CHEMISTRY — BASKERVILIvE. 145 



is no such difference, and that all of the fractions have the atomic weight 

 of old thorium. Charles Baskerville, in a note discussing Meyer and 

 Gumperz's paper, called attention to the necessity of exactly following the 

 procedure described by him for the preparation of the anhydrous chloride. 



This criticism necessitated a redetermination of the atomic weights of the 

 fractions prepared by the original process, combined with a comparative 

 study of the methods which have been suggested for the determination of 

 the atomic weight of thorium, and are based upon the relation of oxide to 

 adhydrous or hyd rated sulphate. 



In the determinations reported on below, the purely physical manipula- 

 tions, as heating and weighing, were carried out with the same precau- 

 tions as have been described in the previous work. For heating the sulphate 

 and oxide to a certain fixed temperature, our electric resistance furnace, 

 with heating space for two crucibles, was used. We had found in the earlier 

 work that the two counterpoised platinum crucibles when heated to 1,200° C. 

 lost in weight at a different rate. This difficulty was easily avoided by 

 heating to only 1,050° C, at which temperature the difference in the loss 

 is immeasurable. It was ascertained by experiment that at this temper- 

 ature thorium sulphate is quantitatively converted into oxide. 



In one experiment the use of counterpoised quartz crucibles was tried. 

 They could easily be brought to constant weight at 1,200° C, but it was 

 found that at this high temperature the thorium-oxide reacted with the 

 silica, apparently forming a silicate. This silicate having a different coeffi- 

 cient of expansion, the crucible cracked in one place. It is apparent from 

 this observation that quartz crucibles can not be safely used for like experi- 

 ments. 



Before weighing, both crucibles were always inclosed in glass-stoppered 

 weighing bottles of thin glass and of approximately the same weight. This 

 precaution was taken in order to avoid errors resulting from the highly 

 hygroscopic properties of both the sulphate and the oxide. 



Leaving out the first possible method for the atomic weight determination, 

 which consists in heating the weighed oxide with sulphuric acid repeatedly 

 and dr^dng at 400° C. to constant weight, there are two more processes 

 available. 



The first one is accomplished by dissolving pure hydroxide in hydrochloric 

 acid in a weighed platinum crucible, adding a slight excess of sulphuric acid, 

 evaporating on the water-bath, and then heating in an air-bath, at 400° C, 

 to constant weight. The sulphate thus obtained is then converted into 

 oxide by glowing. 



Experiments carried out in this way before had given 241 for crude caro- 

 linium and 258 after several crystaUizations of the chloride. 



Two more determinations made in the same way with material prepared 

 by Baskerville' s method resulted in the following figures: 0.9706 gram 



10 — YB 



