PROFESSOR OWEN ON THE GENUS DINORNIS. 151 
The postaxial surface (figs. 4 and 6, pc) has reverse proportions to the preaxial one, 
the longest diameter being vertical. It is divided into a pair of narrow vertically 
concave facets by a still narrower medial tract, the transverse contour being thus rather 
angular than, as in Struthio and most birds, convex. The under border is nearly straight, 
and the transverse extent of the neural margin slightly exceeds that of the hemal one. 
The articulation slopes downward and backward. The hypapophysial process (figs. 4, 5, 
6, 7, hy) is relatively more produced hzmad than in Struthio, and descends vertically and 
slightly backward from the postaxial surface, thus adding to the length of the vertebra. 
The neural canal (figs. 5 & 6, 7) is half the length of the entire vertebra, and its width 
is nearly one third of the breadth. Its area is a full ellipse with the long axis vertical, 
not transverse as in Struthio. 
The ridges on the fore part of the hemal surface of the centrum in Struthio! are 
not present in Dinornis maximus or in D. robustus. The parapophysis is represented 
by the short obtuse ridge, p, figs. 4 & 7; the pleurapophysis (ib. p/) by a similar one 
above, near the middle of the outer wall of the vertebrarterial canal, v. From the 
diapophysial plate (fig. 6, d), completing that wall above, a ridge (fig. 4, 7) extends back- 
ward to the postzygapophysis, 2’; this ridge circumscribes externally the vertical canal 
(figs. 4 & 7, s). On the medial side of this canal is the pneumatic orifice (fig. 7, ¢), 
leading to the cancellous part of the neurapophysis. The canal s is not noted in 
Struthio, and in place of one large hole there are irregular pneumatic foramina?. The 
postzygapophysis is thrice the size of the prezygapophysis, but the antero-posterior hardly 
if at all exceeds the transverse diameter; its aspect is as in Struthio and birds gene- 
rally. The hyperapophysis (figs. 4 & 6, hp) is relatively more prominent than in 
Struthio. The neural spine gains thickness as its base extends backward; its summit 
is broken off in my specimen, exposing the wide-celled pneumatic texture. 
In Dinornis elephantopus the hyperapophyses (fig. 5, hp) are relatively larger and 
higher ; the prezyapophyses are relatively less. 
The odontoid process (figs. 4 & 7, ca) is less than half the size of the atlantal hyp- 
apophysis; its free extremity is obtusely rounded. 
The third cervical vertebra, as in Struthio and birds generally, gains a transverse 
breadth of the neural arch, anteriorly, with concomitant size of the prezygapophyses 
(fig. 11, z), fitting the postzygapophyses of the axis. From this gain results a quadrate 
form of the roof of the vertebra; but, from the less relative length, or greater breadth, 
of this part in Dinornis, the roof (fig. 11) is a transverse quadrilateral, not so oblong as 
in Struthio: in both birds the angles are rounded off 3. 
The centrum is a horizontal wedge, with the edge anterior, concave, and formed by 
the neural border of the preaxial articular surface (fig. 9, ac), which, broad and con- 
cave transversely, is short and almost flat vertically, but here slopes from the vertical 
so much backward that it is on a plane with the contour of the hemal surface of the 
* Mivart, loc. cit. p. 391, fig. 12, ¢. * Th. loc. cit. p. 393. Ib. loc. cit. p. 395, fig. 15, 
