metallic Conductors and their Resistance to Electric Currents. 7 



approaches the fusing point of platinum the tension has almost ceased. The 

 front and back of the instrument are brass ; the sides ai'e glazed, but, except in 

 photometric experiments, the glass is covered inside with slips of bright tin, to 

 lessen the effect of radiation. The top is mahogany, to insulate the screw a ; 

 from its low conducting power this becomes very hot, and therefore exerts a 

 cooling power on the wire much less than what acts at its other end. 



To deduce the temperature from the corrected expansion, I have used the 

 expansibility of platinum given by Dulong and Petit* They assign 



1 



Mean absolute dilatation from 0' to lUO" Cent. = 

 from 0' to 300° Cent. 



37700' 

 1 



36300' 



The corresponding temperatures, being measured by an air thermometer, might 

 require a slight correction for the coefEcient of gaseous expansion, which was 

 Gay Lussac's ; but such refinement is needless in the present research. The 

 expansion-rate of the metal evidently increases with the temperature ; its law is 

 unknown, but we shall probably not err far by assuming 



e = a.t + ^.t-, 

 and the above values give 



0.0000088418 = a x 180' + /3 x (180°)^ 

 0.0000091828 = a x 180= + 3^ x (180')'; 



the degrees being Fahrenheit's, but their origin at the freezing point of 

 water. Hence 



a = log-' (4.68282) ; j3 = log-' (0.72118). 



But since e is the absolute increase, divided by the length, we have 



'~=t{l+ft)^^A 



/being^^log-' (6.03836), and j/ = 4-- 



This quadratic may be solved in each experiment, or its positive root tabu- 



* Annals of Philosophy, 1819. 



