Prof. Clausius on the Conduction of Heat by Gases. 533 

 calculated for 1 second ; we thus get 

 11 x 3600 _ I 

 4,000,000 ~ 100* 



This calculation leads therefore to a conducting-power for heat 

 which is 1400 times smaller than that of lead*. 



If this number can but lay claim to a small degree of accu- 

 racy, so that it can only be regarded as an approximation to the 

 truth, we may at least regard it as proved that the conducting- 

 power of gases for heat, which can be theoretically deduced from 

 the hypotheses respecting the molecular motions of gases which 

 lie at the foundation of this memoir, is much less than that of 

 the metals — a result which entirely accords with observation. 

 The objection that this hypothesis involved so rapid a distribu- 

 tion of heat that local differences of temperature within the gas 

 are impossible, is accordingly completely without foundation. 

 We may even quote, as a fresh argument in favour of this hypo- 

 thesis, the very phenomenon which has been urged with such 

 particular emphasis against it. 



§ 29. The expressions which have been found for K admit, 

 further, of an approximate comparison with each other of the 

 various gases in respect of their conducting-power for heat. 



In the expression (XVII.) the specific gravity a is sufficiently 

 well known ; and the specific heat 7 can be approximately calcu- 

 lated from the experiments of Regnault. If we consider apart 

 the simple gases and such compound gases as possess the same 

 volume as their constituents before combination, we may, as has 

 been said, assume that 7 is the same for all ; and the expression 

 for K thus assumes the form (XVIIL), which contains nothing 



but the fraction -—=. in addition to the numerical factor. 



V (7 



The quantity e, the mean length of excursion of the mole- 

 cules, is not necessarily equal for different gases, and we do not 

 know what proportion the lengths of excursion in various gases 

 hold to each other. Nevertheless there is no obvious reason to 



* Maxwell arrived at quite a different result, namely, that atmospheric 

 air conducts ten million times worse than copper. This was, however, 

 caused only by the occurrence of two oversights in his numerical calcula- 

 tions. In the first place, instead of Peclet's numbers, which express the 

 conducting-power of the metals in French measures, he employs numbers 

 calculated from them by Rankine (Manual of the Steam-engine, p. 259) 

 in order to express the conducting-power in English measures. These num- 

 bers, however, are not quite correct ; they still require to be multiplied by 

 0'4536, the ratio of the English pound to the kilogramme, in order to 

 make them correspond with Peclet's numbers. Maxwell has further em- 

 ployed the numbers which relate to one hour as the unit of time as though 

 thev were calculated for one second. 



