The value and errors of one method of operative cholangiography were assessed prospectively with the aim of reducing further the incidence of unnecessary explorations of the common bile duct and the occurrence of residual stones. Operative cholangiograms were obtained using a cystic duct cannula in 174 cases, the technical failure rate being 2%. All cases were followed up using intravenous cholangiography where necessary. Of 129 judged normal by the operator, 2 were found to have common bile duct stones-false negative rate 1%. Of 45 judged abnormal by the operator, 10 had negative explorations- false positive rate 6%. Anomalies were discovered in 12% of patients. It is concluded that meticulous technique and accurate interpretation are essential for the best results. The high incidence of anatomical variations present is of interest and this alone emphasises the necessity of carrying out the investigation and the need of experienced interpretation.