PROFESSOR OWEN ON THE GENUS DINORNIS. 371 
in a slight degree. Here the upper surface is almost flat, with a feeble mid linear 
rising; where this rises to the anterior convexity there is a linear slit on each side, 
leading to the “ ileo-neural ” vacuities. 
Fragmentary as is this portion of pelvis, it permits of the deduction of some cha- 
racters which support the reference of it to the porphyrian genus Notornis. The 
antacetabular portion included, as I am led to infer, at least half the total length of 
the pelvis. In this respect it resembles Apteryx’; but the superior contour of the 
coalesced iliacs in that bird is convex. The undulated contour of the same part of 
Notornis rather exaggerates that in Aptornis (Pl. XLII. fig. 1), and still more so 
what is more feebly shown in 7ribonyx and some other Rallines, viz. Gallinula nesiotis, 
Scl., Porphyrio, sp., and Ocydromus, sp. But in none of these is the undulation so strong 
as in Notornis. In Gallinula nesiotis I observe the same slit-shaped outlets of the ilio- 
neural canals on each side the descending mid ridge of the coalesced ilia. 
The character of the under surface of the antacetabular centrums in Notornis 
(Pl. XLII. fig. 5) agrees with that in species of Gallinula and Porphyrio. ‘The ridge- 
like shape of the same part in Aptornis and Cnemiornis finds a partial resemblance in 
the first and, sometimes, second sacral centrums of species of Yulica. Only the first sacral 
supports, as in Notornis, moveable ribs in Gallinula nesiotis and Porphyrio celestis ; 
but in some species of Fulica both first and second sacrals have moveable ribs, as in 
Aptornis and Cnemiornis. In the vertebre which I have called ‘ interacetabular,” 
though they are partly in advance of the internal openings of those articular cavities, 
Notornis resembles Aptornis and many existing Rallides, in the absence of transverse 
processes; the neurapophyses rise, from above the double nerve-holes, as vertical walls 
slightly expanding to be lost in the flat iliac roof of that part of the pelvis. In Cnemi- 
ornis oblique ridges extend from the neurapophyses of those vertebra, strengthening, 
as buttresses, the support of the superincumbent bony roof’. 
§ 5. Femur, Tibia, and Fibula of Aptornis defossor. 
Femur.—The difference in the size of this bone, as compared with the femur of Aptornis 
otidiformis, is given in the “ Table of Admeasurements” (p. 378), and may be seen by 
comparing fig. 5. Pl. XLIII. with those cited below*. The differences which the femur 
of Aptornis presents, as compared with that of Dinornis*, are repeated, and in some 
respects exaggerated, in Apt. defossor: the straight subcylindrical character of the pro- 
portionally longer and more slender shaft in Aptornis is better marked in the present 
species. That of Apt. didiformis, viewed in profile, as in the figure of the femur of the 
larger species given in Pl. XLIII., shows a slight bend, convex forward; this is due to 
a minor diminution of fore-and-aft diameter in the lower part of the shaft, and to a 
relatively greater fore-and-aft diameter of the outer condyle in Apt. defossor. The 
1 Trans. Zool. Soc. vol. ii. pl. 54. ? Th. vol. y. pl. 64. fig. 6. 
> Ib. vol. iv. pl. 3. fig. 3; vol. v. pl. 63. fig. 3. ‘4 Tb. vol. iv. p. 10. 
3E 2 
