It is commonly assumed that there is minimal variation between the hips in an individual, but is densitometry of one femur representative of the other? We performed bone mineral density (BMD) measurements of both hips using a Hologic QDR 1000 densitometer. There were 110 patients, all of whom were right handed, and three main groups of subjects: (1) normal volunteers (n = 36); (2) subjects with known hip pathology (n = 36); (3) subjects with medical conditions not affecting the hip (n = 38). The mean age of the subjects was 46 (21-87) years and a standard analysis protocol was followed in all patients. The coefficient of variation (COV) for femurs was 0.9-3%, depending upon the region studied and the BMD. The left femur had a greater BMD 48% of the time and there were variable differences between femurs in each group studied. While the greatest differences were found in people with unilateral hip pathology, all groups had mean differences greater than the COV. It may be acceptable to study only one hip, but the large variation between femurs in individuals should be borne in mind when interpreting data.