1 8 Larmor, Physical Aspect of the Atomic Theo}y. 



becomes increasingly difficult to resist the direct evidence 

 for the simple view that, in many cases, chemical com- 

 bination is not so much a fusion or intermingling of the 

 combining atomic structures, as rather an arrangement 

 of them alongside each other under steady cohesive 

 affinity, the properties of each being somewhat modified, 

 though not essentially, by the attachment of the others ; 

 and that the space formulae of chemistry have therefore 

 more than analogical significance. The many instances, 

 thermal capacity, refractive index, etc., in which the 

 physical properties of the compound molecule can be 

 calculated additively with tolerable approximation from 

 those of its constituent atoms, are difficult to explain 

 otherwise. The crystallographic evidence has already 

 been referred to. 



The Spectrum. 



Yet the spectrum, which the physicist is accustomed to 

 regard as the most complete (though largely undeciphered) 

 index of the structure of the molecule, is totally different, 

 at any rate in the simpler combinations as compared 

 with single atoms, unlimited groups of lines (forming 

 bands) taking the place for the molecule of the single 

 lines of atomic spectra. It may be permissible to believe 

 — it is now in fact widely accepted — that no stimulation 

 of an atom, less violent than complete disruption of some 

 molecule in which it exists, can suffice to excite sensibly 

 its atomic line spectrum. But there seems to be more corre- 

 spondence between the absorption spectra of complex 

 molecules and those of the molecules or radicles of which 

 they are built up. The difference is fundamental between 

 the firm, almost unalterable structures which are the atoms, 

 and the molecules, considered as intimate definite aggre- 

 gations of atoms capable of definite disruption ; it ought 



