140 REPORTS ON INVESTIGATIONS AND PROJECTS. 



started from the double chloride of thorium and potassium prepared 

 by dissolving pure thorium hydroxide in pure hydrochloric acid, 

 adding a calculated amount of potassium chloride, evaporating on 

 water-bath to dryness, and then heating the resulting mass to redness 

 in a current of dry hydrochloric acid gas. This double compound 

 was readily and completely soluble in water ; therefore Nilson con- 

 cluded that it was K^ThCl^. This was reduced by means of metallic 

 sodium in an iron cylinder which was heated, in a large Roessler gas 

 furnace, to white heat. The double chloride was reduced to a gray 

 metallic powder which took fire far below red heat. It was readily 

 soluble in hydrochloric acid, slowly soluble in dilute nitric and sul- 

 phuric acids, and insoluble in a potassium hydroxide solution. The 

 material contained oxygen. This Nilson attributed to occluded air 

 between the particles within the cylinder, but Kriiss later showed 

 that the material used by Nilson was in fact an oxychloride of 

 thorium combined with potassium chloride. 



Our experiments on the reduction of thorium oxychloride by 

 means of metallic .sodium in a closed, specially constructed iron 

 cylinder, from which the air was exhausted, gave us a gray powder 

 with similar properties to that described by Nilson, but much con- 

 taminated with the oxide. It was impracticable to separate this 

 oxide. We also made the following interesting observation : When 

 our steel bomb, with walls of 2 cm. thickness, was heated by means 

 of gas we invariably found thorium carbide as one of the contamina- 

 tions of the metal as well. The gas evidently filtered through the 

 steel. 



Thorium readily forms a carbide when efforts are made to reduce 

 the oxide by means of carbon in an electric furnace. This we have 

 demonstrated in the preparation of considerable quantities of the 

 carbide which are to be used in the investigations. 



Efforts to electrolyze the tetrachloride with carbon and iron ter- 

 minals in a copper furnace, according to the method of Muthmann, 

 who in this manner prepared pyrophoric cerium, praeseodidymium, 

 neodidymium, and lanthanum, were fruitless. Thorium invariably 

 forms the carbide when carbon is present and the metal is reduced 

 at great heat. 



An Heraeus electric oven was substituted for the gas furnace. 

 We obtained the metal with the oxide impurity as described above. 

 It is in a fine state of subdivision, as the temperatures for such fur- 

 naces are limited to the fusing point of platinum, which is much 

 below the melting-point of thorium. The melting-point of thorium 

 has not as yet been definitely determined. 



