540 



PEOFESSOE OWEN ON THE FOSSIL MAMMALS OF AUSTEALIA. 



(for the anterior spinal nerve or a division of it) on each side of the base of the neural 

 arch, the vertebrarterial canal being, as above described, an open groove. In the atlas of 

 proboscidian and of other large placental Mammals, which the present fossil resembles 

 in size, the diapophyses are widely bored by the vertebral artery, which usually perforates 

 also the fore part of the neurapophysis*. 



I recognize, therefore, in the portion of the atlas vertebra, here referred to Dipro- 

 todon, marsupial characteristics ; compared with that of the Kangaroo, its diapophyses 

 are relatively shorter, thicker, terminally more obtuse, not so much expanded or depressed 

 at that part, upon the whole more resembling those in the Wombat and Koalaf . 



The axis or vertebra dentata (Plates XLIII. fig. 1, & XLIV. figs. 1, 2, 3) is entire save 

 the ends of the diapophyses, which have been broken away. The length of the body, with 

 the odontoid process, is 6 inches 3 lines, the height of the vertebra is 8 inches 4 lines, 

 the breadth across the anterior articular surfaces is 5 inches 9 lines. The size of this 

 vertebra thus equals that in the largest Rhinoceros or Hippopotamus, and in length that 

 of a full-sized Elej)hant. The hind surface of the centrum (Plate XLIV. fig. 3, c) is 

 flat, rather rough, transversely elliptic, with a tendency to an angular or octagonal outline. 

 The under surface (ib. fig. 2) expands as it advances to develope the bases of the par- 

 apophyses (ib. pp) ; contracting in advance of these it again expands into the anterior 

 articular processes (ib. and fig. 1, z, z). A low hypapophysis (ib. fig. 2, hy) of a subtri- 

 angular form projects from the middle of the under surface towards the fore part. The 

 anterior articular surfaces (Plates XLIII. & XLIV. fig. l,z) converge to the base of the 

 coalesced body of the atlas, called " odontoid " process. This element (Plate XLIII. 

 fig. 1, ca, and Plate XLIV. fig. 2, ca), 2 inches in length, 1 inch 6 lines in breadth, and of 

 similar depth, is convex transversely and longitudinally below , it has a pair of slightly 

 concave roughened surfaces, meeting above, along the anterior sloping half (Plate XLIII. 

 fig. 1, m), behind which the upper surface rises into a low broad tuberosity (ib. t), bound- 

 ing anteriorly a smoothish elliptical surface (ib. r) occupying the upper part of the 

 rest of the odontoid. A broad deep irregular depression (ib. 6) divides this surface of 

 the odontoid from the anterior articular surfaces of the axis. These surfaces (ib. and 

 Plate XLIV. fig. 1, z, z), of a full oval shape, 3 inches in diameter, are moderately convex. 

 The neurapophyses (Plate XLIII. w), after developing the diapophysis (ib. cZ), contract to 

 a fore-and-aft extent of 1 inch 9 lines, then expand backward to develope the postzyga- 

 pophyscs (z'), in advance of and between which the neurapophyses converge and coalesce 

 to form the base of the neural spine (ib. ns). This expands both forward and backward, 



* Osteological Catalogue of the Museum of the Royal College of Surgeons, 4to, 1853, p. 475, no. 2678, ♦ Atlas 

 of Elephant' (by misprint the vertebral artery is called " raeduUaiy ") ; p. 509, no. 2945, ' Atlas of Rhinoceros 

 bicornis,' " the vertebral artery perforates the diapophysis and then also the neural arch ; " p. 560, no. 3404, 

 Hippopotamus amphihiiiit, " the transverse processes are i)crforated by the vertebral arteries." 



t I am led to believe, after fresh study of Diprotodont fossils, that the one ascribed to a calcaneum in my 

 ' Catalofjue of Fosail Mammalia in the Museum of the lloyal College of Surgeons ' may bo a fragment somewhat 

 rolled and worn of the atlas vertebra. 



