The number of cervical vertebræ1 among the Anseres varies considerably. The smallest number is found, for example, in Plectropterus gambensis, Biziura lobata, and Tachyeres cinereus, where there are only sixteen. In Edemia nigra there are seventeen; among the swans, twenty to twentyfour. The number of true ribs also varies considerably. The smallest number is to be seen in Cereopsis, where there are but five. Tachyeres, Plectropterus, and a number of other genera have seven; there are eight in Tadorna vulpanser, and as many as nine in certain swans and geese.

The sternum, which has a moderate spina externa, but no spina interna, is whole in Cnemiornis, but has two notches or foramina in other Anseres. The coracoids come into contact, but do not overlap at their sternal articulation. The procoracoid is small, and does not reach the clavicle, which, however, reaches the scapula.

The skull has large oval sessile basipterygoid processes. It is holorhinal and desmognathous. There are frequently lateral occipital fontanelles, as in many pluvialine' birds. The palatines are remarkable for the rudimentary character of their inner laminæ, which brings about a resemblance to the gallinaceous birds, as has been pointed out by HUXLEY, and, it may be added, to the parrots. That part of the palatine is only indicated by a not well marked ridge which is totally absent in Bernicla leucopsis, Chen cærulescens. The general direction of the bone, therefore, is oblique ; it is only near to the attachment with the pterygoids that it becomes feathered out in a horizontal direction. That, at least, is the more normal arrangement; for in Mergus the greater part of the bone has its upper and lower surface coincident with the horizontal axis. This, too, is the case with Biziura lobata.

The oval basipterygoid facets for articulation with the

For osteology see PARKER, On the Morphology of the Duck and the Auk Tribes,' Cunningham Memoirs R. Irish Ac. No. 6, 1890, and SHUFELDT, 'On N. American Anseres,' P. U. S. Nat. Mus. xi. p. 215.

2 Absent in Cereopsis, Biziura, Cygnus, and Cnemiornis. See OWEN, Tr. Z. S. ix. pt iii.

pterygoids are placed so far forwards that the anterior ends of those bones articulate with them.

FIG. 224.-SKULL OF Querquedula crecca. LATERAL VIEW.
(AFTER HUXLEY.)

Fr, frontal; Na, nasal; Pmx, premaxilla; Vo, vomer; Pa, palatine; Pt, pterygoid.

The vomer is a thin deepish plate of bone which is more or less intimately connected with the median septum and maxillo-palatines in front. The maxillo-palatines are completely fused across the middle line in many Anseres (e.g. Chen, Hymenolamus); in Biziura and Mergus they come into contact but are not fused. The latter has very un-ducklike palatine in that the somewhat delicate maxillo-palatines diverge from each other after their junction posteriorly as well as anteriorly, the palatine vacuity in front being (for a duck) unusually extensive. In Chen cærulescens, indeed, secondary bony growths have almost completely obliterated this vacuity, a kind of 'false' palate having been formed.

The lacrymal bones are large, having a considerable length of line of union with the skull; they are sometimes (e.g. Chloëphaga) and sometimes not (e.g. Mergus) ankylosed with the orbital margin. Cereopsis has (among the genera which I have examined) a skull

which is peculiar in that posteriorly the lacrymal is free from the orbital wall, but is fused with a process of the frontal at the anterior end of the supra-orbital impression, leaving (as in Chionis and some other birds) a foramen.' In Cereopsis the descending process of the lacrymal curves backwards and comes near to the zygoma; 2 the junction is completely effected, and there is ankylosis, in Dendrocygna.

The ectethmoids of the Anseres are often largely deficient as ossifications; when present they are thin-walled bubblelike structures, coming into relation with the lacrymals. The interorbital septum is generally very complete; but it is largely vacuolate in Mergus and Biziura.

The following table shows the number of cervical vertebra and the condition of the hæmapophyses and catapophyses in a series of anserine birds :

Fused to form a solid bifid hæmapophysis,' as in some other birds.

* They are rudimentary and do not mount upon the hypapophysis.

As a possible appendix to the Anseres must be mentioned three or four species of an extinct genus of birds, Gastornis, flightless and larger than an ostrich. It has been found only in Europe and from Eocene beds. It is placed among the ratites by LYDEKKER and some others; this is largely on account of the coracoid, which is imperfect above, and appeared to LEMOINE and others as probably 'platycoracoidal.' FÜRBRINGER, however, considers that the tuberosity

fig. 8.

This is figured by OWEN in his paper on Cnemiornis, Tr. Z. S. ix. pl. 35,

2 Sometimes joining. Cf. OWEN, loc. cit. pl. 35, fig. 6.

described by the latter as tubérosité préglénoïdienne' is really the broken end of the scapula, which would be thus, as in Didus, ankylosed with the coracoid, and would also form with it an angle approaching to a right angle. The supposed remains of the scapula, on the other hand, are for FÜRBRINGER the acrocoracoid. On this interpretation the shoulder girdle of Gastornis would be a nearly typical carinate shoulder girdle. The length and slenderness of the coracoid too is not a ratite character, but it does ally Gastornis with Cnemiornis (and also for that matter with Phororhacos). In spite of the freedom of the metacarpals (a character only known elsewhere in Archaopteryx), the complete furcula and various points in the bones of the lower limb, pointed out by NEWTON, together with the facts already referred to, seem to point to a greater likeness to Chemiornis than to any other known type. The skull had basipterygoid processes and seemingly teeth in sockets. The pygostyle appears to have been at most very small and probably absent.

ICHTHYORNITHES

Definition. Small toothed birds with carinate shoulder girdle and sternum. Bones of pelvis not united. Quadrate single-headed. Vertebræ amphicœlous.

This group of birds, from the Cretaceous of North America, has been investigated by MARSH, who, in his great work upon the toothed birds, placed it in a group Odontotormæ, as opposed to Odontolca (Hesperornis), on account of the fact that the teeth are implanted each in a distinct socket. Its relationship to other groups is doubtful; but it is probably not greatly misplaced if we consider it in the neighbourhood of the stork and plover tribe, as has been done by FÜRBRINGER. The group contains two genera-Ichthyornis and Apatornis. Of the former MARSH refers to several species, viz. I. dispar, I. victor, I. validus, I. tener, I. agilis, and I. anceps.

Of Apatornis there is but one species known, A. celer.

These two genera comprise a number of small birds, scarcely larger than a pigeon. In their powerful wings and small legs and feet they remind one of the terns, and according to present evidence they were aquatic birds of similar life and habits."

The restoration of Ichthyornis given by MARSH has been extensively copied in various works, in some of which it would appear as if our knowledge of the osteology of the species selected were greater than is really the case. It has been made, for example, to show schizorhinal nostrils and a pelvis constructed after the carinate type, with the ischia and ilia fused. It is not known whether the skull was schizorhinal, as only the calvarium and the lower jaw and a fragment of the upper jaw have been discovered. The skull has well-marked grooves for the supra-orbital glands; the quadrate, as stated in the definition, is single-headed, as in Hesperornis and many Struthiones. The brain, like that of Hesperornis, is small, and the cerebellum is remarkably large as compared with the hemispheres. The teeth of Ichthyornis are implanted in distinct sockets.

MARSH has remarked upon the close resemblance between the lower jaw, with its teeth, and that of the smaller mosasauroid reptiles. In Ichthyornis dispar there are twentyone distinct sockets in each ramus of the jaw. I. victor had the same number of teeth; in I. anceps the teeth were more numerous, and at the same time more slender. The jaws were united, as in Hesperornis, by cartilage or ligament.

The vertebra, as already mentioned, are amphicœlous; but an approach to the typical saddle-shaped vertebræ is seen in some of them. The atlas is notched for the odontoid process of the axis. None of the dorsal vertebræ appear to have coalesced, and there is a pygostyle quite typical in form, but rather small. The shoulder girdle of both Ichthyornis and Apatornis is constructed upon the carinate plan.' There is the same angle between the scapula and the coracoid, and the clavicles are well developed. There are, however, differ

1 SHUFELDT (Notes on the Extinct Bird Ichthyornis,' J. Anat. Phys. xxvii. p. 336) especially compares Ichthyornis with Rhynchops and Sterna.