102 AGRICULTURAL ANALYSIS 



tation is incomplete and the separated liquid is precipitated alike 

 b\ magnesia and phosphoric acid. 



The results of the work show that when stirred with ammonio- 

 magnesium phosphate the final result is -obtained without any loss, 

 and that on calcination the identical weight with which the experi- 

 ment commenced is obtained. The conclusions obtained by Lasne 

 from a large mass of analytical data are summarized as follows : 



(1) The formation of phosphoric acid in the state of pyrophos- 

 phate without any other precaution than previous elimination of 

 silica gives results which are not affected by any systematic error. 



(2) Rapid precipitations cause an excess of weight in the pre- 

 cipitate due to a partial formation of tri-magnesium phosphate, 

 which is only transformed into ammonio-magnesium phosphate by 

 contact for 16 hours with a sufficiently concentrated solution 

 of ammonium citrate, namely, containing ten grams of citric acid 

 in 150 cubic centimeters of the solution. It is necessary, there- 

 fore, in order to obtain rrgorous results, to allow the precipitate 

 to stand before filtering for at least 16 hours. Nevertheless, it 

 must be admitted that this excess of weight is so small as not to 

 warrant the complete rejection of the rapid methods for all in- 

 dustrial purposes when proper precautions are employed. 



(3) The transformation of tri-magnesium phosphate into am- 

 monio-magnesium phosphate takes place very slowly in the pres- 

 ence of ammonium chlorid alone, and there should always, there- 

 fore, in these precipitations, be added the quantity of citrate of 

 ammonia indicated above. 



The precipitation of magnesia in the presence of an excess of 

 ammoniacal phosphate gives, in addition to ammonio-magnesium 

 phosphate, a phosphate poorer in magnesium, and as much poorer 

 as the excess of phosphoric acid is larger. The determination of 

 magnesia by this classical method is always erroneous. 



95- Comparative Accuracy of the Citrate and Molybdate Meth- 

 ods. The general use of the citrate method of determining phos- 

 phoric acid by the German chemists has led Johnson to review 

 some trials of that method in his laboratory made as early as 

 i88o. 72 These determinations have lately been repeated in com- 

 71 Journal of the American Chemical' Society, 1894, 16 : 462. 



