48 PHYLOGENY 



generally held by naturalists, that the rock pigeons must be regarded 

 as the ancestors of domestic races, the question was limited to the 

 point just stated. 



It was known that the two types interbreed freely, under domesti- 

 cation, and it had been reported that checkered pigeons sometimes 

 appeared as the offspring of two-barred pigeons. Moreover, Darwin 

 discovered that the checkers were homologous with the spots com- 

 posing the bars. As the main purpose was to show that variation 

 was present to any extent required for the origin of new species, 

 rather than to trace its course in any specific case, and as variation 

 was supposed to be multifarious, and progress to be guided by na- 

 tural selection of the "fittest," it is not strange that Darwin failed 

 to get the direction of variation, or to realize that in direction is given 

 the key to one of the fundamental laws of evolution. 



As the two color-patterns are alike in having a common element, 

 and differ chiefly in the number of elements, it was natural enough 

 to take the smaller number as the point of departure, and to view 

 tha larger number as "an extension of these marks to other parts of 

 the plumage." (Animals and Plants, vol. i, p. 225.) With the ances- 

 tral type thus determined, and a simple mode of variation pointed 

 out, Darwin could dismiss the problem with these words: "No import- 

 ance can be attached to this natural variation in the plumage." 



Whence and how the two bars arose was not explained. The mode 

 of departure assumed to account for the checkered variety would, 

 however, suggest that the bars themselves originated in the same 

 manner; that is, from one or two spots arising de novo, as chance- 

 variations, and the gradual extension of like spots in two rows of 

 feathers. The one or more original spots, according to the general 

 theory, would first appear as minute rudiments, and then be grad- 

 ually enlarged and intensified by the aid of natural selection, guided 

 by their utility as recognition marks. 



Such a mode of origin would presuppose a plain, uniform gray 

 ancestor, without any spots or bars in the wings, and would raise 

 many puzzling questions that would be beyond the reach of inves- 

 tigation. For example: Why two bars? Why at the posterior end 

 of the wing? Why do the spots taper backwards to a more or less 

 sharp point in the checkered variety, while presenting a nearly 

 square form in typical bars? Why should they have first extended 

 upward, or downward, and in two rather than any other number of 

 rows of feathers? If two rows of feathers were favored long enough 

 to establish the bars for ornamental or other purposes, what freak of 

 natural selection could have then interposed to turn a long- favored, 

 definitely directed extension into a diffuse general extension, and 

 thus to neutralize completely the very effects it was invoked to ex- 

 plain? 



