metallic Conductors and their Resistance to Electric Currents. 19 



is nearly insensible. If from the first we deduce by interpolation the currents 

 and resistances which correspond to the temperature of the vacuum series, we ob- 

 tain one in which radiation and conduction must be the same as in it ; and the 

 difference between the A C- for the same temperature is evidently the measure 

 of the cooling due to air. Comparing them with the corresponding values of T, 

 I find that they are as its first powers. The deficiency from the experiment 

 of Dulong and Petit proceeds, probably, from the air being heated. 



In the vacuum series A C- must be as the combined effects of radiation and 

 conduction. Omitting No. 15 as too low, if we divide AC- by T\ we obtain 

 the numbers 



No. 22 0.75 



21 0.67 



20 0.66 



19, 0.64 



18 0.73 



17, 0.54 



16 0.68 



And if we allow for the residual air and the conduction, we may assume the 

 radiation in this pyrometer to be as the square of the temperature. Hence, 

 f{y) = Gy+ I-y\ and as G is less than H.ar, the equation becomes, 



d'y 



-^^ = my--ny-p, 



in which it must be remembered that m, the quotient of the coefficient of radi- 

 ation by that of conduction, is constant for a given wire, but n and p vary with 

 the current. The integration of this is facilitated by considering, that the effect 

 of conduction must cease at a certain distance from the origin, beyond which y 

 is constant. Let this value of j^ be e, then we must have, 



me--ne=p, (3) 



substituting which, and writing n^ for 2me - n, and u for 6 -y,we have, 



d-u , 



^ = " (^' - "iw)- 



d2 



