1 Lithium salts, used in the treatment of affective disorders, may have adverse effects on glucose tolerance in man, and suppress glucose-stimulated insulin secretion in rats. 2 To study the interaction of these effects with pre-existing diabetes mellitus, plasma glucose and insulin responses to lithium chloride were measured in male Wistar rats made diabetic with intraperitoneal streptozotocin, and in normal controls. 3 In both normal and diabetic anaesthetized rats, intravenous lithium (4 mEq kg-') caused a rise in plasma glucose. In absolute terms, the rise was greater in diabetic (5.2 mmol l ) than in normal rats (2.3 mmol I`).4 Plasma insulin concentrations were reduced by lithium in normal rats, but the low insulin concentrations measured in the diabetic rats were not significantly changed. 5 After intravenous glucose (0.5 g kg-'), lithium-treated diabetic rats showed a second rise in plasma glucose at 60-90 min without any insulin response, while normal rats showed typically reduced insulin responses and initial glucose disappearance rates. 6 Intravenous glucose reduced plasma glucagon concentrations to a greater extent in normal than in diabetic rats, but lithium induced an equal rise in plasma glucagon in both groups, with a time-course similar to that of the hyperglycaemic effect. 7 The hyperglycaemic action of lithium is greater in the hypoinsulinaemic diabetic rats and appears to involve a stimulation of glucagon secretion in both normal and diabetic animals.