402 Prof. H. BufF on the Thermal Conductivity 



be stronger in hydrogen than in other gases /^ Such difference 

 in the behaviour of air and hydrogen, apart from the greater 

 mobility of the latter, could not possibly exist, if the two gases 

 (equal volumes of which, at the same pressure, have the same 

 specific heat) possessed also equal thermal conductivity and 

 equal diathermancy. 



Designating the weight of v cub. centims. of a gas as usual 



by 



273 . vhd 



^ 1000.760(273 + 0' 

 where d is the weight of 1000 cub. centims. at 0° and 760 

 millims. pressure, h the tension, and t the temperature, then 

 the difference of weight, or the upward pressure of v cub. 

 centims. of this gas, which in the cooler medimn at t has risen 

 to the higher temperature t' , will be 



1000 . 760^73 + t 273 + W " 1000 . 760 * (273 + 0(273 + 

 This gives the quotient of the upward pressure through the 

 mass, and consequently the acceleration 



n^ 9{t'-t ) 

 21^ + t' 



If, therefore, different gases obtain equal accelerations for 

 the same difference of temperature t' — t, this ceases to be 

 the case as soon as a change in the difference of temperature 

 takes place. But this latter must always happen if the spe- 

 cific heats of any two gases, and more especially their capaci- 

 ties of absorbing or losing heat in contact with solid bodies, 

 are not the same. So far, then, there is no need of assuming 

 a metal-like conductivity of hydrogen, in order to explain 

 the more rapid abstraction of heat by this gas. It is sufficient 

 to know that, though a bad conductor, its conducting-power 

 is superior to that of other gases. 



Magnus's own experimental results appear to us much more 

 favourable for his hypothesis. The apparatus he employed 

 was essentially as follows * : — Onto a cylindrical vertically 

 placed vessel of thin glass 56 millims. in diameter and 160 

 millims. in height, there was fused another, open glass cylinder 

 of equal diameter, but of only 100 millims. height, in such a 

 manner as to separate both vessels by only a thin glass plate. 

 Through the cork which closes the lower opening of the cylin- 

 drical vessel pass two glass tubes provided with stopcocks, by 

 means of which the vessel may be filled with air or any other 

 gas at any desired pressure. About 50 millims. below the 

 separating plate is a tubulature for the insertion of a ther- 

 * Phil. Mag. [IV.] vol. xxii. p. 5, pi. i. 



