When insulin is incubated in a buffered medium with small pieces of pancreatic tissue from the normal rat, it is rapidly destroyed. The lytic substance released by the tissue into the medium destroys both insulin and glucagon, but has no effect upon growth hormone large protein molecules or upon insulin which is already bound by antibodies in guinea pig anti-insulin serum (GPAIS). When GPAIS is added to the medium, it traps added or secreted insulin before the lytic substance has time to destroy it. Thus, pancreatic insulin secretion in vitro is estimated from the fall in insulin-binding capacity of known amounts of GPAIS added to the medium at the beginning of incubation. The method appears to be specific, since in the presence of glucose no neutralization of GPAIS occurs using pancreatic tissue from the alloxan-diabetic rat or normal tissue incubated in the presence of mannoheptulose. Insulin secretion is directly related to the weight and size of pieces of incubated tissue, and occurs at a constant rate for the first 2 hr. of incubation. Glucose and mannose stimulate insulin secretion, unlike 6 other tested sugars. The stimulant effect of glucose is abolished by anoxia, 2,4-dinitrophenol and cyanide, and is competitively inhibited by 2-deoxyglucose. From the results obtained by this reproducible and simple method, it is concluded that glucose has a stimulant effect on insulin secretion but that the role of glucose in this process has still to be determined.