Forty-five overweight children aged 6 to 13 were randomly assigned to one of four conditions: a behavioural programme occurring on a rapid schedule, the same behavioural programme presented on a schedule of gradually decreasing frequency, a non-specific control procedure and a waiting list control group. Experimental procedures required subjects to attend eight sessions, accompanied by a parent. The behavioural approach was found to lead to significantly greater reductions in obese status as measured by absolute weight loss and percentage overweight for age, sex and height, in comparison to both the non-specific control procedure and the waiting list group during treatment. This difference was maintained at the 11 week follow-up. Comparison between the rapid and gradual scheduling of behavioural sessions revealed little difference in outcome in the long term, other than effects which reflected differences in duration since the onset of treatment. Skinfold measures were found to be less sensitive to change, with differences between groups being evident only in the longer term assessments.