metallic Conductors, and their Resistance to Electric Currents. 13 



Before endeavouring to deduce the values of these constants for eacli 

 wire, it is to be remarked, that the resistances given in the precedin^r tables 

 are too small, and require to be corrected for the heating effect of the current 

 on the rheostat and resistance coils by which they were measured. The 

 thickness of the wire in the former, and the immersion of the others in alcojiol, 

 might seem to guard against this danger ; but, with powerful currents, both be- 

 come warm to the touch. If we assume the truth of (1), the resistance 

 measured is noiA, hut A (1 + iY), supposing all reduced to the freezing point. 

 Now the heat generated is as ^ x square of current ; and I have found by ex- 

 periment, that the temperature of a wire follows the same law under 100'. 

 Hence, for l+b't', we may write 1 -f c.^1 . C', and (1) becomes 



A=a + hT-c.A\C\ 



Each result furnishes an equation of condition, which may be grouped to- 

 gether, and either by minimum squares or ordinary elimination the values of 

 a, b, and c determined. 



If the twenty-two that belong to the wire ^ be thus combined, we have 

 the equations 



442.10 = a -(- J X 674.1 -ex 195855.0, 

 660.53 = a + i X 1447.9 -ex 1002800.0, 

 833.32 = a + bx 2153.8 - c x 3141477.0. 



Hence, 



a = 198.4 ; b = 0.3412 ; c = 0.00003182. 



Computing with these the apparent values of ^, and subtracting them from 



the observed, I have arranged the results according to T. 



