The authors carried out a retrospective review of intraoperative biopsy of the head of pancreas in 47 patients. Vim-Silverman needle biopsy (32 patients), or wedge biopsy (9 patients), or both (5 patients) were used. The correct diagnosis in the 34 patients with carcinoma was made in 76% by paraffin section and in 65% by frozen section. Diagnositic accuracy was 77% with needle biopsy and 38% with wedge biopsy. Complications occurred in six patients and consisted of pancreatitis, fistula and hemorrhage. Two patients required reoperation because of hemorrhage. The case of one patient who died of a pancreaticoperitoneal leak at the biopsy site is described in detail. The overall complication rate was 15%. These results indicate that wedge biopsy carries significant risks and frequently does not detect carcinoma. Needle biopsy is relatively safe and is more likely to provide an accurate diagnosis of carcinoma.