450 RIESS ON THE INCANDESCENCE OF 
On comparing the changes in the thermometer with the radii 
of the incandescent wires, it is found that the former are in pro- 
portion to the squares of the squares of the latter. If we take 
as mean numbers 9394° for the first thermometer, and 1202° 
for the second, as the temperature at which a platina wire of 0°1 
line radius would just be incandescent, we obtain for the altera- 
tion in the thermometer ©, or for the power of the current of 
discharge, sufficient to bring a wire of r-tenths of a line in radius 
to incandescence, the equation 
= nea 
1202 
From which the following series results :— 
Power of current of discharge 
at point of incandescence. 
Wire Calculated. Observed. 
ay 10°0 20 
2. 17°9 20°2 
Be 43°9 43°0 
Se 5°6 5°8 
4. 8:0 8°1 
Dis 32°4 31°0 
The accordance between observation and calculation must be 
considered satisfactory, as to the other considerable sources of 
error in experiments of this kind, the difficulty is added of de- 
termining an equal degree of incandescence in wires of different 
thickness. 
The power of current of discharge in a battery which is re- 
quired to bring a wire to an incandescent state, is proportional 
to the fourth power of the radius of the wire. 
Incandescence in Wires of different Metals. 
The difficulty of determining a fixed degree of incandescence 
in different wires is considerably increased if the wires are com- 
posed of different metals. Not only does the colour of the metal 
and its greater or lesser affinity for oxygen render the observa- 
tion inaccurate, but another circumstance occurs of which men- 
tion will be made at p. 455. It is not easy with some metals 
to keep them solid and in a state of incandescence at the same 
time, which is indispensably necessary to our present purpose. 
The wires with which the following experiments were made were 
applied without being submitted to further chemical examina- 
