Forty-seven patients with type 1 (insulin-dependent) diabetes mellitus and 14 normal subjects had renal volume determined by an ultrasonic technique. Renal volume of 299 +/- 49 ml/1.73 m2 (mean +/- SD) in type 1 diabetic patients with normal urinary albumin excretion exceeded that in the normal subjects (245 +/- 53 ml/1.73 m2, p less than 0.05). Compared with diabetic patients with normal urinary albumin excretion, renal volume was significantly higher in patients with microalbuminuria (372 +/- 24 ml/1.73 m2, p less than 0.05) and patients with clinical nephropathy (352 +/- 48 ml/1.73 m2, p less than 0.05). In a multiple linear regression with HbA1c, urinary albumin excretion, age, diabetes duration and mean blood pressure as independent variables, variations in HbA1c could account for 33% of the variations in kidney volume (n = 47, r = 0.57, p less than 0.01). The other variables played no role. When only patients without clinical nephropathy were included, HbA1c and kidney volume remained significantly correlated (n = 34, r = 0.60, p less than 0.01). In those patients a strong correlation between kidney volume and function, i.e. the glomerular filtration rate was found (n = 34, r = 0.70, p less than 0.01); metabolic control and function were also correlated (n = 34, r = 0.62, p less than 0.01). The urinary albumin excretion accounted for only 6% of the variation of the kidney volume (NS). In nine patients with microalbuminuria the kidney volume could be reduced during 2 years of improved metabolic control by means of insulin infusion pumps.(ABSTRACT TRUNCATED AT 250 WORDS)