Surgery for Hirschsprung's disease is associated with high rate of morbidity, in the form of either constipation or incontinence or a combination of the two. This study investigates the mechanisms responsible for incontinence and/or constipation following the pull-through operation for Hirschsprung's disease. There were 19 children (15 boys and 4 girls), who at the time of study; 16 had undergone Duhamel, 1 Rehbein, and 2 Soave operation. We classified patients according to their symptoms into 3 groups: Group A was incontinent of faeces; Group B was constipated and incontinent of faeces, and Group C was constipated only. The median age at referral was 6 years, and the median period after operation was 5 years. All patients were investigated by intestinal transit study, endoanal sonography and anorectal manometry. Group A had normal or rapid transit study, as opposed to Groups B and C, who had delayed-transit study. On endoanal sonography, all children had an intact internal and an external anal sphincter, below the level of pull-through operation. The anorectal manometry showed a significantly lower resting anal pressure in the incontinent Group A as compared to the constipated children with or without incontinence in Group B or C (38 mmHg versus 57 or 66 mmHg respectively). The rectal pressure was also significantly higher in children in Group A as compared to those in Group B or C (71 mmHg versus 42 or 36 mmHg). The ratio of rectal/anal pressure was higher in incontinent children in Group A, as compared to constipated children in Group B or C. Therefore, constipation can be caused by high anal resting pressure and a weak rectal peristalsis, while faecal incontinence can be secondary to poor compliance and elevated rectal pressure in the presence of normal or low anal sphincter resting pressure. Aperients are the mainstay of treatment of constipation, however, children with incontinence are more difficult to treat. We did not attempt to define the pattern of nerve plexus because of poor results of revision operation for residual hypoganglionic segment and intestinal neuronal dysplasia. Treatment of these children can become more rational, if furnished with detailed functional studies. We advocate investigation of the anorectal function at an early stage in symptomatic children after surgery for Hirschsprung's disease, and less invasive treatment should be considered before embarking on major surgery.