Seven-day individual weighed dietary intakes and anthropometric measurements were determined in 123 children, 2 to 19 yr of age, from 26 poor families in Lima: each included one child who had been malnourished, six were adopting families. Heights and weights were converted to 'ages' based on Boston reference data and local data, then to 'quotients' as percentages of actual ages. Mineral and vitamin intakes were expressed as amounts per day and per 1000 kcal, calories and protein as percentages of FAO/WHO recommendations for age and height age and of modified recommendations based on size of Peruvian children. Regression analyses identified common sources of nutrients and greater dependence of intakes on body size than on age. Analysis of covariance for sex identified different nutrient-growth correlations. Polynomial regression analysis identified percentage protein from animal sources and percentage fat calories as having significant quadratic as well as linear correlations with achieved growth. In males, multiple regression analysis identified percentage protein from animal sources and β-carotene intakes as strongly associated with achieved height and weight, and percentage fat calories as strongly associated with weight quotient/height quotient ratios. In females the correlations were not as strong, possibly because a significant percentage had reached the menarche some time before the survey and were probably no longer growing. Nevertheless, calorie intake, as a percentage of the recommendation for height age, was prominent in the regressions for height quotient and percentage fat calories in that for weight quotient. The very strong association of animal protein intake with male height, despite seemingly generous total protein intakes, is difficult to reconcile with current recommendations. The possible role of vitamin A (as β-carotene) has potentially important implications for food policies.