The lack of a technique that allows mass isolation of intact, viable human islets is part of the reason that islet transplantation has not become available to the human diabetic. This report outlines the history of islet isolation and presents two new technical modifications that have been developed in the dog. Many of the current problems in islet isolation are presented, including the difficulty in obtaining enough human pancreatic tissue with minimal warm-ischemia time; inadequate distention of the pancreas to provide sufficient disruption for maximal enzymatic reaction to release intact islets; inefficient chopping methods; the use of collagenase of variable composition; different digestion methods for obtaining isolated islets; and inefficient methods for separating and purifying the islets from the ductal, acinar, and fibrous components. The first new modification involves distention of the dog pancreas through the venous system of the gland rather than the ductal system. This results in improved intralobular disruption, which improved the yield of isolated dog islets by permitting more efficient collagenase digestion. The second new modification eliminates the concept of isolating intact islets: the dog pancreas is digested by trypsin to a single-cell preparation that is partially purified by Ficoll gradients; further purification of the endocrine cells results from selective aggregation using rotational culture. This process produces pseudoislets that contain all the islet cell types and can be kept in culture for up to 4 wk, releasing their hormones in response to appropriate stimuli. These modifications may assist in the struggle to isolate the elusive human islet for safe and effective islet transplantation in the diabetic patient.