A NEW VIEW OF EVOLUTION 367 



into two sets bearing " pure " dominant determinants and 

 " pure " recessive determinants, but that the practical "purity " 

 is wrought out by a process of germinal selection. 



However this may be, the facts of Mendelism lead us to a 

 renewed confidence in the relative independence of unit char- 

 acters. It looks as if a unit character sometimes behaves like 

 a radicle in chemistry ; it can be replaced en bloc by another, 

 but it cannot compromise with that other. "The outlook," 

 as Bateson says, " is not very different from that which opened 

 in chemistry when definiteness began to be perceived in the 

 laws of chemical combination." 



A New Yiew of Evolution. As is well known, Darwin believed 

 that specific differences and adaptations were slowly brought 

 about by the consistent selection of small continuous variations 

 in a profitable direction. He did, indeed, recognise that large 

 discontinuous variations may suddenly arise, as in the case of 

 the short-legged Ancon sheep. He could not, however, lay 

 stress upon such occurrences, believing as he did that they were 

 of rare occurrence, and therefore very liable to be swamped by 

 intercrossing with the normal forms. 



Over and over again, both before and after Darwin, naturalists 

 had suggested that sudden emergences of new structures with 

 no small degree of completeness, brusque transitions from one 

 position of organic equilibrium to another, might be of evolution- 

 ary importance. We need only mention Etienne Geoffroy 

 Saint-Hilaire and Francis Gal ton. But the difficulty always 

 was, that these discontinuous variations seemed to be of rare 

 occurrence, and liable to be swamped. 



In 1894 Bateson showed in his Materials for the Study of 

 Variation that discontinuity in variation was a fairly common 

 phenomenon, and might, therefore, have played in the past an 

 important role in the origin of species (see Chapter III.). 



Similarly, Hugo de Vries showed in most convincing detail 

 that sudden discontinuous variations or mutations not infre- 



