metallic Conductors and their Resistance to Electric Currents. 



15 



In these equations No. 40 is not included from the uncertainty of P. The 

 comparison with observation gives, 



Here also the conformity to observation is very satisfactory. 



If we divide the values of b by a (assuming a"= 46), we obtain 0.0017, 

 0.0013, and 0.0015, taking the mean of which we have for a platinum wire, 



A = a (1 + 0.0015 X T), (2) 



a being, as before, the resistance at 32°. 



It would be desirable to extend these experiments to other metals ; but I 

 could not find any determination of their expansions similar to that which 

 Dulong and Petit have given for platinum, except what those philosophers 

 have given for iron and copper. The first of these metals, however, I found 

 to oxidate so rapidly in the pyrometer, that I could get no consistent results. 

 It should for this purpose be surrounded by dry nitrogen. Copper also oxi- 

 dates ; but the film of oxide acts as a coating, and protects the interior, so that 

 its diameter does not change up to 900°. The wire used was of the same 

 diameter as the last platinum one, Jg. Each result is a mean of three closely 



When the current was increased to produce ignition, 7"= 919.8, A = 68. 

 but the zero changed, and the wire stretched till its diameter became ^V- 



