While quantitative or immunological studies of sulfated glycoproteins have been carried out in the gastric juice, the actual origin of these molecules is far from being established; indeed sulfate is not detected histochemically in the normal gastric mucosa. Contamination by salivary glands and/or small glands in the pharynx, trachea and esophagus may be a cause of error. Some observations in patients throw light on the actual source of the SGP in gastric juice. In a patient treated for a melanoma, and receiving 3 injections of radiosulfate, the saliva contained a SGP of specific activity 685 000 dpm/mg. There was no histochemical or histoautoradiographical evidence of sulfate in fundic biopsies 6 and 24 h after the injection of the isotope. In 2 esophagectomized patients, there was no evidence of SGP in the juice obtained at gastrostomy, i.e. after interruption of the alimentary tract under the esophagogastric junction. At the same time SGP were present in the salivary secretions. The esophagogastric junction was examined in 2 specimens: a large amount of acidic glycoproteins was detected in the cardiac glands. Sulfate esters were detected mainly above the epithelial junction. Thus esophageal cardiac glands may be responsible for the presence of SGP in the gastric juice. Sulfate esters are also present in the submucosal glands of the thoracic esophagus. SGP in the gastric juice are not secreted by the normal gastric mucosa. Contamination by salivary SGP is more or less avoidable. The glands at the esophagogastric junction represent an unavoidable element of contamination.