The authors conducted a 2–year (1989–1991) community-based longitudinal study in a shantytown in Lima, Peru, to examine the effect of Cryptosporidium parvum infection on child growth during the year following the onset of infection. A cohort of children, aged 0–3 months at recruitment, was followed monthly for anthropometries, weekly for stool samples, and daily for diarrheal status. Data from 185 children in the cohort permitted a comparison of growth in C. parvum-infected and noninfected children. The analyses fitted smooth, flexible curves with a linear random-effects model to estimate growth differences between C. parvum-infected and noninfected children. Children infected with C. parvum experienced growth faltering, both in weight and in height, for several months after the onset of infection, followed by a period of catch-up growth. Younger children took longer to catch up in weight than did older children. Catch-up growth, however, did not occur in children infected between ages 0 and 5 months. These children did not catch up in height, and one year after infection they exhibited an average deficit of 0.95 cm (95% confidence interval (Cl) 0.38–1.53) relative to noninfected children of similar age. Stunted children who became infected also did not catch up in either weight or height, and one year after infection they exhibited a height deficit of 1.05 cm (95% Cl 0.46–1.66) relative to noninfected, stunted children of similar age. These results indicate that Cryptosporidium parvum has a lasting adverse effect on linear (height) growth, especially when acquired during infancy and when children are stunted before they become infected. Am J Epidemiol 1998; 148:497–506.