RICHARDS AND BAXTER. ATOMIC WEIGHT OF IRON. 257 



evaporated, and the salt was three times recry stall ized from a large 

 amount of nitric acid. This acid had been distilled in a platinum still 

 and condenser and collected in a platinum receiver. The mother liquor 

 of the first crystallization vpas analyzed for impurities as follows : After 

 precipitation of the iron as ferric hydrate by the addition of an excess of 

 pure ammonia and filtration, the filtrate was evaporated to dryness and 

 the residue was ignited. Since only a negligible amount of alkaline 

 solid remained, it can fairly be assumed that the final preparation of ferric 

 nitrate must have been free from such impurity. During the latter part 

 of this purification platinum vessels only were employed, and the water 

 and reagents were the purest obtainable. 



The next stejD, that of converting the nitrate into oxide, was carried 

 out in part by gentle heating in an electric oven at about 150°.* The 

 resulting cake of basic nitrate, after removal from the platinum dish in 

 which it was contained during the above operation, was crushed in an 

 agate mortar, and the reaction was completed by ignition in an open 

 porcelain tube at a temperature of about 900° for several hours. The 

 oxide, which was now contained in a large platinum boat, was carefully 

 weighed and again heated. The loss of weight during this second igni- 

 tion amounted to one tenth of a milligram, while a third prolonged 

 heating caused a further loss of three-tenths of a milligram. This differ- 

 ence must have been caused by the escape of included gases, and was sO' 

 small in comparison with the weight of the oxide (seventeen grams), that 

 further heating was deemed unnecessary. 



The determination of the gases contained in this oxide proved to be a 

 matter of considerable difficulty. At first the method used was that of 

 dissolving the oxide in a Sprengel vacuum in fused acid potassic sulphate. 

 Owing to the fact that the sulphuric acid, which was evolved from the 

 fused sulphate, attacked both the rubber connections used for joining the 

 apparatus, and also the stick potash used as a drying agent, it was uncei-- 

 tain whether the small quantity of gas which was collected in the air 

 pump was originally included in the oxide. A more successful method 

 was eventually found by dissolving the oxide in hot hydrochloric acid con- 

 taining stannous chloride. The operation was carried on in an apparatus 

 for determining the gases' evolved during the solution of a substance, 

 which has been already described in papers from this laboratory. f In 

 two experiments half a gram of ferric oxide yielded a mere trace of gas. 

 Cupric oxide treated in the same way evolved considerable gas, less, how- 



* Richards, Am. Cliem. Jour., 22, 45 (1899). 

 t These Proceedings, 28, 200. 



