408 Mr. (jr. J. Stoney on Polarization Stress in Gases. 



into harmonics like the vibrations of a string, or else into 

 quasi-harmonics not to be distinguished from harmonics by 

 observation (see Donkin's 'Acoustics,' § 194), like the trans- 

 verse vibrations of an elastic rod — probably the former. On 

 the more probable supposition that they are true harmonics, 

 the periodic times have been determined with great precision 

 in some cases, notably in the cases of a motion within the 

 molecules of hydrogen, which gives rise to three of its spec- 

 tral lines, and a motion within the molecules of chlorochromic 

 anhydride, which gives rise to 105 of its spectral lines. In 

 hydrogen the motion is repeated as often as 2,280,000,000,000 

 times each second in every molecule, and in the vapour of 

 chlorochromic anhydride rather more than 800,000,000,000 

 times*. Such are the periodic times on the supposition that 

 the motions are resolvable into true harmonics ; and whether 

 the fact be that the components of the motions are harmonics 

 or quasi-harmonics, their periodic times are at all events quan- 

 tities of this order. The general presumption, therefore, is 

 that the periodic times within the molecules of other gases are 

 also quantities of this order. But it is not necessary for our 

 present purpose to establish this. The only circumstance re- 

 lating to these inner motions with which we are here directly 

 concerned is that the energy which is transferred from mole- 

 cule to molecule is employed partly in altering the velocities 

 with which the molecules travel about, and partly in altering 

 these internal motions and (perhaps) collocations, and that 

 the proportion of the energy which is employed in the former 

 way bears on the average a numerical ratio to the whole energy 

 transferred which can be determined experimentally (see Max- 

 well's ' Theory of Heat,' p. 299) and is denoted in the sequel 



10. We now proceed to determine limits between which 

 the thermal and mechanical properties of the gas must lie. 

 For this purpose let us imagine a tube of the kind described 



* The periodic times deduced from the observations are respectively 



T T 



— — and — — , r being the time that light takes to advance 1 millim. 

 76*18 zvU 



in vacuo. (See Phil. Mag. April 1871, p. 295, and July 1871, p. 45. In 

 the former paper read 0-013127714 for 013127714.) The first of these 

 determinations was made by the present author, and the second by 

 the present author in conjunction with Professor Emerson Reynolds, of 

 Dublin; but, before either of these determinations were made, Professor 

 Clifton, of Oxford, had mentioned at the Exeter Meeting of the British 

 Association in 18G9 that he had found two of the hydrogen lines (pro- 

 bably C and h) to be related harmonically. I am not aware that any re- 

 cord of this important observation has been published. 



