[Illustration:  FIG. 6.—­BIRD.  ORNITHOSCELIDAN.  CROCODILE.

(The letters have the same signification in all the figures.  Il., Ilium; a, anterior end; b, posterior end; Is., ischium; Pb., pubis; T, tibia; F, fibula; As., astragalus; Ca., calcaneum; 1, distal portion of the tarsus; i., ii., iii., iv.; metatarsal bones.)]

On comparing the pelvis and hind limb of the ornithoscelidan with that of the crocodile, on the one side, and that of the bird, on the other (Fig. 6), it is obvious that it represents a middle term between the two.  The pelvic bones approach the form of those of the birds, and the direction of the pubis and ischium is nearly that which is characteristic of birds; the thigh bone, from the direction of its head, must have lain close to the body; the tibia has a great crest; and, immovably fitted on to its lower end, there is a pulley-shaped bone, like that of the bird, but remaining distinct.  The lower end of the fibula is much more slender, proportionally, than in the crocodile.  The metatarsal bones have such a form that they fit together immovably, though they do not enter into bony union; the third toe is, as in the bird, longest and strongest.  In fact, the ornithoscelidan limb is comparable to that of an unhatched chick.

Taking all these facts together, it is obvious that the view, which was entertained by Mantell and the probability of which was demonstrated by your own distinguished anatomist, Leidy, while much additional evidence in the same direction has been furnished by Professor Cope, that some of these animals may have walked upon their hind legs, as birds do, acquires great weight.  In fact, there can be no reasonable doubt that one of the smaller forms of the Ornithoscelida, Compsognathus, the almost entire skeleton of which has been discovered in the Solenhofen slates, was a bipedal animal.  The parts of this skeleton are somewhat twisted out of their natural relations, but the accompanying figure gives a just view of the general form of Compsognathus and of the proportions of its limbs; which, in some respects, are more completely bird-like than those of other Ornithoscelida.

[Illustration:  FIG. 7.—­RESTORATION OF COMPSOGNATHUS LONGIPES.]

We have had to stretch the definition of the class of birds so as to include birds with teeth and birds with paw-like fore-limbs and long tails.  There is no evidence that Compsognathus possessed feathers; but, if it did, it would be hard indeed to say whether it should be called a reptilian bird or an avian reptile.