Some Sources of Error in the Determination of Chlorides in Blood and Similar Material.*

In the course of work in the analysis of blood, the authors have employed a number of reagents for the removal of the proteins before proceeding to the determination of other constituents. Two of these, picric acid (1) and nitric acid (2)) appeared to offer peculiar advantages in the subsequent determination of chlorides and other inorganic constituents. However, it was found that the amount of silver chloride that was obtained from volumes of filtrates corresponding to the same amounts of blood was much less when nitric acid was used as the precipitant than when picric acid was so employed. Moreover, the silver precipitate obtained with the latter was always yellow and it retained this color even after prolonged washing with hot water. The color could be removed by washing with concentrated nitric acid but the weight of the residual silver chloride was greater than that of the-precipitate obtained from a nitric acid filtrate corresponding to the same amount of blood. The yellow silver chloride precipitate was not completely soluble in ammonia but left a gelatinous precipitate, suggesting the presence of a purine. The precipitate was filtered out, washed, and suspended in dilute hydrochloric acid. After heating for 30 minutes, the mixture ‘was filtered, the filtrate was almost neutralized and was then treated with sodium acetate, sodium bisulfite, and copper sulfate in the usual manner. The precipitate obtained was filtered out, washed with hot water, and decomposed with hydrogen sulfide. The filtrate from the copper sulfide was evaporated and treated with a little picric acid. On cooling, a lemon-