361 
bones referable by length to this species, two varieties are indicated by metatarsals! ; 
the longer and more slender form, contributed by Dr, Lillie (ante, p. 252), I have now 
reason to suspect may prove to be a variety so well established and so characterized by 
concomitant modifications in other parts of the skeleton as to need, for convenient 
reference, a distinct name, for which that of Dinornis altus may be accepted. 
In comparing the skulls, indicative of eight or nine of the fifteen species of Dinornis 
characterized by limb-bones, we see that they differ in the relative length of the beak?, 
and in that of the terminal osseous part of the upper mandible* to the rest of the 
skull in the sharper or more obtuse termination of both mandibles‘, in the relative 
antero-posterior extent of the temporal fossw®, in the relative flatness or transverse con- 
vexity of the cranium®, in the longitudinal contour of the upper part of the skull’, and 
in a few such minor particulars as the breadth and direction of the postorbital process. 
The sternum of Dinornis, whilst conforming to the apterygian type*, shows two well- 
marked modifications thereof—in the proportions of breadth to length, and in the 
degree of divergence of the lateral processes®, 
The back toe (hallux, i) was a small functionless appendage to the foot in the best- 
developed condition in which it has been found!®, The trace of the ligamentous 
attachment to the functionally developed metatarsus is feeble in the species in which 
the existence of this toe is most certain, It cannot be inferred to have been wholly 
absent in the cases in which the connecting ligaments have left a less definite trace or 
where such is undistinguishable. I have therefore felt obliged to abandon this ground 
of generic distinction". 
One is naturally disposed to group together the thick-legged or strong-limbed species 
of the South Island represented by the five above cited. It would seem that the 
broader type of sternum was associated with such proportions of the limb-bones in 
those species, If such association should be proved and found to be constant, and if, 
in like manner, the narrower form of sternum shown in Dinornis rheides should be 
found to prevail in the six slender-limbed species, we shall have the same ground for 
restricting to them the old generic term. If it could be shown that Palapterya, us 
' Pl. LXXIX, figs. 4 & 5, 
* Compare Pl. LXIV, fig. 1 (Dinornis robustus) with Pl. LXXVITI. fig. 1 (Dinornis casuarinus). 
* Compare Pl, LXXXII, fig. 1 (Dinornis ingens) with Pl. LXXVI. fig. 1 (Vinornis crassus), 
‘ Compare Pl. LXXVI. figs. 2, 3, 8, & 9 (Dinornis crassus) with Pl, LAXVIT. figs. 6-9 (Dinornis ete- 
phantopus), ; 
* Compare Pl. LXII. fig. 1 (Dinornis robustus) with Jaeger, Paliiontologie der Novara Expedition, art. vi. 
Bericht iiber einen fast vollstiindigen Schiidel yon Palapterya ingens, Tat. xxvi. fig. 2. 
* Compare Pl. LXXVII. fig. 2 (Dinornis elephantopus) with Pl. LIL. tig. 1 (Dinornis giganteus). ; 
7 Compare Pl. LXIV. fig. 1 (Dinornis robustus) with Pl, LXAXVI. fig. 1 (Dinornis crassus), or with 
Pl. LXXVIIT, fig. 1 (Dinornis casuarinus). 
* Ante, p. 34, Pl. IX. figs. 2 & 3, and p, 254. ie 
* Compare Pl, LX XU. fig. 1 (Dinornis elephantopus) with Pl. LX XII fig. 1 (Dinornis rheides). ; 
” Pl. XLIX, figs. 4 & 5 (Dinornis robustus). 1 Agta, ps 28s, 
