Conductivity of Metals with Temperature. 255 



tion (14) also reduces to the same curved =©«"'**, if we make 

 m infinite and a and 7 both zero, i. e. if we neglect variations 

 in conductivity and radiation-power ; for the numerator is 



then -7T-, and the denominator is indeterminate indeed, but 



finite. 



My brother, Mr. Alfred Lodge, of St. John's College, Ox- 

 ford "(to whom I am indebted for several suggestions) has 

 drawn for me the curve represented by equation (20) for some 

 arbitrary and rather extreme values of the constants, viz. 



\= — , say = 3QQ, © = 300, and ^= ^ ; that is, the curve 



300e~ so 



0: 



2 — e so 



and it is represented in Plate X., where, for comparison, is 

 drawn also the logarithmic curve 



0=300*"^ 



and also (by simply diminishing all the abscissaa in the ratio 

 3:2) the logarithmic curve which fits the correct curve best, 

 and which would have been assumed to be the correct one 

 from calculation of /j, from observed values of the temperature 

 by the ordinary formula, namely the curve 



0=3OOe~4 



The value of \x is therefore not obtained correctly, at least a3 

 regards its absolute value, by the old process ; and the nature 

 of the divergence between the new and the old curves is ap- 

 parent in the Plate. 



The Method of Calculating the Conductivity and its Variation- 

 coefficient by means of Equation (17) to the Curve of Tempe- 

 rature. 



22. The conductivity — is involved in the constant /ul; so 



that if one knows a, the absolute conductivity can be calculated. 

 For the meaning of /j, refer to equations (17), (15), (11), 

 (10), (9), (6), and (7), which show that it may be written in 

 the following equivalent ways : 



I* 2 = Rm*= f IV' log a = °p- . Pa w log a = 2c 'P'° . . ( 2 1) 



