upon the Electricity of Flame. 543 



was provided with a bellows, by means of which air could be 

 forced through five small openings into the space between the 

 outer cylinder which inclosed the wick and the surrounding 

 chimney. The lamp No. 3 was an old one, not capable of giving 

 out any great amount of heat. 



For the detection of the current a galvanometer was used, 

 which consisted of a pair of astatic needles, around which were 

 wound 16,454 feet of wire. 



The first electric current which I obtained from the flame, was 

 by uniting one end of the wire of the galvanometer with the 

 lamp No. 1 j while the other end, to which was attached a piece 

 of platinum foil, was held in an inclined position in the flame 

 above the chimney. The time of oscillation of the needles was 16", 

 and in the present case the action of the stream excited caused 

 a divergence of 1°. The direction of the current could be 

 altered by a commutator which was introduced into the circuit ; 

 and by properly managing this, the swing of the needle could 

 be so increased, that after a few reversions it reached an am- 

 plitude of 15° or 16°. In this way the action of weaker cur- 

 rents, such as those of small flames, or of the flames of alcohol 

 and sether, was rendered evident. It was, however, desirable to 

 obtain a greater angle than the above ; and for this purpose a 

 pair of needles was chosen, whose magnetism was so nearly ba- 

 lanced that the time of oscillation amounted to 45". Placing a 

 new wick in the lamp, and raising it so as to obtain the greatest 

 flame possible, a deflection of 20° was obtained. The current passed 

 ■moreover from the top to the bottom of the flame. Most of the 

 experiments have been made with the last-mentioned pair of 

 needles. 



The strength of the electric current depends upon the magni- 

 tude of the flame. When a piece of platinumfoil 0-23 millim. thick 

 was brought over the lamp so that a space of 8 millims. separated 

 it from the upper rim of the chimney, the flame being made so 

 great that the entire platinum was encompassed by it, an angle 

 of 20° was obtained. The platinum being allowed to remain in 

 the same position, and the lamp being lowered, the needle fell to 

 16° ; and as the flame was made still smaller, the needle descended 

 correspondingly. Although an increase of flame in the present 

 case caused an increase of the deflection from 16° to 20°, it by no 

 means follows that this was due to an increased electric tension 

 between the portion of the flame in contact with the platinum and 

 that in contact with the lamp. The same might be the result of 

 a diminution of the resistance within the flame. If even a decrease 

 of electric tension accompanied the enlargement of the flame, 

 still a more powerful stream would be exhibited if at the same 

 time the diminution of the resistance to conduction were suffi- 



202 



