370 PROFESSOR OWEN ON THE GENUS DINORNIS. 
I no longer delay, therefore, to communicate a description of a very mutilated portion 
of pelvis received from the sand-beds at the embouchure of the Wanganui river, North 
Island of New Zealand, in the same brittle but unpetrified state as the parts on which 
the genus Notornis was originally founded', under the impression of its being extinct, 
as in the North Island of New Zealand it actually is. 
I am led to refer the specimen about to be described to that genus by reason of the 
relation of its size to that of the skull described and figured in the under-cited Memoir, 
and also to the size of the femur and tibia described and figured in a subsequent 
Memoir’. 
This portion of pelvis includes thirteen confluent sacral vertebra, the rest being 
broken off from the hind end of the series. 
The first sacral offers the usual articular surfaces on the prezygapophyses (P!. XLII. 
fig. 6,z) and on the fore part of the centrum, the transverse concavity of the latter 
(ib. c) being deep, the vertical convexity slight. A circular costal pit (ib. fig. 6, pl) 
impresses each side of the centrum and each diapophysis. A large pneumatic foramen 
opens at the base of each diapophysis. The neural spine, moderately compressed and 
high, is confluent at top with the iliacs: the vertical length of the vertebra is 1 inch 
5 lines; the transverse diameter across the diapophyses is 1 inch. ‘The sides of the 
centrum converge below to a tract from 1} line to 2 lines in breadth (Pl. XLII. 
fig. 5,¢). The second sacral has no free rib; its transverse process, directed outwards, 
contracts an extensive anchylosis with the ilium; the centrum, expanding backward, 
has a broader convex under surface. That of the third sacral continues the expansion 
with a broad, smooth, convex under surface; the lamelliform transverse processes 
incline forward to their terminal coalescence. The more expanded fourth sacral cen- 
trum has a broad, flat under surface, on a level with which the thick, short parapo- 
physes pass directly outward. ‘Those of the fifth and sixth sacrals have a like position, 
size, and direction. The breadth of the flat lower surface is here half an inch. The 
four succeeding vertebre are “interacetabular,” have no parapophyses; the inferiorly 
flattened centrums gradually lose breadth, and are defined by the nerve-canals in pairs, 
as usual, opening obliquely backward, and progressively decreasing in size from the 
seventh to the eleventh (Pl. XLIII. fig. 9,¢,7-11). In this and the succeeding centrums 
the parapophysis reappears, but is broken away in each of the remaining sacrals. 
As much of the iliacs remain as have coalesced with the neural arches of the thirteen 
sacrals here preserved. ‘They meet and form an obtuse continuous smooth ridge above 
the first five sacrals, from which ridge the bones slope, like the sides of a steep roof, to 
their lower, fractured margin, The longitudinal contour above (ib. fig. 9) is slightly 
convex, then rather abruptly sinks to the lower level of the expanded neural arches of 
the seven or eight hinder vertebrae, describing a concavity, which again becomes convex 
' On Diyornts (Part III.): Trans. Zool. Soc. vol. iii., “‘ Cranial characters of the genus Wotornis,” p. 366, 
pl. 56. ? Trans. Zool. Soc. vol. iv. p. 12, pl. 2. figs. 3 & 4. 
