A comparative, experimental study has been done in which 3 methods were utilized to determine the macroscopic antral corpus boundary in proximal gastric vagotomy: anatomical references (A.R.), pHmetry, and Congo red (C.R.). Acute experiments were done on 10 dogs in which, after stimulation of the gastric secretion with 6 pg/kg of pentagastrin intravenously, the macroscopic antral‐corpus boundary was determined at the level of the lesser curvature of the stomach using the 3 methods. At the same time a series of specimens of gastric mucosa was obtained along the whole of the lesser curvature, to determine the relation between these macroscopic results and the microscopic limit. The results indicate that the A.R. gave an average error of 0.20±0.47 cm in relation to the microscopic limit, which was not significant (p>0.30). The average error produced by pHmetry was 1.45±0.36 cm, and that obtained with C.R. was 2.25±0.72 cm, which were both statistically significant (p<0.001). If these results are contrasted 2 by 2, it is found that the error with A.R. is significantly less than with pHmetry and with C.R. (p< 0.001). Likewise, the error with pHmetry is significantly less than with C.R. (p<0.05). We conclude that the determination of the macroscopic antral‐corpus boundary in proximal gastric vagotomy should preferably be done using anatomical references, since this procedure is the most precise in relation to the microscopic line and is also the easiest to perform.