Seventeen patients were studied 3-31 months (median 6.4 months) after mucosal proctectomy and ileal pouch-anal anastomosis for ulcerative colitis (n = 15) or adenomatous polyposis (n = 2). Seven had a triplicated pouch, and ten a duplicated pouch. Clinical bowel function was determined by detailed questionnaire, and correlations sought between clinical end-points such as frequency of defaecation, urgency of defaecation and continence, and the results of laboratory investigations, comprising anal manometry, capacity and compliance of the pouch, completeness of emptying, faecal bacteriology and mucosal inflammation. Frequency of defaecation was found to be inversely correlated with both the capacity (rs = -0.66, P less than 0.01) and the compliance (rs = -0.53, P less than 0.05) of the pouch. Patients who could postpone defaecation for greater than 30 min had higher anal squeeze pressures (P less than 0.05) than patients who had greater urgency of defaecation. Patients with perfect anal continence had higher resting anal pressure (P less than 0.05) and emptied the pouch more completely (P less than 0.01) than patients who experienced minor leakage. The faecal flora of the pouches showed a greater predominance of anaerobes (P less than 0.01) and increased numbers of bacteroides (P less than 0.01) compared with the faecal flora of ileostomies, but the changes in the flora did not correlate with any aspect of bowel function. The best clinical results (i.e. perfect continence, low frequency of defaecation and little urgency) were associated with high anal pressure and with large volume, high compliance and complete emptying of the pouch. The completeness of emptying was similar for both designs of pouch, but the capacity and compliance of triplicated pouches were greater than the capacity and compliance of duplicated J pouches (P less than 0.05), and this was associated with a better clinical result in the triplicated pouches.