Open access colonoscopy for patients with suspected colonic disease is often not practical and some form of patient selection may be necessary. One year's colonoscopic data from our unit were analysed to determine the major indications for the procedure and the diagnostic yield, and to evaluate the suitability of colonoscopy for each indication. The seven major indications were rectal bleeding, iron deficiency anaemia, cancer follow-up, polyp follow-up, abdominal pain, abnormal bowel habit and 'other'. Four hundred and forty-eight procedures were included in the analysis, with rectal bleeding, polyp follow-up and iron deficiency anaemia producing the highest diagnostic yields of 69.1%, 53.3% and 47.7% respectively. Lower yields were obtained for cancer follow-up (21%), abdominal pain (38.2%) and abnormal bowel habit (46.8%). The indication, 'other', produced a combined yield of 66.7%; the majority of patients in this group were known to have colitis. On the basis of these findings we propose that where facilities and expertise do not allow for routine colonoscopy, some form of patient selection should be employed and we believe this selection should take place according to the diagnostic yield for each indication.