Bone scintigraphy has been studied in two groups of patients presenting with low back pain. In one group of 38 patients suffering "nonspecific" back pain, bone scintigraphy and laboratory findings were negative in 24. There were abnormal laboratory findings in all of the remaining 14 and 7 had positive bone scans indicative of clinically significant disease. Selection of patients for bone scintigraphy in this group should therefore be influenced by abnormal laboratory findings and elevation of the erythrocyte sedimentation rate in particular. By comparison, the bone scans were reviewed from another group of patients suffering previously known malignancy. Out of 138 patients, nearly 40% showed a positive bone scan due to subsequently proven metastasis. Bone scintigraphy was positive in a further 14% as a result of osteoporotic rib fracture and vertebral body collapse. In half of these, it was not possible to exclude malignancy by scintigraphy. The present findings indicate that bone scintigraphy is not a useful procedure in patients with long-standing low back pain who have normal radiographs and normal laboratory findings.