415 
in that vertebra, The postaxial surface is more definitely subquadrate, with the 
angles rounded off and the upper and lower borders emarginate (fig. 34, pc). The 
transverse convexity is not greater than the vertical concavity; both are feeble, so that 
the entire surface approaches to flatness; and in a duplicate homologue the flattening 
is greater than in the specimen figured. In both the surface has lost its synovial 
smoothness, through suppression of motion upon the first sacral vertebra. ‘The trans- 
verse dimension does not exceed, as it does in Struthio, the vertical one. 
The neural canal (fig. 54, n) is more depressed than in the twenty-first vertebra, and 
still more deviates from the form shown by the hinder outlet in Mivart’s figure 56 of 
the Ostrich. The parapophysis in Dinornis (fig. 53, p) is represented merely by the 
raised margin of the capitular concavity. The diapophysis is less massive in proportion 
to the rest of the vertebra, and especially the neural spine, than in the antecedent 
dorsal. ‘The neural spine is not carinate along either the fore or the hind border; both 
present a flat rough surface, about two thirds the breadth of each smooth lateral 
surface. A transverse section of the spine thus gives an oblong quadrate figure. A 
pair of depressions at the fore part of the base of the spine intervene between it and the 
prezygapophyses; they answer to the ‘ antero-lateral fosse,’ f1, in Mivart’s figures 59, 
57. A narrower pair of fosse hold a like relation to the postzygapophyses, answering 
to those marked f* in figs. 56 & 67 (‘Mivart’). The fossw, /*, ib. ib., are feebly, if at 
all, represented in Dinornis. | 
The pleurapophysis retains its twofold articulation, but has lost in length; its 
hemapophysis is attached to that of the preceding segment, and this element fails to 
reach its spine (sternum) in the fifth and subsequent dorsals. 
The sternum may be considered, archetypally, as a coalescence of four or more such 
hemal spines, the foremost retaining its connexions with its heemapophyses, which 
are expanded in Strwthio and in birds of flight as ‘coracoids;’ but in Dinornis the 
‘coracoids’ retain the slender proportions of the true thoracic hemapophyses. They 
are also here confluent with their pleurapophyses, which, detached asa ‘ scapula’ from 
its proper centrum, has the proximal end free without articular head, and in Dinornis 
is reduced to the normal form of a rib with diminished proportions. 
Retaining these views of the ‘general homology’ of the sternum, I find its proper 
place of description at the part of the axial skeleton here attained. 
The sternum belonging to the skeleton of Dinornis maximus under description has 
suffered some mutilation; but a detached example of the bone, transmitted from New 
Zealand to Edinburgh (Plate XCVIII.), shows a unique condition of integrity. 
Still regarding, after long practice in the interpretation of avian fossils, the sternum 
as one of the most characteristic and taxonomically instructive parts of the skeleton of 
the bird, I append figures of the natural size of this most perfectly preserved specimen 
of the bone, which is referable to the largest of the known species of Dinormis’. 
| This statement is made on the faith of the sternum transmitted with the rest of the skeleton of Dinornis 
