In order to test the hypothesis that rotavirus infection leads to IgE hypersensitivity to cows' milk, evidence of past rotavirus infection, as reflected by the presence of antibodies, was sought in two groups of children with cows' milk allergy. The groups were defined according to the time to react to milk challenge. In group I there were fifteen children with an adverse reaction within 40 min of milk challenge. The twentyone children in group II had gastrointestinal reactions between 1 hr and 24 hr after milk challenge. The results showed a higher ineidence of antibodies to rotavirus in group II than group I (/'<0 05). Although group I patients had significantly higher levels of total and milk-specific IgE antibodies to cows' milk than those in group II, both patient groups had lower levels of serum IgG, IgA and IgM than a control population. No association between the presence of rotavirus antibodies and IgE hypersensitivity to cows' milk was found. However, an association between non-IgE milk enteropathy and previous rotavirus antibodies was demonstrated.