Art. XIII.—On the Axial Skeleton in the Dinornithida.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 27, 1894, Page 157

Art. XIII.—On the Axial Skeleton in the Dinornithida. By Captain F. W. Hutton, F.R.S., Curator of the Canterbury Museum. [Read before the Philosophical Institute of Canterbury, 3rd October, 1894.] Sir Richard Owen has described and figured different parts: of the axial skeleton in the Dinornithidae, and has distinguished what are now taken to be generic characters in the sterna of Dinornis, Anomalopteryx, Meionornis (= Mesopteryx) and Pachyornis.* Trans. Zool. Soc. of London, passim, and Ext. Birds of N.Z. Mr. R. Lydekker has distinguished the vertebrae and pelves of Dinornis and Pachyornis, and has also described the pelvis in Megalapteryx and Anomalopteryx;† “Catalogue of the Fossil Birds in the British Museum,” 1891. and Dr. Parker has, quite recently, given a full description of the skull, so far as it is at present known, in all the genera.‡ Trans. Zool. Soc. of London, vol. xiii. I now offer, as a further contribution to our knowledge of these birds, descriptions of the vertebral column, pelvis, and sternum of Euryapteryx (—Emeus of Lydekker and Parker); the vertebral column and pelvis of Meionornis (= Mesopteryx); and some remarks on the vertebrae of Megalapteryx. Nomenclature of the Genera.—The genus Dinornis is well established, and all the species admitted into it by Mr. Lydekker are also admitted by Dr. Parker and myself. The sub-genus Tylopteryx appears to be unnecessary, for Parker has shown that the figures given by Owen, and by Jaeger in the palaeontology of the voyage of the “Novara,” are erroneous. I also abandon my attempt to resuscitate the genus Pal-apteryx, as the skull in the British Museum on which it was founded is said by Mr. Lydekker to be made up from the calvarium of geranoides and the praamaxillae and mandibles of casuarind. A pelvis from Te Aute, belonging to Mr. A. Hamilton, is of the true Dinornis type, but much too small for D. struthioides, and I have little doubt but that it belongs to Dinornis dromioides. On the other hand, the pelvis in the Wellington Museum, which I associated with my Palapterya; plena, agrees, fairly well with Mr. Lydekker's description of the pelvis in Megalapteryx. I also abandon the genus Cela. A part-skeleton of associated bones with the skull of curtvs is in the Auckland Museum, and shows that it is congeneric with didinus, although differing somewhat in the sternum and vertebrae, as I shall presently point out. Dinornis oweni is

therefore quite distinct from it, and must be placed near Anomalopteryx parva. We now come to the genus which, includes curius, didinus, casuarinus, and a larger species (Mesopteryx b of Parker), which is probably rheides. It is Syornis and Cela of Reichen-bach; Meionornis of Haast; the typical Anomalopteryx of Lydekker (not of Reichenbach); and Mesopteryx of myself and Dr. Parker. Which of these should be adopted? Reichenbech's names, I think, cannot stand unless subsequently defined by some one else. His only knowledge on the subject was derived from Owen's published papers, and probably he never handled or even saw a moa bone. He gave no characters, but, without assigning any reasons for doing so, he placed in different genera each one of Owen's species, at a time when hardly anything was known about them. Cela (Möhring) is a synonym of Casuarius, and is not brought forward by Reichenbach as a new genus; the three species curtus, casuarinus, and crassus being grouped with Casuarius, while the species of Dinornis are grouped with Dromaius. If I had known this in 1891, when I was writing my paper on the moas of New Zealand, I should have ignored Reichenbach's names altogether. Mesopteryx also cannot be used here, as it has been previously given by H. de Saussure to a genus of Orthoptera.* “Méanges Orthopterologuiques,” fasc. iii., Mantidae, p. 188 (1870). To adopt Anomalopteryx for the generic name would now lead to confusion, as it was used by me—nearly simultaneously with Mr. Lydekker—for Reichenbach's typical species didiformis, and has been so used by Dr. Parker. Also Meionornis (Haast) has precedence of Anomalopteryx (Lydekker), and was fairly well defined. It was only Haast's mistake of identifying didinus as didiformis which led to Mr. Lydekker using the name of Anomalopteryx for this genus. I shall therefore take Meionornis as the correct name, and keep Anomalopteryx for the “Celine group” of Mr. Lydekker, as has been already done by Dr. Parker and myself. For the next group—which includes ponderosa, crassus, and gravis—there is Reichenbach's old name of Emeus, which was only defined by Lydekker in 1891; and Haast's name of Euryapteryx, fairly well defined in 1874. Haast certainly mixed up Euryapteryx with his Palapteryx (Pachyornis of Lydekker), and I did the same. It was Mr. Lydekker who first clearly distinguished the two genera: still, I do not think that this is a sufficient reason for passing over Haast's name. Emeus was not given by Reichenbach as a new genus, but only as an amended spelling of Emeu (Barrère), which is identical with Emou (Vieillot), another synonym of Casuarius. By no canon of nomenclature can a genus Emeus of Reichenbach

bach be allowed. Of the correctness of the name Pachyornis there can be no doubt. As the generic characters depend more on the axial than on the appendicular skeleton, it follows that the species in which the axial skeleton was first described-must be taken as the type of the genus, and not the species to which the generic name was first applied, if the leg-bones only were known. This is necessary to prevent constant changes. General Characters of the Vertebral Column.—So far as is at present known, all the moas have fifty-six vertebræ. Of these, 1 to 21 are cervical; 22 to 24 are cervico-thoracic; 25 to 27 are free thoracic, with sternal ribs articulating with the sternum. No. 28 is sometimes free, sometimes attached to the pelvic vertebrae; both it and No. 29 have floating sternal ribs. The pelvic (or syn-sacral) vertebræ are seventeen in number, from 28 to 44. No. 45 is sometimes anchylosed to those in front of it, sometimes free; Nos. 46 to 53 are free caudals; No. 54 is sometimes free, sometimes anchylosed to Nos. 55 and 56, which always form a single bone. The atlas is broadly oval, the longer diameter dorsi-ventral. The bony bridge for the vertebral artery is always strong. The post-zygapophyses are well marked; there is no trace of a neural, spine. The pre-axial articulating surface is crescentic, the dorsal notch being very large.* I have never seen an atlas resembling the one figured by Owen in his Ext. Birds of N.Z., p. 392. It is much like the atlas in Dromaius, and very different from that of Apteryx. The axis is comparatively short in an axial direction. The neural platform is triangular, with a nearly straight posterior-margin. There is a deep pit behind the neural spine, and well-marked hyperapophyses. The posterior articulating surface is much elongated ventrad, and below it is a strong hypapophysis. This bone also closely resembles that of Dromaius, and is very different from that of Apteryx. In the cervical vertebræ the centra increase in length to No. 16 or 17, and are proportionally longest and narrowest from 9 to 11. In Nos. 2 to 5 the hypapophysis forms a keel, which becomes obsolete on No. 6. A pair of hypapophyses-spring from the ventral surface of the parapophyses in No. 6; they increase in size in each posterior vertebra, and are widest apart at No. 15 or 16. They then approach, and coalesce into a median tubercle on No. 19 or 20. Dorsad, the anterior cervicals, Nos. 3 to 6, expand into a quadrangular neural platform, which is transverse in Nos. 3 to 5, nearly square in No. 6, From this platform rises the neural spine, always divided into two on No. 6, and usually so on Nos. 3 to 5. On No. 7 the two neural spines are continued postaxially

as ridges on to the post-zygapophyses, and from No. 8 to 16 the spine cannot be distinguished from the ridge. But in No. 17 the spines again appear, increase in size on each posterior vertebra, and coalesce into a single spine on No. 21. In these vertebræ a distinct hyperapophysis can be recognized on the post-zygapophyses. This remarkable blending of the neural spines, hyperapophyses, and post-zygapophyses into two longitudinal ridges is very characteristic of the Dinornithidæ, and enables the cervical vertebrae to be distinguished from those of all other birds. Sir R. Owen and Professor Mivart call them the hyperapophyses, and Mr. Lydekker calls them the post-zygapophysial ridges. They will be distinguished in the present paper as the neural ridges. Between the pre-axial ends of the neural ridges there is a transverse bony bar, considered by Lydekker as the neural spine, on the post-axial side of which there is a deep fossa for the ligament. This transverse bar increases in height on the posterior vertebrae until it becomes blended with the neural spines in No. 21. The vertebrarterial canals are larger than the neural canal. In the cervico-thoracic vertebræ the centra are approximately of the same size in all; but they can be distinguished by the hypapophysis which in No. 22 forms a compressed central nodule, extending in No. 23 to the pre-axial face, and in No. 24 covering the whole length of the centrum; usually projecting forwards in front of the pre-axial surface, and getting shallower towards the post-axial surface. In the free thoracic vertebræ the centra have all the same length, but increase in breadth from No. 25 to No. 28. In No. 25 the hæmapophysis forms a deep keel from the pre-axial to the post – axial articulating surfaces, which is sometimes notched in the centre. In No. 26 the hæmapophysis is posterior, but there is also a slight anterior tubercle. In No. 27 the hæmapophysis is posterior only, sometimes single, sometimes divided. In No. 28 it forms two widely-separated posterior nodules. No. 27 can generally be further distinguished from No. 26 by the shape of the centrum, which is more compressed in the latter than in the former. Of the pelvic vertebræ—Nos. 29 to 44–No. 30 always has the rib anchylosed to the centrum. Nos. 31 to 33 have transverse processes which unite with the pre-acetabular portion of the ilia. No. 34 sends stout transverse processes to meet the proximal ends of the ischia and pubes. In the next three—Nos. 35 to 37—transverse processes are-absent, and this gives rise to a deep fossa on each side, in the centre of the pelvis (the pelvic fossæ). Nos. 38 to 44 have transverse processes which coalesce with the posterior portion of the ilia and form, dorsad, the sacrum, or pelvic disc of Sir R. Owen.

A very good series of woodcuts of vertebrae are given by Owen in Trans. Zool. Soc., vol. x., pp. 152–170, and in “Extinct Birds of New Zealand,” pp. 396–414. These, which are referred to Dinornis maximus, really belong to Pachyomis, as has already been pointed out by Mr. Lydekker. A correction in the numbers of these vertebræ will be useful:—- Figs. 8–11 are Vert. No. 4. Figs. 12, 13 are " " 6. Fig. 14 is " " 8. Figs. 15, 16 are " " 17. Fig. 17 is " " 18. Fig. 22–24 are " " 20. Figs. 18–21 are " " 21. Figs. 25, 26 are " " 23. Figs. 27–29 are " " 24. Fig. 30 is " " 25. Figs. 31, 32 are " " 27. Figs. 33, 34 are " " 28. Rings of the Trachea.—Two types of ossified tracheal rings are found with moa-bones: (1) The hoop-like rings of Lydekker, which are slender, smooth, and often elliptical in shape; (2) the tube-like rings of Lydekker, which are thick, rough, deep, and circular in section. These two types have always been considered as generieally distinct, but such is not the case. In a skeleton of Euryapteryx ponderosa, found by Mr. B. S. Booth in the central district of Otago, the trachea was nearly complete, and it had hoop-like rings near the head and tube-like rings near the sternum, one set gradually passing into the other about the middle of its length. This was probably the same in all the moas, as some of the tube-like rings are very small. In the following remarks I shall treat all the free vertebræ which have ribs as thoracic—that is, from No. 22 to No. 28. Those in front of them are cervical; those behind them belong to the pelvis. Genus Dinornis Owen (1843). Type—D. robustus, Owen. Figures.—Atlas, Owen, Trans. Zool. Soc., vol. v., pl. 53, figs. 4–6, and Ext. Birds of N.Z., pl. 62. Axis, Owen, Trans. Zool. Soc., vol. x., p. 150, figs. 4–7, and Ext. Birds, p. 394. Cervical vertebra, Owen, Trans. Zool. Soc., vol. iii., pl. 40, and Ext. Birds, pl. 32. Lydekker, Fossil Birds, Brit, Mus., p. 247, fig. 59. Pelvis, Mivart, Trans. Zool. Soc., vol. x., p. 43, figs. 42, 43. Sternum, Owen, Trans. Zool. Soc., vol. x., pl. 32, and Ext. Birds, pl. 98. Lydekker, l.c., p. 222, fig. 56. Cervitcal Verteberæ. In the axis the odontoid process is long, tapering, and slightly rounded. The pre-axial articulating surface moderate, the height less than the breadth. The neural spine is remarkably high and thick. The anterior cervicals—2 to 6—are short and broad. In No. 5 the length is three-quarters of the

breadth, and the height of the post-axial face of the centrum is equal, to the breadth; the posterior margin of the neural platform is straight or slightly concave. The neural spine is not divided in Nos. 3, 4, or 5. Hyperapophyses are well developed on No. 4, slightly so on No. 5. The faces of the post-zygapophyses in Nos. 4 and 5 are directed more post-axially than in other genera. Middle cervicals, 7 to 17: The post-axial surface of the centrum is broad, the greatest breadth at the ventral margin being more than twice the least vertical (dorsi-ventrad) height; and the ventral breadth less than twice the dorsal breadth. There is generally a pneumatic foramen behind the diapophysis. The neural ridges are more elevated than in any other genus, and converge only slightly pre-axiad; the transverse bar is high, with well-marked pre and post-axial rough ligamentous areas. The faces of the post-zygapophyses are directed more outwards than usual; and usually there is a deep conical pit on each side at the pre-axial end of the neural ridges. The posterior cervicals—18 to 21—have the same characters, but the centra shorten in length considerably, and there are no pits at the pre-axial bases of the neural spines. The neural spine is occasionally single on 19 and 20, and the hæmapophyses become single on 20. Thoracic Vertebræ. The centra have broad post-axial surfaces. The anterior pneumatic foramen is very large, and in Nos. 24 to 28 descends below the rib-facet (in very old birds this foramen becomes obliterated); the posterior pneumatic foramen is large in all. The hæmal ridge (hypapophysis) is not continued to the pos-terior end of the centrum, in 22. In 23, 24, and 25 it extends all along the centrum, but is deeper preaxiad, and often projects forwards as a hook. The transverse processes are short and thick, as also are the neural spines, which in Nos. 22 and 23 are nearly square in section, and have both pre- and post-axial surfaces remarkably rough. The post-zygapophyses are shortr with broadly-oval faces; the channels between them and the buttresses of the neural spine are shallow. The pre-zygapophyses are short, and not expanded at the extremity. Pelvis. The pre-acetabular portion of the ilium is moderately deep, the anterior margin projecting much forwards, and the dorsal margin nearly straight axially. The anterior margin of the acetabulum is in the centre of the ilium. The posterior portions of the ilia begin to diverge at the anterior margins of the acetabula, but soon become more nearly parallel than in any other genus. The post-acetabular portion is short, and does

not project beyond the 45th or 46th vertebra; its lateral surface slopes downwards and outwards; and posteriorly it descends as a sharp ridge below the vertebral column. The ischia and pubes project backwards beyond the ilia. The ventral surfaces of the sacral vertebrae—35 to 37—are short, broad, and flat. The transverse processes of Nos. 38 and 39 are at right angles to the axis, but in all the posterior vertebrae these processes slope obliquely backwardd, thus leaving a triangular fossa on each side behind the usual pelvic fossæ. Sternum. The body is broad, with well-marked coracoid pits, short, costal processes, long and diverging curved lateral processes, and the median process often notched. Genus Megalapteryx Haast (1885). Type—M. tenuipes, Lydekker. Figures.—Pelvis, Lydekker, Fossil Birds in Brit. Mus., p. 278, fig. 63, and fig. 61 (outline). The twelve more or less imperfect vertebræ here described were found in a cave in the Lower Buller district with a cranium and leg-bones of M. tenuipes. The cranium has been described by Dr. Parker in Trans. Zool. Soc. Cervical Vertebræ. In the middle cervical region the centra are long and narrow; the post-axial surfaces, broad. The hypapophyses arise considerably behind the pre-axial surface. The neural ridges are very low and flat, converging rapidly, anteriorly; the transverse bar low, and rising very obliquely. A saucershaped depression on each side at the pre-axial base of the neural ridge. Faces of the post-zygapophyses more horizontal than in Dinornis. The posterior cervicals—19, 20, 21—have the centra short. The hypapophysis is single on No. 19; on No. 20 it forms a short, compressed anterior ridge, which in No. 21 extends backwards over the greater part of the centrum. The neural spine is double in No. 19; single in 20 and 21. Hyperapophyses occur on Nos. 20 and 21, but not on 19. Thoracic Vertebræ. Nos. 22, 26, and 27: The post-axial faces are broad. The anterior pneumatic foramen descends below the rib-facet in Nos. 27 and 26, and in 22 it is slightly developed. The posterior pneumatic foramen is large in 27 and 26, obsolete in 22. The hypapophysis in No. 22 forms a central compressed

keel. The transverse processes are moderate and distinctly ascending. The neural spine in No. 22 slopes forward, and is much compressed; in Nos. 26 and 27 it is short, compressed, straight or slightly curved forwards. The post-zygapophyses form a right angle with each other; their faces are broadly oval. The pre-zygapophyses are moderate. Pelvis. The pelvis is narrow, and the anterior margin of the aceta-bulum is in the centre of the ilium. The pre-acetabular portion of the ilium is deep, and the two meet dorsally in an arched ridge. They begin to diverge far back, behind the acetabula; their lower post-acetabular margins are thick and blunt. The ventral surfaces of the post-sacral vertebrae are slightly longitudinally ridged. Genus Meionornis, Haast (1874). Anomalopteryx (typical), Lydekker (1891). Syornis and Mesopteryx, Hutton (1891). Type—M. casuarinus, Owen. Figures.—Sternum, Owen, Trans. Zool. Soc., vol. vii., pl. 8 and 9; Ext. Birds of N.Z., pl. 73 and 74. Lydekker, Fossil Birds in Brit. Mus., p. 254, fig. 62A. Cervical Vertebræ. The axis has the pre-axial articulating surface narrow, the height being nearly as great as the breadth. The hypapo-physis descends pre-axially more abruptly than in other genera. The odontoid process is very short. The neural spine high and much compressed. The hyperapophyses also compressed. The anterior cervicals—3 to 6—are distinguished by the neural spine being single on No. 3, and only partly divided on No. 4. The hyperapophyses are small on Nos. 4 and 5. The post-axial margin of the neural platform is straight in 3 and 4; concave in 5 and 6. In No. 6 also the platform is slightly arched dorsad. The middle cervicals—7 to 18—have long narrow centra, with the post-axial surface broad ventrad and narrow dorsad; the breadth of the ventral margin is more than the height, and more than one and a half times that of the dorsal margin. Usually there is a pneumatic foramen behind the diapophysis on Nos. 15 to 18. The hypapophyses rise immediately behind the pre-axial margin of the centrum. The neural ridges are high and converge rapidly, their post-axial ends being twice as far apart as their pre-axial ends. The transverse bar is high and rises abruptly. At the pre-axial

base of each neural ridge there is a saucer-shaped depression which is sometimes pierced by a foramen under the zygapophysial bridge. The faces of the post-zygapophyses are more horizontal than in Dinornis. Hyperapophyses are distinct on No. 18. The posterior cervicals—19 to 21—have short, but comparatively narrow, centra on all. The neural spine is usually single on No. 21 only. The hypapophysis is single on No. 19. The hyperapophyses are distinct in Nos. 19 and 20. Posterior pneumatic foramina are developed in all. Thoracic Vertebræ. Post-axial faces broad. In Nos. 22 and 23 the dorsal margin is rather less than the height, while the ventral margin is more than the height; in all the others the sides are parallel; the breadth is about equal to the height in 24 and 25, and greater than the height in the others. The anterior pneumatic foramen descends below the rib-facet in 27 and 28, but is often divided into dorsal and ventral portions by a bony septum. In 26 it is large, and often reaches slightly below the rib-facet, but is not divided; in 24 and 23 it is often small or obsolete, but usually developed again on No. 22. The posterior pneumatic foramen is generally well developed on all. The hypapophysis in 22 and 23 is central; in 24 and 25 it is continuous the whole length of the centrum, and is slightly notched in the middle in No. 25. There is a small anterior tubercle in Nos. 26 to 28. The transverse processes are long, especially in Nos. 25 and 26, and rise slightly. The neural spine is much inclined forward in No. 22; in the others it is long and compressed; straight or slightly curved forwards in 26 to 28. In these three the neural spine is twice as long as the distance from its base to the ventral margin of the centrum. The post-zygapophyses rise rapidly, and their faces are narrowly oval; they form a right angle with each other, but the valley between them is shallow. The pre-zygapophyses are moderate, slightly expanding distally. Pelvis. Pre-acetabular portion of the ilia moderately deep, forming a sharp dorsal ridge, which is curved axially. The centre of the acetabulum is in the centre of the ilium. The ilia begin to diverge at the anterior margins of the acetabula, and are prolonged backwards rather beyond the centrum of No. 46, making a broad pelvic disc: the angle of divergence is between 60° and 70°. The lateral surface of the post-acetabular portion of the ilium slopes downwards and inwards, and posteriorly projects as a sharp ridge below the vertebral column. The ventral surfaces of the sacral vertebrae are

broad, but longitudinally ridged; the transverse processes of the posterior ones are strong, slightly oblique, and leave no second fossa. Sternum. Body nearly quadrate, anterior margin slightly curved dorsally; costal processes long and directed much dorsad, but not pre-axiad; lateral processes long, slender, not diverging much; median process long. No coracoidal depressions, but internally there are deep pneumatic depressions at the antero-lateral corners. Note.—M. curtus differs from the others in the following points: The cervical vertebrae have flatter neural ridges, and often no depression at their pre-axial bases. In the thoracic vertebrae the anterior pneumatic foramen does not descend so low, and is not divided by a septum, and there is no posterior pneumatic foramen in No. 28. The neural spine is shorter, and the faces of the zygapophyses are broader. The sternum has no pneumatic depressions in the antero – lateral corners. Genus Anomalopteryx, Hutton (1891).* “New Zealand Journal of Science,” Nov., 1891, p. 248. Anomalopteryx (Group Cela), Lydekker (1891). Type—A. parva, Owen. Figures.—Owen, Trans. Zool. Soc., vol. xi., pl. 51 to 58. Haast, Trans. Zool. Soc., vol. xii., pl. 31 and 32. The material in the Museum is not sufficient to enable me to give a full description of the vertebral column; but the following is offered as a contribution. The vertebrae resemble those of Pachyornis more than any other genus. Cervical Vertebræ. Anterior cervicals, 4 and 5: The neural spine is divided nearly to the base in both. The post-axial margin of the neural platform is deeply concave in both. Hyperapophyses are strong on No. 4, but not strong on No. 5. Middle cervicals have the centra moderately long; the hypapophyses rise close to the pre-axial border. The neural ridges are flattened, but carry narrow hyperapophyses. The transverse bar is low, and rises very obliquely. There is no depression at the pre-axial base of the neural ridge. The posterior cervicals have the hypapophysis single on both 20 and 21; while the neural spine is double in No. 20, and either single or double in No. 21.

Thoracic Vertebræ. The post-axial surfaces of the centra are narrow. The anterior pneumatic foramen does not descend below the rib-facet. The posterior pneumatic foramen is well developed. The post-zygapophyses form an acute angle with each other, and the pre-zygapophyses are short. Pelvis. Narrow. Centre of ilium half-way between centre and anterior margin of acetabulum. Pre-axial portion of ilia rather low, and sharply ridged; dorsal margin nearly straight axially. The ilia begin to diverge behind the anterior margin of the acetabula, and extend backwards to vertebra No. 46. Ischia generally extending slightly beyond the ilia. Lateral surfaces of post-acetabular portion of ilia the same as in Meionornis. Ventral surfaces of sacral vertebrae long, not ridged. Transverse processes of posterior vertebrae more oblique and stronger than in Meionornis. Sternum. Body quadrate; the anterior margin strongly curved dorsad, the costal processes long and directed dorsad also. Coracoid depressions deep and oval. Shallow pneumatic depressions on the inside at the antero-lateral corners. Lateral processes stronger than in Meionornis. Scapulo-coracoid about 3 ¼ in. in length, of which the coracoid occupies one-third; not much expanded at the anchylosis. Genus Pachyornis, Lydekker. Euryapteryx, in part, Hutton. Section A. Type—P. immanis, Lydekker. The following description is taken from a nearly-complete vertebral column found with the skull, sternum, pelvis (fragmentary), and most of the leg-bones, the latter being typically immanis. Cervical Vertebræ. Anterior cervicals—3 to 6—have the post-axial surfaces of the centra more quadrate than in any other genus, the height being equal to the breadth in No. 4, slightly greater than the breadth in No. 3, and less than the breadth in No. 5. Greatest breadth of No. 3 is 38mm. (1.5in.). The neural spine is divided nearly to the base on Nos. 3, 4, and 5. The zygapo-physial

faces look more dorsi-ventrad than in Dinornis, and they are not so broad. In the middle cervicals—11 to 14—the centra are very short, and the post-axial surface very broad; the width at the ventral margin more than one and a half times that of the dorsal margin, the height less than one-half the ventral breadth. No pneumatic foramen on Nos. 7 to 10. Hypapophyses rise immediately behind the pre-axial border of the centrum in Nos. 13 and 14, and nearly so in the anterior vertebrae. Neural ridges very short, and nearly parallel, steep-sided in the anterior vertebrae, but getting quite flat in the posterior ones; the transverse bar well developed in all. An inter-zygapophysial bridge in 11 and 12. A shallow saucer-shaped hollow at the pre-axial base of the neural ridge. Faces of the post-zygapophyses nearly horizontal (as in Meionornis). Posterior cervicals, 15 to 21: Centra short; pneumatic foramina in all. A deep valley in front of the hypapophyses in No. 17. Neural ridges high, but short, in No. 15; a pit on their inner sides in Nos. 16 to 18. Hyperapophyses distinct in No. 18. Thoracic Vertebræ. Post-axial faces of the centra moderate, narrowest in No. 25. In No. 22 the ventral breadth is considerably more than the height; in No. 24, also, the ventral margin is greater than the dorsal; but in the others the sides are parallel, and the height is about equal to the breadth. The anterior pneumatic foramen is situated entirely above the rib-facet; it is well developed in Nos. 28–25, small in Nos. 24 and 23, and absent in No. 22. The posterior pneumatic foramen is small or absent in Nos. 28–23, but well developed in No. 22. The hypapophysis in No. 22 is a compressed central nodule; in No. 23 it extends over the anterior half only of the centrum; in 24 it is continuous along the whole of the centrum; in 25 it is divided in the centre into two parts: there is an anterior nodule on 26, but not on 27. The transverse processes are rather short; the neural spine in No. 22 is much inclined forwards (in all the others they are broken off). The post-zygapophyses are inclined to each other at about a right angle; the valley between them is shallow. The pre-zygapo-physes are shorter than in any other genus, and they are but slightly expanded distally. Sternum. Resembles that figured by Sir R. Owen in Trans. Zool. Soc., vol. vii., pl. 7, and Ext. Birds of N.Z., p. 72. The body is broad and flat, with strong lateral processes, which diverge much from the central line. Median process short. Anterior border slightly curved; the costal processes well developed.

No coracoid depressions; but distinct pneumatic depressions inside. The pelvis is too fragmentary for description. Section B. Type—P. elephantopus, Owen. Cervical Vertebræ. The cervical vertebra differ from the last in the post-axial surfaces of the centra being narrower, and the neural ridges considerably longer, and the bar between them rising very obliquely. The thoracic vertebra, from No. 28 to No. 24, have well-developed posterior pneumatic foramina. The transverse processes are longer; as also are the pre-zygapophyses, with a deep channel between them. This type is much more common than the first, and is represented in Sir R. Owen's drawings, Trans. Zool. Soc., vol. x., pp. 152–170, and Ext. Birds of N.Z., pp. 396–414. Pelvis. The pelvis is very broad, and the centre of the acetabulum is in the centre of the ilium. The pre-acetabular portion of the ilium is low, and the two meet dorsally in a rounded ridge, which is straight in an axial direction. They do not project in front of the neural spine of No. 28 vertebra. They begin to diverge before the anterior margins of the acetabula, and then form a broad pelvic disc, the angle of divergence being about 70°. The posterior portion of the ilium is very broad and flat, and does not project beyond vertebra 45 or 46. The lateral surface is perpendicular and very shallow, not projecting as a ridge below the vertebral column. The ventral surfaces of the sacral vertebrae are narrowed; the transverse processes of the posterior vertebrae are strong and not very oblique, those of Nos. 38 and 44 being at right angles to the axis. Sternum. Same as in Section A. Genus Euryapteryx, Haast (1874). Emeus, Lydekker (1891). Type—E. gravis, Owen. Cervical Vertebrae. The atlas has the neural canal transversely oval, the horizontal diameter rather more than one and a half times the vertical. The pre-axial articulating surface is crescentic, the margins, if continued, forming a circle. Post-zygapophyses distinct, projecting post-axially, and carrying distinct rounded

hyperapophyses; the articular face looks inwards and slightly ventrad. In the axis the pre-axial articulating surface is crescentic, the height not quite half the breadth, and nearly straight dorsad, at the root of the odontoid process. The postaxial surface is higher than broad, with straight sides, and both dorsal and ventral margins concave. The odontoid process is moderate, tapering, bluntly rounded, the upper surface concave; a slight cup on each side, near the base, for articulation with slight tubercles at the post-axial bases of the neural laminae of the atlas. The hypapophysis is strongly developed post-axially, but fades away about the middle of the centrum, the anterior portion of which is smooth. Neural platform triangular, its post-axial margin slightly concave, and its breadth considerably greater than the length to the tip of the odontoid. Neural spine short and stout. The post-zygapophyses look almost completely ventrad, but slightly postaxiad; the hyperapophyses prominent and rounded, not compressed. Anterior cervicals, 3 to 6: The post-axial margin of neural platform straight in No. 4, concave in Nos. 5 and 6. Neural spine divided to the base in all. Hyperapophyses strong on Nos. 3 and 4 only. Middle cervicals, 7 to 14: Centra moderate or rather short; the post-axial face narrowed dorsad: breadth at ventral margin less than one and a half times that of the dorsal margin; the height slightly greater than the dorsal breadth. Generally a pneumatic foramen below the diapophysis in Nos. 10 to 14; sometimes another behind it in Nos. 12 to 14. The hypapophyses from Nos. 9 to 14 rise some distance behind the pre-axial border of the centrum. The neural ridges rise abruptly, and from Nos. 6 to 12 they converge rapidly, the distance apart of the post-axial ends being considerably more than twice that of the pre-axial; but in Nos. 13 and 14 they are more nearly parallel. The transverse bar is very low, and rises obliquely. There is a deep conical pit on each side at the pre-axial ends of the neural ridges, usually with a foramen at the bottom; there is also usually a foramen in the inner faces of the neural ridges, near their anterior ends. The faces of the post-zygapophyses look more outwards than in Meionornis. Posterior cervicals, 15 to 21: The centra are short; the post-axial face in No. 19 has the height in the centre about equal to the dorsal breadth. There is a pneumatic foramen behind the diapophysis in all, and usually one below it also. The hypapophyses are separated from the anterior border of the centrum by a transverse groove. They are single on Nos. 19 to 21, usually forked on No. 18, and double on No. 17. The neural spine is single on No. 20; double, or forked, on No. 19. Hyperapophyses are well marked on 18 to 21.

Thoracic Verterbae. The post-axial surfaces of the centra are narrow: in No. 22 the height is rather more than the greatest breadth, and they get narrower to No. 26, in which the height is one and a half times the breadth; in No. 28 the height is about equal to the breadth in the middle. The anterior pneumatic foramen descends below the rib-facet in No. 28, slightly so in Nos. 27 and 26, and becomes obsolete in No. 24. There is a posterior pneumatic foramen in all. The hypapophysis in No. 22 forms a compressed central nodule, deepest behind. In No. 23 it is longer, but does not reach either end of the centrum. In No. 24 it is continuous as a keel and projects forwards. In No. 25 there is no notch, but a slight posterior nodule, and the keel does not project forward. There is an anterior nodule in No. 26, but not in 27. The transverse processes are rather long and thick; not ascending. The neural spines are long, that of No. 22 sloping forward but slightly, those of the posterior vertebrae often much curved forwards. The post-zygapophyses make an acute angle with each other, the valley between them deeply hollowed out, and in No. 25 they are deeply channelled on the upper surface. Pre-zygapophyses long and expanded distally. Pelvis. The pelvis is broad, with the centre of the acetabulum in the centre of the ilium. The pre-acetabular portions of the ilia are deep, and produced forwards in front of the neural spine of No. 28 vertebra. Dorsad they meet in a sharp ridge, which is often curved in an axial direction. The ilia begin to diverge at the anterior margins of the acetabula, and form a broad pelvic disc, the angle of divergence being about 90°. The post-acetabular portions of the ilia are produced backwards beyond No. 46 vertebra; their lateral surfaces are perpendicular and short, forming ventrad a sharp ridge which does not descend so much as in Meionornis, but more so than in Pachyornis. Ventral surfaces of the sacral vertebrae as in Meionornis. The transverse processes of the posterior vertebrae, 40 to 44, very oblique. The ischia project considerably beyond the ilia. Sternum. The body is quadrate, or broader than long; with strong, straight lateral processes, which do not diverge much. Median process long. Anterior border slightly curved; the costal processes large, slightly directed dorsad. No coracoid nor pneumatic depressions.

In the absence of illustrations the following keys may assist in distinguishing between the genera:— Middle Cervical Vertebrae. Neural ridges nearly parallel.  Neural ridges high Dinornis.  Neural ridges low.   Transverse bar high; ridges short Pachyornis, section A.   Transverse bar low; ridges long Pachyornis, section B. Neural ridges markedly converging.  Neural depressions* The depressions at the pre-axial ends of the neural ridges. deep Euryapteryx.  Neural depressions shallow.   Dorsal margin of post-axial surface of centra less than two-thirds of the ventral margin Meionornis.   Dorsal margin of post-axial surface of centra more than two-thirds of the ventral margin.    Centra elongated Megalapteryx.    Centra shorter Anomalopteryx. Thoracic Vertebrae. Anterior pneumatic foramen descending below the rib-faced in No. 27.  Posterior faces of the centra of Nos. 26 to 24 broad.   Neural spines short, square Dinornis.   Neural spines short, compressed Megalapteryx.   Neural spines long, compressed Meionornis.  Posterior faces of the centra of Nos. 26 to 24 narrow Euryapteryx. Anterior pneumatic foramen not descending below rib-facet in No. 27.  Post-zygapophyses form an acute angle with each other Anomalopteryx.  Post-zygapophyses form a right angle with each other.   No posterior pneumatic foramen Pachyornis, section A.   A posterior pneumatic foramen Pachyornis, section B. Pelvis. Centra of ribless post-sacral vertebrae (Nos. 35 to 37) forming a mass which is as broad as long, and transversely ridged. Two pelvic fossae on each side Dinornis. Mass of ribless centra longitudinally keeled or ridged: generally longer than broad.  Ilia diverge at an angle of about 65° Meionornis.  Ilia diverge at an angle of about 90° Euryapteryx. Mass of ribless centra much longer than broad, nearly smooth, or slightly ridged.  Lower surface of the centra rounded Pachyornis.  Lower surface of the centra flat, with sharp lateral margins.   Posterior portion of ilium sharply ridged below Anomalopteryx.   Posterior portion of ilium flat below Megalapteryx.

Sternum. Coracoid depressions deep.  Body of sternum broader than long; lateral processes widely diverging Dinornis.  Body of sternum quadrate; lateral processes not widely diverging Anomalopteryx. No coracoid depressions.  Body of sternum broad; lateral processes widely diverging Pachyornis.  Body of sternum quadrate; lateral processes not widely diverging.   Pneumatic depressions deep Meionornis.   Pneumatic depressions none Euryapteryx.