Tlie only assumption made which is unsupported by 

 experimental proof is that of the valency-values assigned to 

 the several elements under discussion. It is now generally 

 accepted that the valency of any given atom depends upon 

 the nature of the atoms with which it combines and is 

 consequently a variable property. 



Sulphur for example is divalent towards hydrogen, tetra- 

 valent towards chlorine and hexavalent towards fluorine. 

 Valency is not even a constant quantity towards the same 

 element, thus manganese forms MnCl 2 , Mn01 3 , MnCl 4 etc. 



No experimental evidence is available concerning the 

 valencies of the atoms of an element in combination with 

 atoms of the same element, and the assumption has been 

 made that these are the same as those which the elements 

 exhibit towards hydrogen &J& in their principal compounds, 

 and which are suggested by their positions in the periodic 

 table of the elements. Sulphur, selenium and tellurium 

 have been assumed to be divalent, arsenic and phosphorus 

 trivalent. Such an assumption does not contravert the 

 theory of variable valency. 



In the case of sulphur, selenium, and phosphorus, the 

 hypothesis would appear to offer a reasonable explanation 

 of the observed facts. In the case of tellurium and arsenic 

 the data are not sufficiently unimpeachable to justify an 

 extension of the argument to their case though it appears 

 probable that they will be found to follow the examples of 

 sulphur and phosphorus respectively. In the case of 

 carbon, silicon, and boron, our inability to visualize the 

 molecule prevents our speculating on these lines. 



A point of considerable interest in this connection is that 

 allotropic forms do not exist among the non-metallic 

 elements which are monovalent, (such as hydrogen and the 

 halogens), nor among those which are monatomic (helium, 



