1530 
Thus we find finally: 
3M : 3M 
UT yp = —u? tt 
nim nim 
to 
INR 
if we call f=rrt the time during which the deviation A is obtained. 
If N represents the number of molecules contained in one gramme- 
molecule, we may also write: 
By all ste ots SRT. 
T 
N .n'm dra? .o.0o.N 
JAP = ee oO, (Gil 
$ 3. In this formula r is certainly the quantity of whose value 
least is known. If nevertheless we wish to make an estimation of 
its value, it is obvious that we must have recourse to the coefficients 
of friction, of conductivity of heat, and of diffusion. Properly speaking 
t is the time between two collisions of a molecule lying close to 
the surface of the Brownian particle, which can, therefore, collide 
besides with surrounding molecules, also with the particle. For 
liquid we shall, however, assume that this time does not differ 
much from that for a molecule entirely surrounded by the solvent. 
For the three coefficients mentioned above Maxwerr has drawn 
up the following formulae for gases: 
fs 1 1 
6 === Oph == = Oke 
a 9 
oO o 
1,6 Ge 
% = OG ostc,rt *) Ky ette ieee ea) 
3 3 
1 1 
De ES cae 
3 3 | 
in which g represents the density, s the mean velocity, and / the 
mean free path. He then determines / for the first two formulae 
from : 
1 
LE en a aS 
mony 2 
For the diffusion the value of / of course gets somewhat more 
intricate. 
If we now ask in how far this theory of Maxwerr, is also valid 
for liquids, we are of opinion that we are not allowed to apply 
equation (5). It seems to us that there are not so many objections 
to the. assumption of equations of the form (4), if only we assign 
the true signification to /. We shall, namely, understand by / the 
1) The factor 1,6 has not been introduced by MAXWELL, but by O. E. Maver. 
