728 Mr. G. Stead on the Separation of 



When the coil was first turned on the Swan spectrum was 

 seen at both poles, but after a few seconds it was replaced at 

 the anode by carbonic oxide bands, and at the cathode by 

 the " four-line " spectrum of hydrogen. There was no trace 

 of the latter at the positive pole, but there were faint indi- 

 cations of the carbonic oxide spectrum at the negative 

 electrode. 



(ii.) Ethylene Dichloride. — Initially the Swan spectrum 

 was visible at both electrodes, but in a short time chlorine 

 lines and the ordinary hydrogen spectrum appeared at the 

 cathode, whilst carbonic oxide bands together with a faint 

 secondary hydrogen spectrum were seen at the anode. 

 Carbon was gradually deposited on the walls of the tube. 



(iii.) Ethylidene Dichloride. — The behaviour of this vapour 

 was precisely analogous to that of ethylene dichloride. Thus 

 there is no difference spectroscopically to be observed between 

 these two isomerides. 



(iv.) Carbonyl Chloride (phosgene gas). — At the anode 

 bright carbonic oxide bands were observed, whilst chlorine 

 lines showed brilliantly at the cathode. The separation of 

 the two spectra in this case was practically complete. There 

 was no deposition of carbon on the walls of the tube. 



(v.) Mercury Dimethyl. — Liquid air was used to freeze out 

 any mercury vapour which might come over from the Topler 

 pump. The liquid was admitted into an exhausted bulb 

 through a tap in the usual way (see previous paper), except 

 that mercury could not be used here to displace the air from 

 the bore of the tap. This air was therefore removed by 

 running a little of the mercury dimethyl itself through the 

 tap, the bulb being then re-exhausted by means of charcoal 

 and liquid air. 



At first when the discharge was passed the mercury lines 

 (5791, 5769, 5461, 4358) appeared very brilliantly at both 

 poles, and they were accompanied on the positive side only 

 by a pure Swan spectrum, i. e. pure in that there was no trace 

 of the carbonic oxide spectrum. No hydrogen lines were 

 seen until the coil had been running some seconds, when they 

 appeared chiefly at the cathode. After a short time the 

 pressure diminished, and carbonic oxide bands became visible 

 at both electrodes. The greater the initial pressure the longer 

 it took before the spectrum of carbonic oxide was seen. 

 Finally a condition was reached in which hydrogen showed 

 mainly at the cathode, and carbonic oxide principally at the 

 anode, whilst mercury lines were visible faintly at both 

 electrodes. 



(vi.) Aniline. — -At the cathode the four-line spectrum of 



