ON THE EPIPTERYGOID—ALISPHENOID TRANSITION

IN THERAPSIDA

T. H. Barry

South African Museum, Cape Town

(With 22 figures in the text)

CONTENTS

PAGE

Introduction. : : - 399

Material . : : : 1 401

The palatoquadrate of Pristerodon . 402

Phylogeny . : : . 405

Placodermi : : : . 405

Crossopterygii . : : . 406

Labyrinthodontia : : AO

Cotylosauria.. : : . 408

Pelycosauria ‘ : ; 410

Therapsida ‘ : 3 Ata

Triconodonta . A : . 418

Living reptiles and mammals - 419

Conclusions ; 4 . - 421

Acknowledgements. 6 AZ 3

Summary . : : , 5 py

References : 3 5 - 424

INTRODUCTION

Mainly as a result of the researches of Gaupp (1902), Allis (1919), Broom

(1907, 1909 and 1914), Fuchs (1912) and Gregory and Noble (1924) it is

now generally accepted that the reptilian epipterygoid and mammalian

alisphenoid, although superficially different in many respects, are homologous.

structures. The views of these workers have received support from embryologists,

comparative anatomists and palaeontologists alike over the last few decades

and the theory has now virtually become axiomatic.

In modern reptiles there is a marked degree of variation in the development

of the epipterygoid, ranging from the primitive, rather solid, construction of

the bone in Sphenodon, to the absence or vestigial development of the bone in

adult Ophidia and Crocodilia with the Lacertilia displaying a thin columnar

structure. The relationship of the branches of the trigeminal nerve to the

bone, remains constant throughout the class, the profundus emerging anterior

to the epipterygoid and the maxillary and mandibular branches posterior to

the bone.

a99

Ann. S. Afr. Mus. 48 (17), 1965, 399-426, 22 figs.

400 ANNALS OF THE SOUTH AFRICAN MUSEUM

In mammals the development of the alisphenoid as a relatively broad and

flattened bone is fairly constant except in Echidna where it appears to be

absent. In mammals, however, the relationship of the branches of the trigeminal

nerve to the alisphenoid varies. This ranges from the ‘reptilian’ type, where

the bone lies between the profundus and maxillary branches, found in Dizdelphis,

through stages where first the maxillary and then also the mandibular branches

pierce the bone, to where the profundus and maxillary branches both emerge

anterior to the alisphenoid.

The palaeontological evidence for the homology of the reptilian epiptery-

goid and the mammalian alisphenoid presented in the early years (mainly by

Broom) naturally reflects the inadequate state of knowledge of the early

reptile fossils at the time. Although seldom stated explicitly there was a strong

tendency to see in the lacertilian condition the basic type from which the

mammalian condition could be derived.

As far back as 1907 Broom stated that palaeontological evidence supported

the view that the alisphenoid of mammals had evolved from the reptilian

epipterygoid concluding:

‘In the very primitive reptiles, of which Procolophon may be taken as a

type, we have a lizard-like columella cranii. In the early types, which

have specialized along the mammalian line, such as the Therocephalians,

we still have a columella cranii. In the even more mammal-like ano-

modonts, such as Dicynodon and Oudenodon, there is a columella cranii, but

no alisphenoid. When we come to examine the Cynodonts—those remark-

able reptiles, so very nearly related to the Mammals as to be regarded by

many as the group ancestral to them—we find a broad flattened bone,

which most anatomists would not hesistate to call the alisphenoid, and

yet there can be little doubt, it is only the columella cranii of the anomodont

flattened out’ (p. 114).

In 1914 Broom offered more direct evidence to substantiate his view

stating:

‘A short epipterygoid is known in Diadectes and a long lizard-like columella

cranil is met with in Procolophon. A short epipterygoid occurs in the Pely-

cosaur Dimetrodon. In Anomodonts the epipterygoid is present as a long

slender rod—in some species rounded and in others much flattened . . .

and its lower end is considerably expanded antero-posteriorly, giving it a

long suture with the pterygoid. In Gorgonopsians the epipterygoid is a

long, much flattened rod which inferiorly has probably similar relations

to the pterygoid as seen in the Anomodonts. In the Therocephalians the

epipterygoid is only satisfactorily known in Scylacosaurus. Here, . . . it is a

relatively short, flattened structure with a very wide base which lies along

the pterygoid. In the Cynodonts, at least in the higher forms as exemplified

by Diademodon and Cynognathus, in the region occupied by the epipterygoid

in the Therocephalians there is a very much larger bone which there can

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 401

be little doubt is also an epipterygoid development. The upper part is

greatly expanded antero-posteriorly and forms much of the cranial wall.

The lower portion of the bone is so much more developed than in the

Therocephalian that it completely replaces the posterior part of the

pterygoid, . . . and extends outwards as far as the quadrate. There can be

little doubt that while this bone is homologous with the epipterygoid of

the lower forms, it is also the homologue of the mammalian alisphenoid’

(p- 30).

In the next thirty years very little advance was made. In 1944 Olson still

describes the structural changes undergone by the epipterygoid in the transition

from the primitive reptilian condition to that of mammals as follows:

‘In primitive reptiles the bone is composed of a restricted basal portion

and a slender ascending ramus, the columella cranii. In somewhat more

advanced forms, the basal portion is expanded into a posterior quadrate

ramus and an anterior pterygoid process. At this stage the ascending

ramus remains a thin rod. In the advanced mammal-like reptiles, the

ascending ramus is much expanded and has attained intimate association

with the periotic behind and the parietal above. The quadrate and

pterygoid processes are somewhat elongated’ (p. 110).

In the twenty years that have elapsed since Olson wrote the foregoing

descriptions of the palatoquadrate of all the principal groups of reptiles and of

the forms ancestral to the reptiles have become available. A restatement of the

course of the evolution of the epipterygoid-alisphenoid has therefore become

necessary. In the following pages a short review of the present state of our

knowledge of this element is given. Although the epipterygoid is a distinct

element in the early tetrapod skull its evolution cannot be discussed in isolation

since both the number of palatoquadrate ossifications and the extent of these

elements vary. An account of the palatoquadrate as a whole will therefore be

given and the evolutionary trends evident in its ossifications discussed.

My interest in the transformation of the epipterygoid (and quadrate

in so far as it affected the epipterygoid) originated while studying serially

ground sections of the skull of the anomodont Pristerodon buffaloensis. This

technique revealed important features of the palatoquadrate complex which

would not have been readily seen in material cleaned in the normal way. As

some of these features contradict vital points in the evolutionary sequences as

interpreted by Broom and Olson, an account of the Pristerodon palatoquadrate

will be presented first and will be followed by a general review of the trans-

formation of the palatoquadrate.

MATERIAL

A skull of Pristerodon buffaloensis Toerien, No. B.P.1. 339, was made available

for sectioning by Dr. A. S. Brink, Assistant Director of the Bernard Price

402 ANNALS OF THE SOUTH AFRICAN MUSEUM

Institute for Palaeontological Research, Johannesburg. It was discovered

in 1956 by Mr. James W. Kitching of the same Institute in Crstecephalus Zone

strata on the farm Kirkvors, today known as De Hoop, approximately two

miles north-west of Murraysburg, Cape Province. Pristerodon belongs to the

family Endothiodontidae of the Sub-order Anomodontia (Sub-class Synapsida ;

Order Therapsida).

THE PALATOQUADRATE OF PRISTERODON

The ossified palatoquadrate of Pristerodon buffaloensis is represented by the

separated epipterygoid and the quadrate (figs. 1 and 4). Viewed laterally the

epipterygoid is L-shaped, with both the vertical and the horizontal limbs well

developed. The vertical limb (or columella cranii) is relatively thin and extends

upward and slightly forward to meet a ventrally extending ridge originating

PILA ANTOTICA

INCISURA

PROOTICA

> |

PARABASISPHENOID ~ wenn ~

EPIPTERYGOID~ ESET | Ce

PTERYGOID ~ ate’ fo

fin/ / Aa \

Asif : TOJUGA

QUADRATE RAMUS EPIPTERYGOID / a, QUADRATOJUGAL

QUADRATE RAMUS PTERYGOID /

QUADRATE’

Fic. 1. Pristerodon buffaloensis. Lateral view of skull reconstructed from serial sections.

Squamosal cut away to show palatoquadrate complex.

PILA ANTOTICA

AUDI TORY CAPSULE

PALATOQUADRATE BAR~

PARS QUADRATA~

Fic. 2. Tropiocolotes tripolitanus. Lateral view of skull of 15:2 mm.

embryo. (After Kamal, 1960.)

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 403

from the under-surface of the parietal. The posteriorly directed horizontal

portion is thickest near the angle of the bone, tapering posteriorly to become

rod-shaped (quadrate ramus of the epipterygoid). The entire base of the

epipterygoid is applied to the dorsal and/or dorso-lateral surface of the quadrate

ramus of the pterygoid. The quadrate rami of both the pterygoid and the

epipterygoid are directed postero-laterally, towards the inner surface of the

quadrate.

The quadrate is relatively large, has the shape of an inverted Y in cross-

section, the upper leg lying in a deep antero-posteriorly directed ventral

pocket of the squamosal. It is suturally attached to two other elements only, the

quadratojugal on the outside and the stapes to which it is attached through a

short horizontal ridge projecting medially from the inner condyle. Antero-

dorsally to this ridge there is a longitudinal groove in the vertical face of the

quadrate, extending backwards from the anterior border for approximately

one-third of the length of the vertical plate (fig. 4). Throughout its length

this groove is seen to follow remarkably closely a course parallel to the posterior

portion of the quadrate ramus of the pterygoid, the latter bone terminating

close to, but free from the inner surface of the quadrate, approximately half-way

along its length. A distinct rounded bulge terminates the groove in the left

quadrate. As the groove follows what could have been the course of a posterior

extension of the horizontal limb or quadrate ramus of the epipterygoid it

would seem reasonable to assume that the groove housed a cartilaginous rod

connecting the quadrate with the epipterygoid.

The occurrence of a solid link between the quadrate and epipterygoid in

an adult Pristerodon, recalls conditions found in the developmental stages of

the skull in many recent reptiles and in developmental stages of Sphenodon

(fig. 3) as well as in the adult. The complex, as reconstructed in Pristerodon,

ASCENDING

PROCESS

PALATOQUADRATE BAR

Fic. 3. Sphenodon punctatus. Embryo. Lateral view of skull. (After Howes

& Swinnerton, 1901.)

404. ANNALS OF THE SOUTH AFRICAN MUSEUM

shows a remarkable resemblance to the palatoquadrate of Zonurus (Condylus),

as featured by Broom (1925), and of the 15:2 mm. developmental stage of the

Egyptian lizard Tropiocolotes tripolitanus (fig. 2), described by Kamal (1960).

Similarly the relationship between the horizontal limb and the quadrate

ramus of the pterygoid in Pristerodon is mirrored by conditions found in Sphenodon,

the base of the palatoquadrate cartilage occupying the dorsal and dorso-lateral

surface of the quadrate ramus of the pterygoid in both forms.

In this connection it is important also to consider the relationship of the

nerves to the complex, and especially that of the trigeminal nerve to the

epipterygoid. In recent reptiles the nerve emerges from the braincase through

the incisura prootica. The profundus branch then passes medially to the

epipterygoid while the maxillary and mandibular rami pass behind the epiptery-

goid. In Pristerodon there are no indications of the routes followed by the

presumed branches of the trigeminal nerve but judging by the nerve foramina

and grooves found in some early gnathostomes and tetrapods it would seem

indicated that present-day relationships had an early origin.

QUADRATE RAMUS

OF EPIPTERYGOID

LONGITUDINAL

GROOVE

ASCENDING

RAMUS OF

EPIPTERYGOID

BULGE |

~ POSITION OF QUADRATE

QUADRATE RAMUS OF PTERYGOID

AREA OF ATTACHMENT

OF STAPES

Fic. 4. Pristerodon buffaloensis. Medial view of palato-

quadrate complex. Left side.

In the phylogenetic review that follows, it will be seen that the basic

form of the palatoquadrate is retained from the early gnathostome condition

up to the early ‘stem reptile’ stage. The number of ossifications in the palato-

quadrate would seem to vary from stage to stage and even within stages and

could possibly be of minor evolutionary importance, especially as cartilaginous

inter-connections are indicated. If this is so it might explain the apparent

inconsistency of an otic process on the epipterygoid ossification in certain

labyrinthodonts while in other labyrinthodonts and later forms it forms part

of the quadrate ossification. After the attainment of the basic reptile stage, or

possibly even already during the labyrinthodont stage the palatoquadrate

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 405

begins to undergo evolutionary change which will have far reaching effects.

The development is channelled into two main streams; one leading towards

the expansion of the epipterygoid and reduction of the quadrate, as exemplified

in those trends showing mammalian affinities, and another resulting in the

retention of the quadrate and reduction of the epipterygoid in trends showing

reptilian affinities.

PHYLOGENY

The first adequately known vertebrates are agnathous forms found in

the Late Silurian. Jawed vertebrates have as yet not been discovered in the

Silurian but the variety of these forms in the Lower Devonian indicates they

must have been undergoing development in the Silurian (Romer, 1955).

PLACODERMI

The early Devonian gnathostomes are predominantly placoderms.

Although widely varied in appearance, the jaw apparatus is usually of a

relatively primitive type. Amongst them the acanthodians are generally

regarded as the earliest and most archaic and their morphology could, therefore,

throw light on the problems connected with the formation and evolution of

the primitive jaws.

According to Watson (1937) the acanthodian palatoquadrate is large

and superficially divisable into two elements, a short horizontal suborbital or

palatal portion and a large vertical postorbital or paraotic portion. In the

genera Climatius and Cheiracanthus the palatoquadrate is ossified as a single

unit, but in Mesacanthus and Ischnacanthus the palatal and paraotic portions

are ossified independently. In Acanthodes (fig. 5), the last surviving member of

the group, and therefore possibly specialized, the palatoquadrate is ossified

as three separate structures, but it seems certain that these bones, in life,

comprised parts of a single palatoquadrate. Anteriorly the paraotic plate

ends in a vertical border, which forms the hind margin of the orbit. From the

top of this border the bone curves downwards posteriorly to end in the thickened

quadrate condyle. Behind the vertical border the paraotic plate is perforated

by a foramen, extremely large in Cheiracanthus, which could possibly have

served for the exit of the maxillary and mandibular branches of the trigeminal.

The palatal portion of the palatoquadrate terminates anteriorly at a point in

line with the middle of the orbit. This is some distance behind the front end

of Meckel’s cartilage and would appear to indicate that the palatoquadrate

was continued forward as cartilage.

The palatoquadrate does not seem to have contact with the neurocranium

in Climatius, but in Mesacanthus, Cheiracanthus and Acanthodes, the paraotic

flange bears an otic process which articulates with the skull behind the post-

orbital process. The palatal part of the bone has a basal articulation.

406 ANNALS OF THE SOUTH AFRICAN MUSEUM

OTIC PROCESS POSTERIOR BONE

ANTERIOR BONE oF pALATOQUADRATE IN PALATOQUADRATE

IN PALATOQUADRATE

spac err TT

auld We

NY Autilt wy

2 CARTILAGINOUS ee

Gilet name oa

=z =

Fic. 5. Acanthodes sp. Reconstruction of the skull from specimens from the Lebach

ironstones. Complete except for squamation. (After Watson, 1937.)

According to Watson (1937) the palatoquadrate in arthrodires is directly

comparable with that occurring in several acanthodians. In Pholzdosteus the

palatoquadrate has two ossifications, an anterior which articulates with the

ventral part of the neurocranium just behind the olfactory capsules, and a

posterior quadrate.

CROSSOPTERYGII

In the Crossopterygii the processes and future subdivisions of the

palatoquadrate become more apparent.

In Eusthenopteron the palatoquadrate is a single ossification with, according

to Jarvik (1954), five distinguishable parts; an anterior pars autopalatina,

articulating with the ethmoidal region and the anterior neural endocranium;

a basal process articulating with the basipterygoid process; an antero-dorsally

directed ascending process which articulates with the suprapterygoid process

of the neural endocranium; a paratemporal process, which articulates with

the anterior end of the otic shelf and a thickened pars quadrata posteriorly.

A thin film of bone, dotted with large fenestrae, makes up the rest of the

palatoquadrate (fig. 6).

Jarvik states that the profundus passed medially to the ascending process

while the maxillary and mandibular branches of the trigeminal probably

passed through the notch between the ascending and paratemporal processes.

Osteolepis macrolepidotus similarly possesses a completely ossified palato-

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 407

ASCENDING PROCESS

TAROT NTI

met lls >

PTERYGOID

Fic. 6. Eusthenopteron foordi. Lateral view of skull with dermal bones

removed. (After Jarvik, 1954.)

quadrate (Watson, 1954). But in Megalichthyes (Watson, 1925), the palato-

quadrate is ossified as a continuous series of bones. Of these the anterior one,

which extends into the basal process, represents the epipterygoid, the posterior

the quadrate.

LABYRINTHODONTIA

The Coal Measure embolomerous anthracosaurs Palaeogyrinus decorus and

Eogyrinus altheyi (Panchen, 1964) both display well developed palatoquadrates.

In Palaeogyrinus (fig. 7) a large facet forms the anterior termination of a massive

buttress, shaped like the mouth of a horn, which tapers forward as the palatal

ramus. The latter appears to have rested on the horizontal palatal ramus of

the pterygoid, but its lower edge cannot be made out. In mesial view the

columella cranii appears as a slender rod extending dorsally until it expands

to form an anteriorly directed process and a second vertical process. Panchen

believes that the anteriorly directed process, the dorso-mesial surface of which

is grooved, formed the dorsal edge of a notch for the profundus and that the

second process probably divided the maxillary and mandibular branches of the

trigeminal nerve. In lateral view the columella is continuous with a sheet of

bone covering the whole of the upper half of the quadrate ramus of the pterygoid

and which runs from the level of the cranial base to the quadrate condyle. A

similar expansion is present in Edops (Romer and Whittier, 1942). Panchen

believes that the epipterygoid may have extended as cartilage further down

the quadrate ramus of the pterygoid and there seems also to have been a

broad process projecting downward and backward from the level of the basis

cranil. There is also a complex articulation for the basipterygoid process of

the basisphenoid formed by both epipterygoid and pterygoid.

The quadrate is small in Palaeogyrinus but is a much more extensive

ossification in Edops.

Panchen states that the considerable ossification in the cartilaginous

palatoquadrate shown by these forms, and probably by Eogyrinus must be

408 ANNALS OF THE SOUTH AFRICAN MUSEUM

eae ee PARIETAL

—

aaa ne

Viz

Z o~

TE ee «

—

EPIPTERYGOID A PTERYGOID

Fic. 7. Palaeogyrinus. Lateral view of suspensorial region as though

sectioned through a plane parallel to the quadrate ramus. (After

Panchen, 1964.)

interpreted as a primitive condition and that a progressive reduction of the

epipterygoid ossification is to be seen in later labyrinthodonts.

Although Triassic labyrinthodonts are not on the phylogenetic line

leading to reptiles, the structure of the palatoquadrate is nevertheless of

interest. It consists of two parts, a horizontal cartilaginous anterior part and

a mainly vertical ossified posterior part (Sushkin, 1899; Watson, 1919 and

1926, and Saéve-Séderbergh, 1936). The vertical part had, at least in Lyrocephalus,

basal, ascending and otic processes. Save-Sdderbergh maintains that all

these processes were embodied in the epipterygoid in the most completely

ossified individuals, but that in Aphaneramma and Platystega the ascending

process only was included in the epipterygoid. The epipterygoid was joined to

the quadrate ossification by means of a cartilaginous link which rested in a

groove of the ramus of the pterygoid.

We may therefore assume that the palatoquadrate survived in laby-

rinthodonts complete from the quadrate to at least the front end of

the epipterygoid.

CoTyLOSAURIA

Whether or not the seymouriamorphs are considered amphibians or

reptiles, the skull of primitive forms such as Seymouria and Kollassia show that

they are not far removed from the anthracosaurian amphibians from which

the seymouriamorphs and more advanced reptiles have been derived (Romer,

1956). The epipterygoid, which is not well known and seems to be slow to

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 409

ossify, extends as a rod-like structure to the skull roof. The primitive quadrate,

when well ossified, ran forward to meet the epipterygoid. When the quadrate

is less fully ossified, a cartilaginous area may have bridged the gap between

the quadrate and the epipterygoid (fig. 8). It is presumed that the tendency

for the epipterygoid to remain unossified is a degenerate feature (Romer, 1956).

The earliest known reptiles date from the Carboniferous but knowledge

of their anatomy is fragmentary. Abundant forms appear in the Upper

Carboniferous and Lower Permian but as these include not only primitive

forms but also more advanced types, it is certain that a very important section

of the early history of the reptiles is still unknown.

In diadectids both epipterygoid and quadrate are greatly developed.

Diadectes itself features a quadrate of considerable height and width but little

length (Romer, 1956). In the South African pareiasaurs both epipterygoid and

quadrate are plate-like. The epipterygoid, which has a slender ascending

process, furthermore stands transversely on the quadrate ramus of the pterygoid

near its articulation with the basipterygoid process (Boonstra, 1934). According

to Romer (1956) the margins of the base of the epipterygoid in pareiasaurians

are unfinished, indicating a cartilaginous connection with the quadrate which

has a similarly unfinished surface facing towards the epipterygoid.

The captorhinomorphs Limnosceles, Captorhinus and Labidosaurus from the

Lower Permian of the American Southwest are among the oldest cotylosaurs

EPIPTERYGOID

OCCIPITAL

Pa Nth

ST an

PTERYGOID

a7,

CARTILAGINOUS PORTION OF

PALATO QUADRATE RECONSTRUCTED QUADRATE

Fic. 8. Seymouria. Lateral view of the skull with dermal bones removed.

Cartilaginous portion of palatoquadrate reconstructed. (After Romer,

1956.)

410 ANNALS OF THE SOUTH AFRICAN MUSEUM

of which we have detailed knowledge. Here the quadrate also consists of a

vertical sheet of bone with an essentially flat outer surface and an epipterygoid,

separated from the quadrate but presumably attached to it in life by cartilage,

with a rod-like ascending process (fig. 9).

t es eee

tas ee ~~,

2? CARTILAGINOUS

PALATOQUADRATE

77 PTERYGOID QUADRATE

Fic. 9. Captorhinus. Lateral view of the skull with dermal bones removed.

_Cartilaginous portion of palatoquadrate tentatively indicated. (After

Romer, 1956.)

Although we have no knowledge of the epipterygoid in the aberrant

early Permian form Bolosaurus (Watson, 1954), the fact that the inner part of

the quadrate continues forward as a ramus, on the outer surface of the pterygoid,

would seem to indicate that the palatoquadrate was basically the same as the

general cotylosaurian type.

The Lower Triassic form Procolophon belongs to the last group of survivors

of the cotylosaurs. In this form the epipterygoid has an expanded footplate, the

anterior extension of which is more fully developed than the posterior portion.

From the posterior portion of the footplate a slender ascending ramus extends

dorsally. The quadrate is large and has a process extending forward which

partially replaces the quadrate ramus of the pterygoid. In Procolophon the

latter is reduced in length as well as in height.

PELYCOSAURIA

If, as is generally accepted, the Pelycosauria represent an early stage in

mammalian history and that they possess ‘. . . many archaic features which

EPIPTERYGOID— ALISPHENOID TRANSITION IN THERAPSIDA 411

illustrate the structure of the primitive reptilian stock’ (Romer & Price,

1940: 1), then it should not be surprising to find that this group still displays

a palatoquadrate complex in which the epipterygoid and quadrate portions

are linked. This condition would merely represent the retention in this group

of the basic gnathostome condition of a unified palatoquadrate as exemplified

in the placoderms, and retained, with modifications, in the crossopterygians,

labyrinthodonts and probably also the early stem reptiles.

The epipterygoid has a wide base closely applied to the lateral surface of

the anterior portion of the quadrate ramus of the pterygoid (fig. 10). Anteriorly

a slender extension curves down along a groove on the dorsal surface of the

palatal ramus of the pterygoid. A continuation of this groove in some cases

beyond the anterior end of the bone suggests a further cartilaginous extension

of the palatoquadrate. The slender rod-like portion of the epipterygoid gains

contact, in Dimetrodon, with the anterior surface of the paroccipital process.

EPIPTERYGOID

Se WS PROOTIC

<a CSE

FOOTPLATE OF \Y

EPIPTERYGOID PTERYGOID

Fic. 10. Dimetrodon limbatus. Lateral view of skull with dermal bones

cut off to show palatoquadrate. (After Romer & Price, 1940.)

The quadrate is essentially a triangular plate varying greatly in extent in

the few forms where it is adequately known. In some cases in which it is

well developed it appears to gain contact with the epipterygoid along the

dorsal margin of the internal surface of the pterygoid (fig. 10), while in others

in which the bone is more restricted in size, it appears to have been continued

forward as cartilage.

THERAPSIDA

Dinocephalia

Of the slightly more advanced mammal-like reptiles, the therapsids, the

Dinocephalia are generally regarded as the oldest and in many respects the

most primitive suborder. In them, according to Boonstra (personal communica-

tion), the epipterygoid is small or reduced except in the titanosuchid genus

412 ANNALS OF THE SOUTH AFRICAN MUSEUM

Anteosaurus. In the latter the bone extends right up to the skull roof, has an

expanded waist and a relatively long anterior footplate. In Jonkeria (titano-

suchid) the footplate is also extended anteriorly but dorsally the epipterygoid

only reaches halfway to the roofing bones while the footplate as well as the

dorsal ramus are reduced in the tapinocephalid Struthiocephalus. In none of

the Dinocephalia do we find evidence of a posterior extension to the footplate

which could have formed a link between the epipterygoid and the quadrate.

According to Watson (1914) the Dinocephalia, as a group, has a very special

importance because alone amongst the therapsids it retained a large quadrate.

This feature as well as the absence of the quadrate ramus of the epipterygoid

indicates affinities with the reptilian rather than the mammalian lines of

evolution.

Anomodontia

The anomodonts are generally regarded to be an aberrant group of

mammal-like reptiles with but weak mammalian affinities. This is borne out

by the structure of the palatoquadrate complex in this group. As in the Dino-

cephalia the anomodonts Aannemeyeria erithrea (Case, 1934), Dicynodon kolbet

(Broom, 1932), D. sollast (Watson, 1948), Lystrosaurus murray: (Broom, 1932),

Daptocephalus leoniceps (Ewer, 1961), Aingoria nowackt (Cox, 1959), Dicynodon

grimbeeki and Pristerodon buffaloensis, all display an epipterygoid with a long,

thin dorsally projecting columellar portion, reminiscent of recent reptiles

(see later). However, in contrast to the latter the base of the epipterygoid is

expanded, to varying degrees, in all of the above, extending for some distance

along the upper edge of the quadrate ramus of the pterygoid. In Kannemeyeria

the base extends from a point close to the quadrate to beyond the front of the

interpterygoid space.

In the anomodonts investigated, there is no direct contact between the

epipterygoid and the quadrate (fig. 11), but there is a distinct probability

that a cartilaginous link existed in life in some of these forms, as has been shown

59 QUADRATO-

JUGAL

QUADRATE

Fic. 11. Dicynodon sp. Lateral view of skull.

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 413

for Pristerodon. However, the general tendency within the group is for the

bar, linking the epipterygoid to the quadrate portion, to become reduced;

an evolutionary trend, as will be shown later, that was present also in the line

that culminated in modern reptiles.

Gorgonopsia

In all known Gorgonopsia the epipterygoid is high and relatively narrow,

with a well-developed footplate. In ‘“Lycaenodon’ (fig. 12), Scymnognathus, Lepto-

trachelus, Cynariops (Boonstra, 1934a) and Aelurognathus (Haughton, 1924), the

footplate has a long tapering anterior extension, with a short posterior extension

ending some distance away from the quadrate. Boonstra found no evidence

in this group of a widening of the vertical portion of the epipterygoid.

ue,

ae > Mp

EPIPTERYGOID

PROOTIC

FOOT PLATES:

Fic. 12. ‘Lycaenodon’. Lateral view of skull with dermal bones cut away.

(After Boonstra, 19344.)

Therocephalia

In the early Tapinocephalus Zone forms such as the pristerognathids

Scylacosaurus and Scymnosaurus (Boonstra, 1934b, 1954), and most of the primitive

therocephalians, the vertical portion of the epipterygoid is still relatively

narrow, but in Glanosuchus macrops (Boonstra, 1954) the epipterygoid has

become dumb-bell-shaped, although still not much widened. The base has

no noteworthy posterior process. However, in the lycosuchid Trochosaurus

major (Boonstra, 1934b) occurring in the same zone, the epipterygoid has

414 ANNALS OF THE SOUTH AFRICAN MUSEUM

developed into a large, broad plate which obscures the lateral opening into

the pituitary fossa and the foramina for the Vth and VIth nerves. The epiptery-

goid here is more than twice as broad as in any known gorgonopsian and has

the appearance of a cynodont epipterygoid. It differs from that known in

Scymnosaurus and Scylacosaurus in having an expanded upper end in addition

to a wide footplate, a condition reaching its greatest development in the later

whaitsids (Boonstra, 1934).

Of the Cistecephalus Zone Therocephalia, Euchambersia mirabilis (Boonstra,

1936) also shows the dorsally and ventrally expanded condition (fig. 13) but

the shaft of the bone does not show much expansion. In the scaloposaurid

Ictidosuchops intermedius (Crompton, 1955), however, the shaft is a broad one,

and ends in a slightly more expanded dorsal portion which fits into a shallow

depression in the antero-dorsal wall of the prootic, consequently forming

part of the true lateral wall of the braincase. The anterior extension of the

footplate is long, but the posterior one is short. At the juncture of the posterior

extension and the ascending ramus there is a shallow notch which most probably

housed the maxillary or both the maxillary and mandibular branches of the

trigeminal nerve (Crompton, 1955).

EPIPTERYGOID ~ 7 bake

PTERYGOID PAROCCIPITAL

QUADRATE RAMUS OF PTERYGOID

Fic. 13. Euchambersia mirabilis. Lateral view of skull with dermal bones

cut away. (After Boonstra, 1936.)

Za,

Fic. 14. Aneugomphius ictidoceps. Lateral view of skull with outlines of

palatoquadrate bones indicated. (After Brink, 1956.)

EPIPTERYGOID —ALISPHENOID TRANSITION IN THERAPSIDA 415

In Thertognathus microps (Boonstra, 1934), Notosollasia laticeps (Boonstra,

1934) and Aneugomphius ictidoceps (Brink, 1956), as in all known whaitsids, the

epipterygoid is broad and flat with greatly expanded dorsal and ventral ends

(fig. 14). The base is long and in most forms posteriorly overlaps the prootic

obscuring the incisura prootica. The hind wall of the epipterygoid in Aneugomphius

shows two deep notches, probably for the exit of the maxillary and mandibular

branches of the trigeminal nerve.

Cynodontia

Brink (1960) states that the epipterygoids of the primitive cynodont

Scalopocynodon gracilis are *. . . Cynodont-like in being broadly expanded, but

Therocephalian-like in their feeble overlap of the prootics’ (p. 145). The

anterior extension of the footplate is long and broad while the remainder of

the ventral margin follows the dorsal margin of the quadrate ramus of the

pterygoid (fig. 15). Brink made the interesting observation that the anterior

margin of the epipterygoid is very thick. This may well support the view that

this portion of the vertical limb is the more stable and is older phylogenetically

than the thinner, posterior section of the limb. The quadrate is in the shape of a

wedge lodged loosely in a cavity anteriorly in the squamosal. The anterior

margin is straight, blunt and vertical, while the posterior margin is sharp

and broadly curved. The quadrate is a much smaller bone than the epipterygoid

and is well separated from the latter.

In Thrinaxodon lorhinus (Parrington, 1946) the epipterygoid is greatly

expanded and dorsally is broadly in contact with the prootic. The quadrate

ramus of the epipterygoid joins the prootic, but does not reach the quadrate,

leaving a large foramen between the epipterygoid and prootic for the maxillary

and mandibular rami. The quadrate, which is about one-third the size of

the epipterygoid, has a wide, dumb-bell-shaped articulating surface and a

thin, slightly curved dorsal process which medially is greatly strengthened by

a pillar-like development which continues to the apex of the bone.

In an unidentified cynodont, which compares closely with Trzrachodon,

Parrington (1946) found two foramina between the epipterygoid and prootic.

A groove comparable with that leading into the single foramen in Thrinaxodon,

leads into the lower, larger foramen from below and behind. The dorsal

smaller foramen is probably new. Parrington states that the significance of this

development lies in the possibility that the maxillary left the skull by the

upper foramen, which is apparently mainly enclosed by the epipterygoid, and

is equivalent to the foramen rotundum, and the mandibular by the lower,

larger foramen, the equivalent of the foramen ovale. The size of both foramina

indicates that blood vessels accompanied the nerves.

In Diademodon mastacus (Brink, 1955) the foramen for the trigeminal is

also divided into two, although incompletely in this species (fig. 16). The

foramen lies slightly above and behind the pituitary fossa. The epipterygoid

itself covers the anterior half of the prootic. The footplate is extensive and

416 ANNALS OF THE SOUTH AFRICAN MUSEUM

extends forward as a very thin tapering process while it reaches the quadrate

with its posterior extension.

In the cynognathids, the thin plate-like epipterygoid overlaps the anterior

superior process of the prootic. The footplate of the base is extensive. It sends

back a posterior extension which occupies the position of the quadrate ramus

of the pterygoid and which reaches the relatively reduced quadrate.

QUADRATE RAMUS

OF PTERYGOID

Fic. 15. Scalopocynodon gracilis. Lateral view of skull with dermal bones

cut away. (After Brink, 1960.)

EPIPTERYGOID

QUADRATE RAMUS OF PTERYGOID

Fic. 16. Diademodon mastacus. Lateral view of skull with dermal bones

cut away. (After Brink, 1955.)

NOTCH

QUADRATE RAMUS

OF EPIPTERYGOID

PTERYGOID

Fic. 17. Diarthrognathus broomi. Lateral view of skull. (After Crompton,

1958.)

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 417

Ictidosauria

In Diarthrognathus broom: the dorsal portion of the epipterygoid is extremely

broad and overlaps the prootic above the prootic incisure. Here the posterior

edge of the epipterygoid is slightly indented for the maxillary and mandibular

branches of the trigeminal nerve (fig. 17). Crompton (1958) states that although

it is difficult to ascertain, it would appear that the quadrate ramus of the

epipterygoid meets the antero-medial edge of the quadrate. The quadrate is

much reduced, S-shaped in occipital view, and the articular surface is markedly

concave. This latter feature, according to Crompton, is of great significance as

in all known therapsids, including Brenotherium and Oligokyphus, the articular

face of the quadrate is either flat or convex, but never concave. It is also

significant that the mammalian malleus (articular) possesses a convex articular

surface which meets a convex articular face in the incus (quadrate).

In Bienotherium (Hopson, 1964) the epipterygoid and the anterior part of

the prootic together form the side wall of the brain case (fig. 18). The quadrate

ramus of the epipterygoid is more vertically orientated and deeper than in

EPIPTERYGOID FRONTAL PARIETAL

ee een i iene PROOTIC

ORBITAL

FISSURE

Fic. 18. Bienotherium yunnenense. Lateral view of braincase reconstructed

from serial sections. (After Hopson, 1964.)

418 ANNALS OF THE SOUTH AFRICAN MUSEUM

the cynodonts and could have reached the quadrate. The ascending ramus is

broad and high with a posterior border which is overlapped laterally, above

the trigeminal foramen, by the prootic. This latter is the reverse of the usual

reptilian condition in which the prootic lies medial to the epipterygoid, and

is ascribed by Hopson to the development of a forward extension of the anterior

portion of the prootic in such a way that it lies lateral to the cavum epiptericum.

In Bienotherium both this extension—the ventro-lateral flange—and the antero-

dorsal border of the prootic extend well forward of the prootic incisure, an

observation which has led Hopson to conclude that the prootic component of

the trigeminal foramen in this form is not strictly homologous with the prootic

incisure in cynodonts, because the former is merely a notch in the anterior

border of the ventro-lateral flange. This conclusion seems to be supported by

the fact that the trigeminal foramen pierces the side wall of the skull anterior

to the depression of the semilunar ganglion. It also indicates that the maxillary

and mandibular branches extend in an anterior direction through the cavum

epiptericum in Bzenotherium, whereas normally, in those forms with broadened

epipterygoids, they pass in a posterior direction.

The trigeminal foramen is slightly constricted at mid-length presumably

indicating the incipient subdivision of the single trigeminal foramen into the

separate foramina rotundum and ovale of mammals.

TRICONODONTA

In a group of Mesozoic mammals, the Triconodonta, the brain case has

an essentially reptilian structure. According to Kermack (1963) Triconodon

mordax and Trioracodon ferox still possessed a cavum epiptericum lying outside

the ossified lateral wall of the braincase (formed by the petrosal) and as in

Morganucodon and possibly all pre-Cretaceous mammals, the alisphenoid

formed the lateral boundary of the cavum. The ramus profundus passed out

through the anterior end of the cavum epiptericum while the maxillary and

mandibular nerves passed out through two foramina formed in the posterior

edge of the alisphenoid.

According to Kermack, the semilunar ganglion in Morganucodon lay inside

the primary wall of the braincase and the mandibular branch of the trigeminal

nerve passed though the foramen pseudovale in the anterior lamina of the

petrosal. The lamina extends much further forward in Morganucodon than it

does in cynodonts. The maxillary branch of the fifth nerve presumably left

the braincase, passing through the notch at the front end of the anterior

lamina. After crossing the cavum epiptericum both branches would have left

it by passing behind or through the alisphenoid in some way reminiscent of

cynodonts.

In basic construction the braincase of the triconodonts differs from that

of an advanced therapsid only in the narrower cavum epiptericum in the

former, a difference due to the relatively larger size of the brain in mammals.

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 419

Lirvinc REPTILES AND MAMMALS

Amongst recent reptiles Sphenodon has virtually retained its larval palato-

quadrate in the adult (compare figs 3 and 1g) but in the lizards Lacerta

(Gaupp, 1906), Cordylus, Eremias and Mabuia (Broom, 1903), Tropiocolotes

(Kamal, 1960), Varanus (Frazetta, 1962), Tupinambis (Jollie, 1960) and many

others, and in the chelonians Emys (Kunkel, 1912) and Chrysemys (Shaner,

1926), the palatoquadrate complex undergoes considerable changes during

ontogeny. Ossifications within the complex culminate in the formation of

widely separated epipterygoid and quadrate portions, while the rest of the

palatoquadrate is resorbed. The quadrate is retained as a comparatively

strong element while the epipterygoid is whittled down to a narrow rod-like

structure (fig. 20). The base of the bone is not expanded. It has, however, a

cartilaginous epiphysis, fitting into the fossa pterygoidei which may be linked

with the quadrate through a strand of connective tissue, both epiphysis and

connecting strand probably representing remnants of the palatoquadrate

cartilage.

EPIPTERYGOID PROOTIC

Fic. 19. Sphenodon punctatus. Lateral view of the skull with dermal

bones cut off to show palatoquadrate.

~_EPIPTERYGOID

cai | QUADRATE

PTERYGOID RROGTIC

QUADRATE RAMUS OF PTERYGOID

Fic. 20. Tupinambis nigropunctatus. Lateral view of skull. (After Jollie,

1960.)

420 ANNALS OF THE SOUTH AFRICAN MUSEUM

Further reduction of the epipterygoid has taken place in the lizards

Agama, Lyriocephalus and Calotes (Ramaswami, 1946), Ophioceps and Anniella

(Jollie, 1960), Phrynocephalus (Siebenrock, 1895), Physignatus, Chlamydosaurus

and Amphibolurus (Beddard, 1905), where the epipterygoid is very short.

However, in the Chamaeleontidae, Dibamidae (Boulenger, 1887) and

apparently most Amphisbaenidae (known only in Trogonophis, Bellairs, 1950),

the epipterygoid is absent. In the Chelonia the epipterygoid is still present

but very small (Parker, 1880) while it is much reduced or vestigial in Ophidia

and Crocodilia, although the embryos of crocodiles still show the ascending

process (Parker, 1883; Shiino, 1914). In birds it appears to be absent (Goodrich

1930), the quadrate being the only part of the palatoquadrate to be retained

in the adult. .

Where present the epipterygoid lies lateral to the lateral head vein,

postero-lateral to the profundus and antero-medial to the maxillary and

mandibular branches and the orbital artery.

In mammals the alisphenoid appears to be lost in Echidna. In this feature,

as will be seen later, Echidna probably shows reptilian affinities. In Dasyurus

the dorsal end of the ala temporalis fuses with the orbital cartilage, but in all

other mammals, as far as is known, the ala temporalis ends freely (De Beer,

1937). In Didelphis the alisphenoid lies between the profundus and maxillary

rami, as in reptiles. In Trichosurus, Mus, Mustela and many others (De Beer,

3 LPR

“XUBRTALISPHENOID \ =

ORBITAL FISSURE’ FORAMEN OVALE

FORAMEN ROTUNDUM

Fic. 21. Felis domestica. Lateral view of skull with jugal arch cut away.

1937) the alisphenoid is pierced by the maxillary ramus (foramen rotundum),

while in Felis (fig. 21), and the majority of mammals the mandibular ramus

too may pass through it (foramen ovale). In the fourth group the maxillary

emerges freely in front of the alisphenoid (De Beer, 1937).

Edinger and Kitts (1954) state that the mandibular branch of the trige-

minal appears to have been variable in its relations to the alisphenoid in

living mammals as the foramen ovale is absent in some of them. Several

EPIPTERYGOID —ALISPHENOID TRANSITION IN THERAPSIDA 421

genera of extant Perissodactyla and Artiodactyla lack the foramen but in

both these orders a separate foramen ovale was the usual condition in the

early Tertiary. Comparison of fossil and recent material suggests a similar

trend in certain families of the Rodentia and Insectivora. They relate the

formation of the foramen ovale with an earlier palaeoneurological development

when the mandibular nerve came to branch off the trigeminal stem intra-

cranially, during or near the origin of mammals. It is only in some later

forms that the nerve used the foramen lacerum for its passage from the brain-

case. This probably indicates that its absence is a secondary condition and

possibly reveals an evolutionary trend within the mammals to abolish the

separate exit of the mandibular nerve.

CONCLUSIONS

The evolutionary history of the palatoquadrate has been followed, as

far as the availability of information on fossil material permits, from the

early gnathostome condition to that in recent reptiles and mammals. It has

been shown that the primitive palatoquadrate was a solid structure, consisting

mostly of several ossifications, with four main processes for its attachment to

the neurocranium. These are the otic, ascending, basal and pterygoid processes.

In tetrapods the ossifications are reduced to two.

During its evolutionary history the palatoquadrate shows a tendency to

become reduced in size, the pterygoid process shortens considerably while the

area between the otic and ascending processes becomes deeply excavated

thereby accentuating and demarcating the epipterygoid and quadrate portions.

The demarcation of these two elements becomes more pronounced as evolution

progresses.

Conditions in recent and fossil forms show that the palatoquadrate of

both recent reptiles and mammals are deducable from the early gnathostome

condition and that they evolved through the placoderm, crossopterygian,

labyrinthodont and early stem reptile stages. In the groups leading to or

showing affinities with the lines leading to modern reptiles, the trend is towards

the progressive reduction of that part of the palatoquadrate anterior to the

quadrate. In most reptiles the epipterygoid is the only part of this area to be

retained but in some recent forms even this too disappears. In those lines

leading towards mammals the opposite occurred and regression of the quadrate

took place while the epipterygoid expanded (fig. 22).

From the foregoing it is clear that the phylogenetic and ontogenetic

evidence do not support the view that the epipterygoid in the primitive reptile

possessed a restricted base. The latter condition is a specialized reptilian one

as is proved by conditions obtaining in adult forms of many recent reptiles.

It is a stage in the gradual reduction of the epipterygoid in this group. Olson’s

(1944) views on the evolutionary development of the epipterygoid, must then

differ from those given here.

422

EVOLUTIONARY

TREND

| Li PELYCOSAURIAN

RHYNCHOCEPHALIAN ( \

SPHENODON

ANNALS OF THE SOUTH AFRICAN MUSEUM

sriy ea PLACODERM STAGE

oe CROSSOPTERYGIAN STAGE

LABYRINTHODONT STAGE

rena COTYLOSAURIAN STAGE

LINES SHOWING LINES SHOWING

REPTILIAN ae MAMMALIAN AFFINITIES

}

+ EVOLUTIONARY

TREND

\ | () THEROCEPHALIAN

}

Ke \

GORGONOPS!AN

PROCOLOPHONID

DICYNODONT

CYNODONT

) a

LIZARD ui

MAMMAL

CHAMELEON

Fic. 22. Schematic presentation of probable evolutionary trend in development of palatoquadrate

leading towards recent reptiles and mammals.

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 423

Although the therapsids show a marked degree of variability in the struc-

ture of the epipterygoid, we find that the groups which show mammalian

affinities have also retained certain basic features of the early tetrapod palato-

quadrate complex. The most noticeable of these is the retention of the extensive

base of the epipterygoid, indicating that a considerable portion of the base of

the palatoquadrate anterior to the quadrate, of the early labyrinthodont and

stem reptile, had been retained.

Two further features are closely linked with the development of the

alisphenoid in mammals. Firstly the expansion posteriorly of the rod-like

ascending process of the epipterygoid and secondly, the increase in size of

the brain in forms approaching the transitional stage.

Backward extension of the ascending process would result in the maxillary

and mandibular branches of the trigeminal nerve being pushed backward by

the epipterygoid as the latter progresses past the incisura prootica through

which the trigeminal emerges.

The increase in size of the brain in mammals and the consequent lateral

expansion of the braincase, could conceivably result in the nerves becoming

trapped between the posteriorly extending epipterygoid and the expanding

braincase. Close contact between the epipterygoid and braincase could force the

nerves to seek an exit through the epipterygoid. Whether this results in the

formation of a notch only, a foramen or two foramina depends largely on the

separate development of the epipterygoid and prootic, the way these bones

make contact and where this contact is made. In this way the maxillary ramus

can emerge through the foramen rotundum while the mandibular emerges

farther back through its own foramen ovale or merely through a notch in the

posterior portion.

If the expansion of the braincase is continued anteriorly the cavum

epiptericum would be obliterated, as it has been in mammals. This could

mean that the pila antotica, which forms the inner wall of the cavum in reptiles,

could have been pushed against the alisphenoid and as the pila will then

virtually have become redundant, there is no need for it to develop in mam-

mals. The possibility that this could have happened is strengthened by the

fact that isolated cartilaginous fragments of the pila antotica have been found

in many higher mammals, e.g. Lepus (Voit, 1909), Homo (Macklin, 1914),

Felis (Terry, 1917), Didelphys (Téplitz, 1920), Halicore (Matthes, 1921) and

Tarsius (Henckel, 1927).

ACKNOWLEDGEMENTS

I wish to record my gratitude to the South African Council for Scientific

and Industrial Research for a research grant during the tenure of which most

of this work was carried out. I should like to thank Professor E. Jarvik, Profes-

sor E. Stensié and the staff of the Palaeontological Institute of the Swedish

Museum of Natural History for their kindness, co-operation and assistance

424 ANNALS OF THE SOUTH AFRICAN MUSEUM

while working in the Swedish Museum of Natural History, Stockholm. My

thanks are also due to Dr. A. S. Brink, Dr. M. E. Malan and Dr. L. D. Boonstra

for reading the manuscript and making valuable suggestions, and Mrs. I.

Rudner for copying some of the drawings used in this paper.

The trustees of the South African Museum are grateful to the Council

for Scientific and Industrial Research for a grant to publish this paper.

SUMMARY

The serial sectioning of the skull of the anomodont Pristerodon buffaloensis

has produced valuable evidence regarding the structure of the palatoquadrate

complex and the evolution of the therapsid epipterygoid. A short review is

given of the structure of the palatoquadrate in the most important known

fossil vertebrate groups.

It is emphasized that the mammalian alisphenoid should not be looked

upon as a transformation of the rod-like epipterygoid, as found in more spe-

cialized true reptiles as Broom and others implied, but that both conditions

are derived from a common ancestral type.

REFERENCES

Aus, E. P. 1919. The homologies of the alisphenoids of the Sauropsida. 7. Anat., Lond. 53:

209-222.

BEDDARD, F. E. 1905. Some notes on the cranial osteology of the mastigure lizard Uromastix.

Proc. zool. Soc. Lond. 1905: 1-9.

Beuiarrs, A. d’A. 1950. ones ations on the cranial anatomy of Anniella, ana a comparison

with that of other burrowing lizards. Proc. zool. Soc. Lond. 119: 887-904.

Boonstra, L. D. 1934 a. Additions to our knowledge of the South African Gorgonopsia, pre-

served in the British Museum (Natural History). Ann. S. Afr. Mus. 31: 175-213.

Boonstra, L. D. 1934 b. A contribution to the morphology of the mammal-like reptiles of the

sub-order Therocephalia. Ann. S. Afr. Mus. 31: 215-267.

Boonstra, L. D. 1938. On a South African mammal-like reptile, Bauria cynops. Palaeobiologica

6: 164-183.

Boonstra, L. D. 1943. Pareiasaurian studies. Part IX.—The cranial osteology. Ann. S. Afr.

Mus. 31: 1-38.

Boonstra, L. D. 1954. The pristerognathid therocephalians from the Tapinocephalus Zone in

the South African Museum. Ann. S. Afr. Mus. 42: 65-107.

BouLEeNGER, G. A. 1887. Catalogue of the lizards in the British Museum (Natural History). 2nd ed.

3. London: British Museum.

Brink, A. S. 1955. A study of the skeleton of Diademodon. Palaeont. Afr. 3: 3-39.

Brink, A. S. 1956. On Aneugomphius ictidoceps Broom and Robinson. Palaeont. Afr. 4: 97-115.

Brink, A. S. 1960. A new type of primitive cynodont. Palaeont. Afr. 7: 119-154.

Broom, R. 1903. On the development of the pterygoquadrate arch in Lacertilia. 7. Anat.,

Lond. 37: 107-111.

Broom, R. 1907. On the homology of the mammalian alisphenoid bone. Rep. S. Afr. Ass. Adv.

Sci. 1907: 114-115.

Broom, R. 1909. Observations on the development of the marsupial skull. Proc. Linn. Soc.

N.S.W. 34: 24-105.

Broom, R. 1914. Croonian lecture. On the origin of mammals. Phil. Trans. (B) 206: 1-48.

Broom, R. 1925. On the origin of lizards. Proc. zool. Soc. Lond. 1925: 1-16.

Broom, R. 1932. The mammal-like reptiles of South Africa and the origin of mammals. London:

Witherby.

EPIPTERYGOID—ALISPHENOID TRANSITION IN THERAPSIDA 425

Case, E. C. 1934. Description of the skull of Kannemeyeria erithrea Haughton. Contr. Mus. Paleont.

Univ. Mich. 4: 115-127.

Cox, C. B. 1959. On the anatomy of a new dicynodont genus with evidence of the position

of the tympanum. Proc. zool. Soc. Lond. 132: 321-367.

Crompton, A. W. 1955. A revision of the Scaloposauridae with special reference to kinetism in

this family. Navors. nas. Mus., Bloemfontein. 1: 149-183.

Crompton, A. W. 1958. The cranial morphology of a new genus and species of ictidosaurian.

Proc. zool. Soc. Lond. 130: 183-216.

De BEER, G. 1937. The development of the vertebrate skull. Oxford: University Press.

Epincer, T. & Kirts, D. B. 1954. The foramen ovale. Evolution. 8: 389-404.

Ewer, R. F. 1961. The anatomy of the anomodont Daptocephalus leoniceps (Owen). Proc. zool.

Soc. Lond. 136: 375-402.

Frazetta, T. H. 1962. A functional consideration of cranial kinesis in lizards. 7. Morph. 111:

287-319.

Gaupp, E. 1902. Uber die Ala temporalis des Saugethierschadels und die Regio orbitalis einiger

anderer Wirbelthierschadel. Anat. Hefte Abt. 1. 19: 155-230.

Goopricy, E. S. 1930. Studies on the structure and development of vertebrates. London: MacMillan.

Grecory, W. K. & Nose, G. K. 1924. The origin of the mammalian alisphenoid bone.

J. Morph. 39: 435-463.

Haucuton, S. H. 1924. On some gorgonopsian skulls in the collection of the South African

Museum. Ann. S. Afr. Mus. 12: 499-517.

HENCKEL, K. O. 1927. Zur Entwicklungsgeschichte des Halbaffenschadels. <. Morph. Anthrop.

26: 365-383.

Hopson, J. A. 1964. The braincase of the advanced mammal-like reptile Bienotherium. Postilla

87: 1-30.

Jarvik, E. 1954. On the visceral skeleton in Eusthenopteron with a discussion of the parasphenoid

and palatoquadrate in fishes. K. svenska Vetensk. Akad. Handl. (4) 5 (1): 1-104.

Joti, M. T. 1960. The head skeleton of the lizard. Acta zool., Stockh. 41: 1-64.

Kamat, A. M. 1960. The chondrocranium of Tvopiocolotes tripolitanus. Acta zool., Stockh. 41:

297-312.

Kermack, K. A. 1963. The cranial structure of the triconodonts. Phil. Trans. (B) 246: 83-103.

KunkEL, B. W. 1912. The development of the skull of Emys lutaria. J. Morph. 23: 633-765.

Mackin, C. C. 1914. The skull of a human fetus of 40 mm. Amer. J. Anat. 16: 317-426.

Mattues, E. 1921. Zur Entwicklung des Kopfskelettes der Sirenen. Das Primordialkranium

von Halicore dugong. X. Anat. Entw-Gesch. Abt. 1. 60: 1-304.

Otson, E. C. 1944. Origin of mammals based upon cranial morphology of the therapsid sub-

orders. Spec. Pap. geol. Soc. Amer. 55: i-xi, 1-136.

PANCHEN, A. L. 1964. The cranial anatomy of two Coal Measure anthracosaurs. Phil. Trans.

(B) 247: 593-637.

PARKER, W. K. 1880. Development of the green turtie (Chelone viridis, Schneider). Rep. Voy.

Challenger 1873-76. & (5): 1-58.

Parker, W. K. 1883. On the structure and development of the skull in Crocodilia. Trans.

zool. Soc. Lond. 11: 263-310.

PARRINGTON, F. R. 1946. On the cranial anatomy of cynodonts. Proc. zool. Soc. Lond. 116: 181-197.

Ramaswam, L. S. 1946. The chondrocranium of Calotes versicolor (Daud.) with a description of

the osteocranium of a just-hatched young. Quart. F. micr. Sci. 87: 237-297.

Romer, A. S. & Price, L. W. 1940. Review of the Pelycosauria. Spec. Pap. geol. Soc. Amer. 28:

I-x, 1-538.

Romer, A. S. 1955. Vertebrate paleontology. Chicago: University Press.

Romer, A. S. 1956. Osteology of the reptiles. Chicago: University Press.

SAVE-SODERBERGH, G. 1936. On the morphology of the Triassic stegocephalians from Spitz-

bergen and the interpretation of the endocranium in Labyrinthodontia. K. svenska Vetensk.

Akad. Handl. (3) 16 (1): 1-181.

SHANER, R. F. 1926. The development of the skull of the turtle, with remarks on fossil reptile

skulls. Anat. Rec. 32: 343-367.

Suuno, K. 1914. Studien zur Kenntnis des Wirbeltierképfes. 1. Das Chondrocranium von

Crocodilus mit Beriicksichtigung der Gehirnnerven und der Kopfgefasse. Anat. Hefte

Abt 1. 50: 253-382.

426 ANNALS OF THE SOUTH AFRICAN MUSEUM

SIEBENROCK, F. 1895. Das Skelet der Agamidae. S.B. Akad. Wiss. Wien. Abt. 1. 104: 1089-1196.

Susuxin, P. O. 1899. Schadel von Tinunculus. Nouv. Mem. Soc. Nat. Moscou. 16: 1-163.

Terry, R. J. 1917. The primordial cranium of the cat. 7. Morph. 29: 281-433.

T6piitz, C. 1920. Bau und Entwicklung des Knorpelschadels von Didelphis marsupialis. Zoologica,

Stuttgart. 70.

Vorr, M. 1909. Das Primordialkranium des Kaninchens unter Beriicksichtigung der Deck-

knochen. Anat. Hefte. Abt. 1. 38: 425-6016.

Watson, D. M. S. 1919. The structure, evolution and origin of the Amphibia— The ‘Orders’

Rachitomi and Stereospondyli. Phil. Trans. (B) 209: 1-73.

Watson, D. M.S. 1926. The evolution and origin of the Amphibia. Phil. Trans. (B) 214: 189-257-

Watson, D. M. S. 1937. The acanthodian fishes. Phil. Trans. (B) 228: 49-146.

Watson, D. M. S. 1948. Dicynodon and its allies. Proc. zool. Soc. Lond. 118: 823-877.

Watson, D. M. S. 1954. On Bolosaurus and the origin and classification of reptiles. Bull. Mus.

comp. ool. Harv. 111: 297-449.

INSTRUCTIONS TO AUTHORS

MANUSCRIPTS

In duplicate (one set of illustrations), type-written, double spaced with good margins,

including Taste 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 4} in. x 7 in. (7# 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 (bold 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:

SmitH, 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:

SmitrH, 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, Fuly 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).