XXVI 



requirements of flight. It is, in fact, so much in the way, 

 that its suppression must have become an absolute necessity 

 for the bircVs comfort. I was much struck by this fact on 

 examining the wing of a dead Gijpaetns, where the carpal 

 covert is evidently out of place in the modern wing. On 

 expanding and closing the wing, as any one can verify in a 

 dead Hawk, the carpal covert is certainly not at home in 

 its present position, and folds over most awkwardly when 

 one closes the wing. 



When the metacarpal bones became fused in process of 

 time, instead of being separate as in the ancient bird, there 

 was obviously no room for the remiges and coverts of the 

 8rd and 4th digits to coexist with those of the second, as the 

 raetacarpo-digitals occupy all the available space to such an 

 extent that their position is very much cramped. 



My belief is therefore that in process of time the feathers 

 of the 3rd and 4th digits were forced back on to the ulna, 

 where we may still trace their presence. To make this clear 

 in the Plate, the feathers are represented in what I believe 

 to have been their original "^o^iixon, and are again represented 

 in their actual situation on the ulna. 



Thus the series 6-8, coloured green in the Plate, of 

 which the first flight-feather possesses a shorter covert, 

 must originally have formed an independent group. This 

 group was derived from a fourth digit, now expunged, the 

 previous existence of which is sufficiently indicated by the 

 shorter major covert — to judge from analogy of the 1st 

 nietaoarpo-digital major covert. 



The next group, preceding the one just dealt with, consists 

 of the remiges 1-5, and includes also the tiight-feather about 

 to become suppressed, and numbered 1', whose place is always 

 indicated by the permanent carpal remex, as we have seen. 

 Tlie number of the quills for this group is, therefore, now 

 six ; remiges 1-4 are coloured blue. The fact that the carpal 

 remex is sometimes situated on the base of the first metacarpal 

 bone could not affect its serial affinity with the group under 

 consideration, which is clearly shown by its shorter major 

 covert. It has therefore been marked x in our Plate (fig. 3), 



