Estimation of age at death from the microscopic structure of the femur

Abstract

Estimating age at death from skeletal remains can be done with relative accuracy when a skeleton is complete, however incomplete and/or poorly preserved skeletons pose a problem in assigning an accurate age range to unknown remains. Techniques for determining age at death from the microstructure of bone have shown to be relatively accurate in North American and European populations but, until recently, had not been attempted on a South African group. This technique is based on the fact that bone remodels and changes throughout an individual’s life. The purpose of this study was to develop standards for estimating age at death, using bone microstructure, that are applicable to a South African population. The sample consisted of 146 individuals (105 males and 41 females) of known age and sex. A 0.2 cm x 1.0 cm sample was removed from the anterior surface of the femur, and slides were prepared according to standard histological methodology. Ten variables, which included the total osteon count (measurable and non-measurable), the average Haversian canal diameter, the average number of lamellae per osteon, the total number of osteon fragments, the number of non-haversian canals, the total number of resorption spaces and the average percentage of osteonal bone, unremodeled bone and fragmental bone were assessed. The relationship between the changes in each of the variables with age was examined. Four variables demonstrated significant correlation with age and included the total osteon count (r = 0.53), the percentage unremodeled bone (r = -0.53), the total number of non-haversian canals (r = -0.55) and the average percentage of fragmental bone (r = 0.55). These variables were then used to calculate single and multiple linear regression formulae to determine age. Coefficient of determination (r2) for multiple regression analyses ranged from r2 = 0.27 to 0.42. The general range of the standard error of the estimate (SEE) for this study ranged between 13.31 and 14.04 years and is similar to the results of previous studies. Various factors may have contributed to low r2 values, such as poor nutrition, mechanical stress to the bones and misreporting of age of the individuals in the sample. The microscopic techniques appear to not be more accurate than macroscopic methods for estimating age at death, but can be useful in situations were remains are fragmentary or in combination with macroscopic methods.