The role of mucosal biopsy in the monitoring of pediatric intestinal allografts is analyzed. We performed a retrospective review of all biopsy, resection, and autopsy material from 22 bowel allografts in 21 patients, followed from 6 months to 3 1/4 years and treated on an immuno-suppressive regimen based on FK 506 (Tacrolimus). There were 579 biopsies, of which 35 were stomal, with two to three fragments taken at each biopsy. There were three explanted bowels and three autopsies. Stomal biopsies proved to be inappropriate for monitoring. Biopsies with three to five pieces of tissue per site, under endoscopic direction, provided the most information. Early cellular infiltrate with lymphoid activation in the absence of epithelial apoptotic figures was not considered sufficient to diagnose rejection although preceded it in most instances. At least two apoptotic figures in a gland or several single apoptotic cells in the presence of a lymphoid infiltrate with activated lymphoid follicles and prominent endothelium correlate best with clinical rejection and response to antirejection measures. Epstein-Barr viral disease is common in this population, and early, late, and noncontiguous bowel involvement can be subtle and difficult to distinguish from rejection, though without the apoptosis. Epstein-Barr virus in situ probes are essential to make the differential diagnosis and the two conditions may co-exist. Mucosal biopsy monitoring appears to be of clinical utility and is part of a program that involves clinical, endoscopic, microbiological, and morphologic assessment.