Whereas surgery often concludes the saga of hypoglycemia, the pathologist has the final word. In children and occasionally in adults, that word usually is nesidioblastosis, a somewhat ill-defined and nebulous diagnosis (1-6). Exactly what is nesidioblastosis? Does it represent any sort of pathology at all? More than 30 years after the initial description of infantile hypoglycemia by McQuarric (7), the pathogenesis of hyperinsulinism remains unclear in most cases. Few resected glands from these hypoglycemic infants show focal lesions: they either consist of a true adenoma exhibiting a compact ribbonlike pattern reminiscent of that observed in islet cell tumors of adults or of focal adenomatous hyperplasia. Surgical resection limited to the lesion is usually followed by disappearance of the hypoglycemia, which indicates that the focal abnormality was the cause of the hyperinsulinism. In most cases, however, no focal lesion is detectable. Instead, a diffuse and disseminated proliferation of islet cells budding off from pancreatic ducts is usually observed and has been proposed repeatedly as the underlying pathological lesion (8-12, 14-17). The term nesidioblastosis, currently used to describe this condition, was originally coined by Laidlaw (18) because pancreatic endocrine tumors originate from nesidioblasts, the stem cells that differentiate from the duct. Subsequently, Yakovac (9) defined nesidioblastosis as a continuous differentiation of (3-cells from the ductular system of the exocrine pancreas. Several recent studies that used quantitative immunohistochemical techniques have demonstrated that endocrine cells in close contact with ducts (nesidioblastosis) can also be observed in normoglycemic infants and young children and thus are not pathognomonic for the disease (15-21). Unfortunately, nesidioblastosis is a morphologic parameter difficult to quantitate, particularly in the neonatal and infantile pancreas in which scattered endocrine cells or small clusters are numerous. Even when careful morphometric studies have been performed, it cannot be certain that these morphologic findings are more prominent in hypoglycemic than in normoglycemic infants. On the other hand, precise quantitative studies have established that the total pancreatic mass of endocrine cells is not increased in hyperinsulinemic infants (30). Furthermore, more numerous small endocrine clusters in hypoglycemic than normoglycemic infants might reflect an abnormality in the formation of true islets (nesidiodys-