422 
vertebra of Struthto (Mivart’s ‘lumbar vertebra’), differs in presenting an unmistakably 
rib-like pleurapophysis, although unanchylosed. The fourth sacral in Dinornis is the 
first which may be said to “ present no indication of a rib,” and which would be entitled 
to the term “lumbar,” according to such character. I view, however, the parapophysial 
element of this transverse process as more probably the serial homologue of the cervical 
part of the preceding pleurapophysis. 
With this explanation the neural arch of the fifth sacral vertebra in Dinornis, as in 
Struthio, advances and crosses the interspace between its own and the preceding cen- 
trum; and the thirteenth vertebra is that in which the arch resumes its normal con- 
nexions, Thus the interlocked part of the sacrum in Dinornis is more extensive than 
in Struthio, and relates to the heavier mass which the pelvis had to transmit upon the 
femora, 
‘The antacetabular part of the sacrum (1st to 6th vertebra in Dinornis) is relatively 
shorter and broader than in Struthio; the postacetabular part is still broader in pro- 
portion to its length; and this part is shorter than the antacetabular part, instead of 
being, as in Struthio, longer. 
More striking differences are presented by the pelvis as a whole (Plates XIX., XX., 
XX.a, XCVI. fig 4). The antacetabular plate of the ischium is relatively longer; 
the postacetabular part is shorter, but much broader in Diwornis than in Struthio 
(Plate XIX. fig. 4): the greater relative breadth of the entire pelvis would seem to 
relate to the larger proportional size of the egg in Dinornis. 
The ischium is shorter and deeper than in Struthio: it unites with the ilium 
anteriorly to bound there the ischiadie notch, which remains open posteriorly, as in 
Struthio, and is not circumscribed by a second terminal union of the ischium with 
the ilinm, as in Dromaius. ‘The obturator interspace, closed behind, as in Struthio, by 
ischial confluence with the pubis, and having its fore part defined by the descending 
process of the ischium, is much narrower in Dinornis, as in Apterya. The pubis does 
not send off so long and well-defined a ‘ pectineal process’+, as in Struthio; its body 
extends backward parallel with the ischium, slightly concaye downward, and ter- 
minates in the vertical expansion joining the ischium without being continued down- 
ward and forward to meet its fellow at the symphysis, a structure which is peculiar, 
among birds, to the genus Struthio. 
The type of the pelvis in Dinornis is that of the Apteryx, not of the Emu or Cas- 
sowary; it differs therefrom in less marked modifications than from the pelvis in 
Struthio and hea. 
The number of terminal sacral vertebrae in Dinornis maximus, answering to those 
defined as ‘sacro-caudals’*, is four. The last of these in Dinornis is the thirty-ninth of 
the vertebral series; in Struthio it is the forty-sixth. 
' For this process in Apteryx australis, sce p. 30, Pls, VIII, & LX. In the two skeletons of the smaller 
Kiyi (Apterye Owenw, Gd.) L have found ossification extending along the ligament attaching the pectincal 
process to the last sacral rib. * Mivart, uf supra, p. 426, fig. 62, 
