compsognathus and his cousin the archæopteryx. One of these first sketches, the ichthyornis, has a row of teeth in each jaw, and displays another strikingly early reptilian or fish-like peculiarity in the joints of its backbone, which are cup-shaped or hollow on either side, exactly like those of a cod. This strange bird must have resembled an emu in many respects, and it might easily have devoured the large ganoid fish of this period with its formidable jaws. Still more reptilian in some particulars is the hesperornis, also found in the Western American chalk. Hesperornis was a huge swimming ostrich, and it had pointed teeth like a crocodile's, set in a groove running down the jawbone. They were supported on stout fangs, in the same way as the teeth of its reptilian allies, the mosasaurians. Like the ostrich, hesperornis had a broad breast-bone, but this breast-bone was destitute of a keel, as is still the case in all the ostrich family. The wings were also very imperfect, like those of the cassowaries. In its tail, hesperornis resembled its predecessor, archæopteryx, so far as regards the lizard-like separateness of the vertebræ, except at the extreme end, where they were slightly massed together into the first resemblance of a plowshare-bone, such as the one I hold in my hand. Thus these two intermediate birds of the chalk period, though slightly more bird-like than their cousins of the oölitic age, still retained, each in its own way, many unmistakable relics of their descent from reptilian or almost amphibian ancestors. As usual, the further back we go, the more do we find all the lines converging toward a common center.
The primitive teeth died slowly and gradually out as time went on. In the still later eocene deposits of the London clay in the Isle of Sheppey, we find the remains of a true bird, known as odontopteryx, in which the teeth have entirely coalesced with the beak, and have assumed the form of bony projections. Strict biologists will tell us that these projections are not teeth at all, because true teeth are not bony in structure, and are developed from the skin of the gums. But such hair-splitting distinctions are of little value from the evolutionary point of view; the really important fact to observe is this, that while hesperornis has teeth in a groove, reptile-fashion, ichthyornis has teeth in distinct sockets, mammal-fashion, and odontopteryx has them reduced to bony projections from the bill, in a fashion all its own, thus leading the way to modern birds, in which the teeth are wholly wanting and the bill alone remains. Indeed, among our existing kinds there are some which still keep up some dim memory of the odontopteryx stage; for the merganser, a swimming fish-eating bird, has bony ridges on its bill, which help it to grasp its prey; and the South American leaf-cutter has a double set of bony bosses on its beak and palate.
The most apparently distinctive feature of birds lies in the fact that they fly. It is this that gives them their feathers, their wings, and their peculiar bony structure. And yet, truism as such a state-
