A prospective study of accuracy of ultrasound in measuring and counting gallstones was undertaken in 30 patients undergoing cholecystectomy. Stones were correctly counted (up to five) in 27 of 30 patients. Ninety-three of 106 stones (88%) examined were measured accurately (with a 2 mm error margin). The size of the smaller stones tended to be overestimated, whereas the size of the larger stones tended to be underestimated. When nonoperative treatment of gallstones is considered, ultrasound can be used as the first examination to identify patients who will not be eligible for one or another protocol on the basis of size or number of stones.