Over a period of one year, 251 dogs were presented to a UK-based dermatology referral clinic. Eighty-five of these were either diagnosed as having symptoms compatible with atopy (58 dogs), or suffered from chronic otitis or recurrent pyoderma. All 85 were placed on a carefully restricted diet for eight to nine weeks in an attempt to establish whether the symptoms were due to food sensitivity. In total, 19 were shown to have food sensitivity, representing 7.6 per cent of all dogs presented to the clinic, and one-third (32.7 per cent) of those dogs with signs compatible with a diagnosis of atopy. In five dogs with proven food sensitivity, otitis was the principal clinical sign and, in two others, recurrent pyoderma. In the population studied, labradors appeared to be predisposed to the condition. Improvement was monitored by asking owners to assess their dog's symptoms on an ordinal scale of pruritus. In those cases in which food sensitivity was confirmed, significant reduction in pruritus occurred. Most of these could be maintained long term on a commercial restricted-component diet. Particular effort was made to ensure owner compliance with the diet trials, using an explanation and model based upon a Venn diagram showing assumed links between atopy and several 'flare factors'. It was found that this approach significantly enhanced client understanding and cooperation. It is concluded that a careful approach, monitored by active clinical audit, will help to establish the true incidence of food sensitivity.