series, by this method, t o be the true result, then the maximum variation below this result is only 0.17 and the maximum variation above is 0 . 2 6 . These differences might well be considered within experimental error. RATIO O F AMMONIA TO CITRIC ACID. Dr. McCandless, the referee on phosphoric acid for the Association of Official Agricultural Chemists, in 1909, suggested that a reagent be used which should have the same ratio of ammonia to citric acid as the pure triammonium salt. This ratio

more »

... s I : 3.765. He made some cooperative tests with three other chemists, in which solutions were obtained having ratios from I : 3.775 to I : 4.189. Some time later the Division of Fertilizer Chemists of the American Chemical Society recommended a solution in which the ratio of ammonia to citric acid was I : 4 . 2 5 . A neutral solution was made up by the conductivity method and the ratio of ammonia to citric acid determined. An average of three determinations gave a ratio of I : 3.785. This would probably have approached still closer to the theoretical ratio if the neutral solution had been made up immediately after making the conductivity measurements. During the several days that elapsed between the determinations of the quantity of dilute ammonium hydroxide solution required to give perfect neutrality and the actual preparation of the neutral solution, the strength of the dilute ammonium hydroxide solution was probably changed. I t seems probable that this method could be used to give more satisfactory results than the present Official Method. The present Official Method for preparing a neutral solution of ammonium citrate is extremely unreliable. The purified litmus method gives somewhat better results, although the limit of error is too great for reliable work. The conductivity method, on the other hand, is reliable and not difficult of operation and the results obtained by different workers agree very closely. COKCLUSIONS.