physial part (51) has, also, retained its slender rib-like form*; it has coalesced with the hamapophysis (52), and the inverted arch is completed, as in the crocodile, by a hæmal spine, as much modified in form by flattening and expansion as is the neural spine represented by the supraoccipital (3). The diverging appendage of the occipito-hæmal arch also retains much of its primitive simple character: a long and slender bone (53) supports two rays (54, 55), and there is an attempt at three at 57, of which one is short, atrophied and anchylosed to the rest. In the two small bones (56, 56) interposed between this and the preceding segment, we recognise the special homologues of the carpal series in the crocodile and fish: in 54 we have the ulna, in 55 the radius, in 53 the humerus, in 57 the metacarpus ; in d 3 and d 4 the rudiments of the digits so numbered in the crocodile (fig. 22) and the mammal (fig. 24). The evidences of the unity of plan in the construction of the scapular limb, whether it be an arm with the prehensile hand, a hoofed foreleg, a wing, or a fin, are admitted by all; the same scapula, humerus, antibrachial, carpal, metacarpal and phalangial bones are readily recognised by the tyro in comparative osteology in the ape, the horse, the whale, the bird, the tortoise and the crocodile. The beautiful simplicity of the fundamental basis of all these adaptations of structure is descanted upon in all our popular teleological treatises. But the higher law governing the existence of these special homologies has attracted little attention in this country. Yet the inquiry into that more general principle of conformity to type according to which it has pleased the Creator of organic forms to restrict the manifestations of the variety of proportion and shape and substance and even relative position of the limbs requisite for the various tasks assigned to the vertebrate species, is one that by no means transcends the scope of the comparative anatomist. And the conclusion to which my comparisons have conducted me is, that one and the same element, viz. the diverging appendage of the occipital vertebra, forms the seat or substratum of all the adaptive modifications of the part called 'anterior' or 'superior extremity.'

The second segment of the skull has for its central element a bone (fig. 23, 5), which in the bird, as in other ovipara, is connate with that (9) which stands in the same relation to the third cranial segment; the proof of the natural distinction of these segments is given by the neural, N 11, N 111, and hæmal, H 11, H III, arches. Probably the circumstance of the bodies of those vertebræ being formed by ossifications of the fibrous capsule of the notochord, representing the external or cortical parts only of such centrums, may be the condition, or a favourable physical cause of such connation. The neural arch of the parietal vertebra retains the same characters which it first manifested in fishes. Besides the neurapophyses (6) impressed by the mesencephalic ganglia and transmitting the trigeminal nerves, besides the vastly expanded and again, as in fishes, divided neural spine (7), the parapophysis (8) is independently developed. It is of large proportional size; and, owing to the raised dome of the neural arch, is relatively lower in position than in the crocodile; it sends downwards and outwards an unusually long 'mastoid' process, and forms a large proportion of the outer wall of the chamber of the internal ear with the bony capsule of which it speedily coalesces.

The hæmal arch of the parietal vertebra (H 11) is more reduced than in the crocodile, and owes much of its apparently typical character to the retention of the thyrohyals (46, 47) borrowed from the branchial arches of the

* The very common modification of form which this element undergoes in becoming expanded into the broad scapula of man and other mammalia, appears to have influenced Öken in his idea of that bone being the homologue of a congeries of ribs.

visceral system, which are feebly and transitorily manifested in the embryo bird. These spurious cornua project freely or are freely suspended, and are the subjects of singular and excessive development, as has been exemplified in the chapter on Special Homology.

The bones (10) of the third neural arch protect a smaller proportion of the prosencephalon than in the crocodile, but maintain their neurapophysial relation to it and to the optic nerves: the neural spines (11) cover a larger proportion of the hemispheres, and, with their homotypes (7), exhibit a marked increase of development in conformity with that of the cerebral centres protected by their respective arches. The parapophysis of the frontal vertebra (12) is relatively smaller in the bird than in the cold-blooded vertebrates, and is rarely ossified from an independent centre; but I have seen this in the emeu, and it appears to have been constantly an autogenous element in the dinornis. The hæmal arch of the frontal vertebra has been transferred backwards to the parietal one; its pleurapophysis (28), which is simple, as in the crocodile, articulating exclusively with the parietal parapophysis (8), though this in some birds unites with that of the frontal vertebra. In the young ostrich and many other birds traces of the composite character of the hamapophysis are long extant; and bear obviously a homological relation to the teleologically compound character of the element in the crocodile : for the pieces, nos. 29, 29′, 30 and 31 ultimately, and in most birds early, coalesce with each other and with the hamal spine (32), the halves of which are confluent at the symphysis.

The centrum (13) of the nasal vertebra is always single, and, when it does not remain distinct, coalesces with the neurapophyses, 14, and pleurapophyses, 20, of its own segment, and sometimes, also, with the rostral production of the frontal centrum (9): it is elongated and pointed at its free termination, and deeply grooved above where it receives the above-named rostrum ; indicating by both its form and position that it owes its existence, as bone, to the ossification of the outer capsule of the anterior end of the notochord. In the ostrich the long presphenoidal rostrum intervenes between the vomer (13) and prefrontals (14). These latter bones manifest, however, as has been shown in the paragraph on their special homology (p. 214), all the essential neurapophysial relations to the rhinencephalon and olfactory nerves: but they early coalesce together, or are connate, as in the tailless batrachians. The neural spine (15) is divided along the middle line; but in most birds the suture becomes obliterated and the spine coalesces with its neurapophyses, with the frontal spine and with those parts of the hæmal arch of the nasal vertebra with which it comes in contact.

The pleurapophyses (fig. 23, 20) of this inverted arch retain their typical connections with the nasal centrum and neurapophyses at one end, and with the hæmapophysis (21) at the other end, and they also support the constant element of the diverging appendage of the arch, no. 24. The hamapophysis (21) resumes in birds more of its normal proportions and elongated slender form: but the hamal spine (22) is largely developed though undivided, and sends upwards and backwards from the part corresponding to the symphysis of the spine, when this element is divided, a long pointed process (22), which joins and usually coalesces with the neural spine (15) and divides the anterior outlet of the hæmal canal into two apertures called the nostrils. The modification of the inferior arch of the nasal vertebra in the lizard tribe is here repeated. The pleurapophysial appendage, 24, connects the palatomaxillary arch with 29, and in the ostrich and a few other birds, also with 5: the second or hæmapophysial ray of the diverging appendage is developed in all birds, as in the squamate saurians; combining the movements

Νι