A mild form of diabetes in young people was recognized in the pre-insulin era but was forgotten, probably because of Joslin's dictum that all young people with diabetes should have insulin as a safeguard against complications. After the introduction of sulphonylureas in the 1950s it was found, most notably by Fajans and Conn at the University of Michigan, that tolbutamide could improve or normalize carbohydrate tolerance in some young non-obese mildly diabetic patients. These experiments were not primarily of genetic interest because diabetes was regarded as homogeneous with young and old patients forming part of the same continuum. The question was whether treatment could prevent young subjects with mild diabetes progressing to a total loss of insulin reserve. By 1973, Fajans had shown that the carbohydrate intolerance of 45 patients diagnosed under age 25 had not progressed after up to 16 years on sulphonylureas. Nearly all (43 out of 45) these subjects had a first degree relative with diabetes. In 1974, under the title 'Mild familial diabetes with dominant inheritance' Tattersall described three families in which diabetes, although diagnosed in adolescence, could be treated with sulphonylureas over 40 years later and was dominantly inherited. Collaboration between Fajans and Tattersall established that 'chemical' diabetes in Michigan was also predominantly inherited and distinct from classical 'juvenile-onset' diabetes. In Paris in 1973 Lestradet also described a non-insulin-dependent form of childhood diabetes and later established that it was dominantly inherited. In 1974, Tattersall and Fajans coined the acronym MODY which was defined as 'fasting hyperglycaemia diagnosed under age 25 which could be treated without insulin for more than two years'.