1014 
In fig. 3a and fig. 36 I show for two particles in how far the 
observed and the calculated distribution agree with each other. 
Fig. 3a refers to the mercury particle N°. 123, fig. 36 to the 
oil particle N°. 158. 
Fig. 3a. Fig. 3b. 
I haye obtained the curved lines by calculating the to be expected 
number of times of displacement between t=O and r= 0.5 by 
means of equation (14), starting from the measured value of 14, 
and by drawing this number as ordinate of the point t= 0.25. In 
the same way the ordinate of r—0.75 gives the number of times 
of displacement between + — 0.5 and 1.0 ete. 
The crosses give the corresponding values found from observation. 
5. I will now proceed to the discussion of the results. I used 
for this 13 series obtained with oil, 18 with potassium mercury 
lodide, and 14 with mercury. ‘Fer some series the time of fall, 
hence also the radius, proved the same, e.g. for N°. 152 and 153. 
Such series | have combined. Everything was recalculated to 17°C. 
For most experiments /= 1.87 10-3, L = 2.24 10-2. 
Table I (p. 1015) gives the results obtained with oil and mercury 
iodide, arranged in descending values of a. 
1 
I will first try to determine whether ——, hence also ¢4 , is pro- ~ 
nd, 
portional to a’, to ak or to a. 
The circles in tig. 4 (p. 1016) represent observations with oil, the 
crosses observations with potassium mercury iodide. For the present 
we shall leave the series with mercury out of consideration. 
