A fifth family, the Cariamidce, or Cariamas, is quite nearly related to 

 those mentioned, and has by some authors beea even included among 

 the FalconidcB ; but the degree of its relationship by no means justifies 

 that view of its affinity. The families Strigidce, Falconidce, Cathartidw, 

 Serpentariidm, and Cariamidce may jjossibly be eventually combined to 

 form an order; but whether this association would be a natural one is an 

 undecided question, beyond the province of this paper. 



Having defined the limits of the family Falconidce, it now remains to 

 treat of this alone ; the subdivisions of the family being the subject of 

 discussion. 



To the present time, the Falconidce have been divided into a greater 

 or less number of so-called " subfamilies", the number varying accord- 

 ing to the author ; those foundiilg their classification on jjurely exter- 

 nal characters finding it necessary to adopt a great many, and those 

 relying upou the anatomical characters carefully avoiding any sub- 

 division at all.* 



To review in this connection all the classifications of the family which 

 have been proposed up to the preseiat time would require far more space 

 than the limits of this memoir will allow ; each author, while following 

 a generally-recognized plan, having his own peculiar views regarding 

 certain details of arrangement. It will, therefore, suffice for the pres- 

 ent to give a mere outline of this generally-adopted plan, and supple- 

 ment n by the modified systtefiis of our latest and best authorities. 



The "subfamilies" usually recognized are the following : (1) " Falco- 

 nince", (2) "Milvince", (3) "Accipitrince^', (4) ^'Cireince^\ (5) ^^Buteonince'\ 

 (6) '^Aquilinw", and (7) '^Polyborince.^' Some authors add " Girewtince^' 

 and ^^PandionincB" ; while previous to the important discovery, made 

 by Professor Huxley, regarding the vultures before alluded to, this 

 supposed family was divided into the so-called subfamilies, (1) " Vul- 

 turincB^\ (2) " Oypinw", (3) ^^ NeopJironinm'", (4) "Oypaetince", and (5) "Gy- 

 poMeracincB'', — some authors grouping two or more of these in one, 

 others recognizing all. They are all, of course, typical Falconidce, thus 

 making a total of fourteen subfamilies into which this family has been 

 divided, when there are in reality but two. 



Mr. George Robert Gray, in his "Hand List of Birds in the Brit- 

 ish Museum ",t divides the J'afcoTOdcB into seven so-called subfamilies, 

 as follows: (!) Polyborinw, Lape., 1839 (=Polybori of the subfamily 

 Falconinw); [2) '^ Buteonince, Swains., 1837"; (3) "Aqiiilince, SwAiNS., 

 1837 "; (including Pandion !) ; (4) '■'■Falconince, Swains., 1837" (including 

 JSarpagusl); (5) " ikftZmwte, BONAP., 1838 " ; (6) ^^Accipitrince, SWAINS., 

 1837" (including Rerpetotheres and Micrastur, both groups of the sub- 

 family Falconidce) and (7) " Gircinw, Be." 



Messrs. Philip Lutley Sclater and Osbert Salvin, in their " Womencla- 

 tor Avium Neotropicalium^W divide the American members of the fam- 

 ily into the following "subfamilies": (1) ^'Pandionince'^ (2) '-Gircinw^', 

 (3) "Buteoninw'\ (4) "Accipitrinw", (5) ^'Falconince'' (6) ''■Milvinw'K (7) 

 ^'Herpetotherince", and (8) "Polyborinw^' 



Mr. E. Bowdler Sharpe, in his recently-published great-work upon 

 the duirnal '^Aocipitres'',^ employs a singularly inconsistent classification, 



*This has, perhaps, been mainly 'due, not to the difficulty of finding suiiEciently good char- 

 acters, but to the lact that the conslusions arriyej at were so opposed to views long estab- 

 lished by usage. ° 



t London, 1869, (vol. I^. 



t London, 1873, pp. 118-123. 



