60 PROCEEDINGS OF THE AMERICAN ACADEMY. 



hydrochloric acid gas or dry air could be passed through the tube con- 

 tainiiisj the weio-hiiia; tube aud boat. 



With the help of this contrivance it was found possible to drive off the 

 amnionic chloride in a current of dry hydrochloric acid, to drive off the 

 excess of acid from the fused maguesic chloride by means of a current of 

 perfectly dry air, and to shut up the pure salt in a weighing bottle with- 

 out the least possible means of access of a trace of aqueous vapor. The 

 details of the method will be described later ; magnesic chloride prepared 

 after this fashiou gives a perfectly clear solution in water. Since this 

 problem was solved, attention was now turned to the preparation of 

 materials for the atomic weight determinations. 



Preparation of Materials. 



The sample of ammonic magnesic chloride which will be hereafter 

 referred to as sample No. 1, was prepared as follows. About five hun- 

 dred grams of ordinary " C. P." magnesic chloride was saturated with 

 hydrogen sulphide, a small amount of ammonia was added, and the whole 

 was allowed to stand in a warm place for several days. To the super- 

 natant liquid after decantation a small quantity of very pure ammonic 

 oxalate was added. The magnesic chloride thus almost wholly freed 

 from calcium was again decanted ; and after more ammonic oxalate 

 had been added, the whole was allowed to stand, and the clear liquid was 

 yet once more decanted. The solution was then evaporated to dryness, 

 and the resulting cake dried in an oven and ignited in a platinum dish. 

 The mixture of magnesic oxide and oxychloride thus formed was washed 

 with the aid of a filter pump for about sixty hours. At the end of this 

 time, although the wash water contained no sodium, the insoluble pre- 

 cipitate was not free from that metal. The precipitate was therefore 

 dissolved in hydrochloric acid, previously distilled in platinum for the 

 purpose, and the solution was filtered. In order to eliminate the sodium, 

 a portion of the magnesium was precipitated by passing into the solution 

 a current of ammonia gas. The precipitate formed b}' this very wasteful 

 process was washed for several days, at the end of which time it was found 

 to be free from any appreciable traces of sodium and potassium, when 

 tested with the spectroscope. 



Ammonic chloride was now prepared by mixing streams of ammonia 

 and hydrochloric acid gas. This gave ammonic chloride mixed probably 

 to a certain extent with various amines, but free from inorganic salts. 

 As the amines must be driven off later, it was not thought worth while 

 to take the trouble of removingr them at this staire of the work. 



