Summary. This review seeks to supply the arguments which support or deny the relationship between the quality of control of blood glucose levels and the course of diabetic microangiopathy. The ideal study is impossible to do in man but most prospective studies suggest that the better the control the slower the rate of progression and severity of lesions. "Scientific" but indirect arguments from biochemical, enzymatic and functional studies have shown that insulin and/or blood glucose control reverse some early diabetic changes that are probably related to the late events. Recent studies suggest that observations on animals that have shown the beneficial effect of treatment are relevant to the problem of diabetic microvascular lesions in man. Heredity influences the development of diabetic microangiopathy in diabetics but retinal and/or glomerular lesions -which are definitely not pathognomonic for diabetes -do not appear in the absence of hyperglycaemia. Muscle capillary basement membrane thickening that seems not to be a specific abnormality was not observed by several investigators in recently diagnosed diabetics. Therefore most of the arguments that have been given to support the point of view that tight control is not justified may be rejected. The opinion of the author is that strict control of diabetes is worthwile in patients with long life expectancy and no psychological, social or cultural handicaps.