54 Journal of the Mitchell Society. [March 



some came lower and some higher than those values. The 

 crystals that seemed to be purest gave values within the lim- 

 its named. From the inability to obtain a constant value, it 

 seemed that the crystals of chloride must be contaminated. It 

 will be remembered that the crystals of chloride were the mid- 

 dle portion in the formation of the chloride: that is, a light 

 "Weisser-dampf" passed to and condensed on the front of the 

 tube, just beyond the boat were the crystals of the chloride 

 used for the work, and there still remained a residue, mixed 

 with the carbon in the boat, that could not be volatilized. In 

 other words, by distilling thorium oxide in chlorine, three 

 fractions were obtained. 



After the first distillation was performed, a determination 

 of the equivalent of each of the end products was made by the 

 sulphate method. The following results were obtained: 



Oxide from non-volatile part. sulphate. At.Wt. 



1.636725 gm. 2.595223 241.44. 

 Oxide from most volatile. 



0.794692 1.309245 212.70. 



As the complexity of thorium had been indicated by Basker- 

 ville and the name "Carolinium" proposed for the heavier ele- 

 ment, we have applied the name of Carolinium to the heaviest 

 portion. For the lightest the name "Berzelium" was used, 

 and Thorium for the middle portion. 



After the redistillation of our carolinium, thorium and ber- 

 zelium fractions three times, a determination of the atomic 

 weight of each was made by the sulphate method. In chang- 

 ing the oxide into sulphate, sulphuric acid (1:1) was used. 

 This was then carefully evaporated. The berzelium seemed 

 to be rapidly changed; but the thorium and carolinium were 

 more persistent, requiring about four evaporations. In every 

 case sulphuric acid was added and evaporated at least four 

 times, to be sure that all the oxide was converted into sul- 

 phate. The carolinium sulphate was perfectly soluble in 



