It is generally held that microaneurysms constitute the characteristic lesion of diabetic retinopathy. First observed by MacKenzie and Nettleship in 1879,1the microaneurysms were rediscovered by Ballantyne and Loewenstein in 19432and have been described by numerous clinicians and pathologists since that time. Noteworthy has been the predominant localization of the diabetic microaneurysms in the central and paracentral regions, their random distribution, and their association with punctate hemorrhages and "exudates" in the deep retina. It has also been assumed that the microaneurysms and other evidence of angiopathy in diabetic retinas have a bearing on the glomerular changes in diabetic nephropathy3-5and that a clarification of the one might yield useful information for the other. As must be evident from the many reviews which have appeared in recent years,6-12the pathogenesis of retinal microaneurysms has been a subject of wide speculation based on some firm evidence but