76 'WHAT IS A SPl^Cll'.S :^ ' 



species, but. as it were, trembling on the ed^e of dis- 

 ruption, ever ready, by the development of [)ronounced 

 preferential mating or by the accumulated incidental 

 effects of isolation prolonged beyond a certain point, 

 to break up into distinct and separate species. 



e. Rtsii/is of A)tifiiial Selection : — These obvious 

 difficulties encountered by a mechanical adherence to 

 definition by Diagnosis naturally lead to the consideration 

 of the further difficulties presented b\' domestic races 

 of animals and plants. The wide structural differences 

 between the forms accumulated by human selection 

 greatly impressed Darwin. Thus h(} wrote to Hooker, 

 September 8, [1856]: — * By the way, I have been aston- 

 ished at the differences in the skeletons of domestic 

 rabbits. I showed some of the points to Waterhouse, 

 and asked him whether he could pretend that they were 

 not as great as between species, and he answered, " They 

 are a great deal more." How very odd that no zoologist 

 should ever have thought it worth while to look to the 

 real structure of varieties. . . .' ^ But if the differences 

 between many of our domestic breeds, and between 

 them and the nearest wild species, are, as is well known, 

 generic rather than specific, why do we not consider 

 such races to be of different species and genera ? Be- 

 cause of the criterion suggested by Lankester ; because 

 we have reason to believe in their descent from common 

 j)arents within the historic period ; because, in spite of 

 their wide differences, they are still syngamic. 



Adz'iDiiages of I he adniissioii tJiat coticliisions based on 



Diagnosis are provisional. 



What is the practical bearing of these criticisms upon 

 the definition of species by Diagnosis and Diagnosis alone. -^ 

 The systematist, confronted by his series of specimens in 

 a museum cannot do otherwise than arrancre them in 

 groups which he will describe and name as species. But 

 much would be gained if he admitted at the outset that 

 his conclusions are provisional, if he said with Dr. Karl 



' More LetUrs, vol. ii, p. 210, LeUer 543. 



I 



