EBAPPBAEAKCE OF SPECIES. 185 



■would almost certainly inherit difiEerent characters from their dis- 

 tinct progenitors. For instance, it is just possible, if our fantail 

 pigeons were all destroyed, that fanciers, by striving during long 

 ages for the same object, might make a new breed hardly distin- 

 guishable from our present fantail ; but, if the parent roek-pigeon 

 were also destroyed, and in nature we have every reason to believe 

 that the parent-form will generally be supplanted and exterminated 

 by its improved ofispring, it is quite incredible that a fantail, iden- 

 tical with the existing breed, could be raised from any other species 

 of pigeon, or even from the other well-established races of the 

 domestic pigeon, for the newly-formed fantail would be almost 

 sure to inherit from its new progenitor some slight characteristic 

 differences." That the reasons here given satisfactorily explain 

 why, in the vast majority of instances, lost species should not reap- 

 pear, most naturalists will admit ; but that they do not in them- 

 selves sufficiently explain why such reappearance may not occasion- 

 ally occur, may be reasonably contended. Thus, in the case of the 

 birds referred to, it would by no means be unreasonable to suppose, 

 even if such instances are a decided rarity, that the parent rock- 

 pigeon, through some special adaptation to surrounding conditions, 

 might have long survived the generations of fantails that were 

 primarily derived from it, and that, at some future period, after a 

 process of selection, a second series of fantails might have been 

 produced practically identical with the first. The irregular dura- 

 tion in length of time of species is well known to geologists, and 

 its importance as a factor in evolution has been insisted upon by 

 Mr. Darwin. It is probably true, as Darwin affirms, that "the 

 parent-form will generally be supplanted and exterminated by its 

 improved offspring ; " but it does not necessarily follow that the 

 offspring will invariably prove to be of an improved stock, and, 

 where this is not the case, there would be nothing very surprising 

 in the survival of the parental form. The tenacity with which cer- 

 tain specific characters adhere to some of the older genera, espe- 

 cially of moUusks, is so great that paleontologists are frequently at 

 a loss to determine just what characters separate practically the 

 newest from the oldest species of a given genus. This is the case 

 with the Nautili, and with a number of the brachiopods. Now, 

 either we have here a retention through almost indefinite periods of 

 primary specific characters, or a reversion (or a series of reversions) 



