Thirty sex offenders were randomly allocated: 10 to receive medroxyprogesterone therapy (M), 10, imaginal desensitization (ID) and 10 both (ID + M). Twenty-four responded for one year though 3 subsequently relapsed. There were no significant differences in response to the 3 treatments. Four patients who did not respond to the initial treatment and the 3 who relapsed responded to further treatment. Most treated with M maintained heterosexual intercourse at pretreatment frequency. Self-reported reduction in anomalous sexual urges in patients receiving M correlated with reduced testosterone levels one month following treatment, demonstrating that the response was specific and validating the assessment by patients' self-reports. Where cost-effectiveness or time constraints are factors influencing treatment of sex offenders, one of these therapies warrants consideration.