26 Larmor, Physical Aspect of the Atojiiic Theory. 



theory, which was emphasised by Maxwell in the seventies,* 

 seems to have lost none of its force in the years which 

 have since elapsed. The transmission of qualities from 

 ancestors to descendants in organic nature, now studied 

 under the name of heredity, is of far too subtle a character 

 to be managed by aggregations composed of discrete 

 atoms of matter as large as we know the Daltonian atoms 

 to be. The identity of the atoms is a safe foundation in 

 chemistry and physics ; but if an attempt is made to 

 carry it over into biology we get lost in two directions, 

 both in the enormous complexity of the organic chemical 

 molecule, and in the infinite delicacy of transmission of 

 characteristics in organic life, the latter being in sharp 

 contrast with the rapidity of the spread of its lower forms, 

 once inoculation has taken place under suitable conditions 

 as to food. We are entitled to conclude, not that there 

 is here any essential contradiction with the atomic theory, 

 but rather that the complexities of the phenomena tran- 

 scend our powers of mental analysis. Will they always 

 do so ? Every new physico-chemical explanation in 

 physiology is fresh evidence that the processes, so far as 

 they can be extricated, are all rational ; wherever the 

 complex of phenomena can be partially disentangled, we 

 find order. The only method of progress is to seize the 

 salient or large-scale manifestations of order as they 

 present themselves, and by correlating them and fitting 

 on new regularities thereby discovered, to go on improv- 

 ing the working scheme of representation. The model 

 is not erroneous, because it is incomplete ; explanation is 

 usually worth more than criticism : an imperfect physical 

 representation which has stood the test of substantial 

 prediction can be improved, but it is not often absolutely 

 refuted. Whatever be the world of reality behind and 



• "Encyc. Brit." Art, 'Atom' ; 'Collected Papers,' vol. ii., p. 461. 



