The validity of the use of H2 gas clearance to measure gastric mucosal blood flow (GMBF) was investigated in the intact stomach of anesthetized dogs. With a modified, more sensitive platinum electrode, we were able to reduce the H2 gas concentration to a nonflammable gas mixture. GMBF was repeatedly measured when the dogs inhaled gas containing 100, 10, and 3-5% H2. GMBF measured using 100% H2 demonstrated close correlation with inhaled gases containing 10% H2 (r = 0.84, slope = 0.89) and 3-5% H2 (r = 0.91, slope = 0.88). At a nonflammable concentration of 3% H2, it was safe to add 20% O2 into the gas mixture. The addition of O2 into the inhaled gas eliminated the transient hypoxia noted otherwise. It was also possible to reduce the gas flow rate from 5-10 to 1-3 l/min. Regression analysis of GMBF determined by inhalation of 5% H2 at these flow rates revealed a significant linear correlation (r = 0.95, slope = 1.13). In 10 dogs GMBF determined by using these low concentrations of H2 at the low flow rates also showed a good agreement (r = 0.93, slope = 0.65) with that measured by radioactive microspheres. These two methods also demonstrated comparable changes in GMBF induced by intravenous infusion of histamine and vasopressin. It was concluded that, with the technique as modified in our laboratory, H2 gas clearance could be a safe and accurate tool for quantitating gastric mucosal blood flow.