ALUMINIUM HYDROXIDE BY ELECTROLYTES. 6 



the colloidal solutions that none pass through filter paper of 

 certain definite brand. But as Schulze pointed out more than 

 twenty years ago, all these methods must give different values, 

 because they are only fit, so to speak, to determine certain 

 definite points on the curves of coagulation. This is due to the 

 fact that coagulation comprises a whole series of changes, the 

 magnitude of suspended particles increasing continuously with 

 the increasing concentration of the active ion. At a certain 

 stage the average particles begin to reflect so much light that 

 the solution appears turbid. When the particles attain such 

 magnitude that they can no longer remain suspended in the 

 medium they separate out as coagulum. For complete precipita- 

 tion still greater concentration of the active ion is necessary. 

 If we could determine the average magnitude of the particles as 

 a function of the concentration of the ion concerned, a clear 

 idea of the phenomena might be obtained. Indeed it may not 

 be impossible to accomplish this by means of the ultramicroscope. 

 It will however be a very laborious task ; and the results, being 

 obscured by various unavoidable sources of errors, may not cor- 

 respond to the pains taken. But if we could find a specific 

 property of the colloidal solution which varied continuously with 

 the growing magnitude of the suspended particles, we might 

 employ it as an index of the degree of coagulation ; and by 

 measuring this property we might attain our object much more 

 readily. 



The inner friction or viscosity of colloidal solutions is evi- 

 dently a specific property corresponding to the foregoing descrip- 

 tion, and being in general quite easy of measurement, seems to 

 be particularly well fitted for the study of the problem under 

 discussion. In order to test this supposition the coagulation of 



