448 RIESS ON THE INCANDESCENCE OF 
follows, that—The power of the current of discharge necessary 
to preduce incandescence in a wire is independent of the length of 
the wire. 
By the aid of the last two paragraphs a clearer insight is ob- 
tained into the nature of the problem mentioned before (p. 444), 
the experimental solution of which (if we substitute for fusion 
the smaller electrical effect of incandescence) has occupied se- 
veral philosophers. 
Let © be the indication of a thermometer placed in the con- 
stant part of the connecting circuit, A, g, x the length, radius 
and retarding power of an interposed wire, g the quantity of 
electricity, and s the number of jars used in the experiments. 
As the indication in the thermometer is proportional to the tem- 
perature of the wire extended within it, we have, if a and 3d re- 
present constant quantities dependent upon the constant part 
of the connecting circuit, 
Let this discharge suffice to produce incandescence in the in- 
cluded wire, the length of which is J, it is required to find the 
length of wire /', which would be brought to the incandescent 
state by the quantity of electricity nq collected in ~s number 
of jars. The indication of the thermometer must necessarily be 
the same as before; we have therefore 
and hence 
2 
N=nr + (n—1) 2. 
Two jars charged to any amount will therefore render incan- 
descent a wire more than double the length of that similarly 
affected by one jar; how much more than double the length 
cannot be determined for all cases; for the length a’ depends 
upon the nature of the wire to be heated, and upon the magni- 
tude of the constant quantity 4, which varies with the condition 
of the connecting circuit. The relation sought approaches so 
much the nearer to the relation of the number of jars used in 
the experiment to each other (in this case therefore as 2 to 1), 
as the wire to be heated is thin, and the connecting circuit is 
