Learning from violent death

Abstract

An attempt was made to build a classification scheme for gunshot residues (GSR) samples originating from four types of ammunition, collected from shooters’ hands immediately after shooting. The secured material was examined with the use of SEM–EDX method in the automatic manner. The obtained results were expressed as frequencies of occurrence of particles assigned to various chemical classes. In order to establish the most discriminative of these features the Mann–Whitney test was performed. Cluster analysis was performed for grouping the analysed samples according to their origin, i.e. the type of ammunition. It has been found that samples of GSR originating from Browning 7.65 mm and Luger 9 mm ammunition can be fairly easy differentiated from the remaining samples, whereas samples of GSR originating from of Makarov 9 mm and these of Tokarev 7.62 mm could not be differentiated using frequencies of occurrence of particles in the selected chemical classes.