Volume 102 1993 > Volume 102, No. 3 > Tooth ablation in old Hawai'i, by Michael Pietrusewsky and Michele T. Douglas, p 255-272
TOOTH ABLATION IN OLD HAWAI'I
Tooth ablation, or the deliberate removal of anterior teeth during life, has been observed in a great many human groups. It has been recorded in archaeological series as far removed as Japan (e.g., Hideji 1986), Taiwan (Lien 1987, 1989), south-east Asia(Sangvichien et al. 1969), Africa (Dahlberg 1963), Europe, Siberia and the New World (Hrdlicka 1940) and Australia (Pounder 1984). The custom has been further documented in the ethnographic literature for a number of near contemporary groups, including Taiwan Aborigines, Africans and Australian Aborigines. The removal of the anterior teeth is usually associated with some rite of passage such as puberty initiation, marriage or death. Tooth ablation was often practised to demonstrate an individual's tolerance for pain and, if performed correctly, left a permanent and visible record. Shaw (1931), quoting Frazer (1910), notes that the extraction of teeth as a sign of mourning was relatively common among Mongoloids. Milner and Larsen (1991) present an excellent review of intentional alteration of teeth, including tooth ablation.
In Hawai'i, the practice of deliberately extracting teeth as a sign of mourning the death of a loved one (manewanewa), is well documented (Handy and Pukui 1972:156). Ellis (1979:119) states that:
When a chief died, those most anxious to shew [sic] their respect for him or his family would be the first to knock out with a stone one of their front teeth.
Sinclair (1971:5-6) relates an episode in the life of Keopuolani, later to become the sacred wife of Kamehameha I. Kamehameha, while on a visit to Moloka'i in 1790, is said to have mourned the death of Kalola, Keopuolani's grandmother, by knocking out some of his teeth.
Ellis (1979:119-20) further notes that only one tooth was removed at a time, although the practice may have been repeated with each subsequent death of an individual of rank or authority, and continued until all the front teeth had been removed. Ellis states (p. 120) that:
there are few men to be seen, who had arrived at maturity before the introduction of Christianity to the islands, with an entire set of teeth; and many by this custom have lost the front teeth on both the upper and lower jaw….
Further, according to Ellis, the method of removal entailed placing one end - 256 of a stick against the tooth and striking the other end with a rock until the tooth had been broken out (Fig. 1).
Figure 1. Hawaiians practised tooth ablation as one of the ways to signify that they were mourning the death of a chief or loved one.
Sometimes, the custom was self-inflicted, but more frequently it was done by another person (Ellis 1979; Buck 1964; Kamakau 1964; Kelly 1978). Ellis further writes that, if the men failed to observe this practice, the women might knock out the teeth while the men were sleeping.
These early accounts of tooth ablation in Hawai'i give the impression that the custom was widespread and frequent. According to Ellis, the custom had disappeared by 1823:
…all the wicked practices, and most of the ceremonies usual on these occasions have entirely ceased. Knocking out of the teeth is discontinued… (1979:124).
Curiously, two other native Hawaiian scholars, I'i (1983) and Malo (1951), writing in the 1800s, do not mention tooth ablation as being practised during their lifetimes.
One of the first studies of Hawaiian skeletal remains to mention this custom - 257 was by Allen (1898), who observed the absence of a number of upper incisor teeth in several specimens he examined. Allen stated (p.43) that:
The natives of the Sandwich Islands were in the habit of knocking out some of the upper front teeth as a sign of mourning for the death of a chief.
In a later study, Chappel (1927) described tooth ablation in Hawaiian skeletal remains he observed during a visit to the Bishop Museum in 1920. In contrast with Allen's findings, Chappel reported that both upper and lower incisors were removed and that the lower incisors were more commonly missing than the upper ones. Chappel further stated that the custom appears to have been more common in males than in females and in specimens from Hawai'i Island. Unfortunately, except for the island of origin, the exact provenance of these specimens is unknown. Chappel assumed that most of the specimens he examined were precontact in origin.
More recent osteological studies report tooth ablation in Hawaiian remains from 'Anaeho'omalu (Pietrusewsky 1971; Pietrusewsky et al. 1990), Keōpū (Collins 1986), South Point (Underwood 1969) and Kalahuipua'a (Kam 1979) on Hawai'i Island. Only sporadic cases of tooth ablation, mostly late in time, have been reported in remains from O'ahu (Pietrusewsky and Douglas 1990; Pietrusewsky et al. 1989b). More interesting is the apparent absence of tooth ablation in early remains from Mōkapu (Snow 1974), Kualoa (Pietrusewsky and Douglas 1989a), Bellows Beach (Pearson et al. 1971), all on O'ahu, and its limited occurrence in remains from Maui (Pietrusewsky et al. 1991) and Kaua'i (Douglas and Ikehara 1991; Pietrusewsky et al. 1992).
Evidence of tooth ablation, ethnographic or skeletal, in other parts of the Pacific is largely lacking. Ellis (1979) compares mourning customs of Hawai'i and other South Pacific areas, but he does not indicate that the custom was practised anywhere but in Hawai'i.
As this brief review demonstrates, there are some uncertain, as well as conflicting, reports regarding tooth ablation in Hawaiian culture. Was tooth ablation widely practised in precontact Hawai'i, as implied in some of the earlier ethnographic accounts, or was it restricted to relatively few individuals from one or a few islands? Can any temporal limits be imposed on the custom? What was the most frequent pattern of ablation practised by early Hawaiians? Was tooth ablation more common in males than in females, as early studies suggest? Was there an age limitation? And, finally, how and why did early Hawaiians intentionally avulse their teeth?
In this paper we examine these and related questions regarding the ritual removal of teeth in early Hawaiians. Where appropriate, tests of significance are - 258 applied to the original data for determining the pattern, frequency and the temporal, spatial and age limits of the custom. The method of extraction and possible reasons for this unique practice are also discussed.
Data that record the presence and absence of teeth in 11 Hawaiian skeletal series are assessed to determine the frequencies of tooth ablation in Hawaiian remains. Representative samples of adolescents and adults of known sex from six islands (O'ahu, Hawai'i, Maui, Kaua'i, Lāna'i, Moloka'i) are assembled.
The skeletal series used and the number of individuals in each is given in Table 1. The approximate location of the skeletal series is shown in Figure 2.
Figure 2. Map of the Hawaiian Islands showing the location of the skeletal and dental series used in the present study.
Although exact provenances are not known for some of these remains, approximate temporal sequences may be inferred by mortuary behaviour and associated artefacts. 'Anaeho'omalu represents 100 individuals interred in burial caves on the desolate west coast of the island of Hawai'i. The remains, excavated in 1969/70 were initially analysed by Pietrusewsky (1971) and then re-analysed in 1990 before their reburial (Pietrusewsky et al.1990). One of the highest frequencies of deliberate tooth extraction in Hawai'i was noted in these remains. Another- 259
TABLE 1: Frequency of Tooth Evulsion by Individual in Hawaiian Skeletal Remains
sample from the island of Hawai'i is a small, historic cluster of 24 burials from Kona ('Auhaukea'ē) analysed by Pietrusewsky and Douglas (1989b). There was a single headstone on this site dated 1894. Lineal descendants report that this cemetery was used over a relatively brief 16-year period.
The island of Maui is represented by two samples. The largest skeletal series excavated in Hawaiian history is from the west coast of Maui at Honokahua (N=712). The burials are believed to be prehistoric in provenance and may extend from relatively ancient to contact times (Pietrusewsky et al.1991). The second sample of excavated material is from Kahoma Stream in Lahaina on the south-west coast of Maui (N=23). The Kahoma burials are believed to represent mixed prehistoric and historic interments that had been disturbed and subsequently reinterred in the early 1950s in a commingled manner (Pietrusewsky et al. 1989a).
O'ahu is represented by three skeletal series. The first is represented by mostly prehistoric burials (41 prehistoric, one historic) from Kualoa on the windward side of the island, an area believed to have been used by the ruling class in the late prehistoric period (Pietrusewsky and Douglas 1989a). The second series from this island is an historic series of 28 individuals interred during the smallpox epidemic of 1854, excavated from the Kaka'ako District in Honolulu and described by Pietrusewsky et al. (1989b). A third series from O'ahu includes 29 prehistoric and five historic burials recovered in three separate excavations from the Wai'anae coast. Five individuals were recovered from this site in 1984 (Collins 1984), 10 more in 1985 (Pietrusewsky and Ikehara 1985) and most recently, 19 in 1988-9 (Pietrusewsky and Douglas 1990).
Two prehistoric skeletal series from Kaua'i are included in the present study. The first sample includes 23 individuals excavated at Keoneloa, a sand dune site on the south-west shore of Kaua'i (Douglas and Ikehara 1991). The second sample includes approximately 30 individuals, of which 14 are used in the present study, from a beach site at Hā'ena on the north shore of Kaua'i (Pietrusewsky et al. 1992).
Finally, we include a few individual specimens described by Chappel (1929) from the islands of Lāna'i (N=83) and Moloka'i (N=5). The latter sample is extremely small, poorly provenanced and probably subject to collector's bias.
Determining whether teeth have been loss before death (premortem tooth loss) or after the death of the individual is relatively easy. Empty, unremodelled tooth sockets indicate the loss occurred after death while tooth absence in the presence of alveolar bone resorption generally indicates premortem tooth loss. Premortem tooth loss may be caused by a number of processes, including dental pathology (tooth decay, dental abscessing, periodontal disease), accidental - 261 trauma, unerupted and/or impacted teeth and cultural modification. Although most premortem tooth loss in archaeological remains is the result of pathogenesis, the purposeful or deliberate removal of teeth, especially of the anterior teeth, as a rite of passage has been recorded for a number of archaeological as well as living human groups. Although difficult to demonstrate conclusively, identification of tooth ablation in archaeological remains frequently relies on the recognition of a repetitive pattern of missing teeth in the presence of otherwise good dental health. Other criteria used by physical anthropologists to identify tooth ablation include the spacing of teeth, symmetry and non-randomness of loss, the absence of significant dental pathology, presence of roots, fractures of the labial portion of the alveolus and ethnographic accounts. Most important among these criteria which are suggestive of a cultural behaviour is the recognition of a specific and repetitive pattern of tooth loss.
Because deliberate tooth ablation relies on the recognition of a specific pattern of tooth loss, information recorded by other researchers on the presence or absence of teeth in different archaeological dental series may be difficult to compare. In the present study, the presence or absence of each tooth, or tooth socket, was scored selecting one of the following categories: present, premortem loss, postmortem loss, root present, ablation, agenesis, tooth erupting, unerupted and impacted. It should be reiterated that this system of data collection relies upon the examiner's knowledge and recognition of a pattern of tooth loss which is consistent with deliberate ablation. Further, this system does not allow us to distinguish between those cases where the tooth was avulsed completely and those situations where root fragments remain, indicating the tooth was not completely removed.
We, together or singly, examined all of the dental remains in this analysis except for the large sample from Honokahua on the island of Maui and the relatively small series from Lāna'i and Moloka'i. Because the information recorded for the Honokahua series was often made by individuals not familiar with Hawaiian osteology, some manipulation of the data was required. Tooth presence data in this latter series were reviewed for each individual with an emphasis on identifying a typical pattern of tooth ablation and then rescoring these observations. In most cases this meant changing observations recorded as premortem tooth loss to ablation and in some cases observations of the presence of roots to ablation. Adopting this method, 18 individual records in the Honokahua series were altered. Similarly, two records in the 'Anaeho'omalu sample and one record each in the Kaka'ako, O'ahu and 'Auhaukea'ē samples were altered. A partial photographic and radiographic record was available for checking the validity of some of the recorded observations of ablation in these series to confirm the presence or absence of a broken root in the alveolus.
Using the Statistical Analysis System (SAS) of computer programs, the - 262 significance of the observed patterns of tooth ablation and its distribution by sex, age-at-death, time period and island (by individual and/or tooth category), is tested using the chi-square statistic (X2). The assumptions of the chi-square statistic are: that the measurement is made at the nominal scale or higher, that there are at least two samples and two mutually exclusive categories into which the samples may be placed, that no category should have an expected frequency of less than one and not more than one in five should have an expected frequency less than five (Norcliffe 1985). In cases where the expected frequency is less than five, a continuity-corrected chi-square statistic (Yates' correction factor) is reported (Thomas 1986).
Five-hundred and eighty-nine individuals from six of the Hawaiian Islands and three time-periods were used in this analysis of tooth ablation. Females (342) outnumber males (247). Adolescent and adult age cohorts are represented. The frequency of tooth ablation by individual observed in each of skeletal series is given in Table 1. These figures indicate there are differences between sexes, sites and islands. The highest occurrence of tooth ablation (sexes combined) occurs in the 'Anaeho'omalu sample (44.0%) followed by samples from Moloka'i (40.0%) and Kaka'ako (28.6%). The high rate of occurrence in the sample from Moloka'i is possibly the result of collector sampling error as well as small sample size. In only two samples ('Anaeho'omalu and Wai'anae) does the frequency of tooth ablation in females exceed that in males.
The overall frequency of tooth ablation, using all 11 Hawaiian series, is greater in males (12.6%) than in females (7.0%) and this difference is statistically significant (X2=5.186, p<0.05, d.f.=1). However, when the difference in the frequency of tooth ablation in males and females was tested for significance in the 'Anaeho'omalu, Honokahua and Kaka'ako series, no significant difference was found.
The observed patterns of tooth ablation (by individual) are summarised in Table 2. When the pattern is examined in individuals possessing both maxillary and mandibular dental arcades, no clear pattern emerges and the sample size is drastically reduced. For this reason our observations are restricted to individual jaws. The most prevalent pattern (Figs 3 and 4) in the lower jaw in both sexes is ablation of all four incisors (6.9% in males, 2.3% in females and 4.2% overall). The second most common pattern in males is ablation of both upper central incisors (3.8%) while, in females, ablation of the mandibular central incisors (1.7%) is the next most common pattern. Ablation of both mandibular and maxillary central incisors occurs in 1.4% (5 males and 1 female) of individuals. Also evident from this table is that the variety of observed patterns is greater in females (14 different patterns) than in males (10 patterns). The chi-square test- 263
TABLE 2: Frequency of Occurrence of Patterns of Tooth Evulsion (by individuals)
for significance was applied to the frequencies in all patterns which have a sex difference and meet the assumptions of the test. A significant sex difference is found only in the frequency of ablation of all four of the mandibular incisors (X2 = 6.465, p<0.05, d.f.=l).
Figure 3. A drawing of a mandible showing that the lower incisors have been knocked out well before this individual's time-of-death. Removal of all four mandibular incisors was the most common pattern of tooth ablation in Hawaiians.
Figure 4. Ablation of all lower incisors in the mandible of a 17-19-year-old female from 'Anaeho'omalu, Hawai'i.
Examination of the patterns of tooth ablation by site reveals that most of the variability occurs in the samples from 'Anaeho'omalu and Honokahua. In the other samples, ablation of the mandibular incisors and maxillary central incisors are the only two patterns observed.
Frequencies of tooth ablation by individual and age group are presented in Table 3. Because no cases of ablation were observed in individuals with a nonspecific age estimate (i.e., individuals whose age could only be determined to be adult), these individuals (N=7) are not included in this table. The highest frequency of ablation occurs in the middle-aged interval (14.8%), followed by the young adults (7.8%). The lowest rate of ablation occurs in the adolescents (5.0%). There is a statistically significant difference in the frequency of tooth ablation by age group (X2 = 10.981, p<0.05, d.f.=3).
Review of the distribution of tooth ablation by age group within each of the sites was undertaken to see if there was variability. In the 'Anaeho'omalu sample, ablation was noted in young and middle-aged adults and a single adolescent. The Lāna'i sample ablation occurs in the young and middle-aged adults. The Moloka'i and Wai'anae samples have ablation only in middle-aged adults. In the Kaka'ako and Honokahua samples young, middle-aged and old individuals are noted to have avulsed teeth. A clear pattern with respect to age does not emerge from these results.
An apparent difference in the frequency of tooth ablation by island, with Moloka'i and Hawai'i Island having the highest frequencies, followed by O'ahu, was noted in Table 1. Because of limited provenance information for each individual within the sites, we cannot eliminate the possibility that the interisland difference is the result of a temporal difference rather than a geographic one. The distribution of tooth ablation by island, using the frequency of ablation per incisor tooth, is summarised in Table 4. The highest frequency is found in the Moloka'i series (27.8 %), followed by the island of Hawai'i (23.2 %) and O'ahu (7.0%). Tooth ablation on the remaining islands is rare (< 3.0%). Here, too, the sketchy provenance and possible collector bias of the Moloka'i sample may be skewing these results. A test of these inter-island differences is significant (X2 = 378.73, p<0.05,. d.f.=5).
The distribution of tooth ablation by temporal period is indicated in Table 5. The designated time periods are admittedly broad, but the practice of interment in shifting sand dunes without stratigraphic control and the frequency of disturbance and mixing of remains in burial cave sites makes dating of these sites problematic. The highest frequency of ablation is noted in individuals from the late prehistoric/early historic time periods (34.9%), followed by the historic samples (20.7%). The differences are significant (X2 = 62.071, p<0.05, d.f.=2). Reference to Table 1 will reveal that the bulk of the late prehistoric sample is from a single site ('Anaeho'omalu) on Hawai'i Island.- 266
TABLE 3: Frequency of Tooth Evulsion by Age (by individual)
Note: Seven individuals with a non-specific age (Adult) are not included in this table.
TABLE 4: Frequency of Incisor Evulsion by Island (by teeth)
TABLE 5: Frequency of Incisor Evulsion by Provenience (by individual)
Figure 5. Frontal view of the mandible of a 35-40-year-old male from 'Anaeho'omalu, Hawai'i, showing ablation of the lower incisors. Root fragments of both lateral incisors remain in the alveolus, indicating a traumatic, rather than extractive, method of removal.
A more detailed analysis of tooth ablation for the purpose of documenting the method of removal was conducted on the 'Anaeho'omalu skeletal sample (Pietrusewsky et al. 1990). Radiographs of the maxilla and mandible in 17 individuals with tooth ablation were taken to document the presence of residual tooth roots. In 11 individuals (64.7%) from one to four root fragments, ranging in size from 3 to 10 millimetres in length, remained in the alveolus (Fig. 5). All alveolar bone was healed. The presence of such a high frequency of root fragments supports the ethnographic descriptions of a traumatic rather than extractive method of tooth removal.
DISCUSSION AND CONCLUSIONS
Examining the distribution of the cultural practice of tooth ablation in 589 individuals in 11 Hawaiian skeletal samples from six islands has provided extensive new information on this unusual observable cultural behaviour and has generated some additional questions. Although problems remain — such as the - 268 lack of exact provenance for some of the skeletal series used in the present study — adequate data now exist for summarising deliberate tooth removal in Hawai'i.
The overall frequency of tooth ablation in these samples is relatively low (9.3%), suggesting that the practice was not as widespread as implied by some of the early ethnographic citations. The infrequent occurrence of tooth ablation in Hawaiian skeletal remains parallels Hrdlicka's observation that the practice was usually restricted to a few, possibly related, individuals in Siberian and Amerindian remains he examined (Hrdlicka 1940:31). The highest frequency of tooth ablation in Hawaiian skeletal series occurs in the 'Anaeho'omalu, Hawai'i cave site burials (44.0%), followed by a small, but likely biased, sample from Moloka'i (40.0%) and an historic sample from Kaka'ako, O'ahu (28.6%). Although the practice appears to be more common in males than in females, this difference is statistically significant only when all the samples are combined; the difference in the incidence of ablation between males and females within each individual sample was not found to be significant. Tooth ablation is never found among Hawaiian children and only one adolescent was observed to have mandibular incisor ablation. Middle-aged individuals have the highest occurrence of ablation, followed by young adults and then old individuals. Examining the distribution of the practice by the subjects' age-at-death within each of the samples produces no obvious patterns with respect to age.
Although several different patterns of ablation are demonstrated in these remains, the most common or classic pattern, in both sexes, is the deliberate removal of all four of the lower incisors. There is a statistically significant sex difference in the occurrence of this pattern of ablation. Allen's (1898), observations that the upper incisors were frequently lost in the limited series of Hawaiian skulls he examined is not supported by the results of our study. Because Allen does not provide good documentation, there are problems with accepting and/or confirming his observations. The second most common pattern of ablation in males is removal of the upper central incisors, whereas, in females, removal of the lower central incisors is the next most frequent pattern. Direct support for the theory of recurrent episodes of ablation of a single tooth cannot be found in these data. The extent of variability in the pattern of ablation suggests that the convention was not highly ritualised but, rather, there was some degree of personal choice. Radiographic analysis documents the presence of residual tooth roots in more than half of the 'Anaeho'omalu sample, supporting a traumatic method of extraction, rather than one which involved pulling or prying out of the teeth. This latter observation supports early ethnographic accounts that report the removal was by tooth-knocking.
Perhaps the most interesting finding supporting Chappel's initial conclusions is that tooth ablation is most common on the island of Hawai'i (accepting that the Moloka'i sample is biased). The interisland differences in the frequency of - 269 tooth ablation are statistically significant. Because ablation is not found in several early sites on O'ahu (before A.D. 1700) but is found in sites that overlap with the contact period, our study suggests that the practice of tooth ablation is possibly associated with the rise in power of the ali'i, especially Kamehameha I, and the interisland conquests occurring in the late 1700s and early 1800s. During these periods, warfare would have resulted in an increased number of deaths of beloved rulers and both Hawai'i Island and O'ahu were seats of Hawaiian power during this period. The appearance of this practice late in Hawaiian prehistory suggests an outside introduction, but we have no evidence to suggest a possible source. Interestingly, Hrdlicka (1940) reached a similar conclusion in his study of tooth ablation in Asian and New World populations.
The pattern of tooth ablation seen in Hawaiian remains appears to be quite different from patterns observed in other groups which have been studied. Whereas the lower incisors are the most frequently targeted teeth for removal in Hawaiian skeletal series, they are not avulsed in Taiwan Aborigines (prehistoric or those living in the ethnographic present), for example. Among Taiwanese, Lien (1987:176) reports that the most frequently removed teeth are the maxillary canines and/or the lateral incisors. The upper canines were also the most frequently removed teeth in prehistoric Jomon skeletons from Japan, although extraction of the lower incisors is also reported there (Hideji 1986). The practice in these cultures is presumably associated with rites of puberty. To our knowledge, tooth ablation has not been documented in other Polynesian populations. A recent study of remains from Tonga (previously described by Pietrusewsky (1969)) and Sigatoka, Fiji, indicate there is no tooth ablation in these series. The recovery and analysis of skeletal populations from other parts of the Hawaiian Islands, with attention to the temporal provenance, may enable a refinement of these conclusions.
We should like to thank Paul H. Rosendahl, Inc. [PHRI], Hilo, Hawai'i, and the Waikoloa Development Co. (Waikoloa, Hawai'i) for their assistance and support in the restudy of the 'Anaeho'omalu skeletal series which prompted the present study. Pareticular thanks to Patricia A. Kalima and Rona M. Ikehara for their help with this latter project. Mark Nakamura, of the Center for Instructional Support of the University of Hawai'i, produced the graphic illustrations used in this paper. Figure 1 was drawn by Larry Telea of PHRI, Hilo, and Julia Devrell drew Figure 3.
1 Denotes a significant difference (p<0.05)
2 Denotes a continuity corrected x2 value
3 Blank cells indicate chi-square assumptions were not met or there is no difference in occurrence of the pattern.
4 Denotes a significant difference (p<0.05)
5 Denotes a significant difference (p<0.05)
6 Denotes a significant difference (p<0.05)
7 Denotes a significant difference (p<0.05)