[Illustration:  FIG. 3.—­HESPERORNIS REGALIS (Marsh).]

But Hesperornis differs from all existing birds, and so far resembles reptiles, in one important particular—­it is provided with teeth.  The long jaws are armed with teeth which have curved crowns and thick roots (Fig. 4), and are not set in distinct sockets, but are lodged in a groove.  In possessing true teeth, the Hesperornis differs from every existing bird, and from every bird yet discovered in the tertiary formations, the tooth-like serrations of the jaws in the Odontopteryx of the London clay being mere processes of the bony substance of the jaws, and not teeth in the proper sense of the word.  In view of the characteristics of this bird we are therefore obliged to modify the definitions of the classes of birds and reptiles.  Before the discovery of Hesperornis, the definition of the class Aves based upon our knowledge of existing birds, might have been extended to all birds; it might have been said that the absence of teeth was characteristic of the class of birds; but the discovery of an animal which, in every part of its skeleton, closely agrees with existing birds, and yet possesses teeth, shows that there were ancient birds which, in respect of possessing teeth, approached reptiles more nearly than any existing bird does, and, to that extent, diminishes the hiatus between the two classes.

[Illustration:  FIG. 4.—­HESPERORNIS REGALIS (Marsh).

(Side and upper views of half the lower jaw; side and end views of a vertebra and a separate tooth.)]

The same formation has yielded another bird Ichthyornis (Fig. 5), which also possesses teeth; but the teeth are situated in distinct sockets, while those of Hesperornis are not so lodged.  The latter also has such very small, almost rudimentary, wings, that it must have been chiefly a swimmer and a diver, like a Penguin; while Ichthyornis has strong wings and no doubt possessed corresponding powers of flight. Ichthyornis also differed in the fact that its vertebrae have not the peculiar characters of the vertebrae of existing and of all known tertiary birds, but were concave at each end.  This discovery leads us to make a further modification in the definition of the group of birds, and to part with another of the characters by which almost all existing birds are distinguished from reptiles.

[Illustration:  FIG. 5.—­ICHTHYORNIS DISPAR (Marsh).

(Side and upper views of half the lower jaw; and side and end views of a vertebra.)]

Apart from the few fragmentary remains from the English greensand, to which I have referred, the mesozoic rocks, older than those in which Hesperornis and Ichthyornis have been discovered have afforded no certain evidence of birds, with the remarkable exception of the Solenhofen slates.  These so-called slates are composed of a fine grained calcareous mud which has hardened into lithographic stone, and in