1906.] on the Passage of Electricity through Liquids. 251 



coagulative powers of sulphates of univalent, divalent, and trivalent 

 metals the relative numbers 1 : 35 : 1023. 



As we saw above, Faraday's experiments show that a univalent ion 

 is associated with one natural unit of electricity, a divalent ion with 

 two such units, and a trivalent ion with three. 



Let us suppose that to effect the coagulation of a region of colloid 

 solution, it is necessary for a certain electric charge, equal in amount 

 to that on the colloid particles present and opposite in sign, to be 

 brought within the region. This can only be done by the chance 

 conjunction of ions which, in the absence of an external electric field, 

 must be supposed to be moving in irregular and changeable ways 

 throughout the liquid. If the chance of one ion entering the region 

 be represented by I/2;, that of two ions entering together will be the 

 product of these separate chances or Ijx', while the chance of the triple 

 event of these conjunctions will be l/x^. 



Now, to obtain an equal amount of electricity, we need the 

 presence of 2 trivalent ions, 3 divalent ions, or 6 univalent ions ; and, 

 if we work out the problem for solutions containing the same number 

 of molecules, we find that the relative coagulative powers of univalent, 

 divalent, and trivalent solutions should stand to each other in the 

 general approximate ratio of 



1 : p : p^ 



The value of the p depends on unknown quantities, such as the 

 effective radius of ionic action, and cannot, at present, be calculated 

 theoretically. But, by putting p equal to 32, it is easy to see that 

 the law of increase of coagulative power with valency is that we have 

 deduced ; for the theoretical numbers 



1 : -d'l : 1024 



agree well with Linder and Picton's mean results. 



[W. C. D. W.] 



