1905} Davis — Atomic Weight of Thokium. 51 



an iron cylinder, and was analyzed for oxygen and other 

 impurities. About two per cent, of an inert gas was found, 

 but it was not deemed necessary to remove it. The chlorine 

 was dried by passing over a train of eight towers and wash 

 bottles containing' sulphuric, and two towers with glass beads 

 and sulphuric acid, then finally through five towers with glass 

 beads and sulphuric acid, this being the apparatus used by 

 Richards 1 in his work on the atomic weight of cobalt and 

 nickel. 



The method of determining the ratio between the chloride 

 and the oxide was to weigh the chloride, dissolve in water and 

 evaporate in a platinum crucible, then ignite to constant 

 weight. The portion of the tube in which the crystals were 

 formed was cut out and slipped quickly into a weighing bottle 

 and then placed in an air bath at 160°C. to drive off the chlo- 

 rine It was then removed and placed in a dessicator and 

 after an hour weighed. After weighing the chloride was dis- 

 solved in water and then the tube returned to the weighing 

 bottle and broug-ht to constant weight. This gave the weight 

 of the thorium tetrachloride. 



The solution of thorium tetrachloride was evaporated in a 

 platinum crucible on the water bath, then ignited to constant 

 weight. In ordei to bring to constant weight the oxide had 

 to be blasted about seventy hours. This gave the weight of 

 the oxide derived from a given amount of the chloride. But 

 no constant relation could be obtained. 



Thinking that perhaps the difficulty lay in not being able 

 to drive off all the chlorine by heat and ignition, we turned 

 our attention to the determination of the halog-en. A long 

 series of experiments investigating a new method for the 

 determination of the halogen was begun. Since silver chlo- 

 ride is soluble in water to a certain extent, we decided to 

 investigate its solubility in alcohol. 



The alcohol used was absolute and freed from aldehyde by 

 fractional distillation until it would not respond to the fus- 



1 Proc. Amer. Acad. Arts & Sci. Vol. 34, No. 13, Pp. 332, 1889, 



