Assessment of Scholander micromethod for gas concentrations versus weighing method.

Three gaseous mixtures in N2, of varying concentrations of CO2 (3.60, 5.79, and 9.36%) and O2 (22.18, 15.12, and 6.82%), respectively, were prepared by precise weighings, which may serve as absolute measures. These mixtures were analyzed with two apparatuses by the Scholander micromethod. Taking into account the influence of several factors (apparatuses, order of analyses, etc, it appears that differences between apparatuses were not significant in five of six cases and the effect of order was significant in O2 (P less than 0.01) but not in CO2 for the set of measurements. When compared with weighings, measurements differed only in one apparatus (P less than 0.02) in two of six cases and by no more than 0.02%. The relation between weighings and measurements was highly significant (P less than 10(-9) and no lack of fit to linearity occurs. Side effects of the factors considered and interactions between them are discussed. In conclusion, the micromethod of Scholander gives accurate results, in very good agreement with the reference values provided that certain technical precautions are observed.