1859-1863] NATURAL SELECTION 127 



which has thus varied ; together with the tendency to its Letter 79 

 inheritance." Any variation, which was of no use whatever 

 to the individual, would not be preserved by this process of 

 "natural selection." But I will not weary you by going on, 

 as I do not suppose I could make my meaning clearer without 

 large expansion. I will only add one other sentence : several 

 varieties of sheep have been turned out together on the 

 Cumberland mountains, and one particular breed is found to 

 succeed so much better than all the others that it fairly starves 

 the others to death. I should here say that natural selection 

 picks out this breed, and would tend to improve it, or 



aboriginally to have formed it 



You speak of species not having any material base to rest on, 

 but is this any greater hardship than deciding what deserves 

 to be called a variety, and be designated by a Greek letter ? 

 When I was at systematic work ,1 know I longed to have no 

 other difficulty (great enough) than deciding whether the form 

 was distinct enough to deserve a name, and not to be haunted 

 with undefined and unanswerable questions whether it was 

 a true species. What a jump it is from a well-marked variety, 

 produced by natural cause, to a species produced by the 

 separate act of the hand of God ! But I am running on 

 foolishly. By the way, I met the other day Phillips, the 

 palaeontologist, and he asked me, " How do you define a 

 species?' I answered, "I cannot." Whereupon he said, 

 " At last I have found out the only true definition, any form 

 which has ever had a specific name ! " . . . 



To C. Lyell. Letter 80 



Ilkley, Oct. 3ist [1859]. 



That you may not misunderstand how far I go with 

 Pallas and his many disciples I should like to add that, 

 though I believe that our domestic dogs have descended from 

 several wild forms, and though I must think that the sterility, 

 which they would probably have evinced, if crossed before 

 being domesticated, has been eliminated, yet I go but a very 

 little way with Pallas & Co. 1 in their belief in the importance 



" With our domesticated animals, the various races when crossed 

 together are quite fertile ; yet in many cases they are descended from 

 two or more wild species. From this fact we must conclude either that 

 the aboriginal parent-species at first produced perfectly fertile hybrids, or 



