The first rib as a method of adult age-at-death estimation in a modern South African sample

An age-at-death estimation method using the first rib may be particularly advantageous as this rib is relatively easy to identify, not easily damaged postmortem, and associated with less mechanical stresses compared to other age indicators. Previously, mixed results have been achieved using the first rib to estimate age-at-death. This study aimed to develop and test an age-at-death estimation method using the first rib. An identified modern black South African sample of 260 skeletons were used to collect age-related data from the first rib. Multiple linear regression analysis equations were created from this data for male, female, and combined samples. When tested on a hold-out sample, equations generated mean inaccuracies of 7-13 years for point estimates. The 95% confidence intervals contained the true age in 11-33% of individuals depending on the equation used, but wider intervals generated using 95% prediction intervals contained true ages for 100% of individuals. Point estimate inaccuracies are comparable to other age-at-death estimation methods and may be useful if single indicator estimation is unavoidable in the case of missing or damaged bones. However, combined methods that use indicators from many areas of the skeleton are preferable and may reduce interval widths.

Automated summary

An age-at-death estimation method using the first rib is advantageous due to its easy identification, low susceptibility to postmortem damage, and less mechanical stresses compared to other indicators. This study developed and tested such a method using a modern black South African sample of 260 skeletons, creating multiple linear regression equations for male, female, and combined samples. When tested on a hold-out sample, the equations generated mean inaccuracies of 7-13 years for point estimates. While the 95% confidence intervals contained the true age in 11-33% of individuals depending on the equation used, wider intervals generated using 95% prediction intervals contained true ages for 100% of individuals. Point estimate inaccuracies are comparable to other methods, but combined methods using indicators from multiple areas of the skeleton are preferred to reduce interval widths.