It has been suggested that death seasonality could be estimated based on the dark/bright appearance of the outermost dental cementum deposit. The few existing studies show variable accuracy rates. The question whether estimating season of death based on dental cementum represents a realistic and achievable goal can arise. This work, based on a software-assisted procedure and direct human observation, aims to critically evaluate the applicability of dental cementum to estimate the season of death in forensic anthropology.

The sample consists of 20 canines from individuals belonging to the 21st century Identified Skeletal Collection, University of Coimbra, Portugal, and with documented biodemographic data. Fifty percent of the individuals died during the summer and 50% during the winter. Five non-decalcified 100μm cross-sections per tooth were prepared, using a standardized protocol. Images of the region of interest, with the acellular cementum, were captured, and imported into FIJI/ImageJ software. Performance of the season of death predictions was very low, with 50% accuracy, score of 0.0000 for Matthews Correlation Coefficient, and F1 score of 61.5% for Summer and 28.6% for Winter, were obtained for the observer assessment. For the software model, performance was equally poor, with 52.6% accuracy, score of 0.0272 for Matthews Correlation Coefficient, and F1 score of 30.8% for Summer and 64.0% for Winter. This work is of relevance to the forensic science community by demonstrating the prospects and limitations of dental cementum as a method for estimating seasonality of death.