Erythrocyte Na/K ATPase activity is decreased in Type I diabetic patients; for Type II diabetic patients, literature data are controversial. Therefore, we have compared this enzymatic activity in 81 patients with Type I diabetes mellitus, 87 with Type II diabetes mellitus and 75 control subjects. Mean erythrocyte Na/K ATPase activity was lower in the Type I diabetic patients (285 ± 8 nmol Pi · mg protein–1· h–1) than in the control subjects (395 ± 9 nmol Pi · mg protein–1· h–1) whereas that of the Type II diabetic patients did not differ from that of control subjects. Sex, age, body mass index, and HbA1 c levels did not influence erythrocyte Na/K ATPase activity. The 25 Type II diabetic patients treated with insulin, however, had lower Na/K ATPase activity than the 62 on oral treatment (264 ± 18 vs 364 ± 16 nmol Pi · mg protein–1· h–1, p < 0.001) but similar to that of Type I diabetic patients. Among the Type II diabetic patients, stepwise regression analysis showed that fasting C-peptide level was the only factor independently correlated with Na/K ATPase activity; it explained 23 % of its variance. In fact, in the insulin-treated patients, those with almost total endogenous insulin deficiency (C-peptide < 0.2 nmol · l–1) had the lower Na/K ATPase activity (181 ± 21 vs 334 ± 17 nmol Pi · mg protein–1· h–1, p < 0.0001). The biological effects of treatment with C-peptide have recently led to the suggestion that this peptide could have a physiological role through the same signalling pathway as insulin, involving G-protein and calcium phosphatase and thus restoring Na/K ATPase activity. The relationship we describe between endogenous C-peptide and this activity is a strong argument for this physiological role. [Diabetologia (1998) 41: 1080–1084]