Twenty-one children under 15-years' old with chronic constipation were treated in a crossover trial lasting 3 weeks. In the first week they received either lactulose or senna, in the next week no treatment, and in the third week the alternative treatment. Patient diaries were kept by parents on the number and character of stools passed, and of side-effects reported, during the 3 weeks. There was a significant difference, in favour of lactulose, in the number of days on which normal stools were passed during the treatment weeks. The number and frequency of the side-effects reported in the senna treatment week were very much higher (p less than 0.001) than in the lactulose week. Lactulose is recommended as an effective and very well tolerated treatment for the constipated child.