During the 50-year period 1928 to 1977, 1,250 Rochester, Minnesota residents experienced 1,355 proximal femur (hip) fractures. The incidence of these fractures appeared to rise during the first 15 years of the study, but this was most likely due to underascertainment of cases in the early period since there was no evidence of a cohort effect. From 1943 onward, there was no significant increase in the incidence rates which remained relatively level for total hip fractures, initial hip fractures alone, and for initial hip fractures due to moderate trauma. Thus, we found no evidence to support the contention that hip fracture incidence has risen dramatically in recent years. Secular changes in incidence cannot account for the observation that hip fracture rates in Rochester are greater than those reported from earlier population-based studies.