THE PROBLEM OF THE ORIGIN OF 8PECIKS. 41 



The orientation of the field is given in all essentials by simple inspect inn i.f the 

 wing, and hence I need not here go into the genealogy and the pn-eni .-tatu> of 

 the older character, beyond the incidental mention of one or two facts for tin- 

 sake of perspective. 



The elemental units of the two color-characters take, of course, the serial order 

 of the feathers bearing them, and so fall in transverse rows. The sequence in each 

 row is, again, that of the feathers, which run above downward. The sequence of 

 the rows, as wholes, runs from before backward, and this anlero-posterior sequence 

 coincides, as will be seen, with the direction of the transformation ,,f the color- 

 pattern. As this direction is that of axial development in the embryo, it seems 

 probable that it is determined by the same general laws in both cases. 



The pathway of transformation to be followed in the future i> already (ire- 

 delineated, in part, by transitional phases, and hence the nature and position of the 

 goal may be forecast. 



With the orientation indicated in a preliminary way, we may turn to the exami- 

 nation of the several plumage stages. The simplest stage for this species (Ocyphaps 

 lophotes) is to be seen in the first or juvenal feathers. It is here that earlier ances- 

 tral features have not been entirely obliterated ; and for the sake of a better per- 

 spective these may be briefly called to mind without citing evidences to be dealt 

 with elsewhere. The ancestral pattern is that of the typical turtle-doves (Turlur 

 turtur, pi. 2, Vol. II; wing, pi. 22, Vol. I) and T. orientalis (pi. 1, Vol. II, and text- 

 fig. 3, this volume). It is everywhere among pigeons, as among many other birds, 

 the point of departure in the evolution of color-patterns. The common element 

 of this uniformly mottled pattern consists of a black spot occupying the center of 

 the feather and a light border at the edge of the feather. The dark center has 

 become pale in the young crested pigeon, and is indicated only by the dotted area 

 in text-figure 10. The light edge is everywhere present, but it is also paler than the 

 same part in the ancestral pattern. 



To this pattern two quite distinct new patterns have here been superadded. 

 One of these consists of longitudinal black spots, or chequers, placed on the two 

 posterior rows of feathers. The other consists of transverse stripes or bands, seen 

 in all the anterior rows. In these longitudinal spots, and transverse bands we have 

 the two distinct characters to which we have looked forward. They are so different 

 that at first they appear to be quite independent of each other. 



The two rows of chequer-like spots will readily be recognized as homologues 

 of the two bars seen in rock-pigeons. Such spots are widely distributed among 

 both wild and tame species, and although they have many specific forms they have 

 everywhere the same origin from the dark central spots of the turtle-dove pattern. 

 There can, then, be no doubt as to the specific value of these spots. That the bands 

 are equally entitled to rank as specific marks will not, I believe, be questioned. 

 Much smaller differences would pass as specific, if the marks stood alone, each in 

 a species by itself. 



If one of two marks contrasting as sharply as these do can be transformed 

 gradually into the other, it is of no vital importance how we rank them. The proof 

 of transitional variation remains the same. It would be but a specious quibble to 

 maintain that two such different characters can not be regarded as specifically 

 distinct because they stand side by side in the same species. It can be seen readily 



