1013 ; 
2vy 
Pet We, ate OD ese LANEN 
— Vs 
vy 
Now in measurements with ammonium chloride I actually found 
values of p departing as much as 50°/, from unity. In this p was 
always smaller than 1, which points to a position of the particle 
with its length in the direction of the electric force. 
For the particles with which I performed my experiments, equation 
(13) was always sufficientiy fulfilled. 
They were: a. electrically sprayed mercury (which will be dis- 
cussed later), 
b. ice oil, density 0.87, 
c. potassium mercury iodide, density 2.56. 
I used the lasi substance in order to get a length of radius lying 
between that of the mercury and oil particles. The time of falling, 
namely, depends besides on the radius, also on the density (equation 6). 
In the observations ¢, must not be too small, the measurements 
becoming too inaccurate in this case, and not too large, because 
then the deviations owing to the Brownian movement have too much 
influence and the number of times of falling required to determine 
Vy, then becomes very great. For this reason I could observe 
particles with smaller radius of a heavier substance than of a 
lighter one. 
For the measurement of A? I used only series for which at least 
100 times of displacement were observed. ScHRODINGER ') calculates 
the relative accuracy of the results, obtained in such a way, for an 
9 
analogous problem at LA when n represents the number of 
n 
elements of the series. In our case the accuracy will not differ much 
from this, 
The distribution of the times of displacement that is to be expected 
follows from equation (11). The chance that ¢ lies between ¢, and ¢, is : 
(2,)2 (2s)a , 
frou mec ade, — | ede, +... (14) 
(a) . (23); 
in none 
ti 
Ea is tare etc. 
1 
The distribution of the t’s is the same. 
1) E. SCHRÖDINGER, l.c. 
