In order to determine the stability of directly measured pH and PCO2, and calculated [HCO3] in stored urine, 11 alkali-loaded normal subjects provided 33 spot and sixteen 24-hour mineral oil-covered, thymol-preserved, refrigerated urine samples. For the spot samples, pH and PCO2 were measured immediately and again at 4 and 24 h. In addition, immediately after voiding, 24 of the spot samples were split into oil-covered and no-oil moieties and analyzed immediately and again at 4 and 24 h. pH and PCO2 measurements of the 24-hour collections were carried out immediately after completion and again 24 h later. The results demonstrated the importance of using oil to limit the escape of CO2 from stored urine. Thus, after 24 h the oil-uncovered subgroup of 24 spot urine samples sustained a 82% decline in PCO2 and 20% fall in [HCO3]. In contrast, the corresponding percentage decrements in the oil-covered samples were 16 and 1%, respectively. The results also indicated that even with oil there is loss of CO2, which increases with time and which shows a statistically significant direct correlation with the baseline level of PCO2. Nevertheless, modest loss of CO2 usually produces only slight decrements in [HCO3] because of the countervailing influence of the resultant increase in pH. We conclude that the use of mineral oil is necessary in order to obtain adequate stability of PCO2, pH and [HCO3] in refrigerated urine requiring several hours or more of storage.