238 NEW CLASSES OF GALVANIC CONDUCTORS. 



lose nothing of its divergence; but it will be completely de-. 

 prived of it. the moment a direct communication between 

 the ground and the flame is established. 

 Both effects These two effects may be seen at once acting in combina- 

 atonee by°wo t * on » ^ preparing two perfectly insulated flames, and guid- 

 flaraes. ing into each one of the wires proceeding from the two ex- 



tremities of the pile. If the insulation be perfect in all 

 points, both the electrometers will indicate after a few se- 

 conds the same state of divergence, as if the poles were not in 

 contact with the flame. Now if one of the flames be made to 

 communicate with the ground, the electrometer of that pole 

 ■will immediately lose all its divergence, and the divergence 

 of the electrometer of the other pole will be a maximum. 

 The alternate contact of the two flames therefore produces 

 the same effect, as if we had immediately touched the extre- 

 mities of the pile itself. 

 Farther proo"s Lastly, that we may be fully convinced of flame being 



of the conduct- an exce l lent con ductor for all the effects of the pile, that do 



ang power of ... . 



flame. not depend on the closing of the circle, the following facts 



should be noticed. 



Bend the wire on the top of the electrometer, so that the 

 point shall terminate' in an insulated flame. Into the same 

 flame insert a wire from one of the poles. If now the oppo- 

 site pole be touched, the electrometer will receive a maxi- 

 mum of divergence corresponding to the case. If afterward 

 the electrometer itself be touched, the pole with which it 

 communicates through the medium of the flame will be dis- 

 charged. Lastly, by touching the flame, we shall discharge 

 at once both the electrometer of this pole, and the electro- 

 meter communicating with the flame. 

 Flame there- These facts prove to a demonstration, that the flame is far 



fore does not £ vom insulating the electric effects of the pile in the cases 

 insulate L-alva- . • : n _,, ....... , , 



frism, and con- indicated. I hey show, that with respect to tnese cases there 

 duct electricity j s no ground for admitting a galvanic fluid, which the flame 



insulates, in opposition to the electric fluid, of which it serves 



as a conductor. 

 Its conducting '■ In the following fact, however, we find an anomaly, which 

 tothat lf f m° r snows u>; » ^' l!at *- ne conducting power of the flame, however 

 tajs. perfect it has appeared to us in the preceding experiments, 



is nevertheless very inferior to that of metals, when the&e 



two 



