the Electricity of Flame. 181 



ing to the above conception, a thermoelectrico-electrolytic 

 element composed of 



Cold platinum | Hot air + Hot air [ Flame-gases + 

 Flame-gases | Red-hot platinum. 



If we fix our attention upon the single members of this com- 

 bination, it is proved by the experiments that an electric exci- 

 tation takes place between cold platinum with hot air, on the 

 one hand, and incandescent platinum with flame-gases, on the 

 other ; while the question is still undecided whether an elec- 

 trical difference exists between the hot air and the flame-gases 

 even without wires or liquids being in contact with those gases. 

 This question is identical with that whether a proper electricity 

 does or does not belong to flame. 



In order to bring this point also to a decision, let us here 

 adduce a few more experiments, wliicli decidedly speak against 

 the existence of a proper electricity of flame. 



On the hypothesis that to the stratum of air A A' (fig. 1) 

 enveloping the flame a certain quantum of positive electricity 

 is brought by the process of- combustion or by mere contact, 

 at least a partial equalization of the electricities must take 

 place, even if we take into consideration the bad conductivity 

 of the two strata of gas, as soon as one or more well-insulated 

 wires are passed ^transversely through the flame. But the 

 potential- difference existing between the electrodes S and B 

 (fig. 1) is not at all or only very slightly altered thereby. 

 Let E be the electromotive force without the transverse wire, 

 and Ed the electromotive force with it (platinum). 



One series of experiments gave 



E =169-0 164-8 168-4 mean 167'4, 

 E D = 162-0 161-2 , mean 161-6. 



consequently a diminution of about 3 per cent. 



A second series, when two very carefully cleaned and well- 

 insulated platinum wires were passed transversely through the 

 flame, gave: — 



E. Ed. E — Ed- 



Experiment I. 168-5 168-0 +0'5 



II. 158-4 158-9 -0-5 



III. 182-0 182-0 0-0 



Each of the above numbers is the mean of five readings ; 

 and the position of the electrodes was somewhat altered from 

 one experiment to another ; hence the difference in the num- 

 bers in the same column. Therefore a partial equalization of 



