We examined the effect of 60-mg (200,000-IU) supplements of vitamin A administered every 6 mo on the incidence of xerophthalmia among preschool children who were free of eye symptoms and signs of vitamin A deficiency. We also prospectively studied the relationship of dietary vitamin A intake with the same endpoint. After 18 mo of follow-up, 400 children developed xerophthalmia during 80,104 child-periods of follow-up. Vitamin A supplementation only modestly reduced the risk of xerophthalmia (relative risk 0.88, 95% confidence interval 0.72-1.07, P = 0.19). On the other hand, total dietary vitamin A intake was strongly associated with reduced risk of xerophthalmia; the multivariate relative risk when children in extreme quintiles were compared was 0.38 (95% confidence interval 0.19-0.74; P for trend over quintiles = 0.002). These results emphasize the need for further data on factors that modify the bioavailability of large-dose vitamin A supplements. Increased consumption of inexpensive vegetables and fruits is highly likely to reduce significantly the risks of vitamin A deficiency, including nutritional blindness in developing countries.