Zool. J. Linn. SOC., 55: 151-175.With 5 figures
September 1974
A new heterodontosaurid dinosaur (Reptilia:
Ornithischia) from the Upper Triassic Red Beds
of Lesotho
RICHARD A. THULBORN, F. L. S.
Department of Zoo1og.y. University of Queensland. St. Lucia, Queensland 406 7.
Australia
Accepted for publication March 1974
(Lycorhinusconsors sp. nov.) is established on a skull
from the Upper Triassic Red Beds of Lesotho. This ornithischian is assigned to the family
Heterodontosauridae of the suborder Ornithopoda. The dinosaurs of the family Heterodontosauridae are reviewed: Grranosaurus atawus Broom (191 1) is considered anomen dubium and the
genus name Heterodontosaurus Crompton & Charig (1962) is held to be a junior synonym for
Lycorhinus Haughton (1924).
Functional and palaeoecological implications of the heterodontosaurid dentition are
discussed. The pattern of tooth wear may reflect a highly specialized jaw action which involved
protraction and retraction of the mandible to produce a grinding effect between upper and
lower cheek teeth. Lycorhinus consors is presumed to be a female heterodontosaurid because it
differs from all other hcterodontosaurids in lacking caniniform tusks. It is suggested that the
tusks of heterodontosaurids were functionally analogous to those of tayassuids and tragulids
and that they were employed as weapons for intra-specific combat and defence. Dental
peculiarities indicate that tooth replacement processes were suppressed in hcterodontosaurids;
replacement of the teeth seems to have been restricted to a brief period each year (presumably
when heterodontosaurids underwent aestivation or hibernation).
A new diagnosis is formulated for the family Heterodontosauridae. The relationships of
early ornithopod dinosaurs are briefly reviewed and a new classification is proposed. Ten
families of ornithopod dinosaurs are recognized; these are ranked in two grades-one (named
Dolichopoda) representing the conservative main stem of the ornithischian phylogenetic tree
and the other (named Brachypoda) comprising the several more advanced lines of ornithopod
evolution.
A new species of ornithischian dinosaur
CONTENTS
Introduction
. . . . . . . . . . . . . . . . . . . . . .
Material and methods
. . . . . . . . . . . . . . . . . . .
Systematics
. . . . . . . . . . . . . . . . . . . . . .
Description
. . . . . . . . . . . . . . . . . . . . . .
Skull bones
. . . . . . . . . . . . . . . . . . . .
Dentition
. . . . . . . . . . . . . . . . . . . . .
Discussion
. . . . . . . . . . . . . . . . . . . . . .
Review of heterodontosaurid dinosaurs
. . . . . . . . . . . .
Relationships of Lycorhinus consois
. . . . . . . . . . . . .
Functional and palaeaecological implications of the heterodontosaurid dentition
Classification of ornithopod dinosaurs
. . . . . . . . . . . .
Acknowledgements
. . . . . . . . . . . . . . . . . . . .
References
. . . . . . . . . . . . . . . . . . . . . .
11
151
152
152
153
154
154
158
159
159
161
162
167
174
174
152
R . A. T H U L B O R N
INTRODUCTION
Triassic ornithischians of the family Heterodontosauridae stand apart from
all other dinosaurs on account of their markedly heterodont dentitions, which
typically include large canine-like tusks. Heterodontosaurids are known from
rare and generally incomplete fossils, perhaps the most familiar specimen being
the skull which was briefly described by Crompton & Charig (1962) under the
name Heterodontosaztrus tircki. Other heterodontosaurids (Lycorhinus
aiigustideizs Haughton, 1924 and, possibly, Geranosaurtts atavus Broom, 1911)
are recorded solely from cranial fragments. The skull described below
represents a new and very distinct species of heterodontosaurid dinosaur. This
new specimen augments knowledge of the heterodontosaurids in general and
has particularly interesting implications for the palaeoecology and classification
of early ornithischians.
MATERIAL AND METHODS
The account is based on a single specimen, numbered B54, which is
preserved in the collection of the Zoology Department at University College
London. This specimen was collected by K. A. Kermack and Frances Mussett
during the course of an expedition (1963-64) from University College to
Lesotho (then Basutoland). Specimen B54 came from the Red Beds formation
of the Stormberg Series between the settlements of Whitehill and Qacha’s Nek
in southern Lesotho (see map, Fig. 1). At the village of Noosi, 5.1 miles east of
Whitehill, the road to Qacha’s Nek crosses a stream running northwards into
the Orange River; the specimen was taken from the bank of this stream, just to
the north of the road and about 20 feet below a small waterfall. The parent
horizon is just below the base of the Drakensberg volcanics and there can be no
doubt that specimen B54 is of Upper Triassic age. Greater stratigraphic
precision is not possible (see Thulborn, 1972: p. 29).
Specimen B54 comprises the skull and articulated post-cranial skeleton of a
single animal. The bones are enclosed in a soft, fine-grained sandstone of dull
wine-red colour, which is interbedded with thin and persistent seams of dark
red shale. Veins and streaks of pale green sediment seem, in places, to form a
distinct cortex over the bones in the specimen. Such local discoloration of the
matrix may be attributed to the reducing action of organic acids liberated
during decay of the soft tissues (Pettijohn, 1957). Much of this soft matrix was
removed with a light hammer and small cold chisels; individual bones were
exposed by lifting off the matrix with a mounted needle. Nearly every bone is
traversed by a network of fine cracks-comparable with the “checkering”
observed by Simmons ( 1 965) in reptiles from the Chinese Trias. In a specimen
of the Red Beds ornithischian Fabrosaurus australis similar “checkering” has
been interpreted as sun-cracking acquired by the bones before their burial
(Thulborn, 1972). But in specimen B54 the bones seem t o have been buried
while enclosed in the soft tissues (as evidenced by their undisturbed state and
by discoloration of the matrix) and their fissures, which extend into the
surrounding sediment, can only be the result of recent weathering. To prevent
their further disintegration the bones in specimen B54 were coated with a thin
(1 : 4) solution of polybutyl methacrylate in ethyl acetate. This transparent
protective coating may be washed off with ethyl acetate.
153
A NEW HETEKODONTOSAURII) DINOSAUK
I
r1 '
LOCALITY
\
\
fl
\
I
50 MILES
Figure 1. Maps showing provenance of specimen B 5 4 (Lycorhinus consors sp nov., holotype).
Shaded rectangle on small map indicates area shown in large map.
SYSTEMATICS
Class Reptilia
Order Ornithischia
Suborder Ornithopoda
Family Heterodontosauridae Kuhn, 1966
Genus Lycorhinus Haughton, 1924
Type species Lycorhinus angustidens Haughton, 1924. Haughton, S. H., 1924:
343-4, fig. 8.
Lycorhinus consors, sp. nov.
(Figs 2 to 4)
Etymology. Latin censors (spouse), in allusion to suspected feminine gender
of the holotype.
Type specimen. Skull with partial skeleton, numbered B54 and preserved in
the Department of Zoology, University College London.
Type locality. Streamside exposure at the village of Noosi, 5.1 miles east of
Whitehill, southern Lesotho.
Stratigraphic horizon. Dull red sandstone at top of the Red Beds formation,
Stormberg Series. Upper Triassic.
*Diagnosis. Unarmoured ornithischian about 1 m long, with slender and
hollow limb bones. Skull deeply triangular in profile, laterally compressed, with
* The diagnosis includes such evidence as is immediately apparent from the partly-prepared
post-cranial skeleton.
154
R . A. THULBORN
large circular orbits at sides. Antorbital vacuity triangular, widely open. Nasals
and frontals narrow, largely confined to skull roof. Premaxilla separated from
maxilla by wide arch-like diastema, extended into robust postnarial ramus.
Lateral face of maxilla vertically concave, without longitudinal ridge. Mandible
fairly stout, with low and obtusely angular coronoid eminence, without lateral
fenestra. Dentaries inflated at front, separated by small and toothless
predentary. Dentition heterodont; implantation thecodont; teeth in simple
linear series. Premaxilla toothless in front, with 2 small conical teeth at rear,
without caniniform tooth. Maxilla with 12 teeth; mandible with 14 teeth,
without caniniform tooth. Crowns of cheek teeth latero-medially compressed,
with small divergent denticles at occlusal margins and with feeble ridges on
buccal surfaces. Tooth wear planar; each cheek crown with a single, flat and
steeply inclined wear facet which is continuous with similar facets on preceding
and succeeding teeth. Sacral vertebrae four in number, not fused. Fore limb
much smaller than hind limb; manus diminutive, probably with phalangeal
formula 2: 3:4:2: 1; metacarpal I1 longer than metacarpal 111. Femur stout,
forwardly arched, with pendent fourth trochanter entirely confined to
proximal half. Distal parts of metatarsals 111 and IV deflected laterally;
metatarsal 111 half as long as tibia. Foot very large, with phalangeal formula
2: 3 :4:5 : O ; digits clawed.
DESCRIPTION
The bones of the post-cranial skeleton are so shattered that their complete
preparation and interpretation will take some considerable time. Consequently
the following account deals only with the skull and the dentition. The
description is based mainly on the bones of the better preserved (left) side.
Skull bones
(Figs 2 and 4)
The skull is severely crushed and lacks the occipital region together with
much of the skull roof. It has a length, from the snout to the ventral tip of the
quadrate, of 79 mm and is typically archosaurian in appearance-triangular in
profile, with a large circular orbit and an extensive antorbital vacuity. In its
shape and basic construction the skull resembles that of Lycorhinus tucki"
rather than that of Fubrosaurus australis (see Crompton & Charig, 1962: fig. 1
and Thulborn, 1970b: text-figs 1 & 8 ) .
Below the narrow and almond-shaped external naris the premaxillu is about
5 mm deep and is gently convex in a vertical direction. The premaxilla is nearly
11 mm long at the tooth row and forms a rounded, overhanging and hood-like
tip to the snout. The pre-narial ramus is represented merely by a slender stump
of bone and probably did not interpose between the nasals for any great
distance. The post-narial ramus is much broader and extends postero-dorsally
towards the lacrimal at about 55" relative to the line of the tooth row. Though
the post-narial ramus is incomplete its robusmess may well indicate that it
reached the lacrimal and separated the nasal from the maxilla. The premaxilla
* My reasons for regarding the name Hererodonrosaums Crompton & Charig (1962) as an invalid
junior synonym for Lycorhinus Haughton (1924) are explained elsewhere (Thulborn, 1970a) and are
summarized in a later section of this paper.
A NEW HETERODONTOSAURID DINOSAUR
155
of Lycorhinus tucki is developed in very similar fashion (see Galton, 1970:
fig. 4) but has the post-narial ramus less steeply inclined relative to the tooth
row (c 35").
The left maxilla has a length of 37 mm. The lateral face of the tooth-bearing
portion is concave from top to bottom and bears no trace of any longitudinal
ridge. Both Lycorhinus angustidens and L. tucki differ in having ridged
maxillae (Thulborn, 1970a). At the tooth row the maxilla is separated from the
premaxilla by a wide and arch-like diastema. Apart from interrupting the upper
tooth row this diastema serves to off-set the tip of the snout in a ventral
direction-so that the inferior margin of the premaxilla lies well below the level
of the maxillary teeth. A short, acute and reflexed dorsal process from the
front of the maxilla defines the antero-ventral angle of the triangular and
widely open antorbital vacuity. The vacuity is bounded postero-dorsally by the
small and triangular lacrimal. This bone is some 5 mm high at the front., acutely
pointed behind, and has its long axis directed up and forwards.
The left jugal is represented mainly by an impression in the matrix. The
anterior ramus of the jugal clearly overlapped the postero-dorsal edge of the
maxilla in normal ornithopod fashion and probably met the lacrimal t o form
the lower rim of the orbit. I t may be inferred, from the pronounced lateral
over-hang at the rear of the maxilla, that the jugal carried the ventro-laterally
directed boss which is known to occur in Lycorhinus tucki (see Galton, 1970:
fig. 4). The sheet-like posterior ramus of the jugal appears to be sandwiched
between the ventral part of the quadrate (medially) and the quadrato-jugal
(laterally), but these bones are very poorly preserved and it is difficult to be
certain of their relationships.
a
or
ni I
'an
Figure 2. Lycorhinus comws sp. nov. Holotype. Skull in left lateral view.
156
R . A . THULBORN
14
Figure 3. Lycorhinus consors sp. nov. Holotype. Left maxillary teeth (above) and left
mandibular teeth (below) in buccal view, x 2 . Teeth numbered from front to back.
\
I
P'
dl
an
Figure 4. Lycorhinus consors sp. nov. Outline reconstruction of the skull in left lateral view.
Shading indicates skull bones actually preserved in the holotype. Unknown parts based on the
figure of Lycorhinus (Heterodontosaurus) tucki as given by Galton (1970).
A NEW HETERODONTOSAURID DINOSAUR
157
The left quadrato-jugal is damaged both in front and below. Its anterior
portion is a thin bar of bone which covers the rear part of the jugal while its
postero-dorsal portion, which overlies the quadrate, is considerably thicker and
appears to bear traces of a vertical ridge. The lower temporal bar is about as
slender as that in Lycorhinus tucki or Fabrosaurus australis and is not as deep
as it is in many post-Triassic ornithopods. The quadrato-jugal has clearly been
displaced, for its rear margin lies well behind that of the quadrate. I t seems
most likely that the quadrato-jugal has been rotated in an anti-clockwise
direction upon the underlying quadrate, and in drawing up the reconstruction
of the skull (Fig. 4) the quadrato-jugal has been rotated in the opposite
direction. Consequently the anterior part has been swung upwards to conceal
the posterior ramus of the jugal and the rear margin of the quadrato-jugal has
been shifted forwards to coincide with that of the quadrate.
The anterior part of the quadrate is expanded into a paper-thin sheet which
occupies the broadly rounded postero-ventral angle of the lateral temporal
opening. This anterior lamina of the quadrate is damaged and it is not clear
how far it invaded the lateral temporal opening. The thick and buttress-like rear
margin of the quadrate extends ventrally as a robust articular process which
terminates just below the level of the mandibular teeth in a rounded and
slightly inflated condyle. The orientation of the quadrate, which runs down
and slightly forwards, may well be natural. The quadrate has a similar
orientation in Hypsilophodon foxii but is deflected to the rear in Lycorhinus
tucki (see Galton, 1970: fig. 4C, D). At its ventral end the lateral face of the
quadrate bears a shallow recess at the posterior margin. This depression
doubtless accommodated a ventral lamina from the overlying quadrato-jugal, so
that a narrow strip of the quadrate would have been visible in lateral view just
above the articular condyle. Galton shows (1970: fig. 4) that the quadrate and
quadrato-jugal have a very similar relationship in Lycorhinus tucki.
The paired frontals and nasals are represented by splinters of bone over the
snout and the orbits. Evidently these skull roofing bones were quite narrow and
did not reach very far down on to the sides of the skull. A fragment of bone
behind the left orbit was originally identified as that part of the postorbital
which, together with the squamosal, formed the upper temporal bar. From the
position of such a fragment it might have been deduced that the lateral
temporal opening was restricted in vertical extent and decidedly smaller than
the orbit. But it seems very much more likely that this same fragment is part of
the left supraorbital which has been shifted back from the orbit into the lateral
temporal opening. Consequently it is impossible to estimate the dorsal extent
of the lateral temporal opening.
A long, gently tapered, slightly arched and rod-like bone at the front of the
left orbit is identified as the left prefrontal. This bone seems to have defined
much of the anterior rim of the orbit and generally resembles the prefrontal in
Fabrosaurus australis (see Thulborn, 1970b). The prefrontal of Lycorhinus
tucki appears to be quite different; Galton shows the prefrontal in this animal
(1970: fig. 4) as a relatively small bone which makes a very limited
contribution to the rim of the orbit. Several small osseous scales, vestiges of the
sclerotic ring system, are preserved behind the left prefrontal. These are
disposed in a roughly circular pattern but are so poorly preserved that their
number and manner of overlapping cannot be determined.
158
R. A. THULBORN
The anterior parts of both mandibles are preserved intact, the seeming
absence of sutures implying that the front part of each ramus is composed
largely of the dentary bone. Each straight and fairly deep ramus has a thickly
rounded inferior margin, is flattened on the medial side and is noticeably
inflated at its anterior end. N o openings, lateral or medial, are visible in either
ramus and the coronoid apophysis is merely a low and obtusely angular
eminence. The mandibles of Lycorhinzis tucki and L. angustidens are somewhat
more stoutly constructed and that of L. tticki, at least, is distinguished by the
presence of an external fenestra (Galton, 1970: fig. 4). The dentaries are joined
in front by the median predentarv-a scoop-like and toothless bone, some
7 mm long, which is extended postero-ventrally as a narrow wedge between the
inflated front ends of the dentaries. The predentary is perforated on each side
by a single large foramen. A fissure running down from the postero-dorsal slope
of the coronoid eminence may possibly have developed along the suture
between dentary and surangzilar bones (this latter having a comparable
forwards extent in Lycorhinns tzrcki). The ventral end of this fissure forks to
enclose a wedge of bone which may possibly represent part of the angular.
Den tition
(Figs 2, 3 and 4)
The dentition is heterodont, the teeth in the premaxilla being quite different
in appearance from those in the maxilla and dentary. In each of these jaw
bones the teeth are set in separate alveoli and are arranged in a single line.
There is no appreciable overlap or imbrication between the crowns of adjacent
teeth.
The ventral part of the right premaxilla is missing but the left premaxilla
bears two teeth which are preceded by an edentulous space of at least 8 mm.
Both of the conical and peg-like teeth are directed ventrally and slightly to the
rear. Their crowns are smooth, completely enamelled and (as their damaged
occlusal ends show) circular in cross-section. The anterior crown is 2.5 mm tall;
the second is almost 4 mm tall. Behind the second tooth the inferior margin of
the premaxilla turns upwards to form part of the arch-like diastema which
precedes the maxilla. There is no trace of any third tooth, nor of any alveolus
for such a tooth.
The left maxilla has the full complement of 12 teeth preserved in situ. In the
following account the teeth are numbered from front to back. Maxillary teeth
1 to 6 have tall blade-like crowns which are successively larger to the rear. In
each case the buccal surface is gently convex from mesial to distal edges and is
completely enamelled. Mesial and distal margins diverge away from the alveolus
and are elaborated into weak ridges on the buccal surface. These marginal
ridges are bordered towards the mid-line of the crown by extremely shallow
furrows. In each crown the rounded occlusal margin bears about ten short,
blunt and divergent denticles which are all roughly similar in size and which are
prolonged on to the buccal surface as feeble ribs. Teeth 3, 5 and 6 have crown
heights of 2.5, 3.5 and 4 mm respectively. The 7th and 8th crowns resemble
the 6th but are somewhat larger and have flatter buccal surfaces. Crowns 9 and
10, the largest in the maxilla, are nearly as long (mesio-distally) as they are tall.
The 9th crown is nearly 3 mm across and is 3.5 mm tall; in the 10th the height
A NEW HETERODONTOSAURID DINOSAUR
159
barely exceeds the mesio-distal length (4.5 mm). Both crowns have flat and
smooth buccal surfaces and finely denticulate occlusal margins which are
angular, rather than rounded, in profile. The 1l t h crown is a little smaller than
the 10th but is ornamented in similar fashion; it has widely divergent margins
and is broadly triangular in profile. The 12th crown, which is represented by
fragments, seems t o have resembled the 1lth.
Left maxillary crowns 7 to 10 have also been exposed in lingual aspect. In
each case the lingual face of the crown is obliterated by a single and very large
wear facet. These perfectly flat wear facets are inclined at about 20" from
vertical and they all lie in a single plane. This unusual pattern of tooth wear is
quite characteristic of heterodontosaurid dinosaurs and has been described in
both Lycorhinus tucki and L . angustidens (see Crompton & Charig, 1962;
Thulborn, 1970a, 1971a).
The left mandible carries 14 teeth. The first crown is a smooth and blunt
conical peg, barely 1 . 5 mm tall, which is situated at the very front of the
dentary (i.e. there is no space for a preceding tooth). The second crown
resembles the first but is taller (2.5 mm) and is very slightly recurved. These
first two teeth lie directly beneath the diastema preceding the maxilla. The 3rd
crown is fractionally smaller than the 2nd and has four minute denticles on the
distal margin. Evidently this tooth is transitional between the non-denticulate
teeth in front and the fully denticulate ones behind. Both edges of the 4th
crown bear small denticles; these are arranged in a divergent pattern and extend
on to the buccal surface as feeble ribs. The 5th crown appears to be a larger
version of the 4th, though its buccal surface is a little flatter and bears weak
ridges at the mesial and distal borders. Though crowns 6 to 9 are damaged they
do show a gradual increase in size to the rear. These appear to resemble the 5th
crown but bear slightly stronger marginal ridges. The 10th crown is represented
by fragments. Crowns 11 and 12 are the largest in the mandible, with
maximum mesio-distal lengths of 2.5 and 3 mm respectively. They are similar
to crowns 8 and 9 but have the marginal ridges even more strongly developed.
The 13th crown is represented by a single fragment. The angular occlusal
margin of the short and broadly triangular 14th crown bears ten blunt,
divergent and rather variable denticles. All the crowns in the middle of the left
mandibular tooth row lack their occlusal margins and no identifiable wear
facets have been preserved.
It is significant that there is no obvious evidence of active tooth replacement
in the dentition. There are no vacant alveoli in the tooth rows and no clear
traces of the alternating tooth generations which are visible in the dentitions of
many other ornithischian dinosaurs (see Edmund, 1957).
DISCUSSION
Review of heterodontosaurid dinosaurs
Before proceeding to assess the relationships of Lycorhinus consors it will be
useful to list, and to comment upon, other heterodontosaurid dinosaurs. These
are (in chronological order):
Geranosaurus atavus Broom, 1911
Lycorhinus angustidens Haughton, 1924
Lycorhinus tucki (Crompton & Charig, 1962) (Heterodontosaurus)
1 60
R. A. THULBORN
Gcv-anosaurusatautts Broom, 1911
Represented by a single poorly preserved jaw from the Cave Sandstone of
South Africa. The left mandible which constitutes the holotype carries the
stumps of nine teeth and Broom shows (1911: plate XVII, fig. 24) that there
are traces of a toothless predentary bone, diagnostic of ornithischian dinosaurs,
at the mandibular symphysis. The teeth are set in separate sockets and it is
possible that the foremost one, which is noticeably larger than the others,
might have been comparable with the mandibular caniniform in Lycorhinus
angustidens and L. tucki (see below). The inadequacy of the holotype
precludes any meaningful comparisons with better known forms and it may be
concluded that Geranosaurus atauus. though probably ornithischian, is
generically and specifically indeterminate and that the name is a nomen
dubium.
Lycorhinus angustidens Haughton, 1924
Established on a tooth-bearing piece of mandible from the Red Beds of the
Stormberg Series at Paballong, near Mount Fletcher, in the Herschel District of
Cape Province, South Africa. The holotype carried 1 1 cheek teeth together
with a very conspicuous “canine” tooth and the appearance of this heterodont
dentition led Haughton (1924) to regard Lycorhinus angustidens as a therapsid.
In 1932 Broom re-figured the holotype, stated that it had begun to disintegrate
and mentioned that the cheek teeth resembled those of dinosaurs. Then in
1962 Crompton and Charig gave a preliminary account of an ornithischian
skull, which they named Heterodontosaurus tucki, from the Cave Sandstone of
South Africa. These authors remarked on several similarities between Heterodontosaurus tucki and Lycorhinus angustidens and suggested that the latter
might be an ornithischian. Evidence from a topotype of Lycorhinus
angustidens has endorsed this suggestion and seems, indeed, to warrant
relegation of the name Heterodontosaurus t o the synonymy of Lycorhinus (see
Thulborn, 1970a). My placing of Heterodontosaurus in the synonymy of
Lycorhinus has received sharp criticism from Galton (1973) who asserts that
“. . . Thulborn (1970a) . . . does not and cannot demonstrate that this
specimen [i.e. the topotype] is identical to the lost holotype of Lycorhinus
angustidens Haughton (1924) . . .”. I have already drawn attention (Thulborn,
1970a) to very striking resemblances between the holotype and the single
topotype of Lycorhinus angustidens and it seems necessary merely to re-state
the following facts: that both specimens have an extremely unusual and very
characteristic dentition, that the two specimens are not greatly dissimilar in
size, that both show the same pattern of tooth wear and that both are from the
same horizon and the same locdity. The topotype compares very closely with
descriptions and published figures of the holotype (Haughton, 1924: fig. 8 ;
Broom, 1932: fig. 104, I) and with latex impressions taken from the holotype.
If one makes the assumption that both specimens are from a single population
the minimum values of V (coefficient of variation) for various linear
dimensions would be as follows:
.4 NEW HETERODONTOSAURID DINOSAUR
~
161
~~~
Linear dimension
Height of mandibular “canine” crown
Mesio-distal length of mandibular “canine” crown
Length of gap between mandibular “canine” crown and 1st post-“canine” crown
Depth of mandible below 3rd post-‘‘canine’’crown
Mesio-distal length of 3rd post-“canine” crown
Mesio-distal length of 4th post-“canine” crown
Mesio-distal length of 5th post-“canine” crown
Minimum V
4
6
7
6
2
5
8
In general it is unwise to assume that a difference in a linear dimension alone
may characterize different taxonomic groups unless it implies a minimum V of
at least 10-values as high as this, or even higher, being known t o occur in pure
races. In short, there seems to be no reason for contesting my assertion that the
specimen described in 1970 (Thulborn, 1970a) is representative of Lycorhinus
angustidens.
Finally, it is worth noting that the validity of the name Lycorhinus
angustidens is not affected by loss, or disintegration, of the holotype. In such
circumstances it would be perfectly sound taxonomic practice to promote the
described topotype to the status of neotype.
Lycorhinus tucki (Crompton & Charig, 1962)
1962 Heterodontosaiirus tucki Crompton & Charig, pp. 1074-77.
1970 Lycorhinus rucki Thulborn, p. 244.
Represented by a skull, together with some post-cranial bones, from the
Cave Sandstone of South Africa. Only the teeth and the anterior part of the
skull have been described in detail (Crompton & Charig, 1962). Galton (1 970)
provides a useful figure of the reconstructed skull and mentions that the
anterior ramus of the pubis is practically non-existent.
There are no major differences between the jaw bones of Lycorhinus tucki
and those of L. angustidens and both forms have very characteristic and
extremely unusual dentitions which incorporate large canine-like teeth.
Lycorhinus tucki is distinguished from L. angustidens by minor differences in
tooth structure. In L. tucki both edges of the mandibular “canine” are finely
serrated, but in L. angustidens only the anterior edge is so ornamented. Fluting
on the buccal surfaces of the cheek teeth is more pronounced in L. tucki. Both
L. angustidens and L. tucki were collected from local developments of the Red
Beds facies within the Cave Sandstone formation. In view of gross and detailed
resemblances between the two forms I reiterate my opinion (Thulborn, 1970a)
that the name Heterodontosaurus should be regarded as a junior synonym for
Lycorhinus (which has priority by 38 years).
Relationships of Lycorhinus consors
All adequately known Triassic ornithischians belong t o the suborder
Ornithopoda. Specimen B54 is no exception, its ornithopod status being
demonstrated by marked disparity in size between fore and hind limbs, by the
elongate and clawed foot, by the pendent fourth trochanter of the femur and
by the short (tetramerous) sacrum.
162
R. A. THULBORN
Specimen B54 is assigned to the family Heterodontosauridae on account of
resemblances in skull structure with Lycorhinus angustidens and L. tucki and
on account of its typically heterodontosaurid pattern of tooth wear. The skull
generally resembles that of Lycorhinus tucki (as figured by Galton, 1970), but
there are several definite differences, particularly in the extent of the
prefrontal, in the architecture of the jaw bones and in the appearance of the
dentition. Lycorhinus consors lacks the longitudinal ridge which is present on
the flank of the maxilla in L. angustidens and L. tucki and has a prefrontal
which is quite different in shape from that of L. tucki. The mandible is
distinguished by its less robust construction and by its inflated anterior tip and
differs from that of L. tucki in lacking an external fenestra. The cheek teeth
differ from those of L. angustidens and L. tucki in having weaker ribs and
grooves on the crown surfaces, in having finer marginal denticles and in lacking
any large cusp-like denticles.
But the most striking feature of Lycorhinus consors is the total lack of
caniniform teeth. Both L. angustidens and L. tucki possess a large canine-like
tooth at the rear of the premaxilla and another at the front of the mandibular
tooth row (Crompton & Charig, 1962; Thulborn, 1970a). In the holotype of 2,.
consors there is no trace of a premaxillary “canine” (nor of any alveolus for
such a tooth) and the foremost mandibular tooth is a minute conical peg. I
suggest that the presence or absence of “canine” teeth in heterodontosaurids
may reflect sex dimorphism (see below) and it would seem unwise to establish
a new taxon for specimen B54 purely on the grounds that it lacks “canine”
teeth. Leaving aside the lack of such teeth, specimen B54 does not differ
significantly enough from L. angustidens or L. tucki to warrant its separation
into a new genus. Consequently this specimen is referred to the genus
Lycorhinus. Specimen B54 cannot be identified precisely with L. angustidens
or L. tucki on account of several differences in the architecture of the skull
bones and in the dentition (i.e. apart from the lack of “canine” teeth). These
differences lead me to conclude that specimen B54 represents a new species of
heterodontosaurid dinosaur, which is named Lycorhinus consors.
I t may be observed that L. consors has been distinguished from L.
angustidens by means of fewer anatomical characters than it has from L. tucki.
But this observation cannot be taken to indicate that L. consors is more nearly
related to L. angustidens than to L. tucki; it merely reflects the fact that L.
angustidens is less well represented by fossil material than L. tucki and that, in
consequence, fewer comparisons are possible.
Functional and palaeoecological implications of the heterodontosaurid
dentition
Heterodontosaurids seem to have resembled all other ornithischians in being
basically herbivorous (see Galton, 1973, for review of feeding adaptations in
ornithischians) and their heavily worn cheek teeth testify to the abrasive and
resistant nature of the heterodontosaurid diet. Lull concluded (1917) that the
anterior parts of the jaws functioned as a food-gathering device throughout the
ornithischians and this generalization evidently holds true for the heterodontosaurids. The condition in the heterodontosaurid premaxilla (where the
teeth are confined to the rear of the bone) is rivalled in the Wealden
A NEW HETERODONTOSAURID DINOSAUR
163
ornithopod Hypsilophodon foxii, where the similarly restricted premaxillary
dentition was preceded by a horny upper beak (Galton, 1973). I t is not
unreasonable to infer that similar upper beaks were present in heterodontosaurids, though incontrovertible evidence, in the form of attachment
scars on the premaxillae, appears to be lacking. All adequately known
ornithischians have the mandibular symphysis braced by a scoop-like predentary bone which probably carried a lower beak-traces of both upper and lower
beaks being preserved in some specimens of hadrosaurs (Morris, 1970). In the
heterodontosaurids it is likely that the lower beak bit against the upper beak,
and against the premaxillary teeth, to function as a cropping device roughly
comparable-though in inverted fashion-with that in living artiodactyls (where
the lower incisors bite against a horny pad in the upper jaw). This conclusion is
supported by the presence of wear facets on the lingual sides of the
premaxillary teeth in Lycorhinus angustidens (see Thulborn, 1970a: fig. 2).
Evidently the “canine” tusks of heterodontosaurids had no very specialized
role in food-gathering (aside from the upper tusks being incidentally involved in
the cropping mechanism just described). The outwardly and upwardly curved
tusks of the warthog (Phacochoerus aethiopicus) are used in digging for food
(Maberly, 1963; Walker et al., 1964) but the practically straight and vertically
orientated tusks of heterodontosaurids can hardly be regarded as convincing
analogues. I t is extremely unlikely that the tusks of heterodontosaurids were
employed in stabbing or puncturing food materials, or that upper and lower
tusks could have been opposed, point to point, in pincers-like fashion; such use
of the tusks would have required an improbably wide gape of the jaws, and the
gape was severely limited by the presence of fleshy cheeks (see Galton, 1973).
Lull noted (in Hatcher, Marsh & Lull, 1907) that the cheeks of ceratopsians
would have prevented use of the maxillary and mandibular teeth in
foodgathering and deduced that the cheek teeth were concerned only with
mastication. The same general arrangement obtains in heterodontosaurids. The
cheeks of heterodontosaurids probably extended to the front of the maxilla, so
that the corner of the relatively small mouth opening was situated just behind
the lower “canine” tusk. The ossified remnants of the hyoid apparatus in the
primitive ornithopod Fabrosaurus australis point to the existence of a large and
active tongue in the earliest ornithischians (Thulborn, 1970b, 1971a).
Presumably the heterodontosaurid tongue was used, like that of Fabrosaurus
australis, to transfer food material backwards in the mouth and to assist in
retaining it within the narrow occlusal zone between upper and lower cheek
teeth. I t is also possible (particularly in view of the limited size of the mouth
opening) that the heterodontosaurid tongue might have been employed as a
prehensile food-gathering device. A prehensile tongue is commonly supposed to
have been present in the Wealden ornithopod Iguanodon (see Casier, 1960).
Though the skull bones of heterodontosaurids are not yet sufficiently well
known t o permit detailed investigation of the jaw musculature it is possible to
deduce the general style of jaw action from the appearance of the dentition and
from the very characteristic pattern of tooth wear. In the heterodontosaurid
dentition there are no obvious traces of the alternating tooth generations which
are visible in the dentitions of other ornithischians (see Edmund, 1957, 1960)
and the flat wear facets of the cheek teeth are all developed in a single steeply
inclined plane. Consequently the rows of cheek teeth appear to have been worn
164
R. A . THULBORN
down as composite units rather than as collections of individual teeth (see
Crompton & Charig, 1962). It is difficult to envisage this planar tooth wear as
the product of vertical movements between dentary and maxillary teeth,
because any minor irregularities in the shape and disposition of the cheek teeth
would inevitably have resulted in some corrugation of the occlusal plane
between upper and lower teeth. I t may be concluded that the pattern of tooth
wear in heterodontosaurids reflects a highly specialized jaw action which
involved back and forth sliding of the mandibles while upper and lower teeth
were in occlusion (Thulborn, 1971a). Such a jaw action would have produced a
grinding or milling effect between upper and lower cheek teeth, serving to
comminute even the most resistant plant tissues. A similar system of jaw action
has been attributed to the hadrosaurs by Ostrom (1961). The prominent tusks
of Lycorhinus angustidens and L. tucki would not have impaired protraction
and retraction of the mandibles: the upper tusk would have cleared the flank of
the predentary by several millimetres while the lower tusk could have been
shifted back and forth in the diastema between premaxilla and maxilla.
I t is obvious that any gap in the heterodontosaurid cheek dentition would
have impaired feeding. But the normal processes of tooth replacement often
produce discontinuities (such as vacant tooth spaces or newly erupted crowns)
in the tooth rows (see Edmund, 1957, 1960). The hadrosaurian solution to this
problem of maintaining an efficient “dental mill” consists of having the cheek
teeth arranged in large batteries. These batteries may have up to six or seven
teeth in vertical series and there may be up to 40 such series in the maxilla or
dentary (Ostrom, 1961). In hadrosaurs the occlusal surface is distributed over
the oldest two or three teeth in each vertical series and the loss of a few teeth
from the margin of the battery, in the course of normal tooth replacement, is
of no great consequence. In a heterodontosaurid, in contrast, there is only a
single line of functional teeth and threatened disruption of the “dental mill” by
tooth loss has apparently been countered by suppression of tooth replacement.
This suppression of tooth replacement in the heterodontosaurid cheek
dentition is evidenced by the absence of alternating tooth generations and by
the lack of any vacant tooth spaces or newly erupted crowns in the rows of
cheek teeth. There seem t o be two possibilities: (1) either heterodontosaurids
did not replace their cheek teeth at all, or (2) they replaced the entire cheek
dentition in a single short episode. I t is highly improbable that heterodontosaurids did not replace their teeth. Evidently heterodontosaurids fed on
very resistant plant materials, which caused a great deal of wear on the cheek
teeth, and it is difficult to imagine how they could have survived for any length
of time without replacing the cheek teeth. Other ornithischians (even the
hadrosaurs and the ceratopsians, with their large batteries of teeth) found it
necessary to replace teeth throughout life and it is unlikely that heterodontosaurids were exceptional.
The conclusion that heterodontosaurids did replace their cheek teeth seems
inescapable. But tooth replacement cannot have followed the normal ornithischian pattern as described by Edmund (1957, 1960) because the shedding
and replacement of teeth in alternating waves would have disrupted the entire
feeding mechanism. Since the heterodontosaurid dentition cannot have been
replaced piecemeal it must, of necessity, have been replaced in toto.
Unfortunately there is practically no tangible evidence relating directly t o
A NEW HETERODONTOSAURID DINOSAUR
165
tooth replacement in heterodontosaurids. No one has described any heterodontosaurid jaw with replacement teeth or with the “special foramina” which
are involved with tooth replacement in many ornithischians (see Edmund,
1957). And no one has described any shed teeth which might be attributed to a
heterodontosaurid. The shed teeth of other ornithischians may, in contrast, be
of relatively common occurrence (see, for example, Thulborn, in press). I t is
impossible, in view of these deficiencies, to speculate on the details of the
tooth replacement mechanism in heterodontosaurids.
Palaeoclimatological evidence (Krynine, 1949; Schwarzbach, 1963) indicates
that the Red Beds environment was semi-arid and was subject to seasonal
rainfall. I t is quite possible that the herbivorous heterodontosaurids underwent
aestivation (or hibernation) when the vegetation cover diminished during the
yearly dry season. It would seem logical to suppose that the entire cheek
dentition would have been replaced during this annual dormancy, so that each
heterodontosaurid would have been provided with a new and unworn “dental
mill” at the start of each season when food became plentiful. The heterodontosaurid cheek dentition may then be regarded as an impermanent and
disposable food-processing mechanism, designed to be rapidly abraded by
dealing with harsh plant materials, then to be discarded and replaced in a single
short episode. Such shedding and replacement of the teeth in brief annual
bursts would have left the heterodontosaurids free to feed normally throughout
the greater part of each year and would have disrupted the feeding mechanism
only at those times when food materials were in short supply. This hypothesis
would account for the fact that no specimen of heterodontosaurid appears to
show any traces of active tooth replacement: replacement would have been
restricted to a brief period in each year and the odds would have been very
much against a heterodontosaurid dying, and being suitably preserved, at
exactly such a period. ‘This same hypothesis would also explain the fact that all
heterodontosaurids described to date have appreciably worn teeth, for the
teeth would have remained unworn only for a very short period following the
annual dormancy. Finally this hypothesis will explain the apparent lack of
worn and discarded heterodontosaurid teeth in the Red Beds; the teeth would
have been discarded in groups at selected sites, rather than distributed one by
one throughout the environment. A similar process of annual dormancy might
be invoked to account for an accumulation of worn and shed teeth found in
association with a specimen of the Red Beds ornithischian Fabrosaurus australis
which had practically unworn teeth in its jaws (see Thulborn, 1971a: 171 and
Galton, 1973: 72 for discussions). I t is quite conceivable that this particular
specimen of Fabrosaunis australis had perished while undergoing aestivation or
hibernation. The heterodontosaurids are not to be regarded as exceptional in
undergoing annual dormancy, for there are well-recorded instances of aestivation and hibernation in other archosaurs, living and extinct. Guggisberg
(1972) mentions that North American, South American and African crocodilians are known to aestivate or hibernate and Ewer (1965) has suggested that
the Lower Triassic thecodontian Euparkeria capensis underwent an annual
dormancy.
It has already been deduced that the prominent “canine” teeth of the
heterodontosaurids Lycorhinus angustidens and L. tucki played no very
significant role in feeding. Before proceeding to investigate the function of the
R. A. THULBORN
166
“canine” tusks it is necessary to examine the possibility that the lack of such
teeth from the holotype of Lycorhinus C O ~ S O Y Smight signify immaturity. The
possibility that the holotype of Lycorhinus consors might be a juvenile
specimen of L. angustidens or L. tucki (or an unknown tusk-bearing
heterodontosaurid) must be discounted for several reasons. First, the holotype
of Lycorhinus censors is roughly comparable in size with other heterodontosaurids. Maximum skull lengths for heterodontosaurids are listed below,
together with the dental formulae (where the letters pr, c and P O signify
pre-“canine”, “canine” and post-“canine” teeth respectively):
Approximate skull length (cm)
Dental formula
PC
c
PO
L. angustidens holotype
L. angustidens topotype
L. aucki holotype
L. consors holotype
11.0.
10.0
9.58
-
1
11t
2
1
13
0
1
2
1
7t
12
0
1
12
2
0
12
0
0
14
* Estimated; - unknown.
The holotype of Lycorhinus consors is not very much smaller than any other
specimen of heterodontosaurid dinosaur and has, aside from its lack of
“canine” tusks, a typically heterodontosaurid dental formula. It must be
emphasized that there is no space in the jaws to accommodate any caniniform
teeth (see Fig. 2); the second premaxillary tooth lies right against the arch-like
diastema and the row of mandibular teeth extends without interruption to the
front of the dentary. In consequence it seems unlikely that the holotype of L.
consors might be an animal in the process of developing or replacing the tusks.
The growth and emplacement of an upper tusk would require profound
modification of the premaxillary tooth row while the development of a lower
tusk would involve loss of several teeth at the front of the dentary. I t is also
unlikely, for these same reasons, that the last premaxillary crown and the first
mandibular crown might represent the tips of erupting tusks. I t may be
concluded that the holotype of Lycorhinus consors is an adult heterodontosaurid which never had any “canine” tusks and which was not about to
develop any such tusks at the time of its death.
I suggest that the lack of “canine” teeth from the holotype of Lycorhinus
consors does not signify immaturity but that the absence or presence of paired
tusks in heterodontosaurids is indicative of gender. Such an example of sex
dimorphism, where males were equipped with tusks while females were without
these teeth, may have been rivalled among the dicynodonts-where tusk-bearing
forms (presumably males) were commonly deposited in the genus Dicynodon
while tuskless (? female) specimens were frequently assembled under the name
A NEW HETERODONTOSAURID DINOSAUR
167
Oudenodon (see Toerien, 1953; Romer, 1956). The same pattern of sex
dimorphism finds convincing parallels among living mammals. In many
mammals the canines of the male are larger than those of the female, the list
including primates, carnivores and ungulates. But there are relatively few cases
where the canines of the male are developed into large tusks while those of the
female are small or absent-the most striking examples being among suids,
tayassuids, tragulids and cervids (Hydropotes inermis, Moschus moschiferus).
The parallel between suids and heterodontosaurids is not a particularly close
one; the tusks of suids curve laterally and upwards while those of heterodontosaurids are practically vertical. The tayassuids (peccaries) form a closer
parallel, for the tusks are generally shorter than those of suids-with an average
length of about 4 cm (Walker et al., 1964)-and are carried more or less in a
vertical plane. The tusks of male tragulids (chevrotains) are generally similar in
appearance to those of the peccaries. Both tragulids and tayassuids use the
tusks in intra-specific fighting and as defensive weapons (Walker et aZ., 1964;
Dorst & Dandelot, 1970) and I suggest that the tusks of heterodontosaurids
were functionally analogous. The tusks were doubtless very prominent features
of the head in male heterodontosaurids and may also have been employed in
display or as a visual threat.
Suids and tayassuids can inflict severe damage with their sharp-edged tusks
and maintain them in fighting trim by honing the upper tusks on the lower
ones. Such tooth-on-tooth wear has been described in detail by Every (1965,
1970) and by Every & Kuhne (1971) under the name thegosis. The clashing
and grinding together of the tusks in suids and tayassuids produces a clattering
noise which may also have a deterrent effect on adversaries. In male
heterodontosaurids the upper and lower tusks are not antagonistic, the upper
one lying well in front of the lower. While it is possible that the upper and
lower tusks might have been brought into contact through protraction of the
lower jaw such a system could hardly have been used to keep the edges of the
tusks finely honed. Since the tusks of heterodontosaurids cannot have been
maintained in fighting trim by means of thegosis-like grinding their edges are
finely serrated (such serration being an odontological feature often encountered on tooth edges which are important in cutting and slicing but which
cannot be sharpened by tooth-on-tooth wear).
According to the suggestions outlined above the tusk-bearing heterodontosaurids Lycorhinus angustidens and L. tucki would be considered as
males while the tuskless L. consors would be regarded as a female. This division
into males and females necessarily leads one to ask if the female L. censors
might be conspecific with one of the male heterodontosaurids. Unfortunately
there is not enough evidence to form any conclusive answer to this question.
Lycorhinus consors differs from L. angustidens and L. t u c k in the architecture
of its skull bones and in the structure of its cheek teeth and it seems wisest, in
view of these differences, to retain L. consors as a separate species until such
time as conclusive evidence is forthcoming.
Classification of ornithopod dinosaurs
All adequately known and determinate ornithischians of Triassic age are
members of the suborder Ornithopoda. In previous works I have adopted a
12
168
R. A. THULBORN
conservative approach to the classification of Triassic ornithopods, including all
determinate forms in the family Hypsilophodontidae (see Thulborn, 1970a, b,
1971b, 1972, in press). However, Kuhn advocated (1966) separation of
heterodontosaurids from hypsilophodontids at family level and this scheme has
been endorsed by Romer (1966) and by Galton (1972). In addition Galton
(1972) has placed the Triassic Fabrosaurus australis and the late Jurassic
Echinodon becklesii in a separate family (Fabrosauridae) on the grounds that
these forms appear to have lacked the fleshy cheeks which were probably
present in the great majority of ornithischians. In the final section of this paper
I give a revised classification for ornithopod dinosaurs. This classification
recognizes the families Heterodontosauridae Kuhn (1966) and Fabrosauridae
Galton (1972) and is particularly intended to clarify the relationships of the
Triassic ornithopods.
Galton (1972) defines the family Heterodontosauridae by four characters:
the inset dentition (implying the presence of fleshy cheeks), the planar tooth
wear, the rudimentary prepubis and the presence of caniniform teeth. The inset
dentition is seen in all ornithischians apart from fabrosaurids and the planar
tooth wear appears to be matched in Pisanosaurus mertii (see Casamiquela,
1967)-which Galton assigns (1972) to the Hypsilophodontidae. The
rudimentary prepubis is encountered in several early ornithischians; it occurs in
Fabrosaurus australis (see Thulborn, 1971b) and in the Liassic ornithopod
figured by Charig (1972) as a young specimen of Scelidosaurus. The fourth
character listed by Galton (1972) must be discounted because the heterodontosaurid Lycorhinus consors lacks caniniform teeth. In short, it is
practically impossible to define the Heterodontosauridae (with the inclusion of
Lycorhinus consors) using the suite of characters listed by Galton (1972). In
formulating a workable diagnosis for this family (see classification given below)
I have taken into account the unusual appearance of the cheek teeth in
heterodontosaurids. Briefly, :he teeth are arranged in an unbroken line, with
the crowns decreasing in size towards both the front and the back of the jaw.
There are no definite traces of the alternating tooth generations which are
visible in the dentitions of other ornithischians and the planar tooth wear is
continuous from one crown to the next. I t has already been suggested that
these adaptations may reflect a highly specialized jaw action which involved
back and forth sliding of the mandibles to produce a grinding effect between
upper and lower teeth. Whatever its functional implications may be the
heterodontosaurid dentition is quite sharply distinguished from other
ornithopod dentitions by the absence of tooth replacement waves and by the
lack of vacant tooth sites or newly erupted crowns in the rows of cheek teeth.
Galton considers (1972) that the hypsilophodontids were derived from an
ornithopod like Pisanosaurus mertii-a poorly known form collected from the
Upper Trias of western Argentina. However, Casamiquela’s illustrations of
Pisanosaurus mertii (1967) are far from enlightening and it is difficult to
support Galton’s contention that this animal is an early hypsilophodontid.
There is no certain record of any hypsilophodontid before the Upper Jurassic.
The fragmentary specimen of “Hypsilophodon ” or some allied form identified
by Newman (1968) in the English Lias (= Lower Jurassic) may possibly be
identical with the so-called juvenile or young form of Scelidosaurus which has
been recovered from about the same horizon (Rixon, 1968; Romer, 1968;
A NEW HETERODONTOSAURID DINOSAUR
169
Charig, 1972). This latter animal would appear, from the illustrations given by
Charig (1 972), to be an ornithopod closely related to Fabrosaurus austrulis.
The dentition of Pisanosaurus mertii does appear to approach that of a
heterodontosaurid in some respects (though the illustrations given by
Casamiquela (1967) are difficult to interpret). The inset rows of cheek teeth do
not show any obvious waves of tooth replacement and the tooth wear may
possibly be of planar type. There are no caniniform teeth in Pisanosaurus mertii
but, as Lycorhinus consors demonstrates, such teeth are not invariably present
in heterodontosaurids. On the basis of this (admittedly slight) evidence I have
tentatively included Pisanosaurus mertii in the family Heterodontosauridae (see
classification below). There is nothing in the anatomy of Pisanosaurus mertii
which militates against heterodontosaurid affinities and the tentative
assignment of this “difficult” animal to the Heterodontosauridae considerably
simplifies the pattern of early ornithopod history. The heterodontosaurids
represent a short-lived off-shoot from the base of the ornithischian phylogenetic tree (Fig. 5 ) . I t is not unreasonable, in the present state of knowledge,
to suggest that Pisanosaurus mertii approaches the heterodontosaurids in its
adaptations related to feeding but may not be directly ancestral to the
heterodontosaurids of the African Trias (Lycorhinus spp.).
The relatively unspecialized and medium-sized to large ornithopods of the
Jurassic and Cretaceous have commonly been deposited in a single family, the
Iguanodontidae (see, for example, Romer, 1956, 1966). It is likely, however,
that the family Iguanodontidae is an artificial assemblage comprising several
groups of ornithopods which have independently attained a graviportal style of
locomotion (Ostrom, 1970; Thulborn, 1971b; Galton, 1972). Galton has
discerned (1972) a minimum of three lines of iguanodontid ancestry: one in
the mid-Jurassic (leading to Iguanodon and Tenontosaurus), one in the late
Jurassic (leading to Cumptosaurus) and one in the late Cretaceous (leading to
Thescelosaurus). A fourth, and very early, line of iguanodontid ancestry may
be represented by the Lower Jurassic Scelidosaurus harrisonii-an armoured
ornithischian which has hitherto been regarded as a primitive stegosaur or an
early ankylosaur. In practically every part of its anatomy Scelidosaurus
harrisonii appears to be a perfectly acceptable ornithopod dinosaur (see
descriptions and figures given by Owen, 1861, 1863) and certain parts of the
skeleton (notably the hind limb) are strikingly similar to their counterparts in
the Lower Cretaceous Tenontosaurus. Consequently I have classified Scelidosaurus harrisonii as a progressive graviportal ornithopod*. In the following
classification I have accorded each of the four groups of graviportal
ornithopods separate family status and have considerably restricted the scope
of the family Iguanodontidae. Galton suggests (1972) that all lines of
iguanodontids (sensu lato) spring from the hypsilophodontids. But since there
is no convincing record of any hypsilophodontid before the Upper Jurassic it
seems quite possible that one or more of the earlier lines leading to the
It may be objected that the armoured Scelidosvrurus harrisonii is somewhat out of place among the
ornithopods on account of its dermal amour. But traces of armour may possibly occur in the Wealdeo
ornithopod HypsUophodon foxU (see Romer, 1956) and it k quite conceivable that armour may have
persisted in some ornithopods at the iguanodontid le&l of organization. To substantiate this possibility
Professor W. J. Moms has recently reported (pers. eomm.) the discovery of scutes in the Upper
Cretaceous Thmelomunu neglectus.
170
R. A. THULBORN
Figure 5. Chart showing stratigraphic ranges and presumed relationships of the families of
ornithopod dinosaurs. Derived, in part, from Galton (1972). Families of dolichopodous
omithopods indicated by three letters, thus: FAB, Fabrosauridae; HET, Heterodontosauridae;
HYP, Hypsilophodontidae; PSI, Psittacosauridae. Families of brachypodous ornithopods
indicated by two letters, thus: CA, Camptosauridae; HA, Hadrosauridae; IG, Iguanodontidae;
PA, Pachycephalosauridae; SC, Scelidosauridae; TH, Thescelosauridae.
The pre-Upper Jurassic history of the Hypsilophodontidae is speculative and it is possible
that some early brachypodous ornithopods might have been derived from the Fabrosauridae.
iguanodontid grade of organization might have arisen from the unspecialized
fabrosaurids (see Fig. 5 ) .
Finally, I would point out that Galton (1972) has misrepresented my
concept of a hypsilophodontid “plexus” at the core of ornithischian phylogeny
(Thulborn, 1971b). The term plexus was introduced into my general survey of
ornithischian history (Thulborn, 1971b) in its original sense (as a rope or plait)
A NEW HETERODONTOSAURID DINOSAUR
171
in order to avoid lengthy analysis of relationships among those ornithopods
showing adaptations for cursorial locomotion. I did not state that known
hypsilophodontids were unrelated, as cited by Galton (1972), but merely
indicated that relationships were difficult to discern. Evidently there were three
major lines of evolution among early cursorial ornithopods: one persistently
primitive-looking (the family Fabrosauridae), one decidedly progressive and
successful (essentially the family Hypsilophodontidae as defined below) and
the third representing a short-lived off-shoot from the base of the ornithischian
phylogenetic tree (the family Heterodontosauridae).
To emphasize the points discussed above I present a revised classification for
ornithopod dinosaurs.
Order Ornithischia
Suborder Ornithopoda
A. Grade Dolichopoda
Small-headed and (presumably) cursorial ornithopods. With tibia as long as,
or longer than, femur; metatarsus elongate, at least half length of femur.
1. Family Fabrosauridae
(Upper Triassic-Upper Jurassic)
Cheek teeth marginal; lateral faces of maxilla and mandible non-recessed;
mandible slender, with very low coronoid eminence; cheek dentition discontinuous, with waves of tooth replacement; jaw articulation at or about level
of tooth row; prepubis rudimentary.
Echinodon; Fabrosnurus; 2 genera from the Upper Jurassic of Portugal
(Thulborn, in press); juvenile or young form of “Scelidosaurus” mentioned
by Rixon (1968), Romer (1968) and Charig (1972).
2. Family Heterodon tosauridae
(Upper Triassic)
Cheek teeth inset; lateral faces of maxilla and mandible recessed; mandible
robust, with moderately high coronoid process; cheek dentition continuous and
unbroken, without obvious waves of tooth replacement; caniniform teeth
sometimes present; jaw articulation below line of tooth row; tooth wear planar;
prepubis rudimentary.
Lycorhinus (Heterodontosaurus); ?Pisanosaurus.
3 . Family Hypsilophodontidae
(Upper Jurassic-Upper Cretaceous)
Cheek teeth inset; lateral faces of maxilla and mandible recessed; mandible
massive, with prominent coronoid process; cheek dentition discontinuous, with
waves of tooth replacement; no caniniform teeth; jaw articulation below level
of tooth row; tooth wear irregular, with randomly formed facets in various
planes; prepubis salient.
Dry osaurus ; Dysalo tosaurus ; Hypsiloph od on ; Laosaurus ; ?Nanosauins;
Parksosaurus; a genus from the Upper Jurassic of Portugal (Thulbom, in
press).
R. A. THULBORN
172
4. Family Psittacosauridae
(Lower Cretaceous)
Skull with upper beak formed by rostra1 bone; cheek teeth inset; lateral faces
of maxilla and mandible recessed; mandible massive, with prominent coronoid
process; cheek dentition discontinuous, with waves of tooth replacement; no
caniniform teeth; jaw articulation below level of tooth row; tooth wear
irregular, with randomly formed facets in various planes; prepubis salient.
Protiguanodon ;Psittacosaurus; ?Stenopelix.
B. Grade Brachypoda
Large-headed and (presumably) graviportal ornithopods. With femur as long
as, or longer than, tibia; metatarsus relatively short, less than half length of
femur.
1. Family Scelidosauridae
(Lower Jurassic)
Skull relatively low and broad, with widely open antorbital vacuity;
supraorbital flattened and incorporated into skull roof; prefrontal forms part of
orbital rim; parietals separate (?); femur with pendent fourth trochanter; well
developed dorsal armour comprising rows of keeled scutes.
? Lusitanosaums ;Scelidosaurus (S. harrisonii).
2. Family Iguanodontidae
(Middle Jurassic-Upper Cretaceous)
Skull relatively deep and narrow, with restricted antorbital opening; one or
two rod-like supraorbitals; prefrontal sometimes excluded from orbit; quadrate
tall, slightly arched to front; fourth trochanter of femur pendent or developed
as ‘‘Crete saillante”; no armour.
?Anoplosaums; Craspedodon; Cryptodraco; Iguanodon ; ?Kangnasaurus;
Rhabdodon; ?Sanpasaurus; Tenontosaurus; ? Vectisaurus.
3 . Family Camptosauridae
(Upper Jurassic-Lower Cretaceous)
Skull low and broad, with restricted antorbital opening; single rod-like
supraorbital; prefrontal excluded from orbit; quadrate short, strongly arched to
front; fourth trochanter of femur pendent; no armour.
Camptosaurus.
4. Family Thescelosauridae
(Upper Cretaceous)
Premaxilla with teeth; ventral surfaces of neck centra flattened; fourth
trochanter of femur pendent; armour of scutes.
Thescelosaurus.
A NEW HETERODONTOSAURID DINOSAUR
173
5 . Family Pachycephalosauridae
(Lower Cretaceous-Upper Cretaceous)
Bones of skull roo€ greatly thickened and fused; upper temporal openings
small or absent; premaxilla sometimes with teeth; mandible very short and
massive; fourth trochanter of femur not pendent.
Pachycephalosaurus;Stegoceras; ? Yaverlandia.
6. Family Hadrosauridae
(Upper Cretaceous)
Premaxillae expanded to form duck-like beak; premaxillae and nasals often
extended as “crest”; no supraorbitals; mandible massive, with very prominent
coronoid process; cheek teeth very numerous, crowded into pavement-like
batteries; no armour.
A na tosaurus ; Bac trosaurus ; Brachy lophosaurus; Cheneosaurus; Claorhynchus; Claosaums; Corythosaurus; Dysganus; Edmontosaurus; Hadrosaurus;Hypacrosaurus;?Hypsibema;Kritosaurus;Lambeosaurus;Mandschurosaurus (Nipponosaurusl; ? Ornithotarsus; Orthomerus; Parasaurolophus;
Procheneosaurus; Prosaurolophus; Saurolophus; Tanius; Thespesius;
Trachodon; Tsintaosaurus;Yaxartosaurus.
The relationships of the ornithopod families listed above are expressed in
diagrammatic form (Fig. 5 ) . It may be noted that I regard the appearance of
the cheek teeth, the site of the jaw articulation and the architecture of the
mandible as key characters in defining the first three families of ornithopods.
Without referring to these characters it is difficult t o provide a workable
diagnosis for the family Heterodontosauridae. Tatisaurus oehleri Simmons
(196S), from the Upper Trias of China, may not be an ornithischian (see
Galton, 1972) and is omitted from the classification. Azendohsaurus laaroussii
Dutuit (1 972) (non ilzandohsaurus laarousii Dutuit (1972), Zapsus calarni?)
appears to be a prosauropod, rather than an ornithischian, and is also excluded
from the classification.
The diagnoses for the families of advanced and graviportal ornithopods
(categories B1 and B6 above) are greatly abbreviated; very comprehensive
diagnoses are given by Romer (1956) for the families Pachycephalosauridae,
Hadrosauridae and Ipanodontidae (sensu lato). I t is not possible to provide
detailed diagnoses for the Thescelosauridae and Scelidosauridae on account of
deficiencies in the known fossil material (Thescelosaurus) and in the literature
(Scelidosaurus;see Owen, 1861, 1863).
Finally, I introduce two new terms-Dolichopoda (Greek, “long feet”) and
Brachypoda (Greek, “short feet”)-which refer to the cursorial and graviportal
ornithopods respectively. The Ornithopoda is a large group which includes both
conservative and progressive forms and which has an immense stratigraphic
range (earliest Upper Trias to uppermost Cretaceous). Cleavage of the
ornithopods into two grades is of particular value in that it permits separation
of the conservative main stem of the ornithischian phylogenetic tree
(Dolichopoda) from the several more advanced lines of ornithopod evolution
(Brachypoda). Romer (1966) listed six families of ornithopods while Galton
(1972) recognized seven. The splitting of Triassic ornithischians and of
174
R. A. THULBORN
“iguanodontids” (sensu lato) into several families has increased the number of
ornithopod families to ten (see above). My grouping of these ten families into
two grades gives a useful generalized impression of ornithopod evolution, which
would be unobtainable from a list of the families alone, and is broadly
comparable with the practice of dividing the theropod saurischians into
coelurosaurs and carnosaurs. The grade Dolichopoda appears to be a natural
group but the grade Brachypoda, in contrast, is decidedly polyphyletic (see
Fig. 5); consequently I have hesitated to accord these categories the status of
infraorders. The adjectival forms dolichopodous (= long-footed) and brachypodous (= short-footed) are essentially anatomical and may usefully be
employed in any formal definition of a taxon. It must be emphasized, however,
that the terms dolichopodous and brachypodous are not synonyms for the
epithets cursorial and graviportal; these latter terms are quite undesirable in any
scheme of classification since they may be applied to a specimen only by means
of a subjective decision. Finally, it is worth noting that the nominal forms
dolichopod and hrachypod would conveniently serve as replacements for the
now restricted terms “hypsilophodontid” (sensu Romer, 1956; Thulborn,
1970a, b, 1971b, 1972) and “iguanodontid” (sensu Romer, 1956, and many
other authors).
ACKNOWLEDGEMENTS
My particular thanks go to Dr Kenneth A. Kermack (University College
London) who generously provided the material basis for this paper and offered
many helpful suggestions. I also thank Professor William J. Morris (Occidental
College, Los Angeles) for permission to mention his discovery of scutes in
Thescelosaurus neglectus.
REFERENCES
BROOM, R., 1911. On the dinosaurs of the Stormberg, South Africa. Ann. S. Afr. Mus., 7: 291-308.
BROOM, R., 1932. The mammal-like reptiles of South Africa and the origin of mammals. London:
Witherby.
CASAMIQUELA, R. M.. 1967. Un nuevo dinosaurio ornitisquio trihsico (Pisanosaurus merrii; Ornithopoda) d e la formacion Ischigualasto, Argentina. Ameghiniana, 5: 47-64.
CASIER, E., 1960. Les Zguanodons d e Bernissart. Brussels: L’inst. toy. Sci. nat. Belg.
CHARIG, A. J., 1972. The evolution of the archosaur pelvis and hind-limb: an explanation in functional
terms. In K. A. Joysey & T. S. Kemp (Eds), Studies in vertebrate evolution. Edinburgh: Oliver &Boyd.
CROMFTON, A. W. & CHARIG, A. J., 1962. A new Ornithischian from the Upper Triassic of South
AfrGca. Nature, Lond.. 196: 1074-7.
DORST, J. & DANDELOT, P., 1970. A field guide to the larger mammals of Africa. London: Collins.
DUTUIT, 3.-M., 1972. DCcouverte d’un Dinosaure ornithischien dans le Trias supkrieur de I’ Alas
occidental marocain. C. r. hebd. Skanc. Acad. Sci. Paris, 275(D): 2841-4.
EDMUND, A. G., 1957. On the special foramina in the jaws of many Ornithischian dinosaurs. Contr. R.
Ont. MS. ZooL Palaeont. 48: 1-14.
EDMUND, A. G., 1960. Tooth replacement phenomena in the lower vertebrates. Contr. R. Onr. Mus. Life
SCL, 52: 1-190.
EVERY, R. G., 1965. The teeth as weapons. Lancet, 1965: 685-8.
EVERY, R. G., 1970. Sharpness of teeth in man and other primates. Postilia, 143: 1-30.
EVERY, R. G. & KUHNE, W.G., 1971. Biomedical wear of mammalian teeth. In D. M. Kermack & K. A.
Kermack E d s ) ZooL J. Linn. Soc., 50, Suppl. 1: 23-7.
EWER, R. F., 1965. The anatomy of the Thecodont reptile Euparkeria capensis Broom. Phil. Trans. R.
SOC.L o n d , 248tB): 379435.
GALTON, P. M., 1970. Ornithischian dinosaurs and t h e origin of birds. Evolution, 24: 448-62.
GALTON, P. M., 1972. Classification and evolution of omithopod dinosaurs. Nature, Lond., 239: 464-6.
A NEW HETERODONTOSAURID DINOSAUR
175
GALTON, P. M., 1973. The cheeks of ornithischian dinosaurs. Lethaia, 6: 67-89.
GUGGISBERG, C. A. W., 1972. Crocodiles: Their natural history, folklore and conservation. Newton
Abbot: David & Charles.
HATCHER, J. B., MARSH, 0.C. & LULL, R. S., 1907. The Ceratopsia. Monogr. U.S. geol. Surv., 49:
1-1 57.
HAUGHTON, S . H., 1924. The fauna and stratigraphy of the Stormberg series. Ann. S. A@. Mus., 12:
32 3-497.
KRYNINE, P. D., 1949. The origin of red beds. Trans. N. Y. Acad. Sci., (2), 22: 60-7.
KUHN, O., 1966. Die Reptilien. Munich: Krailling.
LULL, R. S., 1917. The Triassic fauna and flora of the Connecticut Valley. Bull. U.S. geol. Surv., 597:
105-27.
MABERLY, C. T. A., 1963. The game animals of southern Africa. Cape Town: Nelson & Sons.
MORRIS, W. J., 1970. Hadrosaurian Dinosaur Bills-Morphology and Function. Conrr. Sci., 193: 1-14.
NEWMAN, B. H., 1968. The Jurassic dinosaur Scelidosaurus harhoni. Palaeontology, 11: 4 0 - 3 .
OSTROM, J. H., 1961. Cranial morphology of the hadrosaurian dinosaurs of North America. Bull. Am.
Mus. nat. Hist., 122: 35-196.
OSTROM. J. H., 1970. Stratigraphy and Paleontology of the Cloverly Formation (Lower Cretaceous) of
the Bighorn Basin Area, Wyoming and Montana. Bull. Peabody Mus. nat. Hist.. 35: 1-234.
OWEN, R., 1861. A monograph of the fossil reptilia of the Liassic formations, PartI. London:
Palaeontogr. SOC.
OWEN, R., 1863. A monograph of the fossil reptilia of the Liassic formations, PartII. London:
Palaeontogr. SOC.
PETTIJOHN, F. J., 1957. Sedimentary rocks, 2nd ed. New York: Harper & Row.
RIXON, A. E., 1968. The development of the remains of a small Scelidosaurus from a Lias nodule.
Musfurns J., 6 7: 3 15-2 1.
ROMER, A. S., 1956. Osteology of the reptiles. Chicago: Chicago Univ. Press.
ROMER, A. S., 1966. Vertebrate paleontology, 3rd ed. Chicago: Chicago Univ. Press.
ROMER, A. S., 1968.Notes and comments on vertebrate paleontology. Chicago: Chicago Univ. Press.
SCHWARZBACH,’M., 1963. Climates of the past. (Trans. R. 0.Muir). London: Van Nostrand.
SIMMONS, D. J., 1965. The non-Therapsid reptiles of the Lufeng Basin, Yunnan, China. Fieldiana
(Geol.), 15: 1-93:
THULBORN, R. A., 1970a. The systematic position of the Triassic ornithischian dinosaur Lycorhinus
angustidens. Zool. J. Linn. SOC.,49: 235-45.
THULBORN, R. A., 1970b. The skull of Fabrosaurus australis, a Triassic ornithischian dinosaur.
Palaeontology, 13: 414-32.
THULBORN, R. A., 1971a. Tooth wear and jaw action in the Triassic ornithischian dinosaur Fabrosaurus.
J. Zoology, 164: 165-79.
THULBORN, R. A., 1971b. Origins and evolution of Ornithischian dinosaurs. Nature, Lond., 234: 75-8.
THULBORN, R. A., 1972. The post-cranial skeleton of the Triassic ornithischian dinosaur Fabrosaurus
australis. Palaeontology, 15: 29-60.
THULBORN, R. A., in press. ContribuiCao para a Fauna d o Kimeridgiano da Mina de Lignito Guimarota
(Leiria), Portugal. Teeth of ornithopod dinosaurs from the Upper Jurassic of Portugal, with
description of a hypsilophodontid from the Guimarota lignite. Mem. Serv. GeoL Portugal.
TOERIEN, M. J.. 1953. The evolution of the palate in South African Anomodontia and its classificatory
significance. Palaeont. afr., 1: 49-117.
WALKER, E, P.; WARNICK, F., HANLET, S. E., LANGE, K. I., DAVIS, M.A., UIBLE, H. E. &
WRIGHT, P. F., 1964. Mammals of the world, 2. Baltimore: John Hopkins.
ABBREVIATIONS USED IN FIGURES
an
a0
di
dl
dr
en
ju
la
ma
?angular
antorbital opening
diastema
left dentary
right dentary
external naris
jug
lacrimal
maxilla
orbit
predentary
prefrontal
premaxilla
postorbital
quadrato-jugal
quadrate
Zsurangular
sclerotic plates
Zsupraorbital