348 EXPERIMENTAL INVESTIGATIONS. 



such gases ; and the experiments of the latter show that the 

 effusion or escape of gases through a minute aperture in a 

 plate takes place with velocities inversely as the square root 

 of their specific gravities. In Graham's experiments, however, 

 when the escape took place through capillary tubes, the results 

 seemed subject to no ascertained law, though the compounds 

 of carbon with hydrogen passed through with greater facility 

 than other gases. 



The cooling effects of gases on the ignited wire are de- 

 cidedly not in any ratio with their specific gravities ; thus, 

 carbonic acid, on the one hand, and hydrogen, on the other, 

 produce greater cooling effects than atmospheric air ; and 

 olefiant gas, which closely approximates air, and is far removed 

 from hydrogen in specific gravity, much more nearly approxi- 

 mates hydrogen, and is far removed from air in its cooling 

 effect 



Upon the whole we may conclude, from the experiments 

 detailed in this paper, that the cooling effect of different gases, 

 or rather the difference in the cooling effect of hydrogen and 

 its compounds from that of other gases, is not due to differences 

 of specific heat ; it is not due to differences of specific gravity ; 

 it is not due to differences of conducting powers for electricity ; 

 it is not due to the character of hydrogen in relation to its 

 transmission of sound, noticed by Leslie, for reasons which I 

 have before given ; * it is not due to the same physical 

 characters of mobility which occasion one gas to escape from 

 a small aperture with greater facility than another ; but it 

 may be, and probably is, affected by the mobile or vibratory 

 character of the particles by which heat is more rapidly 

 abstracted. I at one time thought that the effect might have 

 relation to the combustible character of the gas, and that the 

 electro-negative gases were in respect to it contra-distinguished 

 from the electro-positive or neutral gases, but the experience 

 I have obtained from the experiments detailed here induces 

 me to abandon that supposition. 



I incline to think that, although influenced by the fluency 

 of the gas, the phenomenon is mainly due to a molecular action 



* Phil. Trans., 1847. 



