Wei; thts of Radium and Thorium Emanations. 65 



attached by rubber to the diffusion-vessel (fig. 2). A small 

 quantity of soda-lime was placed near the open end of the 

 tube, and separated from the potassium permanganate by a 

 plug of glass-wool, so that the oxygen as it was generated 

 had to pass over it. 



JPiar. 2. 



Scca-Lime KsMr^O? 



Before starting an experiment, the apparatus was exhausted 

 and filled with oxygen by heating the potassium permanga- 

 nate in a Bunsen burner ; this oxygen was then pumped out 

 and the apparatus refilled with oxygen, after which the 

 diffusion- vessel could be filled with pure oxygen whenever 

 required. 



To find the composition of the gas in the diffusion-vessel 

 at the end of an experiment, a suitable volume was withdrawn 

 and transferred to a gas burette, where its volume was 

 measured. The gas was then transferred to a wide glass tube 

 inverted over mercury, where the oxygen was removed in the 

 usual manner by means of phosphorus vapour, and the loss 

 or" volume determined. 



Carbon dioxide. — To prepare this gas, the tube containing 

 the potassium permanganate was removed and replaced by a 

 similar one filled with pure dry sodium bicarbonate, from 

 which carbon dioxide is readily obtained by gently heating. 

 A- in the experiment^ with oxygen, the apparatus was twice 

 exhausted and filled with carbon dioxide before starting. 



The analysis of the gas left at the end of an experiment 

 Phil. May. S. I). Vol. 9. No, 49. Jan. 1905. F 



