PROFESSOR OWEN ON THE GENUS DINORNIS. 169 
to those marked f! in Mivart’s figure 57, of the last dorsal vertebra, are here well 
marked. 
In the sixth dorsal (twenty-first vertebra, figs. 31, 32) the hypapophysis is sup- 
pressed, as in the twenty-second (fifth dorsal) vertebra of Struthio (‘ Mivart,’ figs. 49-51)}, 
with which the present will be compared. The articular surface (fig. 32, ac) thus 
almost “entirely occupies the preaxial end of the centrum ”?: only a few lines breadth 
on each side of the neural half of that surface is non-articular in Dinornis, and may be 
ascribed to the fore part of the parapophysis (ib. p). The vertical as compared with the 
transverse diameter of the preaxial surface is greater than in Struthio. The cha- 
racteristic height of the neural spine in Dinornis (figs. 31, 32, ns) is still more marked 
in this comparison. ‘The pneumatic orifice (fig. 31, pn) between the par- and di-apo- 
physes is the chief one for admission of air into the vertebral substance; but a small 
homologue (ib. pn’) of the posterior pneumatic orifice remains. 
The postaxial surface (fig. 31, pc) is absolutely and much more relatively approxi- 
mated to the postzygapophysis (ib. pz) than in Struthio. The neural canal (fig. 32, 2) is 
transversely, not vertically, elliptical (comp. Mivart’s fig. 51). The sides of the preaxial 
surface are much produced, and the transverse concavity of that surface is proportionally 
deepened. ‘The lower border of the postaxial surface is more produced than in Struthio, 
rendering the lower contour of the centrum in Dinornis more concave (comp. fig. 31 
with fig. 49, Mivart, doc. cit.). ‘The zygapophysial surfaces are relatively more ex- 
tensive in Dinornis, the dorsal vertebre being more securely interlocked in the larger 
terrestrial bird. 
The characteristically broad and massive proportions of these vertebre in Dinornis 
are well brought out in comparing figs. 27-32 with figs. 47-51 of Mivart, loc. cit. 
The minor length and greater thickness of the diapophyses, d, and the much greater 
development of the neural spine are exemplified in fig. 52 as contrasted with fig. 51 
(Mivart, loc. cit.). 
The vertebra in Dénornis, which answers, in rib-character, to that in Struthio sup- 
porting the eighth pair of movable pleurapophyses, is that which supports the seventh 
pair. In both genera it is the hindmost rib-vertebra not confluent with the sacrum. 
In the present skeleton of Dinornis it is the twenty-second vertebra, counting from the 
occiput ; in Struthio it is the twenty-fifth. Of this Prof. Mivart gives four figures *. 
The chief differential characters of its homologue in Dinornis maximus will be exem- 
plified in the two subjoined cuts from the lateral (fig. 33) and postaxial (fig. 34) aspects. 
In the comparison of figure 33 with figure 54 (‘ Mivart’), the deep longitudinal 
concavity of the under surface (¢) of the centrum may be first remarked, due in Dinornis 
to a downward production of the border of the preaxial articular surface (ac) and a still 
greater production in the same direction of the postaxial surface (pc), augmented by the 
development of a pair of hypapophyses (fig. 34, hy). These are not developed in 
? Loc. cit. p. 414, * Ib. p. 414, 5 Loe. cit. p. 419, figs. 54-57, 
