1 74 MM. Elster and Geitel on 



elements; consequently a number of them can be united to 

 form a "flame battery.'' 



§ 11. Thermoelectrical Behaviour of Platinum Wires 

 separated by a Stratum of Hot Air*. 



The phenomena above discussed can be naturally explained 

 both from the thermoelectric and the electric theory; but we 

 soon arrived at the conviction that, so long as experiment was 

 made on the flame itself, a decisive experiment for the one or 

 the other theory could not be instituted, especially if the 

 existence of a peculiar flame-electricity (hitherto excluded 

 from our considerations) were assumed in order to aid in ex- 

 plaining the phenomena. Hence it was necessary to consider 

 the matter from another point of view, and to discover a 

 method by aid of which wires at different temperatures in hot 

 air could be examined as to their respective electrical beha- 

 viour. Of course in this case the hot air would not proceed 

 from a flame, and therefore would not be mixed with the pro- 

 ducts of combustion. 



Starting from these views, we employed the apparatus repre- 

 sented in fig. 4. a b is a fine platinum wire stretched between 

 two copper wires x and y, which can be rendered incandes- 

 cent by a battery of two Bunsen elements B. At the point u, 

 its electricity, and with it also that of the battery, was con- 

 ducted to earth and connected with one pair of quadrants of 

 the electrometer. A second platinum wire, c, was connected 

 with the insulated pair of quadrants, and could be brought to 

 any degree of proximity to the wire a b. This movable wire 

 was placed so as to be as near as possible to the point u; if a 

 thermoelectric difference then arose from the incandescence of 

 the wire a b, it was necessarily announced by the electro- 

 meter. 



There is, however, in this experiment a source of error to 

 be mentioned. As it would be inadmissible, and even (with 

 precision) impossible, to place the wire c exactly opposite to 

 the point u, a difference of potential might also possibly arise 

 from the circumstance that the potential-difference of the 

 points u and v on the stretched wire a b traversed by the cur- 

 rent would be measured through the intervention of the con- 

 ductivity of the heated air. In order to be independent of 

 this, a turn-plate Wi was inserted in the circuit, by which the 

 direction of the current in the wire a b could be altered. If 



* The electric excitation here occurring is taken into consideration in 

 this place only so far as it is immediately connected with the electricity of 

 flame. The general treatment of this phenomenon i* reserved for a future 

 communication. 



