Determination of chloride concentration in cheese: collaborative study

Three samples of ground Gouda cheese containing 1-2% chloride were analyzed by 7 laboratories by 3 methods: oxidation with KMnO4 and HNO3 followed by a Volhard titration; the same but with filtering off the precipitated AgC1 before back-titration; and the general potentiometric method without ashing or oxidation. The data were analyzed by ISO statistics (ISO-R 5725) and by AOAC statistics (Youden), the major differences being the rejection of different values as outliers and in the statement of the precision parameters. The within-laboratory variability (repeatability) is comparable for all 3 methods; the between-laboratory variability (reproducibility) is comparable for the Volhard method with filtration and the potentiometric methods, but the direct Volhard method is inferior. Because of its generality and simplicity, the potentiometric method has been adopted official first action; the Volhard method with filtration has been reinstated official final action as an alternative.