Intestinal biopsy is an essential procedure in the diagnosis of cow's milk protein-sensitive enteropathy. The number and timing of biopsies, however, is still controversial. Twenty-two infants who were clinically suspected of having cow's milk protein intolerance were put on a milk-free diet, during which they all improved. In the first 14 infants (Group I), milk withdrawal lasted 6-8 weeks; in the following 8 infants (Group II), milk withdrawal lasted 3.5-4 months. At the end of the withdrawal period, each patient was submitted to an oral lactose tolerance test and then to a first jejunal biopsy, followed by a cow's milk challenge, and a second biopsy (1-2 days later). None of the infants showed evidence of lactose intolerance, but they all reacted adversely to milk reintroduction. In Group I, the distribution of mucosae from grade 0 (normal) to III was almost the same before and after challenge (p greater than 0.5); this is considered to be due to the severe mucosal damage still present in prechallenge biopsies. In Group II, however, a significant difference was found in the distribution of mucosae before and after challenge (p less than 0.05), and a deterioration could be observed in each case after challenge. Intraepithelial lymphocyte counts and mitotic index determinations before and after milk reintroduction were of no diagnostic value in either group. It is suggested that the diagnosis of cow's milk protein-sensitive enteropathy may rely on two biopsies, the first after a period of 3.5-4 months on milk-free diet and the second following milk challenge.