Rolling/sliding contact fatigue tests were carried out on a two-disc machine. Transverse and longitudinal furrows and conical and spherical dents were made on each specimen surface to trigger surface or near surface crack initiation. The experiment results show that the size of the defect did not have a strong influence on crack initiation, but the height of the edge buildup was the dominant parameter on surface crack initiation. Variables, such as surface roughness, hardness and temperature, were investigated to determine their influence on crack initiation. In the cases presented, surface cracks were easily found for film thickness parameters less than 0.3. When film thickness was greater than 0.5, cracks on the parts of the surface which were far away from the defects were difficult to find. It was noted that for the higher hardness specimen, the crack initiation life was longer. The reason for a crack being initiated on the specimen near the leading edge of the defects was explained by the micro-EHL analysis. It was noted that surface cracks might not always propagate. The propagation rate could also be influenced by roughness, hardness and temperature.Presented at the 46th Annual Meeting in Montreal, Quebec, Canada April 29–May 2, 1991