The reproducibility of simultaneous, long-term, ambulatory gastric pH recordings in the antrum and corpus was investigated in nine healthy subjects who underwent three separate, 27-h gastric double pH-metries. Intraindividual reproducibility for the entire 27-h recording period was good in the corpus (Kendall's concordance coefficient, W' = 0.6393, p less than 0.025) but not in the antrum (W' = 0.4806, NS). Analysis of predefined time periods showed that non-meal daytime pH was reproducible in the corpus (W' = 0.6531, p less than 0.025) but not in the antrum (W' = 0.3395, NS), whereas mealtime pH was reproducible in the antrum (W' = 0.7159, p less than 0.005) but not in the corpus (W' = 0.4954, NS); nocturnal pH was not reproducible in either the antrum or the corpus. These results reflect the functional separation of corpus and antrum and their differing responses to food. Thus, studies of gastric acidity over long periods should be conducted in the corpus, whereas studies of gastric acidity over shorter, meal-related periods should be conducted with a second electrode in the antrum.