Abstract:
Research on the estimation of age at death, sex and stature from skeletal remains has received more attention than methods used to evaluate ancestry. While this may be due to the stigma attached to classifying people into groups, the application, interpretation and precision of non-metric methods used to predict ancestry need to be examined; as these variables are routinely applied to forensic case work in South Africa. The aim of this study was to score fifteen non-metric cranial traits, namely nasal bone structure, nasal breadth, nasal overgrowth, anterior nasal spine, inferior nasal margin, interorbital breadth, zygomaxillary suture shape, malar tubercle, alveolar prognathism, mandibular and palatine tori, shovelshaped incisors, Carabelli’s cusps and the transverse palatine suture shape on a South African sample, with the intent to assess the influence of sex, ancestry and age at death on these facial features. A total of 520 crania were obtained from the Pretoria Bone, Raymond A. Dart and Kirsten Collections in South Africa and included 237 (135 males, 102 females) Africans, 158 (94 males, 63 females) Europeans and 125 (87 males, 38 females) persons of Coloured origin. Data were analyzed using SPSS v.11.5 for Windows. Ordinal regression was used to evaluate the effect the independent variables (age, sex and ancestry) had on the dependent variable (non-metric traits). Results showed that all the variables were associated with ancestral differences among and within groups. Interorbital breadth, nasal bone structure, nasal breadth and shovel-shaped incisors exhibited statistically significant interactions with sex and ancestry, whereas the appearance of the anterior nasal spine, alveolar prognathism, incisor shovelling of the upper incisors, and Carabelli’s cusp morphology were correlated with age at death. If traditional classification methods are used, then these non-metric traits are not a valid prediction of ancestry in South African populations. Future research is to focus on several statistical approaches, including multi-variate analysis, for the classification of non-metric traits. Copyright
