The scintigraphic detection of small nocturnal aspirations of radio-labelled gastric contents is difficult in the presence of high remaining activity in the abdomen, causing a non-uniform background activity. This problem was examined in phantom experiments and a technique for interpolative background correction was further developed. The accuracy of this technique was found to be influenced by the distance between the lung and the abdominal source of activity, and the minimum detectable 'aspirated' activity was determined as 0.1 MBq at a distance of 15 cm and 1 MBq at 5 cm. The interpolative technique for background correction was evaluated on healthy volunteers and laryngectomized patients, examined 10 h after intragastric instillation of 200 MBq of 99Tcm-pertechnetate. After background subtraction, their calculated pulmonary mean net count value was comparable to that registered before the radioactive tracer was administered. No localized accumulation of activity was found in any of these controls. The technique was then applied clinically to 55 patients with chronic respiratory disorders and symptoms of gastroesophageal reflux. Aspiration was detected in 11 patients (20%). Five aspirators had asthma, two a chronic cough of unknown origin, two recurrent pulmonary infections, and one chronic bronchitis and chronic laryngitis respectively. Aspiration was detected among patients with and without demonstrated pathological gastroesophageal reflux.