The aims of this study were to develop a gastroscopic food sensitivity testing (GFST) technique for clinical use in dogs and to determine if the results of GFST were influenced by the feeding of a hypoallergenic diet immediately before the testing period ("unmasking"). The technical requirements for GFST were devised during a total of eight endoscopies performed in four healthy dogs. GFST was performed in anesthetized dogs in sternal recumbency. Food extracts were dripped onto the dependent aspect of the body of the stomach via plastic tubing passed through the endoscope. Changes were observed within 2 to 3 minutes of application, and included localized mucosal swelling and erythema, generalized mucosal erythema, and hyperperistalsis. The influence of "unmasking" was then examined in 6 atopic and 2 healthy dogs, which underwent GFST on three occasions, 4 weeks apart. Before the first and third testing periods, the dogs consumed a commercial dry dog food. For 5 days before the second testing period the dogs were fed a hypoallergenic elemental diet. Oral challenges were performed to identify which of the dogs had clinically overt immediate food sensitivity. Localized swelling was most frequently correlated with positive challenge PO. No positive reactions occurred in response to the negative control extract (lamb). The number of positive GFST results increased after feeding the hypoallergenic diet. In conclusion, these preliminary results indicate that GFST holds substantial promise for the diagnosis of immediate food sensitivities affecting the gastrointestinal tract. The sensitivity of the procedure appears to be enhanced by preceding testing with a hypoallergenic diet.