Empirical evidence is presented to show that the geometry of tool wear, and the presence of built-up-edge, is strongly controlled by the chip curl. Evidence is also given which shows that some free-machining additives affect only the cutting temperature while others may affect the basic wear mechanism. Two simple theories are presented for calculating the rate of tool wear based on diffusion of tool material into the workpiece. While not conclusive, the results show remarkable agreement between predicted and measured wear based on a model wherein tool atoms jump into excess vacancies created in the chip by the plastic deformation of friction.