DOUBLE SALTS 159 



it possible to refer back the so-called ' molecular compounds ' 

 to actual linkings between quite definite atoms." An exami- 

 nation of the formulae of the double salts dealt with in the 

 course of this paper, will show that the majority of them support 

 Werner's hypothesis ; thus, for the double chloride of copper, 

 the double sulphates of the astrakanite type, the co-ordination 

 number is six. But at the same time there are undoubted 

 exceptions, such as FeCl3.3CsCl.H2O, for which the co- 

 ordination number would be seven; CuCl 2 . LiCl. 2H 2 0, for 

 which it would be five. It remains for future work to show 

 how far these and other exceptions can be accounted for, and 

 how far the hypothesis will prove deductively fruitful. At 

 the present time it certainly marks an interesting attempt at 

 simplification. 



