270 patients with painful bone metastases requiring palliative radiotherapy were randomised to receive single fractions of 4 Gy or 8 Gy in a randomised trial. Pain scores and analgesic usage were recorded before treatment and at 2, 4, 8 and 12 weeks. Pain was assessed by patients on a 4 point graded scale using pain charts administered by a central trials office. Response was defined as an improved rating compared to the pretreatment value. Compliance with the pain chart was 72% at 4 weeks. At 4 weeks, the actual response rates were 69% for 8 Gy and 44% for 4 Gy (p less than 0.001), but there was no difference in complete response (no pain) rates at 4 weeks or duration of response between the two arms. It is concluded that 8 Gy gives a higher probability of pain relief than 4 Gy, but that 4 Gy can be an effective alternative in situations of reduced tolerance.