The number of functionally intact B-cells in the islet organ is of decisive importance for the development, course and outcome of diabetes mellitus. Generally speaking, the total B-cell mass reflects the balance between the renewal and loss of these cells. While factors resulting in damage and degeneration of the B-cells have attracted much attention there has been relatively little interest in the kinetics and regulation of B-cell proliferation. This paper briefly reviews the literature in the latter field with special attention to B-cell renewal in various kinds of diabetes. In early fetal life both the endocrine and exocrine pancreatic cells probably arise from a common "protodifferentiated" cell type. The islets subsequently grow according to a regular pattern characterized by a symmetrical distribution of the islet volume in relation to islet diameter. The growth of the islets reflects the replication of both B-, A1- and A2-cells in individual islets. Whether new islet cells are derived also from duct epithelium or by transformation of differentiated acinar cells remains a matter of controversy. The ability of the differentiated B-cell to replicate in adult life has been confirmed in many studies. Factors believed to increase the rate of replication in vivo include a high caloric intake, hypoglycemic sulfonylureas, various hormones and hyperglycemia. Recent studies in vitro have so far confirmed the mitogenic action of a high extracellular glucose concentration. The renewal of B-cells in diabetes has been studied in several animal models. Both in experimental diabetes, such as that induced by alloxan, and in hereditary diabetes, exemplified by the diabetic mutant mouse (gene symbol db), there is initial stimulation of the mitotic rate during development of hyperglycemia followed by a notable decrease in the B-cell renewal. In the diabetic mouse the decreased proliferation of B-cells is paralleled by the appearance of fulminant diabetes. These and other data suggest a limited capacity for B-cell proliferation, which may differ between species. According to this hypothesis, there is an increased risk of developing diabetes once the predetermined potential for B-cell division has been exhausted.