the Electricity of Flame. 165 



Hence, in order to avoid possible errors, it appeared advi- 

 sable to cover as much of the electrodes as was not within the 

 flame with an insulating coat, which could easily be done by 

 fusing the platinum wires into glass tubes. The wire project- 

 ing out of the glass was just long enough to reach from one 

 margin of the flame to the other. 



If, now, two such platinum wires were placed opposite one 

 another in one and the same cross section of the flame, and one 

 of them was continually moved further and further from the 

 other, with its complete withdrawal from the flame a very con- 

 siderable increase of the electromotive force occurred ; it rose 

 from 12 to 192 ; and the electrode which was in the layer of 

 hot air was positive. 



Accordingly, a flame gives the maximum of action when 

 the apes-electrode is introduced into the hottest part of the 

 flame, and the base-electrode into the sensitive hot layer of air, 

 about as represented in fig. 2 a. 



Let it be further remarked that with this arrangement of 

 the experiment a reversal of the polarity of the flame was never 

 observed by us, and that all carburettecl-hydrogen flames ex- 

 hibited the same behaviour. 



§ 5. On the Change of the Polarity of a Flame. 



It was mentioned in § 3 that one and the same flame appears 

 sometimes positive, sometimes negative. Since, then, the 

 wire introduced into the air envelope becomes highly positively 

 electric, it is clear that uncommonly much will depend on how 

 the electrodes are introduced into the flame. If it is wished to 

 have the apex of the flame negative, the apex-electrode must 

 be completely enwrapped in the flame ; in the other case, 

 especially if the base-electrode dips quite into the flame, the 

 electric excitation of the enveloping air stratum may prepon- 

 derate, and consecmently the flame appear positive. If this is 

 the true explanation, it must be possible artificially to change 

 the polarization of a flame. 



Of the experiments with alcohol-flames we note the follow- 

 ing ; the positions of the platinum electrodes are given in 

 fig. 2 a, b, c, d, e. In fig. 2 a the base-electrode B is in air at 

 about J millim. distance from the margin of the flame, and 

 will now be gradually pushed in till (fig. 2 c) it touches both 

 margins of the flame. The apex-electrode S has had this 

 same position from the beginning. The electrode B was then 

 left in this situation and S gradually drawn out of the flame 

 until it was all in air (fig. 2 e). The deflections were as fol- 

 lows: — 



