activating one. This effect is the mcre important as hitherto I have 
not succeeded in the usual way to demonstrate a favourable influence 
of zine. 
Laboratories for Microbiology and Organical Chemistry 
of the Technical University 
Delft, March 1913. 
Physics. — “On the law of partition of energy’. IL. By J. D. van 
DER Waars Jr. (Communicated hy Prof. J. D. van per W aars). 
$ 6. It is obvious that the chance that the value of one of the 
variables p or qg lies between specified limits cannot be represented 
by a normal trequency curve. If however we investigate a region 
of the spectrum, which is very narrow, but yet contains many 
elementary vibrations, then we find another probability curve than 
for one single elementary vibration. If the region is sufficiently 
small, then the radiation will appear to us to be homogeneous. 
Only an observation during a long time (i.e. very long compared 
with one period) will reveal the want of homogeneity by the 
increase and decrease of the amplitude in consequence of beats. In 
order to describe the momentaneous condition we can represent one 
elementary vibration by : 
2at ant 
a sin —— +- b cos —- 
ie if 
and the total vibration of the spectral region by : 
Ant n Aant 
(a) sin Za + (0b) cos pr 
In this expression the separate a’s and 6’s may have all kinds 
of values. The chance that they lie between specified limits is not 
represented by a normal frequency curve. But this does not detract 
from the fact that the chance for a specified value of (2 a) == A, 
is represented by a normal curve, at least if the sum contains a 
sufficiently great number of terms. 
Let us imagine that the decrease of the amplitude of the vibrators 
in consequence of the radiation has such a value, that they are 
perceptibly set vibrating by a great number of elementary vibrations 
whose period does not differ too much from the fundamental period 
of the vibrators, then Maxwera’s law will hold. for the chance that 
the velocity of a vibrating particle lies between specified limits. The 
mean energy of a linear vibrator is probably rightly represented by 
