254 PROCEEDINGS OF THE AMERICAN ACADEMY. 



questions as the possible presence of gases in the oxide, the possibility 

 of incomplete reduction, and the danger of contamination from the con- 

 taining vessels, should obviously receive attention. 



The balance and weights used in the present study vrere the same as 

 those used in other recent determinations of atomic weitrhts.* It is need- 

 less to state that the weights were carefully justified, and that all the 

 usual precautions necessary to procure accuracy in such work were ob- 

 served. All the weighings were reduced to the vacuum standard, and the 

 values of the atomic weights refer to the universally accepted standard 

 = 16.000. 



One of the most important steps in such an investigation is clearly the 

 choice of the substance to be analyzed, and the choice usually requires a 

 careful preliminary qualitative and quantitative study of many compounds 

 of the element under consideration. This work occupied a number of 

 weeks, and in the course of it we found that the preparation of either 

 ferrous or ferric halides in a pure anhydrous state offered so many diffi- 

 culties that as yet no satisfactory method for the purpose has been de- 

 vised. Ferric oxide, on the other hand, can be obtained in a variety of 

 ways, hence the choice of the early experimenters seems to have been 

 a wise one. The first step in a systematic treatment of the subject is 

 obviously a repetition of the discordant work of half a century ago, with 

 especial reference to the possible errors noted above. 



Owing to the fact that most oxides formed from nitrates are known to 

 include considerable quantities of oxygen and nitrogen, f ignition of ferric 

 nitrate as a means of obtaining ferric oxide was at first rejected, and ferric 

 hydroxide Was used instead. Very pure iron ribbon was dissolved in 

 sulphuric acid, and the metal was then deposited electrolytically from an 

 ammouic oxalate solution. After the film had been dissolved in the pur- 

 est nitric acid, the carbon was removed by filtration, and the iron \vas 

 precipitated as ferric hydrate by the addition of an excess of ammonia 

 which had been distilled into pure water contained in a platinum dish. 

 The precipitate was washed thoroughly, collected upon a pure washed 

 filter paper, separated from the filter paper while moist, and finally dried 

 in a platinum dish upon the steam bath. The lumps of dried hydrate 

 were crushed in an agate mortar and finally ignited for several hours in 

 a current of pure dry air. During the latter operation the weighed 

 platinum boat with the oxide was contained in a large porcelain tube 



* These Proceedings, 26, 242 (IWH). 



t Ricliards, These Proceedings, 26, 281 (1891). Richards and Rogers, ibid., 28, 

 200. 



