STUDY OBJECTIVE--The aim was to describe the population based age and sex specific incidence of fractures at different sites in a large English health district. DESIGN--Recording of fractures was accomplished by a specially constructed outpatient index and by record linkage to hospital inpatient information, for the three years surrounding the 1981 census. SETTING--The fracture index was held by the Department of Community Health in Leicester using data from the fracture clinic at the central large district general hospital, supplemented by hospital inpatient data from Trent Region and the two adjoining regions. PARTICIPANTS--The denominator population was the Leicestershire Health Authority resident population. In the three years, 12,711 fractures amongst males and 10,565 amongst females were recorded. MEASUREMENTS AND MAIN RESULTS--The overall estimated annual incidence of fractures was 100 per 10,000 population for males and 81 per 10,000 population for females. Below the age of 55 years all fractures showed a higher incidence amongst males but amongst the over 55s, there was a consistent fall in the male:female incidence ratio with some sites showing a striking female preponderance. The results also show an apparent age specific temporal increase in incidence at certain fracture sites compared with earlier British data, but fracture incidence figures still suggest lower rates in this country than in North America and some Scandinavian countries. CONCLUSIONS--These findings provide population based incidence data on a major public health problem and are consistent with the major determinants of osteoporosis and increase in falls in postmenopausal women. The temporal and geographical variation in fracture incidence remain to be explained.