46 LarmOR, Physical Aspect of the Atomic TJieory. 



A, a. B, and a C, which would be necessary to the imme- 

 diate formation of an ABC; whereas, if ever formed, it 

 would be liable to the normal chance of dissociating by 

 collisions ; it would thus practically be non-existent in 

 the statistical sense. But if an intermediate combination 

 AB could exist, very transiently, though long enough to 

 cover a considerable fraction of the mean free path of the 

 molecules, this will readily be formed by ordinary binary 

 encounters of A and B, and another binary encounter of 

 AB with C will now form the triple compound ABC in 

 quantity. The cognate subject of the dynamics of gas 

 theory illustrates the point, in fact is closely implicated : 

 that theory proceeds by aid of statistics of encounters, 

 yet in its analysis triple and multiple encounters are left 

 aside as negligible in number. 



The principle thus suggested, that immediate mole- 

 cular combinations and dissociations are practically all 

 binary, may have a wider application. It would tend to 

 explain, as above indicated, how it is that in organic 

 chemistry only those types of molecules — perhaps very 

 few compared with what are in totality possible — have 

 any chance of being isolated amid the chances of natural 

 reactions, involving no control whatever of individual 

 molecules, which can of themselves divide into parts or 

 radicles of appreciable persistence, some of them replace- 

 able by other such blocks or parts without allowing time 

 for the dissolution of the whole. That science, in fact, 

 proceeds by searching out and classifying the ways in 

 which complex molecules may be thus definitely dis- 

 sected and reconstructed. 



There is involved, on this view, the proposition that 

 in the ultimate type of chemical interaction each molecule 

 is divided into two parts at most, which it may inter- 

 change with another molecule in the process of double 



