Fructosamine and various measures of blood glucose were compared to glycosylated hemoglobin as indices of glycaemic control in 148 patients with insulin treated diabetes. Fructosamine correlated fairly well with glycosylated hemoglobin (r = 0.67), but around 40 per cent of the patients with glycosylated hemoglobin below upper reference limit had a fructosamine value over upper reference limit and vice versa. The possibility to predict the level of glycosylated hemoglobin from fasting blood glucose, postprandial blood glucose, and self-measured blood glucose was poor. The difference in self-measured blood glucose from patients with high versus low levels of glycosylated hemoglobin was very modest (0.5-2.0 mmol/l). It is concluded that it is reasonable to measure both glycosylated hemoglobin and fructosamine to evaluate glycaemic control in insulin treated diabetic patients. Fasting blood glucose, postprandial blood glucose, and self-measured blood glucose only seem to reflect glycaemic control to a minor degree.