A Bayesian approach to ageing perinatal skeletal material from archaeological sites : implications for the evidence for infanticide in Roman-Britain.

Abstract

The skeletal remains of substantial numbers of perinatal human infants have been excavated from within a variety of archaeological contexts dating to the Romano-British period. It has been argued that the distribution of ages at death of these infants, which appears to exhibit a pronounced neonatal peak, provides evidence for infanticide. This study re-evaluates the osteological evidence for infanticide in Roman Britain by first identifying biases in traditional techniques for estimating the age of perinatal skeletons and then using a Bayesian procedure to reassess the ages at deaths of almost 400 infants from a number of Roman sites throughout England. We conclude that the apparent peak in neonatal mortality shown by earlier investigations is an artefact of regression-based age estimation. The distribution of ages at death in Romano-British infants is similar to a natural mortality profile.