DR. J. VON HAAST ON A NEW SPECIES OE DINORNIS 175 



one. Towards the middle of the series the latter are so much alike that it is only with 

 great difficulty one can assign to them their proper position, especially if they belong 

 to two different sizes of the same species (male and female), or even to allied species. 

 However, the vertebrae from the eighteenth to the twenty-first are very characteristic, 

 and not to be mistaken. I have selected for illustration the last, or seventh, of the 

 dorsal series. 



Atlas. 

 This bone is - 64 inch broad, - 76 inch high, and - 24 inch long. A comparison 

 with the atlas of Dinornis parvus, figured by Sir Richard Owen in vol. xi. pi. li. of the 

 ' Transactions ' of this Society, will show that in Dinomis oweni, even allowing 

 for difference of size, it is much smaller and more slender. This is especially 

 shown at and near the summit of the neural arch where the small articular surfaces 

 are situated. The neural arch is complete at its summit just as in Dinornis parvus. 

 Sir Richard Owen is quite right in surmising that the non-union of the sides of the 

 neural arch, leaving a longitudinal fissure at the upper midline in the atlas belonging 

 to Dinornis robustus (as figured on pi. lxii. figs. 4, 5, op. cit.), may be traced to the 

 fact that it belongs to an immature individual. The Canterbury Museum possesses the 

 atlas of Dinornis robustus as well as those of Dinornis maximus and several other larger 

 species of the Dinornithidse, all evidently belonging to mature individuals, in which the 

 neural arch is always closed. 



Seventh or last Dorsal Vertebra. 



This vertebra, of which the lateral and postaxial aspects are illustrated by figures 

 10 and 11 (Plate XXXI.) for comparison with Sir Richard Owen's figures of the same 

 bone of Dinornis robustus (p. 414, op. cit.), shows, notwithstanding its diminutive 

 size, that it exhibits in every particular the true dinornithic type. 



The neural spine, unfortunately partly broken off, possesses on the remaining anterior 

 and posterior surfaces the rough tract for the attachment of the elastic ligaments. Both 

 the post- and prezygapophyses possess a nearly circular articular surface, which in 

 Dinomis robustus has an oblong form. The neural canal, large for the size of the 

 species, is not so vertically depressed as it is in Dinornis robustus ; in fact it approaches 

 a circular form. The postaxial surface, in comparison to size larger than in Dinomis 

 robustus, is much more produced : on its upper and outer border the parapophyses 

 with the capitular concavity are situated ; moreover, its transverse convexity is much 

 greater than its vertical concavity. The preaxial surface subquadrate, and higher than 

 it is broad, is more flattened in both directions ; a pair of small hypapophyses are 

 produced on both sides at its lower termination. The under surface of the centrum is 

 round, with no sign of a keel. 



vol. xn. — part v. No. 5. — December, 1883. 2 d 



