Abstract Cataractogenesis can be induced by glucose in the rat lens in vitro and in vivo. In vitro, this is done by increasing the amount of glucose in the medium surrounding the isolated lens; within 48 hr considerable globular degeneration is seen subcapsularly, deeper in the equatorial region. In vivo, it is achieved by making the rat diabetic by injecting streptozotocin i.v., which selectively destroys the β-cells of the pancreas; the blood serum glucose level increases markedly, and thence the aqueous humour level and, in turn, the lens concentration. Globular degeneration occurs as in vitro, but not until 6 weeks is a degree of damage observed comparable to that seen in the lenses incubated in vitro for only 48 hr. Lenticular sorbitol and fructose are also markedly elevated as a result of the high glucose levels. If glutathione (GSH) is present in the medium (0·1 m m) or injected s.c. daily into the diabetic rats, there is no evidence of subcapsular globular degeneration of the cortical fiber cells, even though the lenticular levels of glucose, sorbitol and fructose are the same as when GSH was not given; this is true for either the in vitro or in vivo situation. although individual values in the two situations do differ somewhat from one another. When rats were given GSH beginning several weeks after the diabetic state had been induced, the damage subsequently observed was much less than if the rats had been diabetic without GSH for the same total length of time; it was also much less than damage which should have occurred by the time GSH treatment was instituted. It would thus appear that a certain amount of reversal of the globular degeneration is possible, although damage in the equatorial region (wedge-shaped) seems less amenable to rescue by glutathione. The data indicate that glutathione can prevent or diminish the severity of sugar cataractogenesis. and that there would appear to be more steps in sugar cataractogenesis than simply osmotic damage, although this may be the primary event.