GENESIS OF FAULT-BARS IN FEATHERS. 333 



stand for the same thing, and they are here classed as type 4. 

 A wild Cardinalis virginianus brought from Florida threw, at the 

 time of its first moult in captivity, feather-germs (Fig. 3) which 

 were deeply constricted ; when these expanded they presented 

 the extremely wavy appearance shown in PI. XIII., Fig. 9. The 

 complete history of these defects from the time of their appear- 

 ance about August 30 until October 15, 1906, was obtained. 

 This material was secured through the kindness of Dr. Strong. 

 The production of the extreme constrictions in the feathers of this 

 bird is doubtless to be associated with its captivity and confine- 

 ment in a cage. A longitudinal section of one of these feather- 

 germs is to be seen in PI. XII., Fig. 3, and the photo- 

 graph of an entire feather in PI. XIII., Fig. 9. The 

 longitudinal section shows at "a" an indentation of 

 the pulp cavity by the epidermal layers. 



It will be noted that in all of these defects there is 

 a weakening of the feather at successive levels in its 

 length ; or to state it more adequately, in all of these 

 four types certain parts of the feather-vane are ab- 

 sent, weakened; or modified. The three parts of the 

 feather — i. e., barbules, barbs, and shaft — are, how- 

 ever, unequally and rather differently affected. The ^ 

 barbules may be absent (type 1), or merely weakened 

 (type 3) ; the barbs may be absent (type 2), weakened, or kinked 

 (type 4) ; the shaft or rhachis may be constricted and weakened 

 (type 4). See text Fig. 1 and PI. XII., Fig. 3. It will be observed 

 too that all of the above defects — to which we apply the gen- 

 eral term fault-bars — are extended upon the transverse axis 

 of the feather. In one case I have found a weakening on one 

 side of the feather-vane extending across many barbs in such a 

 way as to produce a longitudinal fault-bar. Duerden 1 reports 

 this as a rare occurrence in ostrich plumes. This furnishes us 

 type 5 and is photographed in PI. XIV., Fig. 23. 



From what has been said, the relation between the first four 

 types of defects is apparent. The importance attaching to the 

 equivalence of types 1 and 3, however, merits particular notice. 

 That they are equivalents is certain. The evidence in part is, 



1 Loc . cit. 



