curve. The total numberof parasitic associations for any geologic periodrarely exceeds a dozen, and if we consider thoseassociations for which the identity of the parasiteis well-constrained, based on co-occurrence orcharacteristic trace fossils (categories 1 and 2 inFig. 1), the number is lower still.To illustrate our scheme, we consider thebivalve-capulid gastropod association reported byHayami and Kanie (1980). These authors reportedspecimens of Cretaceous capulid gastropodsattached to the valves of inoceramids. Since theidentity of the interacting organisms is known, thisrepresents “category 1” in our scheme. Hayami andKanie (1980) inferred that this interactionrepresented parasitism because it was ecologicallylong-lived, as evidenced by the presence ofattached specimens, and was analogous to themodern capulids that are found in association withbivalves, gastropods, brachiopods, and annelids.Extant capulids are suspension feeders and can befacultative semi-parasites or parasites: they canFIGURE 1—Temporal distribution of reported parasites on marine invertebrates. Black, stippled, andwhite bars represent categories 1, 2, and 3, respectively. Categories 1, 2, and 3 explained in text. (Datafrom Table 1.)197

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002feed from the inhalant current created by the ciliaryactivity of the host (Thorson, 1965), or by usingtheir pseudoproboscis to divert particles capturedby the host to its own mouth (Pernet and Kohn,1998). Thus, in our scheme, “nutrients” benefit theparasite, and their loss is a detriment to the host.As Table 1 indicates, examples of parasitismsuch as the bivalve-capulid association, where theidentity of the interacting organisms and the natureof the benefit/detriment can be reliably assessed, arerare in the fossil record. Most common are instanceswhere a trace fossil, or a growth abnormality, isinterpreted as having been parasite-induced, wherethe identity of the parasite cannot be ascertained,and where the beneficial effects on the parasite andthe detrimental effects on the host are poorlyconstrained. The case of the Pliocene bivalve,Isognomon, serves as an illustration of such parasiticinterpretations. Savazzi (1995) reported sevendifferent types of anomalies in the hinge region of14 of 24 specimens of Isognomon maxillatus fromthe Upper Pliocene of Italy. The size and shape ofsome of the cavities suggested that they wereproduced by a worm-like organism. According toSavazzi (1995, p. 136), “…a polychaete could be areasonable candidate,” and a “…sipunculid couldalso be a possible candidate.” However, because theobserved teratologies, or malformations, find noexact modern analogs, we categorize the identityof the parasite as “inferred” (category 3).The Isognomon example also illustrates theproblems in assessing the benefit of an associationto the parasite and its detrimental effects on thehost. The benefit to the infesting organism mayseem obvious because the host is at least providingthe infester with a suitable life position orprotection, but unless the infester’s effect on thehost can be shown to be detrimental, such aninteraction could represent mutualism orcommensalism. For Isognomon, Savazzi (1995, p.137) chose parasitism, which he claimed was“much more likely,” because the position of someof the anomalies would have made it difficult forthe infesting organism to access sea water directlywhile allowing it to tap into the digestive systemor the hemocoele of the bivalve. In our scheme,we leave the question of benefit/detriment forIsognomon as undetermined (“?”), in part becauseSavazzi (1995) did not specify whether it involvednutrients, tissues, or something else that wasbenefiting the parasite and causing harm to the host.Although one might infer that the parasite wasstealing nutrients from the host, we consideredthose fossil examples for which no modern analogfor parasitism had been presented, and no explicittest was used to reject commensalism ormutualism, as “undetermined.”In our classification scheme, the vast majorityof inferred cases of parasitism are “undetermined”with regard to benefit/detriment (Table 1). Ingeneral, the inference of parasitism in these casesis based on accepting “reaction” features asdetrimental to the host. Reaction features, such aspearls, galls, blisters, and deformations, do indicatethat the host responded biotically to the infester,but whether such a response placed the infestedindividual at a selective disadvantage relative touninfested conspecifics has rarely been explored.Of course, it is plausible to assume that theenergetic costs associated with the “reaction” mayplace infested individuals at some disadvantage,but such costs may be quantitatively, and thusselectively, trivial. Thus, even the swollen,distorted, and excavated stems of crinoids causedby the reaction to some infester that have beenclaimed as cases of “true parasitism” (Pickett, 1973,p. 342), remain enigmatic and may be moreappropriately treated as commensal (Franzen,1974; Brett, 1978, 1985; Baird et al., 1990).Many of the examples in Table 1 illustrate theproblems faced by paleontologists in distinguishingbetween parasitism, commensalism, andmutualism—problems that have long beenrecognized (Conway Morris, 1981; Savazzi, 1995).Distinguishing between predation and parasitismcan prove equally difficult. Complete boreholes inthe tests of fossil organisms have generally beeninterpreted as predatory, especially whenmorphologically similar holes are produced byextant predators, such as muricid or naticidgastropods. However, when such boreholes precedethe known first appearance of modern predatory198

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESdrilling gastropods, and the causative organismcannot be determined, their interpretation has beenless certain. This has been especially true forboreholes in Paleozoic invertebrates (Fisher, 1962;Carriker and Yochelson 1968; Sohl, 1969). For manyof these boreholes, parasitism has been considereda plausible hypothesis (Buehler, 1969; Ausich andGurrola, 1979; Conway Morris and Bengtson,1994). A parasitic interpretation of completePaleozoic boreholes has become even more tenableafter it was shown that platyceratid gastropods werecapable of drilling echinoderms (Baumiller, 1990,see below), and Table 1 includes several examplesof inferred parasitism by platyceratid gastropods.The long range of platyceratids (Ordovician toPermian), their abundance, occurrence in a broadrange of marine settings, known association with avariety of taxa (crinoids, blastoids, cystoids), andsedentary nature implies that they should beconsidered as the “null-hypothesis” when seekingthe culprit of complete Paleozoic boreholes, suchas those in brachiopods (Baumiller et al., 1999;Kowalewski et al., 2000; Leighton, 2001). Giventhe abundance and range of platyceratids, newlypublished and yet unpublished data relevant to theirlife habit, and the fact that they are our favoriteexample, we will present a review of the data andtheir status as parasites.PLATYCERATIDS ANDCRINOIDSRecord of interaction.—One of the classicexamples of biotic interactions in the fossil recordis that between platyceratid gastropods and crinoids(Fig. 2). The consistent occurrence of platyceratidgastropods preserved attached to the calyxes ofcrinoids was noted by mid-nineteenth centuryFIGURE 2—Examples of platyceratid-crinoid association. All specimens from Middle Devonian strata. Scalebar = 0.5 cm. 1, 3, Arthroacantha carpenteri, UMMP 23915, Arkona Shale, Thedford, Ontario. 2, Corocrinuscalypso, UMMP 24170, Arkona Shale, Arkona, Ontario. 4, Gennaeocrinus variabilis, 224G, S. Virgilis personalcollection, Bell Shale, Rockport, Michigan. 5, Arthroacantha carpenteri, K. Dobson personal collection,Silica Shale, Sylvania Ohio. 6, Corocrinus calypso, UMMP 57528, Arkona Shale, Hungry Hollow, Ontario.199

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002paleontologists (Austin and Austin, 1843; Yandelland Shumard, 1847; Owen, 1862; Meek andWorthen, 1866). The record of this associationextends from the Ordovician to the Permian (Fig.3), and not only is the identity of the interactingorganisms known, but already these early workershad recognized that the gastropods and crinoidsmust have been interacting during life. However,establishing the nature of that interaction proved amore elusive goal.Nature of interaction.—One of the earliestinterpretations of the gastropod-crinoid fossils wasthat of Austin and Austin (1843), who assumed thatthe specimens represented crinoids caught in theact of feeding on gastropods. The predator-preyinterpretation was rejected by Meek and Worthen(1866, 1868), who noted that the irregular shapeof the gastropod margin forming a tight fit to thecrinoid calyx implied a long-term interaction. Thatled them to conclude that the gastropods wererelying on the crinoids for food. By the latter partof the nineteenth century, the fact that gastropodstypically occupy a position over the crinoids’ analaperture led to the inference of coprophagy (Hinde,1885; Keyes, 1888a, 1888b).The interpretation that platyceratids fed oncrinoid waste and thus benefited from theassociation persisted through the twentieth century(e.g., Clarke, 1908; Bowsher, 1955; Lane, 1978;Meyer and Ausich, 1983; Boucot, 1990). Underthat scenario, it has been generally assumed thatthe impact on the crinoid was neutral and, thus,that the interaction represented commensalism(Keyes, 1888a; Bowsher, 1955; Lane, 1978; Meyerand Ausich, 1983). Wood (1980, p. 110) consideredthat the crinoid might have benefited from acoprophagous gastropod by being provided with a“competent elimination system…”—clearlyarguing that this represented a case of mutualism.Thomas (1924, p.451), on the other hand, arguedthat crinoids “were unquestionably the unhappyhosts of a weighty and persistent parasite.”Likewise, Clarke (1921, p. 64) suggested that theinteraction represented parasitism: “… [The]gastropod…must have been obnoxious to it [thecrinoid] as it interfered with the normal alimentaryfunction.” More recently the parasitic interpretationhas gained support from a number of studiesFIGURE 3—Distribution of crinoid genera parasitized by platyceratids during the Paleozoic. Solid barsrepresent co-occurrences of crinoids and platyceratids; open bars represent inferred associations basedon trace fossils (platyceratid growth scars). The upper curve represents Paleozoic crinoid genericdiversity based on data from Sepkoski’s unpublished compendium.200

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESincluding those of Lane (1984), Rollins andBrezinski (1988), Baumiller (1990, 2001), Gahnand Baumiller (2001), and Gahn et al. (in prep.).Although coprophagy has been the mostcommonly invoked behavior for the infestinggastropods, other ideas about how they may havebenefited from crinoids have also been proposed.For example, Lane (1984) suggested that gastropodsfed on crinoid gametes that were shed through theanal vent. This idea is supported by the fact thatwhereas the gametes of modern crinoids are foundon their proximal arms in specialized genitalpinnules, such pinnules have not been found amongPaleozoic crinoids. Lane argued that those crinoidsmust have shed their gametes through the anal vent,which was occupied by the gastropod.Baumiller (1990, 2001) also argued thatgastropods positioned over the crinoid anal openingcould have fed on more than just crinoid excreta;for example, they may have stolen undigestednutrients from their host, and thus have beenkleptoparasitic. He supported this claim ofkleptoparasitism by noting that among extantcomatulid crinoids, the captured and still largelyundigested food travels rapidly to the hindgut(Holland et al., 1991), where most digestion occurs.If this were true of Paleozoic crinoids, infestinggastropods could reach through the anal openinginto the hindgut with their proboscis and extractundigested nutrients. The kleptoparasitic scenariowas recently quantified using a cost-benefitanalysis (Baumiller 2001).Effect of interaction on platyceratids.—Coprophagy, gametophagy, and/or kleptoparasitismrepresent plausible modes of obtaining nutrientsby the infesting platyceratids. Moreover, eachpredicts that a platyceratid occupying the anal vent,and thus having access to nutrients, was at anenergetic advantage relative to a conspecific thathad no such access. A recent description of multigastropod-infestedcrinoids (Baumiller, 2002)provides a test of this prediction and thus a rareopportunity to demonstrate explicitly theadvantages to infesting individuals.The crinoid-platyceratid association isgenerally characterized by a single gastropodpositioned on the crinoid tegmen; only a fewinstances of multi-infested crinoids have beenreported. Baumiller (2002) described twospecimens of the crinoid Arthroacantha withmultiple gastropods attached. On one of these, thelargest gastropod was positioned directly over theanal vent with 6 smaller gastropods attached inclose proximity to the vent. The larger size wasinterpreted as a reflection of greater rate ofgrowth—a direct consequence of access to crinoidwaste, gut, gametes, or any combination of these—and demonstrates the advantage to the infester.Effect of the interaction on crinoids.—Although in the Arthroacantha-platyceratidexample cited above, the infesting organism wasshown to have gained an “advantage” by virtue ofits position over the crinoid anal vent, determiningwhether the interaction involved mutualism,commensalism, or parasitism requires knowledgeof the effects on the host. One feature of the hostthat may be affected by the presence of an infester,and that is readily quantifiable, is size: if infestedcrinoids are smaller than uninfested ones, thenparasitism is supported. Rollins and Brezinski(1988) used this logic to assess whether infestinggastropods had a detrimental effect on crinoids.Using four crinoid ‘stands’, they measured the sizeof infested and uninfested calyces of the crinoidPlatycrinites, and found that infested individualswere on average smaller than uninfested ones,suggesting that parasitized crinoids were negativelyaffected. This elegant approach was hindered bythe small sample size (infested N = 11 and 1,uninfested N = 6 and 3) and the fact that thecomparison involved infested individuals from oneset of crinoid ‘stands’ and uninfested individualsfrom another set of ‘stands’. Due to the smallsample, the average size differences were notstatistically significant. Although statisticallyinconclusive, the Rollins and Brezinski (1988)study was suggestive of parasitism and provided amethodological blueprint for future studies.A large collection of two species of Devoniancamerate crinoids allowed for a statistical test ofthe size effect of infestation on the hosts (Gahn etal., in prep.). By comparing infested and uninfested201

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002specimens of two species of crinoids, Corocrinusand Gennaeocrinus, the authors will show thatinfested individuals were significantly smaller thanuninfested ones (Fig. 4). The differences areinterpreted as indicating slower growth rates orhigher mortality rates of infested crinoids. Thisdetrimental effect of platyceratids on their hostssupports the parasitic, nutrient-stealing hypothesis.Evolutionary consequences of platyceratidparasitism on crinoids.—The above observationsindicate that platyceratid gastropods represented abiological “hazard” to crinoids. In the context ofthe hypothesis of escalation (Vermeij 1977, 1980,1987), organisms that have enemies, be theypredatory or parasitic, are expected to respondevolutionarily to these hazards. Given that theassociation between crinoids and platyceratids wasgeologically long-lived, one might predict a crinoidresponse. A study by Gahn and Baumiller (2001)provides a test of this prediction.Gahn and Baumiller (2001) suggested that onepotentially effective anti-infestation feature ofcrinoids was a long, slender anal tube. Theseauthors argued that shifting the gastropod’s pointof access to the apex of the tube would, amongother things, have made it difficult for gastropodsto position themselves over the anus.To test the effectiveness of the tube as antiinfestationdevice, Gahn and Baumiller (2001)categorized crinoids as 1) tube-bearing, infested;2) tubeless, infested; 3) tube-bearing, uninfested;or 4) tubeless, uninfested—and showed that thedistribution of platyceratids was not independentof the presence of the tube (chi square P = 0.007).Platyceratids were associated almost exclusivelywith tubeless crinoids.Gahn and Baumiller (2001) also explored theevolutionary scenario that the tube evolved inFIGURE 4—Size distribution of platyceratid-infested (dotted curve) and uninfested (solid curve)specimens of Genneaocrinus variabilis from the Middle Devonian Bell Shale, Rockport, Michigan. Thedifference is statistically significant; infested individuals (N = 30) averaged 2.4 mm in height whereasuninfested specimens (N = 396) averaged 4.5 mm (t-test, P

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESresponse to parasitism. A phylogenetic analysisusing 26 genera of monobathrid camerates showedthat: 1) the most parsimonious distribution required4 independent events of tube evolution, and 2) thatthe tubeless sister taxa of tubed crinoids wereinfested by platyceratids, while the tubed crinoidswere generally uninfested.The results of the Gahn and Baumiller (2001)study were consistent with the hypothesis that thetube evolved in response to parasitism; but escalationgoes a step further and invokes an arms race. Thus,adaptations that counter the effectiveness of enemiesshould be rendered ineffective by morphological orbehavioral evolution of those enemies. Didplatyceratids counter the evolution of exaggeratedanal tubes with such changes? Baumiller (1990)reported numerous individuals of the Mississippianbatocrinid crinoids, Batocrinus icosidactylus andBatocrinus irregularis, with circular, cylindrical totapered holes penetrating the plates of the tegmenat the base of a long, slender, multi-plated tube withan anal opening at its apex (Fig. 5). A sectionedspecimen of another tube-bearing Mississippianbatocrinid, Macrocrinus mundulus, infested by aplatyceratid, revealed a hole in the base of the tubedirectly beneath the gastropod. The presence of U-shaped attachment scars, rare instances of multipleholes, and healed (or incomplete) holes was used asevidence that the holes in crinoids did not representpredation, but parasitism. Furthermore, it indicatesthat the tube was not a foolproof strategy of escapefrom infestation by snails, as it could be counteredby drilling. If platyceratid drilling evolved inresponse to the evolution of the tube, the “tit-fortat”would represent a case of escalation.PLATYCERATIDS ASPARASITES ON OTHER TAXACrinoids are not the only echinoderm groupassociated with platyceratids. Several instances ofplatyceratids attached to blastoid calyxes areknown, including two Devonian and twoMississippian examples (Fig. 6). Levin and Fay(1964) described several specimens of Diploblastuskirkwoodensis (Mississippian), each with a smallFIGURE 5—Examples of Mississippian batocrinidcrinoids with drillholes. Scale bar = 0.5 cm.1, Macrocrinus mundulus (P19426), lateral view ofspecimen. Note the long anal tube extendingbeyond the tips of the arms and the platyceratidpositioned on the tegmen between the arm bases.2, Photomicrograph of ground section of samespecimen. 3, Batocrinus icosidactylus (P19402),lateral view showing drillhole at base of broken-offanal tube. 4, Batocrinus icosidactylus, lateral viewshowing two complete drillholes near base of analtube. 5, Batocrinus icosidactylus (P19394), lateralview of partial calyx with complete drillhole. 6,Batocrinus irregularis (P19393), lateral view ofspecimen with a robust tube penetrated by drillhole.All specimens housed in the Field Museum ofNatural History, Chicago, Illinois.203

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 6—Examples of blastoid-platyceratid association (5,6,7) and blastoids with drillholes (1,2,3,4).Scale bar = 0.5 cm. 1, Nucleocrinus verneuilli (OSU 14516) with complete drillhole, Columbus Limestone,Middle Devonian. 2, Orophocrinus stelliformis (MCZ 360) with complete drillhole, Burlington Limestone,Mississippian. 3, Orophocrinus stelliformis (MCZ 144) with incomplete or healed drillhole, BurlingtonLimestone, Mississippian. 4, Heteroschisma canadense (A. Fabian personal collection), Hungry HollowFormation, Middle Devonian. 5, Nucleocrinus sp. with attached platyceratid (J. Topor personal collection),Rockport Quarry Limestone, Middle Devonian. 6, Diplocrinus kirwoodensis with an attached platyceratid(WUPM ), Saint Louis Limestone, Chesterian. 7, Heteroschisma subtruncatus (USNM 481253) with anattached platyceratid, Thunder Bay Formation, Middle Devonian. Repositories: OSU–Ohio StateUniversity Natural History Museum; MCZ–Museum of Comparative Zoology, Harvard University; WUMP–Washington University Paleontology Museum.Platyceras attached to the blastoid calyx. Severalspecimens of Mississippian Pentremites withattached platyceratids have also been reported(Meek and Worthen, 1868; Thein and Nitecki,1974; Kelly, 1984), and an additional dozenspecimens are housed in the collections of theUMMP. Much more rare are cases of Devonianblastoids with platyceratids: such specimensinclude Nucleocrinus and Heteroschisma(Baumiller, 1996). These latter specimens areespecially important to the interpretation of theinteraction between platyceratids and blastoids.Whereas Levin and Fay (1964), noting the positionof the gastropods over the anal vent of specimensof Diploblastus, concluded that they werecoprophagous, which would indicate that theinteraction was either mutualistic or commensal, aparasitic interpretation has recently been offered(Baumiller, 1993, 1996; Baumiller and Macurda,1995). These authors illustrated numerous casesof drilled blastoids. Most commonly drilled werespecimens of nucleocrinids, Heteroschisma, andthe Mississippian blastoid Orophocrinus. As is thecase with holes in crinoids, the blastoid holes aretypically single, circular in plan view with adiameter greater than 1 mm (range 0.3–2.5 mm),complete, and straight-sided to tapered. The knownassociation of platyceratids with blastoids,204

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESespecially with the two drilled taxa, Nucleocrinusand Heteroschisma, and the documented drillingof crinoids by platyceratids led the authors to arguethat the blastoid holes were produced by thesegastropods (Fig. 7). They suggested that the rarepresence of doubly-drilled blastoids and theoccasional incomplete/healed holes argued againstpredation, and instead for a long-term association.In addition, the position of the holes away fromthe anal vent was used to argue against coprophagy;instead, a hypothesis of nutrient-stealing wasproposed. This parasitic hypothesis is yet to betested explicitly, but if it stands up to scrutiny, theblastoid-platyceratid interaction may be worthexamining in the context of escalation.One other example of a platyceratidechinodermassociation is found in the Silurian andinvolves the platyceratid Naticonema and thecystoid Caryocrinites (Clarke, 1908; Bowsher,1955; Kluessendorf, 1983). As in other instancesof infested pelmatozoans, the gastropod is foundattached to the theca of the cystoid; this has beeninterpreted as reflecting the coprophagous habit ofthe gastropod and thus a commensal or mutualisticrelationship. However, based on the evidence fromcrinoids and blastoids, a parasitic interpretation isequally plausible.CONCLUSIONThe list of examples of parasitism in the fossilrecord is extensive. Distinguishing parasitism frompredation, commensalism, or mutualism requires(1) evidence of a long-term relationship betweentwo organisms, (2) demonstration of how theinteraction benefits the parasite, and (3)demonstration of how it is detrimental to the host.Only in exceptional cases can these requirementsbe fulfilled. One example that provides muchinformation on parasitic interactions involvesplatyceratids and crinoids. This association hasrecently been re-interpreted as parasitic rather thancommensal. The fact that platyceratids weregeologically long-lived (Ordovician to Permian),abundant, occurred in a broad range of marinesettings, and are known to be associated with avariety of taxa (crinoids, blastoids, cystoids,brachiopods), implies that the parasitic habit mayhave been common during the Paleozoic.Furthermore, because platyceratids have beenFIGURE 7—Distribution of blastoid genera parasitized by platyceratids during the Paleozoic. Solidbars represent co-occurrences of crinoids and platyceratids; open bars represent inferred associationsbased on trace fossils (platyceratid growth scars). The upper curve represents Paleozoic blastoid genericdiversity based on data from Sepkoski’s unpublished compendium.205

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002shown to be capable of drilling their hosts, it maybe more appropriate to begin with parasitism, ratherthan predation, as the “null-hypothesis” for themany boreholes in Paleozoic invertebrates. Theconsequences of parasitism, like predation, maylead to escalation, and the association ofplatyceratids and crinoids, and possible other hosts,may provide a rich data source for testing thisevolutionary hypothesis.ACKNOWLEDGMENTSThis work was supported by grants from theNational Science Foundation and the PetroleumResearch Fund of the American Chemical Society.It could not have been completed without thegenerous help of the Friends of the University ofMichigan Museum of Paleontology, especially K.Dobson, A. Fabian, S. Hyne, J. Koniecki, R.Rutkowski, D. Smarjese, D. Thompson, the Toporfamily, and S. Vergiels. Discussions with D. Fisherand P. Kaplan helped clarify many of the issues.B. Deline, B. Miljour, and E. Roberts providedinvaluable technical assistance. The manuscriptwas much improved by comments from M.Kowalewski, D. Meyer, and M. Wilson. To all, agreat many thanks.REFERENCESALEKSEEV, A. S., AND L. G. ENDELMAN. 1989. Association of ectoparasitic gastropods with Upper Cretaceousechinoid Galerites, p. 165–174. In Anonymous (ed.), Fossil and Recent Echinoderm Researches. Academyof Sciences of the Estonian SSR, Tallin.AUSICH, W. I., AND R. A. GURROLA. 1979. Two boring organisms in a Lower Mississippian community of southernIndiana. Journal of Paleontology, 53:335–344.AUSTIN, T., AND T. AUSTIN. 1843–1846. Monograph on Recent and Fossil Crinoidea. Bristol and London, 128 p.BABCOCK, L. E. 1990. Conularid pearls, p. 68–70. In A. J. Boucout, Evolutionary paleobiology of behavior andcoevolution. Elsevier, Amsterdam, 725 p.BAIRD, G. C., C. E. BRETT, AND J. T. TOMLINSON 1990. Host-specific acrothoracid barnacles on Middle Devonianplatyceratid gastropods. Historical Biology, 4:221–244.BARRY, C., W. HATCH, AND S. KLIMLEY. 1972. Commensalism and parasitism of shell-borers from the CretaceousNavesink Formation of New Jersey. Geological Society of America Abstracts with Programs, 4(1):7–8.BATES, D. E. B., AND D. K. LOYDELL. 2000. Parasitism on graptoloid graptolites. Palaeontology, 43(6):1143–1151.BAUMILLER, T. K. 1990. Non-predatory drilling of Mississippian crinoids by platyceratid gastropods. Palaeontology,33:743–748.BAUMILLER, T. K. 1993. Boreholes in Devonian blastoids and their implications for boring by platyceratids. Lethaia,26:41–47.BAUMILLER, T. K. 1996. Boreholes in Middle Devonian blastoid Heteroschisma and their implications for gastropoddrilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 123:343–351.BAUMILLER, T. K. 2001. Cost-Benefit analysis as a guide to the ecology of drilling platyceratids. Paleobios, 21:29.BAUMILLER, T. K. 2002. Multi-snail infestation of Devonian crinoids and the nature of platyceratid-crinoidinteractions. Acta Palaentologica Polonica, 47(1):132–139.BAUMILLER, T. K., AND D. B. MACURDA, JR. 1995. Borings in Devonian and Mississippian blastoids (Echinodermata).Journal of Paleontology, 69:1084–1089.BAUMILLER, T. K., L. R. LEIGHTON, AND D. L. THOMPSON. 1999. Boreholes in Mississippian spiriferide brachiopods andtheir implications for Paleozoic drilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 147:283–289.BOUCOT, A. J. 1990. Evolutionary paleobiology of behavior and coevolution. Elsevier, Amsterdam, 725 p.BOUCOT, A. J., AND S. R. MCCUTCHEON. 1986. Ziegler’s blisters in Pentameroides from a Lower Silurian localityin the northeastern part of the Mascarene-Nerepis Belt, southern New Brunswick. Canadian Journal of EarthScience, 23:1437–1442.206

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESBOWSHER, A. L. 1955. Origin and adaptation of platyceratid gastropods. University of Kansas, PaleontologicalContributions, Mollusca, 5:1–11.BRETT, C. E. 1978. Host-specific pit-forming epizoans on Silurian crinoids. Lethaia, 11:217–232.BRETT, C. E. 1985. Tremichnus: a new ichnogenus of circular-parabolic pits in fossil echinoderms. Journal ofPaleontology, 59:625–635.BUEHLER, E. J. 1969. Cylindrical borings in Devonian shales. Journal of Paleontology, 43:1291.CARRIKER, M. R., AND E. L. YOCHELSON. 1968. Recent gastropod boreholes and Ordovician cylindrical borings.U.S. Geological Survey Professional Paper, 593-B:1–26.CHATTERTON, B. D. E. 1975. A commensal relationship between a small filter-feeding organism and AustralianDevonian brachiopods. Paleobiology, 1:371–378CLARKE, G. L. 1954. Elements of Ecology. John Wiley, New York, 534 p.CLARKE, J. M. 1908. The beginnings of dependent life. Bulletin of the New York State Museum, 121:1–28.CLARKE, J. M. 1921. Organic dependence and disease: Their origin and significance. Yale University Press, NewHaven, 113 p.CONWAY MORRIS, S. 1981. Parasites and the fossil record. Parasitology, 82:489–509.CONWAY MORRIS, S., AND S. BENGSTON. 1994. Cambrian predators: Possible evidence from boreholes. Journal ofPaleontology, 68(1):1–23.ELIAS, M. K. 1966. Living and fossil algae and fungi, formerly known as structural parts of marine bryozoans.The Paleobotanist, 14:5–18.FISHER, P. H. 1962. Perforation de fossiles pre-tertiaries attribuees à des gastéropodes predateurs. Journal deConchyliologie, 102:68–78.FRANZEN, C. 1974. Epizoans on Silurian-Devonian crinoids. Lethaia, 7:287–301.GAHN, F. J., AND T. K. BAUMILLER. 2001. Testing evolutionary escalation between crinoids and platyceratidgastropods and phylogenetic analysis of the Compsocrinina (Crinoidea: Monobathrida). Geological Societyof America Abstracts with Programs, 33(6):247.GLAESSNER, M. F. 1969. Decapoda, p. R400–R533. In R. C. Moore (ed.), Treatise on Invertebrate Paleontology,Part R, Arthropoda 4(2). University of Kansas Press, Lawrence, Kansas.GRAHN, Y. 1981. Parasitism on Ordovician chitinozoa. Lethaia, 14(2):135–142.HAYAMI, I., AND Y. KANIE. 1980. Mode of life of a giant capulid gastropod from the Upper Cretaceous of Saghalienand Japan. Palaeontology, 23(3):689–698.HERDMAN, W. A. 1906. Report to the government of Ceylon on the pearl oyster fisheries of the Gulf of Manaar,Part V. Royal Society of London, London, 452 p.HINDE, G. J. 1885. Description of a new species of crinoids with articulated spines. The Annals and Magazine ofNatural History, 5th Series, 15:157–173.HOLLAND, N. D., A. B. LEONARD, AND D. L. MEYER. 1991. Digestive mechanics and gluttonous feeding in thefeather star Oligometra serripinna (Echinodermata: Crinoidea). Marine Biology, 111:113–119.KELLY, S. M. 1984. Paleoecology and paleontology of the Indian Springs shale member, Big Clifty Formation(Middle Chesterian) in south-central Indiana. Ph.D. dissertation, Indiana University.KEYES, C. R. 1888a. On the attachment of Platyceras to Paleocrinoids, and its effects in modifying the form of theshell. American Philosophical Society Proceedings, Transactions, 25:231–243.KEYES, C. R. 1888b. The sedentary habits of Platyceras. American Journal of Science, 36:269–272.KIER, P. M. 1981. A bored Cretaceous echinoid. Journal of Paleontology, 55:656–659KLUESSENDORF, J. 1983. Observations on the commensalism of Silurian platyceratid gastropods and stalkedechinoderms. Wisconsin Academy of Sciences, Arts and Letters, 71:48–55.KOWALEWSKI, M., M. G. SIMÕES, F. F. TORELLO, L. H. C. MELLO, AND R. P. GHILARDI. 2000. Drill holes in shells ofPermian benthic invertebrates. Journal of Paleontology, 74(3):532–543.LANE, N. G. 1978. Mutualistic relations of fossil crinoids, p. T345–T347. In R. C. Moore and C. Teichert (eds.),Treatise on Invertebrate Paleontology, Part T, Echinodermata 2. Geological Society of America and Universityof Kansas, Lawrence.LANE, N. G. 1984. Predation and survival among inadunate crinoids. Paleobiology, 10:453–458.207

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002LEIGHTON, L. R. 2001. New example of Devonian predatory boreholes and the influence of brachiopod spines onpredator success. Palaeogeography, Palaeoclimatology, and Palaeoecology, 165:53–69.LEVIN, H. L., AND R. O. FAY. 1964. Relationship between Diploblastus kirkwoodensis and Platyceras (Platyceras).Oklahoma Geology Notes, 24:22–29.MACKINNON, D. I., AND G. BIERNAT. 1970. The probable affinities of the trace fossil Diorygma atrypophilis.Lethaia, 3:163–172.MADSEN, F. J., AND T. WOLFF. 1965. Evidence of the occurrence of Ascothoracica (parasitic cirriped) in upperCretaceous. Dansk Geologisk Forening, Meddelelser, 15:556–558.MAROUF, A. M. 1999. Parasitism, abnormal growth and predation on Cretaceous echinoids from Egypt. Revue dePaleobiologie, 18(1):69–83.MEEK, F. B., AND A. H. WORTHEN. 1866. Radiata, Echinodermata, Crinoidea. Proceedings of the Academy ofNatural Sciences of Philadelphia, (1866):251–275.MEEK, F. B., AND A. H. WORTHEN. 1868. Geology and Palaeontology, pt 2, Palaeontology of Illinois. IllinoisGeologic Survey, 3:289–565.MERCIER, J. 1937. Zoothylacies d’échinide fossiles provoquées par un crustace: Castexia douvillei nov. gen., nov.sp. Bulletin de la Societe Géologique de France, Série 5, 6:149–154.MEYER, D. L., AND W. I. AUSICH. 1983. Biotic interactions among Recent and fossil crinoids, p. 377–427. In M. F. S.Tevesz and P. L. McCall (eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum, New York.OWEN, R. 1862. Report of a geological reconnaissance of Indiana made during the years 1859 and 1860.Indianapolis, 368 p.PALMER T. J., AND M. A. WILSON. 1988. Parasitism of Ordovician bryozoans and the origin of pseudoborings.Palaeontology, 31(4):939–949.PERNET, B., AND A. J. KOHN. 1998. Size-related obligate and facultative parasitism in the marine gastropodTrichotropis cancellata. Biological Bulletin, 195:349–356.PICKETT, J. 1973. Interspecific relationships among fossil species. Australian Natural History, 17(10):338–343.RODRIGUEZ J., AND R. C. GUTSCHICK. 1977. Barnacle borings in live and dead hosts from the Louisiana Limestone(Fammenian) of Missouri. Journal of Paleontology, 51:718–724.ROLLINS, H. B., AND D. K. BREZINSKI. 1988. Reinterpretation of crinoid-platyceratid interaction. Lethaia, 21:207–217.ROMAN, J. 1952. Quelques anomalies chez Clypeaster melitensis Michelin. Bulletin de la Société geologique deFrance, 2:3–11.SAVAZZI, E. 1995. Parasite-induced teratologies in the Pliocene bivalve Isognomon maxillatus. Palaeogeography,Palaeoclimatology, Palaeoecology, 116(2):131–139.SEILACHER, A. 1968. Swimming habits of belemnites, recorded by boring barnacles. Palaeogeography,Palaeoclimatology, Palaeoecology, 4:279–285.SEPKOSKI, J. J., JR. 1984. A kinetic model of Phanerozoic taxonomic diversity, III: Post-Paleozoic families andmass extinctions. Paleobiology, 10:246–267.SOHL, N. F. 1969. The fossil record of shell borings by snails. American Zoologist, 9:725–734.SOLOVJEV, A. N. 1961. Parasit Canceripustula nocens v pozdneoorskogo morskogo ezha. PaleontologicheskiiZhurnal, 4:115–119.SUTHERLAND, S. J. E. 1994. Ludlow chitinozoans from the type area and adjacent regions. Paleontological SocietyMonograph, 104 p.TASNADI-KUBACSKA, A. 1962. Paläopathologie. G. Fischer-Verlag, Jena, 2, 269 p.TAYLOR T. N., AND J. M. OSBORN, 1996. The importance of fungi in shaping the paleoecosystem. Review ofPaleobotany and Palynology, 90(3–4):249–262.THEIN, M. L., AND M. H. NITECKI. 1974. Chesterian (Upper Mississippian) Gastropoda of the Illinois basin. FieldianaGeology, 34:1–238.THOMAS, A. O. 1924. Echinoderms of the Iowa Devonian. Iowa Geologic Survey, Annual Reports, 29:385–552.THORSONG, G. 1965. A neotenous dwarf-form of Capulus ungaricus (L.) (Gastropoda, prosobranchia)commensalistic on Turritella communis Risso. Ophelia, 2(1):175–210.VERMEIJ, G. J. 1977. The Mesozoic marine revolution: Evidence from snails, predators, and grazers. Paleobiology,3:245–258.208

BAUMILLER AND GAHN—PARASITISM ON MARINE INVERTEBRATESVERMEIJ, G. J. 1980. Drilling predation of bivalves in Guam: Some paleontological implications. Malacologia,19:329–334.VERMEIJ, G. J. 1987. Evolution and escalation. Princeton University Press, Princeton, 527 p.VOIGHT, E. 1955. Artspezifischer parachorismus (?) von serpuliden in kreidebryo. Palaeontologische Zeitschrift,29(1–2):8–20.WELCH , J. R. 1976. Phosphannulus on Paleozoic crinoid stems. Journal of Paleontology, 50:218–255.WOOD, G. D. 1980. Coprolite, urolite, and “vomite.” Maledicta, 4:109–115.YANDELL, L.P., AND B. F. SHUMARD. 1847. Contributions to the geology of Kentucky. Prentice and Weissinger,Louisville, 36 p.209


LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESPALEOBIOLOGY OF PREDATORS, PARASITOIDS, ANDPARASITES: DEATH AND ACCOMODATION IN THE FOSSILRECORD OF CONTINENTAL INVERTEBRATESCONRAD C. LABANDEIRADepartment of Paleobiology, National Museum of Natural History, Smithsonian Institution, Washington, DC20560-0121 and Department of Entomology, University of Maryland, College Park, Maryland 20742 USAABSTRACT—Carnivory is the consumption of one animal by another animal; among invertebrates in terrestrialand freshwater ecosystems this type of feeding can take three forms: predation, parasitoidism, and parasitism.Differences among these three functional modes involve (i) whether the duration of feeding on the prey item is quickor there is an accommodation, coevolutionary or otherwise, between the carnivore and the host prey; (ii) whetherthe prey or host is killed; (iii) whether single or multiple prey or host items are consumed during the carnivore’slifespan, and (iv) the relative sizes of the carnivore and its prey or host. Uniformitarian and nonuniformitarianevidence directly relating to the history of carnivory can be found in exceptionally preserved deposits from themid-Paleozoic to the Recent, but such evidence is relatively rare because carnivores are the least representedtrophic group in ecosystems. Six types of paleobiological data provide evidence for carnivory: taxonomic affiliation,fossil structural and functional attributes, organismic damage, gut contents, coprolites, and indications of mechanismsfor predator avoidance.Only 12 invertebrate phyla have become carnivorous in the continental realm. Six are lophotrochozoans(Acanthocephala, Rotifera, Platyhelminthes, Nemertinea, Mollusca, and Annelida) and six are ecdysozoans(Nematoda, Nematomorpha, Tardigrada, Onychophora, Pentastoma, and Arthropoda). Most of these groupshave poor continental fossil records, but the two most diverse—nematodes and arthropods—have comparativelygood representation. The record of arthropods documents (i) the presence of predators among primary producers,herbivores, and decomposers in early terrestrial ecosystems; (ii) the addition later in the fossil record of the moreaccommodationist strategies of parasitoids and parasites interacting with animal hosts; (iii) the occurrence ofsimpler food-web structures in terrestrial ecosystems prior to parasitoid and parasite diversification; and (iv) arole for mass extinction in the degradation of food-web structure that ultimately affected carnivory. Future researchshould explore how different modes of carnivory have brought about changes in ecosystem structure throughtime. Despite numerous caveats and uncertainties, trace fossils left by predators on skeletons of their prey remainone of the most promising research directions in paleoecology and evolutionary paleobiology.INTRODUCTION“The distinctive element is that in many, but notall, of the predaceous forms the female eats the malewhile mating is in progress, by piercing the cuticleand dissolving and sucking out the body contents aswith any other small insect prey [Fig. 1a]…“…The female [of Probezzia concinna] attacksthe male through the head, the only part of the bodythe proboscis can reach, and the whole process ofinjection of the lytic saliva and intake of theliquefied tissues takes place at this one site. [T]hemale … was reduced to an empty and dry cuticleonly. When the female moved away however, the[male’s] cuticle broke at the membrane betweenthe 6th and 7th abdominal segments, leaving thetorn off terminalia attached to the female in themating position [Fig. 1b].“…In the male of the predaceous midgesdestruction of the head by the saliva of the female,soon after mating is established, appears to resultin a sustained copulatory position maintained untilthe tissues of the male have been entirely consumed,and even longer.”—J. A. Downes (1971a, p. 38–39)(Brackets added.)211

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002THE POET TENNYSON described natureas “red in tooth and claw,” a metaphoric andseemingly apt description of predation. His imageof vertebrates dispatching other vertebrates by theuse of carnassials and talons, however, is clearlyexceptional, considering where the preponderanceof terrestrial carnivory resides; this picture shouldbe replaced by the equally grisly image ofhemolymph-encrusted mouthparts and tarsi—for itis among the invertebrates, and particularly thearthropods, where the overwhelming bulk ofcarnivory occurs in terrestrial settings. The abovequotation from Downes (1971a) about predationin the biting midge Probezzia is emblematic of theinvertebrate world, which forms an important partof the trophic capstone in terrestrial food webs. Inthis contribution, the fossil history of terrestrialinvertebrate carnivory is explored, with a focus onarthropods and especially insects. The discussionincludes examples from both freshwater andterrestrial environments, both of which belong tothe continental realm. This paper begins with adefinition of carnivory and its subcategories, thenaddresses the nature of the fossil evidence, entersinto a discussion of the historical pattern ofcarnivory, delves into four salient issues regardingits evolution, and ends in a summary. The brevityof this review does not allow for adequate coverageof all relevant issues, and it is hoped that theinterested will consult the cited references foradditional information and insight.THE NATURE OF CARNIVORYThere are three basic types of carnivory:predation, parasitoidism, and parasitism (Vinsonand Barbosa, 1987). All three comprise the thirdtrophic level characterized by the consumption oforganisms, especially of herbivores and othercarnivores. The ultimate basis of carnivory is inprimary producers whose energy originates eitherfrom the sun (photoautotrophs) or the chemicalbonds of oxidized compounds (chemoautotrophs).Primary producers provide food, via herbivores andeventually carnivores, for the sustenance of foodwebs. This trophic pyramid, together with a crucialside-loop for organisms that degrade all types ofFIGURE 1—A, Two biting midges of Probezziaconcinna (Diptera: Ceratopogonidae), in copulowith the female simultaneously feeding on the malethrough a single puncture of his head capsule. B,Posterior female abdomen of a closely relatedgenus, Palpomyia, after mating, with attached,torn-off male genitalia in light-grey shading. Bothredrawn from Downes (1971a).dead tissues (saprobes), is characterized by a typical90 percent loss in biomass and energy assimilatedat each interface between levels (DeAngelis, 1992).Interestingly, this decrease in mass and energytransferred between adjacent levels is also evidentfor each trophic type in the fossil record. Recordsof carnivory are much scarcer than records of plantherbivoreassociations (Labandeira, 2002a).212

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESInvertebrates in general, and insects in particular,dominate continental food webs; among insects,approximately 50 percent of all species areherbivorous and 25 percent are predaceous (Hagen,1987; Wilson, 1992). Site-specific studies providesimilar values (Erwin and Scott, 1980).Nevertheless, the comparative scarcity of evidencefor carnivory in fossil assemblages should not beliethe intricate, diverse, and fascinating roles playedby predation, parasitoidism, and parasitism incontinental ecology during the past 420 millionyears. It is for this reason that an understanding ofthese three types of feeding strategies, subsumedunder “carnivory,” is important.The common conception of predation involvesa consumption of prey resulting in the victim’sdeath. Predation is defined as the total or near-totalconsumption of a live organism (Price, 1980), butalso includes suspended zooplankton in aquaticsystems (Merritt and Cummins, 1984), plantorganisms such as seeds (Janzen, 1978), and pollenor spores (Fuchs, 1975). Both seed predation andpalynivory are nutritionally important for manycontinental arthropods (Proctor et al., 1996), andrepresent feeding strategies that are distinct fromtypical herbivory, which is characterized by survivalof the host plant. Conventional continental predatorsare animals that directly and immediately consumeother animals (Hagen, 1987; Sih, 1987), althoughtrophically equivalent modes are seen in fungi(Evans, 1989) and even plants (Juniper, 1986) thatslowly dispatch their animal prey. Organisms thatfeed on invertebrate prey more slowly andincompletely are known as parasitoids (Godfray,1994): in parasitoidism, a relatively long-lived hostis eventually consumed and dies. A third type ofcarnivory is parasitism (Schmid-Hempel, 1998),characterized by multiple hosts that survive andare only partially consumed. Consumption ofvegetative tissues which does not result in the deathof the host plant has also been termed parasitism(Schoonhoven et al., 1998), but is excluded herebecause of its equation with herbivory in theliterature and its status as the contrasting trophicmode to carnivory. Additionally, the consumptionof suspended plankton and benthic films in aquaticsystems by microvorous filter feeders, sievers, andscrapers includes the consumption of minuteanimals, but is often ecologically equated withheribvory and will not be treated here further (seeMerritt and Cummins, 1984).The life histories of parasitoids are ecologicallytransitional between predators—which rapidlyconsume multiple and often unrelated individualsduring a lifetime—and parasites, which feed throughseveral generations on a near-infinite food resourcefrom mostly single, generally surviving individuals(Price, 1980; Vinson and Barbosa, 1987). Thus,parasitoids pursue an ecologically intermediatestrategy of feeding generally on a predictable andfinite food resource consisting of a single host thatis eventually killed; moreover, they are characterizedby a life cycle that includes a free-living reproductivestage whose habitat is separate from that of thedeveloping larva (Yeates and Greathead, 1997).Additional distinctions between predators andparasites help to define the role of parasitoids.Whereas predators are similar in size to their prey,parasites are quite small compared to their hosts.Also, predators typically are generalized feedersthat acquire and dispatch their prey by force; butparasites generally are tissue specialists that areoften coevolved with their hosts. Lastly, predatorsare characterized by fewer and broader ecologicalniches than are specialist parasites, which typicallypartition monospecific hosts into multiplemicrohabitats. The host-accommodation exhibitedby parasites and parasitoids, in contrast to the rapidconsumption of prey by predators, is a major themein the evolutionary history of invertebrate carnivoryin continental ecosystems.TYPES OF EVIDENCEMuch of our understanding of the paleobiologyof carnivory is inferential, relying on imperfectpreservation, circumstantial evidence, and rareinsights gleaned by way of unique and spectacularmodes of preservation. These data are principallyderived from Fossil-Lagerstätten such as silicifiedhot-spring deposits, sideritic nodules withinorganic shales, ambers, and lacustrine oil shalesthat have particular spatiotemporal distributions in213

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002the 420-million-year continental invertebraterecord. These and more typical deposits (Figs.1,2)provide six types of evidence for carnivory:taxonomic affiliation, structural and functionalattributes, organismic damage, gut contents,coprolites, and mechanisms of predationavoidance—all but the last of which are analogousto those cited for recognition of arthropod herbivory(Labandeira, 1998b, 2002a). Of these types ofevidence, taxonomic affiliation is most often used,whereas coprolites from vertebrate or insectcarnivores are rarely reported (Thulborn, 1999).Functional and structural attributes as well asdamage to organisms are more frequent in the fossilrecord, but have been underused for inferring dietsand life habits. Gut contents provide a “smokinggun” by establishing both the carnivore and its prey,but are rarely preserved. Mechanisms of carnivoreavoidance, such as mimicry, spines, large size, larvalcases, and the presence of auditory tymbals foravoiding aerial predators are recognizable fromanalogous patterns in the modern world, but becomemore problematic deeper in the fossil record(Labandeira, 2002a). In fact, the applicability oftaxonomic uniformitarianism for understandingcarnivory decreases into the geologic past, andmust be supplemented or supplanted by other,intrinsic approaches such as structural andfunctional attributes or gut contents.Taxonomic Affiliation.—Obvious invertebrateclades that are pre-eminent carnivores are theChilopoda (centipedes), Araneida (spiders),Odonatoptera (dragonflies and damselflies),Mantodea (mantises), the Raphidioptera,Megaloptera, and Planipennia (snakeflies, alderflies,lacewings, and relatives), the Carabidae (groundbeetles), Dytiscidae (predaceous diving beetles), andAsilidae (robber flies) (Fig. 2J), among many others.For these taxa, which have phylogenetic continuityto the present day, a taxonomic uniformitarianapproach is valid. An extensive literature hasdocumented the obligate dependence of these extantgroups on live prey (Withycombe, 1922; Clausen,1940; Brues, 1972; Sih, 1987; Hagen, 1987; Foelix,1996; Corbet, 1999). A uniformitarian approach isalso applicable to fossil representatives of animalendoparasites such as tylenchid and mermithidnematodes, which consume live internal tissues(Poinar, 1991; Weitschat and Wichard, 1998); andectoparasitic lice (Fig. 2H) and fleas (Fig. 2L), whichfeed on dermal and subcutaneous tissues. Muchdiscussion of extant carnivory is centered onmouthpart structures and their association with diet(reviewed in Labandeira, 1990) and behavior,drawing inferences from functional morphology. Indiverse taxa that encompass multiple diets andfeeding strategies and larval stages—for example,leiodid and meloid beetles, and ants—a finertaxonomic resolution is necessary for certitude indietary assignments.There are instances in which prey andcarnivore are preserved in conjunction, especiallyin single pieces of amber (Weitschat and Wichard,1998; Kutscher and Koteja, 2000a), and frequentlyinvolving ants as predators. Many of theseexamples represent predator-prey interactions thatoccur today. Perhaps the ultimate directdemonstrations of carnivory are those samples ofBaltic amber in which a predatory ant is graspinga scale insect with its mandibles (Kutscher andKoteja, 2000b), or a spider is grasping a capturedant (Weitschat and Wichard, 1998). Other samplespreserve parasites or parasitoids exiting the bodycavity immediately after envelopment by resin(Weitschat and Wichard, 1998; Poinar and Poinar,1999). Lastly, there are amber specimens whereectoparasites such as mites or nematodes areentombed with mouthparts firmly attached to theirhost’s intersegmental membrane (Poinar et al.,1994a, 1997; Weitschat and Wichard, 1998), ordisplaying body distension due to engorgement ofhost hemolymph (Poinar et al., 1991).Structural and Functional Attributes.—Continental invertebrates possess several structuresthat indicate carnivorous habits. Prominent amongthese are mouthparts. Also included are body shape,the presence of attachment suckers, auditorytympanal organs, spider spinnerets, specializationsof the legs, ovipositor type, and cerci. Amongmandibulate hexapods, head mobility andforwardly positioned mouthparts are associatedwith predation (Walker, 1932), as are mandibles214

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITEScharacterized by incisiform and shearing teethrather than broad molar shelves for grinding(Samways et al., 1997). In insects with stylatemouthparts, such as flies with piercing-and-suckingbeaks, serrate or barbed mandibular stylets orblades indicate hematophagy (blood feeding)(Borkent, 1995; McKeever et al., 1991).Additionally, elongate and falcate mandibles thatare distally curved are almost always borne bypredators, some of which display a complicated “trapjaw” mechanism such as larval antlions and manyadult ants (Korn, 1943; Gronenberg, 1996). Othertypes of mouthparts are fashioned as co-coordinatedmultielement structures that can be hurled at preyimmediately anterior of the head, such as the labialmask of aquatic odonatopterans (Pritchard, 1976)or the adhesive labial sling of some terrestrial rovebeetles (Betz, 1996), both of which demonstrateobligate predation. Although it is the mouthparts ofinvertebrates that most directly interact with prey,other external appendicular structures are used tosubdue potential victims, including raptorialprothoracic legs armed with spines in Mantodea,cerci modified into forceps in japygid Diplura(telsontails), ovipositors designed to penetratearthropod hosts for insertion of parasitoid eggs(Fig. 2F), or chelate tarsi used for grasping hairamong ectoparasites of warm-blooded tetrapods(Askew, 1979; Godfray, 1994) (Fig. 2E).Additionally, extrasomatic structures may revealcarnivory, such as nits, or eggs glued to hair shafts,that are deposited by ectoparasitic lice (Weitschatand Wichard, 1998) (Fig. 2K). Other fossilizablestructures produced by carnivores are silk webs thatensnare potential prey (Bachofen-Echt, 1934;Gerhard and Rietschel, 1968; Weitschat andWichard, 1998). The body structures responsiblefor spider silk have a record extending back to theMiddle Devonian (Shear et al., 1989a) (Fig. 2B).Organismic Damage.—Carnivory can result inmany kinds of damage to continental invertebrates.This damage can occur on the prey or host oralternatively on the carnivore itself, and includesteratologies (abnormal growths) (Poinar and Poinar,1999), amputated appendages or broken bodyprocesses (Petrunkevitch, 1942; Rolfe, 1985;Hannibal and Feldman, 1988), small exit holes ofparasitoids from cocoons (Houston, 1987; Bown etal., 1997) (Fig. 2N), prey items wrapped by spidersilk (Weitschat and Wichard, 1998), excavated softtissue of arthropods consistent with predation(Poinar, 1999a), and refuse accumulations ofdiscarded victim carcasses (Abel, 1935; Weitschatand Wichard, 1998). These examples involveunusual removal or alteration of tissue that resultsfrom both unsuccessful and successful predationattempts; and they indicate the behavior of apredator located at a stationary site for an extendedtime. Although plant rather than animal individualsare killed, there is a fossil history of seed predationextending to the Late Paleozoic (Zherikhin, 1989;Labandeira, 2002a), characterized by variousmodes of removal of endosperm or its nutritiveequivalent in plants (Sharov, 1973; Genise, 1995;Mikulas et al., 1998) (Fig. 2M). An analogoussituation is that of palynivory (Fig. 2A)(Labandeira, 1998b, 2000).Gut Contents.—Instances of terrestrialinvertebrates preserved in the digestive tracts ofpredators are extremely rare in the fossil record.Regurgitated gut contents have been identified inmacerated material from the Middle DevonianGilboa deposit in New York State and are attributedto trigonotarbid arachnids (Gray and Shear, 1992).Large carnivorous insects such as protodonatandragonflies should reveal boluses of insect-bearingmaterial, but due to the rarity of preserved and intactsoft tissues, only one specimen is known: a terminalabdomen with a dark, oval mass within the lastsegment (Durden, 1988). The best examples of gutcontents in the fossil record come from mammals ofthe middle Eocene Messel oil shale of Germany. Thisdiverse assemblage contains gut contents from severalarchaic antecedents of extant mammalian orders.They include the bat Palaeochiropteryx that preyedon beetles and especially primitive moths andbutterflies active at night, dawn, or sunset (Franzen,1985) (Fig. 2I); the hedgehog Pholidocercus, whosestomach contents include beetles and other types ofinsects (Von Koenigswald et al., 1992); and theanteater Eotamandua, an avid consumer of termites,among other insects (Storch and Richter, 1992). It215


LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESis clear that anaerobic sealing of these carcassesresulted in exceptional preservation and promotedin particular the preservation of chitin, thusproviding rare dietary data on early insectivorousmammals (Richter, 1992).Coprolites.—Vertebrate-produced coprolitesthat contain identifiable terrestrial invertebrates are,like gut contents, rare in the fossil record. The bestexamples come from the Late Paleozoic. Forexample, from the Upper Carboniferous(Moscovian) Mazon Creek site in north-centralIllinois, Fischer (1979) reported coprolites thatcontained disarticulated sclerites of scorpions,millipedes, and cockroaches, indicating a terrestrialbut unknown predator. Occasionally, coprolites withfragmented insect material contain dismemberedelements that reveal morphological structuresotherwise difficult to observe in insect body fossils(Fig. 2C). For example, from lowermost Permiandeposits in central Kansas, coprolites laden witheuphorberiid millipede sclerites, some stillarticulated, have been described by Hannibal andFeldman (1988). A more recent occurrence ofcarnivore coprolites is described in Richter and←FIGURE 2—Evidence from taxonomic affiliation (F,H,J–L), structural and functional attributes (B,E,F,H,J,L),organismic damage (M,N), gut contents (I), coprolites (A,C), and predation avoidance (D,G) for insectcarnivory in the fossil record. Examples of carnivory are predation (A–D,G,I,J,M), parasitoidism (F,N), andparasitism (E,H,K,L). A, A coprolite containing early land-plant spores from the Lower Devonian (Lochovian)of Wales; from Edwards et al. (1995). Approximate coprollite length: 1.5 to 2.0 mm. B, A fossil spider spinneret,showing a cluster of attached and detached spigots, from the Middle Devonian (Givetian) of Gilboa, NewYork; from Shear et al. (1989a). Horizontal length (excluding spigots): 0.24 mm. C, Vertebrate coprolitecontaining remains of a cockroach, including wings (w), legs (l) and ovipositor (o) from the middlePennsylvanian Mazon Creek locality of Illinois (FMNH PE 54114). D, From the same locality as (C), theinsect Protodiamphipnoa woodwardi Brongniart (Protorthoptera: Cnemidolestidae), with prominent eyespoton forewing; from Carpenter (1971). E, The enigmatic mecopteroid insect, Strashila incredibilis Rasnitsyn(Mecopteroidea: Strashilidae), from the Upper Jurassic (?Oxfordian) of Transbaikalia, Russia, showingectoparasitic structures such as a chelate hind tarsus, suctorial beak, and absence of wings; from Rasnitsyn(1992). F, The parasitoid wasp Leptephialtites caudatus Rasnitsyn (Hymenoptera: Ephialtidae) from theUpper Jurassic (Kimmeridgian) of Karatau, Kazakhstan, with elongate ovipositor (o) and ovipositor valves(ov); from Rasnitsyn (1975). G, The digger wasp Palaeapis beiboziensis Hong (Hymenoptera: Sphecidae)from the Jurassic-Cretaceous boundary of China, bearing an abdomen with contrasting color-banding andtypical of a Batesian model; from Hong (1984). H, The chewing louse Saurodectes vrsanskyi Rasnitsynand Zherikhin (Phthiraptera: Mallophaga: Saurodectidae), from the Early Cretaceous (Berriasian) of Baissa,Transbaikalia, Russia; from Rasnitsyn and Zherikhin (1999). I, Gut contents of the middle Eocene batPalaeochiropteryx tupaiodon Revilliod (Chiroptera: Palaeochiropterygidae) from Messel, Germany, exhibitingbutterfly scales and other insect fragments; from Richter and Storch (1980). Approximate length of scale atright: 90 µm. J, Robber fly (Diptera: Asilidae) from the Piceance Creek Basin of the Green River Formationof Colorado (USNM 501477), showing raptorial forelegs and mouthparts of a single, dagger-like stylet(arrow). K, The nit (egg) of an undetermined sucking louse (Phithiraptera: Anoplura) on a shaft of mammalianhair, from late Eocene Baltic amber of Germany; from Voigt (1952). L, The mammal-parasitizing flea,Palaeopsylla klebsiana Dampf (Siphonaptera: Hystrichopsyllidae) from the same deposit as (K), showingthe socketed antennae, head comb, and maxillary lever/stylet fascicle (arrow) typical of many ectoparasiticinsects; from Dampf (1910). M, Seed of Tectocarya rhenana (Cornaceae), with exit hole of a seed predator,from the Miocene (Aquitanian) Braunkohle of Germany; from Schmidt et al. (1958). N, A bee cell(Hymenoptera: Stenotritidae) from the Pleistocene of South Australia, displaying a small exit hole of aprobable parasitoid; from Houston (1987). All subfigures are redrawn from original photographs or linedrawings, or are from camera-lucida sketches of specimens. Scale bars: solid = 1.0 cm, striped = 0.1 cm.FMNH = Field Museum of Natural History; USNM = National Museum of Natural History.217

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Baszio’s (2001) report of fish coprolites withterrestrial arthropod remains, including insects, fromthe middle Eocene Messel Deposit of Germany.Mechanisms of Predation Avoidance.—Thereare two broad classes of defense strategies by meansof which terrestrial invertebrates avoid predation.First, there are those that involve resemblance of apotential prey’s external body to another organismor aspect of its ambient environment. Second, thereis physical deterrence based on increasing the levelof difficulty for predation, exemplified by thepresence of spines. Instances of the resemblance ofone organism to another in color, overall form, orbehavior are well documented among moderninsects, and are likewise represented in the fossilrecord. The three basic types of protection aremimicry (both Batesian and Müllerian versions),warning coloration, and crypsis (Price, 1997). Formimicry, the fossil record provides few clues todistinguish the Batesian type, where a palatablemimic resembles an unpalatable model, from theMüllerian type, where both the mimic and the modelare distasteful or otherwise negatively affect thepredator. However, in some fossil examples,determination of mimicry can be approached bymodern analogy (Jarzembowski, 1989). Oneexample is the distinctive color banding on thedigger wasp Palaeapis from the Jurassic-Cretaceousboundary of China (Hong, 1984) (Fig. 2G), whichprobably was a model. Another example, the soldierfly “Stratiomys” from the late Eocene of southernEngland (Jarzembowski, 1976), likely was a mimic.Other instructive examples of the use of color andpattern for protection include the disruptivecoloration—with light and countershaded darkpatterns—on the wings of canopy-inhabitingpaleodictyopteroid insects, presumably to enhanceconcealment when viewed from below, or perhapsrepresenting an example of aposematic (warning)coloration for potential dragonfly predators fromabove (Shear and Kukalová-Peck, 1990).Additionally, there is a fossil record of distinctivewing eyespots, which function in extant insects toinflate apparent size and to startle predators (Preston-Mafham and Preston-Mafham, 1993). Such warningcoloration occurs in a protorthopteran from the LateCarboniferous at Mazon Creek (Carpenter, 1971)(Fig. 2D), as well as throughout the later Mesozoicin the planipennian family Kalligrammatidae, abutterfly-like lineage in which some species boreprominent eyespots (Panfilov, 1968; Jarzembowski,1984). Unlike mimicry and warning coloration,crypsis (camouflage) in the fossil record has faceda more checkered interpretation: one of the bestexamples is the resemblance of Late Carboniferouscockroach wings to seed-fern pinnules, first notedby Scudder (1895). Subsequent observers such asPruvost (1919) and North (1931) expanded thetaxonomic scope of this pinnule similarity byextending it to orthopteroid insects. The resemblanceof cockroach forewings to seed-fern pinnules,however, is more likely attributable to structuralconvergence—based on biomechanical principlesfor the support of planated structures—than toprotective camouflage (Shear and Kukalová-Peck,1990; Jarzembowski, 1994). A similar criticism canbe applied to Fischer’s (1979) argument regardingthe subaerial occurrence of the horseshoe crabEuproops and its proposed concealment amid leafyLepidodendron shoots, whose elongate leavesresemble the animal’s carapace spines. The mostperplexing possible occurrence of crypsis is the leaflikecolor patterns on the tegmina of the MiddleTriassic grasshopper Triassophyllum (Papier, et al.,1997), which resembles certain fern leaves withangiosperm-like venation.Evidence of defenses against predatorsinvolving spines, large size, and protective domicilesoccurs sporadically throughout the fossil record.From the Upper Carboniferous, the herbivorousMazon Creek form Gerarus bore robust andradiating prothoracic spines that evidentlyfunctioned for predation deterrence (Shear andKukalová-Peck, 1990), a feature found in moderntropical grasshoppers and other insects. In the samedeposit and throughout the Late Carboniferous,many terrestrial arthropods achieved great size,including arthropleurid myriapods, the arachnidMegarachne, paleodictyopteroids, protodonatandragonflies, and even Diplura and Zygentoma(silverfish) (Dudley, 1998). Although this gigantismis ultimately attributable to atmospheric oxygen218

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESlevels (Dudley, 1998), a predator-prey arms racecould have developed in which not only insectpredators approached the physiological upper limitsof arthropod size, but so did their prey (Vermeij,1987; Graham et al., 1995). A third mechanism ofphysical defense is the construction of larval cases,which may be used to shield against carnivores. Thismode of larval living is standard in fossil and extantcaddisflies (Sukacheva, 1982), bagworms(Weitschat and Wichard, 1998), coleophorid moths(Labandeira, 2002b), and certain chrysomelidbeetles (Poinar, 1999a). Nevertheless, one of themost intriguing recent investigations focuses on thegeochronology of nocturnal bat predation on largemoths, based on these insects’ ultrasound-detectingtympanal structures, or “ears,” that are located ontheir metathoraces, legs, and mouthparts (Göpfertand Wasserthal, 1999; Fullard and Napoleone,2001). These ultrasound-perceiving tympanalorgans, which alert the moth to the presence ofecholocating insectivorous bats, are prime evidencefor predator avoidance strategies. Although otherevidence suggests that this system existed by themiddle Eocene (Richter and Storch, 1980), pre-Eocene bats are unknown, as is the structure offossil moth tympanal ears.PATTERNS OF CARNIVORYAMONG CONTINENTAL PHYLAOf the approximately 40 extant and extinctphyla, only 12 have invaded the continental realmof freshwater and land habitats, and of these only afew have any significant species-level diversity—principally nematodes, molluscs, annelids, and,spectacularly, arthropods. These invertebrates haveexcelled at life on land (Fig. 3), although all have amarine origin, based on physiological, paleontological,or phylogenetic data (Conway Morris, 1981;Little, 1990; Gordon and Olson, 1995; Fortey andThomas, 1998). These twelve phyla are assigned totwo superphyla (Aguinaldo et al., 1997; Ruiz-Trilloet al., 1999): the Lophotrochozoa, characterized bythe presence or a derivative of the trochophore larva;and the Ecdysozoa, characterized by the keydevelopmental innovation of molting. The origin ofterrestrial carnivory is represented in all twelvephyla, but there are in fact multiple independentorigins within subclades of these phyla. For example,within the Crustacea alone—conventionally rankedas a subphylum of the Arthropoda—the Ostracoda,Branchiopoda, Amphipoda, Isopoda, and severallineages within the Decapoda have each developedcarnivory and either a facultative or obligatecontinental existence (Little, 1983).With the exception of the arthropods andprobably nematodes, all invertebrate continentalphyla have poor fossil records. Intriguingly, theAcanthocephala and Pentostoma, extant membersof which are endoparasitic on mammals, do notoccur in the fossil record, but are thought to haveoriginated in marine habitats during the earlierCambrian, based on sister-group relationships(Conway Morris and Crompton, 1982) and aremarkable early Paleozoic record (Andres, 1989;Walossek and Müller, 1994; Walossek et al., 1994).Similarly, the Tardigrada and Onychophora, thoughtoday free-living and carnivorous, have theirearliest occurrences in the Cambrian marine realm(Müller et al., 1995; Conway Morris, 1998), butappear, respectively, in Late Paleozoic compressiondeposits (Thompson and Jones, 1980; Rolfe et al.,1982) and Late Mesozoic amber (Cooper, 1964;Bertolani and Grimaldi, 2000). For othernonarthropodan phyla, early fossil documentationis limited to either systematically unassignable orquestionably attributed Late Paleozoic finds(Nemertinea, Nematoda, Mollusca) or to mid-Mesozoic earliest occurrences of modern classes(Annelida). The phyla with the poorest and mostrecent (Cenozoic) fossil records are thePlatyhelminthes, Rotifera, and Nematomorpha, andtheir fossil representatives are referable to extantfamilies and even genera (Fig. 3) (Wills, 1993).Structural specializations related to feeding onother animals are varied and have evolved multipletimes in both the Lophotrochozoa and Ecdysozoa.Predation and parasitism were common during theearly terrestrialization of these twelve phyla, as thereis no evidence for the direct transfer of aquatic filterfeeding and epibenthic or infaunal detritivory intoterrestrial habitats (Little, 1990). In this realm, the219


LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESLophotrochozoa display the most varied strategiesfor predation: two phyla use cutaneous absorptionthrough the body wall (platyhelminth flukes andtapeworms, acanthocephalans); one group usesmodifications of a distinctive oral device, the radula(pulmonate gastropods); another group producescopious amounts of pharyngeal mucus to ensnareprey, with or without an oral circlet of buccal ciliaor hooks (oligochaete annelids); hirudinean annelids(leaches) use their distinctive oral armature of slicingjaws in conjunction with a suction apparatus; and alast group (other leaches, nemertineans) housesstylets within an eversible proboscis. By contrast,the Ecdysozoa displays a more uniform pattern ofpredation and parasitism. Proboscides withprotractile chitinous stylets involved in fluid feedingare found in some nematodes (one stylet),tardigrades (two stylets), and larval pentastomes andnematomorphs (three stylets), housed in amouthcone or tubular rostrum (Brusca and Brusca,1990). Other ecdysozoan feeding apparatusesinclude the three or more radially symmetrical jawlikestructures in some nematodes; the two pairs ofslicing, chitinous jaws in onychophorans; thechelicerae and pedipalps of arachnids; and especiallythe primitively adducting/abducting mandibles ofcrustaceans, myriapods, and hexapods. Themouthparts of basal mandibulate hexapods havebeen transformed into a bewildering array ofparticulate and invasive or noninvasive system offluid feeding, including filter-feeding, sponging,siphoning, rasping, lapping, and boring (Schram,1986; Labandeira, 1997a; Walter and Proctor, 1999).Lophotrochozoans.—The Lophotrochozoa—represented by the Rotifera, Acanthocephala,Platyhelminthes, Nemertinea, Mollusca, andAnnelida—have not successfully entered thecontinental realm, and remain overwhelminglymarine. The only significant continental clades arethe gastropod molluscs, and annelids (earthwormsand leeches), many of which occur in fresh waterrather than on land. Extant and extinct specioseclades include bivalve molluscs, cephalopods, andnonpulmonate gastropods, some of which haveonly marginally invaded fresh water. Terrestrializedlophotrochozoan phyla range in species diversityfrom about 750 species for the Acanthocephala toabout 20,000 for pulmonate gastropods.Unlike the Ecdysozoa, the Lophotrochozoalack a unifying synapomorphous or otherwiseobvious feature that could account for theirprevalence in terrestrial habitats. Rather, they havebecome terrestrial and subsequently carnivorousvia four basic strategies, each of which hasoriginated multiple times within and among phyla.The first pathway is via endoparasitism—as incestode and trematode platyhelminths (flukes andtapeworms) and acanthocephalans, all of which←FIGURE 3—The fossil history of carnivory by terrestrial invertebrates, expressed as predation, parasitism,mixed strategies, and microvory. Major sources are Conway Morris (1981), Little (1983), Eisenbeis andWichard (1988), Gray (1988), Brusca and Brusca (1990), Wills (1993), Foelix (1996), Selden (1996),and Poinar and Poinar (1999). Phyla are capitalized; classes or approximately equivalent ranks are inlowercase and standard font; and supraphyletic ranks are in lower case and bold. Geochronology isresolved to stage level and is modified from Gradstein and Ogg (1996); abbreviation: Neog., Neogene.A question mark indicates a questionable geochronologic assignment. Missing epoch designations,from bottom to top: Wenlock and Pridoli (Silurian); Gzelian (Carboniferous), Early and Middle (Triassic);and Pliocene and Pleistocene (Neogene).Notes: (1) The Pentostoma and Acanthocephala lack a body-fossil terrestrial fossil record, thoughboth are internal parasites of mammals and other vertebrate lineages that extend to the Paleogene.(2) The Nemertinea and Nematoda have earliest terrestrial occurrences in the Paleozoic that cannotbe assigned to a higher-level clade (Schram, 1973). (3) See Figure 4 for delimitation of lower taxonomicranks. (4)Microvory is defined as the consumption of live or dead small particulate food and includesunicellular organisms.221

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002lack a significant fossil record—or by theectoparasitism of leeches. A second route is throughsmall size and inconspicuousness in free-livingforms, exemplified by rotifers, turbellarianflatworms, and some nemerteans that occur as soiland litter meiofauna. Third, there has been thedevelopment of an external, mineralized, protectiveshell, as in the vast majority of pulmonategastropods. Fourth, terrestrialization has beenaccompanied by infaunalization within soils andother moist habitats, as typified by relatively largeanimals such as oligochaetes and some nemerteans.The only large, free-living exceptions to the aboveoptions are the shelless pulmonates, or slugs, whichare rarely predaceous, exude copious amounts ofslime, and probably are defended by noxiouscompounds in their tissues (Greene, 1975).1. Acanthocephalans. Acanthocephalans areobligate endoparasites of the alimentary tract ofvertebrates, especially teleost fish, although aminority of species infect arthropods such asterrestrial crustaceans and insects as intermediatehosts (Conway Morris and Crompton 1982).Historically, acanthocephalans have beenconsidered to have their origin among thepseudocoelmate “aschelminth” phyla, although ithas been suggested that they are closely related tothe Priapulida; both of these groups were diverseduring the Cambrian (Conway Morris, 1981; butsee Barnes et al., 1993). Unlike the arthropod-likeonychophorans, tardigrades, and pentastomes (seebelow), acanthocephalans lack a continental or pre-Cenozoic fossil record. It has been speculated(Brusca and Brusca, 1990) that they may have hadan evolutionary trajectory similar to thepentastomes—endoparasitism of primitive marinevertebrates and subsequent evolution parallellingvertebrate phylogeny, including terrestrializationalong with an early, land-based, tetrapod stock(Brusca and Brusca, 1990).2. Rotifers. Of the continental phyla, the onewith the smallest individuals is the Rotifera, mostof which are less than a millimeter long. Rotifersfrequently occur in marine and especially freshwaterhabitats, as well as in moist soils and on the wetsurfaces of mosses. Approximately 1,800 specieshave been described, distinguished by the structureof their oral corona—the metrachronally beatingband of cilia encircling the mouth—and the type oftheir mouthpart-like mastax. They are variouslydetritivorous filter feeders, raptorial predators thtingest small animals, or piercers-and-suckers thatfluid-feed on larger prey. A few are parasitic on othermeiofauna, whereas others have entered intomutualistic associations with crustaceans (Bruscaand Brusca, 1990). With regard to the evolution ofvarious carnivorous feeding strategies, the fossilrecord is uninformative, except inasmuch as it candocument the presence of particular clades, such asthe Monogonta in the Eocene (Lutetian) of SouthAustralia (Southcott and Lange, 1971), and theBdelloidea in the Miocene (Aquitanian) ofDominican amber (Poinar and Ricci, 1992).3. Platyhelminths. The Platyhelminthes, orflatworms, include approximately 20,000 namedspecies and constitute one of the simplestmorphological grades of bilaterian metazoans; theyare triploblastic, acoelomate, and dorsoventrallyflattened. The class Turbellaria is free living andgenerally predaceous, although some species aresymbiotic with other invertebrates; by contrast, theclasses Monogenea (single-host flukes), Trematoda(multiple-host flukes), and Cestoda (tapeworms) areall ecto- or endoparasitic on vertebrates and, to alesser extent, cephalopods and crustaceans (Bruscaand Brusca, 1990). One group of flatworms, theAcoela, were once included in the Turbellaria, butnow are considered to be very basal bilateriansunrelated to the lophotrochozoic Platyhelminthes(Ruiz-Trillo et al., 1999). Feeding amongturbellarians is diverse and involves structuralvariations of a well-musculated pharynx for suction,release of mucus for external prey entrapment,capture of small prey by adoral ciliary action, oruse of a copulatory stylet or oral hooks forimpalement of victims. The fossil record of flukesand tapeworms is nonexistent, and that ofturbellarians is confined to the Neogene, with theirearliest occurrence in the late Miocene of California(Pierce, 1960). Resistant egg capsules and cocoonshave been documented from the European Pleistocene(Frey, 1964). A putative turbellarian from the late222

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESPrecambrian of Alaska (Allison, 1975) is not reliablyassignable to the phylum (Conway-Morris, 1981).4. Nemertineans. Nemertineans, orribbonworms, include about 900 species that rangein length from less than 1 cm to nearly 60 m, someof which occur in terrestrial habitats. Nemertineanssuperficially resemble flatworms but are clearlydifferentiated from them by their feedingspecializations, including a complete digestive tractand an unique, eversible proboscis housed separatelyin a hydrostatically-controlled rhynchocoel. Thisproboscis may possess a single large stylet or manysmaller stylets for lancing prey. Nemertineansgenerally prey on small invertebrates, wrapping theireverted proboscis around a prey item, sometimesalso introducing toxic exudates into the victim byrepeated styletal punctures (Brusca and Brusca,1990). The proboscis draws the subdued or nearlydead prey to the region of the mouth and into therhynchocoel. The history of this distinctive modeof feeding, as documented by nemertinean fossils,is represented only by the Late Carboniferous(Moscovian) Archisymplectes rhothon, discoveredby Schram (1973), which probably lived in brackishor fresh water. There are no other fossil occurrences,although with their distinctive morphology (Stricker,1983), mineralized nemertinean stylets should beidentifiable in the fossil record.5. Molluscs. Of the molluscs, only pulmonategastropods (land snails and slugs) have expandedtheir ecological range to fresh water and land.Pulmonate gastropods display a variety of feedingstrategies, including detritivory, carnivory, and,most commonly, herbivory; but they are united bypossession of a unique and variously modifiedradula (Brusca and Brusca, 1990). The radula is abulbous organ, on the surface of which are severalor more rows of rasp-like, chitinous teeth thatscrape, tear, peel, or otherwise remove tissue foringestion. Carnivorous pulmonates have evolvedmultiple times and include forms with unmodifiedradulae that use slime to entangle prey for eventualconsumption; forms with radular teeth modifiedinto lance-like structures; and others with a radulafashioned into a barbed harpoon functionallysimilar to that of marine cone snails (Vermeij,1987). Pulmonates have a significant fossil record(Gray, 1988; Tracey et al., 1993) and occur inCenozoic fluvial deposits and in amber (Larsson,1978; Roth, 1986), but are particularly abundantin Neogene lake deposits (Cohen, 1989). However,the oldest pulmonate fossils come from severalLate Carboniferous localities in North America(Solem and Yochelson, 1979). The taxonomicassignment of these early specimens has beenchallenged (Tillier et al., 1996) based ontaphonomic arguments and the presence of anunacceptably long temporal gap before the nextoccurrences in the late Mesozoic. Without explicittaxonomic consideration of the fossil material,however, the Late Carboniferous date is accepted.6. Annelids. Both predaceous and parasiticlifestyles occur in the two continental classes ofannelids, Oligochaeta (earthworms) and Hirudinea(leeches). Oligochaetes are commonly associatedwith detritivory, although many—especiallyfreshwater—species are predaceous; leeches, on theother hand, are typically identified withectoparasitism on tetrapods, even though manytropical species are predaceous (Brusca and Brusca,1990). Oligochaete trace fossils have beenmentioned in the literature (Buatois et al., 1998),but predaceous forms are known from body fossilsin Baltic amber (Menge, 1866; Schlee and Glöckner,1978), and later from the Pleistocene of the sameregion (Frey, 1964). Interestingly, the fossil recordof leech-like forms may extend to the Early Silurian,with a specimen reported from an obviously marinebiota in Wisconsin (Mikulic et al., 1985). However,the taxonomic assignment of this specimen istentative (Mikulic et al., 1985), and a more reliablerepresentative of the group occurs in a late Jurassiccompression: Hirudella angusta Münster, from theSolenhofen lithographic limestone of Germany(Kozur, 1970). This latter form may not have beenparasitic according to Conway Morris (1981), andwas probably predaceous.Ecdysozoans.—The Ecdysozoa areconsiderably more diverse than the Lophotrochozoa,probably as a result of their success in terrestrialenvironments (Figs. 3, 4). This success is probablyattributable to their very namesake: ecdysis was an223

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002exaptation for life in a subaerial environment wherewater conservation and osmotic regulation of bodytissues were critically important. Almost all phylaof Ecdysozoa have terrestrial members, namely theNematoda, Nematomorpha, Tardigrada,Onychophora, Pentastoma, and Arthropoda, thelatter synonymous with the Euarthropoda of someworkers. The exception is the more removed andexclusively marine Kinorhyncha + Priapulidasubgroup which was more diverse in the Cambrianthan it is today (Conway Morris, 1998). In terms ofpresent speciosity, most of Ecdysozoan phyla arenot very diverse—there are approximately 90species of Onychophora, 110 species of Pentastoma,and 450 species of Tardigrada—particularlyconsidering that all three have their origins inCambrian marine ecosystems older than 500 Ma.These conservative, basal lineages to the Arthropodahave been interpreted by Gould (1995) as exhibitiingprolonged stasis. Alternatively, drastic habitatchanges in the history of these groups suggests thecontinuous evolution of a durable body plan: freelivingonychophorans and tardigrades have shiftedfrom the marine to the continental realm; parasiticpentastomes have undergone profound shifts in hostsfrom unspecified marine fish or perhaps arthropodsto terrestrial vertebrates (Walossek and Müller, 1994;Müller et al., 1995; Poinar, 1996). Interestingly, threeof these phyla were present in the Cambrian,occupying an epibenthic habitat (Müller et al., 1995);and together with the lophotrochozoanAcanthocephala, they subsequently producedlineages that became continental parasites orpredators. The Nematomorpha may have had asimilar history, but lack an adequate fossil record(Poinar, pers. comm., 2002).Two phyla, nematodes and arthropods, dominatecontinental ecosystems today in terms of abundanceand diversity. Both phyla have invaded everypossible continental habitat, including the interiorsand exteriors of virtually all other organisms. About15,000 species of nematodes have been described,although an estimated one million species probablyexist. Arthropods in particular represent the ultimatesuccess story of life on land in terms of the breadthof their feeding strategies, their numerical andtaxonomic dominance in ecosystems, and estimatesof biomass (Wheeler, 1990; Wilson, 1992; Brusca,2000). Arthropods comprise about one milliondescribed species (Brusca, 2000), and estimates ofthe number of undescribed species range fromseveral million to 80 million species (Erwin andScott, 1980; Gaston and Hudson, 1994).1. Nematodes. Of the twelve continentalinvertebrate phyla, the Nematoda, or roundworms,are second only to the Arthropoda in number ofcarnivorous taxa documented in the fossil record.Most specimens occur in amber from the EarlyCretaceous of Lebanon, the middle Eocene of theBaltic region, and the early Miocene of theDominican Republic (Poinar, 1977; Poinar et al.,1994a, 1994b; Poinar and Poinar, 1999). The earliestdescription of these organisms cites elongate bodieson appendage exocuticle of the large EarlyMississippian scorpion, Gigantoscorpio willsi(Størmer, 1963). These minuscule structures werepreserved in a process akin to permineralization andretain some micromorphology. However, thesestructures are probably not nematodes (Wills, 1993;Poinar et al., 1994a). The earliest credible occurrenceis from the Late Carboniferous (Schram, 1973), andis followed by a long absence of specimens fromthe Permian to Early Cretaceous. During the EarlyCretaceous, the nematode fossil record resumes withseveral amber occurrences (Poinar and Poinar, 1999)and a few reports from compression deposits (Voigt,1957; Dubinina, 1972). The Cretaceous andCenozoic occurrences provide examples ofparasitism on known arthropod hosts, principallyinsects, by members of two most speciosenematode clades: the order Tylenchida of the ClassSecernentea and the order Mermithida of the classAdenophorea (Poinar, 1984a, 1993). These twoclades of nematodes represent two of the fourindependent originations of continental arthropodparasitism by nematodes (Blaxter et al., 1998).Individuals are excellently preserved inconjunction with infected beetles and flies trappedin amber, whose ruptured bodies have allowedrelease of nematodes into adjacent resin (Schleeand Glöckner, 1978; Larsson, 1978; Poinar 1984a).Among the several recorded families of224

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESTylenchida, insect hosts include the Staphylinidae(rove beetles; Poinar and Brodzhinsky, 1985),Mycetophilidae (fungus gnats; Poinar 1991a), andDrosophilidae (pomace flies; Poinar 1984a,1984b). In the Mermithida, reported insect hostsare the Cerambycidae (longhorn beetles; Heyden,1860, 1862), Chrysomelidae (leaf beetles; Voigt,1957), Limoniidae (crane flies; Poinar, 1984c),Culicidae (mosquitoes; Poinar, 1984c; Poinar et al.,1994a), and Chironomidae (aquatic midges;Menge, 1866; Schlee and Glöckner, 1978; Poinaret al., 1994b; Poinar and Milki, 2001). Nematodetrails have been found in the Middle Eocene GreenRiver Formation in Utah (Moussa, 1970); and theparasites have been found in association withspiders from Baltic amber (Poinar, 2000), and inPleistocene mammals, including a horse from theBalkans (Dubinina, 1972) and a ground squirreland a mammoth from Russia (Dubinin, 1948).2. Nematomorphs. Nematomorphs, or hairworms, are thin endoparasitic worms that mayreach a meter in length. In the juvenile phase theyinhabit particular species-specific tissues or organsin arthropods, particularly orthopteroid insects(Brusca and Brusca, 1990). As adults they areaquatic and free-living. The obligately parasiticphylum Nematomorpha includes about 250 extantspecies, and has a sparser historical record thannematodes, with only two known fossil occurrences.Members of the Gordidae are preserved in middleEocene compression deposits in Germany and Italy(Voigt, 1938; Sciacchitano, 1955), although thisassignment has been questioned by Poinar (1999b).By contrast, a member of the Chorododidae has beenfound extruding from the anus of a member of thecockroach family Blatellidae in younger Dominicanamber (Poinar, 1999b).3. Tardigrades. Water bears, or tardigrades, area poorly known phylum of arthropod-likeinvertebrates that occur in the oceans, fresh water,and humid terrestrial environments (Dewel andDewel, 1998). They have a fossil record beginningin Middle Cambrian deposits of northern Siberia(Müller et al., 1995). These specimens, however,are demonstrably marine, cannot be assigned to anypost-Jurassic clade, and have been considered tobelong to a stem-group lineage (Müller et al., 1995).The next and only post-Paleozoic fossil tardigragesare members of the Class Eutardigrada found in LateCretaceous ambers of North America—specificallythe Turonian of New Jersey (Bertolani and Grimaldi,2000) and the Campanian of Manitoba (Cooper,1964)—as well as in the Pleistocene of Italy(Durante and Maucci, 1972). Continental, free-livingtardigrades are herbivores or predatory carnivores.They are significantly different from their Cambrianmarine progenitors, as they possess a fourth, hindpairof legs that are positioned laterally rather thanventrally. However, there are similarities betweenextant marine parasitic Tardigrades and theCambrian forms, such as dorsoventral flattening,the presence of anterior cephalic (suction) discs, aninvoluted mouth, and laterally deployed, robustlyclawed legs (Müller et al., 1995). The Cambrianforms most likely are ancestors of extant marinetardigrades—namely the mostly parasiticarthrotardigrades of the Class Heterotardigrada,which are basal within the phylum and exhibit thegreatest range of morphology (Dewel and Dewel,1998). They represent the stock from which thecontinental lineages evolved, especially thecomparatively diverse Eutardigrada.4. Onychophorans. Velvet worms, oronychophorans, are predaceous, terrestrial,caterpillar-like relatives of arthropods withlobopodous limbs. Their head region has hiddenmouthparts consisting of two pairs of cutting bladesthat serve as jaws and slice tissue in a fore-and-aftmotion (Manton, 1977). Onychophorans currentlyhave a Gondwanan distribution and typically preyon terrestrial invertebrates in tropical to subtropicalenvironments. The fossil record of the phylumextends to the primitive form Aysheaia from theMiddle Cambrian Burgess Shale of Alberta (Gould,1995). A hiatus of approximately 225 million yearsseparates these early marine forms from the nextfossil occurrences, the Late Carboniferous andterrestrial Mazon Creek Helenodora in Illinois(Thompson and Jones, 1980), and an undescribedform from Montceau-les-Mines in France (Rolfeet al., 1982). No Mesozoic specimens are known.Two demonstrably terrestrial taxa, however, are225

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002known from mid-Cenozoic Baltic and Dominicanambers (Poinar, 1996); they exhibit features, suchas the absence of foot portions with claws, whichshow a closer link to the Paleozoic forms than toextant taxa. There is no evidence to indicate thatthis evolutionarily conservative lineage was everdiverse; and it may have been trophicallyoutcompeted by the ecologically equivalent andpredaceous centipedes.4. Pentastomids. As structurally distinctiveendoparasites of tetrapod respiratory systems,pentastomids are a unique phylum withresemblances to lobopods, crustacean naupliuslarvae, and annelids (Brusca and Brusca, 1990).Although recent molecular-based studies indicatecrustacean affinities with a divergence date of 350to 225 Ma (Abele et al., 1989), an alternativeinterpretation that suggests an independent, higherrankingorigin is supported by the occurrence ofmarine pentastomids 510-Ma deposits (Gould,1995). Currently, their typical host is a terrestrialtetrapod, often a reptile, but they are also knownto make use of bird and mammal hosts. Somespecies display a larval stage that temporarilyparasitizes an intermediate vertebrate host. In theseinstances the intermediate host with its pentastomidparasites is, in turn, consumed by the ultimate host,a predaceous tetrapod, completing the cycle. Adultsattach themselves securely to the tissues of thelungs and nasal passageways of their host, suckingblood with a simple mouth and a powerfulpharyngeal pump. Given their current obligateparasitism of terrestrial tetrapods, three occurrencesof pentastomids in upper Cambrian and lowermostOrdovician marine deposits (Andres, 1989;Walossek and Müller, 1994; Walossek et al., 1994)are intriguing, particularly as there are nosubsequent fossil specimens. Considering theirmarine origins during the early Paleozoic and theirpresence among abundant conodont faunas, it hasbeen speculated that they were parasites in the gillchambers of marine chordates or basal vertebrates(Walossek and Müller, 1994). This association thencontinued through the terrestrialization of tetrapodsduring the late Paleozoic, and is manifested todayin their amniote descendants.6. Arthropods. The four arthropod subphylawith carnivorous representatives—Chelicerata,Myriapoda, Crustacea, and Hexapoda—occuramong the earliest of continental deposits, from theuppermost Silurian of northwestern Europe to theMiddle Devonian of northeastern North America(Labandeira et al., 1988; Shear and Selden, 2001).This pattern is quite unlike the terrestrial record ofthe other eleven phyla, which occur moresporadically later in the Paleozoic (Nemertinea,Mollusca, Onychophora), or are confined to the lateMesozoic or Cenozoic (Rotifera, Annelida,Nematoda, Tardigrada), or have a very poor or absentfossil record (Platyhelminthes, Acanthocephala,Nematomorpha, Pentastoma). One explanation forthis is the greater preservability of cursorial,megascopic, and structurally distinctive arthropodsthat bear a chitinous or otherwise mineralizedexoskeleton and could be buried in mid-Paleozoicparalic and continental basinal environments.Several factors, in addition to incumbency (Wilson,1992), indicate that arthropods have always beendiverse and abundant in marine and particularly incontinental ecosystems during the Phanerozoic(Brusca, 2000); and thus their taxonomic dominancein modern ecosystems has a longstanding history.The chelicerates, with the exception of themany detritivorous and parasitic mites, arepredaceous; one apomorphic feature that has arisenin the group is the specialized use of silk by spiders(Araneida) to trap prey. Chelicerates use a pair ofclaw-bearing mouthparts, the chelicerae, forpredation (Snodgrass, 1952). In spiders, thechelicerae are modified into horizontally orvertically oriented structures with terminallyarticulating fangs; ducts in these fangs connect topoison glands (Foelix, 1996). A second, posteriorpair of typically grasping mouthparts, thepedipalps, are used to manipulate prey inScorpionida (scorpions), Ambylypygi (taillesswhipscorpions), and Opiliones (harvestmen). Thecontinental fossil record of chelicerates (Størmer,1969; Selden, 1996) begins in the latest Silurian,and by the mid–Early Carboniferous six orders arerepresented, including spinneret-bearing spiders(Shear et al., 1989a, 1989b). From the middle to226

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESlate Carboniferous an additional seven ordersappear in the record, the best known of which isthe Trigonotarbida (Dunlop, 1994). Only theSchizomida (short-tailed whipscorpions) andPalpigradi (whipscorpions) have their known firstappearances in the Cenozoic (Selden, 1996),although they undoubtedly originated during thePaleozoic based on phylogenetic position.One of the most interesting phenomena in therecord of chelicerate interactions with otheranimals is the preservation of phoretic or parasiticassociations in amber—particularly mites and tickson arthropods and vertebrates. For example, mitesare known to occur ectoparasitically onChironomidae in Late Cretaceous Canadian amberand middle Eocene Baltic amber (Schlee andGlöckner, 1978; Poinar, 1985, 1997); theseassociations have no present-day counterparts. Anolder fossil, however—an erythraeid mite on abiting midge (Ceratopogonidae) in EarlyCretaceous Lebanese amber (Poinar et al., 1994)—represents an association that is still extant. Morerecent associations involving modern taxa aredocumented in lower Miocene Dominican amber,such as erythraeid mites on gracillariid and tineidmoths (Poinar et al., 1991), and macrochelid andother mites phoretically attached to drosophilidflies and leiodid beetles (Poinar and Grimaldi,1990; Poinar, 1993). Parasitic interactionsbetween chelicerates and vertebrates can supportparticular biogeographic hypotheses: for instance,a flightless host-unassociated argasid tick withNeotropical affinities in New Jersey ambersuggests dispersal by birds or pterosaurs(Klompen and Grimaldi, 2001); and listrophoridmites on probable rodent hair in Dominican ambersuggest vicariant mammalian colonization of theAntilles during the Cenozoic (Poinar, 1988).In addition to chelicerates, continentalpredaceous representatives of the Myriapoda,Crustacea, and Hexapoda also appear during thelatest Silurian to Middle Devonian interval (Tasch,1957; Shear and Bonamo, 1988; Greenslade andWhalley, 1986; Labandeira et al., 1988)—althoughearly members of diplopods and scorpions havebeen interpreted to be aquatic (Almond, 1985;Jeram, 1990). The continental Crustacean taxainclude the Phyllopoda (Scourfield, 1926) andMalacostraca (Schram, et al., 1978); continentalmyriapods are represented by the Chilopoda (Shearand Bonamo, 1988); and continental hexapods bythe Parainsecta (Greenslade and Whalley, 1986)and Insecta (Labandeira et al., 1988). Unlike theoverwhelmingly carnivorous Chelicerata and theChilopoda, the crustaceans and hexapods representan eclectic mix of feeding strategies, includingaquatic filter feeding, detritivory, fungivory, andherbivory, as well as carnivory (Schram, 1986;Labandeira, 1997a).PATTERNS OF CARNIVORYAMONG HEXAPODARTHROPODSThe Hexapoda comprises two clades, theParainsecta—consisting of the Collembola(springtails) + Protura (proturans) and probably themore distantly related Diplura—and the Insecta,sister-group to the Parainsecta. The Insecta includethe basal Archaeognatha (bristletails), the morederived Zygentoma, and the yet more derivedPterygota, or winged insects, which represents 99percent of extant hexapod diversity (Kukalová-Peck,1990). There are two major lineages within thePterygota; one lineage, the Palaeoptera, ischaracterized by the inability to fold the wings overthe abdomen, and includes the Odonatoptera,Ephemeroptera (mayflies), and the highly diversePalaeodictyopteroidea of the Paleozoic, nominallydivided into four orders. The other pterygote lineage,the Neoptera, is represented by five majorsupraordinal clades: the Pleconeoptera, consistinglargely of the Plecoptera (stoneflies); its sister-group,the Orthoneoptera, comprised of variousorthopteroid lineages such as the Orthoptera(grasshoppers and crickets); the Blattoneopteracontaining the Blattodea (cockroaches), Mantodea,and Dermaptera (earwigs); the Hemineoptera, sistergroupto the Blattodea, representing the hemipteroidorders Psocoptera (booklice), Phthiraptera (lice),Thysanoptera (thrips), and Hemiptera (bugs, aphidsand relatives); and the very successful227

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Endoneoptera, or Holometabola (Haas andKukalová-Peck, 2001). This latter group ischaracterized by a four-phase metamorphosis— egg,to larva, to pupa, to adult—and includes all the ordersfrom Megaloptera (alderflies and dobsonflies) toHymenoptera (sawflies, wasps, ants, and bees) inFigure 4. Holometabolous insects represent about90% of all insect species, and insects represent 86–92% of all known metazoan species (Wheeler, 1990;Brusca, 2000). They have been the mosttaxonomically, if not numerically dominant animalclade since the earliest Pennsylvanian.Continental carnivorous hexapods have twokey structures: mouthparts and ovipositors. Thegeneralized mandibulate mouthparts of theprimitive hexapod (Kukalová-Peck, 1990) havebeen modified into 34 major types, all extant, eachof which represents considerable to subtlevariations on a basic plan (Labandeira, 1997a). Forexample, the broader category of mandibulocanaliculatemouthparts comprises three of the 34mouthpart classes; each of these three ischaracterized by unique features associated withscythe-shaped mandibles and/or associatedmaxillae that are used by larval holometabolans forpursuit or entrapment of prey (Labandeira, 1997a).Other examples include the structurally diversestylate mouthparts of fluid-feeding ectoparasites(Smith, 1985), and the analogous concealed nectarextraction apparatus used by parasitoid adults at acertain stage in their life cycle (Jervis and Kidd,1986). Similarly, the ovipositor has been used byparasitoids, especially hymenopterans, to inserteggs into hosts such as wood-boring larvae.The record of hexapod carnivory (Fig. 4) showsthree basic trends, each corresponding to the riseof family-level lineages representing the threefunctional feeding groups: predation, parasitoidism,and parasitism. The first trend is the expansion ofinsect predators during the Late Paleozoic. Thesecond is the spectacular diversification ofparasitoid groups during the Middle Jurassic tomid-Cretaceous. And the third trend is the radiationof parasitic groups during the Late Jurassic to mid-Cretaceous, lasting into the Paleogene. Twoancillary patterns should also be noted. One is thebeginning of heavy seed predation during themiddle Cretaceous to Paleogene—although theichnofossil and functional morphological evidencefor seed endosperm consumption extends back tothe Late Carboniferous and Early Permian (Sharov,1973; Scott and Taylor, 1983). Post-Paleozoic seedpredators and parasitoids include stylate piercingand-suckingand mandibulate chewing forms, andare concentrated in three clades: pentatomorphHemiptera, “phytophagan” Coleoptera, andchalcidoid Hymenoptera. A few families ofLepidoptera are notable larval consumers ofendosperm (Janzen, 1971), but lack fossil records.The other secondary trend is reflected in the ancientrecord of palynivory (sporivory and pollinivory;Taylor, 1981), which is analogous to predation ongametophytic seeds, representing consumption ofthe sporophyte phase of a vascular plant. There arefour specific pollinivorous evolutionaryassemblages, defined by plant taxa consumed andassociated consuming insects (Labandeira, 1998b,2000). Palynivory and predation and parasitoidismon seeds will not be considered further.Predators.—The earliest terrestrial predatorswere chilopods and several major chelicerate taxa,with first occurrences during the latest Silurian toMiddle Devonian; several additional lineages appearduring the earlier Pennsylvanian (Fig. 3). Amongthe hexapods, the Odonatoptera—withapproximately 20 family-level taxa—were thedominant airborne predators, including somemembers of the Meganeuridae with wingspans upto 71 cm (Durden, 1988). These large forms wereprobably the ecological equivalent of birds, andcaught their victims on the wing, judging by theirwell-developed raptorial mandibles, leg baskets forsecuring prey, and canted thoracic segments(Rohdendorf and Rasnitsyn, 1980; Brauckmannand Zessin, 1989). Ground-dwelling, LatePaleozoic hexapod predators included severalfamilies of protorthopterans, as revealed by taxa withspinose and raptorial forelegs, and robust andincisiform mandibles. Early holometabolan lineagesof the neuropteroid complex—Megaloptera,Raphidioptera, and Planipennia—originated duringthe earliest Permian (Asselian); virtually all of their228

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESlarvae are assumed to have been predaceous, basedon descendant taxa. There is little direct evidencefor freshwater predation, though the immatures ofOdonatoptera and Ephemeroptera, if aquatic in thePaleozoic (Rohdendorf and Rasnitsyn, 1980), werelikely predaceous.During the Mesozoic, more derivedodonatopteran lineages—representing about 80families and subsumed within the extant high-rankedsubgroup, Odonata—replaced the Late Paleozoicand Early Mesozoic lineages (Bechley, 1996;Bechley et al., 2001). This diverse fauna is preservedin the Solenhofen Limestone of southern Germany(Barthel et al., 1994) and similar Mesozoic deposits.Additionally, insectivorous brachyceran Diptera—such as Rhagionidae (snipe flies), Asilidae,Empididae (dance flies), and others—became aerialpredators during the later Mesozoic. This expansionaccompanied the diversification of dipteranmouthpart types, which involved modifications ofthe mandibular, maxillary, and hypopharyngealstylets, as well as the tracheation and dentition ofthe labellum. These changes provided several newways to secure, process, and consume prey. At thesame time, Mesozoic hexapod predators dwellingon the ground were dominated by the Mantodea,neuropteroids, and the superfamilies Caraboidea,Staphylinoidea, and Elateroidea among theColeoptera. Later, during the Cretaceous, vespoidwasps and ants became important ground predators.Among the first well-substantiated predators infreshwater ecosystems were nepomorph Hemipteraand caraboid Coleoptera during the latest Triassicto early Jurassic; these clades undoubtedly subsistedon a trophic base of other, perhaps herbivorous,arthropods (Ponomarenko, 1996). This Mesozoicarray of freshwater and terrestrial predatorsessentially persisted into the Cenozoic, judging bythe geochronologic ranges of major family-level taxa(Labandeira, 1994). Recent evidence, however,indicates a significant diminution of herbivoroustaxa at taxonomically lower levels at theCretaceous-Paleocene boundary in the WesternInterior (Labandeira et al., 2002; cf. Labandeiraand Sepkoski, 1993). This would have significantlyimpacted the herbivore component of food webstructure; and based on theoretical (Dunne et al.,2002) and empirical (Godfray et al., 1999) data,should translate to significant decreases intaxonomic diversity. This impoverishmentprobably cascaded upward to trophically dependentpredators, parasitoids, and parasites.Parasitoids.—One of the biggest transitions incontinental carnivory was the rather rapidappearance of major parasitoid taxa during theMiddle Jurassic to mid-Cretaceous. Of the 74parasitoid families recorded in Figure4, 57 (or 77percent) are Hymenoptera, 12 (16 percent) areDiptera, and the remaining 5 (7 percent) representPlanipennia, Coleoptera, and Lepidoptera. In theHymenoptera, the overwhelming presence ofsawflies (Symphyta) plus seven superfamilies ofapocritan and aculeate lineages (Gauld and Boulton,1988), as well as the superfamilies Tabanoidea,Asiloidea, Empidoidea, Muscoidea, and a few otherlineages in the brachyceran Diptera, truly records amajor shift in the mode of carnivory within terrestrialecosystems (Fig. 4). In the post-Triassic fossilrecord, evidence for parasitism comes fromtaxonomic assignments (Rasnitsyn, 1988), thepresence of small parasitoid emergence holes infossil pupae (Bown et al., 1997) (Fig. 2N), orteratologies of adult insects in amber (Poinar andPoinar, 1999), such as an exceptional case of abraconid larva preserved in the act of emerging froman adult ant (Poinar and Miller, 2002). The transitionto parasitoidism was accomplished by modificationsin the reproductive biology of numerousholometabolous lineages (Godfray, 1994). For theHymenoptera, an important structural change wasthe transformation of a laterally compressed,sawtooth ovipositor—designed for slicing planttissues—into a needle-like piercing or drillingmechanism with a circular cross-section. Thisstructure, initially used for inserting eggs intowood, was transformed into a device for puncturingendophytic larvae—a key innovation for basalparasitoid taxa (Rasnitsyn, 1988). An analogousstructure is the highly telescoped posteriorabdomen of the Diptera. The second importantdevelopment in early parasitoids was a shift inlarval food source, from less nutritious plant tissues229


LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESto more proteinaceous animal tissues, which wereoften supplied by the adult (Malyshev, 1968). Apart of this general shift was the appearance of thelarval cleptoparasitoid habit, in which an insectdevelops at the expense of another (host) insect byconsuming its food supply and depriving the hostof nutrition, thus assuring its eventual death(Eggleton and Belshaw, 1992).Several trophic pathways leading to theevolution of the parasitoid life-habit have beenidentified (Eggleton and Belshaw, 1992). Withinthe Hymenoptera, parasitoidism apparentlyoriginated only once, and most likely grew out ofmycophagy in an earlier Mesozoic symphytanlineage, such as the siricoid sawflies (Whitfield,1998). Siricoid larvae generally inhabit dead woodand feed on symbiotic fungi in tunnels, but someconstituent taxa, lacking such a food supply, deposittheir eggs on other insects that occur in the samehabitat. Thus the initial step toward parasitoidismcould be the usurpation of another species’ fungalfood supply, resulting in the death of the starvedlarval host. Subsequently, there would be a shiftfrom the introduced larva feeding on symbioticfungi to consumption of the host itself. Amodification of this scenario has been proposedby Crowson (1981) to account for the origin ofparasitoids in wood-associated beetle lineages. Onthe other hand, in multiple independent lineages ofDiptera, vertebrate parasitoidism has originated bystep-wise conversion from carrion saprophagy. Thesequence of this transition would have been (i)←FIGURE 4—At left and on the following two pages is represented the fossil history of carnivory for 359hexapod families, expressed as predation, parasitoidism, parasitism, and mixed strategies. Taxonomicranks are orders (capitalized), suborders or superfamilies (italicized, lower-case), and families (regular font,lower-case). A conservative approach is taken whereby assignments to predator, parasitoid, or parasitecategories in extant families are based on a plurality of carnivorous relative to non-carnivorous memberspecies. Trophic assignments of fossil families were inferred from multiple criteria, including relatedness toor inclusion within modern clades with known ecologies, functional morphology, ecological evidence suchas gut contents, or trace-fossil context (see Fig. 2). Many family-level taxa with sporadic or otherwise limitedoccurrences of carnivory among overwhelmingly herbivorous, fungivorous, or detritivorous confamilialswere excluded. Sporivory and pollinivory were not included. Only carnivorous taxa with fossil records areincluded. Geochronology is resolved to stage level and is modified from Gradstein and Ogg (1996), withupdated Lagerstätten ages such as Baltic amber (Ritzkowski, 1997), the Florissant biota (Evanoff et al.,2001), and Dominican amber (Iturralde-Vinent, 2001). Major sources for feeding-type assignments areClausen (1940, carnivorous insects), Janzen (1971, seed predators), Carpenter (1971, 1992, fossil groups),Askew (1979, parasites), Richards and Davies (1977, all insects), Crowson (1981, Coleoptera), Merritt andCummings (1984, aquatic insects), Hagen (1987, insect predators), Gauld and Bolton (1988, Hymenoptera),Lehane (1991, hematophagous insects), Eggleton and Belshaw (1992, parasitoids), Godfray (1994,parasitoids), Schuh and Slater (1995, heteropterous Hemiptera), and Arnett (2000, all insects). Fossiloccurrences are from Bechley (1996), Bechley et al. (2001), Labandeira (1994) and subsequent updates,Poinar and Poinar (1999), and Rasnitsyn (2000; 2002, pers. comm.). Taxonomic classification follows Bechley(1996) and Bechley et al. (2001) for Odonatoptera, Schuh and Slater (1995) for heteropteran Hemiptera,and Naumann et al. (1990) for all other insect orders. Missing epoch designations, from bottom to top:Gzelian (Carboniferous), Early and Middle (Triassic), and Pliocene and Pleistocene (Neogene). Suprafamilialor subordinal designations are: A, Protanisoptera; B, Leptopodomorpha; C, Mantispoidea; D, Tenebrionoidea;E, “Phytophaga”; F, Empidoidea; G, Sciomyzoidea; H, Tephritoidea; I, Ichneumonoidea; and J, Cynipoidea.Asterisks at the bottom of the ranges of the Isotomidae (Collembola) indicate an earliest occurrence in theLochkovian Stage of the Devonian; for the Osmylidae and Nymphidae (Planipennia), first appearances inthe Ladinian Stage of the Triassic; and for the Staphylinidae (Coleoptera) the oldest is from the CarnianStage of the Triassic. A question mark indicates a questionable trophic assignment for the interval indicated.231

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 4 (cont.)232

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESFIGURE 4 (cont.)233

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002initially attacking live but dying animals, orsubsisting on their wounds (facultativeectoparasitism), followed by (ii) initiating wounds(obligate ectoparasitism), and finally (iii) invadingthe host itself and causing its death (parasitoidism)(Zumpt, 1965). In lineages such as carabid beetles,parasitoidism may have developed from predation.Erwin and Erwin (1976) proposed a sequence inwhich an ancestral ground beetle species was apolyphagous predator on insects; it wascharacterized by an active larva that emerged fromeggs deposited in the general habitat of its eventualvictims, and it pupated in an area removed from itsprey. The spatial confinement of a particular preyspecies resulted in monophagy of the beetle larvalpredator, and the reduction of prey seeking resultedin metabolic savings. This in turn favoredconsumption of fewer prey items and pupationwithin the immediate host environment. Anyenvironmental change that provoked host dispersalwould favor a shift from a larval predator to a small,immobile, parasitoid living within a single host.This explanation works best for spatially confinedhosts, such as wood-boring or gall-inhabitinglarvae. There are at least 100 known originationsof parasitoidism in insects, most by way ofmycophagy, then saprophagy, and the fewestthrough predation and herbivory (Eggleton andBelshaw, 1992). Parasitoidism itself has in turnevolved into other trophic modes such asprovisioning polyphagy and predation in somecoleopteran, dipteran, and hymenopteran lineages(Eggleton and Belshaw, 1992).Parasites.—When compared to the relativelygood fossil record of parasitoids, the record of insectparasitism is poor: only about 30 percent of parasitefamilies have fossil occurrences. The reason for thisdeficiency is that, whereas the adult forms ofparasitoids are overwhelmingly free-living and socan enter the fossil record through a variety ofenvironments (many are small wasps), the adults ofmost parasitic organisms are attached to theirvertebrate hosts and are destroyed by the processesof carcass decomposition. Consequently, only inexceptional environments, such as resin flows, areinsect parasites preserved (Lewis and Grimaldi,1997). Fine-grained compression depositsoccasionally contain fossil parasites, such as theenigmatic mecopteroid Strashila from the UpperJurassic of the Siberia (Rasnitsyn, 1992), the possibleflea Saurophthirus from the Lower Cretaceous ofTransbaikalia (Ponomarenko, 1976; but seeLakshminarayana et al., 1984), and the flea Tarwiniafrom the Lower Cretaceous of New South Wales,Australia (Riek, 1970; Jell and Duncan, 1986; butsee Smit, 1972). All three taxa display ectoparasiticfeatures, including hypognathous, piercing-andsucking mouthparts, compact antennae, long legswith grasping claws, and distensible abdomens(Rasnitsyn, 1992). Both Strashila and Saurophthirusare dorsoventrally flattened and morphologicallyconvergent on bat parasites such as the hemipteranPolyctenidae, the dipteran Nycteribiidae, and eventhe dermapteran Arixeniidae (Popham, 1962;Ponomarenko, 1976), which feed on bat wingmembranes and occupy sparse to tomentose hair;the resemblance suggests a similar feeding style forthe Mesozoic fossils, possibly on pterosaurs. In theAustralian flea, of indeterminate family assignment,lateral body compression indicates life in dense furnear the body core, rather than on the wingmembrane; and they likely inhabited mammalianhosts such as monotremes, triconodonts, ormultituberculates. Perplexingly, a literal reading ofthis evidence points to an apparently missing fossilrecord, judging from the relationship of theseparasitic clades to their free-living (and sorepresented by more fossils) sister clades (Lyal,1985; Kristensen, 1999). These phylogeneticrelationships, in conjunction with the fossil recordsof the better-preserved clade, indicate splittingevents ranging from the Early Permian to the LateJurassic (Downes, 1971b; Lyal, 1985; Kinzelbachand Pohl, 1994; Fang et al., 1999). The conflictingdata suggest either that the parasite clades are goodexamples of fossil ghost taxa, or that they originatedmuch later within the crown groups of their sisterclades(e.g., Chalwatzis et al., 1996).The origins of hematophagy during theMesozoic, and the likely initial hosts for severalmajor parasitic insect clades, have generatedconsiderable discussion. The best evidence comes234

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESfrom the record of dipteran parasites, particularlythe Culicomorpha, whose adults are free-rangingand not bound to particular host individuals. Withinthe Culicomorpha, the Ceratopogonidae are wellknown to feed on vertebrate blood and hemolymph;this feeding behavior is considered a plesiomorphicfeature within the family (Downes, 1971a; Borkent,1995). The ceratopogonid fossil record extendsback to the Early Cretaceous and, notably, includesseveral extant genera that feed on reptilian blood(Desportes, 1942; Wirth and Hubert, 1989). In lightof this feeding capacity of ceratopogonids, Borkent(1995) examined the correlations between palpaland antennal sensillae, stylet dentition, and whetherthe host organisms were invertebrates, birds, or largetetrapods—the latter determined by downwindplumes of host-generated carbon dioxide and heat(Rowley and Cornford, 1972; McKeever et al., 1991;Blackwell et al., 1992). He concluded that two orthree species of biting midges from Canadian UpperCretaceous amber were parasitizing largevertebrates, most likely co-occurring hadrosaurs (butsee Szadziewski, 1996). Two of the threeceratopogonid genera he identified are extant andsuck blood from large vertebrates, including reptilesand mammals. This plesiomorphic feature of theCeratopogonidae may extend to the rest of thesuperfamily Culicomorpha, appearing in the formof functional piercing mouthparts in females,subsequently lost in lineages such as theChaoboridae (phantom midges), Chironomidae, andDixidae. The Tanyderidae (primitive crane flies), oneof the oldest dipteran lineages, had stylatemouthparts during the Late Triassic and probablyfed on blood (Kalugina, 1991; Borkent, 1995). Basedon this evidence, some authors suggest that stylatemouthparts consisting of a fascicle of 3 to 5 styletsmay be plesiomorphic for the Culicomorpha, witha modest expansion of the labellar pad as a laterdevelopment (Downes and Colless, 1967; Downes,1971b; Fang et al., 1999). According to these authorsit is unlikely that there was repeated evolution ofthe same mouthpart type. However, Pawlowski etal. (1996) provide a molecular phylogenetic analysisindicating at least two independent derivations ofblood feeding within the Culicomorpha; theyconclude that nectarivory is the primitive feedingstrategy. Similarly, Glukhova (1989) proposed thatthe basal Culicomorpha and related taxa initiallywere saprophagous on carrion, and then, throughthe production of salivary proteolytic enzymes, wereable to feed on the external secretions of livingvertebrates, followed by subsistence on blood. Suchan evolutionary shift, however, seems moreappropriate for the origin of advanced blood feedingin (muscoid) flies (Szadziewski, 1996), and certainSoutheast Asian blood-sucking moths (Bänziger,1975). Another hypothesis is that insecthematophagy originated from inquiline associationsin vertebrate nests by a shift from nest detritivoryof epidermis, feces, sebum, and other exudates toblood feeding (Balashov, 1999). This hypothesis ismost relevant to the origin of ectoparasitism in thePhthiraptera, since many members of the closelyrelated Psocoptera are nest associates of vertebratehosts (Waage, 1979).There are two ecologically different approachesto blood feeding. The first is solenophagy—foundin the Culicidae (mosquitoes), Reduviidae (assassinbugs), Cimicidae (bed bugs), and anopluranPhthiraptera (sucking lice)—in which blood vesselssuch as capillaries are punctured with needle-likestylets (Brown, 1989). In contrast, telmophagy—found in the Ceratopogonidae, Tabanidae(horseflies), Muscidae (house flies), and ticks—ischaracterized by the creation of pockets of laceratedtissue that has been slashed by stylets, which inthese organisms are often modified into serratedblades and used in opposition by scissor-likemovements. The insect then drinks the resultingpool of blood and lymph (Brown, 1989). Membersof the Glossinidae (tsetse flies) have uniquepiercing-and-sucking mouthparts and anintermediate strategy for obtaining blood. Althoughglossinid fossils occur in the North American lateEocene (Cockerell, 1917; Grimaldi, 1992),glossinids today are biogeographically restrictedtoday to sub-Saharan Africa. They are currentlymajor vectors for infectious trypanosomiases thataffect ungulates (nagana) and humans (sleepingsickness) (Lehane, 1991). It has been suggested thatnagana-like diseases may have been vectored by235

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002North American glossinids feeding on titanotheres(Tasnádi-Kubacska, 1962) or on nonhominidprimates (Lambrecht, 1993) during the midCenozoic, and subsequently transferred to hominidsin eastern Africa. Similarly, certain fossil species inthe North American Paleogene belong to thehematophagous culicid genus Culex (a mosquito),and a parallel argument has been advanced for theNeogene transferral of plasmodia, filiariae, andarborviruses such as malaria to early hominids inAfrica (Capasso, 1993).THE ROLE OF CARNIVORY INECOSYSTEM EVOLUTIONIn the preceding overview of the fossil recordof continental invertebrate carnivory, six classesof evidence were discussed (Fig. 2), and distinctivepatterns of trophic evolution were cited among the12 invertebrate phyla (Fig. 3) and 359 insectfamilies (Fig. 4) that occur in freshwater andterrestrial ecosystems. Equally important toaddress, however, are the effects that these patternshad on terrestrial ecosystems. The data cited here,in conjunction with other features of the terrestrialfossil record, generate four questions that arecentral to the role of arthropod carnivory interrestrial and freshwater ecosystem evolution.Which came first: herbivory or carnivory?—There are two hypotheses regarding the trophicstructure of the earliest well-preserved terrestrialecosystems from the latest Silurian to the mid-Devonian. One view holds that these earlycommunities were dominated almost exclusivelyby detritivorous and predatory animals, similar tomodern litter communities (Gray and Shear, 1992).An alternative is that herbivores are as old ascarnivores and detritivores, and all three trophicmodes were present a system analogous to extantdepauperate communities on grass. In addition tocryptogams, bryophytes, early land plants, and thefungus Prototaxites, the animal component of theseearly communities consisted of trigonotarbids,pseudoscorpions, spiders, mites, chilopods,arthropleurids, collembolans, and archaeognathans.Among these arthropods, piercing-and-suckingpachygnathoid mites from the Rhynie Chert havebeen associated with feeding on live algal protoplasts(Shear and Selden, 2001). Collembolans from thesame deposit have been identified as members ofthe Isotomidae (Greenslade and Whalley, 1986), andpossibly the Neanuridae (Hopkin, 1997), modernrepresentatives of which subsist on live fungal sporesand hyphae, dead vascular plant tissue, and detritus.As evidence regarding plant-insect associationsthere are spore-rich coprolites, typically containingmultiple spore taxa and exhibiting stereotyped fecalshapes and sizes, from latest Silurian and EarlyDevonian deposits of the Welsh Borderland(Edwards, et al., 1995). Coprolites containingvascular tissues are also known from the LowerDevonian of Gaspé, Canada (Banks and Colthart,1993). These associations, as well as thosedocumented in the Rhynie Chert (Kevan et al.,1975), indicate an important role for palynivores,exploiting highly nutritious, land-plant sporeprotoplasts. Additionally, evidence for surfacefeeding on plant axes—complete with proliferationof anomalous reaction tissue (Kevan et al., 1975;Banks and Colthart, 1993)—and the occurrence ofstylate punctures surrounded by conical regions ofhistolytic tissue (Banks and Colthart, 1993)—similar to damage produced by modern piercingand-suckingmicroarthropods—indicate a diverseherbivore community of both chewing andpiercing-and-sucking forms. Thus the balance ofevidence supports that the earliest terrestrialcommunities consisted of herbivores, as well asdetritivores and predators—a pattern unlike thedelayed herbivory hypothesis of Vermeij andLindberg (2000) for the marine realm.Is prey death or host accommodation thedominant mode in arthropod carnivory?—Theintroduction to this chapter presents the canonicalview of carnivory at all times and places: predatorsdispatch and consume prey in the manner oflionesses and antelope. This view has been thedominant research paradigm in marine invertebratepaleobiology (Vermeij, 1987; Kelley and Hansen,1996), and justifiably so. Such an approach tomarine invertebrate predation probably reflectsactual differences in the marine versus continental236

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESversions of carnivory—though there is a smallliterature on the more poorly known associationsof marine invertebrate parasites. By contrast,research on the history of terrestrial invertebratecarnivory has been dominated by studies ofparasitoids and parasites, and far fewer of predators,particularly during the past 40 years (Askew, 1979;Price, 1980; Godfray, 1994; Schmid-Hempel, 1998).The work on parasites and especially parasitoids hasbeen a primary source of data for ecological theoriesof community structure, particularly valuable forstudies of tritrophic interactions (Abrahamson andWeis, 1997), for instance, and for food-web analyses(Godfray et al., 1999). In fact, many investigationscenter around whether the parasitic habit promotesdiversification (Wiegmann et al., 1993; Cronin andAbrahamson, 2001), and if individual phylogeniesof hosts and their parasites reflect coevolvedassociations (Lyal, 1987; Kim, 1988; Hafner andNadler, 1988). Recent biological work and the fossilhistory of parasitoids and parasites supports theconclusion that more accommodationistmechanisms, typified by the coexistence of host andconsumer, characterize the post-Triassic history ofcarnivory in freshwater and terrestrial ecosystems.What was trophic structure like beforeparasites and parasitoids?—There is no directfossil evidence for insect parasitism orparasitoidism on arthropods in continentalecosystems before the Early Jurassic. Evidencefrom the phylogenetic relationships of noninsectanparasites and parasitoids would strongly supportthe presence during this interval of freshwater orterrestrial members of the phyla Acanthocephala,Platyhelminthes, Annelida, Nematoda,Nematomorpha, and Pentastoma, whose hostspresumably were vertebrates, arachnids, or insects.The absence of direct evidence for parasitic insectsis countered by phylogenetic relationshipsindicating that some of these life-habits originatedduring the Permian to Late Jurassic (Waage, 1979),in, for example, the Phthiraptera (Kim, 1988),Siphonaptera (Traub and Starcke, 1980), andStrepsiptera (Kinzelbach and Pohl, 1994). Theseorders could have existed for all that time as ghostlineages alongside their earlier occurring, freelivingsister groups, but this is unlikely. In otherclades dominated by parasitoids, which have aconsiderably better fossil record than parasites, therecord reveals explosive diversification for theHymenoptera and to a lesser extent the Dipteraduring the Middle Jurassic to mid-Cretaceous. Thuscontinental invertebrate carnivory from the latestSilurian to Early Jurassic can be characterized asentirely or near-entirely predaceous in nature. Thefossil record indicates that predators were the soleor overwhelmingly dominant carnivores during theearliest Pennsylvanian to the mid-Cretaceous, asevidenced by the dominance of Odonatoptera,some protorthopteran lineages, Titanoptera,gerromorph Hemiptera, Raphidioptera,Megaloptera, Planipennia, and caraboid andstaphylinoid Coleoptera (Fig. 4). The presence ofonly predator guilds of carnivores in continentalecosystems must have had a profound effect onthe regulation of herbivorous and otherheterotrophic organisms (see Eggleton andBelshaw, 1992), particularly since considerableevidence shows that modern trophic webs arestrongly regulated by parasites and parasitoids(DeAngelis, 1992; Godfray et al., 1999).Accordingly, in a world without these two keycarnivore guilds, one would expect simpler foodwebs with linear trophic pathways (Dunne et al.,2002), and the absence of lineages with typicaldefensive strategies to ward off parasites andparasitoids, such as small size, concealment,exoskeletal or other types of mechanical isolation,and behaviors to reduce the risk of attack such asparental or group care and protection. The latterwould include the caste-based, regulated societiesseen today in social insects—and this may explainthe rise of sociality in termites, ants, wasps, andbees (Wilson, 1992) during the Mesozoicdiversification of parasitoids.Is there trophic-level selectivity at massextinctions?—Among terrestrial carnivorousinvertebrates, insects have the best fossil record.The insect fossil record is characterized by twoimportant extinction events. The first is deducedfrom the geochronologic distributions of order-rankclades. This was a catastrophic extinction at the237

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002end of the Permian Period, in which the PaleozoicInsect Fauna was replaced by the Modern InsectFauna through the loss of a high percentage of highrankedclades (Labandeira and Sepkoski, 1993).Among carnivores, this event affected onlypredators, as the other carnivore guilds presumablywere absent. The second event, at the Cretaceous-Paleogene boundary, is more surprising, givenprevious studies that failed to detect any family-levelsignal beyond background rates (Ponomarenko,1988; Jarzembowski, 1989; Labandeira andSepkoski, 1993). A recent examination, however,has taken the explicitly ecological approach ofevaluating insect-mediated leaf damage patterns inwell-preserved and diverse floras across the K-Tboundary in North Dakota (Labandeira et al., 2002).This study demonstrated a pronounced qualitativeand quantitative diminution of insect herbivory(especially among specialists) at the boundary,indicating a significant deterioration of the herbivoreportion of terrestrial food webs. Such a major trophicalteration would have induced extinctions, based onour present knowledge of food-web perturbationfrom both theoretical modeling and empirical fieldobservations (DeAngelis, 1992; Dunne et al., 2002).After the expansion of parasitoidism and parasitismduring the later Mesozoic, this sudden reduction inhost resource probably trophically affected majorcarnivore guilds in continental food chains.SUMMARYThe transfer of energy in terrestrial ecosystems,both in fresh water and on land, culminates in theprocess of carnivory or the consumption of animalsby animals. This consumption may be rapid as inthe case of predators and their prey, or it may beprolonged as in the case of parasitoids with theirrelatively small hosts, and parasites with theircomparatively large hosts. While these three typesof carnivory are essential to the structure of foodwebs in modern ecosystems, their detection in thefossil record presents challenges. Six types ofpaleobiological evidence can reveal the presenceof carnivory: taxonomic affiliation, structural andfunctional attributes (especially mouthparts andovipositors), organismic damage, gut contents,coprolites, and mechanisms of predationavoidance. This evidence typically occurs in wellpreserveddeposits.Twelve invertebrate phyla have adopted acontinental carnivorous existence, representingpredatory, parasitoid, and parasitic life-habits. Ofthese phyla, the most speciose, ecologically varied,and abundant are nematodes and arthropods, bothof which have representative members practicingeach of the three modes of carnivory. Thishyperdiversity is related, particularly in arthropods,to such features as small size, ecdysis, ovipositors,and mouthpart morphology. This latter can involveeither a buccal stylet apparatus or the varied, oftenco-opted multielement mouthpart structures thathave been modified into several functional modesfor obtaining nutrition—including chewing,piercing-and-sucking, and boring. In the terrestrialinvertebrate world, accommodationist strategiessuch as forms of carnivory that result in theprolonged survival of a host, are probably morecommon than the rapid dispatching of prey.Members of all twelve continental invertebratephyla are documented in the fossil record, andarthropods are first represented by predatorscoexisting with detritivores and herbivores. Theyhave their first appearances in the earliest terrestrialbiotas from the latest Silurian to Middle Devonian,but it is not until the Middle Jurassic to mid-Cretaceous, approximately 250 million years later,that parasitoids and parasites are recorded as fossils,even though their phylogenetic relationships mayindicate greater antiquity. The diversification ofthese two functional groups of carnivores addedconsiderable complexity to food webs and may havebeen associated with the origin of insect sociality asa means of warding off attack. Although insect massextinctions at the end of the Permian had a profoundeffect on terrestrial invertebrate predation, recentevidence indicates that an end-Cretaceous extinctionmay have had less profound but more integrativeconsequences on food webs by affecting dependentpredators, parasitoids, and parasites. Thus anintriguing aim for further study is to understand the238

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESfabric of terrestrial food webs in a world in whichpredators were the only carnivores, preceding theemergence of parasitoids and parasites.ACKNOWLEDGMENTSAppreciation is extended to Finnegan Marsh,who formatted the figures. I thank the followingpeople for reviews: Steven Hasiotis, EdJarzembowski, Elena Lukashevich, MihailMostovski, George Poinar, Jr., Alex Rasnitsyn, andDena Smith. This is contribution 92 of the Evolutionof Terrestrial Ecosystems consortium at the NationalMuseum of Natural History. This article is dedicatedto the deceased paleoentomologist VladimirZherikhin of the Paleontological Institute of theRussian Academy of Sciences (Moscow), asystematist and a pioneer in the use of phylogenetic,taxic, and associational data in insect paleoecology.REFERENCESABEL, O. 1935. Vorzeitliche Lebensspuren. Gustav Fischer, Jena, 644 p.ABELE, L. G., W. KIM, AND B. E. FELGENHAUER. 1989. Molecular evidence for inclusion of the phylum Pentastomidain the Crustacea. Molecular Biology and Evolution, 6:685–691.ABRAHAMSON, W. G., AND A. E. WEIS. 1997. The Evolutionary Ecology of a Tritrophic-level Interaction: Goldenrod,the Stem Gall maker and its Natural Enemies. Princeton University Press, Princeton, NJ.AGUINALDO, A. M. A, J. M. TURBEVILLE, L. S. LINFORD, M. C. RIVERA, J. R. GAREY, R. A. RAFF, AND J. A. LAKE.1997. Evidence for a clade of nematodes, arthropods and other moulting animals. Nature, 387:489–493.ALLISON, C. W. 1975. Primitive fossil flatworm from Alaska: new evidence bearing on ancestry of the Metazoa.Geology, 3:649–652.ALMOND, J. E. 1985. The Silurian-Devonian fossil record of the Myriapoda. Philosophical Transactions of theRoyal Society of London, B, 309:227–237.ANDRES, D. 1989. Phosphatisierte Fossilien aus dem unteren Ordoviz von Südschweden. BerlinerGeowissenschaftliche Abhandlungen A, 106:9–19.ARNETT, R. H., JR. 2000. American Insects, Second Edition. CRC Press, Boca Raton, FL, 1003 p.ASKEW, R. W. 1979. Parasitic Insects. Heinemann, London, 316 p.BACHOFEN-ECHT, A. 1934. Beobachtungen über im Bernstein vorkommende Spinnengewebe. Biologia Generalis,10:179–184.BALASHOV, Y. S. 1999. Evolution of haematophagy in insects and ticks. Entomological Review, 79:943–954.BANKS, H. P., AND B. J. COLTHART. 1993. Plant-animal-fungal interactions in Early Devonian trimerophytes fromGaspé, Canada. American Journal of Botany, 80:992–1001.BÄNZIGER, H. 1975. Skin-piercing blood-sucking moths I: ecological and ethological studies on Calpe eustrigata(Lepid., Noctuidae). Acta Tropica, 32:125–144.BARNES, R. S. K., P. CALOW, AND P. J. W. OLIVE. 1993. The Invertebrates: A New Synthesis, Second Edition.Blackwell, London, 488 p.BARTHEL, K. W., N. H. M. SWINBURNE, AND S. CONWAY MORRIS. 1994. Solnhofen: A Study in Mesozoic Palaeontology.Cambridge University Press, Cambridge, UK, 236 p.BECHLEY, G. 1996. Morphologische Untersuchungen am Flügelgeäder der Rezenten Libellen und derenStammgruppenvertreter (Insecta; Pterygota; Odonata). Petalura, 2:1–402.BECHLEY, G., A. NEL, X. MARTÍNEZ-DELCLÒS, E. JARZEMBOWSI, R. CORAM, D. MARTILL, G. FLECK, F. ESCUILLIÉ, M. M.WISSHA, AND M. MAISCH. 2001. A revision and phylogenetic study of Mesozoic Aeschnoptera, with descriptionof numerous new taxa (Insecta: Odonata: Anisoptera). Neue Paläontologische Abhandlungen, 4:1–219.BERTOLANI, R., AND D. GRIMALDI. 2000. A new eutardigrade (Tardigrada: Milnesiidae) in amber from the UpperCretaceous (Turonian) of New Jersey, p. 103–110. In D. A. Grimaldi (ed.), Studies on Fossils in Amber, withParticular Reference to the Cretaceous of New Jersey. Backhuys, Leiden.239

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BETZ, O. 1996. Function and evolution of the adhesion-capture apparatus of Stenus species (Coleoptera,Staphylinidae). Zoomorphology, 116:15–34.BLACKWELL, A., A. J. MORDUE, M. R. YOUNG, AND W. MORDUE. 1992. Morphology of the antennae of two speciesof biting midge: Culicoides impuncatatus (Goetghebuer) and Culicoides nubeculosus (Meigen) (Diptera,Ceratopogonidae). Journal of Morphology, 213:85–103.BLAXTER, M. L., P. DE LEY, J. R. GAREY, L. X. LIU, P. SCHELDEMAN, A. VIERSTRAETE, J. R. VANFLETEREN, L. Y.MACKEY, M. DORRIS, L. M. FRISSE, J. T. VIDA, AND W. K. THOMAS. 1998. A molecular framework for thephylum Nematoda. Nature, 392:71–75.BORKENT, A. 1995. Biting Midges in the Cretaceous Amber of North America (Diptera: Ceratopogonidae).Backhuys, Leiden, 237 p.BOWN, T. M., S. T. HASIOTIS, J. F. GENISE, F. MALDONADO, AND E. M. BROUWERS. 1997. Trace fossils of Hymenopteraand other insects and paleoenvironments of the Claron Formation (Paleocene and Eocene), southwesternUtah. United States Geological Survey Bulletin, 2153-C:41–58.BRAUCKMANN, C., AND W. ZESSIN. 1989. Neue Meganeuridae aus dem Namurium von Hagen-Vorhalle (BRD) und diePhylogenie der Meganisoptera (Insecta, Odonata). Deutsche Entomologische Zeitschrift, N.F., 36:177–215.BROWN, S. J. 1989. Pathological consequences of feeding by hematophagous arthropods: comparison of feedingstrategies, p. 4–15. In C. J. Jones and R. E. Williams (eds.), Physiological Interactions between HematophagousArthropods and Their Vertebrate Hosts. Entomological Society of America, Lanham, MD.BRUES, C. T. 1972. Insects, Food and Ecology. Dover Publications, New York, 466 p.BRUSCA, R. C. 2000. Unraveling the history of arthropod biodiversification. Annals of the Missouri BotanicalGarden, 87:13–25.BRUSCA, R. C., AND G. J. BRUSCA. 1990. Invertebrates. Sinauer, Sunderland, MA, 922 p.BUATOIS, L., M. G MÁNGANO, J. F. GENISE, AND T. N. TAYLOR. 1998. The ichnologic record of the continentalinvertebrate invasion: evolutionary trends in environmental expansion, ecospace utilization, and behavioralcomplexity. Palaios, 13:217–240.CAPASSO, L. 1993. Fossil mosquitoes and the spread of infectious diseases in man’s ancestors. Journal ofPaleopathology, 3:171–201.CARPENTER, F. M. 1971. Adaptations among Paleozoic insects, p. 1236–1251. In E. Yochelson (ed.), Proceedingsof the First North American Paleontological Convention. Allen Press, Lawrence, KS.CARPENTER, F. M. 1992. Volume 3: Superclass Hexapoda. In R. L. Kaesler, E. Brosius, J.Keim, and J. Priesner(eds.), Treatise on Invertebrate Paleontology, Part R, Arthropoda 4. Geological Society of America andUniversity of Kansas, Lawrence, KS, 655 p.CHALWATZIS, N., J. HAUF, P. Y. VAN DER PEER, R. KINZELBACH, AND F. K. ZIMMERMANN. 1996. 18S ribosomal RNAgenes of insects: primary structure of the genes and molecular phylogeny of the Holometabola. Annals of theEntomological Society of America, 89:789–803.CLAUSEN, C. P. 1940. Entomophagous Insects. McGraw-Hill, New York, 688 p.COCKERELL, T. D. A. 1917. Glossina and the extinction of Tertiary mammals. Nature, 103:265.COHEN, A. S. 1989. The taphonomy of gastropod shell accumulations in large lakes: an example from LakeTanganyika, Africa. Paleobiology, 15:26–45.CONWAY MORRIS, S. 1981. Parasites and the fossil record. Parasitology, 82:489–509.CONWAY MORRIS, S. 1998. The Crucible of Creation: The Burgess Shale and the Rise of Animals. Oxford UniversityPress, Oxford, UK, 242 p.CONWAY MORRIS, S., AND D. W. T. CROMPTON. 1982. The origins and evolution of the Acanthocephala. BiologicalReviews, 57:85–115.COOPER, K. W. 1964. The first fossil tardigrade: Beorn leggi Cooper, from Cretaceous amber. Psyche, 71:41–48.CORBET, P. S. 1999. Dragonflies—Behaviour and Ecology of Odonata. Harley, Colchester, UK, 829 p.CRONIN, J. T., AND W. G. ABRAHAMSON. 2001. Do parasitoids diversify in response to host-plant shifts by herbivorousinsects? Ecological Entomology, 26:347–355.CROWSON, R. A. 1981. The Biology of the Coleoptera. Academic Press, New York, 802 p.DAMPF, A. 1910. Palaeopsylla klebsiana n. sp., ein fossiler Flöh aus dem baltischen Bernstein. Schriften derPhysikalisch-ökonomischen Gesellschaft zu Köningsberg in Prussia, 51:248–259.240

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESDEANGELIS, D. L. 1992. Dynamics of Nutrient Cycling and Food Webs. Chapman & Hall, London, 270 p.DESPORTES, C. 1942. Forcipomyia velox Winn. et Sycorax silacea Curtis, vecteurs d’Icosiella neglecta (Diesing):filaire commune de la grenouille verte. Annales de Parasitologie, 19:53–68.DEWEL, R. A., AND W. C. DEWEL. 1998. The place of tardigrades in arthropod evolution, p. 109–123. In R. A.Fortey and R. H. Thomas (eds.), Arthropod Relationships. Chapman and Hall, London.DOWNES, J. A. 1971a. Feeding and mating in the insectivorous Ceratopogoninae (Diptera). Memoirs of theEntomological Society of Canada, 104:1–62.DOWNES, J. A. 1971b. The ecology of blood-sucking Diptera: an evolutionary perspective, p. 232–258. In A. M.Fallis (ed.), Ecology and Physiology of Parasites. University of Toronto Press, Toronto.DOWNES, J. A., AND D. H. COLLESS. 1967. Mouthparts of the biting and blood-sucking type in Tanyderidae andChironomidae (Diptera). Nature, 214:1355–1356.DUBININ, B. V. 1948. Discovery of a Pleistocene louse (Anoplura) and nematodes during the study of corpses ofIndigirsk fossil gophers. Doklady Akademiia Nauk SSSR, N.S., 62:417–420 [in Russian].DUBININA, M. N. 1972. Nematoda Alfortia edentatus (Looss, 1900) iz Kischechinka Verkhne pleistotsenovoiloshadi. Parasitologiya, 6:441–443 [in Russian].DUDLEY, R. 1998. Atmospheric oxygen, giant Paleozoic insects and the evolution of aerial locomotor performance.Journal of Experimental Biology, 201:1043–1050.DUNLOP, J. A. 1994. The palaeobiology of the Writhlington trigontarbid arachnid. Proceedings of the Geologists’Association, 105:287–296.DUNNE, J. A., R. J. WILLIAMS, AND N. D. MARTINEZ. 2002. Network topology and species loss in food webs:robustness increases with connectance. Santa Fe Institute Working Paper, 02-03-013:1–17.DURANTE, M. V., AND W. MAUCCI. 1972. Descrizione di Hybisibius (Isohyps.) basalovoi sp. nov. e altre notizie sutardigradi del Veronese. Memorie del Museo Civico di Storia Naturale di Verona, 20:275–281.DURDEN, C. J. 1988. Hamilton insect fauna, p. 117–124. In G. Mapes and R. H. Mapes (eds.), Regional Geologyand Paleontology of Upper Paleozoic Hamilton Quarry Area in Southeastern Kansas. Geological Society ofAmerica, Lawrence, KS.EDWARDS, D., P. A. SELDEN, J. B. RICHARDSON, AND L. AXE. 1995. Coprolites as evidence for plant-animal interactionin Siluro-Devonian terrestrial ecosystems. Nature, 377:329–331.EGGLETON, P., AND R. BELSHAW. 1992. Insect parasitoids: an evolutionary overview. Proceedings of the RoyalSociety of London, B, 337:1–20.EISENBEIS, G, AND W. WICHARD. 1988. An Atlas on the Biology of Soil Arthropods. Springer Verlag, Berlin, 437 p.ERWIN, T. L., AND L. J. M. ERWIN. 1976. Relationships of predaceous beetles to tropical forest wood decay, Part II:The natural history of Neotropical Eurycoleus macularis Chevrolat (Carabidae: Lebiini) and its implicationsin the evolution of ectoparasitoidism. Biotropica, 8:215–224.ERWIN, T. L., AND J. C. SCOTT. 1980. Seasonal and size patterns, trophic structure and richness of Coleoptera in thetropical arboreal ecosystem: the fauna of the tree Luehea seemannii Triana and Planch in the Canal Zone ofPanama. Coleopterists Bulletin, 34:305–322.EVANOFF, E., W. C. MCINTOSH, AND P. C. MORPHEY. 2001. Stratigraphic summary and 40 Ar/ 39 Ar geochronology ofthe Florissant Formation, Colorado, p. 1–16. In E. Evanoff, K. M. Gregory-Wodzicki, and K. R. Johnson(eds.), Fossil Flora and Stratigraphy of the Florissant Formation, Colorado. Proceedings of the Denver Museumof Nature and Science, 4.EVANS, H. C. 1989. Mycopathogens of insects of epigeal and aerial habitats, p. 205–238. In N. Wilding, N. M.Collins, P. M. Hammond, and J. F. Webber (eds.), Insect-Fungus Interactions. Academic Press, London.FANG, Q. Q., S. MCKEEVER, AND F. E. FRENCH. 1999. Cladistic analysis of tabanids (Diptera: Tabanidae) usingmicroscopic characters of the mouthparts. Memoirs on Entomology International, 14:355–366.FISCHER, D. C. 1979. Evidence for subaerial activity of Euproops danae (Merostomata, Xiphosurida), p. 379–447.In M. H. Nitecki (ed.), Mazon Creek Fossils. Academic Press, New York.FOELIX, R. F. 1996. Biology of Spiders, Second Edition. Oxford University Press, New York, 330 p.FORTEY, R. A., AND R. H. THOMAS. 1998. Arthropod Relationships. Chapman & Hall, London, 383 p.FRANZEN, J. L. 1985. Exceptional preservation of Eocene vertebrates in the lake deposit at Grube Messel (WestGermany). Philosophical Transactions of the Royal Society of London, B, 311:181–186.241

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FREY, D. G. 1964. Remains of animals in Quaternary lake and bog sediments and their interpretation. Archiv fürHydrobiologie, Beihefte, 2:1–114.FUCHS, G.-V. 1975. Die Gewinnung von Pollen und Nektar bei Käfern. Natur und Museum, 104:45–54.FULLARD, J. H., AND N. NAPEOLEONE. 2001. Diel flight periodicity and the evolution of auditory defences in theMacrolepidoptera. Animal Behaviour, 62:349–368.GASTON, K. J., AND E. HUDSON. 1994. Regional patterns of diversity and estimates of global insect species richness.Biodiversity and Conservation, 3:493–500.GAULD, I., AND B. BOLTON. 1988. The Hymenoptera. Oxford University Press, Oxford, UK, 332 p.GENISE, J. F. 1995. Upper Cretaceous trace fossils in permineralized plant remains from Patagonia, Argentina.Ichnos, 3:287–299.GERHARD, S., AND W. RIETSCHEL. 1968. Ein Stück Bernstein und seine Einschlusse. Natur und Museum, 98:515–520.GLUKHOVA, V. M. 1989. Blood-sucking midges of the genera Culicoides and Forcipomyia (Ceratopogonidae).Fauna of the USSR, 139,3,5a, 408. Nauka, Leningrad.GODFRAY, H. C. J. 1994. Parasitoids: Behavioral and Evolutionary Ecology. Princeton University Press, Princeton,NJ, 473 p.GODFRAY, H. C. J., O. T. LEWIS, AND J. MEMMOTT. 1999. Studying insect diversity in the tropics. PhilosophicalTransactions of the Royal Society of London, B, 354:1811–1824.GÖPFERT, M. C., AND L. T. WASSERTHAL. 1999. Hearing with the mouthparts: behavioural response and the structuralbasis of ultrasound perception in acherontiine hawkmoths. Journal of Experimental Biology, 202:909–918.GORDON, M. S., AND E. C. OLSON. 1995. Invasions of the Land: The Transitions of Organisms from Aquatic toTerrestrial Life. Columbia University Press, New York, 312 p.GOULD, S. J. 1995. Of tongue worms, velvet worms, and water bears. Natural History, 103:1–15.GRADSTEIN, F., AND J. OGG. 1996. A Phanerozoic time scale. Episodes, 19:3–5.GRAHAM, J. B., R. DUDLEY, N. AGUILAR, AND C. GANS. 1995. Implications of the late Palaeozoic oxygen pulse forphysiology and evolution. Nature, 375:117–120.GRAY, J. 1988. Evolution of the freshwater ecosystem: the fossil record. Palaeogeography, Palaeoclimatology,Palaeoecology, 61:1–214.GRAY, J., AND W. SHEAR. 1992. Early life on land. American Scientist, 80:444–456.GREENE, A. 1975. Biology of the five species of Cychrini (Coleoptera: Carabidae) n the steppe region of southeasternWashington. Melanderia, 19:1–43.GREENSLADE, P. J. M., AND P. E. S. WHALLEY. 1986. The systematic position of Rhyniella praecursor Hirst &Maulik (Collembola), the earliest known hexapod, p. 319–323. In R. Dallai (ed.), Second International Seminaron Apterygota. University of Siena, Italy.GRIMALDI, D. A. 1992. Vicariance biogeography, geographic extinctions, and the North American tsetse flies, p. 178–204.In M. J. Novacek and Q. D. Wheeler (eds.), Extinction and Phylogeny. Columbia University Press, New York.GRONENBERG, W. 1996. The trap-jaw mechanism in the dacetine ants Daceton armigerum and Strumigenys sp.Journal of Experimental Biology, 99:2021–2033.HAFNER, M. S., AND S. A. NADLER. 1988. Phylogenetic trees support the coevolution of parasites and their hosts.Nature, 332:258–259.HAAS, F., AND J. KUKALOVÁ-PECK. 2001. Dermaptera hindwing structure and folding: new evidence for familial, ordinaland superordinal relationships within the Neoptera (Insecta). European Journal of Entomology, 98:445–509.HAGEN, K. S. 1987. Nutritional ecology of terrestrial insect predators, p. 533–577. In F. Slansky, Jr., and J. G. Rodriguez(eds.), Nutritional Ecology of Insects, Mites, Spiders, and Related Invertebrates. John Wiley & Sons, New York.HANNIBAL, J. T., AND R. M. FELDMANN. 1988. Millipeds from late Paleozoic limestones at Hamilton, Kansas, p.125–131. In G. Mapes and R. H. Mapes (eds.), Regional Geology and Paleontology of Upper PaleozoicHamilton Quarry Area in Southeastern Kansas. Geological Society of America, Lawrence, KS.VON HEYDEN, C. 1860. Mermis antiqua, ein fossiler Eingeweidewurm. Entomologische Zeitung, 21:38.VON HEYDEN, C. 1862. Gliederthiere aus der Braunkohle des Niederrhein’s der Wetterau und der Rohn.Palaeontographica, 10:62–82.HONG, Y.-C. 1984. New fossil insects of Laiyang Group from Laiyang Basin, Shandong Province. ProfessionalPapers in Stratigraphy and Palaeontology, 11:31–41 [in Chinese with English abstract].242

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESHOPKIN, S. P. 1997. The Biology of the Springtails (Insecta: Collembola). Oxford University Press, Oxford, UK, 330 p.HOUSTON, T. F. 1987. Fossil brood cells of stenotritid bees (Hymenoptera: Apoidea) from the Pleistocene of SouthAustralia. Transactions of the Royal Society of South Australia, 3:93–97.ITURRALDE-VINENT, M. A. 2001. Geology of the amber-bearing deposits of the Greater Antilles. Caribbean Journalof Science, 37:141–167.JANZEN, D. H. 1971. Seed predation by animals. Annual Review of Ecology and Systematics, 2:465–492.JANZEN, D. H. 1978. The ecology and evolutionary biology of seed chemistry as related to seed predation. In J. B.Harborne (ed.), Biochemical Aspects of Plant and Animal Coevolution. Annals and Proceedings of thePhytochemical Society of Europe, 15:163–206.JARZEMBOWSKI, E. A. 1976. Report of Easter Field Meeting: the Lower Tertiaries of the Isle of Wight, 27–31.III.1975.Tertiary Research, 1:11–16.JARZEMBOWSKI, E. A. 1984. Early Cretaceous insects from southern England. Modern Geology, 9:71–93.JARZEMBOWSKI, E. A. 1989. Cretaceous insect extinction. Mesozoic Research, 2:25–28.JARZEMBOWSKI, E. A. 1994. Fossil cockroaches or pinnule insects? Proceedings of the Geologists’ Association,105:305–311.JELL, P. A., AND P. M. DUNCAN. 1986. Invertebrates, mainly insects, from the freshwater Lower Cretaceous,Koonwarra Fossil Bed (Korumburra Group), South Gippsland, Victoria. Memoir of the Association ofAustralasian Palaeontologists, 3:111–205.JUNIPER, B. E. 1986. The path to plant carnivory, p. 195–218. In B. E. Juniper and T. R. E. Southwood (eds.),Insects and the Plant Surface. Edward Arnold, London.JERAM, A. J. 1990. Book-lungs in a Lower Carboniferous scorpion. Nature, 343:360–361.JERVIS, M. A., AND N. A. C. KIDD. 1986. Host-feeding strategies in hymenopteran parasitoids. Biological Reviews,61:395–434.KALUGINA, N. S. 1991. New Mesozoic Simuliidae and Leptoconopidae and blood-sucking origin in lower dipterans.Paleontologicheskii Zhurnal, 1991:69–70 [in Russian].KELLEY, P. H., AND T. A. HANSEN. 1996. Naticid gastropod prey selectivity through time and the hypothesis ofescalation. Palaios, 11:437–445.KEVAN P. G., W. G. CHALONER, AND D. B. O. SAVILE. 1975. Interrelationships of early terrestrial arthropods andplants. Palaeontology, 18:391–417.KIM, K. C. 1988. Evolutionary parallelism in Anoplura and eutherian mammals. In M. W. Service (ed.), Biosystematics ofHaematophagous Insects. Systematics Association Special Volume, 37:91–114. Oxford University Press, Oxford, UK.KINZELBACH, R. K., AND H. POHL. 1994. The fossil Strepsiptera (Insecta: Strepsiptera). Annals of the EntomologicalSociety of America, 87:59–70.KLOMPEN, H., AND D. A. GRIMALDI. 2001. First Mesozoic record of a parasitiform mite: a larval argasid tick inCretaceous amber (Acari: Ixodida: Argasidae). Annals of the Entomological Society of America, 94:10–15.KORN, W. 1943. Die Muskulatur des Kopfes und des Thorax von Myrmeleon europaeus und ihre Metamorphose.Zoologischer Jahrbücher (Anatomie), 68:273–330.KOZUR, H. 1970. Fossile Hirudinea aus dem Oberjura von Bayern. Lethaia, 3:225–232.KRISTENSEN, N. P. 1999. Phylogeny of endopterygote insects, the most successful lineage of living organisms.European Journal of Entomology, 96:237–253.KUKALOVÁ-PECK, J. 1990. Fossil history and the evolution of hexapod structures, p. 141–179. In I. D. Naumann,P. B. Carne, J. U. F. Lawrence, E. S. Nielsen, J. P. Spradbery, R. W. Taylor, M. J. Whitten, and M. J. Littlejohn(eds.), The Insects of Australia, Second Edition, 1. Cornell University Press, Ithaca, NY.KUTSCHER, M., AND J. KOTEJA. 2000a. Trace fossils in Bitterfeld amber: excrements or detritus? Polskie PismoEntomologiczne, 69:175–178.KUTSCHER, M., AND J. KOTEJA. 2000b. Coccids and aphids (Hemiptera: Coccinea, Aphidinea), prey of ants(Hymenoptera: Formicidae): evidence from Bitterfeld amber. Polskie Pismo Entomologiczne, 69:179–185.LABANDEIRA, C. C. 1990. Use of a Phenetic Analysis of Recent hexapod Mouthparts for the Distribution of HexapodFood Resource Guilds in the Fossil Record. Ph.D. Dissertation, University of Chicago, 1186 p.LABANDEIRA, C. C. 1994. A compendium of fossil insect families. Milwaukee Public Museum Contributions inBiology and Geology, 88:1–71.243

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002LABANDEIRA, C. C. 1997a. Insect mouthparts: ascertaining the paleobiology of insect feeding strategies. AnnualReview of Ecology and Systematics, 28:153–193.LABANDEIRA, C. C. 1997b. Permian pollen eating. Science, 277:1422–1423.LABANDEIRA, C. C. 1998a. The role of insects in Late Jurassic to middle Cretaceous ecosystems, p. 105–124. In S.G. Lucas, J. I. Kirkland, and J. W. Estep (eds.), Lower and Middle Cretaceous Terrestrial Ecosystems. NewMexico Museum of Natural History and Science Bulletin, 14.LABANDEIRA, C. C. 1998b. Early history of arthropod and vascular plant associations. Annual Review of Earthand Planetary Sciences, 26:329–377.LABANDEIRA, C. C. 2000. The paleobiology of pollination and its precursors, p. 233–269. In R. Gastaldo and W.DiMichele (eds.), Phanerozoic Terrestrial Ecosystems. Paleontological Society Papers, 6.LABANDEIRA, C. C. 2002a. The history of associations between plants and animals, p. 26–74, 248–261. In C.Herrera and O. Pellmyr (eds.), Plant-Animal Interactions. Blackwell Science, Oxford, UK.LABANDEIRA, C. C. 2002b. The paleobiology of middle Eocene plant-insect associations from the Pacific Northwest:a preliminary report. Rocky Mountain Geology, 37 (in press).LABANDEIRA, C. C., AND J. J. SEPKOSKI, JR. 1993. Insect diversity in the fossil record. Science, 261:310–315.LABANDEIRA, C. C., B. S. BEALL, AND F. M. HUEBER. 1988. Early insect diversification: evidence from a LowerDevonian bristletail from Québec. Science, 242:913–916.LABANDEIRA, C. C., K. R. JOHNSON, AND P. WILF. 2002. Impact of the terminal Cretaceous event on plant-insectassociations. Proceedings of the National Academy of Sciences, USA, 99:2061–2066.LAKSHMINARAYANA, K. V., M. S. MANI, AND W. D. EICHLER. 1984. On the relationship of the fossil flea Saurophthiruslongipes Ponomarenko (Siphonaptera: Insecta). Records of the Zoological Survey of India, 81:43–48.LAMBRECHT, F. L. 1993. Tsetse flies and trypanosomiasis during the American Tertiary. National GeographicSociety Research Reports, 21:241–249.LARSSON, S. G. 1978. Baltic amber—a palaeobiological study. Entomonograph, 1:1–192.LEHANE, M. J. 1991. Biology of Blood-Sucking Insects. Harper-Collins, London, 288 p.LEWIS, R. E., AND D. A. GRIMALDI. 1997. A pulicid flea in Miocene amber from the Dominican Republic (Insecta:Siphonaptera: Pulicidae). American Museum Novitates, 3205:1–9.LITTLE, C. 1983. The Colonisation of Land: Origins and Adaptations of Terrestrial Animals. Cambridge UniversityPress, Cambridge, UK, 290 p.LITTLE, C. 1990. The Terrestrial Invasion: An Ecophysiological Approach to the Origins of Land Animals.Cambridge University Press, Cambridge, UK, 304 p.LYAL, C. H. C. 1985. Phylogeny and classification of the Psocodea, with special reference to the lice (Psocodea:Phthiraptera). Systematic Entomology, 10:145–165.LYAL, C. H. C. 1987. Co-evolution of trichodectid lice (Insecta: Phthiraptera) and their mammalian hosts. Journalof Natural History, 21:1–28.MALYSHEV, S. I. 1968. Genesis of the Hymenoptera and the Phases of their Evolution. Methuen, London, 319 p.MANTON, S. M. 1977. The Arthropoda: Habits, Functional Morphology and Evolution. Oxford University Press,Oxford UK, 527 p.MCKEEVER, S., D. V. HAGAN, AND W. L. GROGAN. 1991. Comparative study of mouthparts of ten species of predaceous midgesof the tribe Ceratopogonini (Diptera: Ceratopogonidae). Annals of the Entomological Society of America, 84:93–106.MENGE, A. 1866. Ueber ein Rhipidipteron und einiger andere im Bernstein eingeschlossene Tiere. Schriften derNaturforschenden Gesellschaft in Danzig, N.F., 1:1–8.MERRITT, R. W., AND K. W. CUMMINS (eds.). 1984. An Introduction to the Aquatic Insects. Second edition. Kendall/Hunt, Dubuque, IA, 722 p.MIKUILÁS, R., Z. DVORAK, AND I. PEK. 1998. Lamniporichnus vulgaris igen. et isp. nov.: traces of insect larvae instone fruits of hackberry (Celtis) from the Miocene and Pliocene of the Czech Republic. Journal of theCzech Geological Society, 43:277–280.MIKULIC, D. G., D. E. G. BRIGGS, AND J. E. KLUESSENDORF. 1985. A Silurian soft-bodied biota. Science, 228:715–717.MOUSSA, M. T. 1970. Nematode fossil trails from the Green River Formation (Eocene) in the Uinta Basin, Utah.Journal of Paleontology, 44:304–307.244

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESMÜLLER, K. J., D. WALOSSEK, AND A. ZAKHAROV. 1995. “Orsten” type phosphatized soft-integument preservationand a new record from the Middle Cambrian Kuonamka Formation in Siberia. Neues Jahrbüch für Geologieund Paläontologie Abhandlungen, 197:101–118.NAUMANN, I. D., P. B. CARNE, J. F. LAWRENCE, E. S. NIELSEN, J. P. SPRADBERY, R. W. TAYLOR, M. J. WHITTEN, AND M.J. LITTLEJOHN (eds.). 1990. The Insects of Australia: A Textbook for Students and Research Workers, SecondEdition. Cornell University Press, Ithaca, NY, 1137 p.NORTH, F. J. 1931. Insect-life in the coal forests, with special reference to South Wales. Transactions of theCardiff Naturalists Society, 62:16–44.PANFILOV, D. V. 1968. Kalligrammatids (Neuroptera, Kalligrammatidae) in Jurassic deposits of Karatau, p. 166–175. InB. B. Rohdendorf (ed.), Jurassic Insects of Karatau. Izdatel’stvo Nauka, Moscow [in Russian].PAPIER, F., A. NEL, L. GRAUVOGEL-STAMM, AND J.-C. GALL. 1997. La plus ancienne sauterelle Tettigoniidae,Orthoptera (Trias, NE France): Mimétisme ou exaptation? Paläontologische Zeitschrift, 71:71–77.PAWLOWSKI, J., R. SZADZIEWSKI, D. KMIECIAK, J. FAHRNI, AND G. BITTAR. 1996. Phylogeny of the infraorder Culicomorpha(Diptera: Nematocera) based on 28S RNA gene sequences. Systematic Entomology, 21:167–178.PETRUNKEVITCH, A. 1942. A study of amber spiders. Transactions of the Connecticut Academy of Arts and Sciences,34:119–464.PIERCE, W. D. 1960. Silicified Turbellaria from Calico Mountains nodules. Bulletin of the South California Academyof Sciences, 59:138–143.POINAR, G. O., JR. 1977. Fossil nematodes from Mexican amber. Nematologica, 23:232–238.POINAR, G. O., JR. 1984a. Fossil evidence of nematode parasitism. Revue de Nématologie, 7:201–203.POINAR, G. O., JR. 1984b. First fossil record of parasitism by insect parasitic Tylenchida (Allantonematidae:Nematoda). Journal of Parasitology, 70:306–308.POINAR, G. O., JR. 1984c. Heydenius dominicus n. sp. (Nematoda: Mermithidae), a fossil parasite from theDominican amber. Journal of Nematology, 16:371–375.POINAR, G. O., JR. 1985. Fossil evidence of insect parasitism by mites. International Journal of Acarology, 11:37–38.POINAR, G. O., JR. 1987. Fossil evidence of spider parasitism by Ichneumonidae. Journal of Arachnology, 14:399–400.POINAR, G. O., JR. 1988. Hair in Dominican amber: evidence for Tertiary land mammals in the Antilles.Experientia, 44:88–89.POINAR, G. O., JR. 1991. The mycetophagous and entomophagous stages of Iotonchium californicum n. sp.(Iotonchiidae: Tylenchida). Revue de Nématologie, 14:565–580.POINAR, G. O., JR. 1993. Insects in amber. Annual Review of Entomology, 46:145–159.POINAR, G. O., JR. 1996. Fossil velvet worms in Baltic and Dominican amber: onychophoran evolution andbiogeography. Science, 273:1370–1371.POINAR, G. O., JR. 1999a. Chrysomelidae in fossil resin: behavioural inferences, p. 1–16. In M. L. Cox (ed.),Advances in Chrysomelidae Biology 1. Backhuys, Leiden.POINAR, G. O., JR. 1999b. Paleochordodes protus n. g., n. sp. (Nematomorpha, Chordodidae), parasites of a fossilcockroach, with a critical examination of other fossil hairworms and helminthes of extant cockroaches (Insecta:Blattaria). Invertebrate Biology, 118:109–115.POINAR, G. O., JR. 2000. Heydenius araneus n. sp. (Nematoda: Mermithidae), a parasite of a fossil spider, with anexamination of fossil helminths from extant spiders. Invertebrate Biology, 119:388–393.POINAR, G. O., JR., AND J. BRODZHINSKY. 1985. Fossil evidence of nematode (Tylenchida) parasitism in Staphylinidae(Coleoptera). Nematologica, 32:353–355.POINAR, G. O., JR., AND D. A. GRIMALDI. 1990. Fossil and extant macrochelid mites (Acari: Macrochelidae) phoreticon drosophilid flies (Diptera: Drosophilidae). Journal of the New York Entomological Society, 98:88–92.POINAR, G. O., JR., AND R. MILKI. 2001. Lebanese Amber: The Oldest Insect Ecosystem in Fossilized Resin.Oregon State University Press, Corvallis, OR, 96 p.POINAR, G. O., JR., AND J. C. MILLER. 2002. First fossil record of endoparasitism of adult ants (Formicidae:Hymenoptera) by Braconidae (Hymenoptera). Annals of the Entomological Society of America, 95:41–43.POINAR, G. O., JR., AND R. POINAR. 1999. The Amber Forest: A Reconstruction of a Vanished World. PrincetonUniversity Press, Princeton, NJ, 239 p.245

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002POINAR, G. O., JR., AND C. RICCI. 1992. Bdelloid rotifers in Dominican amber: evidence for parthenogeneticcontinuity. Experientia, 48:408–410.POINAR, G. O., JR., A. ACRA, AND F. ACRA. 1994a. Animal-animal parasitism in Lebanese amber. Medical ScienceResearch, 22:159.POINAR, G. O., JR., A. ACRA, AND F. ACRA. 1994b. Earliest fossil nematode (Mermithidae) in Cretaceous Lebaneseamber. Fundamental and Applied Nematology, 17:475–477.POINAR, G. O., JR., A. E. TREAT, AND R. V. SOUTHCOTT. 1991. Mite parasitism of moths: examples of paleosymbiosisin Dominican amber. Experientia, 47:210–212.PONOMARENKO, A. G. 1976. A new insect from the Cretaceous of Transbaikalia, a possible parasite of pterosaurians.Paleontological Journal, 1976:339–343.PONOMARENKO, A. G. (ed.). 1988. The Cretaceous Biocoenotic Crisis in the Evolution of Insects. USSR Academyof Sciences, Moscow, 230 p. [in Russian].PONOMARENKO, A. G. 1996. Evolution of continental aquatic ecosystems. Paleontological Journal, 30:705–709.POPHAM, E. J. 1962. The anatomy related to the feeding habits of Arixenia and Hemimerus (Dermaptera).Proceedings of the Zoological Society of London, 139:429–450.PRESTON-MAFHAM, R., AND K. PRESTON-MAFHAM. 1993. The Encyclopedia of Land Invertebrate Behaviour.Massachusetts Institute of Technology Press, Cambridge, MA, 320 p.PRICE, P. W. 1980. Evolutionary Biology of Parasites. Princeton University Press, Princeton, NJ.PRICE, P. W. 1997. Insect Ecology, Third Edition. Wiley, New York, 874 p.PROCTOR, M., P. YEO, AND A. LACK. 1996. The Natural History of Pollination. Timber Press, Portland, OR, 479 p.PRUVOST, P. 1919. Le faune continentale du terrain houiller du Nord de la France. Mémorie pour Servir àl’Explication de la Carte Géologique Détaillé de la France. Imprimerie Nationale, Paris, 584 p.PRITCHARD G. 1976. Further observations on the functional morphology of the head and mouthparts of dragonflylarvae (Odonata). Quaestiones Entomologicae, 12:89–114.RASNITSYN, A. P. 1975. Hymenoptera Apocrita of the Mesozoic. Transactions of the Paleontological Institute,147:1–134 [in Russian].RASNITSYN, A. P. 1988. An outline of evolution of the hymenopterous insects (order Vespida). Oriental Insects, 22:115–145.RASNITSYN, A. P. 1992. Strashila incredibilis, a new enigmatic mecopteroid insect with possible siphonapteranaffinities from the Upper Jurassic of Siberia. Psyche, 99:323–333.RASNITSYN, A. P., AND V. V. ZHERIKHIN. 1999. First fossil chewing louse from the Lower Cretaceous of Baissa, Transbaikalia(Insecta, Pediculida = Phthiraptera, Saurodectidae fam. n.). Russian Entomological Journal, 8:253–255.RICHARDS, O. W., AND R. G. DAVIES. 1977. Imms’ General Textbook of Entomology, Tenth Edition. Chapman andHall, London, 1354 p.RICHTER, G. 1992. Fossilized gut contents: analysis and interpretation, p. 285–289. In S. Schaal and W. Ziegler(eds.), Messel—An Insight into the History of Life and of the Earth. Oxford University Press, Oxford, UK.RICHTER, G., AND S. BASZIO. 2001. First proof of planctivory/insectivory in a fossil fish: Thaumaturus intermediusfrom the Eocene Lake Messel (FRG). Palaeogeography, Palaeoclimatology, Palaeoecology, 173:75–85.RICHTER, R., AND G. STORCH. 1980. Beiträge zur Ernährungsbiologie eozäner Fledermause aus der “Grube Messel.”Natur und Museum, 110:353–367.RIEK, E. F. 1970. Lower Cretaceous fleas. Nature, 227:746–747.RITZKOWSKI, S. 1997. K-Ar-Alterbestimmungen der bernsteinführenden Sedimente des Samlandes (Paläogen,Bezirk Kaliningrad). Metalla Bochum, 66:19–23.ROHDENDORF, B. B., AND A. P. RASNITSYN. 1980. Historical Development of the Class Insecta. USSR Academy ofSciences, Moscow [in Russian].ROLFE, W. D. I. 1985. Aspects of the Carboniferous terrestrial arthropod community, p. 303–316. In J. T. Dutro,Jr. and H. W. Pfefferkorn (eds.), Comptes Rendus de Neuvième Congrès International de Stratigraphie et deGéologie du Carbonifère, 5. Southern Illinois University Press, Carbondale, IL.ROLFE, W. D. I., F. R. SCHRAM, G. PACAUD, D. SOTTY, AND S. SECRETAN. 1982. A remarkable Stephanian biota fromMontceau-les-Mines, France. Journal of Paleontology, 56:426–428.ROTH, B. 1986. Land mollusks (Gastropoda: Pulmonata) from early Tertiary Bozeman Group, Montana. Proceedingsof the California Academy of Sciences, 44:237–267.246

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESROWLEY, W. A., AND M. CORNFORD. 1972. Scanning electron microscopy of the pit of the maxillary palp of selectedspecies of Culicoides. Canadian Journal of Zoology, 50:1207–1210.RUIZ-TRILLO, I., M. RIUTORT, D. T. J. LITTLEWOOD, E. A. LHERNIOU, AND J. BAGUÑA. 1999. Acoel flatworms: earliestextant bilaterian metazoans, not members of Platyhelminthes. Science, 283:1919–1923.SAMWAYS, M. J., R. OSBORN, AND T. L. SAUNDERS. 1997. Mandible form relative to the main food type in ladybirds(Coleoptera: Coccinellidae). Biocontrol Science and Technology, 7:275–286.SCHLEE, D., AND W. GLÖCKNER. 1978. Bernstein. Stuttgarter Beiträge zur Naturkunde (C), 8:1–72.SCHMID-HEMPEL, P. 1998. Parasites in Social Insects. Princeton University Press, Princeton, NJ, 409 p.SCHMIDT, W., M. SCHURMANN, AND M. TEICHMÜLLER. 1958. Biß-Spuren an Früchten des Miozän-Waldes der niederrheinischenBraunkohlen-formation. Fortschritte in der Geologie von Rheinland und Westfalen, 2:563–572.SCHOONHOVEN, L. M., T. JERMY, AND J. J. A. VAN LOON. 1998. Insect-Plant Biology: From Physiology to Evolution.Chapman and Hall, London, 409 p.SCHRAM, F. R. 1973. Pseudocoelomates and a nemertine from the Illinois Pennsylvanian. Journal of Paleontology,47:985–989.SCHRAM, F. R. 1986. Crustacea. Oxford University Press, Oxford, UK, 606 p.SCHRAM, F. R., R. M. FELDMAN, AND M. J. COPELAND. 1978. The Late Devonian Palaeopalaemonidae and theearliest decapod crustaceans. Journal of Paleontology, 52:1375–1387.SCHUH, R. T., AND J. A. SLATER. 1995. True Bugs of the World (Hemiptera: Heteroptera). Cornell University Press,Ithaca, NY, 337 p.SCIACCHITANO, J. 1955. Su un gordio fossile. Monitore Zoologico Italiano, 63:57–61.SCOTT, A. C., AND T. N. TAYLOR. 1983. Plant/animal interactions during the Upper Carboniferous. Botanical Review,49:259–307.SCOURFIELD, D. J. 1926. On a new type of crustacean from the Old Red Sandstone (Rhynie Chert Bed, Aberdeenshire)—Lepidocaris rhyniensis gen. and sp. nov. Philosophical Transactions of the Royal Society of London, 214:153–187.SCUDDER, S. H. 1895. Revision of the American fossil cockroaches with descriptions of new forms. Bulletin ofthe United States Geological Survey, 124:1–176.SELDEN, P. A. 1996. Fossil mesothele spiders. Nature, 379:498–499.SHAROV, A. G. 1973. Morphological features and way of life of Palaeodictyoptera, p. 49–63. In G. Y. Bei-Benko(ed.), 24 th Annual Lectures in Memory of N. A. Kholodkovskogo, 24. Academy of Sciences, Moscow.SHEAR, W. A., AND P. M. BONAMO. 1988. Devonobiomorpha, a new order of centipeds (Chilopoda) from the Middle Devonianof Gilboa, New York state, USA, and the phylogeny of centiped orders. American Museum Novitates, 2927:1–30.SHEAR, W. A., AND J. KUKALOVÁ-PECK. 1990. The ecology of Paleozoic terrestrial arthropods: the fossil evidence.Canadian Journal of Zoology, 68:1807–1834.SHEAR, W. A,, AND P. A. SELDEN. 2001. Rustling in the undergrowth: animals in early terrestrial ecosystems, p. 29–51.In P. G. Gensel and D. Edwards (eds.), Plants Invade the Land: Evolutionary and Environmental Perspectives.Columbia University Press, New York.SHEAR, W. A., W. SCHAWALLER, AND P. M. BONAMO. 1989b. Record of Palaeozoic pseudoscorpions. Nature, 341:527–529.SHEAR, W. A., J. M. PALMER, J. A. CODDINGTON, AND P. M. BONAMO. 1989a. A Devonian spinneret: early evidenceof spiders and silk use. Science, 246:479–481.SIH, A. 1987. Nutritional ecology of aquatic insect predators, p. 579–607. In F. Slansky, Jr. and J. G. Rodriguez (eds.),Nutritional Ecology of Insects, Mites, Spiders, and Related Invertebrates. John Wiley & Sons, New York.SMIT, F. G. A. M. 1972. On some adaptive structures in Siphonaptera. Folia Parasitologica, 19:5–17.SMITH, J. J. B. 1985. Feeding mechanisms, p. 33–85. In G. A. Kerkut and L. E. Gilbert (eds.), ComprehensiveInsect Physiology, Biochemistry, and Pharmacology, 4. Pergamon Press, Oxford, UK.SNODGRASS, R. E. 1952. A Textbook of Arthropod Anatomy. Cornell University Press, Ithaca, NY, 363 p.SOLEM, A., AND E. L. YOCHELSON. 1979. North American Paleozoic land snails, with a summary of other Paleozoicnonmarine snails. United States Geological Survey Professional Paper, 1072:1–42.SOUTHCOTT, R. V., AND R. T. LANGE. 1971. Acarine and other microfossils from the Maslin Eocene, South Australia.Records of the South Australian Museum, 16:1–21.STORCH, G., AND G. RICHTER. 1992. The ant-eater Eurotamandua: a South American in Europe, p. 209–215. In S. Schaal andW. Ziegler (eds.), Messel—An Insight into the History of Life and of the Earth. Oxford University Press, Oxford, UK.247

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002STØRMER, L. 1963. Gigantoscorpio willsi, a new scorpion from the Lower Carboniferous of Scotland and itsassociated preying microorganisms. Skrifter Utgitt a v Det Norske Videnskaps-Akademi i Oslo, 8:1–171.STØRMER, L. 1969. Oldest known terrestrial arachnids. Science, 164:1276–1277.STRICKER, S. A. 1983. SEM and polarization microscopy of nemertean stylets. Journal of Morphology, 175:153–169.SUKACHEVA, I. D. 1982. Historical development of the order Phryganeida. Transactions of the PaleontologicalInstitute, 197:1–111 [in Russian].SZADZIEWSI, R. 1996. Biting midges from Lower Cretaceous amber of Lebanon and Upper Cretaceous amber ofTaimyr (Diptera, Ceratopogonidae). Studia Dipterologica, 3:23–86.TASCH, P. 1957. Flora and fauna of the Rhynie Chert: a paleoecological reevaluation of published evidence.University of Wichita Bulletin, 36:1–24.TASNÁDI-KUBACSKA, A. 1962. Paläopathologie, Vol. 2. Gustav Fischer, Jena, Germany, 269 p.TAYLOR, T. N. 1981. Pollen and pollen organ evolution in early seed plants, p. 1–25. In K. J. Niklas (ed.),Paleobotany, Paleoecology, and Evolution, 2. Praeger, New York.THOMPSON, I. D., AND D. S. JONES. 1980. A possible onychophoran from the Middle Pennsylvanian Mazon Creekof northern Illinois. Journal of Paleontology, 54:588–596.THULBORN, R. A. 1991. Morphology, preservation and palaeobiological significance of dinosaur coprolites.Palaeogeography, Palaeoclimatology, Palaeoecology, 83:341–366.TILLIER, S., M. MASSELOT, AND A. TILLIER. 1996. Phylogenetic relationships of the pulmonate gastropods fromrRNA sequences, and tempo and age of the stylommatophoran radiation, p. 267–284. In J. Taylor (ed.),Origin and Evolutionary Radiation of the Mollusca. Oxford University Press, Oxford, UK.TRACEY, S., J. A. TODD, AND D. H. ERWIN. 1993. Mollusca: Gastropoda, p. 131–167. In M. J. Benton (ed.), TheFossil Record 2. Chapman & Hall, London.TRAUB, R., AND H. STARCKE. 1980. Fleas. Balkema, Rotterdam, 420 p.VERMEIJ, G. J. 1987. Evolution and Escalation: An Ecological History of Life. Princeton University Press,Princeton, NJ, 527 p.VERMEIJ, G. J., AND D. R. LINDBERG. 2000. Delayed herbivory and the assembly of marine benthic ecosystems.Paleobiology, 26:419–430.VINSON, S. B., AND P. BARBOSA. 1987. Interrelationships of nutritional ecology of parasitoids, p. 673–695. In F.Slansky, Jr. and J. G. Rodriguez (eds.), Nutritional Ecology of Insects, Mites, Spiders, and RelatedInvertebrates. John Wiley & Sons, New York.VOIGT, E. 1938. Ein fossiler Saitenwurm (Gordius tenuifibrosus n. sp.) aus der eozänen Braunkohle des Geiseltales.Nova Acta Leopoldina, N.F., 5:351–360.VOIGT, E. 1952. Ein Haareinschluss mit Phthirapteren-Eiern im Bernstein. Mitteilungen der GeologischeStaatsinstitut in Hamburg, 21:59–74.VOIGT, E. 1957. Ein parasitischer Nematode in fossiler Coleopteren—Muskulatur aus der eozänen Braunkohledes Geiseltales bei Halle (Saale). Paläontologische Zeitschrift, 31:35–39.VON KOENIGSWALD, W., G. G. RICHTER, AND G. STORCH. 1981. Nachweis von Hornschuppen bein Eomanis waldiaus der ‘Grube Messel’ bein Darmstadt (Mammalia, Pholidota). Senckenbergiana Lethaea, 61:291–298.WAAGE, J. K. 1979. The evolution of insect/vertebrate associations. Biological Journal of the Linnean Society, 12:187–224.WALKER, E. M. 1932. Prognathism and hypognathism in insects. Canadian Entomologist, 44:223–229.WALOSSEK, D., AND K. J. MÜLLER. 1994. Pentastomid parasites from the Lower Paleozoic of Sweden. Transactionsof the Royal Society of Edinburgh, Earth Sciences, 85:1–37.WALOSSEK, D., J. E. REPETSKI, AND K. J. MÜLLER. 1994. Heymonsicambria taylori n. sp. (Articulata: Pentastomida)from Upper Cambrian/Lower Ordovician boundary beds of Newfoundland, Canada. Canadian Journal ofEarth Sciences, 31:1664–1671.WALTER, D., AND H. PROCTOR. 1999. Mites: Ecology, Evolution and Behaviour. University of New South WalesPress, Sydney, 322 p.WEITSCHAT, W., AND W. WICHARD. 1998. Atlas der Pflanzen und Tiere im Baltischen Bernstein. Friedrich Pfeil,Munich, 256 p.WHEELER, Q. D. 1990. Insect diversity and cladistic constraints. Annals of the Entomological Society of America,83:1031–1047.248

LABANDEIRA—PREDATORS, PARASITOIDS, AND PARASITESWHITFIELD, J. B. 1998. Phylogeny and evolution of host-parasitoid interactions in Hymenoptera. Annual Reviewof Entomology, 43:129–151.WIEGMANN, B. M., C. MITTER, AND B. FARRELL. 1993. Diversification of carnivorous parasitic insects: extraordinaryradiation or specialized dead end? American Naturalist, 142:738–754.WILSON, E. O. 1992. The Diversity of Life. Harvard University Press, Cambridge, MA, 424 p.WILLS, M. 1993. Miscellanea, p. 555–560. In M. Benton, (ed.), The Fossil Record 2. Chapman and Hall, London.WIRTH, W. W., AND A. A. HUBERT. 1989. The Culicoides of Southeast Asia (Diptera, Nematocera). Memoirs of theAmerican Entomological Institute, 44:1–508.WITHYCOMBE, C. L. 1922. Notes on the biology of some British Neuroptera (Planipennia). Transactions of theEntomological Society of London, 1922:501–594.YEATES, D. K., AND D. GREATHEAD. 1997. the evolutionary pattern of host use in the Bombyliidae (Diptera): adiverse family of parasitoid flies. Biological Journal of the Linnean Society, 60:149–185.ZHERIKHIN, V. V. 1989. Oligocene seed beetles and acorn weevils (Coleoptera: Bruchidae, Curculionidae) fromthe Bol’shoy Svetlovodnaya River (Northern Primorye). Cenozoic of the Far East, 1989:145–150.ZUMPT, F. 1965. Myiasis in Man and Animals in the Old World. Butterworth, London.249


FARLOW AND HOLTZ— PREDATION IN DINOSAURSTHE FOSSIL RECORD OF PREDATION IN DINOSAURSJAMES O. FARLOW 1 AND THOMAS R. HOLTZ, JR. 21Department of Geosciences, Indiana-Purdue University Fort Wayne,2101 Coliseum Boulevard, East, Fort Wayne, Indiana 46805 USA2Department of Geology, University of Maryland, College Park, Maryland 20742 USAABSTRACT—Predatory theropod dinosaurs can usually be identified as such by features of their jaws, teeth, andpostcrania, but different clades of these reptiles differed in their adaptations for prey handling. Inferences abouttheropod diets and hunting behavior based on functional morphology are sometimes supported by evidence fromtaphonomic associations with likely prey species, bite marks, gut contents, coprolites, and trackways. Very largetheropods like Tyrannosaurus are unlikely to have been pure hunters or scavengers, and probably ate whatevermeat they could easily obtain, dead or alive. Theropods were not the only dinosaur hunters, though; other kindsof large reptiles undoubtedly fed on dinosaurs as well. The taxonomic composition of dinosaurian predator-preycomplexes varies as a function of time and geography, but an ecologically remarkable feature of dinosaurianfaunas, as compared with terrestrial mammalian faunas, is the very large size commonly attained by bothherbivorous and carnivorous dinosaurs. The K/T extinction event(s) did not end dinosaurian predation, becausecarnivorous birds remained prominent predators throughout the Cenozoic EraINTRODUCTIONCARNIVOROUS DINOSAURS (Fig. 1)included some of the biggest, most spectacularpredators of all time, but also numbered in their ranksa diversity of smaller predators. In this paper wesurvey what is known about the diets of theropoddinosaurs, and briefly consider morphologicaldifferences among taxa that presumably affected theway they dealt with prey. We will also consider nondinosauriancarnivores that likely fed upondinosaurs. Finally, we will compare the taxonomiccomposition of herbivores and carnivores indifferent dinosaurian faunas, and examine someecological questions posed by the huge body sizesattained by many predatory dinosaurs.IDENTIFYING DINOSAURIANPREDATORS AND PREYMorphological features.—It is possible toidentify most extinct tetrapods as herbivores orcarnivores from skeletal morphology by judiciouscomparison with extant animals of known foodhabits. Plant-eaters usually have dentitions suitablefor shredding, crushing, slicing, or grinding theirfodder (and patterns of tooth wear consistent withsuch oral processing), capacious guts for housingmicrobes that assist in breaking down plant fibers,and toes that terminate in blunt nails or hooves ratherthan claws (cf. Reisz and Sues, 2000). Carnivores,in contrast, have sharp teeth for ripping, cutting, ortearing flesh, narrower gut regions, and sharp clawsfor restraining and dispatching prey.On the basis of such criteria, most ornithischiansand sauropodomorphs are presumed to have beenprimarily herbivorous (Farlow, 1997; Ryan andVickaryous, 1997; Sander, 1997; Upchurch andBarrett, 2000), although some taxa may have beenmore omnivorous (Barrett, 2000). Most adulttheropods were probably vertebrate-eaters, but thereare exceptions: therizinosaurs were probably planteaters(Russell, 2000), troodontids may have beenomnivores (Holtz et al., 1998), and ornithomimidswere likely filter-feeders that consumed aquatic plantsand/or small invertebrates (Norell et al., 2001).Even among those theropods that clearly weremeat-eaters, there are major morphological251

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 1—Dinosaurian predators and prey. A group of half-grown tyrannosaurids pursue an ornithomimidin the Late Cretaceous of western Canada. The inference that some tyrannosaurids may have lived ingroups is drawn from an Albertosaurus bonebed (Eberth et al., 2001). Drawing by James Whitcraft.differences among taxa that presumably reflectdifferences in attack and feeding behavior and/ordiet. For example, the relatively long and narrow,lightly constructed skull and the laterally compressedteeth of Allosaurus and many other carnivorousdinosaurs (Rayfield et al., 2001) contrast markedlywith the broader, massively constructed muzzle andvery stout teeth of Tyrannosaurus (Farlow et al.,1991; Molnar, 2000) and its kin; and bothmorphologies differ greatly from the very longsnoutedspinosaurs, whose conical teeth haveunusually fine serrations, or lack them altogether(Charig and Milner, 1997; Naish et al., 2001).Spinosaurids and the Triassic-Jurassiccoelophysoids both demonstrate a curvature of thepremaxillary-maxillary tooth row similar to thatseen in some modern crocodylians; perhaps, aswith these extant reptiles, this curvature representsa location in the snout for holding and manipulatingsmaller prey items or the extremities of larger prey.Oviraptorosaurs lack teeth, but have stronglyconstructed skulls that could nonetheless haveadministered a wicked bite to small prey (Ryan andVickaryous, 1997). Theropods are diagnosed bythe possession of a specialized intramandibularjoint between the dentary and postdentary bones(Bakker et al., 1988; Sereno and Novas, 1994;Sereno, 1999; Holtz, 2000). Although thisadaptation has yet to be subjected to rigorousbiomechanical analysis, it seems likely that itserved in part as a “shock absorber” to deal withthe forces generated by struggling live prey and/orthe dismemberment of carcasses.Predatory theropod clades differed in the extentto which the forelimbs and hindfoot were likelyinvolved in capturing and killing prey. Theropodsretained the ancestral dinosaurian condition ofobligate bipedalism, and thus (unlike most othercarnivorous reptiles and mammals) the forelimb wasfreed from the necessity of serving simultaneously252

FARLOW AND HOLTZ— PREDATION IN DINOSAURSas an organ of locomotion and of prey capture. Basalcarnivorous dinosaurs (Sereno, 1993) possessedlong fingered hands with elongated penultimatephalanges, an adaptation associated with enhancedgrasping ability (Hopson, 2001). Many lineages oftheropods retained this condition, and inoviraptorosaur and dromaeosaurid maniraptoransthe forelimbs were especially elongated (Middletonand Gatesy, 2000). In contrast, several groups oftheropod carnivores reduced the size and/or graspingfunction of the hand, such as neoceratosaurs(Gilmore, 1920; Bonaparte et al., 1990) andtyrannosaurids (Carpenter and Smith, 2001).Typical theropod feet have claws, which, whilecurved, do not have the trenchant shape of the manualtalons. In ornithomimosaurs (which are unlikely tohave preyed upon other dinosaurs), in fact, the pedalclaws are relatively straight and more hoof-like.Several taxa of theropods, however, are characterizedby a sickle-shaped ungual on a hyperextensiblesecond digit. These include the dromaeosaurids(Ostrom, 1969), troodontids (Barsbold et al., 1987),the primitive bird Rahonavis (Forster et al., 1998),and the neoceratosaur Noasaurus (Bonaparte andPowell, 1980). As documented in a spectacularlypreserved association (see below), this claw wasused in at least some cases to pierce (and presumablyrip out) the throat tissue of the victim.Taphonomic occurrences.—The circumstancesof preservation of dinosaur skeletons sometimessuggest predator-prey interactions. Shed theropodteeth are frequently found associated with single ormultiple skeletons of herbivorous dinosaurs (Chin,1997). Perhaps the most spectacular taphonomicassociation comprises interlocked specimens of asmall theropod (Velociraptor) with a smallceratopsian (Protoceratops). In this assemblage, thesickle claw of the dromaeosaurid is positioned veryclose to the ventral surface of the cervical vertebraeof the herbivore, and thus would have been withinthe neck of the plant-eater during the final momentsof both animals’ lives (Carpenter, 2000).Bite marks, gut contents, and coprolites.—Dietary inferences based on functional morphologycan sometimes be corroborated by trace fossils. Toothmarks in bone indicate that theropods did indeed feedupon herbivorous dinosaurs, and occasionally on eachother (Hunt et al., 1994; Erickson and Olson, 1996;Jacobsen, 1997, 1998, 2001; Chure et al., 2000). Aremarkable Hypacrosaurus leg bone even has atheropod tooth embedded within it (Fig. 2), as doesa limb bone of an azhdarchid pterosaur (Currie andJacobsen, 1995)!Some theropod skeletons contain the bonyremains of their prey. Specimens of compsognathidshave been found with bones of lizards andendothermic vermin (otherwise known as Mesozoicmammals) inside them, indicating that these smalltheropods ate correspondingly small prey (Ostrom,1978; Chen et al., 1998; Currie and Chen, 2001),unlike their portrayal in a recent multi-million-dollarmotion picture. In contrast, a tyrannosaurid skeletoncontained partially digested bones of juvenilehadrosaurids (Varricchio, 2001). The stomach regionof a beautiful specimen of Baryonyx containednumerous fish scales and teeth (a diet consistent withthe dinosaur’s cranial anatomy), as well as bones ofa young Iguanodon (Charig and Milner, 1997). Ona grislier note, two individuals of Coelophysiscontained the bones of what may be smallerindividuals of their own species (Colbert, 1989).Coprolites presumably made by herbivorousdinosaurs contain fragmented plant materials (Chinand Gill, 1996; Chin and Kirkland, 1998). Incontrast, Chin et al. (1998) described a 44-cm longcoprolite from the Maastrichtian FrenchmanFormation that contained angular pieces of bone.The osteohistological texture of the bony inclusionssuggests that the bone fragments came from asubadult ornithischian. Given the tremendous sizeof the coprolite, its most likely maker wasTyrannosaurus (or a very sick smaller theropod).NON-DINOSAURIANPREDATORS ON DINOSAURSDinosaurs originated in the Late Triassic(Heckert and Lucas, 1998; Hunt et al., 1998), andstarted out as modest-sized animals compared withmany of their non-dinosaurian neighbors; it is verylikely that Triassic dinosaurs frequently fell preyto (or were scavenged by) large phytosaurs and253

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 2—Portion of the fibula of an herbivorous dinosaur (Museum of the Rockies 549, Hypacrosaurus),with a theropod tooth embedded in it (arrow). The exposed portion of the tooth is about 5 mm long.other predatory non-dinosaurian archosaurs(Hungerbühler, 2000). Throughout the laterMesozoic, small-bodied dinosaurs (adults orjuveniles) were likely eaten by crocodylimorphs(including terrestrial cursorial forms; Kirkland,1994) and other large reptiles.The Cretaceous saw the evolution ofcrocodyliforms that were probably large and massiveenough to take even big adult dinosaurs.Sarcosuchus from the Early Cretaceous of Africa isestimated to have reached a total length of 11–12meters and a body mass of 8000 kg, as large as anyknown carnivorous dinosaur (Sereno et al., 2001).Deinosuchus, an alligatorid crocodylian from theLate Cretaceous of the southern and western U.S.,may have been equally big, and bite marks likelymade by this reptile occur in both hadrosaurid andtyrannosaurid bones (Schwimmer, 2002). In the LateCretaceous of the southeastern U.S., Deinosuchusmay have displaced large theropods as the dominantbig predator (Schwimmer, 2002), at least near largerbodies of water.DINOSAUR PREDATOR-PREYINTERACTIONSTheropod food preferences and the intensityof predation.—Because predation by and ondinosaurs often resulted in the destruction of preyitems, it is difficult to quantify the food preferencesof theropods, or to assess the intensity of theirpredation on herbivorous dinosaurs, in the way thatcan sometimes be done for marine invertebrates(e.g., by determining the relative frequency ofdrilled bivalve or brachiopod shells). However,some inferences can be made by examining bitemarks and coprolites.In Late Cretaceous skeletal assemblages fromAlberta and Montana, the incidence of tooth-markedbone ranges from a few percent to about 14%; the254

FARLOW AND HOLTZ— PREDATION IN DINOSAURSnumber is higher for isolated bones than for boneswithin bonebeds (Fiorillo, 1991; Jacobsen 1997,1998, 2001; Jacobsen and Ryan, 1998). There is noindication that tyrannosaurids deliberately crushedbones, in the manner of some mammaliancarnivores, even though their teeth and jaws werestrong enough to handle bone-breaking (Farlow etal., 1991; Erickson et al., 1996; Molnar, 1998;Hurum and Currie, 2000); bone-biting seems to havebeen incidental to feeding on meat. Hadrosaur bonesmore frequently show bite marks than do the bonesof other herbivorous dinosaurs and theropods.The presence of juvenile ornithischian bonesin tyrannosaurid gut regions and coprolites (Chinet al., 1998; Varricchio, 2001) invites speculationthat these large predators preferred to take young asopposed to fully grown individuals of plant-eatingdinosaurs. This would have involved less risk ofinjury to the predator than tussles with large andperhaps dangerous prey. Given that even the largestdinosaurs would have had the capacity to producelarge clutches of eggs every year (as opposed toplacental mammals, in which larger body size isassociated with longer gestation periods; Carranoand Janis, 1991), a stable population size ofdinosaurs would have required a high rate of infantmortality. It seems quite likely that a substantialfraction of these juvenile dinosaurs disappeareddown the gullets of theropods.Trackway evidence.—Fossilized trackwaysoffer clues to predatory behavior by theropoddinosaurs. Thulborn and Wade (1984) described amid-Cretaceous tracksite in Queensland, Australia,in which a host of small bipedal dinosaurs panickedand fled during the approach of a much largerbipedal dinosaur, most likely a large theropod.Whether the bigger dinosaur was actually huntingthe smaller animals is uncertain, but at one point itmade a sharp change in its direction of travelconsistent with the hypothesis that it was trying todrive them in a particular direction.In 1940 Roland T. Bird collected segments ofthe trackways of a sauropod and a large theropoddinosaur in the Lower Cretaceous Glen RoseLimestone at what is now Dinosaur Valley State Parknear Glen Rose, Texas (Bird, 1985). The theropod(very likely Acrocanthosaurus; Farlow, 2001)repeatedly stepped into and deformed the printsmade by the sauropod, and the trails of both animalsmade a turn at the same point, suggesting that themeat-eater was close behind and following the bigherbivore (Farlow, 1987; Thomas and Farlow, 1997).Dinosaur tracksites suggest that at least somedinosaurs were gregarious some of the time(Ostrom, 1972, 1986; Currie, 1983; Lockley et al.,1986; Thulborn, 1990; Lockley, 1991; Lockley andHunt, 1995; Lockley and Meyer, 2000),corroborating interpretations about dinosaursociality based on skeletal assemblages (Coombs,1990; Horner, 1997; Farlow, 2000; Eberth et al.,2001). Conceivably, herding behavior on the partof herbivorous dinosaurs was an anti-predatorstrategy (Day et al., 2002), while group hunting bytheropods may have permitted them to kill prey toolarge for a single hunter to take (Farlow, 1976;Maxwell and Ostrom, 1995).The Paluxy River sauropod trackway collectedby R. T. Bird was one of at least a dozen sauropodtrails that seem to have been made by a group ofthe huge plant-eaters. Bird further believed that agroup of theropods was following this herd—rather than just one carnivore tracking a singleherbivore. Regrettably, the trackway evidence atDinosaur Valley State Park does not clearlysupport Bird’s interpretation, but neither does itfalsify it (Farlow, 1987).Predation vs. scavenging.—Perhaps the bestknown predatory dinosaur, Tyrannosaurus, hasbeen suggested to have been an obligate scavenger(Horner, 1994; Horner and Lessem, 1993; Hornerand Dobb, 1997). Horner (1994) argues that severalmorphological features of Tyrannosaurus wouldhave precluded a predatory lifestyle: 1) relativelysmall size of the eye that would have prohibitedspotting prey at a distance; 2) limb proportionsindicative of slow top running speeds, which wouldhave prevented Tyrannosaurus from chasing andcapturing prey; 3) disproportionately tiny forelimbsthat would have been useless for holding prey; 4)relatively broad teeth that depart from the expectedblade-like configuration for teeth of a predator.We do not find these arguments persuasive. The255

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002size of the orbit of Tyrannosaurus relative to its skullsize is in fact rather large for a reptile of its size(Fig. 3). Furthermore, the dimensions of the orbitsuggest that Tyrannosaurus had a big eye in absoluteterms, which would have increased its lightgatheringcapacity and thus its acuity (Walls, 1942;Dusenberry, 1992). Even though Tyrannosauruslacks the cursorial hind limb proportions of smallertheropods, and was probably not as good a runneras sometimes portrayed (Farlow et al., 1995b, 2000;Christiansen, 1999; Hutchinson and Garcia, 2002),its metatarsus/femur or tibia/femur length ratiosindicate that it was likely as fleet, or faster, thanother big theropods, and certainly faster than theherbivorous dinosaurs that were its likely prey(Gatesy, 1991; Holtz, 1995).Horner’s last two arguments strike us as beggingthe question. Without explicitly saying so, he ishypothesizing that grasping forelimbs are a necessityfor killing prey (which will be news to wolves,seriemas, and secretary birds), and that animals withbroad-based teeth are unable to kill prey with themFIGURE 3—Relationship between skull length (occipital condyle to tip of snout) and anteroposteriordiameter of the orbit in tyrannosaurids (Gorgosaurus, Daspletosaurus, Tyrannosaurus), theropods otherthan tyrannosaurids (Eoraptor, Herrerasaurus, Coelophysis, Dilophosaurus, Syntarsus, Abelisaurus,Carnotaurus, Ceratosaurus, Acrocanthosaurus, Allosaurus, Giganotosaurus, Monolophosaurus, Sinraptor,Yangchuanosaurus, Dromaeosaurus, Velociraptor, Erlikosaurus, Ingenia, Ornitholestes, Saurornithoides,Dromiceiomimus, Gallimimus, Garudimimus, Struthiomimus), extant crocodylians (Alligator, Caiman,Melanosuchus, Paleosuchus, Crocodylus, Osteolaemus, Tomistoma, Gavialis), the extinct crocodylianDeinosuchus, the extinct crocodylomorph Sarcosuchus, and several extant species of the varanid lizardgenus Varanus (acanthurus, bengalensis, dumerili, exanthematicus, gouldii, griseus, indicus, komodoensis,niloticus, olivaceus, prasinus, rudicollis, salvator, timorensis). Note that tyrannosaurids (includingTyrannosaurus itself, represented by the three biggest tyrannosaurid specimens) have orbits (and thereforepresumably eyes) as large or larger relative to skull size than those of other carnivorous reptiles.256

FARLOW AND HOLTZ— PREDATION IN DINOSAURS(which orcas and crocodiles will find surprising).Because the morphology of Tyrannosaurus matchesthe predictions of his hypotheses, Horner concludesthat Tyrannosaurus could not have been a predator,without first testing those hypotheses.The brain of Tyrannosaurus had respectablylarge olfactory bulbs (Brochu, 2000), suggestingthat the sense of smell was quite acute in thisdinosaur. Horner and Dobb (1997) argued that thiswould have allowed Tyrannosaurus to detect theodor of rotting carcasses from afar. This isunquestionably true, but it is also true that a keensense of smell would have been useful for pickingup the scent of live prey, or for behaviors unrelatedto food acquisition (Brochu, 2000).We agree with Horner that Tyrannosaurus isunlikely to have engaged in extended, Hollywoodstylebattles with other large dinosaurs (or hugeapes, for that matter). However, surprise, hit-andrunattacks on healthy victims (Paul, 1988), orculling of sick, injured (Carpenter, 2000), or veryyoung dinosaurs, would seem quite likely. In short,we suspect that Tyrannosaurus and othercarnivorous theropods were, like most extantpredators, opportunistic carnivores, eagerlysearching for carrion (in which activity the largebody sizes of many theropods may have been anadvantage; Farlow, 1994), but also killing preywhenever possible.DINOSAUR FAUNASComposition.—Dinosaurs began as minorcomponents of Late Triassic large-tetrapod faunas(cf. Parrish, 1993; Rogers et al., 1993), but by thebeginning of the Jurassic Period had become thedominant terrestrial large vertebrates. Over theremainder of the Mesozoic Era the taxonomiccomposition of herbivorous and carnivorousspecies in dinosaur faunas varied across time andspace, but two particularly noteworthy faunal suitescan be recognized. One of these has the herbivorousdinosaur component strongly influenced or evendominated by sauropods (e.g., the Morrison andWessex Formations) (Table 1); this faunal type ischaracteristic of much of the world during theJurassic and Cretaceous Periods. Sauropods areabsent or rare and ornithischians dominant in thesecond faunal type (e.g., the Dinosaur ParkFormation), which occurred in the Late Cretaceousof western North America and eastern and centralAsia (Table 1).Neoceratosaurs, basal tetanurans, andcarnosaurs (Holtz, 2000) are the dominant theropodgroups in the first faunal type, and coelurosaurs inthe second. Medium-sized and large theropods inthe first faunal suite come from a variety oflineages, but all large-bodied taxa in the secondtype are tyrannosaurids. In both faunal typespredatory dinosaurs individually are far lessabundant than plant-eaters (Farlow, 1997)Although we do not know which carnivorousdinosaur species specialized on which herbivorousspecies, the marked differences between the twokinds of faunas suggest the possibility of majordifferences in predator-prey interactions betweenthem. For example, adult sauropods wereconsiderably bigger than the largest theropods, whilemost big ornithischians were much closer totyrannosaurids in body size. Even if tyrannosaurspreferred to attack immature individuals of preyspecies, it is easy to imagine a single tyrannosaurkilling an adult hadrosaur or ceratopsian. It is muchharder to visualize a single allosaur slaying an adultapatosaur or brachiosaur. Did large theropods insauropod-dominated faunas attack only immaturesauropods and ignore fully grown adults, or did theyengage in group hunting to haul down big sauropods,or did they mainly scavenge sauropod carcasses?Theropod species in the multi-taxon predatorassemblages typical of sauropod-dominated faunasshow interesting morphological differences fromtyrannosaurids that suggest differences in the styleof predation between carnivores in thesecommunities. In the multi-taxon assemblagesseveral large-bodied theropod taxa (basal tetanurans,spinosaurids, and carnosaurs) possessed verypowerfully built forelimbs terminating in large talons.It is quite likely that these predators employed theirforelimbs as weapons of prey acquisition. In contrast,tyrannosaurids are characterized by greatly reducedforelimbs, and so their style of prey acquisition would257

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002TABLE 1—Comparison of the composition of several dinosaur faunas (X = Xiashaximiao Fm, China,Middle Jurassic; M = Morrison Fm, American West, Late Jurassic [LJ]; T = Tendaguru Grp, Tanzania,LJ; W = Wessex Fm, Isle of Wight, Early Cretaceous [EK]; Y = Yixian Fm, China, EK; C = Cloverly Fm,Wyoming and Montana, EK; B = Bahariya Fm, Egypt, Late Cretaceous [LK]; D = Dinosaur Park Fm,Alberta, LK; N = Nemegt Fm, Mongolia, LK; H = Hell Creek Fm, Montana amd Wyoming, LK).Fauna: X M T W Y C B D N HHerbivores: "+" indicates that the taxon is present and abundant;"r" indicates that the taxon is present but rare.Sauropods + + + + r + rOrnithopods r r + + r + + + +Marginocephalians r + + + +Thyreophorans r + + + r + + + +Theropods: "*" indicates that one or all of the species in the taxon may not have beenstrictly carnivorous; "L" indicates that the taxon is present and includes thelargest theropods in the assemblage; "p" indicates that the taxon is present;"?" indicates that the identification of this taxon in the assemblage istentative at present.Coelophysoids p pNeoceratosaurs p LBasal Tetanurans p p p?Spinosaurids L LCarnosaurs L L L? p L LBasal Coelurosaurs p p p p? pTyrannosaurids p p L L LOrnithomimosaurs* p p? p p pOviraptorosaurs* p? p? p p p p pTherizinosauroids* L p? pTroodontids* p? p p p pDromaeosaurids p p p p? pAvialians* p? p p p p258

FARLOW AND HOLTZ— PREDATION IN DINOSAURShave relied on their powerful jaws alone.Another difference between the faunal types isthe overlap of theropod body sizes. In assemblagespossessing multiple lineages of large-bodiedtheropods there is commonly great overlap in thesize of the carnivores. For example, in the MorrisonFormation the carnosaur Allosaurus, the basaltetanuran Torvosaurus, and the neoceratosaurCeratosaurus would all include individuals of1 tonne or greater body mass. Similarly, theBahariya Formation’s spinosaurid Spinosaurus,carnosaur Carcharodontosaurus, and basalcoelurosaur Deltadromeus all exceeded 2 tonnesin mass (the first two by a considerable margin).The presence of comparable-sized predatorssuggests the possibility of competition among thesetaxa for food, perhaps mitigated by some form ofmorphologically mediated niche partitioning(Henderson, 2000). Similar size overlap occursbetween the adults of medium-sized theropods inthese assemblages, which would additionally havebeen in potential competition with immatureindividuals of the largest-bodied species.In marked contrast, all the larger carnivorousdinosaurs in Late Cretaceous assemblages of westernNorth America and eastern and central Asia aretyrannosaurids, and among these there is typicallyjust one or two species present in potential sympatry.Furthermore, there is often a large discontinuity inadult sizes between the tyrannosaurids and the nextlargest unquestionably carnivorous dinosaurs in thefauna (generally dromaeosaurids), rather than thegradational distribution of adult sizes seen in nontyrannosaurid-dominatedfaunas.Macroecology of carnivorous dinosaurs.—Thesingle most noteworthy feature of most dinosaurs,of course, is their large size. Body size affects or iscorrelated with numerous physiological andecological features of animals (Brown, 1995; Brownand West, 2000). Large animals have bigger homeranges than do smaller species, and carnivoresrequire more habitat space than herbivores (Kelt andVan Vuren, 2001). Farlow (2001) used publishedregressions of home range area against body massin extant predatory mammals, birds, and lizards tospeculate that the home range size of the 2500-kgcarnosaur Acrocanthosaurus would haveencompassed hundreds or thousands of squarekilometers. Kelt and Van Vuren (2001), however,suggested that there may be some upper limit tohome range area in mammals, regardless of bodysize and diet. If true, and if this upper limit holds forother terrestrial vertebrates, it raises the question ofhow gigantic predators like large theropods couldhave survived on relatively small (as compared withindividual animal size) home ranges.Because an individual animal’s home rangearea becomes larger with increasing body size,population density (number of individuals / habitatarea) must decrease (Damuth, 1987; Brown, 1995;Smallwood, 2001), which in turn mandates largegeographic ranges if big-bodied species are to berepresented by enough individuals for long-termviability (Calder, 2000). For trophodynamicreasons carnivores must have lower populationdensities than herbivores, and so the problem ofsufficient habitat space should be particularly acutefor enormous carnivores (Farlow, 1993; Burnesset al., 2001). The huge sizes routinely achieved bycarnivorous dinosaurs are therefore ecologicallypuzzling. Conceivably, theropod gigantism wasfacilitated by a combination of lower foodrequirements than expected for elephantinemammalian meat-eaters, along with elevated ratesof biological productivity under the greenhouseconditions of the Mesozoic Era (Farlow, 1993;Farlow et al., 1995; Burness et al., 2001). In anotherscenario (Carrano and Janis, 1991) the greaterreproductive capacity of herbivorous dinosaursrelative to placental mammals (due to oviparity ofthe former) would allow for more available“packages” of dinosaurian meat that could beconsumed by theropods while still allowing for aviable sustainable population of prey.POSSIBLE DIRECTIONS FORFUTURE RESEARCHAlthough progress in understanding predatorpreyinteractions in dinosaur communities willdepend in large part on fortuitous discoveries ofparticularly informative specimens or assemblages,259

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002we can suggest some approaches that might provefruitful. One matter worth exploring is the incidenceof bite marks in dinosaur bones, or teeth embeddedin bone. Such fossils have already been noted forthe Late Cretaceous of western North America(Jacobsen, 1997, 1998, 2001), but older formationscould also be surveyed. Similarly, paleontologistsshould keep their eyes open for potential theropodcoprolites. With sufficiently large sample sizes oftooth-marked bone and coprolites, it might bepossible to determine which species of herbivorousdinosaurs, and which size classes within thosespecies, were preferentially eaten by which predatorspecies. If we were really lucky, we might even find,say, a bite mark unambiguously made by aTyrannosaurus that had healed, which wouldestablish beyond doubt that these predatorydinosaurs at least sometimes attacked live prey.Unfortunately, distinguishing successful predationevents from scavenging on the basis of toothmarkedbones is probably impossible, because thevictim cannot recover from either.Structural analysis (including computermodeling) of a variety of theropod skulls inparticular faunas (cf. Henderson, 2000; Rayfieldet al., 2001) could be used to test whetherreconstructions of different biting and/or feedingstyles in sympatric theropod species aremechanically feasible, and thus ways in whichcoexisting species could have subdivided theresource base. Such approaches could be combinedwith analyses of tooth shape, tooth cutting edges,and wear and breakage patterns, in both in situ andshed theropod teeth (Farlow et al., 1991; Farlow andBrinkman, 1994; Abler, 1997). The degree of sizeand shape overlap of the skulls and teeth of differentspecies of potentially sympatric theropods couldbe compared with that in modern communities ofpredatory lizards (e.g., the varanids of Australia)and crocodylians.With a better understanding of the systematiccomposition of dinosaur faunas, we could see howthe different composition of the prey base insauropod-dominated and ornithischian-dominatedFIGURE 4—Right lateral view of the pelvis of a moa (Canterbury Museum Av 8317, Emeus crassus)showing an elliptical gouge (arrow) dug by the hind toe talon of Harpagornis, a huge, extinct eagle. Thepaper label is 102 mm long. Photograph courtesy of Richard Holdaway.260

FARLOW AND HOLTZ— PREDATION IN DINOSAURSfaunas affected the structure of predatory dinosaurguilds. Do the two kinds of faunas consistently differin the ratio of the number of herbivore species tocarnivore species, or might they differ in the ratioof individual herbivorous animals to carnivorousanimals? And can any such differences be relatedto the mechanisms by which sympatric carnivorousdinosaur species reduced potential niche overlap?CENOZOIC REPRISEBecause birds are likely a specialized clade oftheropods (Gauthier and Gall, 2001), the fossilrecord of dinosaur predation does not end with theK/T boundary. A diversity of Cenozoic birds hasevolved as important predators of invertebrates andsmall vertebrates. At certain times and places,however, avian faunas have been particularlyevocative of the Mesozoic glory days. For example,the phorusrhacoids of Tertiary South America(Andrews, 1901), and possibly the gastornithidsof Paleogene Europe and North America (Witmerand Rose, 1991), were big, flightless, predatorybirds startlingly reminiscent of their Mesozoictheropod predecessors.The real Cenozoic lost world of dinosaurs,however, was New Zealand. In the absence ofsignificant mammalian competition, a host of largeand small birds, both volant and flightless,dominated the terrestrial vertebrate fauna (Worthyand Holdaway, 2002). Eleven species of moa, turkeyto ostrich-plus in size, clumped through forest andfield, cropping the vegetation like scaled-downsauropodomorphs or ornithischians. No ground-basedtyrannosaur-avatar threatened the moa. Instead theirchief predator was a huge eagle that attacked withtalons from the air (Fig. 4), an entirely different styleof dinosaurian predator-prey interaction than seenin the Mesozoic world. Dinosaurian dominance ofNew Zealand remained unchallenged until about athousand years ago and the arrival of a bipedal,predatory primate far deadlier than any theropod, atwhich time this faraway land, too, finally fell undermammalian sway.REFERENCESABLER, W. L. 1997. Tooth serrations in carnivorous dinosaurs, p. 740–743. In P. J. Currie and K. Padian (eds.),Encyclopedia of Dinosaurs. Academic Press, San Diego.ANDREWS, C. W. 1901. On the extinct birds of Patagonia–I: The skull and skeleton of Phororhacos inflatusAmeghino. Transactions of the Zoological Society of London, 15: 55–86.BAKKER, R. T., M. WILLIAMS, AND P. J. CURRIE. 1988. Nanotyrannus, a new genus of pygmy tyrannosaur, from thelatest Cretaceous of Montana. Hunteria, 1(5):1–30.BARRETT, P. M. 2000. Prosauropod dinosaurs and iguanas: Speculations on the diets of extinct reptiles, p. 42–78.In H.-D. Sues (ed.), Evolution of Herbivory in Terrestrial Vertebrates: Perspectives from the Fossil Record.Cambridge University Press, Cambridge, U.K.BARSBOLD, R., H. OSMÓLSKA, AND S. M. KURZANOV. 1987. On a new troodontid (Dinosauria, Theropoda) from theEarly Cretaceous of Mongolia. Acta Palaeontologica Polonica, 32:121–132.BIRD, R. T. 1985. Bones for Barnum Brown: Adventures of a Dinosaur Hunter. Texas Christian University Press,Fort Worth, TX, 225 p.BONAPARTE, J. F., AND J. E. POWELL. 1980. A continental assemblage of tetrapods from the Upper Cretaceous bedsof El Brete, northwestern Argentina (Sauropoda-Coelurosauria-Carnosauria-Aves). Memoires de la SociétéGeologique de France, N.S., 139:19–28.BONAPARTE, J. F., F. E. NOVAS, AND R. A. CORIA. 1990. Carnotaurus sastrei Bonaparte, the horned, lightly builtcarnosaur from the Middle Cretaceous of Patagonia. Natural History Museum of Los Angeles County,Contributions in Science, 416:1–41.BROCHU, C. A. 2000. A digitally-rendered endocast for Tyrannosaurus rex. Journal of Vertebrate Paleontology, 20:1–6.BROWN, J. H. 1995. Macroecology. University of Chicago Press, Chicago, 269 p.BROWN, J. H., AND G. B. WEST (eds.). 2000. Scaling in Biology. Oxford University Press, Oxford, U.K., 352 p.261

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BURNESS, G. P., J. DIAMOND, AND T. FLANNERY. 2001. Dinosaurs, dragons, and dwarfs: The evolution of maximalbody size. Proceedings of the National Academy of Science USA, 98:14518–14523.CALDER, W. A. 2000. Diversity and convergence: Scaling for conservation, p. 297–323. In J. H. Brown and G. B.West (eds.), Scaling in Biology. Oxford University Press, Oxford, U.K.CARPENTER, K. 2000. Evidence of predatory behavior by carnivorous dinosaurs. Gaia, 15:135–144.CARPENTER, K., AND M. SMITH. 2001. Forelimb osteology and biomechanics of Tyrannosaurus rex, p. 90–116. InD. H. Tanke and K. Carpenter (eds.), Mesozoic Vertebrate Life. Indiana University Press, Bloomington.CARRANO, M., AND C. JANIS. 1991. Scaling of reproductive turnover in archosaurs and mammals: Why are largeterrestrial mammals so rare? Annales Zoologici Fennici, 28: 201–216.CHARIG, A. J., AND A. C. MILNER. 1997. Baryonyx walkeri, a fish-eating dinosaur from the Wealden of Surrey.Bulletin of the Natural History Museum, 53:11–70.CHEN, P., Z. DONG, AND S. ZHEN. 1998. An exceptionally well-preserved theropod dinosaur from the YixianFormation of China. Nature, 391:147–152.CHIN, K. 1997. What did dinosaurs eat? Coprolites and other direct evidence of dinosaur diets, pp. 371–382. In J.O. Farlow and M. K. Brett-Surman (eds.), The Complete Dinosaur. Indiana University Press, Bloomington.CHIN, K., AND B. D. GILL. 1996. Dinosaurs, dung beetles, and conifers: participants in a Cretaceous food web.Palaios, 11:280–285.CHIN, K., AND J. I. KIRKLAND. 1998. Probable herbivore coprolites from the Upper Jurassic Mygatt-Moore Quarry,western Colorado. Modern Geology, 23:249–275.CHIN, K., T. T. TOKARYK, G. M. ERICKSON, AND L. C. CALK. 1998. A king-sized theropod coprolite. Nature, 393:680–682.CHRISTIANSEN, P. 1999. Long bone scaling and limb posture in non-avian theropods: evidence for differentialallometry. Journal of Vertebrate Paleontology, 19:666–680.CHURE, D. J., A. R. FIORILLO, AND A. JACOBSEN. 2000. Prey bone utilization by predatory dinosaurs in the Late Jurassicof North America, with comments on prey bone use by dinosaurs throughout the Mesozoic. Gaia, 15:227–232.COLBERT, E. H. 1978. The Triassic dinosaur Coelophysis. Museum of Northern Arizona Bulletin 57, 160 p.COOMBS, W. P., JR. 1990. Behavior patterns of dinosaurs, p. 32–42. In D. B. Weishampel, P. Dodson, and H.Osmólska (eds.), The Dinosauria. University of California Press, Berkeley, CA.CURRIE, P. J. 1983. Hadrosaur trackways from the Lower Cretaceous of Canada. Acta Palaeontologica Polonica, 28:63–73.CURRIE, P. J., AND P. CHEN. 2001. Anatomy of Sinosauropteryx prima from Liaoning, northeastern China. CanadianJournal of Earth Sciences, 38:1705–1727.CURRIE, P. J., AND A. R. JACOBSEN. 1995. An azhdarchid pterosaur eaten by a velociraptorine theropod. CanadianJournal of Earth Sciences, 32:922–925.DAMUTH, J. 1987. Interspecific allometry of population density in mammals and other animals: The independenceof body mass and population energy-use. Biological Journal of the Linnean Society, 31:193–246.DAY, J. J., P. UPCHURCH, D. B. NORMAN, A. S. GALE, AND H. PHILIP POWELL. 2002. Sauropod trackways, evolution,and behavior. Science, 296:1659.DUSENBERRY, D. B. 1992. Sensory Ecology: How Organisms Acquire and Respond to Information. Freeman,New York, 558 p.EBERTH, D. A., P. J. CURRIE, D. B. BRINKMAN, M. J. RYAN, D. R. BRAMAN, J. D. GARDNER, V. D. LAM, D. N. SPIVAK,AND A. G. NEUMAN. 2001. Alberta’s dinosaurs and other fossil vertebrates: Judith River and Edmonton Groups(Campanian-Maastrichtian), p. 49–75. In C. L. Hill (ed.), Guidebook for the Field Trips, Society of VertebratePaleontology 61 st Annual Meeting, Museum of the Rockies Occasional Paper 3, Bozeman, MT.ERICKSON, G. M., AND K. H. OLSON. 1996. Bite marks attributable to Tyrannosaurus rex: preliminary descriptionand implications. Journal of Vertebrate Paleontology, 16:175–178.ERICKSON, G. M., S. D. VAN KIRK, J. SU, M. E. LEVENSTON, W. E. CALER, AND D. R. CARTER. 1996. Bite forceestimation for Tyrannosaurus rex from tooth-marked bones. Nature, 382:706–708.FARLOW, J. O. 1976. Speculations about the diet and foraging behavior of large carnivorous dinosaurs. AmericanMidland Naturalist, 95:186–191.FARLOW, J. O. 1987. Lower Cretaceous Dinosaur Tracks, Paluxy River Valley, Texas. South-Central Section,Geological Society of America, Baylor University, Waco, TX, 50 p.262

FARLOW AND HOLTZ— PREDATION IN DINOSAURSFARLOW, J. O. 1993. On the rareness of big, fierce animals: Speculations about the body sizes, population densities,and geographic ranges of predatory mammals and large carnivorous dinosaurs. American Journal of Science,293-A:167–199.FARLOW, J. O. 1994. Speculations about the carrion-locating ability of tyrannosaurs. Historical Biology, 7:159–165.FARLOW, J. O. 1997. Trophic groups, p. 755–758. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs.Academic Press, San Diego.FARLOW, J. O. 2000. Tracking dinosaur society, p. 232–241. In G. S. Paul (ed.), The Scientific American Book ofDinosaurs. St. Martin’s Press, New York.FARLOW, J. O. 2001. Acrocanthosaurus and the maker of Comanchean large-theropod footprints, p. 408–427. InD. H. Tanke and K. Carpenter (eds.), Mesozoic Vertebrate Life. Indiana University Press, Bloomington.FARLOW, J. O., AND D. L. BRINKMAN. 1994. Wear surfaces on the teeth of tyrannosaurs, p. 165–175. In G. D.Rosenberg and D. L. Wolberg (eds.), Dino Fest. Paleontological Society Special Publication 7, University ofTennessee, Knoxville, TN.FARLOW, J. O., P. DODSON, AND A. CHINSAMY. 1995a. Dinosaur biology. Annual Review of Ecology and Systematics,26:445–471.FARLOW, J. O, M. B. SMITH, AND J. M. ROBINSON. 1995b. Body mass, bone “strength indicator,” and cursorialpotential of Tyrannosaurus rex. Journal of Vertebrate Paleontology, 15:713–725.FARLOW, J. O, D. L. BRINKMAN, W. L. ABLER, AND P. J. CURRIE. 1991. Size, shape, and serration density of theropoddinosaur lateral teeth. Modern Geology, 16:161–198.FARLOW, J. O., S. M. GATESY, T. R. HOLTZ, JR., J. R. HUTCHINSON, AND J. M. ROBINSON. 2000. Theropod locomotion.American Zoologist, 40:640–663.FIORILLO, A. R. 1991. Prey bone utilization by predatory dinosaurs. Palaeogeography, Palaeoclimatology,Palaeoecology, 88:157–166.FORSTER, C. A., S. D. SAMPSON, L. M. CHIAPPE, AND D. W. KRAUSE. 1998. The theropod ancestry of birds: newevidence from the Late Cretaceous of Madagascar. Science, 279:1915–1919.GATESY, S. M. 1991. Hind limb scaling in birds and other theropods: implications for terrestrial locomotion.Journal of Morphology, 209:83–96.GAUTHIER, J., AND L. F. GALL (eds.). 2001. New Perspectives on the Origin and Early Evolution of Birds. PeabodyMuseum of Natural History, Yale University, New Haven, CT.GILMORE, C. W. 1920. Osteology of the carnivorous Dinosauria in the United States National Museum, withspecial reference to the genera Antrodemus (Allosaurus) and Ceratosaurus. Bulletin of the United StatesNational Museum, 110:1–154.HECKERT, A. B., AND S. G. LUCAS. 1998. Global correlation of the Triassic theropod record. Gaia, 15:63–74.HENDERSON, D. M. 2000. Skull and tooth morphology as indicators of niche partitioning in sympatric MorrisonFormation theropods. Gaia, 15:219–226.HOLTZ, T. R., JR. 1995. The arctometatarsalian pes, an unusual structure of the metatarsus of Cretaceous Theropoda(Dinosauria: Saurischia). Journal of Vertebrate Paleontology, 14:480–519.HOLTZ, T. R., JR. 2000. A new phylogeny of carnivorous dinosaurs. Gaia, 15:5–61.HOLTZ, T. R., JR., D. L. BRINKMAN, AND C. L. CHANDLER. 2000. Denticle morphometrics and a possibly omnivorousfeeding habit for the theropod dinosaur Troodon. Gaia, 15:159–166.HOPSON, J. A. 2001. Ecomorphology of avian and nonavian theropod phalangeal proportions: implications for the arborealversus terrestrial origin of bird flight, p. 211–235. In J. Gauthier and L. F. Gall (eds.), New Perspectives on the Originand Early Evolution of Birds. Peabody Museum of Natural History, Yale University New Haven, CT.HORNER, J. R. 1994. Steak knives, beady eyes, and tiny little arms (a portrait of T. rex as a scavenger), p. 157–164.In G. D. Rosenberg and D. L. Wolberg (eds.), Dino Fest. Paleontological Society Special Publication 7,University of Tennessee, Knoxville, TN.HORNER, J. R. 1997. Behavior, p. 45–50. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs. AcademicPress, San Diego, CA.HORNER, J. R., AND E. DOBB. 1997. Dinosaur Lives: Unearthing an Evolutionary Saga. Harcourt Brace and Company,San Diego, CA, 244 p.263

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002HORNER, J. R., AND D. LESSEM. The Complete T. rex: How Stunning New Discoveries Are Changing OurUnderstanding of the World’s Most Famous Dinosaur. Simon and Schuster, New York, 244 p.HUNGERBÜHLER, A. 2000. Taphonomy of the prosauropod dinosaur Sellosaurus, and its implications for carnivorefaunas and feeding habits in the Late Triassic. Palaeogeography, Palaeoclimatology, Palaeoecology, 143:1–29.HUNT, A. P., S. G. LUCAS, A. B. HECKERT, R. M. SULLIVAN, AND M. G. LOCKLEY. 1998. Late Triassic dinosaurs fromthe western United States. Geobios, 31:511–531.HUNT, A. P., C. A. MEYER, M. G. LOCKLEY, AND S. G. LUCAS. 1994. Archaeology, toothmarks and sauropod dinosaurtaphonomy. Gaia, 10:225–231.HURUM, J. M., AND P. J. CURRIE. 2000. The crushing bite of tyrannosaurids. Journal of Vertebrate Paleontology,20:619–621.HUTCHINSON, J. R., AND M. GARCIA. 2002. Tyrannosaurus was not a fast runner. Nature, 415:1018–1021.JACOBSEN, A. R. 1997. Tooth marks, p. 738–739. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs.Academic Press, San Diego.JACOBSEN, A. R. 1998. Feeding behavior of carnivorous dinosaurs as determined by tooth marks on dinosaurbones. Historical Biology, 13:17–26.JACOBSEN, A. R. 2001. Tooth-marked small theropod bone: an extremely rare trace, p. 58–63. In D. H. Tanke andK. Carpenter (eds.), Mesozoic Vertebrate Life. Indiana University Press, Bloomington.JACOBSEN, A. R., AND M. J. RYAN. 1999. Taphonomic aspects of theropod tooth-marked bones from an Edmontosaurusbone bed (Lower Maastrichtian, Alberta, Canada). Journal of Vertebrate Paleontology, 19(3):55A.KELT, D. A., AND D. H. VAN VUREN. 2001. The ecology and macroecology of mammalian home range area.American Naturalist, 157:637–645.KIRKLAND, J. I. 1994. Predation of dinosaur nests by terrestrial crocodilians, p. 124–133. In K. Carpenter, K. F.Hirsch, and J. R. Horner (eds.), Dinosaur Eggs and Babies. Cambridge University Press, Cambridge, U.K.LOCKLEY, M. G. 1991. Tracking Dinosaurs: A New Look at an Ancient World. Cambridge University Press,Cambridge, U.K., 238 p.LOCKLEY, M. G., AND A. P. HUNT. 1995. Dinosaur Tracks and Other Fossil Footprints of the Western United States.Columbia University Press, New York, 338 p.LOCKLEY, M. G., AND C. MEYER. 2000. Dinosaur Tracks and Other Fossil Footprints of Europe. Columbia UniversityPress, New York, 323 p.LOCKLEY, M. G., K. J. HOUCK, AND N. K. PRINCE. 1986. North America’s largest dinosaur trackway site: implicationsfor Morrison Formation paleoecology. Geological Society of America Bulletin, 97:1163–1176.MAXWELL, W. D., AND J. H. OSTROM. 1995. Taphonomy and paleobiological implications of Tenontosaurus-Deinonychus associations. Journal of Vertebrate Paleontology, 15:707–712.MIDDLETON, K. M., AND S. M. GATESY. 2000. Theropod forelimb design and evolution. Zoological Journal of theLinnean Society, 128:149–197.MOLNAR, R. E. 2000. Mechanical factors in the design of the skull of Tyrannosaurus rex (Osborn, 1905). Gaia,15:193–218.NAISH, D., S. HUTT, AND D. M. MARTILL. 2001. Saurischian dinosaurs 2: Theropods, p. 242–309. In D. M. Martill andD. Naish (eds.), Dinosaurs of the Isle of Wight. Palaeontological Association Field Guide to Fossils 10, London.NORELL, M. A., P. J. MAKOVICKY, AND P. J. CURRIE. 2001. The beaks of ostrich dinosaurs. Nature, 412:873–874.OSTROM, J. H. 1969. Osteology of Deinonychus antirrhopus, an unusual theropod from the Lower Cretaceous ofMontana. Peabody Museum of Natural History Bulletin 30, Yale University, 165 p.OSTROM, J. H. 1972. Were some dinosaurs gregarious? Palaeogeography, Palaeoclimatology, Palaeoecology,11:287–309.OSTROM, J. H. 1978. The osteology of Compsognathus longipes Wagner. Zitteliana, 4:73–118.OSTROM, J. H. 1986. Social and unsocial behavior in dinosaurs, p. 41–61. In M. H. Nitecki and J. A. Kitchell(eds.), Evolution of Animal Behavior. Oxford University Press, Oxford, U.K.PARRISH, J. M. 1993. Distribution and taxonomic composition of fossil vertebrate accumulations in the UpperTriassic Chinle Formation, Petrified Forest National Park, p. 393–396. In S. G. Lucas and M. Morales (eds.),The Nonmarine Triassic, New Mexico Museum of Natural History and Science Bulletin 3, Albuquerque.264

FARLOW AND HOLTZ— PREDATION IN DINOSAURSPAUL, G. S. 1988. Predatory Dinosaurs of the World: A Complete Illustrated Guide. Simon and Schuster, NewYork, 464 p.RAYFIELD, E. J., D. B. NORMAN, C. C. HORNER, J. R. HORNER, P. M. SMITH, J. J. THOMASON, AND P. UPCHURCH. 2001.Cranial design and function in a large theropod dinosaur. Nature, 409:1033–1037.REISZ, R. R., AND H.-D. SUES. 2000. Herbivory in late Paleozoic and Triassic terrestrial vertebrates, p. 9–41. In H.-D. Sues (ed.), Evolution of Herbivory in Terrestrial Vertebrates: Perspectives from the Fossil Record.Cambridge University Press, Cambridge, U.K.ROGERS, R. R., C. C. SWISHER III, P. C. SERENO, A. M. MONETTA, C. A. FORSTER, AND R. N. MARTÍNEZ. 1993. The Ischigualastotetrapod assemblage (Late Triassic, Argentina) and 40 A/ 39 A dating of dinosaur origins. Science, 260:794–797.RUSSELL, D. A. 1997. Therezinosauria, p. 729–730. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs.Academic Press, San Diego.RYAN, M. J., AND M. K. VICKARYOUS. 1997. Diet, p. 169–174. In P. J. Currie and K. Padian (eds.), Encyclopedia ofDinosaurs. Academic Press, San Diego.SANDER, P. M. 1997. Teeth and jaws, p. 717–725. In P. J. Currie and K. Padian (eds.), Encyclopedia of Dinosaurs.Academic Press, San Diego.SCHWIMMER, D. R. 2002. King of the Crocodylians: The Paleobiology of Deinosuchus. Indiana University Press,Bloomington, 220 p.SERENO, P. C. 1993. The pectoral girdle and forelimb of the basal theropod Herrerasaurus ischigualastensis.Journal of Vertebrate Paleontology, 13:425–450.SERENO, P. C. 1999. The evolution of dinosaurs. Science, 284:2137–2147.SERENO, P. C, AND F. E. NOVAS. 1993. The skull and neck of the basal theropod Herrerasaurus ischigualastensis.Journal of Vertebrate Paleontology, 13:451–476.SERENO, P. C, H. C. E. LARSSON, C. A. SIDOR, AND B. GADO. 2001. The giant crocodyliform Sarcosuchus from theCretaceous of Africa. Science, 294:1516–1519.SMALLWOOD, K. S. 2001. The allometry of density within the space used by populations of mammalian Carnivora.Canadian Journal of Zoology, 79:1634–1640.THOMAS, D. A., AND J. O. FARLOW. 1997. Tracking a dinosaur attack. Scientific American, 277(6):74–79.THULBORN, R. A., AND M. WADE. 1984. Dinosaur trackways in the Winton Formation (Mid-Cretaceous) ofQueensland. Memoirs of the Queensland Museum, 21:413–517.THULBORN, T. 1990. Dinosaur Tracks. Chapman and Hall, London, 410 p.UPCHURCH, P., AND P. M. BARRETT. 2000. The evolution of sauropod feeding mechanisms, p. 79–122. In H.-D.Sues (ed.), Evolution of Herbivory in Terrestrial Vertebrates: Perspectives from the Fossil Record. CambridgeUniversity Press, Cambridge, U.K.VARRICCHIO, D. J. 2001. Gut contents from a Cretaceous tyrannosaurid: implications for theropod dinosaur digestivetracts. Journal of Paleontology, 75:401–406.WALLS, G. L. 1942. The Vertebrate Eye and Its Adaptive Radiation. Cranbrook Institute of Science Bulletin 19,Bloomfield Hills, MI, 785 p.WITMER, L. M., AND K. D. ROSE. 1991. Biomechanics of the jaw apparatus of the gigantic Eocene bird Diatryma:implications for diet and mode of life. Paleobiology, 17:95–120.WORTHY, T. H., AND R. N. HOLDAWAY. 2002. The Lost World of the Moa: Prehistoric Life of New Zealand. IndianaUniversity Press, Bloomington, 720 p.265


VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSEVOLUTIONARY PATTERNS IN THE HISTORY OFPERMO-TRIASSIC AND CENOZOIC SYNAPSID PREDATORSBLAIRE VAN VALKENBURGH 1 AND IAN JENKINS 21Department of Organismic Biology, Ecology and Evolution, University of California,Los Angeles, California 90095-1606, USA2Department of Earth Sciences, University of Bristol, Wills Memorial Building,Queen’s Road, Bristol, BS8 1RJ, EnglandABSTRACT—Synapsids include modern mammals and their fossil ancestors, the non-mammalian synapsids, or‘mammal-like reptiles’ of old classifications. The synapsid fossil record extends from the Late Carboniferous tothe present, a span of nearly 300 million years. However, it can be broken into two distinct phases of diversification,separated by about 150 million years. The first phase extends from the Late Carboniferous to the mid-Triassic,includes the first large land predators on Earth, and is almost entirely non-mammalian. The second phase beginsabout 65 million years ago after the demise of the dinosaurs, includes only mammals, and extends to the present.In this overview of synapsid predators, we emphasize terrestrial species of large size, and their adaptations forkilling and feeding, rather than locomotion. Despite fundamental differences in jaw mechanics and toothmorphology, there are significant parallels in the non-mammalian and mammalian radiations of synapsid predators.Both groups evolve sabertooth forms more than once, and both evolve short-snouted, powerful biting forms. Inaddition, both the Late Carboniferous–Triassic and Cenozoic phases are characterized by repeated patterns ofclade replacement, in which one or a few clades evolve large size and seem to dominate the carnivore guild forseveral million years, but then decline and are replaced by new taxa. Moreover, within both ancient and Cenozoicpredator clades, there are parallel trends over time toward increased body size and hypercarnivory that likelyresult from a combination of interspecific competition and energetic constraints.INTRODUCTIONTHE SYNAPSIDA IS a huge clade of amniotesthat is of major evolutionary significance andincludes modern mammals and their fossilancestors—the non-mammalian synapsids, or‘mammal-like reptiles’ of old classifications(Carroll, 1988; Benton, 1998). Synapsids possessonly one temporal opening for the jaw muscles,located low on the side of the skull, as opposed tothe two openings in diapsid reptiles and birds. Thesynapsid skull type gave rise to an astonishinglyversatile jaw system, allowing side-to-side as wellas fore-aft chewing movements (Romer, 1966).Synapsids are currently considered to comprise theparaphyletic pelycosaurs (known as pelycosaurgradesynapsids), the Therapsida, non-mammaliancynodonts (derived non-mammalian synapsids), andmammals (Kemp, 1988; Hopson, 1991; Laurin andReisz, 1995). The synapsid fossil record extendsfrom the Late Carboniferous to the present day, aspan of nearly 300 million years. However, it canbe broken into two distinct phases of diversification,separated by about 150 million years. The first phaseextends from the Late Carboniferous to the mid-Triassic, includes the first large land predators onEarth, and is almost entirely non-mammalian. Thesecond phase begins about 65 million years ago afterthe demise of the dinosaurs, includes only mammals,and extends to the present.The Carboniferous to Triassic fossil record ofnon-mammalian synapsids includes a wide varietyof carnivores. Not only are sphenacodont synapsidsthe first mega-carnivores on Earth, but succeedingsynapsid predators—such as dinocephalians andLate Permian gorgonopsids—made up the firstdiverse carnivore guilds on Earth and pre-datethose of Cenozoic mammals by as much as 200267

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002million years. The 170-million-year hiatus betweenthe last diverse synapsid carnivore community(Middle Triassic cynodonts) and the first Cenozoicpredators (creodonts) was the result of dinosaurprevalence during that interval (Benton, 1998).Predatory non-mammalian synapsids wereprofoundly dominant in their niches throughout thePermo-Carboniferous, a pivotal period in Earthhistory for tetrapod evolution. At this time, thefundamental amniote groups—anapsids, diapsids,and synapsids—had their origin and firstdiversification. In addition, ancient amphibians—chiefly temnospondyls—radiated tremendouslyand commanded inland freshwater aquatic andsemi-aquatic habitats (Carroll, 1988; Benton,1998). Carnivores in any ecosystem applyconsiderable ‘evolutionary pressure’ and areimplicated in such adaptations among prey speciesas crypsis, armor and cranial adornments, tusks,increased cursorial capacity, and greater size(Gittleman, 1989, 1996). This was as true forancient terrestrial ecosystems as it is today. Gaininginsights into the structure and composition ofancient predator guilds should shed light ontetrapod evolution during the Late Paleozoic.Predators, as defined in this overview, areanimals that regularly kill and eat other vertebrates(Savage, 1977). In the interest of brevity, andbecause the fossil record is more complete for largerthan for smaller vertebrates, we will largely restrictour review to terrestrial species of large size—thatis, jackal-size (about 7 kg) and larger. Our primaryfocus will be on heads and teeth and hence killingand feeding behavior, rather than on limbs andlocomotion. This reflects the fact that cranial anddental material is much more abundant in the fossilrecord than are limbs, especially in the case of thePermo-Triassic synapsids. The review begins witha look at overall similarities and differences in theanatomy and feeding mechanics of the two widelyseparated radiations of carnivores. This is followedby a summary of the evolutionary history withineach group and then by an attempt at a synthesis,noting parallel trends in morphology and guilddynamics between the Permo-Triassic and Cenozoicsynapsid predators.NON-MAMMALIAN ANDMAMMALIAN SYNAPSIDCARNIVORES COMPAREDAll but the latest Permo-Triassic synapsidcarnivores differed from Cenozoic synapsids in jawmechanics. Reconstructing the fine details ofmuscles in fossil vertebrates is problematic, butgeneral directions of muscle forces (vectors) canbe estimated. In addition, the overall morphologyof the skull and jaws provides some indication offunction (Parrington, 1955; Olson, 1961; Kemp,1969, 1982). Based on their anatomy, most ancientcarnivores used a predominantly Kinetic-Inertial(KI) system of jaw closure in which the jaws wererapidly slammed shut, disabling the prey by themomentum and ultimately the kinetic energy ofthe jaws and teeth (Olson, 1961). The major jawclosing muscles were the anterior pterygoideus andtemporalis (Fig. 1). In more derived amniotes, suchas modern mammals, the KI system wassupplemented (and supplanted) by the Static-Pressure (SP) system. In SP systems, a strong bitewith nearly closed jaws serves as the killing bite(Olson, 1961). The pterygoideus musculature isreduced; instead, the muscles driving the closureare the temporalis and the masseter (Fig. 1).Modern examples of predators that use KI orSP systems offer insights into differences in jawmusculature and function that can be applied tofossil tetrapods. For example, the rapid and highlydamaging jaw adduction of crocodilians is a goodexample of a KI system combined with a significantSP bite (Cleuren and de Vree, 2000). Incrocodilians, the fast KI bite is generated by anexpanded anterior pterygoideus musculaturewhich, when the jaws are wide open, lies at 90 o tothe long axis of the lower jaw. This is the optimalalignment for the input of forces (muscular) to alever (lower jaw). The temporalis musculature incrocodilians is aligned at 90 o to the long axis ofthe lower jaws when they are almost closed andthus provides the SP component of the bite. Largefelids such as lions use a more purely SP system inwhich the jaws are closed more slowly but withgreat force (Biknevicius and Van Valkenburgh,268

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSFIGURE 1—Kinetic-Inertial (KI) vs. Static-Pressure(SP) bite systems as illustrated by a gorgonopsid(Arctognathus) and a modern cat (puma, Pumaconcolor), respectively. 1, In the KI system, thelargest jaw adductor (P) is the anterior pterygoideusmusculature. Note that it has its maximum leveragefor jaw closing when the jaw is wide open, and caninitiate a rapid closure; A = external adductors. 2, Inthe SP system, the largest jaw adductor is thetemporalis musculature (T) followed by the masseter(M); the pterygoideus musculature is reduced. Inthe SP system the major jaw adductors have theirmaximum advantage when the jaws are nearlyclosed. (1) modified from Kemp, 1982, fig. 40.1996). The anterior pterygoideus musculature offelids is greatly reduced so there is little KIcomponent, but the mechanical advantage (distancefrom fulcrum to effort divided by distance fromfulcrum to bite point) of the expanded temporalismuscles is great because it has a very long momentarm (the coronoid process). Moreover, much of thetemporalis is aligned close to 90 o relative to thisstructure, enhancing muscle force when the jawsare nearly closed (Russell and Thomason, 1993).The SP bite system may have replaced the KIsystem in synapsids because impact stresses onteeth and bones were lessened during the killingbite. In addition, the SP system allowed a firmergrip on prey and better control of teeth at occlusion.The masseter musculature enhanced the bite forceof slicing cheek teeth and reduced the need for largeinternal pterygoideus musculature that couldobstruct air and food pathways.In addition to jaw mechanics, Permo-Triassicmammalian predators differ from their Cenozoicequivalents in tooth morphology. With the exceptionof some very derived, nearly mammalian,cynodonts, Permo-Triassic synapsid predatorsdisplayed little differentiation in dental morphologyalong the tooth row, except in size (Figs. 1, 2).Numerous simple conical to blade-shaped teeth linedthe jaws. Large caniniform teeth were often presenton both upper and lower jaws, thus separating thetooth row into incisors, canines, and cheek teeth.Unlike modern carnivores, there was no tooth-totoothocclusion among the cheek teeth and no preciseshearing between the upper and lower teeth. As aresult, the diversity of dental types among nonmammaliansynapsid carnivores was much lowerthan that expressed among mammalian carnivores.In mammals, the evolution of precise occlusion andside-to-side jaw movements (in association with theevolution of the masseter musculature) allowedmammals to produce more effective grinding as wellas slicing actions. Cenozoic carnivores includespecies with teeth specialized for herbivory (e.g.,Giant Panda), omnivory (e.g., Brown Bear),carnivory (e.g., felids), and bone-cracking (e.g.,hyaenids). By contrast, the radiation of nonmammaliansynapsid carnivores produced onlyspecies specialized for meat-eating, with littleevidence of adaptation for consuming bone or nonvertebratefoods. Included among these were the firstsabertooth predators, foreshadowing an ecomorphthat was to appear repeatedly among mammals.269

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002THE EVOLUTIONARY HISTORYOF PERMO-TRIASSICSYNAPSID PREDATORSExisting knowledge of terrestrial tetrapodpredators from the Late Carboniferous to themiddle Late Triassic has been broadly outlined aspart of stratigraphical studies on these faunas(Anderson and Cruickshank, 1977; Rubidge, 1995)and systematic analyses of amniotes (Laurin andReisz, 1995) and of synapsids in general (Hopson,1991; Hopson and Barghusen, 1986; Reisz, 1986;Kemp, 1988; Sidor and Hopson, 1997; Sidor,2001). The fossil record of non-mammaliansynapsids is more complete than that of anyterrestrial vertebrate group except Cenozoicmammals; it allows their evolutionary history andthe nature of their adaptations to be considered ingreat detail. The early history of synapsidsproduced at least two firsts in Earth history.Synapsids include the first terrestrial amniotes onearth that acted as specialized herbivores (Modesto,1995). But, of equal significance, synapsids includethe Earth’s first large land predators, animalscapable of killing and eating other vertebrates. Thefirst obligate carnivore to appear on land was thesphenacodont pelycosaur-grade synapsidFIGURE 2—Skulls of Permian carnivorous non-mammalian synapsids. 1, sphenacodont pelycosaur-gradesynapsid Dimetrodon. 2, Anteosaur dinocephalian Titanophoneus. 3, Early therocephalian Lycosuchus.4,gorgonopsid Leontocephalus. 5, Chiniquodontid cynodont Probelesodon. (1–4 modified from Kemp, 1982.)270

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSDimetrodon during the Early Permian (Fig. 2.1).At that time, Dimetrodon alone had a craniodentaldesign that was sufficiently robust to disable andconsume large vertebrates; it gave rise to theanatomically progressive therapsids, which includethe saber-toothed genera (Kemp, 1982).So dominant were synapsids as terrestrial apexpredators that no member of any other amnioteclade took this role alongside synapsid killers. Theabundant, large, and formidable temnospondylamphibian Eryops of the Permo-Carboniferous wasperhaps the only possible instance of any nonsynapsidpredatory competition. Other gigantictemnospondyl amphibians were comparable tosynapsids in terms of their habitat dominance, butwere semi- or fully aquatic (Carroll, 1988; Benton,1998), as was Eryops. The history of nonmammaliansynapsids records five ecologicaldynasties of predator communities. Basaltherapsids form a possible sixth but are poorlyknown, and it is not yet feasible to reconstruct acommunity of these carnivores. An ecologicaldynasty of predators is here defined as a singleclade or small number of clades (relative to all thatare present) that fill the large-body-size, highlycarnivorous roles within a community for millionsof years. The five predator dynasties (Figs. 2, 3) ofthe non-mammalian Synapsida, in chronologicalorder, are sphenacodont ‘pelycosaurs’ (Permo-Carboniferous), anteosaur dinocephalians (middleLate Permian), therocephalians (Late Permian),gorgonopsids (Late Permian), and non-mammaliancynodonts (Permo-Triassic). These five dynastiesspan a total stratigraphic range extending from theLate Carboniferous (sphenacodonts) to the MiddleTriassic (chiniquodontid cynodonts).Permo-Carboniferous (approximately 300–270Ma): Pelycosaur-grade synapsids.—In pelycosaurgradesynapsids of the Permo-Carboniferous, theonly large terrestrial predators were Dimetrodon,Ophiacodon (Romer and Price, 1941; Reisz, 1986),and possibly Secodontosaurus (IJ, pers. obs.).Secodontosaurus (Reisz et al., 1992) was a longsnoutedform superficially similar to the muchlarger Ophiacodon. Secodontosaurus has beenconsidered to have been an insectivore that usedits elongate rostrum to reach the invertebrateoccupants of burrows (Reisz et al., 1992). However,its cranial size (280 mm in length) was comparableto that of a mature Komodo Monitor. Its cranialmorphology (narrow, deep, and long) shows a morerobust skull than that of Ophiacodon, a welldevelopedregion for the origin of the anteriorpterygoideus musculature, and stout, sharp teeth.These details reflect true carnivory, not insectivory,for which numerous simple teeth are more typical.Its rostrum was not sufficiently narrow to accessinvertebrate burrows but may have been able toenter those of small tetrapods. Ophiacodon had thelongest skull of any pelycosaur, and some remainssuggest an animal with a skull length in excess of500 mm (Romer and Price, 1941) and a total lengthof more than four meters. It is usually consideredto have been a semi-aquatic piscivore on the basisof its unossified ankle and wrist bones and its dentalarcade of many small conical teeth (Romer andPrice, 1941). Its skull size suggests that tetrapodprey was an option, but this is countered by thestructurally fragile nature of its lower jaws and arostrum that was not particularly strong in bendingand torsion (IJ, pers. obs.). Hence, the only largemega-carnivore in Permo-Carboniferous terrestrialecosystems was Dimetrodon.Dimetrodon species had very deep skulls witha strongly convex dorsal profile and well-developedtemporalis and anterior pterygoideus musculature.The reflected lamina of the angular bone in thelower jaw was the largest of any ‘pelycosaur’ andheralds the expansion of this structure in thetherapsids, where it becomes part of the hearingcomplex (Kemp, 1982). In Dimetrodon it supportedan enlarged anterior pterygoideus musculaturewhich operated a powerful KI bite. Dimetrodonhad a strong but not especially robust lower jawand a characteristic dentition (Romer and Price,1941; Reisz, 1986). The postcanine dentitionincluded fewer but larger teeth in the upper jaw.These teeth are recurved and very sharp withserrated edges; their cross section is ovoid to subrounded,suggesting considerable structuralstrength (Fig. 2.1). The pterygoid flanges in thepalate were huge and robustly designed; their edges271

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 3—Stratigraphic ranges and time duration of the carnivorous groups of non-mammaliansynapsids (modified from Kemp, 1982, fig. 115). A star indicates the presence of a saber-toothed formwithin the lineage.lay at 90 o to each other and were equipped withwell-developed pterygoid teeth. These structuresmay have functioned as intraoral ‘guillotine blades’(IJ, pers. obs.) and allowed the enhanced triturationof ingested food. The canines are very large andextremely robust, but the huge, caniniform incisorsare similar in size to the canines (Romer and Price,1941). There is a pronounced step at the maxilla–premaxilla junction of the upper jaw. The caninesand incisors are situated on opposite sides of thisstep, allowing the equally large lower canines tointerpose between the upper canines and incisors,creating a tight grip on prey. Sphenacodonts werehighly successful predators as shown by a long fossilrecord (about 40 million years) that extends fromthe Late Carboniferous to the Guadalupian (earlyLate Permian) (Benton, 1993). The skull designremained unchanged throughout this long period.Mid-Late Permian (approximately 275–260Ma): Dinocephalians.—The first of three therapsiddynasties belongs to the dinocephalians—large,often gigantic, synapsids from the early LatePermian of South Africa and Russia (Fig. 2.2). Overtheir evolutionary history, they increased in size;their skull bones thickened; the adductor(temporalis) chamber became more vertical inorientation; and they developed reniform (kidneyshaped)palatal tooth bosses of considerable size,robust canines, and very large, procumbent incisorsthat interlaced in a precise manner (Kemp, 1982).Dinocephalians were the largest of all synapsidsand rivaled some dinosaurs in size. The carnivorousmembers belong to the anteosaur clade (Hopson,1991; Hopson and Barghusen, 1986) and includethe South African form Anteosaurus, and thesomewhat smaller Russian forms Titanophoneus(Fig. 2.2), Syodon, Archaeosyodon, Doliosaurus,and Doliosauriscus.Anteosaurus from South Africa had a high,deep skull almost a meter long (Kemp, 1982)equipped with mighty canine teeth, as did theRussian forms Archaeosyodon, Doliosaurus, andDoliosauriscus (Orlov, 1958; Tatarinov, 1974;Chudinov, 1983; Battail and Surkov, 2000). The272

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSbody length of Anteosaurus is estimated at morethan 5 meters (Boonstra, 1954). Large pterygoidflanges indicate a well-developed KI system inanteosaurs, and increased vertical alignment of thetemporalis muscles suggests an expanded SPcomponent of the bite cycle. Unlike later therapsids,in which the postcanine dentition was reduced,anteosaurs retained the three elements of their dentalarcade as fully functional units. Anteosaurspossessed the heaviest canines, incisors, and skullsof all carnivorous synapsids. The postcanine teethare large, trenchant, ovoid in cross section, and hencevery strong. In many anteosaur genera, the caninesare recurved, in contrast with the straighter caninesof all other carnivorous synapsid genera. Thispronounced recurvature is best seen in Syodon fromRussia. In addition to the complete functionalcomplement of marginal dental units, anteosaursalso possessed very prominent palatal tooth bosses.These reniform structures are more conspicuousthan in any other predatory synapsid andaugmented the marginal dentition considerably.Anteosaur head skeletons thus show a trend towarda remarkably expanded dental complement,comprising incisors, canines, postcanines, andpalatal teeth, all of which were large. In concert withhuge size, a powerful SP bite, and exceptionallyheavy, thick-boned skulls, anteosaur dinocephalianswere designed to prey on particularly large animalsand were among the most highly predaceous of allsynapsids (Sennikov, 1996). Potential prey includedthe bull-sized armored pareiasaurs (Lee, 1997) andenormous tapinocephalid dinocephalians (Rubidge,1995). The latter are typified by dome-headedMoschops; it and related genera such asTapinocephalus were the heaviest of all synapsids(Kemp, 1982). Despite their remarkable craniodentalweaponry, dinocephalians do not have a fossil recordas long as that of sphenacodonts and are replaced inthe early part of the Late Permian by earlytherocephalians and gorgonopsids (Anderson andCruickshank, 1977; Kemp, 1982; Rubidge, 1995;Battail and Surkov, 2000).Late Permian (approximately 260–250 Ma):Gorgonopsids.—A number of gorgonopsidecomorphs or ecotypes can be defined, based onskull construction and jaw biomechanics (Fig. 4).Cyonosaurids such as Cyonosaurus (Fig. 4.1) weresmall (skull length about 140 mm) gorgonopsidswith very long snouts, small canine ‘sabers’, slimlower jaws, and gracile skulls. These forms maybe immature or juvenile forms of other genera (IJ,pers. obs.). Cyonosaurids have a long fossil record(relative to other gorgonopsids) and seem to haveFIGURE 4—Gorgonopsid skull morphotypes that probably represented different predatory habits andniche occupation. See text for more information. 1, Cyonosaurus; 2, Gorgonops; 3, Arctops;4, Arctognathus; 5, Rubidgea; 6, Broomicephalus. Scale bar = 100 mm. Skulls not to scale.273

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002filled roles analogous to those of modern small,long-snouted canids such as jackals and foxes thatprey on both arthropods and small vertebrates.Larger gorgonopsids such as Leontocephalus,Scylacops, Aelurognathus, and Gorgonops (Fig.4.2) display fairly long rostra and skulls with a flatdorsal profile, large canine teeth, and up to fourpostcanine teeth (Kemp, 1969; Sigogneau, 1970;Sigogneau-Russell, 1989). In these genera, theadductor chamber shows the characteristic ‘flaring’to produce expanded areas of origin for thetemporalis musculature. The pterygoid flanges aresituated far anteriorly, allowing for long anteriorpterygoideus fibers that powered a strong KI bite(Jenkins, 1998). A number of even largergorgonopsid genera, such as Arctops (Fig. 4.3),display shorter skulls with a very strongly convexdorsal profile that is reminiscent of the sphenacodontDimetrodon. The convex profile is especiallyefficient for accommodating the stresses imposedduring powerful canine biting (Jenkins, 1998). Inespecially large (skull length about 300 mm),convex-skulled genera such as Arctops, the reflectedlamina of the angular bone is expanded and henceprovided an increased insertion area for acorrespondingly enlarged anterior pterygoideusmusculature. This indicates a very powerful KI biteand the capacity to accelerate the jaws shut from avery wide gape. The saber teeth in some specimensare exceptionally long, and would have necessitateda wide gape to clear the tips of the huge teeth(Jenkins, 1998). Another gorgonopsid ecotype isdefined by the short-snouted genus Arctognathus(Fig. 4.4). This small to medium-sized gorgonopsidhas a box-like robust skull with a deep rostrum anda rather flat dorsal profile; the canines are large,areas of bone associated with the anteriorpterygoideus musculature are expanded, and thesymphysis of the lower jaw is enlarged. Thesedetails suggest a rostrum that was strong in bendingand torsion, powerful KI and SP constituents ofthe bite cycle, and a stable lower jaw (Jenkins,1998, 2001a, 2002b; Jenkins et al., 2002).Gorgonopsids attain their most impressivemanifestation in their most derived members, theRubidgeinae. These were the latest-survivinggorgonopsids and are found solely (so far) in SouthAfrica just below the horizon beds of the end-Permian extinction event (Rubidge, 1995).Rubidgeine gorgonopsids are characterized by asuite of cranial features that evolved for anexceedingly predaceous way of life. The skulls areoften huge, exceeding 400 mm in length. Skullbuttressing is extremely pronounced. The rostrumis broad, pterygoid flanges are situated far anteriorly,the lower jaw is massive, and the symphysis isexpanded more than in other gorgonopsid genera.The incisors are very large, postcanine teeth areabsent, and the saber canines are huge—those ofRubidgea atrox (Fig. 4.5) are longer than the teethof the theropod dinosaur Tyrannosaurus rex(Jenkins, 1998). Rubidgeine gorgonopsids aredefined taxonomically by a posterior ventral flangeon the zygomatic arch (Sigogneau-Russell, 1989),which appears to reflect an enlarged zygomaticomandibularismuscle (Kemp, 1969) that stabilizedthe lower jaw and jaw joint during use of the largesaber teeth (Jenkins, 1998). Within the Rubidgeinae,distinct ecotypes are defined by the huge buttressskulledRubidgea, the very wide-snoutedClelandina, and the convex-skulled, long-sabertoothedDinogorgon and Prorubidgea (figured inSigogneau, 1970; Sigogneau-Russell, 1989). A finalgorgonopsid ecomorph is represented by one ofthe rarest rubidgeines, Broomicephalus (Fig.4.6).Broomicephalus is the only known gorgonopsid inwhich maximum skull width exceeds total skulllength (Sigogneau, 1970; Sigogneau-Russell, 1989;Jenkins, 1998; Jenkins and Dyke, in press). It hasan unusually broad rostrum, short skull, especiallyflat blade-like sabers, and the largest jawsymphyseal mass of any gorgonopsid. The widerostrum includes an expanded origin area for theanterior pterygoideus musculature, indicating moreof an emphasis on the KI component of its bitecycle than in some other rubidgeines. Associatedwith this are rather small adductor chambers,suggesting a reduced SP constituent. Wide jawsare very stable and robust (Covey and Greaves,1994); large jaw symphyses increase thepenetration potential of the sabers and allow bothsides of the jaw adductor muscle force to be274

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSutilized. Broomicephalus probably preyed uponunusually large animals (Jenkins, 1998), and itsunusually flat sabers foreshadow the canines ofscimitar-toothed cats such as Homotherium.Despite the complexity and diversity in LatePermian carnivore guilds, all gorgonopsids becameextinct during the end-Permian event. To someextent, gorgonopsid ecological supremacy came atthe expense of saber-toothed carnivoroustherocephalians, which died out before the lastgorgonopsids and were to a large extent replacedby them as top predators (Jenkins, 1998, 2001a;Jenkins et al., in press).Late Permian (approximately 265–250 Ma):Therocephalians.—Therocephalians overlap the tenmillion-yearspan of the gorgonopsids and extendslightly further into the basal Triassic (Fig. 3). Earlytherocephalians include the families Lycosuchidaeand Scylacosauridae, and both were saber-toothed.Their saber canines are modest compared with thoseof large gorgonopsids, although in biggerscylacosaurid genera and in Lycosuchus they arevery pronounced (Fig. 2.3). The therocephalian SPbite is larger than that of gorgonopsids due toenlarged adductor chambers for the temporalismusculature (Kemp, 1982). The anteriorpterygoideus muscle fibers inserted relatively closeto the jaw joint, thus reducing the mechanicaladvantage of the KI system relative to that ofgorgonopsids but increasing the speed of closure(Jenkins, 1998). The palates of early therocephalianshave a large number of discretely located and welldevelopedscarf joints that reduce biting stresses(Jenkins, 1998; Jenkins et al., 2002). Lycosuchushas a shorter snout than most scylacosaurids suchas Glanosuchus, Scymnosaurus, and Ptomalestes,and its sub-rounded shape suggests strength intorsion in association with proportionally largesabers (Jenkins, 1998). The short snout also gives agreater mechanical advantage for bites at the canines.The therocephalian group Euchambersiidae(Hopson and Barghusen, 1986) contains the highlydistinctive genera Euchambersia and Moschorhinus.Both these forms are characterized by very short,robust rostra and canines that are rounded in crosssection.Euchambersia is known from only twopartial skulls; Moschorhinus is much better known.Moschorhinus skull dimensions range from thoseof a large monitor lizard to those of a lion. It is amost unusual therocephalian and shows a largetemporal fenestra, convex palate, and laterallycompressed incisors set in a shallowly curvingarcade, in addition to the short, strong snout(Fig. 5.2). The lower jaw has remarkable caninesand a symphysis that is expanded more than in anyother therocephalian (Jenkins, 2001b). The caninesare as long as the sabers of all except the mostderived gorgonopsids (Fig. 5.1), but are round incross-section, resembling the spikes used to securesome railway cross-ties. There is no modern analogfor these teeth, the closest perhaps being the hyperelongatecanines of the Clouded Leopard (Neofelisnebulosa) (Fig. 8.2). The combination of long, robustcanines and flattened incisors in a shallow arcadesuggests a felid-like mode of attack (sensuBiknevicius et al., 1996), in which the strong canineswould have sustained a long period embedded inprey (Jenkins, 2001b). This is the earliest knowninstance of this design. Moschorhinus went extinctafter gorgonopsids and was a therocephalianoccupying a gorgonopsid-like ecological niche(Jenkins, 2001b). The powerful symphysis, robustsnout, and extraordinary canines show Moschorhinusto have been a truly formidable predator; why it shouldhave survived when equally powerful and similarlydesigned gorgonopsids went extinct is probably dueto some as-yet-unknown feature of its physiology.Once Moschorhinus had disappeared, the roles of toppredators were taken up by cynodonts (Fig. 3).Permo-Triassic (approximately 255–225 Ma):Cynodonts.—Small cynodonts in the basal-mostTriassic (Lystrosaurus Zone) are represented byGalesaurus and Thrinaxodon (Broom, 1932). Theyshow expanded temporalis musculature and thefirst signs of the masseteric musculaturecharacteristic of modern mammals (Kemp, 1982;Hopson, 1991; Hopson and Barghusen, 1986). Thesnout is short and the canines dog-like; thepostcanine teeth are numerous and form a seriesof curving blades (Kemp, 1982), similar to thoseof the Komodo Monitor (Varanus komodoensis).In general skull proportions and size, they most275

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 5—Diagram illustrating the geometricdifferences in skull, jaw and tooth morphologyamong: 1, a very derived gorgonopsid,Dinogorgon; 2, the unusual therocephalianMoschorhinus, which represents a convergencewith the gorgonopsid cranial design; and 3, thelarge Triassic cynodont Cynognathus. Thisillustration charts the change from a largely KI bite(with some SP component)—Dinogorgon—to anSP bite with very little KI component—Cynognathus. Ventral views of the anterior palatesare shown to the right. Note the parabolic incisorarcade of gorgonopsids, and the peculiarly felid-likestraight incisor arcade of Moschorhinus along withits palatal fenestrae (for muscular attachments). InCynognathus an almost complete osseoussecondary palate has formed. The flat sabers ofthe gorgonopsid can be contrasted with the roundelongate ‘spikes’ of Moschorhinus. See text fordetails. Scale bar = 100 mm. Skulls not to scale.resemble modern-day mustelids (weasels) andviverrids (civets), and probably behaved more likemodern hypocarnivorous/omnivorous viverridsthan specialized hypercarnivorous mustelids.The largest carnivorous cynodont,Cynognathus, defines the Cynognathus Zone of theLower-Middle Triassic (succeeding the LystrosaurusZone), and has a disproportionately large, robustskull (Broom, 1932; Kemp, 1982). More than 400mm long, its proportions are similar to those ofextant canids, characterized by a long rostrum. Thedentition is broadly similar to that of Galesaurusand Thrinaxodon, with pointed incisors, welldevelopedcanines, and a long row of sharp, recurvedpostcanine teeth. There are two striking features ofthe Cynognathus skull—both the lower jaw andzygomatic arch are extraordinarily deep vertically(Fig. 5.3). The zygomatic arch is so deep that theorbit appears as a simple hole ‘punched’ throughthe ‘plate’ formed by the zygomatic arch (Kemp,1982). Application of lever mechanics to theCynognathus skull reveals a very powerfultemporalis musculature and an expanded massetericcomponent. There is little KI component to the bite,but the SP constituent must have been extremelypowerful. This and the overall robustness of the skullsuggest a hyena- or bear-like design. However,Cynognathus did not have the bone-crackingpremolars or carnassial blades of modern hyaenids.Its trenchant, recurved postcanine teeth, in concertwith this strong SP bite and huge lower jaw, indicatethat Cynognathus used its postcanine teeth to a muchgreater extent than perhaps all previous synapsidpredators (IJ, pers. obs.).Later cynodont predators are thechiniquodontids of the Middle and Late Triassic.Their skull dimensions (Romer, 1969a, 1969b) rangefrom those of a modern coyote (Probelesodon, Fig.2.4) to those of a large hyena (Chiniquodon andBelesodon). Analysis of jaw muscle biomechanicsindicates that the horizontal component of the jawmuscle vectors in cynodonts increased throughoutthe lineage and served to reduce reaction forces atthe jaw joint (Crompton, 1963, 1972; Bramble,1978). At this stage, the jaws of carnivorouschiniquodontids were functioning in an almostidentical manner to those of modern carnivores.276

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSNON-MAMMALIANSYNAPSID PREDATORS:LONG-TERM TRENDSThe long-term evolutionary trends seen in thecrania of carnivorous non-mammalian synapsidpredators are essentially those documented for thewhole of the Permo-Triassic Synapsida (Romer,1966; Kemp, 1982; Hopson and Barghusen, 1986).They include an increase in solidification of thebraincase wall, reduction in the number of skullbones, expansion of the temporalis musculature anddevelopment of the masseteric musculature,reduction of the pterygoideus component, increasedheterodonty, formation of an osseous secondarypalate, increased brain size, and reduction of thepostdentary bones, which evolve into bones of themammalian middle ear (Kemp, 1982; Sidor, 2001).In terms of skull mechanisms relating to acarnivorous habit, the KI bite system is replaced bythe SP bite system, and the saber-toothed adaptationprevalent in the Late Permian is not seen aftergorgonopsids until the advent of Cenozoiccreodonts. In temporal sequence, convex-skulled,step-jawed sphenacodonts were followed bypachyostotic (thick-boned) skulls and gigantismamong anteosaurs. These in turn were succeededby a variety of saber-toothed forms, both earlytherocephalians and gorgonopsids, which dominatedLate Permian ecosystems. Smaller, rather ‘mammallike’carnivores of the Triassic cynodont cladesupersede the Late Permian saber-toothed predatorcommunity. These animals were close to mammalancestry and very similar to modern mammals intheir skull design (Kemp, 1982).THE EVOLUTIONARY HISTORYOF CENOZOICSYNAPSID PREDATORSSynapsids relinquished the role of dominantterrestrial predator to theropod dinosaurs for theremainder of the Mesozoic. Then, following thedemise of the dinosaurs about 65 Ma, about tenmillion years passed before the first appearance ofa large, specialized meat-eater (e.g., Oxyaena;Gunnell, 1998). Thus, the period to be covered inthis part of the review, the last 55 million years, ismuch shorter than that spanned by the abovedescribednon-mammalian synapsids (80–100million years). Nevertheless, because of its relativerecency, the fossil record of carnivorous mammalsis much richer in detail and diversity than that ofthe ancient synapsids. Here we can provide only abrief summary, but there are several recent reviewsof carnivore history that offer more information (e.g.,Martin, 1989; Hunt, 1996; Werdelin, 1996; VanValkenburgh, 1999). Our emphasis will be onnotable morphological innovations and repeatedpatterns in the history of carnivorous mammals,including convergence on similar ecotypes anddynasty replacement (as was described for Permo-Triassic synapsids). Most of the discussion relieson examples from the excellent North American andEurasian fossil records.Today, the large, terrestrial predators of theworld all belong to a single placental order, theCarnivora. Until very recently, there was onemarsupial predator of reasonable size, the Australianthylacine (Thylacinus cynocephalus), but itsextinction has left the carnivorans (members of theorder Carnivora) as sole occupants of this adaptivezone. In the early part of the Cenozoic, two nowextinctorders, Mesonychia and Creodonta,dominated predator guilds. The mesonychids wereperhaps the first Cenozoic predators, although theirdentition is not highly specialized for meat-eating,and they may have been quite omnivorous.Mesonychids, such as the 62 Ma coyote-sizedDissacus, looked somewhat dog-like externally, buttheir teeth differed markedly from those of canids.The lower tooth row of Dissacus includes a seriesof premolar-like teeth with blunt cusps and weaklydeveloped cutting blades (Fig. 6.1) that seem lessspecialized for carnivory than those of the Triassiccynodont Cynognathus (Fig. 5.3). The firstunquestionable Cenozoic hypercarnivores(specialists on vertebrate prey) were among thecreodonts. Unlike mesonychids, creodonts evolvedspecialized slicing molar teeth, in which upper andlower molars occluded like scissors and created largeshear facets on opposing teeth (Fig. 6.2).277

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Carnivorans also evolved specialized slicing teeth,but unlike creodonts, they have only a single pair,the upper fourth premolar and lower first molar(Fig. 6.3). This single carnassial pair is a keydiagnostic character of the order Carnivora. In atypical canid, the carnassials lie about midway alongthe tooth row, separating somewhat conicalpremolars from more flattened rear molars. Thisvariety of tooth form has allowed great evolutionaryversatility within the order. By enhancing one set ofteeth over another, species can evolve towardomnivory (larger rear molars for grinding; e.g.,FIGURE 6—Lateral view of the lower jaws of 1,Dissacus, a mesonychid; 2, Hyaenodon crucians,a creodont; and 3, Hesperocyon gregarius, a canidcarnivoran. Arrows indicate molar teeth specializedfor cutting. Dissacus is modified from Fig. 1 ofO’Leary and Rose (1995b); the others are fromcolor transparencies.bears), carnivory (larger carnassials; e.g., cats), orbone-cracking (larger, bulbous premolars; e.g.,hyenas) (Ewer, 1973).In both Eurasia and North America, theCenozoic history of mammalian carnivorescomprises a series of five dynasties (VanValkenburgh, 1999). Each dynasty is defined bythe prevalence in body size and diversity of one ormore groups. The taxa are not identical on bothcontinents, but there is considerable overlapreflecting the fact that the two have been connectedseveral times over the last 65 million years. InNorth America, the sequence of large predatordynasties was: 1) creodonts; 2) creodonts andnimravids; 3) canids, amphicyonids, hemicyonineursids; 4) borophagine canids and felids; and 5)canine canids and felids. In Eurasia, the pattern isvery similar, except that hyenas fulfill many of thesame roles filled by canids up until the Pliocene(Werdelin and Solounias, 1991, 1996). We nextreview each of these dynasties, highlightingsignificant morphological advances.Early to middle Eocene (approximately 55–40Ma): Creodonts.—Creodonts appear in the latePaleocene (about 57 Ma) and reach their maximumdiversity in North America in the early to middleEocene (48–54 Ma). Early forms were civet-like,preying on relatively small mammals, birds, andperhaps arthropods (Gunnell, 1998). Over time,creodonts expanded into more hypercarnivorousforms, including the first sabertooth predator sincethe Permian. This early sabertooth, Macheroides, ispoorly represented but appears to have been aboutthe size of a bobcat (Lynx rufus), which is muchsmaller than typical sabercats, though comparableto some Late Permian gorgonopsids (Dawson et al.,1986). There were also a few very short-facedcreodonts in the Eocene, with deep jaws and robustpremolars, that may have been bone-crackers, suchas Patriofelis and Paleonictis. This diverse array ofcreodonts ranged in body size from that of foxes tothat of leopards and they were the dominantcarnivores of the Eocene. They coexisted withsmaller species, feliforms and caniforms that representthe earliest members of the order Carnivora (Flynn,1998), as well as a few omnivorous mesonychids.278

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSThe creodonts included ambulatory, stoutlimbedspecies (oxyaenids) as well as more slenderlimbed,cursorial forms (hyaenodontids), althoughnothing was as cursorial as a modern canid(Gunnell, 1998). This ecological diversity offorms declined to a single subfamily, theHyaenodontinae, in the late Eocene of NorthAmerica. Hyaenodontines were fairly cursorialhypercarnivores with well-developed shearingdentition set into long jaws. They were capable oflarge gapes and may have emphasized KI biting aswell as SP biting, as their masseter muscles werereduced relative to their internal pterygoid muscles(Mellett, 1969). Hyaenodontines persisted in NorthAmerica for about 12 million years, sharing the roleof dominant carnivore with the cat-like Nimravidae.Late Eocene to Oligocene (approximately 40–28 Ma): Creodonts and Nimravids.—The secondCenozoic dynasty was made up of hyaenodontinecreodonts and sabertooth nimravids (Fig. 7).Coexisting with them at somewhat smaller sizeswere more omnivorous species within the familiesCanidae and Amphicyonidae. The familyNimravidae includes the first truly cat-likepredators bearing grappling forelimbs armed withretractile claws, enlarged canines, andforeshortened skulls with no postcarnassialcrushing molars. The convergence in form betweennimravid and some felid carnivorans is so marked(Emerson and Radinsky, 1980) that nimravids wereconsidered to be felids until relatively recently(Hunt, 1987; Bryant, 1991). Almost all nimravidshad enlarged, laterally flattened saber-like caninesthat slid into a sheath on the lower jaw (Fig. 8.3).A single, poorly known early Miocene genus,Dinaeleurus, displays conical, more peg-likeFIGURE 7—Stratigraphic ranges and time duration of the carnivorous groups of large North Americanmammals. Data are from Janis et al. (1998).279

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 8—Lateral views of extinct sabertooth (1, 3, 4) and extant conical-tooth (2) cats. 1, Smilodonfatalis; 2, Neofelis nebulosa (Clouded Leopard); 3, Barbourofelis fricki; and 4 , Thylacosmilus atrox. (1)and (2) are members of the family Felidae, (3) is a member of the extinct family Nimravidae, and (4) isa South American marsupial.canines as is typical of extant cats (Martin, 1998a).The first nimravids appear about 37 Ma in NorthAmerica as two sabertooth genera, Dinictis andHoplophoneus, each of which has a distinct caninemorphology. Dinictis is scimitar-toothed withmoderately elongate canines that are broader fromfront to back than those of Hoplophoneus.Hoplophoneus is a dirk-toothed form with veryelongate, narrow canines. These two saber-likecanine forms appear again in felids in the laterCenozoic alongside species with conical canines(Martin, 1980). Thus, the Oligocene sees the first280

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSappearance of ambush predators with retractileclaws and all three canine tooth types: scimitar,dirk, and conical.The nimravids persisted in North America untilabout 23 Ma, filling the ecological roles held nowby bobcat- to jaguar-sized felids, while thehyaenodontids filled more canid-like roles.Between 23 Ma and 17 Ma, specialized cat-likepredators were absent from North America forabout 5 million years until felids arrived from theOld World (Fig. 7). Notably, nimravids eventuallyreappeared in North America in the late Miocene(circa 11 Ma) in the form of the most bizarre andlast of all nimravids, Barbourofelis. Barbourofelis(Fig. 8.3) had enormously elongate dirk-shapedupper canines and highly reduced cheek teeth. Thecoronoid process on the lower jaw was very smallin order to allow the jaws to open wide enough toprovide space between upper and lower canine tips.Late Oligocene to mid-Miocene (approximately28–15 Ma): Canids, Amphicyonids, HemicyonineUrsids.—In the latter half of the Oligocene, theformerly dominant nimravids and creodontsdeclined in diversity for reasons that are unclear.They were replaced by species that evolved in NorthAmerica (canids, amphicyonids) as well as by newimmigrant taxa (ursids, amphicyonids, mustelids)from the Old World. The elimination of cat-likenimravids around 23 Ma seems to have promotedthe evolution of hypercarnivory in several taxa.Among the canids, short-snouted forms such asEnhydrocyon appeared before the nimravidsdisappeared but were smaller in body size than thesabertooths. During the “cat” gap in the record, alarge, leopard-sized mustelid (Megalictis), as wellas several wolf-like ursids (e.g., Phoberocyon) andamphicyonids (e.g., Daphoenodon), filled thehypercarnivorous niche (Van Valkenburgh, 1991).With the exception of the large mustelid, all thesehypercarnivores appear to have been more cursorialthan either Hyaenodon or any of the Oligocenecanids (Hunt, 1998a, 1998b). Although several ofthese species converged on the cat pattern in havingshorter snouts and reduced postcarnassial molars, asabertooth ecomorph did not evolve among them.One additional group outside the orderCarnivora should be mentioned as possiblepredators and scavengers of this period: theentelodonts, non-ruminant artiodactyls of large size(150–170 kg) with formidable canines, incisors andcrushing premolars (Joeckel, 1990). These large,somewhat boar-like ungulates exhibit heavily wornteeth suggestive of bone-cracking behavior, andthey may have played a somewhat hyena-like rolein the late Oligocene–early Miocene.Mid–Late Miocene (approximately 15–5 Ma):Borophagine Canids and Felids.—Cat-likecarnivorans reappeared in the New World about 18Ma with the immigration of the Old World felidPseudaeleurus (Martin, 1998b). Pseudaeleurus hadconical canines but was joined by the scimitartoothedand larger felid Nimravides and the dirktoothednimravid Barbourofelis around 11 Ma. Thegeneric diversity of felids reaches a maximum ofsix in North America in the latest Miocene (about6 Ma) at which time there were dirk-toothed (tribeHomotheriini) and scimitar-toothed (tribeSmilodontini) sabercats as well as at least oneconical-toothed genus. Alongside the rise in feliddiversity, there were declines in diversity in all otherlarge hypercarnivorous taxa, including hemicyonineursids, amphicyonids, and borophagine canids (VanValkenburgh, 1999). Of these three, the borophaginecanids declined slowest and persisted into thePliocene. The borophagine canids of the middle andlate Miocene were of two types—large cursorialwolf-like predators (e.g., Epicyon) and less cursorialspecialized hyena-like bone-crackers (e.g.,Osteoborus) (Munthe, 1998). Both types are likelyto have hunted in packs, taking down prey as largeas or larger than themselves (Van Valkenburgh etal., in press). The bone-cracking borophagines arevery similar to living hyaenids in cranialmorphology, with dome-like skulls that provideextensive areas for jaw adductor musculature, aswell as short snouts to increase bite force at thecanines (Fig. 9). However, whereas hyenas use theirrobust premolars to crack bones, borophagines usedboth their carnassials and premolars (Werdelin,1989). This activity blunted their carnassials andlikely inhibited their ability to use them as manyspecies do today, to slice through tough skin toaccess muscle and viscera (Van Valkenburgh, 1996).281

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002The last of the Borophaginae, Borophagus, wasextinct by the start of the Pleistocene (1.8 Ma), anda bone-cracking ecomorph did not reappear in NorthAmerica, despite the fact that a hyaenid didimmigrate from the Old World about 3.5 Ma. Thisspecies, Chasmoporthetes ossifraga, did not havethe enlarged bone-cracking premolars of extanthyenas. Instead, it was more wolf-like in itsdentition, except that unlike wolves and like hyenas,Chasmoporthetes had no postcarnassial molars(Berta, 1981). This hypercarnivore is relatively rarein the North American record, vanishing about750,000 years ago.Pliocene to Recent (approximately 5 Ma topresent): Felids and Canine Canids.—This finalturnover event within the carnivore guild isassociated with a major extinction event amongterrestrial mammals that occurred near the Miocene-Pliocene boundary (8–5 Ma). In both the Old andNew Worlds, more than 60 percent of the generawent extinct (Webb, 1983, 1984; Savage andRussell, 1983) at a time when there was significantclimatic shift towards increased aridification.Associated with this was a decline in mesicwoodlands and expansion of seasonally aridgrasslands (Janis, 1993). In North America, thelosers among the predators were the wolf-likehypercarnivorous borophagine canids. In Eurasia,this was paralleled by a significant decline in thediversity of wolf-like hyaenids as well as sabertoothcats. In both the Old and New Worlds, the bonecrackingcanids and hyaenids were less affected bythe extinction event, suggesting that their ability tomore fully exploit carcasses was advantageous.In North America, wolf-like borophagine canidswere replaced by members of the extant subfamilyCaninae. Canines dispersed from North America toAsia about 7–8 Ma and there replaced the wolf-likehyaenids of the Old World. In both regions, thecanines coexisted with sabertooth and conicaltoothedfelids. Over the past 5 million years,sabertooth cats decreased in diversity as conicaltoothedbig cats increased. Although sabertoothnimravids and felids were never very diverse—mostpaleocommunities included two species—they werepresent in both the Old and New Worlds throughoutmost of the past 40 million years, and it is not clearwhy this widespread ecomorph declined toextinction in the Pleistocene. They disappeared firstfrom Africa (circa 1 Ma), then Eurasia (circa 0.5Ma), and finally North America (circa 0.01 Ma),which hints that humans may have played somerole in their demise, but the mechanism is as yetunclear. Because we don’t understand fully thefunctional significance of having elongate flattenedcanines as opposed to shorter, stouter canines, it isdifficult to determine what factors, such as extinctionof a particular size or type of prey, might have had anegative impact on sabertooths. It does seemprobable that sabertooth cats, almost all of whichFIGURE 9—Lateral view of the skulls of twoCenozoic bone-crackers drawn to the same size.1, the extant Spotted Hyena, Crocuta crocuta, and2, the Pliocene borophagine canid, Borophagussecundus (drawn from Matthew and Stirton, 1930).282

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSwere leopard-sized and larger, killed relatively largeprey, and their knife-like canines may have made iteasier for a solitary individual to succeed.The last five million years witnessed theevolution of a number of highly cursorialcarnivorans among a variety of families. Forexample, most of the Caninae are characterized byslender, relatively inflexible limbs, compact feet,and the ability to run tirelessly. Within the felids,long-legged genera arose in both Eurasia (thecheetah Acinonyx) and North America (the cheetahlikepuma Miracinonyx) (Van Valkenburgh et al.,1990). Running hyenas, such as Chasmoporthetes,were present in both the Old and New Worlds, andthere was even a long-legged running bear,Arctodus, in North America. This burst of largecursorial predators in the Plio-Pleistocene is notassociated with a similar trend among theirungulate prey. Instead, the ungulates evolved theircursorial adaptations some 20 million years earlier(Janis and Wilhelm, 1993). Consequently, it seemslikely that the predators were responding to a globalchange in vegetation structure that favored longdistancepursuit over ambush, such as a dramaticreduction in vegetative cover. In any case, the Plio-Pleistocene guilds of large carnivorans display agreater diversity of locomotor types than anyprevious guild.MAMMALIAN PREDATORS:LONG-TERM TRENDSThe long-term trends in morphological changeobserved among the Cenozoic synapsids are notnearly as profound or dramatic as those of thePermo-Triassic. As noted above, the evolutionaryhistory of these ancient synapsids documents thetransition from a reptilian type of craniodental andjaw muscle architecture to that of mammals. Nosuch major advances were made by Cenozoicsynapsids. Instead, the history is more one of earlydiversification into a variety of feedingmorphologies followed by repeated iterations ofthese types by different groups over time. Forexample, among Eocene creodonts, there werespecies similar to cats (both sabertooth and conicaltoothed),bone-cracking hyaenids, wolves, andcoyotes. These fundamental carnivorous feedingtypes evolved multiple times over the last 55million years. The absence of secular trends in thedentition and feeding mechanics of the Cenozoicpredators suggests that the synapsid craniodentalarchitecture that evolved in the Mesozoic was verysuccessful and required little or no tinkering toproduce capable meat-eaters. In addition, the lackof a trend reflects the fact that the material propertiesof the food (prey) probably changed little over theCenozoic; skin, muscle, and bone are assumed tohave been much the same in the Paleocene as thepresent. By contrast, Cenozoic herbivorousmammals do exhibit long-term trends in dentalmorphology that reflect an environmental shifttoward cooler, more arid habitats and associatedtougher vegetation (e.g., grass) (cf. Janis, 1993).However, the Cenozoic cooling trend did havean impact on predators. It produced more openhabitats and favored the evolution of cursorialityin both predator and prey. Thus, over the course ofthe past 55 million years, there are repeated trendstowards longer limbs and more compact joints incarnivores. However, it is not a steady progression;instead, it occurred in steps, with the most obviousanatomical advances made in the early Miocene,and then in the Plio-Pleistocene (Janis andWilhelm, 1993). Whether these anatomical changesactually corresponded to improvements in runningability is not known. It is very difficult to ascertaineither top speed or endurance ability frommammalian skeletons, in part because species withvery different builds, such as spotted hyenas andwolves, display similar locomotor capabilities(Bakker, 1983; Janis and Wilhelm, 1993).Nevertheless, in general, extant predators are moregracile in form than their Eocene and Oligocenecounterparts and this probably reflects a shifttoward speed over brawn.PATTERNS IN THEEVOLUTIONARY HISTORY OFSYNAPSID PREDATORSDespite the striking differences in anatomybetween Permo-Triassic and Cenozoic synapsid283

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002predators, there are a number of parallels. Bothradiations produced saber-toothed forms multipletimes—gorgonopsids and therocephalians in thePermo-Triassic, and creodonts, nimravids, felids,and even marsupials (Thylacosmilus (Fig. 8.4) ofSouth America) in the Cenozoic. Without a doubt,saber-like canines are advantageous, but theirabsence today frustrates our ability to explain theirobvious success. As noted above, it might enhancethe ability to take relatively large, thick-skinnedprey. Both radiations also produced short-snouted,powerful, biting predators (e.g., the ancientMoschorhinus and Arctognathus), although thisform is much more common among Cenozoicpredators, where it appears across a wide array offamilies, including oxyaenids, felids, nimravids,canids, ursids, and mustelids. The lower diversityof short-snouted species in the Permo-Triassic isunderstandable given that most species relied on aKI jaw system for killing and feeding.Both radiations of predators show a pattern ofdynasty replacement. That is, one or a few cladesevolve large size and seem to dominate thecarnivore guild for several million years, but thendecline and are replaced by new taxa. In theCenozoic, most of the dynasties appear to be madeup of a more diverse collection of taxa than thoseof the Permo-Triassic. That is, the Cenozoicdynasties include two to three families whereas thePermo-Triassic dynasties include a single suborderor infraorder (e.g., gorgonopsids). In some cases,such as the gorgonopsids (Fig. 4), a fairly diversearray of ecomorphs is included within the group,rivaling that seen in the Cenozoic. However, inmost cases there does seem to have been lessecological diversity in feeding types within Permo-Triassic predator guilds, based on their similaritiesin tooth form and jaw mechanics.In both the Permo-Triassic and Cenozoic,dynasty replacement is associated with considerableevolutionary convergence, as declining clades seemto leave ecological vacuums that are subsequentlyfilled by new taxa. This is especially apparent in theCenozoic record. For example, during the NorthAmerican “cat gap” (23–17 Ma), several families(e.g., canids, ursids, mustelids) produced relativelyshort-snouted species with large canines and reducedpost-carnassial teeth. A similar trend towardhypercarnivory is apparent in South AmericanPleistocene canids and also occurred underconditions of very low felid diversity (VanValkenburgh, 1991). In this instance, the nearabsence of cat-like species was a result of a priorextinction of the sabertooth marsupialThylacosmilus, and an apparently slowerimmigration of felids than canids into SouthAmerica. Convergence among Permo-Triassicsynapsids is less common, but is evident betweentherocephalian sabertooths and the gorgonopsidsthat replace them, as well as between derivedgorgonopsids, which went extinct at the end of thePermian, and moschorhinid therocephalians,which did not. Another example is that of theconvex dorsal profile that is seen in thesphenacodont pelycosaur-grade synapsidDimetrodon and some gorgonopsids such asArctops. Functional convergence may also underliethe broadly similar skull proportions of thesphenacodont Secodontosaurus and somedolichorostral therocephalians such as Lycideops.Within both ancient and Cenozoic predatorclades, there is frequently a tendency towardincreasing body size and hypercarnivory over time.For example, therocephalians, gorgonopsids, anddinocephalians all tend to get larger in body size,enlarge their canine teeth, and strengthen their jawsand snouts over their history. Among Cenozoicmammals, the tendency toward large size is usuallyaccompanied by a reduction in snout length and postcarnassialmolars. This trend is apparent among allthree subfamilies of canids (Van Valkenburgh et al.,in press), oxyaenids (Gunnell, 1998), hyaenodontids(Mellet, 1977), amphicyonids (Viranta, 1996; Hunt,1998b), and hyaenids (Werdelin and Solounias,1996). The prevalence of this pattern suggests thatit is driven by aspects of the large carnivore niche.Levels of interspecific competition for food arerelatively high within guilds of extant carnivorousmammals (Eaton, 1979; Palomares and Caro, 1999;Van Valkenburgh, 1985, 2000). Because largerindividuals tend to dominate inter- and intraspecificconflicts, such as occur over carcasses,284

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSselection favors the evolution of larger body size.However, large size brings with it a cost—the needto feed on prey as large as or larger than oneself.As shown by Carbone et al. (1999), almost all livingcarnivoran species larger than about 21 kg take preyas large as or larger than themselves, whereassmaller species feed mostly on prey that is 45% orless of their body weight. Using an energetic model,they argued that because of limitations on intakerate and foraging time, it becomes increasinglydifficult to subsist on small prey items as predatorbody mass grows. Consequently, the evolution oflarge body size (>20 kg) in a carnivore must beassociated with craniodental adaptations for killingand feeding on large prey. These include shortenedsnouts to enhance bite force at the canines, as wellas robust canines and large carnassials. Nonmammaliansynapsids seem to have been subjectto these same evolutionary forces, and modifiedtheir skulls to improve the KI bite system as wellas to add the SP bite system.Van Valkenburgh (1999) argued that thistendency toward specialization within carnivoreclades might explain their inevitable declines.Large, hypercarnivorous species may do well inthe short run, but their specialization results in aloss of evolutionary versatility that inhibits theirability to adapt rapidly to changing environmentalconditions. Moreover, families or subfamilies oflarge-sized carnivores tend to be less species-richthan those of smaller carnivores, and consequentlymore vulnerable to extinction. Thus, as clades ofcarnivores evolve toward large size andhypercarnivory, they dwindle in species richness—leading to their decline and ultimate extinction.Thus, the cycle of rising and falling dynasties thatso typifies the fossil record of carnivores seemsbetter explained by the dynamics of competitionamong predators themselves than by changes inprey diversity or adaptations.REFERENCESANDERSON, J. M., AND A. R. I. CRUICKSHANK. 1978. The biostratigraphy of the Permian and the Triassic, Part 5: Areview of the classification and distribution of Permo-Triassic tetrapods. Palaeontologia Africana, 21:15–44.BAKKER, R. T. 1983. The deer flees, the wolf pursues: Incongruencies in predator-prey coevolution, p. 350–382.In D. J. Futuyma and M. Slatkin (eds.), Coevolution. Sinauer Press, Sunderland, Massachusetts.BATTAIL, B., AND M. V. SURKOV. 2000. Mammal-like reptiles from Russia, p. 86–119. In M. J. Benton, M. A.Shishkin, D. M. Unwin, and E. N. Kurochkin (eds.), The Age of Dinosaurs in Russia and Mongolia. CambridgeUniversity Press, Cambridge.BENTON, M. J. 1993. The Fossil Record II. Chapman and Hall, London, 841p.BENTON, M. J. 1998. Vertebrate Palaeontology, 2nd edition. Chapman and Hall, London.BERTA, A. 1981. The Plio-Pleistocene hyaena Chasmoporthetes ossifraga from Florida. Journal of VertebratePaleontology, 1:341–356.BIKNEVICIUS, A. R., AND B.VAN VALKENBURGH. 1996. Design for killing: Craniodental adaptations of predators, p.393–428. In J. L. Gittleman (ed.), Carnivore Behavior, Ecology and Evolution, Vol. 2. Cornell UniversityPress, Ithaca, NY.BIKNEVICIUS, A. R., B. VAN VALKENBURGH, AND J. WALKER. 1996. Incisor size and shape: implications for feedingbehaviors in saber-toothed “cats.” Journal of Vertebrate Paleontology, 16:510–521.BOONSTRA, L. D. 1954. The cranial structure of the titanosuchian Anteosaurus. Annals of the South African Museum,42:108–148.BRAMBLE, D. M. 1978. Origin of the mammalian feeding complex: models and mechanisms. Paleobiology, 4:271–301.BROOM, R. 1932. The mammal-like reptiles of South Africa and the origin of mammals. H. F. and G. Witherby,London, 376 p.BRYANT, H. N. 1991. Phylogenetic relationships and systematics of the Nimravidae (Carnivora). Journal ofMammalogy, 72:56–78.285

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002CARBONE, C., G. M. MACE, S. C. ROBERTS, AND D. W. MACDONALD. 1999. Energetic constraints on the diet ofterrestrial carnivores. Nature, 402:286–288.CARROLL, R. L. 1988. Vertebrate Paleontology and Evolution. W. H. Freeman and Co., New York, 698 p.CHUDINOV, P. K. 1983. Early therapsids. Transactions of the Palaeontological Institute of the Academy of ScienceUSSR, 202:1–230 (in Russian).CLEUREN, J AND F. DE VREE. 2000. Feeding in crocodilians, p 337–359. In K. Schwenk (ed.), Feeding: Form,Function and Evolution in Tetrapod Vertebrates. Academic Press, London.COVEY, D. S. G., AND W. S. GREAVES. 1994. Jaw dimensions and torsion resistance during canine biting in theCarnivora. Canadian Journal of Zoology, 72:1055–1060.CROMPTON, A. W. 1963. On the lower jaw of Diarthrognathus, and the origin of the mammalian lower jaw.Proceedings of the Zoological Society of London, 140:697–753.DAWSON, M. R., R. K. STUCKY, L. KRISHTALKA, AND C. C. BLACK. 1986. Machaeroides simpsoni, new species, oldest knownsabertooth creodont (Mammalia) of the Lost Cabin Eocene, p. 183–196. In K. M. Flanagan and J. A. Lillegraven(eds.), Vertebrates, Phylogeny, and Philosophy. Contributions to Geology, University of Wyoming, Special Paper 3.EATON, R. L. 1979. Interference competition among carnivores: a model for the evolution of social behavior.Carnivore, 2:9–16.EMERSON, S. B., AND L. B. RADINSKY. 1980. Functional analysis of sabertooth cranial morphology. Paleobiology,6:295–312.EWER, R. F. 1973. The Carnivores. Cornell University Press, Ithaca, NY.FLYNN, J. J. 1998. Early Cenozoic Carnivora (“Miacoidea”), p. 110–23. In C. M. Janis, K. M. Scott, and L. L.Jacobs (eds.), Evolution of Tertiary Mammals of North America. Cambridge University Press, New York.GITTLEMAN, J. L. (ed). 1989. Carnivore Behavior, Ecology and Evolution, Vol. 1. Cornell University Press, Ithaca,NY, 620 p.GITTLEMAN, J. L. (ed). 1996. Carnivore Behavior, Ecology and Evolution, Vol. 2. Cornell University Press, Ithaca, NY, 644 p.GUNNELL, G. F. 1998. Creodonta, p 91–109. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolution ofTertiary Mammals of North America. Cambridge University Press, New York.HOPSON J. A. 1991. Systematics of the nonmammalian Synapsida and implications for patterns of evolution insynapsids, p 635–693. In H.-P. Schultze and L. Trueb (eds.), Origins of the Higher Groups of Tetrapods-Controversy and Consensus. Cornell University Press, Ithaca, NY.HOPSON, J. A., AND H. R. BARGHUSEN. 1986. An analysis of synapsid relationships, p. 83–106. In N. Hotton III, P.D. MacLean, J. J. Roth, and E. C. Roth (eds.), The Ecology and Biology of Mammal-like Reptiles. SmithsonianInstitution Press, Washington, D.C.HUNT, R. M. 1987. Evolution of the aeleuroid Carnivora: significance of auditory structure in the nimravid catDinictis. American Museum Novitates, 2886:1–74.HUNT, R. M. 1996. Biogeography of the order Carnivora, p. 485–541. In J. L. Gittleman (ed.), Carnivore Behavior,Ecology and Evolution. Cornell University Press, Ithaca, NY.HUNT, R. M. 1998a. Ursidae, p. 174–95. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolution of TertiaryMammals of North America. Cambridge University Press, New York.HUNT, R. M. 1998b. Amphicyonidae, p. 196–227. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolutionof Tertiary Mammals of North America. Cambridge University Press, New York.JANIS, C. M. 1993. Tertiary mammal evolution in the context of changing climates, vegetation, and tectonicevents. Annual Review of Earth and Planetary Sciences, 24:467–500.JANIS, C. M., AND P. B. WILHELM. 1993. Were there mammalian pursuit predators in the Tertiary? Dances withwolf avatars. Journal of Mammalian Evolution, 1:103–25.JENKINS, I. 1998. Cranial form and function in some Permian carnivorous synapsid (mammal-like) reptiles.Unpublished Ph.D. thesis, University of Cambridge, England, 275 p.JENKINS, I. 2001a. Palaeozoic carnivorous reptiles. Geology Today, 17:36–39.JENKINS, I. 2001b. Finite element modelling approaches to morphology in vertebrate palaeobiology. Journal ofMorphology, 248:246–247.JENKINS, I. In Press, a. Cranial mechanisms and carnivory in therocephalians, the first ‘sabre-toothed’ tetrapods:a finite element analysis. Proceedings of the Royal Society London, B.286

VAN VALKENBURGH AND JENKINS—HISTORY OF SYNAPSID PREDATORSJENKINS, I. In Press, b. Modern anatomical data in the analysis of extinct tetrapods: cranial mechanisms in ancientcarnivorous mammalian relatives. Journal of Clinical Anatomy.JENKINS, I., J. J. THOMASON, AND D. B. NORMAN. In press. Primates and engineering principles: applications tocraniodental mechanisms in ancient terrestrial predators. In Senckenbergiana Lethaea. Special Volume,Senckenberg Museum Press.JOECKEL, R. M. 1990. A functional interpretation of the masticatory system and paleoecology of entelodonts.Paleobiology, 16:459–482KEMP, T. S. 1969. On the functional morphology of the gorgonopsid skull. Philosophical Transactions of theRoyal Society of London B, 256:1–83.KEMP, T. S. 1982. Mammal-like reptiles and the origin of mammals. Academic Press, London, 363 p.KEMP, T. S. 1988. Interrelationships of the Synapsida, p. 1–23. In M. J. Benton (ed.), The Phylogeny andClassification of the Tetrapods, Volume 2: Mammals. Systematics Association Special Volume No. 35.LAURIN, M., AND R. R. REISZ. 1995. A re-evaluation of early amniote phylogeny. Zoological Journal of the LinneanSociety, 113:165–223.LEE, M. S. Y. 1997. A taxonomic revision of pareiasaurian reptiles: implications for Permian terrestrialpalaeoecology. Modern Geology, 21:231–298.MARTIN, L. D. 1980. Functional morphology and the evolution of cats. Transactions of the Nebraska Academy ofScience, 8:141–154.MARTIN, L. D. 1989. Fossil history of the terrestrial Carnivora, p. 536–568. In J. L. Gittleman (ed.), CarnivoreBehavior, Ecology and Evolution. Cornell University Press, Ithaca, NY.MARTIN, L. D. 1998a. Nimravidae, p.228–235. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolution ofTertiary Mammals of North America. Cambridge University Press, New York.MARTIN, L. D. 1998b. Felidae, p.236–242. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolution ofTertiary Mammals of North America. Cambridge University Press, New York.MELLETT, J. S. 1969. Paleobiology of North American Hyaenodon (Mammalia: Creodonta). Contributions toVertebrate Evolution, 1:1–134.MODESTO, S. P. 1995. The skull of the herbivorous synapsid Edaphosaurus boanerges from the Lower Permian ofTexas. Palaeontology, 3:213–239.MUNTHE, K. 1998. Canidae, p. 124–143. In C. M. Janis, K. M. Scott, and L. L. Jacobs (eds.), Evolution of TertiaryMammals of North America. Cambridge University Press, New York.OLSON, E. C. 1961. Jaw mechanisms: rhipidistians, amphibians, reptiles. American Zoologist, 1:205–215.ORLOV, YU. A. 1958. Carnivorous dinocephalians from the fauna of Isheev (Titanoscuhia). TrudyPalaeontogicheskogo Instituta Academika, USSR, 72:1–114 (in Russian).PALOMARES, F., AND T. M. CARO. 1999. Interspecific killing among mammalian carnivores. The American Naturalist,153:492–508.PARRINGTON, F. R. 1955. On the cranial anatomy of some gorgonopsids and the synapsid middle ear. Proceedingsof the Zoological Society of London, 125:1–40.REISZ, R. 1986. Pelycosauria, p.1–102. In P. Wellnhofer (ed.), Encyclopaedia of Paleoherpetology, Part 17A.Gustav Fischer Verlag, Stuttgart.REISZ, R. R., D. S. BERMAN, AND D. SCOTT. 1992. The cranial anatomy and relationships of Secodontosaurus, anunusual mammal-like reptile (Pelycosauria: Sphenacodontidae) from the early Permian of Texas. ZoologicalJournal of the Linnean Society, 104:127–184.ROMER, A. S. 1966. Vertebrate Paleontology. University of Chicago Press, Chicago, 468 p.ROMER, A. S. 1969a. The Chanares (Argentina)Triassic reptile fauna II: A new chiniqoudont cynodont Probelesodonlewis—cynodont ancestry. Breviora, 333:1–24.ROMER, A. S. 1969b. The Brazilian cynodont reptiles Belesodon and Chiniquodon. Breviora, 332:1–16.ROMER, A. S., AND L. I. PRICE. 1941. Review of the Pelycosauria. Special Papers of the Geological Society ofAmerica, 28:1–538.RUBIDGE, B. S. (ed). 1995. Biostratigraphy of the Beaufort Group (Karoo Supergroup). Geological Survey ofSouth Africa. Biostratigraphic Series, 1.287

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002RUSSELL, A. P., AND J. J. THOMASON. 1993. Mechanical analysis of the mammalian head skeleton, p. 345–383. In J.Hanken and B. K. Hall (eds.), The Skull. University of Chicago Press, Chicago.SAVAGE, D. E., AND D. E. RUSSELL. 1983. Mammalian Paleofaunas of the World. Addison-Wesley, Reading, Massachusetts.SAVAGE, R. J. G. 1977. Evolution in carnivorous mammals. Palaeontology, 20:237–271.SENNIKOV, A. G. 1996. Evolution of the Permian and Triassic tetrapod communities of Eastern Europe.Palaeogeography, Palaeoclimatology, Palaeoecology, 120:331–351.SIGOGNEAU, D. 1970. Revision Systematique des Gorgonopsiens Sud-Africains. Cahiers Paleontologie, 417 p.SIGOGNEAU-RUSSELL, D. 1989. Theriodontia I, p. 127. In O. Kuhn (ed.), Encyclopaedia of Palaeoherpetology, Part17B/1. Gustav Fisher Verlag, Stuttgart.SIDOR, C. A., AND J. A. HOPSON. 1997. Ghost lineages and “mammalness”: assessing the temporal pattern ofcharacter acquisition in the Synapsida. Paleobiology, 24:48–68.SIDOR, C. A. 2001. Simplification as a trend in synapsid cranial evolution. Evolution, 55:25–58.TATARINOV, L. P. 1974. Theriodonts of the USSR. Trudy Palaeontogicheskogo Instituta Academika, USSR, 143:5–250 (in Russian).VAN VALKENBURGH, B. 1985. Locomotor diversity within past and present guilds of large predatory mammals.Paleobiology, 11:406–428.VAN VALKENBURGH, B. 1991. Iterative evolution of hypercarnivory in canids (Mammalia, Carnivora): evolutionaryinteractions among sympatric carnivores. Paleobiology, 17:340–362.VAN VALKENBURGH, B. 1996. Feeding behavior in free-ranging, large African carnivores. Journal of Mammalogy,77:240–254.VAN VALKENBURGH, B. 1999. Major patterns in the history of carnivorous mammals. Annual Review of Earth andPlanetary Sciences, 27:463–493.VAN VALKENBURGH, B. 2001. The dog-eat-dog world of carnivores: A review of past and present carnivorecommunity dynamics, p. 101–121. In C. Stanford and H. T. Bunn (eds.), Meat-Eating and Human Evolution.Oxford University Press, Oxford.VAN VALKENBURGH, B., F. GRADY, AND B. KURTÉN. 1990. The Plio-Pleistocene cheetah-like cat Miracinonyxinexpectatus of North America. Journal of Vertebrate Paleontology, 10:434–454.VAN VALKENBURGH B., T. SACCO, AND X. WANG. In press. Pack hunting in Miocene Borophagine dogs: evidencefrom craniodental morphology and body size. In L. J. Flynn (ed.), Vertebrate Fossils and Their Context: ATribute to the Career of Richard H. Tedford. Bulletin of the American Museum of Natural History.VIRANTA, S. 1996. European Miocene Amphicyonidae—taxonomy, systematics, and ecology. Acta ZoologicaFennica, 204:1–61.WEBB, S. D. 1983. The rise and fall of the late Miocene ungulate fauna in North America, p. 267–306. In M. D.Nitecki (ed.), Coevolution. University of Chicago Press, Chicago.WEBB, S. D. 1984. Ten million years of mammal extinctions in North America, p. 189–210. In P. S. Martin and R.G. Klein (eds.), Quaternary Extinctions: A Prehistoric Revolution. University of Arizona Press, Tucson.WERDELIN, L. 1989. Constraint and adaptation in the bone-cracking canid Osteoborus (Mammalia: Canidae).Paleobiology, 15:387–401.WERDELIN, L. 1996. Carnivoran ecomorphology: a phylogenetic perspective, p. 582–624. In J. L. Gittleman (ed.),Carnivore Behavior, Ecology and Evolution. Cornell University Press, Ithaca, NY.WERDELIN, L., AND N. SOLOUNIAS. 1991. The Hyaenidae: taxonomy, systematics and evolution. Fossils andStrata, 30:1–104.WERDELIN, L., AND N. SOLOUNIAS. 1996. The evolutionary history of hyaenas in Europe and Western Asia duringthe Miocene, p. 290–306. In R. L. Bernor, V. Fahlbusch, and H-W. Mittmann (eds.), The Evolution of WesternEurasian Neogene Mammal Faunas. Columbia University Press, New York.288

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONORIGINS AND EARLY EVOLUTION OF PREDATIONSTEFAN BENGTSONDepartment of Palaeozoology, Swedish Museum of Natural History, Box 50007, SE-104 05 Stockholm, SwedenABSTRACT—Predation, in the broad sense of an organism killing another organism for nutritional purposes, isprobably as old as life itself and has originated many times during the history of life. Although little of the beginningsis caught in the fossil record, observations in the rock record and theoretical considerations suggest that predationplayed a crucial role in some of the major transitions in evolution. The origin of eukaryotic cells, poorly constrainedto about 2.7 Ga by geochemical evidence, was most likely the ultimate result of predation among prokaryotes.Multicellularity (or syncytiality), as a means of acquiring larger size, is visible in the fossil record soon after 2 Gaand is likely to have been mainly a response to selective pressure from predation among protists. The appearance ofmobile predators on bacteria and protists may date back as far as 2 Ga or it may be not much older than theCambrian explosion, or about 600 Ma. The combined indications from the decline of stromatolites and thediversification of acritarchs, however, suggest that such predation may have begun around 1 Ga. The Cambrianexplosion, culminating around 550 Ma, represents the transition from simple, mostly microbial, ecosystems to oneswith complex food webs and second- and higher-order consumers. Macrophagous predators were involved from thebeginning, but it is not clear whether they originated in the plankton or in the benthos. Although predation was adecisive selective force in the Cambrian explosion, it was a shaper rather than a trigger of this evolutionary event.THE EARLY WORM CATCHESTHE—WHAT?THE ORIGIN of predation is veiled in as muchuncertainty as is the origin of life. Perhaps evenmore: Life, as we know it today, has a commonorigin, but predation—in the broad sense of anorganism killing another organism for nutritionalpurposes—has originated many times at differentlevels of organismal interactions. We can assume,however, that whenever predatory lifestylesevolved they became a strong evolutionary force.Predation introduces hazard into complacency,expands food webs, redistributes resources,recombines characters, and stimulates responsesthat cascade into an ever-expanding and neverendingseries of evolutionary thrusts and ripostes.Predators and prey may enter into symbioticrelationships and emerge as new organisms.Current theories on a number of major transitionsin evolution (non-cellular to cellular; prokaryoteto eukaryote; non-sex to sex; small to large;unicellular to multicellular; multicellular to tissuegrade;sessile to motile; soft to hard; smooth tospiny) tend to focus on the introduction of predationas a decisive factor.The broad definition of predators alluded toabove is in common use (e.g., Levinton, 1982;Woodin, 1983; Menge, 1995; Abrams, 2000). Itinvolves much more than fanged beasts that pouncewith a roar upon the hapless leaf-muncher. Itincludes organisms eating those that are smaller,of the same size, or larger. It includes grazing,whether the organisms being grazed are grass,plankton, or microbes in mats. The central aspectof the definition of predation is that it kills thevictim. Leaf munching is not predation—notbecause the leaves are plants, but because browsingthem usually does not kill the plant. Parasitism isnot predation, for the same reason. Scavenging alsois not predation, for the “prey” is already dead.Obviously, there are fuzzy lines between predationand other kinds of interactions—browsing andparasitism may kill the victim in the end, and thedistinction between grazing (predation) andbrowsing (not predation) is blurred by the diffuseboundaries between individuals and clonalcolonies. Scavenging and predation are often two289

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002sides of the same behavior, and detritus feedersare bound to engulf countless living microbes. Mostorganisms are not confined to a single mode of life,so the same organism may be predator, scavenger,parasite, etc.—and, of course, prey. Phenomena innature tend to have fuzzy edges, and terminologyshould not lead us to forget that.Neither predator nor prey has to be an animal,so the definition allows for a discussion of theevolutionary mechanisms that might have beenpresent long before crown-group animals werearound. The definition encompasses so much oforganism–organism interactions, however, thatconstraints are necessary to keep the chapter withinbounds. The constraints will mainly be practical—I will deal with processes that either have left directevidence among fossils or at least have the potentialto have influenced the fossil record recounting theearly history of life on Earth, up till about 500million years ago. Also, emphasis will shift fromprokaryote and protist predators during the earlypart of the interval covered, to multicellularsuspension/filter feeders and grazers on plankticand benthic microbes, and finally to macrophagouspredators—animals eating animals.Because successful predation by definitionleads to the death of the prey, its selectiveimportance is considerable. Interactions betweenorganisms are generally regarded as a major factorin evolution (though see Gould, 1985, 2002), andsuch interactions that lead to the failure ofindividuals to reproduce should have the strongesteffect of all. The evolutionary effects are strongerin prey than in predator taxa (Dawkins and Krebs,1979; Vermeij, 1987; Abrams, 2000), because ofthe unequal nature of the interactions—at eachencounter the prey risks it life, the predator onlyits meal (the “life–dinner principle” of Dawkinsand Krebs, 1979). Experimental work in modernecosystems confirms that the introduction ofpredators may cause a rapid evolution of variousdefensive characters in the prey species(Thompson, 1998; Agrawal et al., 1999). When thepredators respond by evolving more efficientmeans of predation, the feedback loop sets up thefamiliar “arms-race” (Dawkins and Krebs, 1979)or escalation (Vermeij, 1987, 1994, 2002) scenario.Nonetheless, our ability to identify theevolutionary effects of predation in the fossil recordis limited, because evolution is the sum of alleffects, and controlled experiments are generallynot possible. We have very incomplete informationabout the nature of the encounters between predatorand prey, and in most cases we know the identityof only one of the participants. We know equallylittle about population structure, competition levels,environmental stress, etc. The fossil record mayyield trends through time, but interpreting them issimilarly difficult. For example, trends towardlarger size in both predators and prey may beinterpreted as causally related but may equally wellbe responses to the same external factors.For the vast stretches of pre-Phanerozoic timecovered in this chapter, matters would seem evenworse. Fossils are scarce, they are mostly microbial,and their mode of life cannot in general be deducedfrom their morphology. For most of the time period,direct evidence of predation is lacking. The fossilrecord is generally silent with regard to animals olderthan about 600 million years, and only indirectevidence suggests that animals or animal-likeorganisms and predatorial modes of life existedearlier. Clearly, formulating and testing ecological/evolutionary hypotheses in this setting is difficultor impossible. Nonetheless, theoreticalconsiderations of the role(s) of predation in earlyevolution, set against paleontological data, help usto interpret the sparse early fossil record and toevaluate hypotheses regarding the role of predationas a driving force in the evolution from an almostexclusively microbial biosphere to one characterizedby multicellular organisms and the complex foodwebs of modern ecosystems. Although thecausalities and triggers of this process are far fromunderstood, the evolution of predatory modes of lifeis likely to have played a central role, certainly indetermining the course of evolution in a number oflineages and perhaps also as a major shaping forcein the radiation of multicellular and unicellularorganisms during the Cambrian explosion, one ofthe most significant and certainly the most manifestrestructuring of the biosphere during Earth history.290

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONPREDATION AND DIVERSITYIN THE EARLY BIOSPHEREAn imaginary biosphere without predatorswould be very different from what we areacquainted with. There would be various kinds ofphoto- and chemoautotrophs making use ofavailable energy gradients to reduce carbon forenergy storage and constructional/physiologicalpurposes. There would be organisms scavengingexcess organic matter, but there would be noorganisms directly interrupting the lives of othersby pilfering their tissues.Leaving aside the question of whether such aShangri-La for primary producers anddecomposers is even theoretically possible, theselective pressures would be very different fromthose that affect most organisms today. Survivalrequirements would center around positioningoneself with respect to chemical, temperature, andlight gradients, and the only need to move wouldbe in order to adopt to shifting gradients—forexample, varying light intensities or redoxboundaries. Although competitive interactionswould not be excluded, they would mostly berelated to relative efficiencies of energy conversionsystems. Under such circumstances diversitieswould be low and stable. The most complex benthicecosystems would likely be layers of physiologicallydifferentiated microbes, i.e., microbial mats. In theplankton, diversities would possibly be even lowerbecause of the movement and mixing of watermasses, which reduces spatial heterogeneity.The cropping principle (Stanley, 1973a, 1976b)suggests in its general form that the introductionof predation into a low-diversity ecosystem willcreate a self-propagating feedback system ofdiversification. Stanley specifically discussed theappearance of cell-eating heterotrophy amongplanktic protists, which in his view may have beenthe driving force behind the eventual burgeoningof multicellular organisms and the Cambrianexplosion. The phenomenon of predation-inducedvariability is well established in different kinds ofecosystems (see references in Stanley, 1973a,1976b; as well as Porter, 1977; Kitchell, 1983;Richards et al., 1999), and we may ask the moregeneral question of whether some of the majordiversity changes in the early fossil record werepredator-induced.Theories to that effect abound. At the base ofthe bush of life, the origin of cells has beeninterpreted as a symbiotic or predatory event(Maynard Smith and Szathmáry, 1995; Scudo,1996; Cavalier-Smith, 2001). The origins ofeukaryotic cells, multicellularity, and hard tissuesare commonly interpreted to be primary results ofpredation (see below). These evolutionaryinnovations clearly had a great effect on diversityby introducing new kinds of organisms withunexploited capabilities of diversification.Less dramatic, but perhaps stronger in longtermeffect, are the diversity effects caused by thedynamics of predator–prey interactions atestablished levels of organization. Although thediversity effects of such interactions are commonlydescribed in the ecological literature in terms ofequilibrium models (where the predator–prey ratiois drawn toward a stable value), this may not be agood description of natural systems. Predators maydrive their prey to local extinction (Katz, 1985) ormake them more susceptible to extinctions by otheragents (Schoener et al., 2001). The net effects ondiversity are dependent on a number of factors,such as the existence of refuges, the selectivity andintensity of predation, etc., but as a general rule,selective predation on dominant species increasesdiversity (Kitchell, 1983). Competitive interactionmay also influence diversity, though its effects mayhave been overstated in the past (Gould andCalloway, 1980; Benton, 1983). In the end,diversity may be less dependent on direct effects,such as those of predator–prey interactions, thanon more-or-less complex cascades of indirecteffects of biotic interactions (Menge, 1995).There are of course also environmental (sealevel, temperature, oxygen level, nutrientavailability) and preservational parameters thataffect diversity, and these may or may not beindependently analyzed. Predation itself may biasthe apparent diversity in the fossil record. Preyeaten by predators may be totally destroyed and291

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002thus escape fossilization, although it is not likelythat this will remove the record of the preyed-onspecies altogether. On the other hand, in somecircumstances predation may enhance thepossibilities of fossilization, thereby boosting thediversity record of the prey. Fecal-pellet transportvia planktic predators is today the dominant modeof transfer of plankton to the sea floor, andindigestible tests of the prey are thereby protectedfrom dissolution by the seawater (Honjo andRoman, 1978; Kitchell, 1983).Sampling artifacts may have considerablystronger effects on fossil diversity curves than isgenerally recognized, and it is not unlikely thatmost of the short-term diversity changes reportedfrom the fossil record are in fact a function ofdifferential preservation in the rock record (Raup,1976a, 1976b; Peters and Foote, 2001, 2002).In conclusion, various effects of predation ondiversity may be postulated, but general diversitydata cannot be used to argue levels of predation. Wemay have to be content with “asking what isplausible in theory and what is interesting in themeasurable fossil record of diversity” (Sepkoski,1996). The “interesting” aspects of diversity may inthis case be related not to clades or grades but toconvergent aptations to predator–prey interactions(spinosity, burrowing habit, tube-dwelling,spiculation, sclerotization, etc.) in comparableenvironments. Such data are generally not directlyavailable from the literature, and will have to bespecially compiled to be useful. Later in this chapterI will discuss, however, how some of the availablediversity curves, in particular for stromatolites andprotists, may be of use as proxies for predatorialactivities in the early evolution of life.SIZE INCREASE AS ARESPONSE TO PREDATIONPredators either penetrate their prey or swallowit. In our imaginary Shangri-La, there would belittle need to get big. On the contrary, efficiency inthe exchange of gases and nutrients is a functionof an organism’s surface-to-volume ratio, and sosmaller organisms have the advantage. Anyincrease in size would have to be accompanied byan exponential increase in surface complexity tokeep the surface-to-volume ratio stable.Enter a predator. Now a large surface area maybecome a vulnerability—the more exposed surface,the more there is for the predator to attack. Increasein size may then be a better option, not only becauseit reduces the surface-to-volume ratio thusprotecting against penetration, but also because itmakes the potential prey more difficult to swallow(cf. Guillard and Kilham, 1977). Conversely, it isadvantageous for a swallowing predator to be largerthan its prey (see Hansen et al., 1994), so a positivefeedback loop is created.Also (with the exception of the large landanimals, for which gravitation becomes the majorobstacle), larger organisms can move faster thansmall ones (Bonner, 1965, 1993). Althoughmovement may be an advantage also for anorganism seeking out suitable energy gradients,there is no compelling reason to move quicklyunless someone else does too. Thus, increasingmotility, a corollary of size, may also be selectedfor in predator–prey interactions.Other effects of larger size are division of laborand hence differentiation of tissues and thedevelopment of organs that would have no functionin smaller organisms: respiratory, digestive,circulatory, and muscular structures, for example.These effects may be seen as secondary to theprimary phenomenon of size increase (Bonner,1965, 1993, 1998). They also have a much widersignificance than merely being involved inpredator–prey interactions, so at these higher levelsof organization the connection between predationand size increase becomes weak.Although increase in cell/body size, at least forsimpler organisms, may thus be a more useful proxyfor predation pressure than taxonomic diversity, acaveat is needed also here. The tendency towardlarger body size and complexity in evolution is soprevalent that it has been regarded as a general law(Cope’s Rule). Whereas individual instances of sizeincrease may be due to specific selection pressures,the general phenomenon does not have to beexplained as anything more than an increase in292

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONvariance during the course of evolution: if you startsmall and simple, the only direction to go is towardlarge and complex (Stanley, 1973b; Gould, 1988;Bonner, 1993, 1998). Thus any event ofdiversification is likely to bring with it an increasein variance and, hence, an average size increase.Two major advances in the early evolution oflife, however, are inseparably connected with sizeincrease: the origin of eukaryotes and the origin(s)of multicellularity. Both of these probably are infact direct results of predation.ORIGIN OF EUKARYOTESWhereas prokaryotic organisms (bacteria andarchea) represent almost all of the biochemicaldiversity of the biosphere, eukaryotic cells are thebasis for most of the structural and morphologicaldiversity, most particularly with regard tomulticellular organisms. In current theory,predation was a main factor behind the origin ofeukaryotes. Molecular and structural evidencesuggest that eukaryotes evolved through a seriesof endosymbiotic events in which prokaryotesengulfed or invaded other organisms, eventuallyleading to an amalgamation of several lineages intodaughter organisms representing a higher level ofco-operational complexity (Margulis, 1970, 1981;Cavalier-Smith, 1987a, 1987b; Martin and Müller,1998; Gray, 1999; Lang et al., 1999; Roger, 1999).In particular, mitochondria and chloroplasts,containing as they do their own genome, showstrong evidence of having been derived from freelivingα-proteocteria and cyanobacteria, respectively(Gray and Spencer, 1996). The probable origin ofthese endosymbiotic relationships is predation bymeans of phagocytosis and the survival of some preywithin the predator (McFadden et al., 1994; de Duve,1995; Roger, 1999). Thus the origin of eukaryotesmay be seen as a direct consequence of predatorialinteractions among prokaryotes (e.g., MaynardSmith and Szathmáry, 1995).The same may be true of the origin of eukaryoticsex. The classic interpretation of sex in eukaryotesis that it arose from a single organism as a means ofreshuffling genomes. Maynard Smith and Szathmáry(1995) propose that alternating meiosis andendomitosis in this organism produced a haploiddiploidlife cycle, and that (isogamous) syngamyeventually replaced endomitosis for the productionof the diploid phase (because of the doubleadvantages of repressing deleterious mutations andallowing for recombination); anisogamy was a laterdevelopment. However, as commonlyacknowledged (Williams, 1975; Maynard Smith,1978; Maynard Smith and Szathmáry, 1995), theevolution of sex is far from well understood. Aradical alternative to the classical model (Walther,2000) is original anisogamy through the fusion oftwo prokaryotic organisms in a predatorial/symbioticevent. This would mean that eukaryotic sex, likeeukaryotes themselves, is the result of predation.Does the fossil record have anything to sayabout this? Although the record is fundamentallyinadequate to illuminate processes at the level oforganelles, some important information about earlyeukaryote evolution is in fact available. Cavalier-Smith (1987a) proposed that the original bacterialsymbiotic host, in order to be capable of engulfingother organisms, must have lost its polysaccharidecell wall and compensated this by evolving aninternal cytoskeleton and sterol cell membranes.Sterols (a group of steroid lipids) are an importantand characteristic component of eukaryote cellmembranes. The degradation products ofeukaryotic sterols, C 27–C 29steranes, have beendiscovered in 2.7 billion-year-old organic mattertogether with 2-methylhopanes, a knowndegradation product of cyanobacterial membranelipids (Brocks et al., 1999; Summons et al., 1999).Thus there is fossil chemical evidence that by thattime at least two of the organismal groups thatparticipated in the symbiotic events leading up toeukaryotic cells were present in the biosphere.With regard to body fossils, the generally largersize of modern eukaryotic cells with respect toprokaryotic cells was used in a pioneering attemptto date eukaryote origins based on the sizedistribution of Precambrian microfossils (Schopfand Oehler, 1976). The earliest fossil nowcommonly attributed to eukaryotes is the 1.85billion-year-old (Hoffman, 1987; Morey and293

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Southwick, 1995; P.F. Hoffman, pers. comm.,2002) Paleoproterozoic Grypania, a coiled,cylindrical organism that may attain half a meterin length and 2 mm in diameter (Han and Runnegar,1992; Runnegar, 1994). Because of its complexityand size, Grypania is commonly interpreted to bea eukaryotic alga. Runnegar (1994) speculated thatit may be a unicellular or coenocytic organismsimilar to some modern dacycladaceans(Acetabularia and relatives). Advanced unicellulareukaryotes, including spinous forms, are presentin the Mesoproterozoic (Samuelsson et al., 1999;Javaux et al., 2001), although higher diversities(several tens of taxa or more) do not appear to havebeen attained until the Neoproterozoic (Vidal andMoczydlowska-Vidal, 1997). The diversificationof unicellular phytoplankton, starting at about 1Ga,is generally considered mainly related to predationtypes and levels (Knoll, 1992; Vidal andMoczydlowska, 1992; Knoll, 1994a, 1994b;Butterfield, 1997, 2001; Vidal and Moczydlowska-Vidal, 1997; Smetacek, 2001). Alternatively,Schopf (Schopf et al., 1973; Schopf, 1999) hasproposed that it reflects the origin of sexualityabout 1.1 billion years ago. Butterfield (2000),however, recently presented convincing evidencefor sexually reproducing multicellular red algaealready at 1.2 Ga, and argued that the origin ofsexuality is linked to that of multicellularity.ORIGIN(S) OFMULTICELLULARITYIn the living biota there are at least 13 lineages,eukaryotic as well as prokaryotic, in whichmulticellularity has been attained independently(Bonner, 1998, 2000). Although the selectivepressures behind multicellularity may be complex,multicellularity as a general phenomenon can beseen as a consequence of size increase (Bonner,1998). As discussed above, size increase in smallunicellular organisms has its primary advantage inpredator–prey interactions. An example is providedby the Myxobacteria, soil-living bacteria the cellsof which aggregate in motile swarms to concentrateenzymes that digest other bacteria (Shimkets,1990). This is a classic predatorial behavior, thoughthis type of aggregating multicellularity (seen alsoin slime molds) is characteristic of terrestrial, notaquatic, organisms (Bonner, 1998, 2000).In the absence of preserved cells,multicellularity in fossil organisms can usually onlybe inferred. Many of the possible multicellularorganisms in the Precambrian fossil record may justas plausibly have been syncytial—consisting of acontinuous protoplasmic mass with numerous nucleibut no cell walls. This does not matter much for ourunderstanding of them, however; in modern animalssome members of a group may be syncytial, othersmulticellular, and syncytial tissues may occur inotherwise multicellular animals. The first largepresumed eukaryote in the fossil record, the 1.85Ga Grypania (see above), has been compared withsyncytial algae (Runnegar, 1994); and megascopiccarbonaceous compressions in 1.8 Ga rocks in Chinaare reported to have preserved cellular tissue (Zhuet al., 2000). Fossils resembling traces of motilemulticellular organisms have recently been reportedfrom 1.2–2 Ga rocks in Australia (Rasmussen et al.,2002). Their mode of feeding is not known, however.Although strictly not multicellular (though seeShapiro, 1988 for a view of bacterial colonies asmulticellular organisms), microbial mat-formingcommunities will be considered in this context.They are very common Precambrian fossils, andthey show a diversity pattern that, it has beensuggested, relates to the evolutionary appearanceof grazing megascopic animals.STROMATOLITE DECLINEAND THE RISE OFGRAZING MACROFAUNAMicrobial mats are accretionary cohesivemicrobial communities, which are often laminatedand found growing at the sediment–water(occasionally sediment–air) interface (Pierson et al.,1992). The communities may be quite diverse andcomplex, involving photo- and chemosynthesizers,autotrophs and heterotrophs, aerobes and anaerobes,the different types occupying different layers in themats. Photosynthesizing cyanobacteria are often a294

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONdominant constituent in the uppermost layers, andthe mats may be sites of considerable primaryproduction. Filamentous mat-building cyanobacteriaare motile; they tend to dominate in areas of highersedimentation rates because they are able to glideupwards through their sheaths to avoid becomingburied by sediment (Des Marais et al., 1992). Thecohesiveness of the mats is mainly due to largeamounts of extracellular polysaccharides, andcommonly also to the presence of filamentousbacteria. This makes the mats effective in bindingsediment. Mat microorganisms also commonlyinduce mineral deposition as a by-product of theirmetabolism (Burne and Moore, 1987). Mats thusmay form buildups, typically assuming the shapeof pillows, low mounds, or columns.Because they act as sediment binders andcommonly precipitate minerals, mats are easilyfossilized, and their fossil record extends over life’sknown history on Earth. Laminated fossil mats,stromatolites, are particularly prominent inPrecambrian sedimentary environments, mostly incarbonate rocks. In their most distinctive form,developing pillow- or column-like structures, theyare easy to recognize; but flat laminated mats maybe difficult to distinguish from non-microbiallayered sediments. Also, because of the simplephysical principles involved in the shaping also ofmore complex stromatolites, distinguishingbiogenic stromatolites from chemical precipitatesis sometimes difficult or impossible (Buick et al.,1981; Grotzinger and Rothman, 1996).A number of metazoans graze on mats, therebyoften disrupting their coherence. Stromatolitestoday are therefore a feature mainly ofenvironments where grazing fauna is restricted(Garrett, 1970; Farmer, 1992; Steneck et al., 1998),such as hypersaline pools or lakes, hydrothermalsprings, ice-covered lakes, and tidal environments.This has inspired the hypothesis that an observeddecline of stromatolites during the Proterozoic iscoupled to the advent of grazing fauna (Garrett,1970; Awramik, 1971; Walter and Heys, 1985;Walter et al., 1992b; Walter, 1994; Awramik andSprinkle, 1999). If true, this would provide a usefulproxy for the evolutionary appearance ofmacroscopic grazers during a time when moredirect evidence for animal life is lacking. In theview of Walter (1994), grazing and burrowingmetazoans are “the simplest and best explanation”for the stromatolite decline in the Proterozoic. Therelationship between stromatolite decline andgrazing fauna is far from simple, however, and anumber of factors have to be taken into account.WHAT IS STROMATOLITE“DIVERSITY”?The idea that increasing levels of grazingwould lead to an overall decrease in diversity ofthe grazed organisms over evolutionary time iscontrary to the expectations from the croppingprinciple (Stanley, 1973a, 1976b) discussed above.This paradox may be only apparent, however,because stromatolite diversity, as measured,reflects the extent of distribution rather than truetaxonomic diversity.Diversity is a taxonomic measure, the basicparameter in a diversity index being number oftaxa. Because the microbiota of stromatolites isonly rarely preserved, the taxonomy ofstromatolites is based mainly on gross morphology,lamina shape, and microstructure (Bertrand-Sarfatiand Walter, 1981). As a crude rule-of-thumb,morphology largely reflects environmentalinfluence, whereas microstructure is moredependent on the taxonomy of the participatingmicroorganisms (Semikhatov and Raaben, 2000).Consequently, although stromatolite taxonomymakes use of Linnean binomina, it is not equivalentto biological taxonomy.Because stromatolite taxa have proven usefulin stratigraphy (Bertrand-Sarfati and Walter, 1981;Grey and Thorne, 1985; Grey, 1994), it is oftenassumed that the taxonomy as applied reflects somemeasure of evolutionary changes in thecomposition of the microbial communities. If so,stromatolite diversity may indeed be used as aproxy for biological diversity. Environmentaltrends through time, however, may also producestratigraphically discernible changes in stromatolitediversity in a way that mimics biological evolution295

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002(Fischer, 1965; Pratt, 1982; Grotzinger, 1990;Riding, 2000). Although not denying the existenceof a Proterozoic decline, Pratt (1982) argued thatit is in part a chimaera: Phanerozoic stromatolitesare widespread but tend to be diluted by the sheerdiversity of reef-building metazoans (Pratt, 1982;Riding, 2000). Unlike their Proterozoiccounterparts, the younger stromatolites havetherefore not been the focus of taxonomic andstratigraphic studies. Walter and Heys (1985),however, found no correlation between stromatolitediversity and number of authors publishing on therespective time interval in the Proterozoic.Nevertheless, it is likely that the stromatolitediversity curves from the Proterozoic reflect notso much real changes in diversity as changes inthe relative abundance of stromatolites. The lowreported diversities of Phanerozoic stromatolites(cf. Awramik and Sprinkle, 1999) may partly reflectthat fact that stromatolite taxonomy is largely apre-Phanerozoic endeavor. Modern stromatoliteseven have a morphological variability similar tothat of Proterozoic ones (Bauld et al., 1992; Walteret al., 1992a), but their more complex fabric andprominent protist components make them pooranalogues of the Proterozoic forms (Riding, 2000).As a measure of possible effects of disruptiveactivities by metazoans, stromatolite abundance inparticular environments may be more significantthan overall “taxonomic” diversity. Walter and Heys(1985) indeed included a measure of abundance,corresponding to the number of basins in which acertain taxon was recorded from a certainstratigraphic interval. Although this gives someinformation on how geographically widespread ataxon is, as a measure of the total relative abundance,the “abundance” as represented in Walter and Heys’s(1985) curves is flawed, as in fact it incorporatesdiversity. The diversity and abundance curves arealmost indistinguishable, and this may be becausethey basically measure the same thing. This “thing”is probably closer to abundance than to diversity.Thus the apparent decline of “taxonomic”diversity in the Proterozoic may be rather an effectof decreasing abundance of well-preservedstromatolites. As such, it may actually be a moredirect measure than true taxonomic diversity offactors that prevent the growth of stromatolites.Measures of stromatolite numbers per unit of rock(“density” of Grotzinger, 1990) or of areal coverof stromatolites in different environments throughtime would be even more appropriate, but thecollection of such quantitative data would be amomentous task.THE CAUSAL CONNECTIONBETWEEN METAZOAN ASCENTAND STROMATOLITE DECLINEDeclining stromatolite diversity in thePhanerozoic had been noted (Fischer, 1965; Cloudand Semikhatov, 1969), and Garrett (1970)proposed that this was due to non-competitiverestriction from grazing and burrowing animals.Awramik (1971) noted a distinct decline in thediversity of columnar stromatolites already in thelate Proterozoic, from a peak in the Upper Riphean(950–675 Ma), and associated this with theevolutionary appearance of bottom deposit feedersand burrowing metazoans in the subtidalenvironment. Data on Proterozoic diversities havesubsequently been improved by various efforts, inparticular those of Walter and Heys (1985), whoincluded also non-columnar stromatolites andcorrected the diversity values for the relativelengths of the stratigraphic intervals and the relativeintensity of study. Their data confirm the patternof late Proterozoic decline, but suggest thatdiversity peaked in the Middle Riphean (1350–1050 Ma), earlier than in Awramik’s 1971 curvebut consistent with his later published curve(Awramik and Sprinkle, 1999) (see Fig. 1).Schubert and Bottjer (1992) noted a briefresurgence of stromatolites in the Early Triassicand attributed this to the dearth of benthic grazersin the aftermath of the end-Permian marineextinction events. A similar effect may be presentfollowing the Late Devonian (Frasnian–Famennian) mass extinction (Schubert and Bottjer,1992; Whalen et al., 1998).Grotzinger (1990) stressed that the data ofWalter and Heys (1985) show the decline of296

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONstromatolite diversity to have set in already at about1000 Ma, whereas the rise of Ediacaran metazoanswas some 400 million years later. Following Cloud(1968b) and Stanley (1976a, 1976b), a view hasbecome prevalent among paleontologists that thefirst crown-group metazoan (i.e., belonging to anextant branch of animals) appeared no earlier thanabout 600 Ma (for a contrary view, see Fortey etal., 1997; Knoll and Carroll, 1999; Valentine et al.,1999; Budd and Jensen, 2000; Conway Morris,2000). This is in more or less stark contrast tomolecular sequence comparisons (Runnegar, 1982;Wray et al., 1996; Nikoh et al., 1997; Bromham etal., 1998; Gu, 1998; Wang et al., 1999; Hausdorf,2000), which suggest that the major animal lineagesdiverged considerably earlier, maybe around 1,500Ma or even earlier (Wray et al., 1996; Bromham etal., 1998; Wang et al., 1999). The considerablespread of the molecular biology dates currentlyreduces their usefulness, but even the severestcritics of the old-divergence estimates based onmolecules (Ayala et al., 1998; Lynch, 1999; Cutler,2000) agree that molecule dates, if anything,support a much older metazoan history than a literalreading of the fossil record suggests.This discrepancy is still unresolved. It istempting to use the stromatolite record as anindicator of cryptic early small and soft-bodiedmetazoans, and thus to overcome a major weaknessof the fossil record of early animal evolution. Thereare some problems with this approach, however,that have to do with size and abundance of thegrazing metazoans.A number of ecological studies of living biotasupport the proposed connection between thedevelopment of modern microbial mats/stromatolites and the absence of grazing orburrowing fauna. For example, Steneck et al. (1998)investigated a stromatolite-reef complex in theBahamas that represents a gradation from astromatolite-dominated back-reef, to a macroalgaldominatedreef flat, to a reef front dominated bycorals, algae, and fish. Stromatolites transplantedfrom their original site had twice as high a survivalrate in the back-reef than in the reef front. Levels ofherbivory by all kinds of organisms were high inthe reef front, but below detectable levels in the backreef.Although the experiments could not be carriedout under total environmental control, the resultssupport the hypothesis that the presence of grazingfauna has a destructive influence on stromatolitefabric. Other examples in support of the hypothesiswere summarized and discussed by Farmer (1992).The problem is that animals less than a fewmillimeters in size tend not to disrupt the fabric ofmodern microbial mats, and so may co-exist withstromatolites (Farmer, 1992). This means that thekind of animals (small and soft), the exclusivedominance of which might have explained a longnon-record of a Proterozoic metazoan clade, wouldprobably be unable to disturb microbial matssufficiently to cause a decline in stromatoliteabundance/diversity. Similarly, to explain thedecline of stromatolites by the actions of animalslarge and active enough to leave trace fossils wouldmeet with the justified objection that trace fossilsfrom the time of stromatolite decline areexceedingly rare or absent. Occasional trace-likefossils do exist in Meso- and Paleoproterozoicrocks (Faul, 1950; Kauffman and Steidtmann,1981; Breyer et al., 1995; Seilacher et al., 1998;Rasmussen et al., 2002), hinting at the earlypresence of animal-like organisms large enoughto displace sediment and disturb stromatolite fabric,but these traces are exceedingly scarce incomparison with the massive stromatolite declinethat can be traced all over the Earth.The pattern of ecological control of modernstromatolites is still persuasive enough to suggestthat grazing metazoans are important for holdingstromatolites and microbial mats at bay. As anexplanation for stromatolite decline during theNeoproterozoic, the grazing hypothesis may beincomplete, but it seems to explain more of thedemise and the present distribution pattern thando alternative or complementary hypotheses, suchas geochemical trends, competition fromeukaryotes, or taxonomic artifacts (Pratt, 1982;Grotzinger, 1990; Riding, 2000). At present,however, the pattern of stromatolite decline canonly be taken as suggestive of widespread andabundant grazing organisms.297

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002PREDATION ON AND BYPROTEROZOIC PROTISTSModern planktic predators are efficient grazerson phytoplankton (Steele, 1974; Stanley, 1976b),and most of the morphology of planktic protists isprobably a response to predation (Smetacek, 2001).The development of spines or other externalprocesses is widespread and is considered to bemainly a way for the potential prey to expand itsexposed surface beyond the size that a predator ofthe same size order is able to handle, or to reduceits nutrient-to-volume ratio (Burzin, 1997;Butterfield, 1997).Planktic ecosystems are often not veryaccessible to paleontological investigations, butfossil data are potentially of great value to testthe several hypotheses that place important phasesof early metazoan evolution in the plankton(Nielsen, 1985, 1995, 1998; Runnegar, 2000). Onepossible source of such data would be thedemonstration of antipredatory devices in earlyphytoplankton, represented by acritarchs (looselydefined as organic microfossils of unknown andprobably varied affinity; cf. Mendelson andSchopf, 1992a), as indicative of the presence ofgrazers in the water column.There are a number of problems in theinterpretation of such data. Acritarchs are a verydisparate group of fossils, and their ecology is inmany cases unknown. Not all are planktic (seediscussion in Butterfield and Chandler, 1992;Butterfield, 1997), and all may not be protists.Processes may be of different kinds and of differentfunctional significance (for example, they may alsobe selected for as a means to increase water frictionor adhesiveness). The presence of process-bearingacritarchs is therefore not a definite indication ofthe presence of predators/grazers. Conversely,however, a biota of simple spheromorphicacritarchs of consistently low diversity would bestrongly suggestive of the absence of selectivepressure from plankton-eaters.Acritarchs undergo a dramatic diversificationnear the Precambrian–Cambrian boundary (e.g.,Moczydlowska, 1991), with a wealth of complexand process-bearing forms introduced. Diversebiotas of Neoproterozoic large process-bearingacritarchs have been discovered during the lastdecades (Chen and Liu 1986; Zang and Walter,1989; Mendelson and Schopf, 1992a; Zang andWalter, 1992; Knoll, 1994b; Vidal andMoczydlowska-Vidal, 1997; Zhang et al., 1998).Occurrences of process-bearing forms before 1 Gaare exceedingly scarce, though weakly spinyacritarchs are known already from about 2 Ga(Hofmann, 1971; Mendelson and Schopf, 1992b).A recently reported 1.5 Ga biota with processbearingacritarchs (Javaux et al., 2001) is a notableexception to an otherwise rather consistent seriesof Paleoproterozoic and Mesoproterozoic simplespheromorphic assemblages. The total curve ofacritarch species (Fig. 1) suggests that diversitieswere low between 2 and 1 Ga and then rose to apeak before a decline during the greatNeoproterozoic ice ages (the “Snowball Earth”episodes of Kirschvink, 1992 and Hoffman et al.,1998). Another peak after the last of these ice ageswas followed by an extinction event and asubsequent Cambrian bloom.Though this evidence is tentative, it may benoted that the rise in acritarch diversity during theNeoproterozoic is an approximate reciprocal of thedecline seen in stromatolite “diversity” (Fig. 1), andthat both trends may reflect an increase of predatorialactivity. An alternative explanation is that this dualpattern reflects a general diversification of protists,which ecologically displace the mat-formingprokaryotes. However, modern microbial matsusually incorporate protists (red, brown, and greenalgae, diatoms, etc.), which help to stabilize thesediment (Bathurst, 1967; Ward et al., 1992; Riding,2000), so there is no evidence that mat-formingprokaryotes and protists are mutually exclusive.Non-acritarch eukaryotes in the Neoproterozoicalso show probable antipredatory morphologies. The“vase-shaped microfossils”, or melanocyrillids(Bloeser, 1985), have flask-shaped tests andresemble modern testate amoebae (Porter andKnoll, 2000); plate-shaped microfossils ofprobably siliceous composition resemble scales ofvarious Phanerozoic groups of biomineralizing298

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONFIGURE 1—Time diagram showing diversities of stromatolites (after Awramik and Sprinkle, 1999) andacritarchs (after Knoll, 1994b), with timings of the predation-related evolutionary steps discussed in thetext (knife tip points to the first appearance of macrophagous predators in the fossil record; othersymbols show only approximate timings). Snowflakes indicate periods of global glaciations (“SnowballEarth”; Kirschvink, 1992, 2000; Hoffman et al., 1998).protists (Allison, 1981; Allison and Hilgert, 1986;Kaufman et al., 1992).THE ROLE OF PREDATION INTHE CAMBRIAN EXPLOSIONFollowing the massive glaciations in the lateNeoproterozoic (Kirschvink, 1992; Hoffman et al.,1998; Eerola, 2001), the biosphere underwent athorough restructuring. On the face of it, it was aburgeoning of multicellular life, but it has becomeclear that the event affected the biosphere at alllevels, and also that the biological events werecoupled with geochemical, oceanographic,tectonic, and atmospheric changes. The end resultwas that a modern type of marine biosphere, withcomplex food webs and diverse feeding strategies,was established for the first time. In thesedimentary record, it is expressed as profounddifferences between rocks below and above thetransitional interval. This “Cambrian explosion,”which culminated between 550 and 540 Ma, hasbeen and is the focus of intense and multifacetedresearch and speculation (recent reviews of the fieldare by Fortey et al., 1996; Butterfield, 1997; Vidaland Moczydlowska-Vidal, 1997; Knoll and Carroll,1999; Valentine et al., 1999; Brasier, 2000; Buddand Jensen, 2000; Conway Morris, 2000; Levinton,2001, p. 443–494; Zhuravlev, 2001). A recentthematic volume (Zhuravlev and Riding, 2001)deals specifically with the ecological aspects of theCambrian explosion.The Cambrian explosion is, in its anatomy,thoroughly dependent on ecological processes. Thatpredation had an important role might seem obvious;and already in the beginning of the last century theproposal was made that the sudden appearance ofskeletal tissues in the Cambrian was due to theintroduction of predators (Evans, 1912). At that time,the common understanding regarding thePrecambrian biota was “not that animals did not existin those early periods of the earth’s history, but thatthe scarcity of creatures having a resistant skeleton,precluded the preservation of their remains in sucha form as to be easily recognizable” (Matthew,1912). Interestingly, after Cloud (1948, 1968a)successfully argued for the opposite alternative, thatthe metazoans did not have a long Precambrianhistory and that the Cambrian explosion was notjust a calcareous dress-up party, the role of Cambrianpredators began to be downplayed. Evans’ idea aboutthe crucial influence of predators for the origin ofskeletons had lived on (Dunbar, 1960; Hutchinson,1961), but when Cloud’s interpretation de-299

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002emphasized the role of skeletons in the Cambrianexplosion, the view started to become prevalent thatpredators were absent or at least of little importancein Cambrian ecosystems (Nicol, 1966; Glaessner,1972; Valentine, 1973; Erben, 1975). As theevidence for Cambrian predators and predationmounted, however (Bengtson, 1968; Bergström,1973; Alpert and Moore, 1975; Bengtson, 1977;Birkenmajer, 1977; Szaniawski, 1982; Whittingtonand Briggs, 1982), this view again gave way to thenow-common acceptance of predators as a majorand important part of the Cambrian ecosystems (e.g.,Conway Morris, 1986; Debrenne and Zhuravlev,1997). Let us look at a couple of questions:1. Could the Cambrian explosion have beentriggered by predators?2. What was happening in the plankton?3. How did macrophagous predation enterinto the picture?4. What were the responses to macrophagouspredation?COULD THE CAMBRIANEXPLOSION HAVE BEENTRIGGERED BY PREDATORS?The Cambrian explosion has attracted as manyexplanation attempts as ever did the demise ofdinosaurs, and no smoking gun has yet turned up.There has been a certain tendency to suggest that theproximal cause for the event is whatever object orphenomenon is under study, and predation has notescaped this trigger-happiness. There seems littlereason to doubt that predators played an early andimportant role in the evolving Cambrian ecosystems(Stanley, 1976a, 1976b; Bengtson, 1977, 1994;McMenamin, 1986; Vermeij, 1987, 1990;McMenamin and Schulte McMenamin, 1990;Crimes, 1994; Butterfield, 1997), but more isdemanded of a trigger for the Cambrian explosionthan that things would have been different without it.The search for a trigger may in fact be unfruitful:Any phenomenon relating to an event such as thiscan belong to one of three causal categories:prerequisite, trigger, and effect; or it could have nocausal relationship at all with the event (Bengtson,1994). Prerequisites for the Cambrian explosion aremany (free oxygen, shelf space, regulatory genes,biominerals, etc.), and so are its effects. All theseare parts of cascades, however, whereas a truetrigger should be independent of them, an analogueto (and as elusive as) “free will”. It must eitherarise “spontaneously” or be introduced from the“outside”; i.e., it must have a timing independentof the integrated biological–chemical–physicalsystem that determines the actual course of theevent. Such a trigger might arise from, say, acosmic event, but may not be in any wayspectacular. An actual trigger is not even neededfor the event to take place; the impetus may insteadcome from a critical accumulation of prerequisiteconditions (see also Kauffman, 1989).Predation is probably as old as (cellular) lifeitself, and it is likely to have existed in manydifferent forms and at many different levels duringthe formative phases of the Cambrian explosion.What we can hope for is a better understanding ofhow predation interacted with other ecological/evolutionary forces to produce the specific biotasand food webs of the Cambrian and—in the end—in what way this came to determine the subsequentevolution of the biosphere.WHICH WAY THE PLANKTONREVOLUTION?Planktic ecosystems represent most of themarine biomass in today’s oceans, and predator–prey interactions are probably the single mostimportant factor in their evolution (Kitchell, 1983;Signor and Vermeij, 1994; Verity and Smetacek,1996; Butterfield, 1997; Smetacek, 2001). Theevolution of diverse and complex acritarchs duringthe Neoproterozoic suggests activities by plankticand/or benthic predators, and the possibilities ofopen oceans even during extreme “Snowball Earth”events (Hyde et al., 2000) may have left the plankticrealm as the only part of the biosphere relativelyuntouched by the global freezing (Runnegar, 2000).Thus animal predators on protist photosynthesizersmay have evolved during theNeoproterozoic, survived the “Snowball Earth”300

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONbottleneck in the plankton, and later reinvaded thebenthic realm as the shakers-and-movers of theCambrian explosion. This idea would be consistentwith a long metazoan prehistory of small animals,which did not leave a fossil record (Fortey et al.,1996; Peterson et al., 1997; Peterson and Davidson,2000; but see Budd and Jensen, 2000). It has someweak points, however.First, the elaborate acanthomorphic acritarchsof the Neoproterozoic are quite large, typicallyhundreds of micrometers (Zang and Walter, 1989),and Butterfield (1997) has argued that most or all ofthese were benthic, and that the only truly plankticacritarchs of that age are the undifferentiated smallspheroidal forms. Secondly, the spiny processes arenot unquestionably antipredatory aptations. Thirdly,the Proterozoic predators need not be animals—theycould be protists having no direct phylogeneticconnection with the Metazoa. Finally, the openoceanversion of “Snowball Earth” (or “SlushballEarth”) has been strongly contested (Hyde et al.,2001; Schrag and Hoffman, 2001).An alternative view holds that animal predatorson phytoplankton had a much later origin. Signorand Vermeij (1994) stressed that the major groupsof Paleozoic zooplankton and suspension-feedersoriginated in the Middle or Late Cambrian anddiversified in the Ordovician radiation. Theysuggested that this indicates a relatively lateexpansion of animals into the pelagic realm.Butterfield (1997) pointed out, however, that thestrong diversification of small spiny acritarchs (e.g.,Moczydlowska, 1991) and the presence of filterfeedingapparatuses on zooplankton (Butterfield,1994) already in the Early Cambrian indicated thatthe zooplankters were a prominent part of theCambrian radiation.In Butterfield’s (1997) view there was little orno animal presence in the Proterozoic plankton, butthe key event that triggered the Cambrian explosionwas “the expansion of metazoan activities into theplankton,” leading to “the evolution of smallmetazoans able to intercept and exploit a significantproportion of ... [the primary] production, therebypermitting the evolution of the large, activemetazoans that define the Phanerozoic.”Both these scenarios place emphasis onplanktic predators. In the former case there was along Neoproterozoic history of planktic/benthicpredation followed by an ice-age bottleneck wherethe predators survived in a planktic refuge, and asubsequent recolonization of the benthic realmduring which time macrophagous predatorsevolved from planktic predecessors into a majorgoverning force in the Cambrian radiation. Thelatter (Butterfield, 1997) scenario implies thatplanktic filter feeders evolved from the benthicfauna near the beginning of the Cambrian andplayed a decisive role by harvesting the primaryproduction of the water column and making itavailable to larger organisms.ONSLAUGHT OF THEFANGED BEASTSThe importance of larger macrophagouspredators is that they represent second- and higherorderconsumers, signifying the advent of complexfood webs and complex interactions betweendifferent kinds of multicellular organisms. In theEdiacara biota, the first possible macrophagouspredators belong to the first skeletal assemblage—the Neoproterozoic Cloudina-Namacalathusassemblage (Germs, 1972; Grotzinger et al., 2000).These sessile organisms enclosed themselveswithin calcareous tubes and calices. Their generalcnidarian-type morphology suggests that theymight have had a predatory lifestyle like mostmodern cnidarians, but this is conjectural. Moresignificantly, there is evidence of predatory shellborers in this assemblage (see below).From the point of view of a possible derivationof the metazoans from planktic predators, it isinteresting to note that the second oldest evidenceof probable macrophagous predators are theprotoconodonts (Missarzhevskij, 1973; Bengtson,1976, 1977, 1983). These animals had slender teethcombined in a complex grasping apparatus(Landing, 1977; McIlroy and Szaniawski, 2000),and Szaniawski (1982, 2002) has convincinglyargued for their close affinity to modernchaetognaths, arrow worms. Chaetognaths are one301

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Figure 2302

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONof the major groups of modern plankticmacrophagous predators, and they are currentlybelieved to occupy a basal position in the bilaterianphylogenetic tree (Telford and Holland, 1993,1997; Wada and Satoh, 1994; Halanych, 1996).The list of other Cambrian macrophagouspredators identified as fossils is now long, in starkcontrast to the commonly held opinion a few decadesago that predators were few or absent in the Cambrian(Glaessner, 1972; Erben, 1975). The quantitativelymost important macrophagous predators appear tobe arthropods (Budd, 2001; Hughes, 2001), thoughthe relative predisposition of arthropods forfossilization, even in “soft-bodied” lagerstätten,may have enhanced their apparent dominance.Conway Morris (1986) analyzed thecommunity structure of the Phyllopod Bed in theBurgess Shale. His main conclusion was that theproportion of predators in Cambrian ecosystemshad been severely underestimated in previousstudies based only on hard-part preservation.Bengtson et al. (1992) tabulated metazoan generathrough the Lower Cambrian and concluded thatalthough predators (including herbivores) were oflow diversity throughout this time period, the basiccomponents of a modern marine ecosystem werepresent already from the beginning of theCambrian. Zhuravlev and Debrenne (1996) andDebrenne and Zhuravlev (1997) reviewed thetrophic structure of three types of Lower–MiddleCambrian benthic environments—reefal, levelbottomopen-marine, and level-bottom dysaerobic(Fig. 2), suggesting short and simple food chainscomparable with those of recent eutrophic areas.ANTIPREDATORY RESPONSESOf the many direct aptations to counterpredation pressure that are available to organisms,only a few types are potentially visible in the fossilrecord. Chemical defense, life-historymodifications, migrations (or indeed any behaviorthat does not leave trace fossils), mimicry, andprotective coloring are all unlikely to leave arecognizable fossil signature, and so must largelybe left as a reminder to the paleontologist that theinformation is incomplete.Skeletons.—The once-common interpretation ofthe Cambrian explosion as a biomineralization event(see quote from Matthew, 1912, above), an“explosion of fossils” rather than of organisms, isnow largely in disrepute. This is partly because ofthe massive evidence for an equally rapid evolutionof non-skeletal organisms; partly because of therealization that biomineralization as such iswidespread among organisms (Lowenstam andWeiner, 1989). Bengtson (1994) formulated fourgeneral conclusions regarding the advent of animalskeletons: 1. Biomineralization has an ancienthistory and was only a prerequisite for the adventof skeletons. 2. Skeletons are constructed using avariety of processes and materials. Minerals aresuitable because they give hardness to the compositematerial, can be produced using exapted pathways,and are physiologically cheap. 3. Whereas the initialchoice of shell mineral usually precludes futureevolutionary switches to other minerals (because ofthe intricate systems developed to modify the growthof the mineral), there is no reliable indication of any←FIGURE 2—Trophic webs in principal Early Cambrian benthic communities according to Zhuravlevand Debrenne (1996; partly based on data from Conway Morris, 1986, and Kruse et al., 1995).1, Reefal. 2, Level-bottom open-marine. 3, Level-bottom dysaerobic (diagram to the left indicates relativeimportance of these environments). D—Deposit feeders; S—Suspension feeders; F—Filter feeders;B—Browsers/herbivores. (Note that all of these categories include various grazing predators accordingto the broader definition of predation used in the present paper, and that the category “predators”corresponds to macrophagous predators on animals.) Arrow width indicates relative biomass. Questionmarks indicate categories that are not preserved but hypothesized to be present; question marks withinparentheses indicate doubtful trophic assignments of taxa.303

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002regularity in the acquisition of skeletal mineralswithin or between clades. 4. The propellant in theevolution of skeletons was organismal interactions,and a primary factor behind the evolution of tubes,shells, sclerites, and spicules, was selection pressurefrom predators.The primary role of ecological selectionpressures does not preclude the possibility thatphysiological (Marin et al., 1996) or geobiochemical(Kirschvink and Hagadorn, 2000) mechanismsinvolved in skeletal biomineralization have commonevolutionary origins. These mechanisms are ofcourse prerequisites for the appearance ofbiomineralized skeletons, but their use in skeletonformation is likely to be exaptational.A survey of the skeletal types appearing in theCambrian biota (Bengtson and Conway Morris,1992) differentiated between spicules, tubes, conchs,external sclerites, toothlike structures, carapaces, andcalcareous reinforcements. All these have defensive/protective potential, though for some of the skeletaltypes other functions, such as internal support, maybe more fundamental than protection.It is difficult in fossil material to ascertain thebest function of a particular structure, and evidencefrom living organisms underscores that intuitivelycorrect interpretations are not always the best. Acase in point is spicules—mineralized structures,often needle-shaped, distributed within the softtissues of most major groups of the Metazoa (e.g.,Rieger and Sterrer, 1975). Many sponges are fullof needle-sharp siliceous or calcareous spicules,which might intuitively seem important to deterpredators (Wainwright et al., 1976; Hartman,1981), but there is in fact little evidence in supportof a protective function for the spicular skeletonof sponges (Bergquist, 1978, p. 94). Experimentalwork on modern sponges (McClintock, 1987;Chanas and Pawlik, 1995, 1996; Dunlap andPawlik, 1998; Waddell and Pawlik, 2000a, 2000b)indicates that predators of various kinds (fish,arthropods, echinoderms) are not influenced intheir selection by the presence of spicules in theprey tissues. Whereas such results could partly bean effect of specialized spongivores having evolvedmechanisms to diminish the potential harmfulnessof sponge spicules (Oshel and Steele (1985)reported such a case concerning an amphipodpredator), even generalist feeders seem undeterredby spicules in the sponge prey (Chanas and Pawlik,1995, 1996; Dunlap and Pawlik, 1998).Vreeland and Lasker (1989) found a similarpattern in spiculated gorgonian octocorals preyedon by a polychaete worm: the polychaete’spreference for a particular gorgonian species wasnot correlated with the sclerite density of the latter.On the other hand, a gastropod feeding ongorgonians showed preference for colonies withshort sclerites over those with long ones (West,1998), and gorgonians respond to predatorsimulatedmechanical damage by generating astiffer cortex with longer sclerites (West, 1997).Similar problems of functional interpretationexist with regard to many of the other skeletaltypes. Reasonable arguments can be made whymost of them should have a primary protectivefunction (Vermeij, 1987, 1990; Bengtson, 1994),but even if that is correct the picture may beobscured by multiple other functions. I will givetwo examples of early skeletal fossils, however,where a primary antipredatory function appearswell supported by the evidence.Tube-dwellers are common among the earlyskeletal animals, and the varying composition andmorphology of the tubes suggest that a number ofindependently derived lineages are represented(Bengtson and Conway Morris, 1992). One of theearliest known animals producing a mineralizedtube, the late Neoproterozoic Cloudina, has beenfound to display boreholes made by a predatory orparasitic organism (Bengtson and Yue, 1992).Predatory boreholes are known from various typesof tubes and shells in the Cambrian (Bengtson,1968; Miller and Sundberg, 1984; Conway Morrisand Bengtson, 1994; Streng, 1999), and theyconstitute one of the most important records forpredation throughout the Phanerozoic (Vermeij,1987; Kelley and Hansen, 1993; Kowalewski etal., 1998). The presence of such borings evenamong the earliest skeletal animals stronglysuggests that protection against predators was aprimary function for these tubes.304

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONThe other example of an early exoskeleton witha clear antipredatory function is that of thechancelloriids. These sessile, bag-shapedorganisms are common from the Early Cambrianto the early Late Cambrian (Walcott, 1920). Theyhad a soft integument beset with compositecalcareous sclerites having sharp, radiating spines(Bengtson and Hou 2001). Since the sclerites wereexternal and non-interlocking, they could not havehad a supporting function, and since the body wassessile and attached, the sclerites would not haveserved to increase friction. Thus there seems to beno other conceivable function for the chancelloriidsclerites than antipredatory: Mechanicalconsiderations suggest that they would preventaccess to the integument by large (i.e., about thesize of the distance between the sclerites or larger)predators, and the corresponding morphology inspiny cacti is known to deter herbivores (Theimerand Bateman, 1992; LeHouerou, 1996).Various instances of probably predatorinflicteddamages to skeletons of Cambrian animals(Pocock, 1974; Conway Morris, 1985; ConwayMorris and Jenkins, 1985; Babcock and Robison,1989; Babcock, 1993; Pratt, 1998; Nedin, 1999)have also been taken as evidence of the protectivefunction of these skeletons.Behavior.—One of the crucial pieces ofevidence for the Cambrian explosion as an allencompassingbiological overhaul (rather than justthe invention of skeletons) has been the dramaticdiversification of trace fossils across thePrecambrian-Cambrian boundary (Seilacher, 1956;Alpert, 1977; Crimes, 1987, 1989, 1992, 1994;Macnaughton and Narbonne, 1999). During thisprocess metazoans expanded their biotopes into theinfaunal realm, somewhat earlier in the clastic thanin the carbonate environments (Droser and Bottjer,1988; Droser et al., 1999; McIlroy and Logan, 1999;Droser and Li, 2001). McIlroy and Logan (1999)interpret this in terms of a positive feedback loopbegun in deeper waters by the increased downwardflux of organic matter through fecal pellets producedby plankton-harvesting metazoan zooplankters(Logan et al., 1995, 1997; cf. Butterfield, 1997, anddiscussion above); bioturbation by deposit-feedingmetazoans would then gradually drive oxygen, labileorganic matter, and nutrients deeper into thesediment, stimulating deeper bioturbation.Whereas deposit-feeders are to a great extentdriven by the availability of organic matter andnutrients, many traces in the Neoproterozoic–Cambrian reflect activities other than depositfeedingor grazing (Crimes, 1992). Vertical burrowscontaining a core of trilobite shell fragments havebeen interpreted as made by sea anemones preyingon trilobites (Alpert and Moore, 1975). Deepdwelling traces (Diplocraterion, Rhizocorallium,Skolithos, etc.) appear to represent protectivebehavior, in effect equivalent to that used by tubedwellinganimals. They may thus have arisen inresponse to predation pressure.Predators on infauna may dig their own holesor be “weasel predators” (Woodin, 1983), enteringthe sediment through the hole made by the prey. Inthe latter case, no trace-fossil evidence is likely tobe preserved. Some evidence for the former type ofpredation exists among Cambrian ichnocoenoses.Associations of arthropod traces and “worm”burrows have been interpreted as instances ofarthropod predation on burrowing infauna(Martinsson, 1965; Bergström, 1973; Jensen, 1990,1997; Pickerill and Blissett, 1999); however, a recentstudy of an assemblage with 29 such associationssuggested that the “worm” burrows were formedafter the arthropod traces and thus that the “worms”more likely were seeking out patches visited by thearthropod (Rydell et al., 2001). The role of infaunalpredation in soft sediments during the Cambrianexplosion is thus poorly understood, partly becausethe trace fossil evidence may be difficult to interpret,but also because the effects of predation incorresponding modern environments are very poorlyknown (Wilson, 1990).SUMMARYSteps in the early evolution of predation.—Theforegoing discussion has dealt with predators atdifferent levels of organization, and predation atdifferent trophic levels. The perspective has shiftedover the time period covered—from the earlyinteraction between prokaryote cells to the late305

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002emergence of macrophagous predators and secondorderconsumers. Also, the amount and quality ofinformation available from the respective stages hasforced a shift in the level of analysis from speculationbased on general biological principles, to hypothesesbased on indirect evidence in the fossil record, andto hypotheses more firmly based on fossil evidence.The theme has been escalation in the sense ofVermeij (1987), but brought to bear on the earlyhistory of life up into the Cambrian (whereVermeij’s story begins). The etymology of“escalation”—from Spanish escala, ladder, andLatin scala, with the same meaning—suggestsstepwise rather than gradual shifts, and Vermeij(1987) stresses the pattern of punctuated equilibria(Eldredge and Gould, 1972) when analyzing theanatomy of escalation. Although not dealing withpunctuated equilibria at the species level, I havefocused on some important steps (or escalations)in which predation is likely to have played a keyrole (Fig. 1). They all represent the attainment of ahigher level of organization (in the sense ofcombining previously existing systems to a newwhole), and so correspond to some of the majortransitions in evolution as discussed by MaynardSmith and Szathmáry (1995).Step 1.—From prokaryotic to eukaryotic.Predation was in all probability the determiningfactor in this event, and the resulting organismscombine characters of predators and prey in a waythat opens new evolutionary possibilities. The timefor this step is poorly constrained to around 2.7Ga (when geochemical evidence suggests that thehost and at least one of the guest symbionts wereavailable).Step 2.—From unicellular to multicellular. Thisstep was taken many times independently, but as ameans of producing bigger organisms it may reflectpredatorial pressures from cell-engulfing eukaryotes.At least in some lineages this happened soon after 2Ga (when the atmospheric oxygen had gone up andthe first non-stromatolite macrofossils appear).Step 3.—The appearance of mobile selectivepredators on bacteria and protists. This is the mostuncertain event of them all, for it may go back asfar as 2 Ga or it may be not much older than theCambrian explosion. The combined indicationsfrom the decline of stromatolites and thediversification of acritarchs suggest that it mayhave begun around 1 Ga.Step 4.—From simple, mostly microbial,ecosystems to ones with complex food webs andsecond- and higher-order consumers. Theappearance of macrophagous predators is thetelltale sign, and it took place no later than a fewmillion years before the beginning of the Cambrian,or around 550 Ma.FUTURE RESEARCHDIRECTIONSThe really interesting new research results arealways the unexpected ones. Any recipe for futureresearch that I may attempt to write will be one forstale cookies—the unpredictable cannot bepredicted. It should be clear from this review of theearly history of predation, however, that there areenormous gaps in our knowledge of the ecologicalinterplay between organisms up to about thePrecambrian–Cambrian transition (after that the gapsare only huge). The partial filling of some of thesegaps is something that one might humbly wish for.For example: Where did the main organismalinteractions take place that led to the Cambrianexplosion of animals? The planktic habitat has ofold been considered difficult to analyze from fossils,both because planktic organisms tend to be fragileand because in order to be preserved at all theyneed to be shifted out of their habitat. Thediscoveries that delicate animal tissues, such asminute arthropod limbs (Müller and Walossek,1985; Butterfield, 1997) and embryonalblastomeres (Zhang and Pratt, 1994; Bengtson andYue 1997; Xiao et al., 1998; Yue and Bengtson,1999; Xiao and Knoll, 2000), may be exquisitelypreserved by carbonization or phosphatization in 3-dimensional detail in rocks of this age spell greatpromise for the investigation of early animals,whether they be planktic or benthic. The extensivephosphorite deposits from the time period still guardmany secrets, and a suitable target in these rocksmay be fecal pellets, today an important medium of306

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONpreservation of planktic prey (Kitchell, 1983).Geochemical methods are becoming ever moresensitive, and the search for characteristicbiomarkers in Proterozoic rocks has started to yieldspectacular insights into the occurrence of organismsin sequences where a morphological record islacking (McCaffrey et al., 1994; Brocks et al., 1999;Summons et al., 1999). The recently realizedpossibility of analyzing isotopic ratios in individualProterozoic microfossils (House et al., 2000) maylet us characterize fossil organisms physiologicallyand thereby throw light on their mode of life. Theuse of lipid ratios in Pleistocene mollusc shells toidentify predators vs. suspension feeders (CoBabeand Ptak, 1999) is a particularly fascinatingextension of biogeochemical methods with greatecological significance, although the currentlyavailable analytical procedures are hardly applicableto the Proterozoic and Cambrian fossil record.Trace fossils are a direct reflection of behavior,and may represent the currently most profitableavenue for research into early predatory andantipredatory behavior. Whereas Cambrian tracefossils and bioturbation are almost ubiquitous, lateNeoproterozoic examples are comparativelyscarce, and reports of trace fossils older than 600Ma have yet to find general acceptance.Nonetheless, several reports of earlier trace-likefossils (e.g., Faul, 1950; Kauffman and Steidtmann,1981; Breyer et al., 1995; Seilacher et al., 1998;Rasmussen et al., 2002) have still to be given abetter explanation. Rather than being ignored asfreakish occurrences, they should be used as searchimages in a more concerted exploration forevidence of possible early adventures into motilemulticellularity and associated behavior.In the Cambrian, the prospects are quite goodfor a deepening understanding of the ecologicalinteractions shaping the biota. The main reason forthis is that the Cambrian is unusually blessed withfossil preservation lagerstätten. New Cambrian lifeforms are being reported from these each year, andwhen the basic morphologic and taxonomicinformation has been obtained, the foundation islaid for spectacular advances in the synecology andautecology of the Cambrian biota. With that, wewill also get a better handle on the ecology of theCambrian explosion itself.REFERENCESABRAMS, P. A. 2000. The evolution of predator-prey interactions: theory and evidence. Annual Review of Ecologyand Systematics, 31:79–105.AGRAWAL, A. A., C. LAFORSCH, AND R. TOLLRIAN. 1999. Transgenerational induction of defences in animals andplants. Nature, 401:60–63.ALLISON, C. W. 1981. Siliceous microfossils from the Lower Cambrian of Northwest Canada: possible source forbiogenic chert. Science, 211:53–55.ALLISON, C. W., AND J. W. HILGERT. 1986. Scale microfossils from the Early Cambrian of Northwest Canada.Journal of Paleontology, 60:973–1015.ALPERT, S. P. 1977. Trace fossils and the basal Cambrian boundary, p. 1–8. In T. P. Crimes and J. P. Harper (eds.),Trace fossils 2. Geological Journal, Special Issue.ALPERT, S. P., AND J. N. MOORE. 1975. Lower Cambrian trace fossil evidence for predation on trilobites. Lethaia,8:223–230.AWRAMIK, S. 1971. Precambrian columnar stromatolite diversity: reflection of metazoan appearance. Science,174:825–827.AWRAMIK, S. M., AND J. SPRINKLE. 1999. Proterozoic stromatolites: The first Marine Evolutionary Biota. HistoricalBiology, 13:241–253.AYALA, F. J., A. RZHETSKY, AND F. J. AYALA. 1998. Origin of the metazoan phyla: Molecular clocks confirmpaleontological estimates. Proceedings of the National Academy of Sciences, USA, 95:606–611.307

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002BABCOCK, L. E. 1993. Trilobite malformation and the fossil record of behavioral asymmetry. Journal of Paleontology,67:217–229.BABCOCK, L. E., AND R. A. ROBISON. 1989. Preferences of Paleozoic predators. Nature, 337:695–696.BATHURST, R. G. C. 1967. Subtidal gelatinous mat, sand stabilizer and food, Great Bahama Bank. Journal ofGeology, 75:736–738.BAULD, J., E. D’AMELIO, AND J. FARMER. 1992. Modern microbial mats, p. 261–270. In J. W. Schopf and C. Klein(eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.BENGTSON, S. 1968. The problematic genus Mobergella from the Lower Cambrian of the Baltic area. Lethaia, 1:325–351.BENGTSON, S. 1976. The structure of some Middle Cambrian conodonts, and the early evolution of conodontstructure and function. Lethaia, 9:185–206.BENGTSON, S. 1977. Aspects of problematic fossils in the early Palaeozoic. Acta Universitatis Upsaliensis. Abstractsof Uppsala Dissertations from the Faculty of Science, 415:1–71.BENGTSON, S. 1983. The early history of the Conodonta. Fossils and Strata, 15:5–19.BENGTSON, S. 1994. The advent of animal skeletons, p. 412–425. In S. Bengtson (ed.), Early Life on Earth,. obelSymposium 84. C,lumbia University Press, New York.BENGTSON, S., AND S. CONWAY MORRIS. 1992. Early radiation of biomineralizing phyla, p. 447–481. In J. H. Lippsand P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum, New York.BENGTSON, S., J. D. FARMER, M. A. FEDONKIN, J. H. LIPPS, AND B. RUNNEGAR. 1992. The Proterozoic–Early Cambrianevolution of metaphytes and metazoans, p. 425–462. In J. W. Schopf and C. Klein (eds.), The ProterozoicBiosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.BENGTSON, S., AND X. HOU. 2001. The integument of Cambrian chancelloriids. Acta Palaeontologica Polonica, 46:1–22.BENGTSON, S., AND Z. YUE. 1992. Predatorial borings in Late Precambrian mineralized exoskeletons. Science,257:367–369.BENGTSON, S., AND Z. YUE. 1997. Fossilized metazoan embryos from the earliest Cambrian. Science, 277:1645–1648.BENTON, M. J. 1983. Progress and competition in macroevolution. Biological Reviews, 62:395–338.BERGQUIST, P. R. 1978. Sponges. University of California Press, Berkeley, 268 p.BERGSTRÖM, J. 1973. Organization, life, and systematics of trilobites. Fossils and Strata, 2:1–69.BERTRAND-SARFATI, J., AND M. R. WALTER. 1981. Stromatolite biostratigraphy. Precambrian Research, 15:353–371.BIRKENMAJER, K. 1977. Trace fossil evidence for predation on trilobites from Lower Cambrian of south Spitsbergen.Norsk Polarinstitutt Årbok, 1976:187–194.BLOESER, B. 1985. Melanocyrillium, a new genus of structurally complex late Proterozoic microfossils from theKwagunt Formation (Chuar Group), Grand Canyon, Arizona. Journal of Paleontology, 59:741–765.BONNER, J. T. 1965. Size and Cycle: An Essay on the Structure of Biology. Princeton University Press, Princeton, NJ, 219 p.BONNER, J. T. 1993. Life Cycles. Reflections of an Evolutionary Biologist. Princeton University Press, Princeton, NJ, 209 p.BONNER, J. T. 1998. The origins of multicellularity. Integrative Biology, 1:27–36.BONNER, J. T. 2000. First Signals. The Evolution of Development. Princeton University Press, Princeton, NJ, 156 p.BRASIER, M. D. 2000. The Cambrian explosion and the slow burning fuse. Science Progress, 83:77–92.BREYER, J. A., A. B. BUSBEY, R. E. HANSON, AND E. C. I. ROY. 1995. Possible new evidence for the origin ofmetazoans prior to 1 Ga: Sediment-filled tubes from the Mesoproterozoic Allamoore Formation, Trans-Pecos Texas. Geology, 23:269–272.BROCKS, J. J., G. A. LOGAN, R. BUICK, AND R. E. SUMMONS. 1999. Archean molecular fossils and the early rise ofeukaryotes. Science, 285:1033–1036.BROMHAM, L., A. RAMBAUT, R. FORTEY, A. COOPER, AND D. PENNY. 1998. Testing the Cambrian explosion hypothesis byusing a molecular dating technique. Proceedings of the National Academy of Sciences of the USA, 95:12386–12389.BUDD, G. E. 2001. Ecology of nontrilobite arthropods and lobopods in the Cambrian, p. 404–427. In A. Yu. Zhuravlevand R. Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.BUDD, G. E., AND S. JENSEN. 2000. A critical reappraisal of the fossil record of the bilaterian phyla. BiologicalReviews, 75:253–295.BUICK, R., J. S. R. DUNLOP, AND D. I. GROVES. 1981. Stromatolite recognition in ancient rocks: An appraisal ofirregularly laminated structures in an Early Archaean chert–barite unit from North Pole, Western Australia.Alcheringa, 5:161–181.308

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONBURNE, R. V., AND L. S. MOORE. 1987. Microbialites: Organosedimentary deposits of benthic marine communities.Palaios, 2:241–254.BURZIN, M. B. 1997. Zachem shipy dokembrijskomu fitoplanktonu? [Why do Precambrian phytoplankton havespines?] Priroda, 1997:98–110 (in Russian).BUTTERFIELD, N. J. 1994. Burgess Shale-type fossils from a Lower Cambrian shallow shelf sequence in northwesternCanada. Nature, 369:477–479.BUTTERFIELD, N. J. 1997. Plankton ecology and the Proterozoic–Phanerozoic transition. Paleobiology, 23:247–262.BUTTERFIELD, N. J. 2000. Bangiomorpha pubescens n.gen., n.sp.: implications for the evolution of sex, multicellularity,and the Mesoproterozoic/Neoproterozoic radiation of eukaryotes. Paleobiology, 26:386–404.BUTTERFIELD, N. J. 2001. Ecology and evolution of Cambrian plankton, p. 200–216. In A. Yu. Zhuravlev and R.Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.BUTTERFIELD, N. J., AND F. W. CHANDLER. 1992. Palaeoenvironmental distribution of Proterozoic microfossils,with an example from the Agu Bay Formation, Baffin Island. Palaeontology, 35:943–957.CAVALIER-SMITH, T. 1987a. The origin of eukaryote and archaebacterial cells. Annals of the New York Academyof Sciences, 503:17–54.CAVALIER-SMITH, T. 1987b. The simultaneous symbiotic origin of mitochondria, chloroplasts and microbodies.Annals of the New York Academy of Sciences, 503:55–71.CAVALIER-SMITH, T. 2001. Obcells as proto-organisms: Membrane heredity, lithophosphorylation, and the originsof the genetic code, the first cells, and photosynthesis. Journal of Molecular Evolution, 53:555–595.CHANAS, B., AND J. R. PAWLIK. 1995. Defenses of Caribbean sponges against predatory reef fish .2. Spicules,tissue toughness, and nutritional quality. Marine Ecology Progress Series, 127:195–211.CHANAS, B., AND J. R. PAWLIK. 1996. Does the skeleton of a sponge provide a defense against predatory reef fish?Oecologia, 107:225–231.CHEN M., AND K. LIU. 1986. [The geological significance of newly discovered microfossils from the Upper Sinian(Doushantuo age) phosphorites.] Scientia Geologica Sinica, 1986:46–53.CLOUD, P. E. 1948. Some problems and patterns of evolution exemplified by fossil invertebrates. Evolution, 2:322–350.CLOUD, P. E. 1968a. Atmospheric and hydrospheric evolution on the primitive earth. Science, 160:729–736.CLOUD, P. E., JR. 1968b. Pre-metazoan evolution and the origins of the Metazoa, p. 1–72. In E. T. Drake (ed.),Evolution and Environment. Yale University Press, New Haven, CT.CLOUD, P. E., AND M. A. SEMIKHATOV. 1969. Proterozoic stromatolite zonation. American Journal of Science,267:1017–1061.COBABE, E., AND A. J. PTAK. 1999. Comparison of in situ mineral-associated lipid compositions in modern invertebrateskeletons: preliminary evidence of dietary and environmental influence. Paleobiology, 25:201–211.CONWAY MORRIS, S. 1985. The Middle Cambrian Metazoan Wiwaxia corrugata (Matthew) from the BurgessShale and Ogygopsis Shale, British Columbia, Canada. Philosophical Transactions of the Royal Society ofLondon B, 307:507–586.CONWAY MORRIS, S. 1986. The community structure of the Middle Cambrian Phyllopod Bed (Burgess Shale).Palaeontology, 29:423–467.CONWAY MORRIS, S. 2000. The Cambrian “explosion”: Slow-fuse or megatonnage? Proceedings of the NationalAcademy of Sciences, 97:4426–4429.CONWAY MORRIS, S., AND S. BENGTSON. 1994. Cambrian predators: possible evidence from boreholes. Journal ofPaleontology, 68:1–23.CONWAY MORRIS, S., AND R. J. F. JENKINS. 1985. Healed injuries in Early Cambrian trilobites from South Australia.Alcheringa, 9:167–177.CRIMES, T. P. 1987. Trace fossils and correlation of late Precambrian and early Cambrian Strata. GeologicalMagazine, 124:97–119.CRIMES, T. P. 1989. Trace fossils, p. 166–185. In J. W. Cowie and M. D. Brasier (eds.), The Precambrian–CambrianBoundary: Oxford Monographs in Geology and Geophysics. Oxford University Press, Oxford.CRIMES, T. P. 1992. The record of trace fossils across the Proterozoic–Cambrian boundary, p. 177–202. In J. H.Lipps and P. W. Signor (eds.), Origin and Early Evolution of the Metazoa. Plenum, New York.309

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002CRIMES, T. P. 1994. The period of early evolutionary failure and the dawn of evolutionary success: The record ofbiotic changes across the Precambrian–Cambrian boundary, p. 105–133. In S. K. Donovan (ed.), ThePalaeobiology of Trace Fossils. Johns Hopkins University Press, Baltimore, MD.CUTLER, D. J. 2000. Estimating divergence times in the presence of an oversdispersed molecular clock. MolecularBiology and Evolution, 17:1647–1660.DAWKINS, R., AND R. J. KREBS. 1979. Arms races between and within species. Proceedings of the Royal SocietyBiological Sciences Series B, 205:489–511.DEBRENNE, F., AND A. YU. ZHURAVLEV. 1997. Cambrian food web: A brief review. Geobios, M.S. no. 20:181–188.DES MARAIS, D. J., E. D’AMELIO, J. D. FARMER, B. BARKER JØRGENSEN, A. C. PALMISANO, AND B. K. PIERSON. 1992.Case study of a modern microbial mat-building community: The submerged cyanobacterial mats of GuerroNegro, Baja California Sur, Mexico, p. 325–333. In J. W. Schopf and C. Klein (eds.), The ProterozoicBiosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.DROSER, M. L., AND D. J. BOTTJER. 1988. Trends in depth and extent of bioturbation in Cambrian carbonate marineenvironments, Western United States. Geology, 16:233–236.DROSER, M. L., J. G. GEHLING, AND S. JENSEN. 1999. When the worm turned: Concordance of Early Cambrianichnofabric and trace-fossil record in siliciclastic rocks of South Australia. Geology, 27:625–628.DROSER, M. L., AND X. LI. 2001. The Cambrian radiation and the diversification of sedimentary fabrics, p. 137–169. In A.Yu. Zhuravlev and R. Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.DUNBAR, C. O. 1960. Historical Geology, 2nd Ed. Wiley, New York, 500 p.DUNLAP, M., AND J. R. PAWLIK. 1998. Spongivory by parrotfish in Florida mangrove and reef habitats. MarineEcology, 19:325–337.DE DUVE, C. 1995. Vital Dust. Life as a Cosmic Imperative. Longman, 382 p.EEROLA, T. T. 2001. Climate change at the Neoproterozoic–Cambrian transition, p. 90–106. In A. Yu. Zhuravlevand R. Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.ELDREDGE, N., AND S. J. GOULD. 1972. Punctuated equilibria: An alternative to phyletic gradualism, p. 82–115. InT. J. M. Schopf (ed.), Models in Paleobiology. Freeman, Cooper & Co., San Francisco, CA.ERBEN, H. K. 1975. Die Entwicklung der Lebewesen. Piper, München, 518 p.EVANS, J. W. 1912. The sudden appearance of the Cambrian fauna. Report of the Session of the InternationalGeological Congress, Stockholm, 1910:543–546.FARMER, J. 1992. Grazing and bioturbation in modern microbial mats, p. 295–297. In J. W. Schopf and C. Klein(eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.FAUL, H. 1950. Fossil burrows from the Precambrian Ajibik Quartzite of Michigan. Journal of Paleontology, 24:102–106.FISCHER, A. G. 1965. Fossils, early life and atmospheric history. Proceedings of the National Academy of Sciences,53:1205–1215.FORTEY, R. A., D. E. G. BRIGGS, AND M. A. WILLS. 1996. The Cambrian evolutionary “explosion”: Decouplingcladogenesis from morphological disparity. Biological Journal of the Linnean Society, 57:13–33.FORTEY, R. A., D. E. G. BRIGGS, AND M. A. WILLS. 1997. The Cambrian evolutionary “explosion” recalibrated.BioEssays, 19:429–432.GARRETT, P. 1970. Phanerozoic stromatolites: noncompetitive ecologic restriction by grazing and burrowing animals.Science, 167:171–173.GERMS, G. J. B. 1972. The stratigraphy and paleontology of the Lower Nama Group, South West Africa. Chamberof Mines, Precambrian Research Unit, Bulletin:1–250.GLAESSNER, M. F. 1972. Precambrian palaeozoology, p. 43–52. In J. B. Jones and B. McGowran (eds.), StratigraphicProblems of the Later Precambrian and Early Cambrian: University of Adelaide, Center for PrecambrianResearch, Special Paper.GOULD, S. J. 1985. The paradox of the first tier: an agenda for paleobiology. Paleobiology, 11:2–12.GOULD, S. J. 1988. Trends as changes in variance: a new slant on progress and directionality in evolution. Journalof Paleontology, 62:319–329.GOULD, S. J. 2002. The Structure of Evolutionary Theory. Belknap, Cambridge, MA, 1433 p.GOULD, S. J., AND C. B. CALLOWAY. 1980. Clams and brachiopods—ships that pass in the night. Paleobiology, 6:383–396.GRAY, M. W. 1999. Evolution of organellar genomes. Current Opinion in Genetics and Development, 9:678–687.310

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONGRAY, M. W., AND D. F. SPENCER. 1996. Organellar evolution, p. 109–126. In D. M. Roberts, P. Sharp, G. Alderson,and M. Collins (eds.), Evolution of Microbial Life. Cambridge University Press, Cambridge.GREY, K. 1994. Stromatolites from the Palaeoproterozoic (Orosirian) Glengarry Group, Glengarry Basin, WesternAustralia. Alcheringa, 18:275–300.GREY, K., AND A. M. THORNE. 1985. Biostratigraphic significance of stromatolites in upward shallowing sequencesof the early Proterozoic Duck Creek Dolomite, Western Australia. Precambrian Research, 29:183–206.GROTZINGER, J., AND D. H. ROTHMAN. 1996. An abiotic model for stromatolite morphogenesis. Nature, 383:423–425.GROTZINGER, J. P. 1990. Geochemical model for Proterozoic stromatolite decline. American Journal of Science,290:80–103.GROTZINGER, J. P., W. A. WATTERS, AND A. H. KNOLL. 2000. Calcified metazoans in thrombolite–stromatolite reefsof the terminal Proterozoic Nama Group, Namibia. Paleobiology, 26:334–359.GU, X. 1998. Early metazoan divergence was about 830 million years ago. Journal of Molecular Evolution,47:369–371.GUILLARD, R. R. L., AND P. KILHAM. 1977. The ecology of marine planktonic diatoms, p. 372–469. In D. Werner(ed.), The Biology of Diatoms. University of California Press, Berkeley, CA.HALANYCH, K. M. 1996. Testing hypotheses of chaetognath origins: long branches revealed by 18S ribosomalDNA. Systematic Biology, 45:223–246.HAN, T.-M., AND B. RUNNEGAR. 1992. Megascopic eukaryotic algae from the 2.1 billion-year-old Negaunee Iron-Formation, Michigan. Science, 257:232–235.HANSEN, B., P. K. BJØRNSEN, AND H. J. HANSEN. 1994. The size ratio between planktonic predators and their prey.Limnology and Oceanography, 39:395–403.HARTMAN, W. D. 1981. Form and distribution of silica in sponges, p. 453–493. In T. L. Simpson and B. E. Volcani(eds.), Silicon and siliceous structures in biological systems. Springer, New York.HAUSDORF, B. 2000. Early evolution of the Bilateria. Systematic Biology, 49:130–142.HOFFMAN, P. F. 1987. Early Proterozoic foredeeps, foredeep magmatism, and Superior-type iron-formations ofthe Canadian shield, p. 85–98. In A. Kroener (ed.), Proterozoic Lithospheric Evolution. Geodynamics Series17, American Geophysical Union.HOFFMAN, P. F., A. J. KAUFMAN, G. P. HALVERSON, AND D. P. SCHRAG. 1998. A Neoproterozoic Snowball Earth.Science, 281:1342–1346.HOFMANN, H. J. 1971. Polygonomorph acritarch from the Gunflint Formation (Precambrian), Ontario. Journal ofPaleontology, 45:522–524.HONJO, S., AND M. R. ROMAN. 1978. Marine copepod fecal pellets: production, preservation and sedimentation.Journal of Marine Research, 36:45–57.HOUSE, C. H., C. D. COATH, J. W. SCHOPF, T. M. HARRISON, AND K. D. MCKEEGAN. 2000. The carbon isotopiccomposition of individual Precambrian microfossils. First Astrobiology Science Conference, April 3–5, 2000.HUGHES, N. C. 2001. Ecologic evolution of Cambrian trilobites, p. 370–403. In A. Yu. Zhuravlev and R. Riding(eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.HUTCHINSON, G. E. 1961. The biologist poses some problems, p. 85–94. In M. Sears (ed.), Oceanography. AmericanAssociation for the Advancement of Science, Publication.HYDE, W. T., T. J. CROWLEY, S. K. BAUM, AND W. R. PELTIER. 2000. Neoproterozoic “snowball Earth” simulationswith a coupled climate/ice-sheet model. Nature, 405:425–429.HYDE, W. T., T. J. CROWLEY, S. K. BAUM, AND W. R. PELTIER. 2001. Life, geology and snowball Earth: Reply. Nature, 409:306.JAVAUX, E. J., A. H. KNOLL, AND M. R. WALTER. 2001. Morphological and ecological complexity in early eukaryoticecosystems. Nature, 412:66–69.JENSEN, S. 1990. Predation by early Cambrian trilobites on infaunal worms—evidence from the Swedish MickwitziaSandstone. Lethaia, 23:29–42.JENSEN, S. 1997. Trace fossils from the Lower Cambrian Mickwitzia sandstone, south-central Sweden. Fossilsand Strata, 42:1–111.KATZ, C. H. 1985. A nonequilibrium marine predator–prey interaction. Ecology, 66:1426–1438.KAUFFMAN, E. G., AND J. R. STEIDTMANN. 1981. Are these the oldest metazoan trace fossils? Journal of Paleontology,55:923–947.311

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002KAUFFMAN, S. A. 1989. Cambrian explosion and Permian quiescence: Implications of rugged fitness landscapes.Evolutionary Ecology, 3:274–281.KAUFMAN, A. J., A. H. KNOLL, AND S. M. AWRAMIK. 1992. Biostratigraphic and chemostratigraphic correlation ofNeoproterozoic sedimentary successions: Upper Tindir Group, northwestern Canada, as a test case. Geology,20:181–185.KELLEY, P. H., AND T. A. HANSEN. 1993. Evolution of the naticid gastropod predator–prey system: An evaluation ofthe hypothesis of escalation. Palaios, 8:358–375.KIRSCHVINK, J. L. 1992. Late Proterozoic low-latitude global glaciation: the Snowball Earth, p. 51–52. In J. W.Schopf and C. Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge UniversityPress, Cambridge.KIRSCHVINK, J. L., E. J. GAIDOS, L. E. BERTANI, N. J. BEUKES, J. GUTZMER, L. N. MAEPA, AND R. E. STEINBERGER.2000. Paleoproterozoic snowball Earth: Extreme climatic and geochemical global change and its biologicalconsequences. Proceedings of the National Academy of Sciences, 97:1400–1405.KIRSCHVINK, J. L., AND J. W. HAGADORN. 2000. A Grand Unified Theory of biomineralization, p. 139–150. In E.Bäuerlein (ed.), Biomineralization. Wiley–VCH, Weinheim.KITCHELL, J. A. 1983. Biotic interactions and siliceous marine phytoplankton: an ecological and evolutionaryperspective, p. 285–329. In M. J. S. Tevesz and P. L. McCall (eds.), Biotic Interactions in Recent and FossilsBenthic Communities. Plenum, New York.KNOLL, A. H. 1992. The early evolution of Eukaryotes: a geological perspective. Science, 256:622–627.KNOLL, A. H. 1994a. Neoproterozoic evolution and environmental change, p. 437–449. In S. Bengtson (ed.),Early Life on Earth. Columbia University Press, New York.KNOLL, A. H. 1994b. Proterozoic and Early Cambrian protists: evidence for accelerating evolutionary tempo.Proceedings of the National Academy of Sciences, USA, 91:6743–6750.KNOLL, A. H., AND S. B. CARROLL. 1999. Early animal evolution: emerging views from comparative biology andgeology. Science, 284:2129–2137.KOWALEWSKI, M., A. DULAI, AND F. T. FÜRSICH. 1998. A fossil record full of holes: The Phanerozoic history ofdrilling predation. Geology, 26:1091–1094.KRUSE, P. D., A. YU. ZHURAVLEV, AND N. P. JAMES. 1995. Primordial metazoan–calcimicrobial reefs: Lower Cambrian,Aldan river, Siberian Platform. Palaios, 10:291–321.LANDING, E. 1977. “Prooneotodus” tenuis (Müller, 1959) apparatuses from the Taconic allochthon, eastern New York:construction, taphonomy and the protoconodont “supertooth” model. Journal of Paleontology, 51:1072–1084.LANG, B. F., M. W. GRAY, AND G. BURGER. 1999. Mitochondrial genome evolution and the origin of eukaryotes.Annual Review of Genetics, 33:351–397.LEHOUEROU, H. N. 1996. The role of cacti (Opuntia spp.) in erosion control, land reclamation, rehabilitation andagricultural development in the Mediterranean basin. Journal of Arid Environments, 33:135–159.LEVINTON, J. S. 1982. Marine Ecology. Prentice-Hall, Englewood Cliffs, NJ, 526 p.LEVINTON, J. S. 2001. Genetics, Paleontology, and Macroevolution, 2nd Ed. Cambridge University Press,Cambridge, 617 p.LOGAN, G. A., J. M. HAYES, G. B. HIESHIMA, AND R. E. SUMMONS. 1995. Terminal Proterozoic reorganization ofbiogeochemical cycles. Nature, 376:53–56.LOGAN, G. A., R. E. SUMMONS, AND J. M. HAYES. 1997. An isotopic biogeochemical study of Neoproterozoic and earlyCambrian sediments from the Centralian Superbasin, Australia. Geochimica et Cosmochimica Acta, 61:5391–5409.LOWENSTAM, H. A., AND S. WEINER. 1989. On Biomineralization. Oxford University Press, New York, 324 p.LYNCH, M. 1999. The age and relationships of the major animal phyla. Evolution, 53:319–325.MACNAUGHTON, R. B., AND G. NARBONNE. 1999. Evolution and ecology of Neoproterozoic–Lower Cambrian tracefossils, NW Canada. Palaios, 14:97–115.MARGULIS, L. 1970. Origin of Eukaryotic Cells. Yale University Press, New Haven, CT.MARGULIS, L. 1981. Symbiosis in Cell Evolution. Freeman, San Francisco, CA, 419 p.MARIN, F., M. SMITH, Y. ISA, G. MUYZER, AND P. WESTBROEK. 1996. Skeletal matrices, muci, and the origin ofinvertebrate skeletons. Proceedings of the US National Academy of Sciences, 93:1554–1559.MARTIN, W., AND M. MÜLLER. 1998. The hydrogen hypothesis and the first eukaryote. Nature, 392:37–41.312

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONMARTINSSON, A. 1965. Aspects of a Middle Cambrian thanatotope on Öland. Geologiska Föreningens i StockholmFörhandlingar, 87:81–230.MATTHEW, G. F. 1912. The sudden appearance of the Cambrian fauna. Report of the Session of the InternationalGeological Congress, Stockholm, 1910:547–559.MAYNARD SMITH, J. 1978. The Evolution of Sex. Cambridge University Press, Cambridge, 242 p.MAYNARD SMITH, J., AND E. SZATHMÁRY. 1995. The Major Transitions in Evolution. Freeman, Oxford, 346 p.MCCAFFREY, M. A., J. M. MOLDOWAN, P. A. LIPTON, R. E. SUMMONS, K. E. PETERS, A. JEGANATHAN, AND D. S. WATT.1994. Paleoenvironmental implications of novel C30 steranes in Precambrian to Cenozoic age petroleumand bitumen. Geochimica et Cosmochimica Acta, 58:529–532.MCCLINTOCK, J. B. 1987. Investigation of the relationship between invertebrate predation and biochemicalcomposition, energy content, spicule armament, and toxicity of benthic sponges at McMurdo Sound,Antarctica. Marine Biology, 94:479–487.MCFADDEN, G. I., P. R. GILSON, C. J. B. HOFMANN, G. J. ADCOCK, AND U. G. MAIER. 1994. Evidence that an amoebaacquired a chloroplast by retaining part of an engulfed eukaryotic alga. Proceedings of the National Academyof Sciences, 91:3690–3694.MCILROY, D., AND G. A. LOGAN. 1999. The impact of bioturbation on infaunal ecology and evolution during theProterozoic–Cambrian transition. Palaios, 14:58–72.MCILROY, D., AND H. SZANIAWSKI. 2000. A lower Cambrian protoconodont apparatus from the Placentian ofsoutheastern Newfoundland. Lethaia, 33:95–102.MCMENAMIN, M. A. S. 1986. The garden of Ediacara. Palaios, 1:178–182.MCMENAMIN, M. A. S., AND D. L. SCHULTE MCMENAMIN. 1990. The Emergence of Animals. Columbia UniversityPress, New York, 217 p.MENDELSON, C. V., AND J. W. SCHOPF. 1992a. Proterozoic and Early Cambrian acritarchs, p. 219–232. In J. W. Schopfand C. Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.MENDELSON, C. V., AND J. W. SCHOPF. 1992b. Proterozoic and selected Early Cambrian microfossils and microfossillikeobjects, p. 865–951. In J. W. Schopf and C. Klein (eds.), The Proterozoic Biosphere: A MultidisciplinaryStudy. Cambridge University Press, Cambridge.MENGE, B. A. 1995. Indirect effects in marine rocky intertidal interaction webs: patterns and importance. EcologicalMonographs, 65:21–74.MILLER, R. H., AND F. A. SUNDBERG. 1984. Boring Late Cambrian organisms. Lethaia, 17:185–190.MISSARZHEVSKIJ, V. V. 1973. Konodontoobraznye organizmy iz pogranichnykh sloev kembriya i dokembriya Sibirskojplatformy i Kazakhstana [Conodont-shaped organisms from the Precambrian–Cambrian boundary beds of theSiberian Platform and Kazakhstan], p. 53–57. In I. T. Zhuravleva (ed.), Problemy paleontologii i biostratigrafiinizhnego kembriya Sibiri i Dal’nego vostoka. Trudy Instituta Geologii i Geofiziki SO AN SSSR. (In Russian.)MOCZYDLOWSKA, M. 1991. Acritarch biostratigraphy of the Lower Cambrian and the Precambrian–Cambrianboundary in southeastern Poland. Fossils and Strata, 29:1–127.MOREY, G. B., AND D. L. SOUTHWICK. 1995. Allostratigraphic relationships of Early Proterozoic iron-formations inthe Lake Superior region. Economic Geology, 90:1983–1993.MÜLLER, K. J., AND D. WALOSSEK. 1985. A remarkable arthropod fauna from the Upper Cambrian “Orsten” ofSweden. Transactions of the Royal Society of Edinburgh. Earth Sciences, 76:161–172.NEDIN, C. 1999. Anomalocaris predation on nonmineralized and mineralized trilobites. Geology, 27:987–990.NICOL, D. 1966. Cope’s Rule and Precambrian and Cambrian invertebrates. Journal of Paleontology, 40:1397–1399.NIELSEN, C. 1985. Animal phylogeny in the light of the trochea theory. Biological Journal of the Linnean Society,25:243–299.NIELSEN, C. 1995. Animal Evolution. Interrelationships of the Living Phyla. Oxford University Press, Oxford, 467 p.NIELSEN, C. 1998. Origin and evolution of animal life cycles. Biological Reviews, 73:125–155.NIKOH, N., N. IWABE, K. I. KUMA, M. OHNO, T. SUGIYAMA, Y. WATANABE, K. YASUI, Z. SHI-CUI, K. HORI, Y. SHIMURA,AND T. MIYATA. 1997. An estimate of divergence time of parazoa and eumetazoa and that of cephalochordataand vertebrata by aldolase and triose phosphate isomerase clocks. Journal of Molecular Evolution, 45:97–106.OSHEL, P. E., AND D. H. STEELE. 1985. Amphipod Paramphithoe hystrix: a micropredator on the sponge Haliclonaventilabrum. Marine Ecology Progress Series, 23:307–309.313

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002PETERS, S. E., AND M. FOOTE. 2001. Biodiversity in the Phanerozoic: a reinterpretation. Paleobiology, 27:583–601.PETERS, S. E., AND M. FOOTE. 2002. Determinants of extinction in the fossil record. Nature, 416:420–424.PETERSON, K. J., R. A. CAMERON, AND E. H. DAVIDSON. 1997. Set-aside cells in maximal indirect development:evolutionary and developmental significance. BioEssays, 19:623–631.PETERSON, K. J., AND E. H. DAVIDSON. 2000. Regulatory evolution and the origin of the bilaterians. Proceedings ofthe National Academy of Sciences, 97:4430–4433.PICKERILL, R. K., AND D. BLISSETT. 1999. A predatory Rusophycus burrow from the Cambrian of southern NewBrunswick, eastern Canada. Atlantic Geology, 35:179–183.PIERSON, B. K., J. BAULD, R. W. CASTENHOLZ, E. D’AMELIO, D. J. DES MARAIS, J. D. FARMER, J. P. GROTZINGER, B.BARKER JØRGENSEN, D. C. NELSON, AND A. C. PALMISANO. 1992. Modern mat-building microbial communities:a key to the interpretation of Proterozoic stromatolitic communities, p. 245–342. In J. W. Schopf and C.Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.POCOCK, K. J. 1974. A unique case of teratology in trilobite segmentation. Lethaia, 7:63–66.PORTER, K. G. 1977. The plant–animal interface in freshwater systems. American Scientist, 65:159–170.PORTER, S., AND A. H. KNOLL. 2000. Testate amoebae in the Neoproterozoic Era: evidence from vase-shapedmicrofossils in the Chuar Group, Grand Canyon. Paleobiology, 26:360–385.PRATT, B. R. 1982. Stromatolite decline—a reconsideration. Geology, 10:512–515.PRATT, B. R. 1998. Probable predation on Upper Cambrian trilobites and its relevance for the extinction of softbodiedBurgess Shale-type animals. Lethaia, 31:73–88.RASMUSSEN, B., S. BENGTSON, I. R. FLETCHER, AND N. MCNAUGHTON. 2002. Discoidal impressions and trace-likefossils more than 1200 million years old. Science, 296:1112–1115.RAUP, D. J. 1976a. Species diversity in the Phanerozoic: a tabulation. Paleobiology, 2:279–288.RAUP, D. J. 1976b. Species diversity in the Phanerozoic: an interpretation. Paleobiology, 2:289–297.RICHARDS, M. G., M. HUXHAM, AND A. BRYANT. 1999. Predation: a causal mechanism for variability in intertidalbivalve populations. Journal of Experimental Marine Biology and Ecology, 241:159–177.RIDING, R. 2000. Microbial carbonates: The geological record of calcified bacterial-algal mats and biofilms.Sedimentology, 47:179–214.RIEGER, R. M., AND W. STERRER. 1975. New spicular skeletons in Turbellaria, and the occurrence of spicules inmarine meiofauna. Zeitschrift für Zoologische Systematik und Evolutionsforschung, 13:207–278.ROGER, A. J. 1999. Reconstructing early events in eukaryotic evolution. American Naturalist, 154:146–163.RUNNEGAR, B. 1982. A molecular-clock date for the origin of the animal phyla. Lethaia, 15:199–205.RUNNEGAR, B. 1994. Proterozoic eukaryotes: evidence from biology and geology, p. 287–297. In S. Bengtson(ed.), Early Life on Earth. Nobel Symposium 84, Columbia University Press, New York.RUNNEGAR, B. 2000. Loophole for snowball Earth. Nature, 405:403–404.RYDELL, J., J. HAMMARLUND, AND A. SEILACHER. 2001. Trace fossil association in the Swedish Mickwitzia Sandstone(Lower Cambrian): Did trilobites really hunt for worms? Geologiska Föreningens i Stockholm Förhandlingar,123:247–250.SAMUELSSON, H., P. R. DAWES, AND G. VIDAL. 1999. Acid-resistant palynomorphs from the Proterozoic ThuleGroup, northwest Greenland. Precambrian Research, 96:1–23.SCHOENER, T. W., D. A. SPILLER, AND J. B. LOSOS. 2001. Predators increase the risk of catastrophic extinction ofprey populations. Nature, 412:183–186.SCHOPF, J. W. 1999. Cradle of Life: The Discovery of Earth’s Earliest Fossils. Princeton University Press,Princeton, NJ, 367 p.SCHOPF, J. W., B. N. HAUGH, R. E. MOLNAR, AND D. F. SATTERTHWAIT. 1973. On the development of metaphytes andmetazoans. Journal of Paleontology, 47:1–9.SCHOPF, J. W., AND D. Z. OEHLER. 1976. How old are the eukaryotes? Science, 193:47–49.SCHRAG, D. P., AND P. F. HOFFMAN. 2001. Life, geology and snowball Earth. Nature, 409:306.SCHUBERT, J. K., AND D. J. BOTTJER. 1992. Early Triassic stromatolites as post-mass extinction disaster forms.Geology, 20:883–886.SCUDO, F. M. 1996. Symbiosis, the origins of major life forms and systematics: A review with speculations. Memoriedella Societa Italiana di Scienze Naturali e del Museo Civico di Storia Naturale di Milano, 27:95–108.314

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONSEILACHER, A. 1956. Der Beginn des Kambriums als biologische Wende. Neues Jahrbuch für Geologie undPaläontologie, Abhandlungen, 103:155–180.SEILACHER, A., P. K. BOSE, AND F. PFLÜGER. 1998. Triploblastic animals more than 1 billion years ago: trace fossilevidence from India. Science, 282:80–83.SEMIKHATOV, M. A., AND M. E. RAABEN. 2000. Proterozoic stromatolite taxonomy and biostratigraphy, p. 295–306.In R. Riding and S. M. Awramik (eds.), Microbial Sediments. Springer, Heidelberg.SEPKOSKI, J. J., JR. 1996. Competition in macroevolution: The Double Wedge revisited, p. 211–255. In D. Jablonski,D. H. Erwin, and J. H. Lipps (eds.), Evolutionary Paleobiology. University of Chicago Press, Chicago.SHAPIRO, J. A. 1988. Bacteria as multicellular organisms. Scientific American, 258:82–89.SHIMKETS, L. J. 1990. Social and developmental biology of myxobacteria. Microbiological Reviews, 54:473–501.SIGNOR, P. W., AND G. J. VERMEIJ. 1994. The plankton and the benthos: origins and early history of an evolvingrelationship. Paleobiology, 20:297–319.SMETACEK, V. 2001. A watery arms race. Nature, 411:745.STANLEY, S. M. 1973a. An ecological theory for the sudden origin of multicellular life in the Late Precambrian.Proceedings of the U.S. National Academy of Sciences, 70:1486–1489.STANLEY, S. M. 1973b. An explanation for Cope’s rule. Evolution, 27:1–26.STANLEY, S. M. 1976a. Fossil data and the Precambrian–Cambrian evolutionary transition. American Journal ofScience, 276:56–76.STANLEY, S. M. 1976b. Ideas on the timing of metazoan diversification. Paleobiology, 2:209–219.STEELE, J. H. 1974. The Structure of Marine Ecosystems. Harvard University Press, Cambridge, MA, 128 p.STENECK, R. S., T. E. MILLER, R. P. REID, AND I. G. MACINTYRE. 1998. Ecological controls on stromatolite developmentin a modern reef environment: a test of the ecological refuge paradigm. Carbonates and Evaporites, 13:48–65.STRENG, M. 1999. Early Middle Cambrian representatives of the superfamily Acrotretoidea (Brachiopoda) fromMorocco. Zeitschrift der deutschen geologischen Gesellschaft, 150:27–87.SUMMONS, R., L. L. JAHNKE, J. M. HOPE, AND G. A. LOGAN. 1999. 2-Methylhopanoids as biomarkers forcyanobacterial oxygenic photosynthesis. Nature, 400:554–557.SZANIAWSKI, H. 1982. Chaetognath grasping spines recognized among Cambrian protoconodonts. Journal ofPaleontology, 56:806–810.SZANIAWSKI, H. 2002. New evidence for the protoconodont origin of chaetognaths. Acta Palaeontologica Polonica,in press.TELFORD, M. J., AND P. W. H. HOLLAND. 1993. The phylogenetic affinities of the chaetognaths: a molecular analysis.Molecular Biology and Evolution, 10:660–676.TELFORD, M. J., AND P. W. H. HOLLAND. 1997. Evolution of 28S ribosomal DNA in chaetognaths: duplicate genesand molecular phylogeny. Journal of Molecular Evolution, 44:135–144.THEIMER, T. C., AND G. C. BATEMAN. 1992. Patterns of prickly-pear herbivory by collared peccaries. Journal ofWildlife Management, 56:234–240.THOMPSON, J. N. 1998. Rapid evolution as an ecological process. Trends in Ecology and Evolution, 13:329–332.VALENTINE, J. W. 1973. Evolutionary paleoecology of the marine biosphere. Prentice Hall, Englewood Cliffs, NJ, 511 p.VALENTINE, J. W., D. JABLONSKI, AND D. H. ERWIN. 1999. Fossils, molecules and embryos: new perspectives on theCambrian explosion. Development, 126:851–859.VERITY, P. G., AND V. SMETACEK. 1996. Organism life cycles, predation, and the structure of marine pelagicecosystems. Marine Ecology Progress Series, 130:277–293.VERMEIJ, G. J. 1987. Evolution and Escalation. An Ecological History of Life. Princeton University Press, Princeton,NJ, 527 p.VERMEIJ, G. J. 1990. The origin of skeletons. Palaios, 4:585–589.VERMEIJ, G. J. 1994. The evolutionary interaction among species: selection, escalation and coevolution. AnnualReview of Ecology and Systematics, 25:219–236.VERMEIJ, G. J. 2002. Evolution in the consumer age: Predators and the history of life. In M. Kowalewski and P. H.Kelley (eds.), The Fossil Record of Predation, Paleontological Society Special Papers, 8 (this volume).VIDAL, G., AND M. MOCZYDLOWSKA. 1992. Patterns of radiation in the phytoplankton across the Precambrian–Cambrian boundary. Journal of the Geological Society, London, 149:647–654.315

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002VIDAL, G., AND M. MOCZYDLOWSKA-VIDAL. 1997. Biodiversity, speciation, and extinction trends of Proterozoicand Cambrian phytoplankton. Paleobiology, 23:230–246.VREELAND, H. V., AND H. R. LASKER. 1989. Selective feeding of the polychaete Hermodice carunculata Pallas onCaribbean gorgonians. Journal of Experimental Marine Biology and Ecology, 129:265–277.WADA, H., AND N. SATOH. 1994. Details of the evolutionary history from invertebrates to vertebrates, as deducedfrom the sequences of 18S rDNA. Proceedings of the National Academy of Sciences, USA, 91:1801–1804.WADDELL, B., AND J. R. PAWLIK. 2000a. Defenses of Caribbean sponges against invertebrate predators, I: Assayswith hermit crabs. Marine Ecology Progress Series, 195:125–132.WADDELL, B., AND J. R. PAWLIK. 2000b. Defenses of Caribbean sponges against invertebrate predators, II: Assayswith sea stars. Marine Ecology Progress Series, 195:133–144.WAINWRIGHT, S. A., W. D. BIGGS, J. D. CURREY, AND J. M. GOSLINE. 1976. Mechanical Design in Organisms.Edward Arnold, London, 423 p.WALCOTT, C. D. 1920. Cambrian geology and paleontology IV:6—Middle Cambrian Spongiae. SmithsonianMiscellaneous Collections, 67:261–364.WALTER, M. R. 1994. Stromatolites: The main geological source of information on the evolution of the earlybenthos, p. 270–286. In S. Bengtson (ed.), Early Life on Earth. Columbia University Press, New York.WALTER, M. R., J. BAULD, D. DES MARAIS, AND J. W. SCHOPF. 1992a. A general comparison of microbial mats andmicrobial stromatolites: Bridging the gap between the modern and the fossil, p. 335–338. In J. W. Schopf andC. Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.WALTER, M. R., J. P. GROTZINGER, AND J. W. SCHOPF. 1992b. Proterozoic stromatolites, p. 253–260. In J. W. Schopf andC. Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.WALTER, M. R., AND G. R. HEYS. 1985. Links between the rise of the Metazoa and the decline of the stromatolites.Precambrian Research, 29:149–174.WALTHER, B. T. 2000. Do life’s three domains mirror the origins of sex? Journal of Biosciences, 25:17–20.WANG, D. Y. C., S. KUMAR, AND S. B. HEDGES. 1999. Divergence time estimates for the early history of animal phylaand the origin of plants, animals and fungi. Proceedings of the Royal Society of London B, 266:163–171.WARD, D. M., J. BAULD, R. W. CASTENHOLZ, AND B. K. PIERSON. 1992. Modern phototrophic microbial mats: anoxygenic,intermittently oxygenic/anoxygenic, thermal, eukaryotic and terrestrial, p. 309–324. In J. W. Schopf and C.Klein (eds.), The Proterozoic Biosphere: A Multidisciplinary Study. Cambridge University Press, Cambridge.WEST, J. M. 1997. Plasticity in the sclerites of a gorgonian coral: tests of water motion, light level, and damagecues. Biological Bulletin, 192:279–289.WEST, J. M. 1998. The dual role of sclerites in a gorgonian coral: conflicting functions of support and defence.Evolutionary Ecology, 12:803–821.WHALEN, M. T., G. P. EBERLI, AND P. W. HOMEWOOD. 1998. Microbial carbonates as indicators of environmentalchange and biotic crises in carbonate systems: examples from the Upper Devonian, Alberta Basin, Canada.Geological Society of America Abstracts with Programs, 30:195.WHITTINGTON, H. B., AND D. E. G. BRIGGS. 1982. A new conundrum from the Middle Cambrian Burgess Shale.Third North American Paleontological Convention, Proceedings, 2:573–574.WILLIAMS, G. C. 1975. Sex and Evolution. Princeton University Press, Princeton, NJ, 210 p.WILSON, W. H. 1990. Competition and predation in marine soft-sediment communities. Annual Review of Ecologyand Systematics, 21:221–241.WOODIN, S. A. 1983. Biotic interactions in Recent marine sedimentary environments, p. 3–38. In M. J. S. Tevesz andP. L. McCall (eds.), Biotic Interactions in Recent and Fossils Benthic Communities. Plenum, New York.WRAY, G. A., J. S. LEVINTON, AND L. H. SHAPIRO. 1996. Molecular evidence for deep pre-Cambrian divergencesamong metazoan phyla. Science, 274:568–573.XIAO, S., AND A. KNOLL. 2000. Phosphatized animal embryos from the Neoproterozoic Doushantuo Formation atWeng’an, Guizhou, south China. Journal of Paleontology, 74:767–788.XIAO S., Y. ZHANG AND A. KNOLL. 1998. Three-dimensional preservation of algae and animal embryos in aNeoproterozoic phosphorite. Nature, 391:553–558.YUE Z., AND S. BENGTSON. 1999. Embryonic and post-embryonic development of the Early Cambrian cnidarianOlivooides. Lethaia, 32:181–195.316

BENGTSON—ORIGINS AND EARLY EVOLUTION OF PREDATIONZANG W., AND M. R. WALTER. 1989. Latest Proterozoic plankton from the Amadeus Basin in Central Australia.Nature, 337:642–645.ZANG W., AND M. R. WALTER. 1992. Late Proterozoic and Cambrian microfossils and biostratigraphy, AmadeusBasin, central Australia. Memoirs of the Association of Australasian Palaeontologists, 12:1–132.ZHANG, Y., L. YIN, S. XIAO, AND A. H. KNOLL. 1998. Permineralized fossils from the terminal Proterozoic DoushantuoFormation, south China. Journal of Paleontology, 72:1–52.ZHANG X.-G., AND B. PRATT. 1994. Middle Cambrian arthropod embryos with blastomeres. Science, 266:637–639.ZHURAVLEV, A. YU. 2001. Biotic diversity and structure during the Neoproterozoic–Ordovician transition, p. 173–199. In A.Yu. Zhuravlev and R. Riding (eds.), The Ecology of the Cambrian Radiation. Columbia University Press, New York.ZHURAVLEV, A. YU., AND F. DEBRENNE. 1996. Pattern of evolution of Cambrian benthic communities: Environmentsof carbonate sedimentation. Rivista Italiana di Paleontologia e Stratigrafia, 102:333–340.ZHURAVLEV, A. YU., AND R. RIDING (eds.). 2001. The Ecology of the Cambrian Radiation: Perspectives inPaleobiology and Earth History. Columbia University Press, New York., 525 p.ZHU S., S. SUN, X. HUANG, Y. HE, G. ZHU, L. SUN, AND K. ZHANG. 2000. Discovery of carbonaceous compressionsand their multicellular tissues from the Changzhougou Formation (1800 Ma) in the Yanshan Range, NorthChina. Chinese Science Bulletin, 45:841–847.317


BAMBACH—SUPPORTING PREDATORSSUPPORTING PREDATORS: CHANGES IN THEGLOBAL ECOSYSTEM INFERRED FROM CHANGES INPREDATOR DIVERSITYRICHARD K. BAMBACHProfessor Emeritus, Virginia Tech;Botanical Museum, Harvard University, 26 Oxford Street, Cambridge, MA 02138 USAABSTRACT—This paper presents new estimates of the genus diversity of predators in each major taxon containingpredators, as well as an estimate of the total genus diversity of predators through the Phanerozoic. Predatorshave never been numerically abundant compared to prey. However, the diversity of predators and the proportionof total faunal diversity composed of predators have both increased over time, implying that ecosystems haveincreased their ability to support either more predators or more specialization among predators. Also, turnoverin diversity dominance among predator groups, with more energetic predator taxa replacing or being added to afauna of less energetic groups, implies that the energy available in marine food webs has increased. The apparentincrease in diversity and biomass of primary producers plus patterns of diversity change in prey taxa supportsthese inferences based on patterns of change in predators alone.INTRODUCTIONTHIS PAPER COMPILES estimates of theglobal genus diversity of marine predators aspreserved in the fossil record through thePhanerozoic. The global picture is the sum of localand regional distributions. No local ecosystem needhave the same relative pattern of either taxonrichness or abundance as the global whole. Butchanges in global patterns can only occur throughthe accumulation of changes in local or regionalpatterns. Therefore, determining pattern, andchange in pattern, at the global level not only givesus information on the history of life in sum, a notuninteresting subject (consider, for example, theamount of attention given mass extinctions), butalso identifies times when change must have beenoccurring at the local or regional level. Globalanalyses may not reveal the detailed story ofchanges in organism-organism interactions in eachlocal setting, but global analyses provide invaluableindices of when local changes must have beenoccurring. They serve as a guide to the timeintervals for which targeted research at local orregional levels could have large dividends.Defining predator.—What do we mean by theterm predator? The American Heritage Dictionaryof the English Language, Third Edition(Soukhanov, 1992) defines predator as “anorganism that lives by preying on other organisms.”Even though preying can be interpreted to meancatching and killing, the definition as stated is toobroad for the sense commonly used by biologists(and used in this paper) because it would includeherbivores, parasites, and suspension feeders aswell as deposit-feeders that gain sustenance frommicrobes. Nor is predation, as used here, simplysynonymous with carnivory (which includesscavenging and some forms of suspension-feeding,such as the feeding of baleen whales). Predation(as used here) and these other trophic systems mayall be mathematically roughly equivalent activitiesin terms of energy acquisition or in their effects onthe populations of the organisms preyed upon, butthe other activities don’t catch the sense ofspecifying those organisms that hunt or trap,subdue, and kill individual animals that have somecapacity for either protection or escape. This latterrestricted categorization is what I mean by the termpredator. Organisms that “graze” on non-motile319

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002colonial animals in a manner similar to the grazingof herbivores on vegetation, suspension-feedingand deposit-feeding animals that may consume tinyplanktonic or interstitial animals as part of theirintake of food, but do not pursue animals asindividual targets, and parasites that may weakentheir hosts, but do not actively subdue and kill themin order to gain their nutrition, are not regarded aspredators in the sense meant by the definition usedhere. Predators, then, are secondary consumers—heterotrophs higher in the food chain than primaryproducers (autotrophs), primary consumers(herbivores), “indeterminate” level consumers(deposit feeders, consumers of microbes anduseable dead organic material from many sources,and suspension-feeders, organisms that also gainsustenance from mixed low levels of the foodchain—phytoplankton and/or animal plankton), orcarnivores that are either purely scavengers or eatentirely passive prey that is as available asvegetation is to herbivores.One implication of restricting the definition ofpredators to metazoans catching other metazoansis that the nature of the food pyramid restricts theabundance of predators. While the relationships arevery complex (May, 1981; Pollis and Winemiller,1996), the biomass of predators is alwaysconsiderably less than that of prey. Although acommon “rule of thumb” is that each level of afood pyramid contains only about ten percent ofthe biomass of the next lowest level, therelationship is, in fact, highly variable. Calculatedvalues of food-chain efficiencies vary from lessthan 0.1% to over 10% (May, 1981).In this paper I use the unpublished database ofthe stratigraphic ranges of genera of marine animalscompiled by the late J. John Sepkoski, Jr. to estimatethe genus diversity of marine metazoan predatorsboth as a group and within many of the major taxathat contain predators. The data are necessarilylimited to taxon richness only, not abundance, so nodiversity metrics that account for the interaction ofabundance and taxon richness are used. But taxonrichness alone does have implications for predatorpreyrelationships. Because the abundance ofpredators must be less than prey (from a factor of0.3 to less than 0.001), we can extrapolate that,unless contravening major secular changes occur inthe average abundance of individual predatorspecies, changes in diversity of predator taxaprobably reflect changes in the abundance of preytaxa. It would be unlikely, at least, that predatordiversity would increase in the face of decreasingbiomass (abundance) of prey. This will be discussedin more detail in later parts of the paper.The groups designated in this paper as prey taxaare restricted to non-predators only. Some predatorsprey on other predators. Although those preyed-uponpredators are themselves prey, they are not countedin the prey groups in the examples in this paper. Butthe inclusion of some predators as prey would extendfood chains and expand the trophic pyramid forpredators even more than in the case consideredhere, only adding to the increase, documented inthis paper, of non-predatory prey alone—thus furthersupporting, not countering, most of the conclusionsof the paper.Tracking Genus Diversity.—Figure 1 illustratesthe history of genus diversity of metazoa throughthe Phanerozoic. Because the metazoa compriseover 90% of the taxonomic richness tabulated incomprehensive synoptic databases on marinediversity, the metazoan diversity path follows thewell-known pattern established in the “consensuspaper” of Sepkoski et al. (1981) and best knownfrom the widely reproduced family diversity curve(Sepkoski, 1981). In the Paleozoic we see the“Cambrian Explosion” (the increase in diversityin the Early Cambrian), the Cambrian “Plateau,”the Ordovician Radiation, and the long interval(starting in the Caradocian) of fluctuating but nontrendingdiversity that lasts until the Late Permian.Diversity changes during the “Paleozoic Plateau”include the end-Ordovician mass extinction,diversity buildup through the Silurian and EarlyDevonian to a Mid-Devonian peak, followed by along slide in diversity through much of the Middleas well as the entire Late Devonian, recovery ofdiversity in the early part of the Carboniferous andthen fairly stable diversity through the rest of theCarboniferous and the first half of the Permian.The Paleozoic ends with the precipitous drop of320

BAMBACH—SUPPORTING PREDATORSFIGURE 1—Genus diversity of marine metazoa through the Phanerozoic. The heavy solid line connectsthe data on number of genera crossing from one time interval to the next. The zig-zag made by thedotted light line represents the turnover in genus diversity within each time interval. As explained in thetext, the figure represents all the data available on diversity in the Sepkoski compilation.diversity associated with the Guadalupian and end-Permian extinctions. The post-Paleozoic ischaracterized by nearly continuous diversityincrease, interrupted by end-Triassic, end-Jurassic,and late Eocene diversity decreases and the sharpera-bounding end-Cretaceous mass extinction. Inanother paper I showed that the post-Paleozoicincrease in diversity fits an exponential model withvery high fidelity (Bambach, 1999).There are some concerns that the large Cenozoicincrease in diversity may be partly artifactual. Rockvolume differences (Raup, 1976; Peters and Foote,2002), the Pull of the Recent (Raup 1979), and alack of standardized sampling in compiling mostsynoptic databases (Alroy et al., 2001) could allinflate estimates of diversity in the near, as comparedto the distant, past. However, alpha diversityincreased between the Paleozoic and later Cenozoic(Bambach, 1977; Powell and Kowalewski, 2001).Also, the number of comparably defined provinceshas not changed much (Bambach, 1990), implyingthat beta diversity has not decreased. Combined,these facts argue strongly that some of the observedCenozoic increase in diversity is real. But moreimportantly, all taxa should be affected similarly bythese potential biases and it is comparisons amongtaxa, not the absolute numbers of taxa as such, thatare the subject of this paper. The changes in diversitydominance through time and the changes inproportional importance at the heart of the analysesin this paper are generally independent of thesepossible biases.Figure 1 accounts for all the data in theSepkoski compilation. The heavy line connects thedata on number of genera crossing each intervalboundary (the number leaving one interval andentering the succeeding interval, calculated as thetotal number of genera in the first interval minus321

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002the number of genus extinctions in that interval,leaving the rest to go on to the next interval) andfollows the path of minimum likely standingdiversity. This is regarded as the minimum diversitybecause in any interval it is possible that originationmight increase diversity somewhat before newextinctions occur and this higher, but unknown,diversity would probably persist throughout aninterval if some of the extinction in the interval isconcentrated as an extinction event at or near theend of the interval. The peaked dotted linerepresents the genus turnover within each interval.The rising part of each peak represents all the genusoriginations reported from the interval. The peak isat the total number of genera reported from the timeinterval. This total includes the genera that areconfined to the single time interval as well as thosethat occur in multiple intervals, whereas the data ongenera crossing from one interval to the next doesnot include short-lived genera confined to singleintervals. The descending part of the peak representsthe number of extinctions (last records) of genera inthe interval. Many diversity curves use total diversityas the recorded data (this is true for publishedillustrations of Sepkoski’s family curve, forexample). If one mentally “connects the dots” oftotal diversity peaks in Figure 1, it is clear that thegeneral shape of a curve connecting total diversitieswould be very similar to the boundary-crossingdiversity emphasized here. Mid-Devonian diversitywould appear higher than Late Ordovician diversity,Carboniferous diversity would fluctuate more, andthe upturn in diversity in the mid-Cretaceous wouldbe sharper than shown by the standing diversity plot,but the general pattern would be the same.Since total diversity (the number of genera ineach interval at the peak of the dotted plot in Fig.1)would never actually be achieved at any time(unless all originations occurred in an intervalbefore any extinctions), whereas the recordeddiversity crossing from one interval to the next isa direct measure of standing diversity at intervalboundaries, I use the boundary-crossing standingdiversity as the preferred representation of diversityand diversity change through time. Boundarycrossingdiversity not only approximates standingdiversity, but changes in boundary-crossingdiversity give a clearer representation of theconsequences of within-interval evolutionarydynamics than do summed total diversity data forwhole intervals. Admittedly, boundary-crossingdiversity is a minimum diversity—for the reasonsnoted above, and because there are likely to becases of genera actually surviving into a newinterval but never (or at least not yet) collected inthe later interval, and also because boundarycrossingdiversity represents the diversity thatfollows immediately after any extinction event thatmay close an interval—but three factors make itmy choice. One is that it follows the general pathof total diversity, so we know that it is not showingus anomalous diversity compared to the total. Thesecond is that it is the only measure we have of anactual standing diversity level. Thirdly, changingdiversity is shown best by comparing diversity atthe start of one interval to that at start of another.Diversity is a function of both origination andextinction. If no originations or extinctions occurin an interval then diversity at the end of the intervalwill be the same as at the start. The same would betrue if every genus that entered the interval wentextinct but that number of genera also originatedduring the interval. Following standing diversityat interval boundaries tells us whether originationand extinction are in balance (little or no changeof diversity from one boundary to the next) orwhether either origination or extinction dominatedduring an interval (origination dominating ifboundary-crossing diversity increases, extinctionbeing more important if boundary-crossingdiversity decreases). If one just follows totaldiversity an extinction event in one interval couldbe masked by an increase in origination in the next.Total diversity could be unchanged between the twointervals because of rapid recovery from anextinction event. By using boundary-crossingstanding diversity, if there is more extinction thanorigination during an interval the standing diversityleaving the interval will be lower than that enteringthe interval. If there is no change in diversity betweenthe starting and ending boundary diversities, thenorigination during the interval matched extinction322

BAMBACH—SUPPORTING PREDATORSTABLE 1—Proportions of Genera Used in Calculating Numbers of Predator GeneraPhylum Class Subclass taxonMinimumproportionof predatorsMaximumproportionof predatorsCnidariaAnthozoa 0.10 0.50AnnelidaPolychaetaEunicemorpha all allPhyllodocemorpha all allProblematica 0.05 0.25MolluscaArthropodaEchinodermataChordataGastropodaProsobranchiaArchaegastropoda 0.00 0.10Mesogastropoda 0.12 0.35Neogastropoda all allOpisthobranchia 0.90 allCephalopoda all allIncertae sedis 0.50 0.90Trilobita 0.10 0.40MerostomataEurypterida all allMalacostraca 0.50 0.90Asteroidea all allOphiuroidea 0.05 0.15EchinoideaEchinacea 0.05 0.15Conodontophorida 0.75 allPlacodermi all allAcanthodii all allChondrichthyes all allOsteichthyes 0.65 0.90Reptilia all allMammalia 0.85 0.95323

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002and diversity overall remained stable, even ifextinction was concentrated in short-term eventsor an event at the end of the interval.Estimating Overall Diversity of Predators.—Now let’s attempt to determine the diversity ofpredators through the Phanerozoic. Several factorsneed to be considered in estimating overall predatordiversity. The data on genera in the database aregrouped at the order level and above, but not allmembers of every higher taxon are exclusivelypredators or not predators. Not only that, buttrophic activity is not known for many seldomencounteredtaxa. I ignored the rare exceptions ingroups whose natural history is well known andthat are overwhelmingly predators or not predators(whale sharks are functionally filter-feeders, but theyare a minuscule portion of chondrichthyan genera)and designated those groups as entirely predatorsor non-predators. For groups that have a mixture ofpredators and non-predator members I determinedfrom descriptions of feeding habits and diet in theliterature (Parker, 1982 and other references as notedbelow in notes on individual taxa) approximatelyhow many taxa in each group seem to bepredominantly predators and, since these numberswere usually at the family level and some were“educated guesses,” I calculated a reasonable rangeof the proportion of taxa in each group that mightbe considered predators. This range is my estimateof what might be thought of as a confidence intervalconcerning the completeness and reliability of thedata available to a single experienced worker. I hopeit is clear that it frames a range of estimation; thebounds are not precisely determined. Future work,especially thoughtful compilations by specialists,will undoubtedly improve our knowledge. Theminimum proportion is the proportion of the taxain the group certain to be predators; the maximumproportion is the maximum proportion of the taxain the group that reasonably might be predators.Some more information on criteria is given withthe discussions of each major group below.Table 1 lists the taxa and the minimum andmaximum proportions chosen in estimating thediversity of predators. The proportions were thenused to determine the number of genera in eachgroup likely to have been predators in each timeinterval. Both maximum and minimum numberswere calculated, plus turnover. All groups weresummed together to get the total diversity range.The range of boundary-crossing diversity plus theturnover for the maximum reasonable diversity forall groups together are shown in Figure 2. Toestimate the average diversity of predators, theFIGURE 2—The range in number of boundary-crossing genera for estimated minimum and maximumnumbers of predators through the Phanerozoic. Turnover in the estimated maximum number of predatorsis also shown (the turnover for the minimum number would always fall inside that of the maximum).324

BAMBACH—SUPPORTING PREDATORSmean of the maximum and minimum range ofdiversity of predators was calculated, along withthe turnover appropriate for the boundary-crossingdiversities, and they are shown in Figure 3a.At this time this is the best estimate we can makeof the genus diversity of predators represented inthe fossil record. Because of uncertainty about howmany genera in many groups are actually predatorsthis estimate is necessarily a compromise, greaterthan the minimum possible number but less thanthe maximum reasonable number. However, it islikely to be closer to the actual value than either themaximum or minimum estimates. The only otherfactor that could alter the pattern of diversity ofpredators would be secular change in the proportionof predators within those groups that have a mixof predator and non-predator taxa. An example atthe class level will be shown below for gastropods,where we can follow change in the proportion ofdiversity contributed by different orders, each witha different proportion of predator genera. Althoughsuch changes in proportion did occur within someor all groups, this would not alter the generalpatterns, even if the change in each group was fromits minimum to maximum proportion as calculatedhere.The diversity of non-predator metazoan genera(calculated simply by subtracting the data onestimated predator diversity from the total of allmetazoa) is shown in Figure 3b at the same scaleas Figure 3a for predators. As expected, nonpredatorsconsistently outnumber predator taxa.The relationship between them will be discussedin the concluding section of this paper.DIVERSITY HISTORIES OFPREDATOR GROUPSBecause constant proportions were used inestimating maximum and minimum diversity ofpredators within each taxon, only the average(mean) of the range of estimated diversities ofpredators was used in the following examples,except for the anthozoa. This mean, or average, isnecessarily parallel to the maximum and minimumpossible numbers of predator genera, but midwaybetween them, as is illustrated by the data on theanthozoa (Fig. 4). Unless a significant shift in theproportion of taxa having predatory habits occurredwithin a taxon during the Phanerozoic, the patternof actual diversity change in predators should berecorded reliably by the average estimateddiversity. If the proportion of predator taxa changedin the same direction as overall diversity (forexample, if both increased), then the real changewould be exaggerated compared to the pattern seenusing a constant proportionality. Only if theproportion of predators in a taxon changed inopposition to the change in overall diversity of thetaxon, and at a greater rate than the change inoverall diversity (both unlikely), would the generaltrend of predator diversity in a taxon be obscuredor misrepresented by the method used here.Anthozoa.—Although anthozoans aresedentary organisms, many use their nematocystsand tentacles to entrap animals (Parker, 1982;Hyman, 1940). Some even extend parts of theseptal filaments out through the mouth and begindigestion externally. However, hermatypic reefbuildingcorals gain most nutrition fromphotosymbionts (zooxanthellae), and many smallpolypanthozoans use mucus and cilliary currentsin suspension-feeding, so many anthozoans are notpredators. But long-tentacled and large-polypanthozoa must be regarded as predators eventhough sedentary. The large polyps of manyrugosans and some tabulates, plus doubts thatPaleozoic corals were hermatypic, suggest thatmany of them had similar “passive-predator” lifehabits to those of living predatory anthozoans.From a survey of the distribution of polyp sizes,between ten and fifty percent of anthozoan generacan be regarded as predators. Figure 4 shows therange in number of genera of anthozoa that couldhave been predators through the Phanerozoic.Although there is a wide range of possible diversityof predatory anthozoans, three features are clear:1) the slow recovery from the Permian extinction,which eliminated the Tabulata and Rugosa, createda low-diversity interval from the start of the Triassicto the mid-Jurassic; 2) Cretaceous and Cenozoicdiversity is not dramatically higher than Devonian325

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 3—Diversity of predators and prey taxa. (a) Estimated genus diversity of predators (the meanof the maximum and minimum diversity estimates; used as the total number of predators for the analysesin this paper). Turnover (based on the origination and extinction proportions of the groups that includepredators and calculated from the estimated mean boundary-crossing diversities of predators) is alsoshown. (b) Estimated diversity of non-predator metazoa (determined by subtracting the estimateddiversity of predators from the total diversity shown in Fig. 1). Turnover is also shown.326

BAMBACH—SUPPORTING PREDATORSFIGURE 4—Maximum, minimum, and mean boundary-crossing diversity estimates for predatoryanthozoans through the Phanerozoic.diversity for predatory anthozoans (and if a higherproportion of tabulates and rugosans were predatorythan for scleractinians, it could even be lower); and3) diversity of predatory anthozoans in the Cenozoicis not markedly greater than in the Late Jurassic orCretaceous, a different pattern from the diversitytrend for either Metazoa or predators overall.The reason the range of estimated diversity ofpredators is shown for the anthozoa (Fig. 4) isbecause the uncertainty of recognizing predators ismaximal for this group. Also, because there iscomplete replacement in the corals at the order levelbetween the Paleozoic (Rugosa and Tabulata) andMesozoic-Cenozoic (Scleractinia), it is possible thatthe proportion of predatory coralline anthozoa (thevast majority of the genera recorded in the fossilrecord) was markedly different in the Paleozoicversus the Mesozoic-Cenozoic. In fact, the apparentevolution of hermatypy in the scleractinia makesthis a case in which a decrease in proportion ofpredator taxa between the Paleozoic and Mesozoic-Cenozoic, rather than an increase, is likely.Polychaetes.—Polychaetes are soft-bodied andhave a poor fossil record. Scolecodonts (phosphatictooth-like structures) are known and have beendescribed, but lägerstetten with soft-partpreservation are also an important source of dataon these infrequently fossilized organisms.Figure5 displays the diversity data known forpredatory (not suspension-feeding) polychaetes.The very stable form of the boundary-crossing(continuing) diversity plus the widely spaced sharpturnover peaks indicates that much of the data are“range-through” from widely spaced occurrences,apparently from lägerstetten. A very strikingfeature, however, is that the maximum diversitypeaks, which are widely spaced and primarilyrepresent “monographic effects” from lägerstetten,form a near-parallel diversity path if the points areconnected (see the dashed line in Fig. 5). It appearsthat the diversity signal is stronger than the biasesof the record and can show through even in a poor327

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 5—Diversity of predatory polychaetes. The dashed line connecting the peaks of maximumdiversity within intervals shows that the diversity revealed by superior preservation in lägerstettenfollows the same general path as the diversity of boundary-crossing genera.fossil record (the records of other taxa to be notedlater also support this idea). The implication is thatpredatory polychaetes became a part of the predatorfauna in the later Ordovician and have maintainedbut not increased their diversity through the restof the Phanerozoic.Gastropods.—Gastropods include manyimportant predators. Several of these taxa makedistinctive drill holes that can be identified in thefossil record. Figure 6a shows the diversity historyof predatory gastropods. From the Ordovician to themid-Cretaceous the diversity of predatorygastropods increased nearly continuously, but verygradually. Since the mid-Cretaceous the diversityof gastropod predators has increased dramatically.This contrasts markedly with the diversity historiesof predator anthozoans and polychaetes.Archaeogastropoda were the dominant Paleozoicprotobranch gastropods and few of them are certainpredators. However, a higher proportion ofMesogastropoda are predators and all theNeogastropoda and most of the readily preservedOpisthobranchia are predators. As shown inFigure 6b, the predatory opisthobranchs increasedin diversity in the Jurassic and again in the Cenozoic;the Mesogastropoda expanded their diversitysomewhat in the Triassic and then more or lessmonotonically from the Early Cretaceous onward;and the Neogastropoda have had an explosivediversification that began in the Late Cretaceous.Taxa containing a higher proportion ofpredators have diversified sequentially, so theproportion of all gastropods that are predators haschanged through time. Figure 7a shows thatpredators were a small part of the total diversity ofPaleozoic gastropods, but increased during theMesozoic to nearly equal the diversity of nonpredatorsin the Late Cretaceous before becomingthe majority of gastropod genera in the Cenozoic.Figure 7b tracks the change in proportion ofgastropod diversity comprised of predators throughthe whole Phanerozoic. Even though the methodused to estimate diversity of predators uses aconstant proportionality for each taxon analyzed(meaning that the proportion of predators within asingle analyzed taxon remains constant and doesnot change over time), the changing mix of the fourdifferent gastropod taxonomic groups used in thisanalysis demonstrates the continuous increase inthe proportion of predatory gastropods from theOrdovician onward. It is interesting to note thatpredatory strategies were added sequentially ingastropods. Drilling, although it had occurred at lowfrequency since the early Paleozoic, first became328

BAMBACH—SUPPORTING PREDATORSabFIGURE 6—Gastropod predators. (a) Estimated genus diversity of predatory gastropods. Note that genusdiversity built up very gradually, but nearly continuously, from the Ordovician to the mid-Cretaceous andhas increased dramatically in the last 100 million years. Note also that turnover peaks are very low beforethe Late Cretaceous. (b) Genus diversity of gastropod predators subdivided into constituent major taxa.329

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 7—Comparison of predatory with non-predatory gastropods. (a) Comparison of genus diversityof all gastropods with the genus diversity of predatory gastropods. (b) Proportion of genus diversity ofgastropods comprised of predators.330

BAMBACH—SUPPORTING PREDATORScommon in some mesogastropod groups (naticids).More “aggressive” strategies effective against moreactive prey, such as “harpooning” prey withmodified radular teeth and injecting poison, aspracticed by cone shells, or using shell features toaid in actively prying open the shells of prey, onlybecame common in the Neogastropoda (Hyman,1967). The Late Cretaceous burst of diversificationof Neogastropods is one of the reasons the MesozoicMarine Revolution noted by Vermeij (1977) isexpressed most strongly in the later Cretaceous.The limitations of the database, which includesno information on relative (or absolute) abundance,must again be emphasized. The increase indiversity of predatory gastropods to the point atwhich they are more diverse than non-predatorsmust say something about the ability of gastropodpredators to specialize, thus adding diversity asprey biomass can support it, but it says nothingabout the abundance of predators per se, nor aboutthe number of predator individuals compared toprey. Tabulations, such as those made by GwenDaley (Daley, 1999; Daley and Kowalewski, 2000)indicate that in Cenozoic deposits most predatorygastropod species are rare compared to suspensionfeeding forms and bivalves.Cephalopods.—The diversity of cephalopods,an exclusively predatory class, is remarkable in nottrending up or down from the later Ordovicianthrough the Cretaceous despite the great volatilityof the group as a whole (Fig. 8a). This volatility isapparent in the fluctuations of continuing standingdiversity coupled with the high rate of withinintervalturnover. Total diversity of cephalopodswithin intervals was often more than twice theboundary-crossing standing diversity. The rapidturnover implies rapid evolution, which is attestedby the utility of cephalopod fossils forbiostratigraphy. Figure 8b shows the shift over timein diversity dominance from all nautiloids (varioustetrabranch groups) in the Ordovician, to aboutequal diversity of ammonoids (phylogeneticallyallied with the dibranch coleoids; Engeser, 1990,1996; Jacobs and Landman, 1993) and nautiloidsin the Late Paleozoic and Triassic, to diversitydominance by ammonoids in the Jurassic andCretaceous. Belemnites, present only in theMesozoic, are also grouped with the coleoids. Theextinction of ammonites and belemnites at the endof the Cretaceous reduced the fossil record ofcephalopod genus diversity dramatically, but thenumber of living families, over 30, is about thesame as the average number of families recordedthrough much of the Phanerozoic (Sepkoski, 1982,1992). Elsewhere I have suggested that theprogressive change in diversity dominance incephalopods may have been related to improvedswimming control driven in part by the evolutionof other swimming competitors such as the variousfish groups (Bambach 1985, 1999). Themaintenance of diversity, but with no increase overtime, implies that cephalopods were able to fillsome portion of ecospace well, but not capture newecospace or specialize beyond their initialcapabilities (Bambach, 1985).Arthropods.—Many different classes in thearthropods contain predators: arthropod-likeanimals of unknown affinities, such asanomalocarids; trilobites; eurypterids; andmalacostracans. The jaws and pincers of many ofthese taxa attest to their predatory nature. Trilobiteshad relatively undifferentiated limbs and have oftenbeen thought to have been predominantly depositfeeders, but recent work has shown that many werelikely predators (Fortey and Owens, 1999; Fortey,2000). The diversity of predatory arthropodsfollows a different pattern than that of thecephalopods. Predatory arthropods have highdiversity in the Early and Middle Paleozoic, whichstarts to fade in the mid Devonian and reaches alow level through the Permian and Triassic beforerising continuously from the Jurassic to theHolocene (Fig. 9a). Relatively high diversity wasachieved earlier than for any other predatory groupwith the arthropod incertae sedis (which includesa variety of “stem group” arthropods, such asAnomalocaris and Opabinia from the BurgessShale); and then trilobites diversified in theCambrian (Figs. 10a, 10b), and although overalltrilobite diversity after the end-Ordovicianextinction never recovered its Late Ordovicianlevel, the success of the eurypterids in the Silurian331

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 8—Cephalopods. (a) Genus diversity of cephalopods through the Phanerozoic. Note the highlevel of turnover. Total diversity within intervals may be three times the boundary-crossing standingdiversity. (b) Pattern of change in diversity of nautiloid, ammonoid, and belemnite cephalopods.332

BAMBACH—SUPPORTING PREDATORSabFIGURE 9—Arthropods. (a) Genus diversity of predatory arthropods through the Phanerozoic. (b) Thereplacement of dominant Paleozoic arthropod groups by the Malacostraca and their subsequent increasein diversity in the Cretaceous and Cenozoic.333

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002and Early to mid Devonian kept total arthropodpredator diversity up until the later part of theMiddle Devonian (Fig. 10c). Adrain et al., 2000demonstrate that Silurian alpha (i. e., within habitat)diversity of trilobites recovered to Ordovician levels.This illustrates the important point that global andlocal patterns may be decoupled, as Adrain et al.also argued. In this case, beta (between region)diversity may have decreased, possibly because theglobal trilobite fauna was re-established from a smallnumber of surviving taxa, but those surviving taxawere able to recapture the full range of life habits(niches) that trilobites had occupied prior to theOrdovician extinction. Those surviving taxa mayhave had some properties that contributed to otherchanges observed in the evolutionary dynamics ofthe group. The trilobites show a very high rate ofturnover in the Cambrian and Early Ordovician,attesting to their well-known high rate of evolutionat that time (which gives them great biostratigraphicutility). Their turnover (compared to their continuingstanding diversity) dropped off noticeably in the laterOrdovician and remained at values proportionallyless than in the Cambrian for the rest of thePaleozoic.Figure 10c illustrates that there was aprogressive shift in diversity dominance ofpredatory arthropods throughout the Paleozoic.Trilobites dominated the Cambrian and Ordovician;the eurypterids and malacostracans became asignificant proportion of total predatory arthropoddiversity in the Silurian and Devonian, althoughtrilobites continued in the majority. Then, as boththe trilobites and eurypterids declined in the laterDevonian, the malacostracans retained theirdiversity, which even increased in the Carboniferousbefore also declining to low end-Paleozoic levels.Malacostracans were the only major arthropodmarine predator taxon to survive the end-Permianextinction, and so they are solely responsible forthe post-Paleozoic increase in diversity of predatorymarine arthropods (Fig. 9b). Although notdocumented in detail in Figure 9b, much of thisdiversification from the Cretaceous to the laterCenozoic has been the radiation of the brachyurancrabs with their varied pincer-type claws. In fact,the highly varied limb differentiation ofmalacostracans clearly contrasts with the nearlyuniform and undifferentiated limbs of trilobites.These features attest to the increase in specializationthat characterizes at least some of the Cenozoicincrease in predator diversification.Echinoderms.—The predatory echinodermsalso have a pattern of higher early to mid-Paleozoicdiversity, lower diversity from the LateCarboniferous through the Triassic, and then a riseto a new peak of diversity in the later Mesozoic(Fig. 11a). Unlike the arthropods and more like theanthozoa, they only maintain, but do not increase,Late Cretaceous diversity levels in the Cenozoic.The infrequent but quite large turnover peaks alsoindicate that the fossil record of predatoryechinoderms is rather poor, but the total diversitypeaks track the same general diversity pattern asthe standing diversity plot. Figure 11b illustratesthat, although some ophiuroids and echinaceanechinoids are also predators, the vast majority ofpredatory echinoderms have always been starfish.Echinoids did not become effective predators untilthe modified and stronger lantern of the Echinaceaevolved (Parker, 1982).Vertebrates.—Predatory vertebrates have adiversity history (Fig. 12a) similar to that ofgastropods or of all predators combined, but notlike that of the Anthozoa, Cephalopoda,Arthropoda, or Echinodermata. Their genusdiversity fluctuated up and down from theOrdovician through the Early Cretaceous butremained relatively low, with boundary-crossingstanding diversity at or below 50 genera throughoutthe 400-million-year interval. It more than doubledin the Late Cretaceous and from there has increasedfive-fold during the Cenozoic.Figure 12b shows that conodontophorids werethe diversity-dominant marine vertebrate predatorsin the Ordovician and Silurian. They were replacedas diversity dominants by the combination of thefish classes as a whole in the Devonian. Thecombined fish groups have remained the diversitydominantvertebrate marine predators, with tetrapodsmaking up only a small proportion of Mesozoic andCenozoic vertebrate marine predator taxa.334

BAMBACH—SUPPORTING PREDATORSabcFIGURE 10—Arthropod predators in the Paleozoic. (a) Genus diversity of predatory trilobites in thePaleozoic. (b) Genus diversity of predatory arthropod incertae sedis and eurypterids in the Paleozoic.(c) Shifts in diversity dominance among various taxa of predatory Paleozoic arthropods.335

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 11—Echinoderms. (a) Genus diversity of predatory echinoderms through the Phanerozoic.(b) Diversity of various taxa of predatory echinoderms illustrating the diversity dominance of the Asteroidea.336

BAMBACH—SUPPORTING PREDATORSabFIGURE 12—Vertebrates. (a) Genus diversity of predatory vertebrates through the Phanerozoic.(b) Comparison of the genus diversity of conodonts, the various classes of fishes combined, and the tetrapods.337

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Conodonts.—Feeding in conodonts is still notfully understood, but recent studies of both growthpatterns and evidence of wear of conodont elements(summarized in Purnell, 2001) indicate thatconodontophorids were not suspension (filter)feeders, but were macrophagous. The combinationof their life habits as swimming animals (shownby fin-like features in well-preserved specimensplus their general facies-crossing distribution), thepresence of large eyes (also shown in wellpreservedspecimens), and the evidence ofmacrophagy suggests they were predators. Thegenus diversity of conodontophorids (Fig. 13a)reached a peak in the Ordovician, droppedmarkedly in the Ordovician extinction, recoveredbut to a plateau at about half the Ordovicianmaximum extending from the Silurian through theEarly Carboniferous (with another sharp hit andrecovery in the Late Devonian), and then declinedgradually to extinction by the end of the Triassic.Agnathans.—Although some living jawlessfish are parasites and can effectively be consideredpredators, there is little direct evidence of the dietof most fossil agnathans. Recent opinions,summarized by Purnell (2001), range frompredation (two claims in the last decade) to filterfeeding(two claims), but strongly favor depositfeeding(nine claims) as a major mode of life foragnathans. On this basis I omit the jawless fish fromthe count of predators even though some livingagnathans are harmful parasites that can beresponsible for the death of their food suppliers.Classes of Jawed Fish.—Three classes ofjawed fish with predominantly cartilaginousinternal skeletons became important in thePaleozoic: the placoderms, acanthodians, andchondrichthyans. As a group (Fig. 13b) these threeclasses had relatively high diversity in theDevonian and Carboniferous, declining diversityin the Permian, slowly rising diversity from thestart of the Triassic through the Early Cretaceous(when they again reached diversity levels close tothose achieved in the Devonian and Carboniferous),and then a major increase in diversity in the LateCretaceous with a slower, but steady, increasethrough the Cenozoic. The Late Cretaceous andCenozoic increase in diversity was restricted to theChondrichthyes, the only one of the three classeswith cartilaginous internal skeletons to survive thePaleozoic. The large increase was predominantly thediversification of the Neoselachians, in which “therelationship between the braincase and the jaws ismodified to form a more maneuverable feedingapparatus” (Carroll, 1988, p. 74). Carroll goes on topoint out that this permits modern sharks to dig intolarger prey and gouge out large pieces of flesh, andalso permits the jaws of bottom-feeding forms toform an effective suction device. When the diversityof conodontophorids is added (Fig. 13c), the patternof high Ordovician through Carboniferous diversity,low Permian through Early Cretaceous diversity, andhigh Late Cretaceous and Cenozoic diversity is evenmore striking and is similar to the arthropods andechinoderms. Note that the bony fishes, theOsteichthyes, are not included in this grouping,which is limited only to those classes of vertebratesthat achieved major peaks of diversity in thePaleozoic. If we consider only the Paleozoic(Fig. 14a), but look at all predatory vertebrateclasses, including the Osteichthyes, we see asequential replacement in diversity dominance fromacanthodians in the Silurian and Early Devonian toplacoderms in the Middle and Late Devonian (withsignificant, but subordinate, contributions from theacanthodians, Osteichthyes, and Chondrichthyes) todiversity dominance by the Chondrichthyes in theCarboniferous and Permian. In the post-Paleozoic(Fig. 14b) the Chondrichthyes have increased indiversity more or less monotonically, with an offsetin the early part of the Late Cretaceous. On theother hand, the Osteichthyes diversified at aboutthe same rate as the Chondrichthyes during theentire Mesozoic, as can be seen by the near 50-50division between the two groups of their combineddiversity, including the “burst” in the LateCretaceous; but the bony fish skyrocket in diversityduring the Cenozoic (Figs. 14b and 15).The Late Cretaceous jump and the rapidCenozoic rise of diversity in predatory bony fishlargely reflects the radiation of the teleosts. Most,but not all, teleosts are predators. Moyle and Cech(1996), who note that teleosts include detritivores,338

BAMBACH—SUPPORTING PREDATORSabcFIGURE 13—Dominant Paleozoic vertebrate predators. (a) Genus diversity of conodontophorids throughthe Phanerozoic. (b) Phanerozoic genus diversity of the predominantly cartilaginous skeletonized fish classes(placoderms, acanthodians, and chondrichthyans) grouped together. (c) Paleozoic dominant vertebratepredator genus diversity through the Phanerozoic (all vertebrates except bony fish and tetrapods).339

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 14—Fish groups. (a) Comparison of the genus diversity of the various fish classes during thePaleozoic. (b) Comparison of the Chondrichthyes and Osteichthyes through the Mesozoic and Cenozoic.340

BAMBACH—SUPPORTING PREDATORSFIGURE 15—Genus diversity of the Osteichthyes through the Phanerozoic.FIGURE 16—Genus diversity of predatory marine tetrapods through the Mesozoic and Cenozoic.341

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002herbivores, and omnivores, as well as carnivores,point out that only about 20% of reef fishes areherbivores. Widely spaced high peaks of turnoverin the Osteichthyes (Fig. 15) again illustrate theimportance of lägerstetten for documenting morecompletely the diversity of less easily preserved taxa,but, as with other groups discussed earlier, thegeneral track of the high points follows the overallpath of boundary-crossing standing diversity.Tetrapods.—Figure 16 illustrates the diversityof marine reptiles and predatory marine mammals.Marine birds, while important today, have too poora fossil record to be used in the analysis (no morethan eight genera are recorded for any interval).The diversity of marine reptiles fluctuated, but fromthe Triassic through the Early Cretaceous standingdiversity never rose far above 10 genera. However,there was a burst of diversification in the LateCretaceous, with continuing standing diversity risingabove 35 genera and total diversity within an intervalreaching 58 genera in the Maastrichtian. The end-Cretaceous extinction eliminated all marine reptilesexcept the sea-turtles, and these have persisted withlittle diversity change through the Cenozoic (fourof the five living genera are carnivorous; only oneis herbivorous; Parker, 1982). Marine mammals firstappeared in the Paleogene and diversifiedsignificantly in the Neogene, reaching continuingstanding diversity levels equal to that of marinereptiles in the Late Cretaceous. High turnover peaksfor total diversity recorded within intervals in theNeogene (as high as 100 genera) suggest thatNeogene marine predatory mammal diversity mayactually have significantly exceeded that of themarine reptiles of the Cretaceous. The close spacingof the high turnover peaks indicates that tetrapodfossils were not limited to infrequently occurringlägerstetten, but were preserved regularly throughtime. The high proportional turnover compared toboundary-crossing standing diversity, then, meansthat many taxa were confined to single intervals andimplies either rapid evolutionary turnover or rarityof taxa (or both).PREDATORS AND PREYReview of Predator Diversity in General.—Asseen in both Figures 3 and 17 the genus diversityof predators was low in the Cambrian, built upduring the Ordovician Radiation, fluctuated withno particular trend up or down until the end of theTriassic, increased slowly through the Jurassic andEarly Cretaceous, and has increased rapidly sincethe start of the Early Cretaceous, interrupted onlyFIGURE 17—Sequence of diversity dominance between groupings of predatory taxa based on thetiming of when the members of each group first reach relatively high diversity. Members of each grouplisted on the figure.342

BAMBACH—SUPPORTING PREDATORSby the devastation of the end-Cretaceous massextinction event and a brief pause in the lateEocene. Within this overall pattern is a progressiveshift in diversity dominance, as also shown inFigure 17. Although the predatory taxa thatdominated in the Cambrian actually increased indiversity during the Ordovician, they ended up onlyequal in diversity to the predator taxa that wereadded during the Ordovician Radiation. Theybecame a minority of predator taxa after the end-Ordovician extinction and were all extinct (exceptfor a few genera of conodonts) by the end of thePaleozoic. Those taxa that diversified during theOrdovician Radiation dominated predator diversitythrough the rest of the Paleozoic. In the post-Paleozoic, the taxa that had diversified first in theOrdovician Radiation recovered the diversity theylost in the end-Permian extinction (and then some),but their total diversity in the Late Cretaceous andCenozoic remained only about equal to that of allpredators from the Ordovician through Devonian.The explosive Cenozoic increase in predatordiversity was predominantly confined to groupsthat diversified strongly only after the Paleozoic,although a few early representatives of those taxawere present in the Paleozoic. This burst ofdiversification in the last hundred million years hasbrought predator diversity to a level some eighttimes greater than at any time in the Paleozoic.Evolutionary Dynamics of PredatorsCompared to Prey.—We have seen several groupsof predators with noticeably high turnover peaks(e. g., cephalopods, trilobites, all fish combined inthe Devonian and Carboniferous, and Cretaceousand Cenozoic Osteichthyes). Although it appearsthat the turnover peaks of predators as a whole, asshown in Figure 3a, frequently may beproportionately larger than the turnover peaks ofnon-predator taxa (Fig. 3b), it is not easy to judgeby eye whether this is the general case. However, ifthe differences in proportion of origination andextinction between the two groups are calculatedon an interval-by-interval basis (Fig. 18), it is clearthat the proportion of origination and extinction ofpredators is consistently greater than that for nonpredators(potential prey). The median proportionof origination for predators (Fig. 18a) is 1.4 timesthat for non-predators, and only 10 of 107 intervalsthrough the whole Phanerozoic (all in the Paleozoic,with six in the Middle and Late Ordovician whenthe overall pattern of diversity is leveling off fromthe Ordovician Radiation) have a lower proportionof origination for predators than for non-predators.From the start of the Jurassic through the Paleogene,27 intervals have the proportion of predatororigination compared to non-predators above themedian and only eight below. The same is true forthe proportion of extinction (Fig. 18b). The medianvalue for proportion of extinction of predators is1.49 times that for non-predator taxa and onlyseven of 107 intervals have a lower proportion ofextinction for predators than non-predators. Fromthe start of the Triassic through the Paleogene, 33intervals have the proportion of extinction forpredators above the median for the comparison tonon-predators and only eight are below.The increased evolutionary volatility ofpredators compared to non-predators thatcharacterizes the Mesozoic and Cenozoic actuallybegan in the Devonian. From the start of the LateDevonian through the Permian, 14 intervals show aproportion of origination for predators higher thanthe median for predators compared to non-predators,and only nine below (Fig. 18a). For the comparisonof proportion of extinction in these intervals, 17 areabove the median and only six below (Fig. 18b).Before the Late Devonian, although the proportionof origination and extinction of predators wasgenerally greater than that of non-predators, onlysix of 38 intervals had a proportion of originationhigher than the median for predators compared tonon-predators and only four of 38 intervals had aproportion of extinction of predators higher than themedian value for the comparison to non-predators.Figure 18 illustrates that predators have alwayshad more volatile evolutionary dynamics than nonpredators.Several factors other than the possibilityof intrinsically higher rates of evolution could havecontributed to the higher apparent evolutionaryvolatility of predators compared to non-predators.Because predators must always be less commonthan prey, their higher frequency of rare taxa, which343

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 18—Proportions of origination and extinction of predators compared to non-predators. (a)The proportion of origination of predators compared to that of non-predators expressed as the ratio ofproportion of origination of predators divided by that for non-predators. (b) The proportion of extinctionof predators compared to that of non-predators expressed as the ratio of proportion of extinction ofpredators divided by that for non-predators.should have a sparse fossil record, would lead tomore reported single occurrences, and this wouldinflate recorded turnover rate within intervals.Dependence on prey, especially when predatorsspecialize on single or a few similar taxa as prey,could also make predators ecologically less stablethan prey species. Fluctuations in prey abundancemight disrupt specialized predators so they couldnot survive, but the prey taxa might recover andpersist. In addition, if predators increased inspecialization through time they would increase indiversity, but not necessarily in abundance. Again,reduction in prey could have a severe “upward”effect in the food chain on the specialized predatorsbecause of the constraints of biomass and efficiencyof energy transfer in food pyramids, forcing moreturnover in predators than prey.Predator and Prey Diversities Compared.—Figure 19a shows the diversity of predators andall metazoa through the Phanerozoic. Predators344

BAMBACH—SUPPORTING PREDATORShave always been a minority of the number ofgenera, but their diversity history has followed thesame general outline as all Metazoa combined. But,as can be seen by the relative positions of the twocurves, the proportion of diversity comprised ofpredators clearly increased during the Ordovicianand also in the Cenozoic. This is illustrated directlyin Figure 19b, which shows the calculatedproportion of global marine diversity comprisedof predators through the Phanerozoic. The diversityof predators as a proportion of total metazoandiversity has increased, but in a stepwise fashion,not continuously. The proportionality follows athree phase sequence of transitions and relativelystable intervals, parallel to that recently reportedby Bambach and Knoll for other parsings of theglobal marine fauna (Bambach and Knoll, 1997;Bambach, Knoll, and Sepkoski, 2002). During theCambrian to mid-Ordovician interval, predatorsincreased continuously as a proportion of allmetazoan genera, from less than five percent inthe earliest Cambrian to twenty percent in the mid-Middle Ordovician before their proportion of totaldiversity dropped slightly as the OrdovicianRadiation leveled off. Then, for the 200 millionyears from the start of the Caradoc to the end ofthe Permian, the proportion fluctuated around amean of 0.15 (15%), never rising over 0.19 or fallingbelow 0.10 despite the changes in total diversity andthe turnover and replacement of taxa that occurredduring that time. Selectivity in the end-Permianextinction left the survivors in the earliest Triassicwith an unusually high proportion of predators(0.275). The apparent trend from lower to higherproportion of predators in the Late Permian mayjust be a fluctuation from a low proportion, butwithin the established range (similar to thefluctuation in the Early Carboniferous); but it couldalso include a “Signor-Lipps Effect” influencerelated to apparent early truncation of some rangesin the highly selective end-Permian extinction. Thenew proportion of predators created by theselectivity of the end-Permian extinction did notpersist. During the Early and early Middle Triassicthe proportion of predator taxa decreased, andpossibly by the beginning of the Carnian, butcertainly in the Early Jurassic (after the recoveryfrom the perturbation associated with the end-Triassic extinction), the proportion of predator taxastabilized—but at a new proportionality. From thestart of the Carnian until the end of the Cretaceousthe proportion of metazoan marine generacomprised of predators averages 0.219 with nointerval higher than 0.238 and only two (at the endof the Triassic and the start of the Jurassic) fallingbelow 0.206. In fact, except for the two intervalsat the Triassic-Jurassic boundary, no other intervalsin the entire post-Paleozoic have a proportion ofpredator taxa that overlaps with any in the 200-million-year interval extending from the LateOrdovician through the Permian. Following the end-Cretaceous extinction another transitional intervaloccupied the Paleocene, and from the Eocene to theHolocene the proportion of metazoan diversitycomprised of predators has always exceeded 0.3,with a mean of 0.36 during the Neogene.Predators and Energetics.—The increase froman average of fifteen percent predator taxa in thePaleozoic to more than thirty percent in the laterCenozoic implies there has been a major change inthe structure of the marine ecosystem over time. Ifthe efficiency of energy transfer between levels inthe food pyramid has stayed nearly constant (andthere is no evidence that it has not), then a reasonablehypothesis for the increased proportional diversityof predators is that either the biomass of prey specieshas increased or the populations of prey taxa havebecome more stable (or both), permitting greaterspecialization by predators on more reliable foodsupplies. Thus, predators could increase theirproportional share of total diversity by specializingon fewer prey taxa. In some predator groups thathave increased diversity markedly in the Cenozoic,specialization is clearly documented. Two examplesare the remarkable niche subdivision achieved bythe genus Conus in Hawaii (Kohn, 1959) and therange of highly specialized claw morphologiesevolved in the crabs (Warner, 1977). Vermeij hasdiscussed these and a number of related conceptsat length, emphasizing especially (and importantly)the coevolutionary responses of escalation(Vermeij, 1987, 1994).345

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abFIGURE 19—Comparison of the diversity of predators with that of non-predators. (a) Boundary-crossingdiversity of predators and all marine metazoa compared through the Phanerozoic. (b) Diversity ofmarine predators as a proportion of the diversity of all marine metazoa.346

BAMBACH—SUPPORTING PREDATORSIn an earlier study, I discussed possibleinfluences on energetics in the marine ecosystem,emphasizing the changes that may have taken placewhen terrestrial biomass initially increased on landduring the Devonian and when angiospermsbecame dominant on land in the Cretaceous andEarly Cenozoic (Bambach, 1999). In that study, Inoted major changes in diversity dominance of bothpredator and prey taxa in the Devonian, whenevolutionary volatility in predators began toincrease, although overall diversity did not. Theincrease in proportion of predator taxa in theCenozoic was also noted, although the evaluationof the number of predator taxa in that paper shouldbe regarded as preliminary to the more complete,but still predominantly estimated, analysispresented here. I argued that both intervals ofchange reflected change in the energetics of theecosystem. These intervals of change map withSignor and Brett’s (1984) “precursor to theMesozoic marine revolution” and Vermeij’s (1977)classic analysis of the Mesozoic marine revolution.However, the two major changes in predators as aproportion of the diversity of the marine faunadocumented in the current study, which occur inthe recovery intervals following the devastation ofthe two era-bounding mass extinctions, do notcoincide with the suggestion of Vermeij thatrevolutions in the history of life are not linked tocrises (Vermeij, 1995, p. 145).When we consider the proportion of diversityof several different groups of taxa within thepredators as a whole, a pattern of more continuousturnover in predator diversity dominance emerges.The three groupings illustrated in Figure 20 wereselected because of some general similarities in thediversity histories of the individual taxa. Trilobites,eurypterids, arthropod incertae sedis, conodonts,placoderms, and acanthodians all reached theirmaximum diversity in the Paleozoic, weregenerally confined to the Paleozoic, and were allextinct by the end of the Triassic. When groupedtogether they comprise over 90% of the diversityof Cambrian predators and decline continuouslyin their share of predator diversity after that(Fig.20a). Predatory polychaetes and cephalopodseach display a fairly even diversity path throughthe bulk of the Phanerozoic; and anthozoans,echinoderms, and chondrichthyans have peaks ofdiversity in the Paleozoic and Cenozoic withmarkedly lower diversity in between, but their post-Paleozoic diversity increase is not dramatic andthey fit with the polychaetes and cephalopods inhaving high Paleozoic diversity with Cenozoicdiversity still in the same general range. Summedtogether (Fig. 20b), these groups, which eachreached a Paleozoic peak of diversity and alsopersisted through the Mesozoic, became thediversity dominant predators in the OrdovicianRadiation (rising from less than 10% to 70% ofpredator diversity during that time). Their share ofpredator diversity averaged about 70% through therest of the Paleozoic, fluctuating from 58% to 75%.In the Mesozoic it decreased some, but remainedover 50%, and averaged slightly below 60%, ratherthan over 65% as in the bulk of the Paleozoic. Notethat this group of taxa maintains a relatively stableproportion of predator diversity through the laterPaleozoic and early Mesozoic, despite many oftheir constituent taxa having lower absolutediversity during that interval, because total predatordiversity was lower during that time. Late in theEarly Cretaceous the share of predator diversityheld by these taxa began to drop continuously, andit is now at a low point of 10%. A third group oftaxa, the gastropods, malacostracans, Osteichthyes,and tetrapods, all have diversity maxima in theCenozoic and little or no Paleozoic diversity.Grouped together (Fig. 20c), these taxacontinuously increased their share of predatordiversity through the Phanerozoic. They comprisedless than 10% of predator diversity until the mid-Devonian, but passed 70% of total predatordiversity in the Paleocene.Many of the taxa in the groups that reachedhigh diversity in the Paleozoic are relatively lowenergy-expendingpredators (starfish, anthozoa).In my earlier paper (Bambach, 1999) I alsodocumented that the turnover in diversitydominance in predators in the Devonian may havebeen a replacement of taxa with lower energeticsby those with higher energetics (lower and higher347

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abcFigure 20348

BAMBACH—SUPPORTING PREDATORSmetabolic rates as well as different levels ofactivity). This would imply that the overallenergetics of the predator fauna increased in thePaleozoic and Early Mesozoic. TheNeogastropoda, Malacostraca, Osteichthyes, andtetrapods are all active, high-metabolic-rateorganisms. In the Late Cretaceous and Cenozoicthese groups diversified dramatically, from astanding genus diversity of 164 genera at theAlbian-Cenomanian boundary to 1426 genera atthe Pliocene-Pleistocene boundary, an increase of8.7 times as compared to a 1.44 times increase(from 114 to 164 genera) for other predatory taxaover the same 100 million year interval. It is thisexplosive diversification that is responsible for themajor increase in predators as a proportion of allmetazoan diversity. These increases among taxawith greater energetics and the increase of theproportional representation of predators imply thathigh-quality food (i. e., prey taxa) in greater amountsbecame available to support this diversification.Four lines of evidence from non-predators alsosupport the idea that the base of the food pyramidhas expanded over time, and that this has beenparticularly important in the post-Paleozoic: (1)The diversity of non-predators (prey organisms)has increased (see Fig. 3b). Accompanying that,(2) the biomass per individual of the common preytaxa has apparently increased as diversitydominance (and presumably abundance, as well)has shifted over time (Bambach, 1993). (3) Thedepth and intensity of bioturbation, much of itcaused by the activities of deposit-feeders, hasincreased over time (Thayer, 1983; Sepkoski,Bambach, and Droser, 1991). Bottjer and Ausich(1986) argue that increased predation pressure,something that fits nicely with the data presentedhere, would have been an important selectivepressure to stimulate deeper burrowing by potentialprey. The deeper burrowing by deposit feeders,however, must reflect an increase in the amount ofburied potential food (Bambach, 1993), unless itis because the organic matter that gets buried hasbecome more refractory with time. In the lattercase, it would take longer for bacterial and fungalattack to make the material useable as food; thus,as sedimentation continued, the depth of burial atwhich it would become useful as food wouldincrease. This would require deeper burrowing fordeposit-feeders to gain access to it as a food supply.However, there is no apparent increase in refractoryorganic material in marine phytoplankton or algae,unlike the increase in woody tissues and cuticlesthat evolved as land plants increased in size andinvaded a wider range of habitats; nor was thereany change in refractory components between thegymnosperm floras of the mid-Mesozoic andangiosperm-dominated floras of the Cenozoic.Therefore, even if burial of refractory material fromland plants has been a factor in increased foodsupply to the oceans, much of that change occurredin the Devonian and Carboniferous, not in morerecent geologic time when the depth of bioturbationincreased across the spectrum of marineenvironments (Sepkoski, Bambach, and Droser,1991). Also, deeper burrowing was always presentin high-productivity settings, even in the Cambrian,and the increase in depth of bioturbation hasoccurred over a wider range of facies as time hasgone on, suggesting that it is related to more food←FIGURE 20—Proportion of predator diversity comprising different diversity-dominant groups of taxa.(a) Diversity of the taxa of marine predators that reached their maximum diversity in the Paleozoic andare now extinct, graphed as a proportion of the diversity of all marine predators. (b) Diversity of the taxaof marine predators that achieved high diversity in the Paleozoic and either maintained, but did notincrease it later, or recovered or exceeded that diversity in the Cenozoic, graphed as a proportion ofthe diversity of all marine predators. (c) Diversity of the taxa of marine predators that had low Paleozoicdiversity and only reached their maximum diversity in the Cenozoic, graphed as a proportion of thediversity of all marine predators.349

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002actually getting buried in more settings (Sepkoski,Bambach, and Droser, 1991; Bambach, 1993). And,as a last point, (4) as determined from thecomposition of oceanic sediments, the diversity andabundance of phytoplankton has increased duringthe post-Paleozoic. Recent work on photosyntheticsystems in marine autotrophs (Kirk, 1994;Falkowski and Raven, 1997) indicates that differenttaxa (red algae, brown algae, dinoflagellates,diatoms, coccolithophorids), common both in thephytoplankton and algal benthos, utilize differentwavelengths of light, and some function underlower light intensities than green algae, implyingthat they can conduct photosynthesis successfullyat greater depths, or in more turbid water, than greenalgae alone, thus potentially increasing the volumeof water in which primary production can occur.As an example, some dinoflagellate, rhodophyte,and diatom photosymbionts in various foraminiferaare known to conduct photosynthesis at depths twoto four times greater than chlorophytephotosymbionts (Goldstein, 1999). Although redalgae (Rhodophyta) and several types of greenalgae (Chlorophyta, Prasinophyta, and Acritarcha)are known from the Proterozoic, and charophytes(which are overwhelmingly fresh-water) and oneproblematic species of dinoflagellate are reportedfrom Silurian rocks, the Mesozoic saw a burst ofnew phytoplankton diversity. Dinoflagellatesappeared consistently for the first time in theMiddle Triassic and radiated in the Jurassic(building from 4 to 31 families); coccolithophoridsfirst appeared in the Late Triassic and radiated inthe Jurassic and Early Cretaceous (building from 2to 14 families); and diatoms first appeared in theLate Jurassic and radiated in the Cretaceous andPaleogene (building from one to 23 families) (datafrom Benton, 1993; older species diversitydiagrams also in Tappan and Loeblich, 1973). Soit may be that primary productivity in the oceansdid, in fact, expand during the Mesozoic andCenozoic, enlarging the base of the food chain tosupport the observed increase in predator diversity.Vermeij and Lindberg (2000) wrote anintriguing paper documenting that herbivory isgenerally present only in fairly derived clades andbecame important late in the history of life. Theyargue that microphagy (including suspensionfeeding),detritus feeding, and carnivory are the basalfeeding strategies. This startling, but, in the contextof phylogeny, clearly well-grounded idea nicelymaps with the dominance of suspension-feeders(brachiopods, bryozoa, crinoids) in the Paleozoicoceans. Vermeij and Lindberg place the origin ofmany important herbivorous clades in the Mesozoicand Cenozoic. The spread of herbivory would haveopened opportunity for added diversification ofpotential prey taxa and also quantitatively enlargedthe base of the trophic pyramid for marine animals.Both factors could have contributed to the increasein specialization and diversity of predators in theMesozoic and Cenozoic noted previously by others(Vermeij, 1977, 1987; Kowalewski et al., 1998) andcensused here.All of the patterns noted here are based ontaxon richness alone. To document the precisemeaning of these changes in diversity will requirequantitative evaluation of the abundance of taxa—both predators and prey. Evaluation of size (or someother proxy for biomass) will also be important.Only then can we know whether the inferencesdrawn from the changes in diversity documentedhere are correct. But until then, this is the bestpicture we have of the general history of marinepredators through the Phanerozoic.ACKNOWLEDGMENTSI want to thank Michal Kowalewski for invitingme to compile the data on predator diversity andAndy Knoll for providing opportunity for me tobegin to appreciate the importance and history ofprimary producers, the foundation on which allanimals depend. Arnie Miller and Geerat Vermeijprovided very useful and constructive reviews ofthe manuscript.350

BAMBACH—SUPPORTING PREDATORSREFERENCESADRAIN, J. M., S. R. WESTROP, B. D. E. CHATTERTON, AND L. RAMSKÖLD. 2000. Silurian trilobite alpha diversity andthe end-Ordovician mass extinction. Paleobiology, 26:625–646.ALROY, J., C. R. MARSHALL, R. K. BAMBACH, K. BEZUSKO, M. FOOTE, F. T. FÜRSICH, T. A. HANSEN, S. M. HOLLAND,L. C. IVANY, D. JABLONSKI, D. K. JACOBS, D. C. JONES, M. A. KOSNIK, S. LIDGARD, S. LOW, A. I. MILLER, P. M.NOVACK-GOTTSHALL, T. D. OLSZEWSKI, M. E. PATZKOWSKY, D. M. RAUP, K. ROY, J. J. SEPKOSKI, JR., M. G.SOMMERS, P. J. WAGNER, AND A. WEBBER. 2001. Effects of sampling standardization on estimates of Phanerozoicmarine diversity. Proceedings of the National Academy of Sciences (USA), 98:6261–6266.BAMBACH, R. K. 1977. Species richness in marine benthic habitats through the Phanerozoic. Paleobiology, 3:152–167.BAMBACH, R. K. 1985. Classes and adaptive variety: The ecology of diversification in marine faunas through thePhanerozoic, p. 191–253. In J. W. Valentine (ed.), Phanerozoic Diversity Patterns: Profiles in Macroevolution.Princeton University Press, Princeton.BAMBACH, R. K. 1990. Late Palaeozoic provinciality in the marine realm, p. 307–323. In W. S. McKerrow and C. R.Scotese (eds.), Palaeozoic Palaeogeography and Biogeography. The Geological Society, London, Memoir 12.BAMBACH, R. K. 1993. Seafood through time: Changes in biomass, energetics and productivity in the marineecosystem. Paleobiology, 19:372–397.BAMBACH, R. K. 1999. Energetics in the global marine fauna: A connection between terrestrial diversification andchange in the marine biosphere. GEOBIOS, 32:131–144.BAMBACH, R. K., AND A. H. KNOLL. 1997. Fundamental Physiological Control on Patterns of Diversification in theMarine Biosphere. Geological Society of America, Abstracts With Programs, 29(7):A-31.BAMBACH, R. K., A. H. KNOLL, AND J. J. SEPKOSKI, JR. In press. Anatomical and Ecological Constraints on PhanerozoicAnimal Diversity in the Marine Realm. Proceedings of the National Academy of Sciences (USA), 99.BENTON, M. J. (ed.). 1993. The Fossil Record, 2. Chapman and Hall, London.BOTTJER, D. J., AND W. I. AUSICH. 1986. Phanerozoic development of tiering in soft substrata suspension-feedingcommunities. Paleobiology, 12:400–420.CARROLL, R. C. 1988. Vertebrate Paleontology and Evolution. W. H. Freeman and Company, New York, 698 p.DALEY, G. M. 1999. Evaluation of stability of Pliocene communities, Yorktown Formation, Virginia. UnpublishedPh. D. dissertation, Virginia Polytechnic Institute and State University, Blacksburg, Virginia.DALEY, G. M., AND M. KOWALEWSKI. 2000. Species accumulation curves and sampling efficiency of paleontologicaldata sets. Geological Society of America, Abstracts with Programs, 32(7):A-95.ENGESER, T. S. 1990. Major events in cephalopod evolution, p. 119–138. In P. D. Taylor and G. P. Larwood (eds.), MajorEvolutionary Radiations. Systematics Association Special Volume Series 42. Clarendon Press, Oxford, England.ENGESER, T. S. 1996. The position of the Ammonoidea within the Cephalopoda, p. 3–19. In N. H. Landman, K.Tanabe, and R. A. Davis (eds.), Ammonoid Paleobiology. Plenum Press, New York.FALKOWSKI, R. G., AND J. A. RAVEN. 1997. Aquatic Photosynthesis. Blackwell Science, Malden, Massachusetts.FORTEY, R. 2000. Trilobite! Alfred Knopf, New York.FORTEY, R. A., AND R. M. OWENS. 1999. Feeding habits in trilobites. Palaeontology, 42:429–465.GOLDSTEIN, S. T. 1999. Foraminifera: a biological overview, p. 37–55. In B. K. Sen Gupta (ed.), ModernForaminifera. Kluwer Academic Publishing, London.HYMAN, L. H. 1940. The Invertebrates: Protozoa through Ctenophora. McGraw-Hill, New York.HYMAN, L. H. 1967. The Invertebrates, Vol. VI, Mollusca 1. McGraw-Hill, New York.JACOBS, D. K., AND N. H. LANDMAN. 1993. Nautilus—a poor model for the function and behavior of ammonoids?Lethaia, 26:101–111.KIRK, J. T. O. 1994. Light and Photosynthesis in Aquatic Ecosystems, 2 nd ed. Cambridge University Press, Cambridge.KOHN, A. J. 1959. Ecology of Conus in Hawaii. Ecological Monographs, 29:47–90.KOWALEWSKI, M., A. DULAI, AND R. FÜRSICH. 1998. A fossil record full of holes: The Phanerozoic record of drillingpredation. Geology, 26:1091–1094.MAY, R. M. 1981. Theoretical Ecology: Principles and Applications. Sinauer Associates, Inc., Sunderland,Massachusetts.351

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002MOYLE, P. B., AND J. J. CECH, JR. 1996. Fishes: An Introduction to Ichthyology, 3 rd edition. Prentice-Hall, UpperSaddle River, New Jersey.PARKER, S. A. (ed.). 1982. Synopsis and Classification of Living Organisms (2 vol). McGraw-Hill, New York.PETERS, S. E., AND M. FOOTE. 2001. Biodiversity in the Phanerozoic: a reinterpretation. Paleobiology, 27:583–601.POLIS, G. A., AND K. O. WINEMILLER (eds.). 1996. Food Webs: Integration of Patterns and Dynamics. Chapman &Hall, New York.POWELL, M. G., AND M. KOWALEWSKI. 2002. Increase in evenness and sampled alpha diversity through the Phanerozoic:Comparison of early Paleozoic and Cenozoic marine fossil assemblages. Geology, 30:331–334.PURNELL, M. A. 2001. Scenarios, selection, and the ecology of early vertebrates, p. 187–208. In P. E. Ahlberg(ed.), Major Events in Early Vertebrate Evolution. Systematics Association Special Volume Series 61, Taylorand Francis, New York.RAUP, D. M. 1976. Species diversity in the Phanerozoic: an interpretation. Paleobiology, 2:289–297.RAUP, D. M. 1979. Biases in the fossil record of species and genera. Bulletin of the Carnegie Museum of NaturalHistory, 13:85–91.SEPKOSKI, J. J., JR. 1981. A factor analytic description of the Phanerozoic marine fossil record. Paleobiology, 7:36–53.SEPKOSKI, J. J., JR. 1982. A compendium of fossil marine families. Milwaukee Public Museum Contributions inBiology and Geology, 51.SEPKOSKI, J. J., JR. 1992. A compendium of fossil marine families, 2 nd ed. Milwaukee Public Museum Contributionsin Biology and Geology, 83.SEPKOSKI, J. J., JR., R. K. BAMBACH, AND M. L. DROSER. 1991. Secular changes in Phanerozoic event bedding andthe biological overprint, p. 298–312. In B. Einsele, W. Ricken, and A. Seilacher (eds.), Cycles and Events inStratigraphy. Springer-Verlag, Berlin.SEPKOSKI, J. J., JR., R. K. BAMBACH, D. M. RAUP, AND J. W. VALENTINE. 1981. Phanerozoic marine diversity and thefossil record. Nature, 293:435–437.SIGNOR, P. W., III, AND C. E. BRETT. 1984. The mid-Paleozoic precursor to the Mesozoic marine revolution.Paleobiology, 10:229–245.SOUKHANOV, A. H. (ed.). 1992. The American Heritage Dictionary of the English Language, Third Edition. HoughtonMifflin Company, Boston.TAPPAN, H., AND A. R. LOEBLICH, JR. 1973. Evolution of the oceanic phytoplankton. Earth Science Reviews, 9:207–240.THAYER, C. 1983. Sediment-mediated biological disturbance and the evolution of the marine benthos, p. 479–625. In M.Tevesz and P. McCall (eds.), Biotic Interactions in Recent and Fossil Benthic Communities. Plenum, New York.VERMEIJ, G. J. 1977. The Mesozoic marine revolution: evidence from snails, predators and grazers. Paleobiology,3:245–258.VERMEIJ, G. J. 1987. Evolution and Escalation: An Ecological History of Life. Princeton University Press, Princeton,New Jersey.VERMEIJ, G. J. 1994. The evolutionary interaction among species: selection, escalation, and coevolution. AnnualReviews of Ecology and Systematics, 25:219–236.VERMEIJ, G. J. 1995. Economics, volcanoes, and Phanerozoic revolutions. Paleobiology, 21:125–152.VERMEIJ, G. J., AND D. R. LINDBERG. 2000. Delayed herbivory and the assembly of marine benthic ecosystems.Paleobiology, 26:419–430.WARNER, G. F. 1977. The Biology of Crabs. Elek Science, London.352

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESTHE FOSSIL RECORD OF PREDATOR-PREY ARMS RACES:COEVOLUTION AND ESCALATION HYPOTHESESGREGORY P. DIETL 1 AND PATRICIA H. KELLEY 21Department of Zoology, North Carolina State University, Raleigh, NC 27695-7617 USA2Department of Earth Sciences, University of North Carolina, Wilmington, NC 28403-3297 USAABSTRACT—Arms races between predators and prey may be driven by two related processes—escalation andcoevolution. Escalation is enemy-driven evolution. In this top-down view of an arms race, the role of prey (withthe exception of dangerous prey) is downplayed. In coevolution, two or more species change reciprocally inresponse to one another; prey are thought to drive the evolution of their predator, and vice versa. In the fossilrecord, the two processes are most reliably distinguished when the predator-prey system is viewed within thecontext of the other species that may influence the interaction, thus allowing for a relative ranking of the importanceof selective agents. Detailed documentation of the natural history of living predator-prey systems is recommendedin order to distinguish the processes in some fossil systems. A geographic view of species interactions and theprocesses driving their evolution may lead to a more diverse array of testable hypotheses on how predator-preysystems evolve and what constraints interactions impose on the evolution of organisms. Scale is important inevaluating the role of escalation and coevolution in the evolution of species interactions. If short-term reciprocaladaptation (via phenotypic plasticity or selection mosaics among populations) between predator and prey is acommon process, then prey are likely to exert some selective pressure over their predators over the short term (onecological time scales), but in the long run predators may still exert primary “top-down” control in directingevolution. On the scale of evolutionary time, predators of large effect likely control the overall directionality ofevolution due to the inequalities of predator and prey in control of resources.“Every animal has its enemies, and Natureseems to have taxed her skill and ingenuity to theutmost to furnish these enemies with contrivancesfor the destruction of their prey…For everydefensive device with which she has armed ananimal, she has invented a still more effectiveapparatus of destruction and bestowed it uponsome foe, thus striving with unending pertinacityto outwit herself.” —Forbes, 1887INTRODUCTIONTHE ARMS RACE CONCEPT is all toofamiliar to those of us who have grown up duringthe latter half of the twentieth century. Morecontroversial, however, is the idea that evolutionaryarms races between interacting species (such aspredator and prey) have characterized the historyof life on Earth. Predation is a nearly universalpressure affecting individual animals as well as theorganization of ecosystems. Darwin’s metaphor ofthe “struggle for existence” indicates that he viewedbiotic struggle (or interactions within and betweenspecies) as a major evolutionary force in the historyof life: “The relation of organism to organism isthe most important of all relations” (Darwin, 1859,p. 477). Arms races between species overevolutionary time are likely wherever individualorganisms have enemies with a capacity forevolutionary response (Dawkins, 1986). Suchinteraction (from an economic standpoint) is arguedto drive or shape the evolution of life (Dawkins,1986; 1995; Vermeij, 1987; 1999).Dawkins (1986) credited H. B. Cott withhaving been the first author to apply the arms-raceanalogy to the biology of predator-prey systems:“…In the primeval struggle of the jungle, asin the refinements of civilized warfare, we see inprogress a great evolutionary armament race—whose results, for defense, are manifested in such353

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002devices as speed, alertness, armor, spinescence,burrowing habits, nocturnal habits, poisonoussecretions, nauseous taste; and for offense, in suchcounter-attributes as speed, … ambush, allurement,visual acuity, claws, teeth, stings, …[and] poisonfangs. Just as greater speed in the pursued hasdeveloped in relation to increased speed in thepursuer; or defensive armor in relation to aggressiveweapons; so the perfection of concealing deviceshas evolved in response to increased powers ofperception” (Cott, 1940).Arms races are usually envisioned as attackdefense“games” that are slowly played out overlong periods of evolutionary time, graduallymolding species interactions, especially inpredator-prey systems (Futuyma and Slatkin,1983a; see also the recent popular account by Levy,1999). Thus it is generally thought that, as the quoteby Cott suggests, offensive adaptation on one sideis countered by defensive adaptation on the otherside and vice versa. To paraphrase Dawkins andKrebs (1979), claws get stronger, so shells getthicker, so claws get stronger still. But is thisreciprocity the only, or even the most likely, waythat the elaborate feeding structures of predatorsand the defenses of their prey evolve?The arms race concept has emerged in severalrelated forms in the modern literature of evolutionaryecology and paleoecology, including through thehypotheses of coevolution and escalation. Both ofthese processes assume that biological factors aremajor agents of natural selection and that organismsare able to respond evolutionarily to selective factorsimposed by other organisms in the environment.Coevolution is defined traditionally as theevolution of two or more species in response toone another (Futuyma and Slatkin, 1983b). Theterm “coevolution” was first used to describe theevolution of species interactions by Ehrlich andRaven (1964). Various definitions have beenascribed to coevolution; in the strict sense the termhas been applied to reciprocal adaptation of species,in which each species evolves in response to theother (Fig. 1). The interacting species may bepredator and prey, competitors, parasite and host,or mutualists. The term “diffuse coevolution” hasbeen applied to interactions involving more thantwo species (for instance, a predator with multipleprey species).“Arms races” may also occur withoutinvolving coevolution, as argued by Vermeij (1983,1987, 1994). Vermeij’s hypothesis of escalationclaims that biological hazards have become moresevere with time, and adaptations to those hazardshave increased in expression. The hypothesis ofescalation considers the most significant selectiveagent to be an organism’s enemies. However,adaptation need not be reciprocal (Fig. 1). Vermeij(1987) has argued that prey respond to theirpredators, but that predators are more likely torespond to their own enemies (for instance, theirpredators) than they are to their prey. Thusadaptation is unilateral. However, coevolution mayoccur if both escalating parties are enemies (suchas in the case of a predator and dangerous prey, orparasite and host). Escalation has two components:the “gap” between an organism and its biologicalenvironment (or hazard) and the level of the hazard.Escalation occurs when the level of the hazardincreases and the gap narrows. In other words,escalation occurs when adaptations to the hazardbecome better expressed.Thus, in coevolution, the claws of the predatorget stronger and the prey’s shell becomes thicker inreciprocal response. In escalation, increased defensein the prey is a response to the stronger claws of itspredator, but the increased claw strength of thepredator is a response to agents other than the prey.Coevolution and escalation clearly are related butthey differ in what agents are responsible forselection when it occurs (Fig. 1) (Vermeij, 1994).The distinction between coevolution and escalationis not always clear in the literature, in part becausethe term coevolution has sometimes been definedto encompass the concept of escalation. Forinstance, prior to Vermeij’s introduction of the termescalation for enemy-driven evolution, thisphenomenon was described as “unilateralcoevolution” (Futuyma and Slatkin, 1983b).The fossil record of invertebrate predator-preysystems has excellent potential for testing arms racehypotheses, but it is not always straightforward to354

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESFIGURE 1—Direction of selective pressures in coevolution and escalation of predator-prey interactions.determine what process was responsible for thepattern we observe. Such tests require a)demonstration that two species interacted aspredator and prey, and b) evidence of evolution ofone or both members of the system in response toone another. Chapters in the first two sections ofthis volume describe various lines of evidence andmethods used to analyze predation in the fossilrecord. In this chapter, we discuss the dynamics ofthe evolutionary processes responsible for the historyof predator-prey systems. We assess the likelihoodof the occurrence of coevolution and/or escalationfor several fossil invertebrate predator-prey systems,in order to address the fundamental question: Arethere any general “rules,” or overriding principles,that govern the ecological and evolutionarytrajectories and outcomes of predator-preyinteractions (see also Herre, 1999)? We also explorethe effect of scale in our perception of processesand patterns of predator-prey evolution.MODELS OF COEVOLUTIONAND ESCALATIONRed Queen Hypothesis.—The extreme view ofan arms race is represented in Van Valen’s “RedQueen Hypothesis.” Van Valen (1973, 1976) arguedthat adaptation by one species has a deleterious effecton all other species within its effective environment;this idea is an extension of Fisher’s (1930) view thatany well-adapted species will experience a“constantly deteriorating” environment, owing to“the evolutionary changes…in associatedorganisms.” As Van Valen (1976, p. 181) suggested:“A change in the realized absolute fitness of onespecies is balanced by an equal and opposite changein the realized absolute fitness of all interactingspecies considered together.” Thus a species mustbe running in place (continually adapting) simplyto survive in the context of a changing biologicalenvironment, even if no change occurs within the355

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002abiotic environment. If species do not adapt theywill eventually go extinct. A world run on Red Queenconditions would involve continuous coevolution,in the form of a series of progressive “linear”improvements or endless cycles of defenses andcounterdefenses (Parker, 1983; Dieckmann et al.,1995) among the interacting species (Thompson,1986; Stenseth and Maynard Smith, 1984).Slatkin and Maynard Smith (1979; see alsoKitchell, 1982) questioned why predators do notevolve the ability to overexploit their prey and whyprey defenses do not outpace predator capabilities.If either side gains the advantage in adaptation theinteraction will terminate. Early attempts atmodeling the process of continuous coevolutionpredicted by the Red Queen Hypothesis assumedthat adaptation was not constrained, thatcoevolving systems had to be stable, and that ratesof change in prey and predator were balanced—this leads to an endless arms race in which neitherspecies is expected to “win the race” (Rosenzweig,1973; Schaffer and Rosenzweig, 1978; Slatkin andMaynard Smith, 1979; Roughgarden, 1983). Thistype of approach was not motivated by abstractconsiderations of predator-prey systems, but byobservations that many morphological traits arerelated to feeding or defense and that predator-preysystems do not seem to undergo frequent severepopulation fluctuations and extinctions in nature(Slatkin and Maynard Smith, 1979; Murdoch andOaten, 1975).Although the arms-race view of continuousevolution is a valuable heuristic tool (Thompson,1986), it has its limitations. Rosenzweig et al.(1987) pointed out that the conditions required forthe continuous evolution predicted by the RedQueen Hypothesis may seldom occur in nature. TheRed Queen requires the existence of phenotypictraits that are independent and boundless(Rosenzweig et al., 1987). In other words,phenotypes take more and more extreme values asa result of directional selection (or cycle endlesslythrough maintenance of multiple defenses andcounterdefenses by frequency-dependent selection).For example, the Red Queen would predict that therelatively weak claw of the crab Calappa shouldevolve to infinitely more efficient prey-crushingshapes that would increase its strength (ormechanical advantage). However, claw shape ofCalappa is under other selective pressures fordigging ability and maintaining respiratorypathways for water currents while the crab is buried(Hughes and Elner, 1989). The additional functionof the claws in digging compromises limitless orinfinite evolution of claw shape (and strength) inthis predator. Calappid crabs also have a highlyspecialized, hooked tooth (an adventitiousstructure) on their right claw that is used to peelthe lip of gastropod shells (Shoup, 1968; Vermeij,1982a), and to break the shell of some mussel prey(Hughes and Elner, 1989). The adventitiousstructure of the claw, which has a high mechanicaladvantage, may serve to compensate for theotherwise weak claw (in terms of what type of preycan be crushed successfully).An evolutionary race may stop if selection forgreater prey defenses or predator offenses shouldbe counteracted by their greater costs or tradeoffswith other competing functions (Futuyma, 1986;see also Vermeij, 1987, 1994; Brodie and Brodie,1999). A potential reason for such costs is thatdefensive characters that increase survivorshipagainst one predator may make the species morevulnerable to others (or to abiotic change). Forexample, the degree of expression of induceddefense (see below), such as shell thickness, in themussel Mytilus edulis is predator specific. Musselsproduce thicker shells in response to the shelldrillingand shell-crushing predators, Nucellalapillus and Carcinus maenas, respectively; theincreased-thickness response of Mytilus is greaterto Nucella (Smith and Jennings, 2000). In contrast,mussels exposed to the crab Carcinus grow largershells than those exposed to Nucella. Thus, theremay be a tradeoff between shell thickness and shellgrowth that may increase the time it takes Mytilusto achieve a size refuge from predation. Thespecificity in response may create conflictingdefenses (Smith and Jennings, 2000). Similarly, thecapabilities of the predator for handling one preyspecies may come at the expense of handling others(Futuyma, 1986). Although adaptation by natural356

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESselection is a powerful force in shaping thephenotypic traits of organisms, constraints (the“spandrels” of Gould and Lewontin, 1979), whichare common and sometimes inescapable featuresof living systems, may also be important in limitingevolutionary response in an arms race.As a result of these costs, tradeoffs, andconstraints, an adaptive “stalemate” may be a morelikely outcome of an arms race than the continuousevolution predicted by the Red Queen (Rosenzweiget al., 1987; Vermeij, 1994). These limits to infiniteadaptation suggest that most of the history ofselection in an interaction may be more appropriatelymodeled with a stabilizing function, rather than thedirectional (i.e., the stronger the claw the better)function usually applied to arms races (Brodie andBrodie, 1999). However, these limits to adaptationcan be broken, leading to directional change thatcan be tracked in the fossil record—if changes inthe rules governing adaptive compromise areintroduced (see Vermeij, 1973; 1987; 1994). In thisway, selection is viewed as an episodic rather thancontinuous process (Vermeij, 1994) acting oninteractions among species.The nature of evolutionary trends predicted bythe Red Queen.—The Red Queen’s prediction oflinear (or “lock-step”; Bakker, 1983) evolutionarytrends in predator and prey has led some authors toconclude that interactions among species are notdriving evolutionary change if the expected patternis not evident in the fossil record (Stanley et al.,1983; Bakker, 1983). For example, Boucot (1990,p. 562) stated that, “after the geologically rapid,initial relation has been established [betweeninteracting species] the fossil record suggests thatthere is no subsequent, coevolutionary change, i.e.,stabilizing selection sets in.” To address this type ofskepticism, DeAngelis et al. (1984) developed acoevolutionary model of the energetics of thepredator-prey interaction between drilling naticidgastropods and their bivalve prey. In contrast toearlier models of coevolution, their model wasdeveloped with its empirical utility in mind. (Thefossil record of naticid predation is extensive andhas been key to tests of the importance of bioticinteractions in evolution—see Vermeij, 1987; Kelleyand Hansen, 1993, 1996.) Their model incorporatedan explicit potential for coevolutionary feedbackthrough size effects (Kitchell, 1986), based on sizedependentvariation in the outcome of successfulpredation (Kitchell et al., 1981). The predator wasassumed to maximize its energy intake per unit timeof foraging and the prey its allocation of energy toreproduction and defense. The models assume thatthere is a tight link between predator and prey. (Theunderlying assumption of optimality theory is thatnatural selection favors those individuals that aremost efficient in their behavior.)In the first version of the model, only theinfluence of increasing naticid predation on theallocation of bivalve energy among reproduction,overall growth in size, and shell thickness wasanalyzed (DeAngelis et al., 1985). Simulation resultsshowed that, as predation intensity increased, aninitial single bivalve defense (represented as a peakin an adaptive fitness landscape) changed to threedifferent strategies (or peaks) that varied in theamount of energy diverted to shell growth andthickness. The three alternative means of dealingwith predation are: 1) postponed reproduction(effectively running the predation gauntlet as theprey tries to grow quickly into a size refuge frompredation); 2) early reproduction coupled with someallocation of energy to thickness increase; and 3)significant allocation of energy into thickness as adefense to minimize selection by the predator(DeAngelis et al., 1985).In a later version of the model (summarized inKitchell, 1990) predator size was allowed to evolvesimultaneously with the prey traits (thusmaximizing both predator and prey fitnesses asinterdependent dynamic responses). A two-wayfeedback was thereby introduced. As traits in theprey varied to increase prey fitness, the prey in turnaffected the adaptive landscape of the predator andcaused it to change its own traits in order tomaximize its own fitness. These changes in thepredator’s traits (size), in turn, affected the prey’sadaptive landscape, causing it to adjust its ownevolutionary trajectory.Simulation results suggested that the potentialexists for both stasis and change within the dynamics357

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002of a positive-feedback system of coevolving bioticinteractions, and the changes may be sudden anddiscontinuous as well as gradual and continuous(Kitchell 1990). This view contrasted markedlywith Futuyma and Slatkin’s (1983b) conclusionthat the “ideal paleontological evidence [forcoevolving lineages] would be a continuous depositof strata in which each of two species showsgradual change in characters that reflect theirinteraction.” In this restricted view of Futuyma andSlatkin, if prey shell thickness is a deterrent topredation, increased predation pressure shouldfavor the evolution of thicker shells in the prey.Similarly, if the predator increases in size (invadingthe size refuge of the prey), general models predictan increase in prey size as the likely evolutionaryresponse. The model results summarized inKitchell (1990) suggest that these examples are notthat straightforward, and general theory can predictprey responses other than increased size or shellthickness, even without any change in the directionof selection. Thus the lack of the intuitivelyexpected “linear” response in arms races betweenpredator and prey cannot be used as evidence thatbiotic interactions were unimportant in evolution.When is it coevolution and not escalation?—Although the analogy of an arms race incoevolution between predators and prey has beenwidely used (and assumed) to characterizepredator-prey interactions, such studies often lackempirical evidence that the predator respondsevolutionarily to its prey (see Brodie and Brodie,1999). Despite the assumption of a tight reciprocallink between naticid gastropods and their bivalveprey in the models of DeAngelis et al. (1984, 1985),there is no empirical evidence of reciprocaladaptation. Kelley (1989) interpreted the changesin thickness of the Miocene bivalves she studiedas a response to naticid predation; species preyedupon most heavily displayed the greatest increasein thickness. However, evolutionary changes in thepredator are more interpretable as defenses againstthe gastropod predator’s own enemies. Kelley(1992) did not find any significant trends in mostcharacters thought to affect predator efficiency. Onecharacter of the predator that could be interpretedas a reciprocal response to evolution in the prey wasa size increase in the predatory genus Neverita (sizewas an important morphological character in themodels of coevolution discussed previously). Kelley(1992) pointed out, however, that the size increasein Neverita also could be interpreted as anevolutionary response to its own enemies. Thisinterpretation is supported by the fact that thepredator’s shell thickness also increased, most likelyas a response to the naticid’s own predators.The study of the interaction between crabs andgastropods in Lake Tanganyika, Africa, alsohighlights the lack of empirical evidence for predatorresponse to prey. In general, Tanganyikan gastropodshave thicker, more ornate shells (unusual forfreshwater species; Vermeij and Covich, 1978) thanother closely related lacustrine taxa from outside thelake. They also display a considerably higherincidence of shell repair (a measure of theeffectiveness of the shell as a defense) in responseto unsuccessful predatory attacks by crabs (West etal., 1991). The endemic Tanganyikan crabs possesslarger, more robust crushing claws than other Africancrabs. A reciprocal coevolutionary arms race wasinvoked to explain the observed pattern (West et al.,1991, p. 605): “To protect themselves fromdurophagous predators, Tanganyikan gastropodshave increased their shell size, strength, andsculpture. Tanganyikan crabs have concordantlyincreased their shell-crushing capacity with largerobust chelae lined with broad molariformdentition”—it is the argument that claws getstronger, so shells get thicker, so claws get strongerstill. While it is clear that prey shell charactersevolved in response to selection from predators, acorrelation between the morphological features ofthe predator and prey does not unequivocallyestablish reciprocal selection and evolutionbetween predator and prey. It is equally probablethat the prey responded evolutionarily to the crabs,but the crabs, instead of responding to their prey,evolved in response to competition with other crabsfor prey, space, and/or mates (an escalationinterpretation; Vermeij, 1978).McNamara and Long (1998) implicated acoevolutionary arms race in the evolutionary trends358

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESseen in cassid gastropods and spatangoid seaurchins. Cassids, such as Cassis tuberosa, captureand immobilize their prey by arching and extendingtheir foot over the top of the urchin’s test (Hughesand Hughes, 1981). McNamara and Long (1998)assumed that evolutionary increase in size of theurchin Lovenia was driven by selection pressure toreach a size refuge from cassid predation, and thatreciprocal selection for increased capture efficiencyof larger urchin prey led to an evolutionary increasein predator size. However, it seems equally likelythat the cassids evolved larger sizes in response totheir own predators or competitors.Another example includes the highlyspecialized appendages of gonodactylidstomatopods that function effectively as hammers(Caldwell and Dingle, 1976). These mantis shrimpattack hard-shelled prey such as gastropods,bivalves, and crabs. Prey are caught and thenhammered by repeated strikes of the raptorialappendages. As with the Tanganyikan gastropodcrabinteraction, a coevolutionary arms race iseasily envisioned in this system. However, theseappendages are also used in agonistic encounterswith conspecifics. The appendages are oftenemployed during territorial contests for sheltercavities, which often result in serious injury toconspecifics (Berzins and Caldwell, 1983). Theconsequences of injury to the raptorial appendagesinclude a reduction in fighting ability, which affectsthe outcome of territorial contests (Berzins andCaldwell, 1983); injury also increases thelikelihood of falling victim to cannibalism andpredation, as well as lost mating opportunities.Which interpretation is correct? The fact that theseappendages are much less developed amongstomatopods that occupy soft substrates (and donot compete for cavities; Dingle and Caldwell,1978) suggests that response to prey is lessimportant than response to enemies.Whether or not escalation (in which mostevolutionary change in predators is driven by theirown enemies and not by their prey) explains theevolution of the predator-prey systems discussedabove, the hypothesis does highlight theuncertainty with which the coevolutionary armsrace analogy applies to some predator-preysystems. The theory of why predators should beexpected to respond to the selective pressures oftheir enemies rather than their prey is discussed inthe next section. Unfortunately, there have beenfew detailed studies of the evolution of speciesinteractions that extend their scope beyond thesupposed tight interaction of predator and prey toconsider also the role of the predator’s enemies.Visualization of selection pressure on predatorand prey.—Brodie and Brodie (1999) developed anapproach in which selection is formalized as thecovariance between traits and fitness. Such anapproach is designed to test whether predatorsshould respond reciprocally to their prey (that is, totest whether prey are actually enemies of thepredator). Selection is stronger when the covariancebetween phenotype and fitness is highest, andselection is weaker when the covariance is low. Inpotential coevolutionary interactions, the covarianceof interest is between the fitness of individuals ofone species (e.g., predators) and the values of traitsat the phenotypic interface in the other species (e.g.,prey) (Brodie and Brodie, 1999).Selection that results from an interactionbetween predator and prey is viewed by regressingpredator fitness (or some function of the predator’straits) on the phenotype of the prey species withwhich it interacts (Fig. 2) (Brodie and Brodie, 1999).In biological terms, the slope of the regression linereflects the rate of increase or decrease in theexpected fitness of the predator as prey morphologychanges. A steep slope (Fig. 2A, 2B) indicates thatthe effects of the interaction are strong (i.e., changein prey morphology drastically affects predatorfitness). The amount of scatter around the regressionline reflects the predictability of predator fitness fora particular prey phenotype (Brodie and Brodie,1999). Selection is strong (and hence coevolutionis likely) when the covariance is high betweenpredator fitness and prey morphology.Without experimental evidence from livinganimals to indicate whether fitness costs associatedwith feeding are likely, an analysis based solelyon fossils can only assume that predators are tightlylinked to their prey. The interaction between359

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 2—Covariance between predator fitness and prey phenotype. Redrawn from Brodie and Brodie(1999, Figure 1, p. 560). A and B, see text. C, low but predictable fitness consequences (low slope butnarrow spread of points). D, low and unpredictable fitness consequences (low slope, wide spread).predatory whelks of the genus Busycon and theirtightly closing bivalve prey, Mercenaria, can beused to highlight the utility of the covarianceapproach to selection. Where they occursympatrically, more than 70% of adult Mercenariamortality often can be attributed to whelk predation(Chestnut, 1952; see also Kraeuter, 2001). Whelksactively use their shell lip to chip open their prey,a mode of attack that often inflicts shell damageon both predator and prey. Breakage must berepaired and this represents a diversion of energyaway from somatic growth or reproduction. Whelksthat often break their shells while attempting tofeed are faced with a potential tradeoff betweensubsequent reproduction and shell repair (Geller,1990). If feeding rate is reduced significantlyduring repair of the injury, breakage also may leadto a net energy loss (or negative energy budget)for the whelk resulting in a reduced growth rate.Breakage of the whelk’s shell may thus have severefitness consequences because it increases the timethe individual spends at a smaller size, leaving itmore vulnerable to its own predators like the stonecrab Menippe mercenaria (Magalhaes, 1948; Kent,1983). The selective consequences for predatorywhelks interacting with Mercenaria of varyingsizes are predictable (Fig. 2A) (Dietl, in review).Predators that select large bivalve prey increasethe likelihood that their shell lip will be broken inthe process of attempting to open their prey.Adaptive response of the predator to the prey ispossible in this system of interacting predators andprey, with predators expected to evolve increasedeffectiveness in opening prey (perhaps representedas a temporal decrease in the frequency of repairedfeeding-inflicted shell breakage). Traces of thepredation event are retrievable from the fossilrecord (Dietl, unpublished data); ongoing work isexamining the origin(s) of the chipping mode ofpredation and the evolutionary history of thepredator-prey interaction.Vermeij (1994) argued that the central issue inthe debate about escalation and coevolution is thenature of selection; most current models of360

DIETL AND KELLEY—PREDATOR-PREY ARMS RACEScoevolution do not consider the process ofselection. “In order to make headway in the studyof coevolution and escalation, we need to studythe sources, frequencies, and cost-benefit effectsof selection. This entails careful observation ofencounters between species, together with anevaluation of the effects of such encounters onsurvival and reproduction” (Vermeij, 1994, p. 232;see also Herre, 1999, and below). It is importantto distinguish when selection occurs and whatagents are responsible (Vermeij, 1994). For aninterpretation of reciprocal adaptation to beappropriate, selection by the prey on its predatorhas to be stronger—with predictable consequences(Brodie and Brodie 1999)—than that caused byother enemies of the predator.Inequalities in interaction.—Slobodkin’s(1974) expectation that prey in an evolutionary racecan “keep ahead” led Dawkins and Krebs (1979)to point out that coevolution (reciprocal adaptation)between predator and prey gives the prey aninherent evolutionary advantage. Thus predator andprey are not equal partners from an economic pointof view in an arms race. They termed thisasymmetry in the evolutionary outcome betweenpredator and prey the “life-dinner principle.” It isa widely cited explanation for why predators maynot respond evolutionarily to prey adaptation(Vermeij, 1982b; Kelley, 1992; Brodie and Brodie,1999). If a predator fails, it only loses a meal (andsome energy and time), whereas failure for the preymeans death (or at least injury). The consequencesof the interaction clearly are more severe for theindividual prey, suggesting that selection by thepredator on the prey for improved defenses isstronger than selection by the prey on the predatorfor improved offensive capabilities. Asymmetry inselection is assumed to result in faster evolutionof prey than predators, leaving predators unable tokeep pace (in the evolutionary race) with their prey.Vermeij (1982b) reasoned that the prey’scontribution to selection for improved predatorcapability is further reduced if a predator preys onmore than one species; the short-term response ofthe predator might be to avoid the increasingly welldefendedprey. Vermeij (1982b) took this to implythat selection for predatory improvement is likelyto be far stronger from the predator’s own enemiesthan from its prey; thus interactions among multiplespecies decouple the coevolutionary responses of agiven pair of species. This was the major impetusfor the formulation of the escalation hypothesis andits view that prey play a minor role in directing theevolution of their predators.Dawkins (1982) proposed the “rare enemyeffect” as a means of balancing the asymmetryinherent in the life-dinner principle and allowingcoevolution to occur. Because rare predatorsrepresent a minor selection pressure on the prey,the predators are thought to “win” or “keep ahead”in the arms race (Dawkins, 1982). However, if apredator is rare, the risk of predation for anyindividual prey is low, and the prey would not beexpected to evolve costly morphological oravoidance defenses. If selection pressures on theprey are weak, coevolution of predator and prey isdifficult to envision (see also Abrams, 1986). Insuch a case, an opposite asymmetry may arise inwhich the prey do not respond to evolution of therare predator.Abrams (1986) has argued, on theoreticalgrounds, that the life-dinner principle does not takeinto account the fact that if a predator loses enoughraces, it will also lose its life. Predators that rarelycatch prey will suffer the same fate (death) as anunsuccessful prey individual. Furthermore, thefrequency or probability of attacks, as well as theiroutcome, has to be taken into account indetermining the net lifetime effect on survival,reproductive output, and fitness. The cost perunsuccessful attack is typically low for predators,but predators typically have a low success rate andmust make many unsuccessful attacks for everysuccessful one. In contrast, the cost to a prey ismuch higher, but the number of attacks experiencedper individual may be much lower over a lifetime.Thus, even though the consequences of oneindividual attack are highly asymmetrical betweenpredator and prey, the net outcome is not readilyobvious (J. A. Rice, pers. comm., 2002).In addition, Brodie and Brodie (1999) pointedout that the life-dinner principle, as formulated,361

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002oversimplifies the process of phenotypic evolution,which is a combination of selection and inheritance.Evolutionary change in a trait is equal to theproduct of the strength of selection on the trait andits heritability (the percentage of variation in a traitthat is controlled by additive effects of genes; seeFutuyma, 1986). Thus if shell thickness of the preyand claw strength of the predator were toexperience exactly the same strength of selection,the species with the greater heritability for thetrait in question would evolve faster. If selectionis unequal, as the life-dinner principle predicts,the rate of evolutionary change may still be equalif heritabilities differ. If selection is stronger onprey than on predators, prey should have lessgenetic variation in their defenses, based onFisher’s fundamental theorem of natural selection,which states that the rate of increase in fitness islimited by the amount of additive genetic variation(Brodie and Brodie, 1999). If this is the case,heritability differences might balance theasymmetry in strength of selection predicted bythe life-dinner principle, resulting in comparablerates of evolution.Brodie and Brodie (1999) noted that a majordifference between predator-prey interactions andother victim-exploiter systems is the intimacy ofthe interaction (i.e., the extent to which each speciesexperiences the consequences of the interaction).In predator-prey interactions, prey that do notsuccessfully escape face the direct consequence ofinteraction with the predator; that is, death (or zerofitness). By contrast, predators, in their interactionwith prey, are able to avoid many of the selectiveconsequences that would occur in a more intimateinteraction; this inequality in interaction may helpexplain why selection on prey is thought to bestronger than on predators (Brodie and Brodie,1999). The consequences for the individualpredator of losing a prey may be strong enough tostart an arms race. But the predictability of thisconsequence for any individual predator is low(Fig. 2B); a predator might capture the very nextprey it interacts with, negating the consequencesof the prior interaction, or may switch to analternative prey (Brodie and Brodie, 1999).Brodie and Brodie (1999) argued that the onlypredator-prey systems likely to result in acoevolutionary arms race (in which selectionpressures between predator and prey aresymmetrical) are systems in which predators interactwith dangerous prey, a conclusion Vermeij (1982b)also advocated. If a prey is dangerous to a predator,the predictability of consequences for the predatoris expected to be high (small residuals) (Fig. 2A)and therefore selection is strong. Coevolution islikely to occur in this system of interacting predatorsand dangerous prey. In this sense predators are“forced” into experiencing selection from dangerousprey (Brodie and Brodie, 1999). This situation leadsinevitably to an evolutionary response in the predatoras long as variation in the predator’s offensive trait(either morphological or behavioral) is present.The dangerous prey concept was used toexplain evolutionary shifts in stereotypic placementof holes drilled by naticids on confamilial prey(Dietl and Alexander, 2000). Because an escapedaggressive prey may become the predator and drillits former attacker (a case of the hunter becomingthe hunted), naticid prey may be considereddangerous to their confamilial predators. Naticidsmay have shifted the position of drill holes onconfamilial prey in order to neutralize thepotentially larger prey foot or more aggressiveretaliatory behaviors of such dangerous prey (Dietland Alexander, 2000). This concept remainsrelatively unexplored in the fossil record ofpredator-prey interactions.The interaction between predatory crabs andtheir hard-shelled prey also may have acoevolutionary component. If we were to playDevil’s advocate, West et al.’s (1991) example ofthe interaction between Tanganyikan gastropodsand crabs discussed earlier could involvecoevolution between predator and prey if theevolution of more powerful claws of the predatorwere a counter-adaptation to increased defensivestrength of the prey’s shell. This could result fromthe increased likelihood of breakage or damage totheir claws (as a consequence of fatigue failure),which occurs commonly in living crab species(Juanes and Hartwick, 1990); in this case, the362

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESTanganyikan gastropods would be dangerous prey.The consequences of claw breakage for thepredatory crab Cancer magister, feeding on thehard-shelled bivalve Protothaca staminea, includelonger prey-handling times (decreasing feedingrates) and reduced growth, molting ability, andmating success (Juanes and Hartwick, 1990).Careful observations are required to determinewhen claw breakage frequency, and hence selectionpressure, increases, in order to clarify the processesshaping these interactions.TIME SCALE AND ECOLOGICALAND EVOLUTIONARY ARMSRACESGeographic mosaic theory of coevolution.—Scale is important in ecology and evolution. Forinstance, Futuyma and Slatkin (1983c) wereconfounded by an apparent lack of directionaltrends in the fossil record of predator-preyinteractions, despite the documentation of adaptationbetween species at ecological scales. Rapid,fluctuating, short-term (tens to hundreds of years)selection (frequency- or density-dependent) thatdoes not result in the production of long-term trendsis thought of as evolutionary “noise” (Thompson,1998). However, these short-term fluctuations maybe viewed as the first steps towards long-termdirectional selection (the microevolutionaryprocesses that operate on ecological time scalesmay have macroevolutionary consequences). Sucha link between scales can be achieved if short-termfluctuating selection leads to adaptive innovations(or changes in the developmental sequence thatbreak genetic covariances—or constraints onevolution—allowing traits to vary independently),or if this selection establishes new mutuallybeneficial partnerships among species (see alsoVermeij, 1994).At the interface between ecology and evolution,Thompson (1988, 1994) has addressed the role thatvariation in the outcome of interspecific interactionson an ecological time scale plays in the coevolutionof species interactions. The existence of geographicvariation both in the expression of defenses and inthe selection pressures caused by different groupsof enemy species is thought to produce a selectionmosaic for evolutionary arms races betweenantagonists, such as predator and prey (Thompson,1994). Thompson termed his hypothesis thegeographic mosaic theory of coevolution. In general,it is hypothesized that coevolutionary dynamics aredriven by components of geographic structure: thereis a selection mosaic among populations that favorsdifferent evolutionary trajectories to interactions indifferent populations, and there are coevolutionary“hot spots,” which are subsets of communities inwhich reciprocal adaptation actually occurs (Fig. 3).Selection mosaics occur when natural selection oninteractions and gene flow between populations vary(Gomulkiewicz et al., 2000); the apparent selectiveasymmetry due to the life-dinner principle inpredator-prey arms races discussed earlier can beeither increased or decreased by consideration ofselection mosaics.Hot spots are communities in which interactingspecies have reciprocal effects on each other’sfitness; they are usually embedded withincommunities in which selection affects only one orneither of the interacting species (coldspots)(Gomulkiewicz et al., 2000). Because thereare differences in outcome among populations, aninteraction between two species may coevolve,affect the evolution of only one of the species in theinteraction, or have no effect on the evolution ofeither of the interacting species (Thompson, 1994).The geographic mosaic theory incorporates twoobservations that have often been used as evidenceagainst coevolution: lack of apparent coevolutionin the interaction between a local pair of species,and lack of biogeographic congruence in theinteracting species distributions (Thompson, 1994).Thompson summarized the basic idea of his theorywhen he wrote that the geographic mosaic theory“suggests that the coevolutionary process is muchmore dynamic than is apparent from the study ofindividual populations or the distribution ofcharacters found in phylogenetic trees (e.g., seep.285 in Brook s and McLennan, 1991). Adaptationsappear and are lost. Some populations becomehighly specialized for the interaction as others363

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002FIGURE 3—A geographic mosaic of coevolution between two interacting species. Circles representbiological communities. Arrows within circles indicate selection on one or both species in the interaction;arrow directions (within circles) represent different (co)evolutionary trajectories. Arrows between circlesindicate gene flow. A, Coevolution occurs in all populations in which the interaction takes place; B,coevolutionary hotspots (in which reciprocal selection occurs) are intermixed with populations in whichreciprocal selection is not occurring (cold spots). From Thompson (1999a, Figure 1, p. S4) with permissionof the University of Chicago Press.remain or become less specialized. Somepopulations may fall outside the geographic rangeof the other species, lose some of their adaptationsfor the interaction, and then later be drawn back intothe interaction. A few populations may temporarilybecome evolutionary ‘hot spots’ for the overalltrajectory of coevolution between the species,whereas other populations act as evolutionary sinks.The overall course of coevolution between any twoor more species is driven by this ever shiftinggeographic mosaic of the interaction” (p. 223).This geographic view of an arms race generatesthree general ecological predictions: populationsof a species will differ in the traits shaped bycoevolution; traits will be well matched in somecommunities and mismatched in others; and therewill be few species-level coevolved traits that aredistributed across all populations of a set ofinteracting species (Thompson, 1999a).Thompson’s mosaic view of coevolution maybe limited to highly specialized species interactions.In predator-prey interactions most morphologicalprey defenses are general adaptations that workagainst a number of different predators, and are notspecific to a single predator species (G. J. Vermeij,pers. comm., 2002). The same is true for theoffensive capabilities of the predator. For instance,size and thickness of the shell are two generaldefenses in molluscs that are effective against shellcrushingcrabs or fish, as well as shell-drilling snails,not to mention boring parasites (Vermeij, 1987).How does this generality in adaptation affect theecological predictions of the mosaic hypothesis ofcoevolution, which assumes a specialized or tightlylinked interaction between predator and prey? Forinstance, although the range of a predator may notoverlap the entire range of its prey, another predatorspecies that selects for the same prey defense mightbe present. In this case, the selection pressure forincreased thickness is always present. Thisselective reinforcement calls into questionThompson’s prediction that few traits will be364

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESdistributed across all populations of a set ofinteracting species. It is easy to imagine how thesame defense can be adaptive locally even if thesource of the predation pressure changes amongpopulations. The other ecological predictions mayhold in specific cases, but this level of variation islikely lost (or masked or misinterpreted as a highdegree of phenotypic lability) in the fossil record.Thompson’s (1994, 1998) mosaic view of theevolution of species interactions suggests howreciprocal evolutionary change can shape speciesinteractions that are asymmetric at the species level(one species interacting with multiple species—i.e., a diffuse interaction) (Fig. 1) but are specificat the local level. Thus, although a predator mayprey on a number of species, a single prey speciesmay be most important in driving the evolution ofthe predator. There is some theoretical backing forThompson’s argument. Abrams (1991) has shownfrom models that interactions between a predator,a prey species, and a third species (either thepredator’s own predator, or a second prey species)often increase the magnitude of coevolutionaryresponse in the predator or prey (or both) toevolutionary changes in the additional species.Thus multiple prey species may intensify theresponse of a predator to its main prey. If empiricaldata are found to support Abrams’ models, one ofthe tenets of the hypothesis of escalation may bebrought into question (i.e., that predators areunlikely to respond specifically to any one preyspecies because of the catholic diet of mostpredators; Vermeij, 1987, 1994).What are the implications of the mosaichypothesis for the fossil record of speciesinteractions?—In general, the mosaic view of theevolution of species interactions incorporatesvariation that occurs over a larger spatial scale (therange of the species) than we usually consider (i.e.,within a local population) when evaluating armsrace hypotheses. As paleontologists, we are usuallyonly able to trace traits that have spread among allthe populations of a species, but this does not meanthat the coevolutionary meanderings or continualreshaping of interspecific interactions areunimportant. Interactions may come and go withinthe ecological context of a local population throughtime. A prey population with a geographic rangethat extends beyond that of its predator may escapeattack for many generations, lose (or more likelydecrease) its level of defense, and then fall victimagain as the range of the predator expands. Just asa lack of evidence for local adaptation is notconclusive evidence that coevolution (or for thatmatter escalation) has not occurred, the lack ofdirectional trends through successive stratigraphicintervals in traits important to an interaction is notevidence that biotic factors were unimportant.Gould (1990, p. 22) commented that “positivefeedback [should drive the evolution of aninteraction] to its furthest point in a geologicalinstant, while most actual events [trends] span tensof millions of years”; positive feedback is thoughtto require “locking of biotic interactions” over theduration of the trend. In other words, how can longtermbiotically driven trends be accounted for ifreciprocal adaptation is expected to occur onlyearly in the history of the interaction (see alsoFutuyma, 1985)? The mosaic view of selectionsuggests that the degree of adaptation to localconditions may increase or decrease depending onthe amount of contact with other populations. Thespatial distribution of genetic variation “determinesthe degree and scale over which populations havebeen (or are) evolutionarily independent, andconsequently free to evolve in response to localvariation in selection” (Grosberg and Cunningham,2001, p. 61). In general, species with restricted geneflow are expected to exhibit local adaptation tospatially varying selection; as the level of gene flowamong populations with different selective regimesincreases, the selective costs of local adaptationincrease (Grosberg and Cunningham, 2001). Thus,the geographic structure of a species may favor longtermtrends in the evolution of species interactions.Kelley and Hansen (2001) further added that thecondition of “locking” or maintenance of specificpredator-prey interactions is not necessary ifescalation (which does not require a strongcoevolutionary component) is the most importantprocess in the evolution of species interactions.Few interacting species have identical365

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002geographic ranges. In addition, the outcome of aninteraction often varies among environments withthe varying context of the other species present. Inmodern habitats predation varies at different scales,from local to regional. Not surprisingly, then, manystudies have shown that predation pressure is alsohighly variable on both large and small spatialscales within time-averaged assemblages in thefossil record (Vermeij, 1980; Geller, 1983;Schmidt, 1989; Hansen and Kelley, 1995;Hoffmeister and Kowalewski, 2001; Nebelsick andKowalewski, 1999; Alexander and Dietl, in press).This sets up the conditions for a possible selectionmosaic of hot and cold spots that have fluctuatedin both space and time.The question that studies on spatial variabilityin predation have neglected to ask is how variabilityin the frequency of hot and cold spots affects therate of evolution of species interactions (or traitsinvolved in the interaction). While it is importantthat we evaluate temporal trends in coevolutionorescalation-related traits among ecologicallysimilar habitats (i.e., because local adaptation iscontext dependent; Vermeij, 1994), in doing so weignore the variation across other populations in theselection mosaic. The degree of connectionbetween populations may change as the frequencyof hot and cold spots (the mosaic of spatiallyvarying selection pressures) changes temporally.If the frequency of hot spots is low relative to coldspots, and the scale of gene flow exceeds the scaleover which selection varies, beneficial adaptivechanges (such as an increase in shell thickness)are unlikely to spread quickly to all populations ofthe species. Further work is needed to determinethe general applicability of the conclusion of Bushet al. (2002) that degree of variation in timeaveragedfossil assemblages mirrors withingenerationvariability of living populations; if theseresults hold, then geographic and temporal mosaicsof change should be observable in the timeaveragedfossil record.The outcomes of interactions also can varyamong environments as changes occur in theeffectiveness of traits involved in an interaction(Thompson, 1988). Thompson has called the rangein effectiveness of a particular trait an interactionnorm—by analogy with reaction norm, which isthe range of phenotypes a given genotype expressesamong environments. In order to understandselection and constraints on the evolution of speciesinteractions, patterns of geographic variation in theoutcomes of such interactions might have to beconsidered (see also Thompson, 1988). Vermeij(1982b) has shown convincingly that predatoryattacks are not always successful. It is this failurethat he argues is driving the evolution of adaptationsto enemies (Vermeij, 1982b, 1987, 2002).Geographic variation in sublethal predation thusprovides an opportunity to understand howcombinations of different traits affect the outcomesof interactions over evolutionary time (Vermeij,1982b), or favor the evolution of new associationsor innovations important to an interaction betweenspecies. How do different environments affectselection on interactions (or the interaction norm)?Thompson has argued that differences in outcomeamong environments could potentially be greatenough to shift the mean outcome along a continuumof antagonism, commensalism, and mutualism.Adaptive phenotypic plasticity.—Heterogeneityin predation pressures among habitats in ecologicaltime also favors the evolution of adaptive phenotypicplasticity. Inducible defenses are phenotypic changesinduced directly by chemical cues associated withbiotic agents. Induced morphological defensesinclude the production of antipredatory structures(bryozoans, cladocerans, rotifers) and protectivevariation in shell thickness (barnacles, gastropods,mussels) in the presence of the predator (Harvell,1984; Lively, 1986; Trussell, 1996, 2000; Leonardet al., 1999). For instance, thicker shells areinduced in the gastropod Littorina obtusata bythe presence of the predatory crab Carcinusmaenus (Trussell, 1996).Induced defenses are likely favored overconstitutive defenses because they account for theunpredictability inherent in most habitats. Inducibledefenses in the prey are thought to evolve via tworoutes: “cost-benefit” (=“inducible” in Fig. 4) and“moving target.” The cost-benefit model (Fig. 4)applies when four conditions are met: 1) Selective366

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESFIGURE 4—Constitutive, inducible (cost-benefit),and moving-target strategies. Circles representdifferent phenotypes; circle size represents growthrate in absence of predation and shading representsdegree of defense. Arrows indicate switching inpresence of predation and dashed arrows indicateswitching in absence of predation. See text forcomparison of cost-benefit (inducible) and movingtargetmodels. From Adler and Karbon (1994) withpermission of the University of Chicago Press.pressure of the inducing agent (the predator) mustbe variable and unpredictable, but sometimes strong.If the antagonist is constantly present, constitutivedefenses should evolve. 2) Availability of nonfatalpredictive cues is necessary to indicate the proximityof the threat of future attack and activate the defense.3) The defense must be effective. And 4) mostdefenses should incur direct allocation costs or othertradeoffs. The prey trade the risk of predation againstthe cost of defense.In the moving target model, changes in defense(against herbivory or predation) are induced as aform of nondirectional phenotypic escape fromadverse conditions rather than a ratcheting up ofdefenses (Fig. 4). Under this approach, preyphenotypes cannot be arrayed along a single axisfrom undefended to well-defended because theprey’s defenses are affected by a variety of extrinsicenvironmental factors other than the presence oftheir predator, each of which has an independentcomponent of effect on the predator (Adler andKarban, 1994). Consequently, there is a tradeoffin the effectiveness of different defensivephenotypes. It is predicted that prey will not fixtheir defense to one predator if that state leavesthem vulnerable to another predator. The tradeoffin the moving target case does not therefore dependon the cost of defense, but on the array of predatortypes that respond differently to different preyphenotypes (Adler and Karban, 1994). A movingtarget defense is favored if environments areunpredictable and uninformative (Adler andKarban, 1994). Thus environments that changefrequently in ecological time might not allow for aspecific inducible defense to be targeted byselection over evolutionary time.Predator responses to inducibly defended preycan br morphological, physiological, or behavioral,such as foraging movement strategies and lifehistory changes (Levin et al., 1990). Reciprocalphenotypic change has been observed in predatorpreysystems. Smith and Palmer (1994) showedthat morphology and claw strength of the predatorycrab Cancer productus was plastic. When crabswere fed mussel prey (Mytilus) with shells, theydeveloped larger and stronger claws than when fedmussels that had had their shells removed. Smithand Palmer (1994) suggested that these short-termadaptive responses to a changing environment(prey phenotype), if heritable, could produceevolutionary changes in claw size. Conversely,mussels respond to the presence of predators (e.g.,crabs) by inducing thicker shells (Leonard et al.,1999; Smith and Jennings, 2000). Agrawal (2001)argued that, because species interactions areintrinsically variable in space and time, and ifreciprocal phenotypic change is the result ofadaptive plasticity for both predator and prey, then367

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002it is reasonably predicted that coevolution mayresult in inducible defenses as opposed toconstitutive adaptations.If trends in predation-related traits involveinducible, rather than constitutive, defenses, ourinterpretation of such trends is complicated.Although evolutionary trends in both constitutiveand inducible defenses result from predation, themechanism differs. In cases of consistently highpredation pressure, natural selection will actdirectly to favor constitutive defenses (Tollrian andHarvell, 1999). However, if predation has a variableor unpredictable impact, inducible defenses willbe favored; in that case, selection acts indirectlyby affecting the norm of reaction (Travis, 1994).As the degree and variability of predation changes,defenses may alternate between constitutive andinducible or perhaps be lost entirely, depending onconstraints. This may affect what direction andwhat rate an evolutionary trend might take.Most studies have focused on plasticity fromthe prey’s point of view, as a defense against itspredators. However, adaptive plasticity not onlyprovides a mechanism through which prey mayavoid predation, but, in turn, may give predators theopportunity to respond reciprocally to overcome theprey defense. Thus, inducible defenses may havethe potential to alter both the short-term dynamicsand long-term evolution of predator-prey systems(Adler and Grunbaum, 1999; Harvell, 1990).However, it is unclear if in the long run predatorswould still exact primary “top-down” control indirecting the prey’s evolution as predicted by theescalation hypothesis (see also Vermeij, 2002).Communities that are dominated by generalizedpredators that are capable of switching amongdiverse prey, such as the crab Callinectes sapidusin marine soft-bottom habitats, often control thedistribution and abundance of a number of cooccurringspecies (Virnstein, 1977; Peterson, 1979;Hines et al., 1990).It is important here to distinguish reciprocalselection or within-generation change in phenotypedistribution from evolutionary reciprocal adaptation,which is a change in phenotype distribution acrossgenerations (Fisher, 1930). The conditions underwhich selection for induced prey defenses andpredator offenses operates in nature remain unclear.Is an inducible “dialogue” between predator andprey played out in ecological as well as evolutionarytime? If selection targets the norm of reaction ofboth predator and prey, coevolution may occur overthe long run (Smith and Palmer, 1994). There isexperimental evidence that partners in an interactioncontinually respond in a reciprocal fashion overecological time (Agrawal, 2001). Reciprocal changein ecological time may thus have resulted from longtermevolution in which the environment (the speciesinteraction) has been variable.FINAL REMARKSThere are two major underlying themes of thispaper. Our first goal was to clarify the conceptualdifferences between coevolution and escalation.The major difference between the two processesis in the nature of selection (Vermeij, 1994).Escalation is enemy-driven evolution. In this view,the role of prey (with the exception of dangerousprey) is downplayed in arms races betweenpredator and prey. In coevolution, prey are linkedtightly to their predator and are thought to drivethe predator’s evolution. Janzen (1980) posed thequestion: “When is it coevolution?” to drawattention to the ways in which the process wasmisunderstood. Similarly, we ask: “When is itcoevolution and not escalation?” The answer to thisquestion depends on the predator-prey system ofinterest. The naticid gastropod predator-prey systemwas initially envisioned as a coevolutionary system(Kitchell et al., 1981), but fossil evidence supportsan interpretation of escalation (Kelley, 1992). Insystems in which prey are dangerous to the predator,coevolution is the appropriate model (e.g.,confamilial naticid predation; Dietl and Alexander,2000; Busycon whelk predation on bivalve prey suchas Mercenaria; Dietl, in review). However, even inthese coevolving systems, the role other enemiesplay in reinforcing the selection pressure exactedby prey should not be overlooked—evolution doesnot take place in an “ecological vacuum” (sensuBoucot, 1983) even when considering a coupled368

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESpredator-prey interaction. These systems do,however, provide more definitive results becausetraces of predation are preservable in the fossilrecord. Crab-gastropod predator-prey systemstypically have been characterized as exhibitingcoevolution (Trussell, 2000; West et al., 1991), butalternative interpretations can be offered (seeabove); the same can be said for the cassidgastropod–sea urchin predator-prey interaction. Insuch cases, the question remains open as to whatprocess was important.Our discussion of coevolution models hasrepeatedly highlighted that approach’s main pointof contention with the escalation hypothesis; thatis, predators are not expected to respond to changesin their prey (the “decoupling” argument).Coevolution models (Abrams, 1986, 1991; Kitchell,1986, 1990) have assumed that positive feedbackbetween interacting populations, in terms of changesin population size or density, or energy intake, isthe same as reciprocal adaptation (Vermeij, 1994).But are changes in population dynamics or energyintake of interacting species an appropriaterepresentative of selection-based processes (see alsoVermeij, 1994)? If progress is going to be made inthe debate about coevolution and escalation wemust search for empirical evidence of evolutionaryresponses in nature. This requires not only adescription of the products of selection in terms ofbirths and deaths, or energy intake, but also anevaluation of how interactions among organismsaffect the opportunity for adaptation (Vermeij,1994). Interactions have consequences in the formof success and failure; traces of predation allowthe ranking of importance of various selectiveagents (Vermeij, 1987). “Those agencies that affecta large number of individuals … should … play alarger role in adaptive evolution than do agenciesthat affect a minority of individuals” (Vermeij,1987, p. 23). Distinction between the two processesin the fossil record will require documentation ofsources, frequencies, and cost-benefit effects ofselection (which for many systems requiresevidence from living animals; Vermeij, 1994).Our second theme is that rates of evolutionarychange of traits important in species interactionsmay vary from very short, rapid changes taking placeon the ecological time scale (tens to hundreds ofyears; Thompson, 1998) to longer spans ofevolutionary time (millions of years; Vermeij, 1987,1992). Trends over evolutionary time are also notnecessarily directional—although directionaltrends in antipredatory characters do occur,including increases in shell thickness (Kelley andHansen, 2001) or ornamentation (Dietl et al.,2000)—but may be highly dynamic in nature(Kitchell, 1990). These findings suggest that a lackof the predicted ideal unbounded, progressive trendis not sufficient evidence to argue against theimportance of interactions among species as adriving force in evolution (Kitchell, 1990). Despitethe apparent antiquity of many predator-preyassociations in the fossil record (Boucot, 1990),there are still few detailed studies testing arms racehypotheses. This does not imply that the inherentdirectionality (at the level of “megatrajectories”; seeKnoll and Bambach, 2001) in evolution isunimportant. Given that economic inequalitiesabound in nature between enemies (Vermeij, 1999),such direction is inevitable. Such directionality doesnot assume that within evolving lineages adaptationis boundless as predicted by the Red QueenHypothesis. There are likely to be periods ofdirectionality in an arms race (increasing meanvalues of traits) but also occasional reversals or evenperiods of stasis (Dawkins, 1986; Kitchell, 1990).Vermeij (1994) also espoused this view of theepisodic nature of selection.The fossil record is the only place where thelong-term effects of interactions among species canbe traced; thus it remains a valuable resource fortesting predictions based on these processes. If thegeographic mosaic process envisioned byThompson is a common feature of the evolutionof predator-prey interactions, then escalation andcoevolution studies may have to incorporate anunderstanding of the population structure of thespecies being studied. A geographic perspective inour approach to species interactions and theprocesses driving their evolution may allow for amore diverse array of testable hypotheses on howpredator-prey systems evolve (see also Thompson,369

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 20021994). How do connections among multiplepopulations affect the processes shaping theevolution of interactions among species? Byidentifying general patterns, we can betterunderstand the constraints that interactions imposeon the evolution of organisms. In addition, if shorttermreciprocal adaptation (via inducibility) betweenpredator and prey is a common process, then preyare likely to exert some selective pressure on theirpredators over the short term (and perhaps in specificcases over the long term; see above).At ecological time scales, reciprocal adaptationis likely to occur (Thompson, 1999b); coevolutionmay also characterize the long-term evolution ofspecific predator-prey systems. However, at thelargest scale of paleontological study, the escalationhypothesis may be the most important descriptionof the evolutionary process (see also Thompson,1999b). Over the long run, then, we believe, it islikely that predators control the overall directionalityin evolution (i.e., evolutionary megatrajectories)because of the inequality of predator and prey incontrol of resources (Vermeij, 1999, 2002). Predatorsof large effect influence community structure byway of their high rates of consumption and theirgeneralized diets (Virnstein, 1977; Peterson, 1979;Paine, 1980; Hines et al., 1990; Birkeland, 1996),and likely are “chiefly responsible for organizingthe economy, for defining the roles and attributesof the entities with which they interact, and forsetting the course of economic change” within thecommunity (Vermeij, 1999, p. 247).The fundamental question raised in theintroduction was whether there are any general“rules” that govern the ecological and evolutionarytrajectories and outcomes of interactions (see alsoHerre, 1999). In order to address this question, wemust recognize that predator-prey interactions arecomplex systems and that multiple factors mayinfluence the outcome of encounters betweenpredator and prey. Thus “it is important tounderstand the interactions among several differenttypes of species in order to provide the context toproperly pose and test evolutionary hypothesesabout any of them” (Herre, 1999, p. 235). Onceother agents of selection are considered, theintuitive expectation of the type of arms racedriving the evolution of a species interaction isoften called into question. Thus it is important toview any predator-prey system within the contextof the other species that may influence theinteraction, and to clearly understand the functionalinterrelationships among them. We have not cometo any firm conclusions on the “rules” governingthe processes of coevolution and escalation in anyspecific predator-prey interaction because in manysystems the question is still open. This does notimply that the governing “rules” that yieldexplanatory power concerning the outcomes ofspecies interactions and their long-term effects inevolution are not important. Instead, as Herre (1999,p. 236) pointed out, “understanding why the ruleswork in the cases that they do is crucial, as is theappreciation that context and scale determine theapplicability of those rules we presently recognize.”Distinction between escalation and coevolution canmost reliably be achieved with carefuldocumentation of the details of the natural historyof different systems (Vermeij, 1994). As Kohn(1989, p. 1095) eloquently stated: “Natural history,in focusing on the individual whole organism in itsenvironment, occupies a central position in thespectra of spatial and temporal scales appropriateto biological science.” Solutions to the conceptualconflicts between the coevolution and escalationprocesses fundamentally depend on growingcollaboration among ecologists and paleontologists.This collaboration remains an attractive butseemingly elusive goal; but with it will come amuch deeper understanding of the processes thathave shaped and continue to shape the evolutionof predator-prey systems.ACKNOWLEDGMENTSThis work was supported in part by a pilotproject grant from the Center for Marine Scienceat the University of North Carolina at Wilmington.We thank R. K. Bambach and J. A. Rice for helpfulreviews and G. S. Herbert and G. J. Vermeij forvaluable discussion. This is ContributionNumber269 of the Center for Marine Science.370

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESREFERENCESABRAMS, P. A. 1986. Adaptive responses of predators to prey and prey to predators: the failure of the arms-raceanalogy. Evolution, 40:1229–1247.ABRAMS, P. A. 1991. The effects of interacting species on predator-prey coevolution. Theoretical PopulationBiology, 39:241–262.ADLER, F. R., AND R. KARBAN. 1994. Defended fortresses or moving targets? Another model of inducible defensesinspired by military metaphors. American Naturalist, 144:813–832.ADLER, F. R., AND D. GRUNBAUM. 1999. Evolution of forager responses to inducible defenses, p. 259–285. In R. Tollrianand C. D. Harvell (eds.), The Ecology and Evolution of Inducible Defenses. Princeton University Press, Princeton.ADLER, F. R., AND R. KARBAN. 1999. Why induced defenses may be favored over constitutive strategies in plants,p. 45–61. In R. Tollrian and C. D. Harvell (eds.), The Ecology and Evolution of Inducible Defenses. PrincetonUniversity Press, Princeton.AGRAWAL, A. A. 2001. Phenotypic plasticity in the interactions and evolution of species. Science, 294:321–326.ALEXANDER, R. R., AND G. P. DIETL. In press. The Phanerozic history of shell-breaking predation on marine bivalvesand gastropods. In P. H. Kelley, M. Kowalewski, and T. A. Hansen (eds.), Predator-Prey Interactions in theFossil Record. Kluwer Academic/Plenum Publishers.BAKKER, R. T. 1983. The deer flees, the wolf pursues: incongruities in predator-prey evolution, p. 350–382. In D.J. Futuyma and M. Slatkin (eds.), Coevolution. Sinauer, Sunderland, MA.BERZINS, I. K., AND R. L. CALDWELL. 1983. The effect of injury on the agonistic behavior of the stomatopod,Gonodactylus bredini (Manning). Marine Behaviour and Physiology, 10:83–96.BIRKELAND, C. 1996. Why some species are especially influential on coral-reef communities and others are not.Galaxea, 13:77–84.BOUCOT, A. J. 1983. Does evolution take place in an ecological vacuum? Journal of Paleontology, 57:1–30.BOUCOT, A. J. 1990. Evolutionary Paleobiology of Behavior and Coevolution. Elsevier, New York.BRODIE, E. D., III, AND E. D. BRODIE, JR. 1999. Predator-prey arms races: Asymmetrical selection on predators andprey may be reduced when prey are dangerous. Bioscience, 49:557–568.BROOKS, D. R., AND D. A. MCLENNAN. 1991. Phylogeny, Ecology, and Behavior: A Research Program in ComparativeBiology. University of Chicago Press, Chicago.BUSH, A. M., M. G. POWELL, W. S. ARNOLD, T. M. BERT, AND G. M. DALEY. 2002. Time-averaging, evolution, andmorphologic variation. Paleobiology, 28:9–25.CALDWELL, R. L., AND H. DINGLE. 1976. Stomatopods. Scientific American, 234:80–89.CHESTNUT, A. F. 1952. Growth rates and movements of hard clams, Venus mercenaria. Proceedings of the Gulfand Caribbean Fisheries Institute, 4 th Annual Session:49–59.COTT, H. B. 1940. Adaptive Coloration in Animals. Methuen, London.DARWIN, C. 1859. On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Racesin the Struggle for Life. Facsimile of the 1 st edition. Harvard University Press, Cambridge.DAWKINS, R. 1982. The Extended Phenotype. W. H. Freeman and Company, Oxford.DAWKINS, R. 1986. The Blind Watchmaker. W. W. Norton & Company, New York.DAWKINS, R. 1995. River Out of Eden: A Darwinian View of Life. Basic, New York.DAWKINS, R., AND J. R. KREBS. 1979. Arms races between and within species. Proceedings of the Royal Society ofLondon, B, 205:489–511.DEANGELIS, D. L., J. A. KITCHELL, W. M. POST, AND C. C. TRAVIS. 1984. A model of naticid gastropod predatorpreycoevolution, p. 120–136. In S. A. Levin and T. G. Hallam (eds.), Mathematical ecology. Lecture Notesin Biomathematics, 54.DEANGELIS, D. L., J. A. KITCHELL, AND W. M. POST. 1985. The influence of naticid predation on evolutionarystrategies of bivalve prey: conclusions from a model. American Naturalist, 126:817–842.DIECKMANN, U., P. MARROW, AND R. LAW. 1995. Evolutionary cycling in predator-prey interactions: populationdynamics and the Red Queen. Journal of Theoretical Biology, 176:91–102.DIETL, G. P. In review. Experimental tests of the strength of interaction between a predator and dangerous prey:Busycon predation on Mercenaria. Journal of Experimental Marine Biology and Ecology.371

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002DIETL, G. P., AND R. R. ALEXANDER. 2000. Post-Miocene shift in stereotypic naticid predation on confamilial preyfrom the mid-Atlantic shelf: coevolution with dangerous prey. Palaios, 15:414–429.DIETL, G. P., R. R. ALEXANDER, AND W. F. BIEN. 2000. Escalation in Late Cretaceous–early Paleocene oysters(Gryphaeidae) from the Atlantic Coastal Plain. Paleobiology, 26:215–237.DINGLE, H., AND R. L. CADWELL. 1978. Ecology and morphology of feeding and agonistic behavior in mudflatstomatopods (Squillidae). Biological Bulletin, 155:134–149.EHRLICH, P. R., AND P. H. RAVEN. 1964. Butterflies and plants: a study in coevolution. Evolution, 18:586–608.FISHER, R. A. 1930. The Genetical Theory of Natural Selection. Oxford University Press, Oxford.FUTUYMA, D. J. 1985. Evolution and coevolution in communities, p. 369–381. In D. M. Raup and D. Jablonski(eds.), Patterns and Processes in the History of Life. Springer-Verlag, Berlin.FUTUYMA, D. J. 1986. Evolutionary Biology. Sinauer, Sunderland, MA.FUTUYMA, D. J., AND M. SLATKIN. 1983a. Coevolution. Sinauer, Sunderland, MA.FUTUYMA, D. J., AND M. SLATKIN. 1983b. Introduction, p. 1–13. In D. J. Futuyma and M. Slatkin (eds.), Coevolution.Sinauer, Sunderland, MA.FUTUYMA, D. J., AND M. SLATKIN. 1983c. The study of Coevolution, p. 459–464. In D. J. Futuyma and M. Slatkin(eds.), Coevolution. Sinauer, Sunderland, MA.GELLER, J. B. 1983. Shell repair frequencies of two intertidal gastropods of northern California: microhabitatdifferences. Veliger, 26:113–115.GELLER, J. B. 1990. Reproductive responses to shell damage by the gastropod Nucella emarginata (Deshayes).Journal of Experimental Marine Biology and Ecology, 136:77–87.GOMULKIEWICZ, R., J. N. THOMPSON, R. D. HOLT, S. L. NUISMER, AND M. E. HOCHBERG. 2000. Hot spots, cold spots,and the geographic mosaic theory of coevolution. American Naturalist, 156:156–174.GOULD, S. J. 1990. Speciation and sorting as the source of evolutionary trends, or “things are seldom what theyseem.” p. 3–27. In K. J. McNamara (ed.), Evolutionary Trends. University of Arizona Press, Tuscon.GOULD, S. J., AND R. C. LEWONTIN. 1979. The Spandrels of San Marco and the Panglossian paradigm: a critique of theadaptationist programme. Proceedings of the Royal Society of London B, 205:581–598.GROSBERG, R., AND C. W. CUNNINGHAM. 2001. Genetic structure in the sea: from populations to communities, p. 61–84.In M. D. Bertness, S. D. Gaines, and M. E. Hay (eds.), Marine Community Ecology. Sinauer, Sunderland, MA.HANSEN, T. A., AND P. H. KELLEY. 1995. Spatial variation of naticid gastropod predation in the Eocene of NorthAmerica. Palaios, 10:268–278.HARVELL, C. D. 1984. Predator-induced defenses in a marine bryozoan. Science, 224:1357–1359.HARVELL, C. D. 1990. The ecology and evolution of inducible defenses. Quarterly Review of Biology, 65:323–340.HARVELL, C. D., AND R. TOLLRIAN. 1999. Why inducible defenses?, p. 3–9. In R. Tollrian and C. D. Harvell (eds.),The Ecology and Evolution of Inducible Defenses. Princeton University Press, Princeton.HERRE, E. A. 1999. Laws governing species interactions? Encouragement and caution from figs and their associates,p. 209–237. In L. Keller (ed.), Levels of Selection in Evolution. Princeton University Press, Princeton.HINES, A. H., A. M. HADDON, AND L. A. WEICHERT. 1990. Guild structure and foraging impact of blue crabs andepibenthic fish in a subestuary of Chesapeake Bay. Marine Ecology Progress Series, 67:105–126.HOFFMEISTER, A. P., AND M. KOWALEWSKI. 2001. Spatial and environmental variation in the fossil record of drillingpredation: a case study from the Miocene of central Europe. Palaios, 16:566–579.HUGHES, R. N., AND R. W. ELNER. 1989. Foraging behavior of a tropical crab: Calappa ocellata Holthuis feeding upon themussel Brachidontes domingensis (Lamarck). Journal of Experimental Marine Biology and Ecology, 133:93–101.HUGHES, R. N., AND H. P. I. HUGHES. 1981. Morphological and behavioral aspects of feeding in the Cassidae(Tonnacea, Mesogastropoda). Malacologia, 20:385–402.JANZEN, D. H. 1980. When is it coevolution? Evolution, 34:611–612.JUANES, F., AND E. B. HARTWICK. 1990. Prey size selection in Dungeness crabs: the effect of claw damage. Ecology,71:744–758.KELLEY, P. H. 1989. Evolutionary trends within bivalve prey of Chesapeake Group naticid gastropods. HistoricalBiology, 2:139–156.KELLEY, P. H. 1992. Evolutionary patterns of naticid gastropods of the Chesapeake Group: an example ofcoevolution? Journal of Paleontology, 66:794–800.372

DIETL AND KELLEY—PREDATOR-PREY ARMS RACESKELLEY, P. H., AND T. A. HANSEN. 1993. Evolution of the naticid gastropod predator-prey system: an evaluation ofthe hypothesis of escalation. Palaios, 8:358–375.KELLEY, P. H., AND T. A. HANSEN. 1996. Naticid gastropod prey selectivity through time and the hypothesis ofescalation. Palaios, 11:437–445.KELLEY, P. H., AND T. A. HANSEN. 2001. The role of ecological interactions in the evolution of naticid gastropodsand their molluscan prey, p. 149–170. In W. D. Allmon and D. J. Bottjer (eds.), Evolutionary Paleoecology.Columbia University Press, New York.KENT, B. W. 1983. Patterns of coexistence in busyconine whelks. Journal of Experimental Marine Biology andEcology, 66:257–283.KITCHELL, J. A. 1982. Coevolution in a predator-prey system. Third North American Paleontological Convention,Proceedings, 2:301–305.KITCHELL, J. A. 1986. The evolution of predator-prey behavior: naticid gastropods and their molluscan prey, p.88–110. In M. Nitecki and J. A. Kitchell (eds.), Evolution of Animal Behavior: Paleontological and FieldApproaches. Oxford University Press, Oxford.KITCHELL, J. A. 1990. The reciprocal interaction of organism and effective environment: learning more about“and,” p. 151–169. In R. M. Ross and W. B. Allmon (eds.), Causes of Evolution: A Paleontological Perspective.University of Chicago Press, Chicago.KITCHELL, J. A., C. H. BOGGS, J. F. KITCHELL, AND J. A. RICE. 1981. Prey selection by naticid gastropods: experimentaltests and application to the fossil record. Paleobiology, 7:533–552.KNOLL, A. H., AND R. K. BAMBACH. 2000. Directionality in the history of life: diffusion from the left wall orrepeated scaling of the right?, p. 1–14. In D. H. Erwin and S. L. Wing (eds.), Deep Time: Paleobiology’sPerspective. Paleontological Society, Allen Press, Lawrence, KS.KOHN, A. J. 1989. Natural history and the necessity of the organism. American Zoologist, 29:1095–1103.KRAEUTER, J. N. 2001. Predators and predation, p. 441–589. In J. N. Kraeuter and M. Castagna (eds.), Biology ofthe Hard Clam. Elsevier Science, New York.LEONARD, G. H., M. D. BERTNESS, AND P. O. YUND. 1999. Crab predation, waterborne cues, and inducible defensesin the blue mussel, Mytilus edulis. Ecology, 80:1–14.LEVIN, S. A., L. A. SEGEL, AND F. R. ADLER. 1990. Diffuse coevolution in plant-herbivore communities. TheoreticalPopulation Biology, 37:171–191.LEVY, C. K. 1999. Evolutionary Wars: A Three-Billion-Year Arms Race. W. H. Freeman and Company, New York.LIVELY, C. M. 1986. Predator-induced shell dimorphism in the acorn barnacle Chthamalus anisopoma. Evolution, 40:232–242.MAGALHAES, H. 1948. An ecological study of snails of the genus Busycon at Beaufort, North Carolina. EcologicalMonographs, 18:377–409.MCNAMARA, K., AND J. LONG. 1998. The Evolution Revolution. John Wiley & Sons, New York.MURDOCH, W. W., AND A. OATEN. 1975. Predation and population stability, p. 1–131. In A. MacFadyen (ed.),Advances in Ecological Research, Vol. 9. Academic Press, New York.NEBELSICK, J. H., AND M. KOWALEWSKI. 1999. Drilling predation on Recent Clypeasteroid echinoids from the RedSea. Palaios, 14:127–144.PAINE, R. 1980. Food webs: linkage, interaction strength and community infrastructure. Journal of Animal Ecology,49:667–685.PARKER, G. A. 1983. Arms races in evolution—an ESS to the opponent-independent costs game. Journal ofTheoretical Biology, 101:619–648.PETERSON, C. H. 1979. The importance of predation and competition in organizing the intertidal epifaunalcommunities of Barnegat Inlet, New Jersey. Oecologia, 39:1–24.ROSENZWEIG, M. L. 1973. Evolution of the predator isocline. Evolution, 27:84–94.ROSENZWEIG, M. L., J. S. BROWN, AND J. L. VINCENT. 1987. Red Queens and ESS: the coevolution of evolutionaryrates. Evolutionary Ecology, 1:59–94.ROUGHGARDEN, J. 1983. The theory of coevolution, p. 33–64. In D. J. Futuyma and M. Slatkin (eds.), Coevolution.Sinauer, Sunderland, MA.SCHAFFER, W. M., AND M. L. ROSENZWEIG. 1978. Homage to the Red Queen, I: Coevolution of predators and theirvictims. Theoretical Population Biology, 14:135–157.373

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002SCHMIDT, N. 1989. Paleobiological implications of shell repair in Recent marine gastropods from the northernGulf of California. Historical Biology, 3:127–139.SHOUP, J. B. 1968. Shell opening by crabs of the genus Calappa. Science, 160:887–888.SLATKIN, M., AND J. MAYNARD SMITH. 1979. Models of Coevolution. Quarterly Review of Biology, 54:233–263.SLOBODKIN, L. B. 1974. Prudent predation does not require group selection. American Naturalist, 108:665–678.SMITH, L. D., AND A. R. PALMER. 1994. Effects of manipulated diet on size and performance of brachyuran crabclaws. Science, 264:710–712.SMITH, L. D., AND J. A. JENNINGS. 2000. Induced defensive responses by the bivalve Mytilus edulis to predatorswith different attack modes. Marine Biology, 136:461–469.STANLEY, S. M., B. VAN VALKENBURGH, AND R. S. STENECK. 1983. Coevolution and the fossil record, p. 328–349. InD. J. Futuyma and M. Slatkin (eds.), Coevolution. Sinauer, Sunderland, MA.STENSETH, N. C., AND J. MAYNARD SMITH. 1984. Coevolution in ecosystems: Red Queen evolution or stasis?Evolution, 38:870–880.THOMPSON, J. N. 1986. Constraints on arms races in coevolution. Trends in Ecology and Evolution, 1:105–107.THOMPSON, J. N. 1988. Variation in interspecific interactions. Annual Review of Ecology and Systematics, 19:65–87.THOMPSON, J. N. 1994. The coevolutionary process. University of Chicago Press, Chicago.THOMPSON, J. N. 1998. Rapid evolution as an ecological process. Trends in Ecology and Evolution, 13:329–332.THOMPSON, J. N. 1999a. Specific hypotheses on the geographic mosaic of coevolution. American Naturalist, 153:S1–S14.THOMPSON, J. N. 1999b. Coevolution and escalation: are ongoing coevolutionary meanderings important? AmericanNaturalist, 153:S92–S93.TRAVIS, J. 1994. Evaluating the adaptive role of morphological plasticity, p. 99–122. In P. C. Wainwright and S. M.Reilly (eds.), Ecological Morphology: Integrative Organismal Biology. University of Chicago Press, Chicago.TRUSSELL, G. C. 1996. Phenotypic plasticity in an intertidal snail: the role of a common crab predator. Evolution,50:448–454.TRUSSELL, G. C. 2000. Phenotypic clines, plasticity, and morphological trade-offs in an intertidal snail. Evolution,54:151–166.VAN VALEN, L. 1973. A new evolutionary law. Evolutionary Theory, 1:1–18.VAN VALEN, L. 1976. Energy and evolution. Evolutionary Theory, 1:179–229.VERMEIJ, G. J. 1973. Adaptation, versatility, and evolution. Systematic Zoology, 22:466–477.VERMEIJ, G. J. 1978. Biogeography and Adaptation: Patterns of Marine Life. Harvard University Press, Cambridge, 332p.VERMEIJ, G. J. 1980. Drilling predation of bivalves in Guam: some paleoecological implications. Malacologia, 19:329–334.VERMEIJ, G. J. 1982a. Gastropod shell form, breakage, and repair in relation to predation by the crab Calappa.Malacologia, 23:1–12.VERMEIJ, G. J. 1982b. Unsuccessful predation and evolution. American Naturalist, 120:701–720.VERMEIJ, G. J. 1983. Intimate associations and coevolution in the sea, p. 311–327. In D. J. Futuyma and M.Slatkin (eds.), Coevolution. Sinauer, Sunderland, MA.VERMEIJ, G. J. 1987. Evolution and Escalation: An Ecological History of Life. Princeton University Press,Princeton, NJ, 527p.VERMEIJ, G. J. 1992. Time of origin and biogeographical history of specialized relationships between northernmarine plants and herbivorous mollucs. Evolution, 46:657–664.VERMEIJ, G. J. 1994. The evolutionary interaction among species: selection, escalation, and coevolution. AnnualReview of Ecology and Systematics, 25:219–236.VERMEIJ, G. J. 1999. Inequality and the directionality of history. American Naturalist, 153:243–253.VERMEIJ, G. J. 2002. Evolution in the consumer age: predators and the history of life. In M. Kowalewski and P. H.Kelley (eds), The Fossil Record of Predation. The Paleontological Society Papers, 8 (this volume).VERMEIJ, G. J., AND A. P. COVICH. 1978. Coevolution of freshwater gastropods and their predators. AmericanNaturalist, 112:833–843.VIRNSTEIN, R. W. 1977. The importance of predation by crabs and fishes on benthic infauna in Chesapeake Bay.Ecology, 58:1199–1217.WEST, K., A. COHEN, AND M. BARON. 1991. Morphology and behavior of crabs and gastropods from LakeTanganyika, Africa: implications for lacustrine predator-prey coevolution. Evolution, 45:589–607.374

VERMEIJ—EVOLUTION IN THE CONSUMER AGEEVOLUTION IN THE CONSUMER AGE:PREDATORS AND THE HISTORY OF LIFEGEERAT J. VERMEIJDepartment of Geology, University of California at Davis, One Shields Avenue, Davis, California 95616 USAABSTRACT—Three properties of predation make this form of consumption an important agency of evolution:universality (all species have predators), high frequency (encounters of prey with predators test both parties often),and imperfection (many predatory attacks fail, enabling antipredatory selection to take place). On long time scales,predators have two principal effects: they influence their victims’ phenotypes, and prey species that are highlyvulnerable to all phases of predatory attacks are evolutionarily restricted to environments where predators arerarely encountered. Although predator and prey can affect each other’s behavior and morphology on timescalescommensurate with individual lifespans, predators have the evolutionary upper hand over the long run, especiallyin the expression of sensory capacities, locomotor performance, and the application of force. Only in passive defenses(armor, toxicity, large body size) does escalation favor the prey. In a review of methods for inferring predation in thegeological past, I argue against the use of whole assemblages, which combine species of contrasting adaptive type.Instead, I strongly favor species-level and clade-level approaches (including examples of clade replacement) inwhich comparisons among places and among time intervals are made within the same adaptive types and the samephysical environments. The available evidence, much of which comes from studies of shell drilling and shell breakage,points to temporal increases in both predator power and prey defenses. Escalation between species and their enemies,including predators, has proceeded episodically against a backdrop of generally increasing productivity andincreasing top-down evolutionary control by high-energy predators during the Phanerozoic, the consumer age.INTRODUCTIONCONSUMPTION OF ONE organism byanother is a universal phenomenon in all livingecosystems. Herbivory (eating plants), predation(a large animal or sometimes a plant consumingpart or the whole of another animal), and parasitism(a small animal, fungus, microbe, or plant drawingnutrients from another organism) are differentmanifestations of this phenomenon. Consumptionis itself a special case of competition for food, orchemical energy. Competition in this broadest senseis the central process common to all economicsystems, and has been with us since the origin oflife. Consumption is a derived form of competition.Parasitism could be a very ancient form ofconsumption, whereas herbivory and predationhave become prominent only during the last 600to 700 million years, the interval encompassing thelatest Proterozoic to the Recent, a phase inecological history we may call the consumer age.These forms of consumption have played animportant role in evolution. They represent a typeof top-down evolutionary control that hasprofoundly affected the phenotypes anddistribution of all species, a control that has likelyintensified through time.In this chapter, I first summarize whyconsumption in general, and predation inparticular, should have such a far-reachinginfluence on evolution and in the history of life.Second, I review and evaluate the kinds ofevidence that are available about predation in thefossil record. Third, I discuss briefly the temporaltrends in predation and what these trends implyabout the history of ecosystems and theirconstituent species. Along the way, I suggestavenues for further research, concentratingespecially on better temporal resolution ofevolutionarily important predation-relatedinnovations and on temporal studies of lineagesand clades in physically similar habitats.375

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002THE IMPORTANCE OFPREDATION IN EVOLUTIONThree properties of predation make this formof consumption an important agency of evolution.These are: (1) universality—all species havepredators at some point in the life cycle of constituentindividuals; (2) high frequency—attacks by, orcontact with, predators occur at a frequency highenough so that the average individual of a victimspecies will have an encounter with a predator oneor more times during its lifetime; and (3)imperfection—predators have low rates of successduring one or more stages of an attack, implyingthat victim species can evolutionarily “learn” frompredators’ failures and adapt accordingly.Predation poses a threat to individuals of allspecies. It is the principal cause of death in manypopulations; for others it is a fate restricted to oneor another life stage. Humans today experienceexceedingly little predation, but in our early historyin Africa and elsewhere, predation by largemammals had far-reaching effects on our alertness,locomotor performance, habitat preferences, andsocial structure (Ambrose, 2001). Colonial (orclonal) animals typically experience partialpredation, a form of consumption in which part ofthe colony survives and regenerates. Partialpredation also occurs commonly among solitaryanimals. Examples include fin-eating and eyegougingby cichlid fish, nipping of clam siphons byflatfish, consumption of autotomized (voluntarilyreleased) tails of lizards and limbs of crabs andsiphons of clams, and the removal of 5×7-cmhemispherical plugs of flesh from large fish andmarine mammals by the cookie-cutter shark Isistiusbrasiliensis (Jones, 1971).Predators can be important agents of selectiononly if the average prey individual encounters themat least once during its lifetime. By an encounter Imean either the recognition by the prey that anattack from a predator is imminent, or therecognition by the predator that an item of food isnearby. The number of encounters and the rate ofencounters can be defined from either the prey’sor the predator’s point of view. The number ofencounters a given prey individual has withpredators is proportional to three quantities: (1) thedensity (number of individuals per unit area) ofpredators, (2) the combined activity (area traversedby an individual per unit time) of the prey and itspredators, and (3) the average lifespan of the prey.Similarly, the number of encounters a predator haswith prey depends on prey density, the combinedactivity of the predator and its prey, and the averagelifespan of the predator. Expressed formally, theserelationships aree y= (n d/A)[(A d/t)+(A y/t)]•t y(1)e d= (n y/A)[(A d/t)+(A y/t)]•t d(2)where subscripts y and d refer to prey and predatorrespectively, e is number of encounters, n is numberof individuals, A is area, and t is time.There is typically a tradeoff (negativecorrelation) between the number of individuals andthe activity level of individuals. However, a givenhigh per-capita food requirement for a predator(that is, a high metabolic rate) translates into intenseactivity, and likely offsets the associated decreasein the density of predators in the determination ofa prey individual’s encounter rate. Greater mobilityof either the predator or the prey will increase thelikelihood of encounters. Higher metabolic rates,which make greater mobility possible, shouldtherefore be associated with higher numbers ofencounters for both parties. Predators that eat orkill whole prey animals will collectively beoutnumbered by the prey, meaning that e y(numberof encounters from the prey’s perspective) willtypically be much lower than e d(number ofencounters from the predator’s point of view).All else being equal, then, predators with thehighest encounter rates with prey should have thelargest evolutionary effects on prey, and theevolutionary dynamics between predators and preyare tilted in favor of the predators. In other words,predators should have more effect on prey than preyhave on predators, at least in the long run. Theseeffects will be evolutionary only if many of theencounters are unsuccessful from the predator’spoint of view; they will be manifested as populationregulation of the prey or as limits to prey376

VERMEIJ—EVOLUTION IN THE CONSUMER AGEdistribution if encounters with prey are lethal or ifprey respond by cowering in refuges.I have shown elsewhere (Vermeij, 1982a) thatunsuccessful predation is an extremely commonphenomenon. Few predators, no matter howsophisticated their technology, are 100% successfulin killing the prey they encounter, capture, orsubdue. Only sophisticated predators like rats, cats,mongoose, and humans that colonize island biotaswhose members had no previous evolutionaryexperience with enemies of comparably highperformance are capable of wiping out entire preyspecies. In all other cases, one or another stage inthe predator’s attack is marked by a high failurerate, and most prey species can persist in refuges,habitats or periods of time in which predators areinactive or absent. Failure rates of more than 90%have been recorded for some raptors (predatorybirds), seastars, and crabs (Dayton et al., 1977;Rudebeck, 1950–1951; Vermeij, 1982b).Unsuccessful predation is essential toantipredatory adaptation. If all attacks weresuccessful, then any prey living long enough toreproduce would have had no encounters withpredators, and therefore could not be adapted tothose predators. Failure provides an opportunityevolutionarily to “learn.” Every instance of failureby the predator “tests” some aspect of a preyindividual’s chemistry, behavior, or morphology.The phase of a predator’s attack with the highestpredator failure rate will be the phase to which theprey is best adapted. For example, crabs mayrecognize and capture most potential prey, but oftenhave low success rates during the subjugationphase. Molluscan adaptation against crabs istherefore concentrated on aspects that conferresistance to breakage or forced entry of the shell.On the other hand, many gastropods and bivalveshave well-developed escape responses to relativelyslow-moving predators such as seastars andgastropods. Here, adaptation may thus beconcentrated in the capture phase.On long time scales appropriate to the fossilrecord, predators have two principal effects. Thefirst is that, as major agents of selection, theyinfluence their victims’ phenotypes. If prey andpredator co-occur and encounter each other, theprey must be adapted to the predator, and alladaptations with respect to such functions asfeeding and mating must be consistent withadaptations to enemies, including predators.Second, prey species that are highly vulnerable tosuccessful attack by predators during all phases ofthe attack are confined to times and places whereencounters with predators are rare, fecundity of preycompensates for losses due to predation, or predatorsare constrained physically in feeding. Environmentswith reduced predation intensities often harborspecies with reduced metabolic capacities enforcedby such limiting factors as low temperatures, lowoxygen levels, low food supplies, high resistance tolocomotion (as in sediments), and high risks topredators. Examples of such refuges today includethe deep sea, caves, the upper intertidal zone, deeplayers of sediment, polar regions, and the bodies ofwell-defended host organisms (Vermeij, 1987).Predation does, of course, occur in theseenvironments, but it does so either at a low rate orduring very restricted times, or both. Predators thusexercise pervasive top-down control over manyspecies, whether this be expressed evolutionarilyor in terms of activity and distribution.A striking, and hitherto undocumented, patterndemonstrates the primary evolutionary influenceof predators over other species. In terms of absoluteperformance, predators of mobile prey haveachieved levels of sophistication beyond those ofanimals with other feeding habits. Organs of visionare best developed among highly active predatorsincluding cephalopods, many trilobites, spiders,decapod and stomatopod crustaceans, beetles andbugs and other insects, fish, and tetrapods. Thesense of hearing is extraordinarily developed innight-hunting owls and bats as well as amongraptors. Distance chemoreception among aquaticinvertebrates is largely confined to predatorygroups in gastropods and sea stars. Most gastropodsthat filter particles, graze algae or detritus, orconsume stationary animal prey sample incomingwater from all around, but they evidently cannotdetect food or enemies chemically at a distance.This capability is reserved for members of the clade377

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Sorbeoconcha and some opisthobranchs, the onlygastropod groups containing aquatic predators ofmobile prey (Lindberg and Ponder, 2001). Thesesensory differences between predators and prey aresurprising, because prey can and do use sense organsto detect predators at an early stage of a potentialattack. Early warning is clearly beneficial in that itallows to prey to take elusive, aggressive, or resistiveaction; yet it is the sense organs of the predators ofmobile prey that evidently have greater acuity.The same difference is seen in the escapepursuitphase of an attack. Within faunas, the fastestanimals on land, in the air, and in the water are toppredators—cheetah, raptors, and tunas. Manyherbivores—gazelles, pronghorn, hares, andhorses—are fast, being capable of sustained highspeed, but their absolute powers fall below thoseof co-occurring (or in some cases formerly cooccurring)predators. (Today, the North Americanpronghorn is the fastest running mammal on thatcontinent, but during the Pleistocene it was likelyovertaken by the North American cheetah.) Evenamong gastropods, not known for their agility, thefastest burrowers are predatory naticid and olividsnails (Vermeij and Zipser, 1986). Performanceneed not be in speed alone; often it is manifestedas maneuverability, or as the rapid deployment ofpart of the predator—claws of crabs, protrusiblejaws of many teleost fish, raptorial appendages ofstomatopods, the adhesive tongue of a salamanderor anteater, and traps in many animals—involvedin rapid capture.In the resistance-subjugation phase, too,predators generally have the edge. Almost allvenomous animals are predators. Venom canfunction defensively, but its greatest developmentis in those animals—snakes, spiders, centipedes,toxoglossan gastropods, and many cnidarians—inwhich rapid subjugation of large or potentiallydangerous prey is paramount. The largest forcesexerted by animals are those in the jaws of toppredatoryvertebrates such as alligators, wolves,and (by inference) the Late Cretaceous theropoddinosaur Tyrannosaurus (Erickson et al., 1996).Intense competition among predators is likelyresponsible for the high sophistication of senseorgans, mobility, and the tools of subjugation.Getting to a prey first, or killing and eating it beforecompetitors get to it, can be critical to the survivaland mating success of predators (Bakker, 1983; VanValkenburgh, 1991; Van Valkenburgh and Hertel,1993). When combat over mates involves the sametools as does subjugation of prey, as is the case incrabs (Lee and Seed, 1992; Smith, 1992; Juanesand Smith, 1995), selection for effective predationrelatedtechnology is especially intense. It is thusultimately competition that is responsible for thegeneral top-down evolutionary and distributionalcontrol exercised by predators (Vermeij, 1987).It is only in the passive defenses where predatorsare often outdone by their prey. The strength ofmolluscan shells often exceeds the maximum biteforces of the strongest predators. Toxicity is a verywidespread form of passive defense in plants andanimals, but it rarely if ever characterizes toppredators. In many ecosystems, even body size inconsumers is exceeded by that of the largest prey.No organisms are larger than certain trees, fungi,and seaweeds; and in many living marine andterrestrial ecosystems, as well as in the dinosaurdominatedcommunities of the Mesozoic, the largestherbivorous tetrapods far exceed their predators inbody size (Burness et al., 2001).By emphasizing the control predators haveover their prey, I do not mean to suggest, as I havein the past (Vermeij, 1982a, 1987, 1994), that preyexercise little influence over their predators.Compelling experimental evidence from a widevariety of predator-prey interactions on land andin the sea strongly indicates that predators inducedefenses in their prey, and, even more intriguingly,that prey characteristics directly and non-geneticallyaffect the performance of predators’ organs offeeding. The presence of chemical cues released bypredatory crabs and snails induces prey molluscs togrow thicker shells, often with relatively smaller,less accessible openings compared to individuals notexposed to such cues. The presence of strong-shelledprey causes crabs to grow claws that are larger, morerobust, and capable of applying larger forces duringfeeding (Appleton and Palmer, 1988; Palmer, 1990;Smith and Palmer, 1994; Smith and Jennings, 2000;378

VERMEIJ—EVOLUTION IN THE CONSUMER AGELeonard et al., 1999; for a general review ofinduced phenotypes see Agrawal, 2001).These important findings suggest that predatorand prey influence the directions of each other’sevolution. At first, adaptation is non-genetic, thanksto a responsiveness of behavior and morphologyto environmental cues. Responsiveness is undergenetic control, but the responses themselves takeplace over the short term within an individual’slifetime without genetic instruction. Extremelyrapid—in fact, almost instantaneous—reciprocaladaptation between predator and prey thusproceeds on timescales commensurate withindividual lifespans, but in the long run, predatorand prey evolve by a process of escalation(Vermeij, 1987)—a genetically based evolutionaryprocess in which predators retain the evolutionaryand economic edge of control over the prey. Withthe evolution of predators, therefore, the biospheretruly entered the consumer age.INFERENCES OF PREDATIONIN THE PASTTwo principal types of data are potentiallyinformative about the evolutionary role thatpredators have played in the past. These are (1)traces of predation, successful and especiallyunsuccessful attacks as chronicled by marks lefton fossils; and (2) functional-morphologicalinferences of predatory performance andantipredatory defense drawn from observations andexperiments on living predators and their prey.Repaired injuries provide direct evidence thatthe organism bearing them withstood significanttrauma, often inflicted by competitors andpredators. The incidence of such repairs in a preypopulation depends on (1) the probability of beingattacked, and (2) the probability that the attack isunsuccessful from the predator’s point of view. Theprobability of attack depends on the number ofpredators and on the prey’s age and lifespan,whereas the probability of failure is a function ofthe size of the predator relative to that of the prey.In general, the proportion of attacks that fail fromthe predator’s perspective should increase withprey size and age, although exceptions are knownamong crabs attacking snails (Vermeij, 1982b). Thelonger a potential victim lives, the more injuries itis likely to accumulate.For nonclonal organisms, such as manyanimals, the only form of predation that leavesreliable indicators of both successful andunsuccessful attacks is drilling. For predationinvolving skeletal breakage, we have reliable dataonly on unsuccessful attacks that are subsequentlyrepaired and that are preserved on a part of theskeleton that we can observe.Broken fragments are abundant in the fossilrecord, and no doubt in part reflect the work ofsuccessful predators. Cadée (1994), for example,made careful observations showing that most of thefragments of molluscan shells in the Wadden Sea(the Netherlands) are the result of predation by gullsand ducks. Merle (2000) inferred the destructiveaction of a large stingray from two largereconstructed early Eocene shells found as fragmentspreserved close together in sediments that indicatequiet-water conditions and rapid sedimentation.I have similarly argued from the types of lethaldamage done by crabs and spiny lobsters that thevast majority of lethally broken shells on modernseashores indicate successful predation (Vermeij,1982b). Holes in fossil shells comparable to thedamage done today by shell-smashing mantisshrimps (stomatopods) may indicate the work ofthese predators as argued by Geary et al. (1991)for the Plio-Pleistocene of Florida, Pether (1995)for shells from the Holocene of southern Africa,and by Baluk and Radwanski (1996) for snails inthe middle Miocene Korytnica Clay of Poland.Similarly, Boyd and Newell (1972) suggested thatmany of the fragments they encountered in Permiansediments in Wyoming chronicled predatorysuccesses. The problem is, of course, that countingsuch fragments is fraught with taphonomicinterference. It is difficult to distinguish betweensharp-edged fragments produced by shell-breakersand similar fragmentation produced by compactionor other diagenetic effects. Moreover, smaller preywould be so thoroughly macerated that diagnosticremains may not be preserved or recognizable.379

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002The picture is not entirely one of gloom anddoom. By examining the number of repairedinjuries in successive size classes of prey, whichin a rough way correspond to age classes, we caninfer something of the shape of the function relatingthe number of unsuccessful attacks and time. If,for instance, the frequency rises in successive sizeclasses, we can reasonably infer that the ratio ofunsuccessful to total number of attacks also does.What can we conclude about samples with nomarks of unsuccessful predation? There are twopossible interpretations. The first is that all attackson those individuals that happened to beencountered by predators were successful, makingthe skeleton an ineffective device against attack.Such an interpretation would be supported weaklyby the presence of co-occurring species with moreeffective defenses that do sustain repaired damage,and more strongly by the presence of fragmentsand the remains of potential predators. In theabsence of such evidence, the second interpretationis supported: shell-breakers were absent. It is alsopossible that some species capable of very rapidmovement or characterized by venom, aggressivebehavior, or distastefulness could lack repairedinjuries simply because their intact skeletons reflectantipredatory adaptations other than passive armor.These same remarks apply to the interpretation ofdrilling.Many paleontologists have assessedfrequencies of predation and frequencies ofunsuccessful predation in entire assemblages(Kelley and Hansen, 1993, 1996; Kelley et al.,2001; Hoffmeister and Kowalewski, 2001). Suchfrequencies are, however, difficult to interpretbiologically, because assemblages contain adiversity of adaptive types. Moreover, assemblagesdiffer in the relative abundances of species as wellas in the proportions of adaptive syndromes, so thatthey are not comparable in any meaningful way.In essence, there is no biological signal in suchassemblages until they are divided into species. Theproportions of different species or of adaptive typesmay be informative in themselves, but inferencesof successful and unsuccessful predation shouldbe made at the level of the species or the population,and comparisons among places or among timeintervals should be restricted to the same speciesor the same adaptive syndrome.There is a large literature showing that highincidences of unsuccessful predation on shellbearingmolluscs are associated with variousdefenses such as a thick-walled shell, strongexternal sculpture (or relief), microstructurespreventing the propagation of cracks perpendicularto the shell’s growing edge, and deep withdrawalof edible tissues into the shell (for a review seeVermeij, 1993). Demonstrated effectivenessagainst drilling predation is provided by a thickshell and by the presence of organic layers outsideand between mineralized shell layers (see, e.g.,Harper, 1994; Kardon, 1998; Kelley, 1989, 1992;Dietl and Alexander, 2000; Kirby, 2001). Forexample, Kardon (1998) showed that manyboreholes in corbulid bivalves stop at conchiolinlayers. Similarly, breaks made by crabs in snailshells characterized by the presences of varices(periodic thickenings) stop at the varices.Partial predation requires a slightly modifiedinterpretation. Part of the victim is eaten, but thevictim lives to repair the damage. This occurs whenthe extended siphons of bivalves are nipped by fish,when the arms of brittle stars or crinoids arenibbled, and when some of the modular units—polyps, zooids, etc.—of colonial animals areconsumed. The observed incidence of regeneration(number of parts per individual or number ofregenerating individuals compared to total numberof individuals examined) is a very conservativeestimate of both the intensity of predation and theselective, or evolutionary, importance of predation.Because regeneration is often rapid, we can recastthe measurement of regeneration as a source-sinkproblem. The source is predatory attacks that causenonlethal injury; the sink is the “removal” ofregenerating structures through completerestoration. Sublethal damage that is notregenerating should not usually be included in thecalculation of the incidence of regeneration inanimals that are no longer living, because this couldhave resulted from post-mortem damage. Theincidence of regeneration, expressed at the380

VERMEIJ—EVOLUTION IN THE CONSUMER AGEpopulation level, is thus the rate of injury timesthe mean regeneration time:i r= (dn/dt)•t r(4)where i is incidence of regeneration, n is numberof injuries, t is time, and t ris average time ofregeneration. Ideally the rate of injury is measuredon a time scale similar to that of the average timeneeded for regeneration.All comparative studies of predation revealstrikingly high variability in the incidence of bothsuccessful and unsuccessful attack. The incidenceof lethal drilling in the lucinid bivalve Ctena bellavaries from 0.l7 to 0.75 among Recent populationson the small island of Guam (Vermeij, 1980). Similarvariation in drilling has been detected amongmolluscan assemblages from the Eocene of thesoutheastern United States (Allmon et al., 1990;Hansen and Kelley, 1995), as well as amongNeogene turritellid gastropods (Allmon et al., 1990;Hagadorn and Boyajian, 1997) and late Miocenemolluscs from Bulgaria (Kojumdgieva, 1975). Dataon repaired shell breakage also reveal enormousvariations on spatial scales ranging from the localto the regional (Vermeij et al., 1980; Vermeij, 1982b;Schindel et al., 1982; Zipser and Vermeij, 1980;Schmidt, 1989; Cadée et al., 1997). Large-scalespatial and temporal patterns in predation musttherefore be dramatic in order to be discerned abovethe “noise” of variation (Vermeij et al., 1981;Vermeij, 1987; Cadée et al., 1997).Much of the history of predation as inferred fromthe fossil record is necessarily told from the prey’sperspective. This is because many kinds ofpredation, notably drilling and breakage, leavediagnostic traces on fossil skeletons, and becauseadaptations against these forms of predation arereadily identifiable in fossils. Unfortunately, therecord of predators is much spottier. Whenpreserved, the weapons of subjugation—teeth,claws, jaws, seastars’ arms and tube feet, and shellopeningdevices of predatory snails—offer clues tothe performance of predators, as do sensory organsand the structures involved in locomotion; but thesestructures are often not preserved or interpretable.Erickson et al. (1996) and Rayfield et al. (2001), intheir studies of tyrannosaurs and allosaursrespectively, have paved the way to a quantitativeevaluation of forces applied by predators. Theyemploy sophisticated mechanical measurements onmodel prey together with simulations to infer suchforces. These approaches can be applied much morewidely to other fossil predators.Measures of competition among predators canbe made in fossil material in exceptional cases. VanValkenburgh and Hertel (1993) used the frequencyof tooth breakage among carnivoran mammals inthe La Brea tar pits (late Pleistocene of Los Angeles)as an indicator of potential competition among bonebreakingpredators. In a series of carefulcomparisons, they inferred substantially higherintensities of competition among Pleistocenecarnivores than among their living North Americanand African counterparts. The incidence ofregenerating or broken chelipeds (claws) in crabscould probably be used similarly as an index ofcompetition (see Smith, 1992; Juanes and Smith,1995; Brock and Smith, 1998), but fossil material Ihave seen is inadequate. Evaluation of predatorperformance in fossils remains an important goal.THE BEGINNINGS OFTHE CONSUMER AGEThe early history of consumption remainsshrouded in mystery, not only because the reliablefossil record of consumers is very sparse, but alsobecause uncertainties abound concerning whichstyles of consumption are primitive and which arederived in the evolutionary tree. Nonetheless, someconclusions about the early history of consumptionseem justified by the available evidence.The likely earliest form of predation ispenetration and consumption of one bacterium byanother, as described by Guerrero et al. (1986).Given that this interaction occurs betweenprokaryotic organisms, this kind of consumption islikely to have originated some time before 2.5 billionyears ago, that is, during the Archean Eon. The originof the eukaryotic cell, involving the evolutionaryunion of separate lineages of prokaryotic organismsinto a single integrated whole that could function381

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002as an evolutionary unit, was accompanied by themajor innovation of phagocytosis, which enablesone organism to envelop—effectively to swallow—another individual completely (Margulis, 1981;Knoll and Bambach, 2000). Eukaryotic organizationevidently existed by 2.7 billion years ago and maywell be even older. Consumption of smallerprokaryotic and eukaryotic organisms by largersingle-celled consumers in the so-called microbialloop may have been the norm through much of theMesoproterozoic and Neoproterozoic eras(Butterfield, 1997). The date of origin ofmulticellular animals remains highly controversial,as does the time of appearance of structures thatwould allow such multicellular animals to ingest orotherwise to consume other multicellular life forms.The earliest evidence of successful as well asunsuccessful predation on animals comes from thelatest Neoproterozoic of Namibia. Bengtson andZhao (1992) described complete and incompleteholes excavated by an unknown attacker in themineralized tubular fossil Cloudina. This kind ofpredation persisted into the Early Cambrian(Conway Morris and Bengtson, 1994). At least twoadditional means of consumption—suspensionfeedingand predation by skeletal breakage—maketheir first appearance no later than the EarlyCambrian. Trilobites of this age show skeletalbreakage and repair consistent with the hypothesisthat anomalocarid arthropods, which at a length ofone meter or more are the largest known animalsof the time, were the culprits (Chen et al., 1994;Briggs, 1994; Collins, 1996; Babcock and Robison,1989). Suspension-feeding animals, whichconsume small prey en masse by collecting andfiltering them out of the water, also appeared inthe Early Cambrian (Vannier and Chen, 2000).Fossils preserved in the Burgess Shale (MiddleCambrian) of British Columbia reveal predationinvolving the ingestion of those undamaged pelagicprey by larger predators (Conway Morris, 1979).Thus, at least four distinct modes of predation wereestablished no later than Middle Cambrian time,about 525 Ma. Many adaptations of prey, notablyskeletonization and the evolution of spines, aredemonstrable antipredatory adaptationsrecognizable from latest Neoproterozoic timeonward (Vermeij, 1989). It is also likely thatevolutionary invasions by many lineages into theinfaunal environment (beneath the surface ofsediments), the pelagic realm, and even theendolithic environment (beneath the surface ofrocks) were prompted by the appearance anddiversification of consumers during the openingphases of the Phanerozoic Eon (Vermeij, 1987).PHANEROZOIC TRENDSAND THEIR IMPLICATIONSPatterns in the intensity of successful andunsuccessful predation and in the adaptationsrelated to predation are discernible at two distinctand complementary levels. The first is the level ofindividual lineages or clades, which are tracedthrough time in physically similar environments,that is, environments in which such characteristicsas thermal regime, productivity, and sediment typeremain more or less the same over the intervalconsidered. This level also accommodates lineagereplacement, as long as the replacing and replacedlineages share a common adaptive syndrome, oradaptive style. The second level ismacroevolutionary. Within assemblages fromsuccessive, environmentally similar horizons, wecan track the relative representation of variousadaptive syndromes. For example, the incidenceof species emphasizing resistance defenses ofvarious kinds can be compared to the incidence ofspecies emphasizing adaptations at the pursuitphase of predation. These macroevolutionarypatterns arise from differences in the rates ofspeciation or extinction of clades differing inadaptive style, and from replacement of cladesemphasizing one style of adaptation with cladesemphasizing another. My interest in predator-preyescalation was aroused by temporal patterns of thiskind in the expression of shell armor in molluscs(Vermeij, 1975). An increase in breakage-resistantspecies, and a decrease in the proportion of specieswith predation-vulnerable architecture, led me tothe hypothesis that breakage became anincreasingly important agency of selection through382

VERMEIJ—EVOLUTION IN THE CONSUMER AGEtime among molluscs in warm, shallow marinewaters (Vermeij, 1977).The history of predation following the earlyphase can be characterized as an enemy-drivenprocess of escalation involving predators, theircompetitors, and their prey (Vermeij, 1987, 1994).Slow methods of predation, reflecting low risks tothe predator during prey capture, have persistedalongside newer, more sophisticated, and notablyfaster modes such as breakage and envenomationthat also increased the size range of available prey.These higher-energy forms of predation, whichemphasize greater sensory acuity and locomotorperformance as well as greater subjugationalpower, particularly characterize productive,energy-rich environments, where prey haveadapted accordingly with highly sophisticateddefenses. As these methods evolved, predators andprey with less well developed weapons anddefenses became restricted or expanded into partsof the biosphere—the deep sea, caves, areas of lowtemperature, the safe bodies of well-defendedorganisms, and such physically constrained habitatsas deep layers in sediments or spaces beneathboulders—where access to energy and nutrients islimited (Vermeij, 1987). Very few habitatreversals—that is, from energy-poor to energy-richenvironments—are known. For example, a fewdeep-sea lineages may have penetrated shallowwaters, but they did so only at polar latitudes andin caves. Many tropical clades have invadedtemperate latitudes, but few if any have gone theother way, possibly with the exception of cladesimmediately following mass extinction.Details of the timing of significantbreakthroughs and of episodes of escalation remainstubbornly elusive, and deserve attention in futurestudies; but the summary below broadly capturesour present understanding. This account nicelycomplements Bambach’s synthesis of the globaldiversity of predators through time (Chapter 12, thisvolume), as well as inferences about the temporalincrease in productivity over the course of thePhanerozoic (Vermeij, 1987; Bambach, 1993, 1999).Predation by drilling is the most thoroughlydocumented form of consumption in the fossilrecord. During the Paleozoic, when most drillingis recorded in echinoderms and brachiopods, a peakin the intensity of (successful) drilling was reachedin the Devonian (Smith et al., 1985). Incidenceswere low from the Carboniferous to the LateCretaceous (Kowalewski et al., 1998, 2000)—generally less than 0.06 for most prey species.Modern high incidences, generally exceeding thoseof the Paleozoic, were established during earlyPaleocene to early Eocene time, and were generallymaintained through the rest of the Cenozoic. Thereare temporary declines in the late Eocene, asinferred from whole assemblages in the GulfCoastal Plain of the United States, and among earlyOligocene turritellid gastropods from the sameregion (Allmon et al., 1990; Kelley and Hansen,1993, 1996; Kelley et al., 2001).My unpublished compilation with PeterRoopnarine indicates that edge drilling, a modifiedand evolutionarily derived form of drilling in whichthe predator excavates a hole where valve marginsmeet in bivalved prey, appeared in the Jurassic(Harper et al., 1998) but did not become commonuntil the Recent.Prey effectiveness against drilling, measuredas the number of incomplete holes divided by thetotal number of holes in a sample of a given species(Vermeij, 1987), has generally increased over timesince the Late Cretaceous within lineages ofmolluscs in which the adaptive syndromehighlights passive resistance to drilling predators.Kelley (1989, 1992) has provided the mostthorough documentation of this trend in her studyof five genera of bivalves in the Chesapeake Groupof Maryland. Although there are a few exceptions,successive populations of species (whichpresumably form lineages, although this remainsto be verified) show increases in effectivenessagainst drilling from the early Miocene (Langhian)Calvert Formation to the middle Miocene(Serravalian) Choptank Formation and the earlylate Miocene St. Mary’s Formation (Tortonian).Higher effectiveness is accompanied by relativelythicker valves. Similar trends appear in some LateCretaceous to Paleogene lineages of slowburrowingbivalves (Kelley and Hansen, 1993).383

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002Assemblage-wide studies of drilling in thesoutheastern United States have been interpretedas being inconsistent with the hypothesis ofescalation, because effectiveness against drillingas measured from entire assemblages shows noconsistent trends from the Late Cretaceous to thePleistocene (Kelley and Hansen, 1993, 1996;Kelley et al., 2001). I consider these inferencesinvalid because the biotic composition, especiallywith respect to the relative abundance ofcontrasting adaptive types, varies greatly amongassemblages. An assemblage numericallydominated by slow-burrowing bivalves, whichadapt to drilling mainly through passive resistance,might show a high aggregate effectiveness againstdrilling, whereas an assemblage in which fastburrowingspecies predominate, whose anti-drillingdefense is expressed chiefly in the pursuit phase ofan attack, would show low effectiveness. Whetherthe inferences of an absence of escalation by Kelleyand her coworkers are wrong as well as invalid willhave to await more detailed assessments of thebehavior of individual adaptive types amongpotential prey molluscs. There are no clear trendsin effectiveness, shell thickness, or other potentialdefenses in turritellid gastropods during the lateCenozoic (Hagadorn and Boyajian, 1997), thoughthis result is compromised in that comparisons werenot made within lineages or within physicallycomparable environments. On the Gulf Coastal Plainof the southern United States, Kelley and Hansen(1993) showed that effectiveness among turritellidsincreased from the Paleocene to the Eocene, whereasamong corbulid bivalves it increased between theLate Cretaceous and Paleocene. Lucinid bivalvesshowed no trend. Again, these conclusions mightchange if individual lineages or similar adaptivesyndromes were followed through time in similarenvironments.Increased prey effectiveness through time, ifconfirmed, would indicate a form of escalation inwhich the prey have gained an adaptive advantageover their drilling predators in the long run. Asdiscussed previously, this would be expected if theprey’s adaptive syndrome emphasized passivearmor. Unfortunately, we can surmise little aboutthe adaptations of the predators. Kelley (1992)pointed to an increase in body size (and thuspotentially in the range of available prey) in thetwo lineages of naticid drillers she studied in theMiocene of Maryland, but found no trends in shellshape or aperture shape. Large-footed naticidscapable of edge drilling bivalves and of envelopinglarge prey in the foot are evidently of post-Eoceneorigin, as are edge-drilling muricids.Hagadorn and Boyajian (1997) claimed thatPliocene naticids in the southeastern United Stateswere more stereotyped and more selective in theposition of drilling sites in turritellid prey than wereMiocene naticids, and suggested that there was lessvariation in the size of the predator (as measuredby drill-hole diameter) relative to prey size in theirPliocene samples. If verified, this result could implyadaptations by the predator, but it could also meanthat the prey’s morphology has constrained predatorsto drill at sites where shells are thickest or wherethe shell surface is least easily reached, so increasingprey effectiveness. Still another possibility is thatapparent evolutionary changes in stereotypy bypredators reflect a replacement of one drillingpredator by another. G. Herbert and G. Dietl (pers.comm.) have evidence for this last scenario in thehistory of drilling in the lineage of the veneridbivalve Chione elevata from the late Pliocene toRecent of Florida. Clearly we need clade-level andlineage-level studies of drilling in space and time,in which both predators and their prey are examinedfrom an adaptational point of view.Climate is a factor potentially complicating theinterpretation of temporal patterns in predation.Hansen and Kelley (1995) found evidence of alatitudinal increase in unsuccessful drillingpredation during middle Eocene time (the CookMountain level in the Gulf Coastal Plain). Bycontrast, my analysis of data on effectivenessagainst drilling suggests strongly that assemblagesfrom subtropical and tropical settings show a higheraggregate effectiveness than do those from coolerwaters (represented especially by assemblages inthe middle and late Miocene of Maryland studiedby Kelley et al., 2001). In their detailed study ofthe Recent eastern American lucinid bivalve384

VERMEIJ—EVOLUTION IN THE CONSUMER AGEDivalinga quadrisulcata, Alexander and Dietl(2001) demonstrated a decreasing effectivenesswith increasing latitude from the tropical FloridaKeys to New Jersey. They found no pattern inAnadara ovalis, a species in which drilling isalmost always successful. These and other datasupport the point that evolutionary trends relatedto escalation must be evaluated within lineages orwithin adaptive syndromes from physically similarenvironments.Confirmed living marine perpetrators that drilltheir prey are mainly gastropods (Naticidae,Cassidae, Muricidae, some Marginellidae, and thedorid nudibranch Okadaia) and octopodcephalopods (Vermeij, 1987; Kabat, 1990; Ponderand Taylor, 1992). The gastropods generally attackother molluscs as well as barnacles, ostracodes,tube-dwelling polychaetes, benthic foraminifers,and even ascidians; but cassids specialize onechinoderm prey. Edge-drillers are known onlyamong polinicine naticids and various muricids(Vermeij, 1987). Octopods attack molluscs anddecapod crustaceans (Boyle and Knobloch, 1981;Dodge and Scheel, 1999).Times of first appearance of these moderndrillers are not well constrained. Undoubted naticidsare known from the earliest Cretaceous(Valanginian) onward (Riedel, 2000), but someauthors would extend their range back to the Carnianstage of the Late Triassic based on thecharacteristically beveled shape of drill-holes of thatage in bivalves from the St. Cassian Formation ofnorthern Italy (Fürsich and Jablonski, 1984). Itshould be noted, however, that beveled holes alsooccur in Paleozoic brachiopods (Smith et al., 1985),and that such holes even today are occasionallydrilled by muricids (Gordillo and Muchástegui,1998). Muricids are known from the Late Cretaceous(Campanian) onward (Merle and Pacaud, 2001), butmay extend back to the late Early Cretaceous(Albian) based on cylindrical holes in Britishbivalves of that age (Taylor et al., 1983). Undoubtedcassids date from the Late Cretaceous(Maastrichtian) (Riedel, 1995), but most drill-holesin echinoderms, presumably made by cassids, areof Cenozoic age. Drill-holes in Maastrichtianechinoids could be the work of parasitic eulimidgastropods (Kier, 1981). Marginellids appeared inthe Paleocene (Coovert and Coovert, 1995), but itis unclear when drilling members of this groupevolved. Octopods have a Jurassic origin, but againwe do not know when octopod drill-holes appear.The identity of pre-Cretaceous drillers remainscontroversial. Platyceratid gastropods, many ofwhich lived on and parasitically drilled Paleozoicechinoderms, may well be responsible for drillholesobserved in Paleozoic brachiopods(Baumiller, 1990, 1993, 1996; Baumiller andMacurda, 1995; Baumiller et al., 1999;Kowalewski et al., 2000; Leighton, 2001). Therewere no gastropods in the late Neoproterozoic orEarly Cambrian. The architects of drill-holes ofthese ages thus remain entirely unknown.Shell drilling, or at least shell puncturing, alsooccurs in land snails. The culprits appear to be variouskinds of beetles. The only systematic survey of thephenomenon is that of Ørstan (1999), and nothing isknown about the history of terrestrial drilling.Shell repair, much of it likely a response toskeletal injury by predators, has generally increasedin incidence through Phanerozoic time (Vermeijet al., 1981). The oldest known examples ofrepaired gastropod shells date from the late MiddleOrdovician (Caradoc) of Sweden (Ebbestad, 1998;Ebbestad and Peel, 1997). The number of repairsper shell in the limited Caradoc and Ashgill (LateOrdovician) fossils examined by Ebbestad and Peel(1997) was only 0.07, a very low value by LateCretaceous to Recent standards. Data from wellpreserved Late Carboniferous, Late Triassic, andEarly Cretaceous gastropods indicate higherincidences, but still are not up to the standards ofthe Late Cretaceous, late Miocene, and Recent(Vermeij et al., 1981; Pan, 1991). Sparse data ongastropods of the broadly circumscribed genusConus indicate that Miocene, Pliocene, and Recentspecies show higher frequencies of repair thanEocene species (Vermeij, 1987). Data on Cenozoicturritellid gastropods from the southeastern UnitedStates and tropical America (Allmon et al., 1990)reveal no general trend, although repair was rare inthe early Oligocene species examined. I re-385

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002examined these data in the light of Allmon’s (1996)careful evolutionary studies of these gastropods,but lineage-specific and genus-specific patternsstill do not emerge.Increases in the frequency of shell repair havealso been noted for several ammonoid lineages incomparisons of Early Jurassic (Toarcian) and MiddleJurassic (Callovian) samples from Europe (Kröger,2000) and for environmentally similar samples ofpycnodontine and exogyrine oysters duringsuccessive intervals of the Late Cretaceous in NewJersey (Dietl et al., 2000). Late Cretaceousammonoids generally have higher incidences ofrepair than their older counterparts (Landman andWaage, 1986; Monks, 2000). Many more suchstudies are needed to assess the history of shell repair.Adaptive trends in prey are observable bothwithin clades and in successive replacements ofclades in specified environments. Apertural typesthat limit entry by predators have evolved multipletimes in gastropods, mainly from the late Mesozoiconward, and species with such apertures account forincreasing proportions of the gastropod fauna inwarm-water marine assemblages through time(Vermeij, 1987). Brachiopods with spines and shellreinforcingfolds (plicae) become increasinglyfrequent in assemblages from the Early Devonianto the end of the Paleozoic (Signor and Brett, 1984;Alexander, 1986, 1990). Antipredatory adaptationsin bivalves—crenulated or scalloped valve edges,thickened valve margins, overlapping valves,cementation of one valve to the substratum, spinesand other high-relief sculpture, and external andinternal organic shell layers—have repeatedlyevolved and become more frequent, especially inthe Mesozoic and Cenozoic (Vermeij, 1987; Harper,1991; Harper and Skelton, 1993a, 1993b; Stone,1998). Among ammonoid cephalopods (EarlyDevonian to end-Cretaceous), increases insculptural relief and in the complexity of thejunction between internal septa and the outer shellwall affect many lineages, and imply increasingresistance to shell-crushing predators (Ward, 1986;Daniel et al., 1997; Saunders et al., 1999).Increasing thickness and rugosity, decreasingnumber, and more complex articulation of barnacleplates indicate post-Eocene escalation in manylineages with drilling and test-breaking predators(Palmer, 1982; Zullo, 1984).Traits associated with high locomotorperformance have also increased in frequency andexpression in most mobile groups through time.Rapidly burrowing molluscs appeared mainly inthe Late Cretaceous and Cenozoic (Vermeij, 1987).Chamberlain (1991) has documented generalincreases in speed of swimming amongcephalopods throughout their long (Late Cambrianto Recent) history (see also Ward, 1986).These changes are paralleled by increases in theinferred biting strength and specialization of skeletonbreakingmarine predators. Morphologicallyspecialized shell-breakers with flattened crushingsurfaces in jaws and claws are known from the EarlyDevonian onward. Earliest to appear were pterygotideurypterids, dipnoan lungfish, and various blunttoothedelasmobranchs. Decapod and stomatopodcrustaceans joined this cadre after the Devonian, butspecializations for crushing or cutting shells aremainly post-Cretaceous inventions. Cephalopods,which have a Late Cambrian origin, may have beenshell-breakers as early as the Ordovician, but knowncrushing jaws occur mainly in the Jurassic andCretaceous (summary in Vermeij, 1987). Hybodontsharks may have been the only shell-crushingpredators of the Early Triassic, immediatelyfollowing the end-Permian disaster (Tintori, 1998).A substantial diversification of potentiallymolluscivorous actinopterygian teleost fish beganin the Norian stage of the Late Triassic, followed bythe spectacular diversification of acanthopterygianfish from the Late Cretaceous onward (Tintori, 1998;Wainwright and Bellwood, in press). Marinetetrapods capable of breaking external skeletonsappeared (as placodont reptiles) in the MiddleTriassic, and have been represented by variousreplacing clades ever since.Predators inserting part or all of their bodiesto force entry into external skeletons withoutdrilling or breakage may extend back to the LateOrdovician, when the seastar Promopalaeasterevidently was able forcibly to insert its stomachinto bivalved prey (Blake and Guensburg, 1994).386

VERMEIJ—EVOLUTION IN THE CONSUMER AGEForcible entry by seastars may not have reappeareduntil the Early Jurassic evolution of asteriids(Blake, 1990). Latrogastropods, many of whichhave the ability to insert their proboscis into preyshells, with or without the use of force or venom,evolved in the Early Cretaceous (Riedel, 2000).Those that use force, including Miocene to Recentbusyconine whelks, are evidently of more recentorigins (Vermeij, 1987; Dietl and Alexander, 1998).The application of venom by toxoglossans (asubgroup of latrogastropods) could have originatedas early as the Late Cretaceous. The labral tooth, astructure enabling some marine latrogastropods toincrease speed of subjugation, arose at least 58times independently beginning in the Campanianstage of the Late Cretaceous (Vermeij, 2001).Among top predators in the sea, there appearsto be a long-term trend evident within as well asamong clades toward increased speed andmaneuverability. Anomalocarids of the Early andMiddle Cambrian could swim, but probably not veryfast (Chen et al., 1994; Briggs, 1994). The earliestcephalopods, housed in chambered external shellsthat initially had their long axis oriented verticallyin the water, were likely to have been slowswimmers. So were horizontally oriented, mineralballastedcephalopods in straight shells, which interms of diversity dominated cephalopodassemblages from the Early Ordovician to the EarlyDevonian. Shell coiling, invented first during theEarly Ordovician but becoming common onlyduring the Early Devonian, probably increased boththe speed and maneuverability of cephalopods, butfurther radical improvements in locomotorperformance required a partial or complete freeingof the jet-propulsion mechanism of the animal fromthe confines of an external shell. This process beganin the Early Devonian with the evolution of earlysquid-like cephalopods, and accelerated during theMesozoic when the ammonoid body and jetpropulsivemusculature may well have extendedbeyond the body chamber of the external shell.Fast, highly maneuverable squid have become thecephalopod norm in post-Cretaceous times(Chamberlain, 1991; Jacobs, 1992). Amongvertebrates, conodont animals of the Ordovician,capable of cruising but not of sprint swimming(Purnell, 1995; Gabbott et al., 1995), weresuccessively replaced as top predators by faster,more maneuverable jawed acanthodians andelasmobranchs, bony fishes, and various Mesozoicand Cenozoic marine tetrapods. Even in the slowworld of gastropods, high speed among predatorsis a geologically young attribute in many clades.On land as in the sea, top predators increasinglyemphasized both rapidity of subjugation andlocomotor performance. Serrated teeth, forexample, efficiently slice flesh, and are both latein origin and evolutionarily derived in such groupsas sharks and dinosaurs (Abler, 1992; Currie,1997). The earliest venomous tetrapod is knownfrom the Triassic (Sues, 1991). Among snakes, theability to engulf prey of weight equal to or greaterthan that of the predator is a derived, probably LateCretaceous, character; and envenomation is evenmore recent, perhaps appearing in the Eocene(Greene, 1997).If top predators exercise disproportionateevolutionary control over the phenotypes anddistributions of prey and of the other species withwhich they interact, then the overall increase inpredatory performance over the long sweep ofPhanerozoic time has some important implicationsbesides the observed predator-prey escalation. As ageneral rule, the highest-performing consumers haveevolved in productive biotas comprising manyspecies with large populations and geographicranges. On land, Late Cretaceous and Cenozoiccenters of such evolution were concentrated in Asiaand Africa; in the sea, the Indo-West Pacific regionseems to have played this role, at least fromOligocene time onward. When predators evolvingin these privileged biotas spread to smaller,biologically less sophisticated biotas, they often haddevastating effects on vulnerable species, andperhaps even drove some of them to extinction.Humans exemplify this trend all too well. Ashominids spread out of Africa, we and ourdomesticated and otherwise dependent species—parasites, commensals, and weeds—have broughtabout the extinction of thousands of species, mainlythus far on islands, in lakes, and on small continents,387

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002but increasingly also on larger continents and in theoceans. An important question for future researchersis whether or how the exercise of top-down controlby high-energy consumers has modified ecosystemsand affected the extinction of species.On the other hand, high-level consumers mayalso be responsible for selection favoring highproductivity among primary producers (landplants and phytoplankton, and attachedseaweeds). By this means, they have increasedsupply not only for themselves, but also for otherspecies, both in the habitats where the consumersoccur and in the many marginal habitats that arenutritionally subsidized by export from the centersof photosynthesis. Again, this hypothesizedgrowth-enhancing role of consumers—which, fromthe modest inception of phagocytosis in theArchean, would have become increasinglyprominent through the Phanerozoic—deservesthorough scrutiny from paleontologists andecologists interested in the reconstruction ofancient ecosystems.REFERENCESABLER, W. L. 1992. The serrated teeth of tyrannosaurid dinosaurs, and biting structures in other animals.Paleobiology, 18:161–183.AGRAWAL, A. A. 2001. Phenotypic plasticity in the interactions and evolution of species. Science, 294:321–326.ALEXANDER, R. R. 1986. Frequency of sublethal shell-breakage in articulate brachiopod assemblages throughgeologic time, p. 159–166. In P. R. Racheboeuf and C. C. Emig (eds.), Biostratigraphie du Paléozoïque 4,Les Brachiopodes fossiles et actuels.ALEXANDER, R. R. 1990. Mechanical strength of shells of selected extant articulate brachiopods: implications forPaleozoic morphologic trends. Historical Biology, 3:169–188.ALEXANDER, R. R., AND G. P. DIETL. 2001. Latitudinal trends in naticid predation on Anadara ovalis (Bruguière,1789) and Divalinga quadrisulcata (Orbigny, 1842) from New Jersey to the Florida Keys. AmericanMalacological Bulletin, 16:179–194.ALLMON, W. D. 1996. Systematics and evolution of Cenozoic American Turritellidae (Mollusca: Gastropoda), I:Paleocene and Eocene coastal plain species related to “Turritella mortoni Conrad” and “Turritella humerosaConrad.” Palaeontographica Americana, 59:7–134.ALLMON, W. D., J. C. NIEH, AND R. D. NORRIS. 1990. Drilling and peeling of turritelline gastropods since the lateCretaceous. Palaeontology, 33:595–611.AMBROSE, S. H. 2001. Paleolithic technology and human evolution. Science, 291:1748–1753.APPLETON, R. D., AND A. R. PALMER. 1988. Water-borne stimuli released by predatory crabs and damaged preyinduce more predator-resistant shells in a marine gastropod. Proceedings of the National Academy of Sciencesof the United States of America, 85:4387–4391.BABCOCK, L. E., AND R. A. ROBISON. 1989. Preferences of Palaeozoic predators. Nature, 337:695–696.BAKKER, R. T. 1983. The deer flees, the wolf pursues: incongruencies in predator-prey coevolution, p. 350–382.In D. J. Futuyma and M. Slatkin (eds.), Coevolution. Sinauer, Sunderland, MA.BALUK, W., AND A. RADWANSKI. 1996. Stomatopod predation upon gastropods from the Korytnica basin, and fromother classical Miocene localities in Europe. Acta Geologica Polonica, 76:279–304.BAMBACH, R. K. 1993. Seafood through time: changes in biomass, energetics, and productivity in the marineecosystem. Paleobiology, 19:372–397.BAMBACH, R. K. 1999. Energetics in the global marine fauna: a connection between terrestrial diversification andchange in the marine biosphere. Géobios, 32:131–144.BAMBACH, R. K. 2002. Supporting predators: Changes in the global ecosystem inferred from changes in predatordiversity. In M. Kowalewski and P. H. Kelley (eds.), The Fossil Record of Predation, Paleontological SocietySpecial Papers, 8 (this volume).BAUMILLER, T. K. 1990. Non-predatory drilling of Mississippian crinoids by platyceratid gastropods. Palaeontology,33:743–748.388

VERMEIJ—EVOLUTION IN THE CONSUMER AGEBAUMILLER, T. K. 1993. Boreholes in Devonian blastoids and their implications for boring by platyceratids. Lethaia,26:41–47.BAUMILLER, T. K. 1996. Boreholes in the Middle Devonian blastoid Heteroschisma and their implications forgastropod drilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 123:343–351.BAUMILLER, T. K., AND D. B. MACURDA. 1995. Borings in Devonian and Mississippian blastoids (Echinodermata).Journal of Paleontology 69:1084–1089.BAUMILLER, T. K., L. R. LEIGHTON, AND D. L. THOMPSON. 1999. Boreholes in Mississippian spiriferide brachiopods and theirimplications for Paleozoic gastropod drilling. Palaeogeography, Palaeoclimatology, Palaeoecology, 147:283–289.BENGTSON, S., AND Y. ZHAO. 1992. Predatorial borings in Late Precambrian mineralized exoskeletons. Science,257:367–369.BLAKE, D. B. 1990. Adaptive zones of the class Asterioidea (Echinodermata). Bulletin of Marine Science, 46:701–718.BLAKE, D. B., AND T. E. GUENSBURG. 1994. Predation by the Ordovician asteroid Promopalaeaster on a pelecypod.Lethaia, 27:235–239.BOYD, D. W., AND N. D. NEWELL. 1972. Taphonomy and diagenesis of a Permian fossil assemblage from Wyoming.Journal of Paleontology, 46:1–14.BOYLE, P. R., AND D. KNOBLOCH. 1981. Hole boring of crustacean prey by the octopus Eledone cirrhosa (Mollusca,Cephalopoda). Journal of Zoology, London, 193:1–10.BRIGGS, D. E. G. 1994. Giant predators from the Cambrian of China. Science, 264:1283–1284.BROCK, R. E., AND L. D. SMITH. 1998. Recovery of claw size and function following autotomy in Cancer productus(Decapoda: Brachyura). Biological Bulletin, 194:53–62.BURNESS, G. P., J. DIAMOND, AND T. FLANNERY. 2001. Dinosaurs, dragons, and dwarfs: the evolution of maximal bodysize. Proceedings of the National Academy of Sciences, USA, 98:14518–14521.BUTTERFIELD, N. J. 1997. Plankton ecology in the Proterozoic–Phanerozoic transition. Paleobiology, 23:247–262.CADÉE, G. C. 1994. Eider, shelduck, and other predators, the main producers of shell fragments in the WaddenSea: palaeoecological implications. Palaeontology, 37:181–202.CADÉE, G. C., S. E. WALKER, AND K. W. FLESSA. 1997. Gastropod shell repair in the intertidal of Bahía La Cholla(N. Gulf of California). Palaeogeography, Palaeoclimatology, Palaeoecology, 136:67–78.CHAMBERLAIN, J. A. JR. 1991. Cephalopod locomotor design and evolution: the constraints of jet propulsion, p. 57–98.In J. M. V. Rayner and R. J. Wootton (eds.), Biomechanics in Evolution. Cambridge University Press, Cambridge.CHEN, J-Y., L. RAMSKOLD, AND G-Q. ZHOU. 1994. Evidence for monophyly and arthropod affinity of Cambriangiant predators. Science, 264:1304–1308.COLLINS, D. 1996. The “evolution” of Anomalocaris and its classification in the arthropod class Dinocarida (nov.)and order Radiodonta (nov.). Journal of Paleontology, 70:280–293.CONWAY MORRIS, S. 1979. The Burgess Shale (Middle Cambrian) fauna. Annual Review of Ecology and Systematics,10:327–349.CONWAY MORRIS, S., AND S. BENGTSON. 1994. Cambrian predators: possible evidence for boreholes. Journal ofPaleontology, 68:1–23.COOVERT, G. A., AND P. K. COOVERT. 1995. Revision of the supraspecific classification of marginelliform gastropods.Nautilus, 109:43–110.CURRIE, P. J. 1997. Theropods, p. 216–233. In J. O. Farlow and M. M. Brett-Surman (eds.), The Complete Dinosaur.Indiana University Press, Bloomington, IN.DANIEL, T. L., B. S. HELMUTH, W. B. SAUNDERS, AND P. D. WARD. 1997. Septal complexity in ammonoid cephalopodsincreased mechanical risk and limited depth. Paleobiology, 23:470–481.DAYTON, P. K., R. J. ROSENTHAL, L. C. MAHEN, AND T. ANTEZANA. 1977. Population structure and foraging biology of thepredaceous Chilean asteroid Meyenaster gelatinosus and the escape biology of its prey. Marine Biology, 39:361–370.DIETL, G. P., AND R. R. ALEXANDER. 1998. Shell repair frequencies in whelks and moon snails from Delaware andsouthern New Jersey. Malacologia, 39:151–165.DIETL, G. P., AND R. R. ALEXANDER. 2000. Post-Miocene shift in stereotypic naticid predation on confamilial preyfrom the mid-Atlantic shelf: coevolution with dangerous prey. Palaios, 15:414–429.DIETL, G. P., R. R. ALEXANDER, AND W. F. BIEN. 2000. Escalation in late Cretaceous–early Paleocene oysters(Gryphaeidae) from the Atlantic coastal plain. Paleobiology, 26:215–237.389

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002DODGE, R., AND D. SCHEEL. 1999. Remains of the prey—recognizing the midden piles of Octopus dofleini (Wülker).Veliger, 32:260–266.EBBESTAD, J. O. R. 1998. Multiple attempted predation in the Middle Ordovician gastropod Bucania gracillima.GFF, 120:27–33.EBBESTAD, J. O. R., AND J. S. PEEL. 1997. Attempted predation and shell repair in Middle and Upper Ordoviciangastropods from Sweden. Journal of Paleontology, 71:1007–1019.ERICKSON, G. M., S. D. VAN KIRK, G. SU, M. E. LEVENSTON, W. E. CALER, AND D. R. CARTER. 1996. Bite-forceestimation of Tyrannosaurus rex from tooth-marked bones. Nature, 382:706–708.FÜRSICH, F. T., AND D. JABLONSKI. 1984. Late Triassic naticid drillholes: carnivorous gastropods gain a majoradaptation but fail to radiate. Science, 224:78–80.GABBOTT, S. E., R. J. ALDRIDGE, AND J. R. THERON. 1995. A giant conodont with preserved muscle tissue from theUpper Ordovician of South Africa. Nature, 374:800–803.GARVIE, C. L. 1991. Two new species of Muricinae from the Cretaceous and Paleocene of the Gulf Coastal Plain,with comments on the genus Odontopolys Gabb, 1860. Tulane Studies in Geology and Paleontology, 24:87–92.GEARY, D. H., W. D. ALLMON, AND M. L. REAKA-KUDLA. 1991. Stomatopod predation on fossil gastropods fromthe Plio-Pleistocene of Florida. Journal of Paleontology, 65:355–360.GORDILLO, S., AND S. N. MUCHÁSTEGUI. 1998. Estratégias de depredación del gastrópodo perforador Trophongeversianus (Pallas) (Muricoidea: Trophoninae). Malacologia, 39:83–91.GREENE, H. W. 1997. Snakes: The Evolution of Mystery in Nature. University of California Press, Berkeley.GUERRERO, R., C. PEDROS-ALIÓ, I. ESTEVE, J. MAS, D. CHASE, AND L. MARGULIS. 1986. Predatory prokaryotes:Predation and primary consumption evolved in bacteria. Proceedings of the National Academy of Sciencesof the United States of America, 83:2138–2142.HAGADORN, J. W., AND G. E. BOYAJIAN. 1997. Subtle changes in mature predator-prey systems: an example fromNeogene Turritella (Gastropoda). Palaios, 12:372–379.HANSEN, T. A., AND P. H. KELLEY. 1995. Spatial variation of naticid gastropod predation in the Eocene of NorthAmerica. Palaios, 10:268–278.HARPER, E. M. 1991. The role of predation in the evolution of cementation in bivalves. Palaeontology, 34:455–460.HARPER, E. M. 1994. Are conchiolin sheets in corbulid bivalves primarily defensive? Palaeontology, 37:551–578.HARPER, E. M., AND P. W. SKELTON. 1993a. A defensive value of the thickened periostracum in the Mytiloidea.Veliger, 36:36–42.HARPER, E. M., AND P. W. SKELTON. 1993b. The Mesozoic marine revolution and epifaunal bivalves. ScriptaGeologica, Special Issue 2:127–153.HARPER, E. M., G. T. W. FORSYTHE, AND T. PALMER. 1998. Taphonomy and the Mesozoic marine revolution:Preservation state masks the importance of boring predators. Palaios, 13:352–360.HOFFMEISTER, A. P., AND M. KOWALEWSKI. 2001. Spatial and environmental variation in the fossil record of drillingpredation: a case study from the Miocene of central Europe. Palaios, 16:566–569.JACOBS, D. K. 1992. Shape, drag, and power in ammonoid swimming. Paleobiology, 18:203–220.JONES, E. C. 1971. Isistius brasiliensis, a squaloid shark, the probable cause of crater wounds on fishes andcetaceans. Fisheries Bulletin, 69:791–798.JUANES, F., AND L. D. SMITH. 1995. The ecological consequences of limb damage and loss in decapod crustaceans:a review and prospectus. Journal of Experimental Marine Biology and Ecology, 193:197–223.KABAT, A. R. 1990. Predatory ecology of naticid gastropods with a review of shell boring predation. Malacologia,32:155–193.KARDON, G. 1998. Evidence from the fossil record of an antipredatory exaptation: conchiolin layers in corbulidbivalves. Evolution, 52:68–79.KELLEY, P. H. 1989. Evolutionary trends within bivalve prey of Chesapeake Group naticid gastropods. HistoricalBiology, 2:139–156.KELLEY, P. H. 1992. Coevolutionary patterns of naticid gastropods of the Chesapeake Group: an example ofcoevolution? Journal of Paleontology, 66:794–800.KELLEY, P. H., AND T. A. HANSEN. 1993. Evolution of the naticid gastropod predator-prey system: an evaluation ofthe hypothesis of escalation. Palaios, 8:358–375.390

VERMEIJ—EVOLUTION IN THE CONSUMER AGEKELLEY, P. H., AND T. A. HANSEN. 1996. Recovery of the naticid gastropod predator-prey system from the Cretaceous-Tertiary and Eocene-Oligocene extinctions, p. 373–386. In M. B. Hart (ed.), Biotic Recovery from MassExtinction Events. Geological Society Special Publication 102.KELLEY, P. H., T. A. HANSEN, S. E. GRAHAM, AND A. G. HUNTOON. 2001. Temporal patterns in the efficiency ofnaticid gastropod predators during the Cretaceous and Cenozoic of the United States coastal plain.Palaeogeography, Palaeoclimatology, Palaeoecology, 166:165–176.KIER, P. M. 1981. A bored Cretaceous echinoid. Journal of Paleontology, 55:656–659.KIRBY, M. X. 2001. Differences in growth rate and environment between Tertiary and Quaternary Crassostreaoysters. Paleobiology, 27:84–103.KNOLL, A. H., AND R. K. BAMBACH. 2000. Directionality in the history of life: diffusion from the left wall orrepeated scaling of the right? Paleobiology, 26 (Supplement to No. 4):1–14.KOJUMDGIEVA, E. 1975. Les gastéropodes perceurs et leurs victimes du Miocene de Bulgarie du nord-ouest.Bulgarian Academy of Sciences, Bulletin of the Geological Institute (Series Paleontology), 25:5–24.KOWALEWSKI, M., A. DULAI, AND F. T. FÜRSICH. 1998. A fossil record full of holes: the Phanerozoic history ofdrilling predation. Geology, 26:1091–1094.KOWALEWSKI, M., M. G. SIMÕES, F. F. TORELLO, L. H. C. MELLO, AND R. P. GHILARDI. 2000. Drillholes in shells ofPermian benthic invertebrates. Journal of Paleontology, 74:532–543.KRÖGER, B. 2000. Schalenverletzungen an Jurassischen Ammoniten—ihre paläontologische und paläoökologischeAussagefähigkeit. Berliner Geowissenschaftliche Abhandlungen (E), 33:1–96.LANDMAN, N. H., AND K. M. WAAGE. 1986. Shell abnormalities in scaphitid ammonites. Lethaia, 19:211–224.LEE, S. Y., AND R. SEED. 1992. Ecological implications of cheliped size in crabs: some data from Carcinus maenasand Liocarcinus holsatus. Marine Ecology Progress Series, 84:151–160.LEIGHTON, L. R. 2001. New example of Devonian predatory boreholes and the influence of brachiopod spines onpredator success. Palaeogeography, Palaeoclimatology, Palaeoecology, 165:53–69.LEONARD, G. H., M. D. BERTNESS, AND P. O. YUND. 1999. Crab predation, waterborne cues, and inducible defensesin the blue mussel, Mytilus edulis. Ecology, 80:1–14.LINDBERG, D. R., AND W. F. PONDER. 2001. The influence of classification on the evolutionary interpretation ofstructure—a re-evaluation of the evolution of the pallial cavity of gastropod molluscs. Organisms, Diversity,and Evolution, 1:273–299.MARGULIS, L. 1981. Symbiosis in cell evolution: life and its environment on the early Earth. W. H. Freeman, San Francisco.MERLE, D. 2000. Première étude taphonomique de la predation affectant de grands mollusques benthiques dansl’Eocene de Gan (Pyrénées Atlantiques, France). Comptes Rendus de l’Académie des Sciences de Paris,Sciences de la Terre et des Planètes, 330:217–220.MERLE, D., AND J-M. PACAUD. 2001. The first record of Poirieria subcristata (d’Orbigny, 1850) (Muricidae:Muricinae) in the early Cuisian of the Paris Basin (Celles-sur-Aisne, Aizy Formation), with comments onthe sculptural evolution of some Palaeocene and Eocene Poirieria and Paziella. Tertiary Research, in press.MONKS, N. 2000. Mid-Cretaceous heteromorph ammonite shell damage. Journal of Molluscan Studies, 66:283–285.ØRSTAN, A. 1999. Drill holes in land snail shells from western Turkey. Schriften zur Malakozoologie aus demHaus der Natur, 13:31–36.PALMER, A. R. 1982. Predation and parallel evolution: recurrent parietal plate reduction in balanomorph barnacles.Paleobiology, 8:31–44.PALMER, A. R. 1990. Effect of crab effluent and scent of damaged conspecifics on feeding, growth, and shellmorphology of the Atlantic dogwhelk (Nucella lapillus (L.)). Hydrobiologia, 193:155–182.PAN, H.-Z. 1991. Lower Turonian gastropod ecology and biotic interaction in Helicaulax community from westernTarim Basin, southern Xinjiang, China. Paleoecology of China, 1:266–280.PETHER, J. 1995. Belichnus new ichnogenus, a ballistic trace on mollusc shells from the Holocene of the Benguelaregion, South Africa. Journal of Paleontology, 69:171–181.PONDER, W. F., AND J. D. TAYLOR. 1992. Predatory shell drilling by two species of Austroginella (Gastropoda:Marginellidae). Journal of Zoology, London, 228:317–328.PURNELL, M. A. 1995. Microwear on conodont elements and macrophagy in the first vertebrates. Nature, 374:798–800.391

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002RAYFIELD, E. J., D. B. NORMAN, C. C. HORNER, J. R. HARNER, P. M. SMITH, J. J. THOMASON, AND P. UPCHURCH. 2001.Carnial design and function in a large theropod dinosaur. Nature, 409:1033–1037.RIEDEL, F. 1995. An outline of cassoidean phylogeny (Mollusca: Gastropoda). Contributions to Tertiary andQuaternary Geology, 32:97–132.RIEDEL, F. 2000. Ursprung und Evolution der “höheren” Caenogastropoda. Berliner GeowissenschaftlicheAbhandlungen (E), 32:1–240.RUDEBECK, G. 1950–1951. The choice of prey and methods of hunting of predatory birds with special reference totheir selective effect. Oikos, 2:65–88; 3:200–231.SAUNDERS, W. B., D. M. WORK, AND S. V. NIKOLAEVA. 1999. Evolution of complexity in Paleozoic ammnonoidsutures. Science, 286:760–763.SCHINDEL, D. E., G. J. VERMEIJ, AND E. ZIPSER. 1982. Frequencies of repaired shell fractures among the Pennsylvaniangastropods from north-central Texas. Journal of Paleontology, 56:729–740.SCHMIDT, N. 1989. Paleobiological implications of shell repair in Recent marine gastropods from the northernGulf of California. Historical Biology, 3:127–139.SIGNOR, P. W., III, AND C. E. BRETT. 1984. The mid-Paleozoic precursor to the Mesozoic marine revolution.Paleobiology, 10:229–245.SMITH, L. D. 1992. The impact of limb autotomy on mate competition in blue crabs Callinectes sapidus Rathbun.Oecologia, (Berlin) 89:494–501.SMITH, L. D., AND J. A. JENNINGS. 2000. Induced defensive responses in the bivalve Mytilus edulis to predatorswith different attack modes. Marine Biology, 136:461–469.SMITH, L. D., AND A. R. PALMER. 1994. Effects of manipulated diet on size and performance of brachyuran crabclaws. Science, 264:710–712.SMITH, S. A., C. W. THAYER, AND C. E. BRETT. 1985. Predation in the Paleozoic: gastropod-like drillholes inDevonian brachiopods. Science, 230:1033–1035.STONE, H. M. I. 1998. On predator deterrence by pronounced shell ornament in epifaunal bivalves. Palaeontology,41:1051–1068.SUES, H. D. 1991. Venom-conducting teeth in a Triassic reptile. Nature, 351:141–143.TAYLOR, J. D., R. J. CLEEVELY, AND N. J. MORRIS. 1983. Predatory gastropods and their activities in the BlackdownGreensand (Albian) of England. Palaeontology, 26:521–553.TINTORI, A. 1998. Fish biodiversity in the marine Norian (Late Triassic) of northern Italy: the first neopterygianradiation. Italian Journal of Zoology, 65 (supplement):193–199.VAN VALKENBURGH, B. 1991. Iterative evolution of hypercarnivory in canids (Mammalia: Carnivora). Paleobiology,17:340–362.VAN VALKENBURGH, B., AND F. HERTEL. 1993. Tough times at La Brea: tooth breakage in large carnivores of theLate Pleistocene. Science, 261:456–459.VANNIER, J., AND J.-Y. CHEN. 2000. The Early Cambrian colonization of pelagic niches exemplified by Isoxys(Arthropoda). Lethaia, 33:295–311.VERMEIJ, G. J. 1975. Evolution and distribution of left-handed and planispiral coiling in snails. Nature, 254:419–420.VERMEIJ, G. J. 1977. The Mesozoic marine revolution: evidence from snails, predators and grazers. Paleobiology,3:245–258.VERMEIJ, G. J. 1979. Shell architecture and causes of death in Micronesian reef snails. Evolution, 33:686–696.VERMEIJ, G. J. 1980. Drilling predation of bivalves in Guam: some paleoecological implications. Malacologia,19:329–334.VERMEIJ, G. J. 1982a. Unsuccessful predation and evolution. American Naturalist, 120:701–720.VERMEIJ, G. J. 1982b. Gastropod shell form, repair, and breakage in relation to predation by the crab Calappa.Malacologia, 23:1–12.VERMEIJ, G. J. 1987. Evolution and Escalation, an Ecological History of Life. Princeton University Press, Princeton.VERMEIJ, G. J. 1989. The origin of skeletons. Palaios, 5:585–589.VERMEIJ, G. J. 1993. A Natural History of Shells. Princeton University Press, Princeton.VERMEIJ, G. J. 1994. The evolutionary interaction among species: selection, escalation, and coevolution. AnnualReview of Ecology and Systematics, 25:219–236.392

VERMEIJ—EVOLUTION IN THE CONSUMER AGEVERMEIJ, G. J. 2001. Innovation and evolution at the edge: origins and fates of gastropods with a labral tooth.Biological Journal of the Linnean Society, 72:461–508.VERMEIJ, G. J., AND E. ZIPSER. 1986. Burrowing performance of some tropical Pacific gastropods. Veliger, 29:200–206.VERMEIJ, G. J., E. ZIPSER, AND E. C. DUDLEY. 1980. Predation in time and space: peeling and drilling in terebridgastropods. Paleobiology, 6:352–364.VERMEIJ, G. J., D. E. SCHINDEL, AND E. ZIPSER. 1981. Predation through geological time: evidence from gastropodshell repair. Science, 214:1024–1026.WARD, P. 1986. Cretaceous ammonite shell shapes. Malacologia, 27:3–28.WAINWRIGHT, P. C., AND D. R. BELLWOOD. In Press. Ecomorphology of feeding in coral reef fishes, p. 33–55. In P.F. Sale (ed.), Coral reef fishes: dynamics and diversity in a complex ecosystem. Academic Press, San Diego.ZIPSER, E., AND G. J. VERMEIJ. 1980. Survival after non-lethal shell damage in the gastropod Conus sponsalis.Micronesica, 16:229–234.ZULLO, V. A. 1984. New genera and species of balanoid barnacles from the Oligocene and Miocene of NorthCarolina. Journal of Paleontology, 58:1312–1338.393


CLOSING REMARKSTHE FOSSIL RECORD OF PREDATION:METHODS, PATTERNS, AND PROCESSESThe papers assembled in this volume showcompellingly the methodological, topical, andconceptual richness of paleontological research onpredation. Such studies span a broad spectrum oforganisms and a wide range of observational scales,from individual interactions to global-scale seculartrends. The fossil record offers us diverse andprovocative evidence of predator-prey interactionsthrough time, ranging from predation traces tofunctional morphology and phylogenetic affinities.In recent years, these diverse data have become abase for posing increasingly sophisticated questionsand for testing increasingly complex hypotheses.As clearly demonstrated in this volume,predation can be defined in a variety of ways. Apredator can be defined very broadly as “anorganism killing another organism for nutritionalpurposes” (Bengtson; see also Labandeira; Lippsand Culver; Brett and Walker), or the definitionmay be restricted behaviorally by excludingorganisms that kill by passive filter-feeding, so thatpredators are only “those organisms that hunt ortrap, subdue, and kill individual animals that havesome capacity for either protection or escape”(Bambach; see also Kowalewski). Predation as aconcept may be restricted to macro-carnivory—with a predator then defined as “a large animal orsometimes a plant consuming part or the whole ofanother animal” (Vermeij)—or can be expandedto include tiny organisms, like foraminiferasubduing and eating larger animals (Lipps andCulver), or even behaviors that involve the killingof plants (e.g., seed predation; Labandeira). Thus,the definition of predation varies significantlyamong researchers. The term predation may alsoinclude related behaviors along a spectrum fromactive predation to scavenging, from lethalpredation to partial (sublethal) predation,parasitism, and amensalism (see Baumiller andGahn; Labandeira; Kowalewski; Vermeij).METHODSThe chapters included in the first part of thevolume review a variety of methods that can beapplied to study the fossil record of predation.Various direct and indirect indicators of predationare available to paleontologists, including tracefossils, coprolites, gut contents, exceptionalpreservational events, taphonomic patterns, andindirect evidence provided by functionalmorphology and phylogenetic affinities(Kowalewski). Because of the highly disparatenature of the data itself, variable research goals,and even personal idiosyncrasies, the methodsused to study ancient predator-prey interactionsare very diverse.Among various lines of evidence, trace fossilsleft by predatory activity offer a particularly richsource of quantifiable data (Kowalewski; see alsoHaynes; Lipps and Culver; Brett and Walker; Walkerand Brett; Baumiller and Gahn; Farlow and Holtz;Bengtson; Dietl and Kelley; Vermeij); and variousanalytical approaches for studying trace fossils canbe fruitful depending on the nature of the materialand the scientific goals of paleontological projects(Kowalewski). Coprolites and stomach contents alsoprovide a wealth of direct data on ancient predatorpreyinteractions and are widely used for studyingthe diet of ancient predators, from marine invertebratesto terrestrial vertebrates (Chin; see also Brettand Walker; Walker and Brett; Labandeira; Farlowand Holtz). In addition to the predation traces studiedby paleontologists, anthropologists have developeda distinct set of qualitative and quantitative methodsfor studying the hunting behavior of hominids, notonly to distinguish between scavenging andpredation, but also to draw the much more subtledistinctions among different butchering behaviors,such as skinning, meat-stripping, or sectioning ofanimal carcasses (Haynes).395

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002The methods part of this volume shows thatthe methodological dimension of research on thefossil record of predation is a rapidly growing fieldof study, with many promising future researchdirections—particularly through laboratoryexperiments, observations in present-dayecosystems, and numerical modeling. Clearly, weneed to continue improving our statistical tools andanalytical strategies and to work together onerecting some general methodological guidelinesfor studying the fossil record of predation.PATTERNSThe chapters included in the second part of thisvolume impressively document the range ofevidence related to predator-prey interactions thathas been amassed by paleontologists. However,these studies also point to the numerous temporaland taxonomic gaps in our current knowledge.The fossil record of microorganisms providesinsights into primary producers and secondaryconsumers in marine ecosystems over the last 3.8+billion years, and points to a long-term increasein the complexity of trophic structures of marineecosystems through the history of life (Lipps andCulver). In particular, the trophic strategy ofendosymbiosis between animals and microbes,especially photosynthesizing algal eukaryotes,seems to have long played an important role inmarine ecosystem development and the evolutionof organisms (Lipps and Culver).The marine invertebrate fossil record offers aspectacular wealth of evidence relating to predatorpreyinteractions, including trace fossils, coprolites,and functional morphology of prey and predators(Brett and Walker; Walker and Brett). Thesepredatory records suggest that long-termevolutionary changes in predation pressure arelinked to episodes of abrupt biotic reorganizationduring and after mass extinctions, punctuatinglonger interludes of relative stability (Brett andWalker; Walker and Brett). Moreover, the evolutionof predation in the pelagic realm may have beenlargely decoupled from its evolution in benthicecosystems. The data from the marine fossil recordmake a strong case for the existence of predatoryattack on shelled organisms as early as the latestPrecambrian and the early Cambrian—with afurther intensification of predation during themiddle Paleozoic paralleled by increases inspinosity and other potentially defensive traits ofthe prey skeleton (Brett and Walker). In contrast,late Paleozoic forms may have taken refuge insmaller size and resistant, thicker-walled skeletons(Brett and Walker). Following the end-Permianmass extinction, the data suggest episodic, butgenerally increasing, predation pressure on marineorganisms through the Mesozoic-Cenozoicinterval. Predation in benthic communities mayhave intensified substantially in the LateCretaceous–Early Cenozoic with the evolution ofneogastropods, varied crustaceans, anddurophagous fish. In the early to mid-Mesozoic,large-predator guilds were filled predominantly byvaried marine reptiles; whereas neoselachiansharks, teleosts, and marine mammals dominatedin similar roles throughout the Late Cretaceous toCenozoic (Walker and Brett).The marine invertebrate record also suppliesevidence relating to parasite-host interactions,with a nearly even distribution of reported casesthrough the post-Cambrian Phanerozoic(Baumiller and Gahn). However, in few of thereported examples can we explicitly distinguishparasitism from predation, commensalism, ormutualism; and only in exceptional cases (suchas interactions involving platyceratids andcrinoids) is the evidence for parasitic interactionssufficiently compelling (Baumiller and Gahn).Terrestrial and freshwater invertebrates alsoprovide a diverse fossil record of predation,parasitoidism, and parasitism—with evidence forcarnivory (i.e., taxonomic affiliation, fossilstructural and functional attributes, organismicdamage, gut contents, coprolites, and mechanismsindicating predator avoidance) occurring from themid-Paleozoic to the Recent (Labandeira).However, only 12 invertebrate phyla have becomecarnivorous in the continental realm and only thetwo most diverse groups (nematodes andarthropods) left behind a comparatively goodfossil record (Labandeira).396

CLOSING REMARKSMajor groups of amniote predators such astheropod dinosaurs and carnivorous synapsids offera continuous fossil record of predator-preyinteractions in the terrestrial realm. The fossilrecord of predatory theropod dinosaurs suggeststhat the taxonomic composition of dinosaurianpredator-prey systems varied notably as a functionof time and geography (Farlow and Holtz). Detailsregarding diet and hunting behavior of theropodscan be inferred from their functional morphology,supported by evidence from taphonomicassociations with likely prey species, bite marks,gut contents, coprolites, and trackways (Farlow andHoltz). Following the K-T extinction, carnivorousbirds (the direct descendants of theropods)remained prominent predators throughout theCenozoic Era (Farlow and Holtz).The fossil record of synapsids points tosignificant parallels between the diversification ofnon-mammalian synapsid predators in the LateCarboniferous–Triassic and the Cenozoicradiation of mammalian predators: both groupsevolved sabertooth forms as well as short-snouted,powerful biting forms (Van Valkenburg andJenkins). Both radiations are characterized byrepeated patterns in which one or a few cladesevolved large size and dominated the carnivoreguild for several million years, but then declinedand were replaced by new taxa (Van Valkenburghand Jenkins). Both non-mammalian andmammalian synapsid clades show trends towardincreasing body size and hypercarnivory over time(Van Valkenburgh and Jenkins).PROCESSESThe last part of this volume includes chaptersthat provide an introduction to some major modelsregarding the origin and history of predation as wellas the evolutionary role of predator-preyinteractions through time.Although data on early life are understandablyscarce, existing data and theoretical considerationssuggest that predation may have played animportant role in some of the major transitions inevolution, including the origin of eukaryotic cells,the origin of multicellularity (as a means ofacquiring larger size), the decline of stromatolites,the diversification of acritarchs, and the Cambrianexplosion (Bengtson; see also Lipps and Culver).Predation may have been a decisive selective forcebehind the transition from simple, mostly microbial,ecosystems to those with complex food webs andhigher-order consumers (Bengtson; Lipps andCulver). Following the Cambrian explosion, thediversity of predators and the proportion of the totalfauna represented by predators have both increasedthroughout the Phanerozoic, implying thatecosystems have increased their ability to supporteither more predators or more specialization amongpredators (Bambach). This pattern may be linked toa secular increase in diversity and biomass ofprimary producers, and changes in the compositionof prey taxa (Bambach).The evolutionary importance of predationremains a hotly debated topic. Arms races betweenpredators and prey may be driven by two relatedprocesses: escalation—enemy-driven evolution, inwhich the role of prey in the evolution of thepredator is downplayed—and coevolution—inwhich two or more interacting species respondreciprocally to one another; prey are thought todrive the evolution of their predator, and vice versa(Dietl and Kelley; Vermeij). In the fossil record,the two processes are distinguished most reliablywhen the predator-prey system is viewed within thecontext of the other species that may influence theinteraction, thus allowing for a relative ranking ofthe importance of selective agents (Dietl and Kelley).Scale is also important in evaluating the role ofescalation and coevolution in the development ofspecies interactions: prey are likely to exert someselective pressure on their predators over ecologicaltime scales, but predators may still exert primary“top-down” control in directing evolution overevolutionary timescales. In the long term, predatorshave two principal effects: they influence preyphenotypes and they restrict prey to environmentswhere predators are rare (Vermeij). Predators likelycontrol overall directionality in evolution becauseof the inequalities of predator and prey in control ofresources (Vermeij; Dietl and Kelley). Indeed,predators have the evolutionary upper hand over397

PALEONTOLOGICAL SOCIETY PAPERS, V. 8, 2002the long run, especially in the expression of sensorycapacities, locomotor performance, and theapplication of force; and only in passive defenses(armor, toxicity, large body size) does escalationfavor the prey (Vermeij). The evidence provided bythe fossil record points to increases over time in bothpredator power and prey defenses. These escalatoryincreases have proceeded episodically (Brett andWalker; Walker and Brett; Vermeij) against abackdrop of generally increasing productivity(Bambach) and increasing top-down evolutionarycontrol by high-energy predators (Vermeij).MICHAL KOWALEWSKI, PATRICIA H. KELLEY, RICHARD K. BAMBACH, TOMASZ K. BAUMILLER, STEFANBENGTSON, CARLTON E. BRETT, KAREN CHIN, STEPHEN J. CULVER, GREGORY P. DIETL, JAMES O.FARLOW, FOREST J. GAHN, GARY HAYNES, THOMAS R. HOLTZ, JR., IAN JENKINS, CONRAD C.LABANDEIRA, JERE H. LIPPS, BLAIRE VAN VALKENBURGH, GEERAT J. VERMEIJ, AND SALLY E. WALKER398

PALEONTOLOGICAL SOCIETY SPECIAL STUDIES, 1980-2002Short Course NotesEchinoderms. Edited by T. W. Broadhead and J. A. Waters, 1980 (Out of Print)Lophophorates. Organized by J. T. Dutro, Jr. and R. S. Boardman, 1981 (Out of Print)Foraminifera. Organized by M. A. Buzas and B. K. Sen Gupta, 1982 (Out of Print)Sponges and Spongiomorphs. Organized by J. K. Rigby and C. W. Stearn, 1983 (Out of Print)Mammals. Organized by P. D. Gingerich and C. E. Badgley, 1984, 234 pp., $20Mollusks. Organized by D. J. Bottjer, C. S. Hichman, and P. D. Ward, 1985, 305 pp., $20Land Plants. Organized by R. A. Gastaldo, 1986, 226 pp., $20Fossil Prokaryotes and Protists. Organized by J. H. Lipps, 1987, 303 pp., $5Short Courses in PaleontologyNumber 1: Molecular Evolution and the Fossil Record. Organized by B. Runnegar and J. W. Schopf, 1988,167 pp. (Out of Print)Number 2: The Age of Dinosaurs. Organized by K. Padian and D. J. Chure, 1989, 210 pp. (Out of Print)Number 3: Arthropod Paleobiology. Organized by D. L. Mikulic, 1990, 315 pp., $15Number 4: Analytical Paleobiology. Organized by N. L. Gilinsky and P. W. Signor, 1991, 216 pp., $20Number 5: Trace Fossils. Organized and edited by C. G. Maples and R. R. West, 1992, 238 pp., $15Number 6: Taphonomic Approaches to Time Resolution in Fossil Assemblages. Organized and edited byS. M. Kidwell and A. K. Behrensmeyer, 1993, 302 pp., $20Number 7: Major Features of Vertebrate Evolution. Organized and edited by D. R. Prothero and R. M.Schoch, 1994, 270 pp., $20Number 8: Siliceous Microfossils. Organized by C. D. Blome, P. M. Whalen, and K. M. Reed, 1995, 185 pp., $20Special PublicationsVolume 1: The Evolution-Creation Controversy: Perspectives on Religion, Science, and Education – AHandbook. (Out of Print)Volume 2: Paleoecology and Taphonomy of Pleistocene to Recent Intertidal Deposits, Gulf of California.Edited by K. W. Flessa, 1987, 237 pp., $5Volume 3: Methods and Applications of Plant Paleoecology. Edited by W. A. DiMichele and S. L. Wing,1988, 171 pp., $5Volume 4: Paleotechniques. Edited by R. M. Feldmann, R. E. Chapman, and J. T. Hannibal, 1989, 358 pp., $20Volume 5: Paleocommunity Temporal Dynamics: The Long-term Development of MultispeciesAssemblies. Edited by W. Miller, III, 1990, 421 pp. (Out of Print)Volume 6: Fifth North American Paleontological Convention – Abstracts and Program. Edited by S.Lidgard and P. R. Crane, 1992, (Out of Print)Volume 7: Dino Fest. Edited by G. D. Rosenberg and D. L. Wolberg, 1994, 500 pp., $20Volume 8: Sixth North American Paleontological Convention Abstracts of Papers. Edited by J. E.Repetski, Jr., 1996, 443 pp., $5Volume 9: Evolution: Investigating the Evidence. (1st ed.) Edited by J. Scotchmoor and D. A. Springer, 2000,357 pp. (Out of Print)Volume 10: Guidelines for the Management and Curation of Invertebrate Fossil Collections Including a DataModel and Standards for Computerization. Edited by R. D. White and W. D. Allmon, 2000, 260 pp., $15Volume 11: Evolution: Investigating the Evidence. (2nd ed.) Edited by J. Scotchmoor and D. A. Springer,2002, 330 pp., $20Prices and conditions subject to change. See next page for ordering information.

THE PALEONTOLOGICAL SOCIETY PAPERSPALEONTOLOGICAL SOCIETY PAPERS (PSP) is a new subscription series opened by theSociety in 1996. One or more numbers will be published annually, subjects variable, lengthvariable, priced individually, format quarto. Individual and library rates are the same. Standingorders are available.❏ PSP #1 BIOLOGY AND PALEOBIOLOGY OF CORALSG. D. Stanley, Jr. (ed.), 1996, 296 pp. $20.00❏ PSP #2 LEARNING FROM THE FOSSIL RECORDJ. Scotchmoor and F. K. McKinney (eds.), 1996, 348 pp. $20.00❏ PSP #3 GEOBIOLOGY OF ECHINODERMSJ. A. Waters and C. G. Maples (eds.), 1997, 355 pp.❏ PSP #4 ISOTOPE PALEOBIOLOGY AND PALEOECOLOGYR. D. Norris and R. M. Corfield (conv.), 1998, 285 pp. $20.00❏ PSP #5 THE EVOLUTION-CREATION CONTROVERSY II:PERSPECTIVES ON SCIENCE, RELIGION,AND GEOLOGICAL EDUCATIONP. H. Kelley, J. R. Bryan, and T.A. Hansen (eds.) 1999, 241 pp. $20.00❏ PSP #6 PHANEROZOIC TERRESTRIAL ECOSYSTEMS $20.00R. A. Gastaldo and W. A. DiMichele (eds). 2000, 306 pp.❏ PSP #7 BRACHIOPODS ANCIENT AND MODERN: $20.00A Tribute to G. Arthrur CooperS. J. Carlson and M. R. Sandy (eds). 2001, 261 pp.❏ PSP #8 THE FOSSIL RECORD OF PREDATION $20.00M. Kowalewski and P. H. Kelley (eds). 2002, 398 pp.Order from: The Paleontological Society – Special Studies Contact: Ilona WeyersCarnegie Museum of Natural History Voice: (412) 622-19474400 Forbes Avenue Fax: (412) 622-8837Pittsburgh, PA 15213-4080, USA Email: weyersi@clpgh.orgHow to order: Orders can be placed by mail, phone, fax, or e-mail.Payment: Payment can be made by check or VISA/MasterCard. Please pay in US$and make checks payable to: The Paleontological SocietyForeign Orders: Add US$ 4.00 postage per book ($20 max., additional books are shipped free) fororders outside North America plus a one-time processing fee of $1.50 per order.Foreign orders must be paid by a check drawn on a New York City Bank,Interntional Money order, or a VISA/MasterCard credit card.Prices are effective on July 1, 2001

View - Kowalewski, M. - Virginia Tech

Create successful ePaper yourself

Delete template?

Save as template?