Early differentiation of gallstone from nongallstone associated acute pancreatitis by imaging methods is often difficult. Timing of surgery in gallstone pancreatitis is controversial, but early surgery requires early demonstration of gallstones. This study assesses the value of easily available clinical and laboratory data in establishing gallstones as the etiology of pancreatitis. In 405 consecutive episodes of acute pancreatitis, data were collected prospectively on 14 clinical and laboratory variables. Gallstones caused 177 episodes and alcohol 135, 93 were due to other or unknown causes. Age, sex, and within 48 hours of admission, serum alkaline phosphatase, aminotransferases, amylase, and bilirubin were all significantly different (all p less than 0.001, chi square) in gallstone and alcohol groups. Multivariate analysis based on five of these variables enabled correct prediction of the presence or absence of gallstones in 50 of a further 56 episodes. This method may help in planning early interventional treatment of gallstone associated acute pancreatitis.