410 Prof. H. BufF on the Thermal Conductivity 



True, what we speak of as a vacuum is not absolutely free 

 from gas ; when compared with a perfect vacuum the diather- 

 mancy of hydrogen may possibly be less. This seems even 

 beyond doubt ; for it would be impossible to raise the tempe- 

 rature of hydrogen if this gas did not possess any absorptive 

 capacity for heat. 



The preceding remarks, as already stated, are not meant to 

 deny that hydrogen possesses a certain degree of thermal con- 

 ductivity, but to show that this conductivity is too small to 

 become distinctly apparent — i. e. beyond the limits of the 

 errors of observation in the experiments performed by Magnus 

 and repeated by myself. 



The numerical results given in Magnus's paper* exhibit, 

 therefore, not the differences in conductivity of the various 

 gases, but in diathermancy, greatly affected though by dis- 

 turbing influences. This is especially the case for air and 

 hydrogen, the figures of which are too high in comparison with 

 the average value found for vacuum. In this manner the 

 supposition must have originated, that hydrogen possesses a 

 conductivity similar to that of the metals. 



Proceeding from this supposed property of hydrogen, Mag- 

 nus thought he had found the key to the explanation of Grove's 

 experiment, that a platinum wire, through which the electric 

 current passes, loses its incandescence in hydrogen, though 

 not in air. But this phenomenon is just as easily explained 

 by the greater diathermancy of hydrogen than of air, more 

 especially if we take into consideration its behaviour in va- 

 cuum, which Grove seems not to have noticed. 



The greatest incandescence of the wire, omnibus paribus, 

 takes place in vacuum ; and the incandescence diminishes in a 

 striking manner when air is admitted ; on admission of hy- 

 drogen instead of air, the white incandescence sinks to a dull 

 red heat. 



The researches of Dulong and Petit on the laws of cooling 

 bodies t have established the fact that a heated body cools more 

 rapidly in gas than in vacuum, and that this rapidity varies for 

 different gases. In so mobile a medium as hydrogen, a very 

 rapid current may be formed without an essential diminution in 

 the loss of heat by radiation. The conditions for rapid thermal 

 motion are thus most favourable. True, for atmospheric air 

 the convection may also be great ; but the loss by radiation is 

 diminished ; the rapidity of thermal motion must therefore be 

 less than in hydrogen. Inasmuch as the diathermancy of the 

 air is less than that of hydrogen, this would of itself explain 



* Phil. Mag. vol. xxii. p. 10. 

 t Ann. Chim. Phys. t vii. 1817. 



