Using a strain gauge, we measured anal sphincter function in 116 chronically constipated and 18 healthy children. Eighteen constipated children were re-evaluated two months later (receiving laxative), and 15 were again studied seven to 12 months later. The anal resting tone varied along the length of the anal canal and was highest at 1 to 1.5 cm from the anal verge. This region was used to study the resting motor activity of the internal anal sphincter, the amplitude of the rectosphincteric reflex after 30 and 60 ml rectal distension, and the rectosphincteric reflex threshold. The mean and resting tone was significantly lower in constipated than in control children (P less than 0.001), but normalized in patients who recovered. Resting motor activity of the internal anal sphincter and the amplitude of RSR were significantly lower in constipated patients than were the corresponding values in control children (P less than 0.001), and remained lower during and after treatment, even in patients who recovered. The length of the anal canal and the RSR threshold were comparable in control and constipated children. Thus, the basic problem in chronically constipated children appears to be an abnormal internal anal sphincter, which is weaker and less responsive to rectal distension than in nonconstipated children.