The association between elevated levels of glycated haemoglobins and diabetes mellitus has been known for twenty years [92]. Since then the determination of glycated haemoglobins has become a valuable tool for the objective assessment of long-term glycaemia in diabetic patients. The marked clinical interest in reliable measurements of glycated haemoglobins has stimulated the development and perfection of the necessary methodology. Limitations of the techniques have led to investigation of the underlying causes. Some of them led to the recognition of processes that were not known to occur in vivo before, such as glycation at sites other than the amino terminus of the beta-chains, modification of haemoglobin by reactants other than glucose or the existence of labile haemoglobin adducts. With ideal methodology these features would have gone unnoticed. Furthermore, the determination of glycated haemoglobin in large populations of diabetic patients has lead to the discovery of new, clinically silent mutant haemoglobins. Today, the routine determination of glycated haemoglobins in diabetic patients probably represents the broadest screening for mutant haemoglobins. The experience with glycated haemoglobins shows that overcoming difficulties in their determination, and progress in biomedical research, are closely intertwined.