548 



Dr. J. C. Willis. On the Lack of 



by natural selection to this force, with the enormous destruction of seed that 

 goes on, is almost inconceivable, however. 



The Process of Evolution. — It follows from the above that, as there is no 

 adaptational selection, or selection of characters better suited to the struggle 

 for existence, the process of evolution in these families from the simpler to 

 the more complex forms cannot have been by the gradual accumulation of 

 infinitesimal variations, unless these can also be selected by the one 

 permanent cause acting upon the families, viz., their plagiotropism. Many 

 of their variations, too, are not dorsiventral, and plagiotropic life would 

 not be likely to select them. The more complex and modified forms 

 are in no way whatever superior to the simpler forms ; both live together, 

 and one shows no sign of exterminating the other. It is all but impossible, 

 however, to imagine that plagiotropic life could select infinitesimal variations, 

 and as I have shown in dealing with the endemics of Ceylon that mutation 

 is the only possible explanation of them, so here I propose that we accept 

 the theory of mutation as the only feasible explanation at present possible of 

 the facts with which we have to deal. Mutations not being liable to go back, 

 it is possible to accumulate them, unless any one of them should prove of very 

 serious disadvantage, in which case it would be eliminated by natural selection. 



Taking the evidence which is here brought forward with that which I 

 have set forth in other papers already quoted, added to the fundamental 

 work of de Vries, it seems to me that a good case is now fully made out for 

 mutation, and that the onus of proof is completely thrown upon the 

 infinitesimal variationists, while a very fair case is made out for mutation 

 without natural selection. Analogy of the conditions and phenomena which 

 we have been considering in this paper with those which occur throughout 

 the vegetable kingdom lead almost irresistibly to the conclusion that 

 mutation without natural selection must be a theory of general applicability, 

 although there is no reason to exclude natural selection from operation upon 

 a smaller scale than that for which it has hitherto received credit — chiefly, 

 be it remarked, it operates by destroying disadvantageous variations, unless 

 they are compensated by advantageous ones at the same time. This subject, 

 however, leads into great issues, which must be left for subsequent con- 

 sideration. A re'sume' of the arguments which have led me to take up the 

 position I hold in regard to this question is given on p. 208* of the third 

 edition of my ' Manual and Dictionary of the Flowering Plants and Ferns, 

 1908.' 



The Mechanism of the Process of Evolution. — So far the reasoning has been 

 straightforward, and we have come to the conclusion that there must have 

 been evolution by means of mutations, and without natural selection in the 



