Conductivity of Metals with Temperature, 259 



two ends being rendered as impervious to heat as possible by 

 means of a thickness of felt. Observations of its excess of 

 temperature are then to be taken at successive intervals of 

 time. If the excess of temperature 6 corresponds to the era 

 of reckoning time, and if the temperature 6 be the excess after 

 a lapse of time t, it follows, since 



riff 



_ =Pa v °(a 9 -l) 

 (see equation (5) and § 3), that 



lo S Jhi = log a^-L + IV0T l0g a ; 

 whence we obtain the constant required, 

 ^ „ 1, l-a- e o 



r ^ a l — a~ e 

 _logioa l-a-Qo 



The Absence of Experimental Results suitable for applying the 

 method of calculation to at present. 



28. To obtain experimental results suitable for applying 

 equation (25), one would have to observe temperatures very 

 accurately, over a considerable range of temperature, on a rod 

 in vacuo with a highly radiating surface. Principal Forbes's 

 experiments were all made in air; and, as he shows, convection 

 had a greater cooling effect than radiation, even on his paper- 

 covered rod. The only experiments which, so far as I know, 

 have been made in a vacuum, are those of Wiedemann and 

 Franz ; but the range of temperature observed by them only 

 went as high as 60° C. Moreover the surface of their rods 

 was silvered, and the loss of heat therefore so small that they 

 themselves did not regard these experiments as at all so satis- 

 factory as those which they made in air. 



Moreover the rods used were not very long, and their cool 



ends were connected up with the enclosure, which would 



doubtless make = at that end, but it would not secure 



d6 

 that -Y- should vanish at the same time : and this was of no 

 aw 



consequence in their method of calculation; but it is in ours 



(see § 17). 



Again, the temperatures, though very carefully observed, 



cannot, I imagine, be regarded as accurate even to the first 



decimal place. I have applied equation (27) to their numbers 



X2 



