Hypomagnesemia and low erythrocyte magnesium content are both common findings in non-insulin-dependent diabetic subjects. Moreover, intracellular magnesium may play a crucial role in modulating B-cell response to glucose by interfering with potassium permeability. Eight elderly, moderately obese, non-insulin-dependent diabetic subjects were treated with either magnesium supplementation (3 g/day) to the diet or placebo. Both treatment schemes lasted 4-weeks and were separated by a 'wash-out' of 3 weeks. At the end of each treatment period, in glucose test (0.33 g/kg for 3 min) and an iv arginine (5 g) test were performed to determine the B-and A-cell responses. Dietary magnesium supplementation vs placebo produced a slight but significant decrease in basal plasma glucose (8.6 +/- 0.3 vs 8.0 +/- 0.1 mmol/l, p less than 0.05) and an increase in acute insulin response after iv glucose (3.7 +/- 2.3 vs - 14.7 +/- 0.9 pmol.l 1. (10 min)-1, p less than 0.01) and after iv arginine (151 +/- vs 81 +/- 15 pmol.l-1. (10 min)-1, p less than 0.01), respectively. Plasma glucagon levels were unaffected by chronic dietary magnesium supplementation as well under basal conditions as in response to arginine. Net increase in acute insulin response after iv glucose and after iv arginine was significantly correlated to the net increase in erythrocyte magnesium content after dietary magnesium supplementation. We conclude that magnesium administration may be a useful adjuvant to the classic hypoglycemic agents in the treatment of non-insulin-dependent diabetic subjects.