The Head Start Program measures children's heights and weights to screen for growth problems such as obesity, wasting, and short stature. At present, little public health use is made of these data. In this paper, the authors present serial cross-sectional nutrition surveillance data from Massachusetts Head Start Programs. Nonrandom samples of local Head Start Programs provided annual screening data from 1988 to 1991 on an average of 2,664 children per year. Height and weight measurements were compared with National Center for Health Statistics (NCHS) reference populations. On average, 87 percent of the children were 36 to 59 months of age and 51 percent were white. From 7.3 to 8.8 percent of children were below the 5th percentile of height for age each year, and from 1.2 to 3.3 percent were underweight, with weight below the 5th percentile for height (P < 0.05 compared with NCHS population). In each year overweight (weight for height above the 95th percentile) was most prevalent, ranging from 9.6 percent to 13.3 percent (P < 0.05 compared with NCHS) and demonstrating a statistically significant upward trend over the 4 years of study (chi-square = 9.21, P < 0.01). The prevalence of overweight and short stature varied by race and ethnicity. A statistically significant upward trend in overweight was seen among Hispanic children (chi-square = 5.99, P < 0.05). Also, children who were 48 months of age or older were more likely than younger children to be overweight (P < 0.05). The prevalence of short stature did not vary significantly by year, sex, or age. The authors conclude that children attending Head Start Programs in Massachusetts are at risk for short stature and are at increasing risk of obesity. These risks vary by race and ethnicity. Further research is needed to determine the generalizability of these findings to other regions and to evaluate social and behavioral correlates of poor nutrition outcomes among Head Start children.