Eighteen infants clinically suspected to be intolerant of cow's milk were placed on a milk-free formula and six to eight weeks later were orally challenged with cow's milk. Following challenge three groups were recognised. Group A: Four infants tolerated oral feeds ofcow's milk and lacked mucosal abnormality or clinical symptoms. Group B: Seven infants had mucosal deterioration but lacked clinical symptoms and tolerated cow's milk. Group C: Seven infants had mucosal abnormality, developed clinical symptoms and were intolerant of cow's milk. The intestinal transudation of IgA was increased in Group A and unchanged in Group Band C : the IgM levels in the duodenal juice was increased in Group A and B but unchanged in Group C : the IgG levels in the juice were increased in all Groups following challenge. It appears that increased transmission of IgA and IgM or IgM alone in the duodenal juice is associated with lack of development of clinical symptoms. Symptoms are present in infants in whom the IgA and IgM levels in duodenal juice remained unchanged after challenge. It is suggested that patients responding to cow's millt challenge with intestinal production of IgA and IgM (or IgM alone) are able to counter balance the deleterious mechanisms leading to clinical cow's milk intolerance whereas those who, for some unknown reason, do not mount a secretory immune response become ill.