Food hypersensitivities contribute to disease exacerbation in a sub-group of children with atopic dermatitis (AD). It has been shown that only selected foods are capable of causing clinical reactions when ingested, whereas other foods, to which the patient is equally sensitive by skin-prick testing, may be tolerated. The purpose of this study was to examine the cutaneous late-phase response (LPR) to food antigens in food-allergic patients with AD and to determine if the skin reacted differently to ‘relevant foods’ (foods eliciting positive skin-prick tests and positive oral challenges) than to ‘non-relevant foods’ (foods eliciting positive skin tests but negative oral challenges). Using blister chambers adfixed to the skin, six children with AD were challenged epicutaneously with foods to which they had previously been shown to be sensitive. Histamine and PGD2 were measured hourly for 10–12 hr in parallel with quantitation of the cellular traffic. There appeared to be no difference in any of the measured parameters between relevant foods and non-relevant foods, and the patterns of the LPR cells and mediators were similar to those previously described with aero-allergens in patients with respiratory allergy. Histamine rose to 13.0±24 ng/ml (P < 0.005) during the first hours, declined to < 1 ng/ml by the fifth hour, and then rose a second time to 6.72 ± 3.4 ng/ml (P < 0.05) during the 12th hour. PGD2 rose to an average of 312 pg/ml (P < 0.05) during the first 3 hr followed by a decline to baseline. The cellular traffic was similar to that observed during the LPR in atopic adults without AD. Neutrophils peaked at 11.2 ± 6.8 × 104 cells but did not reach significance because of background traffic in the control chambers. Eosinophils were significantly increased (P < 0.05) and rose to 2.52 ± 1.7 × 104 cells. Mononuclear ceils (P < 0.05) and basophils (P < 0.38) were also increased but less than either neutrophils or eosinophils. These studies suggest that selectivity in gastrointestinal antigen absorption or differential antigen processing, transport and/or clearing may explain the differences in clinical reactivity to ingested food allergens.