To determine whether operative cholangiography could be performed on a selective basis, 272 consecutive patients undergoing cholecystectomy and routine operative cholangiography were studied prospectively. Appropriate clinical, biochemical and operative data were recorded and the surgeon was asked to estimate the likelihood of common bile duct stones. These factors were correlated with the subsequent cholangiographic and operative findings. From this information, four criteria were selected as indications for performing an operative cholangiogram: an abnormal common bile duct, a wide cystic duct, or elevated levels of alkaline phosphatase or bilirubin documented in the previous six months. Operative cholangiograms would have been indicated by the presence of one or more of these criteria in only 139 of the 272 patients in the study (51 per cent). Only one unsuspected stone would have been missed. It is concluded that a policy of selective cholangiography could have been safely adopted in this series of patients.