Palaeont. afr., 18, 89 - 131 (1975)

THE MORPHOLOGY AND RELATIONSHIPS OF YOUNGINA CAPENSIS BROOM
AND PROLACERTA BROOMI PARRINGTON

by
C. E. Gow

ABSTRACT

Comprchcnsive descriptions o f the osteology of Youngina capensis Broom and Prolacerla broomi
P<llTinglon arc presenled. New de t<lils of the braincase of Prolerosuchus Jergusi Broom are given as
these became necessary for comparative purposes. It is suggested that the initial radiation of
sauropsid reptiles was a Permian event as yet poorly documented. The phylogeneti c position of
Yo III/gil/a bOlh forward and backward in time cannot be narrowly defin ed , though certain
charae!crs sccm spccific<lll y to preclude it from lizard ancestry. Proiacerla, on th e basis of tooth
implanlalion. braincase morphology and postcra nial anatomy is shown to be closest to the
prolcrosllchia n Ihecodonts. It is very definitely not concerned with lizard origins , but would on
<lvailablc cvidcncc seem to be a perfectly good ancestor for the middle Triassi c fo rms
M aOWI/1'1I/1IS and TrlnV.llro/lheus, whi ch latter must cease to be regarded as li za rd ancestors. We
have herc a ralher distinct reptilian lineage which branched off from common ancestral stock
.illsl prior 10 Ihe advcnt of archosaurs.

INTROD UCT ION ....
YOUNG INA

M<llerial and Methods
OSleology
Fllne! ion . . . . . .
Relalionships
Synonymy and diagnosis

PROLACERTA
Review of literature and material
Ma terial and Metho ds
Osteology
Prolerosuckus braincase
Flllle! iOIl
Eco logv .....
Rela I io'nships
Synonymy and diagnosis

DISCUSS IO N
SU MMARY
REFE RENCES
LIST OF ABBR EV IATIO NS

CONTENTS

Page

90
90
97
98

99
99

100
III
112
116
116
118
118
123
123
131

INTRODUCTION are notas generalised as has until now been believed.

The Permian eosuchians, typified by Youngina, are
generally considered the hub oflater diapsid reptilian
divers ification (e .g. Romer, 1956 ).

Youngina is relatively poorly known ; thus part of the
present study will describe the anatomy of this animal
in detail. To begin with it will be necessary to decide
just what we mean by Eosuchia and to try to decide
what is to be understood as comprising the genus
Youngina. Here it is necessary to be very explicit as it is
becoming clear that the Upper Permian was a crucial
period in sauropsid evolution and that relationships

The Eosuchia are an order within the Subclass
Lepidosauria and in the light of the present study must
be added to Romer 's (1956 ) otherwise satisfactory
definition that dorsal dermal armour was present. The
genera comprising Romer's family Younginidae are a
reasonably homogeneous group, with the exception
of Noteosuchus, which shows rhynchocephalian af
finities (Carroll, pers. com.). Of the rest the best
material in existence is that of Youngina (with its tax
onomic variants ), and these with their smooth peg
like marginal dentition can be distinguished from
H eleosaurus (Carroll, in press ) which has serrated

marginal dentition and certain other features which
put it more closely in line with the Archosaurs.

Broom (1914 et seq.) was responsible for the three
generic and five specific names with which what is
mani festly one form (Youngina capensis) is presently en
cumbered. This string of names is misleading; all the
specimens come from the same horizon, differing only
in size, state of preservation and preparation, and
nature of deformation, which factors adequately ac
count for all supposed taxonomic differences. One
can have no doubts about synonymising them all un
der Youngina capensis to stress the probability that this
was a single monospecific genus. This work will pre
sent a detailed description of this animal based on all
useful material.

Pro lacerta (Parrington, 1935), on the basis of its in
complete lower temporal bar, has been considered to
represent a stage between the Eosuchia and the
Squamata, but this concept must be reject~d. Acritical
re-examination of the true position of Prolacerta is
necessary. Until now Pro lacerta (which includes Pricea)
has been known from a few skulls and some cervical
vertebrae. Largely on the basis of an incomplete lower
temporal bar it has been regarded as ancestral to the
Squamata. The present study describes the whole
skeleton and the animal is shown to be Archosaurian
in almost every respect.

The braincase of Proterosuchus is here re-described in
some detail for comparative purposes.

YOUNGINA

MATERIAL AND METHODS
The descriptions of Youngina are based on

material in the Bernard Price Institute (BPI) and one
partial skull from the Transvaal Museum (TM). The
several specimens in the Rubidge Collection (RC)
were examined. Professor Crompton kindly provid
ed copies of his unpublished drawings of Youngina
rubidgei; these, though useful, have not been used
here, and there are one or two points where we
differ. These points of difference are those which it
has been possible to prepare more fully. A cast and
photographs of the Chicago specimen (Youngoides
romeri ) were made available to the author.

All Youngina material comes from the Dap
tocephalus zone of the Karroo. The state of preserva
tion is rather different from that of the millerettids
(Gow, 1972 ) of the same age, which might imply
some ecological differences in life. While the
millercttids are articulated and distortion free, the
younginids are disarticulated and the skulls are dis
torted.

Most of the Youngina material has in the past been
subjected to mechanical preparation which had
caused some damage and also restricted the scope of
further work. Fortunately it was possible to prepare
a good braincase and skeleton in formic acid; other
details were pieced together from mechanical
preparation of selected areas of different skulls.

Material referable to Youngina capensis Broom 1914.
A.M. N.H. 5561: Younginacapensis Broom 1914. Skull

and vertebrae, in American Museum of Natural
History.
Locality: New Bethesda, Graaff- Reinet.
Collected: R. Broom.

u.c. 1528: Youngoides romeri Olson and Broom 1937.
Skull in University of Chicago.
Locality:Towerwater, Murraysburg.

R. C. 90: Y oungopsis rubidgei Broom and Robinson
1948. Very good skull.
Locality: Doornkloof, Graaff- Reinet.
Collected:J. W. Kitching, 1939.

R. C. 91: Youngoides minor Broomand Robinson 1948.
A poor partial skull. A second undescribed
specimen with the same number is also Youngina.
Locality: New Bethesda, Graaff- Reinet.
Collected: Kitching Bros., 1940.

T.M. 1490: Youngopsis kitchingi Broom 1937. A very
poor skull.
Locality: New Bethesda, Graaff- Reinet.
Collected:J. W. Kitching.

T.M. 200: Younginacapensis Broom 1922. Fragments of
postcranial skeleton, now even fewer than when
described by Broom, but certainly of Young ina.
Locality: New Bethesda, Graaff- Reinet.
Collected: I. Venter, circa 1920.

T.M. 3603: Posterior half of a large skull, now acid
prepared.
No catalogue data.

R. C. 625 and 626: Two poorly preserved skulls.
Locality : Well wood , Graaff- Reinet.
Collected: S. H. Rubidge, 1936 -45.

R.C. 714: Incomplete skull.
Locality: Ganora, New Bethesda, Graaff-Reinet.
Collected: S. H. Rubidge, 1940.

K. 106: Badly flattened skull in the collections of the
Geological Survey, Pretoria. -
Locality: Blaaukranz.
Collected: A. W. Keyser.

B.P.1.375:Skull.
Locality: von der Waltzhoek, Graaff- Reinet.
Collected:J. W. Kitching, 1946.

B.P.1.2459: Small flattened partial skull.
Locality: Klipplaat, Richmond.
Collected:J. W. Kitching, 1951.

B.P.1.2871 : Skull.
Locality: Beeldhouersfontein, Murraysburg.
Collected:J. W. Kitching, 1958.

B. P.1. 2859: Skull and skeleton.
Locality: Doornplaas, Graaff- Reinet.
Collected:J. W. Kitching, 1964.

OSTEOLOGY OF YOUNGINA

The Skull

Dnma/ bonf'1 of the skull roof

PTf'1llaxil/a (Figures 1 and 6). The premaxilla is
fairly well represented only in B.P.I. 2871. There

N ------t.

2cm

Figure I. Youngina capensis B.P.I. 287 J.

appear to have been three premaxillary teeth. The
important feature of the premaxilla is its medial ex
tent in the palate: the premaxillae run well back in
side the maxillae in a strong contact and meet in the
mid line.

Maxilla (Most Figures). The maxilla sheaths most of the
slender, rather shallow snout, and has an extensive
contact with the palatine. The teeth are conical, sharply
pointed and not serrated; they thicken lingually at the
base much as described for Milleretta ( Cow, 1972); they are
sub- thecodont and become firmly ankylosed in deep pi ts.
An exact count is not possible but there are about30 tooth
positions with alternate replacement so that one gets a
count of about 20 functional teeth of different ages. An
important aspect of the maxillary tooth row is that it
extends back to below the postorbital bar (c.f.
Proterosuchus Cruickshank, 1972, and crocodilians, but
not lizards).

SeptomaxiLla. Detailsofthis bone are totally lacking.
Nasal (Figures 1 and 6). The relationships of the

nasals are clear except anteriorly where there is some
doub t as to the exact position of the sutures with the
premaxillae, but this is oflittle consequence.

Lacrimal (Figures 1 and 6). The lateral exposure of
the lacrimal is reduced to a short distance in front of
the orbit. A single foramen runs the full antorbital ex
tent of the bone (as seen in section in R. C. 625 ).
. Prefrontal. As figured, requires no special descrip

tIOn .
Frontal (Figures 2 and 6). The frontals are the only

.skull roofing elements which have roughened sur
faces, notably between the postfrontals and above the
orbits . The frontals are rigidly sutured to theparietals.

Parietal (Figures 2, 4, 5 and 6). The parietals enclose a
large pineal opening and send out posterolaterally
directed wings. The area bordering the upper tem
poral opening is deep and has a shallow pocket in the
posterior corner affording a strong area of origin for
part of the jaw adductor muscle complex. The wings
are joined by a slight ridge between skull table and oc
ciput. Each lateral wing forms important at
tachments; anteriorly there is an extensive firm con
tact with the squamosal (Figures 4 and 6): applied to
the posterior surface are a post-parietal and tabular,
while the tip is sheathed by a supratemporal whose ex
act relationships will be discussed separately. The
lateral wings roof large post-temporal fossae and
there is no firm contactwith the supraoccipital.

Postfrontal. The postfrontal is firmly sutured
between frontal and postorbital.

Postorbital. The postorbital is overlapped above by
the post-frontal (this region, the anterior border of
the upper temporal opening, is deep); ventrally it has a
rigid sloping contact with the jugal , and posteriorly a
broad flat sheet has an extensive overlap onto the
squamosal.

Jugal (Figures 2, 5 and 6). This is a straightforward
elementhavinga longslopingcontactwith the maxilla
and short rigid connections with post-orbital and
quadratojugal. Where it forms the anterior border of
the lower temporal opening there is a distinctive
depression (indicated in Figure 6, lateral view). Inter
nally the jugal has the ectopterygoid abutting against
it. It has not been possible to expose the inner surface
of this region of the jugal.

Q.uadratojugal (Figures 2 and 6). This is an element

ST -~~'/I

_
___ ",2iij

Cl'::ffi====-_ . BPI. 3859.

. penslS. . . 2 Youngma ca FIgure .

F :J--v \

2cm

. PI375. . pensls B ... . 3 Youngma ca FIgure .

lCffi

tvIX

L --+:177/r-.. !

PRF

PTGD

PSP

(A)

pp T

SO ST

par. proc. SQ
EO Q

q.ra. pt.

2cm

Figure 4 . Youngina capensis T.M. 3603 . A, Dorsal ; B, Ventral; C, Occipital ; D, Internal detai l of
frontoparietal region .

critical to phylogenetic discussion . The contact with
the jugal presents no problem; itis the relationships to
squamosal and quadrate which are important. With
B. P. I. 3859 (Figure 2) it was possible to remove part of
the squamosal to reveal the full extent of the
quadratojugal beneath it. T.M . 3603 provides a check
in that the upper two-thirds of the inner surface of the
squamosal are clean and there is no sign of any dorsal
continuation of the quadratojugal. Unfortunately no
specimen shows the contact of quadrate and
quadratojugal.

Squamosal (Figures 2, 4, 5 and 6). The squamosal
overlies part of the quadratojugal below, then rises up
in contact with the quadrate, finally turning inwards to
cap the quadrate, running under a facet of the post
orbital and lapping against the parietal wing.

Post parietal, Tabular and Supratemporal (Figures 4, 5
and 6) . There is reasonable certainty regarding the
presence of these three elements. Postparietal and
tabular are simply applied against the back of the
parietal. The supratemporal on the other hand is more
complex; it lies against the back of the parietal wing
and extends beyond its tip, turning down and
forwards to form a hook which lies against the top of
the pterygoid flange of the quadrate . This hooked
supratemporal possibly supported a carti laginous
pad which could rotate against the paroccipital
process .

Sclera. These are present in B.P. I. 2871 and suggesta
ring of ± 12 flat plates .

The palate. The palate of Youngina has to date been
poorly known and imaginatively drawn. Olson's
(1936) reconstruction is poor and has been used by
Romer (1956): in this, the anterior of the palate is quite
wrong as is the epipterygoid stuck in the basal articula
tion .

Vomer (Figures 2 and 6). The vomers are fused in
the midline . Anteriorly they disappear above the
palatal extensions of the premaxillae which are also
united in the m idline. Relationships with palatine
and pterygo id are clear. Median and lateral borders
of each vomer bear a row of small teeth: the median
row divides into two along the posterior third of its
length . Medial to the tooth rows where the vomers
meet th ey are upturned to form what in palatal view
is a narrow smoothly rounded trough. There are
two or three larger teeth grouped at the tip of each
vomer.

Palatine (Figure 2). The relationships of the
pa latine are clear; it bears four rows of teeth, one a
continuation of the lateral row on the vomer, the
others of corresponding rows on the palatine. The
foot in contact with the maxi lla projects downwards
so mewhat from the level of the palate.

EctojJter.Yf!,oid (Figures 2 and 6). The ectopterygoid
overlaps onto the palatal surface of the pterygoid,

extending well down the pterygoid flange. This ele
ment is poorly known in more primitive reptiles, but
by analogy with Millerettids it appears that the in
fraorbital fenestra in Youngina is formed by reduc
tion of the ectopterygoid. The extent of the surface
abutting against the jugal is not clear.

pterygoid (Figures 2, 3, 4, 5 and 6). The palatal por
tion of the pterygoid bears a double row ofteeth on the
mesial edge continuous with a double row on the
vomer. Three rows of teeth fan out from near the basal
articulation and continue onto the palatine. The
strongly down-turned flange of the pterygoid bears a
row of 6 or 7 robust teeth. The articulation with the
basisphenoid is freely movable. The quadrate ramus
of the pterygoid is a rigid triradiate structure with its
edges directed laterally, mesially and dorsally, enclos
ing a dorso-mesial depression (for the protractor
pterygoideus). There is a considerable overlap
between pterygoid and quadrate, and a marked sup
portingridge on the quadrate.

The palatoquadrate
Epipterygoid (Figure 5 E). Both epipterygoids are par

tially exposed in T.M. 3603. The upper portio!, of the
shaft is missing; there is no sign of an attachment area
on the parietal. A broad footplate rests on the
pterygoid very much as illustrated for Milleretta (Cow,
1972, Figure 8) and, by analogy, probably caps the basal
articulation.

Quadrate (Figures 3, 4, 5 and 6). The quadrate isheld
laterally and dorsally by the squamosal and internally
by parietal and supratemporal above and the
pterygoid below. Anteriorly it is braced by the
quadratojugal. l'his is a rigid system. The pterygoid
ramus is a tall flat sheet. The ridge supporting the
quadrate ramus of the pterygoid extends to the back of
the quadrate, as is clear from Figures 3, 5 and 6; lateral
to this and sloping in from the back of the outer con
dyle is an abrupt depression. The stapes would have
passed across this region. As is plain from the
reconstructed lateral view the quadrate is held ver
ticallyand the term "otic notch" hardlyapplies.

The braincase (Figure 5)
Supraoccipital. The relationships of this element are

largely evident from the figures. The supraoccip~al
runs in under the parietals, this area being bridged by
the postparietals. The bone slopes obliquely down
and backwards.

Opisthotic. The structure and relationships of the
opisthotic are clear from the figures. The point to note
here is that the paroccipital process articulates in a
pocket formed by the supratemporal (Figure 5 E).

Exoccipitals and Basioccipital (Figures 4, 5 and 6).
These form a fused unit. The exoccipitals almost meet
above the foramen magnum behind the supraoc
cipital. The overlap between parasphenoid and
basioccipital is clear from Figure 5 C.

Prootic (Figure 5C, 0 and E). The region of the
fenestra ovalis is poorly ossified (and this is a large

POF

2cm

(D) (El

lcm

Figure 5. Youngina capensis T.M. 3603. A, Lateral view; B. Mesial
view of quadrate and cheek region; C, Saggital section
through braincase; D, Braincase dorsal view; E,
Braincase lateral view.

specimen), so that the prootic appears susp~nded
from a strong suture with the supraoccipital whileven
trally it rests on the basisphenoid. The system of inner
ear canals converging at the fenestra ovalis from
supraoccipital, opisthosthotic and prootic is well dis
played in the specimen.

Basisphenoid (Figures 3, 4, 5 and 6). Basisphenoid
material is poor. The nature of the basipterygoid
processes is clear and the articulation possible here is
very similar to that described for millerettids (Cow,
1972); the low clinoid processes too are very similar.

Parasphenoid (Figures 3, 4, 5 and 6). The
parasphenoid is edentulous. It overlaps the basioc
cipital extensively and has a pronounced semicircular
lip posteroventrally (seen in section in Figure 5 C).

Stapes (Figure 6). The stapes is a thin rod pierced
near the proximal end by a stapedial foramen. The
foramen is bounded by an extremely thin bridge
which is bowed slightly outwards; such a foramen
could well disappear in later forms.

The lower jaw (Figure 6)
There is no good lower jaw material: the region of

the retroarticular process is eroded off in all
specimens. The tooth~bearing rami are rather slender
while the region of the adductor fossa is considerably
deeper.

PMX

. . .
. .. . .

II---~-' . MX ·· .. ----f.

" . . " , .. :.

.. " . .
." ."

' - •• "; 0"

-." .. -

...... ". ".
'. .
. . . " , .'

-H>--' ...... · MX .
. -- - .

. . 95 ·

~,,-'r'---,-,-, PRF .' '.

POF

. ,.".

Figure 6. Youngina capensis . Composite skull reconstruction.

ThepostcranialskeletonOfYOUNGINA .
The skeleton ofB.P:1. 2859 was jumbled together in

a tight bundle behind the skufl and part of it had
already rotted away. Before acid preparation could
commence the skull which had been extensively
mechanically prepared had to be removed along a
natural joint. The forelimb too was prepared
mechanically and removed. The remainder was then
disentangled by the repeated use of formic acid. The
skeleton comprises a nearly complete set of presacral .
vertebrae apparently lacking only the atlas, axis and
first sacral. The second sacral and first caudal are at
tached and there are two other caudals and two
haemal arches (thought absent by Broom, 1922).
There are several plates of middorsal armour. There
are several ribs. All elements of both girdles are
represented. The left forelimb is complete, also the
right humerus. The left femur and the proximal end of
the tibia and a presumed fourth metatarsal are all that
remain of the hind limb. Thus it is only the hind limb
which is incomplete.

Vertebrae (Figure 7) and ribs
The neural arches are broad and flat with moderate

ly tall spines. The zygapophyses are horizontal, the
transverse processes are pronounced, slanted, and
have facets for single rib heads. The centra are
amphicoelous and notochordal and invisibly fused to
the neural arches. Intercentra are present.

The first sacral is missing; in the second the ribs are
directed forwards and the distal ends . bifurcate, the
more anterior branch articulating with the iliumand
the posterior branch forming a continuation of the
caudal transverse processes and a site of attachment
for caudal musculature. The caudal vertebrae with
their strong transverse processes and deep haemaL
arches indicate a deep ' and powerful tail. A small .
number of symmetrical middorsal scines are preserv
ed (Figure 8), sufficient to indicate that there was one to
each vertebra but there is no indication as to the extent
of the row.

Girdles and limbs (Figures 9and 31)
The pectoral girdle comprises a T-shaped In

terclavicle (with anteroventral notches for the
clavicles), and a scapulocoracoid anteriorly notched
to leave a gap between it and the clavicle. The ends of
the humerus are rotated at 900 to each other. There is
an entepicondylar foramen and a pronounced
ectepicondylar groove. Radius and ulna require no
additional description. The wrist is practically intact:
there is a large ulnare and a rather small radiale, an in
termedium, lateral and medial centrales, and five dis
tal carpals. The digits are complete and bear powerful,
laterally compressed, curved claws. The pelvic girdle is
rather similar to that of Howesia(Broom, 1906). The il
ium has a short, almost vertical blade, the pubis a
pronounced thickened outturned antero-ventral

(el

(E 1

(A)

lcm

OJ . " ... . : .: .. ;;

(F 1 (Gl

Figure 7. Youngina capensis B.P.1. 3859. A, Cervical vertebra; B, Anterior dorsal vertebra; C, Se
cond sacral and first caudal vertebrae; D, Proximal caudal vertebra; E, Proximal
haemal arch; F, Distal caudal vertebra; G, Distal haemal arch.

(Cross hatching of neural spines and articular facets does not indicate damaged surfaces .)

\!%)", ... '.'.'
.. .. : )~:;;: .

.< : ... '\.

lcm

Figure 8. Youngina capensis B.P.I. 3859. Mid-dorsal scutes .

process, and the ischium extends well back. The
acetabulum is confined to the ilium. The notch
between pubis and ischium would, according to
Romer (1956, p. 318), indicate incomplete ossification
and notan incipient thyroid fenestra.

The femur is primitive, though with a marked cur
~ature which seems natural. The tibia is unfortunately
Incomplete. There is a single metatarsal which is
almost certainly the fourth.

The following are the limb element measurements
In mm:

Humerus 23
Ulna 16 Radius 18
Femur 34

The skeletal fragments described by Broom (1922)
were alm<?st certainly correctly assigned to Youngina,
and as he Illustrated a humerus and tibia it is possible
t? ~educe that the femur and tibia were of closely
similar length, as shown in the reconstruction (Figure
10).

The ankle described by Broom (1922) is now ap
parently irretrievably lost, but it is quite clear from
Broom's figure and text that it is very like that of early
rhynchocephalians and archosaurs.

YOUNGINA FUNCTION
~ith its long, low, and rather narrow snout, sharp

pOinted teeth and dermal sculpturing, the Youngina
skull has a rather crocodilian appearance. Belying this

1!f!r
I I
I I

" \ I

2cm

\l
I

I
I
I
I

'.
---'

(8)

Figure 9. Youngina capensis B.P.!' 3859. A, Left forelimb ; B, Femur,
tibia and fourth metatarsal in anterior view; proximal and
distal ends of femur ; C, Right aspect of pectoral girdle; D,
Right pelvic girdle.

impression are the apparently unspecialised terminal
nares.

Broom (1922) figured the humerus and tibia of
Youngina; combining measurements from these with
data available from the present study we can conclude
with a fair degree of certainty that the tibia was shorter
than the femur, but not dramatically so-this points
towards a terrestrial quadrupedal existence.

The tail is important, but as the interpretation of tail
function is important with Prolacerta as well, and as this
is a complex and neglected field, some general
remarks about tails are in order at this point. Several
parameters control the shape and weight distribution
of the reptilian tail and some may change according to
their position along its length. One may cite: height,
breadth and thickness of neural spines; breadth across
the transverse processes; position, depth, breadth and
thickness of chevrons, and the rate of reduction in size
of all the above from root to tip of tail. Clearly a detail
ed study of tail morphology and function would fill an
important gap in present knowledge.

The Youngina tail has very low neural spines, which
could possibly be simply the retention of a primitive
captorhinomorph character. The chevrons on the
other hand are long in proportion. Looking for com-

.. 98 .·

Scm
' . -.'

Figure 10 .. Youngina capensis. ·Reconstruction of the skeleton.

parisoris we find .that in Sphenodonthe chevrons are
substantial but balanced . by tall neural spines.
Something dosely approaching the Youngina condi
tion is seen in one of our local lizards, Cordylus
giganteus, a fairly large spiky armoured lizard of the
grassveld. Pairs live in burrows of their own making,
and judging by the even distribution of burrows are
strongly territorial. In this animal neural spines ·are
very poorly developed throughout and the chevrons
are comparable iii size to those of Young ina. This
similarity of tail structure, however, may be just that
and no more, anditis interestingthatSphenodonandC.
giganteuswhich have moreor less the same way oflife in
several respects should have such different tails. C.
giganteus would probably fare poorly as a swimmer.
This may be a case where it is not possible roexplain
the observed differences in functional terms.

It should be noted that in swimming reptiles (e.g.
crocodiles, monitors) the distal caudal vertebrae have
tall neural spines which balance longchevrons in sup
porting a flattened, moreor less symmetrical, tail.

There can be little doubt that Youngina was a
terrestrial animal. Girdle measurements relating to
body sections are given in Table 1.

. " . Crocodylus Youngina
niloticus capensis

YOUNGINARELATIONSHIPS
While the skull of Youngina is essentially Permian

(braincase, palatal dentitiOn} it exhibits certain ad
vances which foreshadow the Triassic early thecodont
grade (considerable palatal .extent of premaxillae,
suborbital vacuity, and slender stapes ). This should
not be construed to imply a direct relationship
between Youngina and a possible Prolerosuchus-Prolacer
ta ancestor. These are simply time grades of reptilian
evolution, and while it is useful to know what the
"Eosuchian" skull looks like there are other small lit
tie known U . . Permian eosuchians which rria y prove to
be more directly · on specific lines to the Triassic
Thecodonts (e .g.H eleosaurus) .

One must concede that the diapsid upper temporal
opening is a rigid distinctive feature shared by
Youngina and its kin plus all later diapsids. The one
serious drawback to Youngina as a thecodont ancestor
is in the structure and relationships of the quadrato
jugal. Both the tall "archosaurian" quadratojugal ot
e.g. Proterosuchus and the small articulating slip of a
bone inProlacerta are readily derivable from the sort of
condition seen in millerettids (Gow, 1972, Figures 8
and 22 ) but not from the arrangement in Youngina.

Table I
Sphenodon Chamaeleo Prolacerta
punctatus dilepis parringtoni

Measurements in mm.

Inte rgleno id
width IGW 241 28 34 II 32

Interacetabula r
width lAW 127 19 28 9 30

P",o" 1 }
girdle depth 17 8 18 33 18 50

Pe lvic
girdle dep th

Taken togethe r
as girdle
depth GO 17 8 22 34 16 46

Ra tios
IGW : IAW 2 :1 3 :2 1:1 1 :1 1:1
IGW :G D 4 :3 3 :2 1 :1 1 :2 3 :5
IAW :GD 2:3 1 :1 1 :1 1 :2 3 :5

Bo dy Shape Do rsoventra II y Round La tera lly
compressed bodied compressed

99 '

. This is notto suggest arelatio'nship between milleret- ·.· . has undescribed material. of equivaleni:;ige. collecred
tids and thecodonts. Thisquadratojugalobjection . by Kitchirig. in Antarctica.l Pro[acerta.broomfwasfirsr
does not (lpply to-rhynchocephalian derivation from ' described ' by Pairing-ton .' inT935. parr-jng-ton' s .
Youngina, but it does rejecra too Close relationship specimen lacks ,the posterior spur of t~ejug<il.With .·
between rhynchbcephalians and earlythecodonts . -· thisex:ception; ' Pa~ington's~ n3constrUction IS ' good

It seems likely that where the .smallPermian 'artdthesp~cimen isanacceptabletype. o, ." _ '. '.' ...
eosuchiansare :concerrteddetails of thepOstcranial - ' Gafl1p(1945)describ~daseconds:kul1andsix cer
skeleton Will prove significant in assigningaffiniti·es.vical veftebrae. This iSc'tn .excdl~nrspecimen and .'
The girdles and limbs of Youngina{the pelvic girdle in . . Carnp'srec6ri~ttuctions<'ie;on i:hewhble, good; .

. partiCliladare . strikingly . similar to those of the .-.BroomandRobinson (l948)desctibedathirdskull
rhy'nchocephalianHowesia (Broom, 1906); . . ,: .. , . as Pric.er;dongieep5; hut the:bniy respect in which this '

. - differs frpm thepr,evio.us t\VoskulIsi.sinthe, n~tureof . '
YO UN G [N.(~SYNO NYMY . 'AND ·n fA G N 0 SIS its -preseryallon;suikiriglydi fferenttypesQf pteserva-

His riOt d.esi.rab.l.e at this stage to consi.derthe ord. inal .. tion ofthe .sali1e~ sp~ciesare typicaLofthe~ L,Y5tiosaurus
zone;asall farilil-iar'-Viththe 'zonef'ossilswillknQw;"

and farnilialstatus of Younginl1 . Future work must . . A fourtns.· ku.· ··.l1;irt :th.· .~co. lle.c. tion. s: of ·the ·.· South ·,
decide whether the Order Eosuchia' will remain as a
depositcjiYfoiun.~. eblt.edpr.imitivediapsidso.fdoub.t" . - AfriCanMuseum; .CapeTowri.,:hasbeenwoikecionby

. Crompton who h. as.·. kindlycgiVeht.Ni Writercopiesof - "
ful allni ty. Certainly several unrelated families are , his detailed drawings. .'. . ....... ........., .
represented in the Youngiriidae as presently con- '. . . In 1970 Barry I:'fughescameacro.ss. so.rn¢patently .
'stituted:Younginaalmostcertainlycannotberegarded d I' h II ' f h d . sauropsi . materia in t ecqc<::tionso .. t eBernar ' '.'
as more thartone ·p·alaeontologic.species . . ··· . . . .... d' b' . 1

PticeInstitut.c; this.saonpr.o.ve. to e . . pr.oa(;~rtid,and Genus . YOUNGINABroom(1914 · skull, · 1922 ' 1
skeleton} .'. .....•. . .... . . .. ...•. ' .... . .. .. . .... '. .' •. . . theInstirute'smaterialcannow be jsted~s f<:~Uows: : ....

Species YO UN GIN ACAFENSIS Broom 't914 .. ' 471 SkulL Prieea longiceps Broom and Robirison .. '.
. 'd - I d · 1948. .... .' . . ." .
Youngol esromeriO sonan BrooTn.1937 . Locality:Heuning~ans,B~r.ghe~sdoi-p. , .

. - ~oungops~skit~~~ng~:room 19~7 b' Cbllected:j,W.Kitchip.g. · .
1 ;~~gOpSlsrU l get room an Ro lOson -4196 VerysmaUbiok~n<inddi~tor-tedskull III

Y 'd " B ' . d R ' b" .' nodule. '. . '" .' .'. . .'.:
oungot es mmor roOm an .0 Ill.son .. ' '. Local.ity· : T. w .. eefo. ntein_ ' Bethulie: .· . _, 1948 . . ' . .

Small up-per Perrnia~ diapsid.rep·tiles occ. urrin.g in the ' Collected:J.W;Kitching. .. . '. '.'
. 2675 .. A skull andskdeton 'n:iirim sacrum; pelvis, tail -

D aptocephalus zone ofthe Karroo beds of South Africa. . . . '. . .
Skull rather shallow snotHed with terminal exterrial -. . . and hind legs and .including a third s'captll<i . . '
nares. Some roughening ofthe frontalsparticularly ' . Locality : Hairismitl1. .· · . - .
b h b' S· I I l"k d . I Collected :J.·W.Kitching. . . . .

a ove t e or: Its. Impe p. ate~ I e qua ratoJuga no '. 2676 Ail almost complete" articulated postcranial:
taller than lower third of quadrate. PostparietaIs; • skeleton, lacking theheadaiid {irstthree cer-
tabular IS and sUPdrahtemPkOdr~ls PI'lresr,ent,the l~tter wf ihth . . .vicals. '.' ... ' . •. . .,. ' .
ventra y recurve · 00 s Isla y or reception 0 teLocality : H~risrnith.
paroccipital processes. Marginal . teeth conical . and . CoIletted:J :W: Kitching . . '
subthecodont with alternate replacement with ap- .·· .400'5 . Tibia,fib(a,ankleandpart6ffoot.
proximately 20 functional maxillary.teeth. Maxillary ' Locality : OldBrickfieldsD6nga; Barrismith: ..
tooth row extendingwell b~ckbeneaththeorbit. Can - . . .. .
siderable palatal extent of prernaxilla, particularly :' . Collected:J~ ",,:Kitclling. " ..
lateral to the internal choanae. Vomer bearing teeth
on both edges, with two or three anterior median . MATERIALANo' :METHODS • OF .•. •.
"vomerine fangs".Paraspherioid edentulous. In ~ '. '. " .' ' PRESENT STUDY ',' ....• . '. . .'.
fraorbitalfenestrapresent. .' . . . . . ." .. ' .• . B. P .1.2675 wasalmostcomple.telydisartitulated .'

Vertebrae ai11phicoelous and notochordal with with acetkacid ;the~kulnsnow ehtifely ieducedloits
broad flat arches. A middorsal rowofscutes, onereI' . elerne~ts, and dnlfonehandhasbeenleftintact, .
vertebra. Second sacral rib bifurcated distally(firstnot ' . prepared from aile side only: As this tbokan inor- '.
known). Tail with low neural spines and deep haemal . dinately king time and as the material is riote.'nti'rely '.
arches. _Body . dorso-ventrally compressed in ' suited to acid preparatioo;B.P.L2676 wasfirstpre"
crocodilian .. fashion . . Proximal limb segments pard. mechanically' in lateral aspect; acid was used
marginally longer than distal. . . . on the sacnim and tail which werideff articulated, and ·

' only~ the p'elvis .andhind lirnbshavebeencbmpletely .

. PROLACERT A ..

,REVIEW OF LITERATURE AND MATERIAL
. All the PrQlacerta material comes from the Lystrosaurus
zone of South Africa oT LowerTriassicage.(Colbert

cleaned in acid. • B: P: 1..4005 was ent,ireiy,unenable to .
. acid preparation and has been completely cleared of
'. matrix to yield ;:(usefUladdition<ila.nkle.B.P,1. 4 71, '

the. ty'pe of Pricea, would Qe-ruinedJ:?yacid ·but·it has ,
'. been possible tb re_mov~the anterior thirdofthe lower

......• -.;

.. - ~ ..

100

jaws with a saw cu t between upper and lower teeth and
so to prepare mechanically the anterior of the
palate - notably the premaxillae and tips of the
vomers . This material then yields the complete os
teology of Prolacerta.

OSTE O LO GY O F PROLACERTA
The skull

Though the fo llowing description refers to B.P.1.
2675, points of difference with the interpretations of
other workers are noted.

Dermal bones of the skull roof
Premaxilla (Figures 11 and 12 ). In lateral aspect the

premaxilla is somewhat downturned. There appear to
be five tooth positions . The premaxillae extend back
beyond the tips of the maxillae in the palate and unite
in the midline beneath the ti ps of the vomers.

M axilla (Figures 11, 12, 13, 13A and 32 ). Figure 13
clearly shows the structure of the maxilla with its
thickened tooth-bearing margin dropping away inter
nally to an extremely delicate vertical sheet with inter
nal thickening behind the nasal capsule. The maxilla is
excluded from the nasal opening by the premaxilla.
There are 25 ( + 2?) maxillary tooth positions and ac
tive alternate replacement. The thecodont teeth are
laterally compressed, with sharp unserrated edges
and are recurved ; they are deep rooted, extending the
full depth of the thickened a lveolar portion of the

roo psp. ------t';

PTGD

PSP

maxilla which is extremely thin walled .
Marginal dentition (Figures 13, 13A and 32). It is es

sential that the marginal dentition be described in
full detail. Thecodont teeth are known in only one
group of lizards, the Cretaceous mosasaurs (Ed
mund, 1967; Russel, D. A., 1967) : these teeth are
distinguished by having enamel-coated crowns one
third the length of the dentine base; resorption pits
affect the tooth base and the adjoining bone and
arise in the posterior half of the lingual surface of
the tooth .

By contrast the teeth of Prolacerta appear entirely
coated by enamel and are held in deep a lveoli by bone
of attachment; resorption pits are medially situated in
the lingual base of each tooth, and do not affect the
adjoining bone of the mandible. In Figure 32A the
g:oove for the dental lamina can be seen clearly,
Figure 13A supplements the photographs, giving the
pattern of resorption pits and replacement teeth.
Among known dentitions this is typical only of the
Archosauria.

Nasal (Figures 11 and 14). Little can be added to
~amp's description. The nasal is overlapped con
Siderably by the maxilla . A clear suture with the
frontal is seen in Figure 14A.

Lacrimal (Figures 11 and 12). The lacrimal is ex
posed as a narrow strip on the side of the snout.
There are two lacrimal foramina as indicated .

PMX

2cm

:..+,....;-q- --- N

1.'P.fr---- PRF

J'~a---- L

PAL

EPT

POF

Figure II . Prolacerta broomi B.P.I. 2675 . Skull in palatal .and dorsal views .

2cm

Figure 12 . Prolacerta broomi B.P.1. 2675. Skull in lateral and occipital views.

~--- ... , , ,
-\

I ., ,

I
I
I

,
. /
I

\ . \ ,
/ ,

. I'
f
f . ,

./
I

. I
'/

. /

I
I
' 1

E
u

/
I
I ,
0. ,

I , .
I
/.

, I

I '

I , .
" "

)
~~-~"
..

, .
, ....... __ !
I'
I
I ,.
/ ,

" I ' ,

" I
I '

I , .
I·
I
'-

Figure 13 . Pro lacerta broomi B.P.1. 2675. Maxillae in external and internal aspects.

. ,
"

' 1
' . 1
. /
/

. /
. ,

. ,"'1'
. /

. ' .. ,
'. ~'-'J'

";:~

lOl

... '-'

~ . .

. , ........ :-
.. , .. ' .. ..... . ..

1 em - __ II::::!I __ . =. ::JiI .. -=:' ===i -' '. '''<~''~ '~ ~;'-:'',-« ~! .,
- . Figure 13A. <Prolacertabroorrii B. P.I : 267 5 .. Leftinaxillary dentition . Read in COnj~ctionwith ·

. srereopbQtographsof Figure 32 .

PreJfQfltaLjFigure~ 11 .~ 12 aria ·.14J.·. ·Thisis :an exc . . ·.·Poslfront~1(FigUre~11, 12 and 14) . This.li~s againsta
.' tremdy thin walled eiernenfrirnming pan of the or~ .···.· suhstantialfacet fQrmedby frOhtaJand . parietal.

.' . ' .. bitand~arPjng r'o~nd ' ontcithe sI;lout;- .' •.... . .... •... .' . '.' Posteriotly it has adeepcontact wi th ibe postorbital, ".
Frontal (Fig~re 14}: The.t:elati6nships ofthefmn- which latter runs right up ::agairist 'the parietal" ex- .

. ta~ar~clear. frQ[l1' the ' figure~ supraOrqital thickening . chiding the postfrontal from thetemporal fossa. · .
lsmdicated riYC; - ... ' . .. . .' ...... ;" ' ...•.•. " . Postorbital{Figures 12and15G). The contact with the

Parietar . (Figure -14); S:P.L ·2675 is the . only postfrontal is describedabove. The vemraJprong in-
Pro lacerta skullwi'th .' asubstantialpinealopening ' serts ina gnjove.inthejugal (Figures 15Hand I ), while '
midway along the . parietal suture;· in . the 'other -' ' the posterior tip is received into a depression on the
desciibedspecirnensthere is no sign of this opening; -squamosal (Figure 15 i2:.) , . .' . .. ..
this'mustbetakentobeavariablethaiacter. Dorsal- Jugal {Figures 12 and 15). The right jugal is present
ly the _paiietalscutve ' laterally . over the · brai~ . area. intactwhich clears up all doubt regarding the extentof

'. PosteroJateraIry . vertically . flattened wings . run out the· posteriorspur.Theportion rimming the orbit is
and back, and these are connected in the occipital . substantial butthe spur and the area above it behind

.. pio:nebY:I: thin bridgn,vhichslides' ciyerthe :s\.Ipraoc- . thepos torbi tal are extremely thin. . .'. .
cipitaLThesupratemporalliesalongthe top 'of the . Quadratojugal (Figures 12 and 15). The quadrato
pari.etalwing .,mdis heIdproximaUyin a small . jugal is a small bemand flattened rod articulating .
notch : " '.' . - . between facets on the quadrate (Figure lSA) aild the

· squam6~al (Figures ·15£ and F), with af6ramen--
" betweenitand thequadr<.lte. . . .' ..•.. '. .'. .
. ,Squamo.s2zl(Figures 11,12 arid 15),Facetsforpostor- .

bital and quadratojugalhave·alreadybeendescribed:
Posteriorly thereisa mesiaUydirectedhemispherical .

. . facet in .· which the head .ofthequadratearticulates . .
· Above thisis a long concave facet which receives the
paroccipital process; (this i~alm6st ventrallydirecte.d .

. " . and It seems likely that squamosal ; quadrate, and '
.. par-occipita:! process meetat this pOint-thisexplana

tionlends perspective to Figure 15F), Thesquamosal is
held againsttheparietalaridsupratemporal.by a ,fifth
facet · which is vertical and follows the curve of the

' .. supratemporal. '. '. '. '. .

. - .:

. .-..
: .. : lcm ··

_ . . . . .... 0 ·.. .

Fi gure 14 . Pfoldceri'abrl)omi B: P .1.2615 ~ Nasal s, frontalS ,parietals .
.and supratemporaIS.A; Dorsal (inCludesprefronra1 );

'· 8 , Left latetal; C,Vemral ; D,OtcipitaL . . .

Supratemporal (Figures> :n, . 12 and - 14 ). The
· supratemporal is a thin, cUrved Fad notched into the
· top of the parietal wing and extending beyond it: it

is exposed indorsaland 'occipitaLview but entirely
· obscured laterally by the squamosal. '.' .

. .' Sclera1platei Scleroti'cs are absent in this specimen.
Camp gives acountonl~12 forhis; . . . .. ' ' .
. Thepalate{Figures 11 ; 16and 17 ) '. '. '.' . '.' .'

Vomer. The left varner is (omplete (Figure 16A) but
wasi)'ing out ofpositiori.It bearsthreedistinctrows of ·
teeth .It is 'dear that the vomers would be . united
anteriorly for about half their length, -a~shown by

. Camp. Posteriorly the vomer 'slightlyoverlaps the

. pterygoid and palatine; alsofiHing the' gap between
the tips of these: eh;ments (Camp illustrates a gap here

· butthisisimprcibablel. - '. _

.. : . .

S.P . _ H

(El

Icm - --

(Hl

~ :: ....... .
'. ..

(()

Fi'gure IS. Prolacerta broomi B.P.I. 2675. A-D, Quadrate; A, Lateral; B, Mesial; C, Ventral ; D,
Dorsal; E, Squamosal, quadrate and quadratojugal; F, Squamosal in internal aspect and
quadrato jugal ; G, Posto rbital; Hand J, Jugal ; J , K and L, Epipterygoid in latera l, mes ial
and ventral views.

,'\ ,

, I I,
" V

E Icm

Figure 16. Pro lacerta broomi B.P.I. 2675. A, Vomer; B, Palatine; C, Palatine in lateral view; D,
Dorsal view of palatine foot; E, Palatal extent of premaxilla.

103

(Fl

104

t\
I I
I •
, I

Figure 17 . Proiacerta broomi B. P. I. 2675. Pterygoid and Ectopterygoid: palatal, dorsal, mesial and
lateral views (last slightly rotated ).

Palatine. The mesial portion of the palatine is an ex
tremely thin sheet; laterally the bone thickens and
turns downwards, this thickened edge bearing a single
row of teeth continuous with those of pterygoid and
vomer. The foot of the palatine is applied against the
alveolar border of the maxilla and has a small dorsal
process which rests on top of the alveolar ledge (Figure
16B and D). Above this latter process is the palatine
groove which transmits the maxillary artery and
superior alveolar nerve (Figure 16C)which then enter
a foramen in the maxilla (Figure 13), branches also
possibly running forward in a groove behind the
alveolar ledge to enter the more anterior foramen.

Ectopterygoid. Only part of the left ectopterygoid is
present, the process abutting against the jugal having
been lost. However, this element is present in Camp's
specimen and well illustrated by him. It is also present
in Crompton's specimen, whose drawing agrees with
that of Camp. This element overlaps the pterygoid
from above (not from below as shown by Camp ); it
curves gently forward and outwards, continuing the
edge of the pterygoid flange, and terminates in a sub
stantial footincontactwith the jugal.

pterygoid. The palatal portion of the pterygoid is ex
tremely thin, except where it turns up into a ridge
bordering the interpterygoid vacuity. On the mesial
border there is a continuous row of teeth, while from
just anterior to the basal articulation a double row
runs towards those of the palatine; the flange is also
toothed. A notch on the lateral edge of the pterygoid
receives the palatine. The articulation for the
basi pterygoid process is deep, hemicylindrical, and

directed obliquely posteriad. The quadrate ramus is
deeply concave on its mesial surface with a pit situated
anteriorly. The ornate margins of this process (Figure ·
17) are genuine. The area of origin of the pterygoman
dibularis is marked by a ridge on the inwardly sloping
lower portion of the quadrate ramus.

Epipterygoid. The lower half of each epipterygoid is
preserved. It is very similar to that of Youngina, in
cluding a facet capping the basal articulation. Camp's
description and figure of a thin quadrate process is
clearly an artifact.

Quadrate (Figures 11, 12 and 15). The articular sur
face of the quadrate comprises inner and outer con
dyles with a facet above the outer condyle for attach
ment of the quadratojugal (Figure 15A). The
pterygoid ramus is extensive, with a depressed area of
articulation with the pterygoid (Figure 15B). Figure
15 D indicates the areas of cartilaginous attachment to .
squamosal and paroccipital process.

The braincase (Figures 12 , 18, 19,34 and 35).
Supraoccipital. The supraoccipital (Figure 18A-D) is

broad, as in Youngina. There is no suggestion of contact
with the parietals; post-temporal fossae were present.

Opisthotic. The opisthotic is firmly united to
supraoccipital and prootic. Its ventral ramus borders
the foramen ovale anteriorly and the jugal foramen
mesially: Above the jugal foramen is a facet for the
reception of the exoccipital. The ventral ramus has a
loose mesial contact with a ventrally directed process
of the basioccipital. The paroccipital process is flatten
ed, as in Varanus, but meets the squamosal/quadrate

B ;;;;::::,.-5

l c m

~ .. c\ . '. . . . .

. F ··· ... G

• . ,j, .. . • < 8 . '!;,
Figure 18 . Prolacerta broomi B.P.1. 267 5 . A - D, Supraoccipital ,

prootic and opisthotic ; A, Dorsal ; B, Ventral; C,
Anterior ; D, Posterior ; E-G, Basi- and exoccipitals;
E, Anterior; F, Posterior; G, Lateral; H, Dorsal (ex
occipital ring removed ).

joint more or less horizontally, and not vertically as
does that of Va ran us.

Exoccipital and basioccipital (Figure 18E-H). These
elements are fused and they completely ring the
foramen magnum. The braincase is difficult to assem
ble as so many joints involved cartilage, and this is par
ticularly true of the basioccipital/parasphenoid
relationship (Figure 11 ). These bones separated freely,
though it seems the medial prong of the parasphenoid
was applied beneath the basioccipital. The lateral
prongs of the parasphenoid, however, span the same
width as the ventral tips of the opisthotics and it seems
very likely that parasphenoid, basioccipital , and
opisthotics were held together in cartilage at this
point, the basioccipital contributing reasonably dis
crete basal tubera.

105

lcm ---

Figure 19. Prolacerta broomi B.P.1. 2675 . A, Lateral view of brain
case ; B, Dorsal view of para-basispheno id; C, Ven
tral view of para-basisphenoid; D, Anterior view of
para -basisphenoid .

Before leaving the occiput it should be noted that
Camp's ( 1945) reconstruction of this area is complete
lywrong. Robinson 's (I 966 )is better, butthe tips of the
paroccipital processes are out of position, the tops of
the exoccipitals are the wrong shape and do not meet,
the top of the quadratojugal is shown touching the
outer surface of the squamosal instead of the inner,
and the quadrate squamosal contact is too low.

Prootic( Figures 18 and 19). The prootic has extensive
sutural contact with supraoccipital and opisthotic.
The pila antotica contacts the dorsum sella of the
basisphenOid and there is an additional ventral con
tact with a delicate posterior ridge of parasphenoid (cf.
description of Proterosuchus braincase to follow ). There
is a deep trigeminal notch and a pronounced crista
prootica which continues on to the opisthotic; within ,
the recess formed by the crista and somewhat
posterior of the trigeminal notch is a small foramen
for the facial (VII ) nerve. This is the earliest known
appearance of the crista prootica s.s. (see Youngina )
which is particularly significant as a point of origin for
the protractor pterygoideus . (Internal canals in sU{Jraoc
cipital, opisthotic and prootic are clearly dlspIay
ed-should comparison prove necessary.) Medial
to the crista prootica is a further depression above the
foramen ovale (involving prootic and opisthotic) for
the anterior semicircular canal (Oelrich, 1956, p . 15 ).

Para/Basisphenoid (Figures 11 and 19 ). This element
is completely edentulous. The parasphenoid rostrum,
V-shaped in cross section, extends a little farther

106

forward than the pterygoids. Posteromedially there is
a light overlapping contact with the basioccipital,
while posterolaterally a gap exists which, when closed
by cartilage, would connect basisphenoid and
opisthotic. Prominent basi pterygoid processes
(orientation Figure 190) correspond with deep
pockets in the pterygoids already described. The vi
dian (parabasal ) canal is a deep groove.

The only foramina are those for the internal
carotids. The course of the abducens (VI) is marked by
a groove on the posterior dorsal surface between alar
process and dorsum sella. The alar process meets the
pila antotica, its lateral edge continuous with the crista
prootica. Behind the dorsum sella the bony sheet
(parasphenoid ) is W-shaped in cross section, with tall
lateral walls.

The posterolateral extremities of the parasphertoid
do not contact the ventral processes of the opisthotics
in the region where the sternohyoideus muscles would
attach. This area would have been filled in by cartilage .
A distinctive ridge on the posterolateral wall of the
parasphenoid terminates in a knob which meets the
prootic below and in front of the lagena recess.

(A)

Between this point of contact and the alar process
there is a sharp smooth edge of parasphenoid; bycon
trast the corresponding edge of the prootic is broad
and channelled, indicating the presence of a connec
ting membranous sheet.

Stapes (Figures 12 and 18). The stapes is a thin rod,
the stapedial foramen having been lost.

Ceratobranchial. A long thin ceratobranchial I is
present - this has been described by Camp.
The lowerjaw (Figure 20).

The elements of the lower jaw, like those of the skull,
are rather delicate. The Meckelian canal runs to the tip
of the dentary. There is a large Meckelian fossa on the
lingual surface. The glenoid, with two distinct facets,
was probably reinforced anteriorly by a wad of car
tilage. The large broad retroarticular process bears
muscle attachment scars (pterygomandibularis). Twenty
seven tooth positions are indicated, the teeth being
somewhat smaller than the maxillary teeth. The
symphysis is extremely light and was probably
movable. Therelationshipsofthe bones are clear from
the figures, though some notes are in order. The sur
angular contributes the lateral glenoid facet, and

2em

A~.~· .• · •••. ~·Zl$·~"J\bhl)~
SPL

PR ART A

(8)
Iem - --

Figure 20. Prolacerta broomi B. P.1. 2675. A, Reconstructions ot
the lower jaw; B, Do rsal view of posterior half of left
ra ITIUS .

sends a projection mesially to contact the prearticular.
The angular, which forms the postero-ventral

border ofthe jaw, terminates beneath the glenoid. The
tip of the coronoid is barely raised above the dorsal sur
faceofthejaw.

The articular contributes the mesial glenoid facet; its
recurved tip forms the site of attachment for the
depre550r mandibulari5. There is good evidence for a
tympanic crest on the lateral border. The chorda tym
pani/posterior condylar artery foramen is present.
There is an incipient angular process.

The p05tcranial5keleton
The nearly complete articulated skeleton, B.P.1.

2676, lacks the first three cervicals and of the forelimbs
only the proximal ends of the humeri are present.
B. P. I. 2675 which contains the forelimbs and skull is
truncated anterior to the pelvis. The overlap between
these two specimens then is only in the vertebral
column and the scapulocoracoids.

It must be stated at the outset that scales in the
drawings are true for the smaller specimen C~.P.1.
2675) while the larger specimen has been reduced by
1I6th to bring it to the same size: this figure was arrived
at by measuring the scapulocoracoids and this reduc
tion produced a very good fit of the vertebrae. While
this appears to be satisfactory it is unfortunate that an
approximation was necessary, though the critical
proportions of the skeleton are of a far greater order of
magnitude than any error of scale which may have
resulted.

B. P. I. 4005 yielded additional information on the
hindlimb, and here there is no problem of equating
size with undistorted tibiae for comparison.

Camp has described the first six cervical vertebrae
and suggested they might indicate aquatic habits-a
significant remark in view of the close affinity of
bipedal lizards for water (Neill, 1971). Prolacerta was
clearly a bipedal runner with a large tail to counter
balance the weight of the body. That the head and neck
were highly manoeuvrable is shown by thickening of
the tips of the neural spines in the shoulder region.
Like the skull, the skeleton is extremely light, with
hollow limb bones, and centra thin walled but sup
ported by a delicate web of internal bony struts (there
is no sign of pneumatopores). The ankle and hooked
fifi:h metatarsal are similar to those of rhyncho
cephalians and Protero5uchu5, yet not much different
from those of Euparkeria. Intercentra are one primitive
character in an otherwise very advanced, typically
Triassic, skeleton.

-The whole animal is so I closely comparable to
MacrocnemUJ (Peyer, 1937; Kuhn Schyder, 1962) as to
assure very close relationship. It seems ridiculous to
think of these distinctive thecodonts as lizards: they
are yet another group which has dispensed with the
lower temporal arcade. Glevo5auru5 hud50ni (Robinson,
1973) is an undoubted sphenodontid which has an in
::omplete lower temporal arcade. Clearly this feature
evolved independently in squamates, thecodonts and
rhynchocephalia.

107

Axial5keleton
There are 26 presacral vertebrae, including

proatlas, deeply amphicoelous but not notochordal.
There are two sacrals and a string of 13 proximal
caudals plus several isolated more terminal caudals.
Intercentra are present only as far back as between the
first two caudals, after which they are succeeded by
very long, flat haemal arches. The vertebrae are very
lightly constructed: there is no sign of separation
between centrum and neural arch; the centra are
hollow with an intricate system of internal supporting
struts.

Atla5 and axi5 (Figure 21). The atlas complex is
primitive. The proatlas connects the atlas arch to the
exoccipital above the foramen magnum. The paired
atlas arch elements just meet in the midline anteriorly
and rest on a large pleurocentrum. The atlas intercen
trum articulates between the occipital condyle and the
axis intercentrum. This gives the appearance of being
a highly mobile joint. The axis has a long neural spine
typical of this bone, and it has a facet for the first
cervical rib.

C ervical5 (Figure 21 , (7)). The following six vertebrae
may be termed cervicals; they are greatly elongated.
The last three cervicals and first three dorsals have
thickened neural spines -clearly points of attachment
for important neck muscles. In the neck region
capitular facets are all on the centra and do not involve
the intercentra. Rather long, slightly inclined
zygapophyses would permit great freedom of move
ment, both lateral and dorsoventral.

Anterior dorsal5. A lateral ridge leading towards the
tubercular facet becomes increasingly pronounced
posteriad until the first dorsal, which is characterized
by transverse processes with their leading edges almost
at right angles to the column (Figure 21, (l0)). The
length of the neural spines decreases to a minimum on
1 0 before picking up again over the back. This is also
the pointatwhich the rib facets start to merge, forming
a continuous articular region. This process is com
plete by 15.

P05terior dorsal5. Here there is a slight but distinct
alternation in the length of the neural spines. By 19
all trace of two rib facets has gone leaving a single
circular facet (Figure 22, (21)).

Sacral5 (Figure 22). The narrow-necked transverse
processes (or pleurapophyses) widen into substantial
facets which touch and interlock slightly as shown. The
transverse processes of the second sacral bifurcate into
two functionally distinct processes; the anterior
process is thick and broad and abuts against the ilium.
The posterior process is thin and narrow and does not
touch the ilium; it in fact constitutes an additional
caudal transverse process and would no doubt
strengthen the attachment of tail muscles, helping to
support and control the large powerful tail which is
such an important organ to a bipedal lizard-like
animal. (This seems to be a rather different arrange
ment from that described for Protero5uchu5 by
Cruickshank, though I am not convinced that the

108

1-3

2cm

Figure 21. Prolacerta broomi B. P.1. 2675. Vertebrae as numbered.

posterior process contacts the ilium as he has shown. )
The beginnings of a Prolacerta-like condition can be
seen in Y oungina, and several modern lizards trend the
same way (Hoffsetter and Gasc, 1969, p.265 ).

Caudals (Figures 22 and 27). The string of 13 prox
imal caudals all have strong transverse processes
diminishing gradually in size in the same way as do the
neural spines and haemal arches. The chevron bones
or haemal arches are laterally compressed and
anteroposteriorly broadened: they are extremely long
(1 t x the depth of the preceding vertebra ). The enor
mously powerful tail indicated by the osteology
almost certainly relates purely to the animal's bipedal
mode oflocomotion and active feeding movements of
the head and neck, though the possible importance of
lateral threatdisplaycannot be overlooked.

Ribs (Figure 22 ). The cervical ribs have extremely
slender shafts tapering to fine points; they have two
distinct articular facets and an anterior dorsal process.
Ribs 10 and 11 (in the pectoral region) are rather stout
proximally; here the two rib heads have merged into a
single articular facet. For the rest the ribs have slender
shafts circular in section . From 19 to 26 the ribs are

rather weak with expanded circular attachment facets,
shafts very slender proximally, expanding a little dis
tally.

Pectoral girdle andJorelimb (Figures 23 and 30 ). The
scapulocoracoid is extremely delicate, particularly
that portion of the scapula anterior to the shaded line
in the figure, which is paper thin. The posterior third
of the suture between the two elements can be seen on
the mesial surface. A coracoid foramen is present im
mediately in frontof the screw-shaped glenoid.

Clavicles appear to elaborate with age (F is larger
than G and therefore probably older, or possibly a
male ); unfortunately most of the medial shaft is miss
ing. The vertical shaft is grooved on the inner surface
where it would lie against the scapula. A prominent
anteriorly directed boss appears to form with age in
the angle between the two shafts.

The interclavicle is somewhat damaged though its
essential structure is clear-it has a narrow medial
shaft and a broad crosspiece notched anteriorly and
with depressed facets on the ventral surface which
receive the clavicles.

J09

2cm

Figure 22. Pro lacerta broomi B.P.I. 2675. 21st Vertebra; Sacrum and two proximal caudals in
lateral and ventral aspect; Ribs as numbered.

The humerus is primitive and simple, though the
entepicondylar foramen has been lost; there is a deep
ectepicondylar groove.

Radius and ulna are long and slender. The small
wrist elements preserved are as shown, but as they are
somewhat disarticulated nothing definite can be said
regarding the missing elements. The phalangeal count
for the hand is the standard reptilian one.

Pelvic girdle and hindlimb (Figures 24, 31 and 33 l. All
the elements of the pelvic girdle have smoothly round
ed margins. All three contribute to the acetabulum
which has no upper rim as such (though it is rimmed
anteriorly by a ridge angled towards the posterior tip
of the iliac bladel-this could be to allow the femur to
be raised rather high during bipedal locomotion. The
pubo-ischiadic plate is broad; the central area is ex
tremely thin. There is a thickened out-turned antero
ventral process of the pubis, with the obturator
foramen immediately behind it. Girdle measurements
relating to body sections are given in Table 1. p. 98.

The femur has marked posteriad curvature at the
distal end. The head is concave and would thus have
been capped by an epiphysis which would effectively
increase the length of the femur. There is an internal

trochanter, with intertrochanteric fossa, very much
restricted to the head. The shaft is devoid of muscle at
tachment protruberances and scars. The tibia is large
and robust, particularly at the proximal articulation;
the distal end is concave and poorly ossified,
suggesting the involvement of cartilage at this point.
The fibula is rather slender and is laterally com
pressed.

The ankle comprises a proximal assemblage of
astragalus and calcaneum in loose articulation with a
large foramen between them, and a centrale in firm
contact with the mesial surface of the astragalus. There
are four distal elements of which the first three are
small and featureless while the fourth is large with a
lateral facet for the fifth metatarsal and a terminal facet
for the fourth; it lies in the junction between astragalus
and calcaneum.

There is a standard phalangeal count of 2:3:4:
. 5 : 4. The fifth metatarsal is hooked with a ven
trolateral scar indicating the insertion of the gas
trocnemius.

The astragalus has a large proximal facet which
receives the tibia; this is followed by a smooth narrow
neck before a second facet at right angles to the first.

110

A B c

F
. ':.;-"',:, ' : '

.• • 1
.J

Figure 230 Prolacerta broomio (A - E and G, BoPoi. 2675; F, BoPoi. 2676 0) A, Scapulocoracoid; B,
Humerus mesial view; C, Right forelimb ; D, Humerus in lateral view; E, Interclavi
cle ventral view; F, Dorsal limb of clavicle, mesial and lateral views: G, Dorsal limb
of clavicle, mesial and lateral views o

F ~

2cm

Figure 240 Prolacerta broomi BoPoi. 2676 0 A, Right pelvic girdle lateral aspect; B, Left pelvic girdle
internal aspect; C - F, Left femur ; C, lateral view; D, Posterior view; E, Mesial view;
F, Proximal end; G, Right hind limb flat; H, Ankle flexed o

The calcaneum below this is of equivalent thickness,
thinninga little from here towards its lateral edge. One
can thus envisage a mass of cartilage here for reception
of the fibula. The astragalus bears a marked depres
sion in its dorsal surface as indicated.

Figure 24G shows the hind-foot in one plane while
in H a natural positioning has been attempted with the
ankle flexed at the meso tarsal joint.

Some of the more important length measurements
are as follows in mm:

Skull (occipital condyle to snout) 67
Neck (condyle through ninth vertebra) 100
Trunk (glenoid to acetabulum) 130
Femur 54
Tibia 58
Humerus 42
Radius . 37

PROTEROSUCHUS BRAINCASE
The braincase in B.P.1. 3993 (Figures 25, 26, 36 and

37) has suffered from the vertical compression
which has affected the whole skull; no corrections
have been made in the drawings except in Figure 26B
which has been restored to symmetry.

al. pr pro.

p.a.

al. pro bsp.

bpt. pro ------'

f.m.--~

BO----'i'~

bt.-----'I

p. pr. pro.

I-----BO

CAl

(8)

Figure 25 . Proterosuchus vanhoepeni B.P,1. 3993, (Braincase); A,
Lateral view; B, Somewhat oblique occipital view.

The prootic has an open trigeminal notch between
well developed alar process and pila antotica: it has a
long posterior process running along the paroccipital
process of the opisthotic. A distinct ventral process
meets the alar process of the basisphenoid; behind this
a second process from the basisphenoid contacts the
prootic . Anteriorly the prootics are united
in the midline by a substantial sheet of bone, stretch
ing from one pita antotica to the other, which roofs

111

the pituitary fossa. This sheet is continuous with the
dorsum sella; the exit for the VIth cranial nerve can be
clearly seen (Figure 26A) between them. This
condition also occurs in Euparkeria. The
basisphenoid has several notable features. In lateral view

Vl----,~

bpt.pr.

OP---ffl~

so
al. pro pro.

s:::::::2+-- OP

7=-=------ P a.

---al. pro bsp.

~--bpt.pr.

(A)

V. C.

~*---b.t.

Figure 26 . Proterosuchus vanhoepeni B.P.1. 3993, (Braincase); A,
Anterior view; B, Ventral view.

(Figure 25A) can be seen the prominent process com
ing in behind the ventral process of the prootic; and
thin bone behind it is folded inwards, as indicated also
in Figure 25B. This arrangement floors the foramen
ovale. The large postero-Iateral processes of the
basisphenoid bearing the grooves of muscle attach
ment scars run out beyond and below the ventral
processes of the opisthotics -this is indicated in
Figure 25B, a somewhat oblique view. In ventral
aspect (Figure 26B) the basisphenoid runs anterior to
and closely applied against the prominent basal
tubera otthe basioccipital, while just anterior to this is
a deep depression. The open videan canals lie mesial to
the posterior edge of the basipterygoid processes,
while immediately in front of them are a pair of
foramina (present also in Prolacerta) which possibly
transmitted a branch of the palatine artery.

The ear region poses a problem. The foramen ovale
is very small. A stapes if present would have been
restricted to a channel between opisthotic and prootic
and would thus have run obliquely backwards to the
head of the quadrate, indicating a tympanum in the
position postulated by Ewer (1965) for Euparkeria.
There is in any case no room for the attachment of a
tympanum on the retro-articular process.

112

Palatine
Cruickshank ( 1972) shows the palatine forming the

entire lateral margin of the suborbital fenestra. In fact
itonlyrunshalfwayback. (B.P.1. 3993 andP.fergusi.)

Streptostyly
Cruickshank suggested that the quadrate of

Proterosuchus might be movable. This is clearly not
possible in an animal with a solid lower temporal bar.
Variations in the position of the quadrate are ap
parently due solely to post-mortem distortion just as
in Youngina .

PROLACTERTA FUNCTION
(a) Cranial kinesis. It is well known that most if not all

predatory squamates have kinetic skills. These are
generally speaking small animals which swallow their
prey whole. Prolacerta, though not a squamate, is func
tionally similar. A small, lightly built predator, it is the
epitome of the sort of animal to which cranial kinesis is
of advantage. In Prolacerta the skull can move in a ver
tical plane relative to the braincase, pivoting on the
paroccipital processes, the palate being guided by the
basal articulations of the basisphenoid. The quadrate
is free to move anteroposteriorly as a result of the loss
of the lower temporal bar; this is the only accep
table explanation for the retention of the quadrato
jugal as a delicate strut articulating between
squamosal and quadrate. (One can readily accept that
slight refinement to the quadrate controlling
mechanism in a more advanced form would render
the quadratojugal superfluous, but there is no reason
to believe that this type of quadratojugal mechanism
formed part of the lizard evolutionary story.)

Certain less obvious movements are also possible.
These involve the jugal and postorbital, and may be
connected with either streptostyly (fore and aft
movements or quadrate spreading) or muzzle
movements, or both.

Some lateral movement of the postorbital along
its vertical contact with the postfrontal would have
been possible though slight. The postorbital!
squamosal contact is loose and would allow free
movement in anteroposterior and dorsoventral
planes. The tongue and groove contact between
postorbital and jugal would allow sliding movement,
and itseems logical to assume a fair degree of the same
sort of sliding at the long curving jugal-maxillary junc
tion. The great reduction of jugal and postorbital in re
cent lizards and the loss of contact between them, as for
example in Varanus, would suggest thatjugallpostorbital
sliding is necessary in the early stages of the development
of mesokinesis. Movements at the upper extremities of the
postorbital in Prolacerta might be related more to
quadrate movements. Although there is no mesokinetic
joint as such, yet the bones of the snout are extremely thin
and the fenestrae exochoanis very long, so that muzzle
flexure by actual bending of particularly the nasals almost
certainly occurred. Figure 27 . Prolacerta broomi, Reconstruction of the skeleton.

It is interesting to note the movable joints present in
lizard skulls (Frazzetta, 1962) which are not present in
Prolacerta. In all lizards which have a mesokinetic
(fronto-parietal) joint, the skull roofis broadest at the
straight transverse fronto-parietal suture. In
millerosaurs, younginids and Prolacerta this marked
broadening is lacking and the fronto- parietal suture is
W -shaped, effectively preventing flexure at this point.
In this context the keuhneosaurs are intermediate in
morphology, though apparently not yet mesokinetic
(Robinson, 1966). Associated with mesokinesis in
lizards is antero-posterior movement of the
pterygoids. This is reflected in a sliding basal articula
tion and a rod-like epipterygoid which pivots at both
ends . In Prolacerta there is no suggestion that the basal
articulation is moving in this direction and the .
epipterygoid has an antero-posteriorly elongated
footplate which could not conceivably have pivoted.
Relative movement between pterygoid and quadrate
certainly occurred.

Slight spreading of the pterygoids and hence of the
quadrates and lower jaw rami is a possibility, but until
both the movements and their functional importance
can be convincingly demonstrated in living lizards,
there is no purpose served by discussion at this stage.
Frazzetta (op. cit. ) has shown that the lizard quadrate
moves forward as the jaws open; it would thus be
reasonable to assume that the same is true of Prolacerta.
Now while kinesis in Prolacerta is not as complexas that
described by Frazzetta for lizards, there appears to be a
very simple explanation for quadrate protrusion.
Upward movement of the skull on the braincase as the
jaws open is slight, and any upward movement of the
whole head will not alter the fact that the lower jaws are
dropped through a considerable arc relative to the up
per.

This means that without protrusion of the quadrate
the lower teeth move posteriad relative to the upper,
and consequently move forward during the bite, as
shown in the following diagram.

ABC represents the skull. The snout moves upwards
hinging at A. This movement is slight and for present
purposes can be ignored.

113

BDrepresents the lowerjaw, hingingat B.
CF represents the extent of posteriad movement of

lower teeth relative to upper in an akinetic skull .
Protraction of the quadrate during jaw opening

results in the tip of the jaw moving to position E. As
the jaws close and the quadrate is retracted E moves
through an arc to C, as would D in an akinetic
system.

Clearly the teeth and the bite force move upward
through an arc. Representing the bite force as a
chord on this arc it is clear that in the akinetic system
this force CD has a forward as well as an upward
component. In the kinetic system CE moves
posteriad throughout the closing cycle. This cir
cumstance is governed by two variables, the angle of
the jaw opening e and the extent of protrusion DE.

e as used above is greater than any angle of jaw
depression illustrated by Frazzetta, but it is still
necessary to determine DE for live lizards to decide
whether natural movements fall within the limits set
by these variables. It certainly seems plausible to
suggest that there could be an advantage in having
the teeth and the bite force directed backwards dur
ing jaw closing for an animal capturing active prey.
It is surely more than coincidental too that ECD is
the shape of a generalised carnivorous thecodont
tooth .

Compare the akinetic system (A) with the kinetic system
(8) p. 114. In both the muscles contract from length ACto
length AE. In both a decreases to ~ . But, whereas in A the
direction of pull shifts from AC to AE, in B AE and AC lie
on the same line, with C moving to E as B moves to Bl.
Note that in B quadrate retraction B B' = adductor
shortening CE = degree of jaw protrusion D D'. As any
biological system can be expected to represent a com
promise between conflicting ideals, B B' is probably too
short to allow perfection of system B though the system yet
holds considerable advantage by virtue of the extra
linkage and more nearly constant line of action of the ad
ductors.

Streptostyly seems to hold an irresistible fascination
which has led to some curious statements in the
literature. Some of these merit brief consideration.
Walker (1961 ) suggested that the quadrate of
Stagonolepis (a pseudosuchian) might have been
movable and have played a part in that animal's
shovel-snouted foraging mechanism. Ewer (1965)
suggested that some movement of the quadrate may
have been possible in Euparkeria, a thecodont, and
suggested the presence of an additional wholly un
necessary and improbable protractor muscle at
taching to the mesial surface of the quadrate ramus of
the pterygoid, i.e. that portion which lines the throat.
Cruickshank (1972 ) has suggested that the
Proterosuchus quadrate may_~ave been movable-this
to account for the variaoility in the slope of the
quadrate exhibited by several skulls. The fact remains
that Proterosuchus has a solid lower temporal bar and
the quadrate always bears the same relationship to it ; it

114

therefore seems more reasonable to attribute this
variability to distortion of the specimens. It would be
unexpected to find a streptostylic quadrate in any
diapsid , as meaningful movement is only possible
once the lower temporal arcade has been breached as
in Prolacerta.

There is merit in Robinson's (1966 ) suggestion that
the quadrate (of lizards ) might rock back and forth
during chewing movements thus aiding swallowing.
Whether this is true of any lizards will require ex
perimental determination. In particular it is necessary
to determine what sequence of muscle actions is
responsible for quadrate retraction and to what extent
these movements are independent of jaw adduction.

(b) The post cranial skeleton. It is no exaggeration to say
that the combination of characters which make
up. the skeleton of Pro lacerta (and Macrocnemus) is
unIque .

This is a small, very light-boned animal, un
questionably built for speed and/or agility. Limb dis
parity per se is a vague and misleading term. What is
important as a clue as to whether an animal was
terrestrial quadrupedal, aquatic or bipedal, is the

relative lengths of femur and tibia: in a terrestrial
quadruped these are subequal in length, in aquatic
animals the tibia is markedly shorter and in bipedal
animals the tibia is the longer bone. Hence we can say
that Prolacerta, on the basis of limb proportion, was
bipedal (at least at speed). This, however, is the only
immediately apparent indicator ofbipedality; there is
no indication, particularly in the pelvic girdle, of sup
porting modifications for this mode oflocomotion.

Looking for comparisons then we can look first at
the lizards. Several lizards are known to adopt a
bipedal stance and/or gait at times. This subject has
been well covered in a delightful chapter by Neill
(1971 ), whose main thesis it is that bipedality arose in
arboreal lizards as a predator confrontation and es
cape mechanism and is only employed by them as
such. The bipedal lizards are characterized by long
slender tails which are important in counterbalancing
the weight of the head and trunk during this rather un
gainly gait (described by Snyder, 1949, 1954, 1962).
One aspect of this crude bipedal gait which has
seemingly escaped notice is the rotation which occurs

at the foot in contact with the ground as the opposite
limb is swung forward in stiff-legged fashion: this can
easily be seen in the scratch marks of the toe im
pressions left by a basilisc chased across a damp clay
surface. This is worth recording for the implications it
may have for ankle structure, though Prolacerta itself
has a typical early thecodont ankle and itis remarkable
that very little osteological modification, limb dispari
tyexcepted, is seen in the early stages ofbipedalism.

A disconcerting aspect of bipedal lizard
morphology from the point of view of this discussion
is that they all have short necks and trunks; however,
we can immediately remark that perhaps the deep
rooted tail of Prolacerta is designed in part to
counteract the added weight of the long neck. There
are at least two other factors to consider here.

Leaving the lizards for the moment it is pertinent at
this stage to introduce another tentative comparison.
Both the long neck and the large flattened chevrons of
Pro lacerta invite comparison with the incipiently
bipedal prosauropod dinosaurs.

We are now at the stage where we can embark on
detailed argument on the function of the Prolacerta
skeleton. To start with the tail. The bipedallizards have
long slender tails with rather insignificant chevrons;
one aspect of the function of this tail has been
overlooked. While it is clear from Snyder's (op. cit.)
drawings and photographs that the distal part of the
tail is flung out to counterbalance the body, itisequal
ly clear that the base of the tail is held in line with the
pelvis and both it and the trunk are thrown over to the
same side as the limb executing the powerstroke. This
probably ensures that the direction of pull of the
femoral retractors places the arc of travel of the femur
in the required plane, and it also helps distribute the
necessary weight over the sacral region (weight against
which the limb is pushing). As far as the use of the tail as
a counterbalance to the weigh t of the body is concern
ed this is something which requires further explana
tion as we are dealing with a state of dynamic
equilibrium in which weight distribution clearly
changes considerably through the stride sequence.

The above remarks serve to highlight the impor
tance of the second sacral rib of Pro lacerta doubling as a
caudal transverse process tending to lock the base of
the tail to the sacral region; this also suggests a reason
for the impression of relative rigidity conveyed by the
Pro lacerta chevrons (one may note by way of contrast
that in the varanids which use their tails as lashes the
chevrons are centrally situated on the centra, thus not
impeding flexure in anyway) .

The caudal chevrons of Prolacerta, in proportion,
must beamongthe largest on record. Atthe base of the
tail they are 1,5 times as long as the depth of the
preceding vertebra and this ratio gradually decreases
to unity. The only acceptable explanation for this is
that of concentrating the bulk of the muscle mass ven
trally at the proximal end of the tail. There is almost
certainly more than a simple weight factor involved
here; as the pelvis shows no evidence of improvement

115

of the locomotor musculature for bipedal locomotion
it is quite likely that the main femoral retractor, the
caudiJemoralis longus, would be considerably enlarged
in an animal of this size (the size limit of bipedal
lizards). Femur morphology shows that the
caudifemoralis inserted near its proximal end, an
arrangement which sacrifices power for speed.

The shape of caudal chevrons in reptiles is rather
variable; as this is an aspect largely ignored one can
find little help in the literature, but comparison with
the incipiently bipedal dinosaurs is striking, in that the
chevrons are rather large and · lateraffy compressed-:
while these prosauropods have long necks, there is not
much skull weight to counterbalance, and in any case
one cannot compare locomotion in early thecodonts
and dinosaurs too closely. It is instructive to note,
however, that inSaltoposuchus, a bipedal dinosaur with
large head and carnivorous dentition, the cervical
vertebrae are extremely short.

A deep tail must inevitably raise the possibility of its
being an adaptation to swimming. Fortunately that
possibility is easy to discount in this case. In the
crocodile which is adapted to an aquatic existence the
chevrons are of more moderate size, narrow, and cir
cular in section; the tail is dorso-ventrally symmetrical
with tall neural spines extending almost to the tip. In
Pro lacerta the neural spines become insignificant in the
distal half of the tail. In the extinct marine reptiles
caudal chevrons become very small and insignificant.
A piece of circumstantial evidence against an aquatic
existence is worth mention. Macrocnemus occurs in
association with a highly adapted aquatic fauna which
tends to underpin the non-aquatic nature of
M acrocnemus.

The long neck of Pro lacerta may account in part for
the extra weight at therootofthetail. The long neck isa
typically thecodont specialisation common to
Proterosuchians, certain dinosaurs and birds. This is a
specialisation which considerably enhances the
predatory capability of the animal. The thickened
neural spines at the base of the neck and in the pectoral
region are sites of attachment for powerful, quick
acting muscles concerned with raising the head and
neck and wi th lateral movements .

With regard to the limbs and girdles the latter are
extraordinarily primitive for such an advanced
animal, conveying no hint of the animal's obviously
rapid, often bipedal, gait. Though the hind limb is
long with tibia longer than femur, yet it shares a
primitive ankle pattern with Proterosuchus and
rhynchocephalians . The hammate process of the fifth
metatarsal shows no sign of turning inwards un
derneath as is the case in bipedal lizards (Snyder,
1954). The forelimb is not reduced and probably func
tioned during slow locomotion, basking (Figure 27 )
and tree climbing.

116

ECOLOGY

Possible environment and habits
The Lystrosaurus zone presents many problems of in

terpretation of ecology of its fauna. A regional
sedimentological study is an essential prerequisite as a
framework in which to place the fauna, but this is as yet
lacking. This is a zone notable for its omissions ;
there is, for instance, little evidence of vegetation.
Another notable omission is the total absence offish to
date, though these must have been present in an en
vironment which supported labyrinthodonts. Of the
many invertebrates , including insects which must have
been present, only millipedes have so far been record
ed .

Considering its early thecodont ancestry it is pos
sible that sharp blade-Ilke teeth were the only option
available to Prolacerta; these are small in relation to the
size of the skull, compared with the teeth of, for exam
ple, Euparkeria. (This applies as well to Proterosuchus. )
Unfortunately no living reptiles have comparable
dentitions with the possible exception of Varanus
komodoensis . Prolacerta might be envisaged as feeding on
small prey which it would grab, kill, and swallow
whole, such as the young of the many small synapsids,
procolophonids and labyrinthodonts known from the
Lystrosaurus zone, as well as insects. It seems very likely
that the prolacertids were replaced by true
(pleurodont) lizards.

The rather larger Proterosuchus on the other hand
probably fed by tearing pieces of meat off a carcass
much as does Varanus komodoensis, as the teeth of these
two animals are closely comparable in size relative to
the skull, in shape, and in ' having finely serrated
posterior edges. Prolacerta lacks serrations on the teeth
and as these are present in H eleosaurus and Proterosuchus
they have presumably been secondarily lost in
Prolacerta.

Stealth more than agility would be required in hun
ting. Why then this light, swift creature? Presumably
no contemporaries could match Prolacerta for speed,
though some Galesaurids and Therocephalians
would certainly have eaten them, given a chance. The
most likely explanation is that it was to an extent ar
boreal-here lightness and agility are of advantage.
This would support Neill's (op . cit. ) argument for a
close link between an arboreal existence and the
evolution ofbipedality (in animals of this size).

RELATIONSHIPS OF PROLACERTA
Comparisons of all the possibly significant

features' of Pro lacerta cover a very wide field indeed.
Some of them are rapidly discountable, most
flounder on inadequate knowledge of many forms,
but in the end the reasonable options are few and a
stable picture seems to emerge.

Some of the earlier Permian diapsids should be
mentioned. Petrolacosaurus is a Pennsylvanian form
technically diapsid but remote in time and lacking the
strong morphological features which link the late Per-

mian and Triassic forms. Areoscelis merits attention as
the only fairly well known protorosaurian.
Relationship to Pro lacerta has been argued by Camp
(1945 ), Romer (1947) and Vaughn (1955). The only
striking parallel is in the elongation of the cervical
vertebrae, but this feature is common to several un
related groups. Areoscelis has in fact a curious specialis
ed postcranial skeleton (e.g. the elongate limbs )
behind a primitive Skull-certainly not a good
generalised ancestor. The double coracoid is another
point against it.

One animal which must be revived in this discussion
is the Russian Permian M esenosaurus. Here is an animal
with powerful carnivorous dentition (mode of im
plantation unknown ), a lower temporal opening with
lower temporal bar, several primitive characters such
as the maxilla entering the external naris, which
should merit consideration with early thecodonts, but
about which too little is known.

An early Permian diapsid is Mesosaurus, a curious,
little known, highly specialised form, but which with
Petrolacosaurus stands as a caution that diapsid saurop
sids became established early on. Hopefully, the
Brazilian material, reputedly of excellent potential,
will yield a good account of this animal.

One arrives then at the Eosuchia, the millerosaurs ,
younginids, rhynchocephalians (sphenodontids ),
squamates and archosaurs (Thecodontia ) for present
purposes, and the problem of relationships of these
groups.

The millerosaurs, which mayor may not have direct
bearing on later phylogenies, nevertheless exhibit cer
tain morphological details which are not shown by, for
instance, the younginids but which go a long way to
explain the derivation of early fhecodont characters
(particularly one may mention the quadratojugal and
its relationships).

Among the younginids, Youngina as described
above gives an apparently satisfactory rhyncho
cephalian ancestor, but it is necessary to note here that
all the known rhynchocephalians have basically
crushing dentitions and virtually akinetic skulls.

H eleosaurus (to be described by Carroll) is a good
thecodont ancestor with its blade-like marginal denti
tion and advanced femur indicating bipedality.

Much use has been made here of the term
"thecodont" as Pro lacerta is so obviously a thecodont
while the squamates are typically acrodont or more
commonly pleurodont (Edmund, 1969). This applies
to the earliest recognised true lizards, the
keuneosaurids, which are defined as having sub
pleurodont teeth.

Paliguana, contemporaneous with Prolacerta, is a
poor specimen, yet it has very much the morphology
and proportions of the keuneosaurids. One or two
probable lizard skeletons are known from the Permian
of South Africa, of which Palaeagama (Broom, 1926 )
has been described; unfortunately the temporal
region is badly damaged. Carroll believes he can
reconstruct it to look much like the keuneosaurids.

The lizard-like animals mentioned in this paragraph
are to be described by Carroll. There is thus some
evidence to suggest that the Squamata had evolved
before the end of the Permian, though more collecting
is needed to strengthen this .

To place Pro lacerta then weare left with two options.
Detailed comparisons with primitive thecodonts
prove instructive, as do comparisons to the
Protorosauria (sensu Romer, 1956). Comparisons
between Prolacerta and M acrqcnemus are so close there
can be little doubt of a dose relationship. Other
Protorosaurs are so poorly known that they cannot be
drawn into this relationship with certainty.

Proterosuchus (Cruickshank, 1972) is the one animal
close to Pro lacerta known in sufficient detail for
meaningful comparison. Much of the material has
also been available to the writer when additional in
formation was required. Proterosuchus occurs con
sistently lower in the Lystrosaurus zone than does
Prolacerta, but this probably does no more than
reflect an ecological separation. Proterosuchus is a
larger, more heavily built, and quadrupedal animal,
while Prolacerta is small, light and bipedal.
Proterosuchus has an antorbital fenestra which puts it
among the Archosauria, but Cruickshank has
suggested with some justification the derivation of
Proterosuchus from a "rhynchocephalian" ancestor;
such an animal would be required not to possess art
antorbital fenestra, and this step would not have
been far back in time.

Strip Proterosuchus of its archosaurian label and
compare with Prolacerta. The skull morphology and
detailed relationships of the bones are remarkably
close. The writer does not accept Cruickshank's inter
parietal, nor the jugal process of the quadratojugal
which he shows. The palatine does not form the entire
lateral margin of the suborbital vacuity as he shows,
bl.lt only runs halfway back. Cruickshank failed to note
that the teeth of Proterosuchus are serrated on their
posteriormargins.

Having disposed of these points the dlITerences are
now only of degree and quite consistent with the basic
differences of size and build.

Proterosuchus has apparently a proportionately
longer snout, achieved by compressing the orbit, with
a more downturned premaxilla; it has a very odd
retroarticular process (for which a functional inter
pretation is required). It is difficult to accept
Cruickshank's assertion that the quadrate was strep
tostylic. Surely this can only occur once the lower tem
poral bar has been breached?

The postcranial skeletons of the two animals pre
sent a very similar morphological grade. The girdles of
Proterosuchus are heavier, interclavicle and ilium hav
ing distinctive shapes. The bifurcations of the sacral
processes were one of the first points of strong
similarity noted in this study.

Similarity extends to the hind foot and ankle,
though there is some doubt regarding the phalangeal
count of Proterosuchus. Cruickshank has called the

117

astragalus intermedium when it is really a compoun
ding of intermedium and tibiale, and the centrale he
calls astragalus. Names aside, the feet of Prolacerta and
Proterosuchus are identical down to the hooked fifth
metatarsal .

The vertebral columns are very similar, though
Proterosuchus does not have such elongate cervicals. This
similarity extends to the large blade-like caudal chevrons,
though these are only two-thirds the length in
Proterosuchus .

All the above gave sufficient grounds for expecting
to find an explanation for the braincase of Prolacerta in
Proterosuchus. One may expect braincases to be fairly
conservative at this level; also, while that of Prolacerta is
poorly ossified in places, this is consistent with a lil?ht
kinetic skull, while in the larger Proterosuchus the bram
case is well ossified. This comparison, detailed
elsewhere, is so striking in every detail as to put the seal
ona very close relationship .

The quadratojugal may be one of the more impor
tant elements in deciding lizardlrhyncho
cephalian/archosaurian relationships. Gow (1972)
has described the relationships of this bone in
millerosaurs in detail, and this pattern could veryeasi
ly give rise to what we see in Prolacerta and Proterosuchus
and the later archosaurs, e.g. Euparkeria (Ewer, 1965)
where the quadratojugal rests on a facet on the lateral
condyle of the quadrate running up to meet the
squamosal above, lateral to a quadrate foramen .
From the reduced condition in Pro lacerta it is of little
consequence whether this bone is lost or retained by
M acrocnemus. It is difficult to see how this quadrato
jugal arrangement could be derived from Youngina.

The Triassic reptiles of Monte San Giorgio (Kuhn
Schnyder, 1963 and others) are a diverse collection
mostly of specialised marine forms; it is thus sur
prising to find amongst them an animal very close in
deed to Pro lacerta, namely M aero en emus bassanii
Nopsca. Owing to their poor mode of occurrence
these fossils are not as well known as one would like,
particularly in details of skull structure. Nonetheless
Kuhn- Schnyder (1963) described a skull of
M acrocnemus in sufficient detail to conclude con
vincingly that this animal is closely allied to
Prolacerta. In that paper he suggested the possibility
that the lower temporal opening of diapsids evolved
before the upper. The writer has suggested (Gow,
1972) this might apply in the origin of squamates. In
view of the standard morphology of upper temporal
openings of younginids and early thecodonts and
bearing in mind the presence of two temporal
openings in the upper Carboniferous Petrolacosaurus
these suggestions are open to doubt. It seems likely
that the lower temporal bar has been lost (along with
the quadratojugal? ) in Macrocnemus and Tanystropheus .
Prolacerta constitutes the ideal starting point for the
derivation of these forms.

In Kuhn-Schnyder's 1963 paper there is an ex
cellent photograph of an almost complete skeleton of
Macrocnemus. The resemblance to Prolacerta is so close

liS

as hardly to bear detailing and should dispel all doubt
regarding the close relationship of these animals.
Further checks against Peyer's (1937) original detailed
description of Macrocnemus with its many excellent
ph~tographs also serve as a good standard of com
panson.

Size, proportion and morphology are closely com
parable in the two animals. The few concrete
differences are predictable and only slight advances
occur in M acrocnemus.

As far as the M acrocnemus skull is known it is very
similar to that of Pro lacerta. Doubt exists as to the
nature of the snout and the position of the external
nares. The hind margin of the jugal is not known with
certainty, nor can the presence of a quadratojugal be
shown, though it seems certain that no lower temporal
bar was present. Details of squamosal, quadrate and
upper temporal fenestra are identical to the condition
in Prolacerta. Dentitions are very similar in the two
animals. What can be seen of the palate of Macrocnemus
agrees well, and the shape of the basisphenoid is iden
tical.

The vertebrae and ribs agree very closely, even to
the divided second sacral transverse process and the
large flattened caudal chevrons.

Ossification of the girdles in M acrocnemus shows a
considerable reduction over that in Pro lacerta but
this is almost universal in middle Triassic reptiles
and to be expected and involves just those areas that
are extremely thin in the girdles of Prolacerta. The il
ium and inter-clavicle are still virtually in
terchangeable between the two animals. Limb
proportions are the same in both animals, both have
hooked fifth metatarsals and the same digital for
mulae. Slight differences in ankle morphology
reflect only the time gap.

It is thus clear that Pro lacerta is representative of
an as yet poorly known group of early Triassic
prearchosaurian thecodonts which in all probability
gave rise to the later M acrocnemus and Tanystropheus.
These can then be regarded as a sterile group which
paralleled the Squamata in the loss of the lower tem
poral arcade.

PROLACERTA-SYNONYMY AND DIAGNOSIS
Subclass: Incertae sedis

Probably most workers would favour
retention in the Leeidosauria

Order Parathecodontia nov.
Thecodont reptiles of diapsid origin in which the

lower temporal arcade has been breached. Very close
ly allied to proterosuchian thecodonts in most aspects
of skeletal morphology but precluded from the
Subclass Archosauria by the almost certainly primitive
absence of antorbital fenestrae.

Family Prolacertidae Parrington 1935.
Small agile light-boned bipedal animals. Quadrate

streptostylic and quadratojugal reduced or absent.
Pro lacerta broomi Parrington 1935.
= Pricealongiceps Broom and Robinson, 1948.

Lower Triassic prolacertids from the Lystrosaurus
zone of South Africa. Girdles large, unfenestrated,
plate-like structures. Jugal spur and quadratojugal
present.

M acrocnemus bassanii N opcsa 1930
European Middle Triassic prolacertids. Girdles

reduced. Posterior border of jugal smooth,
quadratojugal absent.

Family Tanystropheidae Romer 1956.
Tanystropheus longobardicus Meyer 1852.
Skull very similar to that of Prolacertidae but with

heterodonty in juveniles, in which the teeth in the
posterior half of both jaws are tricuspid. Greatly
elongate neck with longest neck vertebra five times as
long as the longest trunk vertebra. Considerably
larger than Prolacertidae-in excess oHour metres in
length.

DISCUSSION

The reptilian braincase
Although it is now known that the incomplete

lower temporal bar is not confined to lizards but is
present also in the sphenodontid Glevosaurus
(Robinson, 1973) and certain millerettids (Gow,
1972), and in spite of the fact that Pro lacerta has
blade-like carnivorous theodont dentition, it seems
desirable to establish the differences between
squamates and archosaurs on as many features as
possible, particularly those of a predictably conser
vative nature. Clear differences in braincase
morphology would provide this additional distinc
tion.