162 MM. Elster and Geitel on 



in future designate this as the "electrolytic" theory, and. in 

 correspondence therewith, speak also of " electrolytic " excita- 

 tion by the flame. 



(3)' Its explanation is found in a thermoelectric difference 

 of the electrodes (Buff)*. 



The theories which are deduced from the unipolar conduc- 

 tivity of flame we may be permitted to pass without notice, 

 since, as will result from the following investigation, sources 

 of error may have prevailed here, causing the conclusions 

 drawn from them to appear doubtful. 



Besides these differences in respect of theory, however, the 

 different observers also adduce experiments which are abso- 

 lutely irreconcilable with one another. This goes so far that 

 even in regard to the question whether the positive or the 

 negative electricity is that which is proper to flame no unity 

 prevails. 



The reason of this lies in the fact that all the observers have 

 overlooked a point that plays a very essential part in the elec- 

 tricity of flames, namely the behaviour of the shell of air 

 which immediately envelopes the flame. 



The maximum of electromotive force is always found when 

 one wire is introduced into the latter, and another into the 

 interior of the flame, as will be shown in the following. At 

 the same time, however, with the electrodes in this position 

 the resistance of the hot layer of gas separating the wires is 

 uncommonly greater ; so that we can hardly reckon on mea- 

 suring the intensity of the current with the aid of a multiplier ; 

 it is perhaps a consequence of this circumstance that all the 

 observers who have investigated flame-electricity by means of 

 that instrument have lost sight of the point above named. 



As hitherto, so far as is known to us, the electricity of flame 

 has not been examined with an electrometer permitting exact 

 measurements, it appeared to us of importance to test the 

 electric behaviour of flame with Thomson's quadrant-electro- 

 meter, and eventually to verify the correctness of one of the 

 theories above cited. 



§ 2. Apparatus and Method. 



In order to keep the charge of the needle of the electro- 

 meter constant, it was connected with one of the poles of a 

 Zamboni's battery consisting of 2400 pairs of plates, the other 

 pole of which was led to earth. The double deflection pro- 

 duced by a normal daniell varied 4 or 5 scale-divisions during 



* Liel). A/utAxsx. \\ 1 (1851) \- xc. p. J ( L864). 



