of a penguin, where the bones show greater traces than in any other birds of their distinctness, and are furthermore shorter than is the rule. In dinosaurs generally those bones are separate, but not in Ceratosaurus, where the degree of fusion is almost exactly that of Aptenodytes.

Furthermore in some dinosaurs, as in birds, the intermedium is prolonged upwards as the ascending process of the astragalus.

The proportions of the long bones of the hind limb are distinctly birdlike in some dinosaurs. In Laosaurus, for example, the femur, as in birds, is shorter than the tibia, the reverse occurring in most forms. In the same animal and in others the fibula is commencing to degenerate; it is decidedly smaller than the tibia.

In their skulls the dinosaurs show no marked approximation to birds; there are nevertheless one or two features which may be remarked upon in this connection. The earliest known forms from the Trias have perhaps the most birdlike forms of skull. MARSH comments upon the lightness and avian appearance of the skull of Anchisaurus; it has moderately developed basipterygoid processes instead of those of such great length that are apt to characterise the dinosaurs. The great extension backwards of the premaxillaries in some dinosaurs is an avian characteristic; this is seen especially well in Diplodocus and Claosaurus. In the former animal, as in some others, the two vomers diverge widely posteriorly, as in many birds; and in the restoration of the under surface of the skull in this dinosaur the vomers have a very birdlike appearance.

Finally the height at which the transverse processes of the vertebræ are borne seems, as HUXLEY has pointed out, to suggest birdlike respiratory organs, while the hollowness of many of the bones in many dinosaurs points in the same direction. It is, however, undoubtedly in the pelvis and hind limb that the most striking likenesses to birds are shown by the dinosaurs. It has been attempted to put this down merely to bipedal progression. It may be said,' remarked Professor HUXLEY, that all birds stand upon

their hinder feet, and that, as the Ornithoscelida did the same, the resemblance of structure arises from a resemblance of function. But I doubt if the majority of the Dinosauria stood more habitually upon their hind limbs than kangaroos or jerboas do; and, unless there was some genetic connection between the two, I see no reason why the hind limbs of Ornithoscelida should resemble those of birds more than they resemble those of kangaroos.' In addition to this it may be pointed out that Hallopus, which appears to have been very probably a leaping dinosaur, has not the specially ornithic form of limb; it has large pubes and no postpubes.

A recent description by Mr. E. T. NEWTON of the skull, brain cast, and cast of the auditory organ in a pterodactyle, Scaphognathus Purdoni, shows certain most interesting resemblances between the pterosaurians and birds. It is possible that this pterosaurian, like Pteranodon, possessed a horny beak and no teeth. But the presence or absence of a beak or of teeth is no more distinctive of birds (cf. Archæopteryx) than of reptiles. The skull shows more positive points of likeness. In the first place, the bones of the pterodactyle cranium are early ankylosed and well ankylosed, being in this particular avian and not lacertilian. The large size and backward extension of the single premaxillary bone (=two fused premaxilla) agree with that of birds and contrasts with that of lizards. The palate shows also certain interesting resemblances, more especially to both emu and cassowary. As in the struthious birds and in lizards also the palatines are borne off from the middle line by the pterygoids; the latter bones, moreover, as in the emu, articulate at their posterior ends with both quadrate and basipterygoid processes. The vomer too is birdlike in being pushed backwards, owing to the extent of the premaxillæ, and in being thin, apparently single and bifurcate posteriorly. Other general resemblances in the skeleton are the development of air cavities in the bone, the large size of the orbit -which may, however, in the pterodactyles have had some

'On the Skull, Brain, and Auditory Organ of a New Species of Pterosaurian,' Phil. Trans. 1888, B, p. 503.

relation to nocturnal habits-and the presence of a keel upon the sternum.

It is possible, though far from certain, that Scaphognathus had not that characteristic reptilian bone, the os trans

versum.

Finally, the general shape of the scapula and the angle that it makes with the coracoid are birdlike in the pterodactyles.

The brain of the pterodactyles seems also to have presented avian characters; the optic lobes are pushed aside by the large cerebellum, which had well-developed floccular lobes. In the reptile's brain the optic lobes intervene between the cerebrum and cerebellum.

The pelvis of the pterodactyles has some likenesses to that of birds. The ilium has an extension in front of as well as behind the acetabulum; and, if the opinions of SEELEY are to be agreed to, there is a rather backwardly directed pubis, more or less fused with the ischium, and a long and thin forwardly directed piece, the prepubis of dinosaurs and the pectineal process of birds.

The main difficulty, however, in the way of comparing pterodactyles and birds is in the fact that both can fly, and that each has acquired the power of flight by a different method. Having acquired the power of flight it seems clear that certain of the points of resemblance between them may easily be due to that mode of life and may have been independently arrived at.

159

THE CLASSIFICATION OF BIRDS

PROFESSOR NEWTON'S article 'Ornithology' in the 'Encyclopædia Britannica,' and the preliminary sketch of Dr. GADOW in Bronn's Thierreich,' contain a digest of, and criticisms upon, the main schemes of classification of this group which have as yet appeared. I shall, therefore, refer the reader to those works for the history of the subject. There can be no question, in my opinion, that birds must be primarily divided into two great divisions, viz. Saururæ and Ornithuræ, the first to contain Archæopteryx and possibly Laopteryx, the latter the rest of birds, both living and extinct. As to the Ornithura, while there is a very general agreement with the main subdivisions-no one probably will quarrel seriously with the divisions adopted in the present workno one has (to my mind) satisfactorily arranged the different groups with reference to each other. More especially does it appear to me that the majority of ornithologists are in error concerning the position of the picarian and passerine birds.

In considering a scheme of classification it is clear that we must bear in mind indications of the descent of birds. Existing schemes have savoured too much of a mere sorting by combining in various ways characters which are distinctively bird characters. However unsuccessful the construction of phylogenetic trees has been, it is abundantly plain that that must be the line to take in arranging a group scientifically. It follows, therefore, that in sketching, at any rate, the main outlines of our scheme attention must be paid only, or chiefly, to those characters which birds have inherited from their reptilian ancestors.

Now this at once lands us in a difficulty, which has been too lightly regarded by many systematists. Phylogenetic schemes used to be boldly linear, and even so recently as the attempt of FÜRBRINGER the family tree savours a little too much of the linear arrangement. Now the imperfect remains of birds that have come down to us from tertiary times show that the modern types of birds were fully differentiated even then in addition to a few extinct forms, such as Odontopteryx toliapicus (if this be not a steganopod). But beyond that point there is the most scanty record of bird life, limited to the Cretaceous Ichthyornithidae and Hesperornithidae, with a few obscurer forms, and to the Jurassic Archaeopteryx. So emphatically were all these creatures birds that the actual origin of Aves is barely hinted at in the structure of these remarkable remains. Moreover, at least in the case of Ichthyornis, they depart fully as widely from any bird with the required 'mixed' characters as any living group, while Hesperornis can with safety be relegated to the neighbourhood of the existing divers. We get, therefore, no help whatever from the Cretaceous birds, and, if any, only the scantiest assistance from Archæopteryx, in determining what are archaic characters in birds. There are no criteria by which we can assert with any degree of safety the relative positions of this and that existing group; nor has the study of the comparative embryology of birds as yet advanced sufficiently far to give any results, except in isolated characters; such indications are the relatively primitive character of the basipterygoid processes, at least in certain groups; for the gulls which are without them when adult, have them as young chicks, as have in the adult condition most of their near allies, the Limicolæ. It may, therefore, so far be inferred that the gulls are a modification of the limicoline type and not vice versa.

It would be perhaps held that any type in which a number of undoubtedly reptilian characters had survived would be on a lower level of organisation than other types in which fewer or no such characters could be discovered. But the few specially reptilian features in the organisation