July 20, 18S3.] 



SCIENCE. 



77 



be rigorouslj' true at the same time, as appears 

 from the following considerations. The most 

 l)robablc distribution of the component molec- 

 ular velocities of a gas in equilibrium is the 

 same as that of errors of observation. This 

 distribution is l)rought about by fortuitous 

 molecular encounters, and its permanence is 

 insured by reason of them. But in case the 

 l)rogressive motion of a molecule gives rise to 

 radiations, those molecules whose velocities 

 are the greater arc the hotter, and consequent- 

 l}- radiate more heat to other molecules than 

 they receive from them. They therefore lose 

 part of their progressive energy before the next 

 encounter. The whole effect would be to 

 retard the motion of those molecules whose 

 kinetic energy is greater than the mean, and 

 accelerate those whose kinetic energy is less. 

 This would cause a constant interference with 

 the distribution of velocities according to the 

 law of probabilities; and the interference would, 

 so far as we are at present able to form an esti- 

 mate of its amount, be sufficient to cause the 

 kinetic energy of each molecule to approach 

 indefinitely near its mean value during the 

 time in which it describes a very small fraction 

 of the mean path between two successive 

 molecular encounters. If this is the case, the 

 kinetic energy of any molecule does not differ 

 for any appreciable time from its mean value, 

 and is in effect the same during the whole 

 path, so that there is no such distribution of 

 velocities as has been assumed. In case the 

 interference with the assumed law is not so 

 complete as this, it must apparently exert an 

 important influence upon the distribution of 

 velocities, especially in the ease of rarifled 

 gases, in which the encounters are compara- 

 tively infretiuent. 



Again : if the progressive motion of the mole- 

 cules can originate radiations consisting of 

 transverse vibrations, it would appear highly 

 imi)rol)able that their rotary motion should not 

 also do the same. But, as has been sliown in 

 a former paper,' the kinetic energy of transla- 

 tion differs from that of rotation for imperfect 

 gases ; and the temperature cannot be simply 

 proportional to the mean rotary energy, though 

 it might possibly be proportional to the sum of 

 the rotary and translatory energies combined. 



But aside from these difficulties, which may 

 serve to show the intrinsic improl)ability of the 

 supposition that the progressive motion of the 

 molecules originates radiations, we seem to 

 reach pretty decisive evidence against the sup- 

 position, when we consider the specific heats 



' An elti'inlon of the theorem of Ihe virlal, etc. — (5c. proc. 

 Ohiomech. in$t., March, l!i83.) See also SciENCii, I. 66. 



of solid bodies, or when we consider the nature 

 of the radiation itself as revealed b}' the spec- 

 troscope. 



The experimental law of Dulong and Petit, 

 and the analogous results of Neumann,' show 

 that in solid bodies we must consider the tem- 

 perature to be measured more nearly by the 

 energ\- of the atom than by that of the mole- 

 cule. Now, it is hardly supposable that the 

 translatory motion of a gaseous molecule fshould 

 originate radiations, while that of a solid should 

 not. We shall not, at this stage of the dis- 

 cussion, consider the spectroscopic evidence as 

 to the nature of the motions which originate 

 radiations, further than to notice that the char- 

 acteristic spectra of gases appear wholly inex- 

 plicable, on the supposition that the}- are 

 originated b}- translatory motions, with veloci- 

 ties distributed according to the law of proba- 

 l)ilities, or with velocities reduced b^- radiation 

 to au approximate equalitj-, as it has been 

 shown the}- might be ; for even the simplest 

 gases have spectra consisting of at least several 

 lines. 



If these reasons compel us to distrust the 

 supposition that radiations originate in the 

 progressive or rotary motions of the molecules, 

 does the supposition that radiations originate 

 in the vibrators motion, with respect to each 

 other, of the atoms in the molecule, afford a 

 better explanation of the facts? Such a mo- 

 tion, analogous to the clastic vibrations of a 

 bell or other sonorous bod}', might very readily, 

 perhaps, be shown, in case of a complex mole- 

 cule, to have such a relation to the molecular 

 encounters, and thus to the mean kinetic 

 energy of translation, that its energy would 

 be directly proportional to it for each given 

 gas. In case this were established, such vibra- 

 tions, considered as the physical cause of radia- 

 tions, would explain the phenomena of gases 

 as well as the supposition that they are due to 

 the progressive kinetic energ}' ; and they might 

 possibly be shown to explaiu those of solids 

 also. 



But there is at least one difficulty, in the waj' 

 of accepting this supposition, which seems in- 

 superable in the case of monatomic molecides ; 

 for, if radiations could only originate in the 

 vibrations of atoms with respect to each other 

 within the molecule, monatomic molecules 

 could not radiate heat at all. and could not 

 have a temperature. That this shouUl be true 

 is not only inconceivable, but contrary to the 

 known fact that monatomic mercury gas has 

 a perfectl}' ascertainable temperature : hence 



Exf'erimental-i>hynikt 



