An assessment of different metrical methods of sex estimation using long bones

Abstract

Sex estimation from long bones metrics represents an alternative in an archaeological context, especially when the pelvic bone or cranial bones are too fragmented. This study presents an evaluation of univariate and bivariate methods, applied to an Italian reference collection housed at the Museum of Anthropology of the University of Florence, with measurements recorded in the Goldman Data Set. This collection consists of 76 individuals who died between 1900 and 1901. A total of 19 measurements were taken from long bones, including the humerus, radius, femur, and tibia. By studying sexual dimorphism from long bones using T-tests and MANOVA, the differences between male and female means were significant for many variables. The anteroposterior diameter of the radius (RAPD) was found to be the most sexually dimorphic variable (p<0.001), while the combination of this same variable and the maximum length of the radius (RML) was found to have the highest discriminatory power (p<0.001). Following a review of existing methods, those based on the most sexually dimorphic variables identified were selected. Regarding the univariate and bivariate methods, between 79% and 86% of individuals were correctly classified, suggesting that the methods evaluated, initially developed on non-Italian populations, could be potentially applied to an Italian sample.