Volume 104 1995 > Volume 104, No. 4 > An anthropological study of the burials in Marae Te Tahata, Tepoto (Tuamotu Archipelago, French Polynesia), by Eric Conte and John Dennison, p 397-428
AN ANTHROPOLOGICAL STUDY OF THE BURIALS IN MARAE TE TAHATA, TEPOTO (TUAMOTU ARCHIPELAGO, FRENCH POLYNESIA)
This present study concerns the physical anthropological aspects of burials discovered between 1984-85 in the marae Te Tahata on the island of Tepoto. The results of the research undertaken here have been reported only briefly, (Calaque and Conte 1984), and although a publication of the collective results of all the work conducted is at present being compiled, it is useful to make known in advance the analysis of the skeletal material discovered at this site.
Tepoto, situated at 14° south latitude and 141° west longitude, (Fig.l), is a small atoll, having a surface area of 4 km2, with an ancient in-filled lagoon which is overrun with grass and inundated during the rainy season. Sixty-two inhabitants were counted in the 1988 census. Isolated in the north-east of the Tuamotu archipelago, Tepoto and its neighbour Napuka, a few kilometres away, are the two atolls of the group closest to the southern Marquesas islands.
This isolation, accentuated by the dangers to navigation in the area, the scant natural resources, and the aggressive reputation of their inhabitants, was a repelling factor. Thus these two islands, which appear on maps described none too flatteringly as “The Disappointment Islands” — a name given to them by their European discoverer Byron in 1765 — remained beyond those outside influences which were to bring profound modifications to the Polynesian way of life, until later than the other atolls of the archipelago. For these reasons, the islands were among the last, (in 1878), to be evangelised. Very sporadic visits by the Catholic priests, however, left the population plenty of opportunity to perpetuate certain ‘pagan’ practices until relatively recently, and in some cases even up to the present day.
Also, up until the first half of the 20th century, the French administration had only a theoretical influence on the everyday life of the inhabitants of the atolls. The copra trade, which was a determining element in the socioeconomic transformation in the Tuamotus, (inclusion in commercial networks, introduction of fiscal relations, an opening-up of the islands and more contact with the outside world), did not have a real effect until around the 1920s.- 398 - 399
These various factors explain why material culture in particular was maintained on Napuka and Tepoto in traditional life, as well as in other areas of social life such as spirituality. During the 1929-30 and 1934 archaeological and ethnographic expeditions to the Tuamotu archipelago, Emory stayed several months on both Napuka and Tepoto (Emory 1934). He saw the opportunity of observing technical activities that had disappeared elsewhere, and of collecting significant, surviving oral traditions. Among other things, he made an incomparable documentation on marae ritual, notably ceremonies centred around the beginning of the turtle season, from informants who had witnessed them before the arrival of the missionaries. During his stay on Tepoto in 1934, Emory made an inventory of marae, and a sketch of the sole monument then still in a good state of preservation, the marae Te Tahata (Fig.2), which he published with a brief description (Emory 1947:15).
In 1982, Eric Conte, ethnoarchaeologist at the Department of Archaeology of the Centre Polynésien des Sciences Humaines, Tahiti, made a preliminary tour of Napuka while studying traditional fishing techniques (Conte 1988). During a second visit to Napuka in 1983, he was able to go to Tepoto for several days, and to visit the marae Te Tahata.
This marae, like most in the region, was renowned as an ancient site consecrated to the ritual consumption of turtles. The presence of a number of turtle bones on the surface, in the vicinity of ovens in which turtle meat was cooked, attested to the likelihood of this claim. It therefore seemed interesting to juxtapose the wealth of information amassed by Emory in regard to the rituals practised on the marae in this part of the Tuamotus, and those found by Conte in his own research, together with information furnished by the archaeological excavation and investigation of this site.
The abundance of surface remains, which included an assemblage of coral cysts laid out in a number of alignments, whose state of preservation satisfactorily demonstrated that the site had not been touched by cyclones at the turn of the century, made the marae Te Tahata a favourable site for investigation. From this perspective, the first archaeological expedition was organised in 1984. It concentrated mainly on Napuka, where work was conducted on marae Marokau, and the marae on Tepoto. The stay on the latter island lasted from 22nd May to 6th July, and was almost exclusively devoted to research conducted on the marae Te Tahata. Isabelle Calaque, a student at the Université de Paris I, assisted Conte.
Firstly, Conte and Calaque set about clearing the marae completely of the few centimetres of superficial deposit, to expose the features partially covered, and eventually to find those completely covered. This clearance- 400 - 401
enabled the location of some 36 lithic structures and numerous turtle bones, and resulted in the discovery on 29th May of numerous scatters of human bone and the skeleton of an infant (Burial 3). From this point on, the investigation was oriented in a formerly unforseen direction, and expanded in scope.
In total, 22 skeletons of adults and infants of both sexes were revealed in 1984, certain of them having been buried with traditional objects (adzes of Tridacna maxima shell and shell Pandanus tectorius scrapers). For a thorough anthropological study and ultimate dating, the ideal would have been to uplift them, with an assurance of their imminent return to their homeland for reinterment in the same spot. However, an assembly of the village, after more than three hours' debate, decided that it would be unacceptable to remove the bones. It was therefore decided to include in the next expedition a physical anthropologist charged with the analysis of the skeletons in situ.
In October to November 1985, a second expedition to Tepoto was organised, with the participation of John Dennison of the Department of Anatomy and Structural Biology at the University of Otago, and Joseph Tchong from the Department of Archaeology at the Centre Polynésien des Sciences Humaines. While Dennison studied the skeletons uncovered in 1984 and reburied at the end of that first field season, Tchong and Conte proceeded with a number of test pits. One of these (test pit 9) was extended and a further area was excavated, adding a further ten human skeletons as well as that of an associated young dog.
The skeletons are for the most part associated with the coral cysts, arranged in two parallel lines (Fig.3). The analysis of the various apparent associations (skeletons, lithic structures interspersed with the skeletons, etc.) are not discussed in the present article, which is concerned only with the physical characteristics of the 32 human skeletons discovered in the marae Te Tahata.
GENERAL DESCRIPTION OF THE BURIALS
Copra sacks had been placed directly over the skeletons prior to reburial at the end of the 1984 field season. When the burials were disinterred in 1985, it was found that a feltwork of roots from germinating seeds within the sacks had penetrated the bones. Other roots from coconut palms frequently traversed the long axis of limb bones or forced their way through crania, which made exposure of intact bones a rare event.
The burials appeared to be in discrete groups. Burials 1–12 lay in two parallel lines — Burials 1, 2, 4, 6, 9, 10 and 11 in the first line, and Burials- 402 - 403
3, 5, 7 and 8 in the second line. Burial 12 lay somewhat apart from the second line, separated from the others by several large coral rocks. These burials were of five adult females, three adult males and four children, one of whom shared a grave with an adult female. The two deepest burials, 10 and 11, lying side by side, possessed grave goods — a pandanus leaf scraper (kotore) by the right arm of Burial 10, and a Tridacna shell adze on the stomach of Burial 11.
A second section of burials, 13–22, also lay in two lines: the first comprising Burials 13, 17, 18, 22, and the second Burials 14, 15, 16, 19, 20, 21. Buried in this section were four adult males and six children under the age of ten years old. Very few of these burials were complete.
Of the eleven further burials exposed in 1985, one female adult (Burial 23) was found lying parallel to, and about two metres from, Burial 1. The other ten burials (24–33) formed a third line parallel to the two lines of Burials 1–12. In this section were three female adults and six children, buried parallel. Burial 25, probably the oldest individual interred, supported a kotore on her left shoulder. The tenth burial was a young dog lying curled on its right side.
Overall, most of the skeletons were complete, and in articulation. Generally, the bones were in good condition but had been subjected to invasion by roots. Of the 29 burials whose position of interment could be established, all were supine except Burial 21, a child lying on the right side. Most burials lay extended, feet towards the west: the heads of 16 of them were looking to the right, while 12 looked left and three looked straight ahead. The arms of 14 lay alongside the body, two had the right elbow bent so that the right hand rested on the thigh, five had left hand on thigh, two had both hands on thighs, two had arms folded across their chests, two had hands resting under their heads, and one had her hands folded back onto her shoulders. Two females had legs crossed — Burial 4 right over left, Burial 23 left over right; six, all children, had legs bent at the knee. Only three (Burials 10, 11 and 25) had any grave goods.
The burials all lay at a fairly shallow depth in coral sand. The calcareous soil looked a brownish-grey when first opened. Rocks and blocks lay about. Surrounding the marae there is scant grass cover and palm leaf litter. ‘Inland’ from the marae there is thicker bush and tree vegetation, such as pandanus, Messerschmidia argentea, Guettarda speciosa, and Polypodium sp. fern.
AGE AT DEATH
The age at death ranged from infancy up to the forties. The 32 individuals (14 adults and 18 subadults) are grouped by age and sex in Table 1.
Three individuals had a retrievable pubic symphysis; for the remainder, - 404 tooth eruption (according to Schour and Massler's 1941 chart), tooth wear and epiphyseal fusion were used as criteria in the assessment of age. In this particular situation, where heavy use of anterior teeth apparently meant less molar wear, some age estimates may be in error, on the low side.
A survivorship curve (Fig.4) shows the high early childhood mortality and high mortality in the 20–30 year age group. For comparison, the survivorship curves from the findings of Lovejoy et al. (1977:292) in the much larger population from Libben, Ohio, and from Houghton's (1981) study of the people from the Namu site on the Pacific island of Taumako have been superimposed.
The mean age at death for the entire population was 17.6 years, while for the adults alone it was 29.8 years. The highest mortality among the subadults occurred between birth and five years, i.e. infancy and early childhood. The lowest mortality occurred during early adolescence. As with Pietrusewsky's (1976:100) Hane, Marquesas, population, (whose mean age at death was 24.7 years) the Te Tahata people showed a sharp drop off in life-expectancy after 30 years, the end of the first decade of adult life.- 405
Sex ratio = 7 males/11 females = 64
Subadult/adult = 18 subadults/14 adults = 129
Table 1: Age and sex distribution according to burial number
Stature estimates were calculated, using the equations derived for Polynesians (Houghton et al. 1975). The mean stature for seven males was 1751 mm (5′ 9″), and for nine females 1657 mm (5′ 5″). Pietrusewsky's (1976:134) Hane males ranged from 5′ 3″ to 6′, while the females tended to the upper limit of 4′ 9.5″ to 5′ 5″.- 406
Figure 5. Typical arch shape, shovel-shaped central incisor and third molar fused crowns of Burial 14.
Stature measurements were taken from volunteers among the present Tepoto adult population. The measurements showed that the earlier (and younger) population were apparently taller than the present population. Certainly their skeletons were robust.
Root invasion made the recording of cranial measurements rather haphazard. Therefore because of the small sample size generally available, much caution must be exercised in interpreting the values.
Cranial indices were calculated from the measurements. All crania were long-headed (Montagu 1960:605), except for Burial 6, which was borderline medium head length. A similar finding in female crania was reported in the Hane people by Pietrusewsky (1976:109). All crania had average length-height values, except for Burial 6, whose head length was shorter in relation to height. All were high-vaulted. Only Burial 6 furnished an upper facial index, whose ratio between nasion—prosthion height and bizygomatic breadth indicated a tendency to high facial height, and average upper facial width. Cranial capacities were calculated using Lee's formulae l0 and 11 (Wood-Jones 1929:51), which represent both male and female capacities as large.
In summary then, measurements reveal these crania to be long, high-vaulted, of average upper facial width and of large capacity. All the adult crania from the posterior view showed the pentagonal outline typical of Polynesians. A marked mid-sagittal ‘keel’ was often present.
Female mandibular dimensions tended to be smaller than in the male. All mandibles were definitely “rocker” type (Marshall and Snow 1956), a feature occurring in about 80% of Polynesians (Pietrusewsky 1969; Snow 1974), but perhaps only in 2–3% of other groups. A striking feature was the angular, alveolar arch shape (Fig.5) — quite different from the smooth curve of later New Zealand Polynesian arches (Kean 1988, personal communication) yet similar to the arch shape in people from New Zealand's early occupied site of Wairau Bar (Houghton 1988, personal communication) The Te Tahata mandibular measurements were compared with those of other populations. The results appear in Table 2.
Roots passing within the long bone diaphyses restricted the number available for measurement, but indices were calculated from those few. Radio—humeral indices (Montagu 1960:620) were similar to those presented by Pietrusewsky (1976:133). The index of length of arm expressed as a percentage of length of leg is of similar value to what Pietrusewsky has found for other Pacific populations. The femora were proportionately long in relation to the tibiae. The mean male tibio—femoral ratio of 80.9 was a low value among Homo sapiens. The relatively short tibia is a characteristic of both Mongoloid and Polynesian (Houghton et al. 1975; Eveleth and Tanner - 408 1976), however the Polynesian has a relatively long forearm compared with the Mongoloid (Houghton 1981) and the radio—humeral ratio of 77.5 falls among Polynesian proportions (Houghton et al. 1975; Snow 1974). The combination of short tibia and long forearm appears to be unique among Polynesians (Houghton 1981). The upper femoral shafts appeared to be more flattened antero-posteriorly in the female than in the male. A comparison with indices from other populations is given in Table 3.
Table 2: Comparative male mandibular measurements from other populations. The New Zealand and Namu data are from Houghton (1981), and the remainder are from Pietrusewsky (1985:408).
The femora were generally bowed. All bore an oval fovea capitis femoris, typical of Polynesians. All tibiae observed showed squatting facets on their antero-inferior aspects. In the limb bones, marked muscle ridging was usually present.
Limited time, governed by weather conditions and the fact that the vertebral column and pelvis were always enmeshed in a network of roots, precluded much investigation.- 409
Table 3: Comparative infracranial indices from various populations. Values in parentheses are sample sizes. The Hane data are from Pietrusewsky (1976), and the remainder are from Houghton (1981).
With our time constraints, tooth measurements could only be sampled from the population. Two males (Burials 9 and 14) and two females (Burials 4 and 6) were selected. The means of crown mesio—distal and bucco—lingual dimensions did not compare at all favourably with those of other populations (Dennison 1979). This is probably a reflection of the small sample size, although this result must stand on its own merits. Nevertheless, sexual dimorphism was evident, in that male crown dimensions were greater than female dimensions.
The pattern of dental wear is quite distinctive. The notable feature is- 410 - 411
severe wear on the anterior teeth (especially those of the mandible), whereas wear proceeds more slowly on the molar teeth. An example is shown in the mandible of Burial 15 (Fig.6). This wear occurred with equal intensity in males and females. Dennison therefore examined the dental wear in volunteers from the adult population while measuring their stature. In the older members of the group sampled, anterior inferior tooth wear was quite marked, whereas it was not evident in the younger people.
Wear on the anterior teeth and corresponding lack of wear on the molars made ageing by tooth wear difficult. Yet there was often good dental evidence to age children, from the presence of the secondary dentition in situ beneath the primary dentition.
There were few variations in dental morphology. A shovel-shaped upper left central incisor tooth is seen in Burial 14 (Fig.5), and Carabelli's cusps were present in Burials 1, 14 and 22. A peg-shaped third molar was seen in Burial 16, and two peg-shaped premolars in Burials 1 and 2. Burial 14 (Fig.5) showed a rare developmental anomaly — a double germination bud in both the upper right and upper left third molar teeth, maturing into fused crowns. The cusp pattern of the first mandibular molar tended to be Y5, while for the second mandibular molar +5 seemed more common (cf. Suzuki and Sakai 1973).
There seemed a higher incidence of calculus build-up (evidence of severe periodontal disease) in the males than in the females, yet this was still below the level found in the Hane people (Pietrusewsky 1976:131). Caries was present only in the posterior teeth of the few males in which it was studied, but it was more evenly spread throughout the dentition of the females where the incidence was higher. The overall caries level was similar to that in the Hane population.
Displacement due to overcrowding seemed very high in this small population. The percentage of impacted third molars in females (20%) was the same as the 15–20% reported in modern Americans and Western Europeans (Gibson and Calcagno 1993), but the 42% in males was more than twice as high (Fig.6).
Three teeth, the lower left and right first molars from Burial 16, and the lower left second molar from Burial 25, showed shearing damage.
HEALTH AND DISEASE
These people were tall and robust. Several females showed evidence of previous pregnancy (Houghton 1975). Such findings indicate that the level of nutrition was adequate.
Only in Burial 12 was evidence of limb joint arthrosis found, at the distal- 412 - 413
end of the tibia and both ends of both ulnae, but to a minimal degree. This individual showed level 2 (on a scale of 5) arthrosis of the first two cervical vertebrae. But as this was only one of a very few vertebral columns that could be extricated, a general comment for the population cannot be made. There was no evidence of trauma.
The only case of inflammatory disease was found in Burial 9, where on the disto-medial surface of the right fibula there was a lesion 21.2 mm long, penetrating half the thickness of the cortical bone (Fig.7). This lesion was reminiscent of a necrotic perforation of tertiary yaws (Steinbock 1976:142, 158). However absence of any other focus forces the diagnosis of nonspecific infection.
Thirteen second metacarpals were x-rayed antero-posteriorly for measurement of cortical thickness and assessment of Harris lines, (Harris 1933).
Harris lines, or lines of arrested growth, are indicators of a disease episode or malnutrition, which in turn may indicate a seasonal shortage of food (Gindhart 1969) during the growth period. Lines of enamel hypoplasia also, are related to systemic disease seriously affecting nutrition by cessation of ameloblastic activity (Sarnat and Schour 1941). These were recorded from the vestibular surfaces of teeth.
A ranked (male) comparison of cortical thickness with other groups is presented in Table 4. The Te Tahata sample is ranked very low, being ahead only of Nebira (Houghton 1982), a diseased population. Female groups are also included in the table but are not ranked. For the Te Tahata group the sexual difference in cortical thickness was not statistically significant. Burials 9 and 16 were omitted from the series as they were obviously unhealthy.
Nordin's score (Barnett and Nordin 1961) was used as a normalisation of cortical thickness measurement for bone size, presenting the fraction of bone diameter consisting of cortical bone. Tables 5 and 6 present raw data which may bear a relationship to metacarpal cortical thickness. It has been shown by Simpson (1979:95) in his New Zealand Maori series, that in the male, tibial length and Harris lines have a weakly negative correlation while Harris lines and enamel hypoplasia were the only other significant relationship. Neither of these relationships existed in the female, but there was a strongly negative relationship between age and cortical thickness and a lesser one with Harris lines. Age and pregnancy correlated strongly.
In the Te Tahata series statistical analysis was not performed on this small sample size. However, in Table 5 visual inspection of the values from Burial 9 (that showed a non-specific infection) and from Burial 16 (with a high number of Harris lines and lines of enamel hypoplasia and a low Nordin's - 414 score), revealed some measure of correlation, which did not relate to stature. No such visual relationship is seen in Table 6.
Table 4: Second metacarpal cortical thickness comparison (mm). n = sample size. S.D. = standard deviation.The New Zealand Maori and Taumako data are from Simpson (1979), and the Nebira (New Guinea) data from Houghton (1982). These are of prehistoric adults aged 20–45 years. The other data are from Garn (1970), representing living populations aged 25–45 years.- 415
Table 5: Male second metacarpal-related data
Table 6: Female second metacarpal-related data- 416
The dating of archaeological bone has always been fraught with problems related to burial environment, soil water, soil acidity and humidity. In this coralline sand site, rootlet invasion was a particular problem. Thus the smaller metacarpals, metatarsals and phalanges were chosen as dating samples because less likelihood of contamination could be presumed by their better state of preservation.
Two metacarpals, from Burial 7 (ANU–5461) and from Burial 12 (ANU–5462) were sent to Canberra, Australia, for radiocarbon dating by the accelerator mass spectrometer (AMS) method. Head writes:
The bone apatite fraction was very small for each sample, which is a good indicator of the degree of preservation, or the young age of the bone (or both). Results for apatite and collagen fractions are in agreement within the limits of measurement, but I would take the collagen results as the most reliable (and the most accurate). I would be quite confident in placing age limits between 1850 and 1950 AD for these two samples. There is a chance that the bones may be as old as 1800 AD, but there is not much evidence for this. (Head 1987: personal communication)
A series of 30 bones was sent to the Institute of Nuclear Sciences in Wellington for measurement of the radial profile of fluorine across a transverse section by nuclear microprobe, and deduction of age by comparison with the optimum profile calculated from diffusion theory, as described by Coote and Sparks (1981), and Coote and Nelson (1987).
From the fluorine profile dating method all the burials, apart from Burial 19, could fall from 100 years B.P. onwards, which would agree with the radiocarbon dating results. A cursory examination of the fluorine profile results might suggest that all burials occurred close to A.D. 1850, which might suggest a single causal event, an idea supported by two graves shared by adult female and child, and a grave shared by a child and young dog.
We have examined a population of 32 individuals comprising seven adult males, seven adult females, three probably female adolescents, 13 children and two infants.
The burials appeared to be in two discrete groups — 1–12 and 23–33 in four parallel lines, while Burials 13–22 lay further away, in two lines. In two burials an adult female and child were interred together in a quite poignant manner.
All burials faced seaward. All burials were supine except for one child lying on its side. Most burials were extended. The common belief has been that extended burials are characteristic of early eastern Polynesian contexts. - 417 Leach and Leach (1979:211) point out that while this feature appears to be common at Maupiti, and at Wairau Bar in New Zealand, it is much less so at Hane, Ha'atuatua, and Palliser Bay, New Zealand, where more complex burial postures were employed.
Grave goods were found in only three burials—mother-of-pearl pandanus scrapers (kotore) with female Burials 10 (of undetermined age), and 25 (the oldest female interred), and a Tridacna adze with male Burial 11, who was in the oldest male age group, was the tallest individual, and who lay to the right of Burial 10. Sinoto (1970:128) states that burial offerings for people of high status were present in the Marquesan initial periods, but in Phase IV (Classic into Historic), offerings of goods besides food became widely practised.
Dog burials are relatively uncommon in Polynesia but are recorded for Palliser Bay, New Zealand (Leach and Leach 1979:209), Lotofaga in Samoa (Davidson 1969:239) and Hane, Marquesas (Sinoto 1970:110). Sinoto reported only dog teeth in Phase I initial settlement, and no dogs in Phase III expansion, but their prevalence in Phase II suggested they were man's companion rather than his food. Dogs were, and are possessed by the inhabitants of most of the Tuamotus. They were recorded by the Spaniards at Anaa in 1606(Quiros 1904: vol.1:201). Emory (1975:39) said that while the Tahitians coveted the Tuamotuan dogs' long hair for ornamentation, the Tuamotuans prized them as a source of food. In many of the islands, and especially on Napuka and Tepoto, the dog is still eaten, baked in the ground oven, a greater delicacy than pig. However the burial of this particular dog among human graves would suggest a special companionship rather than a food item.
The cause of death could not be established for any individual. Only one case of osteologically involved infection was found, (Fig.7). The average age at death of the thirteen adults was 29.8 years, (31.0 for six females and 28.8 for seven males). This compares with 24.7 years for the Hane people, although the value for this group may have been distorted downwards by the mass burial of eleven subadults and seven adults after what was probably a sea tragedy (Pietrusewsky 1977:99). For 80 adult males from Namu the average age at death was 32.0 years while for 57 females the age was 28.6 years (Houghton 1981). It is perhaps unjustifiable to make comparisons with ages at death from outside the tropics (and certainly of populations living at different time periods), but from New Zealand the age at death from the oldest known population, from Wairau Bar in Marlborough was 27.9 years, while for both a Chatham Island population and a Wairarapa population average age at death was 37.0 years (Houghton 1976:7).- 418
Table 1 shows, with highly reduced numbers, a rather similar pattern of mortality to the Namu people (Houghton 1981) — high early childhood mortality and high mortality in the third decade of life. The female deaths occurring between 15 and 20 years may relate to pregnancy, whereas male deaths are greater between 20 and 25 years, perhaps demonstrating risks in the male life role.
In 1977 Lovejoy et al. reported the survivorship curve of a 1,327-strong prehistoric Amerindian population from Libben, Ohio. This curve is reproduced in Figure 4, together with the curve from the Namu population, for comparison with the curve of the Te Tahata people. The latter, after an initial sharp drop (11 of 32 people died before reaching age 7 years), showed a similar pattern to the 200 individuals of Namu — a lower mortality in later childhood and early adolescence and a higher mortality in the 20–30 year age group by comparison with the Libben people.
Tooth wear was greater on anterior teeth than posterior teeth where one would have expected it, given that atoll life would involve a high marine content in the food supply, albeit deep sea fish. Such a wear pattern would be possible if leaves were being drawn across the occlusal surface of the anterior teeth. Emory wrote that the sprouting aerial roots of the pandanus were chewed after being baked — “the sugarcane of the Tuamotus” (Emory 1975:24). An elderly man commented (Arai 1985, personal communication) that in former times pandanus was chewed, and he remembered doing so as a boy. Dried octopus was eaten by tearing pieces off with an upward jerk of the head. Pandanus and octopus are still eaten today, but to a lesser extent than in former times. This may explain why wear was heavy on the anterior teeth. Pietrusewsky (1976:131) also noted a high degree of dentine exposure among the Hane people, but it was more evenly spread throughout the dentition.
This population did not lack ready food. The people were well nourished, robust in physique, and tall. The minimal joint arthrosis observed would suggest that their environment was not physically demanding upon them. Metacarpal cortical thickness was not great — perhaps it did not need to be. The numbers of Harris lines and enamel hypoplastic lines were relatively low, indicating perhaps relatively good health in childhood.
These people appear to have enjoyed a life span reasonable by the standards of prehistoric times. From the skeletal evidence, they would seem to have been healthy when they died. But we are left with the questions: Why did they die? What caused adult female and child to be buried together simultaneously (to judge from the deliberate positioning of their bodies)? Aubert de la Rue (1958:109) commented that in 1842 Dupetit-Thouars counted a population of 20,000 in the Marquesas, which closely neighbour - 419 Tepoto, but by 1924 that population had dropped to 2,094 — their demise hastened by syphilis, tuberculosis and epidemics. 1850, the date from which most of these burials appear to emanate, was the high-point of epidemics in Polynesia and especially in the Marquesas. Arai (1985, personal communication) reported having heard that a boatload of “whites” had arrived at Tepoto to take on water, and had introduced an epidemic that had caused numerous deaths. Also, Emory's (1947:9) informant, Te Uru, reported that during his youth the marae Te Tahata was no longer used because those to whom it belonged were all dead and it was abandoned.
Although this evidence is circumstantial, it can explain the deliberate multiple burials of adult and child, and child and dog, the lack of bony evidence for cause of death, and the high mortality in the 20–30-year age group; and the time of the epidemic corresponds with the measured date of death of many of the individuals buried beneath marae Te Tahata.
We are grateful to John Head in Canberra, and Graeme Coote in Wellington for their collaboration in the dating study, and we thank Philip Houghton, Gareth Jones, Kazumichi Katayama and Michael Pietrusewsky for critical comments on the text.
Burial 1: This burial marked the right end of a line. The body lay outstretched, with head turned to the left. The left arm was alongside the body with the forearm bent so that the left hand was under the head. The right forearm lay across the stomach. The legs were straight with the toes pointing forwards. The cranium was at the ground surface. The remains were in reasonably good condition although rather fragmentary. Most of the limb bone extremities were missing due to plant root invasion. From cranial morphology this was a female Polynesian. Her age at death was about 18 years (no fusion of: lateral epicondyle of left humerus, proximal right radius, distal right ulna, distal second metacarpal; but lower left and right third molar teeth were recently erupted). Her height was 1652 mm. Teeth were generally in good condition although lower right third molar was impacted, and wear was heavy on anterior teeth, more particularly the lower incisors.
Burial 2: The body lay to the right of Burial 1, in an extended position, arms by sides, toes forwards, head slightly inclined to the right. Again the remains, although in good condition, were fragmentary. From cranial morphology, this was probably a female who from lack of fusion proximally and distally in radii, tibiae and clavicles, distal femur, and distal metatarsals, wide-open roots of the upper left second molar tooth, and lower left third molar being found at the - 422 bottom of its socket with the crown formed, was estimated as aged between 12 and 14 years at death, with a height of 1596 mm. Anterior tooth wear was not so marked.
Burial 3: The bones were intact although largely unfused. This child of less than six years with an intact milk dentition and a mandibular angle of 130° (the formed crowns of some permanent teeth were observed) lay in front of Burial 2. The child lay supine, with the head inclined to the right, the arms folded across the chest and the knees bent outwards.
Burial 4: This skeleton, in good condition, was fairly complete. The body was outstretched, with head turned slightly to the left, mouth open, arms by sides, legs straight but right tibia and fibula crossed over left. Cranial and femoral morphology indicated that this was a Polynesian. Although discriminant function analysis made this a male, cranial morphology and a possible preauricular groove of pregnancy determined this to be a robust female of height 1695 mm, aged about 27 years from pubic symphysis. There was evidence of dental caries. The mandibular canines were both displaced antero-laterally by adjacent teeth in the arcade. Inferior incisal tooth wear was severe.
Burial 5: The intact skeletal remains, in good condition, of a four-year-old child (based on dental evidence) lay in front of Burial 4 in an extended supine position, arms by sides, head inclined downwards to the right.
Burial 6: This complete skeleton, to the right of Burial 4, lay in an extended supine position, head facing forwards, left arm by side, right arm bent at right angles across stomach. From femoral morphology this was a Polynesian. From pelvic morphology she had borne children, and had reached her late thirties from dental comparison with Burial 4. Her height was 1657 mm. Again there is severe dental incisal wear, more marked in the mandibular teeth. The molar teeth are also quite worn.
Burial 7: An outstretched burial lay in front of Burial 6 and to the right of Burial 5, and shared a common grave with Burial 8 to the immediate right. The arms were by the sides although the right forearm was bent so that the hand lay over the pelvis, to accommodate the bent left leg of Burial 8. Cranial morphology and dental evidence indicated that this was a Polynesian female in her twenties. Her height was 1637 mm. The lower anterior teeth were missing but the maxillary anterior teeth showed much use. No pathology was evident. Both femoral extremities and the cranial vault were damaged by roots. Her head was turned horizontally to the right, looking into the face of Burial 8.
Burial 8: This seven-year-old child had been buried on its back, forming an angle with Burial 7. The head was turned to the left, looking towards Burial 7. The right arm was by the side but bent at the elbow so that the right hand lay on the pelvis. The left arm was bent back so that the hand rested on the left shoulder. The right leg was outstretched but the left leg bent at the knee (the femur lay parallel against the radius and ulna of Burial 7's right arm). The feet lay over the right pelvis of Burial 7. The child's height was 1289 mm.
Burial 9: This supine outstretched individual lay alongside Burial 6. The body was outstretched with the head slightly inclined down and to right. Both arms were - 423 by the sides although the right hand rested on the pelvis. From cranial and femoral morphology this was a Polynesian whose teeth wear gave an age in the thirties. Discriminant function analysis and morphology revealed a male, who was very robust (compared with the relatively more gracile Burials 4 and 6) with a stature of 1766 mm. Tree root penetration of right femoral head made measuring difficult, otherwise the skeleton was fairly intact. A bony lesion, 21.2 mm long, that showed evidence of pre morte healing was present on the disto-medial surface of the right fibula. Absence of any corroborative lesions from other bones precludes a diagnosis of a lesion from yaws infection. The cortex of the second metacarpal was thinner than in other males. Lower left and right canine teeth were severely displaced, and lower right lateral incisor tooth was displaced posteriorly.
Burial 10: Again an extended supine burial, lying alongside Burial 9, arms by sides, left arm slightly bent so that left hand lay over pelvis palm downwards. The head, fragmented by root penetration, was turned left and slightly downwards. The lower limb bones in particular, had suffered post mortem fractures. As with the other burials this female Polynesian (from cranial and femoral morphology) was buried in coral sand highly penetrated by coconut palm (Cocos nucifera) roots but rocks had been placed on the three sides of the head. Medial to the right humerus were placed two semicircular pandanus leaf scrapers kotore, made of mother of pearl, forming a circle. Her height was estimated as 1699 mm. Only the right half mandible was present, yet the incisors showed greater wear than the molars. Dental caries was present in lower left central incisor, lower right first premolar and first and second molars; lower right third molar was congenitally absent.
Burial 11: This relatively undamaged outstretched skeleton, lying somewhat apart from the main group, in the second row, was that of a robust male Polynesian (from cranial and femoral morphology) aged in his thirties from dental comparison, and of height 1804 mm. Attachment areas on tibia and humerus indicated that this was a well-muscled individidual. The head with open mouth was turned to the right. There was no sign of alveolar erosion in the maxilla, but erosion was evident in the molar region of both sides of the mandible. Dental wear was heavy on the mandibular incisors, the maxillary incisors were lost before death. This being a rare case where atlas and axis vertebrae were accessible, a grade 2 (on a 1–5 scale) osteoarthrosis was assessed. Otherwise, as in the other skeletons, no osteoarthritic changes were clearly seen. Tibial squatting facets were obvious. The arms were by the sides. The left elbow was bent so that the hand lay over the stomach. A Tridacna adze lay on the stomach area with the blade facing to the right.
Burial 12: The head of this individual had been exposed and the remainder was excavated immediately prior to examination. The skeleton lay between, and roughly parallel to, Burials 6 and 9. The head was inclined 45° to the left with the mouth wide open. The body was outstretched, arms by sides, hands prone on femoral trochanters. Femoral and cranial morphology showed this to be a Polynesian, and a male from pelvic and cranial evidence and general - 424 robusticity. His height was 1717 mm. From dental comparison with Burial 9, age was assessed as late twenties. While the maxillary dentition was almost complete, the mandibular dentition was represented only by lower left first and second molars and lower right second molar, and there was consequently much mandibular resorption. Calculus was present particularly on the buccal surface of the anterior teeth.
Burial 13: This burial marked the start of a discrete section to the north-north-west of Burials 1 to 12. The skeleton was outstretched, the head looking to the right and downwards at a 45° angle, the left foot crossed over right foot. The bones, in good condition, were those of a seven-year-old child (from the wear facets on the lower right first molar tooth, and the unfused limb members).
Burial 14: This outstretched burial lay directly in front of Burial 13. The arms were alongside the body but both slightly bent at the elbows so that the hands lay prone on the pelvis. The head was turned to the left, mouth slightly open. Some bones, particularly the cranium, had suffered root invasion. The “rocker” mandible indicated a Polynesian. The generally robust nature of the bones indicated a male whose height was 1738 mm. He was in his early twenties (the upper third molar teeth had erupted and showed slight wear. These teeth were particularly interesting in that their crowns were both fused, a rare developmental anomaly caused by the maturing of two germinal buds on each tooth. Badly impacted lower third molar teeth were still inside the bone). Lower left central incisor tooth was rotated on its long axis while lower right canine was displaced. Wear was heavier on the mandibular incisors.
Burial 15: This was again an extended burial, to the right of Burial 14. The head faced forwards. The lower jaw had dropped down out of articulation. Although the molar teeth looked younger, this Polynesian male was aged in his early thirties as all the cranial sutures were almost obliterated. The height of this robust man was 1766 mm. Again many bones, especially the cranial vault, had been invaded and fractured by roots. The interior of the frontal bone had been abraded away until only the outer table remained. The maxillary incisal region was lost after death, but wear was heavy on the mandibular incisors. Again both lower third molars were severely impacted.
Burial 16: An outstretched burial lying to the right of and parallel with Burial 15 although the tibiae were at the level of the latter's humeri. The head was turned to the left and slightly upwards, the mouth wide open. The arms were by the sides, and slightly bent at the elbow so that the hands were prone over the pelvis. The bone of this Polynesian male was in good condition although fractured by roots. The right antero-infero-lateral aspect of face and mandible had lain exposed at ground level, from bleaching. His height was 1721 mm and his age was in the early twenties (the third molars had erupted but showed little wear). The cortex of the second metacarpal was the thinnest of those measured. The larger number of Harris lines and lines of enamel hypoplasia suggested extensive childhood illness. Dental overcrowding was evident in the mandible — the canines were pushed forwards; the third molars were impacting at 45° - 425 against the second molars. The first molars were sheared buccally, revealing dentine. The maxillary third molars were growing vertically but were overhung by the second molars, while the upper left first molar had tipped buccally with loss of crown, (the upper right first molar was lost). Calculus was extensive on the mandibular teeth.
Burial 17: This highly fragmented, disturbed burial had been almost destroyed by roots. A femur shaft and the left half of a mandible with intact dentition revealed the burial to be that of a six-year-old child.
Burial 18: This was an extended burial, head turned to right, arms by sides, toes facing forwards. The bones were all present, but the cranium was fragmented and all limb bones were fractured by roots. Left and right tibiae, including unfused epiphyses, measured 159 mm. There was sufficient dental evidence to assess the child to be five years old.
Burial 19: The burial had been severely disturbed. The head was turned to the left. A coral lump lay across the right facial area. One permanent first molar tooth crown gave an age of about four years. The left tibia, including distal epiphysis measured 157 mm.
Burial 20: Most bones were present in a good state of preservation but damaged. The cranial vault on the right parietal region was fractured. The head was turned to the left; arms by sides, the left being slightly bent at the elbow, the hands prone. The torso was turned slightly left; the femora were turned to the left, the right knee was maximally flexed while the left knee made a 90° angle between femur and tibia. The toes were at right angles to the legs. The child was approximately eight years old (lower right central and lateral incisor teeth were close together, lower left central and lateral incisors were behind). Application of the left tibia length, including proximal epiphysis, of 188 mm gave a stature of 1364 mm although its validity in one so young is questionable. There was heavy dental wear, particulary among the lower right milk molars. Right upper and lower teeth were more worn than those of the left.
Burial 21: Here was a five-year-old child (from dental evidence) laid on the right side. The cranium, badly damaged, was facing right. The arms were bent so that the hands were lying anteriorly to the chin. The legs were slightly flexed at the knee; the toes were pointed down and to the right. The left ulna including the distal epiphysis measured 134.6 mm. Dental measurements were also made.
Burial 22: This extended burial marked the northern extremity of this cemetery section. A large stone lay to the right, another against the right foot, and a third beyond the feet. The head was turned to the right, the cranial vault being damaged. Arms were by sides, palms of hands facing forwards. The right tibia was slightly everted so that the right foot was turned outwards, while the left foot was turned slightly left. While the bone was in good condition it was badly fragmented. This was a Polynesian, and male (based on the square ‘chin’ region and a pronounced nuchal crest without nuchal lines), aged mid- to late twenties (based on wear facets of the third molar teeth). His height was 1742 mm. The lower right third molar tooth was displaced laterally from lower right second - 426 molar, and lower left first premolar appeared misaligned from the second premolar. Lingual calculus was particularly noticeable in the mandibular teeth, as was incisal wear.
Burial 23: A newly excavated outstretched burial lay in brown-grey sand about two metres to the left of, and parallel to, Burial 1. The head was turned 60° to the right and looking slightly downwards. The right arm was alongside the body, palm down. The left arm was bent at the elbow so that the left palm was prone across the groin. The left tibia and fibula crossed the right. The toes faced forwards. Some limb bones were fractured by roots. From cranial morphology and general gracility this was a female of height 1662 mm. She was aged approximately 17 years (the upper left third molar tooth had not erupted, but there was a well-formed socket for the lower right third molar). Mandibular incisal tooth wear was marked.
Burial 24: This burial marked the left end of a line of newly-excavated burials west and parallel to the 1-12 burial section lines. This was an extended burial, the (very much damaged) head turned to the right, mouth open, arms by sides, right palm upwards, left palm downwards, toes forwards. The burial was of a Polynesian female in her early twenties (based on wear facets on the crowns of the recently erupted third molar teeth) of height 1647 mm, who from the preauricular groove had borne children.
Burial 25: The complete skeleton in good condition lay outstretched, parallel and to the right of Burial 24. The head was inclined at an angle of 45°down and to the right. The arms were by the sides, the left palm upwards, the right palm downwards. The toes faced forwards. A semicircular pandanus leaf scraper, kotore, lay, straight surface inferiorly, on the left shoulder. This female Polynesian (from general morphology) was aged in her forties (from tooth wear). Her height was 1647 mm. Tooth wear was severe — upper left canine had sheared steeply bucco-lingually; the lower left second molar had also sheared, while the lower left third molar lay horizontally inside the jaw, impinging on the root of the second molar, forcing the latter to tilt distally so that the mesial surface of the tooth formed the occlusal surface.
Burial 26: This was the badly fragmented skeleton of an outstretched infant. The head was turned to the right looking towards Burial 27 close by. The left elbow was bent so that the hand was beneath the chin.
Burial 27: An extended burial lay parallel to Burials 24 and 25, and close to the right of Burial 26. The head was turned to the left and looking down at 45° to Burial 26. The right arm was bent at the elbow so that the right hand rested on the left clavicle, while the left arm was parallel to the body but bent back at the elbow so that the left hand was on the left shoulder. Every long bone was fractured; the calvarium was broken parietally and frontally. From cranial morphology this was a Polynesian female. She was in her late twenties from dental wear facets on the third molar teeth. Her height was 1620 mm. Dental wear was severe in both upper and lower incisal regions, tapering off in the molar regions where calculus was present on the buccal surfaces. Caries was present distally on the - 427 upper right canine tooth, and was rampant medially on the upper right third molar which also bore calculus but showed only slight wear facets.
Burial 28: This approximately nine-year-old child (upper right central and lateral incisor teeth were of equal height while upper left central and lateral incisors were slightly behind) lay about two metres to the right of Burial 27 and parallel to it. The head of the complete, outstretched skeleton was turned slightly to the left and looking down at 45°. The right arm was along the body, while the left arm was bent 90° at the elbow so that the left hand lay on the right thigh. The toes pointed forwards.
Burial 29: A four-year-old child was interred next to Burial 28 forming an angle of about 10°with it. The skeleton was complete but fragmentary. The head was turned to the left, arms by sides, legs slightly flexed outwards at the knees, toes together. Age determination was based on the ‘collar’ on the upper right first molar tooth and the slight ‘collar’ on the upper right central incisor.
Burial 30: Here was a three-year-old child (there was a slight ‘collar’ on the upper left first molar tooth) lying about 1.5 metres to the right of Burial 29. The delicate skeleton was complete and the epiphyses were undisturbed. The head was turned 45° to the right and looking down at an angle of 30°. However, the mandible faced forwards. The right arm was by the side, while the left elbow was slightly bent so that the hand was on the hip. The right leg was straight but the left was slightly bent so that the tibia was angled inwards.
Burial 31: A two-to-three-year-old child lay about one metre to the right of Burial 30. The complete, undisturbed, extended skeleton was in good condition. The head was turned to the right and looking downwards at an angle of 45°. Arms were by the sides, palms down. Age determination was based on the lower left first molar tooth being well down inside the mandible.
Burial 32: This was the articulated skeleton of a young dog lying curled on its right side, its legs flexed beneath it. Although fragmentary, the bones were recognisable and identified by the pelvis shape. Non-fusion of the olecranon process showed this to be a pup. It lay to the left of Burial 31 and due south of Burial 1's cranium.
Burial 33: This was represented by fragmented right parietal and frontal regions of the infant cranium facing to the left, and two fragments of long bone. The burial lay close to Burial 27 and immediately behind the burial row.
- 428 Page is blank