76 
the alisphenoid and the basisphenoid is preserved, proving by its position 
that the large flange, directed outward from above and somewhat behind 
the basisphenoid process, belongs to and is part of the basisphenoid. 
It may be of interest also to note that in this skull the ophthalmic branch 
of the trigeminal nerve (V) is enclosed in bone in its forward course and 
does not occupy an open channel as it appears to do in Edmontosaurus. 
Further, indicating an unusually perfect preservation of structural detail, 
the separation of the fenestra rotunda from the fenestra ovalis by a hori- 
zontal bar of bone is excellently shown. 
To the preparation of this skull by C. M. Sternberg, its discoverer in 
the field, are due many details of structure that less skilful and painstaking 
work would not have revealed. 
With a better understanding of the stephanosaurinid skull certain 
errors in the description of the skull of Cheneosaurus (Figure 39J) as it 
appeared in the pages of the Ottawa Naturalist in 1917, can now be rectified. 
What was regarded as prefrontal is certainly the expanded prolongation 
backward of the lower, external part of the premaxilla, and the supposed 
sutural line running forward from the lower end of the lachrymal (see 
original figure) is evidently a fracture in the bone. The convex, upper 
surface of the dome is nasal met in front by the upper part of the pre- 
maxillary roofing the nasal passage. The bone above the orbit called 
supraorbital in the first instance is the prefrontal, and the frontal is similar 
to the frontal in both Stephanosaurus and Corythosaurus in being small 
and excluded from the orbital rim by the intervention of the prefrontal 
and postfrontal. 
The posterior height of the skull shows a marked difference in five 
genera of crestless or flat-headed hadrosaurs (Hadrosaurince) from the 
Cretaceous of the west of this continent, viz., in Kritosaurus (horizon 
uncertain = ?Edmonton formation), in Edmontosaurus (Edmonton forma- 
tion), in Gryposaurus (Belly River formation), and in ^‘Claosaurus’^ 
(annectens) and Diclonius both from the Lance formation. In 
Kritosaurus Brown, from the Ojo Alamo beds of New Mexico, 
the quadrate is of remarkable length, in Diclonius Cope, from 
Dakota, it is singularly short, the two representing the extremes of skull 
elevation and depression in the Hadrosaurince {Trachodontince of Brown). 
In these five genera, in all of which, with the exception of Kritosaurus, 
the skull is fully known from excellent material, the proportionate lengths 
of the quadrate and skull may be expressed in numbers as follows: Krito- 
saurus I — over 2; Gryposaurus 1 — 2J; Edmontosaurus I — over 2|; ^‘Clao- 
saurus” 1 — nearly 3; and Diclonius 1 — nearly 4. From this comparison 
it is seen that in Kritosaurus the posterior height of the skull (length of 
quadrate) relative to the horizontal length of the same is the greatest, 
that Kritosaurus, Gryposaurus, Edmontosaurus, and “Claosaurus” form a 
series, in the order named, in which the quadrate is successively reduced 
in length in about the same ratio, and that the greatest difference in the 
height of the skull is found between ^‘Claosaurus” and Diclonius. It 
would appear, therefore, that as time progressed the skull in the Hadro- 
saurince, as a general rule, became lower, culminating in the greatly 
depressed and very long skull of Diclonius in the closing days of the Creta- 
ceous. The posterior height of the skull in Edmontosaurus is greater than 
the average among the genera of flat-headed hadrosaurs in which the head 
is known. 
