Antimony and glass pH electrodes show almost identical experimental errors in continuously measuring buffer solutions at constant temperature over 24 hr. These errors are lower than the nominal quantization error of the instruments and are not properly described by the 24-hr drift determination. The addition of food particles to the solutions can induce severe reading artifacts. The longer response time reported in vitro of antimony electrodes when moving from pH 1 to pH 7 (3.4 sec vs 0.8 sec with glass electrodes) is irrelevant during in vivo pH-metry studies, because we found that the greatest absolute difference between raw fast acquired (4-6 sec) consecutive pH readings of two commonly used devices was 0.7 pH units in circadian profiles obtained from 413 subjects with various clinical conditions. In our in vivo studies, gastric acidity was monitored continuously with two side-by-side minielectrodes, which were variously combined (antimony-glass, A-G; antimony-antimony A1-A2; glass-glass, G1-G2) and applied on groups of 27 subjects matched for clinical condition. The 24-hr pH means and the 24-hr [H+] means calculated from the acidity profiles obtained with the three electrode combinations, lie on the identity line in each group. Using the Bland-Altman technique for assessing measurement agreement, the differences between the 24-hr pH means and the 24-hr [H+] means obtained with the three combined systems are similar (P = .903 and P = 0.824, respectively) and their 95% confidence limits are comprised within the range (+/-) of the reading error of the measuring systems (namely, +/- 0.3 pH units and +/- 12 mmol/liter in terms of [H+]).(ABSTRACT TRUNCATED AT 250 WORDS)