182 C. Fairbanks — lodometric Method for the 



(1*4:2 sp. gr.) and YOO'^'^ of water. This solution, it is said, 

 practically remains unaltered for months ; but it is much safer 

 to filter each day the portion to be used. 



In order to obtain a standard solution of phosphorus, 1*2 

 grams of microcosmic salt were dissolved in one litre of water 

 and 50^^^ of this solution were evaporated, and the residue 

 ignited and weighed as sodium metaphosphate. 



Iron free from phosphorus, for admixture with known 

 amounts of phosphorus in the test experiments, was prepared 

 by igniting ferrous oxalate. The oxalate was precipitated by 

 oxalic acid from ferrous sulphate dissolved in warm water. 

 The precipitate was then thoroughly washed with cold water, 

 dried and ignited. 



The ferric oxide thus obtained was dissolved in hydrochloric 

 acid, evaporated to dryness, and dried on a hot plate. The 

 ferric chloride was treated with nitric acid and the solution 

 evaporated twice. In this way all but a trace of the chloride 

 was destroyed, 



Since in the ordinary determination of phosphorus in iron, 

 potassium permanganate is added, as Dr. Drown says,^ to insure 

 the complete oxidation of the phosphorus to the orthophos- 

 phoric acid, it seemed best to make sure that this treatment 

 was not detrimental in the process to follow. Enough perman- 

 ganate was therefore added in my experiments to distinctly 

 color the already boiling solution of ferric nitrate, and then a 

 little tartaric acid was used to clear the solution. 



As too much free acid prevents the full precipitation of the 

 ammonium phospho-molybdate, the ferric nitrate solution was 

 treated with ammonia until a precipitate or the ruby wine color 

 appeared. Then dilute nitric acid was added until the color of 

 the solution was a clear amber. Generally four or five grams 

 of ferric oxide were used in each determination. 



The ferric nitrate and phosphorus solutions, in volume not 

 less than ISO^'^ nor more than 300% were drawn into an ordi- 

 nary 500*^° flask. The flask was heated on the steam bath to just 

 85° C. — very carefully, since at higher temperatures there is 

 danger of precipitating molybdic acid — and immediately 40^° 

 of the filtered ammonium molybdate solution were so added 

 as to wash down the thermometer and the sides of the flask. 

 The flask was corked, wrapped in a towel, and shaken for five 

 minutes. The contents were filtered at once on an asbestus 

 felt, and the precipitate washed three or four times, first with 

 a 10 per cent solution of nitric acid, and then with a 1 per cent 

 solution of potassium nitrate. 



If the nitric acid and potassium nitrate solutions wash the 



* Trans. Am. Inst, of Min. Eng., vol. xviii, p. 93. 



