The presence of cow's milk specific antibodies of immunoglobulin E and G4 classes were studied in 47 children with a positive clinical history of cow's milk allergy. The children were challenged with cow's milk orally. The clinical diagnosis was verified by immediate reactions in 25 patients while 22 had late reactions or were provocation test negative in spite of the clinical history. There was no relation between levels of cow's milk specific IgG4 antibodies and provocation test results, i.e. neither with immediate or late reactions. Total IgE was elevated above +1 SD for age in 31 of 41 tested patients. Of these, 29 had immediate type reactions to cow's milk, wheat flour and/or egg white, while only two of 10 children with IgE of less than +1 SD had a demonstrable allergy to any of these foods. The sensitivity of the total IgE determinations for the diagnosis of food allergy was 94% and the specificity was 80%. Specific IgE antibodies to cow's milk were demonstrated in 11 of 14 children with immediate reactions and in three of 15 who were provocation test negative or had only late reactions. This means a sensitivity of 79% and a specificity of 80%. At least one of the four patients with specific IgE but negative provocation test results had earlier shown an immediate reaction when challenged with cow's milk, indicating that the specific IgE antibodies were not truly "false" positive reactions but a consequence of previous allergy. Our results confirm an association between elevated total serum IgE and food allergy and an association between positive RAST to cow's milk and positive provocations in young children. We did not find any evidence for specific IgG4 antibodies playing a role in these patients.