A Comparison of two methods for estimating age at death from bone microstructure
MetadataShow full item record
AbstractThis thesis was undertaken to assess the accuracy of two means of estimating age at death using microstructural quantification. The techniques that were chosen for comparison were the core method of Thompson and Kerley's method which requires a complete cross section of bone. It is important to assess if Thompson's method is as. accurate as Kerley's in estimating age at death because it requires the use of a much smaller piece of bone and is therefore less destructive. This has important archaeological implications because museum specimens or other rare skeletal materials can not be subjected to extensive damage. Prior to the undertaking of this thesis a comparative assessment of the two techniques had not been completed The accuracy of both methods was assessed on the femora from twenty four individuals of known age at death. These individuals were aged 21 to 90 years. Kerley's method was assessed using eighteen cross sections. Thompson's method was assessed using eighteen cross sections and six bone cores. The results of this research demonstrate that the two methods provide different age at death estimates. Kerley's method provides the best accuracy in estimating age at death both for the complete sample and for subsets of the sample. These subsets were males, females, individuals younger than 45 years of age and individuals older than 45 years of age. The accuracy of Kerley's method varies as a function of field size and the sex of the individual being examined. Certain combinations of these factors provide age estimates that are less accurate than provided by Thompson's method. Kerley's technique was assessed using field sizes similar to those suggested to have been used in his original research, in addition to a smaller field size. It became apparent that the application of a correction factor to compensate for differences in field size was not fully effective. For males, a disproportionately large number of osteons directly adjacent to the periosteal surface created corrected osteon counts in the small field size which exceeded those from the large field size. This created many errors in age estimation. The same phenomenon was not noted for females and age estimates from the smaller field size produced better age estimates for females. Different field sizes produced variation in the accuracy of estimating age at death in males and females suggesting that each sex remodels bone uniquely at the subperiosteal region. This conclusion needs to be assessed further in future studies. Both Kerley's and Thompson's methods are poor at assessing the age at death of the very old. Microstructural assessments of age at death might be limited by the nature of bone remodeling which varies extensively among the elderly. This large spread of values creates errors in age estimation when regression equations are derived from samples which include very old individuals. This error increases with higher order regression models and proved to be particularly problematic with Kerley's revised regression equation for fragments. The form of the regression equations for Kerley's and Thompson's methods are not consistent with the theoretical changes expected between bone parameters and age. This also leads to error in age estimation.
Bibliography: p. 114-125.