Six infants suspected of food allergy during breastfeeding were evaluated using prick tests, total IgE, RASTs and intestinal permeability measurements during fast and provocation with mother's milk. An elimination diet was undertaken in mothers, removing first cow's milk protein (CMP), then, when inefficient, all foods suspected on the clinical history or a positive prick test in the child, followed by oral challenges in mother's diet with the corresponding food. The sole CMP-free diet in mothers always proved insufficient. In four, an additional diet excluding two to three other foods cleared the symptoms. Oral provocations in mother's diet with those foods were positive in all. In two, mothers turned down a diet excluding more than four foods, symptoms cleared while feeding the child with an extensively hydrolysed formula, whereas challenges with mother's milk induced immediate reactions. Intestinal permeability was altered during provocation tests with mother's milk sampled before maternal diet. Food allergy during breastfeeding may be due to multiple foods and the inefficacy of the sole CMP elimination in mothers does not rule out food sensitization.