As an experimentalist interested in arteriosclerosis, I look upon diabetes mellitus as a natural experiment in atherogenesis. This narrow view may distress clinicians engaged in the management or prevention of diabetes, but they should agree that progress in the prevention of atherosclerosis will benefit diabetic human beings most of all. In 1947 I began a laboratory study of the effect of experimental diabetes on the development of vascular disease in animals. The premise that the presence of diabetes aggravates the vascular disease is so attractive that even after five years of fruitless work it is sometimes tempting to try again. It was not possible to demonstrate that this was a valid hypothesis. Although certain capillary lesions were demonstrated in diabetic rats, no evidences of major vessel disease were produced in either rats, dogs, or monkeys, whether fed cholesterol or not. A second hypothesis concerning atherogenesis has been studied and contested for almost half a century and yet has eluded the critical experiment which will properly evaluate its verity. In simple terms this hypothesis states that atherosclerosis, a variant of arteriosclerosis, is a consequence of abnormal lipid metabolism. The immediate problems posed are those of describing the kind and extent of abnormal lipid metabolism necessary for this consequence, a description of the contribution or weight of abnormal lipid metabolism in this effect, and finally a description of the mechanism by which the lipid abnormality develops. There is now a great deal of acceptable information on these points, but for the present I shall consider some evidence that my associates and I have obtained in studies of diabetic human subjects. If from the relation of these observations, any suggestions or implications are made for the management of diabetes, they are fortuitous.