SummarySera from 119 patients with possible food allergies were tested against a panel of thirteen food allergens by the RAST technique. The results were compared within vivotests. 79% of foods causing symptoms gave a positive RAST to the specific food. Symptoms were grouped according to their time of appearance after taking the food; ‘immediate’ up to 1 hr and ‘non‐immediate’ more than 1 hr afterwards. Almost all those with ‘immediate’ symptoms were already aware of the foods causing them and there was a 100% correlation of the RAST result with these. Only a few of those with ‘non‐immediate’ symptoms were previously aware that these foods were responsible, and 64% of these gave a positive RAST. The majority of patients with a positive RAST result had total IgE in excess of 300 u/ml, had specific IgE antibodies against one or more common inhalant allergens, were under the age of 30 years and had a combination of asthma and eczema.We found the RAST method a useful and safe guide upon which to base a clinical investigation of food allergy, especially for patients whose symptoms appeared more than 1 hr after the food and in whom the relationship between their symptoms and food was not apparent. The RAST technique was surprisingly successful in identifying the foods which caused these ‘no