For 16 years the double-blind, placebo-controlled food challenge (DBPCFC) has been used at the National Jewish Center for Immunology and Respiratory Medicine to determine whether adverse reactions to foods do occur in children. The objective of these studies was to explore these reproducible adverse reactions and to characterize them. Although skin testing was performed on all subjects, a history of an adverse reaction to food and to subsequent DBPCFC were the only criteria for entry into this study. Of 480 children studied, 185 (39%) have had positive DBPCFC results. In these 480 children, 245 (24%) of 1014 DBPCFCs showed positive results. Egg, peanut, and cow milk accounted for 73% of the positive DBPCFC reactions, but many foods produced reactions. Skin test results were positive in most children with a positive DBPCFC reaction, but the large number of patients with asymptomatic hypersensitivity limited the accuracy of a positive skin test result alone as a predictor of clinical symptoms during food ingestion. Evaluation of results in this large number of children for a prolonged period revealed reproducible patterns of symptoms, timing, and incriminated foods. Placebo reactions were rare. The procedure was safe. Twelve youngsters with a negative DBPCFC result subsequently had positive reactions to open challenges when large amounts of the challenge food were used. In each of these cases the reactions were limited to areas of direct contact with the food or could be explained by the larger amount of food ingested during the open challenge. Multiple food hypersensitivity has been a rare finding. The DBPCFC should be the "gold standard" for both research and clinical diagnostic evaluations until it is superseded by methods that have yet to be developed.