Studies on diabetes mellitus in tropical zones indicate that its traditional link with overnutrition depends not only on the economic level, but also on some ethnic, social and cultural factors. At present, we insist on the unexpected relationship between diabetes mellitus and undernutrition either in some major infantile forms (described in India and Nigeria) with calcareous pancreatitis, or some less severe forms observed in Africa. This tropical diabetes mellitus occurs in some patients with normal weight or inferior to normal; it is not very ketogenic, responding to glucagon stimulation, and seems more frequent in the chronic malnutrition areas. So, it is tempting to utilize diabetes mellitus as an indicator of nutritional disorder or of dietary toxic factors. However, we ought to consider it within a multifactor surroundings associating genetic determinism and the other factors of tropical aggressiveness.