1879.] 
L. Schwendler— On Electric Light Measurements. 
85 
The consumption of power per unit of measured or external light is an 
increasing function with the number of lights produced by a given current 
in a single circuit. 
Supposing, of course, always that the sum of the resistances of n arcs 
is equal to the resistance of one arc, and that the other resistance in the 
circuit, in which no light is produced, has remained constant throughout. 
If we had a material infinitely conducting, of infinite strength, and 
with a melting point at least as high as that of carbon, then surely the 
division of the light would be perfectly economical, up to any limit, inas¬ 
much as we might then use linear electrodes. 
In practice we can only try to approach this limit. Up to the present 
time, there appears to be no better material for electrodes than carbon, 
either natural or artificial. But this is no reason why an effort should not 
be made to try to find a material for electrodes more accommodating to 
the division of the electric light than even carbon. The above, limited 
strength, limited electric conductivity and limited melting point of the 
material of electrodes, constitutes only one of the difficulties, which stand 
in the way of an unlimited economical division of the electric light. 
A second cause is for instance the fact that in each arc an E. M. F. 
is established in opposition to the original E. M. F. and by no means to bo 
neglected against it. This secondary E. M. F. established in each arc, 
appears to be a function of the current which passes the arc, most likely pro¬ 
portional to that current. Hence, if for a given current passing one arc this 
secondary E. M. F. be e, then the same current through n arcs, successively 
connected, would produce an E. M. F. equal to n e. This secondary E. M. F. 
n e is to be subtracted from the original E. M. F., and internal resistance 
of the original E. M. F. plus resistance of leading wires having remained 
constant, we necessarily have to decrease the total resistance of the n arcs in 
order to work with the same current as before. This merely means a 
decrease of the total length of the n arcs, or which is the same, an increase 
of internal light or decrease of the measured or external light. A parallel 
connection of the n arcs with reference to the poles of the given original 
E. M. F. would certainly produce only one secondary E. M. F. instead 
of n, arid for this reason it might be better to use the parallel circuit for 
the division of the electric light. But there are other very important 
objections to this solution. In the first case, as can be easily shewn, 
the variation of one arc has a far greater influence on the variation of 
the others, in parallel, than in successive circuit. Further, the length 
of each arc must be made very much smaller, in parallel circuit than in 
consecutive circuit. 
Another reason against an unlimited economical division of the elec- 
