634 3PR0F. j. CossAr ewArt On thij! 



It is inconceivable that the smidl protoptile forerunners of the 

 wing-quills of Ducks were evolved from enlarged scales, which as 

 they increased in length became fimbriated. The more the 

 history of feathers is studied the more untenable becomes the 

 belief in their origin from scales. When special attention is 

 directed to the phases through which feathers pass during their 

 development, it becomes evident that in an attempt to trace their 

 origin we are not concerned with the wing-quills or other contour 

 feathers, but with the feather papillee, and still more with the 

 filaments, in which the nestling feathers are developed. It is 

 conceivable that by the appearance of new and dominant factors 

 in the germ-plasm, scale papillse, instead of developing into scales, 

 developed into feather filaments, but it is difficult to imagine 

 how scales by growing longer and splitting were transformed into 

 double-shafted nestling feathers. There is little chance of the 

 geological record ever throwing any light on the origin of either 

 feather papillae or feather filaments, hence we must make the 

 most of the embryological record. Thus far the embryological 

 record has afibrded no support to the view that scales in whole or 

 in part were concerned in the making of feathers. 



When birds first acquired feather filaments it is impossible to 

 say, but we may assume that when the remote ancestors of birds 

 and mammals became warm-blooded any outgrowths from the 

 skin which prevented a loss of heat would from the first count in 

 the struggle for existence. 



A hint as to the nature of the coat in the progenitors of 

 mammals we have from Armadillos, and an indication of the kind 

 of coat worn by the Pro-Aves is obtained by studying the deve- 

 lopment of nestling feathers, more especially of the pi-otoptiles of 

 feather- footed birds, The Six-banded Armadillo has an extensive 

 scaly covering, but only a sparse coat of hair ; birds, with few 

 exceptions, have during development a more or less complete 

 coat of filaments*, and pi'obably in Owls the foot-feathers still 

 bear the same relation to scales as in the primeval birds. In the 

 developing Armadillo a stage is eventually reached when haii'S 

 are seen projecting from under the developing scales or from the 

 skin between the scales. The number of hairs developed in con- 

 nection with scales varies. There are often groups of three hairs 

 projecting from under the tail scales of mammals ; in the Six- 

 banded Armadillo there are often three or four hairs to a scale, 

 but sometimes only one; in the small Argentine Armadillo Ghlmny- 

 dojjJwrus, as text-fig. 12 shows, fifteen hairs may be developed in 

 connection with one foot-scale. 



In Armadillos and other mammals having scales as well as 

 hairs there is no evidence that the hairs are ever developed from 

 scales. 



* All the birds I have examined, with the exception of the House Sparrow, had a 

 more or less complete coat of protoptiles — in the Rook a nestling coat is said to be 

 absent, but even in Rooks there are a few well-developed nestling feathers. 



