Patients with and without gastroesophageal reflux cannot be separated on the basis of lower esophageal sphincter (LES) miximal pressure alone, suggesting that more than this single factor is involved in normal closure of the LES. The physics of the closure mechanism was investigated in vivo in one and in vitro in two models. In anesthetized dogs the gastric and esophageal pressures were measured during gastric infusion while the LES gave way to reflux. In vitro sphincters were simulated on flaccid rubber tubes and excised canine gastroesophageal specimens. In model I the sphincter was simulated by mounting the specimen in a chamber and applying external pressure. In model II a ligature with graded tension was applied around the specimen. Station pullback manometry and pressure variations during constant flow perfusion were studied for these sphincter models. An opening pressure and a closing pressure for the LES were defined. The wall tension of the stomach as a force contributing to sphincter opening was recognized, introducing the degree of gastric distension as an important factor in sphincter function. The sealing property of the mucosa was estimated. The length of the sphincter was suggested to contribute to sphincter competence by decreasing the importance of gastric wall tension in sphincter opening and by improving the mucosal seal.