ABSTRACT Hunger during school may prevent children in

developing countries from benefiting from education. Although many countries have implemented school feeding programs, few programs have been rigorously evaluated. We conducted a randomized, controlled trial of giving breakfast to undernourished and adequately nourished children. The undernourished group comprised 407 children in grades 2–5 in 16 rural Jamaican schools (weights-for-age £ 21 SD of the National Center for Health Statistics references) and the adequately nourished group comprised 407 children matched for school and class (weights-for-age>21 SD). Both groups were stratified by class and school, then randomly assigned to breakfast or control groups. After the initial measurements, breakfast was provided every school day for 1 school year. Children in the control group were given one-quarter of an orange and the same amount of attention as children in the breakfast group. All children had their heights and weights measured and were given the Wide Range Achievement Test before and after the intervention. School attendance was taken from the schools ’ registers. Compared with the control group, height, weight, and attendance improved significantly in the breakfast group. Both groups made poor progress in Wide Range Achievement Test scores. Younger children in the breakfast group improved in arithmetic. There was no effect of nutritional group on the response to breakfast. In conclusion, the provision of a school breakfast produced small benefits in children’s nutritional status, school attendance, and achievement. Greater improvements may occur in more undernourished populations; however, the massive problem of poor achievement levels requires integrated programs including health and educational inputs as well as school meals. Am