To test the hypothesis that diabetes is a form of accelerated aging, the following observations were made. 1) The incidence rate of diabetes mellitus had its peak at around 50 years of age with a gradually decreasing rate thereafter. This was clearly different from the manner of incidence of such disease as arteriosclerosis which increased with advancing age. 2) 100g of the oral glucose tolerance test performed on elderly subjects aged 60 to 89 years revealed high incidence of abnormal tolerance, 21% diabetic and 53% borderline types. 3) The insulin secretory capacity to glucose load of subjects was not different from that of young and middle-aged subjects from 20 to 49 years old. Therefore, decreased tolerance to glucose load could not ba ascribed to deficient insulin secretion. 4) No abnormality of glucagon response to glucose load was found in the old. 5) Serum beta-N-acetylhexosaminidase activity was not increased in elderly subjects, again contrasting with the increased activity found in diabetics. 6) Both glycolytic and gluconeogenic enzyme activities were decreased in the liver of aged rats. 7) No specific abnromality in insulin secretory response was observed in Werner's syndrome which might be considered to be a model for aging. All the above observations do not support the aforementioned hypothesis. Abnormality of glucose tolerance frequently observed in elderly subjects appears to be caused by other pathogenesis than diabetes mellitus.