80 M. E. Wiedemann's Investigations on 



disturbance of the vibratory motion can occur in the place in 

 question ; at other times sudden changes of phase appear, by 

 which the vibrations are altered in a quite indeterminate 

 manner. Hence that difference of phases up to which inter- 

 ferences are still observable gives a measure for the time 

 during which a regular motion takes place at the luminous 

 point. The source of light is formed of the great number of 

 atoms or molecules which lie in the vicinity of that point. 

 Their internal vibratory motion will in every case remain un- 

 disturbed only so long as they do not come into one another's 

 sphere of action — that is, during the time, nearly, that elapses 

 between two collisions. But this period, for the different mo- 

 lecules of the same gas, lies between nil and infinity. A 

 preponderating number of them will within a certain very 

 short time experience no collisions. Therefore the rays emitted 

 at the beginning and end of this period from all these particles 

 may interfere, and only those issuing from the few others 

 illumine regularly the field of vision in the apparatus used. 

 Hence the interferences are sharp. But the longer the time 

 which elapses between the emission of the two interfering 

 rays, and the higher the interference-bands we observe, the 

 fewer molecules contribute to bring them about, the more mo- 

 lecules illumine the field of vision regularly, and the less sharp 

 do the bands become. Lastly, with phase-differences corre- 

 sponding to intervals of time greater than that which is ne- 

 cessary for passing through the mean path-length, they will 

 very quickly disappear, since most of the molecules collide 

 within this period and thus undergo irregular changes of 

 phase. 



There are measurements by Fizeau and Foucault, and more 

 recently by J. J. Miiller, respecting the higher interferences. 

 With sodium-light, interferences corresponding to a difference 

 of path of more than 50,000 wave-lengths were not to be seen ; 

 with hydrogen those of 20,000 were distinctly visible, and the 

 highest possible phase-difference was not yet reached. There- 

 fore, with sodium the vibrations must remain regular up to 

 50,000 double vibrations at the most — that is (since the sodium- 

 line corresponds to about 500 billion vibrations in a second), 

 during the time 



T== 50000__ 10 d 



SooTo 12 second ' 



or during a 10000-millionth of a second. We find the time r 

 between two collisions, if we divide the mean length of path 

 of the gas in question by the mean velocity. 



If (since we can only have to do with the order of magni- 



