254 BULLETIN 50, UNITED STATES NATIONAL MUSEUM. 



out the close relationship suggested by the external appearance. Even 

 the external characters, when closely examined, show man}^ points of 

 difference. Thus, the divisions of the tarsal envelope are essentially 

 different, the Paridse lacking the longitudinal space between the anterior 

 and outer postero-lateral plates; the basal phalanx of the middle toe 

 is united in the Paridas, almost if not quite to the whole length of the 

 basal phalanx of the outer toe, but only for about half this length in 

 the Corvidse; the feathers of the antrorse nasal tufts are different in 

 the two groups, being bristly throughout in the Corvidae, " with lateral 

 branches reaching to the very tip," while in the Paridte they are 

 "broader, with the shaft projecting considerably beyond the basal 

 portion, or the lateral branches are confined to the basal portion and 

 extended forward" (Baird), 



The family Corvidaj is so nearly cosmopolitan that only New Zealand 

 and portions of Polynesia are without representatives. The group is 

 most developed, however, in the northern hemisphere. America 

 possesses nearly half the genera and species of undoubted "' Corvidae 

 enumerated by Dr. Sharpe in his catalogue of the Corvidse in the 

 British Museum. 



A somewhat singular fact in connection with the distribution of this 

 family on the Western Hemisphere consists in the circumstance that 

 while members of the subfamil}^ Garrulinge extend from the northern 

 limit of forests almost to the southern extremit}^ of the temperate 

 districts of South America, no part of the continental portions of 

 tropical America being without its representatives, the subfamily 

 Corvinfe reaches its southern limit in the Greater Antilles and on the 

 highlands of Honduras, no peculiar species occurring south of Central 

 Mexico or Jamaica, the Greater Antilles lacking any representation of 

 the Garrulinre and the Lesser Antilles being without a single member 

 of either group. 



KEY TO THE GENERA OP COKVID^. 



a. Tail much shorter than wing, the latter long and jjointed, with primaries exceed- 

 ing longest secondaries by more than length of tarsus, the sixth, seventh, and 

 eighth longest. (Subfamily Corvinse.) ^ 



«The genera Strepera, Slruthidea, Picathartes, Glaucopis, Heteralocha, Creadion, 

 Palculia, Graculu^', Pyrrhocorax, Corcorax, and Podoces are excluded from the above 

 enumeration as being more or less doubtfully members of the family. 



&Dr. R. Bowdler Sharpe (Cat. Birds Brit. Mus., iii, 1877, 4) gives the subfamily 

 Corvinfe wider limits. He includes with the Corvinse, as here defined, the Garru 

 linje, separating only the Fregilina\ The latter are restricted in their range to the 

 Palaearctic Region and Australia, and comprises the genera Graculus Brisson ( =Fregi- 

 lus Cuvier), Pyrrhocorax Vieillot, Corcorax Lesson, and Podoces Fischer. As already 

 stated on page 253, the limits of the family are as yet by no means satisfactorily 

 defined and it may here be stated that the subfamily divisions are equally unsettled. 

 So far as the American forms are concerned, however, the groups Corvinge and Gar- 

 rulinse are sufficiently distinct, the only question being as to whether they should 

 rank as subfamilies or groups of inferior grade. 



