2 WILLIS ; EVIDENCE AGAINST THE ORIGIN OF 



An argument that has already been used* in favour of this 

 theory of origin of species by mutation, and a point which has 

 always been a great difficulty upon the theory of natural 

 selection of infinitesimal variations, is the apparently useless 

 nature of the differences that divide so many species and varie- 

 ties. Thus, to take the simple instance that comes almost 

 first to hand in the flora of Western Europe, in the genus 

 Ranunculus, the species R. bulbosus differs from B. repens 

 mainly in the fact that the former has the sepals reflexed, 

 while the latter has them spreading. Now it is practicaUy 

 almost impossible to suppose that these differences can have any 

 important bearing upon the success in life of the two species. 

 Very similar results are to be attained by comparing almost 

 any species of any genus. Almost always the points of dis- 

 tinction are such as seem to have no importance in the life 

 of the species. The mere fact that the two Ranunculi above 

 quoted live in almost identical habitats, side by side, would of 

 itself seem to form a most cogent argument against the charac- 

 ters having arisen by natural selection of minute variations. 



When we consider the enormous amount of work that up to 

 about 1895 was put into the study of adaptations, the fact that 



The numbers are very fairly evenly grouped about the mean length. 

 Now it is generally supposed that evolution takes place by there existing 

 a great number of offspring showing similar variation in a character 

 that may be of value in the struggle for existence. Those that have it 

 best marked will survive, those that have it poorly marked will go imder, 

 and so generation by generation the character will increase, till at last 

 by progress in this and other characters we may get a plant so distinct 

 that it will be a new species. 



The other theory is that of mutation, on which the new characters 

 are supposed to arise at one step, as in fact they often do in the produc- 

 tion of new garden varieties. Once they have appeared the new 

 characters are hereditary, and the new form does not go back to the 

 old one. Evidence has been accumulating in favour of this view for 

 some time now, but that set forth in this paper seems to afford a final 

 argument for it. Be it specially noted that on the theory of natural 

 selection of the infinitesimal variations every part mvist be shown to 

 have some use, while on that of mutation this is not necessary, though of 

 course variations that are actually harmful may be killed out in the 

 struggle for existence. 



The position of matters at the present time is very fairly and lucidly 

 stated in a book, " Recent Progress in Variation, Heredity, and Evolu- 

 tion" (London, John Murray, 7s. 6d.) by R. H. Lock, formerly 

 Scientific Assistant at Peradeniya. 



* Willis, Flowering Plants and Ferns, 2nd. ed., p. 118 



