BY H. I. JENSEN. 909 



of perhaps 0*50 %, whilst in the trachytes it would be practically 

 absent. By taking 0*0025 gram of microcosmic salt containing 

 about 0-0005 gram of PoO,-, precipitating in the usual way with 

 ammonium molybdate, redissolving the precipitate and precipi- 

 tating as magnesium phosphate, a precipitate weighing 0050 

 gram was obtained, equivalent to 0-0030 gram of VoO-. 



This I attribute to the following cause, viz., the ammoniacal 

 solution containing the magnesium phosphate in standing 24 

 hours takes up the carbon dioxide from the air and silica from 

 the glass vessel, with the result that one weighs basic magnesium 

 silicate and carbonate with the phosphate. Error from this 

 cause can only be avoided by letting the solution stand in a 

 platinum vessel in a carbon-dioxide-free atmosphere. 



It was also noticed that, in following Washington's method of 

 decomposing the rock with nitric and hydrofluoric acid, in the 

 part used for determining PgOg, no precipitate was obtained 

 with P2O5 without warming, although the rocks might contain 

 apatite equivalent to between 005 and 0-25 % of PoO-; further, 

 on warming, too much molybdate precipitate is generally obtained. 

 These irregularities are jwohahly due to the presence of traces of 

 HF, which hinder the formation of the phosphomolybdate; and, 

 on warming, to the formation of a certain amount of silico- 

 molybdate; but the matter needs looking into. 



In the face of these difficulties and as P0O5 was an unimpoi tant 

 constituent in the rocks which I was examining, I did not 

 consider it worth my while to devise a method for overcoming 

 the difficulties. It is very possible that inexperienced analysts 

 often follow the text-book methods without enquiring into their 

 accuracy; and, not observing the many precautions necessary, 

 get high results for PoO -. The amount of this constituent given 

 in many analyses of trachyte, phonolite and granite seems 

 absurdly high. The smaller the actual quantity of PoOg in a 

 rock, the more exaggerated the error becomes. 



As accuracy in rock-analysis is daily becoming more important, 

 it would be well if some chemist could take up the matter of 

 devising a good laboratory method for determining P2O5. 



