Fifty patients with recurrent post-cholecystectomy pain underwent hepatobiliary scintigraphy. Time-activity curves were generated from regions of interest over the liver, bile duct, duodenum, and bowel. Patients were divided into obstructed and unobstructed groups with endoscopic retrograde cholangiopancreatography criteria. Measurements from the liver, duodenum, and bowel curves contributed little to the analysis. The washout phase of the bile duct curve showed intermittent emptying in both obstructed and unobstructed groups. Less than one-third of peak activity remained in both groups at 90 minutes. Retention fell more rapidly in the later portion of the sequence in patients with obstruction. Quantification was essential, since differences were subtle and could not be appreciated visually. In the absence of hepatocellular disease, the most reliable criterion was the time at which maximal bile duct activity occurred. A cutoff level of 29 minutes or more was used in the diagnosis of obstruction. A sensitivity of 93% with an adequate specificity of 64% and an overall accuracy of 80% was achieved in the prediction of obstruction.