The basic wear mechanisms operating when two identical steel surfaces are rubbed against each other were studied to determine material parameters essential for wear resistance. Three simple model alloys, with the same basic properties as tool steels, were developed, containing three different predetermined volumes of M 7C 3 carbide, having approximately the same hardnesses after hardening and annealing, and approximately the same composition of the matrix. Unlubricated sliding wear tests were performed in air, using a pin-on-ring type machine. Normal force, sliding speed and sliding distance were varied. Friction force and temperature were recorded during the test and changes in weight of specimens were measured. The worn surfaces were carefully examined by scanning electron microscopy, in an attempt to classify the different wear mechanisms. It was found that corrosive wear dominates at low sliding speeds (2 m/min). Material annealed to a lower hardness had a lower wear resistance, irrespective of carbide content. The wear is characterized as mild. At high sliding speed (100 m/min) and especially for high normal forces, the wear was dominantly by a severe adhesive mechanism. Tempering to a lower hardness gave better wear resistance, which indicates that the room temperature hardness is not significant when a high contact temperature is reached. The influence of the carbide content was complex. The results indicate that a carbide free material is the most wear resistant, because of the more extensive occurrence of corrosive wear. Abrasives such as carbides in the more carbide rich alloys may possibly tear up protective corrosive layers and expose the steel to adhesive wear.