Massive retroperitoneal necrosis may follow life-threatening acute pancreatitis. At delayed operation, the surgeon may not be able to delineate dead pancreas from dead adipose tissue. The question arises: has "gloved hand" debridement resulted in pancreatectomy? The histologists report only "necrotic debris, of uncertain origin." To obtain objective data, pancreatography was performed in 13 patients, 10 weeks to 23 months after onset of massive pancreatic necrosis. Each patient had required delayed laparotomy for debridement and external drainage at some earlier stage of their illness. Pancreatography was correlated with the clinical assessment of diabetes and steatorrhea. Except in specific cases involving internal fistulae, pancreatography has not been previously reported in such patients. The results demonstrate that the main pancreatic duct usually maintained its normal length and configuration. Necrosis or stricture of the main duct, if it occurred, was more likely to be followed by diabetes. Steatorrhea was clinically detected in a single patient only. The necrotic tissue, up to several kilograms in wet weight, is largely dead adipose tissue. The pancreas, especially its head, is resistant to necrosis, much more resistant than is the retroperitoneal fat.