MOLECULAR PHYLOGENETICS AND EVOLUTION Molecular Phylogenetics and Evolution 23 (2002) 171–188 www.academicpress.com Extraordinary haplotype diversity in haplodiploid inbreeders: phylogenetics and evolution of the bark beetle genus Coccotrypes Bjarte H. Jordal,a,* Benjamin B. Normark,b Brian D. Farrell,c and Lawrence R. Kirkendalld a School of Biological Sciences, University of East Anglia, NR4 7TJ, Norwich, UK Department of Entomology, Fernald Hall, University of Massachusetts, Amherst, MA 01003, USA Museum of Comparative Zoology, Harvard University, 26 Oxford St., Cambridge, MA 02138, USA d Department of Zoology, University of Bergen, Allegt. 41, N-5007 Bergen, Norway b c Received 9 July 2001; received in revised form 19 November 2001 Abstract Regular inbreeding by sib-mating is one of the most successful ecological strategies in the bark beetle family Scolytinae. Within this family, the many species (119) in Coccotrypes are found breeding in an exceptional variety of untraditional woody tissues different from bark and phloem. Species delineation by morphological criteria is extremely difficult, however, as in most other inbreeding groups of beetles, perhaps due to the unusual evolutionary dynamics characterizing sib-mating organisms. Hence, we here performed a phylogenetic analysis using molecular data in conjunction with morphological data to better understand morphological and ecological evolution in this sib-mating group. We used partial DNA sequences from the nuclear gene EF-a and the mitochondrial genes 12S and CO1 to elucidate patterns of morphological evolution, haplotype variation, and evolutionary pathways in resource use. Sequence variation was high among species and far above that expected at the species level (e.g., 19% for CO1 within Coccotrypes advena). The tendency for exhaustive sequence variation at deeper nodes resulted in ambiguous reconstructions of the deepest splits. However, all results suggested that species with the broadest diets were clustered in a single derived position—another piece of evidence against specialization as a derived evolutionary feature. Ó 2002 Elsevier Science (USA). All rights reserved. Keywords: 12S; CO1; Dryocoetini; EF-1a; Maximum-likelihood; Maximum-parsimony; Morphology; RASA; Scolytinae; Seed feeding; Specialization 1. Introduction Our empirical understanding of the origin and evolution of haplodiploid clades is poor. Most extant haplodiploid clades are very ancient and divergent from their closest relatives, and our understanding of their origins is hampered by uncertainty about phylogenetic topology (e.g., pinworms, iceryine scale insects), by uncertainty about reconstruction of ancestral ecological characters (e.g., Hymenoptera, monogonont rotifers), or by both sources of uncertainty at once (e.g., mite clades, Thysanoptera, whiteflies) (Normark et al., 1999). An exception to this generalization is the haplodiploid, inbreeding clade of scolytine beetles within the tribe * Corresponding author. Fax: +44(0)1603592250. E-mail address: b.jordal@vea.ac.uk (B.H. Jordal). Dryocoetini. We now have a precise understanding of its basal relationships (Jordal et al., 2000, 2002; Normark et al., 1999), across which many important ecological characters remain constant, yielding an exceptionally clear picture of the ecological context in which haplodiploidy arose in this case. However, important sources of uncertainty remain in our understanding of the evolution of the clade. We now have clear evidence that the basal lineage in the clade is the genus Ozopemon Hagedorn (Jordal et al., 2002), and the preponderance of the evidence suggests that the remainder of the clade consists of two monophyletic sister taxa: (a) the tribe Xyleborini and (b) the genus Coccotrypes Eichhoff (including Dryocoetiops; see Jordal et al., 2000, 2002). Two of these taxa are fairly easy to characterize ecologically: all Ozopemon are phloem-feeders and all Xyleborini feed on ambrosia fungi cultured on the walls of tunnels bored into xylem. 1055-7903/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved. PII: S 1 0 5 5 - 7 9 0 3 ( 0 2 ) 0 0 0 1 3 - 1 172 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 However, Coccotrypes is not so easily characterized. Though it has barely more than 10% of the number of species found in Xyleborini, the range of their feeding habits is vastly greater. Some of them, like Ozopemon, have retained the ancestral character of feeding in phloem, but almost all of them have dwarf males like Xyleborini. Clearly, Coccotrypes is a critical group for understanding the evolution of this haplodiploid clade. Coccotrypes is a fairly large genus of sib-mating beetles, with currently 119 species assigned to it (Bright and Skidmore, 1997; Wood and Bright, 1992). Most species are distributed in humid tropical forests of Asia and Africa, and recent introductions of several species into the Neotropics has made the genus circumtropical (Bright and Peck, 1998; Wood, 1982). Among the most widespread species, we find a few species that also are well adapted to drier conditions in warm temperate areas, where they feed and breed in small seeds, in particular palm seeds. The best known species, the ‘‘ivory button’’ beetle Coccotrypes dactyliperda (Fabr.) is well known as a pest of palm seeds (Blumberg and Kehat, 1982; Hamilton, 1993; Herfs, 1950). Overall, the species of Coccotrypes have adopted a wide range of unusual resources for feeding and breeding. They have been collected from many different fruits and seeds (Browne, 1961), leafstalks (Beaver, 1979a; Jordal and Kirkendall, 1998), phloem, and the pith of twigs (Browne, 1959). Coccotrypes is the only bark beetle genus known to breed in ferns (Browne, 1973; Gray, 1970, 1972) and in mangrove radicles (Browne, 1961; Woodruff, 1970). In contrast to the narrow specialization of most scolytines on a particular host tissue—usually phloem, as the common name ‘‘bark beetle’’ implies—fewer than 10 species of Coccotrypes use phloem exclusively (Browne, 1959). In contrast, many species of Coccotrypes are host-tissue (and hosttaxon) generalists, capable of breeding in two or more widely different tissues, usually in a wide range of tree species. The reproductive biology of the entire haplodiploid clade (1400 spp) is characterized by strongly femalebiased offspring sex ratios, in which a single dwarfed male develops in close association with his many sisters (Kirkendall, 1993). In all cases where experimental or cytological studies have been reported, the males are produced by arrhenotokous parthenogenesis (Entwistle, 1964; Herfs, 1950; Ueda, 1997) and are haploid (Takenouchi and Takagi, 1967; B.B.N. and B.H.J., unpublished flow cytometry data for C. dactyliperda). In conjunction with regular sib-mating, haplodiploidy contributes an advantage to colonizing females, which can either mate with a brother before dispersal, or produce haploid sons parthenogenetically, one of which she can eventually mate with (Kirkendall, 1993). Sibmating species of Scolytinae contribute a disproportionately high fraction of the bark and timber beetle faunas on tropical islands, and the wide distribution and ecological success for sib-mating species might be explained by their extraordinary reproductive biology (Jordal et al., 2001). Inbreeding species complexes often tend to be taxonomically difficult to resolve, and inbreeding groups of Scolytinae are no exception (Wood, 1982). Until recently, Coccotrypes and Poecilips Schaufuss were considered two genera difficult to distinguish (Browne, 1961) and the latter was synonomized when complete character intergradation was found (Wood, 1973). Long lists of generic and specific synonymies (Wood and Bright, 1992) reflect the problematic classification of this genus, whose species are regarded particularly difficult to identify (Wood, 1986). A similar pattern of ambiguous morphology and taxonomic synonymies is also observed in the inbreeding Cryphalini and Xyleborini (Wood and Bright, 1992), demonstrating a general taxonomic problem with sib-mating species. The genus Coccotrypes is of great interest due to its remarkable sib-mating habits, its extremely great lability of ecological characters, both within and between species, compared with other scolytine beetles, and its pivotal position in the evolutionary history of a haplodiploid clade. Here we present the first phylogenetic study of the genus, using characters from mitochondrial and nuclear DNA, morphology, and behavior. 2. Material and methods 2.1. Specimens The species analyzed, the GenBank accession numbers, and their geographical sources and main food resources are listed in Table 1. Samples were collected into 100% ethanol or acetone by the senior author or persons credited in the acknowledgements. Vouchers are deposited at the Museum of Comparative Zoology, Harvard University or at the Department of Zoology, University of Bergen. We included 26 specimens from 19 species of Coccotrypes which were selected to represent the overall variation in morphological and ecological characters. Two species of Dryocoetiops, which are known to be part of a paraphyletic Coccotrypes (Jordal et al., 2000, 2002, and the outgroups Ozopemon and Dryocoetes were also included. 2.2. Morphological and behavioral characters Adults of each species were extensively examined and compared, and 28 morphological characters were coded for phylogenetic analysis. Because larval characters are only marginally informative for scolytine genera (Lekander, 1968), larvae were not examined. In 173 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Table 1 Collecting area for the species sequenced in this study, along with their main food resource (monophagous resources in boldface), and GenBank accession numbers Species Alias Collection Food source Acc. EF-1a Acc. 12S Acc. CO1 Coccotrypes aciculatus Schedl Coccotrypes advena Blandford dry16 scd24 Equador Uganda — AF186664 AF444047 AF259818 AF444059 AF187116 Coccotrypes advena Blandford dry11 Uganda AF444072 AF444044 AF444056 Coccotrypes advena Blandford scd32 Japan AF186668 AF259821 AF187120 Coccotrypes advena Blandford scd36 Borneo AF259868 AF259824 AF444068 Coccotrypes advena Blandford dry20 Costa Rica AF444076 AF444050 AF444062 Coccotrypes cardamomi Schaufuss scd38 Japan AF259869 AF259826 AF375308 Coccotrypes carpophagus (Hornung) Coccotrypes cf. cardamomi scd43 dry12 USA, Florida PNG AF259872 AF444073 AF259829 AF444045 AF444069 AF444057 Coccotrypes cyperi (Beeson) scd11 Costa Rica AF259863 AF259813 AF375309 Coccotrypes cyperi (Beeson) dry15 Queensland Seed seed, fruit, petiole, phloem seed, fruit, petiole, phloem seed, fruit, petiole, phloem seed, fruit, petiole, phloem seed, fruit, petiole, phloem seed, fruit, petiole, fern, phloem palm seed petiole, palm seed, more? seed, fruit, petiole, phloem, pith seed, fruit, petiole, phloem, pith palm seed palm seed palm seed petiole, more? mangrove radicle mangrove radicle seed, fruit, petiole, pith small seed small seed mangrove radicle phloem (also under bark of petioles) petiole phloem, petiole, fruit petiole fruit, seed, petiole, phloem phloem phloem pith pith (also pith of petiole) phloem phloem AF444074 AF444046 AF444058 AF186659 AF444078 AF444075 AF444071 AF259867 AF259815 AF444052 AF444048 AF444043 AF444053 AF444049 AF259823 AF187111 AF444064 AF444060 AF444055 AF444066 AF444061 AF375310 AF259866 AF259874 AF259864 AF259871 AF259820 AF259831 AF259814 AF259828 AF438513 AF444070 AF444065 AF438515 AF186669 AF259875 AF259822 AF259832 AF187121 AF438518 AF444077 AF259865 AF444051 AF259819 AF444063 AF444067 AF259817 AF259873 AF439741 AF186670 AF259817 AF259830 AF438492 AF259825 AF187113 AF444054 AF438507 AF187122 AF259870 AF439740 AF259827 AF438491 AF438514 AF438506 Coccotrypes Coccotrypes Coccotrypes Coccotrypes Coccotrypes Coccotrypes Coccotrypes dactyliperda (Fabricius) dactyliperda (Fabricius) cf. distinctus (Motschulsky) sp. near rhizophorae fallax (Eggers) fallax (Eggers) gedeanus (Eggers) scd14 dry22 dry17 dry09 scd13 dryl8 scd35 Argentina USA, Florida Costa Rica PNG Queensland Bangladesh Borneo Coccotrypes Coccotrypes Coccotrypes Coccotrypes graniceps (Eichhoff) impressus Eggers litoralis (Beeson) longior (Eggers) scd31 scd54 scd12 scd40 Japan Thailand Bangladesh Borneo Coccotrypes marginatus (Browne) Coccotrypes medius (Eggers) scd34 scd55 Singapore Singapore Coccotrypes petioli (Browne) Coccotrypes variabilis (Beeson) dry21 scd30 Borneo Japan Dryocoetes affaber (Mannerhein) Dryocoetes autographus (Ratzeburg) Dryocoetiops cf. eugeniae (Schedl) Dryocoetiops coffeae (Eggers) scd20 scd53 dry19 scd37 USA, NH Japan Malaysia Borneo Ozopemon brownei Schedl Ozopemon uniseriatus Eggers scd39 dry04 Borneo PNG — — Note. Alias refers to voucher specimens (see Section 2). addition, 1 developmental character and 2 behavioral characters were compiled from the literature and checked against the location, structure, and occupancy of the gallery systems from which the beetles were collected (Appendix A). These characters were included in the ‘‘morphology’’ data partition. The evolution of each character was traced on the total-evidence trees, and consistency indices for characters from different body regions were tabulated and compared. Particular attention was paid to the evolution of character 30, resource use. 2.3. DNA sequencing and alignments DNA was extracted and amplified using primers and protocols described elsewhere (Jordal et al., 2000; 174 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Normark et al., 1999). Sequences were assembled and edited in Sequencher 3.1 (Gene Codes Corp., Ann Arbor, MI). Of the two different copies of EF-1a, we used the copy that has only one intron, between coding positions 753 and 754 in scolytine beetles (Danforth and Ji, 1998; Normark et al., 1999). The intron was aligned and used in the analyses. For CO1 and the EF-1a coding region, alignments were unambiguous due to the lack of any insertions or deletions in the coding region. Alignments of 12S and the EF-1a intron were done in ClustalX (Thompson et al., 1997) under four different gap cost ratios (2, 4, 8, and 16) with transitions and transversions weighted equally and with opening and extension gaps equally costly. Maximum-parsimony (MP) trees were inferred by heuristic search (see below) for each alignment. For the EF-1a intron, the strict consensus tree for each alignment was compared by inspection with the strict consensus tree for the EF-1a coding region, and the alignment yielding the most congruent consensus tree (gap cost 4) was chosen for further analysis. Eight base pairs consisting of AT-repeats unique to C. litoralis were deleted from the EF-1a intron matrix. For 12S, an elision matrix was constructed (Wheeler et al., 1995) and the strict consensus trees for the individual alignments were compared with the strict consensus tree for the elision matrix. The alignment yielding the most congruent consensus tree (gap cost 8) was chosen for further analysis. Two inserted sites unique to C. longior were deleted; the matrix was reanalyzed (again with gap cost 8), resulting in a 12S matrix with no gaps occurring in that region. After alignment, the combined matrix was 2239 bp in length, including 324 aligned sites of 12S and 65 aligned sites of EF-1a intron. Due to the high degree of consistency between different alignments, the few remaining gaps were each treated as a fifth character state. 2.4. Optimal outgroup analysis A preponderance of molecular and morphological evidence suggests that Coccotrypes + Dryocoetiops constitutes a monophyletic group whose sister group is the tribe Xyleborini (Jordal, 2001). However, the precise relationship between Xyleborini and Coccotrypes + Dryocoetiops is unclear; one may be paraphyletic with respect to the other (Jordal et al., 2000). We are sufficiently persuaded of the monophyly of Coccotrypes + Dryocoetiops to treat them as monophyletic in this paper, but for rooting their phylogeny we have conservatively chosen the unequivocal outgroup Ozopemon (Jordal et al., 2000, 2002; Farrell et al., 2001). As a further outgroup to test for sensitivity to outgroup composition in a functional ingroup/outgroup analysis (Watrous and Wheeler, 1981), we have selected the sister group of the haplodiploid clade, the genus Dryocoetes (Normark et al., 1999). The program RASA (Lyons-Weiler, 1998) was used to explore putative plesiomorphy contents of the different outgroup genera (Lyons-Weiler et al., 1998), separate and in combination, using all data unweighted. The outgroup composition that results in the highest increase in phylogenetic signal for the ingroup in a rooted analyses is supposed to provide the best estimate of the ingroup root. 2.5. Phylogenetic analysis MacClade was used for editing matrices, translating to amino acids, translating to standard format, and calculating substitution frequencies. Paup* 4.0 (Swofford, 1999) was used for calculating pairwise sequence transitions and transversions, base composition, sequence divergences, and tree statistics and for all phylogenetic analyses. Under the maximum-parsimony criterion, we performed 100 random-addition replicates of heuristic searches for each of the data partitions and the combined matrix. Bootstrap support for individual nodes was assessed by 100 bootstrap replicates (Felsenstein, 1985) of 10 random-addition heuristic searches each. For the least variable data partition, the morphology matrix, we set maxtrees to 1000 during bootstrapping. Maximum-likelihood (ML) was used on the molecular data only, with parameters estimated on the topology resulting from the weighted parsimony analysis. We used a general time reversible model in which the proportion of invariable sites and a gamma distribution of rates of substitution were estimated. Several iterations were done to optimize these parameters, until the topology did not change. We used the incongruence difference length (ILD) test (Farris et al., 1995a,b) to measure level of incongruence between data sets. Although we would ultimately combine all data in a total-evidence approach (Kluge, 1989, 1998), we used the ILD test to evaluate eventual lack of resolution or topological conflicts. TreeRot (Sorenson, 1999) was used to calculate partitioned Bremer support indices (Baker and DeSalle, 1997; Baker et al., 1998) showing the relative contribution from each data partition to the combined analysis. We used heuristic searches with 20 randomaddition sequences for each node in each of the partitions. We further explored the relative performance of each data partition using RASA to measure phylogenetic signal (Lyons-Weiler et al., 1996) in the ingroup for each of the data partitions and compared these measures to the overall bootstrap support and saturation plots for the same partitions. We used this information to inform differential weighting between partitions. 175 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 3. Results 3.1. Sequence structure and divergence There was a slight (EF-1a coding region) to strong (mitochondrial and EF-1a intron) AT bias for all gene fragments (Table 2). This bias was also associated with generally high substitution rates, in particular in the mitochondrial 12S and CO1 third codon positions and in the nuclear noncoding EF-1a intron (Fig. 1). While there was no significant heterogeneity in base composition between taxa, among-site rate variation was pronounced within the protein-encoding EF-1a and CO1 (Table 3), reflecting different rates of substitution in the three codon positions. A direct comparison between the EF-1a coding region and the CO1 highlights the threefold higher divergence in the latter, for the ingroup (Table 3). The maximum intrageneric CO1 divergence was 20.4% (uncorrected), and most pairwise comparisons had more than 15% divergence, with slightly lower levels for 12S. Even between individuals of the morphologically nearly invariable C. advena complex, we observed a remarkable 19.0% divergence in the CO1 sequences. High variation was also observed between individuals of the mangrove beetle, C. fallax (9.5%). In contrast, the morphologically distinguishable species C. cyperi and another mangrove beetle C. litoralis were only 2.6% divergent, indicating more rapid morphological and ecological evolution in this clade. The high rates of substitution in CO1 third positions resulted in high levels of homoplasy (Fig. 1). More than half of the informative characters for these sites were changing eight times or more over the total-evidence tree. This same subpartition also revealed evident saturation from the plot of transitions vs transversions, as did 12S (Fig. 1). On the contrary, we observed a complete absence of saturation of transitions in the least variable molecular data partition, EF-1a, for the ingroup. More than 80% of the variable characters from the latter partition changed fewer than four times through the totalevidence tree. Despite the high sequence divergence found in the much more rapidly evolving EF-1a intron, this subpartition mimicked the saturation plot for the coding region (Fig. 1), albeit with a much lower relative rate of transitional substitutions (Table 3). Saturation in both of the mtDNA gene partitions was also evident from exhaustive sequence divergence observed at deeper nodes (Table 3). This trend was not observed in the EF-1a intron which instead revealed considerable increase in sequence divergence also from the ingroup to the primary outgroup Ozopemon. The exhaustive sequence variation and marked saturation in transitional substitutions in the two mtDNA gene partitions also coincided with the lowest phylogenetic signals measured by the RASA regression method (Table 4). Table 2 Base composition for partial gene sequences Locus Position A C G T AT-bias EF-1a 1 2 3 1+2+3 Intron 0.29 0.30 0.23 0.27 0.28 0.17 0.22 0.26 0.22 0.13 0.34 0.15 0.13 0.21 0.08 0.20 0.32 0.40 0.30 0.52 0.49 0.62 0.63 0.57 0.80 CO1 1 2 3 1+2+3 0.30 0.19 0.48 0.33 0.18 0.24 0.16 0.19 0.26 0.16 0.03 0.15 0.25 0.41 0.33 0.33 0.55 0.60 0.81 0.66 12S All 0.39 0.19 0.09 0.33 0.72 3.2. Separate analyses As a means to assess the relative contribution from each data partition to the total-evidence hypothesis, we analyzed each partition separately (Fig. 2). Clearly, the EF-1a data, which showed the lowest substitution frequency (Fig. 1) and the second highest RASA signal (Table 4), had also the highest overall bootstrap support and the highest proportion (52%) of partial Bremer support values (Table 5). Adding the EF-1a intron to the analysis of the coding region resulted in higher resolution with increased bootstrap support. Although CO1 also contributed much to the total Bremer support, half of this support was confined to two nodes involving conspecifics (nodes 10 and 23 in Fig. 3). EF-1a resulted in a consensus topology most similar to the consensus resulting from morphological data (Fig. 2). With the inclusion of the mtDNA gene fragments, which had the lowest RASA signal and the fewest bootstrap-supported nodes, the overall support did not improve. If Ozopemon was used as the sole outgroup, the number of supported nodes actually decreased with the addition of mtDNA data. Separate analyses of amino-acid-translated CO1 sequences did not perform much better than CO1 nucleotides (cf., Pruess et al., 2000; Simmons, 2000). However, when CO1 amino acids were substituted for nucleotides in the combined analysis, the number of ingroup nodes with more than 50% bootstrap support increased from 12 to 15. 3.3. Combined analyses: optimal outgroup analysis The ingroup alone contained significant and high phylogenetic signal (tRASA ¼ 11.82) for all data combined (Table 4). The outgroup designation with the highest increase in signal for the ingroup was the phylogenetic sister group Ozopemon alone (tRASA ¼ 19.48), followed by all outgroups simultaneously (tRASA ¼ 17.54). However, because both of the Ozopemon species had significantly long branches as 176 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Fig. 1. DNA substitution patterns of the three gene fragments analyzed. Numbers of nucleotide changes were estimated over the tree topology depicted in Fig. 3. Plots of transitions vs transversions reflect pairwise divergences. For EF-1a, open bars and filled dots indicate third positions for the exon and hatched bars and open dots indicate the intron. For CO1, open bars and dots indicate first + second codon positions and hatched bars and filled dots indicate third positions only. indicated by the taxon variance ratio plot (Lyons-Weiler and Hoelzer, 1997), we should be cautious with interpreting topologies where only Ozopemon are used as outgroups, and hence we continue here with exploring the effect of outgroup composition on the ingroup rooting, with preferentially all outgroups included. 177 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Table 3 Sequence variation for different subsets of the genes used in this study Locus Position Total characters Informative characters ti/tv ratiosa Max divergence C. advena Max divergence Coccotrypes Max divergence Cocc + Ozop Max divergence Cocc + Dryo EF-1a 1 2 3 All coding Intron 285 285 284 854 65 9 6 139 154 44 14.3 0.6 4.3 – 1.1b 1.1 1.4 10.6 4.3 15.2 2.4 1.4 17.7 6.2 34.8 2.1 1.4 27.9 10.5 41.3 3.3 2.2 27.3 10.6 41.3 CO1 1 2 3 All coding aa 332 332 332 996 332 71 14 295 380 38 3.5 1.3 1.0 – – 9.3 1.8 48.0 19.0 3.6 12.3 2.7 49.6 20.4 6.6 11.1 2.4 48.2 20.5 6.3 13.3 2.4 45.0 19.5 7.5 12S All 324 112 1.5 11.1 18.7 19.9 19.5 Note. Cocc, Coccotrypes; Dryo, Dryocoetes; Ozop, Ozopemon; aa, amino acids a ti/tv signifies the relative proportions of transitional to transversional substitutions. b ti/tv ¼ 1.8 when averaged over MP tree. For other subsets, tree-based ratios differed marginally from the average pairwise estimates in the table. Table 4 Phylogenetic signal for 25 ingroup taxa and eventual outgroups as measured by RASA and the number of ingroup nodes supported by 50 and 90% bootstrap support Partition bobs b0 tRASA df > 50 BP > 90 BP EF-1a, ingroup 12S, ingroup CO1 nuc, ingroup CO1 aa, ingroup Morphology, ingroup All data, ingroup —rooted all outgroups —rooted Dryocoetes —rooted Ozopemon 7.46 7.77 5.73 9.26 11.04 8.95 15.32 13.04 13.99 3.41 4.82 3.70 6.58 6.83 5.29 10.48 8.70 8.39 13.26 9.68 7.36 8.91 14.55 11.82 17.54 14.16 19.48a 272 272 272 272 272 272 272 272 272 16 8 6 8 7 – 14 16 15 7 2 3 1 3 – 10 10 9 a Outgroup with the highest putative plesiomorphy content. 3.4. Combined analyses: ingroup rooting Choice of outgroup affected the placement of the ingroup root. Each of the three outgroup compositions resulted in a different root. In the unweighted MP analysis of 29 taxa the ingroup root consisted of one large basal polytomy (indicated by Bremer support of zero in Fig. 3). The topological conflict between the two most parsimonious trees may be a result of conflict among outgroups. When using only Dryocoetes species as outgroups, C. longior was the basal taxon supported by 71% bootstrap support. When only Ozopemon species were used as outgroups, a clade of seed feeders (C. cf. distinctus, C. impressus, C. carpophagus, and C. dactyliperda) was basal, supported by 58% bootstrap support, a result similar to the ML analysis with both outgroup genera present (Fig. 4). When the rapidly evolving mtDNA gene fragments were downweighted by a factor 0.5, C. longior, C. petioli, and C. marginatus made up the basal clade (all outgroups). If weighted 0.3 or less, C. longior again recovered as the sole basal taxon—a result also consistent with amino acid translation of CO1 (in combination with remaining data). Using C. longior as a functional outgroup resulted in functional ingroup rooting by the seed-feeding clade. 3.5. Combined analyses: ingroup topology Some relationships were consistent across the various outgroup configurations, including the following. The four species feeding on small seeds (the seed-feeding clade) made up the best-supported larger clade, and a fifth seed-feeding species, C. aciculatus, grouped with these in the mtDNA and morphology analyses (Fig. 3). The two petiole-feeding specialists C. petioli and C. marginatus clustered as expected. The two pith-feeding Dryocoetiops species clustered together with high support and clustered further with two, primarily petiolefeeding, species of Coccotrypes. However, a Kishino– Hasegawa (parsimony) variance test (Kishino and Hasegawa, 1989) could not reject monophyly of Coccot- 178 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Fig. 2. Separate unweighted MP analyses of the 25 ingroup taxa and 4 outgroup taxa having no missing data partitions. Bootstrap support values are shown on nodes with more than 50% support. Collecting areas are indicated by their initials for the nonmonophyletic Coccotrypes advena (see Table 1). Morphology: strict consensus of 3 trees of length 87 steps each, CI ¼ 0.57, RI ¼ 0.80. mtDNA: 1 tree of length 3040 steps, CI ¼ 0.28, RI ¼ 0.37. EF-1a: strict consensus of 17 trees of length 628 steps each, CI ¼ 0.56, RI ¼ 0.62. B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Table 5 Partitioned bremer support indices for each of the four data partitions Node EF-1a 12S CO1 Morphology All combined 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Total 34.5 2.5 11.2 0.5 15.5 0.5 6.6 7.5 8.2 15.2 0 7.5 7.8 15 )0.5 )0.5 0 6.5 13.5 )0.5 2.5 4.5 21.5 8.5 12.5 1.5 202 14.5 )2.5 1.3 1.5 1.5 1.5 0.4 0 3.8 13.5 0 )0.5 )1.5 )2.5 )1.5 )1.5 0 )2.5 )2.5 )1.5 )2 )1.8 1 )1.5 2.5 )1.5 18.2 7 9 8.7 )1 9 )1 2.6 2.2 4.1 48.3 0 )4 6.3 6.5 1 1 0 1.5 30 1 )0.5 )1.4 27 )3 )7 5 152.3 7 )1 3.8 )1 7 )1 3.4 )3.7 )2.1 )3 0 1 1.4 5 1 1 0 0.5 2 1 1 )0.3 3.5 )3 )4 )1 18.5 63 8 25 0 33 0 13 6 14 74 0 4 14 24 0 0 0 6 43 0 1 1 53 1 4 4 391 29 0.64 708 0.55 Inf. ch. 198 Tot/Inf. ch. 1.02 112 380 0.16 0.40 Note. Values are estimated from a topology based on combined data for 25 ingroup taxa and 4 outgroups, with no missing partitions (see Fig. 3). rypes, to the exclusion of Dryocoetiops. C. gedeanus grouped with C. cyperi and C. litoralis. The latter species had haplotypes very similar to those of C. cyperi (EF-1a coding region was identical, but intron differed by one 11-bp insertion) and was nested within the latter in many analyses, including some analyses of partial gene fragments, and in the ML analysis. On the contrary, and perhaps the most remarkable result of the analyses, was the consistent nonmonophyly and long branches of C. advena, a widespread resource generalist. The latter result holds for all gene partitions when analyzed separately, but not morphology. The general absence of well-supported deep nodes was not associated with incongruence among data partitions. Only when each of the mtDNA data partitions was tested separately against EF-1a was incongruence significant (P ¼ 0:03 [12S], 0.01 [CO1]). However, because the mtDNA gene partitions are linked in the mitochondrial genome, they must ultimately be combined, and as such, the combined mtDNA partition was not significantly different from any other partition (P ¼ 0:20 [EF-1a], 0.74 [morphology]). 179 3.6. Morphology and behavior Table 6 shows the relative performance of morphological and behavioral characters on the topology depicted in Fig. 3. Among those characters with maximum CI values, only characters 1, 16, 23, and 27 could potentially vary within the ingroup. The phylogenetic utility of proventricular characters (Nobuchi, 1969) is little known, but we note here that the shape of the anterior plate and the placement of marginal bristles define groupings also suggested by our current best estimate of the phylogeny (Fig. 3: node 7, char. 5; node 15, char. 1). We also note that the characters with the worst performance were all describing cuticular granules and ridges (char. 12, 14, 15, and 17). Resource use (char. 30) performed equally to the average for all characters combined. 4. Discussion 4.1. Unusual evolutionary dynamics under sib-mating Perhaps the most striking result from our study is the unusually high sequence divergence and morphological stasis in many lineages. Variation in the genus Coccotrypes in fact equals that reported for the entire Prodoxidae (Brown et al., 1994) or Silphidae (Dobler and M€ uller, 2000) for mtDNA or for the Noctuidae (Mitchell et al., 1997) for EF-1a. Not only is the observed CO1 divergence among the highest reported for a genus, but the 19% uncorrected divergence among the five C. advena haplotypes (22.3% Kimura two-parameter (k2) corrected) is the highest intraspecific CO1 variation known to us for any insect or most organisms (but see Hedin, 2001; Rocha-Olivares et al., 2001). Such high intra specific variation is not unique to C. advena, however. Although lower variation was found between the two C. fallax haplotypes (9.5%), considerable variation was found between ecologically and morphologically near-identical species, such as C. petioli and C. marginatus (18.1%) and C. dactyliperda and C. carpophagus (16.1%). Some other sib-mating scolytine genera contain similar genetic structure as revealed by unpublished CO1 haplotypes of Xylosandrus morigerus (Blandford) (16.0%) and Hypothenemus eruditus Westwood (16.7%) (B.H.J. and L.R.K., unpublished). One might argue that such high levels must reflect morphologically cryptic species, in particular because outbreeding scolytine beetle species—and even species complexes—typically encompass less than 5% sequence divergence at COI (Cognato and Sperling, 2000; Kelley and Farrell, 1998). The consistent nonmonophyly of C. advena further supports this view, suggesting that the Ugandan populations constitute a separate species. On 180 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Fig. 3. Most parsimonious tree for 25 ingroup taxa and 4 outgroup species, with EF-1a and morphology weighted 2 and mtDNA weighted 1. Length 4540 steps, CI ¼ 0.34, RI ¼ 0.46. Numbers in gray refer to node numbers (see Table 5). Values above nodes are bootstrap support values and below are Bremer support values for weighted/unweighted analyses. Bremer support of zero indicates an unresolved or a slightly conflicting node. Arrows point to the placement of taxa for which only mtDNA sequences were available, if included in the equally weighted analysis. Vertical bars show resource preferences. the other hand, none of the markers used were congruent with respect to the topological placement of the different C. advena specimens, pointing toward possible events of introgression and incomplete lineage sorting. Only further sampling of other populations and additional nuclear markers can resolve this ambiguity. However, there are reasons to expect sib-mating lineages to show more incongruence and higher levels of intraspecific DNA sequence divergence than outbreeders. Sexual selection is thought to be one of the major components driving species diversification in a variety of outbreeding organisms (Schluter, 2000), and the loss of sexual selection in sib-mating scolytines may reduce divergence rates between species. This also involves the B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 181 Fig. 4. Most likely tree (score )17489.89) found in the ML analysis of all molecular data partitions combined for 29 taxa (and no missing data partitions). Parameter settings used with the GTR + C + I model were empirical base frequencies, I ¼ 0.496, and C ¼ 0:487 with four rate categories and substitution frequencies A $ G 11.46, A $ C 4.85, A $ T 8.92, T $ G 1.00, C $ T 66.90, and C $ G 1.93. Parameters were approached through iterations of parameter estimates, starting with the topology depicted in Fig. 3. To the right are drawings from the internal side of one of the eight proventricular plates from the following species, top to bottom: Dryocoetes affaber, Ozopemon uniseriatus, Coccotrypes cf. distinctus, Dryocoetiops coffeae, and Coccotrypes gedeanus. Vertical bars indicate similarity in proventricular characters 1 and 5 (see Table 6, Appendix A). loss of specific mate recognition characters which define most outbreeding species. Indeed, the characters most commonly used to distinguish outbreeding bark beetle species are secondary sexual characters (Jordal, 1998) which are presumably important for exchange of tactile cues during mate selection (Page and Willis, 1982). The slowdown in morphological evolution in species of Coccotrypes and other sib-mating beetles is therefore largely expected. Still, morphological differences could develop between shallowly diverged sib-mated lineages because, under zero outcrossing, the fixation of mutations in any lineage is virtually instantaneous. This may 182 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Table 6 The relative fit of morphological and behavioral characters as measured by the consensus index (CI) over tree topology in Fig. 3 Proventriculus 1 1.00 2 0.5 Head 3 0.33 4 0.33 5 0.75 0.58 7 1.0 8 1.0 9 1.0 10 0.67 21 1.0 22 0.5 23 1.0 24 1.0 11 0.5 12 0.17 13 0.4 14 0.25 25 0.6 26 0.5 27 1.0 28 0.5 0.67 Pronotum 15 0.29 6 1.0 Elytra 16 1.0 17 0.25 0.61 18 1.0 19 0.5 20 0.4 Legs 0.73 0.72 Behaviour and development 29 1.0 Average morphology: 0.63 0.89 30 0.67 31 1.0 Average all: 0.66 Note. Indices vary marginally between alternative tree topologies. Numbers in boldface are average CI for each body section and for behavior. See Appendix A for description of characters 1–31. explain the distinct morphological differences between the genetically close C. cyperi and C. litoralis. Another well-studied ecological mate recognition character is pheromone production, where outbreeding species find conspecific mates by sorting diverse pheromone plumes during dispersal. Pheromones are not known from sibmating species, as expected, because females mate with their brother before dispersal. Consequently, the huge genetic divergences that we observe in some sib-mating lineages rarely represent ecologically segregated cryptic species. While close relatives in Coccotrypes sometimes demonstrate distinct ecological segregation (see below), the overwhelming impression is one of ecological niche overlap in sympatry. It is not unusual to find closely related species breeding in the same unit, for instance the petiole specialists C. petioli and C. marginatus or the seed specialists C. dactyliperda and C. carpophagus (B.H.J. and L.R.K., pers. obs.). We also expect to find the same pattern in species of Dryocoetiops where all use pith of tiny twigs or of petioles and in species of Ozopemon where all breed under bark of large logs. The observation of a marked increase in niche overlap for many sib-mating sister species contrasts with the sometimes clear evidence for allopatric distribution or sympatric resource division between closely related species in general (Barraclough and Vogler, 2000) and for outbreeding bark beetles in particular (Cognato and Sperling, 2000; Jordal, 1998; Kelley and Farrell, 1998). If the overall outcrossing rate for a population is low, and if hybridization is physically possible, the relaxed selection on mate recognition characters may also permit more frequent matings between these ecologically similar and sympatric ‘‘species.’’ Judged by field observations (B.H.J., L.R.K.), broods from multiple colonizations sometimes coalesce, providing opportunities for outcrossing. The genetic outcome of this process could sometimes mimic the principles of artificial breeding programs in agriculture, in which a cross between inbred ‘‘pure strains’’ creates hybrids, and then the new hybrid strain is inbred until it ‘‘breeds true,’’ i.e., until it is homozygous. A long history of such low-frequency reticulation events may be the cause of the extreme haplotype diversity and incongruence between morphological characters and gene trees for some Coccotrypes lineages. We also note that hybridization in nature of genetically highly diverged individuals seems only possible for subclonally reproducing lineages. For instance hybridizing Daphnia species in the galeata complex are as much as 19% diverged (K2 corrected) in their CO1 sequences (Schwenk et al., 2000), consistent with our view that hybridization may also be possible between the different lineages of C. advena and related species. We are currently working on assessing more closely the intrapopulation genetic variation within and between sites, but it is clear that variation is considerable for most sib-mating lineages studied (B.H.J. and L.R.K., unpublished). Whatever is causing the extreme haplotype variation in Coccotrypes, is also resulting in certain drawbacks of phylogeny reconstruction as discussed below. 4.2. Gene utility and phylogeny reconstruction The most rapidly evolving mitochondrial genes showed high levels of substitutional saturation and homoplasy between the most divergent lineages. Although high substitution rates rarely contribute mere noise in a data matrix (Wenzel and Siddall, 1999; Yang, 1998) and often provide the majority of phylogenetic signal in terms of node support (Allard et al., 1999; Baker et al., 2001; Bj€ orklund, 1999; K€allersj€ o et al., 1999), the negative effect of noise from mtDNA was in B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 this study demonstrated by the reduced number of bootstrap-supported ð> 50%Þ nodes (16 vs l2) when mtDNA nucleotides were added to the combined data matrix. By comparison, using amino-acid-translated CO1 data, or downweighting mtDNA by a factor of two or more, we increased the number of supported ingroup nodes to 15 or more. Taken together with the many bizarre relationships found in the mtDNA phylogeny, the lack of support and negative contribution to the total-evidence tree suggest that the mitochondrial genes are far too rapidly evolving to be useful for phylogeny reconstruction above species-group level in Coccotrypes. A commonly used method to accommodate poor resolution is downweighting rapidly evolving sites— third positions or transitions—to elevate signal from more conservative sites (Swofford et al., 1996). However, combined analysis of first and second positions yielded only slightly better results than third positions alone for our CO1 data, and transversions performed no better than transitions for all mtDNA data. Despite the appeal of this type of differential weighting, studies showing that such character weighting does not increase node support or decrease incongruence (Baker and DeSalle, 1997; Baker et al., 2001; Broughton et al., 2000; Durando et al., 2000; Fu, 2000; Garin et al., 1999), and that it performs even worse when second positions are especially favoured (Xia, 1998), are now accumulating. The sometimes limited utility of mitochondrial genes is in contrast to the usefulness of certain rapidly evolving nuclear genes (e.g., Campbell et al., 2000). Nuclear genes usually outperform mtDNA genes in combined analyses (Baker and DeSalle, 1997; Baker et al., 2001; Durando et al., 2000; Remsen and DeSalle, 1998), especially at deeper splits (Springer et al., 1999), and the value of nuclear protein-coding genes to the resolution of deeper nodes of insect genera cannot be overstated (Durando et al., 2000). Also in this study, the nuclear gene contributed most of the phylogenetic signal within a combined data matrix which otherwise consisted of highly diverged mitochondrial sequences. Thus, higher weights could be given to the best-performing gene to also recover the deepest splits (see Reed and Sperling, 1999). With higher weights applied to EF-1a and morphology, topologies converged toward the one in Fig. 3 or the one with C. longior as a single basal taxon when mtDNA data were excluded altogether. As noted previously (Jordal et al., 2000, 2002), Dryocoetiops appear to be nested within Coccotrypes. Morphological differences between the two genera are small and not consistent, in particular the separation of the procoxae and shape of the pronotum (Wood, 1986; but see Appendices A and B). Although more nuclear markers and species of Dryocoetiops must be 183 sampled before reaching a final conclusion, synonymy of this genus seems likely when such data become available. The two preferred topologies (Figs. 3 and 4) are also similar to our previous study where C. longior, C. marginatus, and the seed-feeding clade were among the most basal taxa (Jordal et al., 2000). However, we could not confidently place the basal node of the ingroup due to a presumably strong conflict among the outgroups. Because there is no consistent rule on how to root phylogenetic trees, and because general recommendations differ among authors (Lyons-Weiler et al., 1998; Smith, 1994; Watrous and Wheeler, 1981), it is unclear which rooting should be preferred. Hence we prefer to err on the side of inclusion by using both Dryocoetes and Ozopemon as outgroups. The two ingroup roots found in the MP and ML analyses using both outgroup genera (Figs. 3 and 4) are not very different, and both alternatives place the seed-feeding clade close to the base in each topology. Furthermore, the topology depicted in Fig. 3 is identical to the Coccotrypes topology reconstructed as a part of an analysis of 80 dryocoetine and xyleborine EF-1a sequences (B.H. Jordal, unpublished), demonstrating topological stability with increased sampling. On the other hand, we note in the ML analysis that the resulting basal position of the seedfeeding clade is similar to the result obtained with Ozopemon as the sole outgroup in the MP analysis. Despite these ambiguities, the combined data still provided evidence for certain evolutionary patterns in resource use that we discuss below. 4.3. Implications for resource-use evolution The expansion of niches and extreme resource generalism in sib-mated lineages is a paradox, given the presumed loss of genetic variation due to inbreeding. However, we have demonstrated huge amounts of haplotype diversity in populations, hinting at a broad array of genotypic variation conducive to ecological adaptability and evolutionary change. The resolution of the genetic variation paradox may lie in the ability of sib-mating Coccotrypes to regain ‘‘lost’’ genetic variation through occasional outcrossing/hybridization events. During the long periods of inbreeding that separate such events, any beneficial mutations enabling use of a particular resource will be immediately fixed in the inbred lineages on that resource. If that resource is locally abundant such lineages may proliferate, and when outcrossing/hybridization events do occur, such alleles may be captured by other inbred lineages in the same or different species. Selection for ecological adaptability in opportunistic generalists such as Coccotrypes may act to promote distant outcrossing and to eliminate any vestiges of a specific mate recognition system. 184 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 The clear preference for small resources set some sibmating groups apart from most outbreeding bark and timber beetles (Kirkendall, 1993). Smaller resources such as seeds and petioles are scattered in the forest and even more ephemeral than the logs and branches used by other scolytines (Beaver, 1979b). Because larger, longer-lived resources attract more beetle individuals over a longer time period than do smaller resource units, the probability of finding a mate for an outbreeding individual decreases with decreasing resource size. Small resources such as petioles (Beaver, 1979b; Jordal and Kirkendall, 1998) and palm seeds (Kirkendall, 1993) have usually one or a few conspecific colonizing females (inbreeders) or colonizing pairs (outbreeders) per resource unit. Because colonizing females of inbreeders carry with them sperm from their natal nest, these breed independently of colonization densities. This is the most obvious reason that outbreeders are generally less frequent in small resources. Another important factor is the near twofold advantage of sib-maters’ female-biased offspring sex ratios, allowing demographic viability in less productive resource units (Beaver, 1979a; Jordal and Kirkendall, 1998). Some of the highly unusual resources used by Coccotrypes are seemingly evolutionary novelties with little or no subsequent diversification. Perhaps the most unusual such novelty is bracken fern (Gray, 1970, 1972), with three associated species of Coccotrypes. We were not able to obtain DNA from the single monophagous species in ferns, C. pteridophytae (Schedl), but judged by morphological characters, this species seems closely related to C. cardamomi, a generalist species with preference for seeds (Browne, 1961) and occasionally found in bracken fern (Gray, 1972). A third species found in ferns (Asplenium), C. confusus (Eggers), is most probably the closest relative of two other generalists, C. medius and C. cylindricus (Eggers) (see Browne, 1973). Based on the widely separated phylogenetic placement of C. medius and C. cardamomi, it seems likely then that fern feeding evolved at least twice in Coccotrypes. The seeds and radicles of viviparous mangroves also have a limited number of associated beetles (Beeson, 1961; Browne, 1961; Woodruff, 1970) and three species of Coccotrypes breed exclusively in such tissues. In at least two of these (C. litoralis and C. rhizophorae [Hopkins]), the mangrove association has evolved from the use of fruits and seeds as part of their more generalized diet. With the lack of EF-1a data, we were not able to assess the placement of C. fallax in this study. However, judged by proventricular and other morphological characters in conjunction with mitochondrial data, C. fallax is not associated with the ‘‘generalist’’ clade (node 15 in Fig. 3). Hence, two or three independent origins of mangrove seed feeding occurred. Because there is no example of speciation within each mangrove- feeding lineage, and because close relatives of C. litoralis and C. fallax have diverged little in genome and morphology, the association with mangroves appear to be recent in each case. Seeds of various size and form are clearly the most productive resources in terms of brood size (Kirkendall, 1993) and the number of species capable of breeding in this type of resource (Browne, 1961). Despite its productivity, seed feeding has only evolved (in Scolytinae) in two inbreeding clades (Coccotrypes and Hypothenemus) and in a few outbreeding species. Nonetheless, more than half of the species in Coccotrypes show slight or strong preferences for seeds, although only a minor fraction are specialists on this resource. These specialists feed on small seeds (primarily palm seeds) and have followed the recent expansion of palms throughout the tropics. Although we have sampled only a minor fraction of putative seed specialists, a majority of the species seems to constitute a monophyletic group (cf. Figs. 3 and 4) and as such demonstrates evolutionary consistency of this habitat. Given the high-energy endosperm provided by large seeds (e.g., dipterocarp seeds) and the frequent but facultative use of larger seeds by species of Coccotrypes, why is there apparently so little specialization upon this resource? The most important sources of large seeds are those of dipterocarp trees growing in southeast Asia, the area where Coccotrypes originated during the Miocene (Jordal et al., 2000). Since dipterocarps exhibit multispecies gregarious flowering and mast fruiting (Ashton et al., 1988), dipterocarp seeds are not available every year, and it is not possible for multivoltine insects such as Coccotrypes to specialize on them. They can be utilized only opportunistically, by resource generalists. To specialize, a species is constrained to use predictably available resources—for instance, in the case of Coccotrypes, palm seeds, mangrove radicles, Moraceae petioles, or bracken ferns. When dipterocarp seeds are artificially made permanently available, as in ilipe nut (oil-seed) plantations, many different Coccotrypes species converge to destroy much of the crop (Browne, 1973), suggesting preference for this type of resource when available. Regardless of the ambiguity associated with our phylogenetic hypothesis, all reconstructions indicate an evolutionary trend from specialization upon one kind of host tissue toward a more opportunistic behavior using many different tissue types (Figs. 3 and 4). And other generalist species, not sampled for DNA—C. cinnamomi (Eggers), C. nitidus (Eggers), and C. papuanus (Eggers) (Browne, 1961)—have morphological features indicating a clear relationship to the generalist-dominated sub clade (node 15). Our study thus adds to the increasing bulk of evidence that specialization does not necessarily evolve from more generalized lifestyles (Thompson, 1994). B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Acknowledgments We especially thank Roger A. Beaver for identification of specimens and James Lyons-Weiler for help with RASA analyses and permission to use a prerelease of RASA 3.0 Turbo. We also greatly appreciate the additional specimens collected by Akira Ueda, Hideaki Goto, Andrea Sequeira, Geoffrey Morse, and Swee Peck Quek. B.H.J. was funded by a Norwegian 185 Research Council Grant (123588/410), BDF by an USDA National Research Initiative grant (97-353024226), and LRK by a Norwegian Research Council SUP grant (128388/420). Field work was supported by the Putnam Expedition Fund of the Museum of Comparative Zoology, and we thank members of the parataxonomist training centre in Madang, Papua New Guinea for helpful assistance with collecting and logistics. Appendix A. Morphological and behavioral characters coded for Coccotrypes, Dryocoetiops, and outgroups Proventriculus 1. Anterior margin of anterior plate: (0) straight or slightly curved; (1) angular. 2. Anterior plate with tubercles: (0) absent; (1) present. 3. Median suture: (0) oval; (1) linear or diverging. 4. Median denticles: (0) pointed, sharp; (1) small, rounded; (2) absent. 5. Marginal bristles: (0) tuft of few; (1) longitudinal rows of > 10; (2) longitudinal rows of widely spaced, soft bristles; (3) absent. Head 6. 7. 8. 9. 10. 11. 12. 13. 14. Labial palpus segment 1: (0) shorter than 2 + 3 combined; (1) longer than 2 + 3 combined, barrel shaped. Labial palpus, segment 3: (0) longer than segment 2, narrow; (1) subequal to segment 2, broad. Maxillary fringe with setae: (0) few and coarse; (1) many, medium coarse; (2) numerous fine and soft. Maxillary palpus segment 3: (0) shorter than 1 + 2 combined; (1) longer than 1 + 2 combined. Antennal club: (0) more than basal 1/3 corneous; (1) less than basal 1/4 corneous; (2) hairy to base. Setose area of anterior face of club with transverse suture: (0) visible (1) not visible. Epistoma with granules: (0) present; (1) absent. Aciculation in frons: (0) absent; (1) narrowly, from below level of eyes; (2) broadly, from vertex. Median, longitudinal carina in frons: (0) absent; (1) present. Pronotum 15. 16. 17. 18. 19. Pronotal asperities on anterior half: (0) absent; (1) low or granulate; (2) sharply elevated. Row of asperities along apical margin: (0) absent; (1) present. Asperities on basal disc: (0) present; (1) laterally only; (2) absent. Pronotal disc: (0) shining; (1) reticulate. Dorsum: (0) flat; (1) lightly curved; (2) dome-shaped in middle; (3) dome-shaped posteriorly. Elytra 20. 21. 22. Interstrial setae: (0) thin; (1) stout; (2) spatulate. Strial punctures (0) in rows; (1) confused at least on declivity. Strial setae: (0) absent; (1) tiny, recumbent; (2) erect, at least half length of interstrial setae. Legs 24. 25. 26. 27. 28. Probibiae with: (0) 4 socketed teeth, (1) 5 teeth; (2) 7 teeth. Mesotibiae with: (0) 4 socketed teeth; (1) 5 teeth; (2) 6 teeth; (3) 7 teeth; (4) 8 teeth. Metatibiae with: (0) 4 socketed teeth; (1) 5 teeth; (2) 6 teeth; (3) 7 teeth; (4) 9 teeth. Procoxae: (0) contiguous, prosternal process pointed; (1) separated, prosternal process broad/blunt. Mesocoxae separated: (0) approximately width of scapus; (1) twice or more the width of scapus. Males 29. Male development: (0) similar to females: (1) smaller, with adult characters (2) larviform. Behavior 30. 31. Feeding on: (0) phloem; (1) pith; (2) leafstalks; (3) small, hard seeds; (4) fruits; (5) mangrove radicles. Mating system: (0) outbreeding; (1) inbreeding. 186 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Appendix B. Data matrix of the 31 morphological characters described in Appendix A Dryocoetes affaber Dr. autographus Ozopemon brownei O. uniseriatus Coccotrypes aciculatus C. advena (B) C. advena (CR) C. advena (J) C. advena (U) C. advena (U) C. cardamomi C. carpophagus C. cf. carpophagus C. cf. rhizophorae C. cyperi (CR) C. cyperi (Qld) C. dactyliperda (Arg.) C. dactyliperda (FL) C. cf. distinctus C. falax (Bangl) C. fallax (Qld) C. gedeanus C. graniceps C. impressus C. litroralis C. longior C. marginatus C. medius C. petioli C. variabilis Dryocoetiops coffeae D. cf. eugeniae 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 1 0 0 3 0 0 1 1 2 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 1 0 1 1 1 0 0 1 1 2 1 0 1 1 0 0 0 0 0 0 1 0 1 1 0 1 0 1 1 0 0 0 1 1 2 0 0 1 1 2 0 0 1 1 2 0 1 0 0 0 0 1 0 0 0 1 {01} 1 {01} 0 1 0 1 1 0 0 0 1 1 2 1 0 1 1 0 0 0 0 0 0 0 0 1 1 3 0 0 1 0 0 0 0 1 1 3 1 0 1 1 1 0 1 1 0 0 0 1 1 0 0 00000 00000 10111 10111 11210 11211 11211 11211 11211 11211 11211 11210 11211 11211 11211 11211 11210 11210 11210 11211 11211 11211 11211 11210 11211 11211 11212 11211 11212 11211 11211 11211 References Allard, M.W., Farris, J.S., Carpenter, J.M., 1999. Congruence among mammalian mitochondrial genes. Cladistics 15, 75–84. Ashton, P.S., Givnish, T.J., Appanah, S., 1988. Staggered flowering in the Dipterocarpaceae: New insights into floral induction and the evolution of mast fruiting in the aseasonal tropics. Am. Nat. 132, 44–66. Baker, R.H., DeSalle, R., 1997. Multiple sources of character information and the phylogeny of Hawaiian drosophilids. Syst. Biol. 46, 645–673. Baker, R.H., Wilkinson, G.S., DeSalle, R., 2001. Phylogenetic utility of different types of molecular data used to infer evolutionary relationships among stalk-eyed flies (Diopsidae). Syst. Biol. 50, 87– 105. Baker, R.H., Yu, X., DeSalle, R., 1998. Assessing the relative contribution of molecular and morphological characters in simultaneous analysis trees. Mol. Phylogent. Evol. 9, 427–436. Barraclough, T.G., Vogler, A.P., 2000. Detecting the geographical pattern of speciation from species-level phylogenies. Am. Nat. 155, 419–434. Beaver, R.A., 1979a. Leafstalks as a habitat for bark beetles (Col.: Scolytidae). Z. Angew. Entomol. 88, 296–306. 100 100 010 010 112 110 111 111 111 111 101 112 111 112 111 111 112 112 112 100 100 101 101 112 111 010 010 111 010 101 100 100 0 0 1 1 0 0 0 0 0 0 {01} 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 1 2 1 2 2 2 0 0 0 0 0 1 2 1 1 2 2 2 2 2 2 2 2 1 2 2 2 0 1 0 1 2 2 0100 0100 0003 0003 1002 0200 0201 0201 0201 0201 0100 1002 0101 0100 0101 0101 1002 1002 1002 0011 0011 0101 0001 1002 0101 0110 0000 0000 0000 0000 0002 0002 0 0 0 0 1 2 2 2 2 2 0 0 1 0 1 1 0 0 0 0 0 0 0 0 0 {12} 0 0 0 2 0 0 01001 01024 12001 12002 02000 00001 00001 00001 00001 00001 00001 02000 01001 01001 00001 00001 02000 02000 02000 00001 00001 00001 01001 02001 00001 00001 02101 01011 02101 00001 01002 01002 2010 5010 1002 2002 0011 1011 1011 1011 1011 1011 1011 0011 101? 111? 1011 1011 0011 0011 0011 2111 2111 1011 1001 101? 1011 1011 1011 111? 1011 1011 211? 211? 0 0 0 0 3 {024} {024} {024} {024} {024} {34} 3 3 2 {01234} {01234} 3 3 3 5 5 {0124} 3 3 5 {02} 2 {012} 2 {034} 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Beaver, R.A., 1979b. Non-equilibrium ‘island’ communities: A guild of tropical bark beetles. J. Anim. Ecol. 48, 987–1002. Beeson, C.F.C., 1961. The Ecology and Control of the Forest Insects of India and the Neighbouring Countries, 2nd ed. Government of India, New Delhi. Bj€ orklund, M., 1999. Are third positions really that bad? A test using vertebrate Cytochrome b. Cladistics 15, 191–197. Blumberg, D., Kehat, M., 1982. Biological studies of the date stone beetle, Coccotrypes dactyliperda. Phytoparasitica 10, 73–78. Bright, D.E., Peck, S.B., 1998. Scolytidae from the Galapagos Islands, Equador, with descriptions of four new species, new distribution records, and a key to species. Koleopt. Rund. 68, 233–252. Bright, D.E., Skidmore, R.E., 1997. A Catalog of Scolytidae and Platypodidae (Coleoptera), Supplement 1 (1990–1994). NRC Research Press, Ottawa. Broughton, R.E., Stanley, S.E., Durrett, R.T., 2000. Quantification of homoplasy for nucleotide transitions and transversions and a reexamination of assumptions in weighted phylogenetic analysis. Syst. Biol. 49, 617–627. Brown, J.M., Pellmyr, O., Thompson, J.M., Harrison, R.G., 1994. Mitochondrial DNA phylogeny of the Prodoxidae (Lepidoptera: Incurvarioidea) indicates rapid ecological diversification of yucca moths. Ann. Entomol. Soc. Am. 87, 795–802. B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Browne, F.G., 1959. Notes on two Malayan scolytid bark-beetles. Malay. For. 22, 292–300. Browne, F.G., 1961. The biology of Malayan Scolytidae and Platypodidae. Malay. For. Rec. 22, 1–255. Browne, F.G., 1973. The African species of Poecilips Schaufuss. Rev. Zool. Bot. Afr. 87, 679–696. Campbell, D.L., Brower, A.V.Z., Pierce, N.E., 2000. Molecular evolution of the wingless gene and its implications for the phylogenetic placement of the butterfly family Riodinidae (Lepidoptera: Papilionoidea). Mol. Biol. Evol. 17, 684–696. Cognato, A.I., Sperling, F.A.H., 2000. Phylogeny of Ips species (Coleoptera: Scolytidae) inferred from mitochondrial cytochrome oxidase I sequence. Mol. Phylogenet. Evol. 14, 445–460. Danforth, B.N., Ji, S., 1998. Elongation Factor-1a occurs as two copies in bees: Implications for phylogenetic analysis of EF-1a sequences in insects. Mol. Biol. Evol. 15, 225–235. Dobler, S., M€ uller, J.K., 2000. Resolving phylogeny at the family level by mitochondrial Cytochrome Oxidase sequences: Phylogeny of carrion beetles (Coleoptera, Silphidae). Mol. Phylogenet. Evol. 15, 390–402. Durando, C.M., Baker, R.H., Etges, W.J., Heed, W.B., Wasserman, M., DeSalle, R., 2000. Phylogenetic analysis of the repleta group of the genus Drosophila using multiple sources of characters. Mol. Phylogenet. Evol. 16, 296–307. Entwistle, P.F., 1964. Inbreeding and arrhenotoky in the ambrosia beetle Xyleborus compactus (Eichh.) (Coleoptera: Scolytidae). Proc. R. Soc. Lond. A 39, 83–88. Farrell, B.D., Sequeira, A., O’Meara, B., Normark, B.B., Chung, J., Jordal, B., 2001. The evolution of agriculture in beetles (Curclionidae: Scolytinae and Platypodinae). Evolution 10, 2011–2027. Farris, J.S., K€ allersj€ o, M., Kluge, A.G., Bult, C., 1995a. Constructing a significance test for incongruence. Syst. Biol. 44, 570–572. allersj€ o, M., Kluge, A.G., Bult, C., 1995b. Testing Farris, J.S., K€ significance of incongruence. Cladistics 10, 315–319. Felsenstein, J., 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39, 783–791. Fu, J., 2000. Towards the phylogeny of the family Lacertidae—Why 4708 base pairs of mtDNA sequences cannot draw the picture. Biol. J. Linn. Soc. 71, 203–217. Garin, C.F., Juan, C., Petitpierre, E., 1999. Mitochondrial DNA phylogeny and the evolution of host-plant use in Palearctic Chrysolina (Coleoptera, Chrysomelidae) leaf beetles. J. Mol. Evol. 48, 435–444. Gray, B., 1970. Observations on the first scolytid, Poecilips pteridophytae (Coleoptera: Scolytidae), to be found in association with a fern (Pteridophyta). Can. Entomol. 102, 578–585. Gray, B., 1972. Observations on Poecilips cardamomi (Schaufuss), the second species of Scolytidae to be found in bracken fern (Col.). Entomol. Tidskr. 93, 229–237. Hamilton, W.D., 1993. Inbreeding in Egypt and in this book: A childish perspective. In: Thornhill, N.W. (Ed.), The Natural History of Inbreeding and Outbreeding: Theoretical and Empirical Perspectives. University of Chicago Press, Chicago, pp. 429– 450. Hedin, M.C., 2001. Molecular insights into species phylogeny, biogeography, and morphological stasis in the ancient spider genus Hypochilus (Araneae: Hypochilidae). Mol. Phylogenet. Evol. 18, 238–251. Herfs, A., 1950. Studien an dem steinnussborkenk€afer, Coccotrypes tanganus Eggers. Hofch. Wiss. Praxis. 3, 3–31. Jordal, B.H., 1998. A review of Scolytodes Ferrari (Coleoptera: Scolytidae) associated with Cecropia (Cecropiaceae) in the northern Neotropics. J. Nat. Hist. 32, 31–84. Jordal, B.H., 2001. The origin and radiation of sib-mating haplodiploid beetles (Coleoptera, Curculionidae, Scolytinae). Ph.D. Thesis, Department of Zoology, University of Bergen, Bergen. 187 Jordal, B.H., Beaver, R.A., Kirkendall, L.R., 2001. Breaking taboos in the tropics: Inbreeding promotes colonization by wood-boring beetles. Glob. Ecol. Biogeogr. 10, 345–358. Jordal, B.H., Beaver, R.A., Normark, B.B., Farrell, B.D., 2002. Extraordinary sex ratios and the evolution of male neoteny in sibmating Ozopemon beetles. Biol. J. Linn. Soc. 75, 353–360. Jordal, B.H., Kirkendall, L.R., 1998. Ecological relationships of a guild of tropical beetles breeding in Cecropia leafstalks in Costa Rica. J. Trop. Ecol. 14, 153–176. Jordal, B.H., Normark, B.B., Farrell, B.D., 2000. Evolutionary radiation of an inbreeding haplodiploid beetle lineage (Curculionidae, Scolytinae). Biol. J. Linn. Soc. 71, 483–499. K€allersj€ o, M., Albert, V.A., Farris, J.S., 1999. Homoplasy increases phylogenetic structure. Cladistics 15, 91–93. Kelley, S.T., Farrell, B.D., 1998. Is specialization a dead end? The phylogeny of host use in Dendroctonus bark beetles (Scolytidae). Evolution 52, 1731–1743. Kirkendall, L.R., 1993. Ecology and evolution of biased sex ratios in bark and ambrosia beetles. In: Wrensch, D.L., Ebbert, M.A. (Eds.), Evolution and Diversity of Sex Ratio: Insects and Mites. Chapman & Hall, New York, pp. 235–345. Kishino, H., Hasegawa, M., 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. J. Mol. Evol. 29, 170–179. Kluge, A.G., 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Syst. Zool. 38, 7–25. Kluge, A., 1998. Total evidence or taxonomic congruence: Cladistics or consensus classification. Cladistics 14, 151–158. Lekander, B., 1968. Scandinavian bark beetle larvae; descriptions and classification. Rep. No. 4, Department of Forest Zoology, Royal College of Forestry, Stockholm. Lyons-Weiler, J., 1998. RASA for the Mac. Available from http:// loco.biology.unr.edu/archives/rasa/rasa.html. Lyons-Weiler, J., Hoelzer, G.A., 1997. Escaping from the Felsenstein zone by detecting long branches in phylogenetic data. Mol. Phylogenet. Evol. 8, 375–384. Lyons-Weiler, J., Hoelzer, G.A., Tausch, R.J., 1996. Relative Apparent Synapomorphy Analysis (RASA) I: The statistical measurement of phylogenetic signal. Mol. Biol. Evol. 13, 749–757. Lyons-Weiler, J., Hoelzer, G.A., Tausch, R.J., 1998. Optimal outgroup analysis. Biol. J. Linn. Soc. 64, 493–511. Mitchell, A., Cho, S., Regier, J.C., Mitter, C., Poole, R.W., Matthews, M., 1997. Phylogenetic utility of Elongation Factor-1a in Noctuidea (Insecta: Lepidoptera): The limits of synonymous substitution. Mol. Biol. Evol. 14, 381–390. Nobuchi, A., 1969. A comparative morphological study of the proventriculus in the adult of the subfamily Scolytoidea (Coleoptera). Bull. Government For. Exp. Sta. 224, 39–110. Normark, B.B., Jordal, B.H., Farrell, B.D., 1999. Origin of a haplodiploid beetle lineage. Proc. R. Soc. Lond. B 266, 2253– 2259. Page, R.E., Willis, M.A., 1982. Sexual dimorphism in ventral abdominal setae in Scolytus multistriatus (Coleoptera: Scolytidae): Possible role in courtship behaviour. Ann. Entomol. Soc. Am. 76, 78–82. Pruess, K.P., Adams, B.J., Parsons, T.J., Zhu, X., Powers, T.O., 2000. Utility of the mitochondrial Cytochrome Oxidase II for resolving relationships among black flies (Diptera: Simuliidae). Mol. Phylogenet. Evol. 16, 286–295. Reed, R.D., Sperling, F.A.H., 1999. Interaction of process partitions in phylogenetic analysis: An example from the Swallowtail butterfly genus Papilio. Mol. Biol. Evol. 16, 286–297. Remsen, J., DeSalle, R., 1998. Character congruence of multiple data partitions and the origin of the Hawaiian Drosophilidae. Mol. Phylogenet. Evol. 9, 225–235. 188 B.H. Jordal et al. / Molecular Phylogenetics and Evolution 23 (2002) 171–188 Rocha-Olivares, A., Fleeger, J.W., Foltz, D.W., 2001. Decoupling of molecular and morphological evolution in deep lineages of a meiobenthic harpacticoid copepod. Mol. Biol. Evol. 16, 1088–1102. Schluter, D., 2000. The Ecology of Adaptive Radiation. Oxford University Press, Oxford. Schwenk, K., Posada, D., Hebert, P.D.N., 2000. Molecular systematics of European Hyalodapnia: The role of contemporary hybridization in ancient species. Proc. R. Soc. Lond. B 267, 1833–1842. Simmons, M.P., 2000. A fundamental problem with amino-acidsequence characters for phylogenetic analyses. Cladistics 16, 274–282. Smith, A.B., 1994. Rooting molecular trees: Problems and strategies. Biol. J. Linn. Soc. 51, 279–292. Sorenson, M.D., 1999. TreeRot.v2. Boston University, Boston. Springer, M.S., Amrine, H.M., Burk, A., Stanhope, M.J., 1999. Additional support for Afroteria and Paenungulata, the performance of mitochondrial versus nuclear genes, and the impact of data partitions with heterogeneous base composition. Syst. Biol. 48, 65–75. Swofford, D., 1999. PAUP*. Phylogenetic Analysis Using Parsimony (*and other methods) version 4. Sinauer, Sunderland, MA. Swofford, D.L., Olsen, G.J., Waddell, P.J., Hillis, D.M., 1996. Phylogenetic inference. In: Hillis, D.M., Moritz, C., Mable, B.K. (Eds.), Molecular Systematics. Sinauer, Sunderland, MA. Takenouchi, Y., Takagi, K., 1967. A chromosome study of two parthenogenetic scolytid beetles. Ann. Zool. Jpn. 40, 105–110. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The ClustalX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 24, 4876–4882. Thompson, J.N., 1994. The Coevolutionary Process. Chicago University Press, Chicago. Ueda, A., 1997. Brood development of an inbreeding spermatophagous scolytid beetle Coccotrypes graniceps (Eichhoff) (Coleoptera: Scolytidae). Jpn. J. Entomol. 65, 677–687. Watrous, L.E., Wheeler, Q.D., 1981. The outgroup comparison method of character analyses. Syst. Zool. 30, 1–11. Wenzel, J.W., Siddall, M.E., 1999. Noise. Cladistics 15, 51–64. Wheeler, W.C., Gatesy, J., DeSalle, R., 1995. Elision: A method for accomodating multiple molecular sequence alignments with alignment-ambiguous sites. Mol. Phylogenet. Evol. 4, 1–9. Wood, S.L., 1973. New synonymy in American bark beetles (Scolytidae: Coleoptera). Gt. Basin Nat. 33, 169–188. Wood, S.L., 1982. The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Gt. Basin Nat. Mem. 6, 1–1359. Wood, S.L., 1986. A reclassification of the genera of Scolytidae (Coleoptera). Gt Basin Nat. Mem. 10, 1–126. Wood, S.L., Bright, D., 1992. A catalog of Scolytidae and Platypodidae (Coleoptera) Part 2: Taxonomic index. Gt. Basin Nat. Mem. 13, 1–1553. Woodruff, R.E., 1970. A mangrove borer, Poecilips rhozophorae (Hopkins) (Coleoptera: Scolytidae). Florida Dept. Agric. Entomol. Circ. 98, 1–2. Xia, X., 1998. The rate heterogeneity of non-synonymous substitutions in mammalian mitochondrial genes. Mol. Biol. Evol. 15, 336– 344. Yang, Z., 1998. On the best evolutionary rates for phylogenetic analysis. Syst. Biol. 47, 125–133.