20 THE CLASSIFICATION OF BIRDS 



the .ZEgithomorphas and the Coraciomorphas be regarded as de- 

 scended from arboreal birds, the Pelargomorphae and the Sphoe- 

 niscomorphfB may be regarded as descended from aquatic birds. 

 It were, perhaps, idle to speculate upon the precise sequence of 

 events, but on the principle that it is very much easier indepen- 

 dently to lose a character than independently to acquire one, we 

 may hazard the guess that the Sphaaniscomorphas and the Pelar- 

 gomorphge acquired webbed feet before they were isolated from 

 each other, and that the webbed feet of the Anseres, the Tubi- 

 nares, the Pygopodes, and the G aviso date back prior to the 

 isolation of the ancestors of the Ciconiiformes, the Falconiformes, 

 the Ralliformes, and the Charadriiformes from each other. 



The Class Aves may be divided into five Subclasses, each of 

 which is supposed to represent the surviving descendants of 

 an ancestral species from which the various species contained in 

 it have inherited certain characters by which they may be col- 

 lectively diagnosed. There were probably other contemporary 

 species, whose descendants were unsuccessful in the struggle for 

 life, whose numbers gradually decreased until they became extinct. 



The five subclasses of existing Birds may be diagnosed in 

 a fairly satisfactory manner. 



There are two very small groups of Birds which have already 

 been mentioned that possess so many unique characters that it 

 seems probable that their isolation dates far enough back to 

 entitle them to rank as subclasses. The Penguins and the 

 Ostrich and its allies can be diagnosed by so many characters 

 that we may venture to place them at the two extremes of our 

 classification the former under the name of Sphaeniscomorphge 

 and the latter under that of Dromaaomorphas. It is quite true 

 that the Spkwiiiscidcc and the Alcidce have very many characters 

 in common, and that in some remarkable points the Struthiones 

 resemble the Crypturi; but some of the characters in which one 

 group resembles the other may in both cases have been acquired 

 by inheritance from very remote ancestors, whilst others may 

 have been independently acquired by the" influence, direct or 

 indirect, of almost identical environments. 



Having thus disposed of the Sphasniscomorphna and the 

 Dromaaomorphaa, the rest of existing birds may be divided into 

 three subclasses which may possibly be more nearly allied to each 

 other than they are to the two subclasses which we have already 

 mentioned. Be that as it may, they are at least more simple in 



