Kinetic studies on the aldose reductase protein (AR2) have shown that it does not behave as a classical enzyme in relation to ring aldose sugars. These results have been confirmed by X-ray crystallography studies, which have pinpointed binding sites for pharmacological "aklose reductase inhibitors" (ARIs). As with non-enzymic glycation reactions, there is probably a free-radical element involved derived from monosaccharide autoxidation. In the case of AR2, there is free radical oxidation of NADPH by autoxidising monosaccharides, enhanced in the presence of the NADPH-binding protein. Whatever the behaviour of AR2, many studies have showed that sorbitol production is not an initiating aetiological factor in the development of diabetic complications in humans. Vitamin E (alpha-tocopherol), other antioxidants and high fat diets can delay or prevent cataract in diabetic animals even though sorbitol and fructose levels are not modified; vitamin C acts as an AR1 in humans. Protein post-translational modification by glyc-oxidation or other events is probably the key factor in the aetiology of diabetic complications. There is now no need to invoke AR2 in xylitol biosynthesis. Xylitol can be produced in the lens from glucose, via a pathway involving the enzymes myo-inositol-oxygen oxidoreductase, D-glucuronate reductase. L-gulonate NAD(+)-3-oxidoreductase and L-iditol-NAD(+)-5-oxidoreductase, all of which have recently been found in bovine and rat lens. This chapter investigates the molecular events underlying AR2 and its binding and kinetics. Induction of the protein by osmotic response elements is discussed, with detailed analysis of recent in vitro and in vivo experiments on numerous ARIs. These have a number of actions in the cell which are not specific, and which do not involve them binding to AR2. These include peroxy-radical scavenging and recently discovered effects of metal ion chelation. In controlled experiments, it has been found that incubation of rat lens homogenate with glucose and the copper chelator o-phenanthroline abolishes production of sorbitol. Taken together, these results suggest AR2 is a vestigial NADPH-binding protein, perhaps similar in function to a number of non-mammalian crystallins which have been recruited into the lens. There is mounting evidence for the binding of reactive aldehyde moieties to the protein, and the involvement of AR2 either as a 'housekeeping' protein, or in a free-radial-mediated 'catalytic' role. Interfering with the NADPH binding and flux levels--possibly involving free radicals and metal ions--has a deleterious effect. We have yet to determine whether aldose reductase is the black sheep of the aldehyde reductase family, or whether it is a skeleton in the cupboard, waiting to be clothed in the flesh of new revelations in the interactions between proteins, metal ions and redox metabolites.