Annals of the South African Museum = Annale van die Suid-Afrikaanse Museum

QH LIEUWE DIRK BOONSTRA

1 GIRDLES AND LIMBS OF THE
PRISTEROGNATHID THEROCEPHALIA

April 1964 April
Volume 48 Band
Pattie so5 1: Deel

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from The Librarian, South African Museum, Cape Town
(Cash with order, post free)

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van Die Bibliotekaresse, Suid-Afrikaanse Museum, Kaapstad
(Kontant met bestelling, posvry)

OUT OF PRINT/UIT DRUK

1(1-2), (1, 3> 5> 7)> 3(1), 5(2, eo.
t.—p.1.), 6(1, (=o) 8, 9(1), 10(1, 3)> 11(7),
21, 24(2) 31(1-2), 44(4)

Price of this part/ Prys van hierdie deel

R1.25

Printed in South Africa by In Suid-Afrika gedruk deur
The Rustica Press Pty. Ltd. Die Rustica-pers Edms. Bpk.
Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap

THE GIRDLES AND LIMBS OF THE PRISTEROGNATHID
THEROCEPHALIA

By
LrzuwE Dirk BoonstRa
South African Museum, Cape Town

(With 50 text-figures and one Plate)

CONTENTS

PAGE

Introduction ee ee wa eT
Kinnosaurus paucidens gen. ae sp. nov. eek 22
Material .. Ss ae aie 46), LOA
Historical .. ws 50 20
General Mornholosy we 2a 56. LAD
Pectoral Girdle a ae oe 126
Humerus ns A 36 LAAs}
Ulna and Raghne a des aol NGO)
Carpus and Manus .. ee ao HS{o)
Pelvic Girdle .. she ee ee MGT
Femur j eo co. | Ue
Tibia, Fibula aad pee she See 13d)
Systematic Descriptions of Specimens .. 136
A. Pectoral Girdle .. Be sa 0GXS)

Bs Elumerusi. - po WANG}

C. Lower Forelimb and Forefoot Se 150,

D. Pelvic Girdle ies at Soe iD Gs

E. Femur 3 bo poy REX)

F. Lower Hind Aaa acd Pes eo
Discussion .. a ae a Lod!
Summary .. ; a es LOS
Aetuawledeenients) a a8 oo OR
References .. ie ae ae Sig) HOR

INTRODUCTION

In the Tapinocephalus zone of the Karoo the oldest Therocephalia known
are well represented. The Therocephalia form the main component of the
carnivores, the only other carnivores being the very large anteosaurs represented
by five genera and a few small to medium-sized Gorgonopsia represented by
10 genera, of which 30 specimens are known in ail.

The Therocephalia are represented in the Tapinocephalus zone by the
families Pristerognathidae, Lycosuchidae, Scylacosauridae and Scaloposauridae.
Of these the pristerognathids are the dominant family and are well represented
in numbers of specimens as well as in the number of distinct species. Hitherto
16 genera with 28 species have been described. Most of the specimens consist of
partly preserved snouts, with good complete skulls few in number. Of the
postcranial skeleton little is known, for only rarely are some of the postcranial

121
SMITHSOMIAR
istiumion 40 SUN 2 2 1064

Ann. S. Afr. Mus. 48 (5), 1964: 121-165, 50 figs. and 1 Plate.

122 ANNALS OF THE SOUTH AFRICAN MUSEUM

bones preserved and then mostly associated with cranial parts difficult to
classify.

In this paper I present an account of the girdle and limb material in the
South African Museum mostly collected by myself since 1928.

This collection includes material which necessitates the establishment of
the following new genus and species.

ZINNOSAURUS new genus

Type-species: Kinnosaurus paucidens.

Diagnosis: A medium-sized pristerognathid with the dental formula 1.5, c.1,
pc.2; the incisors are fairly weak and well spaced and the postcanines are weak
and lie far apart; the postcanines situated on a thin lappet of the maxilla
clearly demarcated from the general maxillary surface. Scapula with a well-
developed flange-like process on its posterior border immediately above the
glenoid for the scapular head of the triceps. The glenoid widely open and facing
much laterally. Humerus with greatly expanded epicondyles, thin and sheet-
like. Femur with broadly oval caput and strong thickened internal trochanter.

Affinities: Nearest relative Glanosuchus.

Ainnosaurus paucidens new species
(Figs. 1, 19, 28, 46d and 46e)

Holotype. S.:A.M. 12185, skull with lower jaw, scapulo-coracoid, humeri, ends of
femora.

Locality: Meyerspoort, Beaufort West, South Africa.

Horizon: High Tapinocephalus zone, Lower Beaufort Beds, Karoo.

I have great pleasure in naming this new genus after H. Zinn, who has, as
techinical assistant, accompanied me on numerous collecting trips in the
Tapinocephalus zone.

The scapulo-coracoid, humeri and ends of femora of the holotype are
described in the appropriate sections of this paper.

MATERIAL

In the South African Museum catalogue 219 specimens of pristerognathids
from the Tapinocephalus zone are listed. Of these only the following 37 specimens
have parts of the girdles and limbs preserved. Taken together one gets a good
idea of the family character of these structures.

The material studied consists of the following specimens listed in the order
of the date of acquisition by the South African Museum collection.

S.A.M. 4335. Pristerognathus sp. Beaufort West. High? Tapinocéphalus zone. Coll.

Whaits 1916? Incomplete coraco-scapulae, distal and proximal ends of a

humerus and the proximal halves of an ulna and radius. No skull.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 123

S.A.M. 5018. Pristerognathus sp. Abrahamskraal, Prince Albert. Low Tapino-
cephalus zone. Coll. Haughton 1916. Two incomplete ilia, proximal and
distal halves of two femora and of two humeri, proximal ends of tibiae,
radii and ulnae. No skull.

Fic. 1. Zinnosaurus paucidens gen. et spec. nov. S.A.M. 12185.
Type <4. Skull. a, lateral. b, dorsal.

S.A.M. 9005. Scymnosaurus major. Holotype. Klein-Koedoeskop, Beaufort West.
Low Tapinocephalus zone. Coll. Boonstra 1929. Interclavicle with proximal
parts of both clavicles, distal ends of both humeri, complete radius and
ulna, associated with a good anterior half of the skull.

S.A.M. 9084. Scymnosaurus ferox. Rietkuil, Beaufort West. Low Tapinocephalus
zone. Coll. Boonstra 1929. A coracoid, proximal and distal ends of a
humerus and proximal ends of an ulna and radius associated with a good
skull.

124 ANNALS OF THE SOUTH AFRICAN MUSEUM

S.A.M. 11458. Pristerognathus sp. Mynhardtskraal, Beaufort West. Low Tapino-
cephalus zone. Coll. Boonstra and Avenant 1939. An interclavicle, proximal
ends of clavicles, parts of coracoids, poor ends of femur, humerus, radius,
ulna. No cranial material.

S.A.M. 11459. Scymnosaurus sp. Buffelsvlei, Beaufort West. Low Tapinocephalus
zone. Coll. Boonstra and Marais 1939. Part of carpus and tarsus associated
with a poor snout.

S.A.M. 11557. Scymnosaurus sp. Die Cypher, Beaufort West. Low Tapinocephalus
zone. Coll. Boonstra 1940. A scapula, humerus, femur, interclavicle,
clavicle, radius, and partial manus. Without skull.

S.A.M. 11558A. Scymnosaurus? Die Cypher, Beaufort West. Low Tap
zone. Coll. Boonstra 1940. A pair of well-preserved ischia.

S.A.M. 11695. Scymnosaurus sp. Seleryfontein, Fraserburg. Low? Tapinoeee
zone. Coll. Boonstra and Jooste 1946. A coraco-scapula, interclavicle,
clavicles and humerus. With good anterior half of skull.

S.A.M. 11794 Cynariognathus sp. Seleryfontein, Fraserburg. Low? Tapinocephalus
zone. Coll. Jooste 1947. Imperfect pelvis, femora, humeri, epipodials and
manus. With a poor skull.

S.A.M. 11888. Therioides cyniscus. Holotype. Vindragersfontein. Beaufort West.
Low Tapinocephalus zone. Coll. Boonstra and Zinn 1948. Parts of pectoral
girdle, humerus, radius, ulna and part of manus associated with a nearly
complete skull.

S.A.M. 11934. Pristerognathus? Steenboksfontein, Laingsburg. Low? Tapino-
cephalus zone. Coll. Boonstra 1951. Distal end of femur and proximal ends
of tibia and fibula. Without cranial association.

S.A.M. 11936. Pristerognathoides sp. Bosluiskraal, Laingsburg. Low? Tapino-
cephalus zone. Coll. Boonstra 1951. Proximal and distal end of a femur
associated with a snout.

S.A.M. 11942. Ptomalestes avidus. Holotype. Steenboksfontein, Laingsburg. Low ?
Tapinocephalus zone. Coll. Boonstra 1951. Part of pectoral girdle, humeri,
ulnae and radii associated with a good skull.

S.A.M. 11957. Scymnosaurus sp. Abrahamskraal, Prince Albert. Low Tapino-
cephalus zone. Coll. Le Roux 1923. Acetabular parts of ischium and pubis
and proximal end of femur. Without skull.

S.A.M. 12051. Alopecognathus sp. Rietfontein, Lamgsburg. Low? Tapinocephalus
zone. Coll. Boonstra and Fourie 1957. Part of scapula, proximal end of
humerus, proximal and distal ends of femur, ends of radii and ulna and
part carpus, tibia and fibula and a complete pes associated with a snout.

S.A.M. 12102. Pristerognathoides sp. Kalkkraal. Prince Albert. Low Tapino-
cephalus zone. Coll. Boonstra and Zinn 1957. Most of the pectoral girdle,
ends of humerus, ulna and radius associated with a complete skull.

S.A.M. 12112. Pristerognathus? Skoppelmaaikraal, Laingsburg. Low Tapino-
cephalus zone. Coll. Botes 1957. Part of ilium, proximal end of humerus and
other fragments without any skull parts.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 125

S.A.M. 12118. Scymnosaurus sp. Palmietfontein of Kruidfontein, Prince Albert.
Low Tapinocephalus zone. Coll. Boonstra and Zinn 1957. Proximal end. of
femur, part scapula.

S.A.M. 12185. <innosaurus paucidens gen. et spec. nov. Meyerspoort, Beaufort
West. High Tapinocephalus zone. Coll. Boonstra and Zinn 1959. Part
coraco-scapula, one complete humerus and one proximal half, proximal
and distal end of femur and proximal end of ulna and radius and distal
end of radius associated with a good skull.

S.A.M. 12193. Scymnosaurus sp. Fortuin of Dalajalon, Beaufort West. High
Tapinocephalus zone. Coll. Boonstra and Zinn 1959. Fairly complete
pectoral girdle, humerus, ulna and radius associated with piece of jaw
containing teeth.

S.A.M. 12204. Pristerognathoides sp. Plaatjiesrivier, Beaufort West. High
Tapinocephalus zone. Coll. Boonstra and Zinn 1959. Fairly complete
pectoral girdle, humerus, ulna and radius associated with the posterior
third of a skull.

S.A.M. 12262. Scymnosaurus sp. Seleryfontein, Fraserburg. Low? Tapinocephalus
zone. Coll. Boonstra and Jooste 1959. Femur, radius, ulna and manus.
S.A.M. Kai8. Alopecognathus sp. Lammerkraal, Prince Albert. High
Tapinocephalus zone. Coll. Boonstra, Zinn and Gow 1959. Weathered

pelvis.

S.A.M. K233A and B. Alopecognathus sp. Palmietfontein, Beaufort West. High
Tapinocephalus zone. Coll. Boonstra, Zinn and Gow 1959. Two incomplete
pectoral girdles.

S.A.M. K223C. Pristerognathoides sp. Palmietfontein, Beaufort West. High
T apinocephalus zone. Coll. Boonstra, Zinn and Gow 1959. A nearly complete
pelvic girdle associated with a skull.

S.A.M. K227. Pristerognathus sp. Louisrus of Dalajalon, Beaufort West. Tapino-
cephalus zone. Coll. Boonstra and Zinn 1959. Two ilia found in association
with a poor snout.

S.A.M. Ka3r. Alopecognathus angusticeps Lammerkraal, Prince Albert. High
Tapinocephalus zone. Coll. Boonstra and Gow 1959. The major part of an
articulated skeleton articulated to a skull.

S.A.M. K234. Pristerognathus sp. Palmietfontein, Beaufort West. High Tapino-
cephalus zone. Coll. Boonstra, Zinn and Gow 1959. Part of pelvis associated
with a snout.

S.A.M. K238A. Pristerognathus? Lammerkraal, Prince Albert. High Tapino-
cephalus zone. Coll. Boonstra and Zinn 1959. A pelvis and pes.

S.A.M. K245A. Pristerognathid? Rietfontein of Vlakfontein, Beaufort West.
High Tapinocephalus zone. Coll. Boonstra, Zinn and Gow 1959. A pelvis
and part of the pes.

S.A.M. K306. Pristerognathus sp. Kranskraal, Beaufort West. Low Tapino-
cephalus zone. Coll. Boonstra, Zinn and Gow 1960. Distal and proximal
ends of two femora.

126 ANNALS OF THE SOUTH AFRICAN MUSEUM

S.A.M. K317. Pristerognathus sp. Bulwater, Beaufort West. Low Tapinocephalus
zone. Coll. Boonstra and Gow 1960. Proximal and distal ends of femur
with parts of skull.

S.A.M. K339. Alopecognathus sp. Klipbanksfontein, Beaufort West. Tapzno-
cephalus zone. Coll. Boonstra and Zinn 1962. Incomplete pectoral girdle
associated with anterior two-thirds of a skull.

S.A.M. K352. Scymnosaurus sp. Skoppelmaaikraal, Laingsburg. Low? Tapino-
cephalus zone. Coll. Boonstra 1962. Femur (isolated).

S.A.M. K353. Scymnosaurus sp. Klein-Koedoeskop, Beaufort West. Low
Tapinocephalus zone. Coll. Boonstra 1929. Proximal end of a large femur.

HIsTORICAL

-Hitherto little was known of the postcranial skeleton of the Pristerognathi-
dae.

In 1929 Broom published a figure of the shoulder girdle of Pristerognathus
minor as seen in ventral view. Broom does not state on what specimen this
restoration was based and I have not been able to trace the specimen. In the
sequel I am reproducing figures (fig. 2) of the pristerognathid pectoral girdle
drawn from a model I have carved out of a rectangular block of plaster which
has enabled me to be quite sure that the lateral, anterior and ventral views are
really at right angles to each other. My ventral view differs considerably from
that given by Broom, particularly in the disposition of the scapulo-coracoid.

In 1932 Broom published as a front view of the pelvis of a therocephalian
(possibly Pristerognathus minor) what is in fact a ventral view, but again I have
not been able to trace the specimen on which the drawing was based.

The specimen on which Seeley founded Theriodesmus phylarcus, from the
Endothiodon zone, has by some recent compilators been referred to the Pristero-
gnathidae. Comparison with the pristerognathid material described in this
paper shows that Theriodesmus is not a therocephalian at all, but really a
gorgonopsian.

GENERAL MORPHOLOGY
Pectoral Girdle (fig. 2)

This general account is compiled from facts derived from 18 specimens
representing 6 pristerognathid genera described in the systematic part of this
paper. The material as a whole is not well preserved and in even the best
specimens the constituent bones have been displaced relative to each other and
distorted mostly by dorso-ventral compression. The reconstructions presented
here in semi-diagrammatic form have been drawn from a composite model
carved out in a block of plaster.

As in all primitive reptiles the pectoral girdle is composed of 11 bones—
one unpaired and five paired, but in no case is the cleithrum preserved and its
presence is only indicated by the facet on the scapula to which it was applied.
There is no ossified sternum.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 129),

The girdle consists of two sets of bones—replacement and dermal bones.
The replacement bones, forming the so-called scapular girdle, form a scapulo-
coracoid composed of a dorsal plain blade-like scapula, lying practically
vertically but curving slightly round the thorax, and a ventral antero-posteriorly
elongated coracoidal plate, composed of a large precoracoid and a smaller
posterior coracoid, curving gently inwards towards the middle line. The pair of
scapulo-coracoids are held in position relative to each other by their connexion

Fic. 2. Diagrams of the pristerognathid pectoral girdle drawn from a model based on a number of
specimens. a, lateral. b, anterior. c, ventral.

CO—coracoid (posterior). CL—clavicle. CM—cleithrum. g—glenoid. ICL—interclavicle. PGO—
precoracoid (anterior). SC—scapula. scf—foramen supracoracoideus. tr—origin of the scapular head
of the m. triceps (in Zinnosaurus the origin is from a distinct tubercle).

with the clavicular girdle of dermal bones consisting of an unpaired medially
disposed interclavicle and a pair of bracing clavicles. Ventrally the coracoidal
plate rests on the upper face of the interclavicle. Ventrally the spatulate end
of each clavicle curves below around the anterior end of the interclavicle to
fit into an oval hollow on the under face of this bone. Dorsally the stem of the
clavicle sweeps upwards externally of the anterior end of the precoracoid and
the lower part of the scapula and is then applied to the anterior edge of the
scapular blade. Here it meets the cleithrum which is in all probability a splint-
like bone applied to the upper anterior scapular edge.

On the posterior border of the scapulo-coracoid, at the junction of the
scapular and the coracoid, lies the simple, antero-posteriorly shortened, glenoid

3

128 ANNALS OF THE SOUTH AFRICAN MUSEUM

cavity. The glenoid has a dorsal scapular facet which faces ventro-posteriorly
and but slightly externally, and a ventral coracoid facet which faces dorso-
externally and slightly posteriorly. The nature of the glenoid cavity prevents the
humerus from assuming any appreciable downward disposition and also limits
an anterior disposition, but it can be freely directed upwards and backwards.

The area of origin of the scapular head of the triceps muscle from the
postero-lateral surface of the scapula above the glenoid is very indistinct, except
in one specimen, the type of the new genus <innosaurus, where a prominent
tubercle is developed very similar to that present in the anteosaurian dino-
cephalian genus, Eccasaurus. On the anterior edge of the scapula there is no
indication of an incipient acromion process. The scapula blade is flat with no
indication of any spine.

The precoracoid takes no part in the formation of the glenoid cavity and
just enters into its anterior rim. Immediately anterior to this lies the foramen
supracoracoideus. The anterior extent of the precoracoid is great, forming a
large surface for the origin of the m. supra-coracoideus. Above the precoracoid
lies the thin anterior plate of the scapula from whose outer face the m. scapulo-
humeralis originates.

The coracoid is a smaller but more heavily built bone than the precoracoid,
with a large strong glenoid facet. There is no special process as in the pelycosaurs
for the origin of the coracoidal head of the triceps and this head was probably
absent.

The clavicle of the pristerognathids is very distinctive, differing greatly
from both that of the more primitive pelycosaurs and of the other contemporary
therapsids. It is peculiar in that the ventral spatulate end curving round to the
under face of the interclavicle is sharply bent backwards and extends far
posteriorly along the under face of the interclavicle. The dorsally sweeping
stem is also relatively stronger.

The interclavicle is a long, well-developed bone with the anterior spatulate
end rounded in outline, not much expanded laterally and curving upwards
only very slightly. In these features it differs markedly from the interclavicle
in both pelycosaurs and the other therapsid contemporaries.

The cartilaginous sternum presumably lay above the spatulate posterior
end of the interclavicle and extending to the coracoids.

Humerus (fig. 3)

A dozen fairly complete humeri and numerous proximal and/or distal
ends are available for study, but bad preservation due to both erosion and
post-mortem deformation makes it difficult to get a good picture of the nature
of the pristerognathid humerus. It is particularly difficult to determine the
angle subtended by the proximal and distal ends and thus the rotation on the
shaft.

The pristerognathid humerus varies from a fairly light bone with
moderately expanded ends and a fairly long and slender shaft to a moderately

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 129

heavy bone with greatly expanded ends and a short thickened shaft. The rota-
tion of the ends on the shaft apparently varies from 10° to 25°.

The proximal surface has the processus medialis and processus lateralis
indistinctly demarcated from the caput which is narrowly oval in outline. The
facet of the caput curves a little on to the dorsal surface of the humerus. The
delto-pectoral crest is fairly weak and this subsides into the shaft without
continuing as an oblique ridge in the direction of the entepicondylar foramen.

cap

a b Cc

Fic. 3. Diagrams of the pristerognathid humerus. a, dorsal. 6, ventral.
¢, proximal.

advl—anterior dorso-ventral line. bf—bicipital fossa. br—origin of the m.
brachialis. cap—caput humeralis. d—insertion of the m. deltoideus. dpc—
delto-pectoral crest. ect—ectepicondyle (radial). ecf—ectepicondylar foramen.
ent—entepicondyle (ulnar). enf—entepicondylar foramen. 1d—insertion of the
m. latissimus dorsi. lml—latero-medial line. of—fossa for olecranon (trochlear).
pl—processus lateralis. pm—processus medialis. rc—radial condyle or capitel-
lum. tr—origin of the medial head of the triceps. uc—ulnar condyle.

The bicipital fossa is deep with a strong rounded posterior rim whose posterior
face forms a rectangular area for the origin of the medial head of the triceps.

The proximo-dorsal surface of the humerus is divided by the anterior
dorso-ventral line (ADVL) into two parts. Anteriorly (preaxially) to this line
lies the roughly triangular surface for the insertion of the m. deltoideus. On the
surface posterior (postaxially) of the ADVL there is a weak oblique latero-
median line (LML). Anterior to this line the area of insertion indicates a
strong m. latissimus dorsi. Posterior to the LML a well-developed rectangular
area indicates the origin of a strong medial head of the m. triceps.

130 ANNALS OF THE SOUTH AFRICAN MUSEUM

Distally the epicondyles are usually moderately developed, except in
Kinnosaurus, where the epicondyles have widely expanded thin flanges of bone.
In this genus the ectepicondyle, which is fused to a well-developed supinator
flange, forms a widely expanded thin sheet of bone. A rounded foramen pierces
the supinator-ectepicondylar flange obliquely. The entepicondyle in this form
forms an extensive distal sheet of bone indicating well-developed flexors.

Distally the condyles are weakly developed. This is particularly so in the
case of the radial condyle whose articulatory face lies practically wholly distally
with hardly any ventral surface and no indication of a ventral bulbous swelling
(capitellum) as is present in most of the contemporary therapsids. On the
dorso-distal surface the trochlear fossa is shallow, which is related to the virtual
absence of an olecranon on the ulna.

The elbow joint in the pristerognathids thus differs markedly from that of
the other therapsids.

The distal position of the distal humeral condyles makes a more upright
dispostion of the limb possible. With the weak development of the olecranon a
deep trochlear fossa is not necessary for the extension of the epipodial.

Ulna and Radius (fig. 4)

The pristerognathid ulna has its proximal end greatly expanded. Proxi-
mally the lateral corner is developed into a short but thickened olecranon with
a rugose surface for the reception of the m. triceps. The sigmoid face for the
reception of the humerus is shallow but the coronoid process is well developed.
Dorsally the sigmoid rim is concave to receive the head of the radius, and the
sigmoid face of the ulna and the proximal face of the radius form a common
articulatory facet for the reception of the humerus. This facet is shallow and, the
humerus having hardly a capitellum, the elbow joint forms a poor hinge joint.

The sigmoid face of the ulna lies on the proximal end of the bone and the
coronoid process on the medial edge is also proximally situated, so that in the
elbow joint the propodial and epipodial meet end-on without the ulna curving
round the distal end of the humerus.

Carpus and Manus (fig. 4)

The proximal row of carpals consists of three elements—a robust radiale, a
lighter elongated ulnare and a small laterally compressed intermedium. A
pebble-like pisiforme lies laterally of the ulna-ulnare articulation.

In the middle row there are two centrals—one lying between the radiale
and the first and second distals, the central one lying wedged in between the
ulnare and radiale and distally articulated with the third distal.

There are four distals—the fourth and fifth are fused and articulate with
the fourth and fifth metacarpals.

The metacarpals are well developed—the first is very short and looks
very much like the first phalanx, the second is nearly twice as long as the first,
the third about 23 times as long, the fourth 3 times as long and the fifth just

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 131

over 3 times as long. The phalangeal formula is 2, 3, 3, 3, 3. The proximal
phalanges are all fairly short and robust and there is no indication of any
further reduction. The ungual phalanges carried talon-like claws and have a
ventral thickening proximally.

The first digit is the shortest, then come the second and fifth of equal length,
then the third and fourth which are again of equal length.

in dorsal view.

Go Fic. 4. Diagram of the priste-
rognathid epipodial and manus

C,, C,—centrals. I—inter-
medium. P—pisiforme. R—

(0)
yd radius. RE—radiale. U—ulna.
UE —uIlnare. 1-5—distals. I-V
Mo » sys 4 metacarpals.
5 5

V

The purchase of the forefoot on the ground is thus more meso-post-axonic
than pre-axonic.

The digits as a whole are short and the metacarpals and carpus in com-
parison long. The manus can be considered semi-digitigrade.

Pelvic Girdle (fig. 5)

Pelves are not well represented and this account is compiled from half a
dozen specimens representing four genera.

132 ANNALS OF THE SOUTH AFRICAN MUSEUM

The pelvis is much lower than the pectoral girdle and the pubo-ischiadic
plate shorter than the interciavicle, but longer than the coracoidal plate.

The three bones of each side are not firmly ankylosed and the junction
between the two sides is weak between the pubis but stronger at the ischiadic
keel.

The acetabulum is large and nearly circular in outline; all three bones
take part, with the ilium contributing about half. The acetabulum faces mainly
outwards; it is shallow, with strong rims for the attachment of the joint capsule

d D

Fic. 5. Diagram of the pristerognathid pelvis. a, lateral. b, ventral.

a—acetabulum. ap—anterior process of the iliac blade. ife—area of origin of

m. ilio-femoralis. ifi—area of origin of m. ilio-fibularis. it—area of origin of

m. ilio-tibialis. IL—ilium. IS—ischium. isk—ventral keel of the ischia. ist—

ischial tuber. P— pubis. pf— pubic foramen. pp—posterior process of the iliac

blade. pt— pubic tuber. sab—supra-acetabular buttress. san—supra-acetabular
notch.

and ligaments. Dorsally lies a strong supra-acetabular buttress formed by the
ilium, overhanging the acetabulum and forming the main attachment of the
joint capsule. At the dorso-posterior corner of the acetabulum, just posterior to
the buttress, lies the supra-acetabular notch.

At the level of the iliac buttress the iliac blade is antero-posteriorly
expanded to form a high anterior and a somewhat lower posterior process. In
most specimens the anterior process is much shorter than the posterior process,
but in two specimens the lower corner of the anterior process is prolonged to
form a long but weak process. In the other specimens there is an indication
of this process. More and better-preserved specimens may prove that this
elongation is normal for the family. The outer face of the iliac blade is both
dorso-ventrally as well as antero-posteriorly convex. In antero-dorsal direction
runs a shallow groove and near the dorsal edge in a vertical line with the
posterior end of the buttress there is also a slight hollow. One can thus speak
of an undulating surface. There is no eversion of either anterior or posterior

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 133

_edges. The areas of origin of the ilio-fibularis, ilio-femoralis and ilio-tibialis are
thus indefinitely determinable.

The ischium in its acetabular part is massive and forms here, the strongly
thickened postero-ventral segment of the strong acetabular rim. From here it
tapers posteriorly and forms a thickened upper edge but without a distinct
ischial tuber. Extending medially the ventral surface is flattened anteriorly
but just behind the junction with the pubis it carries a strong and prominent
ventral keel.

The pubis in its acetabular part is fairly massive and here forms a fairly
strong acetabular rim. From here it tapers anteriorly and forms a thickened
lateral edge but without a marked pubic tuber. The ventral surface of the
pubis is fairly flat. Just medial to the acetabular rim, near its posterior edge,
lies a well-developed pubic foramen.

The pubo-ischiadic plate is broad and large as in the pelycosaurs, the
antero-posterior dimension of the pubis being relatively larger than in other
therapsids and both anterior and posterior pelvic openings broad and not
V-shaped as in other therapsids; but the posterior opening is incipiently
V-shaped because of the ischial keel.

Femur (fig. 6)

The pristerognathid femur is represented in 20 specimens of 5 separate
genera. Ten femora are fairly complete and there are a couple of dozen poor to
good distal and proximal ends unconnected and usually lacking a shaft.

The femur, always longer than the humerus, is a fairly light bone with
only slightly expanded ends and with a long, fairly slender shaft. There is a
considerable twist on the shaft so that the proximal head and the distal condyles
subtend an angle of 30°-40°. The distal preaxial condyle lies a little further
ventrally than the postaxial distal condyle and a little proximal of its fellow.

The preaxial border is more concave than the postaxial border due to
the femoral head being directed somewhat preaxially.

Proximally the caput femoris is terminal but somewhat preaxially directed ;
it is antero-posteriorly elongated, thick, preaxially with a rounded edge;
postaxially it tapers and curves gently towards the external trochanter, into
which it flows. In its thickened preaxial part the caput is convex and rounded
in its postaxial part. The caput femoris is much smaller than the acetabulum.

The external trochanter forms the proximo-postaxial rounded corner of
the femur; it flows with a rounded curve into both the caput and the postaxial
edge.
On the dorsal proximal surface lies a well-developed ridge near the preaxial
border of the femur; it lies parallel to the preaxial border and commencing
from the caput runs for a short distance distally and then fades away. This
ridge serves as the area of insertion of the m. pubo-ischio femoralis internus.
Between this ridge and the external trochanter lies the area of insertion of the
m. ilio-femoralis.

134 ANNALS OF THE SOUTH AFRICAN MUSEUM

On the ventral proximal surface a sharp, prominent longitudinal ridge is
developed. Lying in the middle of the bone, it commences a short distance
distally of the caput and extends distally for a variable distance to fade away
into the general ventral face of the shaft. This ridge is the only part of the
primitive Y system of adductor ridges preserved in the pristerognathids and
represents an internal trochanter.

d D € d

@xt

ec en ec
Fic. 6. Diagrams of the pristerognathid femur. a, dorsal. 6, ventral.
c, anterior. d, posterior. e, proximal.

ad—insertion of adductor muscle. cap—caput femoris. ec—ecto-

condyle. en—entocondyle. ext—external trochanter (major). ft—

—origin of m. femoro-tibialis. g—origin of the gastrocnemius.

if—insertion of the m. ilio-femoralis. int—internal trochanter.

ifo—intertrochanteric fossa. pifi—ridge on which is inserted the
m. pubo-ischio-femoralis internus.

Between the internal and external trochanters lies a relatively small
intertrochanteric fossa.

The distal condyles lie terminally facing very slightly ventrally and are
well modelled. The intercondylar sulcus is very shallow. Dorsally the inter-
condylar fossa is very shallow, but ventrally the gastrocnemic fossa is quite well
developed.

Tibia, Fibula and Pes (fig. 7)

The posterior epipodial and pes are not as well represented in the material
at my disposal as the corresponding structures of the forelimb.

The posterior epipodial is longer than the anterior one, as we have already
seen is also the case in regard to the propodial.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 135

Both tibia and fibula are lightly built, slender bones, lighter and slenderer
than the radius and ulna. In both epipodials the pairs of bones are approximately
of equal length. Proximally the tibia and fibula form a common terminal
articular facet to meet the femoral condyles end on. Distally the tibia articulates
with a rounded face on the astragalus and the fibula with an elongated facet
on the proximal end of the calcaneum.

tf Fic. 7. Diagram of the
pristerognathid epipodial
and pes in dorsal view.
AS—astragulus. CA—cal-
@ caneum. C—central. F—
4 fibula. T—tibia. tf—tarsal
LS D foramen. 1-5—distals. I-V
) ( ue — metatarsals.
ES

As in pelycosaurs and therapsids generally the pristerognathid tarsus has a
pair of proximal elements—an astragalus and a calcaneum.

The astragalus is a robust, roughly ovoid bone with a rounded face for
articulation with the tibia; laterally it has an elongated facet articulating with
the calcaneum and distally a curved facet facing the centrale.

The calcaneum is a larger but more lightly built bone. It is a sheet of bone,
thickened proximally to form an elongated facet for the reception of the fibula

4

136 ANNALS OF THE SOUTH AFRICAN MUSEUM

and with a thickened medial edge abutting against the astragalus, but notched
for the passage of the penetrating tarsal vessels, and thickened distally where it
carries a facet to receive the fourth distal carpal. There is no indication of a
tuber-like heel.

One fairly small centrale is present. There are five distal tarsals, but in
one specimen the small fifth distal is fused to the large fourth distal as is usual
in therapsids.

The fourth distal is always large; the third and second distals, when well
developed, have flattened upper faces and are roughly squarish in outline; the
first distal is usually large, articulating with both first and second metatarsals,
but in one specimen it is a small, pebble-like bone.

The first metatarsal is a short, fairly squat bone, the second, also short, has
a more constricted shaft; from the third to the fifth the metatarsals become
progressively longer with well-expanded ends.

The digits are, as in the manus, short, with the phalangeal formula 2, 3, 3,
3, 3. The first phalanx in all the digits is usually fairly long, but in one specimen
that of the first digit and in another that of the fifth digit is greatly shortened,
being disc-like without a constricted waist.

The second phalanx of the second to the fifth digits is short with a con-
stricted waist.

The terminal phalanges are long, curving, claw-like. The first digit is
short and the other four of about equal length. The purchase of the hind foot
thus lies in the postaxial part of the foot.

SYSTEMATIC DESCRIPTIONS OF SPECIMENS

A. PECTORAL GIRDLE
Alopecognathus (figs. 8-10)

The specimen 12051 includes a scapular blade; K339 has a nearly complete
scapula; Ke23A and K223B each consists of a fairly complete girdle, but both
lack the scapular blades and the upper part of the clavicles and have undergone
distortion and displacement of the constituent elements; K231 includes anearly
complete girdle, but is distorted by dorso-ventral compression. In the accom-
panying figures the distortion has been corrected and the symmetry restored.
In none of these is the cleithrum preserved.

The girdle is a structure of considerable size; high, broad, with its ventral
element—the interclavicle—long. The scapula, as restored, is dorsally directed
or is tilted slightly backwards and curved slightly to follow the contour of the
thorax. The area of origin of the scapular head of the triceps is indistinct—
there is no ridge, mound or tubercle. The coracoidal plate is long, but in two
specimens the anterior edge of the procoracoid is concave, whereas in the other
it forms an even convex curve. The clavicles have a long, posteriorly directed
ventral spatulate end in two specimens, whereas in another this part of the
clavicle is quite short. The interclavicle is large, with a rounded anterior

‘yernel “9 ‘“[eajUOA “Y ‘o[paIs ]e10}00g
“EX ISS CIN W'S ‘Sdaousnsup snyywusosedopp “Ol *O1q

‘MOIA [BIJUDA UL J[PsIsS [esr0ojog “fx ‘qSsay
WYV'S ¢

[e197] “9 "[eayUdA ‘Dp °

sgaoysnsup snyzousosgojy °6 ‘og "VE2SNI “INV

Q]PIIS [e10JI9g “$x
"S ¢ sdaaysnsup snyjousosaqojy

"9 “Oly

=

a
hi ok
.
.

Secagneee--="

4
4
1
4
s

138 ANNALS OF THE SOUTH AFRICAN MUSEUM

expansion and a long broad posterior spatula joined by a fairly narrow neck
at the level of the posterior coracoids.

Priesterognathoides (figs. 11 and 12a)

In 12102 the right half of the pectoral girdle is preserved together with an
incomplete interclavicle and parts of the left half of the girdle. The girdle is
partially disarticulated but only slightly distorted in its fall backwards when the
soft tissues decomposed.

In 12204 a disarticulated girdle has a good scapula, coracoid, procoracoid,
interclavicle and the ventral spatulate ends of the clavicles.

The scapula is apparently tilted somewhat backwards and shows a marked
curvature around the thorax. No origin of the triceps can be determined and
there is definitely no tubercle. The coracoid is short, but the procoracoid well
developed. The interclavicle is very similar to that of Alopecognathus.

Pristerognathus (fig. 12b and c)

In 4335 there are a pair of incomplete scapulo-coracoids. The coracoid
is short and the procoracoid well developed. No scar or tubercle for the scapular

Fic. 12. Pectoral girdles. 4. a, Pristerognathoides

Fic. 11. Pristerognathoides sp. S.A.M. sp. S.A.M. 12204 in lateral view. b, Pristerognathus

12102.x4. Pectoral girdle in sp. S.A.M. 4335 in lateral view. c, Pristerognathus sp.
lateral view. S.A.M. 11458 in ventral view.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 139

head of the triceps can be traced. In 11458 there is a good interclavicle which has
a large anterior expansion, but the posterior spatula is only slightly expanded.
The ventral spatulate ends of the clavicles extend well posteriorly. The pro-
coracoid extends anteriorly of the interclavicle.

Scymnosaurus (figs. 13-17)

Five specimens have parts of the pectoral girdle preserved. 8034 and 12193
have a well-preserved coracoid each. Both are massive, and that of 12193 is
probably of S. major. In the type specimen of S. major (g005) there are the
anterior two-thirds of a large interclavicle and the ventral spatulate ends of
both clavicies, which, though massive, have relatively a short posterior extent.
In 11557 there are a disarticulated scapula, interclavicle and clavicle. The
scapula is robust but short; the glenoid facet faces much posteriorly and
laterally; the facet for the cleithrum is clearly shown, but the origin of the
scapula head of the triceps is from the smooth postero-lateral surface just
above the glenoid.

A left clavicle is stoutly built; the dorsal end clasping the anterior border
of the scapula is expanded and strong. The ventral spatulate end is large and
extends far posteriorly along the circular head of the interclavicle. Both the
articulating surface of the clavicle as well as the hollowed face on the inter-

Fic. 13. Bones of pectoral girdles.
x 4. a, Scymnosaurus major. S.A.M.
go05. Type. Ventral view of inter-
clavicle and clavicle. b, Scymnosaurus
ferox. S.A.M. 9084. Right coracoid
in ventral view. c, Scymnosaurus sp.
S.A.M. 12193. Right coracoid in

ventral view.

I40 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fic. 14. Right scapula of Scymnosaurus sp. S.A.M. 11557. 4. 4, posterior. b, lateral. c, anterior.
- d, internal.

Fic. 15. Scymnosaurus sp. S.A.M.
11557. X 4. Clavicle and inter-
clavicle. a, ventral. b, lateral.

141

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA

‘o[P4Is [e10}00d
JO MOIA Jero}eT “FX “G6gi1 ‘Wy's ‘ds snunvsoums -L1 ‘org

‘O[psis [e10;90d
JO MOIA TeUZA “fx “S6QII “Wy's ‘ds snunvsoumids +91

OLY

142 ANNALS OF THE SOUTH AFRICAN MUSEUM

clavicle bear longitudinal striae, and the backward movement of the clavicle is
limited by a thickened border on the clavicle which abuts against the anterior
edge of the interclavicle. The dorsal stem of the clavicle and its ventral spatulate
end subtend what is nearly a right angle.

The anterior end of the interclavicle is large and nearly circular in outline,
and a strong median ridge separates the two facets for the clavicles. The posterior
end is greatly expanded but short, and is remarkably lightly built.

In 11695 there is a fairly complete shoulder girdle, but with the constituent
bones displaced and distorted by dorso-ventral pressure. The anterior head of
the interclavicle and ventral spatula of the clavicle are both relatively smaller
than in 11557, and the dorsal stem of the clavicle is longer and more slender,
with a bifurcated end. The scapular blade is fairly tall but is relatively lightly
built.

Theriowdes (fig. 18)

In 11888, the holotype specimen, much of the pectoral girdle is preserved,
but here again disarticulation and distortion have made reassembly and
restoration of symmetry, as shown in the figures, difficult. The girdle is very like

Fic. 19. Zinnosaurus pau-
cidens gen. et spec. nov.
S.A.M. 12185. Type.
Fic. 18. Thertoides cyniscus. Pectoral girdle. S.A.M. x4. Lateral view of
11888. Type. x 4. a, ventral. 5, lateral. scapulo-coracoid.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 143

that in Alopecognathus, but smaller and of a lighter build, with a proportionately
large anterior expansion of the interclavicle.

Ainnosaurus paucidens gen. et spec. nov. (fig. 19)

In 12185, the holotype specimen, there is a well-preserved right scapulo-
coracoid without scapular blade, and a left scapular blade. This scapulo-
coracoid differs very markedly from all those hitherto considered. The scapular
blade has its upper anterior part expanded and on this outer face lies a facet for
the upper end of the cleithrum; lower down the cleithrum is applied to the
anterior edge of the scapula.

Immediately above the upper rim of the glenoid there lies a process on
the posterior face of the scapula for the origin of the scapular head of the triceps.
This process forms a strong, prominent flange of bone, laterally compressed
and dorso-ventrally elongated.

The glenoid is a widely open cavity facing largely outwards. Both its dorsal
scapular rim and ventral coracoidal rim are sharp and prominent. Anteriorly
there is no rim on the procoracoidal border of the glenoid so that in its anterior
movement the head of the humerus would ride in a broad groove.

The procoracoid extends far anteriorly.

B. HUMERUS
Alopecognathus (fig. 20)

In 12051 there is a good, apparently undistorted, proximal half of a
humerus. In K231 both humeri are preserved but both are quite obviously
distorted and in the figures I have tried to correct the distortion. The result
is not very convincing and the two humeri have quite distinctive outlines.

In 12051 the central part of the caput quite definitely flows over onto the
dorsal surface and this would obviously affect the nature of the shoulder joint

€

Fic. 20. Alopecognathus angusticeps. Humerus. ¥ 4. S.A.M. 12051: a, ventral. b, dorsal. S.A.M.
K231: ¢, ventral. d, dorsal.

144 ANNALS OF THE SOUTH AFRICAN MUSEUM

as the humerus would tend to be directed more horizontally. In K23r the
caput is terminal. In 12051 the delto-pectoral crest is weak with little indication
of the pectoralis insertion, and the processus medialis, caput and processus
lateralis flow into one another.

In Kg231 the distal condyles are mostly terminal with only a little extension
onto the ventral face. Both supra-trochlear fossa and brachialis fossa are
deeply excavated. The radial condyle is not developed into a rounded
capitellum.

Cynariognathus (fig. 21a)

In 11794 both humeri are imperfectly preserved. The bone is lightly
built and the distal condyles mostly terminal without a bulbous capitellum.

Pristerognathoides (figs. 21b, ¢ and d and 22)

In 5018 both humeri are incompletely preserved and in 12204 isa fairly
good left humerus. In other specimens, e.g. 12102, only poorly preserved
humeral ends are present. In 5018 the humerus is long, with both ends well
expanded and the shaft slender. The caput, mainly terminal, curves slightly onto
the dorsal surface. The distal condyles are a little ventral of terminal, and the
capitellum is moderately swollen. I have not been able to locate an ectepi-
condylar foramen. In 12204 the ulnar condyle is separated from the olecranon
fossa by a well-developed ridge which would appear to limit the extension of the
epipodial. The delto-pectoral crest is weak and the area for the insertion of the
deltoideus small.

Fic. 21. Humerus. x}. Cynariognathus sp. S.A.M. 11794. a, ventral.
Pristerognathoides sp. S.A.M. 5018, b. dorsal. c, ventral. d, posterior.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 145

Pristerognathus

A number of humeral ends of Pristerognathus are poorly preserved and
warrant no description except to state that the humerus is very similar to
that of Pristerognathoides—but smaller.

Ptomalestes (fig. 23)

In 11942, the holotype, both humeri are preserved. I am including figures
of both the right and left humerus which show how much they have suffered
from post-mortem distortion. If the distortion is corrected the humerus would
be a fairly robust bone, moderately long with well-expanded proximal and

d

Fic. 22. Humerus. Pristerognathoides sp. S.A.M. 12204. X 3. a, dorsal. 6,
ventral. c, posterior. d, S.A.M. 12102, ventral.

ae aaa

; os Ae ‘

Fic. 23. Humerus. Ptomalestes avidus. S.A.M. 11942. Type. x }. a, dorsal. b, ventral. c, posterior.
d, proximal. e, ventral of right humerus.

qd

146 ANNALS OF THE SOUTH AFRICAN MUSEUM
distal ends and a fairly short shaft; both caput and distal condyles are terminal,
with the capitellum weak and little ventrally placed. A ridge separates the
ulnar trochlea from the shallow supra-trochlear fossa.

Fic. 24. Humeri. x 4. Scymnosaurus major. S.A.M. 9005. Type.
a, dorsal. 6, ventral. Scymnosaurus ferox. S.A.M. 9084.
c, posterior. d, ventral.

Scymnosaurus (figs. 24-26)

Of this genus I have two very well preserved distal ends, and one good
complete humerus, which I believe to be undistorted, and this specimen must
be taken to give us the best idea of the pristerognathid humerus as represented
by the largest genus of this family of primitive Therocephalia.

In 9005, the type specimen of Scymnosaurus major, the good distal humeral
end is, considering the bulk of the animal, only moderately robust. The epi-

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 147

7 : 5D ~ -

Fic. 25. Scymnosaurus sp. S.A.M. 11 . <4. Humerus. a, dorsal. b, ventral. c, posterior. d, anterior.
5: 9oy 19 55 3 ! >?
é, proximal.

Fic. 26. Scymnosaurus sp. S.A.M. 11695. x4. Humerus. a, dorsal. 6, ventral.
c, anterior.

148 ANNALS OF THE SOUTH AFRICAN MUSEUM

condylar expansions are modest, with the confluent supinator flange and
ectepicondyle slightly more prominent than the entepicondylar edge. The
condyles lie chiefly terminal. The radial condyle forms only a weak capitellum,
which only slightly enters the ventral surface. The ulnar condyle forms a
shallow trochlea, which dorsally has a ridge separating it from the very shallow
supra-trochlear fossa hardly functioning as an olecranon fossa. Ventrally there
is hardly a coronoid fossa, but the large oval entepicondylar foramen opens
into a deep hollow bounded postaxially by a strong rounded ridge. The distal
epicondylar edges are strong but not much expanded and their rugose surfaces
give a strong origin for the flexors and extensors. In 9084, which is smaller and
lighter, being a humerus of the less bulky Scymnosaurus ferox, the ectepicondylar
flange is more expanded and curves downwards to form a deep groove. In
11557, which is most probably also of Scymnosaurus ferox, there is a well-preserved
left humerus. Its terminal distal condylar face is less rounded than in 9084 and
has no capitellum to speak of, and in its postaxial part it is concave instead of
convex, indicating the presence of considerable joint cartilege. Its ectocondyle
is without the flange present in 9084. Its supratrochlear fossa is deeper than in
9084 with a strong preaxial border.

11557 is the only Scymnosaurus humerus with a good proximal two-thirds
preserved. Ventrally the bicipital fossa is deep and well demarcated from the
surface of the short shaft. The delto-pectoral crest is fairly strong but the
pectoralis insertion is not localized but diffuse. The postaxial border of the
bicipital fossa is formed by a robust rounded ridge which on its postaxial
face gives a strong face for the origin of the medial humeral head of the triceps.
Dorsally both the anterior dorso-ventral line and latero-medial line are
moderately developed and the areas of insertion of the deltoideus, latissimus
dorsi and both the scapulo-humeralis anterior and posterior are well
developed.

Proximally the caput is terminal with only a slight extension onto the
dorsal surface; it is a very flat oval and flows evenly into both the processus
medialis and lateralis.

The incomplete humerus of 11695 shows a greater twist on the shaft but is
otherwise very similar to that of 11557.

Therwoides (fig. 27)
The one humerus known in the type, 11888, is fairly short with moderately

expanded ends but otherwise shows no special features to distinguish it from
those of the other medium-sized pristerognathids so far described.

Almnosaurus gen. et spec. nov. (fig. 28)

In the type material (12185) there are a practically complete right
humerus and the proximal half of the left humerus. Both have suffered from
distortion. I am including figures of the right humerus with the distortion
corrected.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 149

The caput is strap-shaped but in its middle part tends to overflow slightly
onto the dorsal face. The processus medialis and processus lateralis flow gently
into the caput. The delto-pectoral crest is long, with its proximal edge thin and
its distal end somewhat thickened for the reception of the pectoralis.

cn “3

Fic. 27. Theroides cyniscus. S.A.M. 11888. Type. x 4. Humerus.
a, dorsal. b, ventral. c, posterior.

= | ; Lo :

Fic. 28. Zinnosaurus paucidens gen. et spec. nov. S.A.M. 12185. x4. Right
humerus with distortion corrected. a, dorsal. b, ventral. c, posterior. d, anterior.

The distal end is greatly expanded. This expansion is mainly due to the
development in both epicondyles of thin sheets of bone. This is particularly so
in the ectepicondyle, where the confluent supinator process and the epicondyle
form an extensive antero-ventrally sweeping sheet of bone.

The distal condyles are terminal with very little capitellar development,
but the ulnar condyle is quite well developed with a dorsal ridge separating it
from the very shallow supra-trochlear depression.

150 ANNALS OF THE SOUTH AFRICAN MUSEUM

Cc. THE LOWER FORELIMB AND FOREFOOT

Alopecognathus (fig. 29)

In 1205! proximal and distal ends of both ulnae and radu and one good
carpus are preserved.

Proximally the head of the radius fits into the concave sigmoid rim of the
ulna so that its facet lies nearly as far proximally as that of the sigmoid facet of
the ulna which is also situated termina!ly. The olecranon is robust but does not
extend proximally as a process.

Fic. 29. Alopecogna- Fic. 30. Cynariogna-
thus angusticeps. this sp. S.A.M.
S.A.M. 12051. x#. 11794. +4. Right
Epipodial and epipodial and partial
carpus. manus.

In the carpus the radiale is a stout bone roughly rectangular in outline.
A longitudinal ridge separates a larger preaxial dorsal face from a smaller
postaxial face. Distally a ridge separates the facets for the two centrals.

The intermedium is a laterally flattened small bone wedged in between
radiale and ulna and ulnare.

The ulnare is a long element with expanded ends and a long constricted
waist. Its proximal facet is much smaller than the distal ulnar facet.

Both centralia are well-developed bones, each articulating with a pair of
distals.

There are four distals—the fused fourth and fifth articulating with the
fourth and fifth metacarpals.

Only the proximal ends of the five metacarpals are preserved.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA I51

Cynariognathus (fig. 30)

In 11794 both anterior epipodials are preserved but the preservation is
not very good. Both the ulna and radius are lightly built and the ulna has no
olecranon process. In the figure the proximal ends of the two bones are shown as
lying next to each other but in life the proximal head of the radius is applied
to the edge of the sigmoid cavity of the ulna so that the two bones have their
proximal facets forming a confluent articulating face. The radial facet meets the
capitellum and in extension the ulna rides on the trochlear facet of the humerus.

Fic. 32. Ptomalestes

avidus. S.A.M.
Fic. 31. Pristerognathoides ? S.A.M. K357. 11942. Type. X 4.
Nat. size. Partial manus. Epipodial.

Only the right manus is in part preserved—the digits are missing.

In the carpus 9 bones are preserved. The proximal row consists of a long
slender ulnare shaped somewhat like a metacarpal, a large laterally compressed
intermedium and a short, broad radiale with distally two well-developed
concave facets for the two centralia.

The two centrals are well-developed bones of irregular shape each with a
good facet for articulation with the radiale.

There are four distalia; the first three articulate with the first three
metacarpals, whereas the fourth articulates with both the fourth and fifth
metacarpals.

152 ANNALS OF THE SOUTH AFRICAN MUSEUM

Of the five metacarpals the last three have only their proximal ends
preserved. They are short, with well-expanded ends. No. 1 is small and light,
and from the second to the fifth they increase rapidly in size so that No. 5 is
quite a strong element.

Pristerognathoides (fig. 31)

In a number of specimens ulnae and radu are preserved, exceptionally
complete but mostly represented by proximal and distal ends. They call for
little comment, except that in all these there is no definite olecranon process,
and that proximally the two facets form a confluent terminal articulating face
for the humeral condyles.

‘
'

Ae

Dee

Fic. 33. Scymnosaurus major x 4. mae 34. Scymnosaurus sp. S.A.M. 11557.

Epipodial of S.A.M. goos. x 4. a, radius and manus in dorsal view.

Type. Carpus of S.A.M. 12193. b, lateral view of no. IV _ terminal
phalanx.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 153

In K357 there is a good carpus with all five metacarpals present. Here the
intermedium is a weak element and articulating with the ulna and ulnare there
is a well-developed pebble-like pisiforme. The ulnare is elongate, the radiale
robust, and the two centrals well developed. There are four distals, with the
fourth articulating with both the fourth and fifth metacarpal.

The first two metacarpals are short, squat bones and the other three much
longer, with the fifth the longest. The proximal phalanx of the fifth digit has
its proximal end greatly expanded.

Ptomalestes (fig. 32)

In the type specimen of Ptomalestes (11942) both anterior epipodials are
preserved. Both the radius and ulna are lightly built and the ulna has no
distinct olecranon process. In the figure the proximal head of the radius is
shown moved out of its articulation with the proximal end of the ulna.

Scymnosaurus (figs. 33-35)

In this genus the epipodial, carpus and manus are represented by some
very good specimens.

The epipodial in the type specimen of Scymnosaurus major (9005) is repre-
sented by a well-preserved left ulna and radius (a mirror image is shown in the
figure). In 9084 there are good proximal ends of the right ulna and radius in
articulation. In 11557 there is a good radius, in 12193 proximal and distal
ends of both ulna and radius and in 12262 a good complete radius and ulna.

In all these specimens the olecranon process of the ulna extends little
proximally of the sigmoid face, and the proximal head of the radius lies anteriorly
(morphologically dorsally) of the proximal end of the ulna and fits securely
against the sigmoid edge forming a confluent articulating face for the feeble
capitellum and the shallow trochlea of the humerus.

The carpus is preserved in part or nearly complete in 12193, 11459, 11557
and in 12262.

In 12193 the radiale is a robust bone; in dorsal view oval in outline with a
convex dorsal and proximal face; concave distally for the first centrale and
with a convex facet for the second centrale; ventrally it has a deep oblique
groove.

The ulnare has its dorsal face with an outline like that of a metacarpal.
Ventrally it has a longitudinal groove and its preaxial edge is convex incontrast
to the dorsal concave preaxial edge.

The two centrals are robust; the first is nearly circular in outline with a
penetrating foramen near its preaxial edge. The proximal central has its
postaxial edge deeply concave, facing the concavity of the ulnare, thus forming
a passage for the vessels penetrating the carpus.

Of the distal carpals only the proximal parts are preserved with the fourth
distal strongly developed. |

In 11459 the two centrals, distals together with metacarpals 1, 2 and 3,

154 ANNALS OF THE SOUTH AFRICAN MUSEUM

and the proximal phalanx of the second digit are preserved. The first central
has a groove and a penetrating foramen near its preaxial border. The meta-
carpals increase in length from 1 to 3.

In 11557 there is a nearly complete carpus, but the manus is incomplete
with the preserved bones partly displaced and disarticulated. In the figure

SOY

iis

Gy

Fic. 36. Theriodes
cyniscus. S.A.M.
V1Ioss. Lype: <4:
Dorsal view of epi-
podial and partial
manus.

Fic. 35. Scymnosaurus sp.

S.A.M. 12262. x 4. Epipo-

dial and manus in dorsal
view.

the constituent bones are shown in natural relation. The radiale is fairly robust,
with a flattish face for the radius; its dorsal surface shows two faces—the
preaxial face nearly rectangular and the postaxial face more squarish, meeting
at an angle of about 130°; the facet for the first central is slightly concave and
for the proximal central convex.

The ulnare is elongated, with a dorsal outline resembling that of a
metacarpal.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 155

The intermedium is a small, laterally compressed bone. The first central
has a depression near its preaxial border and the proximal central is hollowed
out near its proximal wedge-like end.

The first distale is missing, numbers 2 and 3 lie proximally to their respec-
tive metacarpals and the large fourth articulates with both the fourth and fifth
metacarpals.

The first metacarpal is missing; from number 2 to 5 the metacarpals
increase in length with the fifth about 14 times the length of number 2.

The phalangeal formula is 2, 3, 3, 3, 3, 3. The ungual phalanges are
curved, narrow, pointed and claw-like. The others are short, broad and squat,
with well-expanded ends.

In 12262 the forefoot is very well preserved and is complete except that
part of the radiale and the tip of the ungual phalanx of the second digit have
been lost in transit from the field. The general structure is as in all the pristero-
enathids described above, viz. carpal formula 3, 2, 4 plus an ulnar pisiforme, and
phalangeal formula 2, 3, 3, 3, 3. The fourth and fifth metacarpals articulate
with the large fourth distal carpal; the metacarpals increase in length in post-
axial direction and the distal ends of the third and fourth are particularly well
modelled, with a dorsal trochlear fossa and well-rounded distal corners. The
third and fourth digits are the longest and strongest; the first digit is relatively
weak and the fifth strong. The weight is thus carried somewhat more post-
axially than preaxially.

Therioides (fig. 36)

In the type specimen (11888) much of the right forefoot is preserved. The
carpus is complete; the last four metacarpals are present but only the second
digit is completely preserved.

The structure is typically pristerognathid but the metacarpals are relatively
long and slender.

D. PELVIC GIRDLE
Alopecognathus (fig. 37)

In K218 a fairly large pelvis is preserved. This is, however, badly weathered
and all that can be determined is that the pelvic plate is large with a distinct
ischial keel and that the outer face of the ilium is convex.

In K231 most of a fairly large pelvis is preserved, but with the edges of
the ilia incomplete. The pelvic plate is long and broad with both ischium and
pubis strong; just more than half is contributed by the ischium; the ilium is
large, with its length nearly as long as the ventral plate (91%). The pelvis is
low and long, with the height 79 per cent of the length.

The ilium is not completely preserved on either side, but the two sides are
complementary so that the figure could be compiled fairly accurately. The
antero-posterior length of the iliac blade is great but its supra-acetabular height
moderate; the height is a little more than half the length (52%). The posterior

156 ANNALS OF THE SOUTH AFRICAN MUSEUM

process of the iliac blade is long but fairly low, whereas the anterior process
is fairly short but high; although its ventral corner apparently has a long, low
elongation.

The outer face of the iliac blade is in general convex both antero-posteriorly
as well as dorso-ventrally. The m. ilio femoralis (gluteus) thus originated from
a convex area whereas in tetrapods this area is usually concave. The supra-
acetabular ridge is strong. The ilium forms more than half of the acetabulum.

4

4
/

\ ,
lees I,

/
sea CO
~ ™ TL p

Fic. 38. Lateral views of pelves.
x 4. a, Cynariognathus sp. S.A.M.

Fic. 37. Pelvis of Alopecognathus angusticeps. 11794. b, Pristerognathoides ? S.A.M.
S.A.M. Kag1. x4. a, lateral view. 5018. c, Pristerognathus? S.A.M.
b, ventral view. Ke227.

The pubis is massive in its acetabular part, with a strong dorso-lateral
edge, terminating in a moderate pubic tuber somewhat outwardly directed.
The two pubes form a weak symphysis. A fairly large foramen pierces the bone
near its posterior edge just medial to the strong acetabular rim. The ischium is
longer than the pubis: its acetabular part is strongly developed and its rounded
dorso-lateral edge ends in a moderate tuber. The two ischia form a strong
symphysis which develops a strong medial keel.

The acetabulum is large and faces mainly outwards.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 157

Cynariognathus (fig. 38a)

In 11794 there are preserved, of the right half of the pelvis, the acetabular
proximal parts of the three constituent bones, whereas on the left there are a
fair ischium and the proximal part of the pubis. The iliac blade is not preserved.
Although of lighter build, the pelvis is essentially as in Alopecognathus. The
pubic foramen is large.

Pristerognathoides (figs. 38b and 39)

In 5018 the two ilia are partly preserved and the figure incorporates
features of both sides. The ilium is very similar to that of Alopecognathus, however
its outer face, though also convex, is less so than in Alopecognathus.

Fic. 39. Pelvis of Pristerognathoides sp. S.A.M. K223C. x 4. a, lateral. b, ventral.

In K233C the pelvis lacks only the posterior ends of the ischia. The pelvis
is fairly low but quite long, the height being 75 per cent of the length as recon-
structed. The blade of the ilium is both low and very short, the height being
74 per cent of the length. The outer face of the iliac blade is peculiar. It is not
hollowed out as is usual in all other therapsids, but cannot be described as
convex either. Dorso-anteriorly there is a pronounced bulge separating two
hollows so that the surface can be described as undulating. The anterior process
is short but high, with its anterior edge cut back at the level of the anteriorly
hollowed-out area.

The posterior process is also short but much lower than the anterior
process.

The supra-acetabular ledge is strong.

Both ischia and pubes are robust in their acetabular parts, with strong
circum-acetabular borders. Anteriorly the pubis has a broad, thin anterior
edge with the outer corner showing little of a pubic tuber. The pubic foramen
is large but the symphysis is weak. The ischia form a fairly strong ventral keel.
In its pubic part the pelvis has a flat floor, which in its ischial part becomes
shallowly V-shaped.

158 ANNALS OF THE SOUTH AFRICAN MUSEUM

Pristerognathus (fig. 38c)

In K227 there is a pair of beautifully preserved ilia. Much smaller than
the ilia in the forms so far described, these ilia are nevertheless very similar.
The undulating outer face is clearly shown; the anterior process is high, with its
anterior edge notched; the posterior process is relatively longer and lower.
The supra-acetabular height of the ilium is about two-thirds of the length of the
iliac blade.

In K238A there are a partial ilium and most of a pair of ischia.

Scymnosaurus

Of this genus only the acetabular parts of an ilium and a pubis which
indicate a moderately robust pelvis with an acetabulum of moderate size
directed outwards are preserved (11957).

In 11558A a pair of good ischia are preserved.

d b

Fic. 40. Pelvis of a pristero-
gnathid. S.A.M. Ke45A.
x4. a, lateral. 5, ventral.

a | ra d e

Fic. 41. Femora. x }. Alopecognathus sp. S.A.M. 12051. a, dorsal. b, ventral. c, anterior.
Cynariognathus sp. S.A.M. 11794. d, dorsal. e, ventral. f, anterior.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 159

An unidentified pristerognathid (fig. 40)

In K245A there is a good pelvis, basically as in the above described forms,
but the ilium is rather peculiar, with a long, low process directed anteriorly.

E. FEMUR
Alopecognathus (fig. 41a, b and c)

In 12051 a proximal and two distal femoral ends are preserved. The ends
are moderately expanded, shaft fairly strong, somewhat dorso-ventrally
flattened, and fairly straight.

The well-rounded caput, although terminal, is directed somewhat pre-
axially. The external trochanter is not prominent and is situated well proximally.
The pubo-ischio femoralis internus ridge is well developed, not near the anterior
border. The internal trochanter forms a sharp ridge. The inter-trochanteric
fossa is small and shallow.

In Ke31 both femora are present and, although badly preserved, are
essentially similar to that figured here.

Cynariognathus (fig. 41d, e and f)

In 11794 both femora are preserved, but are both somewhat weathered and
distorted. The curvatures in the figures on the long axis are thus artificial.

The ridge forming the internal trochanter is long and not so far proximally
situated.

Soe

4
i
'
'
'
1
!
‘
'
'
'
i)
‘
|

~ bee ~ < ~ es —

Fic. 42. Femora. x 4. Pristerognathoides sp. S.A.M. 5018. a, dorsal. b, ventral. c, anterior.
S.A.M. 11936. d, dorsal. e, ventral. f, anterior.

160 ANNALS OF THE SOUTH AFRICAN MUSEUM

Pristerognathoides (figs. 42 and 43)

The femur in this genus is represented by the distal and proximal ends of
both femora in 5018 and K306 and one femur in 11936.

In 5018 the internal trochanter forms a long ridge extending on to the
shaft, but it does not extend far proximally, so that the intertrochanteric fossa
is long. The pubo-ischio femoralis internus ridge is low. The distal end is well
expanded. In 11936 the femur is more robust. |

In K306 the right femur is a larger, heavier bone than the left and is
certainly not of the same individual and may be of a different species. The distal
ends are only moderately expanded.

;

Fic. 43. Femora. x 4. Pristerognathoides sp. S.A.M. K306. Right: a, dorsal.
c, ventral. e, anterior. Left: b, dorsal. d, ventral. f, anterior.

¢

(€

Pristerognathus

In this genus the femur is represented by poor ends. In K317 the internal
trochanter forms a short, thickened ridge proximal to which lies a long, deep
inter-trochanteric fossa.

Scymnosaurus (figs. 44, 45, 46a, b and c, 47)

In this genus the femur is fairly well represented. In 11557 there is a good,
well-preserved femur, in 11597 a proximal end, in 12118 a proximal end, in
12193 a distal and proximal end, and in 12262 and K352 a poorly preserved
but nearly complete femur.

In the three complete femora there is a greater amount of twist or rotation
of the two ends on the shaft relative to each other than in the femur of all the
other pristerognathids as hitherto described. This twist or rotation is an anti-

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 161

clockwise or postaxial movement of the proximal expansion in relation to the
distal end. The result is that the caput, though terminal, is directed somewhat
dorsally, whereas in the hitherto described forms this direction runs somewhat
anteriorly.

Another difference is that the pubo-ischio femoralis internus ridge in
Scymnosaurus lies on the morphological anterior edge of the bone and not some-

Fic. 44. Left femur of Scymnosaurus sp. S.A.M. 11557. < 4. a, dorsal. 6, ventral (projection onto
plane of distal condyles). c, ventral (projection onto the plane of the ventral face of the proximal
end). d, anterior. e, proximal.

what away from this edge onto the dorsal surface. The straight, long axis of
11557 is normal whereas the downward curvature of the proximal end in both
12262 and K352 would appear to be due to post-mortem distortion. The internal
and external trochanters are as in the already described Pristerognathids.

K353 consists of an isolated, well-preserved proximal end of a large femur
which may be that of Scymnosaurus major. Although it has an internal trochanter
and pubo-ischio femoralis internus ridge as in the other pristerognathid femora,
both these structures are feebly developed and appear to be out of proportion
to the size of the bone.

Ainnosaurus (fig. 46d and e)

In 12185 a proximal and distal femoral end is preserved. The caput is well
rounded and somewhat anteriorly directed. The internal trochanter forms a

162 ANNALS OF THE SOUTH AFRICAN MUSEUM

robust ridge. The external trochanter is weak and the inter-trochanteric fossa
short but deep.

The pubo-ischio femoralis internus ridge is very weak; it forms the anterior
border of the bone and runs into the anterior edge of the caput.

F. THE LOWER HIND LIMB AND PES
Alopecognathus (figs. 48 and 49a)

In 12051 the left epipodial and pes are completely preserved and in K231
both epipodials and an incomplete pes.

C. ee i

Fic. 45. Femora. x 4. Scymnosaurus sp. S.A.M. 11957. a, ventral. Scymnosaurus sp.
S.A.M. 12118. 6, ventral. Scymnosaurus major. S.A.M. 12262. c, dorsal. d, ventral.
e, anterior.

The tibia is stout proximally with a well-developed facet for the femur.
The fibula is much slenderer and curved to give a wide spatium interosseum.
The proximal facet of the fibula is terminal and not applied to the outer corner
of the femoral condyle but end-on.

The astragalus is a robust, rounded bone with a convex facet for the tibia.

The calcaneum is a large but lightly-built flat bone thickened proximally
at the facet for the fibula. Preaxially it is notched for the passage of the intertarsal
vessels.

The central is quite small.

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 163

There are five distals. The first is large and articulates with both first and
second metatarsals. The fourth is the largest whereas numbers 2, 3 and 5 are
small.

The metatarsals increase in length from number 1 to 5.

The digital formula is 2, 3, 3, 3, 3, with the fourth the longest, but due
to the short phalanges all the digits are short. The ungual phalanges are curved
and carried sharp claws.

Fic. 46. Femora. x4. Scymnosaurus major ? S.A.M. K352. a, dorsal.

b, ventral. c, anterior. Zinnosaurus paucidens gen. et spec. nov. S.A.M.

12185. d, ventral surface of proximal end. e, anterior face of proximal
end.

i ‘ 8 ‘ a i : oo

Fic. 47. Scymnosaurus major ? S.A.M. K353. x}. Proximal end of femur. a, dorsal.
b, ventral. c, anterior. d, posterior.

164 ANNALS OF THE SOUTH AFRICAN MUSEUM

Cynariognathus (fig. 49b)

In 11794 a somewhat crushed tibia and fibula are preserved, with features
as shown in the figure. Distal to the tibia lies a bony element of quite intricate
shape, not at all like the astragalus preserved in Alopecognathus and probably
is not this bone at all.

Fic. 48. Epipodial

and pes of Alopecog- Fic. 49. Epipodials and pes. x 4. a, Alopecognathus Fic. 50. Pes of a

nathus. S.A.M. angusticeps. S.A.M. K231. b, Cynariognathus sp. pristerognathid.
12051. Xt. S.A.M. 11794. c, Scymnosaurus sp. S.A.M. 11459. S.A.M. 245A. x4.

3
Scymnosaurus (fig. 4.9¢)

In 11459 part of a pes is preserved. I have identified a small central.
Distals 2 and 3 are small, 4 is much larger, and 5 is still a small, separate
element. Parts of all five metatarsals are present. Number 1 digit has a very
short first phalanx.

DIscussION

It would be of interest to compare the girdles and limbs of the pristero-
gnathid Therocephalia with those of the other therapsids of the Tapznocephalus
zone. Unfortunately these structures in the contemporary therapsids are
inadequately known. Only in the Dinocephalia have they been adequately
described, and in the Dromasauria two specimens have these structures

GIRDLES AND LIMBS OF THE PRISTEROGNATHID THEROCEPHALIA 165

preserved. Of the Gorgonopsia, they have been described in one genus. Of the
Dicynodontia nothing has as yet been published. Of the other therocephalian
families represented in this zone, a shoulder girdle of a lycosuchid has been
figured, of the Akidnognathidae a manus, and of the Scaloposauridae a brief
account has been given of parts of a hind limb.

I have in recent years collected some gorgonopsians and dicynodonts from
the Tapinocephalus zone in which parts of the girdles and limbs are preserved.

As soon as these specimens have been prepared a comparative discussion
will be presented.

SUMMARY

Descriptions are given of the girdles and limbs of the pristerognathid
Therocephalia from the Tapinocephalus zone of the Karoo. The descriptions are
based on 37 specimens in the South African Museum which have parts of the
girdles and limbs preserved. Taken together this material gives a good idea of
the family character of these structures. A new genus and species of pristero-
enathid therocephalian, Zinnosaurus paucidens, is described.

ACKNOWLEDGEMENTS

The specimens collected in recent years were obtained on Museum
collecting trips, which were in part financed by the C.S.I.R. For these grants
we are grateful.

Some of the later specimens have been prepared for study by Mr. C. Gow,
to whom our thanks are due.

The Trustees of the South African Museum are grateful to the C.S.I.R.
for a grant to publish this paper.

REFERENCES

Broom, R. 1929. On some recent new light on the origin of mammals. Proc. Linn. Soc. N.S.W.
54, 8: 688-694.

Broom, R. 1932. The mammal-like reptiles of South Africa. London: H. F. & G. Witherby.

Huene, F. vor. 1956. Paldontologie und Phylogenie der niederen Tetrapoden. Jena: Fischer.

Sre.ey, H. G. 1888. Researches on the structure, organization and classification of the fossil
Reptilia. III. On parts of the skeleton of a mammal from Triassic rocks of Klipfontein,
Fraserberg, South Africa (Theriodesmus phylarchus, Seeley), illustrating the reptilian inheri-
tance in the mammalian hand. Phil. Trans. R. S. (B) 179: 141-155.

Ann. S. Afr. Mus., Vol. XLVIII Plate T

A. Photograph of a habitat group with life-sized models modelled by the author and exhibited
in a diorama in the South African Museum. Scymnosaurus feeding on Brachypareia.

B. Photograph of life-sized models made by the author and exhibited as a habitat group in the
South African Museum. The Therocephalians (1. to r.) Pristerognathoides, Therioides and Alope-
cognathus feeding on a cadaver of Moschops.

INSTRUCTIONS TO AUTHORS

MANUSCRIPTS

In duplicate (one set of illustrations), type-written, double spaced with good margins,
including TABLE OF CONTENTs and Summary. Position of text-figures and tables must be
indicated.

ILLUSTRATIONS

So proportioned that when reduced they will occupy not more than 43 in. X 7 in. (74 in.
including the caption). A scale (metric) must appear with all photographs.

REFERENCES

Authors’ names and dates of publication given in text; full references at end of paper in
alphabetical order of authors’ names (Harvard system). References at end of paper must be
given in this order:

Name of author, in capitals, followed by initials; names of joint authors connected by &,
not ‘and’. Year of publication; several papers by the same author in one year designated by
suffixes a, b, etc. Full title of paper; initial capital letters only for first word and for proper
names (except in German). Title of journal, abbreviated according to World list of scientific
periodicals and underlined (italics). Series number, if any, in parenthesis, e.g. (3), (n.s.), (B).
Volume number in arabic numerals (without prefix ‘vol.’), with wavy underlining (black type).
Part number, only if separate parts of one volume are independently numbered. Page numbers,
first and last, preceded by a colon (without prefix ‘p’). Thus:

Situ, A. B. 1956. New Plonia species from South Africa. Ann. Mag. nat. Hist. (12) 9: 937-945.

When reference is made to a separate book, give in this order: Author’s name; his initials;
date of publication; title, underlined; edition, if any; volume number, if any, in arabic numerals,
with wavy underlining; place of publication; name of publisher. Thus:

Brown, X. Y. 1953. Marine faunas. 2nd ed. 2. London: Green.

When reference is made to a paper forming a distinct part of another book, give: Name of
author of paper, his initials; date of publication; title of paper; ‘In’, underlined; name of
author of book; his initials; title of book, underlined; edition, if any; volume number, if any,
in arabic numerals, with wavy underlining; pagination of paper; place of publication; name
of publisher. Thus:

SmitH, C. D. 1954. South African plonias. Jn Brown, X. Y. Marine faunas. 2nd ed. 3: 63-95.
London: Green.

SYNONYMY

Arranged according to chronology of names. Published scientific names by which a species
has been previously designated (subsequent to 1758) are listed in chronological order, with
abbreviated bibliographic references to descriptions or citations following in chronological
order after each name. Full references must be given at the end of the paper. Articles and
recommendations of the International code of zoological nomenclature adopted by the XV International
congress of zoology, London, July 1958, are to be observed (particularly articles 22 and 51).
Examples: Plonia capensis Smith, 1954: 86, pl. 27, fig. 3. Green, 1955: 23, fig. 2.

When transferred to another genus:

Euplonia capensis (Smith) Brown, 1955: 259.
When misidentified as another species:
Plonia natalensis (non West), Jones, 1956: 18.

When another species has been called by the same name:

[non] Plonia capensis: Jones, 1957: 27 (= natalensis West).

TTT
3 9088 01206 5850