There is little unequivocal evidence that nutritional supplementation of undernourished children has a beneficial effect on their mental development. The effects of nutritional supplementation, with or without psychosocial stimulation, of growth-retarded (stunted) children aged 9-24 months were assessed in a study in Kingston, Jamaica. 129 children from poor neighbourhoods were randomly assigned to four groups--control, supplemented only, stimulated only, and supplemented plus stimulated. A group of matched non-stunted children (n = 32) was also included. The supplement comprised 1 kg milk-based formula per week for 2 years, and the stimulation weekly play sessions at home with a community health aide. The children's development (DQ) was assessed on the Griffiths mental development scales. Initially the stunted groups' DQs were lower than those of the non-stunted group, and those of the control group declined during the study, increasing their deficit. Stimulation and supplementation had significant independent beneficial effects on the children's development. Estimates of the supplementation effect ranged from 2.2 (95% confidence limits-1.4, 5.7) for the hand and eye subscale to 12.4 (5.4, 19.5) for the locomotor subscale and those for the stimulation effect from 6.4 (2.8, 10.0) for hand and eye to 10.3 (3.3, 17.3) for locomotor. The treatment effects were additive, and combined interventions were significantly more effective than either alone. These findings suggest that poor mental development in stunted children is at least partly attributable to undernutrition.