102 Evolution and Adaptation 



and a species, this source of doubt would not so perpetually 

 recur." 



The point here raised in regard to the systematic value of 

 the new forms is the question that first demands our attention. 

 We must exclude all those cases in which several original 

 species have been blended to make a new form, because the 

 results are too complicated to make use of at present. The 

 domesticated races of dogs appear to have had such a mul- 

 tiple origin, the origin of horses is in doubt ; but the domesti- 

 cated pigeons, ducks, rabbits, and fowls are supposed, by 

 Darwin, to have come each from one original wild species. 

 The great variety of the domestic pigeons gives perhaps the 

 most striking illustration of changes that have taken place 

 under domestication ; and Darwin lays great stress on the 

 evidence from this source. 



It seems probable in this case, (i) that all the different 

 races of pigeons have come from one original species; (2) that 

 the structural differences are in some respects as great as those 

 recognized by systematists as specifically distinct; (3) that the, 

 different races breed true to their kind; (4) that the result 

 has been reached mainly by selecting and isolating variations 

 that have appeared under domestication, and that probably 

 some, at least, of these variations were fluctuating ones. 



Does not this grant all that Darwin contends for ? In one 

 sense, yes ; in another, no ! The results appear to show that 

 by artificial selection of some kind a group of new forms may 

 be produced that in many respects resemble a natural family, 

 or a genus ; but if this is to be interpreted to mean that the 

 result is the same as that by which natural groups have 

 arisen, then I think that there are good reasons for dissenting 

 from such a conclusion. Moreover, we must not grant too 

 readily that the different races of pigeons have arisen by the 

 selection of fluctuating variations alone, for this is not estab- 

 lished with any great degree of probability by the evidence. 



In regard to the first point we find that one of the most 



