2G0 



DETEEMIXATIOX OF niOSrilORK OXIUE. 



determining- the weight of precipitate obtained — that is, l^y 

 an indirect method. Working- in the same way, we obtained 

 3297 ±0003 for the percentag-e of pliosphoric oxide. Since 

 a knowledge of the composition of the precipitate wouhl throw 

 light on the volumetric process previoirsly described, and 

 possibly explain why 'the ratio of sodium hydroxide to 

 phosphoric oxide was so much higher than ratios found for 

 other modifications, the complete analysis of the precipitate 

 was undertaken. 



Method. — The scheme adopted for the analysis of the 

 yellow precipitate was as follows : Precipitates were prepared 

 and dried in i)artial vacuum after washing- with acetone, as 

 recommended by Neubauer and Lucker,* and were weighed. 

 Some of these piecijiitates were dissolved in ammonia 

 solution, the phosphoric acid precipitated as ammonium 

 magnesium phosphate, the filtrate acidified with acetic acid, 

 and the molybdic acid precipitaied as lead molybdate. 

 Considerable difhculty was experienced in effecting a satis- 

 factory separation of the ammonium magnesium phosphate 

 from molybdic acid. Two precipitations freiniently failed 

 to remove all molybdenum from the phosphate precij^itate, 

 and the separation is undoubtedly more difhcult to effect 

 quantitatively than when ammoniiun phospho-molybdate is 

 precipitated by one or other of the modifications of the 

 molybdate-magnesia method. For this reason, the sulphate- 

 molybdate method of precipitation is not to be recommended 

 for irse when it is desired to weigh as magnesium pyro- 

 jdiosphate. Ammonia in the precipitate was determined by 

 distillation into standaid acid in the usual way, while water 

 was estimated by drying- in a cuiTent of air at 175 degrees. 



Tlie folhnving- data were ol)tained : — 

 Ammonia. 



Mean: 2-18 + 0003. 



Zc'it. Anal. Chew., li. 161 (1912). 



