Cataracts, the world's leading cause of blindness, are an enormous public health problem in both developing and industrialized countries. Identifying the risk factors responsible for cataract formation is a difficult and complicated problem because a realistic causal model in cataract formation would not be a simple linear sufficient cause paradigm (e.g., one exposure-one cataract type). A more complex model depicting each risk factor as a component cause, or part of a sufficient cause such as the sufficient/component cause model proposed by Rothman (62), is a more realistic way to summarize how multiple risk factors act in cataract etiology. Moreover, even this model has shortcomings, especially in explaining cataract etiology. It ignores the obvious importance of time to cataract formation and the way different component causes may act on different etiologic branches of cataract formation, e.g., nuclear sclerosis, posterior subcapsular cataracts, and mixed. Despite the complexity in identifying cataract risk factors, attempting to do so provides new hope in dealing with the morbidity, mortality, and cost of this disease. The evidence is overwhelming that age, trauma, and intraocular inflammation are important cataract risks. However, these exposures either are inevitable or are not major contributors to the population attributable risk. On the basis of both coherence and predictive performance, undernutrition is an important risk factor that can be altered. However, work still needs to be done in two areas related to this risk. First, regarding individuals in developing nations, the question must be asked about which nutrients (or lack thereof) are the culprits. The epidemiologic evidence that antioxidants are the missing nutrients is far from overwhelming. For developed nations, the obvious question still to be answered is whether the results of the Linxian Cataract Studies (11) and the India-US Case-Control Study (12) can be generalized to industrial nations. To help answer this question, an intervention study sponsored by the National Institutes of Health is now underway (K. Kupfer, Director, National Eye Institute of the USA, Bethesda, Maryland, personal communication, 1993). Ultraviolet radiation, especially ultraviolet B radiation, is an important risk for cortical cataracts, and one study (27) has even demonstrated a dose-response relation. However, the public health implication of this finding is not clear. Isolating the risk of ultraviolet B radiation exposure as a cause of cortical cataracts (and, in general, not of other types) indicates that the risk is small on a public health scale. This is because cortical cataracts are well tolerated and frequently require no treatment at all. The evidence that links ultraviolet B radiation to other cataract types comes mainly from ecologic studies and needs to be verified by analytic studies that are specifically designed to study the association. The strength of the association, consistency of studies, coherence, and biologic plausibility all indicate that both systemic and topical steroids are significant risk factors for the formation of posterior subcapsular cataracts. Given that most people are not chronic steroid users, the population attributable risk is low; however, the relative risk of those unfortunate enough to require chronic steroid use is high. The evidence is accumulating that cataracts can be added to the list of illnesses that are at least partially attributed to smoking. Although consistency among studies has not been obtained, this is certainly a plausible cause, and dose-response relations have been demonstrated (53). At this point, nuclear sclerosis is the most important cataract type associated with smoking. More work needs to be done to assess the role of smoking on other cataract types and to assess the risk of those who stop smoking. Retrospective studies that examine diabetes as a risk for cataracts are almost inevitably marred by hte selection bia