The long-term benefits of early childhood supplementation and the extent to which catch-up growth occurs following linear growth retardation remain controversial. Stunted children (height-for-age < -2 so of NCHS references, n = 122) recruited from a survey of poor neighborhoods in Kingston, Jamaica, participated in a 2-yr randomized, controlled trial of supplementation beginning at ages 9-24 mo. A group of 32 non-stunted children from the same neighborhoods was also followed. Four years after the intervention ended, when children were 7 to 8 y old, there were no effects of supplementation on any anthropometric measure. From the end of the trial until follow-up, the children who had been supplemented gained 1.2 cm less (P < 0.05) than the non-supplemented children, approximately the same amount as they had gained during the trial compared with the non-supplemented children. After adjustment for regression to the mean, the height-for-age of stunted children (supplemented and non-supplemented combined) increased from enrollment to follow-up by 0.31 Z-score (95% CI 0.17, 0.46). The height-for-age of the non-stunted children also increased (0.96 Z-score; 95% CI 0.70, 1.22). Our results suggest that some catch-up growth is possible even when children remain in poor environments. Longterm benefits of supplementation to growth may not be achieved when intervention begins after age 12 mo in children who have already become undernourished.