The ability of hockey helmets to limit localized pressure was examined by mounting selected samples of Canadian Standards Association (CSA) certified helmets onto a humanoid headform and subjecting them to single puck impacts to the left temporal area at three different velocities. The peak pressure on the headform was determined from Fuji pressure-sensitive film that was applied to the headform before impact. Peak headform acceleration and the severity index (SI) were also measured. For five of the six helmet models, the peak pressure was less than 15.0 MPa at 27.8 m/s, less than 16.0 MPa at 33.4 m/s, and less than 25.0 MPa at 38.9 m/s. In the remaining model, the largest peak pressure was 27.8 MPa at 38.9 m/s. None of the helmets tested was able to limit the localized pressure on the headform to below 3.1 MPa, the reported fracture tolerance of the temporoparietal area of the human skull. Peak headform accelerations were at or below 250 g for all helmets at 27.8 and 33.4 m/s, but exceeded 275 g in four models at 38.9 m/s. SI values were all less than 800, indicating a concussive head injury risk of less than 2 to 5%. The SI and peak g measures were not useful for determining the focal injury risk caused by puck impacts, suggesting that additional test procedures should be included in future certification standards.