A prospective study of patients with symptomatic cholelithiasis was undertaken to determine the effectiveness of identifying clinically significant choledocholithiasis with selective cholangiography. Between 1991 and 1995, 262 patients presented to the senior author (K.W.M.) with acute or chronic cholecystitis. Sixteen patients had a preoperative endoscopic retrograde cholangiopancreatography (ERCP) for an elevated alkaline phosphatase or total bilirubin greater than twice the normal value or an ultrasound finding suspecting choledocholithiasis. Ten of the ERCP patients had choledocholithiasis, with eight patients having successful clearance by ERCP. Ninety other patients had intraoperative cholangiography for abnormal serum liver biochemistries, a history of jaundice or pancreatitis, or a dilated common bile duct (CBD) (>6 mm) on ultrasound. Fourteen of the intraoperative cholangiography patients and the two remaining ERCP patients had choledocholithiasis requiring CBD exploration for clearance of their stones. There were no false-positive cholangiograms, and there were no bile duct injuries in this series. With 100 per cent follow-up of at least 2 years, only one patient required ERCP clearance of a retained CBD stone 13 months after cholecystectomy. The positive predictive value and the negative predictive value for the selective cholangiography criteria are 23 per cent and 99 per cent, respectively. In conclusion, clinically significant choledocholithiasis can be found effectively with selective cholangiography. Also, utilizing selective cholangiography reduces the number of routine cholangiograms by 60 per cent.