the Electricity of Flame. 163 



the time of the investigation, and amounted in the mean to 

 112. To this normal daniell, putting its electromotive force 

 = 100, all the data given in the following communication are 

 referred. The scale was placed at 2 metres distance from the 

 mirror ; and the deflection left and right from the position of 

 rest was measured by means of a suitable turn- plate. 



In order to convey the electricity of the flame to the quad- 

 rants, of which one was, as usual, connected with the earth, 

 straight Avires were mostly employed, or electrodes from 

 liquids, which, fixed in suitable stands well insulated upon 

 cakes of resin, could be raised and lowered at pleasure. 



The experiments were made with well-insulated Bunsen 

 burners and with alcohol flames. The flames of the former 

 issuing from apertures of the usual width proved too flicker- 

 ing, and therefore the measurements too uncertain. On this 

 account a very small Bunsen burner was prepared from a glass 

 tube of 4 millim. width. Its upper extremity was surrounded 

 by a platinum sheath, in order to avoid colouring the flame 

 by the gradually heating glass. When one electrode [the 

 ''base-electrode "] (B, fig. 1) was immersed in the foot of the 

 flame, and the other in its apex [the " apex-electrode "] (S, 

 fig. 1), with the turn-plate in one position the foot of the flame 

 was connected with the earth, and the apex insulated ; with 

 the other position the reverse took place. 



§ 3. Longitudinal Polarization of the Flame. 



Hankel states that when one platinum wire is introduced 

 into the apex of a flame and one into its base, a galvanometer 

 indicates an electric current passing from above downwards. 

 From this it might be inferred that the flame is polarized 

 lengthwise. 



The corresponding experiment with the electrometer gives 

 apparently the same result ; but in this case two very strikino- 

 points are to be remarked. 



The first point is this : — -If the experiment be arranged as 

 represented in fig. 1, the apex mostly appears, as in Hankel's 

 experiment, negative to the base ; but very often, and appa- 

 rently without any external cause, the reverse takes place. 



The second is that when the metal from which the flame 

 issues is connected with that quadrant which is conducted to 

 earth, while the insulated electrode is introduced at different 

 heights s above the base, by suitably shiftiDg the insulated 

 electrode within a cross section the appearance of a constant 

 potential within the flame is easilv attained. 

 M2 



