102 EXPERIMENT STATION RECORD. 



ascribed to variations in the length and intensity of ignition of the pre- 

 cipitate. Intense ignition appeared to increase the loss of phosphoric 

 acid by decomposition of the metaphosphates, and this loss increased 

 witl "'e increase in weight of the precipitate. In the author's exi^eri- 

 mcii' • bore no definite relation to the correction given in Neubauer's 

 table, ihe indications are that the precipitate of Mg(]SrH4)4(POj)2 is 

 very variable, and consequently that the proportion of metaphosphate 

 in the ignited residue is also variable. Experiments are reported which 

 indicate that the difference in weight between the gently ignited and 

 strongly ignited precipitate gives a factor which may be utilized in cal- 

 culating tlie relative proportion of metaphosphate and pyrophosphate 

 present, and thus furnishes a more exact means of determining the per- 

 centage of phosphoric acid present. As a rule this method gave more 

 reliable results than either Wagner's method or Neubauer's method 

 with correction. 



The results by the ^liircker method were very variable, sometimes 

 showing a plus error and sometimes a minus error. In many cases it 

 was found that the precipitate contained a considerable amount of 

 molybdic acid. The error due to formation of metaphosphate and vol- 

 atilization of phosphoric acid was much smaller than in case of the 

 Wagner method. This is exjilained by the fact of the simultaneous 

 formation of Mg(NH4j4(P04)2 and Mg;j(P04)2, which yield on ignition 

 magnesium pyrophosphate without loss of phosphoric acid when com- 

 bin(!d in proper proportion. If the amount of i)hosphoric acid present 

 is not very large it is completely precipitated in the above forms in 

 proportions which yield on ignition pure pyrophosphate, and the results 

 generally show a plus error (due to other impurities). If phosphoric 

 acid is present in large amounts the addition of the large amount of 

 magnesia mixture, which is then necessary, favors the formation of an 

 excess of Mg(NH4)4(P04)2 and results in a minus error. The experi- 

 ments reported indicate that although many of the precipitates contain 

 molybdic acid, accurate results may be obtained by this method when 

 precautions are taken to delay precipitation, as for instance, by the 

 addition of citric acid.' 



In order to get the best results, as Marcker has shown, the solution to 

 be precipitated should not contain more than 2 gm. of phosphoric acid, 

 and sufficient hydrochloric acid should be used so that precipitation 

 does not commence until about one-half of the required magnesia mix- 

 ture has been added. The precif»itate should be strongly ignited in 

 order to remove the molybdic acid which may be present. 



The results of a number of experiments are reported, which indicate 

 that the large amounts of ammoniara chlorid resulting from the use of 

 hydrochloric acid in the solution of i^hosphates rich in iron in no way 

 affects the determination of phosphoric acid by the molybdic method. 



Methods of phosphate analysis, O. Grothe (Amer. Fert., 4 



1 Ztsehr. analyt. Cbem., 32 (1893), p. 64. 



