﻿Distribution of Particles in Colloidal Suspensions. 643 



v then to take the mean of these counts. The volume to which 

 these counts correspond is governed by the diameter of the 

 circular field and by the depth of focus of the objective, both 

 of which were determined ; its value was 2*1 x 10 ~ 7 cm. 3 . 

 The depths were measured (by means of the micrometer 

 screw which moved the cell) from an arbitrary level, which 

 was found to be about -021) cm. from the surface. The 

 depth reckoned from the surface we denote by ?/, the number 

 of particles per unit volume (i. e., the numerical concentration) 

 by n. 



Several complete sets of observations were taken in order 

 to practice the method. The results for the final sets are 

 given below : — 



y- 



023. 



033. 043. 



063. 



083. |- 103. 



123. 



143. 



Number of counts 



•20 



20 20 



20 



40 



20 



20 



40 







Total number of particles counted 



4 



9 



15 



30 



72 



40 



41 



82 



Number per count 



•20 



•45 



•75 



1-50 



1-825 



200 



2-05 



fl-flfi 







7iXl0~ 6 



•95 



?-14 



3-fil 



7-14 



8-69 



9-52 



9-76 



976 















Professor Perrin's calculations are explicitly based upon 

 the assumption of the application of the laws of perfect 

 gases to dilute solutions. 



It is easy formally to extend them to solutions of any 

 concentration. This was done by one of us for true solutions 

 in 1917 *. 



Fig. 1. 



St>Vol 



Sj>Vol 



\ * N 



^ 



dK^qtvHf* 



ooWffo 



Salvenr 



Imagine a column of solution to be put into connexion 

 with a column of pure solvent at two points through semi- 

 permeable membranes, the difference of depths of these points 



* Porter, Faraday Society. Discussion on Osmostic Pressure, 1917. 



2T2 



