Although enteropathogenic strains of Esckerickia coli are a frequent cause of disease in human infants, it has proven difficult to establish experimental infections in animals with these organisms. The most successful results so far have been obtained by feeding cultures to mice treated with antibacterial drugs prior to the test in order to suppress their indigenous intestinal flora (1). Useful as it is, this method is obviously artificial and fails to reproduce the conditions under which the disease is commonly observed in man. The availability at The Rockefeller University of a colony of mice (NCS) maintained under so called "specific pathogen-free" conditions, and especially free of E. coli, prompted us to investigate whether these animals would prove more receptive to human strains of E. coli than mice raised under usual conditions.The experiments reported in the present paper were carried out with a culture of E. coli, serotype 026: K60 (026: B6). This type has been found to be associated with infantile diarrhea and is also occasionally recovered from animals (2); the strain used here was isolated from an infant. While the ultimate purpose of our study was to determine whether NCS mice can be used to differentiate virulent enteropathogenic strains of E. coli from ordinary strains, it seems worthwhile reporting at this time a finding of more general importance. It was found namely that the age of the animal host has a profound effect on the ability of the bacteria to colonize the intestine and on their persistence in this organ. Colonization uniformly took place when very young mice received E. coli per os, whereas adult animals usually failed to become infected under the same conditions. Indeed the population of E. coli in the gastrointestinal tract abruptly fell to very low levels around the time of weaning.