Parallelism in the evolution of the Permian reptilian faunas of the Old and New Worlds

a*

II E> R.ARY

OF THE
UNIVERSITY
Of ILLINOIS

590.5

FI
v. 37-38

BIOLOGi

^,est Date stamped below

L161— O-1096

Parallelism in the Evolution of the Permian
Reptilian Faunas of the Old and New Worlds

Everett C. Olson1
Research Associate, Division of Paleontology

INTRODUCTION

When vertebrate remains were first encountered in the San
Angelo Formation of Texas, we could not avoid speculations about
resemblances of the newly found reptiles to mid-Permian forms of
Russia and South Africa. Doubtless part of the speculation was
prompted by a subconscious feeling that the fossils from these
areas should show resemblances. It soon became apparent, however,
that in size, at least, some of the North American reptiles did
rival the largest of the relatively gigantic forms of the Old World
mid-Permian. A more thorough study has shown that there is a
notable similarity of adaptive types and that evolution, which
appears to have been independent in the Old and New Worlds
during the Permian, has produced a number of interesting paral-
lels. Furthermore, it would appear that there is a rough time
equivalence of the rocks in which these adaptive counterparts
appear in the Old and New Worlds.

By the time that deposition of the San Angelo beds of north
central Texas was taking place, the old, familiar, earlier Permian
deltaic fauna had disappeared, at least in the Texas area, and had
been replaced by a strikingly different assemblage. At about the
same time, in Russia and South Africa, there appeared assemblages
unknown from any earlier deposits. The Old World faunas have
been studied for many years and described and discussed in nu-
merous publications. The San Angelo forms have been known but
a short time. Their general structure and relationships to earlier
reptiles were considered in some detail in an earlier paper (Olson

1 Professor of Vertebrate Paleontology, University of Chicago

385

386 FIELDIANA: ZOOLOGY, VOLUME 37

and Beerbower, 1953). Field work in 1953 and 1954 has added
greatly to the knowledge of the described genera but has increased
the generic list of seven only by one or two possible additions. In
1953 a find was made in the Flower Pot Formation, which directly
overlies the San Angelo Formation in Texas (Olson, 1954). The
specimens from this formation, with one or two exceptions, appear
to be congeneric with members of the San Angelo assemblage. Field
work has now been sufficiently extensive that it appears unlikely
that the list of genera will be materially increased by additional
work in the Texas area. It is thus a rather small array of genera
that provides a basis for comparisons with the more extensive as-
semblages of Russia and South Africa. Even so, the comparisons
yield much that is of interest in the areas of evolution and zoo-
geography. It is the principal purpose of this paper to outline the
evolutionary history that led to the development of these assem-
blages, to compare the assemblages, and to attempt to evaluate
their meanings.

There are, of course, many problems in such a study, and the con-
clusions must be considered in light of the difficulties. There is
the problem of correct evaluation of the role of each species in the
economy of the ecological system, the problem posed by the lack of
information about the totality of the systems, the problem of corre-
lation between continents, and that posed by the incomplete nature
of the record of ancestral stocks. An effort has been made through-
out this paper to point out the specific nature of these difficulties
and the ways in which they have affected interpretations.

THE GROUPS OF ANIMALS CONSIDERED

Since the San Angelo and Flower Pot assemblages are almost
exclusively reptilian, comparative studies are restricted to this class
of animals. The major groups and pertinent subgroups that play
an important part in the study are given in the classification on
page 387.

The Diadectomorpha and Captorhinomorpha are often grouped
into a single order of primitive reptiles, order Cotylosauria, in spite
of the fact that the association has been repeatedly questioned by
prominent students in the field. In the present analysis, the two
groups play very different roles, and for this reason, if no other, they
must be considered separately. Early captorhinomorphs were small
and were lizard-like in general proportions. They appear to have
been fast-running, carnivorous to omnivorous, terrestrial animals

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 387

Class Reptilia

Subclass Parareptilia1
Infraclass Diadecta

Order Diadectomorpha
Family Diadectidae
Family Pareiasauridae
Family Procolophonidae
Subclass Eureptilia

Infraclass Captorhina

Order Captorhinomorpha
Suborder Captorhinomorpha
Infraclass Synapsida
Order Pelycosauria

Suborder Ophiacodonta
Suborder Sphenacodontia
Suborder Edaphosauria
Family Edaphosauridae
Family Caseidae
Order Therapsida

Suborder Titanosuchia
Suborder Tapinocephalia
Suborder Anomodontia
Suborder Theriodonta
Infraorder Gorgonopsia
Infraorder Therocephalia

and to have played an important role in the early faunas. Later in
their history, large herbivores, up to five or six feet in length,
developed. Diadectes and other closely related genera represent the
Diadectomorpha in the late Pennsylvanian and early Permian.
They were heavy-bodied, slow-moving herbivores. They were suc-
ceeded later in the Permian by the massive, herbivorous pareia-
saurs, and the lightly built, small procolophonids. The latter appear
to have been adaptively similar to some of the smaller and earlier
herbivorous captorhinomorphs.

Pelycosaurs were primitive mammal-like reptiles. The earliest,
the ophiacodonts, appear to have arisen directly from primitive
captorhinomorphs and to have accentuated the carnivorous poten-
tialities of the ancestral stock. From the ophiacodonts came the
active, highly predaceous sphenacodonts, of which Dimetrodon is
the best known representative. The source of the edaphosaurians
is less certain. The genus Edaphosauria, genotype of the family
Edaphosauridae, occurs very early, near the beginning of the pely-
cosaur record. The caseids appeared first in the early Permian and

1 Classification after Olson (1947).

388 FIELDIANA: ZOOLOGY, VOLUME 37

became prominent only in the upper beds of this time. Both
families consisted in large part of rather ponderous, slow-moving,
terrestrial to semi-aquatic herbivores.

The therapsids, which were advanced mammal-like reptiles,
probably arose from the sphenacodonts. There seems to be no doubt
that this was the origin of the titanosuchids and theriodonts. The
source of the tapinocephalids and anomodonts is less certain, but a
sphenacodont ancestry seems most probable. If this was the case,
later sphenacodont history set the stage for radiation into herbiv-
orous modes of life as well as for increased adaptive spread in the
carnivorous realm. This concept is important in the interpretation
of the early radiation of therapsids and in evaluation of the dif-
ferences between Old and New World radiations in the mid-Permian.

The fact that only reptiles can be considered in our comparisons
limits the interpretation of faunal evolution. Undoubtedly both
fish and amphibians, among the vertebrates, invertebrates, and
plants, played a role in the modifications of the reptiles. Xenacanth
sharks, primitive actinopterygians, dipnoi, and crossopterygians,
among the fish, and lepospondylous and apsidospondylous amphib-
ians are components of the Permian vertebrate faunas. The roles,
both minor and major, that they have played in a Permian chrono-
fauna have been discussed for the Clear Fork of North America
elsewhere (Olson, 1951). It may be assumed that they were equally
effective in less well-studied faunas and that their impact was in
large part a function of the adaptive types present in a given area.
Fortunately for our purposes, the taxonomic groups were widely
dispersed in the Pennsylvanian and early Permian, and similar adap-
tive types within the higher categories occur in the various areas
that have been studied. Thus, the fish and the amphibians
probably played rather similar roles in the economy of contemporary
ecological structures in the areas with which we are concerned.
There seems fair justification for an assumption that, in a broad
sense, the effects of fish and amphibians were more or less constant
in the various areas during the time considered and that differences
in reptilian evolution were dependent largely on other factors.
Knowledge of directly associated invertebrates during this time is
very slight. Understanding of the plants is somewhat better, but no
co-ordinated studies to evaluate the effects of botanical differences
have been made. It is beyond the scope of this report to attempt
an evaluation of the effects of differences in the plants, but it seems
probable that they were of considerable importance.

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 389
GEOLOGICAL AND GEOGRAPHICAL DISTRIBUTION OF REPTILES

The pertinent data are summarized in Table 1. There are,
however, certain problems of contemporaneity of faunas that need
to be considered, and important points of spatial and temporal
distribution that require special emphasis. These can be treated by
a systematic summary of the temporal sequences:

1. Late Pennsylvanian (Stephanian) : The striking resemblances
between the North American and European middle to late Stephanian
vertebrate faunas have been discussed in some detail by Romer
(1945) in his comparison of the North American faunas with that
from Kounova in Bohemia. The general aspects of these faunas
show clearly that they can be considered ancestral to the early
Permian faunas of both North America and Europe. The similarities
between the two continents are so great that it seems highly probable,
as Romer argued, that there was direct and intimate connection
between the two areas. Whether or not this type of fauna spread
beyond the limits of the two continents is not known.

2. Early Permian (Autunian): The most complete vertebrate-
bearing early Permian section in North America is found in the red
beds of north central Texas. An excellent fauna has also come
from the Abo (Cutler) of New Mexico, and a small, but most in-
teresting series of assemblages is known from the Dunkard Formation
of Ohio, West Virginia, and Pennsylvania. The Texas section may
be extended into Oklahoma, where sporadic finds indicate a verte-
brate fauna similar, in many respects, to that of Texas, but different
in some important details.

The Texas section and probable time equivalents of the Abo
and Dunkard are shown below:

Texas New Mexico Ohio, West Virginia,

Pennsylvania

( Choza

Clear Fork Group J Vale

' Arroyo
Clyde

Belle Plains

Wichita Group J Admiral ^ Dunkard

Putnam
( Moran ? ?

The faunas of Wichita age in Texas, New Mexico, Ohio, West
Virginia, and Pennsylvania, while similar in higher categories, show
many detailed differences. They indicate a geographic variation

390

FIELDIANA: ZOOLOGY, VOLUME 37
Table 1.— DISTRIBUTION OF REPTILES

North America

Western Europe

Russia

Africa

Late

Permian

(Thuringian)

Therapsids

Therapsids

Therapsids

Diadectomorphs

Diadectomorphs

Diadectomorphs

Pareiasaurids

Pareiasaurids

Procolophonids
Pareiasaurids

L

Therapsids
Titanosuchids

Therapsids

Titanosuchids

M

Tapinocephalids
Anomodonts

Tapinocephalids

Pelycosaurs
Sphenacodonts
Edaphosaurians
Caseids

Captorhinomorphs

'Therapsids
Anomodont-like

Anomodonts

Middle

Pelycosaurs

Theriodonts

Permian
(Sazonian,

Ophiacodonts
Diadectomorphs

Gorgonopsians
Therocephalians

Kazanian)

E

Procolophonids
Pareiasaurids

Pelycosaurs

Sphenacodonts
Diadectomorphs

Pareiasaurids

?Captorhinomorphs

C

Pelycosaurs

C

1

h
o

z

Sphenacodonts
Edaphosaurians
Caseids

e
a

a

Captorhinomorphs

Pelycosaurs

r

V

Sphenacodonts
Edaphosaurians

F

a
1

Caseids
Captorhinomorphs

o

e

Diadectomorphs

r

(basal only)

k

Diadectids

Early

Permian

Pelycosaurs

(Autunian)

G

r
o
u
P

A

r
r
o
>•
o

Ophiacodon's
Sphenacodonts
Edaphosaurians
Edaphosaurids
Captorhinomorphs
Diadectomorphs

Diadectids

Pelycosaurs

Pelycosaurs

W

Ophiacodonts

Sphenacodonts

i G

Sphenacodonts

Edaphosaurians

c r

Edaphosaurians

Edaphosaurids

h o

Edaphosaurids! Diadectomorphs

i u

?Caseids Diadectids

t p

Captorhinomorphs

a

Diadectomorpha

Diadectids

Pelycosaurs

Pelycosaurs

Ophiacodonts

Ophiacodonts

Pennsylvanian

Edaphosaurians

Sphenacodonts

(Stephanian)

Edaphosaurids
Captorhinomorphs

Edaphosaurians

Edaphosaurids

Captorhinomorphs

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 391

in North America, during the early Permian, as great as that be-
tween North America and Europe at the beginning of the period.

The only known continuous sequence of early Permian verte-
brates is from Texas, and it is from this area that most of our
knowledge of faunal evolution in the early Permian has come. There
is a recognizable, reasonably coherent fauna throughout the early
Permian, but there are important changes in composition and these
have an important bearing on the constitution of the mid-Permian
assemblage. The most important among the changes are as follows:

Captorhinomorpha: Very primitive, carnivorous to omnivorous
genera occur in the lowest part of the section (Proterothyris, Romeria) .
These are replaced by the well-known genus Captorhinus in the
middle Wichita. A larger genus, Labidosaurus, is present with
Captorhinus during the Arroyo. With the inception of the Vale,
the captorhinomorphs began a rather extensive adaptive radiation
that produced herbivores found in the Vale and Choza. There are
four genera known from the Vale: Captorhinus, Captorhinikos,
Captorhinoides, and Labidosaurikos. All except Captorhinoides occur
in the Choza. It is significant that after the very early Wichita
all captorhinomorphs were either omnivores or herbivores and that
the early carnivorous potential was not realized within the order.

No captorhinomorphs are known in post-Stephanian times in
Europe. The record during the Autunian is poor, yet it seems
probable that had the captorhinomorphs undergone a development
comparable to that in the New World some record would have come
to light. This absence, if real, may have had an important bearing
on the evolution that produced the mid-Permian fauna of Europe.

Diadectomorpha: In the Texas beds there is but a single genus
of this group, Diadectes. Diasparactus, Diadectes, and possibly a
third genus are known from New Mexico, and Desmatodon is present
in the Dunkard. So far as is now known, Diadectes did not persist
beyond the lowest Vale time. At least two genera of diadectids are
present in the Autunian of western Europe. Presumably this family
formed an important element of the fauna.

Pelycosauria: Ophiacodonts are prominent during the Wichita
and persist in very reduced numbers into the early Clear Fork.
Thereafter they are unknown. These primitive pelycosaurs probably
came from the early captorhinomorphs and have emphasized the
carnivorous potentials of that group. Sphenacodonts persist into
the mid-Choza, into the youngest vertebrate-producing beds of the
Clear Fork. There is considerable diversity of genera and species

392 FIELDIANA: ZOOLOGY, VOLUME 37

in the Wichita and the Arroyo, but thereafter only a single species,
Dimetrodon gigashomogenes, is present.

The two families of the edaphosaurians have very different
histories and must be considered separately. There is diversity of
the Edaphosauridae in Wichita times. One genus, Edaphosaurus,
persists through the Arroyo. A single specimen of the family
Caseidae is known from the Abo of New Mexico, but the family is
not known from the Texas section until the beginning of the Vale.
Three temporally successive species of Casea occur in the Vale and
Choza. A second genus, Cotylorhynchus, is present in the Hennessey
Formation of Oklahoma, a Vale or Choza equivalent.

Sphenacodonts and edaphosaurids are known from the western
European Autunian. As in North America, the genus Edaphosaurus
is present, but in Europe this genus alone represents the family.
Three genera of sphenacodonts are recognized by Romer and Price
(1940). Haptodus, which is at about the same level of development
as Sphenacodon and Dimetrodon, is much the best known. Caseids
have not been found in the Old World.

The fauna of the European Autunian resembles that of the very
early Permian of North America in the presence of Edaphosaurus
and sphenacodonts. The case for continuity of the continents,
however, is not particularly strong, for the two faunas could have
been derived independently from the common Stephanian fauna.
Also, the degree of divergence of the mid-Permian faunas of the
Old and New Worlds indicates that separation had occurred well
down in the early Permian.

No other Autunian faunas have been described, although Price,
in informal communication, has indicated that there may be an
Arroyo equivalent in Brazil. The latest early Permian assemblage
in North America includes sphenacodonts, caseids, and captorhino-
morphs. Of these, only the first is known from the early Permian
of Europe.

3. Mid-Permian. The mid-Permian faunas of Russia and
South Africa differ markedly in constitution from the earlier Permian
assemblages of either North America or western Europe. Therapsids
and diadectomorphs are the principal reptiles. The diadectomorphs
are pareiasaurs and procolophons, rather than diadectids. There
are only remnants of pelycosaurs. No captorhinomorphs are known
from Russia. One family, Milleretidae, has been reported from
South Africa, but, as Broom (1938), who described the material,
has noted, the assignment is open to doubt.

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 393

The North American assemblage. — The families of reptiles in the
North American mid-Permian are the same as those in the late
Clear Fork, except for the family of one genus, Dimacrodon, which
has therapsid affinities. The change, at the generic level, is striking.
It seems improbable that the mid-Permian fauna was derived directly
from the known fauna of the Choza. Of the reptiles, only Coty-
lorhynchus, which is present in the Hennessey of Oklahoma, has been
found in earlier beds. Most striking is the strong tendency toward
large size. There existed forms that approximate pareiasaurs,
titanosuchids and tapinocephalids in general dimensions and bulk.
There appear to be strong resemblances in the types of changes
and adaptations in the New and Old Worlds, but changes that
stemmed from very different immediate ancestors. The reptiles of
Russia and South Africa have been thoroughly described in such
general works as Romer (1945), Broom (1932), von Huene (1940),
and Olson (1944). The North American materials have been but
recently discovered and have only been described briefly (Olson
and Beerbower, 1953). The salient features of this fauna are as
follows :

Seven genera of reptiles have been identified from the San
Angelo and Flower Pot Formations of Texas. In addition, xenacanth
sharks and a problematic amphibian jaw have been found. The
ranges and numbers of specimens are shown in Table 2. It has
been noted that the mid-Permian reptiles were strikingly larger
than the largest of their immediate predecessors. Table 3 gives
examples of the differences, based on comparisons of available
materials.

The identified genera and their principal characters are as follows :

Rothia: This is a large captorhinomorph which probably ran
from five to six feet in length. The skull was heavy and the palate
was set with long, irregularly spaced teeth. The animal was un-
doubtedly a herbivore and probably was somewhat like the edapho-
saurids in habits. There is no certain ancestor among known
captorhinomorphs.

Steppesaurus: This is a large sphenacodont, about one-fifth to
one-fourth larger than the largest known specimens of Dimetrodon.
Like Dimetrodon, it had greatly elongated neural spines. The spines
and general structure suggest that it arose from Dimetrodon. Dime-
trodon was widespread through Oklahoma and Texas during the
Clear Fork and appears to have crossed boundaries between adjacent
ecosystems.

s

o

< 23

z x

o o

Oh CQ

£ 2

° 3

M g
W

PQ
W

>

PQ
Eh

B

M

W^H

OPh

fa

g

s

C

O

c
.£ 'E3

'So rt

s s

S a

C e8
O £

CO

OS 03 «* ototwjto .^ +^ cC

"3. "3, "3. "3. "3, "3. "3, Eh.oo-j-

g .g

£2 J2
3, a

o o o

o .2 .2

.S .£ .5 .£

cd OS oj OS

a, "H. — —

t "O T) "O

o o o o

o o o o

CJ3 q3 CR "G

2

ft "X

0)

JO

cp

a

o?*

o oS
E

2 §

os B

0)

CO

+J f-.

]2S

rt ^3

bfl E »h

e« g oj ©

T3 g T3 g
— O o> oj

co co oS c3

l, in bo M «-, t-

* #

^ O O CM O H O t-h CM CM

o

►J

^ CO O M H Oi CO i-H © O

T3 © »— I CM CM iH © © lO ,— I

^oooo o oooo
o

03 05

S A

^ 5

a c

C c e s

* r $ 55

(5.

E-i

If II

Q O O ^

M

tx

394

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 395

Tappenosaurus: This carnivorous reptile was the giant of the
fauna. It probably stood about 5 feet at the shoulder and, assuming
a sphenacodont type of tail, was at least 18 feet in length. What
is known of the skull indicates sphenacodont affinities, but there is
no known ancestor in the earlier Permian.

Cotylorhynchus: This genus occurs in the Hennessey Formation
of Oklahoma. The San Angelo-Flower Pot representative is about
one-fourth larger than the Hennessey species. Cotylorhynchus was
a herbivore of caseid affinities.

Caseoides: This genus is somewhat larger than Casea halselli
from the Choza but could well have been derived from the genus
Casea. It is one of the commonest genera in the San Angelo-Choza
and probably was an important element of the fauna. Its adaptive
characters appear to be similar to those of the late Vale and Choza
species of Casea, for it was evidently a slow-moving, herbivorous
inhabitant of flood plains and margins of bodies of standing water.

Angelosaurus: This genus has been tentatively assigned to the
Caseidae. It was a large, clumsy herbivore with short limbs and
relatively enormous feet. The body structure tended to be massive
throughout. There is no known ancestor of Angelosaurus in the
earlier Permian.

Dimacrodon: This genus has been tentatively assigned to the
Therapsida. Unfortunately it is known largely from jaws and skull
fragments. It has a pair of enlarged "canine" teeth and small
post-canine teeth. There were no "incisors" and the broad sym-
physis of the lower jaw probably was covered by a horny beak.
There are no certain ancestors in the earlier Permian. It is quite
possible that this "therapsid" was from a sphenacodont stock and
that it evolved independently of the Old World therapsids.

There are uncertainties in the correlations between the three
areas that are our chief concern and at present there is no way of
resolving the difficulties. Presumably the greatest hope lies in the
use of botanical remains, but the materials necessary have not as
yet been obtained. The North American deposits have been placed
in the mid-Permian on the basis of local stratigraphic and faunal
evidence (see Olson and Beerbower, 1953). The same in general is
true for the Russian and South African deposits. The North
American deposits are certainly no younger than earliest mid-
Permian, but they lie above typical terrestrial Leonard deposits.
This assignment compares favorably with that made for Zone I
(Rhopalodon Zone) and, perhaps, Zone II (Titanophoneus Zone) of

396

FIELDIANA: ZOOLOGY, VOLUME 37

Table 3— MEASUREMENTS OF GENERA OF SAN ANGELO

REPTILES AND COMPARATIVE MEASUREMENTS OF

VARIOUS GENERA FROM EARLIER PERMIAN

Based on largest known specimens from both San Angelo
and earlier Permian ; measurements in centimeters.

Rothia

Genus

Probable total
length

Skull
length

Rothia

Labidosaurus .
Labidosaurikos
Captorhinus. .
Captorhinikos .

165

90

165

35

50

Steppesaurus

26
20
30

8
10

Genus

Jaw length

Ulna length

Steppesaurus .
Dimetrodon . . .

59

39

27
20

Genus

Est. total
length

Tappenosaurus

Femur Humerus Pelvis
length length height

Centrum

length
dors. vert.

Tappenosaurus .
Dimetrodon

540

328

50 45

27 24

55

7.6
5.2

Genus

Dimacrodon

Jaw length

Dimacrodon

(no comparative material)

20 (est.)

Cotylorhynchus
Genus and species

Humerus
length

Cotylorhynchus
Cotylorhynchus

hancocki (San Angelo) . . 39 . 0
romeri (Hennessey) .... 31.5

Caseoides

Genus and species

Humerus
length

Femur
length

Prox. width
femur

Caseoides

Casea broilii (Lower Vale)

Casea nicholsi (Upper Vale) ....
Casea halselli (Middle Choza) . .

15.7

. 7.2
10.0

14.5
8.5

8.5

7.3
2.5
3.9
6.3

Humerus
Genus and species length

Angelosaurus

Centrum
Femur length
length dors. vert.

Pubo-isch.

symphysis

length

Angelosaurus. . .
Casea nicholsi . .
Casea halselli . . .

18.9
10.0

21.6
8.5

3.9
2.3
2.9

19.5
12.2
19.2

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 397

Russia. The earliest Russian beds in which vertebrates are found
are generally considered slightly older than the earliest producing
beds of the African Beaufort Series (Tapinocephalus Zone). This
relationship is based largely on local stratigraphic evidence and upon
the contained faunas. More important than the precise time rela-
tionships between the areas, however, is the fact that more or less
equivalent evolutionary levels have been reached in Russia and
South Africa and that a somewhat similar adaptive stage is recorded
in the San Angelo and Flower Pot Formations of North America.
There is an event equivalence, even though some time differences
may exist.

EVOLUTIONARY RESEMBLANCES AND DIFFERENCES BETWEEN THE
MID-PERMIAN OF THE OLD AND NEW WORLDS

If the taxonomic differences between the mid-Permian reptiles
of the Old and New Worlds are ignored, and adaptive types con-
sidered, striking resemblances become apparent. Most obvious is
the comparable large size of a number of the animals in both areas.
This relative gigantism is spectacular in North America, coming as
it does shortly after the disappearance of the typical, deltaic fauna
of the early Permian, where proportions are more modest. The
giants of the earlier Permian were Dimetrodon and Cotylorhynchus.
The marked increase of the mid-Permian Steppesaurus and Coty-
lorhynchus over their predecessors has been noted. More amazing
is the great size of Tappenosaurus. Equal size is recorded in titano-
suchids, tapinocephalids, some theriodonts, and the pareiasaurs in
the mid-Permian of Russia and South Africa. Here the size seems
less striking, in the lack of knowledge of earlier faunas.

Resemblances go far beyond the mere coincidence of size, and
extend to more specific adaptations. The adaptive resemblance of
Tappenosaurus to the titanosuchids is clear. It can hardly be denied
that they represent parallel trends from a common, but somewhat
remote sphenacodont ancestry. The analogy between Steppesaurus
and such theriodonts as the gorgonopsians and early therocephalians
is evident. Steppesaurus is an adaptive relict in the mid-Permian,
for it changed little, in an adaptive sense, from Dimetrodon of
earlier times, although it was involved in a very different ecological
situation. The stability of Dimetrodon itself under changing con-
ditions is well shown by the persistence of the genus for nearly the
full span of the early Permian, and the persistence of a single species,
D. gigashomogenes, through most of the Clear Fork. A similar genus,

398 FIELDIANA: ZOOLOGY, VOLUME 37

Steppesaurus, appears to carry on this tolerance, as it occupies a
theriodont-like place in the mid-Permian fauna. The two known
carnivores of the San Angelo and Flower Pot find close parallels in
the principal carnivores of the mid-Permian of Russia and South
Africa, the titanosuchids and the theriodonts.

The situation with respect to herbivores is more difficult to
visualize, for there is less actual resemblance in form, and evidence
of the ecological positions in the economy of the fauna must be the
principal basis for judgment. The closest morphological resem-
blance, so far as known structures are concerned, is between Dimac-
rodon and some of the early anomodonts. In the broad symphysis
of the jaws, the presumed horny beak, the large "canines" and
small post-canines, they are very similar. It seems highly im-
probable, from the differences in the rest of the faunal elements
and geographic remoteness, that there was anything but a rather
remote common ancestry, so that this would appear to be a valid
case of parallelism, or, perhaps, convergence.

The remaining herbivores are captorhinomorphs or caseids in
North America, in contrast to the anomodonts, tapinocephalids,
pareiasaurs, and procolophonids in the Old World. Angelosaurus
and Cotylorhynchus were large, slow-moving vegetarians that must
have provided an important food supply for the large carnivores.
In this sense they were associated with the vegetation on the one
hand and the predators on the other in much the same way as were
the tapinocephalids and pareiasaurs. Neither of the North American
genera had a skeletal structure as massive as that of these two types,
nor had they assumed as upright a pose. Nevertheless, they repre-
sent marked departures from the earlier herbivores of Texas and
their differences appear to be directly related to their place in the
changed ecological situation. The development of the pareiasaurs
and tapinocephalids similarly can be presumed to have taken place
during the evolution that led to the large Old World carnivores.

Rothia and Caseoides, both of which are more frequently en-
countered in the deposits than the larger herbivores, fill the position
of the smaller plant-eaters, occupied in large part by the anomodonts
in the South African and Russian mid-Permian.

Here the analogy ends, for we are faced with the puzzling fact
that at present almost no traces of animals smaller than Rothia
have been found in the San Angelo or the Flower Pot. This is true
in spite of the fact that types of deposits, which, in earlier sequences,
have yielded an abundance of small vertebrates, have been explored

OLSON: EVOLUTION OF PERMIAN REPTILIAN FAUNA 399

thoroughly. During the last year, beds in swamps and swamp
margins that were rich in large vertebrates and in plants have been
combed, without the discovery of small reptiles, amphibians, or fish.
These studies have led to the interpretation of the San Angelo and
Flower Pot faunas as marginal phases of better-integrated faunal
units that included all elements necessary to a balanced economy
(Olson and Beerbower, 1953). There are no counterparts of pro-
colophonids, small anomodonts, or small theriodonts in the San
Angelo-Flower Pot fauna.

ORIGIN AND EVOLUTION OF THE NEW AND OLD WORLD
MID-PERMIAN FAUNAS

A coherent interpretation of the origin and evolution of the
faunas requires certain assumptions. It is assumed that the faunas
in the Old and New Worlds arose from a common ancestral Penn-
sylvanian stock. This stock is considered to have been spread,
with continuity, over both the Old and New Worlds. Such evidence
as we have concerning the Pennsylvanian supports this hypothesis.
It is assumed, further, that the common characteristics of the fau-
nas of the Old and New Worlds appeared at more or less the same
time, during the late part of the early Permian. We know that the
elements had not appeared in North America prior to this time and
such evidence as there is of the early Autunian of Europe gives no
suggestion of their presence in Europe. A coincidence of conditions
favoring increased size and other adaptive similarities and evolu-
tionary thresholds that permitted exploitation of these conditions
must have occurred in both areas. In other words, in both the Old
and New Worlds there existed in the late part of the early Permian
comparable opportunities and stocks able to take advantage of
these opportunities. While the opportunities may well have been
much the same, the stocks seem to have been very different.

What is known of the early Permian faunas in western Europe
gives some indication that differences may have been initiated by
that time. So far as we know there were no caseids and no capto-
rhinomorphs present. These, however, are dominant elements of
the late early Permian in North America, and captorhinomorphs
were abundant during all of early Permian. Sphenacodonts,
edaphosaurids, and diadectids were present in the early Permian of
the Old World. In North America diadectids and edaphosaurids
gave way to the caseids and captorhinomorphs during the late part
of the early Permian. In the Old World the edaphosaurids failed,

400 FIELDIANA: ZOOLOGY, VOLUME 37

but the procolophonids and pareiasaurs, presumably diadectid
descendants, give indication that this group was successful.

The predominant reptilian types of the early Permian in western
Europe were the sphenacodonts. It is from this stock that the
therapsids appear to have stemmed. If this is the case, there must
have been a rapid and pronounced adaptive radiation of this group
late in the early Permian, to produce both herbivores and carnivores.
In the absence of captorhinomorphs and caseids, which maintained
a balance with carnivorous sphenacodonts in North America, such
a radiation would be a most probable event. The only reptilian
competition to the herbivore radiation would have been supplied
by the relatively unprogressive diadectomorphs. The radiation
would compare in part to the rapid adaptive changes in the capto-
rhinomorph stock in North America, a radiation that appears to
have filled the gap left by the disappearance of the diadectids and
edaphosaurids. It would compare as well, once herbivores had
been established, with the caseid radiation that began in North
America during the Vale.

The difference in stocks of the herbivores may be supposed to
have had an important impact on the carnivores. In the Old World,
under the hypothesis advanced above, they stemmed from the
advanced and highly active sphenacodonts, whereas in North
America, except for Dimacrodon, herbivores developed from the
relatively unprogressive captorhinomorphs and caseids. Selective
pressure for activity in the Old World must have been far in excess
of that in the New. The keynote to the origin and evolution of the
therapsids and finally to evolution of the mammals, appears to have
been continued emphasis on activity. Of course there were many
bypaths, in which ponderous and slow therapsids developed, but
fundamentally the level of activity in a therapsid-dominated fauna
must have been higher than that in one where pelycosaurs played
the primary role.

We do not as yet know, and may never know, the history of
reptiles after the early mid-Permian in North America. There is
little indication in what is known of the early mid-Permian fauna
that it was likely to approach and pass the therapsid threshold
necessary to the radiation seen in the Old World. As long as the
isolation, which must have existed to preserve the pelycosaur-
captorhinomorph lines in North America, persisted, adaptive modi-
fications of this stock without "progressive" modifications of funda-
mental structures would appear to be the most probable course
of events.

REFERENCES

Broom, R.

1932. The mammal-like reptiles of South Africa and the origin of mammals.

London. 376 pp.
1938. On a new type of primitive fossil reptile from the upper Permian of

South Africa. Proc. Zool. Soc. London, (B), 108: 535-542.

Huene, F. VON

1940. Die Saurien der Karroo-, Gonduana- und verwandten Ablagerungen in
faunistischer, biolog. scher. und phylogenetischer Hinsicht. Neues Jahrb.
Min., Geol., Pal., Beil.-Bd., Abt. B, 83: 246-347.

Olson, E. C.

1944. Origin of mammals based upon cranial morphology of therapsid sub-
orders. Geol. Soc. Amer., Spec. Papers, no. 55, 136 pp.

1947. The family Diadectidae and its bearing on the classification of reptiles.

Fieldiana, Geol., 11: 3-53.
1951. The evolution of a Permian vertebrate chronofauna. Evolution, 6:

181-196.
1954. Vertebrates from the Flower Pot Formation, Permian of Texas. Jour.

Geol., 62: 512-513.

Olson, E. C. and Beerbower, J. R.

1953. The San Angelo Formation, Permian of Texas, and its vertebrates.
Jour. Geol., 61: 389-423.

Romer, A. S.

1945. Vertebrate paleontology (2nd ed.). Univ. Chicago Press, Chicago.
661 pp.

1945. The Late Carboniferous vertebrate fauna of Kounova (Bohemia)
compared with that of the Texas redbeds. Amer. Jour. Sci., 243: 417-442.

Romer, A. S. and Price, L. I.

1940. Review of the Pelycosauria. Geol. Soc. Amer., Spec. Papers, no. 28,
537 pp.

401