From 1988 to 1991, 284 patients with prostatic cancer and painful bone metastases were treated with either radiotherapy or strontium-89 (200 MBq). Patients were first stratified according to suitability for local or hemibody radiotherapy, then randomly allocated that form of treatment or strontium-89 (i.v. injection). After 4, 8 and 12 weeks pain sites were mapped, toxicity monitored, and all additional palliative treatments recorded. There was no significant difference in median survival (after > 80% had died); 33 weeks following strontium-89 and 28 weeks following radiotherapy (p = 0.1). All treatments provided effective pain relief; improvement was sustained to 3 months in 63.6% after hemibody radiotherapy compared with 66.1% after strontium-89, and in 61% after local radiotherapy compared with 65.9% in the comparable strontium-89 group. Fewer patients reported new pain sites after strontium-89 than after local or hemibody radiotherapy (p < 0.05). Radiotherapy to a new site was required by 12 patients in the local radiotherapy group compared with 2 after strontium-89 (p < 0.01), although there was no significant difference between hemibody radiotherapy (6 patients) and strontium-89 (9 patients) in this respect. Platelets and leukocytes fell by an average 30-40% after strontium-89 but sequelae were uncommon, and other symptoms rare.