Controversy persists regarding the role of intraoperative biopsy in the treatment of resectable pancreatic and periampullary masses. The authors report their experience with intraoperative biopsy in 37 patients in whom preoperative diagnosis was unsuccessful. Tru-cut needle biopsy was positive in 86 per cent of patients with carcinoma of the pancreas, but was negative in all patients having periampullary neoplasms. Incisional biopsy was positive in 75 per cent of all patients, its accuracy being approximately equal in both groups of patients. Duodenotomy with open biopsy was performed in eight patients having periampullary lesions and was accurate in 88 per cent. Biopsy of lymph nodes not contraindicating resection was performed in 11 patients and was positive in only 18 per cent. Resections based on clinical findings suspicious of malignancy with either a negative biopsy or no biopsy attempted resulted in a 13 percent resection rate of patients having benign disease. These results indicate that while histologic determination of malignancy can be obtained in most resectable pancreatic and periampullary lesions, at least 12 per cent of such tumors remain undiagnosed by standard operative biopsy techniques. Pancreatoduodenectomy should proceed in these patients based on their clinical status and the surgeon's degree of experience with the procedure.