334 Profs. J. J. Thomson and R. Threlfall. [May 6, 



the experiments the gauze was placed at the ends and the turnings in 

 the middle, in the other half, half the tube was filled with copper 

 turnings and the other half with gauze. The air was sucked through 

 a tube containing pieces of pumice moistened with potash, and through 

 a bottle half filled with the same substance, the other end of the tube 

 was stopped by an indiarubber cork coated with paraffin, through 

 which a glass tube passed which conducted the nitrogen to a series of 

 bottles and tubes. These bottles and the porcelain tube were made 

 quite air-tight ; this was tested by putting the tube through which 

 the air passed on its way to the copper in connexion with an air- 

 pump, and exhausting down to a pressure of about 20 mm. of mercury, 

 even with this exhaustion there was no appreciable leak through the 

 whole arrangement of porcelain tube, drying-tubes and bottles, 

 discharge-tubes and connexions. The porcelain tube was connected 

 by a piece of thick- walled indiarubber tubing with a series of bottles 

 and tubes containing purifying reagents. After leaving the copper 

 the nitrogen passed through a potash solution in a bottle, then 

 through two large tubes filled with carefully prepared pumice 

 moistened with potash, it then bubbled through sulphuric acid which 

 had been boiled down with sulphate of ammonia to about one-fourth 

 of its original volume, it then passed through two large U^ 11 ^ 8 

 filled with phosphorous pentoxide divided up into a number of layers 

 by asbestos plugs, it then went into a large bottle about one-fourth 

 filled with phosphorous pentoxide. All the phosphorous pentoxide used 

 was tested and found to be free from free phosphorus. The gas after 

 leaving the phosphorous pentoxide bottle passed through thick- walled 

 indiarubber tubing into the discharge-tube. The volume of the tubes 

 and bottles was very large compared with that of the discharge-tube, 

 and as our consumption of nitrogen was slow the gas we used had 

 stood over the phosphorous pentoxide for several days at least and 

 often much longer. On the other hand, the gas had only been in 

 contact with indiarubber for a short time, for the indiarubber bungs 

 in the bottles were all coated on the inside with paraffin, and the only 

 long piece of tubing was that leading from the last drying bottle to 

 the discharge- tube ; gas that had stood in this for more than a few 

 minutes was always sucked out, and was never used for filling the 

 discharge-tube. 



The oxide of copper formed in the tube was reduced from time to 

 time, in most cases by passing hydrogen through the tube ; the 

 hydrogen was generally prepared by pouring sulphuric acid on zinc, 

 but as we suspected that a trace of sulphur dioxide which we detected 

 had its origin in this source, we endeavoured to reduce the copper 

 oxide by electrolytically prepared hydrogen ; we could not, however, 

 produce the hydrogen fast enough in this way, and so we finally 

 reduced the copper oxide by carbon monoxide prepared from pure 



