1332 
The values for the constants & x and D ete., from which these 
ts have been calculated, are borrowed from the “Recueil de con- 
stantes Physiques” published in 1913 by the Société francaise de 
Physique. As much as possible the values have been taken at 
room temperature. 
We shall now examine what value r should have, if equation (3) is 
to agree with the observations. If we take some deviations of gamboge 
particles in water measured by PErrin and CHaUDESAIGUES *), and if we 
suppose that the radius of a water molecule is equal to that of a 
molecule of watervapour, we find the order 2.10~1!! for rt from 
equation (3). Hence we see that we obtain a better agreement when 
we calculate t from the coefficient of friction or of conductivity of 
heat than from the diffusion. The great difference in mobility of 
Brownian particles in water and in glycerine is in harmony with 
the great difference that rt; shows for these two substances. The tx 
and tp for glycerine does not show such great differences. 
§ 4. For gases the supposition which we introduced for liquids, 
that, namely, the / for the effective molecules which almost touch 
a suspended particle, is the same as for a molecule which is in the 
middle of the gas or the liquid, will certainly not agree with reality. 
Here we shall sooner be allowed to disregard the mutual collisions 
of the molecules, and only take the collisions of the molecules 
against the Brownian particle into account. 
Bearing in mind that the quantity g in equation (2) consists of 
terms proportional to u’, v’, and w’, the components of velocity of 
the molecules, we see that g will change for two causes, 1. the 
collisions of the molecules with the particle, 2. in consequence of 
the movement of a number of molecules away from the particle, 
which molecules will thus get outside the “effective layer’, whereas 
other molecules with other values of u’, v’. and w’ will enter 
this layer. 
We shall be justified in assuming that q has got an entirely new 
value, when '/, 2’ molecules, whose velocity was directed towards 
the particle, have struck against the particle, and */, 7’ molecules 
which possessed opposite velocity, have left the effective layer and 
have been replaced by others. Thus +r in equation (3) becomes 
the time in which the particle has met with ‘/, 7’ collisions. If we 
denote the time between two collisions for the particle by tv’, then 
we get t= '/, n’ x’, and 
1) Perrin and CHAUDESAIGUES. C. R. 146 and 147. 
