A parallel rise in hip fracture incidence and the rate of prescription of drugs affecting postural stability have been investigated by a case-control study. Other risk factors were also considered. One hundred and seventy-three cases and 134 hospital emergency surgical controls were interviewed and briefly examined, and drug data were corroborated by letter to general practitioners. The analysis revealed no significant difference between consumption of benzodiazepines in cases and controls, major tranquillizers, diuretics or other anti-hypertensives. However, they differed significantly in their average body weight and incidence of stroke. It is suggested that hip fractures on the whole occur in a particularly frail group of elderly people and that the increased fracture incidence over the last 20 years results from the survival of a more frail or less active group than hitherto.