886 



HON. R. J. STRUTT ON THE LEAST POTENTIAL DIFFERENCE 



found that the cathode fall in ordinary nitrogen was extremely inconstant. Thus in 

 one case the value of this quantity, at first 315 volts, gradually rose to 410 volts 

 during the passage of the current for many hours. He found that this variation in 

 the value of the cathode fall was connected with the presence of traces of oxygen in 

 the nitrogen traces too small to be removed by the ordinary chemical absorbents, 

 such as alkaline pyrogallol. I was anxious to try whether the same cause was opera- 

 tive in my case. 



The method employed by WARBURG for getting rid of the last traces of oxygen 

 from his nitrogen was to transport metallic sodium electrolytically through the heated 

 glass of his discharge tube, so as to form a clean surface of that metal on the inside. 

 At the high temperature used the sodium rapidly absorbed all traces of residual 

 oxygen. 



Such a method was obviously inapplicable to my apparatus, put together as it was 

 with sealing wax joints. I used, therefore, a somewhat different method, which, 

 though troublesome to carry out, gave good results. 



This method was to bubble the sample of gas employed repeatedly through the 

 liquid alloy of sodium and potassium. 



The alloy must be manipulated in the absence of air to prevent its surface 

 becoming fouled. The vessel in which it was contained is represented in the next 

 figure. 



a, and b are glass bulbs blown on the side limbs of a Y-shaped tube f. On the 

 bottom limb of this Y there is a stopcock g. The lower part of the limb is of capil- 

 lary bore. 



The bulbs a and b can be placed in communication by means of the stopcock c. 

 d and e are bulbs containing phosphoric anhydride to constitute an additional 

 safeguard against the access of moisture to the alloy. The exit tubes from these 

 bulbs lead respectively to the pump-reservoir and to the sparking vessel. 



The alloy was prepared in a test-tube k, by melting some sodium and adding potas- 

 sium until the product remained liquid on cooling. A few drops of rock oil were from 

 time to time placed on the surface of the alloy. The vapour from this sufficiently 

 guarded against the access of air. 



To introduce the alloy into the apparatus, c was opened and g closed. The air was 

 then pumped out from the system of tubes. 



The test-tube containing the alloy was then brought up as shown in the figure, the 

 end h of the Y being under the surface of the alloy, g was then cautiously opened 

 so that the alloy was sucked up into the bulbs a and b. When it had risen about 

 half-way up them, as shown in the figure, g was closed, thus preventing the entry <>!' 

 any more of the alloy. In this way the alloy could be introduced without any 

 contamination. It had all the appearance of clean mercury. This apparatus was 

 used to absorb traces of oxygen from nitrogen in the following way : c being open 

 and in communication with the pump reservoir, the gas was admitted, c was then 



