CONCLUSION. 



187 



parts of the earth ; and it is to the fossil forms of these birds that we must 

 look eventually for the intermediate types between them and the less 

 specialized Mesozoic birds. 



For the present, at least, it seems advisable to regard the Odontornithes 

 as a sub-class, and to separate them into three orders, according to the 

 characters given below. These orders are all well marked, but evidently 

 are not of equal rank. Archceopteryx is clearly separated much more 

 widely from both Ichthyornis and Hesperornis than are these two genera 

 from each other. The free metacarpals and long tail of Archceopteryx are 

 significant characters. Gegenbaur and Morse have shown, however, that 

 young birds of existing species have the metacarpals separate, and this is 

 true for all these birds up to a certain age. Hence this character is of 

 less importance than the presence of true teeth, since in no recent birds, 

 young or old, have these been found. The length of tail is perhaps a char- 

 acter of more value, but even this is a variable feature in modern birds. 



Sub-class ODONTORNITHES (or Aves Dentate), Marsh. 

 Order Odontolc;e, Marsh. Odontotorjle, Marsh. Saurur^e, Haeckel. 



Genus Hesperornis, Marsh. 



Teeth in grooves. 

 Lower jaws separate. 

 Vertebrae saddle-shaped. 

 "Wings rudimentary. 

 Metacarpals wanting. 

 Sternum without keel. 

 Tail short. 



Ichthyornis, Marsh. 



Teeth in sockets. 

 Lower jaws separate 

 Vertebrae biconcave. 

 Wings large. 

 Metacarpals ankylosed 

 Sternum with keel. 

 Tail short. 



Arch(Bopteryx,'von Meyer. 



Teeth in — 

 Lower jaws 

 Vertebrae — 



Wings small. 

 Metacarpals separate. 



Sternum ? 



Tail longer than body. 



That the three oldest known birds should differ so widely from each 

 other points unmistakably to a great antiquity for the class. Archceopteryx, 

 'Hesperornis, and Ichthyornis, are all true birds, but the reptilian characters 

 they possess are convergent toward a more generalized type. No Triassic 

 birds are known, and hence we have no light on this stage of the develop- 

 ment of the class. They will doubtless be found, however, and, if we may 



