Spectra in Compound (rases. 731 



discharge is a most important factor. Now small intensity 

 implies low (electrical) temperature ; and it is known that 

 the effective temperature of the anode is much less than 

 that of the cathode. Therefore it appears likely that there 

 are at least two causes of the differences between the anode 

 and cathode spectra of gases, viz.: — 



(a) The inequality of temperature (and potential gradient) 

 at the two electrodes. 



(7>) The electrochemical character of theelements concerned. 



Kayser, in his great Hanflbuch der Spectroscopies somewhat 

 boldly asserts that (a) alone is effective*. He is led to this 

 conclusion, apparently, by an experiment of Morris- Airey's f 

 which seems to be antagonistic to (b). A discharge was 

 passed for some time through a tube containing hydrogen 

 and a little chlorine. The tube was then divided into two 

 portions, and the chlorine spectrum was found to show the 

 same intensity at the anode of each half tube. It is claimed 

 that this proves there is no appreciable separation of hydrogen 

 and chlorine by a process of the nature of electrolysis. But 

 this conclusion does not appear to me to be justifiable. Might 

 it not be almost as well said that, because in the electrolysis 

 of dilute sulphuric acid bubbles of oxygen can be seen only 

 at the anode, therefore no oxygen exists anywhere else in the 

 voltameter ? Just as oxygen may be present in the liquid 

 without being detected by eye, so may the chlorine be dis- 

 tributed throughout the tube, and yet not show its spectrum 

 except where its ions are being discharged, viz. at the anode. 

 Further, even granting that this experiment does prove the 

 absence of a process resembling liquid electrolysis, it is still 

 far from legitimate to assume that (a) is the only cause, for 

 the scope of (6) is not by any means limited to pure electrolysis 

 (i. e. a process subject to Faraday's laws). 



That (a) alone is quite incapable of explaining the facts 

 seems perfectly clear. Otherwise how is it that in no case 

 has an electronegative element been found spectroscopically 

 only at the cathode, or an electropositive one only at the 

 anode ? Why should electronegative elements in nearly 

 every case show a preference for the positive pole, and electro- 

 positive elements for the negative pole ? When a substance 

 which, like chlorine, is normally electronegative is known 

 from chemical evidence to be acting as an electropositive 

 element, it is seen by the spectroscope at the cathode instead 

 of as usual at the anode. If appreciable separation occurs at 



* Kayser, Handbuch der Spectroscopie, vol. ii. pp. 174, 175. 

 t Morris-Airev, Drude Ann. i. 1900, p. 4(36 ; also Phil. Mag. r51 xhx 

 p. 307(1900). 



