ARCHAEOLOGY ON KAPINGAMARANGI ATOLL University of Otago Studies in Prehistoric Anthropology.Volume 16 ARCHAEOLOGY ON KAPINGAMARANGI ATOLL A POLYNESIAN OUTLIER IN THE EASTERN CAROLINE ISLANDS BY Foss Leach & Graeme Ward University of Otago, Studies in Prehistoric Anthropology, Volume 16 © 1981 Department of Anthropology, University of Otago Copyright in all countries signatory to the Berne Convention (Cover design: Canoe shed3 Werua isletj i Kapingamarangi) Earlier volumes in the series Studies in Prehistoric Anthropology are listed below, and may be obtained from: The Secretary, Department of Anthropology, University of Otago, PO Box 56, Dunedin, New Zealand. Those out of print are marked with an asterisk(*). Prices quoted are in NZ$. Volume 1* B.F.Leach.1969. The Concept Prehistoric Studies. of Volume 2* H.M.Leach.1969. Subsistence New Zealand. Patterns Volume 3* J.I.Kennedy.1969.Settlement Islands, 1772. in Volume 4* D.T.Bayard.1971.Non Nok Tha. Volume 5* P.J.F.Coutts & M.Jurisich.1972. An Archaeological New Zealand. Survey of Ruapuke Island, Volume 6* M.Pietrusewsky. 1974.Non Nok Tha, remains from the 1966 excavations N.E.Thailand. Volume 7 C.F.W.Higham. 1975. Non Nok Tha: Price: $3.50 Volume 8* F.F.W.Higham, G.M.Mason, & S.J.E.Moore. 1976. Upper Clutha Valley. Volume 9 D.T.Bayard. 1976. The Cultural Relationships Polynesian Outliers. Price: $3.50 Volume 10 P.J.Wilson. 1976. At our Volume 11 A.Fox. 1978. Tiromoana Pa, Te Awanga, Excavations 1974-75. Price: $4.50 Volume 12 P.Bellwood. 1 9 7 8 . A r c h a e o l o g i c a l Research at Lake Mangakaware, Waikato, 1968-70. Price: $4.50 Volume 13 D.T.Bayard. 1980. The Pa Mong Archaeological 1973-75. Price: $5.00 Volume 14 C.F.W.Higham & Brian Vincent. 1980. Gabriels an Archaeological Survey. Price: $4.00 Volume 15 Amphan Kijngam, Charles Higham, & Vvarrachai Wiriyaromp. 1980. Prehistoric Settlement Patterns in Northeast Thailand. Price: $8.00 Volume 16 Foss Leach and Graeme Ward. 1981. Archaeology on Raping amarangi Atoll - a Polynesian Outlier in the Eastern Caroline Islands. Price: $ 7.50 ii Wit's Similarity the in in South The 1968 Prehistoric East Excavation. the Human at Non Nok the Bay of Faunal Skeletal Tha, Remains, of the Beginning.Price: Hawkes $2.50 Bay Survey, Gully - CONTENTS ACKNOWLEDGEMENTS LIST OF FIGURES LIST OF TABLES INTRODUCTION HISTORICAL SKETCH ENVIRONMENTAL SETTING General Comments Climate and Ocean Currents Pre-European Vegetation Land Fauna and Avifauna Marine Resources Lithic Resources Driftwood Freshwater ARCHAEOLOGICAL SITES ON KAPINGAMARANGI The Site Survey Culturally Modified Soils Earthworks and Culturally redeposited Coral Block Masonry Traditional Sites Historic Sites Summary THE EXCAVATIONS Scope and Methods Adopted The Excavation at Tiroki(KAl) The Excavation at Muri-Harau (KA2) iii Page The Excavation a t P u t a u (KA3) 42 The E x c a v a t i o n a t N g e i h o - H e r e u (KA4) RADIOCARBON DATING A N D T H E CHRONOLOGY OF SETTLEMENT Level I: T r a n s i e n t Occupation (71000 - 700 B P ) Level II:Formation of A r t i f i c i a l Islet(700-300BP) Level III:Intensive Residential Level IV:Modern H o r i z o n 46 50 53 53 Occupation(300-lOOBP)55 (100BP - P r e s e n t ) 55 THE A N A L Y S I S OF M A T E R I A L S RECOVERED Methods A d o p t e d 57 Fish 57 Shellfish 61 Crustaceans 63 Birds 64 Pig, Rat, Turtle and Human 65 Stone and Pumice 66 HUMAN SKELETAL MATERIAL 66 67 Shell Adzes 67 One-Piece Fish hooks 71 Trolling Lure shanks 73 Coconut Grater Heads 74 Pottery 75 Shell Arm Rings 76 Perforated Shells and Shell Beads 78 Drilled Sharks teeth 78 Sea Urchin Spine Files 79 Pumice Abraders 80 Worked Pearl Shell 80 IV Page Japanese and European Artefacts 81 KAPINGAMARANGI AND OTHER POLYNESIAN OUTLIERS Linguistics 85 Drift Voyaging 86 Comparison with Nukuoro 88 CONCLUSIONS Land History of the Atoll 91 Origin of the Founding Population 93 The Prehistoric Fishermen of Kapingamarangi 97 APPENDICES 1: D.T.Bayard. Marine Species in Kapingamarangi orthography, with some proto-Polynesian equivalents 101 2: Graeme Ward. Mollusc survey on Kapingamarangi 107 3: Foss Leach and Janet Davidson. The Analysis of fish bone from Kapingamarangi 114 4: Graeme Mason. Kapingamarangi Crustacean Remains 123 5: Simon Holdaway. Analysis of Archaeological shell from Kapingamarangi 125 6: W.R.Dickinson. Sand Temper in a sherd from Kapingamarangi 134 7: C.A.Landis. Sand Temper Petrography of a sherd and some Japanese tiles from Kapingamarangi 135 8: Michiko Intoh and Foss Leach. Physical Analysis 136 of Kapingamarangi Pottery 9: Graeme Mason. Kapingamarangi Rocks 138 10:Phil Houghton. The Human Biology of Kapingamarangi 14 2 145 v ACKNOWLEDGEMENTS The Kapingamarangi research was financed by the Historic Preservation Office of the Trust Territory of the Pacific, the University of Otago, and the Australian Institute of Aboriginal Studies. This support is gratefully acknowledged. We would like to thank Scott Russell and Ross Cordy of the Historic Preservation Office in Saipan for their assistance in a number of ways in organising the expedition, and also Narsi Kostka and the Historic Preservation Committee in Ponape. We are especially grateful to the Catholic Mission in Ponape, and in particular to Father Cavanagh and Father Zewen for their generous hospitality and interest in the work. Thanks are also due to the Chief Magistrate and Council on Kapingamarangi for their enthusiastic support of the research, and to the people who assisted us on the island. Mike Lieber helped us with a number of organisational matters, and this assistance was very welcome. The laboratory analysis of material recovered from Kapingamarangi was assisted in various ways by Suzanne Moody; specialist help was obtained from Ron Scarlett, Walter Cernohorsky and Graeme Mason. Hank Jansen and Charlie McGill provided much needed help with radiocarbon dating, and the illustrations in this monograph are the work of Sandy McElrea, Martin Fisher, and Richard Newell. To all these people, we express our gratitude. Janet Davidson provided a great deal of help and advice on many matters, and we are very grateful to have been able to draw on her experience with archaeological material from Polynesian outliers. This manuscript benefited from her critical eye. Finally, we would like to thank Koro Monop and his wife Temoe, our adopted parents on Kapingamarangi, who made our visit to their island not only comfortable but memorable. vi LIST OF FIGURES Page 1 Location Map of Kapingamarangi 5 2 Kapingamarangi Atoll, showing Islets 6 3 Kapingamarangi Wind Conditions through the year 7 4 The Islet of Pumatahati from a distance 16 5 Coral stone facing to taro swamp on Werua islet 17 6 Coral slab walls on Werua islet 17 7 Stone lining to fresh water well on Werua islet 19 8 The stone of Utamatua on Touhou islet 19 9 Stone fish trap between Tetau and Nikuhatu islets 20 10 Land building sea wall on Pepeio islet 21 11 Land building sea wall on Herekoro islet 21 12 Vertical Photograph of Touhou islet, 1942 23 13 A Japanese radio ship off Hare islet, 1942 24 14 Oblique aerial photograph of concrete structure on Nunakita islet, 1943 25 15 The same building in 1980 25 16 Concrete house piles at Japanese installation on Nunakita islet 26 17 Aerial photograph of Japanese weather station on Nunakita islet, 1943 26 18 Concrete ablution structures at Japanese installation on Nunakita islet 27 19 Propellors and engines from Japanese sea planes on Hare islet 27 20 Map of northern half of Touhou islet showing location of modern houses and excavation squares 30 21 Square layout and Section conventions 32 22 Vertical photograph of Werua islet in 1942 34 23 General view of excavation in progress at Tiroki 35 vii Page 24 The stratigraphic sequence at Tiroki(KAl) 35 25 Section diagram of the excavation at Tiroki(KAl) 36 26 General view of the excavation at Muri-Karau(KA2) 39 27 Section diagram of the excavation at Muri-Harau(KA2) 40 28 General view of the excavation at Putau(KA3) 43 29 The excavation at Putau(KA3) nearing completion 43 30 Section diagram of the excavation at Putau(KA3) 44 31 General view of the excavation at Ngeiho-Hereu(KA4) 47 32 The stratigraphic sequence at Ngeiho-Hereu(KA4) 47 S e c t i o n d i a g r a m of t h e e x c a v a t i o n at 48 33 Ngeiho-Hereu(KA4) 34 Profile of Touhou islet, showing the four excavations 50 35 Correlation between the four sites excavated 52 36 The evolution of Touhou islet over the last 700 years 54 37 Tridacna shell adzes from Kapingamarangi 68 38 Surface collected Tridacna shell adzes 69 39 One-piece shell, bone and metal fish hooks from Kapingamarangi 71 40 Shell trolling lure shanks from Kapingamarangi 73 41 Shell coconut grater heads from Kapingamarangi 74 42 Metal and ceramic artefacts from Kapingamarangi 76 43 Shell arm rings from Kapingamarangi 77 44 Files, shell beads and industrial pearl shell from Kapingamarangi 79 45 A selection of European artefacts from the excavation 81 at Muri-Harau(KA2) 46 Glass beads from the excavation at Muri-Harau(KA2) 80 47 European artefacts recovered at Muri-Harau(KA2) 82 viii from the excavation Page 48 Ceramic artefacts from the Japanese settlement on Nunakita and Parakahi 83 49 The southern tip of Hare in 1942 91 LIST OF TABLES 1 Historical chronology of Kapingamarangi 4 2 Pre-European Vegetation on Kapingamarangi 9 3 Avifauna on Kapingamarangi 11 4 Radiocarbon Results for Kapingamarangi 51 5 Minimum numbers of fish organised into families and chronological periods 58 6 Fish remains arranged in order of decreasing abundance 59 7 Shellfish minimum numbers organised into families and chronological periods 62 8 Bird Minimum numbers from the excavations 64 9 Minimum numbers of pig,turtle, rat and other mammals from the excavations 65 10 C o m p a r i s o n of a d z e forms from N u k u o r o and Kapingamarangi 69 11 Trends in the ratio of parrotfish to cod and groper in Kapingamarangi 100 12 Basic results of fishbone analysis, Tiroki(KAl), Minimum numbers 116 13 Basic results of fishbone analysis, Muri-Harau(KA2), 118 Minimum numbers 14 Basic results of fishbone analysis,Putau(KA3), Minimum numbers 119 Basic results of fishbone analysis,Ngeiho-Hereu (KA4), Minimum numbers 121 Crustacean remains from Kapingamarangi 124 . 15 16 ix Page 17 The shellfish species identified from Kapingamarangi arranged in taxonomic order 12 7 18 Minimum numbers of shellfish from Tiroki(KAl) 130 19 Minimum numbers of shellfish from Muri-Harau(KA2) 131 20 Minimum numbers of shellfish from Putau(KA3) 132 21 Minimum numbers of shellfish from Ngeiho-Hereu(KA4) 133 22 Physical analysis of pottery 136 23 Average stature figures for Kapingamarangi and various other Oceanic groups 14 3 24 Blood group analysis of Kapingamarangi 144 -1INTRODUCTION In 1978 the Chief Magistrate of Kapingamarangi wrote to the Historic Preservation Office of the Trust Territory of the Pacific drawing its attention to the fact that recent erosion on Touhou islet had uncovered artefacts of kinds not known to the present people, and stratified up to 4m below the present ground surface. He requested that a suitable expert should be sent to Kapingamarangi to investigate these finds, and carry out a general programme of research into the ancient history of the atoll, and make recommendations as to any future work which should be done. The Historic Preservation Office contracted the senior author of this report to organise an expedition to the atoll for these purposes. Short notice prevented a trip to the area before the southern University vacation of 1979-80. A social anthropologist-Mike Lieber - has been working on Kapingamarangi periodically for a number of years, and his knowledge of the language and customs was seen as potentially useful to the proposed archaeological research; consequently, the Historic Prservation Office contracted him independently to carry out research into traditional land use, and generally assist the archaeologist with organisational matters. Kapingamarangi is an isolated atoll, 750 km from the port of Ponape, and shipping schedules are far from reliable. However, a ship normally reaches the atoll once a month. The research team was available from the period 16 November 1979 to February 16 in the hope that at least two full months could be spent on the island. As it turned out, mechanical problems with shipping resulted in delays which meant that only one month of useful work could be carried out on the island. Consequently, the excavation programme had to be substantially limited in scope. In addition, Mike Lieber had arranged to overlap his leave from University of Chicago with the visit by the archaeologists. However, he was forced to depart on the same ship which they arrived on. These logistic problems were unfortunate, but other than the time restriction did not adversely affect the research programme. Following the fieldwork, three crates of archaeological residues were shipped to Otago University for analysis. These took seven months to arrive in New Zealand, and two were badly damaged by forklifts somewhere in transit. Unfortunately some material was lost, including a modern stone collection, and some of the most important bones, not to mention items of equipment and personal belongings. Lack of liability on the part of shipping companies for such damage and loss is a little known hazard which must be added to the list of difficulties involved in archaeology in remote Pacific islands. An added precaution was taken in the case of the radiocarbon samples, which were air freighted to New Zealand at some expense. These went missing at Nauru airport, and were the cause of much anxiety until they were ultimately found. Over the ensuing 12 months, the materials were analysed in the Otago Archaeological Laboratories. -2HISTORICAL SKETCH Kapingamarangi has had a long history of irregular contact with Europeans which is thought to have begun with Fernao de Grijalvares in 1536. Emory, who has carried out detailed analysis of historical information(1965:12ff) considers that despite many sightings and possible stopovers, significant contact with Europeans did not really begin until the last two decades of the 19th century. The fleeting nature of these early contacts is reflected in the numerous alternative names and spellings for the island which were cited on various maps and publications until relatively recently(see Baker,1951:14; Emory, 1965:12-13;Schlaginhaufen,1929:219). Some of these are listed below: Kapingamarangi Kapinmarangi Baki ramarang Kabeneylon Kapenmailang Makarama Pikiram Pigiram Kapinga Malany Pyghirap Pyghiram Puguirame Pikiraf Dos Pescadores Tenuv Constantine Greenwich Guru inichi The first estimate of population size of about 150 people was made in 1883(Eilers, 1934:6), but it is likely that the pre-European population was somewhat larger than this. Emory considers the climax population to have been about 450(1965:66). About 1870, some castaways from the Marshall islands arrived on Kapingamarangi, killed a large number of people, and dominated the islanders until 1872 when the schooner Matautu took them back to their homeland(Majuro and Ebon). An upper limit for a traditional craft on such a journey would be about 3 0 people, although there are reports of grossly overloaded Marshallese outriggers with 40 to 50 aboard(Lewis, 1972:273). The number of canoes which arrived on Kapingamarangi ranges in the telling from three to eight, suggesting the arrival of between 90 and 240 people. It is hard to credit how such a number could subjugate a population at least twice this size on their home territory. This incident, along with traditional knowledge of Kapingamarangi on Lukunor(in the Mortlock islands) and Nukuoro to the north (Emory, 1965:52), and Ontong Java(Luangiua) to the south (Schlaginhaufen, 1929:219), suggests that despite the isolated position of this atoll, some contact with other areas might have occurred from time to time in the prehistoric era. The cultural significance of any such contact is a matter for archaeological research to decide. From 1536 to 18 99, Kapingamarangi was nominally a Spanish territory, although their influence on the island -3appears to have been minimal. The Germans purchased the Caroline Islands from the Spanish in 18 9 9(presumably without consulting the people on Kapingamarangi), and this began a period of administration from Rabaul which lasted until the First World War. During this time, the Japanese took over Micronesia, and a weather station and seaplane base vere established on the island during the Second World War. The Japanese base was bombed in July 194 3. American administration of the island under the aegis of the Trust Territory of the Pacific Islands began in 1945, lasting until 1979. Kapingamarangi is now part of the newly formed Federated States of Micronesia, and is administered from Ponape to the north. There have been a number of scientific surveys of Kapingamarangi over the years, beginning with the German South Seas Expedition of 1910, although Schlaginhaufen made a brief visit there in 1908. Hambruch, who was writing up the material from the Ponape district, died before this was completed, and the Kapingamarangi section was eventually written up by Eilers(1934). An interesting economic survey was carried out in 1946(Gantt, 1946; Fosberg, 1946); and Elbert visited the island in the same year and collected linguistic and folk-lore material(Elbert, 1948;1949).Two major anthropological expeditions to the island were organised by the Bishop Museum in 1947 and 1950, and the results were published in three detailed bulletins(Buck, 1950; Miller, 1953; Emory, 1965). In 1954, a further expedition took place, organised by the Pacific Science Board, and this resulted in three further monographs describing the geography, geology, and terrestrial bioecology of the island(Niering, 1956;McKee,1956;Wiens,1956). Finally, from 1965 onwards, a social anthropologist has carried out a great deal of fieldwork on the island, and on its satellite village on Ponape,Porakiet(Lieber, 1968a;1968b;1970;1974;1977a;1977b; Lieber and Dikepa, 1974) . The combined effect of these researches has resulted in an impressive body of literature on the island, its resources, and its people. Until the Otago archaeology expedition, however, nothing concrete was known of the prehistoric occupation of the island. The various programmes of scientific research on Kapingamarangi are listed in Table 1. -4TABLE 1: HISTORICAL CHRONOLOGY OF KAPINGAMARANGI FEDERATED STATES OF MICRONESIA 1979 TRUST TERRITORY OF PACIFIC ISLANDS (American Admin) 1979-80 Otago University Archaeological Expedition.Leach & Ward.4 0 days 1965- Social Anthropological research by Lieber 1954 Pacific Science Board Expedition Wiens,Niering & McKee. 8 weeks 1950 Bishop Museum Expedition,II. Emory & Miller. 5 months 1947 Bishop Museum Expedition I. Buck,Elbert,Lanthrop & Emory. 3 months 1946 Visit by Elbert 1946 Economic Survey.Gantt & Fosberg 3 days 1943 Island bombed by US 1910 German South Seas Expedition Hambruch,Kramer, etc. 6 days 1908 Visit by Schlaginhaufen.3 days 1899 Germans purchased Carolines 1870 Marshallese incident 1536 Discovery by Fernao de Grijalvares 1945 JAPANESE PERIOD 1914 GERMAN PERIOD 1899 SPANISH PERIOD 1536 -5ENVIRONMENTAL SETTING General Comments Kapingamarangi is one of the most isolated areas of land in the Pacific Ocean, and notoriously difficult to navigate to. The only useful guideline for prehistoric navigators is the 'bird arc' of 7.5°. Commenting on a traditional voyage from Pulusak(in the Western Carolines) to Kapingamarangi, a distance of some 74 4 km, Lewis observed: "It is tempting to surmise that, for such a difficult passage(across variable currents) and devoid of any additional reef or other safety 'screen', this 7.5° expanded landfall arc might represent something like the limit of navigational feasibility"(Lewis,1972:231). Modern navigators to Kapingamarangi experience the same difficulty in making a predictable landfall. Wiens(1956:21) recounts how during their expedition to the island in 1954, the Captain of 10 3 9* 1 o cD o IE o« '\JOroluk (^Truk QPakin AntoOPonape Western Caroline atolls o Mortlock Is, £> Ngatik ^ ^ oPingelap O Kusaie 5• ° North o Nukuoro o kapingamarangi -Equator ' CZZ? Admiralty Is, New Ireland - oTvkuu - QNukumanU • 5° SouthQ^pOntong Java Bougainville New Guinea Figure 1: Location Map of Kapingamarangi. Some greater circle distances from the atoll are: Nukuoro=215 km, Nuguria=511 km, Rabaul=651 km, Ponape=748 km. Ocean currents in the vicinity of the atoll are thought to be generally westward in character - experience shows them to be highly erratic. -6- t h e i r steamship s e t course from Nukuoro for a point 29 km to the east of Kapingamarangi to compensate for the p r e v a i l i n g westward d r i f t , but as a r e s u l t of a change in current d r i f t to an e a s t e r l y d i r e c t i o n , they found themselves in e r r o r by some 51 km to the e a s t ! This i s not an i s o l a t e d example, and a t t e s t s the i s o l a t i o n of t h i s t i n y a t o l l from the r e s t of the Pacific world(See Figure 1 ) . Kapingamarangi i s about 100 km to the north of the equator, and i s considerably c l o s e r to the major landmasses of Papua New Guinea to the south than i t i s to i t s modern a d m i n i s t r a t i v e centre of Ponape to the n o r t h . For t h i s reason, the Germans considered i t appropriate during t h e i r administration to include the i s l a n d in the d i s t r i c t of the Bismarck Archipelago, r a t h e r than with the Eastern Caroline p a r t of t h e i r colony (Schlaginhaufen,1929:220) . _ The a t o l l i s roughly egg shaped, with a maximum dimension of about 12 km. The t o t a l area i s j u s t over 82 km2, of which | tf V, £• a / / / > 7 Jr y ^^ jr ® £i o 46 c ^\\ ~ X^.* *Xv/ J v J&. » a o m d 79 79 0 73 46 f ^ s^ *X X-^ J Pungupungu ft 73 }Wi^* 46 16 16 Ma tiro fStJl'prtgt Y^J^'Poe /&] Tetau IS X 37 I ^ P j Takairorigo rt o I 16 H J %m Cs 37 1 Spot soundings metre 5 North S * ue X N J6°33'ii9Bii 1"5 68 0 tf^s,^* 37 2& 69 <3 \ ^I Fft \| Tout,ou a V^ i1 A o n 0 9 " .^dr^' » * \ ^" 5 0 • 1 ttfen 1 1 F i g u r e 2: Kapingamarangi a t o l l showing 34 i s l e t s on t h e c u r r e n t Hydrographic C h a r t . Only 32 of t h e s e a r e now s e p a r a t e d by s e a . The southernmost land d i s t r i c t on Hare was a s e p a r a t e i s l e t u n t i l r e c e n t l y , and both Herengaua and Herekoro a r e now j o i n e d onto Hare(See a l s o F i g u r e 4 9 ) . Matukerekere was p r a c t i c a l l y l e v e l l e d t o t h e sea i n a s e v e r e storm i n 1858(see Wiens,1956:19), b u t i s growing a g a i n . The anchor symbol shows t h e l o c a t i o n of an anchorage n e a r Hare where a J a p a n e s e schooner was sunk i n 1943(See a l s o F i g u r e 1 3 ) . -7the reef and land ac count for a mere 2 0.4 km . The lagoon reaches a depth of 8 0m in places. By comparison, the nearby atoll of Nukuoro to the north is rather smaller in size(28.5 k m 2 ) , but with a slightly deeper lagoon(reaching 108m). Dry land on Kapingamarangi total s about 112 hectares, and the 32 islets (See Figure 2) range in size from 2.1 km in length(Hare) to only 3 0m long (Matukerekere). More than half the population of about 3 00 live on Touhou islet, which is about 300 x 150m in size, covering 3.72 hectar es. Apart from about 20 people, the remainder live on Werua islet, immediately to the north. Climate and Ocean Currents Precise information about the climate of this atoll is not available, but several reconstructions have been made from patchy observations over many years, and the comments here rely on Wiens(1956:18-20,85ff) and Anon.(1944:3-4).The weather on Kapingamarangi is controlled largely by the seasonal movement of the doldrums belt, which lies immediately to the west and south(Thomas,1965:30). In May and June the belt moves northward across this area, and is accompanied by frequent heavy showers and generally stormy weather. While the doldrums lie in the north from July through September, generally fine weather prevails. The doldrums move south again in October and November, causing squally conditions interspersed with good weather. From December through April the Northeast Trades predominate, with good weather except for occasional tropical storms. Wind frequency through the year is shown in Figure 3, and this illustrates the influence of the Northeast Trade winds from about November to April; the remainder of the year is characterised by variable winds. D -CZ3 T\ D Wind speed knots Wind frequency percent /J 40 60% %)—c calm 1 I 17 >27 Figure 3: Kapingamarangi wind conditions through the year(based on Anon.,1944: 4). A=December to February, B=March to May, C=June to August, and D=September to November. -8Rainfall on Kapingamarangi is highly -erratic from year to year, largely because storms in the region are small in diameter, and the number which actually pass over the island varies. Average rainfall values ranging from 198 to 274 cm have been published(Wiens,1956:18; Anon.1972:10). Temperature varies from about 24° to 34°C, with a mean annual average of about 28°C. Relative humidity varies from 57 to 90%, with an average of about 78%. Kapingamarangi lies in a region of the Pacific close to the boundary of the south equatorial current(west to northwest direction in this region), and the equatorial counter current (easterly direction). Experiences by the Masters of various ships over the years testify to very strong currents running past the atoll - some claim these to be westward, others eastward(Wiens,1956:21). It is a notoriously difficult place to make a landfall at, which suggests an erratic character in the ocean currents. The Pilot Charts for the region give a predominantly westward current of 16 to 64 km per day(Anon.n.d.). Waves and swell come from the northeast from December through April, and largely from the southeast for the remainder of the year. Although hurricanes(typhoons) never occur on Kapingamarangi, violent storms and tidal waves do. In addition, severe droughts occur periodically. Pre-European Vegetation Trying to reconstruct in any detail the vegetation pattern before the arrival of Europeans is very difficult. This has been attempted by Niering (1956), Wiens(1956), and Emory(1965), but it has to be admitted there there could be some important errors of omission and commission. A summary of the position is given in Table 2, from which it will be seen that the flora was very restricted. Eighteen species of trees(of which only about 10 are very common), four shrubs, and 18 herbs are thought to have been available to the prehistoric inhabitants. Of particular importance are those plants which may have been brought by the founding population, or by later visitors to Kapingamarangi after first settlement. Of these, particular mention must be made of the taro. There are three species on the island today - Cyvtosperma ohamissonis, the small swamp taro, known as puraka, which is believed to have been introduced from Nukuoro in Spanish times; Alocasia maovorrhiza, the large dry land taro, again introduced recently along with the name(ngaungau) from Nukuoro; and finally, Colooasia esculenta, the small common dry land taro, known as tava, which is thought to have been present in the prehistoric period(Emory,1965:7-8). However, Thilenius believed differently, claiming that taro was absent(Thilenius,1903:17). This issue could not be finally resolved without positive evidence, such as from pollen spectra. The fact that the name for taro is tava, rather than the more normal taro (by which it is known on Nukuoro), suggests that it was present before contact with other communities in Micronesia, such as people on Nukuoro. -9TABLE 2: PRE-EUROPEAN VEGETATION ON KAPINGAMARANGI This is largely derived from Emory(1965:8-9).Those marked with an asterisk(*) are thought by Niering(1956) to have been introduced by prehistoric Polynesians; the remainder can be thought of as indigenous. The percentage frequencies are taken from Niering(1956:Figures 28,29), and are therefore based on modern vegetation density. Those marked thus(t) are in the Niering list, but not mentioned by Emory. Trees 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 % Frequency Cooos nuoifeva Guettarda speciosa Morinda oitrifolia Premna obtusifolia Pandanus tectorius Barringtonia asiatioa Calophyllum inophyllum Hibiscus tiliaaeus Cordia subcordata Avtocarpus altilis Messerschmidia argentea Terminalia samoensis Pisonia grandis Thespesia populnea Pandanus dubius Ochrosia oppositifolia Pongamia pinnata Soulamea amara 100 97 97 97 97 63 51 48 45 42 39 27 15 12 6 6 3 3 * t t 97 36 27 6 t * * * * * t * Shrubs 1 2 3 4 Soaevola frutescens Allophylus timorensis Pipturus argenteus Sophora tomentosa Herbs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Asplenium nidus Lepturus repens Stenotaphrum mioranthrum Thuarea involuta Cassytha filifovmis Triumfetta prooumbens Fleurya ruderalis Ipomoea pes-aapvae Ipomoea littoralis Nephrolepis hirsutula Wedelia biflora Digitaria miorobachne Euphorbia chamissonis Ipomoea tuba Jussiaea suffrutioosa (Ludwigia octovalvis) Pteris tripartita Colooasia esculenta Psilotum nudum 54 36 36 33 18 18 12 12 12 9 9 6 6 6 6 6 3 3 t t t* t t* t * -10Thus, we should think of the prehistoric cultivation of taro on Kapingamarangi as making a striking contrast to the extensive wet taro(puraka) cultivation of today, which covers about 10 hectares(9.1% of the total land area,see Wiens,1956:36). Whether the supposed prehistoric Coloaasia esculenta was cultivated in swamp pits similar to those of today, or as an essentially dry land form of horticulture, with perhaps some kind of irrigation assistance, is not very easy to answer. Breadfruit(Artocarpus altilis) has the standard Polynesian name(kuru), rather than a term borrowed from Micronesia, and this strongly suggests that it was either on the island when the founding group reached Kapingamarangi, or it was brought by the earliest settlers. It will be noted in Table 2 that two herbs are thought by Niering to have been introduced by man in the prehistoric period, but are not considered by Emory. Digitaria miorobaohne is a grass which colonises open fern country, and Ludwigia oatovalvis is a fairly common cultivation weed, especially in Fiji and Samoa, and also common in swampy places(Parham,1972). Why Niering believes them to have been introduced before the arrival of Europeans is not clear. Other plants of some economic significance today which (arrowroot), have been introduced recently are Taooa pinnatifida Musa spp(banana), Carica papaya (papaya), and Curcuma domestica (turmeric). This latter introduction is especially interesting, as it came from Nukuoro, where the plant possesses ritual importance as it does elsewhere in Polynesia. Land Fauna and Avifauna Amongst the land animals of Kapingamarangi, those of potential economic importance to prehistoric people would have been land crabs such as Cardisoma rotundum and Coenobita spp., and the coconut crab Birgus latro. Both skinks and geckos occur on the island, but are unlikely to have been of any real value. The Polynesian r a t ( E a t t u s exulans) occurs, and there are a few pigs and one dog on the island today. The question of when these animals were first introduced will be discussed later. Only a few birds occur on and about the island, but some of these may have had prehistoric economic value. A list is given in Table 3 of the main birds which are known on the atoll. Of these, Aplonis opaous is very common, and Gygis alba subsp., and Anous spp. moderately so. It is interesting that large flocks of Fregata minor subsp., as well as of Sula sula rubripes are on several of the islands. Emory notes that of the common birds, only Sula sula rubripes was ever eaten at all. Observations during our expedition were to the contrary, and the meat of the frigate bird, and others, was prized. -11TABLE 3: AVIFAUNA ON KAPINGAMARANGI Aplonis opacus Fregata minor subsp. Sula sula rubvipes Pluvialis dominica fulva Anolls stolidus pileatus Anous tenuirostris marcusi Gygis alba subsp. Phaethon lepturus dovotheae Puffinus Iherminieri dichrous Numenius phaeopus variegatus Crocethia alba Sterna fuscata oahuensis Thalasseus bergii pelecanoides Sula leuoogaster plotus Callooalia inqwieta inquieta Eudynamis taitensis Gallus gallus Demigretta sacra sacra Arenaria interpres interpres Micronesian starling Frigate bird Red-footed booby Pacific golden plover Noddy tern White-capped noddy tern White tern White tailed tropic bird Dusky shearwater Whimbrel Sanderling Sooty tern Crested tern Brown booby Carolines swiftlet Long-tailed NZ Cuckoo Domestic chicken Reef heron Turnstone Marine Resources The sea life around Kapingamarangi has unfortunately not been studied in any detail by marine biologists. It represented a major source of food for the prehistoric inhabitants, and some background comments should therefore be made. In order to identify the archaeological materials it was necessary to make a comparative collection during the expedition, and although the preparation of this should not be thought of as a systematic survey, some familiarity was gained with marine life on the atoll. Over 100 species of fish belonging to 37 families were caught and prepared for osteological study. This work clearly showed the lagoon to be abundant in a wide variety of fish, and it is hardly necessary to go outside the lagoon in order to catch them. The overall size of this resource makes it very unlikely that it could have been depleted to any significant degree by sustained exploitation by an atoll population, even with the level of skills possessed by the Kapingamarangi fishermen. Some species of fish, however, could dwindle in supply, particularly those which are readily taken in the shallows between islets - this includes members of the families Muraenidae(moray eels), Diodontidae(porcupine fishes) and Tetradontidae(puffer fishes). These factors are of importance in assessing the significance of any trends in the nature of prehistoric fish catches. It is interesting that lagoon fish have apparently learned to keep at a greater distance from divers with spears than previously; however, a good fisherman finds no difficulty in bringing home a good catch after an expedition on a fine day. The most reliable and productive method of fishing today is certainly diving with locally made spear guns. This is made possible by the recent appearance of -12underwater goggles; and although some diving certainly took place before their appearance, this method of fishing is bound to have been far less productive in the past. A list of the species caught is given in Appendix 1, along with their local names. From this it will be seen that the most abundant species are members of the cod and groper family (Epiniphelidae), followed by unicornfish(Acanthuridae), parrotfish(Scaridae), sea perches(Lutjanidae), and various kinds of trevally(Carangidae). The most sought after fish, however, are the pelagic game fishes, and in particular the tuna(Thunnidae and Katsuwonidae), and to a lesser extent the wahoo(Aoanthoeybium solandri) . An especially important fish in chants is the rainbow runner(Elegatis bipinnulatus), which is also a surface feeding fish, but a member of the Carangidae family. It is a common European belief that all these fish can only be caught in very deep water outside atoll reefs - this is simply untrue. Naturally, they only occur some distance off land of high islands, but they regularly enter lagoons of atolls, and can be readily caught there. Fishing expeditions outside the reef take place for more complex sociological reasons than supposed access to wealthier fishing grounds. Sharks are very common around the reef, both inside and out. The people do not fear them, but swim freely amongst them, and even play games with them. Although sharks are caught, they are seldom eaten today, and one informant suggested their use as fertiliser for taro beds after cooking them. There are two species of crayfish at least, and these are reasonably common, but not highly thought of as food. Marine crabs are also numerous. Two species of marine turtle occur, but these are fairly rare and small. The only browsing grounds for Zostera sp. sea grass are off Hare and Ringutoru, and these are only thin and patchy. This also accounts for the general rarity of the long toms(Belonidae). Cetaceans are occasionally stranded on the reef, probably mostly killer whales{Orcinus oroa), and these are butchered for oil and meat. Shellfish resources are not as rich as they are on many other islands in the Pacific, including atolls(See Appendix 2, and Table 17). The most significant species is undoubtedly Tridaana maxima, and specimens are quite abundant, especially immediately to the seaward of the reef platform. Bippopus hippopus is the only other species of clam on Kapingamarangi, and although this is less abundant, enormous specimens occur in deep water in the lagoon. The main location of normal sized specimens is in the sandy areas between islets, and in coral thickets in fairly shallow waters of the lagoon near land. Of the two species, only Tridaana maxima is now considered worthy of eating. Specimens of Turbo spp., especially Turbo argyrostomus, are common in some areas of the reef platform. There are of course many other species of shellfish, but only a few can be considered as of any real economic significance. Of particular importance is the extreme rarity of specimens spp.), or of black mussel(Atrina vexillum) of pearl shell {Pinotada in the Kapingamarangi lagoon. -13Buck comments on these as follows: "The pearl oyster is scarce and the black mussel{woro) can be obtained only in the main passage through the reef. As they were sought for fishhooks, the flesh is probably eaten"(Buck,1950:49). Despite many days of searching for specimens by expert divers all over the lagoon during our expedition, only one specimen of each was located, and then only after the offer of a substantial monetary reward. The pearl shell was located in only moderately deep water on the southwest side of the lagoon, and the black mussel was found in very shallow water off Hare. There is a possibility that the black mussel was a recent plant from another island(perhaps relating to the modern handicraft industry), as it was found in a very odd ecological habitat for this species. We can be quite certain that these two types of shell do not constitute a significant resource, either for food or for industrial material. The ecological reasons for this dearth are not so easy to detect. Pearl shell is moderately common on Nukuoro, an atoll which is fairly similar in general structure. Of potential interest is the fact that specimens of Spondylus varians are very common in the Kapingamarangi lagoon. These are known locally as kainap, and many are close to world record size(290 mm being not uncommon ) . This shellfish, being a sedentary creature in adulthood, would be a competitor with Pinotada spp. as far as habitat is concerned, though not necessarily for food supply(Cernohorsky, 1981: pers.comm.) Another potentially useful shellfish elsewhere in the Pacific is Troehus nilotiaus, but no specimens of this were recovered, either in the modern collections or archaeologically. Further information on shellfish occurrence on Kapingamarangi is given in Appendix 2 and Table 17. Lithic Resources Essentially all of the rock forming the islets on the atoll is composed of calcareous skeletal debris, and rock fragments cemented by calcium carbonate. It varies in texture and degree of cementation; some being extremely friable, some well cemented and dense(see McKee,1956). However, surprisingly enough, other types of rocks do occur on Kapingamarangi beaches. It is unlikely that these fragments belong to the rocky basement supporting the coral atoll itself. Instead, it is believed that these arrive primarily in the roots of floating trees, and in other driftwood from elsewhere. Samples of rock from driftwood trees were collected during the expedition, but unfortunately all but two of the samples were lost in transit. These are described in Appendix 9(specimens 79.64 and 79.294). One of these rocks had an estimated weight of at least 150 kg (specimen 79.64), although normally such rocks are more like 2 kg in size. It is quite certain that significant quantities of rocky material have been arriving on Kapingamarangi from high islands elsewhere in the Pacific for as long as the island has been above the sea. These rocks are of considerable economic -14importance as they are naturally favoured over limestone blocks for cooking stones in ovans. In prehistoric times, these fragments could also have been very useful for industrial activities. The relative abundance of such rock was a considerable surprise, and points to the need to be very careful about leaping to conclusions of prehistoric trade and exchange based on evidence of the foreign origin of rocks on an island. There is a flourishing exchange network operating today, whereby rocks for cooking ovens are imported in sacks on almost every ship which visits the island. It would be interesting to study the details of this network. Large flat rocks, used as anvils for making Pandanus sp. mats, are a feature of nearly every household on Kapingamarangi, and these appear to have been imported fairly recently, perhaps from Ponape. In the historic era, stones have also arrived as ballast in ships now wrecked on the island. For example, the well known Japanese ship sunk off Hare(see Figure 13) contains large ballast boulders, some of which are now in use on the island. Finally, pumice occurs naturally on the beaches, and this has uses as abrasion tools. Driftwood Very large logs and trees are regularly washed up, and on atolls where large trees are either rare or absent, these can be of great economic importance. Samples of a number of logs were taken during the expedition, but these have yet to be identified. Freshwater As mentioned above, Kapingamarangi has a highly variable rainfall, and the amount of freshwater available is therefore of critical importance. During the well known drought of 1916-1918, about 90 people died(Emory,1965:20; Wiens,1956:20). This represented a large proportion of the population. Most of the islets possess a lens of fresh or brackish water sitting on top of the sea water which occurs in the base of the loose sediments making up each islet. Wells are dug to reach this water, and pits are dug to it for cultivating the recently introduced swamp taro (Cyrtosperma ohamissonis) or puraka. This water would normally be about 30cm above sea level, but can reach 120 cm at times(see McKee,1956). -15ARCHAEOLOGICAL SITES ON KAPINGAMARANGI The Site Survey The site survey was carried out with the aid of an outrigger canoe with outboard motor over several days in the company of a local informant who was knowledgeable about land boundaries and island lore. Land boundary maps were available from the report prepared by Lieber on traditional land use(Lieber, 1979), and this proved most useful as the assistant always knew exactly which part of an islet we were at. This enabled detailed records to be kept on tape recordings, selections from which were later transcribed into a running diary. Several problems became evident shortly after the survey began, and the initial areas were covered a second time for the sake of consistency. These problems can be enumerated as follows: Firstly, informants kept referring to places where there had been occupation during living memory but for which there was a complete absence of normal archaeological evidence. Kapingamarangi residents are meticulously tidy, discarding into the sea all their rubbish. Consequently, there is an absence of midden in occupied areas. Imported coral gravel {kilikili) can be identified as culturally deposited in two different ways. The grain size can be a distinguishing characteristic against the background of humified loam, or coarser coral rubble. In addition, the kilikili can be culturally modified by subsequent activities, such as the incorporation of midden, charcoal and ash; all of which again contrast with the general background. Present residents try to keep kilikili surfaces as clear as possible, and cooking activities are carried out in an area separate from the dwelling. This area can show archaeologically as a small mound feature or as black soil brought to the surface by land crabs. These types of evidence certainly were found on the islets, but cultural soils generally were poorly developed. This contrasts with nearby islands such as Nukuoro and Ngatik, where cultural soils are readily distinguished, and have a significant component of midden as well as artefacts and industrial debris. The writers saw more artefacts and industrial debris during two hours on each of these islands, than during the entire site survey of Kapingamarangi. This first problem of site identification was confined to the low islets of the atoll; that is everywhere except Touhou and Werua. Most surfaces are within one metre of the high tide, so that exceptional tides and stormy weather will certainly remove sediments and deposit rubble over site surfaces. It is strongly suspected that some areas of low islets have been physically stripped of occupation horizons by the sea many times over the past millennium(see Figure 4 ) . Secondly, the high islet of Touhou is at the opposite end of the spectrum. Here there is so much culturally deposited soil that the whole islet has to be considered as a single archaeological site. The eastern side of Touhou has an exposed deep section running along its entire seaward side. This shows -16- Figure 4: The islet of Pumatahati. This is a typical example of the low lying character of islets on Kapingamarangi. Archaeological sediments of any age do not occur on islets like this, which are periodically stripped by storms and tidal waves. very uniform modification of the imported soil throughout its length, and could not be used to identify different activity areas. With the exception of modern cooking places the same continuity is evident on the surface of this islet too. The islet of Werua lies between these two extremes of the low-lying islets on the one hand and Touhou on the other. Against this background it was clearly not possible to employ the normal system of site numbering; rather, each islet was annotated in the diary in terms of the types of sediments found, and the character of any further archaeological evidence. The archaeological sites found on Kapingamarangi can be placed into several distinct categories as follows. Culturally Modified Soils Cultural soils were by far the most common feature seen. Coral gravel has been brought onto the islets as kilikili from beaches and deposited around habitation areas, an activity that can be observed today on Touhou and Werua. This is a very common practice in Polynesia, and the normal process of habitation alters the appearance of the gravel by the incorporation of charcoal, ash, sparse food refuse, and other debris. Consequently, kilikili can be distinguished from naturally deposited gravel by this cultural modification. -17- Figure 5: Coral stone facing to taro swamp on Werua islet Figure 6: Coral slab walls on Werua islet -18A few areas showed rather darker cultural soils which are most likely to be associated with sustained cooking activities rather than house foundations. This kind of soil, however, was only rarely found during the survey. Genuine midden areas, such as shell or fishbone dumps, are more or less absent on Kapingamarangi. No doubt past communities did what the contemporary inhabitants can be seen doing - throwing their rubbish into the sea. Small quantities of midden do become incorporated into and become covered by kilikili. Again, this process can be observed on the modern surface, and was also evident in the excavations. Earthworks and Culturally Redeposited Soils Large quantities of natural sediments have been thrown up during excavation of fields for taro cultivation and redeposited in the form of high ridges surrounding the fields. The cultivations are dug down to the surface of the freshwater lens which occurs on top of the sea water which permeates the substrate of these islets. The ridges can reach an exceptional height of 6 m, such as those on Werua. They afford protection from high seas. Lower ridges are inadequate in this role. A severe storm, along the margins of a typhoon further to the north, in December 1979 resulted in sea water getting into many taro fields, thereby killing the crops. Many of the taro field ridge sections were examined without finding unambiguous evidence that the original excavations had been made through an earlier cultural soil, or buried natural horizon. This was a very surprising discovery, and two possible interpretations can be suggested. Either, all these excavations are amongst the earliest archaeological features on Kapingamarangi, or periodic sediment stripping by the sea means that whatever taro pits are present at any one time are always the oldest archaeological features present, and are continually re-cut into natural material. Coral Block Masonry Stone working, using coral blocks, occurs on Kapingamarangi in several places, but is not as spectacular as that found on some other Polynesian islands. The roads on Touhou and Werua are outlined with coral slab uprights, as are some domestic areas. Houses are only rarely outlined with actual curbing. Some taro fields have dry stone facing to the interior, such as on Ringutoru and Werua(see Figure 5 ) , and there is also some dressed coral slab facing on the roadside on Werua (Figure 6 ) . In addition, wells are invariably lined with coral slabs(Figure 7 ) . Land boundaries are designated by coral stone slabs which might be free standing or incorporated into roadside curbing. A traditionally important coral stone slab rests in an upright position to the east of the present church on Touhou(see Figure 8 ) . -19- &j# :»*»«< Figure 7: Stone lining to freshwater well on Werua islet Figure 8: The stone of Utamatua on Touhou islet. -20- Figure 9: Stone fish trap between Tetau and Nikuhatu islets Free-standing sea walls have been built on both the seaward and lagoon sides of Touhou. Much of the seaward wall was destroyed by high seas recently, and was being reconstructed during the period of fieldwork. Although communal activity was involved on at least one occasion, the responsibility for reconstruction was that of individual families whose land holdings were being threatened, The discontinuous character of the partially rebuilt walls had promoted the formation of small beaches of coral gravel, which was the source of some kilikili being carried in baskets and bags to be deposited around houses. Finer materials are obtained from sources on other islets. Fish ponds and fish traps were recorded during the survey. One large pond, approximately 20 m in diameter, occurs on the reef at the southern side, near the entrance of the atoll. This was said to be used to store fish which are caught nearby, but was in a state of disrepair when visited. Five fish traps were said to have been present before the recent high seas. The only one now extant occurs in the channel between Tetau and Nikuhatu. This consists of a double line of boulders arranged in a funnel shape in the shallows(Figure 9 ) . It is used for channelling small fish(te tawin) into the neck, across which a fine net is placed. Another important type of coral stone construction is that of the land-building wall. This is built between two blocks of land where there is a shallow channel, and it allows the sea to throw up sediments behind it, thus creating new land. -21- Figure 10: Land building sea wall on Pepeio islet. Note the low-lying nature of the land on the islet, and the effects of recent erosion. Figure 11: An old land building sea wall on Herekoro islet which is now well consolidated and colonised by trees and shrubs. -22Several examples of this type of wall were found, notably those at Pepeio, and at the southern end of Hare at Herekoro and Herengaua(see Figures 10 and 11). Traditional Sites There are three notable traditional sites in Kapingamarangi. The first is situated on the islet of Tipae, and is known as the landing place of Utamatua, one of the founding ancestors of the Kapingamarangi people. On the north side of Tipae is a group of rough coral boulders forming a structure approximately 7m long and lm high. These are certainly culturally placed, and although they now form an untidy pile, informants say that they were carefully arranged until fairly recently, when they have been disturbed by people seeking coconut crabs. The second traditional place of interest is the tiny sand patch called Matuke near Nunakita in the north. The patch, which could once have been an islet, is now a rocky outcrop only about 20 m in diameter. Utamatua is said to have moved Matuke after being challenged by an earlier resident on the atoll, and thus became the rightful owner of the land. Utamatua may well have shrewdly awaited the onset of a storm before taking up this challenge, because a high sea would certainly lend some assistance in moving an islet in this vicinity. In the centre of Touhou is a large coral slab upright known as satuhatu. It is said to mark the burial place of Utamatua(Figure 8 ) . This stone was earlier adjacent to the cult house(Emory, 1965:206 ,-Figure 39), but this has been replaced by the present day christian church. Historic sites Apart from recent evidence of occupation on the atoll by the Kapingamarangi themselves, the most obvious historic sites are those relating to the Pacific conflict of 1942-1945, when the Japanese established military bases. There is also evidence of the more recent US Air Force weather observatory, and finally the remains of ill-fated Japanese fishing vessels on the outer reef. About 100m offshore from the Luawa land block lagoon side of Hare lies the broken up remains of a bomber(Emory, 1965:22) in about 10m of water. After American divers exploded the bombs remaining in the and this has further modified the wreck. on the B-17 the war, aircraft An area on Hare was the main gasoline and munition storage dump for the Japanese seaplane base situated on the lagoon side of Hare. Here still stands a square concrete water tank. There is a series of bomb craters in the vicinity including one massive one, apparantly representing a direct hit upon the Figure 12: V e r t i c a l photo of Touhou i s l e t taken i n 1942. An American I n t e l l i g e n c e r e p o r t i n d e n t i f i e d 70 b u i l d i n g s from t h i s photograph of a p r e f a b r i c a t e d t y p e , s i m i l a r t o t h e Quonset h u t . F o r t u n a t e l y , i t was l a t e r r e a l i s e d t h a t they were domestic houses of Kapingamarangi p e o p l e , o t h e r w i s e they may have been bombed when t h e Japanese i n s t a l l a t i o n s were d e s t r o y e d in 1943. A causeway i s under c o n s t r u c t i o n between Touhou and Werua t o t h e l e f t of t h e photograph. main munition dump. This bombing r a i d followed an a e r i a l reconnaissance survey of Kapingamarangi in 1942(see Figure 12). About 150m offshore from Hare(see Figure 2) i s the wreck of a Japanese naval vessel which remains i n t a c t and in r e l a t i v e l y good condition. This was recognised by US i n t e l l i g e n c e as a 38m two masted a u x i l l i a r y schooner, with an extensive radio antenna system. I t was sighted in the lagoon several times between October 1942 and October 194 3. The ship appeared to be one of a type of radio ship used by the Japanese to provide communication in s t r a t e g i c areas and r e p o r t weather conditions (Anon., 1944:1). The vessel r e s t s in about 12m of water. There i s a double b a r r e l l e d a n t i - a i r c r a f t gun on a bow platform, and there i s a marked f l a r e to the bows. Of p a r t i c u l a r importance i s the fact t h a t the ship c a r r i e d a b a l l a s t of rocks. Several large water-rounded rocks were recovered from the wreck by diving, and c o n s t i t u t e the only l o c a l source of rock, other than t h a t a r r i v i n g in the roots of d r i f t i n g t r e e s ( s e e Figure 13). In an area between the US bomber and the Japanese vessel j u s t mentioned i s a small Japanese m i l i t a r y s i n g l e - s e a t a i r c r a f t . I t l i e s in about 4m of water, and i s now much scavenged by divers. -24At the northern end of Hare is a line of wrecked Japanese sea planes resting in the sand of the lagoon, foreshore. There are at least seven rotary engines, each with a triple-bladed propellor(Figure 19). Immediately inland is a scatter of tangled aluminium wreckage representing some of the fuselages and wings of these aeroplanes. Aircraft identified by US Intelligence are the Nakajima Type 97 fighter seaplanes, and Kawanishi Type 97 four-engined flying boats which were used in operations against Rabaul and Kavieng(Anon.,1944:1). The islet of Nunakita has abundant evidence of the Japanese occupation in the form of building foundations and cultural debris such as porcelain scattered over the ground surface(see Figures 14-18). An American Air Force weather observation station was situated until recently upon the southern islet of Pumatahati. An extensive cleared area is being recolonised by low trees, but Figure 13: A Japanese radio ship off Hare islet photographed in 1942. This ship was later sunk, and its rock ballast is collected by divers for modern use on the island. -25- Figure 14:Oblique aerial photograph of Japanese concrete two-story building near a boat basin on Nunakita islet in September 1943. The purpose of the structure puzzled American military intelligence(see Anon,1944:9). Figure 15: The same building in 1980. It shows some damage from strafing with plane machine guns, but is otherwise still intact. The bottom story is solid except for a narrow vertical shaft. Emory states that this building housed instruments in connection with the weather station(1965:21). -26- Figure 16: Concrete house piles at the Japanese installation on Nunakita islet * \ * • * * •••, : ' ?• •• • %, • * *r & «r% 4 m** .» 'v -' Figure 17: Aerial photograph of Japanese weather station on Nunakita islet, September 1943 after bombing by the American military. In the foreground is a 15m observation tower, and other buildings are to the left(after Anon,1944:9), -27- r.lfr- ^ " " , - r f Figure 18: Concrete ablution structures at the Japanese installation on Nunakita islet. A water storage tank is in the foreground. Figure 19: Propellors and engines from Japanese sea planes on Hare islet. This base was destroyed in bombing raids on the island in September 1943 (see Anon,1944:1,10) . -28rubbish dumps are still evident, along with three station buildings in good condition. Finally, there are two shipwrecks of modern Japanese fishing vessels on the Kapingamarangi reef. One was driven on to the north-western reef during the last few years. Everything of value in the wreck has been salvaged by the people of Kapingamarangi. The other vessel foundered on the.north-eastern side of Nunakita about 20 years ago, and has been almost totally destroyed by pounding seas. Remains of the engine and winches can be seen on the reef in the surf, and the steering gear is on the inner shore. The reef flat above the surge zone is scattered with rusting iron fragments. Summary Kapingamarangi is not rich in deeply stratified archaeological sites, and surface finds of artefacts are very rare. During the site survey one piece of pottery, and five shell adzes were collected. Considering the effort put into the survey, this lack of return is remarkable; and along with the absence of stratified midden debris is perhaps an indication of the tidiness of the Kapingamarangi people. Today, roads are swept, leaves are collected from around houses, and all debris is committed to the sea.A significant buildup of cultural debris is present only on Touhou, and for this reason, excavations were concentrated there. -29- THE EXCAVATIONS Scope and Methods Adopted Within the available time of one month it was not possible to explore more than a few lines of research; consequently, strictly limited objectives were set and the excavation strategy followed from these. First priority was given to exploring the overall sequence of occupation of Kapingamarangi; that is a diachronic approach to the atoll's prehistory. It was necessary to abandon any hope of examining equally important synchronic questions, such as those relating to settlement pattern - this would have required large-scale area excavations, which was simply not feasible. It was hoped that in the course of an overall diachronic study, sufficient midden debris would be found to assess the character of Kapingamarangi prehistoric subsistence economics; and that enough portable artefacts would be recovered both to define the early material culture and to assess the possibility of contact with other populations, such as those on Nukuoro and islands further afield in Micronesia and Melanesia. The site survey left no doubt that the place best suited to documenting Kapingamarangi prehistory was Touhou islet. Here there was every indication of a long sequence of intensive habitation. Culturally introduced soils reached a depth of about 4m over much of the centre of the islet, and there was a good chance of establishing a detailed occupation sequence. Some general comments are appropriate about the specific choice of location for the excavations. Factors influencing choice of area to excavate vary depending on what the archaeologist wants to learn by digging. For instance, investigation of the settlement pattern in the period immediately before European contact would involve one strategy, while an attempt to define fishing equipment of the same period would involve another. In most parts of Oceania, the choice of location for an excavation to determine the latter would be closely related to whatever could be learned from modern informants about the location of traditional men's houses, canoe sheds, and so on. This is what might be called the 'direct historical approach'. The relevance of this strategy, however, to locating and excavating the same type of cultural material from earlier periods becomes increasingly questionable, the more removed the periods are from the present day. This is especially so, on islets suspected of having gone through significant changes in their geomorphological character, during the supposed period of human habitation. Archaeologists differ in their opinions as to just where the cut-off point in time is. The present authors regar'd an upper limit of about 300 years as the point at which one must begin to employ quite different site location strategies. As explained above, archaeological sites on Kapingamarangi are not discrete entities in the way they frequently are in other -30- • KA3 C 0 £^ *KA1 KAi« nS D gnn w1 JD ° a D' naif r n1 ] D nnn 50 Metres. Figure 20: Map of the northern half of Touhou islet, showing the location of modern houses and excavation squares. The dashed line represents the profile given in Figure 34. -31Pacific islands. This is especially so on Touhou, where the entire islet is an archaeological site. Normally one has the benefits of surface indications from sites which are spatially separate from each other - that is, it is possible to gain sample knowledge from several periods by examining the presentday surface. This is not generally possible on islets which are artificially built up, and some other strategy must be employed. Eroding sections of course are some help, but fail to give any clues to what is in the centre of the islet - for this very reason it is tempting to dig a hole in exactly this spot. Geomorphological assumptions, such as the general process of lagoonward shoreline progradation on atolls, provide one basis for choice of location for examining different periods; but this is not sufficient on its own. There is little point in digging in a location hoping to find a 600 year old men's house, if there is a good chance that the sea existed at this point 300 years ago. One should not underestimate the value of less concrete factors, such as the 'pricking of one's thumbs', or subtle variations in cliraatological aspect on an island, in choosing where to dig. It has to be admitted that 'pot-luck' may also be a significant factor on islets like Touhou. Some consideration was given to all these factors in selecting locations for excavation, without giving undue weight to any one of them. The four areas excavated were deliberately chosen to form an east-west transect of the islet. This general approach had been adopted on Nukuoro by Davidson(1971:Figure 3 ) , and was a very successful way of revealing the overall history of the island. It was hoped that a similar approach would work for Kapingamarangi also(see Figure 20). Since it was not possible to explore areal distribution of evidence in the month available for the research, small areas of 2 x 2m were laid out for excavation. It was hoped that at least three and possibly four of these could be completed in the time available. Working in a confined space in coral g r a v e l ( k i l i k i l i ) s o i l s is not without its hazards. For one thing, one cannot always determine the precise origin of individual lenses of gravel, which are sometimes re-deposited from earlier horizons following the digging of postholes, ovens and the like. As other archaeologists excavating in kilikili soils have discovered, there are numerous pseudomorphs and ghost structures which begin and end with bewildering complexity. It is difficult enough to trace these in plan during area excavation - it is well nigh impossible in a 2 x 2m square. Some compromise was necessary, and the unit level technique with 20cm spits was adopted as most suited to the circumstances. Some particularly important surfaces were exposed as they appeared and recorded separately; but with these exceptions microstratigraphy was sacrificed. All sections were plotted in detail, but some general observations should be made here to explain how these are to be interpreted. -32Cultural stratigraphy to the archaeologist working on coral islands tends to be a series of fine distinctions of shades of grey. The layers are largely coral g r a v e l ( k i l i k i l i ) , some finer aggregates than others, some with lenses of charcoal, some stained more than others, and so on. Most of these distinctions have little or no recognised pedological meaning. Often the lines observed in sections are undiscernible in plan during excavation, and in any event these lines and lenses can appear and disappear in an astonishing fashion which defies simple stratigraphic interpretation. In this situation, it is all too easy to lose sight of the wood for the trees. It is therefore important not to become so baffled by the fine stratigraphy during excavation that one fails to recognise major changes which denote significant breaks in the function of a site, or periods of surface stabilization, when these occur Arriving at a reliable interpretation of the cultural meaning of these major changes is another matter - it is obviously very difficult in the context of a 2 x 2m square. For these reasons, only changes which are thought to be of some real significance are mentioned in the following sections. By and large the stratigraphic changes recorded Square Layout Section Conventions .. I^.U * ' ,!• ! I.jj.l... ..'."ilj'.'ltf' --.'•, Humified Topsoil, N A A B "^I'.y.'v.'r • . .':.-.'•.?:-"-''.f •"•'••.• t Coral Gravel layer with consolidated surface or other significant change, D C Sand and or gravel lens. - 4 — 1 metre — I * HWSf££TI£&&?&» <?1&?SSS>*\ Ash lens Charcoal Concentration. Coral Boulder. Figure 21: The four excavations were laid out in the manner shown above. Sections are illustrated in Figures 25,27,30, and 33 using a series of conventions described above. -33probably represent periods of surface stabilization; this explains the somewhat unusual simplified conventions adopted for recording stratigraphy, and presenting them in section diagrams(see Figure 21). In our opinion, such conventions are entirely justified, and may be more meaningful in the sense that they assist cross-comparison with other archaeological sites. When the overall stratigraphic sequence in each site had been established with reasonable confidence, charcoal samples for radiocarbon dating were taken from the sections. The recovery technique involved trowelling out the spit matrix, sieving it through one-quarter inch(6.35mm) screen, and hand sorting on plastic sheets. All bone, shell, charcoal, noncoralline stone, and artefacts were recovered, and bagged unwashed. The method of analysis of this debris is described below. In this manner, four excavations were carried out, and approximately 54 cubic metres of material processed. In each case, the excavation square was covered with a large parachute to keep the exposed area dry. The Excavation at Tiroki(KAl) This area is in the vicinity of the highest part of Touhou, about 70m inland(see Figure 20). Dark cultural material was exposed at the surface, and it was an excellent location in which to establish a deep stratigraphic history of the islet. A single potsherd was recovered on the surface, the only one found on Kapingamarangi(see Figure 42), and this was an additional factor in choosing this spot for excavation. According to Lieber(1979:122) informants on the island disagree about the ancient activities on this plot of land. However, one of the oldest men claimed that this plot was used to dump trash from the northern side of the cult house area. This was located where the present christian church is, and is shown on Figure 20. It is not much further from the site of the cult house to the sea - why it is thought that trash would be dumped on this spot rather than thrown in the sea as it normally would is not explained. There is also supposed to have been a 'burial house' in the vicinity of Tiroki, otherwise the plot was used mainly for growing coconut and pandanus. The name Tivoki literally means rubbish heap. The plot of land is marked by several upright coral slabs, which outline an approximate rectangle. The excavation square was situated near the northeast corner, and it proved to have about 4.5 m of deposit(see Figures 23 and 24). Using a bench mark on the concrete wharf(+1.050mASL), a datum was set up at Tiroki(+4.122mASL), and all relevant information was referred to this point during the excavation. Full details of this excavation and the others we^e recorded in the diary. -34- F i g u r e 2 2 : V e r i c a l photograph of Werua i s l e t t a k e n i n 1942. Note t h e i n d i v i d u a l p l o t s of t a r o being c u l t i v a t e d i n t h e low swampy ground i n t h e c e n t r e . This a r e a has been excavated t o t h e f r e s h w a t e r l e n s near sea l e v e l . Immediately behind t h i s swampy ground i s a r i d g e of sediments which have been thrown up t o a h e i g h t of 6m, and a f f o r d c o n s i d e r a b l e p r o t e c t i o n from bad storms which a r e prone t o washing sea water i n t o t a r o p a t c h e s which can t o t a l l y destroy the crop. A s e r i e s of ten reasonably well defined l a y e r s were encountered in the excavation, together with a number of d i s t u r b a n c e s . These are shown in the section diagram, Figure 25. The layers may be described as follows(NB: depth figures vary, and are therefore only approximate). Layer l(20am): humified top s o i l and c o r a l g r a v e l , with root h a i r s . Layer 2(30am): loose c o r a l g r a v e l , with large q u a n t i t i e s of fishbone. This m a t e r i a l i s c o n s i s t e n t with the t r a d i t i o n a l information on the function of t h i s s i t e . Layer 3 (30am): The surface of layer 3 was c l e a r l y indicated by a compacted l i g h t grey s u r f a c e , marking a period of surface stabilization The body of the layer consisted of c o r a l gravel with some f i s h bone. -35 Figure 23: The excavation at Tiroki(KAl) was on a low mound about 4m ASL, and sited where a single potsherd was found on the surface(see Figure 42). The entire depth of coral soil to below freshwater was culturally deposited. Only the top 50cm has cultural material consistent with traditional knowledge of the activities on this plot. In the background is a parachute, used to cover the excavation during downpours which can arrive with little warning. Figure 24: The stratigraphic sequence at Tiroki(KAl). The excavation went down to 4.6m from the surface, and cultural material was found to 0.5mASL. The curious varve-like features in Layers 5 and 7 are difficult to see from this angle of view. A large oven feature is clear on the main West wall. KA-1 North -Ji..,-..1M. ; , . ' - ' . . . . " ' - •' - . ' South L . I I , . , 1.^111 '- ' - ' • Excavation Limits INZ 5135) 679BP±76 Lens of fishbone 18 19 <& o-\ yy 20 ci? ££(* Freshwater level ( -0-55m ASL ) • 21 - 22 : 23 • 2L -1 - « Figure 25: Section diagram of the excavation at Tiroki(KAl). The large oven feature shown on the West wall is clearly seen in Figure 24. -37Layer 4(25 am): compacted brown sandy soil, cut into by a number of features, indicating the use of the area for cooking activities at this time. .Layer 5(110 am): This was marked at the surface by a zone of humification, representing an old stable surface, but not a compacted one. The main body of the layer was very unusual, consisting of at least 100 varve-like silty structures, highly suggestive of water action. In the body of the layer, a large wet patch was evident showing that water still accumulates in the layer. Layer 6(10 am): humified loam. dark brown to black charcoal stained Layer very similar in composition to Layer 5. 7(15 am): Layer 8 (40am): This layer marks the beginning of white coral sand and rubble which characterize the natural soils. Four bands of sandy silt occurred in the layer, representing points of surface stabilization. Cultural material rare. Layer 9( 50 am): White angular coral rubble, and flake-like sand. The layer is very natural looking; however, a lens of culturally deposited fishbone occurred in the layer, together with the distal end of a pearl shell lure shank(see Figure 40). Layer 10( > 110 am): flake-like white sand, and large coral boulders. This is natural beach material, representing the earliest phase of Touhou prehistory. Features: During the excavation, one massive oven structure was uncovered. This shows on both the South and West sections(see Figure 25). It had been used a number of times, and had quite a complex structure to it. Two independant charcoal samples from two different parts of this oven complex gave similar ages of <250 years BP. A large wet patch occurred in the body of Layer 5, indicating that water either congregates or flows in this spot for some unknown reason. Two aspects of this stratigraphic sequence are particularly important. Firstly, there is a clear distinction between naturally and culturally deposited coral gravel at the junction between Layers 7 and 8, occurring at 1.4mASL. However, cultural material, including artefacts, midden, and charcoal, was found to a depth of 0.5mASL in these naturally deposited soils. This shows that people inhabited Kapingamarangi atoll when Touhou islet was barely above the sea, and must have consisted at that time of a low sandy flat, periodically washed by exceptionally high tides, and certainly storms. A radiocarbon date of 679 + 78 BP, represents a time before the build-up of the cultural soils which form Touhou islet today. Three dates in the bulk of this artificially formed horizon indicate that the bulk of this build-up, at this part of Touhou anyway, occurred within the last 250 years. -38Secondly, the soil complex represented by Layers 5 and 7 is somewhat enigmatic, but of potential importance. Water action is strongly suggested by the varve-like appearance of the section, and perhaps also by the large wet patch in the body of the layer. One obvious possibility is mulching activity in a taro patch. This has the effect of storing water, and could perhaps form features of the kind observed. This interpretation can only be advanced with reservation, however, for two reasons. On the one hand it would imply quite a different economic function of Touhou islet. As Emory has pointed out(1965:122), Touhou was not a place for taro cultivation, in the historic era, when the wet fields were not owned individually, but grouped together and situated on other islets - largely Werua, Hare, Ringutoru, and Torongahai. This same pattern exists on Nukuoro also, but there is some suggestion there that some taro patches until relatively recent times were individually sited in the vicinity of family houses(Davidson,1981:pers.comm.). Such a possibility could also be entertained for Kapingamarangi, especially since it is believed that only Colocasia esculenta was cultivated in the pre-European period, and a change of plot location is quite feasible. It is most interesting that the height of this soil complex is considerably above the main water table( 1.4 to 2.7mASL). The modern cultivation pattern is to establish the taro field at the level of the GhybenHerzberg freshwater lens(see Figure 22). Of course this pattern is of relevance to the cultivation of swamp taro,Cyrtosperma ohamissonis, and may not necessarily be most appropriate to a largely different kind of crop. The freshwater lens can reach a maximum height of about 1.7m above the salt water table(McKee, 1956:24), but would normally be only about 30cm above. On the whole the origin of this layer is somewhat mysterious, but the interpretation of a dry-land taro patch is cautiously put forward. Some cultural material was recovered in the layer, decreasingly common with depth. The subject of this layer is returned to later. The few artefacts recovered in the excavation are described in detail below. At this point, mention can be made of the find of a supposed distal fragment of a pearl-shell lure shank in Layer 9. This is of some significance. It was intimately associated with a radiocarbon sample which yielded an age of 679 + 78 BP. This establishes the existence of this unique lure form to a period quite early in the Kapingamarangi sequence. Only two shell adzes were found(see Figure 37), and these are fairly late in age(Spits 3 and 8 ) . The faunal assemblage is substantial, especially from the upper levels, entirely vindicating the modern name of the site location. If the fish bone is any indication, Spits 1 to 3 at this site represent a more intensive dumping pattern than anywhere else excavated(see Tables 12-15). -39The E x c a v a t i o n a t Muri-Harau(KA2) The s e c o n d a r e a c h o s e n f o r e x c a v a t i o n was n e a r t h e p r e s e n t l a g o o n s h o r e , i m m e d i a t e l y i n l a n d from s e v e r a l modern c a n o e s h e d s , and n e a r t o an i m p o r t a n t m e n ' s h o u s e . The v i c i n i t y of a c a n o e s h e d i s o f t e n assumed t o be an i d e a l p l a c e i n which t o f i n d e v i d e n c e of f i s h hook m a n u f a c t u r e ( F u j i m u r a and A l k i r e , 1 9 7 9 : 7 9 ) , and a t r a d i t i o n a l m e n ' s h o u s e i s c o n s i d e r e d a good c h o i c e a l s o ( D a v i d s o n , 1 9 7 1 : 9 ) . Muri-Harau l i t e r a l l y means ' b e h i n d t h e c a n o e h o u s e ' , and L i e b e r s u g g e s t s t h a t t h e r e was one and p o s s i b l y two c a n o e h o u s e s h e r e a b o u t AD 1 8 0 0 ( L i e b e r , 1 9 7 9 : i , 9 1 ) , as well as a ' b u r i a l h o u s e ' ( s e e a l s o Emory,1965:85). A g e n e r a l p r o c e s s of s h o r e l i n e p r o g r a d a t i o n i n t h e d i r e c t i o n of t h e l a g o o n i s now w i d e l y d o c u m e n t e d f o r i n h a b i t e d c o r a l a t o l l s , and an e x c a v a t i o n s q u a r e i n t h e v i c i n i t y of M u r i - H a r a u s h o u l d t h e r e f o r e n o t be t o o c l o s e t o t h e modern b e a c h , i f d e p o s i t s of more t h a n a c e n t u r y o r s o a r e t o be f o u n d . A s p o t 25m back from t h e p r e s e n t s h o r e l i n e was c h o s e n ( s e e F i g u r e 2 0 ) . The M u r i - H a r a u p l o t of l a n d i s q u i t e a c o m p l i c a t e d s h a p e w i t h f i v e s i d e s . The e x c a v a t i o n was s i t u a t e d on t h e w e s t e r n h a l f , a t 4.2m t o t h e w e s t of t h e p r e s e n t day roadway a r o u n d t h e i s l e t . A datum a t +1.923mASL was e s t a b l i s h e d , and a l l information r e l a t e d to t h i s p o i n t . Figure 26: The e x c a v a t i o n a t Muri-Harau(KA2) i n p r o g r e s s . The c u l t u r a l m a t e r i a l d e p o s i t e d h e r e reached a depth of only about 2m, and was l a r g e l y r e c e n t i n o r i g i n ( u p t o 300 y e a r s ) . The s i t e i s immediately back from t h e p r e s e n t lagoon s h o r e , and c l o s e t o a modern canoe shed. In t h e background,the sediments a r e being s i e v e d and s o r t e d on p l a s t i c s h e e t s . KA-2 meters ASL North East 2- South West Sp.t Layers .•:• Wff. • ; '» ! : .*J :f^} ';'.-1 • V>. ^:y• ;f;l^|-v;:---i-lfW'l^r^L:->:^ !; i ,i, *?. tT-TT". ..: . . . , - , , . . , » . " .-'IM I .:J1^J , ( ,MP'I..,.,... ? U J H'I^..IJI.J I ® © 1 - ^ ^ HIJ1.1^.^T^-UTJJ?JI• '.wvyv.;-',;>• .....i. r....' ,• I?.-!U.AI:I w. ^'.AL I U U / . J . - J I . u.»rJui» ^ M ^ ^ . ^ L L ^ V - ^ ^ U . < 1 r t^iMJj.J.'l J - . M . I r J J J J . L-T.• ."JM . . , " U I J | , L W ;"^/,!,:;,liwwi. * •• - • «• »'—• : i:Ji'': :,:.,;!-,-.-,l.-'.!-:,'LT5 'I'jjkitLr• »• • • -i i-'-'-j"'-—• -•• ' vnj..-y-iw<-''i'•-r.• •»'• • r^aaiSr^s^S^p^^^ffiip-' ^ ^ ^ p > ^ ^ f e ^ ^ ^ ' --• .'.• '.-• <'• •;'« .. - j_ 2^ I ife. O I ^ © J&&*?*" ««*£$ ' aw** . .., :tf ; v >5»^i > - - - Seawater level (-0.42 m ASL) Excavation Limits Figure 27: Section diagram for the excavation at Muri-Harau(KA2). European artefacts were recovered from as deep as the stained zone in Layer 7. Some of these may date to the early European contact phase. Note that gravel lenses occurring in Layer 7 slope down to the present lagoon shore. This shows the general process of shoreline progradation over the past 300 years. -41A series of seven moderately well defined layers were discerned in the excavation(see Figures 26 and 27). These may be described as follows: Layer 1( 10 am):This was a complex layer consisting of patchy humified coral gravel with several zones of compaction. Layer 2( 15 am): Loose layer of yellow sandy coral gravel with irregular compaction. Layer 3( 20 cm): compaction. Grey stained coral gravel with irregular Layer 4( Grey sandy-silty coral gravel. Layer 5( 10 am): Charcoal rich coral gravel. Layer 6( Grey sandy loam 15 am): 25 cm): Layer 7(>100am): Complex natural yellow beach sand layer. The top 10-15cm was a zone of discoloration, either from the leaching action of the upper cultural soil, or initial occupation debris on the clean sand. Scattered through the layer were a number of distinctive gravel zones. These are tilted down in the direction of the lagoon, and are progressively more horizontal later in the site's history. This is clear evidence of progressive shoreline progradation. Features: Several significant features were detected. A hearth structure occurred in Layer 1, and another in Layer 5. A large disturbance, possibly a huge post hole, was also cut into Layer 5. This site has two clearly defined phases, separated by the junction between Layers 6 and 7. European material was found right to this junction, and some of it may belong to the early European contact period. A charcoal sample which gave the single radiocarbon date for the site of <250 BP(see Figure 27) was located in Layer 5. Some of the artefacts found are illustrated in Figures 45-47; a few might be mentioned here. An American 5C coin(1964), and a Japanese 10 sen piece(1942) were recovered from Layer 2, and a Japanese 1 sen piece(1927) was found in Layer 3. One pearl shell button(Figure 47, #F) was found on the interface between Layers 6 and 7, and may well be early European in age. A quantity of glass was found as deep as Layer 5, and various glass beads were scattered through Layers 3 to 5. Again, some of these items could be moderately early, possibly back to Spanish times. Two metal fish hooks were found from Layer 4(Figures 39,45). These are made in Polynesian style, again suggesting some age. A small amount of economic debris was found throughout the site, extending below the main junction into Layer 7. This was analysed by Layer rather than by spit as in the other excavations -42The Excavation at Putau(KA3) The third excavation was situated on a plot of land known as Putau. Lieber suggests(1979:128), that about AD 1800 this area had a burial house as well as other houses for residence, and a cooking area represented by a mound in the North-east corner of the plot. Erosion processes have taken away some of the scarp on the seaward side, and in 1979 some human skeletal remains were exposed in the South-east corner of the land. These were recovered by Lieber, and are described in Appendix 10. The mound is a most interesting feature. The surface layer is a fine black ash, and in this respect is very similar to mounds on both Ngatik and Nukuoro. Only one such mound now exists on Kapingamarangi, although it appears that others may well have existed once. They are still fairly numerous both on Ngatik and Nukuoro. They clearly represent an area of concentrated cooking activity, and an excavation in the vicinity could therefore be expected to yield valuable economic information, and probably industrial material also. The Putau area is a four sided parcel of land, and the excavation was situated in the north-east corner at a position 2.6m back from the present scarp, and 7.8m from the north boundary, directly on top of the mound. A datum was establishd at a somewhat lower point than the top of the mound at 2.953mASL, and all material is referred to this point. The location of the site is shown on Figure 20, and general views of the excavation are given in Figures 28 and 29. A series of ten reasonably well defined layers were observed(see Figure 30), and these may be described as follows: Layer 1( 20 am): Layer 2( 60 am): humified top soil. black sandy loam, with some coral rubble. Layer 3( 70 cm): dark brown sandy loam. This is a complex cultural soil, richer in ash and charcoal towards the bottom. Layer 4( 50 am): light brown coral gravel. This layer was generally rather sterile, but rich in fish bone at the top. There was loose coral rubble at the base. Layer 5( 60 am): loose white coral gravel. This is a complex cultural soil with various lenses and patches of blacker material. Layer 6( 20 am): dark grey sandy soil with patches of white coral gravel and sand. Layer 7( 20 am): clean white sand. Layer 8( 15 am): black cultural soil -43- Figure 28: Laying out the excavation at Putau(KA3) on a low mound of black soil about 3.5mASL. All of this deposit is artificial. To the right can be seen the sea wall under reconstruction, following a massive storm in 1979. Figure 29: The excavation at Putau(KA3) nearing completion at a depth of 4.1m. The consolidated basement of coral at this site was just below sea level, and cultural material was recovered from its surface. The reflection in the water is of the parachute covering the excavation. KA-3 metres ASL Spit North East South West Layers mA^'fW'k *;-WMWWJ$^X»#vw WT^T j ^ . , :•• v..l^W^)?•p' 3- •••' • . •: ' • ( NZ 5i«) < 250 BP © <2> —i ! r-i" •. - . . . . . - ' ••- . ' , . . i....... i,.;,,.-,^,...,,/ ,....,..• . - ^ ' .•••• . ' . i rrra 2~ <tasfe ,-r_ 3 :• ; -••• —ftai. ^5^ 0 1 - (NZ51U) • -•••i• k ' TLI -5 2 5 7 B P ±5 6 6 ^ ^.'-• -',..,;......• • .• .-• ™ •L'.ViJ' : --.v_u.» ".• i •. - .. ' " - ™ . • - ,'••. ' . . ' A -•- ' - '•• • >-.... "• ^ ^ ^ ' IJL i./ij^- >• - 1 -.yj J vtfit*- :• --.' T- -• 7 8 •• -.'• '• ' , ' \ 7 7 9 -^ssp^«*^^ c : ;.»<^ • 13 • •i •i'ii:-»i».',nvMl'.'.'.,-'j|iMj.i' ••|J—"j.* yrt^^nVH^V/^J^^^^vL^J^i'w^LMMw- % —" © ~ 15 ?kmwi&&wm® w&®@$^ j^^s^Jta^^w^p^sp^^^w5?^^^5££&££SS5£? 2& 16 1 'F^TTi• .:^ > 1 .•• • -•.!.• • , • • L ^ - , . . . . ^ r — ' • - •-• • • if• u.-^WiJy.T*.'J.- . •• _ 17 Remnant of old sea wall ? © riuLikHinng 18 . . . . ,._• . - Compacted Coral Basement' 1 9 |y 21 \ 22 Figure 30: Section diagram of the Putau(KA3) excavation. A rich cultural layer was present below sea level. -45Layer 9( 60 cm): Layer 10(20 am): white coarse angular coral gravel and sand. grey sandy coral gravel - cultural soil. Features: Few features were noted in this excavation. However, in the top part of Layer 9 some flat coral slabs were present face down, and these may have been the remnants of an early retaining wall of some kind. This is a complex archaeological site. Its function as a cooking area does not extend very far back in time, probably only to the base of Layer 3. Below this point, the cultural layers are far less rich in charcoal, and the surface of Layer 5 particularly is marked by the appearance of very natural looking material. The bulk of the deposit below the surface of Layer 4 is interpreted as part of the artificial build-up of Touhou islet. Two points are notable about this material. Firstly,it does not consist of carefully selected kilikili coral gravel, but is rather a patchwork of sand and coral rubble and gravel. This was probably obtained from the nearby reef flat area. The material should therefore be interpreted as deliberate build-up or fill material, rather than fine quality kilikili, such as is deposited in and around house floors. Secondly, there are a number of layers of richer cultural material which occur during this period of build-up; and this shows that the process was a progressive one rather than a single concentrated operation. It is most interesting that cultural material occurs right on the indurated coral basement of Touhou islet. As with the evidence from Tiroki(KAl), this suggests that people were present on Kapingamarangi at a time when Touhou was virtually awash with the sea. Clearly, these people must have inhabited some other islet at this time, and only temporarily visited Touhou. The fluctuating sea level is shown on Figure 30, and it will be noticed that the lowest cultural layer occurs below sea level. Tidal fluctuations on Kapingamarangi vary from about 60 to 120cm through the lunar cycle(Wiens,1956:22;Niering,1956: Figure 9 ) , and the coral basement at Putau is about -0.60mASL. Thus, Layer 10 is always slightly under sea water. For this reason, the material in Layer 10 is not necessarily in a position of primary deposition, but is interpreted as part of the general tide wash from activities being carried out on Touhou when the islet existed as a low sandy flat, barely above the sea. Four of the five radiocarbon samples give a clear picture of the chronology of events at Putau. These are shown on Figure 30, and are further discussed below. It may be noted here that the upper horizon(above Layer 4 ) , consisting of the cooking mound function, dates back to about 250 to 300 years only. The period of artificial build-up of Touhou precedes this by a further 400 years to about 700BP, and the lowest cultural horizon(below Layer 8 ) , is somewhat earlier again. The charcoal sample from Layer 10, unfortunately does not provide a reliable estimate of the age of this layer.The reason for the divergent result is obscure, but being immersed in sea water,with circulating modern carbon, for up to a millennium must have something to do with it. -46Artefacts were again rare in this excavation, but some are important enough to mention at this point. A further distal fragment of the characteristic Kapingamarangi pearl shell lure shank was recovered from the gravel lens shown towards the base of Layer 4 in Figure 30. This item was unfortunately amongst the material lost in transit with the freight containers, and is therefore not illustrated. The closely associated radiocarbon sample dated to 5 35 + 5 7 BP, provides another good estimate of the antiquity of this lure form. Four Tridaana sp. adzes(Figure 37), and a Cassis sp. coconut grater head(Figure 41) were found in the excavation. Of particular note too are two triangular-sectioned Tvidaona sp. shell arm rings, found in Layer 10(see Figure 43). A large quantity of midden was recovered in the excavation, and this is described below. The Excavation at Ngeiho-Hereu(KA4) The fourth excavation was carried out more or less in the centre of Touhou(see Figure 20), in a parcel of land known as Ngeiho-Hereu, meaning north of the cult house. This is the largest and most important compound on Touhou. Lieber's records show the existence here at about AD 1800 of residential houses, burial houses, sacred houses connected with the cult house, and a large cooking area(Lieber, 1979:95-96). In the late 19th century, a men's house was also established here. The cooking area was once marked by a mound, comparable to that at Putau, but this was levelled during the Japanese period. There was a reasonable chance that an early men's house or canoe shed might be located deeply buried in this vicinity, considering the likelihood that the early lagoon shore once existed this far inland. The three excavations carried out so far had failed to yield much evidence of fish hook manufacture it was hoped that this final excavation would. An excavation square was laid out as near as could be judged to where the cooking mound once existed, 6.7m north of the curbing of the present pathway in the centre of Ngeiho-Hereu. A datum was established at 2.572mASL, and all material related to this point. A series of six well defined deep cultural layers were found in the excavation, and these are described as follows(see Figures 31-33). Layer 1( 35 am): Humified black sandy coral gravel Layer 2( 70 am): Grey ash. This layer evidently represents the remaining part of the cooking area function of this site. Layer 3( 80 am): Black coral gravel layer. This material is definitely kilikili gravel, showing the site to have had a residential function at this time. -47- Figure 31: The area at Ngeiho-Hereu(KA4) cleared of surface rubble immediately before excavation. A mound feature relating to cooking activities in this vicinity was levelled during the period of Japanese occupation of the island. Figure 32: The stratigraphic sequence at Ngeiho-Hereu(KA4). The excavation reached 3.7m in depth. Towards the base can be seen irregular coral boulders which may have been part of an early wall structure. Cultural material in this site extended to below sea level, and into the fresh water lens. KA-4 North East South West metres ASL Figure 33: The section diagram of the excavation at Ngeiho-Hereu(KA4). The boulders in Layer 5 are probably part of an early wall structure of some kind. Layers 3 and 4 are residential kiZ'l'ki'l-i, while Layers 1 and 2 relate to the traditional cooking activities of this site. -49Layer 4( 70 cm): Yellow-grey coral gravel layer. Again, the coral gravel is kilikili in origin. Layer 5( 60 am): Yellow sandy coral gravel with numerous coral boulders, and some lenses of cultural soil. Layer 6(>70 cm): Yellow sand with natural coral gravel. Features: The site had quite a few lenses of ash, charcoal, and gravel in various layers; but other than this it was a relatively featureless site. A large charcoal lens occurred in Layer 3. The large numbers of coral boulders in Layer 5 may have originally constituted some kind of wall structure, but this is not certain. Two radiocarbon samples provide age estimates of 354 + 56 BP for Layer 3, and 548 + 78 BP for Layer 5. The kilikili gravel in Layers 3 and 4 suggests that the area was primarily residential at this time, and was followed by the cooking activities of the upper 2 50 years or so. The earliest cultural remains in Layers 5 and 6 in the site were deposited in what is essentially natural material. Layer 5 may represent deliberate building up of the land above sea level, using nearby rocky sandy fill. The distinctive tilted strandlines found in the basal sediments at Muri-Harau(KA2) did not occur in this site; this suggests that the lagoon shoreline has been further to the west than this part of Ngeiho-Hereu for at least 600 years. Only a few artefacts were recovered in the excavation. Worthy of note at this point are a Conus sp. shell arm ring(see Figure 43), and a shell coconut grater head(see Figure 41), both from Layer 3; three one-piece fish hooks(Figure 39) from Layer 4, and four adze fragments from Layers 3 and 4(Figure 37). The presence of fish hooks in Layer 4 may indicate the existence of a men's house in this vicinity about 400 to 500 years ago. A considerable quantity of midden material was recovered, and this is described below. -50RADIOCARBON DATING AND THE CHRONOLOGY OF SETTLEMENT Touhou islet, small though it is, has an area of about 3.72 hectares. The four excavations carried out constitute a sample of about 0.04% of this area. On Nukuoro, Davidson worried that her 4 0 m 2 of excavations was not a very large sample(Davidson,1971:27); the 16 m 2 explored on Kapingamarangi may seem cause for even less security, when it comes to overall interpretation. However, the excavation and analysis of 54 cubic metres of archaeological debris represents a sizeable quantity of material to work with, and it is therefore believed that a reasonably reliable picture can be put together of the overall chronology of settlement on this islet, and perhaps for Kapingamarangi as a whole. In Figure 34 a profile of Touhou is illustrated. This shows that the land is somewhat higher on the seaward side, sloping away to the lagoon shore. The excavations went to well below sea level in each case, but only at Putau(KA3) was an indurated coral basement struck. This shows that Touhou is really only a pile of gravel perched precariously on the inner edge of a reef of coral rock. Cultural material was present almost to the base of each excavation, and this demonstrates that almost the entire volume of Touhou islet above sea level(estimated at about 96,000 m 3 ) has been redeposited by human agency. The extent of this build-up qualifies Touhou for the appellation 'artificial island' in the same sense as those recorded in the Solomon Islands at Taumako, and along the coast of Malaita. This particular character of Touhou is a striking contrast to the remaining low-lying islets in the Kapingamarangi atoll chain. The fact that Touhou continues to exist in its present elevated form is due quite simply to the continued human maintenance of a coral block wall along the seaward margin of the islet. In order to establish the sequence of events involved in building up Touhou to this present form, the deposits in the excavations must be correlated. To assist this process, a series of twelve charcoal samples were radiocarbon dated. The position of all samples is shown on Figures 25,27,30, and 33. The samples of charcoal-rich soils East 77777 Coral basement Figure 34: Profile of Touhou islet showing the four excavations, and following the transect illustrated on Figure 20 Vertical scale exaggeration five times horizontal. Each excavation went to below sea level. -51- were of course contaminated with coral, and pre-treatment was therefore necessary. The samples were sieved through one-quarter inch (6.35mm) mesh, and all larger coral discarded. The charcoal rich residue was treated with 2% HsPOi* until all CO2 evolution from the coral had ceased. The sludge was then washed, neutralised, combusted wet, and counted for 1000 minutes. The results are given in Table 4 below: TABLE 4: RADIOCARBON RESULTS FOR KAPINGAMARANGI Provenance T=5568 T=5730 T=5730* 6 13 C NZ5135 KA1 Layer 9 679 + 78 699 + 81 695 + 65 -24.2 NZ5136 KA1 Layer 3/4 <250 <250 <250 -25.7 NZ5137 KA1 Layer 3/4 <250 <250 <250 -23.2 NZ5138t KA1 Layer 6 <250 <250 <250 -24.2 NZ51395 KA2 Layer 5 <250 <250 <250 -23.3 NZ5140 KA3 Layer 8 714 + 58 734 + 59 710 + 59 -23.1 NZ5141 KA3 Layer 4 535 + 57 551 + 59 560 + 64 -23.8 NZ5142 KA3 Layer 10 558 + 57 574 + 59 591 + 59 -25.3 NZ5143 KA3 Layer 2 NZ5144 KA3 Layer 3 257 + 56 264 + 58 362 + 58 -25.2 NZ5145 KA4 Layer 3 354 + 56 364 + 58 430 + 58 -25.0 NZ5146 KA4 Layer 5 548 + 78 564 + 81 573 + 86 -25.5 <250 <250 <250 -24.4 NB: * = Corrected for Secular effect Nominal ages are as follows: NZ5138t 223 + 56 229 + 58 303 + 68 NZ5139§ 117 + 56' 120 + 57 152 + 89 Although these nominal ages cannot be taken too seriously, they are entirely reasonable estimates for their contexts. The only additional absolute dating evidence comes from the coins in KA2. Coins dated to 1964 and 1942 were found in Layer 2; thus 1964 provides a terminus -post quern for Layer 2. In addition, the 1927 coin found in Layer 3, provides a terminus -post quern for Layer 3. a. m 1/1 KA-1 L_ lKA-3 0 Level m 2 1 __.^-- 2 100 3 Level m \ 3 ^--—' 4 <250BP 5 6 7 <250BP 8 <250BP 257BP+56 wn' --*, * -^ 8 9 10 ^^ 2 N, ' \ 3 9 4 10 5 <250BP 11 535BP+57 — • 7 " 13 8 12 14 9 13 15 14 16 7%BP±58- 700 - — •— " " 17 " * 18 V 19 18 20 19 21 20 "21 -10 679BP+78 11 558BP+57 (suspect) 22 \ \ \ 3- KA-2 V \ \ \ 35^'BP±56 \ \ \ <250BP 12 \ ••" " " \ \ 11 16 \ 6 12 15 nnn 1 x>^ 5 7 Level I KA-4 4 S Level n 1/1 1 \ 13 \ \ \ s .14 7 15 8 16 5U8BP±78 9 17 \ \ 18 \ 19 Figure 35: The correlation between the sediments in the four sites excavated. 10 \ \ 11 — OmASL -53With one exception, these results form a very satifactory series. The result for NZ5142 is not considered reliable. In view of the importance of the cultural material below sea level in Putau(KA3) Layer 10, it was regarded as essential to attempt to date a charcoal sample from this provenance, despite the fact that any material here had been washed continuously with sea water for the duration of Touhou prehistory. Most archaeologists would agree that this is tempting fate, and the result is not satisfactory. In our opinion, the sample has been contaminated with modern carbon, although the precise mechanism is obscure. Taken together, the dates show that the Touhou archaeological deposits have formed fairly rapidly, and that the prehistoric sequence for Touhou,at least, is not much older than about 700 years. Taking into account the sequence of cultural and natural stratigraphy outlined above, the following reconstruction is offered(see Figures 35 and 36). Level I: Transient Occupation ( ? 1000 to 700 BP ) This period is represented by sediments and associated material found at two sites only: Tiroki Layer 10 (KA1 Spits 17-22) Putau Layers 9 and 10 (KA3 Spits 17-21) In both cases, the sediment matrix is essentially natural, but cultural material is mixed into the sediments. Putau at this time was only 0.25mASL, but Tiroki was slightly higher at 0.5 0mASL. There would not have been a natural freshwater lens on Touhou at this time, and it is most unlikely that the islet was permanently occupied. However, there certainly were people on Kapingamarangi atoll somewhere, because they left scattered cultural material in these early sediments on Touhou. These people must have had a village or villages on one or more of the other islets. It is quite likely that Touhou functioned as a small outpost, in much the same way as Matukerekere does today. That is, it is periodically visited, temporary shelters are made, and it is used as a base for fishing expeditions. The lower age of this period can only be guessed at. Level II; Formation of Artificial Islet ( 700 to 300 BP) This period is represented by sediments in three sites: Tiroki Layers 8 and 9 (KA1 Spits 13-16) Putau Layers 4 to 8 (KA3 Spits 8 -16) Ngeiho-Hereu Layers 3 to 6 (KA4 Spits 7 -19) This building up process was obviously quite complicated. At Ngeiho-Hereu, for example, Layers 5 and 6 represent redeposited natural sediments, while Layers 3 and 4 are clearly imported -54- Figure 36: An idealised picture of the evolution of Touhou islet over the last 700 years. Vertical scale exaggeration five times horizontal. Note that the lens of fresh water could not have always been present on the islet, but would have begun to form as the islet grew in size. -55residential kilikili coral gravel. At Putau, most of this fill material is redeposited natural, but lenses of rich cultural soil occur as well. Finally, at Tiroki, the fill is distinctly natural in appearance, but again lenses of cultural material are found. Touhou islet by the end of this period had been built up to a height of 2.OmASL next to the open sea(KA3), 1.8mASL on the lagoonside(KA4), and there was a hollow in the middle of the islet(see Figure 36), where the height was only about 1.4mASL(KAl). Level III: Intensive Residential Occupation ( 300 to 100 BP) This period is represented by material in the four sites excavated: Tiroki Layers 3 to 7 (KA1 Spits 3-12) Muri-Harau below crust Layer 7 (KA2 Spits 7-11) Putau lower half of Layer 2 (KA3 Spits 3- 7) Ngeiho-Hereu Layer 2 and top of Layer 3 (KA4 Spits 3- 6) These sediments are the product of residential activities consistent with those reconstructed for Touhou at about AD 1800 by Lieber(1979). In particular, the deposits at Putau arid Ngeiho-Hereu throughout this period are ash and charcoal rich gravels indicative of specialised cooking areas. On the lagoonside at Muri-Harau, the sediments are natural, and show that the high tide shoreline a century ago was still further inland than it is today. The sediments at Tiroki for the early part of this period(Layers 5 and 7 especially) are distinct, and somewhat enigmatic. One possibility is that they represent the accumulated material in a waterlogged low lying area in the centre of Touhou where cultivation of taro took place. The deposits are well above the Ghyben-Herzberg fresh water lens, and this area could not be thought of as an example of the kind of taro patch which is found on Kapingamarangi today. These are essentially designed for the wet cultivation of Cyrtosperma ahamissonis . A possibility which has to be suggested is that this area of Touhou was used for the cultivation of Coloaasia esculenta, the only taro form believed to have been on Kapingamarangi before European contact. This taro could have been grown in a dry-land garden, with mulching to improve water retention properties of the soil. Level IV: Modern Horizon ( 100 BP to Present) The upper six layers of Muri-Harau, and the crust of Layer 7(KA2 Spits 1-7) contain abundant European cultural items such as those made from glass, iron, and brass; and it appears that the bulk of this site should be placed in a relatively recent time bracket. For convenience, the upper two spits of the other three sites are also placed in this period. -56This designation is somewhat arbitrary, especially in the case of Ngeiho-Hereu, where some of the upper deposits have been removed in the last 30 years; but it serves to separate these superficial deposits from earlier ones for the purposes of midden analysis. -57THE ANALYSIS OF MATERIAL RECOVERED Methods Adopted When the archaeological material arrived at Otago, it was all rebagged using a uniform system of labelling such as KAl/A/Spit 6. Paper bags were used as they permit slow drying which greatly increases the strength of archaeological materials. At subsequent stages of rebagging, after drying was complete, self-sealing plastic bags were used for permanent storage. All identifications are written on these bags, as this greatly simplifies further more detailed studies in the future. Bags were boxed according to site. Particular items such as artefacts were washed and catalogued in a book, beginning at Ka7 9.1. Next, each paper bag was tipped into a one-quarter inch(6.35mm) mesh sieve, and the retained material was sorted with tweezers in a large plastic tray into: Bone Charcoal Pumice, stone etc Shell Artefacts and worked material and then it was rebagged, again in paper bags. The residue was also rebagged and kept. Bone material was then sorted into the main groups(fish, rat, bird, etc), and rebagged. The methods used in processing individual components are described in the various Appendices attached. Fish The fish remains form a large proportion of the midden material recovered from Kapingamarangi. The method of identification is described in Appendix 3, and need not be enlarged upon here. A minimum number of 1214 fish were identified, about equal numbers being in Tiroki(KA1),Putau(KA3), and Ngeiho-Hereu(KA4), and only about one third of this number in Muri-Harau(KA2). The results can be grouped according to fish families, and also according to time periods, following the chronology described above. These reorganised results are given in Table 5. About 30 species of fish were identified, and another three to five species are present in the midden in small numbers but are not in the comparative collection. As might be expected, only a few species are ever very numerous. The information is pooled and reorganised further in Table 6. From this, it will be seen that fishes in the families Epinephelidae (cod and groper family), and Scaridae(parrotfish) are equally common. These are followed by Order Anguilliformes(almost entirely family Muraenidae, the moray eels), and families Balistidae (triggerfish), Carangidae(trevally etc), Nemipteridae(entirely Monotaxis gvanoculis), Lutjanidae(snapper etc), and Holocentridae (squirrel fishes). TABLE 5. MINIMUM NUMBERS OF FISH, ORGANISED INTO FAMILIES AND CK :RONOLOGICAL PERIODS Site Period Class Elasmobranchii Order Anguilliformes Belonidae Exocoetidae Holocentridae Sphyraenidae Acanthocybiidae Thunnidae/Katsuwonidae Carangidae Epinephelidae Lutjanidae Caesiodidae Nemipteridae Lethrinidae Mullidae Kyphosidae Coridae Scaridae Siganidae Acanthur idae Scorpaenidae Diodontidae Tetrodontidae Balistidae Total I II III 1 9 11 KA1 IV I % 2 10 12 21 3.06 5.36 9 17 1 1 1 1 2 KA2 III IV 1 2 5 9 £ % 6 5.83 11 10.68 KA3 II III IV 3 1 9 18 5 1 13 7 1 3 1 11 1 6.63 0.51 1 5 8 57 5 14 3.57 25 6.38 97 24.74 15 3.83 1 2 1 8 2 21 4 3 2.91 9 8.74 23 22.33 4 3.88 2 2 3 1 16 18 40 10 18 4 1 25 10 2 6.38 2.55 0.51 2 10 1 12 11.65 1 0.97 1 1 13 11 2 1 1 9 17 40 9 7 6 1 3 33 1 4 3 54 26 2 2 2 12 9 6 1.53 87 22.19 1 0.26 13 3.32 1 2 2 7 4 10 1.02 2.55 1 8 11 22 5.61 1 11 22 162 197 392 1 I 0.97 2 1.94 14 13.59 2 2 3.32 4 5 4.85 9 10 13 90 103 9.71 1 9 5 55 1 3 1 5 1 28 2 28 250 2 3 4 4 10 8 1 3 1 1 5 4 4 2 7 2 6 1 l % 18 4.63 39 10.03 5 1 16 1.29 0.26 4.11 23 5.91 29 .7.46 57 14.65 23 5.91 1 0.26 17 4.37 14 3.60 2 0.51 3 5 11 2.83 76 19.54 2 3 0.77 1 0.26 10 2.57 1 0.26 42 10.80 6 68 43 389 KA4 II III IV 10 30 4 12 2 5 2 6 13 9 35 18 14 5 1 8 26 3 6 11 197 £ % 16 4.85 47 14.24 2 0.61 6 19 5.76 3 2 8 4 1 0.30 21 6.36 15 4.55 54 16.36 22 6.67 2 3 1 1 3 3 18 10 1 1 19 5.76 10 3.03 1 0.30 2 0.61 14 4.24 54 16.36 1 0.30 4 1.21 7 1 5 4 11 3 2 1 4 2 4 7 2 19 76 57 330 2.12 0.61 5.76 Total 52 118 7 1 62 2 1 61 78 231 64 1 73 35 5 2 33 231 2 22 1 26 13 93 1214 I 00 I -59TABLE 6: FISH REMAINS ARRANGED IN ORDER OF DECREASING ABUNDANCE I Ill no. % no. 75 15.99 63 19.75 90 23.26 231 9 23.08 81 17.27 60 18.81 81 20.93 231 Anguilliformes 1 2.56 48 10.23 38 11.91 31 8.01 118 Balistidae 3 7.69 41 8.74 19 5.96 30 7.75 93 Carangidae 2 5.13 27 5.76 26 8.15 23 5.94 78 Nemipteridae 1 2.56 27 5.76 15 4.70 30 7.75 73 Lutjanidae 1 2.56 29 6.18 17 5.33 17 4.39 64 Holocentridae 9 23.08 34 7.25 9 2.82 10 2.58 62 Vevuod II no. % Epinephelidae 3 7.69 Scaridae no. % IV % Total Thunnidae/Katsuwonidae 2 5.13 29 6.18 19 5.96 11 0.26 61 Elasmobranchii 3 7.69 20 4.26 19 5.96 10 2.58 52 Lethrinidae 1 2.56 16 3.41 9 2.82 9 2.33 35 Coridae 1 2.56 13 2.77 8 2.51 11 2.84 33 Diodontidae 2 5.13 11 2.35 5 1.57 8 2.07 26 Acanthuridae 6 1.28 4 1.25 12 3.10 22 Tetrodontidae 2 0.43 2 0.63 9 2.33 13 5 1.07 1 0.26 7 2 0.43 Belonidae 1 2.56 Mullidae Sphyraenidae Kyphosidae 1 1 0.31 2 0.52 5 1 0.31 1 0.26 2 1 0.26 2 0.21 Siganidae 2 0.63 2 Scorpaenidae 1 0.21 1 Exocoetidae 1 0.21 1 Acanthocybiidae 1 0.31 1 Caesiodidae 1 0.31 1 Total 39 Shannon's H (bits) Volume Fish statistic (m3) density MN/m 3 469 319 387 1214 3.30 3. 65 3. 53 3. 46 5. 20 20. 20 19. 20 9.60 54.20 7. 50 23. 22- 16. 61 40. 31 22.40 -60Fish in these eight most common groups account for over 78% of the total. Of particular significance in addition to these eight are specimens of Thunnidae and Katsuwonidae, accounting for just over 5% of the entire catch(Minimum Number,MN=61). These are the main representatives of the pelagic game fish, primarily caught with lure hooks, trolled from canoes. Other fish which would probably have been caught in this way are Aoanthooybium (MN=1), family Sphyraenidae(MN=2), and Elegatis solandri (MN=5). Taken together with the tuna families, bipinnulatus this type of fish represents 5.68% of the total catch - a small proportion compared with the great importance attached to these fish today on Kapingamarangi. This feature is returned to later. Three overall observations should be made at this stage. Firstly, it will be seen from Table 6 that the general abundance of fish in the archaeological sites increases more than five-fold during the sequence(7.5, 23.2, 16.6, 40.3 MN/m 3 ). Even taking into account the fact that the sediments forming Touhou accumulated at different rates, this dramatic increase must partly reflect increasing population size on the islet. Secondly, there are some significant changes through time in the relative abundance of different types of fish. For instance,the clearest trend is in the cod and groper family ( Epinephelidae) , which increases some threefold in its relative importance(7.7,16.0,19.8,23.3% of the catches at each period). By contrast, the parrotfish(Scaridae) are by far the most numerous early in the sequence(23.1%), and their importance falls somewhat through time, though not uniformly. Moray eels are not important early in the sequence, but increase in popularity. Thirdly, there could be an' areal factor relating to site function, which complicates the task of interpreting changes through time. From Table 5 it will be seen that the main fish types are not uniformly abundant from one site to another. A number of different reasons could be advanced for the observed disuniformity. Different families would not necessarily have equal access to the different fishing areas in the atoll, nor would all families have fishermen with equal skills in all kinds of fishing methods. One therefore has to be careful not to interpret too closely the trends observed, because this factor is difficult to disentangle. Pooling results helps to minimise this effect. As for catching methods, the main fish types would have been caught by rather different means. The cod and groper family, although caught in the lagoon, are normally in somewhat deeper water, and would have been primarily caught with baited 'hook. The parrotfish, on the other hand, can be taken in quite shallow water around coral thickets, but since their diet is chiefly vegetable matter, invertebrates, molluscs, and coral, they are only rarely caught on a baited hook. These fish can be speared of course, but above all they must have been taken with various kinds of nets. The eels represented could have been caught by a number of different methods too, but given the elaborate trapping methods used by Kapingamarangi -61fishermen in the historic era for catching eels, this interpretation is preferred. In the time trends observed above, therefore, we should also read into the patterns some changes in fishing technology. In particular, netting seems to have dominated the early part of the sequence, and baited hook techniques appear to increase progressively in importance. Trap fishing seems initially fairly rare, but also gains in popularity. The pattern of relative abundance of the different kinds of fish was assessed using Shannon's H statistic(Table 6 ) , and the results are fairly uniform for the different periods. By comparison with the statistics for shellfish(see below), the values obtained indicate a more even emphasis on a greater number of target species. Shellfish The shellfish recovered in the excavations were identified to species level wherever possible, and minimum numbers calculated. The details of this process are given in Appendix 5. More than 70 species were recovered, but of these only a few can be considered to have had any real economic significance. A summary of the shellfish remains is given in Table 7. It will be readily apparent from this that all but about 13% can be accounted for by shells belonging to three families. By far the most common shellfish is Turbo argyrostomus (nearly 68%), followed by Donax sp.(14.5%), and Cypraea spp.(6.9%). Only 59 specimens of Tridaona cf maxima were found, and only eight of Hippopus hippopus. It is very likely that the bulk of these latter two shells were on site for industrial reasons, rather than as food. Today, the esteemed food of Tridaona maxima is gathered on the reef, and the shells discarded on the spot as the meat is separated. Although the modern inhabitants have the benefit of steel knives for this purpose, this pattern of exploitation must have prevailed in antiquity too. It is quite likely that during the prehistoric period, this species constituted the most important component of the shellfish diet - this is not reflected in this Table of results. Hippopus hippopus, on the other hand, is somewhat more difficult to deflesh without a fire, and is more likely to have been brought home if it was eaten at all. Their small number in the sites probably does reflect their unimportance in the diet, and they are regarded with distaste today. As pointed out in Appendix 5, most of the specimens of Cypraea spp. were very small, and few can be considered to have been items of food. Specimens of Donax sp. too, could easily have been imported in the kilikili gravel; again, their economic significance is questionable. Finally, shells of Turbo sp. are a favourite home for hermit crabs, and some specimens must have been imported by them from the surrounding beaches. Even so, concentrations of these shells in the sites strongly suggest that the bulk of these were indeed collected by people for food. It is interesting that neither Nerita spp. nor Donax sp. figure prominently in the diet TABLE 7: MINIMUM NUMBERS OF SHELLFISH ARRANGED IN FAMILIES AND CHRONOLOGICAL PERIODS Period Site Trochidae Turbinidae Neritidae Cerithiidae Strombidae Cypraeidae Naticidae Cassidae Cymatiidae Bursidae Muricidae Magilidae Buccinidae Vasidae Harpidae Mitridae Conidae Terebridae Ellobiidae Arcidae Pinnidae Pteriidae Pectinidae Spondylidae. Carditidae Cardiidae Tridacnidae Mesodesmatidae Donacidae Tellinidae Psammobiidae Trapeziidae Veneridae IV KA1 KA2 KA3 KA4 2 22 401 3 49 2 4 21 4 262 1 1 1 1 1 E 2 56 5 1 1 68 547 57 2 26 267 1 1 1 4 4 % III KA1 KA2 KA3 KA4 0.14 37.19 301 3.87 6 0.14 1 1.77 15 18.15 4 0.07 0.07 0.07 5 5 0.34 0.34 6 17 5 0.41 1.16 0.34 1 1 1 2 1 2 1 1 1 6 1 1 5 7 3 1 6 4 5 4 7 1.43 0.07 0.14 3 2 (bits) 4 12 1 11 18 0.75 1.22 458 1 2 1 8 31.14 0.07 0.14 0.07 0.54 1471 2. 36 17 5 1 0.41 6 8 450 1 2 1 8 Totals Shannon's H statistic 21 1 2 % II KAl KA3 KA4 I E % KAl KA3 E % 1 1 0.07 1 11 0.37 1 4 0.64 10 3 52 139 559 1051 73.81 181 860 1147 2188 73.65 12 472 484 77.07 5 21 1.47 4 40 32 8 2 76 2.56 2 13 15 2.39 1 0.07 3 1 1 0.16 3 6 0.20 98 3.30 1 0.16 16 100 3 134 9.41 19 79 1 13 12 38 2.67 10 38 59 107 3.57 2 32 34 5.41 9 4 1 5 0.17 1 1 0.16 1 6 17 5 E 3 1 2 4 1 4 5 2 25 56 7 1 1 1 1 1 2 1 0.07 0.07 0.07 0.14 0.07 1 2 1 1 1 34 1 1 0.07 0.14 0.07 0.07 0.07 2.39 0.07 0.07 1 2 1 2 2 1 1 1 1 3 3 14 3 10 1 14 5 2 1 3 1 7 6 3 3 1 14 13 0.21 0.07 0.98 0.91 9 97 6.81 1 0.07 6 1 13 1 1 34 199 1 1 12 1 0.07 5 1424 1. 55 2 . 4 1 2 14 95 1 5 4 5 0.17 1 0.03 2 0.07 3 0.10 1 1 4 27 3 0.03 0.03 0.13 0.91 0.10 20 0.67 5 0.17 1 0.03 12 0.40 1 0.03 8 0.27 28 0.94 2 0.07 328 11.04 2 0.07 18 0.61 9 0.30 2972 1. 61 2 2 1 2 2 0.32 1 6 1 6 0.16 0.96 7 7 1.11 1 6 1 8 0.16 1.27 55 1 3 57 1 4 9.08 0.16 0.64 1 1 0.16 628 1. 40 Total % 18 0.28 4270 65.75 169 2.60 10 0.15 259 3.99 446 6.85 7 0.11 1 0.02 9 0.14 2 0.03 8 0.12 10 0.15 1 0.02 1 0.02 2 0.03 i 13 0.20 £ 51 0.79 l 9 0.14 1 0.02 75 1.15 2 0.03 8 0.12 1 0.02 28 0.43 2 0.03 34 0.52 67 1.03 2 0.03 940 14.47 4 0.06 25 0.38 1 0.02 19 0.29 6495 -63at any time during the prehistoric period. This is a striking contrast to archaeological remains in some sites in the Solomon Islands, where these shells appear to increase in importance, as larger species become locally depleted through exploitation(see Leach and Davidson,n.d.). Despite the fact that only a few species are economically important, or perhaps because of it, there are some interesting patterns which can be observed through time. The first feature which can be noted is that the general pattern of shellfish exploitation is quite different in the last 100 years(Period IV), than anything in the previous centuries. This is most clearly seen by the values of Shannon's H statistic for the four periods (see Table 7 ) . In ecological studies, this is used to measure species diversity against environmental stress(Wilhm,1968). The value of H=2.36 for Period IV shows that the pattern of diversity is markedly less than in the prehistoric period, where an average of about H=l.5 prevails. This is largely due to near equal quantities of Turbo sp. and Donax sp.(37.2 and 31.1% respectively), and high proportions of Cypraea spp. also(18.2%). In earlier periods, by contrast, Turbo sp. occupies a position of far greater importance, and the scree of percentages from the level for this species falls off much more steeply. In economic terms this implies that although the prehistoric people collected a wide range of species, their prime targets were few in number, and illustrates a highly conservative approach to shellfish gathering. As pointed out above, the pattern of exploitation of fish, on the other hand, is quite different, and shows greater enterprise. Leaving aside this apparently major shift in emphasis in the last 100 years, some more subtle trends can be observed in the earlier levels. The figures for Turbo sp. decrease somewhat (77.1 to 73.6 to 73.8%), as do those for Cypraea spp.(5.4 to 3.6 to 2.7%). The main contrasting shellfish making up the difference are specimens of the family Strombidae which rise in frequency (0.2 to 3.3 to 9.4%). This might suggest that two of the main target types of shellfish were becoming harder to obtain in the course of time. A total of 6494 shellfish in 54.2 m 3 of excavated deposits is not a high density in comparison with some occupation layers in other Pacific islands, and with the possible exception of Tridacna maxima, shellfish do not seem to have been of major significance in the early diet on Kapingamarangi. Crustaceans The analysis of crustaceans is given in Appendix 4. Only a few fragments were recovered in the excavations, although conditions are considered favourable for their preservation (see Leach and Anderson,1979 b) .This is taken to be good evidence that crustaceans were not very important as an item of prehistoric diet. '-64Crayfish are represented by only two mandibles in Level II, and coconut crab is only found in Level IV. Unfortunately, very little is known of the spread of the coconut crab through the Pacific, but this result should not necessarily be taken to indicate their absence on Kapingamarangi prior to the last 100 years. The exoskeleton of this animal appears to be richer in protein than many other crustaceans, and their archaeological survival is therefore much less certain. The lack of crayfish though is rather puzzling; however, neither of the two species found on Kapingamarangi today figures prominently in the modern diet either. Birds. About 70 fragments of bird bones were recovered in the excavations - a very small number compared with some Pacific island sites. Sufficient of these could be identified to show that the frigate bird, red-footed booby, and noddy tern were definitely items of prehistoric diet(Table 8 ) . These all appear early in the sequence in Level II. The chicken is certainly present by Level IV. Bones of Gallus sp. were recovered in close association with the human skeletal material described in. Appendix 10. The age of this material is unknown, and the chicken bones were unfortunately lost in transit. The human bones are thought to be definitely prehistoric, and it is possible the chicken bones were too - the tarsometatarsus had a strongly developed spur, characteristic of the prehistoric chicken, and not so frequently present with modern chickens. sp. have been identified from Nukuoro. No bones of Gallus TABLE 8: BIRD MINIMUM NUMBERS FROM THE EXCAVATIONS Period Site Bird iv III II i KA1 KA2 KA3 KA4 KA1 KA2 KA3 KA4 KAl KA3 KA4 KAl KA3 ? sp. 1 Gallus gallus Fregata minor 1 1 7 3 3 subsp. 6 1 An&us stolidus piliatus Sula sula rubripes 1 2 1 -65Pig, Rat, Turtle, and Human The minimum numbers of these are listed in Table 9. With the exception of rat, the bones of these animals were relatively rare. It is notable that not one fragment of dog was found. During the excavation at Putau(KA3) a pig's tooth was tentatively identified deep in Level II in a prehistoric context. This important find unfortunately was lost in the freight, and cannot now be confirmed. It is possible that the tooth was actually from a specimen of Pseudobalistes flavimavginatus. This fish grows to a large size on Kapingamarangi, and some of the teeth have a superficial resemblance to pig incisors. All remains of rat were certainly of Rattus exulans, and are quite frequent relatively early in the sequence. It is notable that none were found in Level I. This reinforces the view that Touhou islet at this time was barely above the sea, unsuitable for permanent human habitation, and had not been colonised by rat at this time. The general rarity of turtle bones is interesting, and probably shows that the sea grass flats of Zostera sp. around Kapingamarangi were never more extensive than they are today. Only one human tooth was found in the excavations. It is unusual not to find small pieces of human bone and teeth in Polynesian archaeological sites. TABLE 9: MINIMUM NUMBERS OF TURTLE AND VARIOUS MAMMALS IN THE EXCAVATIONS Period Site Sus sarofa IV KAl KA2 KA KA3 KA4 KA1 1 Turtle 2 Homo sapiens Rattus exulans Mammal 3 ? sp. * = 1 tooth III KAl KA2 KA3 KA4 II KAl KA3 KA4 I KAl KA3 Total 1 1 5 1 1 4 1 10 1* 2 3 2 3 1 7 1 4 1 6 17 20 1 64 1 4 -66Stone and Pumice Some stone material was recovered, but it is not thought that all was retrieved by any means. Without washing on the spot, it can be very difficult to distinguish between dense burnt coral and stone. When the lowest levels of Putau(KA3) were excavated, it was well below sea level, and buckets containing a slurry were wet sieved in the sea. This revealed some most interesting foreign rocks(see Appendix 9 ) . In addition, pumice was seen at various levels in all sites, and although samples were taken, this was not systematic. On the whole, pumice seemed to occur as concentrations at different levels, and may well reflect the periodic arrival of significant quantities on beaches, corresponding with activity from a distant volcano. Of course, this pumice is not now in primary deposition, since it was imported onto Touhou with the gravel material while building up the islet. Its historical significance would therefore be very difficult to establish. The rocks in the sites were almost certainly derived from the roots of drifting trees, which still regularly arrive on the island. In fact, several of the stones described in Appendix 9 were taken from driftwood during the expedition. Ultimately it may prove possible to identify the sources of these stones, and this may be useful information to assess the possible areas from which drift voyages took place to Kapingamarangi during the prehistoric period. The occurrence of such drifted stone should also serve to introduce a note of caution in interpreting possible exchange networks of such things as oven stones in Pacific archaeology. HUMAN SKELETAL MATERIAL Human bone, representing six individuals,was collected by Lieber from the foreshore after it eroded from the section immediately to the north of the excavation at Putau(KA3) following a storm and exceptional tides. Unfortunately the remains are not directly associated with archaeolgical sediments, and their age is therefore uncertain. However, they are certainly of some antiquity, as burial plots in this vicinity are not known by the modern inhabitants. It is fairly safe to assume therefore that they are prehistoric in age. The material has been studied by Houghton(see Appendix 10), who clearly identifies the people as Polynesian in their biological relationships. There is no sign of yaws, and the people appear to have been well nourished. -67ARTEFACTS Shell Adzes A total of 23 shell adzes were found. Only 12 of these were recovered in controlled excavations, the remainder were surface finds(see Figures 37 and 38). The shell species cannot be accurately determined for any, although Tridaana cf.maxima is probable for all. Some of the adzes are quite small in size, although there is no chronological significance which can be attached to this. Specimens of both large and small(Figure 37: #B, #J) were found in Level I at Putau(KA3), showing the presence of both forms early in the sequence. One adze is unusual(Figure 37:#H) in having a definite waisted appearance, otherwise these adzes are perfectly ordinary in the general context of island cultures anywhere in the Pacific. It is possible that the waisted specimen is a coconut grating head(see below). During the Bishop Museum expedition, 2 9 shell adzes were recovered(Buck,1950:165); these were all of Tridaana shell, and most were the common ventral lip form. A few are of the thicker variety made from the solid interior, as is probably a specimen figured here(Figure 37,#L). One other borderline example was also found, but is not figured. It is interesting that Buck also collected a Cassis sp. chisel, but could not obtain information as to its use(Buck,1950:172). This type of tool is absent on Nukuoro(Davidson,1971:68). The single most important feature of the adze assemblage is the total lack of adzes made from Mitra sp. or Terebra sp. It is possible that this adze form was manufactured in the past, and that the excavations simply failed to produce any examples. One point in favour of this view is that an old informant when questioned about the possibility of such adzes, proffered the name toki poko for those made from Mitra sp.( pu manga roro) . Lieber has identified this term as that applied to gouging adzes(Lieber and Dikepa, 1974), and this strengthens the case. However, it is also clear from discussions with old men that considerable influence from Nukuoro has been felt in adze technology, particularly in methods of hafting, during the historic era; it is therefore possible that this is where the knowledge of this adze form derives. On Nukuoro, adzes of either Terebra sp. or of Mitra sp. are quite numerous(Davidson,1971:52). A comparison between the two assemblages is given in Table 10. One is forced to the conclusion that if adzes of Mitra/Terebra were manufactured in the past on Kapingamarangi, they were extremely rare. Suitable large specimens of these shells are quite common on the island. This matter is of some potential importance, since one of the few aspects, of culture-historical significance which has been found amongst the different types of shell adze in the Pacific, is that this form of adze developed relatively late in time. In Taumako, in the Solomon islands, for example, an antiquity of no more than 1000 years is suspected. The relative abundance of this adze type on -68- i i i — i i i cm F i g u r e 37: Tvtdaona s h e l l adzes from Kapingamarangi. The provenances a r e as f o l l o w s : KAl:A(Spit 3 ) , G ( S p i t 8);KA2:D & E(Layer 5-6);KA3:B(Spit 2 1 ) , C ( S p i t 9 ) , J ( S p i t 20),M(Spit 7 ) ; KA4:I(Spit l l ) , K ( S p i t 1 2 ) , N ( S p i t 8 ) , 0 ( S p i t 6 ) ; Surface f i n d s : F & H ( P u t a u ) , L(Werua). -69- Figure 38: Surface collected Tridaana shell adzes from Kapingamarangi. TABLE 10: COMPARISON OF ADZE FORMS FROM NUKUORO AND KAPINGAMARANGI (see Davidson, 1971: 52) Nukuoro Terebra Mitra Cassis maoulata sp. Kapingamarangi 69 2 sp. 5 1 Tridacna maxima (ventral lip) 68 23 Tridacna maxima (solid interior) 21 Total 165 24 Note: The Cassis sp. example from Kapingamarangi is a coconut grater, as perhaps the Nukuoro specimens are too. Buck (1950: 17 2) records the Cassis sp. chisel on Kapingamarangi, but it is not known on Nukuoro. -70Nukuoro, and their presence in the earliest levels here suggests that the island was settled from some area where the form had already been developed. In Kapingamarangi, on the other hand, their apparent absence through the sequence might indicate an origin for the founding group from some area where the adze was not known, or not yet developed. The striking difference in character between the adze assemblages from Kapingamarangi and its nearest neighbour Nukuoro is compelling evidence of cultural isolation through the prehistoric period, and perhaps indicates different origins for the two founding groups. One-Piece Fish Hooks The archaeological sites were practically bereft of evidence of fish hook manufacture, considering the quantity of material excavated. The few examples recovered are illustrated in Figure 39, and the debitage from manufacture in Figure 44. There are some interesting features of this small collection. Of the three definitely prehistoric hooks found, two are made of bone(possibly turtle), and the other is pearl shell. The three are bracketed by two radiocarbon dates(NZ5145 and NZ5146) to between 350 and 550 years BP(See Figure 33). Most of the few fragments of worked pearl shell were found in the same provenance (Figure 44). The authors saw more fragments of worked pearl shell scattered over the surface of Nukuoro in a few hours, than in the sustained archaeological work on Kapingamarangi. This dearth of shell fish hook material in the sites unquestionably reflects the extreme rarity of Pinotada spp., and of Atvina vexillum in the Kapingamarangi lagoon. Clearly, some hooks were made from pearl shell when it was available, but if fish hooks were common in the past they must have been made from some less durable material which has not survived archaeologically. The large number of fish of the family Epinephelidae in the sites could really only have been taken with baited hooks - most must have been made from wood, turtle shell, bone or coconut shell. The information given by Eilers on fish hooks(1934:73ff) is instructive in this respect(Buck's comments[1950:231] are largely a repetition from Eilers). Small hooks are said to have been made from coconut shell, turtle shell, pearl shell, and black mussel. Large hooks were said to have been made from wood, pearl shell, and black mussel. Of these, in our opinion, coconut shell must have been especially common; and this view is supported by the illustration in Eilers1 work of 16 such hooks on a special hook holder (1934:74; see also Buck,1950:240). This view is also held by Anell(1955:95), although apart from a few more recently collected items, again he largely relies on Eilers' earlier observations. In passing, an error should be noted in Anell. Specimen #6 in his Figure 4 from Kapingamarangi, is said to be of pearl shell, and specimen #7 of coconut shell, following Eilers (1934:94) . In fact, the specimens are cited(Eilers,1934:75,76) as being made of pearl and black mussel respectively. Beasley(1928:101) is able to add very little to the subject, as his observations for Kapingamarangi depend on a single and rather crude wooden hook collected by Commander Erskine about 1850. Turtle bone fish hooks are relatively rare in the Pacific islands(some are t • i i i i ic m Figure 39: One-piece fish hooks from Kapingamarangi. They are made from bone (B,E), metal(F,G), Pinatada sp.(A,c,D,I), and Atrina vexillum(U). Both metal hooks are from KA2(Layer 4), the two bone hooks are from KA4(B from Spit 10, and E from Spit 13). Specimen A is from KA4(Spit 10),D was found while digging a grave on Werua,C and I were surface finds from the eroding section at Putau(KA3), and H is a modern handicraft item, made in 1980. •-72known from Funafuti[Hedley,1897:266]), and again their presence on Kapingamarangi attests the difficulty of finding suitable material for hook manufacture, as turtle bone is not very strong. Three types of hook are evident in the small collection. Firstly, the plain rotating hook with slightly incurved point is represented by specimens A,B,C, and E(Figure 39). The method of line attachement obviously varied, as one specimen is notched externally on the upper shank, another is plain, and yet another has a slight inner projection. Secondly, specimen D has a short pointed shank, and a markedly incurved point. The example is notched at the base of the shank and has several grooves along the lower margin. These features are closely in line with the method of snooding illustrated by Eilers (1934:75, specimen #22). This hook was found by an old man while digging a grave on Werua, and he was unequivocal in his name for the hook - matau kina - the hook for the rainbow runner, Elegatis bipinnulatus. A hook type with an equally incurved point from Nukuoro(Type VII) is claimed to have been used for catching an important fish called gina (Davidson,1967:187), which may well be the same species.Thirdly, specimen I is unusual in possessing a snooding knob on the inner proximal end of the shank. This is almost a perfect match of the hook illustrated by Eilers(1934:76,#27), and copied by Anell(1955:93,#6). It has a fairly fresh appearance, and is not thought to be very ancient. It is a rather crude shape, with only bare attention given to final finishing. In our opinion it is unlikely to have been useful for actually catching fish. It is quite likely to have been a handicraft item to satisfy the voracious collecting habits of Europeans at any time in the last century. Eilers comments that souvenir manufacture was firmly established on Nukuoro by 1910(1934:229). It would be wise to wait until this particular hook form was found in a firm prehistoric context before commenting on possible parallels with inner shank knobbed hooks from elsewhere in the Caroline islands(see Anell,1955:92). The Kapingamarangi people are very quick to learn new shapes from books, which they then turn out as handicraft items for sale to modern collectors. It should be noted that Buck showed the fish hook illustrations in Eilers to people in 1947(Buck, 1950:232), and they showed him how to reproduce the lashing from these illustrations. It would not be at all surprising if specimen I illustrated here was not a direct copy of the one illustrated by Eilers. Where the original idea may have come from is open to conjecture. The final stages of this process of morphological degeneration in the face of continuing European demand for curios stripped of their functional significance is evident from specimen H. This was made with steel tools by an old man in 1980 from Atrina vexillum, and has fibreglass cordage. Finally, two metal hooks were recovered in the Muri-Harau excavation. One of these is shaped in typical Polynesian style, and is made of strong brass or bronze; the other is made of softer coated brass, and is quite a crude shape. -73Trolling Lure Shanks Shanks for trolling lures are stylistically an important artefact from Kapingamarangi. Four specimens are illustrated in Figure 40. Three are of dense white shell, possibly Tvidacna sp., but it could be Spondylus vavians (specimens A,B,C). The fourth is of Pinatada sp.(specimen D ) . These shanks are very finely made, and are distinctive in having a groove running nearly the full length of the dorsal surface of the shank, although in the case of the pearl shell example this groove is much shorter. Portions of two other pearl shell lure shanks were also found. One came from the gravel lens towards the base of Layer 4 in the Putau site(KA3, see Figure 30), and the other was found in the eroding section north of Putau(KA3). These were i i i 1 i 1 i crn Figure 40:Shell lure shanks from Kapingamarangi.A:found while digging a grave in the Ngeiho-Hereu area(KA4). B,C:eroded from scarp at Putau area(KA3). D:KAl,Spit 15, next to radiocarbon sample NZ5135.A11 have the characteristic groove along the dorsal surface, unique to Kapingamarangi. -74unfortunately lost in the freight transport. Both also had relatively short grooves, and were comparable to specimen D. Specimens B and C are a rather rounded triangular shape in cross section, but #A is distinctly sharp-edged with a sub-triangular section. Buck saw one similar shank during his visit to the island(1950:237), and another is in the Bishop Museum(C.10170). In addition, Buck regarded a set of 9 very similar items which he recovered as a necklace(1950:273ff). These do not have the characteristic dorsal groove, although they appear identical in all other respects. Eilers illustrates a pearl shell lure (1934:77), and vividly describes the dorsal grooving, likening the distal end to a fishtail. The barbed point is made of turtle shell. The crude untrimmed nature of the proximal end of this specimen led Buck to comment on its Marshallese affinities(1950: 238); something which some authors believe was quite deliberate on the part of the artisans, as it is thought to enhance the fish luring qualities(see Beasley,1928:95;Anell,1955:152). It is possible that this feature was introduced to Kapingamarangi during the Marshallese invasion of about 1870. As for the grooved feature on these Kapingamarangi lures, no cultural parallel can be suggested. The pearl shell versions, although not fully developed in style, establish the form as thoroughly diagnostic of Kapingamarangi, with a respectable antiquity of about 700 years. The lure shanks recovered archaeologically on Nukuoro, are of more widely recognised types. Coconut Grater Heads Three specimens of these are illustrated in Figure 41. The one made from Cassis sp is bracketed by two radiocarbon samples (NZ5141 and NZ5140), and has an age of about 600 years. The second example(B) is also of shell, but the species is obscure. It was possibly made from a juvenile specimen of Spondylus vavians. This one is closely associated with another radiocarbon sample (NZ5145), and has an age of about 350 years. Finally, the third less certain example(C) is made from Atrina cf, vexillum , and is a surface find. It is interesting that no clear evidence for bivalve coconut hand scrapers was seen. This is surprising in view of their very common occurrence in islands to the south, V A-A_>}|_£_A_/i. _* J Figure 41: Shell coconut grater heads from Kapingamarangi. A-.Cassis sp. from KA3(Spit 14). B: from KA4(Spit 9). C: Atrina cf,vexillum surface find near Putau(KA3). -75and presence on Nukuoro. Non-bivalve coconut grater heads on Nukuoro, although common, appear to come into the archaeological record relatively late, possibly no earlier than 300 years (Davidson,1971:72); they are at least twice this age on Kapingamarangi. Buck collected a number of adze-like objects made of Tridacna sp. which old men recognised as coconut grater heads for stools. The main distinguishing feature is the constricted sides for lashing onto the stool. It is possible that one of the adzes illustrated here(Figure 37,#H) really functioned in this manner. In addition, Eilers illustrates a stool(1934:122) with a bivalve shell attached, identified as Cardium sp. Pottery A single sherd of what is thought to be prehistoric Oceanic pottery was found on the surface of the Tiroki site(KAl) immediately before excavation. This is the subject of three technical reports(Appendices 6,7, and 8) which examine its physical and mineralogical properties. The sherd is illustrated in Figure 42(#E). It is difficult to arrive at a hard and fast conclusion about this sherd. One possibility which was considered was that it was a fragment of a Japanese field tile(dokan). These are actually quite common in Micronesia, but should be readily distinguished from their curvature and glazing on both surfaces. To be thoroughly certain, however, samples of these tiles from Kapingamarangi were subjected to mineralogical investigation(Appendix 7 ) . The sherd is quite different in composition to these tiles. The mineralogy of Spanish pottery from Pacific contexts(see Leach and Davidson,n.d.) is again quite different. In general physical appearance it is quite similar to prehistoric pottery to the south and east, but it is exceptionally hard. Since it was found on an atoll, we can be quite certain it was imported; and the fact that it was not found deeply buried in an archaeological site does not rule out the possibility that it arrived with very early visitors to the island, perhaps with the founding group. Of course it may also have been imported during the historic era. Indirect contact with populations to the south was reasonably frequent during the period of German administration(1899-1914), and the sherd does resemble the well-fired modern trade pottery from the Siassi area of New Britain, for example. On the other hand, the mineralogy is consistent with(but not proof of) a derivation from the Solomons area. Both in physical properties, and mineralogy, the sherd is quite distinctive from known Micronesian potteries. I I I I ' I cms Figure 42: Ceramic and metal a r t e f a c t s from Kapingamarangi. Item E i s a s u r f a c e c o l l e c t e d pot sherd from t h e T i r o k i s i t e ( K A l ) . The remainder a r e from t h e Japanese s e t t l e m e n t a r e a s on Nunakita and P a r a k a h i . Aicopper o b j e c t . B:Tin bronze i n c e n s e burner.C:50mm US Browning a i r c r a f t machine gun c a s e . D: Lead o b j e c t . E: Pot s h e r d . F , G , H , I : Japanese g i a z e d f i e l d t i l e s ( d o k a n ) . Shell Arm Rings Fragments of three b r a c e l e t s or arm rings were found(see Figure 43). One of these i s made from a large specimen of Conus sp and i s similar to an example i l l u s t r a t e d by Buck(1950:274), but much l a r g e r . Buck's specimen has an outside diameter of 4 3mm, while the archaeological one i s nearer 70mm. Buck regards t h i s smaller specimen as an ear ornament, but the l a r g e r one would f i t over a small hand. I t i s about 360 years in age. The other two rings are most i n t e r e s t i n g . They were both found in the basal c u l t u r a l layer of Putau(KA3) below sea l e v e l , and in the t i d a l sea water(see Figure 30). This context i s believed to be amongst the e a r l i e s t c u l t u r a l material on t h i s i s l e t , g r e a t e r than 700 years BP. Both specimens are t r i a n g u l a r in c r o s s - s e c t i o n , and are d e l i c a t e l y fashioned from a white s h e l l . This i s probably Tridaona s p . , but could also be Spondylus vavians . One similar type of ring was found on Nukuoro in an early context(NU5,Level 16). This was also t r i a n g u l a r in cross s e c t i o n , and probably of Tridaona sp. The remaining Nukuoro examples are mainly of a b i v a l v e , such as Codakia s p . , although one i s thought to be Troohus sp. Thus, the t r i a n g u l a r - s e c t i o n e d Tridaona sp. arm ring continues to show i t s e l f to be associated with archaic c u l t u r e s in the Pacific i s l a n d s . The d e t a i l s of i t s Oceanic development and spread have yet to be demonstrated. -77- i i i i i i . i cm Figure 43: Shell arm rings from Kapingamarangi. A: Tri-daena sp. 8cm inside diameter from KA3(Spit 21). B: Tridaona sp. 7cm inside diameter from KA3 (Spit 20). C: Conus sp. 6cm inside diameter from KA4(Spit 9). -78Perforated Shells and Shell Beads A considerable number of shells of Nerita sp. were perforated; however few if any of these are thought to be cultural. Some have natural breaks, and many are drilled out neatly, but are probably the boreholes of Muricid spp. One or two holes may have been made by grinding on a flat surface such as a coral slab, and these do suggest human manufacture. The distinction between human and natural performations is difficult to be certain about. In Taumako in the Solomon Islands, most of the shell necklace units had been punched out to form the hole. The reason why they are certain to have been cultural was because they were found in situ around human burials. It is interesting that although shells of Polinices spp. were not uncommon in the sites excavated, none were perforated. These shells are a favourite for necklaces in islands to the south, and also in the Kiripati islands. Eilers illustrates a necklace string of perforated Melampus luteus from Kapingamarangi(Eilers,1934:120), and drilled shells are an important aspect of the modern handicraft industry on the island. Only two shell beads were found, and both are from Ngeiho-Hereu(Spit 13) dating to about 500 years BP. They cf. ducalis show flashes of orange-red, characteristic of Spondylus (see Figure 44). Such beads are widespread in the north Pacific, including early Nukuoro(Davidson,1971:48), and they also occur in early and late contexts in various parts of Melanesia. Drilled Sharks Teeth Two small shark's teeth which had been neatly drilled were found in the Putau site(KA3, Spits 8 and 10). The age here is between 300 and 5 00 years BP. Spears with shark's teeth were described to Hambruch during his visit in 1910, but none were actually observed(Eilers,1934:120). It is quite possible that these were the spears with shark's teeth carried by the Marshallese castaways of 1870, described in vivid detail to Hambruch(Eilers,1934:130). Commonly, the presence of drilled shark's teeth units is taken as an indication of weaponry, but they have a variety of potential uses on small islands in the Pacific. For instance, individual teeth were attached to small sticks and used for tattooing and lancing on Funafuti (Hedley,1897:300), and Vaitupu(Kennedy,1931:Plate 19). Both Buck(1950:278) and Eilers(1934:118), however, are adamant that tattooing was never practiced on Kapingamarangi. Despite this, tattooing is understood by Thilenius(1903:38) to have been introduced to Ontong Java from Kapingamarangi. Similar units were recovered in the excavations on Nukuoro, where incidentally tattooing was part of the culture, and Davidson prefers a more mundane interpretation than weaponry also (1971:79) . -79- \ i 1 i i i i cm F i g u r e 44: F i l e s , s h e l l b e a d s , and i n d u s t r i a l p e a r l s h e l l from Kapingamarangi. The f i l e s a r e made from sea u r c h i n spines.D:KA1(Spit 8 ) . A and B:KA3(Spit 1 2 ) . C:KA3(Spit 8 ) . The s h e l l beads a r e both of Spondylus cf. ducalis, and both a r e from KA4(Spit 1 3 ) . The remaining items a r e p i e c e s of i n d u s t r i a l p e a r l s h e l l . J and K : s u r f a c e f i n d s worked w i t h s t e e l t o o l s f o r modern h a n d i c r a f t s incliiding p s e u d o - t r a d i t i o n a l f i s h hooks. The p e a r l s h e l l was probably imported from Nukuoro. Items G-I and L-N a r e p r e h i s t o r i c . K A l : I ( S p i t 2 ) , L ( S p i t 1 6 ) . KA2:N(Layer 4 ) . KA4:H(Spit 1 5 ) , G ( S p i t 5),M(Spit 1 3 ) . Sea Urchin Spine F i l e s Four well-used f i l e s made from s l a t e pencil sea urchin spines were recovered(see Figure 44). Three were found in the Putau(KA3) s i t e dating to about 500 to 600 years BP, and the fourth a t Tiroki(KAl) dating to about 200 years BP. The l a t t e r i s hollow ground, and r a t h e r c h i s e l - l i k e ( # D ) . The other three are q u i t e d i s t i n c t in being wedge-shaped, and these must have functioned as abrasive implements for smoothing a f l a t surface. Item #B has a wear facet on the very t i p , and i s not a reamer, as might be thought from the i l l u s t r a t i o n . The reduced diameter shown i s e n t i r e l y n a t u r a l . These f i l e s can not be i n t e r p r e t e d -80- as those for manufacturing one-piece fish hooks , where the reaming operation necessary gives quite a different wear pattern. Instead, they are most suited to finishing off some delicate flat surface. A possibility is the flat areas on trolling lure shanks, or arm rings. No comparable specimens have been found on Nukuoro, where abrasive tools appear to have been made from coral instead. Pumice Abraders Among the numerous small pieces of pumice recovered, quite a few show evidence of having been used for grinding various items. These are especially common in the Putau site(KA3) from Spits 11 to 16, and Ngeiho-Hereu(KA4) in Spit 6. Similar finds were also made on Nukuoro(Davidson,1971:71). Worked Pearl Shell Fragments of industrial pearl shell were extremely rare in the excavations. Only six definite pieces of Pinatada sp. were recovered, in addition to the few artefact pieces. These are illustrated in Figure 44. © eO cm Figure 46: Glass beads from the excavation at Muri-Harau(KA2). They are from Layer 3(B), Layer 4 (A ,C,D,E,F) , and Layer 5(G). Some are dark blue(A,B), some light blue(C,E,F,G), and one yellow(D). All are rather squat in crosssection, with the exception of #D, which is a long narrow bead. -81Japanese and European Artefacts P r a c t i c a l l y no modern c u l t u r a l material was recovered from the surface layers of e i t h e r Tiroki(KAl), Putau(KA3), or Ngeiho-Hereu(KA4); but such items were very common in the excavation a t Muri-Harau(KA2). A s e l e c t i o n of items i s i l l u s t r a t e d in Figures 45,46, and 47. The discrepancy between these occurrences i s r a t h e r i n t e r e s t i n g , and the absence from deeper l a y e r s even more so - i t serves to re-confirm a somewhat b a t t e r e d f a i t h in s t r a t i g r a p h i c i n t e g r i t y for a Pacific a r c h a e o l o g i s t . So commonly, Pacific archaeological s i t e s have been severely and continuously disturbed by land crabs and human digging and re-deposition a c t i v i t i e s t h a t close s t r a t i g r a p h i c i n t e r p r e t a t i o n s can be q u i t e meaningless. The absence of obvious c u l t u r a l mixing in the case of European a r t e f a c t s on Kapingamarangi, shows t h a t t h i s problem i s not a s i g n i f i c a n t factor on t h i s i s l a n d . European material i s found in l a y e r s p r e c i s e l y where the land i s being b u i l t towards F i g u r e 45: A s e l e c t i o n of European a r t e f a c t s from t h e excavation a t Muri-Harau (KA2).A:brass c l a s p ( L a y e r 2).B:copper canoe patch(Layer 2 ) . C : b r a s s screw(Layer 2 ) . D:iron n a i l ( L a y e r 1).E and F : b r a s s f i s h hooks(Layer 4 ) s e e a l s o Figure 39. G:brass p o i n t ? f i s h hook(Layer 4 ) . H:brass c l a s p ( L a y e r 3 ) . I : ? c a s e i n comb (Layer 3 ) . J : c o t t o n s t r i n g ( L a y e r 2 ) . -82- cms Figure 47: European and Japanese artefacts recovered from the Muri-Harau(KA2) excavation. A:American 5£ coin(dated 1964) from Layer 2.B:Japanese 10 sen piece, (dated showa 17=1942) from Layer 2. C: Japanese 1 sen piece(dated shove 2=1927) from Layer 3. D: metal button labelled "best ring edge" from Layer 4.E:?shell button from Layer 3.F: button of Pi-nctada sp. from Layer 6. -83- Figure 48: Ceramic artefacts from the Japanese settlements on Nunakita and Parakahi. There are parts of a food bowl(A), high voltage insulators (B,I,J,K,L), rice bowls(E,F), plates(G), tea cups(H,M), and personal containers (C,D). the lagoon today. The absence in the other excavations must also indicate that vertical land building on Touhou had practically ceased by the protohistoric and historic period.At all these excavations the loose surface rubble was cleaned away before digging commenced. Of the artefacts found, only a few need specific mention. Of note, are two metal fish hooks, and a fragment of a third {Figure 45, items E,F, and G ) . One of these is rather crude, but the other more complete piece illustrates the typical Polynesian flare for stylish fish hook curvature. The point #G is from a larger hook/ and is very similar to examples from Nukuoro(Davidson,1967:182,186) . About 30 glass beads were found, some of them very tiny. All but one were found in Layer 4. There are several types evident, and three colour types are readily distinguished dark blue, lighter turquoise blue, and yellow. There are two sizes - a larger, rather squat form, and a variable sized smaller variety. The yellow bead is longer than the others. -84Unfortunately, almost nothing is known of the characteristics of early European trade items in the Pacific islands, and this is an area where research should begin. It is to be expected for instance that glassware, including beads, from England, Spain and Germany would have distinguishable elemental characteristics. The background work which would aid this identification has not yet commenced. It is tempting to attribute some of the Kapingamarangi glass beads to contact with Fernao de Grijalvares in 1536; however, it must be noted that they are quite different in character to glass beads found on Taumako, in a context which almost certainly equates with the visit there by Quiros in 1606. Necklaces of glass beads of various colours and sizes were observed in 1910 on Kapingamarangi(see Eilers,1934:119) . Glass beads have not been found archaeologically on Nukuoro. In passing, it may be mentioned that similar sized tiny glass beads of about 1mm diameter have been found archaeologically on Banaba island (Lampert,1965:19,23) to the east of Kapingamarangi. -85- KAPINGAMARANGI AND OTHER POLYNESIAN OUTLIERS Linguistics The Kapingamarangi language(KAP) is Polynesian in character, and most commonly thought of as closest to that spoken on Nukuoro(NUK), despite the fact that the two languages are not mutually intelligible(Carroll,1965:193). Carroll's study indicated a cognate percentage of 59% between these two languages, although he noted that this could be inflated because his KAP informant lived on Nukuoro and had used that language for some time(ibid.:194). Pawley has examined the relationship between the Polynesian outlier languages in detail, and draws attention to the fact that KAP and NUK "share a considerable number of morphophonemic innovations and lexical items apart from other known languages these shared features require either the assumption of genetic sub-relationship, or very extensive borrowing" (Pawley,1967:283-284) . The question which linguists are unable to answer, of course, is precisely when such borrowings took place - for instance, could they be 19th century in origin, before the languages were recorded in detail ? Pawley suggests at least three Samoic Outlier(SO) subgroups (i) Northwestern Outliers KAP and NUK (ii) Central Outliers Takuu(TAK),Luangiua(Ontong Java,LUA) Sikaiana(SIK) (iii) New Hebrides Outliers West Futuna(WFU) and Mele-Fila(FIL) He also raised the possibility the members of the first two putative groups belong to a wider group with Ellice Islands(ECE) dialects(1967:282). Bayard's study suggests that Caroll's figures are consistently inflated by about 12% (1976:61), and in his assessment the nearest values to KAP have a somewhat different order, as follows(after Bayard,1976:30): Luanguia(LUA) Takuu(TAK) Rennell(REN) Pileni(Taumako,PIL) Tongic(TO) Nukuoro(NUK) Samoa(SAM) 44.3 4 3.7 4 3.7 42.0 37.1 35.7 35.5 -86Thus, in the case of the NUK-KAP cognate figure, Bayard's findings are 23.3% lower than that obtained by Carroll. Of the relationship between these two languages in particular, Bayard comments: "The two northern atolls seem dissimilar in their correspondences despite a relatively high figure shared between them. Even allowing for the possibility of inflated percentages, KAP appears consistently higher to the central atolls of TAK, LUA, and SIK than does NUK, which shows high figures only to SAM, TON, and KAP." (Bayard,1976:34). Drift Voyaging As pointed out above, Kapingamarangi is one of the most isolated areas of land in the Pacific, and one which seafarers continue to find difficult to make a landfall at. Part of this difficulty is the apparently unpredictable nature of the surrounding ocean currents; while these are generally westward in direction, it is clear that they are eastward at times too ! In view of this, it is extremely unlikely that two-way navigation between this island and any other was ever a significant factor influencing the cultural history of the population. Some drift voyages, however, probably arrived on the island from time to time; indeed the original landfall must have been just such a journey. In considering the founding settlement of Polynesian outliers, Ward and others have contributed a great deal of information on the possible role of drift voyaging(Levison et al. ,1973; Ward et aZ.,1973), and Bayard has recently reviewed this material(1976). The relevant simulation voyaging assumed a point of origin from one of five possible areas Tonga(five islands) Wallis area(four islands) Ellice area( two islands) Tokelau(one island) Samoa(two islands) These experiments clearly showed that the northern outliers had a higher possibility of contact from the Ellice area, with the Wallis area a close second(Ward et al.,1973:338). -87For example, the Kapingamarangi figures are Ellice Wallis Tokelau 56% 35% 9% Those for Nukuoro, on the other hand are: Ellice Tokelau Wallis Samoa 36% 31% 18% 15% One has to be very careful not to misinterpret this information. Ward et al. have rightly warned that caution should be used in interpreting their Table, as these figures indicate relative orders, not exact probabilities(ibid.:337-8). In fact, when the number of actual landings is taken into account, the contact probabilities of these drift voyages to Kapingamarangi is negligible(Ellice=0.9%, Futuna & Wallace=0.4%, remainder=zero); moreover, the figures to Nukuoro are even less from these places(see Levison et al.j1973:57).A total of 59 successful landings were made to Kapingamarangi, compared with 4 79 to Ontong Java, and this clearly shows the relative isolation of Kapingamarangi atoll(it is not clear from Ward et al.,1973: 334 whether these landings are from the full 100,000 attempted drift voyages, or a smaller number). Despite this high percentage figure for Ellice Islands cited above, it would be a serious error to assume that the information on drift voyaging in any way points to an origin of the Kapingamarangi people from this island area. It would be interesting to have 'reverse' voyaging attempted from Kapingamarangi, as this is bound to be much more enlightening on likely sources of drift voyages to Kapingamarangi. It is strongly suspected that if this were done, significant contact with islands to the south would be indicated. As Levison et aZ-.^point out in summary: "our probabilities to the more distant Outliers, such as Nukuria and Nukuoro[and presumably Kapingamarangi], are negligible, but would probably be high or very high if voyages were started from Ontong Java or Sikaiana" (Levison et al.3 1973:58). -88Comparison with Nukuoro The nearest neighbour to Kapingamarangi is Nukuoro, some 215 km distant; the next closest land is also a Polynesian outlier, Nuguria at 511 km(see Figure 1 ) . Despite the proximity of Nukuoro, some interesting differences in prehistoric economy and material culture have been revealed by the archaeological research. It is worthwhile to summarise briefly some of the more notable points of comparison between these two islands. Fishing: this was clearly a very important activity on both islands at all periods. Detailed comparison must await further analysis of the Nukuoro assemblage, now • underway. Shellfish: on both islands, the role of shellfish appears not to have been very important in the past; and Turbo sp. is apparently equally predominant in the archaeological sites. Birds: in neither island could birding be seen as a very significant economic activity. Gallus gallus appears late in the prehistoric sequence on Kapingamarangi, but has not been found archaeologically on Nukuoro. Pig: This is not an important aspect of the economy on either island, and only appears in late archaeological contexts, assumed to be historic in origin. Dog: this is an important point of difference between the two islands. Dog was never present on Kapingamarangi, yet it occurs in early deposits on Nukuoro. It apparently died out fairly early in the sequence there, and was re-introduced in the historic period. The lack of dog on Kapingamarangi is good evidence for isolation between the two islands. Rat:these are common in archaeological deposits on both islands, but different species are involved - Rattus exulans on Kapingamarangi, and Rattus rattus cf. mansorius on Nukuoro. This difference is again compelling evidence of isolation between the two islands, and also points to contact between Nukuoro and islands further to the north or west. Turtle: not important in either island. Taro: the information here is not conclusive, but there appears to be a notable difference. Although three species of aroid are present on Nukuoro today, early accounts mention only two - Coloaasia esulenta(taro), and another which may have been either Alooasia maarorrhiza as stated by Kubary(1900:108-9), or Cyvtosperma ohamissonis, as implied by Eilers(1934:189). The name pura or puraka is given by both Kubary and Eilers. It seems most likely that Coloaasia and Cyrtosperma were present in pre-European -89- times, and that Alooasia with its modern name ngaungau is a more recent introduction. On Kapingamarangi, By comparison, only Colocasi-a esaulenta, known as tara, is thought to have been present in the prehistoric period, and the other two tare forms were apparently introduced during the historic era, together with their Nukuoro names. Thilenius(1903:17), on the other hand, claims that not even the taro was present on Kapingamarangi before the advent of the European;this point of view is also put forward by Ray(1917:188). Turmeric: The cultivation of this plant and its associated ritual was important in pre-European times on Nukuoro (Kubary,1900:75). The plant was introduced onto Kapingamarangi in the historic period. Breadfruit: this was almost certainly present on both islands prior to European contact(unusual for atolls), and bears the standard Polynesian name in both places. On Kapingamarangi, a specialised preservation technique occurs. This is not known on Nukuoro. The absence of the paku-kuru conserve on Nukuoro, again suggests isolation between the two islands. Shell Adzes: those made from Terebra and Mitra spp. are as common as the ventral lip Tridacna sp. form on Nukuoro, and inner hinge Tridacna sp. adzes are about one-third as numerous as either. On Kapingamarangi, the ventral lip Tridacna sp. adze almost totally dominates. Hinge adzes are rare, and gastropod adzes are either absent or insignificant in quantity. This difference is not related to any environmental factors, and isolated development is again indicated. It is also notable that turtle bone adzes are common in the Nukuoro archaeological sites(Davidson,1971:76), but absent on Kapingamarangi. One-Piece Fish hooks: on Kapingamarangi, the near absence of Pinctada sp. meant that non-durable forms dominated, such as those made from coconut shell and turtle shell. The fish hook forms are simple, but distinct from those on Nukuoro. This probably reflects the influence of material constraints, rather than lack of contact as such. On Nukuoro, on the other hand, there is an abundance of pearl shell by comparison, and many hooks were made of it in several distinct forms. Metal hooks, based on prehistoric styles, were made on both islands in the early historic period. Trolling Lure Shanks: these are present in small numbers in both islands. On Nukuoro, several types were made which are typical of many in the Pacific. A unique form was developed on Kapingamarangi, again indicative of isolation. -90Abrasive Files: on Nukuoro these are made of coral; and on Kapingamarangi they are of sea urchin spines. Arm Rings: the triangular-sectioned Tridaona sp. arm ring is present in early contexts on both islands. Later examples are known on Nukuoro made from a large bivalve, and from Conus sp. on Kapingamarangi. Coconut Grater Heads: the use of unfashioned bivalves for this purpose appears to have been common throughout the sequence on Nukuoro, yet rare or absent on Kapingamarangi, Specially made heads of Pinotada sp. appear relatively late on Nukuoro, and on Kapingamarangi specially made heads of several kinds of shell are present throughout the prehistoric sequence. Loom Belt Weaving: this was present on both islands; on Nukuoro it is essentially a female activity, as it is generally in the Caroline Islands. On Kapingamarangi, however, it is a male prerogative, in common with almost all Polynesian outliers in Melanesia(Riesenberg and Gayton,1952:350). Moreover, hibiscus fibre was used on Kapingamarangi, and banana fibre on Nukuoro(Buck, 1950:148) . Tattooing: This was practiced on Nukuoro, on women especially(Kubary,1900:80-81); but was almost certainly absent on Kapingamarangi. The points of similarity mentioned above probably reflect the general family resemblance which one could point to for any pair of Polynesian islands in the Pacific, regardless of how distant from each other. The differences observed, however, are so great as to suggest substantial if not total isolation from each other, once primary settlement had been achieved. -91- CONCLUSIONS Land H i s t o r y of t h e Atoll Kapingamarangi a t o l l h a s n o t been a s t a t i c s e r i e s of i s l e t s o v e r i t s r e c e n t g e o l o g i c a l h i s t o r y . On s u c h a l o w l y i n g s p e c k of l a n d , t h e s e a r a g e s a c o n t i n u o u s war of a t t r i t i o n o v e r t h e b a r e l y c o n s o l i d a t e d c o r a l g r a v e l s w h i c h make up e a c h i s l e t . T h e s e g r a v e l s a r e m e r e l y s i t t i n g on t h e l e e w a r d s i d e of t h e main c o r a l r e e f , and sudden s t o r m s can wreak havoc w i t h t h e i r s t a b i l i t y . T h i s i s e v i d e n t f r o m b o t h t h e known r e c e n t h i s t o r y o f t h e i s l a n d , and from k n o w l e d g e o f l a n d - b u i l d i n g p r o c e s s e s p o s s e s s e d by t h e a t o l l ' s inhabitants. Matukerekere, the tiny i s l e t to the s o u t h , was d e v a s t a t e d by a s t o r m e s t i m a t e d t o h a v e t a k e n p l a c e i n 1 8 5 8 , a n d i s now s l o w l y r e - e s t a b l i s h i n g i t s e l f . T h e l o n g i s l e t of Hare h a s t h r e e l a n d d i s t r i c t s on i t s s o u t h e r n p a r t , which were s e p a r a t e i s l e t s u n t i l r e c e n t l y . Only one of t h e s e i s l e t s h a d b e e n c o l o n i s e d b y v e g e t a t i o n i n 1 9 4 2 , t h e Luawa d i s t r i c t ( s e e F i g u r e 4 9 ) . I t i s l i t t l e wonder t h a t t h e P o l y n e s i a n h a s f o r i t s common m e a n i n g ' u n c u t ' , w o r d f o r i s l a n d o f motu, ' n o t - d i v i d e d ' o n K a p i n g a m a r a n g i - a double-entendre nicely d e s c r i b i n g t h e t e n u o u s hold t h e l a n d h a s o v e r t h e s e a on t h i s a t o l l . Figure 49: The s o u t h e r n t i p of Hare i s l a n d a s i t was i n 1942. Three small i s l e t s can be seen, one of which i s i n t h e p r o c e s s of j o i n i n g onto t h e main block of land.By 1950, f u r t h e r land j o i n i n g had taken p l a c e ( E m o r y , 1 9 6 5 : 4 ) , and by 1980, only t h e small i s l e t t o t h e r i g h t was s t i l l s e p a r a t e . This p r o c e s s of p e r i o d i c i s l e t r e c l a m a t i o n and s u b d i v i s i o n i s b e l i e v e d t o have been underway f o r a t l e a s t as long as people have l i v e d on Kapingamarangi. -92The people on Kapingamarangi are thoroughly aware of this power the sea has to take away their land, and to build it up, and they take advantage of it. They engage in some very clever examples of land building of their own, in which they harness marine power to do the work. By judicious wall-building, the sea can be encouraged to make new land, by subtle alterations in the currents this causes. Even apparently insignificant lines of rocks placed on the reef can have a substantial land-building effect. Knowledge of this kind of process could easily be acquired by observing the effects of making fish traps with rocks in the shallows. By consolidating newly formed land with more substantial walls on its periphery, a somewhat more permanent 'islet' can result. Apart from Touhou islet, almost all of the present land in Kapingamarangi is very low to the sea, and no archaeological deposits could be found on other islets which indicate any lengthy stratigraphic history. A distinct possibility therefore is that the land has come and gone many times at different stabilizing points on the reef over millennia; and that the present configuration of islets is merely one chapter in the atoll's complex geomorphological history. Although uncertain, it is strongly suspected there are no culturally deposited sediments anywhere on Kapingamarangi•atoll which are older than those on Touhou islet. Yet, when cultural debris was first deposited on Touhou, about 800 to 1000 years ago, this land was barely above sea level, and could not have sustained domestic habitation in the islet's present form. In short, the people leaving this first cultural evidence must have been in residence on some other islet in the atoll chain. Just how much longer the full prehistoric sequence of occupation is for Kapingamarangi will never be known, because the islets these pre-Touhou people lived on have, in our view, subsequently been stripped of cultural soils. Some early massive storm which could have had this effect, may well have been the deciding factor encouraging people to take up residence on this slowly consolidating patch of sand that was to become Touhou in due course. Over the next 500 years, the people imported massive quantities of coral gravel onto this islet, and built it up to its present-day height of 4m above the sea. An added advantage of this relatively high artificial island is that it would serve to strengthen the fresh water reservoir in its basal sediments. This would act as a useful buffer for the water supply in times of drought. Kapingamarangi is in an area of the Pacific with unreliable rainfall, and any means to lengthen the available buffer would have been worthwhile. It is doubtful though that the people realised this advantage that island building bestows. It is notable that in the course of the formation of Touhou islet, the density of fish bone material increases in the sites five-fold - this must surely reflect in part the growth of the population in residence on the islet. The early people also took the precaution of building a large sea wall on the ocean side of the islet, and this is rapidly replaced each time it gets washed away by the inevitable periodic storm or exceptional tide. -93Thus characterised, the artificial islet of Touhou is seen as an appropriate adaptation to life on an atoll where the sea constantly challenges the permanence of land. Origins of the Founding Population An intriguing question is where precisely did the earliest Kapingamarangi people come from ? Unfortunately, it is not possible to be certain on this point. In the future, the best possibility for establishing a satisfactory answer to it may be by acquiring additional knowledge from human skeletal material, not only from Kapingamarangi, but from other areas in Micronesia, Polynesian outliers, and especially in Melanesia. This type of information is seldom incorporated in discussions of the prehistory of individual islands, and this is a little like observing the stage set of a theatre without the players it leads to imperfect knowledge of the play. At this stage we can be quite confident that the founding group came from an area already inhabited by Polynesians - this is evident from the scant skeletal material recovered, and of course the language spoken at the close of the prehistoric period. However, we do not know exactly how long people have been on Kapingamarangi, nor very much about their early material culture. It appears that the island has been inhabited for at least 700 to 1000 years. So far as is known, Nukuoro to the north was inhabited for a similar length of time, but here it appears that a reasonable sized islet existed when occupation first began. It is therefore unlikely the Nukuoro prehistory extends back much beyond what is presently known. The same cannot be said for Kapingamarangi, as it has been argued above that there must have been a population in residence elsewhere in the atoll before the earliest sediments on Touhou. Thus, present evidence favours Kapingamarangi being settled before Nukuoro, and this cuts out one possible island for the source of the founding group on Kapingamarangi. Of course, these arguments about landform history depend on the assumption of land-sea stability over the last thousand years or so. However, these atolls are not far from the unstable west Pacific zone of tectonic movements and volcanism, and earthquakes are not unknown on Kapingamarangi(Wiens,1956:12). Assuming tectonic stability, present thinking would be against occupation of these northern atolls earlier than about 1500 years at the outside. The suggested independant settlement of Kapingamarangi and Nukuoro is more positively indicated by the differences outlined above in material culture and economy between the two islands. Of special importance in this respect is the different species of rat on the two islands. Traditionally, the people on Kapingamarangi trace their descent from a man called Utamatua and his wife Roua, who came from an island called Tamana, although another island Nikuhetau is also mentioned in connection with the pair(see Eliers,1934: 126ff; Emory,1965:28,35). Tamana is described as one of three islands, the others being Pakihi and Pararu(Elbert,1949:240). There is a Tamana island in the Kiripati group, and Nukufetau -94is known in the Ellice group. The identity of the other two islands is unknown. Nukufetu and Kapingamarangi are both mentioned in origin myths on Nukuoro(Eliers,1934:129); and in Ontong Java, Kapingamarangi[Makarama] is mentioned in origin tales(Thilenius, 1903:31). These traditions surely indicate knowledge of other islands surrounding Kapingamarangi; but they should not be interpreted literally on the question of origins. It is more likely they contain a measure of information gained from the arrival of rare castaways on the atoll from elsewhere. To interpret the Utamatua story literally one must take into account that the story is primarily about the creation of Touhou islet(Elbert,1949:240), and that a man called Korae was already living on Turuaimu, when Utamatua arrived on Kapingamarangi. One may well ask who was this earlier inhabitant Korae, and where did he originate ? - rather than pay too much attention to the origin of the interloper, Utamatua. This is probably a good example of what is so frequently found in the analysis of similar myths amongst the New Zealand Maori; their role is essentially a charter for modern land ownership, rather than an accurate record of history, which can be interpreted literally. Probably the soundest conclusion that can be reached on present knowledge is that the first group on Kapingamarangi arrived from somewhere to the south, probably from another Polynesian outlier in Melanesia. It can be noted in passing that artificial island building, such as that involved at Touhou, is not uncommon on the coastal areas of Malaita, and of course on Tauamko, a Polynesian outlier.On the basis of the linguistic evidence, Bayard reaches the conclusion of "primary settlement... from TAK,NKM, NKR, or SIK"(1976:53), and this view is not at all at variance with the archaeological evidence. What is somewhat more questionable though, is the view of significant secondary contact with Nukuoro, Kiripati, and the Ellice Islands in the pre-European era(ibid.) In our view, the archaeological evidence indicates considerable isolation on Kapingamarangi once settlement had been established. Contacts with the outside world when they occurred do not seem to have made a significant impact on the course of Kapingamarangi culture-history. If the linguistic evidence demands a measure of borrowing between Nukuoro and Kapingamarangi, then this may well have followed contact with Europeans, and the resulting increased mobility of Polynesians, during the 19th century. A probable exception to the long history of isolation on Kapingamarangi concerns the manufacture of a specially wrapped food preserve on the island. This has been described in great detail by Eilers(1934:72),Buck(1950:35ff), and Emory(1965:134). Eilers maintains that this conserve, to bako , was made with pastes of both breadfruit and pandanus, and draws attention to the similar pandandus conserves of Nauru(1934: 72; see also Hambruch,1914:113ff). Buck also notes the similarity of the carefully tied roll of the breadfruit conserve, paku kuru,with rolls of pandanus from both Kiripati and the Marshall islands. However, according to his informants, pandanus was never treated in this way on Kapingamarangi(Buck, -951950:30-31). This type of conserve is unknown on Nukuoro, as it is on all other Polynesian outliers - how did it come to be on Kapingamarangi ? This may well indicate an instance of secondary contact with an island to the east in the Nauru or Kiripati or Marshalls area. The possibility of some prehistoric contact with this area is further supported by the type of roof thatching on Kapingamarangi. People on the island claim that this .was formerly of coconut leaves, and that pandanus thatching was introduced by a castaway couple from Tarawa in the Kiripati group(Buck,1950:66;Emory,1965:52). On linguistic grounds alone, an origin for the Kapingamarangi population from the Kiripati - Marshalls area is out of the question. These two instances of close cultural similarity must be examples of borrowing in the pre-European period, suggesting occasional drift voyages from the outside world to the east. The other difficulty with the linguistic evidence is that archaeological findings now confirm that a number of Polynesian outliers(Anuta, Tikopia and Taumako at least), have been occupied for up to 3000 years, while the language time depth appears only to be about 1000 years. For Taumako, at least, there is no reason to propose that the earliest people were not from Polynesia itself. Thus, relationships established on the basis of languages spoken in the 19th century, may merely document the last chapter in a very complex story; and while linguistics confirms the 'backwash' theory of Polynesian outliers, it is poorly equipped theoretically to cope with what now seems likely - multiple backwashes, perhaps from several areas. On linguistic grounds, the Ellice Islands appears to be a prime candidate as the direct source of a number of Polynesian outliers. It is very important in the future to see whether the Ellice Islands has an occupation time depth of 3000 years, and until this is established it cannot be considered as the source for the earliest populations on some of the outliers. So far as Kapingamarangi is concerned, this problem over relative time depths does not arise - it is unlikely that the full prehistoric sequence would be more than 1200 to 1500 years. The main difficulty in deriving the earliest Kapingamarangi population from the Ellice group is the lack of close resemblance between material culture in the two regions at European contact. A long list of items could be drawn up, particularly in the realm of non-durable material, which simply do not occur on Kapingamarangi. This includes tattooing, reef sandals, turtle bone adzes, ray skin rasps, a special kind of rat trap, certain kinds of small hand nets, the teka dart, drums, and so on(see Hedley,1897 ,-Kennedy,1931; Koch,1961). Their absence on Kapingamarangi is made more significant by the fact that many of these items are present on other Polynesian outliers to the south, such as Taumako. A further point concerns the remarkable fishing technology in the Ellice Islands devoted to catching palu, the ruvettus fish(Kennedy, 1931:12ff). It is difficult to believe that the masterly fishermen of Kapingamarangi, once they knew of such a valuable fish for the taking, would not exploit -96the resource to some degree. This fish has a circumtropical distribution, even occurring in the Mediterranean(Gosline and Brock,1960:253). It was not collected by the fishermen assisting with the comparative collection during the expedition, and this is bound to reflect a lack of knowledge of its habits, rather than a lack of presence of the fish. This type of fish is only caught in depths of 150 to 450m off the sheer face of the reef(Kennedy,1931:29). Bones of these fish are readily identified, although they may well have a lower survival rate than those of other species. This is because of the oily nature of the bone, rather like that of tuna and barracouta, though more so. The bones can even be eaten with the flesh(see Titcomb,1972:159). Lack of survival archaeologically cannot be the full story, however, as less durable cartilage from elasmobranchs was common in the sites. We can be confident therefore that the Kapingamarangi fishermen do not now, and never did, know how to catch this valuable fish. Although it has yet to be established just when the people in the Ellice Islands learned this technology, the lack of it on Kapingamarangi does seem significant. This discrepancy is pointed out with some reservations, however, because it appears that ruvettus fishing was known on Takuu, Nuguria, Ontong Java, Nukumanu, and Sikaiana. This conclusion is based on the presence of the fairly distinctive large wooden hook with a detachable strongly incurved point(Anell,1955:231). Large, but dissimilar hooks were made on both Kapingamarangi and Nukuoro(ibid.:230 ;Buck,1950:234; Beasley,1928:101), but in both places were said to be used for catching sharks, although these would also function for other large deep water fish. It is not known how long this fishing technology was present on these Polynesian islands in Melanesia, and it is not universal amongst outliers by any means. The Taumako people, for example, had only a poorly developed fishing technology, and ethnographic specimens of the ruvettus hook are unknown from this area. On the whole, the evidence of ruvettus fish is difficult to understand fully at present. It would probably be wise to reserve judgement in the meantime, and await better documentation directly from bones of the fish archaeologically. At this stage it would be a mistake to attempt to be too precise on the question of the origins of the founding settlers of Kapingamarangi. Very little is known of the prehistory of the closest Polynesian outliers in Melanesia, but there is every sign of complexity. For instance, although traditions on Ontong Java identify Kapingamarangi as their original homeland, a close biological relationship between these two populations seems very unlikely. This has been vigorously opposed by Shapiro (1933:271), and there is no reason to update this view. Similarly, close biological affinity between Kapingamarangi and Nukumanu has been equally strongly opposed (Schlaginhaufen,1929:282), although in this case the relationship between those people studied, and the earlier inhabitants of Nukumanu may be questioned. The only Polynesian outlier where it has been possible to examine early skeletal material in -97any detail is Taumako in the Solomons. Biological affinity between Kapingamarangi and this island could certainly be proposed; however, the affinity is probably due to general family resemblance between Polynesians, rather than direct contact in this case. There are several artefact forms on Taumako which have been important to the population there for nearly 3000 years(see Leach and Davidson,n.d.); it seems unlikely that these would not occur on Kapingamarangi, had there been direct contact from this area. In addition, the severe incidence of yaws on Taumako makes it quite likely that settlement from this island would have involved transportation of the disease. In summary, there is every reason to believe that the populations on Nukuoro and Kapingamarangi were seeded by immigrants from different areas. Present evidence favours a southern source for Kapingamarangi, although it is not possible to be more explicit than this. It is evident that after initial settlement, occasional contact occurred with islands to the east, perhaps with Nauru or Banaba. The Prehistoric Fishermen of Kapingamarangi Pitting one's wits against the sea and the fish it contains is of foremost importance to all men on Kapingamarangi. Endless hours are spent discussing the whys and wherefores of different kinds of fish, their habits, and how to catch them. In this respect, the Kapingamarangi fisherman is little different to his counterpart in modern European society - however, for skill in fishing and knowledge of fish habits, the men on Kapingamarangi are second to none. It is hardly surprising that in Kolonia on Ponape, the men of the Kapingamarangi village of Porakiet have established themselves as the mainstay of the fish market, and their prowess as fishermen is supreme. Above all others, the type of fish which is seen as the most worthy adversary is the pelagic game fish, which is frequently caught by trolling a lure in the surface of the sea over deep water. This type of fish is a voracious carnivore, feeding on small fish which it drives to the surface, and then seizes from beneath. Frequently, fish of this kind work in packs to drive a shoal of smaller fish together. Such packs are easily recognised either from the birds which follow them about, or by the surface turbulence caused by small fish darting to and fro in their attempts to escape. If the fisherman manages to get into such a position with a lure, an abundant catch is assured, because the game fish in their feeding frenzy will simply attack anything moving in the water. In addition, packs of game fish can be encouraged together from a distance to some extent, by trolling a lure through the water; its action disturbs the water, and emulates a small struggling fish at the surface, and it is therefore an irresistible attraction. The species of fish which can be taken by this method of fishing, vary from one part of the Pacific to another. On Kapingamarangi, the main ones are as follows: -98Agrioposphyraena barracuda Barracouta Aaanthoaybium 'solandri Wahoo, Jack Mackerel Neothunnus macropterus Yellowfin tuna Katsuwonus pelamis Skipjack,striped tuna takua, and te atu3 These fish are known as te ono, te mala, names which are almost universal in the Pacific, and which reconstruct at least to proto-Polynesian(see Appendix 1 ) . In the Kapingamarangi language, special terms exist which relate to small variations in size and characteristics, clearly showing the central importance of these fish. It is hardly surprising that when fishing for our comparative osteological collection began on Kapingamarangi, the first two species of over 100 eventually collected were takua and te mala. As far as the modern Kapingamarangi fisherman is concerned, to find out anything useful or interesting about the prehistory of fishing on the atoll would be to illustrate the varying prowess of his ancestors in catching these types of fish - the evidence shows that while much dreaming of these fish may have taken place in the past, far more were probably caught in the men's house than ever in a canoe ! This discrepancy between the spectacular bonito or tuna fishing activities of people in a number of areas of the Pacific in the historic period, and the pattern of catches revealed by the archaeological record is a considerable mystery at present. Even in the temperate waters of New Zealand, where bonito or tuna fishing was not really feasible, the same discrepancy can be seen in the counterpart fish - the kahawai, Arripis trutta. Special lures for these fish are amongst the most common items in 19th century ethnographic specimens collected from the New Zealand Maori; yet both the hook and bones of the fish are extremely rare in archaeological sites. The same phenomenon is true for Kapingamarangi. Of all the fish which could have been caught by trolling with lures, only five in every 100 fish actually caught, belonged to this type in the past. It would be interesting to know just how frequently these fish are caught today; our suspicion is that the true situation is more like the Pygmy elephant hunters in Africa - there is much talk of elephants, and there is a specialised vocabulary about elephants, but the diet largely consists of items other than elephants. It would be a mistake therefore to hurriedly seize upon this dearth of game fish in the archaeological record, and argue that the prehistoric fishermen of Kapingamarangi were far less skilled than their modern descendants. When Peter Buck visited the island in 1947, one of his ambitions was to go "on a fishing expedition the purpose of which was to view the method of line fishing for tuna"(Buck,1950: 219); but after several days away from Touhou, during which -99many fish other than tuna were caught, he failed to observe this activity at first hand. He commented that "fish seemed to be scarce"(ibid.:248), and was told that "the day was wrong (huaaitu) because it was too bright and the fish were not swimming near the bottom"(ibid.:24 9 ) ; eventually rather sadly, they paddled through the reef channel and the four miles back to Touhou. He adds "we had another demonstration of the value of direct observation"(ibid.:249), which barely disguises the disappointment he must have felt. Buck concludes that "bonito fishing by trolling with a lure appears to be on the wane on Kapingamarangi"(ibid.:237). On the contrary, this experience could be viewed in another way altogether. Catching pelagic fish is largely a matter of seizing the opportunity when and if it arises. These fish are rather unpredictable in their occurrence, and travel far and wide. They can easily be caught whenever they are actually present, either outside the reef or inside, but one must be close at hand or actually in position when the opportunity arises. To go out specifically to catch only these fish is to invite an empty stomach in the evening. In point of fact, as every modern European fisherman knows, when fishing for the common and mundane fish, it pays to have a lure either permanently trolling along behind, or immediately to hand, in order to catch the pelagic fish when they turn up. When a flock of birds is seen, one hurriedly pulls up the baited lines, and dashes over to the vicinity of the birds, to try and catch some of the pack. More often than not, the pack disappears as quickly as it made its presence known, and baited lines are dropped again to carry on the main, but less spectacular fishing activity. Thus, the catch of even the most expert fisherman, will largely be composed of something other than pelagic fish. This is exactly what Buck observed, and exactly what the archaeological record demonstrates. What then can be seen about the main body of the catch during the prehistoric period ? The most interesting aspect of this is the abundance of parrotfish in the earliest levels and the relative absence of members of the cod and groper family. In the troical Pacific area the Scarids or parrotfish fill a similar role to that of Labrid fish in more temperate waters - that is, they are abundant and easy to catch. For this very reason, they are not especially sought after, and in fact are considered inconsequential or even despised by the expert fisherman. It is ironical, but hardly surprising that Labrids formed the mainstay of prehistoric fish diet amongst many communities in early New Zealand, and Scarids form the basis of fish diet of many people in tropical Oceania. It is interesting that in the case of Labrids, some lines of evidence suggest that these fish offer security in times of economic or environmental stress(Leach and Anderson,1979a), and are therefore more abundant during periods of hardship. A similar hypothesis might be advanced for Scarids too. Seen in this way, the early archaeological evidence on Kapingamarangi might show somewhat less environmental familiarity than in later periods. Scarids are primarily herbivores, and as such do not -100normally take a baited hook. They are therefore caught primarily by netting, and many can be taken in this way in relatively shallow water near coral thickets. Catching this type of fish is a 'safe' activity, guaranteed to succeed. On the other hand, baited line fishing, such as is appropriate to the cod and groper family, requires somewhat more initiative and skill in deeper water, is less secure, more adventuresome, and not always guaranteed to produce a lot of fish. A suitable index therefore of the real skill of prehistoric fishermen is not so much how frequently pelagic fish are caught, but the relative quantities of cod and groper as compared with parrotfish.The relevant figures for the different periods on Kapingamarangi are given in Table 11 below: TABLE 11: TRENDS IN THE RATIO OF PARROTFISH TO COD AND GROPER IN KAPINGAMARANGI Period Cod & Groper Parrotfish Ratio 0-100 BP 23.26% 20.93% 1.11 100-300 BP 19.75% 18.81% 1..05 300-700 BP 15.99% 17.27% 0.93 7.69% 23.08% 0.33 700-?1000 BP This simple ratio shows a uniform rise in the relative importance of baited line fishing over the last millennium, and must indicate a dramatic improvement in fishing skills and technology over this period. With the near absence of suitable shell for making one-piece bait hooks, this increasing abundance of fish caught on hooks made from less durable materials such as coconut shell, is a striking testimony of the prowess and enterprise of the prehistoric fishermen of Kapingamarangi. -101- APPENDIX 1 MARINE SPECIES IN KAPINGAMARANGI ORTHOGRAPHY, WITH SOME PROTO-POLYNESlAN EQUIVALENTS D.T.Bayard For ease of comparison, terms in the text have been rendered in a standard "Polynesian" orthography. However, the orthography used on Kapingamarangi and in Lieber and Dikepa's KAPINGAMARANGI LEXICON recognises the distinction made between long (aspirated) and short (unaspirated) consonants in the KAP language (Lieber and Dikepa, 1974: xvii-xviii). This distinction has arisen largely as a result of the disappearance of a medial vowel between two identical consonants (*C1VC1 > ClCl); hence PPN *totolu becomes ttolu <tolu> in contrast with tolu <dolu> (Pawley, 1967:286). The following long/short pairs of phonemes are distinguished in the KAP orthography: p/b (<PPN*p); mm/m (*m) ; ww/w (*w) ; t/d (*t) ; nn/n (*n) ; k/g (*k) ; nng/ng (*ng); hh/h (*f/*s); 11/1 (*l/*r). Where possible, ProtoPolynesian equivalents have been added (from Walsh and Biggs, 1966; Biggs, 1979; Pawley and Green,1971); however, exhaustive search and investigation of irregular reflexes would doubtless increase the number of proto-forms presented here. The family numbers given below follow the order in Munro (1967: vii-xv). The fish and shellfish species, together with the PN names, are those collected and identified by Leach and Ward during their expedition. FISH SPECIES Family Species Car charhinus spallanzani Triaenodon apicalis Amphotistius kuhli Megalops cypvinoides Chanos ahanos Gymnothorax undulatus Tylosurus indicus Rhyneorhamphus georgi Arnoglossus cf.intermedins Ostiahthys murdjan Ostichthys cf. adustus Holoaentrus cf. tiereoides Holocentrus ruber 65 Sphyraenidae Sphyraena jello Agrioposphyraena barracuda 66 Mugilidae Valamugil seheli 72 Scombridae Grammatorcynus bicarinatus 73 Acanthocybiidae Acanthocybium solandri 75 Thunnidae Neothunnus macropterus 76 Katsuwonidae Katsuwonus pelamis 88 Carangidae Caranx melampygus Carangoides cf.laticoudis Caranx cf. melampygus Caranx sexfasciatus Caranx cf. ignobilis Carangoides oblongus Elegatis cf. bipinnulatus 97 Epinephelidae Cephalopholis miniatus Plectropoma maculatum Cephalopholis oyanostigma Cephalopholis sp. Epinephelus cf. tauvina Aethaloperca rogaa 4 5 7 15 21 42 47 48 53 57 Carcharhinidae Triakidae Dasyatidae Megalopidae Chanidae Muraenidae Belonidae Hemirhamphidae Bothidae Holocentridae "PN" Name KAP Name hokoulu mokolewe te hai ? te nga matakalati te kiha te aku moana te iha te paipai malau pungu malaupunga luli malau te ta tapatu te ono kanae ika-pou te mala takua te_atu alala laloale motomoto (no name) te kata te kata lehu laloale te kina mataili ? tai-awa matapaku (no name)_ ngatala talo taea uli hogoulu mogo-lewe di hai di ngaa mada-galadi di gihaa di agu moana di iha di baibai malau-pungu malaupunga luuli malau di daa dabaduu di ono ganae iga-bou di mala dagua di adu (gada-)alaala laloale-modomodo gada di gada-llehu laloale di gina madaaili daiawa madaabagu ngadala-daalo daaea-uli PPN Term *fai *ise *malau *malau *malau i *'ono *kanahe *maratea? *takua * 'atu H O M 97 Epinephelidae Epinephelus Epinephelus Promiarops Epinephelus Epinephelus 106 Lutjanidae 107 108 110 114 Aphareidae Caesiodidae Nemipteridae Lethrinidae damelii fuscoguttatus lanoeolatus fuscoguttatus sp. Plectvopoma maoulatum Pleotropoma maoulatum subsp, Lutjanus mono stigma Lutjanus semioinatus Maoolor sp.(not M.niger) Lutjanus cf . aoatesi Macolor niger Lutjanus janthinuropterus Aprion viresoens Lutjanus sebae Lutjanus rivulatus Aphareus cf. rutilans Caesio coerulaureus Monotaxis granocutis Lethrinus fletus 117 Mullidae Parupeneus 124 Kyphosidae 125 Platacidae Pavupeneus macronemus Kyphosus cf. vaigiensis Platax orbicularis 129 Chaetodontidae 130 Pomacanthidae 138 Coridae 140 Scaridae Chaetodon cf . cf. janseni semeion Heniochus varius Pygoplites diacanthus Epibulus insidiator Ch.tvilobatus, Ch.undulatus Thalassoma lunare Cheilinus fasciatus Eemigymnus melapterus Callyodon cf. niger Bolbometopon bicolov Cetossarus putchellus Bolbometopon muraticus Scarops jordani manukuou pulungame ? te mohia ngatala te ngatala-moko-te te tonu tonu matahale tangau ka tangau pulapula te kuku te tata ? te hina tangau poto[ina?] te utu te taea tangau poto[ina?] matamata te uli te mu te muka te kala te matuailau ngatala akau kanikani hatu tihitihi tihitihi talehala ? ? te lipai te lali te lali talinga mea toliki te huhu paheni uli paheni melenge te huhu manu-guou bulu-nga-mee di mohia ngadala-dee ngadala-mogo-dee di donu donu-madaahale dangau gaa dangau bulabula di gugu di dada di hina dangau bodoina di udu di daaea dangau-bodoina madamada di uli di muu di muu-gaa di gala ? di madu-ai-lau • ngadala-agau ganigani-hadu dihidihi dihidihi dalehala di liba-i di lali di lali dalinga-mea dooligi di huhu baaheni-uli baaheni melenge di huhu *tonu *tonu * 'utukao? i *huli *muu *muu ? *kaloama *matu *tifitifi *tifitifi *'ufu *'ufu £ OJ ' 140 Scaridae 157 Siganidae 159 Acanthuridae 169 Scorpaenidae 175 Diodontidae 17 8 Tetrodontidae 180 Balistidae 182 Ostraciidae Chlorurus mic rorhinos Hipposaarus I ongiceps latus Siganus cf. v ermicu Siganus rostr atus Siganus ahrys ospilos Naso sp. Naso annulatu Aaanthurus cf . gahhm Cyphomyoter v lamingi Aaanthurus cf . gahhm Naso uniaorni s Naso cf. brev irostris Aaanthurus ol ivaaeus Calliaanthus cf. lituratus Naso annulatu s Aaanthurus cf . gahhm Aaanthurus cf . pyroferus Saorpaena car dinalis Pterois volit ans Saorpaena car dinalis Diodon cf. hy strix Arothron hisp idus Arothron stel latus Pseudobaliste s flavimarginatus Hemibalistes chrysopterus Meliahthys bu niva Ostraaion tub eroulatum palakia ulahi pongongo paua pongongo uli lapia te imu-alo hupaka te kelu palangi kelu akau te ika hepu te lipa ta hukumea ? imu-walu te lipa kono ehe te lipa te nohu te hauihau (no name) toutu te wali te ete ? te imu te humu hapula humu hakapekepeke tekemomo balagia ulahi bongongo baua bongongo-uli labiaa di imu-alo hubaga di gelu balangi gelu-agau di iga hebu di liba-daa hugu-mea imu-walu di liba-gono-ehe di liba di nohu di hauihau te keke te hono di gee di hono *'ulafi *palangi *nofu doudu di wali di ede di imu di humu habula *sumu humu hagabegebege *sumu degemoomoo TURTLES *kea *fonu SHELLFISH SPECIES Species Tvidacna maxima Hippopus hippopus Cassis aornuta nautilus Lambis lambis Lambis S D . Atrina vexillum Gafrarium peotinatum Asaphia violascens Codakia sp. Area ? sp. Cardium ? sp. Cardium sp. Spondylus varians Terebra aveolata, T. arenulata Name in text su kima pahua su kima kima te pu loto waihanga te (h)atili te waeanga te waeanga poto noua te wolo te pipi te pipi mea te pipi ta matau te koeho te pipi ta matau te hala te kai nap te pu ka tolotolo Lexicon baehua gima buu (not present) waeanga (not present) wolo bibi bibi bibi goeho bibi (not present) (not present) buu ADDITIONAL SHELLFISH FROM KAPINGAMARANGI LEXICON (Lieber and Dikepa 1974) pearl oyster baa univalves in general buu Melampis luteus buu-hau trumpet shell buu-iliili mollusc sp. golo-golo Vermetid gastropod unga-goo Nerita pioea unga-hihi Nerita polita bulolo Turbo spp. alili PPN Term r puupuu [ pipi r pipi 'pipi f pipi c puupuu *paa *puupuu *puupuu *puupuu *'ungakoo * 'unga-+*sisi *palolo *'alili o I PLACE NAMES APPEARING IN THE TEXT AND IN (1965) AND THEIR LEXICON EQUIVALENTS Emory Lexicon Emory Lexicon kapingamarangi torongahai ringutoru rikumanu turuaimu pepeio nunakita hukuniu parakahi werua touhou taringa pungupungu matiro matuketuke ramotu sakenge matawhei gabingamalangi dolongaahai lingudolu ligumanu duluaimu bebeioo nunagida huguniu balagahi weelua touhou daalinga bungubungu madiiloo madugeduge laamodu tagenge matawwei hukuhenua hepepa tipae tetau nikuhatu takairongo tangawaka hare herengaua herekoro tirakau tariha tiahu tokongo tiraukame pumatahati matukerekere huguhenua [hebeba?]* di bae teedau niguhadu tagailongo tangaawaga hale helengaua halegoloo di laagau taliha di ahu togongo di laagau mee bumadahadi madugelegele *not in Lieber's list -107- APPENDIX 2 MOLLUSC SURVEY ON KAPINGAMARANGI Graeme Ward INTRODUCTION In order to obtain some indication of the range of shellfish available to the prehistoric inhabitants of Kapingamarangi, informal littoral surveys were conducted during the expedition. Since Touhou and Werua are the main inhabited islets, the area around them was the initial focus of the study. It soon became clear, however, that neither the ocean nor lagoon shores of either were areas rich in molluscs; consequently, attention was shifted further afield. Some comments on present and past exploitation are included. METHODOLOGY Since the aim was to obtain data concerning the range of shellfish likely to have been available in the past, semi- or non-quantitative survey techniques were used. On the areas of the ocean reef east of Touhou and Werua, two transects were walked at low tide between the shore and the reef-edge. On each, populations of molluscs within a given area were noted, and some collections made, at set distances along each transect. It was not possible to cover a similar distance offshore from these or any islets on the lagoon side, since the water deepened more quickly with distance from the shore. Nevertheless, several areas up to 50 metres off the lagoon shore of Touhou were investigated at low and intermediate tides using basic diving gear. Subsequently, the lagoon shores of several other islets and inter-islet channels were inspected by walking the littoral and diving off-shore to a distance of 100 metres. Finally, much of the central, northern and southern lagoon-side reef areas were systematically searched by crews of up to seven divers on three occasions. OCEAN SIDE REEF FLATS Two transects of the reef flat oceanward from Touhou and Werua were made on foot at low tide in the company of one older and one younger Kapingi man. On each transect, at approximately each 10 metres, note was made of the immediate environment and • -108a collection, or at least identifications, made of all molluscs in the vicinity of a 4 square metre area. The results of each study are summarized below. Transect 1, Northeast of Touhou Islet Shore: Substrate of exposed coralline rock with large coral rubble, steeply sloping from eroded islet edge. Some wall building in progress, using coral heads and larger coral rubble. No molluscs. + 10 metres: Smooth coral limestone substrate with scatters of small pebbles of coral gravel. No molluscs observed. + 20 metres: Patch of coral gravel about 8 x 4 m exposed at low tide, resting on coralline rock substrate. Under rubble from 100 to 300 mm in diameter were found three small gastropods, Conus and Cypvaea spp. Nearby, a black, feathery-limbed starfish and several crabs were found. + 30 metres: Water-sculptured coral limestone flat under 20 cm water. Under flat coral stone c.20 cm in diameter were found a small Cypvaea sp., a Mitra sp. and a sea urchin. + 40 metres: A slightly raised coral rock flat with algal growth and scattered coral gravel. No molluscs observed. + 50 metres: A similar situation to that of the previous point. At the base of a coral stone were two small cowries {Cypvaea spp.). Also in the vicinity were several cone shells occupied by hermit crabs. + 60 metres: Same as last two areas but with a slightly more broken substrate. No molluscs observed. + 70 metres: As for previous areas but with a greater density of coral gravel. This area produced several concentrations of small molluscs from under coral rocks. A group of several tiny gastropods, five Cypvaea annulus, and several small Conus sp. were found.In the same situations were hermit crabs inhabiting a variety of gastropod shells, including Mitva, Stvombus and Dvupa spp. This area was otherwise notable for the presence of a number of small black and white speckled sea cucumbers. + 80 metres: Same substrate. One loose coral stone but no molluscs. Several Cevithium nodulosum shells inhabited by hermit crabs. + 90 metres: Similar substrate. Two Cypvaea annulus. + 100 metres: Similar substrate but lacking larger coral rocks. No molluscs; some sea cucumbers. + 110 metres: Similar substrate but undulating; no large rocks, and no molluscs. + 120 metres: As above. -109+ 130 metres: Similar substrate. Nothing found under coral rocks. Sea cucumbers in the survey area. + 140 metres: Similar area; large stone had algae at base. Nothing else but large brown and large black and white sea cucumbers. + 150 metres: A slightly higher area only 25 metres from the reef edge with a scatter of fine gravel over the coral limestone which is occasionally pitted with holes of up to 4 0 cm in diameter. In one of these, among the light gravel filling it, were found two Cypraea annulus. + 160 metres: Middle of raised area representing the algal crest of the reef edge and containing isolated patches of algae and bright corals being alternately covered and exposed by shallow water; substrate broken and uneven. Small gastropods including Cypraea spp. and Turbo sp. were found in holes. + 170 metres: Substrate of solid coral limestone, strongly water-worn with clefts overwashed with foaming water; no gravel or coral rocks. Only sea cucumbers seen. Reef edge at commencement of downslope: More broken substrate with a larger proportion of live coral. Several colourful gastropods including Turbo and Cypraea spp. but mostly small individuals. Larger Turbo argyrostomus and Tridaana maxima were seen further out. Many small crabs at the reef edge. The area approximately 20 metres back from the reef edge appeared to be that exhibiting the greatest concentration of molluscs in readily accessible situations on what was generally a disappointingly barren reef flat. It was decided to obtain a greater sample of this area. + 160 metres from shore: Raised area of reef flat, broken substrate with gravel and coral rocks. At the base of a coral head Drupa sp. and other small gastropods were seen. + 160 metres: Fifty metres further north on similar substrate. One large Cypraea and a small Conus sp. + 160 metres: Forty metres further north the ridge was slightly higher and less awash. Substrate pitted with holes. Collected four Cypraea spp., small Tectus and Turbo spp. and a live Mitra sp. from a rubble-filled hollow. + 160 metres: Another 50 metres further north. Similar substrate. Observed several small gastropods including Vasum, Drupa and Cypraea spp. These last sample areas were situated opposite the shallow channel between Touhou and Werua. The second transect commenced toward the reef edge approximately 200 metres further on, near the southern extremity of Werua Islet, and proceeded in towards the shore. The reef flat was only about two-thirds of the width of the flat at the place of the first transect. -110Transect 2, Southeast of Werua Islet + 110 metres: At the reef edge in the surge zone; water 50 cm deep. Substrate broken. Collected one large Turbo argyvostomus. + 100 metres: Water-worn coral limestone substrate. Collected large Conus sp. and small Tridaona maxima. + 90 metres: Slightly higher area of ridge coloured by light green algae. Several gastropods present, including small Troahus maeulata, Turbo, Cypraea and Conus spp. Water too deep for closer search. + 80 metres: Separately ridged area, pocked with hollows containing rubble. Hollows produce tiny gastropods and sea cucumbers. + 70 metres: More level substrate, less hollows and rubble. Some tiny gastropods. + 60 metres: Similar substrate, some larger coral rocks. Under one rock were small Conus and Cypraea spp. + 50 metres: Similar substrate, some algae. Several tiny gastropods. + 40 metres: Similar area. No molluscs. + 30 metres: Similar area. Two tiny gastropods under coral. + 20 metres: Even substrate of coral limestone, some scatters of fine rubble. No molluscs. + 10 metres: At edge of raised coral limestone protruding into shallow water. No molluscs, some crabs. Shore: Comprising predominantly coralline sand pockets among raised limestone. No bivalves or other molluscs found. The reef flats on the ocean side of Touhou and Werua were remarkably unproductive of edible shellfish. This is in keeping with the paucity of molluscan habitats on the reef flat. The inner shore and supra- and inter-tidal zones comprised largely coarse gravels unsuitable for filter feeders and these had been disturbed and, in some areas, extensively redeposited by the recent storm waters. There were some coral rocks in this zone at the south of Werua but these too lacked the expected range of small gastropods such as Nerita spp. and the various limpets common to such areas elsewhere in the Pacific. On the reef flat itself, only small deposits of fine gravels and sands were located and the substrate of hard coralline rock was generally exposed. There were few living corals or algae areas. There were no beds of sea grass providing suitable habitats for bivalves and small gastropods. Further out, in areas covered by water at all but the lowest tides, little live coral was present; again, few suitable molluscan habitats are present. Only at the small algal crest and the donwslope of the reef edge were suitable areas and these were relatively inextensive or difficult of access, even at -Illlow tide. While turbulent conditions prevented a inspection of the outer margin of the reef edge, gastropods and Tvidaona maxima were collected or reef margin. It seems likely that this area is a rich molluscan habitat but might not always have for exploitation. detailed several large observed at the relatively been available No similar systematic inspections were made of the ocean side reef flats of other islets of the Kapingamarangi Atoll. The possibility remains relatively untested, therefore, that the paucity of molluscs observed on the Touhou and Werua transects was a function of over-exploitation or some other factor related to these islets in particular. However, while a greater variety and denser populations of gastropods and some bivalves were observed on islets to the north and south, casual observations suggested that most lacked extensive areas of suitable habitats also. A notable exception were the areas of exposed coralline reef rock found on Nunakita and several of the smaller southern and northern islets; these areas had relatively dense populations of small gastropods in their intertidal and suprasurge zones, mainly Nerita spp., small Turbo sp. and some limpets. LAGOON SIDE LITTORALS The lagoon littorals of Touhou and Werua were found to be relatively poor in molluscan resources. In many places the sandy shore was built on with coral rock foundations for houses. In other places, as on the shore opposite canoe houses, the sand is frequently disturbed. Along much of the shore, especially of Touhou, the water is relatively polluted, the substrate silty and overlain with debris. Nevertheless, there are areas of fine sand and some seagrass among which some small bivalves were observed. The molluscan habitats of the lagoon shore were better observed along some of the other islets. Along the southern shore of Ringutoru, for example, there were extensive areas of fine coralline sand and patchy sea grass as well as large clumps of living coral. The shore beaches provided habitats for small bivalves including Donax and Tellina spp. Offshore, in sandy bottom areas and in patches of sea grass were the bivalves Asayhis, Gafrarium, Quidnipagus, Codakia, Soutaroopagia, Spondylus and Cardium spp. - all used throughout the Pacific as food - and a variety of gastropods including Cerithium, Cypraea, Oliva, Tevebra, Mitra spp., as well as various murids and drupes. A similar range of bivalves and gastropods was observed in the shallow water offshore from Hare and several others of the southern islets. Generally, these habitats were easily accessible and some of the species occurred in patches in considerable number, especially the sand-dwelling filter feeders,Donax and Tellina spp. Elsewhere in Oceania, these small molluscs are collected in large numbers and provide a dependable staple food resource. They were not observed being consumed on Kapingamarangi, and informants indicated that only the larger clams, Tvidaona maxima and Hippopus hippopus_, and occasionally the larger Turbo argyrostomus were used as food today. -112INTER-ISLET CHANNELS The channel between Tetau and Nikuhatu islets was inspected in some detail. It showed little sign of being a substantial molluscan habitat except toward its lagoonward extremity, where greater similarity to the inner reef areas were noted. INNER REEF MARGINS The larger part of the lagoon reef margin of Kapingamarangi atoll comprises not sandy islet shores but intertidal reef. These areas, extensive in the north, west and southwest of the atoll, and complemented by some perpetually but shallowly submerged reefs in the centre-west and southwest of the lagoon, are extremely rich marine ecosystems. A considerable range of molluscan habitats are found here, including those of the larger molluscs. Among the significant gastropods found were large cones, cowries, Trochus maculata and Turbo spp., Lambis and Strombus spp. and the conspicuously burrowing Mitra and Terebva spp. Most significant of the bivalves were the large clams, Hippopus hippopus and Tvidaena maxima. The latter were frequently found partially embedded in coral interstices along the reef margins to several metres in depth. Several areas of these reef margins and the submerged inner reefs were surveyed. CONTEMPORARY SHELLFISH EXPLOITATION Shellfish does not appear to be a very significant food resource to contemporary people on Kapingamarangi. Women and children were not seen collecting small bivalves on the sandy shores, or Nerita spp. and other small gastropods from outcrops of coral rock, or exploiting the variety of small molluscs inhabiting the sandy and sea grass flats, as they so commonly are on other tropical Pacific islands. No recent middens of discarded shells were seen. It is unlikely that seasonality or matters of access to suitable areas were involved in influencing the apparent lack of use of this potentially important resource. The only shellfish gathering which was observed was that done by men. For example, an elderly informant collected Turbo sp. shells from the reef edge and prepared their flesh for consumption; and an acknowledged expert fisherman collected tiny cowries from the reef flat for use in modern ornamental handicrafts. These examples, however, are insignificant when compared with the economic importance of clam shells. In particular, Tvidaena maxima was a prime target for predation as a food source. The large gastropod Turbo argyrostomus was one of the few molluscs claimed to be used for food today on Kapingamarangi (cf. Buck, 1950: 49). To prove his claim that this species collected at the reef edge was a delicious foodstuff, the elder of the two informants assisting in the transects of the reef flat proceeded to cook them on the shore. The shells were placed upside down upon a coral rock slab resting on the ground. A fire was kindled on top of them using dry coconut fronds. After it had burned for three or four minutes, he used a palm midrib to remove the shells from the fire. Each was hit against a nearby stone to knock the operculum from its resting place without damaging the shell mouth. The hot shell was held in one hand with the aid of -113coconut fibre, and a small stick used in the other to prize out the flesh. The small piece of gut was removed and only the muscle was eaten. The flesh was succulent and tasty. Collection of the two large clams, Tridacna maxima and Uippopus hippopus (cf. Buck, 1950: 49), was observed on several occasions at the inner reef margins. Divers used a knife, quickly and deftly thrust between the open valves of the clam to make it release its hold on the substrate; another twist or two of the knife within the shell would separate the mollusc from the two valves; these would tumble away into deeper water as the diver rose to the surface with the mollusc flesh. The gut was removed from the adductor muscle and fleshy mantle with the teeth and the flesh washed in sea water for later consumption. These clams were a favourite, easily obtained and prepared midday meal for fishing crews. Limes brought in the canoe were squeezed into a bowl to marinate the mollusc flesh. Left for an hour or so while the serious fishing proceeded, they provided an easy luncheon for a fishing group who would not have to land on shore or prepare a fire. In view of the area of reef at Kapingamarangi atoll, it appears unlikely that casual exploitation at this level would ever produce a shortage of clam meat. While it seems that Tridacna maxima is the major, if not the only species frequently sought for consumption as food today, other shells, especially the small cowries, Monetavia moneta and Cypvaea annulus are collected and used in ornamental handicrafts of the type offered for sale in Ponape and elsewhere. Buck (1950: 49) listed five molluscs as being of economic importance on Kapingamarangi: pahua, kima, avivi, pu, waenga; he also mentioned another: wovo. These are Tridacna. maxima, Hippopus hippopus_, Turbo argyrostomus,gastropod generally, and small members of the Strombidae family; and Atrina vexillum. Of these, the last is rather unlikely as it is extremely rare. In addition, Hippopus hippopus is barely considered even as poor food today. This list of shellfish of economic importance in recent times is short by comparison with other areas in the Pacific. -114- APPENDIX 3 THE ANALYSIS OF FISH BONE FROM KAPINGAMARANGI Foss Leach and Janet Davidson Comparative material for Pacific Island fishes is not readily available to assist the identification of archaeological finds. Over the past few years, specimens of about 200 species have been caught and prepared to try to rectify this situation. Selected cranial bones are mounted on boards according to anatomy, and are now available for use at the Otago Archaeological Laboratories. The method of processing archaeological assemblages of fish bones has been described in detail elsewhere (Leach and Davidson, 1977; Leach, 1979) and need only be summarised here. (1) The fishbone is sorted into Dentary Articular Quadrate Maxilla Premaxilla Special bones (particularly diagnostic) Residue and rebagged. (2) Dentaries (and each anatomical class in turn) are tipped out, sorted, and rebagged by species using the comparative collection. The identifications are written on the bag labels and entered in forms for the site. (3) Minimum numbers are calculated for each unit in the assemblage (in this case spits), taking no account of mismatched left and right pairs. A number of special bones, such as spines of Biodon sp., occur in large numbers on each fish, and although they are counted, they only give a minimum number of 1 for the assemblage unit, regardless of total number. Experience over the years identifying fish bone from many archaeological sites has consistently indicated that quadrates and articulars add very little to the overall minimum numbers in cases where a reasonable assemblage is available for study. Consequently, although these bones are still separated and rebagged, they only need to be identified in cases of small assemblages. The Kapingamarangi assemblages were quite sizeable, and these bones were therefore not identified. In addition, time prevented the identification of the maxillae at this stage. It is doubtful whether these will ultimately make any difference to the results presented here. The special bones were of considerable importance. In particular, pharyngeal clusters of Scaridae, -115Coridae, etc. quite commonly yield far higher minimum numbers than those from the standard cranial anatomy. The Kapingamarangi assemblages were no exception to this rule. In addition, the characteristic terminal caudal vertebrae of Thunnidae and Katsuwonidae were far more numerous than cranial bones. Spines were the main portions indicating the presence of Balistidae and Diodontidae. For the Kapingamarangi fish bone, this general identification procedure was very successful, and bones from only two or three species could not be matched to genus level, at least, in the . comparative collection. Identifications of course vary in level of certainty from one family to another. For instance, members of the family Scaridae are very difficult to identify to genus, let alone to species. In other families, on the other hand, many species and genera have quite diagnostic features, and more precise identifications can therefore be made. The identifications written on bags often refer to the closest species available in the comparative collection. For instance, Holoaentrus cf. ruber should be understood to mean that this species is the most similar amongst available material, and that the genus is fairly certain. By contrast, cf. Holoaentrus ruber implies that although this is the most similar specimen, the bone may belong to some other genus, for example Ostiahthys. This method gives somewhat uneven initial results through the taxonomic order, but it is possible to pool these later into families, for example, for comparative purposes. It is worth recording the more precise identifications as one proceeds, as these can also be of specific use later. It will be noted that specimens of Thunnidae and Katsuwonidae could not be distinguished. This is because identifications largely relied on vertebrae and tail elements, and although these are distinctive of the two families, they are very difficult to identify further. Elasmobranchs have not yet been identified. This will eventually require X-rays of the vertebrae. At this stage a minimum number of 1 is entered for any spit, regardless of how many vertebrae or teeth were present. The Kapingamarangi fish bone was processed in this way, and the basic results have been pooled to genus level and tabulated below. The taxonomy throughout follows that of Munro (1967) and the family numbers are consecutive from 1 = Orectolobidae. No attempt has been made at this stage to work out size frequency characteristics of the more numerous fish, but several particularly odd features were noted during analysis. Firstly, a few very large specimens of Caranx ?sp. were in the assemblages. Secondly, more than half of the specimens of Lethrinus ?sp. identified were enormous in comparison with anything in the comparative collection. Representatives of family Belonidae (Long toms), on the other hand, are very small specimens, probably indicating that patches of sea grass (their main habitat where they catch small fish) were never more extensive on Kapingamarangi than they are today. Finally, almost all specimens of families Thunnidae/Katsuwonidae in the assemblages are rather small, and some are very small. TABLE 12. BASIC RESULTS OF FISHBONE ANALYSIS, TIROKI Levels Spits IV 1 2 3 Elasmobranchs 1 1 1 Muraenidae cf.Gywnothorax spp. 3 7 6 Holocentridae cf. Holocentrus cf. Ostiohthys esp. ruber esp. murdjan 2 l 5 l 12 l l l 3 4 Sphyraenidae cf. Agrioposphyraena esp. barracuda Thunnidae, Katsuwonidae cf. Neothunnus maoropterus cf. Katsuwonus pelamis and 2 Carangidae cf. Caranx e s p . ignobilis3 melampygus and sexfasaiatus cf. Elegatis bipi-nnulatus Epinephe1idae cf. Epinephelus/Cephalopholis cf. Pleotropoma spp. Lutjanidae cf. Lutjanus cf. Maoolor spp. spp. Nemipteridae cf. Monotaxis Lethrinidae cf. Lethrinus Mullidae cf. Parupeneus granooulis spp. spp. 1 7 spp 25 2 24 6 5 5 13 5 23 9 6 1 Table 12 continued Levels Spits IV 1 2 Coridae cf. Eerrrigymnus melapterus cf. Cheilinus spp. Scaridae Siganidae cf. Sigarcus 1 19 35 spp. Diodontidae cf. Viodon 1 1 Balistidae Total 5 6 7 III 8 9 10 11 12 II 13 14 15 16 I 17 18 19 1 5 1 1 19 3 Total 1 1 2 3 1 1 1 1 87 1 2 e s p . hystrix 4 1 2 Acanthuridae cf. Naso e s p . bvevirostris Tetrodontidae cf. Arothron 3 4 3 1 2 5 8 1 I spp. 4 3 1 1 4 7 4 1 82 124 108 14 1 1 1 7 7 9 1 14 6 5 4 5 1 4 1 1 1 1 1 10 M 22 ' 392 TABLE 1 3 . BASIC RESULTS OF FISHBONE ANALYSIS, MURI-HALAU (KA2), MINIMUM NUMBERS Levels Layers IV Elasmobranchs Muraenidae c f . Gyrmothorax 1 2 1 1 3 4 1 spp. 2 3 5 1 II 7 6 1 1 4 2 Holocentridae cf. Holocentrus spp. Thunnidae/Katsuwonidae Carangidae c f . Cavanx s p p . c f . Elegatis biipinnulatus Epinephelidae c f . Epinephelns/Cephalopholis c f . Pleatropoma spp. Lutjanidae c f . Lutjanus spp. Nemipteridae c f . Monotaxis granooulis Lethrinidae c f . Lethrinus spp. 1 1 1 1 1 3 1 1 4 1 1 2 5 2 5 1 1 1 2 3 1 1 1 1 1 3 3 1 2 1 Coridae cf. Cheilinus spp. Scaridae 1 1 1 3 1 3 4 2 1 1 1 1 1 1 Acanthuridae c f . Naso s p p . Diodontidae c f . Diodon spp. 1 Balistidae 1 1 4 1 1 1 1 0 8 19 26 20 7 13 Total TABLE 14. BASIC RESULTS OF FISHBONE ANALYSIS FROM PUTAU (KA3), MINIMUM NUMBERS Levels Spits 3 III 4 5 1 1 4 4 IV 1 2 Elasmobranchs Muraenidae cf. Gymnotkorax Belonidae cf. Tylosurus spp. esp. 1 6 1 7 1 3 2 8 II 9 10 11 12 13 14 15 16 1 1 5 3 1 1 1 1 1 1 2 1 1 1 indicus 1 Holocentridae cf. Ostiahthys cf. Holoaentrus esp. murdjan esp. ruber Thunnidae/Katsuwonidae cf. Neothunnus macropterus cf. Katsuwonus petamis 1 1 Nemipter id ae cf. Monotaxis esp. sebae 4 1 1 1 39 1 5 1 2 14 5 2 3 18 1 1 23 2 spp. 2 2 1 2 1 2 1 3 1 1 2 1 3 3 1 1 1 2 1 vivescens 2 1 1 1 1 1 2 1 1 1 1 1 2 1 1 1 1 19 8 2 5 5 3 1 2 2 1 3 1 6 1 1 1 1 3 4 39 18 2 1 1 1 1 1 1 18 3 2 2 2 17 2 1 1 coaerulaureus granoeulis H I Carangidae cf. Caranx spp. c f . Caranx e s p . ignobilis c f . Elegatis bipinnulatus Lutjanidae c f . Lutjanus esp. c f . Macolov s p p . c f . Aprion esp. 1 1 1 1 and Epinephelidae c f . Epinephalus/Cephalopholis spp. c f . Plectropoma 1 2 1 Total 4 1 Exocoetidae Caesiodidae c f . Caes'io 1 1 1 I 17 18 19 20 21 1 1 2 2 2 1 1 1 Table 14 continued Levels Spits Lethrinidae c f. Le thrinus IV 1 2 3 4 III 5 6 7 spp. 8 II 9 10 11 12 13 14 15 16 4 1 2 I 17 18 19 20 21 Total 14 3 Mullidae cf. Parupeneus Coridae c f . CheiVinus spp. spp. Scaridae 3 1 4 1 1 2 1 1 1 2 1 1 11 1 1 1 20 3 4 3 5 3 8 1 8 4 1 1 1 2 Acanthuridae cf. Naso spp. 76 3 Scorpaenidae cf. Soovpaena cardinal-is Diodontidae c f. Diodon spp. Tetrodontidae cf. Arothron Balistidae Total 1 to o 1 1 1 1 1 1 1 1 1 10 1 spp. 6 6 37 1 2 2 1 24 15 6 11 12 6 2 1 1 2 1 4 42 11 59 20 12 15 26 34 20 16 48 15 3 1 3 6 389 I TABLE 15: BASIC RESULTS OF FISHBONE ANALYSIS FROM NGEIHO-HEREU (KA4), MINIMUM NUMBERS Levels Spits Elasmobranchs IV 1 2 3 III 4 5 1 1 1 1 1 1 6 7 II 9 10 11 12 13 14 15 16 17 18 19 8 1 1 1 1 1 1 1 1 1 Total 1 16 Muraenidae cf. Gymnothorax Angui11i forme s spp. 2 3 1 1 3 4 ? Family, species A ? Family, species B ? Family, species C 4 1 4 2 4 1 2 1 2 33 11 2 1 11 1 2 1 Belonidae cf. Tylosurus Hoi oc en t r id ae cf. Holocentrus cf. Ostiahthys indious 1 e s p . ruber spp. Acanthocybiidae cf. Aoanthoaybium 4 2 3 2 1 1 2 1 1 1 18 1 1 M solandri Thunnidae/Katsuwonidae cf. Neothunnus macropterus cf. Katsuwonus pelamis 1 and 1 2 Carangidae cf. Caranx melampygus cf. Caranx e s p . ignobilis cf. Carangoides oblongus cf Carangoides latiaaudis 1 2 1 1 2 1 1 2 1 1 1 1 1 2 1 3 1 2 21 1 1 3 1 4 9 1 1 1 1 1 Epinephelidae cf. Epinephelus/Cephalopholis cf. Pleotropoma spp. L u it j a n i d a e cf. Lutjanus cf. Aprion i e s p . sebae se virescens spp. 2 33 1 22 22 1 1 22 1 1 4 4 1 22 3 11 1 33 11 22 3 3 3 3 4 4 3 3 2 22 2 1 22 11 11 1 1 1 1 1 1 1 1 3 34 4 2 2 2 2 33 1 11 1 1 39 15 19 3 ' Table 15 continued Levels Spits IV 1 2 3 4 III 5 6 II 9 10 11 12 13 14 15 16 17 18 19 7 8 1 1 3 Total Nemipteridae cf. Monotaxis granoaulis l l 1 2 3 spp. 2 1 1 1 1 1 1 1 1 2 19 1 •2 Lethrinidae cf. Lethrinus 10 Mullidae cf. Mulloidichthys auriflamma l Kyphosidae cf . typhosus spp. 1 Coridae cf. Hemigymnus cf. Cheilinus spp. Scaridae 2 1 melapterus 2 1 4 1 1 6 1 3 2 4 13 5 2 1 2 12 1 7 1 2 2 3 1 1 54 Siganidae 1 cf. Siganus spp. 4 Acanthuridae 1 Diodontidae cf. Diodon spp. Tetrodontidae cf. Arothron spp. 1 1 1 1 1 1 2 Balistidae cf.Rhinecanthus aeuleatus cf .Meliehthys buniva cf.Pseudobalistes flavimarginatus ToiaZ i 3 26 31 2 11 1 9 18 38 2 1 1 1 2 1 3 15 1 25 17 15 29 29 22 21 20 10 9 330 M M I -123- APPENDIX 4 KAPINGAMARANGI CRUSTACEAN REMAINS Graeme Mason The Kapingamarangi crustacean remains all belong to the Order Decapoda Suborder Reptantia, the super sections Macrura (crayfish), Anomura (hermit crabs etc.), and Brachyura (true crabs) all being represented. The most frequently occurring parts comprise cheliped and pereopod (pincher and leg) fragments among which the dactylus is commonest. By anatomy the remains are distributed thus: Cheliped Left dactylus Right dactylus Left propodus Right propodus 5 11 0 3 Pereopod All dactyli Miscellaneous 11 4 Mandible Fragments 2 Miscellaneous fragments 11 Total 47 The absence of comparative material and, at present, a number of taxonomically important works, coupled with the non-existence of modern reviews of many major groups (particularly the Xanthidae, the largest single crab family in the Indo-Pacific and one which contains many edible species) has severely limited identification to a level well below what the writer would wish, so this report must be regarded as preliminary only. It should also be remarked that chelipeds and pereopods are rarely used as taxobases. Super Section Macrura Family Palinuridae (spiny crayfish): Super Section Anomura Family Coenobitidae Birgus latro (coconut crab): 2 mandibles tip of propodus,right cheliped -124Coenobita sp. (hermit crab): L cheliped dactylus tip R cheliped propodus Coenobitid: 3 dactyli Super Section Brachyura Family Xanthidae (Black-fingered crabs) Xanthid spp.: 7 cheliped dactyli The balance of the material may include some Portunid crabs (swimming crabs) and a few small species whose identity is at present in doubt. The absence of the large edible terrestrial Gecarcinids (especially Cardisoma spp.) and Grapsids (Sesarma spp.Geograpsus spp.) is noteworthy as they are known at present on the island. The scarcity of Birgus remains is disappointing in view of the light they may cast on human movement in the Pacific area. In this context, Cardisoma longipes (Gecarcinidae) has a rather patchy Pacific distribution which may be perhaps partly explained (it is large and palatable) by human transport and maybe, over exploitation locally. Use of terrestrial, littoral and sublittoral crustaceans is clearly demonstrated. The distribution of crustacean remains is shown in the table below. Table 16. Crayfish Coconut crab Hermit crab Xanthids Crustacean Crustacean remains from Kapingamarangi IV III II I K M KA2 KA3 KA4 KAl KA2 KA3 KA4 KAl KA3 KA4 KAl KA3 2 1 2 1 1 7 2 1 4 2 3 1 5 9 4 2 -125- APPENDIX 5 ANALYSIS OF ARCHAEOLOGICAL SHELL FROM KAPINGAMARANGI Simon Holdaway The modern comparative material collected by Leach and Ward during their expedition was sorted into taxonomic groups and identified with the assistance of Cernohorsky of the Auckland Museum. This material, together with the existing comparative collection of tropical shells already in the Otago Archaeological Laboratories, formed the basis for identifying the archaeological residues from Kapingamarangi. Once the shells had been separately bagged, each bag was sorted into pieces relevant to minimum number estimates, and residue. The latter was rebagged for storage. Diagnostic material was sorted into eight categories and rebagged. Donax spp. Lambis spp. Tvidaona spp. Turbo spp. Strombus/Mitra/C'onus/Trochus/Drupa spp . Other Gastropods Other Bivalves Cypraea spp. Each of these categories was later separately treated to obtain accurate identifications, and minimum numbers. Some comments on the main individual shells found are appropriate, since their treatment varied somewhat. Turbo argyrostomus: Here, both the number of opercula and the number of shell apexes were counted and recorded. The minimum number per square and spit was the larger of the two numbers. Operculae were counted when half or greater was present. It is assumed that all individuals are Turbo argyrostomus, although it is possible that some of the operculae are from other Turbo species. Rhinoolavis asper: Minimum numbers were calculated from the apex. Again, whole shells were more easily identified. Strombus luhuanus: The apex was used for minimum number calculations; all other fragments were ignored. Lambis spp.: Only the apex was used for minimum numbers. Separate fragments containing siphonal appendages which allow -126identification, but not including an apex were not included for minimum number calculation. It is noteworthy that there are very few complete shells of Lambis spp. Cypvaea spp.: All shells were identified to species level where possible. Only complete or near complete shells were used for minimum number calculation. Most specimens were very small, and the majority were probably not collected as food. Drupa spp.: The apex was not always very clear, so only complete or near complete shells were used for minimum numbers. Mitra spp.: The apex was used for minimum numbers. Conus spp.: The apex was used for minimum number calculation. Many of the specimens, especially the smaller ones, are unlikely to have been of economic importance.All, however, were identified where possible. As species identification relies on the aperture and siphonal canal shape, there is a bias toward complete shells in the minimum number calculations. Individuals identified only to the level of Conus sp. were not used for minimum numbers, as these were invariably worn, consistent with a beach sand origin. Terebra spp.: The apex was used for minimum numbers. The whorl shape is quite diagnostic, so shell fragments could also be identified. Other Gastropods: The majority of shells in this category were Nerita spp. Whole shells or nearly whole shells were used for minimum numbers. Most of the Nerita were identified to genus level only, as there are many species of Nerita for which comparative material was not available. Minimum numbers of other gastropods were assessed from either the apex or whole shells. Atrina sp. and Pinotada sp.: Very few fragments of these were found, but because of their importance were booked as present in any spit in which even tiny pieces were recovered. This includes the few artefactual pieces. Tridaona spp. and Hippopus sp.: Identification to genus level was achieved for nearly all fragments. Where complete valves were present, species identification and valve siding were possible. Minimum number estimates are based on valves with a hinge present. Donax spp.: Identification was to genus level only. Minimum numbers were calculated as half the number of shells present without attention to left or right valves. Broken valves were counted only when the hinge was present. Other Bivalves: All bivalves were identified to species where possible. Minimum numbers were calculated as half the total number of valves with hinges (or near complete valves). In the case of broken valves, an individual was counted when the majority of the hinge was present. It should be noted that dead specimens of Glyoydonta mavioa are common in coral sand and not all those in the sites are considered to be economic in origin. -127The full set of tables of these identifications is a sizeable document, and for convenience only a summary is presented here. In Table 17, the actual species identified are listed in their appropriate taxonomic position. In Tables 18-21 the archaeological shells are grouped according to family for ease of comparison and economic assessment. In view of this summary treatment here, some additional comments should be made, It is notable that no specimen of. Troohus nilotious was recovered on Kapingamarangi, either archaeologically, or in the modern collection. Troohus sp. generally appeared to be rare. Figures for Strombidae, of course, include both Strombus spp. and Lambis sp. Of these, the former were four times as numerous, In the family Tridacnidae, only two species are positively identified - Tridaona maxima and Hippopus hippopus. Apart from two small fragments of Spondylus varians3 all material in this family was Spondylus duoalis. TABLE 17: THE SHELLFISH SPECIES IDENTIFIED FROM KAPINGAMARANGI ARRANGED IN TAXONOMIC ORDER FOLLOWING CERNOHORSKY (1972: 25-30) Modern 27 Trochidae Troohus Troohus Troohus maoulatus histrio * * sp. 2 9 Turbinidae Turbo argyrostomus 34 Neritidae Nerita polita Nerita undata Nerita sp. 81 C e r i t h i i d a e Cerithium aluoo Cerithium oolumna Cerithium nodulosum Rhinoolavis asper Rhinoolavis sp. 90 Strombidae Strombus luhuanus Strombus mutabilis Strombus gibberulus gibbosus" Lambis chiragra ohiragra 102 Cypraeidae Cypraea depressa Cypraea argus Cypraea oarneola Cypraea annulus Cypraea• caputserpentis Cypraea sourra Cypraea isabella Cypraea poraria Cypraea teres Cypraea lynx Cypraea testudinaria Cypraea arabioa (juvenile) Cypraea mauritiana (juvenile) Archaeological * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * -128Table 17 continued Cypraea sp. Monetaria moneta Monetaria sp. 108 Naticidae Poliniaes melanostomus Poliniaes sp. 109 Cassidae Casmavia erinaoeus Malea pomum 110 Tonnidae Tonna perdix 112 Cymatiidae Cymatium nicobaricum Cymatium sp. Distorsio anus , 113 Bursidae Bursa granularis Bursa sp. 116 Muricidae Morula granulata Chiooreus brunneus Drupa morum Thais armigera Thais tuberosa Thais sp. 117 Magilidae Quoyula madreporarum 119 Buccinidae Cantharus undosus 123 Olividae Oliva miniaaea 124 Vasidae Vasum sp. 126 Harpidae Harpa amouretta 129 Mitridae Mitra stiotioa Mitra mitra Mitra litterata 123 Conidae Conus lividus Conus floridulus Conus arenatus Conus ebvaeus Conus ranunculus Conus tulipa Conus sponsalis Conus pulicarius Conus miles Conus flavidus Terebridae Terebra affinis Terebra areolata Terebra sp. Terebra crenulata Terebra maoulata modern * archaeological * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * -129Table 17 continued 135 Architectonicidae Eeliacus variegatus 239 Ellobiidae Melampus flavus 315 Arcidae Barbatia deoussata 321 Pinnidae Atrina (Atrina) vexillum 322 Pteriidae Pinctada margaritifera Pinctada maculata (juvenile) Pteria penguin Pteriidae ? genus 326 Pectinidae G'Ioripallium pallium 33 0 Spondylidae Spondylus ducalis Spondylus varians 354 Carditidae Beguina semiorbiculata 357 C a r d i i d a e Cardium elongatum 358 Tridacnidae Tridacna maxima Tridacna sp. Eippopus hippopus 360 Mesodesmatidae Atactodea striata 365 Donacidae Donax sp. 366 Tellinidae Quidnipagus palatum Scutarcopagia scobinata 367 Psammobiidae Asaphis violascens 370 Semelidae Semele sp. 372 Trapeziidae Trapezium bicarinatum 380 Veneridae Gafrarium pectinatum Glycydonta marica Periglypta reticulata Periglypta puerpera modern archaeological * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * TABLE 1 8 . MINIMUM NUMBERS OF SHELLFISH FROM TIROKI Levels Spits Trochidae Turbinidae Neritidae Cerithiidae Strombidae Cypraeidae Cassidae Cymatiidae Muricidae Magilidae Buccinidae Mitridae Conidae Arcidae Pinnidae Pteriidae Spondylidae Carditidae Cardiidae Tridacnidae Mesodesmatidae Donacidae Tellinidae Psairanobiidae Total IV 1 2 I 3 8 1 14 2 22 3 9 2 4 4 1 4 4 1 1 1 1 (KA-1) III 8 5 2 15 1 31 52 5 2 I 9 10 11 12 46 1 16 57 65 301 1 1 1 3 6 1 15 4 13 II 14 91 3 6 16 56 1 22 12 181 1 3 1 1 1 1 1 2 4 1 1 6 1 1 1 1 1 1 1 2 4 5 25 1 2 3 5 4 4 1 8 1 20 42 62 1 3 18 1 8 13 21 38 68 54 20 2 62 76 378 4 1 12 2 10 2 Tot, 1 2 1 1 1 13 2 8 1 1 34 1 1 1 24 241 22 703 1 1 2 1 1 1 4 2 2 2 1 2 1 12 1 I 17 1 516 15 1 19 20 1 1 1 3 1 3 4 19 2 3 1 1 6 12 1 69 1 2 1 1 E 15 119 4 62 9 36 1 1 3 1 2 o I TABLE 19. MINIMUM NUMBERS OF SHELLFISH FROM MURIHALAU (KA-2) Levels Layers Trochidae Turbinidae Neritidae Cerithiidae Strombidae Cypraeidae Naticidae Cymatiidae Muricidae Magilidae Mitridae Conidae Terebridae Ellobidae Arcidae Pteriidae Spondylidae Cardiidae Tridacnidae Donacidae Psammobiidae Trapeziidae Veneridae Total 1 2 45 2 1 27 7 76 59 1 1 5 1 2 1 3 41 17 1 4 66 1 1 4 1 5 1 IV 4 2 219 17 3 46 1 5 6 36 1 33 5 6 9 8 6 2 3 5 2 1 1 2 401 49 2 21 262 1 1 4 4 6 17 5 III 7 1 52 5 1 3 453 54 2 37 275 1 2 4 4 6 17 5 1 5 1 9 5 11 506 2 1 9 154 1413 16 13 1 i 1 2 2 1 4 1 2 87 1 41 2 63 1 5 2 1 227 143 210 1 2 1 176 2 780 2 40 102 1 43 5 1 6 4 7 450 2 1 8 97 1259 total 3 1 4 56 U) h-1 I TABLE 20. MINIMUM NUMBERS OF SHELLFISH FROM PUTAU (KA-3) Levels Spits iv 1 2 I Trochidae 30 26 56 Turbinidae 2 3 5 Neritidae Cerithiidae 1 1 Strombidae 1 1 Cypraeidae Naticidae Cymatiidae Bursidae Muricidae Magilidae Mitridae Conidae Terebridae 2 5 7 Arcidae Pectinidae Spondylidae 2 1 3 Cardiidae 1 1 2 Tridacnidae Mesodesmatidae Donacidae Tellinidae Psammobiidae Veneridae Total ill 3 4 5 6 7 £ 8 II 9 10 11 12 13 14 15 16 I 2 3 51 142 212 17 10 26 35 139 158 2 2 2 2 8 1 13 3 3 1 2 42 1 2 53 100 5 12 1 1 4 4 2 1 1 5 1 1 4 3 1 1 1 1 4 1 6 7 6 1 2 3 2 1 4 2 2 2 71 26 28 4 3 2 1 5 2 2 2 1 1 1 7 3 10 26 2 7 1 1 1 2 4 3 2 38 1 1 17 4 1 6 1 17 1 1 2 1 6 2 68 13 2 2 4 1 2 64 7 1 4 1 1 1 1 1 3 38 37 75 35 63 19 75 101 293 188 165 283 172 117 1 1 1 1 1 1 10 1 1 10 14 147 860 1 5 40 3 1 2 19 4 38 4 1 2 1 2 1.. 2 3 1 14 3 1 14 1 1 1 7 1 6 1 13 1 10 15 199 1 1 1 1 12 1 5 I 17 18 19 20 21 Z 1 66 1 1 1 37 2 11 1 6 1 1 1 1 3 2 3 9 13 1 1 5 1 1 4 1 10 1 57 222 4 3 16 1 3 90 472 3 13 1 1 1 32 1 2 1 7 1 6 1 1 1 13 1 1 1 59 65 33 178 1260 97 68 93 250 1 6 3 55 1 3 1 98 606 total 13 1527 66 4 121 83 5 4 2 3 2 4 20 3 38 1 14 15 23 1 261 2 16 6 2234 I TABLE 21. MINIMUM NUMBERS OF SHELLFISH FROM NGEIHO-HEREU (KA-4) LeveIs Spits Trochidae Turbinidae Neritidae Cerithiidae Strombidae Cypraeidae Naticidae Cymatiidae Magilidae Vasidae Harpidae Conidae Terebridae Arcidae Pinnidae Pteriidae Spondylidae Carditidae Cardiidae Tridacnidae Donacidae Tellinidae Psammobiidae Veneridae Total 1 27 IV 2 I 3 68 41 4 3 2 III 5 II 6 I 73 483 559 1 1 2 1 3 6 3 9 1 8 9 10 11 1 203 141 263 263 121 1 3 11 1 1 1 2 1 56 19 3 3 5 6 10 1 12 13 14 15 16 57 6 35 5 39 2 21 3 4 1 8 12 1 7 3 2 2 1 1 4 1 5 1 1 1 2 3 1 1 4 4 2 1 1 1 2 1 1 2 2 4 6 3 9 1 1 1 2 26 10 1 16 4 9 1 1 1 1 1 221 155 354 321 154 96 71 1 7 1 7 1 1 3 12 2 28 47 75 5 8 81 505 599 1 5 1 1 . 1 4 11: 32 3 79 59 1 2 1 1 1 10 1 2 1 3 4 E 1 1 1 17 1 1 1 1 3 4 1 1 1 1 1 54 34 10 3 4 1 2 14 95 1 5 4 1 1471 Total 1 1774 34 3 82 68 1 2 1 1 2 10 1 13 0 4 4 1 8 21 104 1 5 4 2145 I h-1 U) I -134- APPENDIX 6 SAND TEMPER IN A SHERD FROM KAPINGAMARANGI W.R. Dickinson A thin section of a sherd from Kapingamarangi atoll in the Caroline Islands was sent for study by B.F.Leach. The temper in the sherd is a volcanic sand of complex composition perhaps derived from somewhere in the Solomon Islands, but the sherd is possibly from ware introduced during this century from Japan. The composition of the sand temper is approximately as follows: 15% quartz grains, 40% feldspar grains, 20% volcanic rock fragments, 15% chert and microgranular rock fragments, 5% pyroxene grains and 5% opaque iron oxide grains. This sand is unlike any observed from Palau, Yap, or Truk. The presence of quartz implies exposure of granitic or metamorphic rocks, or sediments derived from such sources, in the provenance. Petrographic Report WRD-8 3 -135- APPENDIX 7 SAND TEMPER PETROGRAPHY OF A SHERD AND SOME JAPANESE TILES FROM KAPINGAMARANGI C.A.Landis A solitary sherd collected from Kapingamarangi by B.F.Leach was examined by W.R.Dickinson. In his report (Appendix 6 ) , Dickinson described the sherd temper as being of complex volcanic origin, possibly derived from the Solomon Islands. In addition he pointed out that the presence of quartz in the sherd (estimated 15%) "implies exposure of granitic or metamorphic rock, or sediments derived from such sources, in the provenance". In view of its exotic temper petrography, Dickinson also suggested the possibility that the sherd could be "ware introduced during this century from Japan". Japanese tiles do occur on Kapingamarangi. In order to test the possibility that the sherd may be such recent material, four Japanese tiles from the island were studied petrographically and compared with the Kapingamarangi sherd. The four tile samples (79.425 to 79.428 inclusive) are nearly identical. Their temper consists almost entirely of quartz and alkali feldspar (estimated proportions 60:40 to 40:60). The quartz occurs as very angular single crystals up to 1 mm in maximum dimension, cracked but unstrained, and generally coarser than the associated feldspar. Much of the feldspar is albitic in composition and is marginally vitrified. The more highly vitrified grains show no twinning and may be a potassic feldspar. Minor amounts of bleached biotite and rhyolitic lithic fragments are also present. In addition, two zircon grains were identified. The temper of the Kapingamarangi sherd (79.1) is essentially as described by Dickinson. In addition I would point out that volcanic lithic grains (estimated 25% of total temper) are largely siliceous in nature (rhyolite-dacite) and that some of the associated quartz and feldspar is surely of the same provenance. Other mineral grains (calcium-plagioclase and augite) are also common and imply a more basic igneous provenance. A source in the Solomons or New Guinea seems entirely plausible. It is very likely that the sample originated in a Pacific margin environment. -136- APPENDIX 8 PHYSICAL ANALYSIS OF KAPINGAMARANGI POTTERY Michiko Intoh and Foss Leach The single piece of pottery from Kapingamarangi is a rim sherd from a large, near vertical-sided bowl with a plain rim. The sherd varies in thickness from 10 to 12 mm. From the sherd curvature, the internal diameter of the original rim is estimated as about 16 cra. The paste is 'fairly fine textured, and the sherd shows flashes of red and black, characteristic of a firing atmosphere which is only partly controlled, such as occurs in an open fire where both oxidising and reducing conditions prevail, This feature gives the sherd the appearance of a typical specimen of Oceanic pottery. The sherd is exceptionally hard for an example of Oceanic pottery. In this respect it is more similar to modern field tiles or other European pottery vessels, and this origin must be considered. Two obvious possibilities are early Spanish pottery or a piece of Japanese waste-pipe or field tile. Sherds from such pipes are quite common in Micronesia, and were exported from Japan during the period of Japanese administration (see Figure 42). Spanish pottery occurs in Pacific island archaeological sites, but its characteristics are not well known. To shed further light on this, water absorption tests were carried out on the sherd and other specimens for comparison. Total volume of sand temper (%) was estimated by point counting thin sections. All work was rigorously cross-checked against known standards to achieve maximum reliability. The results are given in the table below. Table 22. Physical Analysis of Pottery. Specimen Dokan 79.425 Dokan 79.426 Dokan 79.427 Dokan 79.428 WA 11.13 10.93 12.16 11.88 11.53 ±0.29 AP 22.27 22.02 24.09 23.58 22.99 ±0.50 BD 2.00 2.01 1.98 1.98 1.99 ±0.01 SG 2.57 2.58 2.61 2.60 2.59 ±0.01 ST 26.8 25.2 32.0 25.0 27.25 ±1.63 WT 11 11 11 11 Kahula Spanish 18.95 34.01 1.79 2.72 18.8 12 Kapingamarangi 79.1 10>5Q Mean and SE of mean 21>45 2>Q4 2>6Q 32<6 1Q_12 WA = water absorption (%), AP = apparent porosity (%), BD = bulk density (gm/cc), SG = specific gravity, ST = sand tempering (% v/v), WT = wall thickness (mm) -137Is the sherd from a Japanese field tile ? The water absorption tests show the specimen to have rather similar characteristics to pieces of dokan waste-pipes collected on several islets on Kapingamarangi. The degree of sand tempering is also similar. However, the general appearance of these tiles is completely different. Firstly, dokan are glazed on both inner and outer surfaces; secondly, the curvature is quite different, with an internal diameter of 7 cm; thirdly, the firing conditions of dokan are highly controlled and give a uniform red colour. Origin from a dokan tile can be ruled out with certainty. Is the sherd from a piece of Spanish pottery ? Unfortunately very little is known of the seventeenth and eighteenth century Spanish pottery in the Pacific. Of material examined at first hand, the glazed potteries can be ruled out, leaving a uniform red, fine paste pottery,similar to the well known 'flower-pot' type of modern European pottery. A sherd of this pottery from Kahula in Taumako is included in the Table above for comparison. It is clearly quite a different ware, being much more porous than the Kapingamarangi sherd. In addition, Spanish pottery was kiln fired, giving a well controlled firing atmosphere and surface appearance. Spanish origin can also be ruled out. Is the sherd a specimen of Oceanic pottery ? The water absorption results for the sherd were not especially typical of known Oceanic potteries. The hardness of the sherd is reflected in the low apparent porosity of 21.45%. Pacific island potteries range from mean porosity values of 28% to nearly 50% for different areas, with Micronesian pottery generally being in the upper half of this range. The closest pottery to the Kapingamarangi sherd is that from Yanuca in Fiji, rare specimens of which can be as low as 20.16%. The values for other characteristics in the Table are well within the ranges for many Pacific potteries. In conclusion, based on these simple tests the Kapingamarangi sherd shows itself to be different from known Japanese or Spanish pottery which occurs in the Pacific. It is from a hard and well fired pottery vessel which has been fired in a relatively uncontrolled atmosphere. It is unlike the biscuit wares of Micronesia, and is more like an unusually hard specimen of pottery from further to the south. -138- APPENDIX 9 KAPINGAMARANGI ROCKS Graeme Mason Being a coral atoll of some geological maturity, Kapingamarangi is unlikely to have washed onto its shores locally derived stone material, with the exception of young, well cemented limestones. The source of non-bioclastic hard rocks is assumed to be either material washed ashore in drift wood or places beyond the island tapped by human agency. Differentiation of material according to these alternatives has not been attempted as only in one instance (79.551) is there any evidence of human modification of a natural clast. In shape, the clasts (ranging in size from 2 - 1 5 cm, most commonly about 5 - 6 cm in greatest dimension) are generally well rounded, but include a fair number which are sub-angular, possibly as a result of heat fracturing and subsequent weathering. In most cases the rocks have been identified only on hand specimen characteristics, backed by cursory examination of some thin sections. No detailed petrography has been attempted. LIMESTONES 79.547. Ka-2, layer 2. Burnt pebble of dense compact carbonate, grey-white aphanitic, homogeneous; shows worm borings. 79.548. Ka-2, layer 2. Yellowish white, dense, finely granular carbonate pebble, showing irregularly nodulose surface. Under the hand lens, dark sub-angular opaque particles to 0.25 mm set in grey-white aphanitic groundmass. Microscopically, the texture suggests that the rock may be a concretionary nodule developed in the lower part of a soil profile. 79.552. Ka-2, layer 1. Pale buff-white compact carbonate rock, finely granular, containing black opaque grains to 0.25 mm. Sedimentary texture. 79.554. Ka-3, spit 20. Mineralised coral fragment showing secondary calcite and solution pitting. 79.556. Ka-3, spit 20. Grey-white dense aphanitic limestone. -13979.557. Ka-3, spit 20. Grey-white dense aphanitic limestone METAMORPHIC ROCKS 79.64. Part of large boulder washed ashore with log in 1978. White granular compact rock showing secondary veins containing minute well developed quartz crystals, minor amounts of dark opaque grains. In thin section, well rounded quartz grains are visible showing dusty margins and clear overgrowths in optical continuity with the grains. Strain extinction present in many grains, some boundaries sutured. Biotite present in minor amounts. 79.83. Surface find, Sakenge Islet. Fine grained, light grey-black mottled rock with fine white veins prominent, sulphide particles (0.25 mm) visible. Mylonitic structure present. In thin section, the specimen is a slightly crushed, poorly sorted greywacke with clasts up to medium sand size. Feldspars clouded with ragged plumes of secondary minerals developed at their edges. The rock appears to be a mylonized greywacke. 79.296. Part of modern pandanus anvil, Pepeio Islet. Burnt,light grey,compact,coarsely granular rock, weathering to rusty brown to chocolate brown, predominately sub-angular to sub-rounded quartz grains up to about 2 mm, numerous minute black opaque grains visible. Linear concentrations of small dark minerals present. In thin section, shows granoblastic texture, quartz grains comprising bulk of the rock. The boundaries of the sub-eguant grains are sutured and contain indeterminate dark brown opaque material, occasional irregular patches of chlorite.The rock is a metamorphic quartzite. VOLCANIC ROCKS 79.546. Ka-2, layer 5. Fine grained, grey to black aphanitic rock, vesicular, in thin section, an intermediate to basic volcanic rock showing tablets of plagioclase with concentric composition zoning at the edges, the cores of which are eroded and with many inclusions. Smaller laths of feldspar are set in a turbid, possibly glassy groundmass making up 10 to 20% of the section. 79.43a. Ka-2, layer 5. Fine grained, green-grey rock, feldspar and dark mineral grains visible but not determinable in hand specimen. Weathers to light grey. In thin section, ragged patches of opaque minerals are present. Plagioclase feldspars fall into two size ranges, one set 0.3 mm long and the other set 0.05 mm. The rock is holo crystalline and appears as an intermediate to basic volcanic rock. 79.43b. Ka-2, layer 5. Dark grey aphanitic rock, feldspar laths to about 1 mm long, dark blebs to about 0.25 mm visible. Weathers to light grey. Thin section shows two sets of plagioclase laths at 0.3mm and 0.05 mm long set in a groundmass of dusty, finely crystalline material difficult to resolve. Rock appears to be an -140altered volcanic of basic composition 79.549. Ka-2, layer 2. Grey-brown rock showing yellow-green circular inclusions up to 2 mm, fine laths of feldspar visible. In thin section, an interlaced mesh of plagioclase crystals packed interstitially with dusty and opaque weathering products can be seen. A volcanic origin is probable. 79.550. Ka-2, layer 7. Dense, black, aphanitic rock containing sparse, randomly oriented feldspar laths and light coloured, sub-equant grains. Weathers to an orange-brown friable material. In thin section the rock appears to be an indurated crystal tuff. 79.551. Ka-2, layer 7. Dark grey to black, mottled, aphanitic rock with sparse, minute vesicles. In thin section, augite phenocrysts up to 0.3 mm diameter are set in a noncrystalline mesh of very fine feldspar laths about 0.05 mm long. Opaque minerals about 3 to 5%. A basic volcanic rock. 79.553. Ka-2, layer 1. Compact, dark grey to black rock, finely granular texture with some lineation visible and small pockets of yellow alteration products. In thin section, an intermediate to basic volcanic rock showing a very fine grained, holocrystalline texture, with some orientation of the small (about 0.06 mm) feldspar laths, which show a marked structure with interstitial tabular and equant crystals of hornblende. 79.555. Ka-3, spit 21. Dark greenish-grey, compact rock, showing small (0.25 mm) blebs of black material, weathering to brown-grey material. In thin section, an intermediate to basic volcanic rock showing tabular plagioclase up to 0.6 mm long set in very fine grained, turbid matrix, with abundant small (0.04 mm) feldspar tablets. 79.294. Washed up on Werua Islet in 1980. Grey-black, compact rock, weathering to pale reddish-grey. Finely vesicular texture, groundmass aphanitic, small phenocrysts of black pyroxene or amphibole visible. In thin section, an intermediate to basic volcanic rock containing abundant tabular plagioclase with concentrically zoned clear overgrowths up to 0.6 mm in diameter, accompanied by very pale green hornblende phenocrysts up to 0.4 mm, all set in a holocrystalline groundmass of small tabular feldspar crystals and minute opaque grains. 79.295. Surface find. Dark blue-grey, compact rock, fine grained texture, indistinct lineation marked by dark material. In thin section, an altered volcanic rock showing highly oriented, essentially sub-parallel feldspar laths up to 0.3 mm long and abundant interstitial holocrystalline material containing many minute, euhedral prisms of pyroxene. Opaque minerals comprise about 10% of the slide. -14179.44. Ka-2, layer 4. A dark brownish-black, compact, vesicular rock. Matrix fine grained with sparse phenocrysts of pyroxene up to 1 mm long. In thin section, this coarse grained basic volcanic rock largely comprises a coarse intergrowth of plagioclase feldspar and pyroxene. -142- APPENDIX 10 THE HUMAN BIOLOGY OF KAPINGAMARANGI Phil Houghton Some useful data on the living population of Kapingamarangi has been provided by Schlaginhaufen(1929), and Emory(1965:64-79). Emory's data relates to measurements and observations made by Buck in 1947 on 59 adult males(Buck,1950:4), whilst Schlaginhaufen studied some 34 adult males in 1908. Schlaginhaufen obtained a mean stature of 1711 mm, and Buck 1723 mm, with a sitting height ratio of 53.2. These are tall people in comparison with neighbouring Micronesian and Melanesian islands(see Table 23), and in this regard are characteristically Polynesian. The high sitting height ratio is again Polynesian, showing relatively long trunk, and short lower limbs. The other traditional anthropometric measurements obtained by these workers, and the various indices derived therefrom, have little biological meaning, and it would not aid clarification of biological relationships within Oceania to repeat them. Schlaginhaufen collected hair samples, and also made general comments on hair form amongst the population. The women showed straight hair, as did the children, with men showing a tendency to strongly waved hair; and he comments that truly curly hair, that is the corkscrew type, was not seen on Kapingamarangi. Again, these comments are in harmony with the suggested Polynesian affiliations of the islanders, though of course such a description of hair form could emerge from more northern Micronesian islands as well. Photographs of 14 individuals are supplemented by two group photographs depicting a further 20 people in Schlaginhaufen's work(1929). With broad lips and noses, long flat-sided faces with vertical profile and mandibular prognathism revealing the characteristic Polynesian form of jaw growth, these people may be safely categorised as being of predominantly Polynesian stock. Miller studied the blood groups of 87 males and 82 females, all believed to be of pure Kapingamarangi ancestry, during the 1950 expedition(Miller, 1953). His data are presented in Table 24. The absence of group B supports the Polynesian relationships. The prehistoric human material obtained from the present Kapingamarangi excavations is small in quantity. Some six individuals are represented, mostly by a few fragments of long bones, but with rather more extensive remains of a mature adult male, and jaw fragments of a six-year old. Fortunately, these few pieces include some of the more informative parts of the skeleton from the point of view of biological relationships within Oceania. -143The mature adult male shows the characteristic Polynesian rocker jaw, from which it can be deduced that his general cranial morphology was Polynesian. Femoral shafts from four people are present, and all show the platymeric flattening characteristic of Polynesians. The average platymeric index for these four is 73, which compares with Eastern Polynesian values of 65.2 for the New Zealand Maori, 71.6 for Hawaiians, and a value of 77.5 from Taumako, a Polynesian outlier in the Solomons. In general, Homo sapiens shows an index of over 80. While the biological basis for this distinctive Polynesian femoral feature has not yet been determined, it may be assumed that the lower limb morphology in general for this group was Polynesian. A permanent upper first molar from the child shows a Carabelli cusp and a hypoconulid, and it has a crown area of 131. However, no reliable conclusions concerning relationships can be drawn from a single tooth. None of the long bone fragments showed evidence of yaws, a condition known to have been rife on at least one Polynesian outlier, Taumako(Houghton, n . d . ) , and the skull fragments of the adult male showed no bony evidence of anaemia, in the form of spongy hyperostosis or cribra orbitalia. A second metacarpal from this man had a cor ical thickness of 5.4mm, suggesting a wellnourished individual. The remaining skeletal fragments were unsuited to radiological examination. TABLE 23: AVERAGE STATURE FIGURES(cm) FOR KAPINGAMARANGI AND VARIOUS OTHER OCEANIC GROUPS(Extracted from Shapiro,1933) POLYNESIAN OUTLIERS Kapingamarangi Nukuoro Taumako Takuu Nukumanu Ontong Java 171.1 171.4* 175.lt 165.7 163.9 163.7 EAST MICRONESIA Kiripati Marshall Islands 169*. 5 163.6 POLYNESIANS Tonga Samoa Society Islands Fiji Maori Hawaii 173.0 171.7 171.4 170.9 170.6 169.5 ATOLL POLYNESIA Funafuti 167.1§ -144Table 23 Continued WEST AND CENTRAL MICRONESIA West Carolines Woleai Ponape Kusaie Mortlocks Truk Yap Palau 163.8 162.0 163.2 162.3 164.4 160.2 163.5 161.7 MELANESIANS Santa Cruz St Matthias Tanna New Caledonia 160.3 160.1 164.5 166.4 NB: * taken from Schlagenhaufen(1929:224) t from Houghton(n.d.) § from Hedley(1897:235) TABLE 24: BLOOD GROUP ANALYSIS OF KAPINGAMARANGI (after Miller, 1950:41-42) AB Males Females Totals A 40 32 B 72 - - 0 47 50 97 Total 87 82 169 -145- REFERENCES Anell,B. 1955 Contribution to the history of fishing in the southern seas. Studia Ethnographica Upsaliensia 9 Anon. 1944 Kapingamarangi. Joint Intelligence Centre, Pacific Ocean Areas Information Bulletin(JICPOA) 26-44. 1972 Ponape Almanac. Instructional n .d, Pilot chart of the Pacific ocean.#1401W,#2601W, Anon Serves Centre,Ponape Anon, U.S.Hydrological Baker,R.H. 1951 Office. The avifauna of Micronesia, its origin,evolution, and distribution. University of Kansas Museum of Natural History 3(l):l-359. Bayard,D.T, 1976 The Cultural Relationships of the Publications> Polynesian Outliers. University of Otago, Studies in Prehistoric Anthropology 9. Beasley,H.G. 1928 Pacific Island Records: Fish hooks. Seeley,Service and Co., London. Biggs,B. 1979 Proto-Polynesian word list I I . U n i v e r s i t y Working Papers in of Auckland Anthropology. Buck,P.H. 1950 Material Culture of Kapingamarangi. B.P.Bishop Museum Bulletin 200. Carroll,V. 1965 An outline of the structure of the language of Nukuoro. Journal 74(2) :192-226. Carroll,V.(Ed.) Adoption 1970 of in Eastern the Polynesian Society Oceania.ASAO Monograph 1. University Press of Hawaii. Cernohorsky,W.O. 1971 Marine shells Sydney. of the Pacific J.Pacific Publications, Cernohorsky,W.0. 1972 Marine shells Sydney. of the Pacific JJ.Pacific Publications, -146Davidson,J.M. 1967 An archaeological assemblage of simple fish-hooks from Nukuoro atoll. Journal of the Polynesian Society 76 (2) :177-196. Davidson,J.M. 1971 Archaeology on Nukuoro atoll: a Polynesian outlier in the Eastern Caroline islands. Auckland Institute and Museum Bulletin 9. Eilers,A. 1934 Elbert,S.H. 1949 Inseln um Ponape. Ergebnisse der Sudsee-Expedition 1908-1910. Friederichsen, de Gruyter & Co,Hamburg. Uta-Matua and other tales of Kapingamarangi. Journal Emory,K.P. 1965 of American Kapingamarangi: Polynesian atoll. Folklore. 62 (245) :240-246. social and religious life of a B.P.Bishop Museum Bulletin 228. Fosberg,F.R. 1946 Botantical Report on Micronesia. Economic of Micronesia 13. U.S.Commercial Co. Survey Fujimura,K. and Alkire,W.H. 1979 Archaeological test excavations on Faraulep, Woleai, and Lamotrek in the Caroline Islands of Micronesia. Unpublished Manuscript. Gantt,P.A. 1946 Livestock survey of Micronesia. Economic of Micronesia IS. U.S.Commercial Co. Gosline,W.A. and Brock,V.E. 1960 Handbook of Hawaiian Hawaii. Green,R.C. 1970 Hambruch,P. 1914 Hedley,C. 1897 Fishes. Survey University Press of Settlement pattern archaeology in Polynesia, pp 13-32 In: Green,R.C. and Kelly,M.(Eds.) Studies in Oceanic Culture History 1. Pacific Anthropological Records 11. Nauru. Ergebnisse der Sudsee-Expedition,1908-1910. Friederichsen, Hamburg. The Ethnology Memoir III. of Funafuti. Australian Museum -147Houghton,P. n.d. The people of Namu. In: Leach,B.F. and Davidson, J.M.(Eds.) q.v. Kennedy,D.G. 1931 Field notes on the culture of Vaitupu, Ellice Memoirs of the Polynesian Society, Volume 9. Koch,G. 1961 Kubary,J.S. 1900 Die Matevielle Kultur Volkerkunde, Berlin. der Ellice-InseIn.Museum Islands. fur Beitrag zur Kenntnis der Nukuoro - oder Monterverde -Inseln. Mitteilungen der Geographische Gesellschaft 16:71-138. Lampert,R.J. 1965 An archaeological investigation of Te Aka village, Ocean island: a preliminary report. Unpublished Manuscript. Leach,B,F. 1979 Excavations in the Washpool valley, Palliser Bay. pp 67-136 In: Leach,B.F. and Leach,H.M.(Eds.) Prehistoric man in Palliser Bay. National Museum Bulletin 21. Leach,B.F. and Anderson,A.J. 1979a The role of labrid fish in prehistoric economics in New Zealand. Journal of Archaeological Science 6:1-15. Leach,B.F. and Anderson,A.J. 1979b Prehistoric exploitation of crayfish in New Zealand, pp 141-164 In: Anderson,A.J.(Ed.) Birds of a Feather. British Archaeological Reports S62. Leach,B.F. and Davidson,J.M. 1977 Fishing methods and seasonality at Paremata(N160/50). 20(3): N.Z.Archaeological Association Newsletter 166-175. Leach,B.F. and Davidson,J.M. n.d. Taumako: a Polynesian outlier in the Eastern Solomon Islands . Unpublished Manuscript. Levison,M., Ward,R.G., and Webb,J.W. 1973 The settlement of Polynesia: a computer Australian National University Press. Lewis,D. 1972 Vie, the Press. navigators. simulation. Australian National University -148Lieber,M.D. 1968a Lieber,M.D. 1968b The nature of the relationships between kinship and land tenure on Kapingamarangi atoll. Unpublished PhD thesis. University of Pittsburgh. Porakiet: a Kapingamarangi colony on Ponape. Eugene, University of Oregon, Department of Anthropology. Lieber, M.D. 1970 Adoption on Kapingamarangi. pp 158-205 In: Carroll, 1970 q.v. L i e b e r , M.D. 1974 Land t e n u r e on1 Kapingamarangi. pp 70-99 I n : Lundsgaarde , 1 9 74 q . v . Lieber,M.D.(Ed.) 1977a Exiles and migrants of Hawaii. Lieber,M.D. 1977b Lieber,M.D. 1979 in Oceania. University Press The process of change in two Kapingamarangi communities, pp 35-67 In: Lieber,1977a q.v. Kapingamarangi land history. Unpublished for Ponape Historic Preservation Committee. report 196pp. Lieber,M.D. and Dikepa,K.H. 1974 Kapingamarangi lexicon. University Press of Oceania. University Press of Hawaii. Hawaii. Lundsgaarde,H.P.(Ed.) 1974 Land tenure McKee,E.D. 1956 in Geology of Kapingamarangi atoll, Atoll Research Bulletin 50. Merrill,E.D. 1946 Plant life New York. of Miller, R.E. 1953 Health report Bulletin 20. the Pacific Caroline islands . worZd.MacMillan Co., of Kapingamarangi. Atoll Research Munro,I.S.R. 1967 The fishes of New Guinea. Department of Agriculture, Stock and Fisheries, Port Moresby, New Guinea. Niering,W.A. 1956 Bioeoology terrestrial of Kapingamarangi atoll, Caroline islands aspects. Atoll Research Bulletin 49. -149Parham,B.E.V. 1972 Plants Pawley,A. 1967 of Samoa. DSIR Information Series 85 The relationships of Polynesian outlier languages Journal of the Polynesian Society 76(3):259-296. Pawley,A. and Green,K. 1971 Lexical evidence for the proto-Polynesian homeland. Te Reo 14:1-35. Ray,S.H. 1917 The people of Greenwich atoll,Western Pacific Ocean. Man 17 (130) :187-190. Riesenberg,S.H. and Gayton,A.H. 1952 Caroline island belt weaving. Southwestern of Anthropology. 8:342-375. Journal Schlaginhaufen,0. 1929 Zur Anthropologie der mikronesischen Inselgruppe Kapingamarangi(Greenwich-Inseln).Arohiv der Julius Klaus-Stiftung fur Vererbungsforschung, Sozialanthropologie und Rassenhygiene 4(3):219-287. (Translation by K.J.Dennison). Shapiro,H.L, The physical characteristics of the Ontong 1933 Javanese. Anthropological papers of the American Museum of Natural History.23(3):225-278 Thilenius,G, Ethnographische Ergebnisse aus Melanesien. 1903 Reisebericht. Die Polynesischen Inseln an der Ostgrenze Melanesiens. Kaiserliche LeopoldinischCarolinische Deutsche Akademie der Naturforscher 80 Thomas,W.L, 1965 Titcomb,M. 1972 Vokes,H.E. 1980 The variety of physical environments among Pacific islands, pp 7-38 In: Fosberg,F.R.(Ed.) Man's Place in the island ecosystem. Bishop Museum Press. Native use of fish of Hawaii. in Hawaii. University Press Genera of the bivalvia. Paleontological Research Institution, Ithica, N.Y. Walsh,D.S. and Biggs,B. 1966 Proto-Polynesian word list I. Te Reo Monographs. -150Ward,R.G., Webb,J.W., and Levison,M. 1973 The settlement of the Polynesian outliers: a computer simulation. Journal of the Polynesian Society. 82 (4) :330-342. Whitmore,T.C. 1966 Guide to Islands. Wiens, H.J. 1956 Wilhm,J.L. 1968 the forests of the British Oxford University Press. Solomon The geography of Rapingamarangi atoll in .Carolines. Atoll Research Bulletin 48. the Use of biomass units in Shannon's formula. Ecology. 49:153-156. eastern