Skeletal muscles in patients with non-insulin-dependent diabetes mellitus (NIDDM) are resistant to insulin; i.e., the effect of insulin on glucose disposal is reduced compared with the effect in control subjects. This defect has been found to be localized to the nonoxidative pathway of glucose disposal; hence, the deposition of glucose, as glycogen, is abnormally low. This defect may be inherited, because it is present in first-degree relatives to NIDDM patients two to three decades before they develop frank diabetes mellitus. The cellular defects responsible for the abnormal insulin action in NIDDM patients is reviewed in this article. The paper focuses mainly on convalent insulin signaling. Insulin is postulated to stimulate glucose storage by initiating a cascade of phosphorylation and dephosphorylation events, which results in dephosphorylation and hence activation of the enzyme glycogen synthase. Glycogen synthase is the key enzyme in regulation of glycogen synthesis in the skeletal muscles of humans. This enzyme is sensitive to insulin, but in NIDDM patients it has been shown to be completely resistant to insulin stimulation when measured at euglycemia. The enzyme seems to be locked in the glucose-6-phosphate (G-6-P)-dependent inactive D-form. This hypothesis is favored by the finding of reduced activity of the glycogen synthase phosphatase and increased activity of the respective kinase CAMP-dependent protein kinase. A reduced glycogen synthase activity has also been found in normoglycemic first-degree relatives of NIDDM patients, indicating that this abnormality precedes development of hyperglycemia in subjects prone to develop NIDDM. Therefore, this defect may be of primary genetic origin. However, it does not appear to be a defect in the enzyme itself, but rather a defect in the covalent activation of the enzyme system. Glycogen synthase is resistant to insulin but may be activated allosterically by G-6-P. This means that the defect in insulin activation can be compensated for by increased intracellular concentrations of G-6-P. In fact, we found that both hyperinsulinemia and hyperglycemia are able to increase the G-6-P level in skeletal muscles. Thus, insulin resistance in the nonoxidative pathway of glucose processing can be overcomed (compensated) by hyperinsulinemia and hyperglycemia. In conclusion, we hypothesize that insulin resistance in skeletal muscles may be a primary genetic defect preceding the diabetic state. The cellular abnormality responsible for that may be a reduced covalent insulin activation of the enzyme glycogen synthase. The compensation results in nearnormalization of glycogen synthesis, but the price diabetic subjects must pay to obtain this is hyperglycemia.