of ferric and of other phosphates. 389 



presence of iron and phosphoric acid in the solution, measured 

 portions were now removed with the object of determining the 

 ratio of Fe 2 3 to P 2 5 . 



The P 2 5 was determined by Woy's modification of the molyb- 

 date process (Chem. Ztg. Vol. xxi. pp. 441 and 469, 1897) and 

 his results have been extensively verified by Sherman and Hyde 

 (Jour. Amer. Chem. Soc. Vol. xxn. p. 652, 1900). 



Expt. (1) 100 c.c. gave 2-228 grms P 2 5 24Mo0 3 = "0879 P 2 0,. 

 „ (2) 100 c.c. „ 2-224 „ „ „ =-0878 „ 



(3) 250 c.c. after reduction of the iron to the ferrous 



N 

 state required 34*5 c.c. — potassium permanganate. 



N 

 „ (4) 250 c.c. after reduction required 34'5 j^ potassium 



dichromate. 



In 100 c.c. this number gives 11 grm Fe 2 3 . We have 

 therefore in 100 c.c. '0891 P 2 5 to -11 Fe 2 3 . Dividing these 

 numbers by their respective molecular weights we have 



P 2 5 : Fe 2 3 :: 61 : 68 or as 1 : Ml, 



thus shewing that the solution contains the compound FeP0 4 

 with a minute quantity of Fe(OH) 3 . 



Further experiments were now carried out with a different 

 sample of the dialysed solution and a sufficient supply was 

 prepared and evaporated to dryness on the water bath, the dried 

 residue tubed and separate portions taken for quantitative 

 analysis. For this purpose the P 2 5 after precipitation by the 

 molybdate process was dissolved in ammonia and weighed as 

 magnesium pyrophosphate. 



The sample was found to contain water to the extent of 

 5 40 per cent, expelled on ignition. 



The iron was determined as in the former experiments. 



The percentages work out as follows : — 



Fe 2 3 = 53-69 

 P 2 5 = 4071 

 HoO= 5-40 



99-80 



which gives the molecular ratio P 2 5 : H 2 : Fe 2 3 as "29 : 3 : '33 

 and the formula 2FeP0 4 , H 2 indicates that the substance as 

 dried on the water bath is a hydrated ferric phosphate somewhat 

 basic from the presence of a small quantity of ferric hydrate. 

 This small quantity of ferric hydrate is to be expected both from 



