Introduction The pathogenesis of diabetic retinopathy remains unknown, partly due to our ignorance of the effect of long-standing diabetes on retinal tissue; for this reason, an experimental approach to this problem seems warranted. In the past1 the efforts to produce diabetic retinopathy in animals met with only limited success and often the results could not be reproduced. One reason for this failure to obtain retinopathy experimentally was because the animals were killed too soon after the diabetes was produced. Therefore, the purpose of this study was to keep the diabetic animals alive as long as possible in order to simulate the conditions existing in the human eye, since it is well known that the retinopathy develops in patients after diabetes has been present for many years. In vitro studies2-7 carried out in a number of tissues of animal origin have shown that diabetes affects certain steps in intermediary