Twenty-five newly presenting, untreated, white, non-insulin-dependent diabetic (NIDDM) subjects were studied within 72 hours of diagnosis. They were allocated to three groups according to their body mass index [BMI] (lean BMI < 25.0, n = 9; overweight BMI 25.0 to 30.0, n = 6; obese BMI > .30.0 kg/m², n = 10). All three groups exhibited equivalent hyperglycemia. Eleven normal control subjects were also studied. The degree of activation of skeletal muscle glycogen synthase (GS) was used as an intracellular marker of insulin action, before and during a 240-minute insulin infusion (100 mU/kg/h). Fractional GS activity did not increase in the lean (change, −0.9 ± 3.3%), the overweight (−1.9 ± 2.7%), or the obese (+2.2 ± 1.6%) NIDDM subjects during the insulin infusion and was markedly decreased compared with the control subjects (change, +14.6 ± 2.4%, all P < .001). Glucose requirement was also significantly decreased in all three NIDDM groups (103 ± 23 v 81 ± 14 v 53 ± 14 mg/m²/min, respectively) compared with the control subjects (319 ± 18 mg/m²/min, all P < .001). There was a significant negative correlation with BMI (r = −.51, P < .01), but the difference in glucose requirement between the lean and obese NIDDM groups was not significant. Muscle GS activity at the end of the euglycemic clamp correlated with glucose requirement (r = .53, P < .001), and a similar correlation was observed between the insulin-induced change in muscle GS activity from basal and glucose requirement (r = .47, P < .005). Fasting plasma nonesterified fatty acid (NEFA) concentrations were elevated in the overweight and the obese NIDDM subjects (732 ± 78 and 635 ± 43 μmol/L, respectively, both P < .02) compared with the control subjects (426 ± 68 μmol/L). During the insulin infusion, plasma NEFA, glycerol, and 3-hydroxybutyrate concentrations all failed to suppress normally in the three NIDDM groups. Fasting serum insulin increased with BMI (4.6 ± 1.3 v 8.5 ± 1.4 v 11.1 ± 1.8 mU/L, respectively). The difference between lean and obese NIDDM was significant (P < .02). Insulin secretion at euglycemia was assessed following an intravenous (IV) bolus of glucose (0.5 g/kg). First-phase responses were abnormal in all three NIDDM groups (1.3 ± 1.3 v 1.7 ± 1.2 v 0.5 ± 1.3 mU/L) compared with the control subjects (50.1 ± 7.5 mU/L, all P < .001). The second-phase response in the lean NIDDM (−126 ± 34) was significantly lower than both the overweight (250 ± 18, P < .05) and the obese (228 ± 120 mU/L min, P < .02) NIDDM subjects, and also the control subjects (628 ± 182 mU/L min, P < .005). Thus, these studies demonstrate the failure of insulin to activate skeletal muscle GS in white NIDDM subjects and failure of first-phase insulin secretion during euglycemia. Both abnormalities were independent of the degree of obesity.