This thesis examined differences in characteristics of cut marks produced by katanas and machetes on three bone types (rib bones, innominate bones and long bones) and also assessed whether the identified diagnostic traits survived after being subjected to fire. As there were few published investigative, methodological, analytical or reporting standards for the identification of cut marks produced by hacking weapons, this study also developed such standards. Using both qualitative and quantitative analysis, the research included a pilot study, which identified eight diagnostic morphological traits produced by these weapons on all three bone types and specific striation patterns on kerf walls (walls of a cut mark). The pilot study also devised a standardised process for the examination, analysis, and identification of cut marks produced by hacking weapons.

Eight diagnostic traits of cut marks were identified, including:​ flaking, feathering, peeling, micro-peeling, micro-curvature, scoop defect, exit notch and chattering. Characteristics that were statistically significantly more likely to be present in machete cut marks, when compared with katana cut marks, included the presence of a scoop defect for the rib bone, micro-peeling for cuts made to innominate bones and chattering for flat and long bones (p < 0.01). Statistically significant traits of cut marks produced by a katana, when compared with a machete, include the presence of micro-curvature for the flat bone (p = 0.04) and the presence of exit notches on cuts made to innominate and long bones (p < 0.01). The statistically significant characteristics associated with weapon type from the bivariate associations were considered for the regression models developed for each bone, so that they could be used for prediction purposes for the blind study to identify the weapon type used.

Although not statistically significant, there were clinically significant (a noticeable, repeated and important trait present but not statistically significant) diagnostic traits that differentiated cuts produced by an experienced katana user (using a smooth slicing action) from an inexperienced katana user (using a hacking action), with micro-curvature, flaking and a scoop defect more likely to be produced by the latter.

The use of Scanning Electron Microscopy (SEM) revealed marked differences in the striations produced by katanas and machetes, and also in those produced by the two different katanas. This has been identified as being due to the sharpening materials and methods, rather than the weapon type alone. Striations produced by the traditionally polished (sharpened) katana exhibited a smooth, parallel, primary step-like pattern, with a second smaller striation pattern on top of the primary striation, located along the step-like edge. In contrast, those produced by the display katana (factory machine sharpened) exhibited rough and mainly-parallel striations with sharp edges. Striations produced by the factory sharpened machete left mainly-parallel striations that had a smooth rolling hill type of appearance on long bones.

In all burnt samples, the burning of any part of a cut mark resulted in the diagnostic features also becoming burnt and absent. The edge of the cut marks exhibited a roughened straight edge that could be distinguished from thermally induced fractures and no striations were present on any of the burnt bone due to the burnt or charred surface.

The bivariate associations and statistical models established during the pilot study, along with the pilot study methodology, were tested for accuracy and reliability in a blind test. The weapons identified to have produced the rib bone cuts in the blind study were all correctly identified based on the statistical model alone. Including the Scanning Electron Microscopy (SEM) analysis strengthened the support for the weapon identification of cuts made on the rib bones. As in the pilot study, all of the innominate bone samples were correctly identified using the statistical models alone. There were no striations present in the innominate bones, so the SEM traits were not used as additional support for identification of the weapon types in cuts made to this bone. Six of the nine long bone samples were correctly identified using the statistical models alone, and when the SEM analyses were also considered this improved the accuracy of the predictions, enabling eight of the nine samples to be correctly identified. The blind study demonstrated that the combination of statistical models and SEM analysis enables accurate and reliable identification of weapon types used to produce cut marks on different bone types.