Acute pancreatitis was experimentally produced in dogs to study the effect of the disease on glucose tolerance. The k value (glucose disappearance coefficient measured in percentage decrease of glucose/min) calculated from the high-dose intravenous glucose-tolerance test was used to evaluate the glucose tolerance of each dog. Thirty dogs were allotted to 3 groups of 10 dogs each as follows: group I--nonsurgical control dogs; group II--surgical control dogs; and group III--pancreatitis-affected dogs. To increase their susceptibility to diabetes, 50% partial pancreatectomies and ductal catheterizations were performed on group II and III dogs. Saline solution was infused into the ductal systems of group II dogs, and staphylococcal alphatoxin was infused into the ductal systems of group III dogs to produce pancreatitis. The results indicated that (1) high-dose intravenous glucose-tolerance test was an effective tool for determining decreased glucose tolerance in dogs; (2) glucose tolerance of group III dogs was markedly decreased compared with that of group I and II dogs; (3) staphylococcal alpha-toxin produced signs of moderately severe pancreatitis; and (4) 50% partial pancreatectomy and saline solution infusion produced clinical and clinicopathologic signs of mild pancreatitis. To determine if a simplified k value (calculated using 2 or 3 blood samples) could closely approximate the standard k value (calculated using 6 blood samples), simplified k values were derived from the 5- and 60-minute blood sample values. These values closely approximated the standard k values, indicating the simplified value may be used in the clinical situation. The standard k value, however, is preferred for investigative work.