254 PROFESSOR H. G. SEELEY ON THE STRUCTURE, ORGANIZATION, 
in tlie first four about 3 ’5 centims. ; the fifth is imperfectly preserved, but probably 
shorter. The antero-posterior extent of the basal expansion of each bone is about 
2’4 centims., convex from front to back. It is also convex transversely, and wider 
behind than in front; the posterior transverse measurement in the second bone is 
1'5 centim., the corresponding anterior measurement is about 1 centim., but in the 
later bones the surface appears to be becoming a rounded boss. 
The neural arch is comparatively small, and is rapidly reduced in size in the later 
vertebrse. The strong pre-zygapophysial processes extend upward and forward as far 
as the middle of the centrum of the preceding vmrtebra; they are convex externally, 
with the articular facet vertical, and looking inward, so as to receive the post- 
zygapophysial wedge between them. This wedge in the earlier caudals is about 
1 ‘8 centim. wide ; the bone above it rises into a neural spine, which is broken away, 
except for an indication that its moderately convex sides converged forward into a 
sharp anterior ridge which inclines backward as it ascends. The transverse width 
over the pre-zygapophyses in the earlier caudals was about 4 centims. 
As the rounded pre-zygapophysial ridge descends it becomes constricted a little 
from side to side, and deeply concave on its anterior margin, to form the inter¬ 
vertebral foramen, and less concave on its posterior border, and then is directed 
outward, to form the base for the transverse process. This process is here a separate 
ossification or caudal rib which is attached high upon the side of the vertebra, partly 
on the neural arch and partly on the centrum ; the depth of this attachment appears 
to increase to the fourth and fifth caudal vertebra, where it exceeds 3 centims. The 
third and fourth processes are between 5 and 6 centims. long. The superior surface 
of the process is convex from front to back, and narrows as it extends downward and 
outward to the compressed free end, which is I'o centim. wide, and less than half a 
centimetre thick. Its anterior and posterior margins are sharp, with these ridges placed 
superiorly, so that the inferior surface is the more convex. In the later vertebra) the 
transverse processes appear to be inclined a little backward ; they obviously have 
small basal attachments, but appear to have been lost from tbe last five ver¬ 
tebrse. The centrum was deeply concave in the early vertebi'se, but there is an 
appearance like that seen in the vertebrse referred by Owen to D. tigriceps (Plate 16, 
fig. 2), only less marked, as though the substance of the notochord were in process 
of ossification, a condition which I regard as showing that relative depth or flatness 
of the articular face of the centrum can have no value as a generic character 
(Plate 16, fig. 3). The articular face of the last centrum preserved is only slightly 
concave, with a moderate central depression. 
The neural canal is small, but it does not appear to decrease in size as it extends 
backward. 
All record of the sacrum with which the specimen was associated had been lost. 
But, at my request, Mr. A. Smith Woodward, F.G.S., made inquiries from which it 
results that Mr. Barlow, the “mason” wdio developed the specimens, indentifies the 
