=68. Comparative Accuracy of the Citrate and Molybdate Methods.=—The general use of the citrate method of determining phosphoric acid by the German chemists has led Johnson[60] to review some trials of that method in the Yale laboratory made as early as 1880. These determinations have lately been repeated in comparison with the ordinary molybdate methods with the result that in sixty-seven determinations on bone-dust, superphosphate, cotton-hull ashes, cottonseed-meal, tankage, bone-char, phosphatic guano, and phosphate rock, only three citrate results differed from those obtained by the molybdate method by more than three-tenths of one per cent. The greatest discrepancy between the two methods was 0.41 per cent, and the average difference was 0.09 per cent. The citrate method was found to give poor results when iron and alumina were present in considerable quantity. Ignited precipitates by the citrate method were found to contain as high as four per cent of lime, and iron and alumina in small quantities when these bodies were abundant in the original substance. In the molybdate method the rapid precipitation from solutions at 65° was found to give unsatisfactory results and it was found necessary to conduct the process at temperatures between 40° and 50°. With a relative excess of nitric or a relative deficiency of molybdic acid some phosphoric acid may easily escape precipitation. The chief objection to precipitating at 65° is found in the fact that in presence of considerable iron and alumina some of these bodies may be found in the yellow precipitate, whence they pass to the final ammonium magnesium phosphate. The citrate method, therefore, only gives safe results by compensating errors which in every class of phosphates must be empirically determined. The molybdate method gives results too high when iron and alumina are present in considerable quantity and the yellow precipitate is obtained at temperatures above 50°. On the other hand, if there, be a great relative excess of nitric acid the results may be too low unless the filtrates from the yellow precipitate be mixed with additional molybdic solution and digested until no further precipitate is formed. Comparative determinations made, by both methods, by the Association of German Experiment Stations have led to the conclusion that both give practically the same results when each one is conducted with the proper precautions peculiar to it.[61] In the latter part of 1892, at the general meeting of the Association, it was declared that the citrate method, after having been subjected to repeated tests, was found to be satisfactory, changing the composition of the solution so that it might have 1,100 instead of 1,000 grams of citric acid and four liters of twenty-four per cent ammonia to each ten liters. The data afforded by the citrate method, when applied to an artificial mixture of known composition, were more satisfactory than those obtained by the molybdic process.