Fiabilité comparée de deux méthodes d'estimation du nombre minimum d'individus (NMI) Ã partir d'un vrac d'ossements

In each case of discovery of commingled human remains, the assessment of the minimum number of individuals (MNI) found, is the major point of the forensic anthropologic analysis. The calculation of this MNI is generally done by counting the number of similar bones, but some authors proposed a “pairing MNI” method that should improve the rate of detected individuals. The criteria used to recognize the pairs seem to be subjective (size, shape and colour) and are susceptible to generate two kinds of errors: the creation of a pair that has no biological reality, or the non-recognition of a real pair. In order to test the reliability of this method, we used a bone sample which real number of individuals was known and we applied the “pairing MNI” method. The results clearly show that this method lacks some reliability.