AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 76:183-188 (1988)

Age and Sex Biases in the Preservation of Human Skeletal Remains

PHILLIP L. WALKER, JOHN R. JOHNSON, AND PATRICIA M. LAMBERT

Department of Anthropology, University of California, Santa Barbara, California 93106 KEY WORDS Mortality profiles, Demography, Vital statistics, American Indians, Archeology ABSTRACT Inaccuracies introduced through biases in preservation are a major source of error in paleodemographic reconstructions. Although it is generally assumed that such biases exist, little is known about their magnitude. To investigate this problem, we studied age and sex differences in the preservation of skeletal remains from Mission La Purisima and a prehistoric cemetery (Ca-Ven-110). Comparison of mortality profiles obtained through analysis of skeletal remains and burial records from the mission indicates that biases in preservation can be very significant in poorly preserved skeletal collections. The Purisima burial records show that most of the people interred in the cemetery were either infants or elderly adults. The skeletal remains, in contrast, are predominantly those of young adults. The burial records and skeletal collection produced comparable sex ratios. These results show that age biases in preservation are much more important than sex biases. This conclusion is supported by data on the completeness of the skeletons from La Purisima and Ca-Ven110. At both sites, the remains of young adults were better preserved than those of children or elderly adults, and the completeness of male and female skeletons was comparable. The use of archeological skeletal remains to make paleodemographic reconstructions rests on the assumption that the age and sex distributions of the skeletal sample reflect those of the burial population they are derived from and that any biases introduced through differential preservation can be recognized and accounted for (Ubelaker, 1984). Although osteologists are generally aware that the comparatively rapid disintegration of the poorly calcified remains of children, elderly adults, and females distort mortality profiles of skeletal collections, little is known about the magnitude of these biases. To investigate this problem, we analyzed skeletal collections from La Purisima Mission and a prehistoric southern California cemetery. The skeletal collection from the Purisima Mission cemetery provides an unusual opportunity to assess the effects of differential preservation on the demographic structure of a burial population. Comparison of mortality profiles based on the burial records and skeletal remains from the mission cemetery indicates that major discrepancies have been introduced through the rapid disintegration of the poorly calcified remains of children and elderly adults. Our studies of the preservation of skeletal remains from Ca-Ven-110 provide additional evidence for age-related biases in preservation.
SKELETAL PRESERVATION AT MISSION LA PURISIMA

A comparison was made of skeletal remains recovered from the Purisima Mission cemetery near Lompoc, California and burial records kept by the Franciscan priests who operated the mission cemetery. The records indicate that it was used as a burial place for the mission's Indian neophytes between 1813 and 1849. After the abandonment of the mission during the latter part of the nineteenth century, the exact location of the cemetery was forgotten.
Received November 7, 1985; accepted January 11, 1988.

1988 ALAN R. LISS, INC.

18 4

P.L. WALKER ET AL.

TABLE 1. Age distribution of individuals listed in the Purisima cemetery burial records

TABLE 2. Age and sex distributions derived from analysis of burial records and skeletal collections from Mission La Purisima'

'Numbers in parentheses indicate percent of total population within each age group for each source.

In 1964, Richard Humphrey (1965) did archeological work to relocate and determine the extent of the cemetery in an area scheduled for development. During these excavations, the remains of over 50 individuals were discovered in the impact zone. These burials were poorly preserved, and only 32 of the skeletons were in good enough condition to remove for analysis. The poor preservation of the skeletons seems to be

explained by the sandy soil of the cemetery, which allowed water to permeate the burials during the rainy season and then evaporate during the summer months. This periodic soaking and drying has resulted in the disintegration of fragile bones such as vertebrae and ribs, which are nearly absent from the collection. Poor preservation often made age and sex determinations difficult. Pelvic remains were

PRESERVATION BIAS IN PALEODEMOGRAPHY

185

available for only a few of the burials. In most cases, it was necessary to use cranial morphology and long-bone size and robusticity as the basis for sex determination. Age estimates were made based on the extent of dental wear, cranial suture closure, and the severity of osteoarthritis. Because of the limitations of these aging criteria, it was only possible to assign the burials to one of the following classes: <18, 18-24, 25-45, and >45 years of age at death. The actual age at death of the neophytes buried in the cemetery was determined using the Libro de Bautismos and Libro de Entierros, in which the mission priests recorded all births and deaths. Although the priests sometimes entered an estimate of a person's age in the Libro de Entierros at the time of their burial, these age estimates are often unreliable (Cook and Borah, 1979). The age estimates we have used were calculated by adding the age of a person in the year of their baptism to the difference between the year of their baptism and the year of their death (see Table 1). Thirty three persons for whom age could not be determined were deleted from our tabulations. Not everyone listed in the Libro de Entierros was interred in the mission cemetery. Some neophytes died and were buried while visiting in distant villages or at other missions; these were omitted from our calculations. Conversely, some who are not listed in the Libro de Entierros were buried at the Purisima cemetery and are included in Tables 1 and 2. This latter group consists primarily of 15 neophytes from Mission Santa Ines who died at Purisima following the Chumash Revolt of 1824. Their burials are recorded in the Libro de Entierros of Santa Ines even though they were interred at Purisima. Analysis of the burial records and skeletal collection from the mission cemetery produced similar sex ratios. The mission records show that females accounted for 53% (n = 536) and males for 48% (n = 484) of the adults over 17 years of age buried in the cemetery (Table 2). In the skeletal collection, males and females each accounted for 50% of the adult population. The chi-squared test shows that the difference between the sex ratios derived from the burial records and skeletal collection is not statistically significant (X2 = 0.005; d.f. = 1; P = 0.92). Comparison of the age distribution of the burial population recorded in the Libro de

Entierros and the skeletal collection from the

Purisima cemetery reveals major discrepancies. People who were younger than 18 years of age at death accounted for 32% (n = 470) of the people recorded as buried in the cemetery. Only 6% (n = 2) of the individuals in the skeletal collection were younger than 18 years of age at death. The chi-squared test indicates that the difference between the age distribution of the people listed in the burial register and the people whose remains were recovered from the cemetery is highly significant (X 2 = 43.3; d.f. = 3; P = 0.00001). However, the La Purisima skeletal sample presents certain problems. The excavated burials represent a small portion of the cemetery population, and sampling error cannot be ruled out as a possible explanation of the discrepancy between the burial records and the skeletal collection. For this reason, a second method was used to assess age and sex biases in preservation. The completeness of each burial was evaluated based on the presence of 14 long bones: claviculae, humeri, radii, ulnae, femora, tibiae, fibulae. A completeness score was computed for each individual by dividing the number of these bones present by 14. For example, a burial with only 2 of 14 bones was assigned a score of .14 (2/14). Males appear to be slightly less complete than females according to this index (males, 27%; females, 32%). However, both of the elderly individuals (>45 years) in the Purisima collection are males. If these two are excluded from the sample, males and females are equally well preserved (males, 31%; females, 32%). The indices by age group are as follows: < 18 years, 0%;18-25 years, 38%; 26-45 years, 24%; >45 years, 7%. These figures indicate that the skeletons of juveniles and elderly individuals at the Purisima cemetery are less well preserved than those of young and middle aged adults. These data are consistent with our conclusion, based on comparison of the burial records and skeletal collection, that elderly people are likely to be underrepresented in poorly preserved skeletal collections.
SKELETAL PRESERVATION AT CA-VEN-110

To further evaluate the significance of age and sex biases in preservation, we analyzed a larger, better preserved collection from the Late-Middle Period cemetery site of Ca-Ven110 in Ventura County, California. This site

186

P.L. WALKER ET AL.

'The preservation index is the average proportion of 14 long bones preserved in the burials from each age group. The calculation of the index is described in the text.

was excavated in December of 1986 by archeologists from California State University at Northridge. Data were collected on the presence, absence, and completeness of the 14 bones listed above as part of the osteological analysis of the burials from the site. An estimate of how much of each bone was preserved was calculated by dividing the observed length of the preserved portion by the expected total length of the complete bone. The expected length of incomplete bones was determined based on the length of the opposing element when this was present and complete, and was estimated from population averages when this information was not available. Missing bones were recorded as 0% present. The average completeness of an individual's long bones was then calculated by summing percentages for all bones and dividing by 14. Indices of preservation were calculated by averaging individual burial scores for each age and sex class. To remove biases caused by the disturbance of burials through erosion, only those individuals with at least 7 of 14 bones present were used in the calculation of the indices. The age of the Ca-Ven-110 burials was determined based on public symphysis morphology, epiphysis fusion, tooth eruption, and dental wear. Equations derived through regressing tooth heights against pubic ages were used to age some of the older adults ( Walker and Dean, 1988). Individuals were

assigned to the following age groups: 0-10, 10-20, 20-40, 40-60, >60 years. Although older individuals are more difficult to accurately age, and their absolute ages may be somewhat in error, the age group >60 years nonetheless contains the oldest individuals in the sample. Sex determination was based on pelvic features, cranial morphology, and discriminate analysis of long-bone measurements. In no instance was the latter criterion used exclusively. Only individuals identified as discrete burials in the field were used in the analysis. A pattern of differential preservation similar to that at Purisima is apparent in the Ca-Ven-110 material. The indices show that the best-preserved individuals belong to the 20-40-year-old age group, and that infants, children, and the elderly are the least well preserved (Table 3). The indices do not provide any clear evidence of sex differences in preservation (Table 4).

DISCUSSION

There are nearly always more males than females in skeletal collections from archeological sites (Weiss, 1972). This has been explained in part by the comparatively rapid disintegration of lightly built female skeletons (Bennike, 1985). This tendency for female skeletons to be lost from the arche-

PRESERVATION BIAS IN PALEODEMOGRAPHY

187

ological record more rapidly than those of males is thought to be especially pronounced among postmenopausal women who experience osteoporosis after the cessation of ovarian function (Raisz, 1982). The close correspondence between the sex ratio of the Purisima skeletal collection and that recorded in the burial records shows that sex differences in rate of skeletal disintegration have not significantly biased this collection. The burial records show that most of the people buried in the Purisima cemetery were either infants, children, or elderly adults. The skeletal remains excavated from the cemetery, in contrast, are predominately those of young adults. The underrepresentation of young children in the skeletal collection is most likely a result of the comparatively rapid disintegration of their incompletely calcified bones (Gordon and Buikstra, 1981). We have considered the possibility that a separate cemetery, not sampled during the archeological excavations, was maintained for children at the Purisima cemetery, and that this explains the underrepresentation of children in the skeletal collection. This is unlikely for several reasons. We have baptismal records for all of the children listed in the burial records, so there is no reason for them not to have been buried in consecrated ground along with other members of the church. Although underrepresented, the remains of children are in fact present in the skeletal collection from the cemetery. This would not be the case if burials were strictly segregated according to their age. Finally, there is no mention of a children's cemetery in the burial records or other historical sources. The underrepresentation of elderly people in the mission skeletal collection is probably explained at least in part by a decrease in the resistance of their bones to disintegration due to the accelerated loss of bone calcium in old age. The Ca-Ven-110 data provides additional evidence for age biases in preservation. In this cemetery, as at Purisima, juvenile and elderly individuals are less complete and more fragmented than young and middle-aged adults. The Ca-Ven-110 collection is also similar to Purisima in the lack of sex differences in preservation. It is interesting to note how closely the age distribution of the La Purisima skeletons parallels age-related changes in the

skeletal mass of living individuals. Skeletal mass increases until about the age of 20 years, remains relatively constant in young adults, and then begins to decrease after the age of about 40 or 50 years (Garn, 1970). This correspondence supports the hypothesis that age-related changes in the skeleton significantly alter the probability that a person's remains will be preserved in the archeological record.
CONCLUSIONS

Comparison of mortality profiles derived from the Mission La Purisima skeletal collection and the mission burial records reveals major inconsistencies (Fig. 1). Infants and elderly adults are markedly underrepresented in the mission skeletal collection, and are less complete in the prehistoric material. These discrepancies can be attributed in large part to the susceptibility of the remains of people from these age classes to disintegration in the ground. Although the Purisima collection may be an extreme example in which differential preservation has greatly skewed the age structure of a skeletal collection away from that of the original burial population, most collections of skeletal remains from cemetery sites are probably biased to some extent in favor of young adults, whose remains are comparatively resistant to disintegration. Evidence from the Ca-Ven-110 cemetery supports this conclusion. If conditions in the depositional environment remain constant, the magnitude of this age-related preser-

Fig. 1. Comparison of mortality profiles derived from analysis of the burial and skeletal collections from the Purisima cemetery.

188

P.L. WALKER ET AL.

vation bias should be roughly proportional to the length of time a group of burials has been in the ground. Older cemeteries would thus be expected to contain fewer infants and elderly people than more recent cemeteries, even if the age structures of the original burial populations were identical. Acceptance of age-specific differential preservation as a significant source of bias in skeletal collections provides a basis for evaluating certain classes of hypotheses about differences in the age structure of burial populations exposed to similar depositional environments. If children or elderly adults are more common in a skeletal collection from an ancient cemetery than from a recent one, this is strong evidence that people from these age classes actually were more common in the ancient burial population. If, on the other hand, infants or elderly people are more common in a skeletal collection from a recent cemetery than they are in an ancient one, much less can be inferred about differences in the original age structure of the two burial populations. Such a difference would be expected due to differential preservation, even if the age structures of the two burial populations were identical. Our data suggest that in poorly preserved skeletal collections such as the one from Mission La Purisima, the underrepresentation of immature individuals can be so great that little evidence remains regarding the original age structure of the burial population. Osteologists should, consequently, be extremely cautious when attempting to reconstruct the demographic structure of prehistoric populations from collections of poorly preserved skeletal remains.

The mission burial records and skeletal collection produced comparable sex ratios when all age classes were pooled. Similarly, sex differences in fragmentation were not demonstrated at Ca-Ven-110. From this, it appears that biases in mortality profiles resulting from the relatively rapid disintegration of gracile female skeletons may not be very significant.
ACKNOWLEDGMENTS

We thank Mike Imwalle for helping us check some of the data in the Libro de Entierros and the students who assisted in the analysis of the Purisima skeletal remains. This research was supported by National Science Foundation grant BNS 85-07836.
LITERATURE CITED Bennike P (1985) Paleopathology of Danish Skeletons: A comparative study of Demography, Disease and Injury. Copenhagen: Akademisk Forlag. Cook SF, and Borah W (1979) Essays in Population History: Mexico and California, Volume 3. Berkeley: University of California. Garn S (1970) The Earlier Gain and the Later Loss of Cortical Bone. Springfield, Illinois: Charles C. Thomas Gordon CG, and Buikstra JE (1981) Soil pH, bone preservation, and sampling bias at mortuary sites. Am. Antiquity 46(3):566-571. Humphrey RV (1965) The La Purisima Mission cemetery. Ann. Rep. Univ. Calif. Archaeo. Surv. 7:179-192. Raisz LG (1982) Osteoporosis. J. Am. Geriatr. Soc. 30:127-138. Ubelaker DH (1984) Human Skeletal Remains: Excavation, Analysis, Interpretation (revised edition). Washington, D.C.: Taraxacum. Walker PL, and Dean GL (1988) Dental dimensions as indices of wear rate and age at death. Am. J. Phys. Anthropol. 75(2):285. Weiss KM (1972) On the systematic bias in skeletal sexing. Am. J. Phys. Anthropol. 37:239-250.