Document Type



Height and weight data, together with the information on maternal age, height, pregravid weight, parity, and infant’s gestation, were analyzed in a population of approximately 3,000 low-income Negro boys and girls in Philadelphia. The relative weight of each infant at each age interval was determined by a regression of weight on observed height, using as constants the values obtained by analysis of the entire age-sex cohort. Using bivariate and multivariate regression techniques, the degree to which later deviations from expected weight could be attributed to earlier deviations was determined. The influence of maternal variables was similarly estimated. It was found that a knowledge of the maternal variables studied contributed very little to a prediction of the relative weight either at birth or at the subsequent intervals. Likewise, the relative weight at birth accounted for only 3% of the observed deviations at 4 months. In contrast, 25-30% of the observed deviations at 12 months could be attributed to the 4-month deviations. This relationship persisted to a significant degree through the 48 month measurement. The implications of this strong correlation between the 4-month relative weight and the relative weight at 12 and 48 months are discussed in relationship to the possibility of earlier intervention in the course of early childhood obesity.