46 OUTHOGENETIC EVOLUTION IN PIGEONS. 



ing forward. At the junction of the two halves of the crescent we have a broad 

 re-entrant angle on the distal side, which vanishes as we run up to the series and 

 sharpens into a minute notch as we go in the opposite direction. Throughout 

 this row the lower half of the crescent is a little wider than the upper half, and 

 this extra width (0.5 to 1 mm.) is added evenly to one side, namely, the proximal 

 concave side. The result is that the lower half preserves its symmetry with the 

 upper half on the distal side, while on the proximal side its curve is retained, but 

 advanced by 0.5 to 1 mm. beyond that of the upper half, thus producing a sharp 

 salient angle in the shaft-line, opposite the re-entrant angle of the distal side. 

 The lower half differs also from the upper in having a slightly stronger inclination. 

 This difference is quite small, but once recognized in this band it may readily be 

 detected in lesser and vanishing degrees in the preceding band. 



Now, every one of the peculiarities here pointed out the greater width of the 

 posterior band and of the lower half-elements in this band, the crescentic form of 

 the elements, the distal notch and the proximal angle, the inclination of the half- 

 elements increasing with their width all these peculiarities are so many transi- 

 tional phenomena, passing phases of a transformation still in progress. The 

 process, although always flowing onward in one direction, advances more rapidly 

 in the upper than in the lower part of the field, and hence the crest of the wave, 

 if such a metaphor be allowable, sweeps over the feathers of each row in regular 

 sequence from above downward. This accounts for the differences in width between 

 the upper and lower half-elements, for the approximations to the typical band 

 being closer in the upper than in the lower parts of the rows, for the obliquity of 

 the band elements, and for the gradual rectification of the latter as the transfor- 

 mation comes nearer and nearer to completion. 7 



Having seen that the first step in the anterior bar has peculiarities which are 

 partially intelligible when approached through the bands that lie in front of it, 

 it remains to examine its relations in the bar itself. 



In the second feather we find (pi. 17, first row), in the lower web, an oblique 

 mark, strongly band-like in appearance, but in several respects falling short of the 

 type to a slightly greater degree than the corresponding mark in the first feather. 

 Its inclination is 29, 6 greater than that of the first mark; its distance from the 

 apex of the feather is 2 mm. greater; and its outline is not so sharply defined, the 

 pigment spreading a little on either side. These differences, though small, are not 

 to be credited to chance, for they are never reversed or wiped out in fluctuations 

 from individual to individual. In the upper web is to be seen just a shadowy 

 indication of the counter-mark, barely stronger than in the first feather. The 

 proximal edge of this upper mark falls 2 mm. nearer the tip of the feather than the 

 lower mark, and thus we get a lack of symmetry of the same nature but more 



7 This change from an oblique to a transverse position, in which the rotation of the element follows the direction 

 of the hands of a watch when we face the right wing, or the contrary direction when we face the left wing, is a phe- 

 nomenon that would repay further study. Is the rotation towards a position of more perfect equilibrium, depending 

 upon the ontogenetic distribution of pigment in a symmetrical structure ? Why, then, in the fore part of the wing, 

 is the rotation carried beyond the vertical or transverse line, so that the half-elements lean slightly forward? And 

 why in so many birds, with essentially the same symmetry of feather, are the half-elements permanently inclined so 

 as to form V-shaped figures pointing sometimes backward, sometimes forward? These are only a few of the great 

 multitude of forms which take their departure from bars and bands. Feather structure may have much to do with 

 the primary phenomenon of transverse bands; but certainly it can not be responsible for ornamental figures in which 

 many feathers may cooperate, contributing very unlike parts to a single design, with a symmetry overriding that of 

 the single feathers, and following lines that regard only the form of the bird as a whole. 



