Cross-platform program for likelihood-based statistical comparisons of mortality profiles on a triangular graph

The construction of mortality profiles to investigate age-at-death patterns is a typical component of most faunal analyses. While many methods exist for constructing and comparing mortality profiles, plotting the percentages of juvenile, prime, and old individuals on a triangular graph, or ternary diagram, remains a popular method for making comparisons. Typically, these comparisons are made visually, but because sample sizes are often small, any differences may be the product of sampling rather than meaningful contrasts in depositional history. To overcome this problem, here we present a likelihood-based method for making statistical comparisons of mortality profiles on a triangular graph, and we make available a cross-platform computer program that implements the method. Although we developed the method with mortality profiles in mind, in principle, it can be used to analyze any kind of artifact for which there are three distinct categories.     

Evaluation of uniradicular teeth for age-at-death estimations in a sample from a Pampean hunter-gatherer cemetery (Argentina)

Age-at-death estimation is one of the most important aspects of bioarchaeological and forensic investigations. A set of analysis carried out with the osteological sample recovered from Chenque I site (Lihué Calel National Park, La Pampa province, Argentina) aims to test the applicability of multiple methods to obtain reliable information about its demographic composition and structure. In the present paper the results of the evaluation of the structure of uniradicular teeth are presented and discussed. It has been previously stated that chronological age is highly correlated with pulp dimensions in these teeth, because dentine deposition continue during all life. Strong correlation has been also identified with periodontal retraction and apical translucency. The analysis of the structure of these teeth offers useful information in order to obtain age-at-death estimations of the individuals they belonged to. Dental age-at-death estimations of the control sample were obtained evaluating some skeletal markers (pubic symphysis and auricular surface of coxae), and then compared with the dental analysis. The formulae whose results showed strong correlations with the skeletal estimations were applied to another sample, composed of teeth that were not associated with any skeletal marker of age. The evaluation of premolars and lower central incisors offers estimations that are consistent with those obtained from the evaluation of the pelvic bones, so that the application of these formulae is an alternative method to obtain estimations in archaeological samples from similar contexts.     

Culling profiles: the indeterminacy of archaeozoological data to survivorship curve modelling of sheep and goat herd maintenance strategies

The comparison of survivorship curves derived from seven different models aiming to reconstruct ancient sheep and goat herd maintenance strategies (e.g. optimization of wool, meat, and milk production) shows that many of these models cannot be distinguished statistically. This observation renders the current theoretical framework for reconstructing ancient herd maintenance strategies problematic, due to the possible indeterminacy of model data analysis. In order to assign empirically observed age-at-death data to a model of herd maintenance strategy, it is suggested that a direct fit of observed data to survivorship curves be forgone in favor of a binning procedure highlighting the differences between fewer and more distinguishable models. The incorporation of high-resolution sexing and taxonomic determination to coarse-grained age-at-death models may go a long way towards solving the current problem of indeterminacy.     

Age at death estimates for the protohistoric Arikara using molar attrition rates: A new quantification method

Dental attrition is positively related to ageing and has often been used to determine age at death of animals and prehistoric/historic humans. A newly designed molar-attrition-scoring standard was applied to a protohistoric skeletal sample, the Arikara (N = 143) from the North American Plains. Several criteria are used in producing a molar tooth crown wear (MTCW) score. These include: (i) the size of enamel wear facets, (ii) the amount of dentine and secondary dentine exposure, (iii) the height of the tooth, (iv) the enamel rim thickness, and (v) the amount of pulp-cavity exposure in the advanced stages. Age-at-death estimates were generated from ordinary least-squares (OLS) analysis, with estimated age regressed independently on the total molar attrition score (TMAS), mean molar-1 tooth crown wear (ȲM1TCW), mean molar-2 tooth crown wear (ȲM2TCW), and mean molar-3 tooth crown wear (ȲM3TCW). In all analyses (with the exception of the third molar), the coefficient of determination (r²) revealed a strong relationship between molar attrition and age. Multiple-regression analysis revealed a significant sex difference in molar-attrition rates. Besides intercept and slope differences, the female scatter showed greater molar attrition variation than the male scatter. There was no statistical difference between mean upper molar attrition score (UMAS) and lower molar attrition score (LMAS). There was no significant difference in molar attrition rates by side, and side by molar tooth.     

Molar crown height: not always a reliable method for the evaluation of age-at-death

In this study, we tested the validity of molar crown height, which changes according to the degree of tooth wear, for the evaluation of age-at-death. The sample consisted of 372 first and second molars (lower and upper) from 157 individuals of known sex and age-at-death. For each molar, we measured the height of the two cusps most subjected to wear (protocone and hypocone for the upper M1 and M2; protoconid and hypoconid for the lower M1 and M2). The correlation between crown height and age-at-death was assessed by linear regression analysis. The resulting models were not very robust since a significant correlation was only found for a small part of the sample, at best (maxillary M1) around 35%. The result slightly improved when bucco-lingual diameter (BL diameter), ante-mortem tooth loss and dental caries were considered, mainly for the maxillary M2 for which the model using age-at-death and BL diameter as independent variables explained 47% of the sample (p < 0.001).