655 



are least to be trusted. Then the influence of small differences of 

 temperature may viz. still be obserbed ; this influence disappears 

 after a long time. Also, because we can only determine theoreticallj 



300 400 



Fig. 1. 



the most probable number, the observations after a long time, which 

 deliver larger numbers, have more importance. 



It is not to be denied however, that there seems to be a systematic 

 error. Perhaps this may be sought in the fact of the particles not 

 being exactly equal, or in a slow change of the electrical double- 

 layers at the particles and the wall, in consequence of the numerous 

 collisions. Further observations of the behaviour of colloidal particles 

 with respect to a fixed wall, will have to decide whether this 

 explanation is tenable. 



A remarkable circumstance is, that the curves practically coincide 



