In a hospital-based case-control study of 1441 patients with age-related cataracts and 549 controls, we studied associations between types of cataract--nuclear, cortical, posterior subcapsular, and mixed--and a number of physiologic, behavioral, environmental, and biochemical variables. Using polychotomous logistic regression analysis, we found an increased risk of cataract with lower educational achievement (all types of cataract), decreased cloud cover at place of residence (all types), use of aspirin less than once a month (posterior subcapsular and mixed), diets low in selected nutrients (posterior subcapsular, nuclear, and mixed), higher blood pressure (nuclear and mixed), lower body mass index (nuclear and mixed), use of cheaper cooking fuels (cortical, nuclear, and mixed), and lower levels of an antioxidant index based on red blood cell levels of glutathione peroxidase and glucose-6-phosphate dehydrogenase and plasma levels of ascorbic acid and vitamin E (posterior subcapsular and mixed). All risks cited were significantly different from those for the other cataract types, a finding that emphasizes the need to investigate the epidemiology of specific types of cataract.