foil ih i Dt Fw
Li Wen

Sabai of ail

, rary et

o athald Bw Mead beh

Nate REL ee EP OO
ete ty A ba Te hb Peder ne | ~
ete be tape be bee

rad 7!
one ene aes : se cy a, . - rr
Pete Ne Rebate VPS G5 - fies. <0 -H-P toh : cme! ( 2 ? bh Re eee , PRR ae AE NEES SS arte
RR oy PARRA RP. pei Pete tut aoe eet . . " $beh ora j wii Raha Orh= Parakey hah oe PPP teal Poems
oboe hoFc). gu Pade Pe hee Salk Pee} ~ 4 eS Oe oe 4 4 ea he : 5
ie eee “ o s

Gems <a oF ~
TPS

Pele PP ost rein Fe RI Peet at
Wiereacs, cs V-8-fn2e-tedte? at=t-¥ Pata
. BPs - fs Gee Dale -<G aoe
BR Fe fe BP gS Sete Ore Pet. aGeotat oF-0!

a es eee
~~ ~

PRP kk Roker nai im Ph
Ted 2 §- FeP- “

we he teh LetytoteP = | >
es ae
bo tm Pah De Pe De >

a j w “

iuvyug ha LIBRARIES SMITHSONIAN _INSTITUTION NOILNLILSNI

2 “ z WW |

wn” a” -_

a o = a

4 <% = <

pe. faa] =~ oO

Oo na re) me

= = ; = -!

TITUTION a NOILALILSNI _NVINOSHLIWS ~°3 IyVvug rh
S) onthe a=. =
= o . = o
5 = Lys = >
s ts SN a oa
— ae WM Qr ai

tyvudgdIT LIBRARI! ES) SMITHSONIAN INSTITETION (NOMAIT=M

ba < ? = < bs = at.

a, e : : : gL

J aH Ww a oe hehe
“oe 5 AP g Yi
ath: 2 : 8 Gf

= > = ~ >
Y F 2 7) ‘a =

TITUTION NOILNLILSNI NYVINOSHLINS S3ZIYVYEGIT LIBRARIES
w = : Ww) =
ud 77) S. = w
— ~ ’ —

: = XY = 5
< | Ss , (=
Oe “ a WS a =
= Ss. WY. S 3
ae SS = 9

IyVvudit Ww IBRARIES _ SUHSGNIANS INSTITUTION NOILNLILSNI
i ; =
~ re) Moss Oo
oe = =
ps) >. =) E
> ae eed yy > _

a "me *d 2) =

a mig ce oE

o Se A ee zZ
NOILALILSN! NVINOSHLINS SSIY¥VUEIT LIBRARIES

”- = w” = re

= << = as he

= = 4 ae

: oO oon Gs :

wn 7p) OW uy WY

Oo pe O % ac

Zz = =z ise

> = > =

a w”) z Ww

SMITHSONIAN INSTITUTION NOILNLILSNI

3 ” = ”

op) a” —_

= oc. ze x

: = : :

= a of 0

O acs Z re) es

Pa ae za aad

STITUTION — NOILALILSNI NVINOSHLINS ~°4 1yYvuad IT _tl BRARIES
2 oy: 9 a
\= ay to a
= > = >
b = b =
z D z ¥,

;uvudii _t! B RARI ES SMITHSONIAN INSTITUTION NOILNLILSNI
< + = ea * g
ax = z =
2) . SAS Sg ” GB tig Y
= ee = = >’
sp) aes w + £e —

STITUTION pn NOILNLILSNI_NVINOSHLINS Sa IYVYEIT LIBRARIES

| >
= GE) = Nyy = GE) 5 Ga™

SMITHSONIAN

NVINOSHLINS S31IY

w ?
NVINOSHLINS S3IYY
= oxy on
Si ee oe
a Load
: -
= me
2 3
SMITHSONIAN
= ‘ a
S) | =
a : =
ae | , et
- GY f=
ey _
= ao
NVINOSHLINS S31YW:
=o es o ,
< = a
= +
= ct
rs z
PO MITHSONIAN _
s =
us A
sl
_
©. =
om 9 %
Sa1uv
iy z i
3 Z
a
> ra
= =
Y
o ee
SMITHSONIAN
¢ =
= <
2k Zz
g 3 |
2 :
2 a Oe

> nw .
= Es ud :
= ASS =
a AS <x
S WY =
oO bat
a a |
= -
= a
5 Be)
= & >
7 =
= ae
NVINOSHLINS |S3 ! uy
= y
ea
cri
wm
oO
=
=
SMITHSONIAN INSTITL
Oo ys hae Se
et $ ” =
of ee pe ae bred }

*.

i s ik ety SE Rng en: TRE ce
w i “A a,
-RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3iuVvudit_LIBRAI
nf Zz a S ° ” Za
lu th ia nw SS uJ 7)
cas ve o : eA =
i... os ao a : c. oO =
a a = = - a
le 2 5 “i 5 = 2
es z a a se =
LALILSNI _NVINOSHLINS ~°4 iyvud 17 tl BRARI ES_ SMITHSONIAN,_INSTITUTION e NOILAL
- ° Oo _ Oo Z on O
= 2 - 5 2 5
> be > i > e
2 = 2 = "2 =
on w var nn” on ”
D ms es oF z on z
RARITIES SMITHSONIAN INSTITUTION NOILNLILSNI _NVINOSHLIWS,S3 iuvugit _t! BRA
, = i = tt, : = py, = RN
= é = UE 4 ae oN
Z i: 2 07 JY" © Z xe NS
a ie S : Ol:
_ NVINOSHLINS SAIYVUSIT LIBRARIES SMITHSONIAN INSTITUTION NOILNL
= “ Z ey ke Z a
= oe sc o = a
fe hee Pa t is. <
oc Cc oa Cc ce
= oO = co 2 oO
2) S ~ e) =. Oo sea
= a | = tall = —
SRARIES SMITHSONIAN INSTITUTION NOILONLILSNI NVINOSHLINS S3IYVYUdtT LIBRA
= - = a x! = £9
= oO 2) a ae Oo
: » E oN 5 tS :
“es > 2 > \ n= =
i 2 . t
! +4 >) ~ > 8 \ \ a ps)
Lon pee o ss a ae % Fe ae
ae oe Zz ee On
ILALILSNI NVINOSHLINS S3I1YVYSEIT LIBRARIES SMITHSONIAN
Ze w =< wee 72) = Ww
es = . aS = < =
4 : Pe foe 5 ie a > =e
Y, o = pip 2 = S =e
yy “fy = ee os Hp sw O 2 3
iif * 2 Gy = 2 = 2
: = > ‘ = > = >
w” = w” Zz Tp) ; Fe

BRARIES SMITHSONIAN _INSTITUTION NOILALILSNI NVINOSHLINS SJIYVUdIT

RIES
NI

Gio

ILNLILSNI NVINOSHLINS S3IuYVvugt

NOILNLILSNI
NOILALILSNI

LIB

INSTITUTION NOILN.

4

JtuYvVudsil LIBRARIES

INSTITUTION NOILNLILS
SaINYVYdIT LIBRARIES

INSTITUTION

INSTITUTION

Sa1uvudl

S

RARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SS1uYVusIT_LIBR/

w = 2) rs wn aan = "
= = = WY; = = = ff
g 2 8 be 2 g 2
z 2 ae ahha 2 a
= : = es >" = >" = oe
“” ss ee w” = w
_NVINOSHLINS S3IUVUGIT_LIBRARIES SMITHSONIAN INSTITUTION | NOILN
Zz = > =

as. ud Sell S01 uw San ws <i
2 Ye EDI GD! G%.2 ER = 6
Si WON Se ebia ie Nl COM ZAGE 2) c (OR

ZOI,6 4

ANNALS OF THE
SOUTH AFRICAN MUSEUM

ae ed
’

VOLUME 59

%

se sib

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME 59

PRINTED FOR THE
TRUSTEES OF THE SOUTH AFRICAN MUSEUM

EQ 71—1972

=)

Ss

t—

| é ae ale
ae y r J ¥

_ THE RUSTICA PRESS (re) LTD.,

: Y { ‘ A pa

NB 7 7 ¥
F ef . / a - ‘
= *
? ct
7 we
ae =.
~ a — °
5 7
— 1
SS c >
) -
- ’
2 = Ge ;
= = yee
~ <
ce = =
= >

=
f
: '
&
74
1
a
ei A iw
ee :
5 1 J
‘ >
a
Zz ~ i ‘
5) ar
+
aa
a
Z
i
= j
a
re )
5
Vive ee =
r .
Z
-
j
= i
)
~~
;
{
.
;
i
7
hy
. —_—
r
ee
}
7

@ PRINTED IN THE REPUBLIC OF souma AFRICA Any

LIST OF CONTENTS

BoonstrA, L. D.

Discard the names Theriodontia and Anomodontia: a new classification of the
Therapsida (published September 1972) a 5

Boonstra, L. D.
The early therapsids (published December 1971)

Harcu, E. H.

Development of Trachurus trachurus (Carangidae), the South African maasbanker
(published May 1972) ait a

Haicu, E. H.

Larval development of three species of economically important South African fishes
(published March 1972) , i. bi si

HENpDEY, Q.B.

The evolution and dispersal of the Monachinae Sn gays Spa ah sige
March 1972) ; : ;

HeEnveEy, Q.B. & REPENNING, C. A.
A Pliocene phocid from South Africa (published March 1972)
HEnNpDEY, Q.B.
A Pliocene ursid from South Africa (published March 1972)
Hooyer, D. A.
A Late Pliocene rhinoceros from aie cvomr eae mate Province publiened August
1972) . i ae : : ee
McKenzig, K. G.
A new species of Paradoxostoma (Crustacea, Ostracoda) from South Africa (published
May 1972) ne ae
Maier, W.

Two new skulls of Parapapio antiquus from Taung and a suggested ppylogenere
arrangement of the genus Parapapio (published December 1971) 2

ROELEVELD, M. A.

A review of the pede Uc puateped?) of southern Africa Ae aes ee
TOFD). :

139

47

O58

7s

115

151

133

193

. ig cis sti tus a ha
ak See ua oe

ot

— =
7

oe ‘ i | ait a - ; t. dusts 2 | nk | i : ir Pe bee

—
— y 1a

#3 &

eit ible edt) secidsarit sto

aupghena sy.” | ri AD eee
; tt af 3, te oa) wna. ved mt , marley
at 4
fear ¢
bee
: -ippiadaiiclaue’ wh Howe ard .e@be servis) go dedeart. >) Cortana RM "
ey : r- ia ; : ‘e < ; ‘
TEPY area ri<) . Dtevagia & Se, TG T peat ‘aeghlen sania Na aisle
: a3 ~ yrep whllenpastl taihiaidng oo, nareny edit. eres
: 3 - te 2 - x Moa
ibeBigh® BoAsibel ugh ears” mewn cscs hy: inboagbeto) ee al edhe
=
a p

Pa 4 ,
malig ‘- _
=, . Hs
Ve —
A : Aa
a
" i" i j
, 4 : ae ay a
: ) Ow eee ere
, aa Pe eT ee
‘" a. .
> ; |
7 : af f he

. ‘

Sares
ir ny
n

. 44 ’
7 7 Ad

ff o
a

ee (

Uy

q j rit
ypb all Fes, (
Or AL

Be OTsG E

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
December 1971 Desember
Parts) Deel

TWO NEW SKULLLS OF PARAPAPIO ANTIQUUS
FROM TAUNG AND A SUGGESTED PHYLOGENETIC
ARRANGEMENT OF THE GENUS PARAPAPIO

By
WOLFGANG MAIER

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 2(1, 3> 5s 7-8), BG 2 5» i= Dis), 51-2, 5; ic)s
6(1, t-=p:i:); , TKU=2))5 8, g(1-2), 10(1),
Tot (2 5» 7s E=jO0))p 21, 24(2), 27; Bn(i=3). 33> 38

Price of this part/Prys van hierdie deel
R1,80

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum

1971

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

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG
AND A SUGGESTED PHYLOGENETIC ARRANGEMENT
OF THE GENUS PARAPAPIO

By
WoLFcaNnc MAIER
Dr. Senckenbergische Anatomie der Universitat Frankfurt a./M.

(With 1 plate, 4 figures, 3 tables)
[Ms. accepted 30 July 1971]

CONTENTS

PAGE
Introduction . ‘ : : : : I
Description . : : : . 2 “ 2
Discussion ; : d , : . ‘ 4
Classification of the genus Parapapio . , : II
Summary : : : i : : ; 14
Acknowledgements . ‘ : : : ; 14
References : : : ‘ : : : 14.

INTRODUCTION

The travertine caves near Taung, Cape Province, are the type locality
of Australopithecus africanus. In addition, some Pleistocene cercopithecoids
and a considerable number of other fossil animals were found there. The fossil
material, however, was not collected systematically and is not as well known
as that of the australopithecine-bearing dolomite caves of the Transvaal.
Peabody (1954) has compiled a faunal list for the Taung sites.

The fossil cercopithecoids from ‘Taung were first mentioned by Haughton
(1925), who proposed the name Papzo antiquus for the material available at
that time. Gear (1926), describing these and additional new specimens in more
detail, distinguished another species, Papio izodi. Broom (1940) included both
species in the genus Parapapio, which had been created by Jones (1937) for a
primitive baboon-like form from Sterkfontein. Only Freedman (1957) recog-
nized the clearcut differences between both taxa; he considered that only the
first species belonged to Parapapio, P. antiquus Haughton, 1925, while the latter
was a primitive true baboon, Papio 1zodi Gear, 1926. In 1957 Freedman also
described the colobid Cercopithecoides williamsi and the small Parapapio jonesi
from Taung, and in 1961 the same author described Papio wellsi, another true
baboon found at this site.

Unfortunately most of the Taung caves were mined out by the early
fifties and consequently have yielded no further fossil material. In 1952, how-
ever, Mr. James Kitching of the Bernard Price Institute for Palaeontological
Research, Johannesburg, was able to rescue the last few primate specimens
from the Taung dumps. The present author was kindly allowed to prepare
this material, which appeared to consist of two fairly complete female skulls

I

Ann. S. Afr. Mus. 59 (1), 1971: 1-16, 1 pl., 4 figs, 3 tables

2 ANNALS OF THE SOUTH AFRICAN MUSEUM

of Parapapio antiquus. These two new specimens were embedded in a fine-grained
pinkish breccia, which seems to be more calcified in M.3079 than in M.3078,
the hardness of the former approximately corresponding to that of the pink
cercopithecoid breccia (Brain’s Upper Phase I) of the Makapansgat Lime-
works. Both specimens are housed at the Bernard Price Institute for Palaeonto-
logical Research, Johannesburg.

DESCRIPTION

The measurements of the two new specimens are incorporated in Tables
1 and 2, which at the same time provide comparative data. Unless otherwise
stated, the technique of measuring is in accordance with the definitions of
Freedman (1957), and most of the comparative data have been extracted from
the publications of the.same author.

(1) Specimen M.3078 (Pl. 1)

This is a fairly complete and undistorted cranium with the third molars
only newly erupted and not having been in occlusion. ‘The muzzle is complete
as is the left half of the braincase; the right half of both the upper face and the
braincase together with most of the cranial base have been eroded away. ‘The
incisors, the canines and the right P? were lost before fossilization, but their
alveoli have been preserved. The remaining teeth are in relatively good
condition. The morphology and size of both skull and teeth indicate that this was
a young female of Parapapio antiquus.

The proportions of this cranium are similar to those of Tvl. 639 (Transvaal
Museum, Pretoria; Freedman 1957: Fig. 48), the muzzle being short in
relation to the braincase. The braincase itself is fairly flat in the frontal region,
but drops relatively steeply in the parietal region. The nuchal plane is therefore
situated deeply in the backward prolongation of the alveolar margins. As far
as can be seen, the mastoid processes must have been well developed, whereas
nuchal crests are absent. ‘The nuchal line runs backward as a straight continua-
tion of the jugal arch, the inion therefore being in a lowered position. The
temporal crest shows the typical course met with in other specimens of
Parapapio antiquus: it is well pronounced in its frontal part, exhibiting only a
slight notch behind the orbit; hence, it overhangs the postorbital constriction
of the lateral wall of the braincase (temporal fossa), resulting in a wide post-
orbital breadth when seen in dorsal view. On the parietal bone the faint
temporal line very gradually converges toward the midline, but approaches
to within only about 15 mm of it.

The anterior root of the jugal arch starts above the distal half of the second
molar. ‘The zygomatic part of the jugal arch is comparatively strong and broad,
showing clearly the area for the insertion of the masseter muscle. The temporal
part of the arch is narrower, but exhibits a strongly developed tubercle fronto-
lateral to the articular fossa. This fossa is remarkably deep and distinctly con-
cave transversely. Posteriorly it is bounded by a very small postglenoid process.

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG 3

Medially, the glabellar region is very undeveloped, the nasal line running
as a nearly straight continuation of the frontal outline. Laterally there exist
shallow excavations between the supraorbital arcus and the cranial vault. The
arcus are barely prominent, but possess distinct supraorbital notches. The
left orbit does not seem to be disproportionately large and is fairly well
rounded. The interorbital and nasal region show a straight contour, which is
a typical feature of Parapapio antiquus as compared with female skulls of the
other species of this genus.

The muzzle appears to be quite narrow and slender in this specimen,
because the maxillary crests are not strongly developed. Hence, although the
muzzle dorsum drops steeply towards the sides, the canine fossae are com-
paratively well excavated. There are 4 to 5 infraorbital foramina on each side,
opening separately just at the posterior end of the maxillary crests. The pre-
maxilla protrudes considerably, indicating a well-developed incisor row; the
lateral wings of the premaxilla do not reach the nasal bone. ‘The nasal aperture
shows a typical ovoid outline.

The maximum breadth of the muzzle and of the ovoid tooth arch lies
across the anterior half of the second molars. ‘The palate seems to be short, the
posterior margin lying between the last molars. The greater palatine foramina
are slit-like and they are situated between the second and the third molars.
The incisive fossa opens between the canines. The angle between the pharyngeal
face of the base of the braincase and the palate is 135°; in two new female skulls
of Parapapio broom: (M.3056 and M.3070) it is 127° and 122°.

Due to the immaturity of the new specimen, the alveolar processes are
quite undeveloped, resulting in a comparatively low facial height.

The alveoli of the (missing) medial incisors are about 5,5 mm in breadth,
but those of the lateral incisors only about 4 mm, thus indicating the specialized
broadening of the former ones. The alveoli of the canines measure about 6
by 7 mm; in the male specimen T.22 (Transvaal Museum, Pretoria) of
Parapapio antiquus these dimensions are 9 by 9 mm, proving that the present
skull is that of a female. Both premolars are well developed and comparatively
elongated. In the last two upper molars, the distal pair of cusps, and particularly
the disto-buccal cusps are conspicuously reduced in size.

(2) Specimen M.3079

This specimen is not as complete as the first one. It comprises only a
fairly well-preserved facial skeleton and the frontal part of the calvaria. The
few remaining teeth are very worn and although damaged to some degree,
show that it was a very old animal. The front teeth were lost before fossilization,
but their alveoli are still visible. The canine alveoli are comparatively small,
thus indicating that the present cranial fragment is that of a female. The skeletal
parts show some minor cracking, probably causing some slight distortion.

The muzzle of this specimen appears to be altogether heavier and stouter
than that of the first specimen. The maxillary crests are more prominent, the

4 ANNALS OF THE SOUTH AFRICAN MUSEUM

muzzle dorsum consequently being broader and more flattened, as is typical
for the species (Freedman 1957). Most breadth measurements of this frag-
mentary cranium are distinctly greater than in M.3078, whereas the length
measurements are very similar. The degree of excavation of the canine fossae
is nearly identical in both specimens. In the present cranium the orbits seem
to be more flattened and the supra-orbital arcus more developed, resulting in a
more conspicuous ophryonic groove. These features are shared with specimen
Tvl. 639, which also represents an old female. The temporal crests are very
strong in M.3079, resulting in a very great intertemporal breadth. The zygo-
matic bone of the left side is partly damaged, but appears to have been very
strong. The anterior root of the jugal arch, as in Tvl. 6309, is also situated above
the anterior part of the third molar; the tooth rows of these two old specimens
were thus shifted relatively more forward than in the younger specimen. The
same age differences are to be observed in the height of the face, the older
specimens being distinctly higher. In living primates this downward and forward
growth of the alveolar processes is well known to occur during adulthood
(Scott 1967).

There is only one premolar left in the present cranium, and this has been
partly damaged; it is fairly similar to those of M.3078 and T. 17 (Transvaal
Museum, Pretoria). The existing first molar is very worn and extruded and
thus comparatively long. However, M? and Mj? are also very long, and they
belong to the top of the known size range for this species. As far as can be
seen in both of these posterior molars, there is considerable reduction in the
breadth of the distal cusps, which is typical for Parapapio antiquus.

Discussion

The two new skulls, described above, add in many respects to our knowledge
of the fossil species Parapapio antiquus, which so far has been found only at
Taung. These two specimens confirm that the peculiar shape of the muzzle,
with its straight nasal and its well-developed maxillary crests and canine
fossae, is very characteristic for this taxon. As the material comprises a young
adult and a very old female skull, we can appreciate some of the morphological
differences due to age. In the young specimen the muzzle is more slender and
the face is narrower and less high than is the case in old specimens (Figs 1 and
2). Skull M.3078 exhibits, for the first time, morphological details of the articular
and infratemporal fossae of P. antiquus.

The lateral and oblique position of the temporal crests in P. antiquus
indicates a’ backward-orientation of the temporal muscle, which may be
correlated with some specialization of the masticatory function, i.e. a stressing
of more anterior parts of the dentition. Interestingly, some of the tooth charac-
ters of P. antiquus seem to support this kind of functional interpretation (see
below). This specific course of the temporal crests is already met with in the
comparatively young animal M.3078, whereas the typical flattening of the
muzzle dorsum is not yet evident.

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG

PARAPAPIO ANTIQUUS

M. 3078
etm ee

ee at) My 9079
Fic. 1. Pantographs of the young female specimen M.3078 and of the old female M.3079 are

superimposed to show differences due to age. Both are shown in norma dorsalis, being orientated
on the occlusal plane.

PARAPAPIO ANTIQUUS

Tvl. 639
56 604 (male)

10mm

Fic. 2. Pantographs of four of the most complete skulls of Parapapio antiquus are superimposed

in norma lateralis. Specimens Tvl. 639 and 56 604 are reversed. The picture demonstrates
an increasing degree of tooth row declination and facial height with increasing age; it shows

also the small degree of sexual dimorphism in the known specimens.

ANNALS OF THE SOUTH AFRICAN MUSEUM

*‘yessucdeyey, WIJ s[[Nys MoU OM} SUIPNOUT yy

*LG61 uewpee1,j IOV x

-— — GE z GE re VE v — 9& — VE VE of Ba gee MUTA
*qioesjUuI—"T IejosATy

(‘s sn10}—9uT] IepOoATY)

os — 6S ce rhe) j BG 6 — zQ — — 6S 1S 0eJ JO yYySOYy [eonIa\
9& re Go z Lo j Lo G — 6z Lo Go (92) Lo * —- yysroy ounjsode jeseyy
gI z GI 6 Li I Li 9 61 Li gI Li QI on * yypeoiq oinjsode jeseyy
Qz 6 Lo j Lo j Go G Go Qz Qz — Qe Vz "+ * yypRoiq [eIIqIO
Vz S Ve & Ve o 3G e 3G &% BG — 0% oS es YSU Tear@
Ge z ol & GE ref jift G ol EL 99 o (gL) (69) * Ypeoiq [eyIq4so yeuIIyxY
rh z 6 6 6 I 6 c 6 6 L — (01) 8 * + -UypeRoiq [ey1q1019}UT
&v I G& a Iv Ti 16 v — o& && — (08) (08) "+ WOTUTYYT — UOTse Ny
99 z 09 re 99 oe 6S v — (VQ) a (Z9) VS vS ‘ uoryysorg — uorAydey¢
98 6 oL & 64 G 89 Go) = 189 (14) (89) = (99) 99 "+ WOTYISOIg — UOTSeN]
ov 6. && z gs a && v — GE of GE L& of ’ eI ‘S1Op YIpPRoiq ojzznPy
Sv € gs 6 9& I 9& v GE L& — — 6& VE "eq ‘Jue YpRoiq a[zznyPy
3G v oS z gv a LY v gv — — LV 9b cP * es ‘Jue ypeoiq o[zznpY
gz j cr I &% z Li c cI 61 — gl gI Li ‘oq ‘que 7YSI0y o[zznypy
66 v 96 I 96 I 16 6 — 16 (06) — _- z6 7 + eraqepy —uoTuyT
gl z bL j afl I iL z = ol — — — (oL) * + perodura} vuNUIxe py
€S v 2S 6 LS I GG G oG bS 6S = (19) (€S) ‘yey UOA19}UT LUNUUTUTTA
oy j 99 z -— —— rh j — EL — --- — (oL) ; eljoqeyy —uorseg
99 I -— —- -— —- 6S z — reke) = - ~- (9S) ‘+ + eusa1g —UOIseg
gol re 96 v 96 I (98) I —- | — — —_ ~~ (99) ‘ ‘yi peaiq onewosdzig
Val I g6 z — — oo! I _- (oor) -—— — — — “0 oes yi Suz] [eseg
1gI g obi g (oGi)| ; i 661 - (oF1) (1% 1) — — Ger ‘ ‘ - yy Susy] wunurxeyy

soyeul 4, SO[PULOY soTeul SoTRUIoy
4 Wuo001g orgodvivg snnbyun orgodvivg

~ mooiq *g pue snnbyuv ordodoing Jo syuouoinseou [eu “I ATAV

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG vf

Although the morphology of the female skull of P. antiquus is fairly well
known, metrical data are still poor, and the present knowledge of the male
skull is very unsatisfactory (Table 1). In overall size, female skulls of Parapapio
antiquus are very similar to those of P. broomi, whereas male skulls of the latter
species are considerably larger than those of the former, suggesting a lesser
degree of sexual dimorphism for P. antiquis (Maier 1971). Figure 2 shows
a craniogram of the most complete male skull known so far (University of
California, Museum of Paleontology Specimen No. 56 604; unfortunately still
partly embedded; see Figure 4 in Freedman 1965); the muzzle is only slightly
longer and more declined than in the super-imposed female craniogram.

The relatively large numbers of teeth permit a statistical analysis to be
made. This is, however, true only for premolars and molars, the front teeth
still being virtually unknown. Table 2 provides comparative data for the
other species of Parapapio. Although the length of the tooth row is very similar
in P. antiquus and P. broomi, there seem to be some discrepancies in the dimen-
sions of individual teeth. P4-M? are distinctly longer in P. antiquus, whereas

110
105
mesial (maximum) breadth | 1
length 100 00
95
90
110
105
distal_breadth
length 100 100
95
90
P. broomi
Te ae ae P. whitei
mio ae P. antiquus
se cccvccesccscecces Pp 1 .
jonesi ty . He s

Fic. 3. Breadth/length indices of the last upper premolar and the upper molars in the four

species of the genus Parapapio. The values for P. broomi are considered to be 100; those of the

other species are related to P. broomi. The diagrams demonstrate the aberrant tooth proportions
of P. antiquus. For exact data see Table 2.

ANNALS OF THE SOUTH AFRICAN MUSEUM

‘qsu9] 4100}= TF

“Upesiq 4j00j=— gq

“SON[LA 9SLIOAR = YX

*suoundoods Jo 1squinu=—u

*LS6] UeuUpsely Wor somsry y

“TL6T Jorey Wor suoundeds Surpnjour soyeuiay puv soyeu estidwI0d seorpuy ,

= sy | im LIS | = 98h | € i USD AS | oe = a = 9°6b (8) (6) ‘N— 9
== €‘€€ € oe 9°6€ (6 = €‘O€ it, = OLE 9] 8*SE 6'SE 8‘8E == Gus (O‘L€) ILE sIN— id
1'%8 6'L € || €6L | OFF I || 878 Ho OL OYE | oil fy fife GL CL S‘6 €'8 8°L OL qP
L‘ZOl 8°6 € 0°96 8°71 I €‘66 L‘ol II || 9*€0I L‘Ol | OO) v‘or v‘Or S‘Il Ul (LOT) TOI * qui
(g =u) 9°6 € || (=u) €‘€l @/Qz=8)} O18 |e @=) vO! LE 86 9°6 rol eT vO! 601 €‘Ol I eA
v'16 6°8 @ | £16 vi) WS Vor | €r|| Tics v'6 9} 9'6 1°6 7.6 8°6 an S*6 1‘6 qP
SO SOE CN SOO a EOE i ea) censta HO 1] 2) fai S01 OIL s'Il Aap (rol) | Sor qui
(¢ =u) €Ol by || (Z=) LAY 9 || (t7=4)] TIT LI || (9=u) CTI L| 60! OTT oI CT (L‘T1) ere}! ST I “WN
5'96 0'8 Ci) 216 8°6 S| 7@16 88 =| Or] 4H S8 L'8 OES L‘L = O1OT | (8's) | (o%6) 0'8 qP
I‘Sol 88 ¢ 6°86 S‘Ol S €‘66 S*6 EW. Sa v°6 ¢ = 1'6 re 9‘O1 (S*6) (Z‘6) ‘8 qui
(9=u) L‘g 9 || (¢=u) L‘ol 9 || (C¢I="4) | 96 PI || (¢=u) 0‘OI L | (6‘8) 76 €‘Ol 601 (86) OTT 0‘OI I iW
Seek | LiL Ne Gra S7o%} a STW Zt 8 IE ETI) UE 2 Ee 78 oa = (o'8) | (o%8) €'8 q
(L=) 0°9 9 || (L=u) 69 9 || (€7=4) | v9 rl || G@=4) OL Oo) I 7G) 89 [fe = Te (89) 69 I rd
STE Tee 9 8901 6'L € ‘ZIT VL OL |) SLL [ti C == os = = 6‘L = PL ao
— aS ¢ ae VL v = 9‘9 Ol a= 66 I = a = = = = 8°6 (I ed
— 6L € ma O'91 I ee a iG = =e = = = = = = = = Y
0'S6 8‘¢ € 6'S6 O'L € PE0I 1'9 ry || 9*T0L | (€‘9) I = — = == (¢€‘9) = = q
a 19 (é = (Se I oa 6S v — (79) I = = re = (79) = == I 3)
es O%L 4 = ae |e Se 9°8 € = = ak Te = =e = = se q
przl | 19 I = eel ae rcOR I O98 (6 = aoe elie |b ates = ci = = =e z= q
—- 6'v (G) — ae = = 8‘¢ v — == = 59 | tine aa = = = SS =s I |
os Otel = re [oan OTE | = ae | a ales a = = cz = = q
S‘I0l L'9 I ras = ——ill 1G 76 6°9 Z = == = — = —— = = =, = q
— 9‘9 Z ee = a cal (SI v = a — rt = = = ra — — I il
OO! 1/4 4 u || OOT 1/4 x u || OOL T/& x u || OOT 1/4 4 U | *b69 96 | «8099S | «STL OL LVL | 6LOE'W | 8Log'W
4 1sauol "J %10114YM “J 4 IMOOAG “J snnbijuo oidpdpvavg

ySe1seds o1dpdpdvg snore JO soyeuloy UI Y}99} Joddn Jo WU-uT syUOWIOINSvOP “7 ATAV

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG 9

TABLE 3. Statistical analysis of the tooth lengths of P*, M!, M? and M#
in the females of Parapapio antiquus and P. broomi

Standard deviation s?* Student’s t-Test Degrees of Significance
for the sample means* | freedom probabilities*
a, Parapapio antiquus | P. broomi t ; df t.99
ee 0,136 0,070 3,9 18 2,878 0,01
M! 0,743 0,366 1,4 19 2,861 0,2-0,1
M? 0,169 0,252 17 22 2,819 0,1
M3 0,345 0,226 3,0 17 2,898 0,01

* After Simpson, Roe & Lewontin 1960.

M® seems to be reduced. The Student t-test proved these differences to be
significant for P* and M? (Table 3).

The breadths of the premolars and molars being nearly identical in both
taxa, the breadth/length index could be expected to express the different
degrees of elongation (or reduction). The index also makes possible a com-
parison with the other species, P. jones: being absolutely smaller and P. whiter
larger than the previous two. Figure 3 shows the relations diagramatically,
the indices of P. broomi being expressed as 100. P. jonest exhibits comparatively
high values, especially for M! and M3, this possibly being a primitive feature.
P. whiter is very similar to P. broomi, showing a slight tendency to elongation,
especially in M®. Again, P. antiquus appears to be very aberrant with its gradual
increase of the index in mesio-distal direction. The indices for the distal
breadths of the molars show that P. antiquus possesses the highest degree of
reduction of the distal cusps in all molars. Absolutely this reduction is most
pronounced in M®?, whereas the relative value is lowest for M?.

Judging from tooth size and morphology, P. jones: could tentatively be
regarded as the most primitive of the fossil cercopithecids of South Africa,
possibly being closely related to their common ancestor. New finds of male
skulls, however, show clearly that this taxon was well advanced in some respects:
its high degree of sexual dimorphism would exclude it from being a direct for-
runner of P. antiquus and its well-pronounced maxillary crests and different
cranial proportions from being a direct ancestor of both P. broom: and P. white
(Maier 1971).

On its teeth alone, however, P. jones: provides a model for understanding
the evolutionary alterations within the genus Parapapio. Compared with
P. jonest, P. antiquus shows P4 and M! very much elongated; Mz? being still
longer, while M® is very similar in both taxa—apart from the conspicuous
reduction of the distal cusps in P. antiquus, which cannot be understood simply
as a consequence of small size.

Io ANNALS OF THE SOUTH AFRICAN MUSEUM

ty)
aoe QO
Ww — iw)
iS) (= )
= (e) oF
(eo) Ww) Ww)
mat = >
g © §
a 2 fe)
: g 8
a o o
KROMDRAAI
SWARTKRANS
ff E 9
8 Oo E
= O
2 fe)
° fe) pl
a a. 5
a. a. =
© © =
Bian hemi 8S
oe hege
STERKFONTEIN - 5 a
a. 2 °
a ene
ra o a
TAUNG

MAKAPANSGAT

Fic. 4. Phylogenetic diagram of the genus Parapapio and the hypothetical origin of the genus
Papio. The diagram is based on the South African evidence only. P. jonesi occurs at all the men-
tioned sites; the forms on the left side occur only at Sterkfontein and Makapansgat, those on
the right only at Taung, Swartkrans and Kromdraai. Less important sites have been disregarded.

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG II

P. broomi and P. white: are very similar in tooth indices, underlining the
close coincidences in their cranial morphology. Both these taxa are similar to
P. jonesi in the proportions of P* and M?, whereas M1! and M? seem to be
distinctly elongated. Summarizing, one can state that P. antiquus shows a
progressive elongation of the ‘P4/M?-field’ of the tooth row, whereas in both
P. broomi and P. white it is mainly the third molar which is increased. In some
regards, these conclusions need to be confirmed by additional observations.

These peculiarities of the dentition could possibly be interpreted func-
tionally: in P. antiquus, the centre of gravity of the chewing activity is shifted
forward as compared with that of the related species. This would necessitate
a more oblique direction of the temporal muscle, and would, in turn, explain
not only the morphology of the temporal crests but possibly even the low
position of the occipital region of the braincase as observed in P. antiquus.

CLASSIFICATION OF THE GENUS Parapapio

Present mammalian systems are based mainly on the methods of com-
parative morphology and this is especially true for fossil forms. The classification
of fossil cercopithecoids is fraught with many difficulties, and even that of the
extant taxa has not yet been satisfactorily established. External and soft-part
characters have proven most valuable for systematic purposes within this
superfamily (Pocock 1925), the teeth and the skeletons being very uniform
within the whole group (Remane 1960; Schultz 1970). As far as possible,
ecological and functional aspects should be considered also and, in the case of
fossils, it is important that there be an appreciation of the time factor as well.
The present state of knowledge as to the classification of the Cercopithecoidea
was discussed recently by the present author (Maier 1970).

All attempts to obtain absolute data on the ages of the South African
australopithecine caves have so far been unsuccessful (Tobias & Hughes 1969).
However, recent finds have resulted in surprising changes of the chronology
of the North and East African Pliocene and Pleistocene fossil sites. Olduvai
Bed I has been dated at about 1,8 m.y., while the deposits at Omo, Kaiso,
Kanapoi, Koobi Fora, Chemeron and Laetolil might reach back some 3 to 4
m.y. (Maglio 1970). These sites with their varied fossil faunas provide a good
basis for comparison with the richly fossil-bearing cavern breccias in southern
Africa, but not much work has so far been done in this field. Hendey (1970)
has pointed out the major difficulties in comparing South African fossil faunas
at the present stage of knowledge.

Whereas Kurtén (1960, 1968) suggested a Middle Pleistocene age for the
South African ape-man cave deposits, we must now contemplate a Lower
Pleistocene age for them. Ewer (1963: 343) reviewed the then available evidence
and concluded that ‘Kromdraai and Swartkrans may correspond to the gap
between [Olduvai] I and II, while Makapan and Sterkfontein belong to the
period covered by the older deposits of Olduvai I and Omo’. This tentative

I2 ANNALS OF THE SOUTH AFRICAN MUSEUM

correlation was repeated by the same author in 1967. According to Cooke
(pers. comm.) the pigs and elephants from Makapansgat compare well with
forms from Kanapoi and Lower Omo beds, indicating an absolute age of some
2,5-3,5 m.y. for that site. Sterkfontein is estimated by Cooke to be about 2,5
m.y. and Swartkrans about 2,0 m.y. As in East Africa, the Pleistocene sites of
South Africa will, most probably, have to be dated further back than was
previously thought.

This evidence, so far based mainly on Suidae, Elephantidae and Carnivora,
seems to be supported by the primate evidence. (Unfortunately, the rich East
African cercopithecoid material has not yet been described comprehensively,
but Mrs. Meave Leakey of the Kenya National Museums will shortly publish
a monograph.) Arambourg (1947) recorded Dinopithecus brumpti from the Omo
deposits, Butzer (1971) added Colobus sp., Cercopithecus sp., Parapapio sp.,
Papio sp. and Simopithecus sp. from this site. R. E. F. Leakey (1969) described
Papio baringensis from the Chemeron Beds which shows much similarity with
Papio robinsoni, and in 1970 the same author recorded from Koobi Fora (-| 2,5
m.y.) Cercopithecus sp., Papio sp. and Simopithecus sp. According to L. S. B.
Leakey (1965), large forms of Papio and Simopithecus are known from Olduvai
Beds I-IV. As far as I could see during a recent visit to the Kenya National
Museums, Nairobi, the genus Parapapio occurs as well, both at Olduvai and
Koobi Fora, the material most probably belonging to the species P. jones. A
mandible from Kanapoi (+ 4 m.y.) has recently been referred to this species
as well (Patterson 1968). A few small teeth from Lothagam would fit approxi-
mately some specimens from the ‘grey breccia’ of Makapansgat, being referred
also to P. jonesi. Most of these better known sites seem to have three cercopithecid
forms side by side: a small Parapapio, a large Papio (Dinopithecus and Gorgopithecus
possibly being only synonyms) and a very large Simopithecus.

The South African cave deposits show a different arrangement, which
may, however, be due partly to geographical separation and a different mode
of deposition (Ewer 1967). The older sites at Makapansgat and Sterkfontein
have so far yielded only various types of Parapapio and a comparatively small
and primitive Szmopithecus (Maier, in press), but no true baboon of the genus
Papio whatsoever (Freedman 1957). Very small and primitive forms of baboons
appear only at Taung, while, besides Parapapio and Simopithecus, Papio is
abundantly represented in the younger sites of Swartkrans and Kromdraai.
The fossil colobids of both East and South Africa are too different for useful
comparisons to be made.

Pending more detailed information about the cercopithecoid material
from the Lower Pleistocene of East Africa, the preliminary evidence seems to
suggest rough contemporaneity of the more important South African faunas.

The faunal comparison of the South African sites is complicated by their
geographical distance and by evident palaeo-ecological differences in the
surroundings of the ancient deposits (Ewer 1956a). Thus, at Makapansgat the
environment was probably more varied and less dry (Ewer 19564; Wells 1967),

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG 13

whereas Taung ‘was distinctly more desert-like than . . . the other deposits’
(Ewer 1957: 139). Zoogeographically Makapansgat shows more affinities with
central Africa than the other sites. Considering these difficulties, Ewer (1957:
141) concluded: “The probable time sequence of the deposits is Sterkfontein
and Makapansgat close together, with the former very probably being the
earlier; then Swartkrans and lastly Kromdraai, while the Taung deposit is
most probably closest in time to Sterkfontein and Makapan.’ From the mor-
phological evidence of the cercopithecoids, Freedman (1957) considered the
Taung deposit to be the oldest, followed in order by Sterkfontein, Makapansgat,
Swartkrans and Kromdraai. Wells (1967) and Cooke (1967) seem to assume
that Taung is slightly younger, the latter author giving a sequence Makapans-
gat, Sterkfontein and Taung for his ‘Sterkfontein Faunal Span’ (see his ‘Table I).

Recently Wells (1969) more clearly expressed his conviction that
Makapansgat may be earlier than Sterkfontein, whereas ‘Taung may be even
closer to the Swartkrans-Kromdraai ‘faunal span’. It seems to be very neces-
sary that the newly prepared elephant material from Makapansgat be studied
by experts who are well acquainted with the East African forms.

Freedman (1957: 248) stated that the more important South African
fossil sites originate from ‘a geologically short period just following the Plio-
Pleistocene boundary’, the time of depositing between the oldest (Taung) and
the youngest (Kromdraai and Cooper’s) breccias not being longer than about
250 000 years. As, according to Simpson (1944), the minimum time span for
the evolution of a new species amounts to about 0,5 m.y., Freedman (1957: 244)
concluded: “Therefore . . . it seems quite obvious that the faunal changes
between the sites could not be due to in sttu evolution’ and that ‘it would there-
fore seem that the most obvious and probable cause of the faunal replacements
was successive migrations into and out of the areas as a result of local and/or
distant environmental changes’. The recently suggested evidence of the very
great age and long duration of these deposits would yield, however, a satis-
factory temporal frame to explain the evolution and radiation of the numerous
Pleistocene Cercopithecoidea in South Africa, without entirely discarding the
possibility of some faunal shifting.

The existing classification of the genus Parapapio was elaborated in the
studies of Broom (1940) and Freedman (1957). Based mainly on the occurrence
of different-sized molars, these authors established four species: the small-sized
and, as it appears now, widespread Parapapio jonesi, two medium-sized forms,
P. antiquus and P. broomi, which ‘are remarkably similar in tooth size but differ
very considerably in skull shape’ (Freedman 1957: 158), and finally the large-
sized P. whiter.

The small Parapapio jonest could easily represent the generalized common
ancestor of the Papionini sensu stricto. Its small teeth are unspecialized as com-
pared with progressive features in the other species of Parapapio. Occurring in
all of the australopithecine caves of South Africa, this taxon has now been
recorded from various places in East Africa as well, probably covering some

14. ANNALS OF THE SOUTH AFRICAN MUSEUM

2-3 m.y. of the Pleistocene fossil record. New finds have shown, however, that
the male skull is quite advanced, although retaining its primitive teeth (Maier
1971).

In cranial shape, tooth specialization and the small degree of sexual
dimorphism, Parapapio antiquus seems to differ more from P. jones: than does
P. broomi, and might thus be an earlier offshoot, possibly being somehow
adapted to the drier ecological conditions prevailing at ‘Taung. Such an environ-
ment could also have stimulated the evolution of the small true baboons
Papio wellsi and P. izodi: ‘Parapapio antiquus is very similar in size and dental
morphology to Papio izodi, and these two species may represent a morphological
stage not far from the point at which the genera Parapapio and Papio started
diverging from a common stem’ (Freedman 1957: 245).

The nature of the molar specializations and the very similarly elongated
male crania indicate a monophyly of both Parapapio broom: and P. whiter. ‘The
teeth of the latter species seem to be relatively larger than in the similarly sized
P. broom. Further material may close the existing size gap, but since both forms
occur in the same blocks of the ‘Upper Phase I’ breccia of Makapansgat, they
cannot form a chronocline. As we do not know their postcranial skeleton, it is
not possible at the moment to assign different ecological niches to these
apparently sympatric species. In Figure 4 an attempt is made to plot the
evidence in the form of a phylogenetic diagram.

SUMMARY

Two new female skulls of the fossil cercopithecid Parapapio antiquus from
Taung, Cape Province, South Africa, are described and discussed. The teeth
of this species especially show some significant differences from the other three
species of Parapapio. A phylogenetic arrangement of the genus is suggested.

ACKNOWLEDGEMENTS

In the first place, I have to thank Mr. J. W. Kitching for placing his
material at my disposal; I owe thanks to Miss J. Roets and Mr. B. Maguire for
much help with the manuscript and to Mr. J. Henderson for help with the
Statistics.

REFERENCES

ARAMBOURG, C. 1947. Contribution a l’étude géologique et paléontologique du bassin du Lac
Rodolphe et de la Basse Vallée de l’Omo. 2¢ partie: Paléontologie. Mission scient. Omo 1:
231-562.

Broom, R. 1940. The South African Pleistocene cercopithecid apes. Ann. Transv. Mus. 20: 89-100.

Butzer, K. W. 1971. The lower Omo Basin: geology, fauna and hominids of Plio-Pleistocene
formations. Naturwissenschaften 58: 7-16.

Cooke, H. B. S. 1967. The Pleistocene sequence in South Africa and problems of correlation.

In BIsHoP, W. W. & CLARK, J. D., eds. Background to evolution in Africa: 175-184. Chicago;
London: University of Chicago Press.

TWO NEW SKULLS OF PARAPAPIO ANTIQUUS FROM TAUNG 15

Ewer, R. F. 1956a. The dating of the Australopithecinae: faunal evidence. S. Afr. archaeol. Bull.
II: 41-45.

Ewer, R. F. 19566. The fossil carnivores of the Transvaal caves: two new viverrids, together
with some general considerations. Proc. zool. Soc. Lond. 125: 259-274.

Ewer, R. F. 1957. Faunal evidence on the dating of the Australopithecinae. Jn CLARK, J. D.,
ed. Third Pan-African Congress on Prehistory, Livingstone 1955: 135-142. London: Chatto &
Windus. -

Ewer, R. F. 1963. The contribution made by studies of the associated mammalian faunas.
S. Afr. F. Sci. 59: 340-347.

Ewer, R. F. 1967. The fossil hyaenids of Africa—a reappraisal. Jn BIsHOP, W. W. & CLARK, J. D.,
eds. Background to evolution in Africa: 109-123. Chicago; London: University of Chicago Press.

FREEDMAN, L. 1957. The fossil Cercopithecoidea of South Africa. Ann. Transv. Mus. 23: 121-262.

FREEDMAN, L. 1961. New cercopithecoid fossils, including a new species, from Taung, Cape
Province, South Africa. Ann. S. Afr. Mus. 46: 1-14.

FREEDMAN, L. 1965. Fossil and subfossil primates from the limestone deposits at Taung, Bolt’s
Farm and Witkrans, South Africa. Palaeont. afr. 9: 19-48.

Gear, J. H. S. 1926. A preliminary account of the baboon remains from Taungs. S. Afr. 7. Sci.
23: 731—-747-

Haucuton, 8. H. 1925. A note on the occurrence of a species of baboon in limestone deposits
near Taungs. Trans. R. Soc. S. Afr. 12: lxviii.

HEnpDEyY, Q.B. 1970. A review of the geology and palaeontology of the Plio-Pleistocene deposits
at Langebaanweg, Cape Province. Ann. S. Afr. Mus. 56: 75-117.

Jones, T. R. 1937. A new fossil primate from Sterkfontein, Krugersdorp, Transvaal. S. Afr. 7.
Sci. 33: 709-728.

KurtTEn, B. 1960. The age of the Australopithecinae. Stockh. Contr. Geol. 6: 9-22.

KurtEn, B. 1968. Dating the early stages of hominid evolution. Jn kuRTH, G., ed. Evolution und
Hominisation: 75-81. Stuttgart: Fischer.

Leakey, L. S. B. 1965. Olduvai Gorge, 1951-61. Cambridge: University Press.

LEAKEY, R. E. F. 1969. New Cercopithecoidea from the Chemeron Beds of Lake Baringo, Kenya.
Fossil Vert. Afr. 1: 53-69.

Leakey, R. E. F. 1970. Fauna and artefacts from a new Plio-Pleistocene locality near Lake
Rudolf in Kenya. Nature, Lond. 226: 223-224.

Mactio, V. J. 1970. Early Elephantidae of Africa and a tentative correlation of African Plio-
Pleistocene deposits. Nature, Lond. 225: 328-332.

Matrer, W. 1970. Neue Ergebnisse der Systematik und der Stammesgeschichte der Cercopithe-
coidea. £. Sdugetierk. 35: 193-214.

Matsr, W. 1971. New fossil Cercopithecoidea from the Pleistocene cave deposits of the Maka-
pansgat Limeworks, South Africa. Palaeont. afr. 13: 69-107.

Maier, W. The first complete skull of Simopithecus darti from Makapansgat, and its systematic
position. 7. hum. Evolution. (In press.)

PATTERSON, B. 1968. The extinct baboon Parapapio jonesi in the early Pleistocene of northwestern
Kenya. Breviora 282: 1-4.

Preasopy, F. E. 1954. Travertines and cave deposits of the Kaap Escarpment of South Africa,
and the type locality of Australopithecus africanus Dart. Bull. geol. Soc. Am. 65: 671-706.
Pocock, R. I. 1925. The external characters of the catarrhine monkeys and apes. Proc. zool.

Soc. Lond. 1925: 1479-1579.

Remanez, A. 1960. Zahne und Gebiss. Jn HOFER, H., SCHULTZ, A. H. & STARCK, A., eds. Primatologia.
Handbook of primatology. 3: 637-846. Basel; New York: Karger.

Scuuttz, A. H. 1970. The comparative uniformity of the Cercopithecoidea. In NAPIER, J. R. &
NAPIER, P. H., eds. Old world monkeys: 39-51. New York; London: Academic Press.

Scott, J. H. 1967. Dento-facial development and growth. London; New York: Pergamon Press.

Smpson, G. G. 1944. Tempo and mode in evolution. New York: Columbia University Press.

16 ANNALS OF THE SOUTH AFRICAN MUSEUM

Stmpson, G. G., Roz, A. & Lewontin, R. C. 1960. Quantitative zoology. Rev. ed. New York;
Burlingame: Harcourt, Brace.

Tosras, P. V. & Hucues, A. R. 1969. The new Witwatersrand University excavation at Sterk-
fontein. S. Afr. archaeol. Bull. 24: 158-169.

WELLs, L. H. 1967. Antelopes in the Pleistocene of southern Africa. Jn BIsHop, w. w. & CLARK,
j. D. eds. Background to evolution in Africa: 99-107. Chicago; London: University of Chicago

Press.
WE ts, L. H. 1969. Faunal subdivision of the Quaternary in southern Africa. §. Afr. archaeol.

Bull. 24: 93-95.

Ann. 8. Afr. Mus., Vol. 59

Plate 1

WW MY

tify
Uy

Vy
ty

Parapapio antiquus M.3078 (female) lateral and basal view. Note the straight contour of the

muzzle dorsum, the low position of the occiput and the reduction of the third molar. Scale
unit 10 mm.

INSTRUCTIONS. TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) o.
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm X 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

BuLitoucu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHer, P.-H., DuvaL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Kool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn sCcHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 80.

shins Al = hot

oS Se —

ae
al

= Psi 4 athe x TY

3 une : Sa MA oe" Hite

i

Pi x uct) DeSales

iL - ; > ae \ ’ ; Ae ry

‘ ae y

ares’. & Pav ml rar aes eke “eke Lins th sates i) ii “4 ‘Rovaathiad
“Coe yey oh Mare OVA ee sghlfue Pare, «yee SEPARA ret oe) Conse
Pe mF ally Ny "thd, iy cbvhdt Pape PB ras ‘hig est gtosey ae Neer” wert a Raia
" : Mod ist ee A ia NR Apter td AO at), fal prays 3) ae mae .
7 a staged) 7 pa eo on, apts, ny Teas ada tied Cheah pala
Be a Poh aun ed OY, eT W. nan gtiol «o.ghhl HEN Beeler Pied

iy a ar) s fy | AZ
mo) : % . :
a 7
A y " < 1
bs a ite. te:
os ] 4 = j
4 i ; a
x RS A ;
a 3
y i t : & ,
< 2 F 7 a 4 iy } q
: fre hs “POL ee AE atest? ra MP? PUSS AE i FF
> > es ; , ; }
y = Ys ; vA ead .
i Ei ell i ;
; P ~ 2
* :
i >;
=
a - : =) i aes
a
RA y rahe A
= eS ae PON ¥ 3 oi | ¥
; | ames bes i: Seating: beta Cine aa
{ af oy
. os Ne
>
- a ” j yet , » ’ -
< a" , + : «,
ou a ext 4 a
» = xf . et { \ .
. 2 :
Mm Sinericc Ca Ec 7 , bie) he ‘
= r - . bi
“Y
‘ = \7 gf : 3 ; oath
< Yar = F r te ' ey x
of 7 ~
N =
"
- ~— uc =
‘ i
se: Se
p= x
: .
Sa 3
i “~
ra a ,
f fi ds. eae
ns a) 7
> sk
ry - ip
.
- ‘4 p
» ‘ = €
¥
Z
7
.
- =
=
a 1
* ub ;
hee
ad — 7
,, \
¥ ow para
~
i
fa
at | i
( 4
‘
: 4 Oy phe eii
4
a
r
- Eb
~ \
'
sk
eT
.
re z
'
‘ iT
, ;
a - : ft
, a b«
4 4

507,68

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
December 1971 Desember
Fat 2 Deel

SB;

= y

Ut
12 \ al NS
pe >
2 ¥

a,
‘ewe

THE EARLY THERAPSIDS

By
L. D. BOONSTRA

- le
OWNER; a
— ” *
7 atthe.

MAR 17 1972. |
Cape Town Kaapstad

C423; >
MS RANIES

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, (1, 3, 5 7-8), Bie 5» ip) 51-2, 5; 79);
6(1, t.—p.i.), é 7(1-3), 8, Q(1-2), 10(1),
11(1-2, 5> 7s t.—p.1.), 21, 24(2), 275 Sita), 33> 38

Price of this part/Prys van hierdie deel
R3,30

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum
1971

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

THE EARLY THERAPSIDS
By
L. D. BoonstrRa :
South African Museum, Cape Town
(With 3 figures)
[MS. accepted 8 September 1971]

CONTENTS

PAGE
Introduction . ; 5 : : F 5 17
Diversity : : : ; , : ‘ 18

Structural variations
Temporal fenestra ; : : : ; 19
Position of the quadrate ; ; ; : 23
Insertion of the temporal muscle . : ; 24
Reflected lamina of the angular . ; ; 25
Marginal teeth . : : ‘ : : 25
Palate. 4 : : : ; : . 26
The locomotor apparatus : é : . 26
Diverse structures . ; , ; , : 29

Morphological series
Series I : : : , ; ‘ : 31
Denies 117 P : : ; F ; 33
Series III. : : a‘ ; : ‘ 35
Series ITV. ‘ : ‘ : ‘ ? 36
Summary of morphological series . : ‘ 40
Early therapsid history. 4 : : : 41
Subsequent history . s : : : ; 41
References F : : ; : : : 44

INTRODUCTION

As early therapsids I consider those forms that have been recovered from
Zones I and II of the Russian succession and from the Tapinocephalus zone of
the Beaufort beds of the Karroo System. They thus range from the top of the
Lower Permian to the end of the Middle Permian.

In this paper I am stressing, firstly, the great diversity of forms with which
we are so suddenly confronted in one of the many explosive faunal develop-
-ments that have so repeatedly occurred during the long history of animal life
and apparently gainsaying the dictum natura non facit saltum.

Secondly, I shall attempt to arrange this assemblage into a number of
morphological series, before attempting a phylogenetic arrangement, because
I feel that this is a safer procedure, bearing in mind that this fauna with which
we are confronted as a fait accompli consists of contemporaries and geologically
speaking of the same age with no one the ancestor of any other.

Thirdly, I shall consider the derivation of these therapsids from the ante-
cedent fauna of pelycosaurs with which in similar fashion we are confronted in
Carboniferous-Permian times.

17

Ann. S. Afr. Mus. 59 (2), 1971: 17-46, 3 figs.

18 ANNALS OF THE SOUTH AFRICAN MUSEUM

DIVERSITY

Of the pre-Upper Permian assemblage of therapsids we know at least 70
well-established genera which have on taxonomic criteria been brigaded into
19 families. This explosive radiation is much greater than the earlier Carboni-
ferous—earlier Permian radiation of the pelycosaurs with its 8 families.

In the early therapsids the size variation is as much as that between a rat
and a hippopotamus, with weights from about 500 g to two tons. In shape they
vary from light and slender to massive and plump. Some are agile, others
ponderous. There are long as well as short tailed forms, long snouted and
extremely short snouted species lived side by side; locomotion varied from
slinking, walking to running, with the body slung between the spread-eagled
limbs or carried fairly high on more upright supports. A few were insectivorous,
some carrion eaters, others predaceous carnivores and many herbivores; some
feeding on soft marsh plants, whereas others, roaming on to higher ground,
subsisted on more fibrous shrubs. In one group the teeth were largely replaced
by horny sheaths analogous to those of tortoises.

This diverse assemblage has been classified into the following 19 families:
Eotitanosuchidae

Brithopidae

Anteosauridae

Titanosuchidae

Tapinocephalidae

Styracocephalidae

Estemmenosuchidae

Phthinosuchidae
9g. Hipposauridae

10. Galesuchidae

11. Otsheriidae

12. Venyukoviidae

13. Dromasauridae

14. Endothiodontidae

15. Dicynodontidae

16. Alopecodontidae

17. Pristerognathidae

18. Lycosuchidae

19. Scaloposauridae

The various authors who have established these 19 discrete families have
done so on the basis of determined differences of a structural nature.

These differences, although often considerable, are also limited, and it is
because of these limitations accompanied by certain basic similarities and
trends that these families have been brigaded into one order—the Therapsida.

A profitable evaluative discussion can best be started by considering

firstly those points of basic similarity and then the extent of the differences and
variations.

Se a Moc oe ce ie ee

THE EARLY THERAPSIDS 19

Structural features common to all the known early therapsids are:

1. A single temporal fenestra lying below the posterior process of the post-
orbital, but the participation of this process and of the various other bones
forming the temporal border varies.

2. The pterygoids tend to meet in the middle line behind the interpterygoid
vacuity, which is thereby variously reduced, and applied to the basi-
cranium with consequent loss of the primitive freely movable joint and the
development of a longitudinal basicranial girder.

3. The jaw articulation never lies in a plane posterior to that of the occipital
condyle and the anteriorly directed slope of the occiput is reduced, often
becoming nearly vertical or even sloping backwards.

4. There is always a reflected lamina of the angular.

5. The septomaxilla, with a foramen, has a more or less well-developed
lateral facial exposure but small in Otsheria and brithopids and phthino-
suchids and not exposed laterally in dicynodonts.

6. There is no supratemporal and the anterior coronoid is always lost and
sometimes both are absent.

7. ‘The lacrimal never reaches the nostril.

8. The maxilla is always deep.

g. The squamosal flares out laterally and posteriorly to various degrees.

10. The quadrate ramus of the pterygoid reaches the quadrate.

11. The vertebrae are amphicoelous and there are no dorsal intercentra.

12. The girdles and limbs are adapted for an early stage of a quadrupedal
gait. ‘The main adaptations are: loss of the supraglenoid buttress and
foramen, loss of one central in the tarsus and one distal, number of
phalanges reduced to varying degree, the glenoid is reduced in length and
the humerus untwisted to varying degrees, the iliac blade is heightened,
and an anterior process developed to varying degrees, the femur loses the
Y system of ridges and develops a greater trochanter.

STRUCTURAL VARIATIONS
TEMPORAL FENESTRA

It is obvious that the origin of the m. capiti-mandibularis in the captorhino-
morphs (and all other anapsids) could only have been from the inner or under
surface of the temporal or cheek bones. Fox has reported that in a captorhinid
examined by him the central part of the temporal covering is composed of very
thin bone with a concomitant thickening peripherally of this weak area and
infers that the attachment of the muscle was mainly on the thickened parts
and that the thinning centrally was due to this area becoming non-functional
and thus liable to fenestration.

Such fenestration has in fact taken place in the pelycosaurs. The area of
bone-resorption is mainly situated at the junction of squamosal and jugal, but
the fenestration in the pelycosaurs has been far from uniform. In fact in one
species of Ophiacodon there is not a single but a double fenestra. Moreover, there

20 ANNALS OF THE SOUTH AFRICAN MUSEUM

is considerable variation in the participation in the border of the fenestra by
the bones of the cheek. The dorsal border or upper temporal arch is always
formed by the postorbital and the squamosal. The lower border or zygomatic
arch is mostly formed by the jugal and squamosal in varying proportions, but
in all the three suborders there are forms in which the quadratojugal enters
the lower border of the fenestra. Is this of sufficient importance to query the
homology of the temporal fenestra ?

The temporal fenestrae of the pelycosaurs lie mainly laterally in the
cheeks and are separated from one another by a broad flat intertemporal skull
table. In the pelycosaurs the original capiti-mandibularis has divided into a
major medial mass, the temporal and a lateral mass, the masseter. The temporal
originated from the inner face of the bones above the fenestra, viz. parietal,
postorbital, postfrontal and squamosal. The masseter arose from the inner
face of the zygomatic arch, i.e. from the jugal and squamosal lying below the
fenestra. The function of the fenestra is undoubtedly to enlarge the adductor
chamber for the bulging of the temporal muscle during contraction.

The temporal fenestra of the earliest therapsids, apparently homologous
to that of the higher sphenacodonts, when first encountered already shows a
number of modifications in divergent directions.

In Hipposaurus the fenestrae are still small and are separated by a wide
intertemporal table, the posterodorsal flange of the postorbital, meeting the
squamosal, lies in a horizontal plane with the temporal muscle originating, in
part, from its ventral face.

In the early Galesuchidae the fenestra is larger both in length and width,
but otherwise essentially as in the higher sphenacodonts.

In Phthinosuchus the fenestra is greatly enlarged both in height and length
due to the outflaring of the squamosal laterally as well as posteriorly. (It
extends forward into the jugal.) The postero-dorsal flange of the postorbital
lying horizontally is, however, shortened and laterally flanked by a horizontally
disposed lappet of the squamosal which on its ventral face provides a large area
of the origin for the temporal muscle.

In Eotitanosuchus the fenestra is enlarged, extending forward into the jugal,
and a groove on the outer edge of the postorbital indicates that the temporal
in part arose from the lateral face of this bone.

In the Brithopidae, and even more so in the Anteosauridae, the fenestra
is enlarged by a lateral as well as a posterior outflaring of the squamosal which
greatly increases the size of the adductor chamber. The dorsal flange of the
postorbital now shows a well-developed lateral face and the temporal muscle
now arose in part from a ridge on the dorsal edge of the postorbital, confluent
with a postero-lateral edge on the squamosal. Moreover, in the Brithopidae
and Anteosauridae, the intertemporal skull table is much reduced in width
and the original horizontally lying upper face of the postorbital now faces
appreciably laterally and is practically excluded from the skull table.

In the Titanosuchidae the temporal fossa is only of moderate size. The

THE EARLY THERAPSIDS 21

posterior flange of the postorbital, lying mainly vertical and applied to the
outer face of the parietal, is greatly reduced and lying low down in the skull
reaches the squamosal as a tapering splint.

The intertemporal skull table is reduced in width and the parietals form a
fairly wide and fairly high sagittal crista.

The origin of the temporal muscle mass is mainly from the outer face of
the posterior postorbital flange and extends up the lateral face of the parietal
to the edge of the crista and posteriorly to the upper part of the squamosal,
whose lateral edge is continued as a ridge on to the parietal.

The jugal is excluded from the fairly deep lower temporal arch, apparently
due to the downgrowth of the strong postorbital.

In the Tapinocephalidae the greatly varying pachyostotic thickening of
the skull bones affects the nature of the temporal fenestra, the adductor chamber
and the degree of participation of the bones forming the borders of the
fenestra.

As to position of the fenestra, the one extreme is seen in Riebeeckosaurus
where the two fenestrae are separated by only a sharp parietal crista; in the
other extreme the intertemporal width is so great in Criocephalus that the fenestra
is not visible in dorsal view. In all the lower arch, formed solely by the squa-
mosal, is very deep so that the fenestra is situated high up in the cheek. In
general the strong postorbital bar makes the distance between orbit and
fenestra great. As to shape, the fenestra is slitlike in some moscopines and tapino-
cephalines with the fore-aft diameter one-third of the dorso-ventral, whereas in
Avenantia it is longer than high, with the struthiocephalines in an intermediate
position.

The adductor chamber is roomy in Avenantia, moderately so in the
struthiocephalines but antero-posteriorly compressed in the tapinocephalines
and in Moschops and Crocephalus. As to the circum-fenestral bones, the Tapino-
cephalidae have one feature in common in that the jugal is wholly excluded
from the lower arch, being pushed anteriorly by the thick postorbital bar
and the forward growth of the squamosal due to the quadrate moving
anteriorly.

In the tapinocephalids considerable variations occur in the upper temporal
arch. In some of the struthiocephalines (where the pachyostosis is moderate)
the dorso-posterior flange of the postorbital and the upper flange of the squa-
mosal do not meet, being thus separated by the parietal. In the other tapino-
cephalids (where the pachyostosis is greatly developed) the junction of the
postorbital and squamosal is pushed down to the lower half of the fenestra.

In the tapinocephalines (where the pachyostosis is great) abnormal over-
growth of both the frontal and postfrontal bones caused these bones to enter
the dorso-anterior part of the rim of the fenestra.

In the moschopines (where the pachyostosis is in some respects even
greater) only the postfrontal enters the dorso-anterior part of the rim of the
fenestra. In all the tapinocephalids the main origin of the temporal muscle

22 ANNALS OF THE SOUTH AFRICAN MUSEUM

mass is from the lateral surfaces of those parts of the parietal, postorbital and
squamosal lying well within the upper part of the temporal fossa.

The rim of the fenestra thus lies lateral to the area of origin, i.e. the more
fibrous part of the temporal muscle, and any bulging of the muscle mass on
contraction could hardly have occurred through the fenestra, which is in any
case small. The body or fleshy part of the muscle lies lower down and is covered
by the deep lower arch (squamosal). The fenestra thus seems to have lost its |
primary function! The forward position of the lower jaw articulation with the
concomitant great depth of the squamosal arch greatly lengthens the muscle
mass and this increased length would compensate for a decreased ability to
bulge locally.

In the Styracocephalidae the pachyostosis has caused a great reduction
of the size of the temporal fenestrae, which are situated widely apart. Differen-
tial bone thickening has resulted in the rim of the fenestra being formed solely
by the postorbital and squamosal.

A forward shift of the jaw articulation as in the tapinocephalids with the
deep squamosal low arch has affected the working of the temporal muscles as
described above for the tapinocephalids.

In the Estemmenosuchidae the fenestra is large particularly in length due
to a forward extension into the jugal as well as a posterior outflaring of the
squamosal. The dorsal flange of the postorbital lying in the skull table is
shortened and fails to reach the squamosal. The intertemporal width is large.
The temporal muscle thus in part originates from the lateral face of the parietal.

In the Otsheriidae the temporal fenestra is large, due to the outflaring of
the squamosal both laterally and posteriorly as well as the reduction of the
width of the intertemporal skull table. It is still primitively bounded by the
three bones—postorbital, squamosal and jugal, but both the upper and the
lower arch are modified.

In the upper arch the posterior flange of the postorbital only provides a
narrow edge to bound the upper border of the fenestra and the temporal
muscles arise in part from the latero-ventral edge of this splint-like flange.

The lower arch is fairly shallow but is deeper than broad, thus lying
vertically, with a large contribution from the jugal. The postero-ventral corner
of the squamosal is prolonged ventrally in the form of a pedicel to hold the
quadrate in a position low down in the skull and also far posteriorly. The
adductor chamber is roomy and the temporal muscles short but bulky.

In the Dromasauridae a single skull of Galeops from the Tapinocephalus
zone is inadequately known.

The temporal fossa is short but deep and apparently bounded by the
postorbital, squamosal and jugal.

The squamosal has a long ventrally directed pedicel similar to that of
Otsheria. ;

In the Endothiodontidae and Dicynodontidae the oldest known forms
from low down in the Tapinocephalus zone already have the temporal region,

THE EARLY THERAPSIDS 23

which is so typically unique for all the Dicynodontia and basically retained
throughout the long history of this group.

Of all the early therapsids the Endothiodontidae and Dicynodontidae
show the greatest modification in the temporal region from the primitive
pelycosaur condition.

The temporal fenestra is greatly enlargéd. ‘The fore-aft diameter is uniquely
lengthened due to the anterior position of the orbit accompanied by the
slenderness of the postorbital bone and the posterior flaring of the squamosal.
The medio-lateral diameter is enlarged due to the reduction of the inter-
temporal width of the skull table.

The upper postorbital-squamosal arch is fairly primitive except for the
greatly lengthened dorsal flange of the postorbital which is somewhat bent
down laterally from the horizontal. ‘The temporal muscle arises in part from
the lateral edge and under surface of the postorbital and the upper edge of the
squamosal.

The greatest modification is seen in the structure of the squamosal.

In the zygoma the squamosal, originally lying in a vertical plane, is bent
down laterally to lie in a horizontal plane with the original dorsal edge now
forming the lateral edge of the bar. In addition the squamosal extends far
anteriorly to terminate in a plane ventral to the orbit and the jugal is almost
completely excluded from the lateral face of the zygoma.

The downward growth of the postero-ventral corner of the squamosal to
form a pedicel, first seen in Ofsheria, is also greatly modified. In Otsheria the
face of this pedicel is lateral. In the Dicynodontia this face is now directed much
anteriorly with the original posterior edge turned outwards to form a sharp
lateral edge. To the lower part of this oblique face the quadratojugal is applied.
Above the quadratojugal is the area of origin of the masseter mainly from fascia
attached to the sharp lateral squamosal edges.

In all the four early therocephalian families (Pristerognathidae, Lyco-
suchidae, Alopecodontidae and Scaloposauridae) the temporal fenestra is large
and faces more dorsally than laterally and the adductor chamber is very
roomy. Here also the posterior flange of the postorbital is greatly reduced and
lies as a small splint lying vertically and applied to the lateral face of the
parietal, which now forms the greatest part of the upper border of the temporal
fenestra.

The intertemporal width is greatly reduced and this part of the skull table
is normally developed into a sagittal crista of varying width and height.

Here the temporal muscle had its main origin from the lateral face of the
parietal.

POSITION OF THE QUADRATE

The foregoing comparison of the temporal fenestra and its arches in the
early therapsids drew our attention to the origin of muscles of the capiti-
mandibularis mass.

24 ANNALS OF THE SOUTH AFRICAN MUSEUM

The function of these adductors is related to the position of the jaw
articulation and the insertion on the lower jaw. These two aspects will now be
considered. In the pelycosaurs the quadrate is situated far posteriorly just
posterior to the plane of the occipital condyle.

In the earliest Gorgonopsia the quadrate lies just anterior to the plane of
the occipital condyle.

In the Eotitanosuchidae the quadrate apparently lies in the plane of the
condyle.

In the Brithopidae the quadrate has shifted somewhat anteriorly to the
plane of the condyle.

In the Anteosauridae the quadrate has shifted still further anteriorly and
due to the backward tilt of the occiput lies very far anterior to the upper edge
of the occiput.

In the early therocephalian families the quadrate still lies in the primitive
posterior position.

In the Titanosuchidae, Tapinocephalidae, Styracocephalidae and Estem-
menosuchidae the quadrate lies very far forward of the plane of the condyle
and still more of the plane of the upper occipital edge.

In the Otsheriidae the quadrate would appear to have been situated
somewhat anterior to the plane of the condyle.

This is also the position of the quadrate in the early Endothiodontidae and
Dicynodontidae.

INSERTIONS OF THE TEMPORAL MUSCLE

The primitive nature of the insertions of the m. capitimandibularis is still
evident in the sphenacodonts and this condition is basically retained in the
early therapsids.

The most significant change is seen in the Gorgonopsia and the Thero-
cephalia, where the dentary developed a prominent free-standing dorso-
posteriorly directed coronoid process for the reception of the temporalis.

No forms are known in which this development is incipient. Low down in
the Tapinocephalus zone it is simply there fully developed in the oldest Gorgonop-
sia and ‘Therocephalia.

As has already been mentioned above, the origin of the subdivided capiti-
mandibularis is in the early Gorgonopsia still of primitive nature, but that in
the earliest therocephalians it is already highly specialized in a mammalian
direction. This very definite difference in origin of the muscles is remarkably
not accompanied by any noteworthy change in the insertion.

The development of the coronoid process in these two groups thus appears
to have been caused by a pull exerted by the adductors in a primitive way in
the case of the gorgonopsians on the one hand and by an advanced mammal-
like way in in the therocephalians.

The presence of a coronoid in the Gorgonopsia can thus at most be con-
sidered as a parallel development and not one in a mammalian direction.

THE EARLY THERAPSIDS 25

THE REFLECTED LAMINA OF THE ANGULAR

This structure is a feature common to all the early therapsids and is con-
cerned with the insertion of the anterior pterygoid and superficial masseter muscles.

This is also the condition in the higher sphenacodonts and held as strong
evidence of their consanguinity with the therapsids.

In the early Dicynodontia the structure of the reflected lamina differs
somewhat from that of the other early therapsids. This is probably associated
with a difference in the origin of the anterior pterygoid muscle for we know
that in the early Endothiodontidae and Dicynodontidae, but not in the Otsherii-
dae, the lateral pterygoidal flange is greatly reduced.

THE MARGINAL TEETH

Of the oldest therapsid families, 10 have a carnivorous dentition and 7
are herbivorous, with the adaptations showing a quite remarkable diversity.

In the primitive pelycosaurs the tooth row is long and consists of simple
pointed teeth. In the maxilla a pair of teeth well back in the row are enlarged
as ‘canines’. The replacement is distichial.

In the advanced sphenacodonts the enlarged ‘canines’ are situated near
the front of the maxillary row. The functional replacement is by a member of
the same tooth family but the upper canines are replaced alternately.

In the early therapsids the tooth row is reduced in all the carnivorous
families, but is secondarily lengthened in the herbivorous Titanosuchidae,
Tapinocephalidae and Styracocephalidae. In the Otsheriidae and Venyukovii-
dae the row is still fairly long but highly specialized. In the Endothiodontidae
and Dicynodontidae development of horny sheaths radically reduces the
marginal teeth.

In the early therapsids the upper canine when present is the first tooth in
the maxilla in all the families, except the Scaloposauridae and Alopecodontidae.

A lower canine is present, except where secondary lost as in the herbivorous
Endothiodontidae, Dicynodontia and Tapinocephalidae but persists in the
herbivorous Titanosuchidae and Styracocephalidae. In the Lycosuchidae
there are a pair of upper canines replaced alternately but functionally by a
member of the same family.

In the pelycosaurs there appears to be no limit to the tooth replacement.
This is also the case in the Titanosuchidae and probably also in the Tapino-
cephalidae. In the other families there is evidence of limited replacement in
the Gorgonopsia and Therocephalia. The condition in the other early therapsid
families is unknown.

In the early therapsids the upper teeth in occlusion lie lateral of the lower
teeth, but in the Anteosauridae, Titanosuchidae, Tapinocephalidae and Styraco-
cephalidae the incisors intermesh, so do the canines in the Titanosuchidae and
the whole battery in the Tapinocephalidae.

In the sphenacodonts the teeth are simple and pointed.

In the early therapsids considerable variations have arisen.

26 ANNALS OF THE SOUTH AFRICAN MUSEUM

In the Eotitanosuchidae the primitive condition is retained.

In the Brithopidae and Anteosauridae the incisors are progressively leng-
thened and in the latter the postcanines become bulbously spatulate.

In the carnivorous gorgonopsian and therocephalian families the distal
edge of the incisors, canines and post-canines becomes serrated.

In the herbivorous Titanosuchidae, ‘Tapinocephalidae and Styracocephali-
dae the incisors develop a talon and heel; in the Tapinocephalidae the canine
and the postcanines develop a similar talon and heel, but in the Styracoce-
phalidae only the postcanines. The canine remains fairly normal in the
Titanosuchidae and Styracocephalidae. In the Titanosuchidae the long row
of postcanines are spatulate with serrated edges.

In Otsheria and Venyukovia the teeth become bluntly conical.

In the Dicynodontia the incisors disappear, the upper canines present or
absent and the lower canine always absent. There are no postcanines in the
Dicynodontidae and in the Endothiodontidae they are reduced, and displaced
medially from the jaw margin.

PALATE

In the sphenacodonts the pterygoids do not meet in the median line
posterior to the interpterygoid vacuity. The quadrate ramus is deep and
strong. The transverse ramus is strong, prominent and dentigerous. The
choana is long and situated anteriorly. There is no suborbital foramen or
fenestra. The posterior end of the vomers is spatulate and the vomerine bar
lies low down in the skull.

The early therapsids manifest considerable variations from the primitive
pelycosaur palatal structure. They all have one advance in common, viz. that
the basipterygoid joint is no longer freely movable. The quadrate ramus
becomes weaker in the Gorgonopsia, Therocephalia and Dicynodontia. The
anterior ramus is (generally) reduced; greatly so in the Dicynodontia. The
transverse ramus is progressively weakened in the series Otsheriidae—Venyu-
koviidae—Dicynodontia and becomes edentulous in practically all the
therapsids.

The choana is somewhat shortened in the Brithopidae and Anteosauridae
but in the dicynodontian families it is both shortened and pushed backwards
by the enlarged palatal process of the premaxilla. Only in the therocephalian
families is a well-developed suborbital fenestra present.

Only in the Eotitanosuchidae, Gorgonopsia and Dicynodontia is the
vomer well raised above the general palatal level.

THE LOCOMOTOR APPARATUS

In the 19 known families of early therapsids the structure of the girdles
and limbs is not adequately known in 11 of these families. Any comparative
consideration must thus be tentative.

An overall picture of the locomotor apparatus in the other 8 families
discloses considerable adaptive radiations, but in all there is an advance

THE EARLY THERAPSIDS 27

beyond the crawling habit of the pelycosaurs to a slinking habit in the brithopids,
anteosaurids and hipposaurids and a more upright walking gait in the endo-
thiodontids, dicynodontids, the 3 therocephalian families and the titanosuchids
and tapinocephalids.

We may commence by attempting to give a picture of the diversity
exhibited in the structure of the girdles and limbs in these early therapsids.

In the hipposaurids the procoracoid has not been ousted from the glenoid;
in all the others it has been ousted.

The procorocoid is enlarged in the hipposaurids and the Dinocephalia,
small in the Dicynodontia and moderate in the other families.

Only in the Dicynodontia is an acromion process developed on the scapula.
This feature, typical of the mammals, is however no evidence of affinity of the
Dicynodontia to the mammals, but rather a case of parallelism as it is also
found in the contemporary Pareiasauridae—a cotylosaur family with no
affinity to the mammals. An ossified sternum is developed in the Dicynodontia
and Gorgonopsia but in none of the other groups. In all the early therapsids
the axial muscles have been forced off the outer face of the ilium and the iliac
height is increased.

The anterior iliac process is incipient in the hipposaurids, moderate in
the pristerognathids and brithopids and anteosaurids, well developed in
titanosuchids and tapinocephalids and great in the endothiodontids and dicyno-
dontids but undeveloped in the dromasaurids.

Only in the endothiodonts and dicynodontids has the acetabulum moved
to the anterior pelvic border.

The pubo-ischiatic plate retains its great primitive length in hipposaurids,
pristerognathids and anteosaurids. ‘The pubic part is shortened in titanosuchids
and tapinocephalids and greatly so in endothiodontids and dicynodontids,
where a pubo-ischiatic fenestra replaces the pubic foramen present in all the
other families.

The pelvic symphysis is strongly ossified in the hipposaurids, pristerog-
nathids and anteosaurids but weak in all the other families.

Humerus

All the early therapsids have lost the primitive strap-like caput of the
humerus and there has been an untwisting of the proximal and distal ends
relative to each other. These ends remain expanded to various degrees, but are
greatly reduced in hipposaurids.

In hipposaurids no epicondylar foramina are present. In endothiodontids,
dicynodontids, Tapinocephalidae and Anteosauridae there is no ectepicondylar
foramen but it is present in brithopids, titanosuchids and Therocephalia.

Femur

In all the early therapsids there has been a preaxial shift and a shortening
of the caput femoris but to varying degrees in the various families. The distal
condyles have shifted to lie in the same plane and this distally so that the knee

28 ANNALS OF THE SOUTH AFRICAN MUSEUM

joint becomes a simple hinge well adapted to a more upright disposition of the
limb.

Only in the Titanosuchidae and Tapinocephalidae has the femur become
greatly broadened.

Forefoot

The primitive phalangeal formula of 2, 3, 4, 5, 3 has been reduced to
2, 3, 4, 4, 3 in the hipposaurids and to 2, 3, 4, 3, 3 in the anteosaurids and to
2, 3, 3, 3, 3, in all the other families of the early therapsids.

Hindfoot

In the tarsus the primitive medial central has been lost in all the early
therapsids, where this structure is known, and the phalangeal formula reduced
from the primitive 2, 3, 4, 5, 4 to 2, 3, 4, 4, 3 in the hipposaurids, 2, 3, 4, 3, 3,
in brithopids and anteosaurids and 2, 3, 3, 3, 3 in all the other early therapsids.

In the pristerognathids the astragalus tends to overlie the calcaneum and
in the hipposaurids a sustentaculum tali is developed as well as a tuber calcis.

Now, what does this rather great diversity in the structure of the locomotor
apparatus signify ?

The main variations are apparently towards the acquisition of a greater
degree of active movement than that of crawling—on the one hand that possible
in a slinking habit and on the other in a more upright walking habit.

Is the improved locomotor ability correlated in any way with an improved
masticatory ability?

In the herbivorous families the achievement of a walking gait would
increase the area that can be grazed and the ability to reach higher ground
would bring these reptiles into contact with hardier and more fibrous plants
than those flourishing in more marshy terrain.

In the Otsheriidae and Venyukoviidae the bluntly conical teeth appear
to be adapted to a coarser fare.

The horny jaw sheaths and plates of the Endothiodontidae and Dicyno-
dontidae together with the fore-and-aft sliding of the lower jaw would greatly
help in the shearing and milling of fibrous vegetable matter.

In the Titanosuchidae and Tapinocephalidae and Styracocephalidae the
intermeshing talon-and-heel teeth showing considerable abrasion are obviously
well adapted for piercing and crushing tough fibrous plants.

In the Lotitanosuchus—Anteosaurus series of carnivores the progressive
development of a formidable battery of long pointed intermeshing incisors
together with the strong canines would enable these reptiles to execute a
strong piercing and jerking bite into the flesh of the larger herbivores. The
progressive decrease in the role of the postcanines would accompany this
method of biting.

A slinking habit of locomotion indicates that these carnivores did not run
after their prey but rather lay in ambush and then pounced.

THE EARLY THERAPSIDS 29

The early gorgonopsians with their moderate anterior teeth and reduced
postcanine series and slinking but agile locomotory ability probably pounced
on the small contemporary Dicynodontia or could have at times been carrion
eaters.

The early therocephalians with their limbs well adapted to a more upright
walking and running gait could pursue and overcome even some of the larger
herbivores. With the postcanines greatly reduced in some genera, the front
part of the jaws was mostly in action pulling and tearing out lumps of flesh.

The scaloposaurids with a long tooth row and small canines and with
some cuspidate postcanines were better adapted as insectivores.

The variations in the adductor muscles in the above groups of early
therapsids appear to be well correlated to both the varied dentitions and modes
of locomotion.

DIVERSE STRUCTURES

The quadratojugal has variable relations in the early therapsids. In all it
is, however, much reduced in size and never enters the lower temporal arch
as it does in some members of all three of the pelycosaurian suborders.

Primitively a surface bone of the postero-lateral corner of the skull,
flanking the quadrate, it first tends to move medially in some of the higher
sphenacodonts to rest on the quadrate above the lateral condyle as a bone of
reduced size.

This process is seen continued in the Gorgonopsia, Therocephalia,
Brithopidae and Anteosauridae. Whereas in the Dicynodontia the quadrato-
jugal becomes a plate applied to the antero-lateral face of the everted squamosal,
in the Titanosuchidae and Tapinocephalidae the quadratojugal, variable in
size and shape, still forms part of the lateral skull surface. Does this indicate
an origin from different pelycosaurian ancestors ?

The lacrimal, primitively a long bone stretching from orbit to nostril, is
reduced to an anterior circumorbital bone in all the therapsids. It is reduced
in some sphenacodonts but also in the edaphosaurian, Mycterosaurus.

The supratemporal is absent in all therapsids and in all one coronoid is
lost, but both coronoids are absent in the Dicynodontia.

In the early therapsids the preparietal is a new acquisition in only the
Endothiodontidae and Dicynodontidae as well as the Hipposauridae and
Galesuchidae. In the primitive dicynodontian family, Otsheriidae, there is
however no preparietal. It is also absent in the possible gorgonopsian forerunner,
Eotitanosuchus and Phthinosuchus.

In the early Gorgonopsia the preparietal lies anterior to the pineal foramen,
but forms its anterior border in the early Dicynodontia which implies a different
raison détre.

The dorsal process of the premaxillaries varies in length in the pelycosaurs,
being long in ophiacodonts, short to medium in sphenacodonts and edapho-
saurs.

30 ANNALS OF THE SOUTH AFRICAN MUSEUM

It is also variable in the early therapsids, being long in eotitanosuchids,
brithopids and anteosaurids; very long in the Titansuchidae and Tapino-
cephalidae. In the dicynodontian families it is long in the primitive Otsheriidae
and Venyukoviidae, but short in the more advanced early Endothiodontidae
and Dicynodontidae.

Arranged into series as to length of premaxillary process we have:

Gorgonopsia
Short to medium—Sphenacodontia ————> Theracephalia
Short in Captorhinomorpha Advanced ai ea
long in Ophiacodontia — Eotitanosuchidae
and
Dinocephalia
Primitive Dicynodontia.

Together with the development of a reflected lamina of the angular we
see a reduction in the role of the posterior mandibular bones in the higher
sphenacodonts and in all the early therapsids and this reduction is more pro-
nounced in those therapsids where the dentary develops a coronoid process.

The braincase has its sidewall largely open in the Therocephalia and
Dicynodontia, but much less so in the Anteosauridae, Titanosuchidae and
Tapinocephalidae. In the Therocephalia there is no downward directed flange
of the parietal whereas in all the other early therapsids, where known, it is
present.

The sphenethmoidal complex is weakly ossified in the early Therocephalia
but well developed in the Dinocephalia and in those early Dicynodontia where
it has been studied. In the Dicynodontia it lies far anteriorly and has no con-
tact with the prootic, whereas in the Dinocephalia contact is made above the
lateral fenestra.

The fenestra ovalis lies low down in the skull in all the early therapsids,
but there are considerable variations in the structure of the stapes. The dorsal
process of the stapes is reduced in the Tapinocephalidae and absent in all the
other early therapsids. A stapedial foramen usually present is absent in the
hipposaurids, brithopids, anteosaurids and in all the early Dicynodontia where
the stapes has been described.

The exoccipital apparently does not enter the floor of the braincase in
Captorhinus. This is definitely the case in the early Endothiodontidae and
Dicynodontidae and the early Gorgonopsia, whereas in Dimetrodon and all the
Dinocephalia it forms the whole posterior part of the brain floor.

The prootics do not meet in the middle line in Captorhinus, but do meet in
Dimetrodon. ‘They meet in all the Dinocephalia, where known, but not in the
Dicynodontia and just meet in the Therocephalia.

The dorsum sellae is very high in Captorhinus and high in Dimetrodon. In
the former it is formed by the basisphenoid, whereas in the latter by the prootic.

THE EARLY THERAPSIDS 31

In the Dinocephalia the upper part of the dorsum sellae is formed by the
prootic, and in the Therocephalia the prootic just enters, whereas it is excluded
in the early Dicynodontia. The sella turcica is deep in Captorhinus and Dimetrodon.
This is also the case in all the Dinocephalia but is shallow in the ‘Therocephalia
and Dicynodontia.

The quadrate ramus of the pterygoid’ is strong in the sphenacodonts and
is greatly strengthened in the Dinocephalia but weakened in Therocephalia,
Gorgonopsia and Dicynodontia.

MORPHOLOGICAL SERIES

In the foregoing the extent of the variations observed in the assemblage of
early therapsids has been given in some detail. The result being that one cannot
see the wood for the trees.

We must now consider whether these divergencies can be arranged in some
orderly manner on a basis of possible consecutive ascending morphological
stages.

The early therapsids form a fauna of discrete types of animals living
together during a definite interval of time. They can thus, broadly speaking,
be considered as contemporaries and thus some cannot be conceived as being
ancestral to others.

What we can, however, attempt to do is to determine the possibility of
arranging the animals exhibiting these various structural features in series,
one derivable from the others in a morphological sense.

SERIES I
Captorhinidae — Pelycosauria — Eotitanosuchidae — Brithopidae — Anteo-
sauridae.

Consecutive steps in the following features:

(a) Temporal fenestra

Absent in captorhinids — small or double in pelycosaurs — large in eotitano-
suchids —> larger in brithopids — very large in anteosaurids. This progressive
increase is mainly due to lateral and posterior outflaring of the squamosal.

(b) Intertemporal skull table

Wide in pelycosaurs — still wide in Eotitanosuchus > greatly reduced in
brithopids — but less reduced in anteosaurids.

(c) Posterior process of postorbital

Horizontal surface bone in pelycosaurs — just starting to tilt down in Eotztano-
suchus —> tilting progressively increased in brithopids and anteosaurids to
culminate as a bone lying nearly vertically flanking the parietal inside the
temporal fossa.

32 ANNALS OF THE SOUTH AFRICAN MUSEUM

(d) Area of origin of the temporal muscle on the postorbital

Under surface of postorbital in pelycosaurs —~ moving to lateral edge and
dorsal face in Eotitanosuchus — on the morphological dorsal face in brithopids
and anteosaurids which progressively becomes a functionally lateral face.

(e) Insertion of temporal muscle

Notwithstanding the changes in the origin of the muscle in this series, the
insertion on the mandible remains constant from sphenacodontid to anteosaurid.

(f) Jaw articulation

In captorhinids this lies posteriorly in a plane with the occipital condyle and
level with the alveolar border — this is still the position in most pelycosaurs,
but in the higher sphenacodonts it has shifted downwards — posterior and
low in eotitanosuchids —> shifted both anteriorly and ventrally in brithopids
and anteosaurids.

(g) Reflected lamina of the angular

Absent in captorhinids — still absent in most pelycosaurs, but developed in
the higher sphenacodonts — progressively better developed in eotitanosuchids,
brithopids and anteosaurids.

(h) Marginal tooth row and ‘canines’

In captorhinids the tooth row is long, without ‘canines’ — in pelycosaurs long
to very long, canines absent or variously present in the three pelycosaur
groups, but strong in KHothyris and most sphenacodontids — row reduced in
eotitanosuchids, but strong definite canine present — progressive reduction
of number of post-canines in brithopids and anteosaurids, canines very strong
and incisors progressively lengthened to culminate in the very long inter-
meshing set of Anteosaurus.

(2) Septomaxilla, maxilla and lacrimal

In captorhinids the septomaxilla is intranarial, the lacrimal enters the narial
border and the maxilla is low — these relations are retained in nearly all the
pelycosaurs but in Mycterosaurus, Sphenacodon and Dimetrodon the lacrimal fails
to reach the naris and the maxilla becomes high. This may be related to the
greater development of ‘canines’ in these three genera, but other pelycosaurs
have enlarged ‘canines’ without affecting the primitive relations of the maxilla
—> in eotitanosuchids, brithopids and anteosaurids the septomaxilla extending
backwards becomes a bone of the lateral surface, the lacrimal fails to reach the
nostril and the maxilla is high.

(j) Dorsal process of the premaxilla

Short in captorhinids moderately lengthened in some pelycosaurs but still
fairly short in sphenacodonts —> greatly lengthened in eotitanosuchids, britho-
pids and anteosaurids.

THE EARLY THERAPSIDS 33

(k) Braincase

Insufficiently known in this series, but in both Dimetrodon and Anteosaurus the
sphenoidal complex is well ossified and the exoccipital and the prootic enter
the floor of the braincase and the prootics meet in the middle line in the
dorsum sellae.

(l) Locomotor apparatus

Insufficiently known, but in all brithopids and anteosaurids the femur has
become a long, slender curved bone.

SERIES II :
Styracocephalidae

Brithopidae —- Titanosuchidae — Tapinocephalidae
Estemmenosuchidae

(a) Temporal fenestra

In brithopids large — slightly reduced in titanosuchids — moderately to very
greatly reduced in tapinocephalids (except in Avenantia and Riebeeckosaurus).

(b) Intertemporal skull table

Moderately wide in brithopids > so also in titanosuchids — moderately to
enormously widened in tapinocephalids (except in Avenantia and Riebeeckosaurus).

(c) Posterior process of the postorbital

In brithopids long and high with good contact with the squamosal — in
titanosuchids reduced to a splint and just meeting the squamosal — in tapino-
cephalids shortened still further so that in some forms it fails to reach the
squamosal.

(d) Area of origin of the temporal muscle on the postorbital

In brithopids from the fairly large tilted (dorsal) face — in titanosuchids this
area is reduced and the origin transferred more on to the parietal — this is
carried further in the tapinocephalids where the total area is small (except in
Avenantia and Riebeeckosaurus).

(e) Insertion of temporal muscle

The primitive pelycosaurian position is retained throughout the series.

(f) Faw articulation

In brithopids anterior to the plane of the occipital condyle — in titanosuchids
still further anteriorly — in tapinocephalids very far anteriorly.

(g) Marginal teeth

In brithopids moderately long pointed incisors, well-developed canines, fairly
long postcanines row of bluntly conical teeth.

34 ANNALS OF THE SOUTH AFRICAN MUSEUM

In titanosuchids we find a radical change to a herbivorous dentition;
strong pointed canine is retained, the strong intermeshing incisors have
developed a piercing talon and crushing heel and the very long postcanine
row has cuspidate spatulate crowns.

In tapinocephalids this process is carried farther in that the canine has
disappeared as such and the very long series consists of isodont talon-and-heel
teeth, all intermeshing but the anterior teeth are weaker than the incisors of
the titanosuchids.

(g) Dorsal process of the premaxilla

In brithopids this is of moderate length intercalated between the nasals, greatly
lengthened in both titanosuchids and tapinocephalids and nearly reaching the
frontal.

(h) Braincase

Little known in the brithopids.

In both titanosuchids and tapinocephalids the lateral wall is well ossified
and the sphenoidal complex strongly ossified; the exoccipital and prootic
enter the floor of the braincase; the prootics meeting in the middle line form
part of the dorsum sellae.

() Locomotor apparatus

In brithopids the girdles and limbs are fairly lightly built whereas in both
titanosuchids and tapinocephalids they are massive to very massive.
Styracocephalidae

With a persistent canine and the development of weak talon-and-heel
incisors and postcanines and a secondary broadened intertemporal skull table
and reduced temporal fossa. Styracocephalus can be derived from the titanosuchids
as a branch somewhat divergent from the tapinocephalid branch.

Estemmenosuchidae

It is difficult to place Estemmenosuchus. The shagreen of oes teeth is
reminiscent of early pelycosaurs.

The broad intertemporal region with the upper part of the postorbital
lying on the dorsal surface overhanging the temporal fenestra are eotitanosuchid
features as is the large temporal fenestra. The incisors and canines are like
those of the brithopids.

The great downward and forward shift of the quadrate and the long
series of postcanines parallel features of both the titanosuchids and the tapino-

cephalids.

THE EARLY THERAPSIDS 35

SERIES III

Pelycosauria — Phthinosuchidae — Hipposauridae — Galesuchidae.

(a) Temporal fenestra

This is small or double in the pelycosaurs —> suddenly very large in Phthino-
suchus > but only moderately enlarged in ‘Hipposaurus > then again large in
the galesuchids. Clearly not a consecutive series.

(b) Intertemporal skull table

Wide in pelycosaurs —> remains wide in Phthinosuchus -> becomes very wide in
Hipposaurus > but somewhat reduced in the galesuchids. Again not a con-
secutive series.

(c) Posterior process of the postorbital

In the whole series it persists as a horizontal surface bone overhanging the
temporal fenestra. Long in pelycosaurs — short in Phthinosuchus — very long in
Hipposaurus —- long in galesuchids. In Phthinosuchus it is almost entirely excluded
from the edge of the skull table due to the development of a lappet of the
squamosal extending far anteriorly and lying laterally of the postorbital.

(d) Area of origin of the temporal muscle on the postorbital

In the whole series the origin remains on its under surface, but in Phthinosuchus
also from the under surface of the squamosal lappet.

(e) Insertion of the temporal muscle

Partially inserted on the upper and outer face of the dentary in pelycosaurs >
this primitive insertion retained in Phthinosuchus > but in hipposaurids and
galesuchids mainly on the strongly developed coronoid process.

(f) Faw articulation

Posterior position in pelycosaurs —> shifted anteriorly in Phthinosuchus, Hipposaurus
and galesuchids. But situated far ventrally in Hipposaurus.

(g) Dorsal process of the premaxilla

Moderately long in pelycosaurs > unknown in Phthinosuchus — very short in
Hipposaurus and the galesuchids.

(h) Marginal teeth

Postcanines progressively reduced in the series. A single well developed canine
present in Phthinosuchus, Hipposaurus and the galesuchids.

(2) Locomotor apparatus

Unknown in Phthinosuchus and the early galesuchids. In Hipposaurus the anterior
iliac process remains weak as in pelycosaurs, an ossified sternum is developed;

36 ANNALS OF THE SOUTH AFRICAN MUSEUM

the limbs have become long and slender as in anteosaurids. The tarsus is greatly
specialized with the development of a tuber calcis and a sustentaculum tali; the
phalangeal formula only slightly reduced, probably 2, 3, 4, 4, 3.

(7) Preparietal

Absent in pelycosaurs and in Phthinosuchus, suddenly present in Hipposaurus and
all other gorgonopsians.

(k) Vomer
Lying in general plane of palate with broad posterior end in pelycosaurs >
raised or vaulted, with broad posterior end in Phthinosuchus — raised or vaulted
in Hipposaurus and galesuchids, posterior end tapering and intercalated between
palatines.

SERIES IV

Pelycosauria —> Otsheriidae —- Endothiodontidae — Dicynodontidae.

(a) Temporal fenestra

In pelycosaurs small or double, bounded by postorbital, squamosal and jugal
in the higher sphenacodonts >in Otsheria large, bounded by postorbital,
squamosal and jugal — in endothiodonts and dicynodonts very large mainly
due to posterior outflaring of the squamosal, with jugal participation in its
border reduced.

(6) Intertemporal skull table

Wide in pelycosaurs—> moderate in Otshera and in endothiodonts and
dicynodonts.

(c) Posterior process of the postorbital

A horizontal surface bone in pelycosaurs and fairly long — in Otsheria showing
only an edge as a surface bone and fairly long — in endothiodonts and dicyno-
donts slanting downwards laterally and very long.

(d) Area of origin of the temporal muscle on the postorbital

In pelycosaurs from its undersurface — in Otsheria from its latero-ventral edge >
in endothiodonts and dicynodonts from the dorsal surface now lying at a slant.

(e) Squamosal

In most pelycosaurs the postero-lateral corner of the squamosal lies in the plane
of the alveolar border, but in the higher sphenacodonts (and Edaphosaurus) it
lies far ventrally and the lower temporal arch, in which the jugal plays a large
part, lies in a vertical plane.

In Otsheria the process of the squamosal lies far ventrally and the lower arch,
in which the jugal plays a large part, still lies in a vertical plane.

THE EARLY THERAPSIDS 37

In endothiodonts and dicynodonts both these features are suddenly radically
changed. The ventral process, still extending far ventrally, is uniquely everted
to present a sharp lateral edge and an anterior face to which the quadrato-
jugal is applied as a flat sheet of bone. The anterior process of the squamosal,
now forming most of the lower temporal arch, is also everted and now lies in a
nearly horizontal plane with its morphological dorsal edge facing laterally.

(f) Temporal muscles

In sphenacodonts the origin of the temporal muscles is mainly from the under
surface of the bones of the skull roof —> in Otsheria it is partly shifted to the
edge of the postorbital >in the early endothiodonts and dicynodonts the
unique and radical changes in the nature of the squamosal is due to the radical
changes in the areas of origin of both the temporal and masseter. Noteworthy
is that the masseter lying medially of the zygomatic arch in the pelycosaurs
now has its origin from the antero-lateral face of the squamosal below the
zygomatic arch.

The insertion of these muscles has not changed much from the pelycosaur
condition and no coronoid process is developed, but there is already an indica-
tion of a lateral flange on the dentary well developed in some later Dicynodontia.

(g) Reflected lamina of the angular

In the higher sphenacodonts we have the first development of this structure >
it is unknown in Otsheria, but in Venyukovia it is well developed — it is present
in the early endothiodonts and dicynodonts but in nature differs considerably
from that in the other contemporary therapsids indicating a difference in the
insertion of the anterior pterygoid muscle. A fenestra in the lower jaw is known
in Ophiacodon but in no other pelycosaur; it is also present in Venyukovia and the
endothiodonts and dicynodonts.

In some pelycosaurs with ‘canines’ the maxilla is high — but in Otsheria
without canines it is suddenly very high as it is in endothiodonts with or without
canines. If the increase in maxillary height is due to the presence of canines, as
has been maintained, then Otsheria must have inherited this feature from an
ancester with canines.

(h) Faw articulation

This lies posteriorly in the series pelycosaur — Otsheria + endothiodonts and
dicynodonts.

In pelycosaurs it lies high up (except in the higher sphenacodonts and
Edaphosaurus) —> very far ventrally in Otsheria and endothiodonts and dicynodonts.

(2) Septomaxilla, lacrimal and maxilla

In pelycosaurs the septomaxilla lies internarially — in Ofsheria there is a small
lateral face — but in the endothiodonts and dicynodonts it is again internarial.
The lacrimal fails to reach the naris in some pelycosaurs (e.g. Dimetrodon

38 ANNALS OF THE SOUTH AFRICAN MUSEUM

and Mycterosaurus) — in Otsheria it is greatly shortened (but still long in Venyu-
kovia) — in endothiodonts and dicynodonts it has become a short bone of the
anterior orbital border.

(j) Snout

In nearly all pelycosaurs the snout is long with the orbit and nostril far apart >
in Otsheria it is greatly shortened (but still fairly long in Venyukovia) — in the
early endothiodonts and dicynodonts the snout is very greatly shortened with
the naris very near the orbit.

(k) Dorsal process of the premaxilla
Moderately long in pelycosaurs — long tapering intercalation between nasals
in Otsheria (and Venyukovia) —> but very short in endothiodonts and dicynodonts.

(1) Palatal face of the premaxilla

Absent in pelycosaurs and choana extending far anteriorly — well developed
in Otsheria and choana pushed posteriorly but not reaching the palatine (very
well developed in Venyukovia, choana pushed back but still long, makes contact
with the palatine) well developed in endothiodonts and dicynodonts,
choana pushed backwards and greatly shortened, it sometimes makes contact
with the palatine.

(m) Vomer

In pelycosaurs paired and lying in general plane of palate broadened pos-
teriorly + unpaired in Otsheria lying low down broad posteriorly (paired in
Venyukovia) —> raised above (vaulted) general plane of palate, reaching inter-
pterygoid vacuity in Dicynodontia.

(n) Lateral ramus of pterygoid
Strongly developed in pelycosaurs — quite strong in Oftsheria (but weak in
Venyukovia) —> absent in endothiodonts and dicynodonts.

(0) Marginal teeth

In pelycosaurs the tooth row is long, with ‘canines’ in some forms, pointed in
most —> moderately long row in Otsheria, with incisors enlarged, no ‘canine’,
postcanines spatulate (in Venyukovia ‘canine’ present, teeth bluntly conical,
some with crushing face) >in early endothiodonts and dicynodonts the
anterior part of the jaws is edentulous with development of horny sheaths,
strong upper canines present or absent, reduced postcanines shifted away from
jaw margin in endothiodonts but absent in dicynodonts.

(p) Braincase

The exoccipital and prootic form floor of braincase in Dimetrodon and prootics
meeting in middle line in the dorsum sellae — unknown in Otsheria — in early
endothiodonts and dicynodonts the exoccipital and prootic do not enter the
floor and the prootic does not enter the dorsum sellae.

THE EARLY THERAPSIDS 39

The sphenoidal complex is well developed in Dimetrodon, situated far pos-
teriorly but the lateral wall is widely open — unknown in Otsheria — very well
developed in endothiodonts and dicynodonts but situated very far anteriorly
with the result that the lateral wall is widely open because in addition the
prootic has little anterior development.

(q) Locomotor apparatus

Primitive in pelycosaurs —- unknown in Otsheria — in the earliest endothiodonts
and dicynodonts it is already highly specialized. The scapula has a well-
developed acromion process; there is a strongly developed ossified sternum,
the ilium has an enormous anterior iliac process; the pubis is greatly reduced;
there is a large pubo-ischiatic fenestra; the phalangeal formula is reduced to

2, 3, 35 3, 3-
Galeops

This very imperfectly known form from the Tapinocephalus zone shows a
few features similar to those of the early Dicynodontia.

The temporal fenestra is short but high; the squamosal has a long down-
wardly directed process; in the lower jaw there is no coronoid process but a
reflected lamina of primitive form is developed and a fenestra pierces the jaw;
the jaws are edentulous.

With its large procoracoid and the absence of an acromion process Galeops
is more primitive than the other early dicynodonts.

Therocephalia

We, as yet, know no forms that could provide a morphological step inter-
mediate between the pelycosaurs and the four earliest therocephalian families
(Alopecodontidae, Lycosuchidae, Pristerognathidae and Scaloposauridae).

The big morphological gap will be evident if we, in summary, list the
advances shown in these early therocephalians.

The temporal fenestra is immediately very large with outflaring squa-
mosals, and the parietal is always intercalated between the postorbital and
squamosal; the intertemporal skull table is narrow and developing a sagittal
crista; the posterior process of the postorbital is reduced to a small splint applied
to the lateral parietal face; the temporal muscle no longer arising from the
under surface of the roof bones and is inserted on a strong coronoid process;
one or two strong canines developed (except in the scaloposaurids); small
teeth anterior to the large canine in alopecodonts but absent in the other
families and the tooth row generally shortened sometimes radically; dorsal
process of premaxilla always very short; jaw articulation shifted slightly
forwards; but the exoccipital and prootic less prominent in the brain floor
than in Dimetrodon and the sphenoidal complex less developed; a large sub-
orbital fenestra present; the whole locomotor apparatus well developed in
adaptation to a more upright walking gait and the phalangeal formula reduced

to 2, 35 3> 35 3.

40 ANNALS OF THE SOUTH AFRICAN MUSEUM

SUMMARY OF MORPHOLOGICAL SERIES

The series primitive sphenacodont (Haptodus?), eotitanosuchids, brithopids
to anteosaurids, undoubtedly forms a morphological ladder with its bottom
end resting further down among the captorhinomorphs with Anteosaurus on
the highest rung.

The series primitive brithopid (Sydon?), titanosuchids to tapinocephalids,
styracocephalids and estemmonosuchids, is clearly a line closely related to but
diverging from the first series. That this series started from a primitive brithopid
appears very probable but can be queried. The tapinocephalids are a very
mixed lot but undoubtedly closely related and in various ways a rung up the
ladder above the titanosuchids, but the picture is complicated by the develop-
ments seen in the other two related forms—Styracocephalus and Estemmenosuchus.
This series also terminates at the top of the Tapznocephalus zone.

In the series primitive sphenacodont, Phthinosuchus, Hipposaurus to gale-
suchid gorgonopsians, the position of Phthinosuchus as intermediate between
sphenacodonts and hipposaurids is very uncertain but a fairly close but less
specialized form than Phthinosuchus would fit the bill. Moreover, Hipposaurus
does not quite fit in as an antecedent stage to the galesuchids.

If those objections are valid then the phthinosuchids, hipposaurids and
galesuchids form a triradiate branch arising from a sphenacodont group
close to that from which the Dinocephalia is derived.

In the dicynodont series the morphological step from any known pelycosaur
to Otsheria is very great and I find it difficult to visualise how such a step could
have taken place; but the transition from Otsheria to endothiodonts and
dicynodonts is small and obvious.

In the pelycosaur—therocephalian series no intermediate stages are known
and the gap is very wide, but nothing that the discovery of some more primitive
forms would not bridge.

In short, the Dinocephalia and Gorgonopsia can be derived from a
sphenacodont near to Haptodus and the Dicynodontia and Therocephalia from
two other as yet unknown primitive pelycosaurs.

The foregoing morphological analysis can also be piesentean in numerical
form. For the various groups under consideration here I have tabulated the
primitive reptilian characters determinable in each.

Arranging these in numerical order we get the following percentages:

Captorhinomorpha 100

Sphenacodontia 88
Eotitanosuchia 68
Dinocephalia 60
Gorgonopsia 56
Therocephalia 52
Otsheriidae 40

Dicynodontia 32

THE EARLY THERAPSIDS 41

EARLY THERAPSID HISTORY

From the foregoing there is no doubt that on purely morphological grounds
the therapsids must be derived from the captorhinomorphs by way of the
sphenacodont pelycosaurs.

Now, does this fit in with the known history of the early tetrapods ?

We can begin the story with the primitive anthracosaurs which were the
first tetrapods to successfully achieve an amphibious life. These are best known
from the Lower Carboniferous of Scotland where the prevailing climate was
warm and moist and eminently suitable for an existence partly in water and
partly on land. |

At the close of the Lower Carboniferous times the Scottish climate changed
radically. The elevation caused by the Hercynian Foldings made the climate
arid and thus unsuitable for these amphibians.

We now find the amphibian history continuing in central Europe and
North America where during Upper Carboniferous times swampy conditions
in a warmer climate prevailed.

Swampy conditions continued into Lower Permian times in central
Europe and North America, but slowly changed to drier conditions and this
sparked off the explosive development of the earliest cotylosaurs especially in
North America, closely followed by the rise of the pelycoasurs also mainly in
North America but with representatives in central Europe.

At the end of the Lower Permian the climate over North America became
more and more arid and the cotylosaur—pelycosaur explosion came to an
abrupt end.

In parts of Europe, however, the Lower Permian climate remained cool
and favourable for the continued existence of the sphenacodonts and during
the Middle Permian this cool climate continued in Cisuralian Russia where
the first therapsids made their appearance in deltaic conditions and from there
spread to southern Africa where the favourable flood plain conditions existed
in a fairly cool to warm climate.

It would thus appear that for every major step in the phylogeny a change
of scene was necessary.

This seeming capriciousness can, however, be reasonably accounted for.

During Carboniferous— Permian times the western part of the northern
hemisphere formed a single continent—Laurentia—and, notwithstanding the
upheavals caused by the Hercynian Foldings and the presence of Tethys, there
would at this time have been fewer barriers for the transmigration of tetrapods
than at the present time.

SUBSEQUENT HISTORY

Arising in late Ecca (Lower Permian) times the therapsids formed a firmly
established order of reptiles with four distinct suborders at the beginning of
early Beaufort (TZ apinocephalus zone— Middle Permian) times.

42 ANNALS OF THE SOUTH AFRICAN MUSEUM

During the whole of the Tapinocephalus zone (2 200 m of sediments) little
further development took place.

The Dinocephalia were fully developed at the base of the zone—only
Styracocephalus is first encountered above the lowest of the tripartite subdivisions
of the zone. This greatly diversified suborder dominated the vertebrate life of
the Middle Permian, consisting as it did of a family of large carnivores (Anteo-
sauridae) and three families of large herbivores (Titanosuchidae, Tapinoce-
phalidae and Styracocephalidae). Life during this time must have been easy
with a cool moist climate in an area of low relief consisting of large expanses of
fresh-water pools separated by low uplands with periodic floodings. But
abruptly at the end of Tapinocephalus zone times the life span of the Dinocephalia
was cut short, thus ending one of the first four developmental trends of the
early therapsids. |

This sudden extinction was apparently not caused by any radical change
in the environmental conditions. The succeeding Endothiodon zone lies con-
formably on the Tapinocephalus zone without any radical change in lithological
character—the only noteworthy feature being the increase of the number of
purplish bands indicating more periods of somewhat drier conditions. ‘There
was also no sudden development of competitors or antagonists.

The only reason for the sudden extinction of the Dinocephalia I can
advance is that they went to seed in too favourable living conditions, aggravated
by the pathological pachyostosis induced by a pituitary hypertrophy.

The Gorgonopsia, represented by two families from low down in the
Tapinocephalus zone, had by then already developed all the characters typical
of this suborder and during the Middle Permian show no further development.
They constituted a very minor element in the fauna of these times. This sub-
order of rather primitive therapsids is represented in the higher zones of the
Beaufort beds to form a much more important element in the fauna.

The Hipposauridae, forming a very distinctive family, is not represented
in the Endothiodon zone, but in the Cirstecephalus zone we know three further
genera. Thus they become extinct at the top of the Upper Permian.

The suborder is further represented in the Endothiodon zone by 13 genera
and in the Cistephalus zone, by 48 genera and these have been subdivided into
as much as 17 discrete families, indicating that during the Upper Permian this
suborder really went to town and during this period constituted an important
element of carnivorous forms in the fauna.

Their span of life came to an abrupt end at the close of the Permian.

This second developmental trend of the early therapsids thus had a life-
span extending through the whole of the Middle and Upper Permian. During
this period the gorgonopsians manifest but minor variations and retain such a
uniform morphological pattern that the subdivision into separate families can
at most be considered as of taxonomic convenience.

They form an interesting group of fairly long-lived primitive therapsids
suddenly present at the beginning of the Middle Permian with their distinctive

THE EARLY THERAPSIDS 43

cachet fully developed in one bound and wholly sterile.

It is of interest to note that the gorgonopsians, apparently of Cisuralian
origin, have only three genera in the Russian Upper Permian. Inostrancevia is a
giant gorgonopsian, whereas the aberrant Proburnetza is very similar to Burnetia
of the Karoo.

The Dicynodontia, with two families in the Middle Permian, numeri-
cally rich in individuals, formed a significant element in the fauna as
the sole assemblage of small herbivores preyed on by the smaller to medium
sized carnivores of those times. With a fully developed morphological pattern
from the base of the Middle Permian they waxed exceedingly until becoming
extinct in the Middle Trias.

The basic pattern, fully developed at the beginning of the Middle Permian,
remains unchanged during their long span of life. But in the Upper Permian a
mass of small variations occur as witnessed by the fact that over 200 species
have been named. During these times the dicynodonts were extremely abundant,
by far outnumbering all their contemporary therapsids and constitute the
bulk of the herbivores on which the gorgonopsians preyed.

In the Lower ‘Trias the lystrosaurs were very nearly the only herbivores
preyed on by some small cynodonts and the predaceous Chasmatosaurus.

In Cynognathus zone times we get the kannemeyerids forming the end of
this line of development, which during its long span of life continued basically
unchanged and sterile.

The Dicynodontia, with the Cisuralian form Otsheria as starting point,
had but five descendant genera in Russia, but since the Upper Permian have
spread to Scotland, China, Indo-China, North and South America, India and
Antarctica.

Thus, notwithstanding their innate inability to escape from their con-
fining basic structural pattern they were very adaptable herbivorous reptiles,
well able to fit into all the vicissitudes of the invironmental changes encountered
from the moist Middle Permian to deep into the arid Trias in six of our present
continents.

The Therocephalia with four families and 18 genera in the Tapinocephalus
zone form an important element of the contemporary fauna forming the small
to medium sized insectivores and carnivorous predators of their time.

The Pristerognathidae continue into the Upper Permian with three
genera in the Endothiodon zone and one in the Cistecephalus zone.

The Alopecodontidae have no representatives in the Endothiodon zone but
there are two genera in the Cistephalus zone. |

The Scaloposauridae continue into the Trias with two genera in the
Endothiodon zone, 14 in the Cistecephalus zone and one in the Lystrosaurus zone.

In the Upper Permian five new families make their appearance, indicating
a continued virile variability. In Euchambersia we have the first poisonous reptile
with a poison gland and an appropriate fang.

In the specialized Whaitsiidae we find a tendency towards the develop-

44 ANNALS OF THE SOUTH AFRICAN MUSEUM

ment of a secondary palate and the closure of the suborbital fenestra.

Of the Akidnognathidae and Ictidosuchidae there are in the Upper
Permian nine genera of small advanced therocephalians.

In the Cynognathus zone there are nine genera of small to medium-sized
Bauridae more advanced than the Scaloposaurids from which they arose.

In the Red Beds there follow the Tritylodontidae and in the Cave Sand-
stone we have the Diarthrognathidae.

The Cynodontia, appearing for the first time in the Cistecephalus zone, if
not developed independently and directly from the pelycosaurs can only be
derived from the Therocephalia. :

In this versatile assemblage the later branches exhibit various trends
towards the mammalian condition particularly in regard to the braincase, the
dentition, the secondary palate, the reduction of the posterior mandibular
bones and the establishment of an articulation of the dentary directly to the
squamosal.

Of the four suborders of the therapsids, developed from pelycosaur ances-
tors before the beginning of the Middle Permian, the Dinocephalia, Gorgonopsia
and Dicynodontia have been proved phylogenetically sterile.

Only the more plastic and versatile Therocephalia developed upper
branches approaching structural levels very close to that of the first mammals
and it seems reasonable to assume that one or more of these trends did actually
culminate in the first mammals.

This phylogenetic success of the Therocephalia can be attributed to a
number of factors. In the initial stages most of the primitive conservative
structural patterns were bred out and a great amount of plasticity was retained.
by developing at a moderate tempo without any extravagant variations. In
some of the later branches, as exemplified by Euchambersia and the whaitsiids,
such sterile developments were however not avoided, but there was always a
plastic core retained.

From the start excessive size was avoided, except in some pristerognathids
and lycosuchids which soon petered out. Together with the small size there
was the acquisition of an agile locomoter potential and this ability to live an
active mobile life ensured their ultimate success.

REFERENCES

A full bibliography has been given in

Boonstra, L. D. 1969. The fauna of the Tapinocephalus zone (Beaufort beds of the Karoo).
Ann. S. Afr. Mus. 56: 1-73.

Fox, R. C. 1964. The adductor muscles of the jaw in some primitive reptiles Univ. Kans. Publs.
Mus. nat. Hist. 12: 657-680.

THE EARLY THERAPSIDS 45

= EXPLOSION OF THERAPSIDS IN THE KAROO

Cave Sandstone

Red Beds
Molteno Beds

Cynognathus Zone

Lystrosaurus Zone

Cistecephalus Zone

Endothiodon Zone

Tapinocephalus

Zone

Ecca Series

Fic. 1. The therapsid explosion in the Karoo.

This diagram very effectively illustrates the main features of the faunistic history of the
therapsids during Permo-Triassic times.

The therapsids, arising in Cisuralian Russia in the Lower Permian, spreading southwards
entered the Karoo Basin at the beginning of the Middle Permian as a well-established and
diversified order of reptiles. In the cool, moist, equitable climate then prevailing in the Karoo
they quickly established themselves as the dominant land vertebrates.

This initial burst slackened off towards the end of the Middle Permian.

But in the Upper Permian, with its more varied climate, in which periods of warmer drier
conditions alternated with fairly cool and moist periods, in a further and greater expansive
burst the therapsids attained their maximum faunistic development.

With the drastic ecological changes in the Lystrosaurus zone (lowest Triassic) with its heavy
rains and marshes in a warmer climate, the therapsids all but petered out.

Later in the Triassic, in drier and sometimes arid conditions, only those therapsids far advanced
in a mammal direction extended the life span of the therapsids, but now they were faunistically
overshadowed by the sauropsid explosion which then came under way.

46 ANNALS OF THE SOUTH AFRICAN MUSEUM

‘pats EXPLOSIONS OF THERAPSID SUBORDERS IN THE KARCO

Gemiead
Red Beds

Molteno in
ee

Cynognathus
Zone

Cistecephalus
Zone

Cynodontia

Dicynodontia Therocephalia
Ecca .
Series

Fic. 2. The explosion of the therapsid suborders in the Karoo.

The Dinocephalia, arising in Russia in the late Lower Permian, entered the Karoo Basin as
a diversified suborder and immediately became the dominant therapsids. But they quickly shot
their bolt and from the middle of the Middle Permian declined rapidly to become extinct before
the Upper Permian. Thus ended the first sterile therapsid trend towards an actively mobile
life on firm land of some altitude.

The Gorgonopsia and Dicynodontia, entering the Karoo Basin fully fledged, but in a sub-
ordinate faunistic role, became the dominant therapsids in the Upper Permian. The predatory
Gorgonopsians came to an abrupt end as the second sterile trend towards life on drier ground
when the marshes of the Lystrosaurus zone made life for their herbivorous prey impossible on
dry land, coupled with the predators’ inability to pursue the surviving lystrosaurs into the
marshes. The herbivorous Dicynodontia, adapted to upland life, found the swampy conditions
of the Lystrosaurus zone all but impossible and only the lystrosaurs could adapt themselves, but
this temporary success proved their final undoing and in the later dry to arid Triassic only the
kannemeyerids could eke out an existence.

Thus ends the third therapsid attempt towards an active mobile life on firm land.

The Therocephalia entered the Karoo Basin as a diversified suborder with two families in a
strong predaceous role and two in an insectivorous role. The predators flourished in the Middle
Permian but were ousted from this role by the gorgonopsians in the Upper Permian. The more
insectivorous families continued successfully into the later Triassic, but where overshadowed
faunistically (except for their cynodont offshoot) by the upsurging sauropsids. This fourth trend
of the therapsids towards upland active life proved to be genetically fertile in that they gave
rise to the first mammals in the Upper Triassic.

Tapinocepha

St

Se

Dicynodontid

DIC

\q

Otsheriid

/

diagramatic phylogeny of the early therapsids.
bhenacodontid represented by Haptodus
)titanosuchid represented by Eotitanosuchus
ithopid represented by Syodon

teosaurid represented by Anteosaurus

racocephalid represented by Styracocephalus
upinocephalid represented by Tapinocephalus
ithinosuchid represented by Phthinosuchus
pposaurid represented by Hipposaurus
lesuchid composite of the early genera
rcosuchid represented by Trochosaurus
isterognathid represented by Glanosuchus
aloposaurid represented by Blattoidealestes
sheriid represented by Otsheria

cynodontid represented by Dicynodon
idothiodontid composite of early genera
omasaurid represented by Galeops.

i eT es own

DINOCEPHALIA GORGONOPSIA THEROCEPHALIA DIC YNODONTIA

Galesuchid Scaloposaurid Dieynodontid

Fic. 3. A diagramatic phylogeny of the early therapsids.
Sphenacodontid represented by Haptodus
Eotitanosuchid represented by Eofitanosuchus
Brithopid represented by Syodon
Anteosaurid represented by Anteosaurus

Titanosuchid represented by Jonkeria

Styracocephalid represented by Styracocephalus

Tapinocephalid represented by Tapinocephalus
3 Phthinosuchid represented by Phthinosuchus

Hipposaurid Pristerognathid Endothiodontid

Po.

Hipposaurid represented by Hipposaurus
Galesuchid composite of the early genera
Lycosuchid represented by Trochosaurus
Pristerognathid represented by Glanosuchus
Scaloposaurid represented by Blattoidealestes
Otsheriid represented by Otsheria
Dicynodontid represented by Dicynodon
Endothiodontid composite of early genera
Dromasaurid represented by Galeops.

Lycosuchid Dromasaurid

Te chid Anteosaurid

eS

Otsheriid

Brithopid Phthinosuchid

Eotitanosuchid

Sphenacodontid 9) ke

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.
Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm X 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, 5, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

Bu.tLtoucn, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHEerR, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHer, P.-H., DuvaL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 1960. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4.): I-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

‘

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 80.

ey

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 ~ Band
March 1972 Maart
Part), 3 Deel

LARVAL DEVELOPMENT OF THREE SPECIES
OF EGONOMICALLY IMPORTANT
SOUTH AFRICAN FISHES

By
E. H. HAIGH

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

5 Ai gia D> 7-8), ATID, By, t.—p.i.), 5(2, By es)
6(1, t.—p.i.), 7(1-3), 8, (1-2), 10(1),
Il (1-2, 5, 7, t--p.i.), 21, 24(2), 27, 31(1-3), 33

Price of this part/Prys van hierdie deel

R2,20

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 03 8

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

LARVAL DEVELOPMENT OF THREE SPECIES OF
ECONOMICALLY IMPORTANT SOUTH AFRICAN FISHES

By
E. H. Haigh

South African Museum, Cape Town

(With 11 figures and g tables)

[MS. accepted 1 September 1971]

CONTENTS

PAGE

Introduction . : ; : ; ‘ 47
Material and methods : ; , : 49
Description : 5 ; Se nae 4 50
Merluccius capensis . , : 50
Thyrsites atun. ~. : ; , , 55
Helicolenus dactylopterus . ; : 60
Distribution. : : ; ; : 69
Summary ‘ ‘ : : 5 . 69
Acknowledgements. ; : : 5 69
References : : : : i = 69

INTRODUCTION

This paper is the first of a series of studies on larval fish development.
Mainly economically important species will be described although other
species of interest will also be included.

The species described in this paper are Merluccius capensis, the Cape
stockfish, Thyrsites atun, the snoek, and Helicolenus dactylopterus, the jacopever.

The Cape stockfish or hake is of great economic importance in South
Africa. The annual trawled catch has ranged from 68 o19 223 to 70 686 775
kilo over the five years from 1961 to 1965 (Ann. Rep. Div. Sea Fish. S. Afr. 33).
At the end of 1970 the trawled landings were reported to be 64 million kilo.
Irvin & Johnson (1963) give an account of the economic importance and
habitat of the adult. Although the adult is almost exclusively demersal, the
larvae are caught in plankton nets, indicating a pelagic mode of life. However,
the number of larvae, especially the older, larger forms, is relatively small in
the samples.

The snoek is an important predator of the pilchard Sardinops ocellata and
related pelagic fish. It is a large rapacious carnivore, with excellent, firm flesh
and thus also an important seasonal food fish in South African waters. Annual
catches vary between 7,2—9 million kilo per annum.

The jacopever is uncommon in fishmarkets around the coast, the larger
number being caught in trawls and used in fishmeal manufacture. Between
1961 and 1965 the landings ranged between 857 427 and 1 487 624 kilo (Ann.

47
Ann. S. Afr. Mus. 59(3), 1972: 47-70, 11 figs, 9 tables.

48 ANNALS OF THE SOUTH AFRICAN MUSEUM

Rep. Div. Sea Fish. S. Afr. 33). Smith (1953) reports its flesh to be palatable.
However it does not constitute a major portion of the fishing resources of the
country. :

Davies (1949) states that Helicolenus inhabits waters between go and
360 metres. It is a bottom-dwelling fish usually on the continental shelf.
Superficially it is very like the Tristan da Cunha scorpaenid Sebastichthys
capensis and several authors have reported both genera to be viviparous.
Specimens between 3,5 and 4,2 mm standard length had well-developéd jaws
(indicating functionality) and several head spines. Davies (1949) suggests a
November spawning season, Ahlstrom (1961) on the other hand suggests a
winter and early spring release of young of Sebastodes spp. on the west coast
of the United States. Moser (1967) states that there are two broods released in
Sebastodes paucispinis, one in autumn and one in spring. The majority of samples
in our collection were caught in late spring, October. These represent a com-
plete range in sizes. As the collection is not very large, no conclusions can be
drawn from this.

The three species described in this paper are classified as follows:

Order: Gadiformes Perciformes Scorpaeniformes
Suborder: Gadoidei Scombroidei Scorpaenoidei
Family: Merlucciidae Gempylidae Scorpaenidae
Genus: Merluccius Thyrsites Helicolenus
Species: capensis atun dactylopterus

Castelnau 1861 Euphrasen 1791 Delaroche 1809

The family Merlucciidae is distinguished by a separate caudal fin.
According to Norman (1937, 1966) it has one genus comprising seven species,
three in the northern and four in the southern temperate zones. Although
Gilchrist (1921) and Barnard (1925) doubt the distinction made between the
European species Merluccius merluccius and M. capensis, Norman (1937) confirms
Regan’s (1908) distinction. Ginsburg (1954) sheds more light on the American
species of the Merlucciidae. |

The Gempylidae is a small family comprising 10 genera each with only a
small number of species. Its taxonomic history seems to have been untroubled.

The taxonomy of the family Scorpaenidae needs world-wide revision.
Eschmeyer (1969) gives a good review of the Atlantic Scorpaenidae, synony-
mizing Helicolenus maculatus with H. dactylopterus and separating the Tristan da
Cunha species from H. dactylopterus. However, he does state that the gradient
in characters is rather disjointed and that a conclusive synonymy needs a more
comprehensive study of material.

Eggs and larvae of Merluccius merluccius have been described by Ehrenbaum
(1909) and D’Ancona (1933); of M. bilinearis by Kunz & Radcliffe (1917); of
M. productus by Ahlstrom & Counts (1955). Marak (1967) describes the early
pro-larvae of M. albidus and distinguishes them from the pro-larvae of M.
bilinearis. Fischer (1959) describes eggs and pro-larvae of M. gayi from Chile

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 49

as do Santander & Castillo (1969) from the coast of Peru. Hart’ & Marshall
(1951) report a larval Merluccius capensis between 19° and 22°S, extending the
possible range further north than this present survey. Matthews & De Jager
(1951) described the development of the egg and pro-larva of 2,35 mm for
M. capensis. Larvae of the genus Merluccius ana a basic similarity in pigmenta-
tion, allowing for easy recognition.

Larvae of species of the family Gempylidae have been described: Gempylus
serpens by Jones (1960); Neszarchus nasatus and Gempylus spp. by Voss (1954)
and Thyrsites atun by Regan (1914-16). De Jager (1955) described the develop-
ment of artificially fertilized eggs and resultant larvae of Thyrsites atun up to
the age of nine days and a length of 3,9 mm. In the present paper the develop-
ment from 4,6 mm to 25 mm is described, thus completing the description of
Thyrsites atun development.

In the family Scorpaenidae larval development of two species of Sebastodes
has been described by Ahlstrom (1963) and Sebastes marinus has been described
by Bigelow & Welch (1924). A paper on the distribution of Sebastodes spp. in
Californian waters was published by Ahlstrom in 1961. In this paper he dis-
cussed briefly the distinguishing features of some scorpaenid larvae in the eastern
North Pacific. However, Moser (1967) gives the complete development of
Sebastodes paucispinis and gives illustrations and a list of characters of early
stages for 14 further species.

MATERIAL AND METHODS

Specimens were obtained by research vessels of the Division of Sea
Fisheries, Cape ‘Town, using N1ooB and Ni1ooH plankton nets, from 1950 to
1967. Samples were fixed and stored in formalin which was replaced by 70%
ethyl-alcohol. Specimens were stained, using the methods of Hollister (1934),
Davies & Gore (1935) and Moran (1956) but modified slightly by reducing
the clearing time in KOH and reducing the concentration of the KOH used.
This was done in order to preserve the pigment of the specimens. As pigments
are inclined to fade, more than one larva in the size range was used in order to
obtain the most characteristic pigmentation pattern. Stained specimens were
preserved in glycerin and ethyl-alcohol.

Measurements were taken as follows:

standard length (s.l.): tip of lower jaw to caudal peduncle

snout: tip of lower jaw to anterior edge of eye

eye diameter

head length: tip of snout to posterior edge of cleithrum

trunk: tip of snout to posterior edge of anus and not
to anal fin insertion

depth: at pelvic fin insertion

pelvic and first dorsal spine lengths for Thyrsites atun only

50 ANNALS OF THE SOUTH AFRICAN MUSEUM

All proportions presented as percentage of standard length. All lengths
cited in text are the standard lengths of the specimen.

Some specimens that should have been well ossified did not absorb stain
properly. This was most probably due to decalcification of the bone by formalin.
Vertebral counts include urostylar complex which is counted as two.

DESCRIPTION

Merluccius capensis

Merluccius capensis is characterized by 130—140 scales in longitudinal series,
13-14 gillrakers in lower part of anterior arch and a pectoral with 13-14 rays
reaching to beyond the origin of the anal, while the pelvic extends nearly to
the vent. Depth is 60% of length and headlength is 32-36% of length. The
maxillary extends to below the posterior edge of the pupil or beyond and is
less than half of head-length. D:10-11; 35-40. A:37-40 (Norman 1937).

Pigmentation

The general pigmentation pattern of M. capensis is similar to that of
M. merluccius as described by Ehrenbaum (1909) and D’Ancona (1933). The
major pigmentation on the head consists of one or two large stellate mela-
nophores situated at the postero-dorsal edge of the brain and anterior to the
first dorsal fin—occipital spot. The dorsal peritoneal wall is always darkly
pigmented with both stellate and closed chromatophores.

The tails of most larvae examined bear three areas of pigmentation. An
anteropostanal spot situated latero-ventrally just behind the anus; a larger
mediopostanal spot, covering the whole lateral surface of the tail, midway
between the anus and the caudal fin; and one or two stellate melanophores
comprising the caudal spot situated latero-ventrally on the caudal peduncle.

These pigmented areas are characteristic of the species and are to a
greater or lesser extent augmented by smaller stellate and contracted melano-
phores at the dorsal aspect of the head and ventral aspect of the abdomen. These
vary a great deal in intensity within a size group but generally increase in size
with age during larval life.

In the early post-yolk-sac stage larvae (Fig. 1a) the head bears a pigment
spot on the anterior edge of the brain which becomes obscured as the larvae
get larger and ossification commences. The cerebral area of the head is further
dotted with a varying number of melanophores. With further development,
melanophores may also appear on the jaw, around the eyes and on flog opercular
surface of the head (Figs 1d, 2a).

Peritoneal pigmentation remains fairly constant throughout development.
The melanophore anterior to the first dorsal fin becomes augmented by smaller
stellate melanophores along the sides of the first dorsal and later the second
dorsal fin. |

The anteropostanal, mediopostanal and caudal spots remain relatively
the same size throughout the larval life but in prejuvenile and juvenile stages,

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES

Fic. 1. Merluccius capensis
Early larval stages showing position of major pigmentation areas and ossification development.
Measurements indicate standard length.

they appear to become smaller and more discrete. They probably break up to

form the smaller melanophores that abound on the dorsolateral sides of the
juvenile (Fig. 2c).

Small, discrete pigment spots are present on the pectoral and pelvic fin-

51

52 ANNALS OF THE SOUTH AFRICAN MUSEUM

buds of some of the larvae. As soon as ossification of the pelvic fins is complete,
pigmentation appears on both fins. No caudal fin pigmentation was observed.

Pigment spots appeared on the dorsal fin late in the larval and early
juvenile stages. Pigmentation was also present on the dorsal head and abdominal
surfaces in juveniles.

Ossification

The premaxilla, maxilla, mandible and cleithrum are ossified in larvae
3,6 mm long. The first traces of the gill-arches can be seen in slightly larger
specimens. The branchiostegal rays then ossify progressively from dorsal to
ventral (Fig. 1b). The premaxilla at 5,5 mm has between four to six teeth and
the dentary four to six. The supracleithrum and posttemporal appear as slender
rods. By 6,1 mm the epihyal and ceratohyal are well formed and bear six
branchiostegal rays. The first traces of the preopercle, cranial and opercular
ridges are ossified and there are eight mandibular teeth. By 7 mm the hyomandi-

TABLE I

Ossification of skeletal elements of Merluccius capensis

Size I Te Neural Haemal Pectoral Pelvic
inmm Dorsal Dorsal Anal Caudal Vertebrae spines spines rays rays
4,5 a a a iene ar 3 ren aa ai
oe hee Te ae i 6 <r “ aa a 2
6,0 — — — 8 a 6 — — 3
6,6 2-3 — — 12-14 8-13. 8-10+2-5 3-5 — 5
6,9 3 = = 15 124-5) 8-5 4 =e 6
7,8 5-7 6-8-2455 1517-20-21) 41-42) 50-4 eae aa 7

Seed
9,0 6-7 26 30 19-21 47-49 54 24 i 9
10,2 8 34 33 28 57 54 27 = 9
12,0 9 35 35 30 57 54 29 He 9
14,1! 11 32 39 40 57 54 29 | 9

bular can be clearly distinguished and the quadrate has started ossifying in the
articular region. The pterygoid has also started forming and the seventh
branchiostegal is complete. Other bones have thickened and widened con-
siderably. By 9,6 mm parietal and frontal bones are well formed, the nasals
have started ossifying and so have the lacrymals. The first trace of the operculum
is present, underlying the opercular ridges. The visible bones in the head have
now been fully laid down and further development takes place to reach the
juvenile stage between 14 and 16 mm (Fig. 2d).

The anterior neural spines ossify first, followed by the anterior centra. At
6,8 mm the first seven centra and five to eight haemal spines are formed.
However, neural and haemal spines of some of the caudal vertebrae also show
signs of ossification at this stage. During this growth-period the centra are
rapidly laid down and at 9,6 mm all the vertebrae except the five preceding
the last centrum have been fully formed. The last vertebra is ossified but the
urostyle is still cartilaginous. At this stage the hypural and epural elements
show some degree of ossification. Haemal spines are only found from the 26th

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 53

27 26 5 24 23 d al wd
iM eeee
: = Nau KE RRSES
= 7 8 iS

Fic. 2. Merluccius capensis

a—b. Late larval stages. c. Juvenile. d. Lateral view: 1. Lacrymal. 2. Maxilila. 3. Premaxilla.

4. Dentary. 5. Circumorbitals. 6. Pterygoid. 7. Quadrate. 8. Articular. 9. Preopercle. 10. Sub-

opercle. 11. Branchiostegal rays. 12. Opercle. 13. Cleithrum. 14. Postcleithrum. 15. Pelvic

girdle. 16. Pelvic rays. 17. Pectoral rays. 18. Anal rays. 19. Caudal rays. 20. Second dorsal fin.

21. First dorsal fin. 22. Supracleithrum. 23. Posttemporal. 24. Cranial bones. 25. Hyomandibular.
26. Frontal. 27. Nasal.

centrum at this stage, indicating the first 24 to 25 vertebrae to be abdominal.
By 13,2 mm all the parapophyses of the abdominal vetebrae, except the first
six, have enlarged somewhat. In the juvenile there are 18 abdominal vertebrae
bearing parapophyses expanded lateroventrally (Fig. 2d).

54 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fin formation takes place in the following sequence: pectoral fin bud,
caudal buds and pelvic bud between 3,0 and 5,0 mm. In size range 5,0 to
6,0 mm the pelvic rays ossify, then the caudals start, followed by the first
dorsals at between 5,5 and 6,8 mm. The anal and second dorsal rays follow
almost immediately. Pectoral rays only appear much later. By 6 mm the first
five pelvic rays have appeared in the pelvic bud. The caudal buds appear at
about 4,5 mm. The ventral caudal rays appear first, followed at 6 mm by the
dorsal caudal rays. Several caudal rays seem to ossify simultaneously. In some
larvae of 6 mm the basal buds of the first four dorsal rays are also evident.

Although the anal fin starts ossifying only after the first dorsal, its develop-
ment proceeds faster and by 6,8 mm between six and eight rays are ossified in
the anal while only five dorsal rays can be distinguished. The pelvic now has
six ossified rays and the caudal has 14.

The pelvic attains its full complement of seven rays at approximately
g mm. The caudal, dorsal and anal fins have an almost complete number of
rays at 13,2 mm but as so few larvae over this size were obtained there is no
certainty about the exact size when full ray number is attained. Ahlstrom &
Counts (1955) give 16 mm as the size where the full complement of both dorsal
and anal fin rays is present in M. productus.

Seven pectoral fin rays were observed at 14,1 mm. The juvenile of 37 mm
has a full‘number of 14. M. productus larvae only develop pectoral fin rays at
24 mm and later (Ahlstrom & Counts, 1955).

Changes in body form and growth rate

The eye diameter is larger than the snout length in the earlier stages but
this difference is gradually diminished and by 6,0 mm they are approximately
equal in length. The snout becomes longer than the eye diameter from 10,5 mms.1.

TABLE 2
Mean measurements in mm and proportions (% of s.l.) of Merluccius capensis
Size group No. Snout 1. Eye diameter Head 1. Trunk l. Depth
% % % % %
25-349 — 2 0,13 4,57 0,31 11,12 0,64 22:44 9 in7r 6393" org mraates
355-4,49° ~ 9) 0,25. 6,57 0,32 8,10 1,07 25,708" 1,73 "43595" At commaa an
4;5—5,49 21 0,48 9,91 0,53 11,07 1,43 28,39 2,86 46,69 1,40 29,55
5,5-0,49 11 0,62 10,32 0,62 10,32 1,89 31,08 2,80 48,65 1,69 30,60
6,5-7,49 18 0,65 8,65 0,65 °§8,65 1,96 97,10 93,23 46:54 1,66) e2egg
7,5-3,49 13 0,71 9,67 0,73 9,36 2,30 30,24 53,66 46,61 / Blog mgenen
8,5-9,49 I) 50,90) | *10,00'; -0,90.) 416,00 2,40 26,66 4,20 46,66 2,10. 929)e3
9:5-10,49 8 0,97 8,86 0,90 9,22 2,92 30,01 4,60 47,55 2,40 25,02
sie a 1,49 6 = 1,05 9,73 0,90 8,33 3,20 30,09 4,90 45,35 2,40 22,71

13,5-14,49 3 1,30 9,33 1,10 7,88 3,70 26,57 6,20 44,63 3,00 21,60
‘ 20,40 1. 1,80 ~8,82 1,50 7,35 “6,60 932,95 9,60. 47,05 4,20 meou6
26,60 1 2,10 7,89 1,80 6,76 7,20. 27.06>° 11,40. 42,85). *5,1ossamanrs
36,30 1 3,80 ‘10,46 2,80 7,71 11,30 © 31,12) ‘17,00 46,83" (7,50 neoeen
41,00 1 4,50 10,97 3,00 7,31 13,00 31,70 19,00 46,34 8,00 19,51

46,00 1 5,00 10,86 3,00 6,52 13,50 29,34 21,20 46,08 8,00 17,39
+ Two size groups not obtained in the samples.
* Specimens no longer fall into size groups.

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 55

The head is 24%, of the standard length in the smallest stages (2,9 to
4,4 mm), increasing to 28% at 4,5 to 5,4 mm and remaining 25 to 32% of
standard length for the rest of development from 5,5 to 46 mm. The head
shows an average rate of increase of 0,36 mm for each millimetre increase in
standard length.

The proportion of the trunk to the standard length varies between 42 and
48% but is usually about 45°, throughout development and can be used as
a taxonomic character (Ahlstrom & Counts, 1955). The trunk shows an average
increase of 0,47 mm/mm increase of standard length. Up to a length of 9,75 mm
the depth increases 0,23 mm/mm increase in standard length, but between
9,75 mm and 13,2 mm the increase is only 0,127 mm/mm increase in standard
length. Thus the hake changes from a rather deep tadpole-like post-larva to a
slender, evenly sloping juvenile (Fig. 2c).

Thyrsites atun

The snoek is characterized by the maxilla reaching slightly beyond the
anterior border of the eye, the long mandible projecting the upper jaw and
reaching to the posterior third of the eye. Both j JAN carry large canines. The
caudal is deeply forked.

D: 18-21 + 10-12 + 6. A: 1-2 + 8-11 + 6. Pectoral: 2 + 11. Pelvic: 1 + 5.
Vertebrae: 34-35. Depth +7 (Smith 1953; Fowler 1936; Beaufort &
Chapman 1951: 199).

Pigmentation

Standard pigmentation in larvae of Thyrsites atun between 4,0 and 10,0
mm consists of a variable number of small stellate and closed melanophores
over the snout and cerebral areas of the head. A dark area of pigmentation,
standard in all larvae, is found on the antero-dorsal and lateral areas of the
peritoneum. Small scattered spots are also to be found posteriorly, above the
anus. The pigmentation is darker in smaller specimens of four to six milli-
metres, becoming more diffuse and evenly distributed as the fish grows larger.
When the snoek larvae are 11 to 15 mm long, abdominal pigmentation consists
of scattered stellate melanophores.

Most characteristic and stable are the two areas of tail pigmentation. On
the ventral surface of the tail, midway between anus and anal fin, is a smaller
spot consisting of only one melanophore. Also ventrally situated is the posterior
pigment area between the anal and caudal fin. This spot consists of several
stellate melanophores clustered together, and covers two to three times the
area of the anterior pigment spot. However, these areas of tail pigmentation
do not appear in the specimens figured by De Jager (1955). I have had occasion
to examine the specimens of De Jager and his figures agree reasonably well
with the specimens. It would appear that rapid migration of pigment takes
place in early larval life. The pigment pattern described above is only evident
in the last stage of De Jager’s larvae at 3,9 mm but unfortunately is not obvious

56 ANNALS OF THE SOUTH AFRICAN MUSEUM

in his Figure 18. The two tail pigment areas only disappear in juveniles over
16 mm and constitute a good diagnostic feature for the larvae of the species.
The characteristic black dorsal pigmentation of the adult snoek begins to
show up in larvae of 6,0 mm. On each side of the dorsal spines a thin line of
black pigment appears, which becomes thicker and more obvious as the larvae
grow. This line is characteristic of the Gempylidae. None of the fins shows pig-
mentation during development (Fig. 3). |

Ossification

At 4,6 mm the premaxilla with four teeth, the maxilla, and the mandible
with two teeth are ossified as well as the cleithrum. Ossification of the pre-
opercle has begun and two spines may be distinguished on the margin as well
as one small spine originating in the middle of the preopercle in line with the

a 46 mm

Fic. 3. Thyrsites atun

Early larval stages showing position of major pigmentation areas and ossification development.
Measurements indicate standard length.

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 57

most dorsal spine. The first gill arches and five branchiostegals are formed.

- Between 5 and 6 mm the premaxilla develops four more teeth and the
anterior canines also start developing. The mandibular teeth increase to six.
The palatine is clearly visible. Preopercular spines increase to four and the
branchiostegals to six. Traces of the pterygoid, quadrate, symplectic, and opercle
are ossified as well as the posttemporal and postcleithrum. The ceratohyal can
also be faintly discerned. By 8 mm traces of the hyomandibular are laid down
and the preopercular spines increase to five. By 10 mm palatine teeth develop.
The cranial bones have by now become clearly ossified and traces of the sub-
opercle are present. By 11,5 mm traces of the nasals can be seen, the fangs are
well developed (two on each side) and three palatine teeth are present. The
subopercle is quite clear and the hyomandibular is well formed. By 20,0 mm
the jaws are heavily ossified, the preopercle has assumed a more median position
and the six preopercular spines are not as obvious as before. The quadrate is
well ossified, obscuring part of the symplectic. Frontal and parietal bones are
well formed but there are still wide gaps between them. The bones of the pectoral
girdle have become much wider and flanges have developed on the post-
temporal and cleithrum.

There are no traces of ossification in the vertebral column at 4,6 mm, but
between 5,4 and 6,0 mm the basic elements of the neural spines appear in
some specimens. At 6,6 mm at least three, usually more, neural spines are
ossified. Between 6,6 and 7,2 mm the first centra of the vertebrae become ossified
and by 9,75 mm at least 18 vertebrae are ossified. By 8,15 mm the first haemal
spines appear. ‘The ossification of the vertebral spines exceeds that of the centra.
The full complement of the haemal spines is ossified at 11,4 mm. Of these,
11 are shorter abdominal and 14 to 15 are larger caudal spines. Only at 13 to
14 mm is the full complement of neural spines laid down and by 14 mm the
vertebral centra are also fully ossified.

The vertebral centra ossify from the dorsal and ventral peripheries inward,
except for the last three vertebrae where ossification proceeds dorsally from the
ventral periphery. In these three vertebrae the neural spines also form later
than in the other vertebrae. The snoek has 21 abdominal and 13 to 14 caudal
vertebrae.

The urostyle begins to turn up at 7,2 mm and the first traces of urostylar
and hypural ossification are evident at 9,0 mm. At 11,0 mm the urostyle is
fully ossified and bears two dorsal and three ventral hypural elements. The
haemal spines of the ultimate and penultimate vertebrae have broadened to
support the caudal fin.

The pectoral lobe is evident at 4,5 mm but the first fin ossification appears
in the pelvic spine at 5,4 mm. By 6,0 mm the initial ossification of the pelvic
girdle has started and by 8,15 mm the pelvic spine has achieved its characteristic
serrate appearance and proportional full length. The pelvic rays develop
gradually until the full complement of five is attained at 14,0 mm.

The first dorsal fin starts ossifying anteriorly between 5,5 and 6,0 mm

58 ANNALS OF THE SOUTH AFRICAN MUSEUM

and this proceeds rapidly posteriorly. At 7,2 mm there are 10 to 13 rays present,
the most anterior having become hard and serrate. The second dorsal fin has
traces of 10 to 15 rays ossified at 8,15 mm and at 9,5 mm the distinction between
the first and second dorsal is clear. ‘There are 19 to 20 spines and 12 to 16 soft
rays. However it is not until later in the juvenile stage that the finlets differen-
tiate from the second dorsal. The only distinction that can be made, even at
20 mm, is that the last five rays are more widely separated than the rest.

TABLE 3

Table of skeletal elements of Thyrsites atun larvae at different stages

Size No. of Dorsal Haemal Neural
in mm specimens Pectoral Pelvic Caudal Gil Anal spine spine Vertebrae

551 2 a) fs) O Oo Oo fo) O O Oo

594 6 Oo O O O O O O O O

6,0 5 Oo I Oo 4-8 Oo a) ) O Oo

6,6 6 0-4 I O-4 9-13 O Oo O 0-34 Oo

7,2 6 4-6 I 5-8 10-13 Oo 3-6 Oo 7 6

80 6 ia I 10 15 +10 7-10 5-10 sig) 15

955 6 12 I 17 19 +14 II-12 16-19 20-27 18-20
11,4 5 13 1+2 17+5 19 +17 13 23-25 28-32 25-28+2
14,0 I 13 1+5 17+12 19-20+16-17 1-2+14 23-25 34 34-36
20,0 I 14 1+5 17+10 21 +16 2 +15 26 35 36-37

The next fin-rays to ossify are those of the caudal and pelvic fins which do
so almost simultaneously between 6,0 mm and 6,6 mm. At 8,0 mm there are
seven to nine pectoral rays and at 9,5 mm 11 to 13, while the pectoral girdle
has started to ossify. At 11,5 mm the coracoid, scapula and radials have formed.
In the caudal fin the ventral fin-rays are completed first, followed shortly after
by the dorsal rays at 9,5 mm. Between 10 mm and 15 mm the secondary caudal
rays appear. By 14 mm the caudal fin is fully ossified except for some small
secondary fin-rays. The caudal has eight ventral and nine dorsal primary rays.

The last fin to start ossifying is the anal fin. By 7,2 mm only traces of the
first three to six rays are laid down and by 8,15 mm at least the first seven rays
are ossified. ‘The full complement of 16 rays is only laid down by 14 mm. In the
specimens examined, the large majority had only one anal spine; a few had
anal fins with two spines. As with the second dorsal fin, the anal finlets are not
differentiated until late in the juvenile development.

The dorsal spines increase rapidly in actual and relative length between
6 and 7,5 mm and remain at 13 to 14% of standard length till 14 mm length
is reached. After this the relative length decreases somewhat to about 8% of
standard length by 21 mm.

Changes in body form

The very young snoek larva has a large head and short abdomen with a
fairly long tail (Fig. 4a). The head is 30,6% of standard length at 5,4 mm and
increases to 35,2°% by 8,5 mm, more or less retaining that proportion till the
juvenile stage is reached where it diminishes proportionately as standard length
increases. The head increases by 0,42 mm/mm increase in standard length

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 59

a 9,12mm

Fic. 4. Thyrsites atun
Larvae showing increase in size of area occupied by viscera.

between 4,08 and 14,25 mm. After this size is reached, increase in head length
appears to decrease and reach 0,225 mm/mm increase in length.

The snout is 37° of the head length. Initially the snout is only 0,og mm
longer than the eye diameter, but this difference increases with age until at
21,0 mm the snout length is 1,05 mm greater than the eye diameter (Table 4).

TABLE 4

Mean proportions of Thyrsites atun larvae presented as % of s.l.
Size range of

standard length No. of larvae Head Depth Trunk 1st Dorsal
in mm
4,08-5,49 II 30,6 21,4 42,0 eae
5,5-6,49 7 32,6 21,2 45,1 7.4
6,5-7,49 22 32,6 20,6 47,9 11,0
7,5-8,49 13 35,2 21,0 55,2 13,2
8,5-9,49 11 37,8 20,8 64,8 14,7
9,5—-10,49 | 36,8 20,0 66,03 13,7
10,5-11,49 3; 3755 19,7 71,2 14,7
12,6 I 3597 20,2 69,0 13,0
13,5 I 3555 18,8 64,4 13,3
14,0 2 36,2 21,2 74,8 13,8
20,17 2 33,0 16,2 76,2 Tear
21,0 I 31,4 16,1 7393 8,5

The ante-anal length or trunk length is initially 42,0% of the standard
length. This, however, does not remain constant but increases as shown in
Table 4. As far as can be judged from material available, the area occupied

60 ANNALS OF THE SOUTH AFRICAN MUSEUM

by the intestine and viscera increases rapidly as the larva reaches 7 mm (Fig. 5
and Table 4). This rapid increase of an average of 0,98 mm/o,g5 mm
increase in standard length continues till 11 mm s.]. is reached when
the rate. of increase of the trunk slows down somewhat. At this stage
(11 mm s.].) the anus has reached the origin of the anal fin, i.e. its
juvenile position. It is possible that the rapid rate of visceral increase is
linked to the change in the diet of the larvae. The larvae cease to feed on
phytoplankton and become predatory on other fish. Head and eyes of larval
fish have been found in the gut of snoek larvae as small as 8 mm.

Proportional depth of the snoek larvae remains fairly constant, between
18 and 21% ofstandard length, until 14,0 mm is reached, then it drops gradually
to 16% of standard length at 21,0 mm. Actual depth increase is of the order
of 0,17 mm for each mm of standard length increase up to 13 mm s.l. After
this, accuracy of calculation breaks down due to the small number of available
specimens, but the rate appears to be less and in the order of 0,056 mm/mm
increase in s.l. Generally the larvae seem to become slimmer as they reach
the juvenile stage.

b 204 mm (hh f

Fic. 5. Thyrsites atun
Late larval and juvenile stages. Measurements indicate standard length.

Helicolenus dactylopterus

The pectoral of Helicolenus dactylopterus is emarginate on the dorsal edge,
with two unbranched rays followed by eight branched and nine unbranched

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 61

rays, the latter being free from the membrane for at least one-third of their
length. Suborbital keel smooth with one spine small or absent, the mouth large
with villiform teeth on the jaws, vomers and palatine. ‘The maxilla reaches to
below the hind margin of the eye. The spination of the adult head is as follows:
1 nasal, 2 supraorbital, 2 parietal, 1 pterotic, 2 posttemporal, 2 opercular, 1
small suborbital, 5 preopercular, second the-longest (Fig. 6c). Soft dorsal higher
than spinous dorsal. Pelvic reaches almost to vent. D: 12 + 12-13. A: 3 + 5.
Pelvic: 1 + 5. Vertebrae: 24-25.

Smaller specimens have black pigments near the end of the spinous dorsal.
Pigment on body of juveniles in vertical bands (Eschmeyer 1969: 92-99).

Pigmentation

All specimens from 3,5 mm have a clearly pigmented peritoneum with
scattered melanophores on the posterior aspect of the head. This peritoneal
colouring is still visible at 20 mm on the dorsal surface of the peritoneum.
The pectoral fins in this species are unpigmented whereas some other scorpaenid
larvae in the collection have variously pigmented pectoral fins.

Ossification

Even the smallest larvae obtained had well ossified head spines and jaws
as well as a cleithrum. By 4,5 mm (Fig. 7a) the premaxilla and maxilla are
clearly defined, as is the lower jaw. The parietal, pterotic and posttemporal
head spines are developed. The opercle is small and lightly ossified while the
preopercle has three well-developed primary spines on the outer edge and two
secondary spines on the median ridge. The middle of the primary spines is the
longest. Four branchiostegal rays are present.

By 6,0 mm the frontal, parietal and pterotic bones are ossified but still
easily distinguished. The hyomandibular has developed and the cleithrum is
wider and better ossified than the 4,5 mm stage. A small subopercle is present.

TABLE 5
Mean measurements of Thyrsites atun larvae in mm

Length Longest
Size rangeof No.of | Standard Head Snout Eye Depth Snout Pelvic dorsal

standard length larvae length s.l. diameter to anus _ spine spine
4,08-5,49 I! 4,08-5,4 1,55 0,588 0,501 1,08 Pe SMR ate (0,3) aR (0,3)
555-6,49 7 5,7—-6,0 1,80 0,681 0,608 1,26 2,69 0,35 0,45

6,5-7,49 22 6,6—7,35 2,30 0,860 0,680 1,43 3534 0,77 0,76
758,49 13 755-8,42 2,88 0,93 O87 170 4552 1,24 1,09
8,5-9,49 11 855-045 3:42 1,32 0,96 1,88 5,84 1,73 1,32
9,5—10,49 7 9,6-10,24 3,60 1,52 1,12 1,97 6,48 2,01 1,34
10,5-11,49 3 10,5—-10,8 4,00 1,56 1,04 2,11 7,60 2,26 1,58

11,5-13,5 2 12,6 4,50 1,80 IAW Behe) O70) |: 2.10 1,65
13,5 4,80 1,80 105.0 92555 040 2.40 1,80

13,5-15,5 2 13,8 4,8 2,10 1,26 3,00 10,20 2,40 1,80
14,25 5.4 1,95 1,50 . 92595. 10,80, 93,00 2,10

15,5 & over 3 20,10 6,6 2,46 1,08 = 3,40. 15,9 2,4 1,8
20,25 6,75 2,70 165 3,15 15,85 3,45 2,7

21,0 6,80 2,70 1365)" 3540." 15,40. 2380 1,80

62 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fic. 6. Helicolenus dactylopterus

a. Larva at 10 mm standard length showing external features only. b. Dorsal view of head

showing position of spines. c. Lateral view of head showing position of cranial bones and spines:

1. Dentary. 2. Angular. 3. Articular. 4. Quadrate. 5. Branchiostegal rays. 6. Preopercle and

five spines. 7. Subopercle. 8. Opercle. 9. Postcleithrum. 10. Posttemporal and spines. 11. Hyo-

mandibular. 12. Pterotic and spine. 13. Parietal and spines. 14. Frontal and supraorbital spine.

15. Premaxilla. 16. Maxilla. 17. Nasal and spine. 18. Lacrymal and suborbital spine. 19. Cir-
cumorbitals. 20. Prefrontal. 21. Pterygoid. 22. Cleithrum.

Gillrakers and gill-arches have started ossifying. A ceratohyal, quadrate and
traces of the pterygoid are visible.

By 6,6 mm the lacrymal and endopterygoid are formed and the lacrymal
bears a spine. Frontal and parietal are fused. Supratemporal and spine are
evident. Six branchiostegal rays are ossified. By 7,35 mm the lacrymal is larger
and bears two spines while the first and second suborbital are well ossified,
obscuring the pterygoid to some extent. Thirteen gillrakers are present and
the sutures between parietal, pterotic, posttemporal and hyomandibular have
become indistinct. In the size range 6,6 to 7,35 mm the head spines present are:
supraocular, parietal, pterotic, supratemporal, preorbital, suborbital, three
primary and two secondary preopercular. Between 7,35 and 9,0 mm ossification
of the head proceeds to near juvenile condition. New elements added include
nasal spines, two extra primary preopercular spines and traces of a preocular

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 63

C 6,6mm

Fic. 7. Helicolenus dactylopterus

Early larval stages showing position and relative length of head spines. Measurements indicate
standard length.

spine. By 10,2 mm the parietal spine has become bifid as in juveniles. T'wo
opercular spines become evident at the posterior edge of the opercle as in
adults. From late larval stage at 10,0 mm and juvenile stage between 15 and
20 mm the relative size of the head spines decreases and they become far less
conspicuous. The secondary spines on the preopercular disappear and the

64 ANNALS OF THE SOUTH AFRICAN MUSEUM

oe sands , V//,
‘ Rs ne LAA

\
\\

=

tan Mined
ae a

b 10,20 mm

Fic. 8. Helicolenus dactylopterus

a. Larva. b. Early juvenile showing completed major ossification. Measurements indicate
standard length.

second primary spine lengthens considerably to become as long as the third
and eventually the longest spine, as in the adult condition. A small spine
develops behind the supraorbital, also behind the supratemporal. Rows of
villiform teeth develop on the upper and lower jaws in the late larval stage.

In the trunk and tail ossification of the neural spines commences between
5,0 and 5,4 mm while that of the centra starts between 5,5 and 6,0 mm and
that of the haemal spines soon after. Ossification proceeds anteroposteriorly
in sequence.

At 5,4 mm there are two neural spines and at 6,0 mm 5 to 18- Between
6,0 and 6,6 mm ossification proceeds rapidly and 17 to 20 neural spines and
14 to 16 haemal spines are formed, as well as 15 to 20 complete and 2 to 3
incomplete centra (see ‘Table 6). Twenty vertebrae are fully ossified at 7,35 mm
(Fig. gd) and by 8,5 mm 24 centra, with a full complement of 18 to 19 haemal
and 23 neural spines ossified by standard length 9,5 mm. Between 10,0 mm

and the juvenile stage the neural arches ossify completely and the spines —

become broader and stouter.

At 4,8 mm there are four ossified caudal rays, five or six at 6,0 mm. The
notocord becomes heterocercal at 6,6 mm. The full complement of 15 primary
caudal rays is ossified by 7,35 mm as well as five to six secondary rays. The

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 65

TABLE 6

Range of skeletal elements in vertebral column of Helicolenus dactylopterus

Length mm Haemal spine Neural spine Vertebrae No. of specimens
504 a 0-3 Tia i
597 0-9 0-15 0-5 +0-10($) 3
6,0 0-13 5-18 5-6+ 5-9(3) 6
6,6 14-16 19-23 15-21-+2($) 3
7:2 16-17 19-23 20-244 2() 4
7,9 15-18 19-23 21-24-+ 2(4) 2
8,1 13-19 18-23 23-24 2
.9,0 19 23-24 24-26 2

10,5 19 24-25 26-27 4
15,0 19 24-25 26-27 5

(4) refers to partly ossified vertebrae.
TABLE 7

Average number of ossified elements in fins of Helicolenus dactylopterus larvae

Size in mm Caudal Dorsal Anal Pectoral Pelvic
Spines Rays Spines Rays lobe
4,00 4 a sag i oF
5.40 6 ic ne oe 4 “a
6,60 2+15+3 8 + 114 1+6 12 =
7,60 3+154+3 8+3+12 2+6 17 1+3
8,50 4+15+5 5+8+12 2+6 7 1-+4
9,50 6+15+6 12 + 13 3+5 18 I+5
10,50 8+15+9 12 + 13 3+5 19 +5

(Bold face denotes present but unerupted spines)

urostyle and four hypural elements also ossify between 6,6 and 7,35 mm. By
10,2 mm the caudal fin takes on a juvenile aspect and more secondary rays
are present (Table 7).

The pectoral fin starts ossifying between 5,0 and 5,4 mm. Ossification
proceeds dorsoventrally and is completed by 10,2 mm when there are 19 pec-
toral rays. Differentiation into the characteristic Helicolenus pattern of two
unbranched plus 8 to g branched plus 8 to 9 unbranched rays only takes place
in late juvenile or early adult stage.

Dorsal and anal fins ossify between 6,0 and 6,6 mm. At 6,6 mm 11 to 12
dorsal rays are lightly ossified and 6 to 10 dorsal spines visibly ossified but not
erupted. The anal fin has six lightly ossified rays and one unerupted spine.
Three posterior dorsal spines have erupted by 7,35 mm and by 10,0 mm the
unpaired fins have become fully differentiated; all the dorsal spines have
erupted between 9,0 and 9,5 mm. The 12 dorsal spines are shorter and stouter
than the 12 to 13 dorsal rays while the three anal spines are as long as the five
rays except for the second spine which is longer. Between 10,0 mm and 15 to
17 mm the fin supports develop fully.

The pelvic fin is evident at 7,0 mm and the spine is first to ossify. One
spine and three rays are complete at 7,35 mm. One spine and five rays present
the full complement and are present at 10,2 mm. Ossification of skeletal parts
seems to proceed at widely differing rates but as little is known about prevailing
conditions under which growth took place, no deduction can be made.

66 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fic. 9

Distribution of Merluccius capensis from 1951 to 1966.

TABLE 8

Average measurements of Helicolenus dactylopterus larvae in mm

Size range No. of Standard Head Snout ‘Eye Depth
of s.l. specimens length length length diameter
3,50-4,49 15 3,99 1,38 0,37 0,42 1,36
4,5-5,49 27 4,82 1,81 0,57 0,60 1,68
5,50-6,49 10 5,92 2,28 0,91 0,78 2,13
6,50-7,49 14 6,92 2,70 0,94 0,92 2,48
7,50-8,49 9 8,10 3,23 1,10 1,06 2,86
8,50-9,49 4 8,92 3,20 1,12 1,20 3,60
9,50-10,49 2 9,90 4,05 1,35 1,50 3,60
10,50-11,49 4 11,10 4,27 1,42 1,18 3,82
11,50-12,45 2 12,00 4,05 1,65 1,50 4,20
12,50-14,95 I 12,60 4,20 1,20 1,50 4,20
16-17 I 17,10 6,00 1,80 2,40 6,00
18-19 2 — 6,90 1,95 2,40 6,30
20-21 2 — 7,20 2,10 2,85 6,60

Trunk
length
2,04
2,51
3,21
3,70

4,66 -
5,30
6,00
6,75
7:35
7,20
10,80
11,25
13,05

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 67

Body proportions of Helicolenus dactylopterus larvae in % of standard length

Size range
in mm

355-4549
455-549
595-649
6,5-7,49
7,5-8,49
8,5-9,49
9,5-10,49
10,5-11,49
11,5-12,49
12,5-13,49
16,5-17,5*
18,5-19,5*
20,5-21,6

* Size ranges not available in collection.

377

10,0

Distribution of Thyrsjtes atun from 1951 to 1956 and 1960 to 1965.

Eye
%
10,7
12,8
13,1
13,3
13,1
13,4
15,2
12,8
12,5
11,9
14,0
11,8
1355

TABLE 9

Head
%
3459
36,4 .
38,6
39,1
40,0
36,1
40,9
38,5
337
3353

3551
3255
3452

Fic. 10

Trunk

0/
/0

51,7
59,5
542
53,6
57,8
59,6
60,3
60,8
61,0

GA.

63,1
61,8
62,1

roa

Depth

%
34,3
33,9
36,0
35,9
35,0
40,3
36,4
34,5
40,0
3353
3551
30,9
31,4

Bint

No. of

specimens

16

23
11

14

ore — NH Oh O

Cee

68 ANNALS OF THE SOUTH AFRICAN MUSEUM

oS oO (OO

Fic. 11

Distribution of Helicolenus dactylopterus from 1962 to 1965.

Changes in body form

During development the body form changes from a rather deep anteriorly
large to a more evenly proportioned shape with a depth approximately 30%
of the standard length. The relative head length remains fairly constant during
development, varying between 34 and 40% of the standard length. Head
length increases + 0,39 mm for each mm increase in standard length. The
snout is 8 to 14% of standard length and eye diameter 11 to 15% of s.1., both
remaining constant throughout development. The rapid increase in snout
length during size range 3,4 to 5,49 mm could however indicate the growth
and ossification of jaw elements. The trunk length increases 0,6 mm per mm
increase in s.]., increasing to 0,9 mm per mm increase in s.]. at 9,0 to 10,0 mm
s.l. Relative trunk length increases gradually during development from 50 to
51% of standard length between sizes of 4 and 5 mm, to 61 to 62% of s.l.
at 19 to 21 mm.

2 2 heme Acai

LARVAL DEVELOPMENT OF THREE SPECIES OF SOUTH AFRICAN FISHES 69

DIsTRIBUTION

The area covered by the research vessels of the Division of Sea Fisheries
on the pilchard research programme has varied since its inception. During
1951 and 1952 the area worked lay between 32° and 35°30’S and was delimited
by the 200 fathom depthline to the west. Approximately the same area was
worked between 1953 and 1957. In 1958 the eastern limit of the work area
was extended round Cape Point to 19°30’E. This area was worked until the
end of 1960, when the eastward delimitation was extended to 21°E. During
these years the westward delimitation extended to 16°31’E. From July 1963
to December 1965 the area covered by the research vessels was between
32°10’ to 36°10’S and 16° to 21°30’E. Stations lists are obtainable from the
Annual Reports of the Division of Sea Fisheries for the relevant years.

SUMMARY

The larval stages of Merluccius capensis, Thyrsites atun and Helicolenus
dactylopterus are described. All three species are economically important in
South Africa. The taxonomy of each species is revised according to latest
opinions. A brief description of distribution is included.

ACKNOWLEDGEMENTS

The author wishes to thank the Division of Sea Fisheries’s sea-going staff
for the collection of the study material, Fisheries Development Corporation for
financial assistance and Dr. N. A. H. Millard and Mrs S. Bruins for reading the
manuscript and other assistance.

REFERENCES

AutsTtrom, E. H. 1961. Distribution and relative abundance of rockfish (Sebastodes spp.) larvae
of California and Baja California. Rap. P.-V. Réun. Cons. perm. int. Explor. Mer 150: 169-176.

Autstrom, E. H. 1963. Kinds and abundance of fishes in the Californian Current region based
on egg and larval surveys. Rep. Calif. coop. oceanic Fish. Invest. 10: 31-52.

Autstrom, E. H. & Counts, R. C. 1955. Eggs and larvae of the Pacific hake Merluccius productus.
Fishery Bull. Fish Wildl. Serv. U.S. 56: 295-311.

BARNARD, K. H. 1925. A monograph of the marine fishes of South Africa. Ann. S. Afr. Mus. 21:
1-418.

Beaufort, L. F. DE & Cuapman, W. M. 1951. The fishes of the Indo-Australian Archipelago. 9.
Leiden: Brill. ;

BicELow, H. B. & WEtcn, W. W. 1924. Fishes of the Gulf of Maine. Bull. Bur. Fish., Wash.

40 (1): 1-567.
D’Ancona, U. 1933. Uova larve e studi giovanili di Teleostei. Fauna Flora Golfo Napoli 38:
177-384.

Davies, D. H. 1949. Preliminary investigations on the foods of South African fishes. Investl
Rep. Fish. mar. biol. Surv. Div. Un.S.Afr. 11: 1-36.

Davigs, D. D. & Gore, U. R. 1936. Cleaning and staining of skeletons of small vertebrates.
Publs. Field Mus. nat. Hist. (Tech.) 4: 1-15.

De JAGcER, B. v. D. 1955. Development of the snoek ( Thyrsites atun), a fish predator of the pilchard.
Investl Rep. Div. Fish. Un.S.Afr. 19: 1-16.

EHRENBAUM, E. 1gog. Eier und Larven von Fischen. 2. Teil. Nord. Plankt. 10: 217-413.

EscHMEYER, W. N. 1969. A systematic review of the scorpion fishes of the Atlantic Ocean
(Pisces: Scorpaenidae). Occ. Pap. Calif. Acad. Sci. 792 1-143.

Eupurasen, B. A. 1791. Scomber atun och Echeneis tropica beskrifna. K. svenska VetenskAkad. Nya
Handl. 12: 315.

7O ANNALS OF THE SOUTH AFRICAN MUSEUM

FiscHER, W. 1959. Huevos, crias y prelarvas de la merluza (Merluccius gayi), Guichenot. Revta
Biol. mar. 9: 229-249.

Fow er, H. W. 1936. The marine fishes of West Africa. Bull. Am. Mus. nat. Hist. 10: 630-631.

Gitcurist, J. D. F. 1921. Deep sea fishes procured by the S.S. ‘Pickle’ (Part 1). Rep. Fish. mar.
biol. Surv. Un.S.Afr. 2 (Spec. Rep. 3): 41-79.

GrnsBurG, I. 1954. Whitings on the coasts of the American continent. Fishery Bull. Fish’ Wildl.
Serv. U.S. 56: 187-208.

Hart, T. J. & MarsHatt, N. B. 1951. Breeding ground of pilchards off the coast of South-West
Africa. Nature, Lond. 168: 272-273..

Ho uster, G. 1934. Cleaning and dyeing fish for bone study. Zoologica, N.Y. 12: 89-101.

Irvin & JoHnson LimitTED. 1963. South African fish and fishing. Cape Town: Irvin & Johnson.

Jones, S. 1960. On the snake mackerel Gempylus serpens Cuvier from the Laccadive Sea. 7. mar.
biol. Ass. India 2: 85-88.

Kunz, A. & Rapcurrre, L. 1917. Notes on embryology and larval development of twelve
teleostean fishes. Bull. Bur. Fish., Wash. 35: 87-134.

Marak, R. R. 1967. Eggs and early larval stages of the off-shore hake, Merluccius albidus. Trans.
Am. Fish. Soc. 96: 227-228.

MatTHEws, J. P. & Dr Jacer, B. v. D. 1951. The development of the Cape stock-fish Merluccius
capensis. Investl Rep. Div. Fish. Un.S.Afr. 132 1-10.

Moran, J. F. 1956. Differential staining of bone and cartilage in toto of fish. Proc. Indiana Acad.
Sci. 65: 234-236.

Moser, H. G. 1967. Reproduction and development of Sebastodes paucispinis and comparison
with other rockfishes off Southern California. Copeia 1967: 773-797.

Norman, J. R. 1935. Coast fishes. Part I. The South Atlantic (including the Cape Verde Islands,
West Africa, South Africa, Ascension Island, Tristan da Cunha and Gough Island).
‘Discovery’ Rep. 12: 3-58.

Norman, J. R. 1937. Coast fishes. Part II. The Patagonian region (including the Straits of
Magellan and the Falkland Islands). ‘Discovery’ Rep. 16: 3-150.

Norman, J. R. 1966. A draft synopsis of the orders, families and genera of recent fishes and fishlike verte-
brates. London: British Museum (Natural History).

Recan, C. T. 1908. Descriptions of the new or little known fishes from the coast of Natal
Ann. Natal Mus. 1: 1-6.

Recan, C. T. 1916. Larval and post-larval fishes. Nat. Hist. Rep. Br. Antarct. Terra Nova Exped.
(Zool.) 4: 125-156.

SANTANDER, H. & De Castitxio, O. S. 1969. Desarrollo y distribucion de huevos y larvas de
merluza, Merluccius gayi (Guichenot) en la costa peruana. Boln Inst. Mar Peru 2: 79-107.

SmitH, J. L. B. 1953. The sea fishes of southern Africa. 4th ed. Cape Town: Central News Agency.

Voss, N. A. 1954. The postlarval development of the fishes of the family Gempylidae from the
Florida current. I. Netiarchus Johnson and Gempylus Cuv. & Val. Bull. mar. Sci. Gulf Caribb.
4: 120-157.

FO EF OES

INSTRUC PIONS: TOVUAUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS
To be reducible to 12 crn X 18cm (1g cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.
For journal articles give title of article, title of journal in italics (abbreviated according to the

World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,

volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

BULLOUGH, W. 8S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHER, P.-H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region
of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In scHULTZE. L,
Koologische und anthropologische Ergebnisse einer Forschungsreise 1m westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51).
The Harvard system of reference to be used in the synonymy lists, with the full references
incorporated in the list at the end of the article, and not given in contracted form in the synonymy
hst.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, b; Liste: 11. Turton, 1932: 8o.

—\

a

“ene ei ud vege ‘ia
bist, 4 a mex A ahah

Way
t
:
Lise
r
‘ 5
i
a gas
; 1

mse 776s

_ ANNALS OF THE SOUTH AFRICAN MUSEUM
_ ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 ~ Band
March 1972 Maart
Part 4 Deel

A PLIOCENE PHOCID FROM
SOUTH AFRICA

By
Q. B. HENDEY & C. A. REPENNING

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 2(1, 3> 5 7-8), 3(1=2; 5> pals); 51-2, 5; 7-9),
: 6(1, tps), fey) 8, G(t—2); 10(1),
i Tio: Be 7s t=pel-)5 21, 24(2), 27; S1(I-9), 33> 38

Price of this part/Prys van hierdie deel
R4,00.

Trustees of the South African Museum © Trustees van die Suid-Afrikaanse Museum

1972

=

ISBN 0 949940 04 6

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

A PLIOCENE PHOCID FROM SOUTH AFRICA

By
Q. B. HENDEY

South African Museum, Cape Town
&
CHARLES A. REPENNING
U.S. Geological Survey, Menlo Park, California
(With plates 2-18, 2 figures and 7 tables)

[MS. accepted 9 September 1971]

CONTENTS

PAGE
introduction 76 1''* . s L : j a |
Systematics ; d ; : : 5 mIe7o
The Langebaanweg seal . : : F ; ers

Description
The skull . : : : : : 3 2? 36
The postcranial skeleton. ‘ c é A iss
Discussion : F : ; , ; , » nae
Summary i : : : : : : . “96
Acknowledgements . ; é ‘ j 3 21 f<g6
References . : ¢ ’ ; : 2 - 996

INTRODUCTION

Although Illiger recognized the basic features of the seals of the world in
1811 by separating them from the sirenians and placing them in a separate
order, the Pinnipedia, their classification is still in a state of flux. In 1880 Allen
divided the pinnipeds into two major groups: the ‘walkers’ and the ‘wrigglers’
which Smirnov (1908) subsequently named the superfamilies Otarioidea and the
Phocoidea. The Phocoidea contains only one family, the Phocidae (Gray 1825,
but defined with its present contents by Brookes, 1828), usually known as the
‘earless seals’ or ‘true seals’. Subsequent to Kellogg’s (1922) introduction of the
subfamily Lobodoninae (respelled Lobodontinae by Hay, 1930) to include the
Antarctic phocids, the family Phocidae was considered to include four sub-
families: Phocinae for the northern seals, Monachinae for the genus Monachus,
Cystophorinae for the genera Cystophora and Murounga, and Lobodontinae.
Scheffer (1958) reduced the rank of the Antarctic phocids to that of a tribe,
Lobodontini, within the family Monachinae. King (1966), in possibly one of the
most detailed explanations of a change in pinniped classification, abandoned
the subfamily Cystophorinae, placing the genus Cystophora in the Phocinae and
the genus Mzrounga in the Monachinae. Most recently (at this writing) McKenna

7X

Ann. S. Afr. Mus. 59 (4), 1972: 71-98, 17 pls, 2 figs, 7 tables.

72 ANNALS OF THE SOUTH AFRICAN MUSEUM

(1969), in possibly one of the briefest explanations of a change in pinniped
classification, stated that the formal taxon order Pinnipedia has been aban-
doned.

As here used, the order Pinnipedia contains three families: Odobenidae,
the walruses; Otariidae, the sealions; and Phocidae, the true seals who must
‘wriggle’ on their bellies in terrestrial locomotion, their hind limbs permanently
extended behind them. The family Phocidae contains two distinct subfamilies,
the Phocinae, inhabiting temperate and arctic waters of the Northern Hemi-
sphere, and the Monachinae, inhabiting parts of most oceans except the Arctic
Ocean.

For the most part, the characters identified by King (1966) may be used to
separate the living members of the two phocid subfamilies. Since it is assumed
that the two subfamilies derive from a common ancestor, it is to be expected
that fewer of these characters will be useful for familial identification in the
fossil ancestors of the living true seals. Such is the case with the Pliocene phocid
from South Africa.

Since 1958, when the first discoveries from the quarries of the African
Metals Corporation were recorded by Singer & Hooijer (1958), tens of thou-
sands of vertebrate specimens have been recovered from these Pliocene phos-
phate deposits near Langebaanweg, Cape Province. At least 60 mammalian
species, as well as a full complement of birds and cold-blooded vertebrates, have
been recognized (Hendey, 1970a, 1970). The fauna is both marine and ter-
restrial, apparently representing accumulation during a prolonged period of
marine, estuarine, and terrestrial deposition in the area.

The dating of the Langebaanweg fauna is a major problem presently
being investigated (Hendey, 1970); Maglio & Hendey, 1970). On the basis of an
admittedly limited number of comparisons with faunal elements from better-
dated localities in East Africa, we believe the Langebaanweg deposits bearing
the main body of higher vertebrate remains to be perhaps 4 to 5 million years
old. ‘The most recent interpretation of the most reasonable definition of Pleisto-
cene, and the approximation of its beginning between 2,6 to 3 million years ago
(Savage & Curtis, 1970), suggest that the Langebaanweg fauna should proba-
bly be called late Pliocene.

The Langebaanweg fauna includes a pinniped which was first reported by
Boné & Singer (1965). These authors tentatively referred it to the otariid genus
Arctocephalus, but more recently discovered specimens show clearly that it is a
monachine phocid which belongs to a previously unrecorded species of the
extinct genus Prionodelphis, heretofore known only from Pliocene deposits in
Argentina. In many respects this seal from Langebaanweg is similar to the
extant Monachus and its ancestors, notably Pliophoca etrusca from the late Pliocene
of Italy (Tavani, 1942). However, similarities to the extant phocids of the
Antarctic seas are equally well marked.

In view of the small number of recorded specimens of Prionodelphis rovereti
(Frenguelli, 1922, 1926), from Argentina, the Langebaanweg seal, represented

A PLIOCENE PHOCID FROM SOUTH AFRICA 73

by a wide variety of specimens, is clearly important in that it provides the first
good evidence of the antiquity and ancestry of the monachine seals in the
Southern Hemisphere.

SYSTEMATICS

As here used, the subfamily Monachinae includes the same genera of living
seals that were included in this subfamily by King (1966). Except for the addi-
tion of the genus Mirounga, this agrees with the definition of the subfamily as
originally defined by Trouessart (1897 :373).

As noted by Scheffer (1958:111), when he reduced the Lobodontinae to
tribal rank within the Monachinae, the major difference between these Antarc-
tic seals and Monachus is one of geography. King (1966) omitted reference to
Scheffer’s tribe Lobodontini when discussing the monachine relationship of
Mirounga, with good reason as the genus does not conform to the geographic
distinction mentioned by Scheffer. Recognition of any tribal subdivision of the
Monachinae now seems pointless.

King (1966:397) noted that with some features otherwise typical of the
monachine seals Monachus was an exception and regarded this genus as being
not quite so advanced. Such exceptions to otherwise typical features of the
Monachinae are even more evident in the Pliocene seal from Langebaanweg.
At this stage in the evolution of the phocid seals the subfamily Monachinae can
be distinguished from the subfamily Phocinae by the following characters of the
skull.

Subfamily Monachinae

Diagnosis. Seals having a mastoid bone without a prominent posterolater-
ally projecting rounded crest but, instead, having a posterolateral surface
curving uniformly from the region of the parietal suture down to the region of
the stylomastoid foramen; mandible with an extensive symphyseal surface that
is elongate and smoothly oval in outline and that firmly articulates over the
entire depth of the chin.

As will be shown in the following report, some postcranial bones of the
Langebaanweg seal exhibit monachine features while other are simply phocid,
with no subfamilial characteristics.

Genus Prionodelphis
Type. Prionodelphis rovereti Frenguelli, 1922.

Known distribution. Pliocene of the South Atlantic Ocean.

Comment. The type species, Prionodelphis roveretit, was described as a squalo-
dont cetacean from a few isolated teeth found in Pliocene deposits in Entre Rios
in Argentina. A mandibular fragment bearing one tooth was later found at
the same locality which led Cabrera (1926:390) to the conclusion that the
animal was a pinniped, a conclusion supported by others (Frenguelli, 1926;

74 ANNALS OF THE SOUTH AFRICAN MUSEUM

Kraglievich, 1934; Kellogg, 1942). We are aware of no additional material.

The material from South Africa now makes it possible to provide a better
definition of the genus.

Diagnosis. A generalized monachine seal lacking the shortened rostrum and
crowded teeth of Monachus monachus and the squared premaxillaries with aligned,
upper incisors of M. tropicalis and M. schauinsland:; postcanine teeth low-cusped
as in M. schauinslandi, M. tropicalis and Pliophoca etrusca and distinctly narrower
in occlusal outline than those of M. monachus; upper fifth postcanine! with
recurved crown; ascending ramus of premaxilla strong, terminating against
nasals and prominently visible in lateral view separating maxilla from nasal
aperture; pre-orbital processes prominent; forehead broad in supra-orbital
region; osseous nasal septum strongly developed; dental formula 2.1.5/2.1.5;
tympanic bulla covers petrosum.

‘THE LANGEBAANWEG SEAL
Prionodelphis capensis n.sp.

Holotype. An incomplete skull with left canine and fourth postcanine, and
right third postcanine (South African Museum No. L 15695).

Referred material. An incomplete skull (L 12695); temporal bone (L 15652) ;
mandible fragments (L 7556, L 12299); one lower and two upper incisors; three
lower and four upper canines; and nine lower and ten upper postcanines.

Various elements of the postcranial skeleton have been recovered, of which
the following have been selected for description: vertebrae (L 7563, L 15680,
L15849A1 & Ag, L 15396, L 15857); scapula (L 2160); humeri (L 2157,
L 4638); ulnae (L 2161, L15682); radu (L 2935, L 12869), innominate
(L15849A), femur (Lio1gr1); tibiae (L 2138, L10128/9); calcaneum
(L1o118); astragali (L 10130, L 10993); navicular (L 15851); entocuneiform
(L 10134); metapodial (L 10996); first phalanx (L 10999); second phalanx
(L 10205).

All specimens are housed in the South African Museum, Cape Town.
Except for the two incomplete skulls, which are too fragile to cast, casts of the
more significant specimens are housed in the U.S. Geological Survey, Pacific
Coast Center, Menlo Park, California.

Locality and horizon. ‘The holotype and most of the referred material is from
horizon 2, ‘E’ Quarry, Langebaanweg. Some postcranial elements are known
from horizon 1. These horizons are thought to be broadly contemporaneous
(Hendey, 1970b). A few fragmentary remains from ‘C’ and Baard’s Quarries
are excluded from this report, but they apparently represent the same species.

Comparative material. Skulls of all living phocid species except Pusa caspica
have been available for comparison either in the South African Museum or in
the Pacific Coast Center of the U.S. Geological Survey. Postcranial material has
been somewhat less available, but comparisons were made with postcranial
1 Called P, by Frenguelli, 1922: 496.

A PLIOCENE PHOCID FROM SOUTH AFRICA 75

elements of Monachus schauinslandi, Hydrurga leptonyx, Lobodon carcinophagus,
Mirounga angustirostris, M. leonina, and all living phocine genera.

Diagnosis. Prionodelphis capensis differs from the type species in that it has
only one anterior accessory cusp on the lower postcanines instead of two, and
there is a greater reduction in size of the posterior root in the second to fourth
upper postcanines. The cheek-teeth of the South African species approach more
closely a three-cusped tooth pattern, and in addition, are slightly larger and
more laterally compressed (Table 1).

TABLE 1. Average length-width ratios for postcanines 2 to 4 of some phocids.

Upper/
Length Width W/L x 100 Lower
Hydrurga leptonyx
meu N =O) ae fF * Mane oOo nam 8,8 mm 49 0,925
iEawer (N= 6)) . : ‘ o) 7G 9,3 53
Prionodelphis capensis
igen i(IN <7)... : : +) Lae 755 54 1,200
Momeni — Oy) ay oat | he) | a. 1G,0 6,8 45
Prionodelphis rovereti+
IDEE TON InrreheWiMt yg fev, eee or 8,0 65 1,204
Mawern(NI— 2 )iriies hi. Uti. nn. =v yprlg3o 7,0 54
Monachus schauinslandi
EVN, =O) ea a uw 1230 7,9 66 1,047
Mawem NO) Posh a se, 2ST 7,6 63
Monachus monachus
Wppers(N — 6)". 4 ; TR 4) Q,1 70 1,011
Lower (N= 6) . ‘ : sl. T2536 8,0 63

1 From Frenguelli, 1922.

DESCRIPTION

The assessment of the fossil remains listed above is somewhat hampered by
their fragmentary nature. The holotype is composed of about 60 individual
pieces, including three teeth, found scattered over a wide area in the excavation
No. LBW 1969/1 (South African Museum departmental records). Numerous
small pieces could not be restored to the skull, although they undoubtedly
belong, and others presumably remain in unexcavated parts of the deposit.

The second partial skull (L 12695) was similarly fragmented. Although the
partially restored snout region is less complete than that of L 15695, parts of the
braincase of the second specimen were also recovered. Many of the individual
pieces are extensively abraded, probably having suffered in a manner similar to
that described for an alcelaphine skull recovered near by (Hendey, 1970a: 82).

As with much of the referred material, the temporal bone (L15652) also
came from the excavation LBW 1969/1, but from approximately 75 cm below
the holotype. It probably belongs to another individual.

Most of the isolated teeth were recovered intact, and they vary from un-
worn to extremely worn.

76 ANNALS OF THE SOUTH AFRICAN MUSEUM

Few elements of the postcranial skeleton were recovered intact, but in some
cases sufficient numbers of a particular bone are known to enable a complete
assessment of its characteristics.

THE SKULL

The skull of the Langebaanweg pinniped is in many ways unique, and it
exhibits a set of characteristics which makes it impossible to assign it to any
previously known phocid species (Plates 2-8).

As a whole, the skull appears convincingly to be that of a monachine seal.
The lack of swollen or crested mastoids, the broad and flat dorsal surface of the
petrosal apex, and the deep and oval mandibular symphysis rule out any known
phocine seal, while the incisor formula and cheek-tooth pattern strongly
suggest a monachine seal. Although the postcranial bones in general also appear
monachine, they possess some features that are characteristic of living phocine
seals, such as an entepicondylar foramen on the humerus.

In general features, the skull most resembles those of Monachus and Hydrurga,
although it is considerably more gracile and less elongated than the latter. Of all
extant genera its teeth most resemble those of Monachus. However, except for the
lack of great vertical exaggeration of the cusps, the teeth are also quite similar to
those of Hydrurga. Of the extinct genera they are, as far as comparisons are pos-
sible, most similar to Prionodelphis rovereti and, less so, to Pliophoca etrusca. ‘The
basic pattern of the cheek-teeth appears to be that of earlier members of the
family Phocidae.

Unlike the usual pattern in the Antarctic monachines, the premaxilla
terminates against the nasal bone. Although it shares this characteristic with
Monachus and several phocine seals, it differs markedly, as does Pliophoca
etrusca, in the massiveness of the ascending ramus of the premaxilla. When the

TABLE 2. Dimensions of the skull of Prionodelphis capensis from Langebaanweg.

L15695 L12695 L7556 L12299

Rostral width 5 : 48,0* 61,0* — —
Distance between pre- aie oe

cesses of maxilla ; 94,0* = == =
Minimum diameter of inigearbitin

region . 36,0* — = =
Distance benveca: pre- aeecillerts nea

jections and posterior limit of

nasal aperture. ; ; ; 58,0* 51.0* = =
Lengths of nasals. : : 62,0* — — ==
Distance between external seanen

margins of fifth postcanines. 74,5* 87,0* ee =
Alveolar length of upper postcanine

SELies: an 72,5 7258 = =
Alveolar length i lees saiiranae

SEHES" =5,2 Bete — — 66,5 ==
Depth of taandible behind M, ai Neds, — — 27,9 35.6

* Estimated.

~ ‘sotias ouIuvoysod szaddn jo yysug] 1e[ooATY -/, -sguruesysod yyy jo sursireur
~~ ss IepOaATe [eUIN}X9 UIIMJoq JoURISIC] *Q “STeSBU JO yysuaT ‘GS ‘ornjsode [eseu Jo WUT] zor19ysod puv suonsefosd Areypixeuroid usaMjoq 90UeysIC] * ‘uo1so1
[e1Iq.10-19}UT Jo Ja}OWINIP UNUUTUTYY “& “eT]EXeUT Jo sassao0id [ey1q40-o1d usIMyoq 9OULISIC| *S “YIPIM [eAISOY “1 + s}USWMOINSVOUT jetuess Jo wieiseIg “1 “Oly

G

A PLIOCENE PHOCID FROM SOUTH AFRICA

78 ANNALS OF THE SOUTH AFRICAN MUSEUM

skull is viewed laterally the premaxilla is visible along its entire length rather
than being partly hidden behind the maxilla, largely within the nasal opening,
as in living monachines. The nasal opening is elongated antero-posteriorly, and
of the southern phocids most resembles Hydrurga and Lobodon, rather than
Leptonychotes and Ommatophoca in which there is foreshortening of the ante-
rior part of the snout. In this respect it is also similar to Monachus and Pliophoca.
The two halves of the premaxilla have fairly prominent projections at their
most anterior point of contact, dorsal to the incisors. Of the extant monachines
M. tropicalis, M. schauinsland and Hydrurga have similar projections, but in
Hydrurga inflation of the alveolar region of the incisors renders the projections
less prominent, and in the two Monachus species the projections are quite widely
separated. In the fossil species there is a marked step between the most anterior
maxilla-premaxilla contact and the premaxillary projection. This is best seen
in the specimen L 12695.

Judging from the size of the alveoli, the lateral incisor is only slightly
larger than the medial one. The latter is situated slightly anterior to the other.
The relative size and position of the upper incisors are as in M. monachus and
Hydrurga, and unlike the specialized condition found in Lobodon and Lepto-
nychotes, in which genera the lateral incisor is much larger than the medial one
and is situated in line with it.

Two upper incisors are known, both left medials judging from their size.
The structure of the crown resembles that of the incisors of MZ. monachus and
Pliophoca etrusca. ‘The cross-section of the canines is only slightly elongated
antero-posteriorly, and in this respect P. capensis differs markedly from the Ant-
arctic phocids. The canines L 11686 (Plate 9B) and L 12695 have large bulbous
roots, a feature also seen in some of the postcanines. ‘This is a characteristic of
old age in most or all pinnipeds.

There are five upper postcanines and at least one specimen of each dental
category is known from the assemblage. The first, and smallest, is single rooted
(Plate gC). The remainder are all double rooted, with the posterior root being
larger than the anterior one in the second, third, and fourth teeth, and the sizes
reversed in the fifth. The latter is also the second smallest of the teeth and is set
slightly separate from the others immediately below the infra-orbital foramen.
The postcanine tooth rows curve outwards posteriorly as in Ommatophoca, and
are not diverging straight lines as in Leptonychotes and Hydrurga; this pattern
is rather close to that of M. schauinslandi as well as some other phocid species.

The postcanines resemble those of Pliophoca etrusca, although narrower and
more gracile, and except for being much less massive, they are also similar to
those of M. monachus (Plate 9D). They are quite distinct from the highly specia-
lized teeth of extant Antarctic phocids; of this group the teeth of Leptonychotes
are Closest to the fossils, but they are nevertheless significantly different. The
basic pattern of the first to fourth postcanines is similar to the corresponding
teeth of M. monachus. There is a prominent central cusp with two smaller cusps
situated anteriorly and posteriorly, with an additional small projection on the

A PLIOCENE PHOCID FROM SOUTH AFRICA 79

most posterior part of the cingulum. The enamel is generally rugose, a condition
which is found in all monachine seals except Mirounga, but which is also found
in some phocine seals. The second, third and fourth upper postcanines have a
marked inflation of the postero-internal cingular region, a condition also
evident in some teeth of P. rovereti, and P. etrusca, and these teeth are broader
posteriorly than they are anteriorly. The first upper postcanine has the internal
cingulum inflated, and the maximum transverse diameter is at about the mid-
point of the tooth. The upper postcanine of P. roverett described by Frenguelli
(1922:493 and figs. 1b, 1c) is shorter and relatively broader than those of P.
capensis. It lacks the distinct anterior accessory cusp present in the latter species
and has a larger posterior root.

The fifth postcanine of P. capensis lacks the posterior accessory cusp and
cingular projection, and the anterior accessory cusp is much reduced, being
barely discernible; the principal cusp is strongly recurved (Plate 9E). It resem-
bles the cheek teeth of Leptonychotes in this respect. Both the fifth postcanines
recovered to date are completely unworn, and it seems probable that this tooth
did not occlude with the lower fifth postcanine. Leptonychotes also has a non-
occluding upper fifth postcanine. This tooth, in P. capensis, is very similar to one
of the original P. roveret: specimens illustrated (as a P,) by Frenguelli (1922: fig.
2A). The principal differences are in the smaller size of the P. rovereti specimen,
in that its anterior accessory cusp is situated higher up the crown, and also in
that there is no cingulum on the buccal surface of the Langebaanweg specimens.
The two P. capensis specimens differ from one another only in that one is slightly
shorter and somewhat broader than the other. In the postcanines such differ-
ences distinguish upper from lower teeth, but judging from the size of alveoli in
mandibles and maxillae known, these teeth can only be upper fifth postcanines.

Of the seven second, third and fourth upper postcanines known, two are
unworn and five show wear angled from the principal cusp to the posterior
cingulum.

The general features of the maxilla of P. capensis correspond most closely to
those of Hydrurga. The fossil seal has very prominent preorbital processes. The
one preserved in the holotype projects outwards and downwards as in Hydrurga,
although in L 12695 it has an outwards and upwards inflection. The presence
and form of the preorbital process is variable throughout the Pinnipedia, but
within the monachine seals it is virtually absent in M. schauinslandi, M. tropi-
calis and Leptonychotes and present but variably developed in other extant
monachine species.

In the Antarctic phocids the jugal terminates lateral to the infra-orbital
foramen, whereas in P. capensis, P. etrusca, Monachus and several other seals it
terminates above this foramen.

The infra-orbital foramen is oval shaped as in Hydrurga and Lobodon, but the
orientation of the longitudinal axis of the foramen differs in that it is directed
upwards and outwards in P. capensis, whereas the axis is upwards and inwards
in Hydrurga and Lobodon.

80 ANNALS OF THE SOUTH AFRICAN MUSEUM

The shape of the nasals is not perfectly known, but from a reconstruction of
this region in the holotype (Plate 2), it appears that they do not correspond in
shape to those of other monachines. Instead they are broader in the frontal
region than between the maxillae.

The osseous nasal septum is a very stout bone which terminates at or near
the most anterior limit of the nasals. Its proportions resemble those of M.
schauinslandi, M. tropicalis, Hydrurga, Ommatophoca and Mirounga, but not M.
monachus, Lobodon and Leptonychotes.

The supra- and post-orbital regions of the frontal bones are essentially
similar to those of Hydrurga, Lobodon and Ommatophoca, and are not parallel-sided
as in Monachus, Leptonychotes and Mirounga, nor as in P. etrusca. In the holotype
there are the beginnings of a sagittal crest towards the posterior part of the
frontals. There are two step-like projections on the frontals above and behind the
preorbital processes of the maxilla. Similar features are present in Ommatophoca
and are less distinctly represented in Hydrurga.

The second partial skull (L 12695) belongs to an aged individual and is
somewhat more robust than that of the holotype (Plate 5). The difference in
size of the two specimens may in part be due to the ages of the individuals
concerned, but may also reflect sexual dimorphism in the species. Apart from
size, the most striking difference between the two specimens is in the form of the
nasal aperture. That of L 12695 is actually slightly shorter than that of L 15695,
although it is, as would be expected, wider and higher. A similar and probably
related allometric feature is the correspondence in the length of the postcanine
tooth rows.

The braincase of P. capensis is not known, although parts of the nuchal
region of L 12695 were recovered. The nuchal crest is fairly well developed,
being more similar to that of Hydrurga than other species, although, like the
sagittal crest, it is considerably less prominent than that of Hydrurga. Unlike
Hydrurga there is no marked concavity of the supra-occipital. As in the Antarctic
seals, Monachus monachus, and Pliophoca etrusca, the nuchal crest extends anteriorly
across the temporal, terminating near the external acoustic meatus, rather than
uniting with an enlarged jugular process of the exoccipital as in M. schauins-
land: and M. tropicalis.

The basi-cranium is largely unknown, but a number of specimens of the
well-ossified temporal bone have been recovered. The most complete (L 15652)
has the mastoid and most of the tympanic intact, and the post-glenoid process
is still attached (Plate 6). That part of the ectotympanic which projects under the
acoustic meatus has been broken off and lost.

Although the general appearance of the tympanic region is most reminis-
cent of Monachus because of the slight inflation of the bulla, some of its features
strongly indicate a relationship between Prionodelphis capensis and the Antarctic
phocids. These include the posterior extent of the bulla and the rounded apex of
the petrosum.

As noted by King (1966:387), the posterior wall of the bulla in phocine

A PLIOCENE PHOCID FROM SOUTH AFRICA 81

seals (except some individuals of Erignathus), and in species of Monachus, is
located rather far forward so that the posterior part of the petrosum is exposed in
ventral aspect (without recourse to peering through the posterior lacerate
foramen), whereas in the Antarctic monachine seals, including Mirounga, the
bulla covers the petrosum and essentially separates the mastoid from the pos-
terior lacerate foramen by almost contacting the exoccipital. The latter condi-
tion is very evident in P. capensis, strongly suggesting an affinity with the
Antarctic seals.

The dorsal (cerebellar) surface of the temporal resembles that in some
living Antarctic seals (Plate 7). The apex of the petrosum is broad and rounded
with low relief, as in Lobodon and Leptonychotes, and differs greatly from the point-
ed apex found in Monachus. It is not the globular structure typical of the pho-
cine seals. However, the petrosal apex of P. capensis is smaller than in the living
Antarctic seals, suggesting that this seal was less well adapted for directional
underwater hearing, according to the interpretation of Repenning (in press).
The cerebellar fossa is relatively large, as in Leptonychotes, and as in the latter,
the squamosal extends medially to the edge of the cerebellar fossa and to the
internal facial canal.

The ventral (external) side of the temporal most resembles that in Hydrurga.
The external opening of the carotid canal is located well forward of the posterior
limit of the bulla, as in Hydrurga, Leptonychotes and Lobodon, but in contrast to the
more posterior location in Ommatophoca, Mirounga and Monachus. In P. capensis
this foramen faces noticeably ventrally, as in Hydrurga, but the general outline of
the bulla is more similar to that in Lobodon. The stylomastoid foramen is rather
widely separated from the external cochlear foramen (Burns & Fay, 1970:374).
A similarly wide separation is found in Hydrurga.

Dissection of the middle ear was not undertaken.

The mandible is rather unspecialized and resembles that of Hydrurga and
Monachus (Plate 8). It differs from that of Hydrurga in its smaller size and in having
the symphyseal region relatively narrower transversely. In Hydrurga the two
lower incisors are situated side by side, but in the fossil the medial incisor lies in
an almost horizontal position above and behind the lateral incisor, in a manner
comparable to that in most other species of phocid seals.

A single isolated lower medial incisor (L 15444A) is known. It is similar in
size to that of Leptonychotes but has a pronounced step on the lingual surface of
the crown which, in lateral view, resembles that of Monachus.

The lower canines are similar to the uppers, but have a straighter root and
are more rounded in cross section.

The five lower postcanine teeth are situated close to one another: the
alveolar walls between the teeth are as narrow as, or narrower than, those
between the two roots of one tooth. In the mandibular ramus L 7556, the teeth
are positioned in much the same way as in the P. etrusca specimen described by
Tavani (1943: fig. 6a). The first postcanine is single rooted, and the remainder,
which are more or less equal in size, have two roots. These teeth apparently all

82 ANNALS OF THE SQUTH AFRICAN MUSEUM

have a crown pattern similar to the first to fourth upper postcanines (Plate 9G).
Nine isolated lower postcanines are known, one of which is identified as a first
lower and two are thought to be second lowers. Both the latter are worn on
their posterior surfaces. Of the remainder, three show most wear on their anterior
surfaces, and four, including the first lower, show no perceptible wear at all.
Judging from the wear on the upper and lower postcanines, it appears that they
functioned as crushing agents, although two of the lower teeth show signs of a
transverse shearing action.

The lower teeth are differentiated from the uppers by the fact that they
are more slender, with little or no inflation of their internal cingula. The lower
postcanines of P. rovereti (Frenguelli, 1922:499, figs. 2B, 2C) are shorter and
relatively broader than those of P. capensis, and have a variably developed
second anterior accessory cusp which is not present in the Langebaanweg
species.

TABLE 3. Dimensions of the teeth of Prionodelphis capensis from Langebaanweg.

UPPER LOWER
A-P A-P
dia- Transverse dia- Transverse
No. meter diameter No. meter diameter
Incisors Incisors
Med. L 11689 6,1 553 Med.| L15444A 3,8 B28
Med L 15381B 6,3 5,2
Canines L11686 10,1 755 Canines| L 15437 Q,2 754.
L 15241 €. 10,2 75 L 15743 9,0 753
L 15630B/2 | ¢. 10,1 c. 8,0 L 13152 9,6 7,8
L 15695 10,1 74
Post- Post-
canines canines
Ist L 11687 9,0 6,1 Ist L 15580 957 5,6
2nd L 15611 13,9 C3750 and | L15680B/2 14,5 6,4
Pend | L15736A 14,2 6,5
3rd L 15695 14,5 755
4th L 15695 13,4 74
5th L 12562 9,8 6,4
5th L 15429 10,4 555
? L 12557 12,9 8,1 & L 15420/1 ¢. 15,0 6,5
? L 12556 13,7 7,8 i L 15413B 1554 6,4
? L 15630B/2 13,8 752 ? L 10160 14,8 732
? L 15664 14,1 7,8 ? L 15444B 15,2 7,0
? L 15771 15,0 71
ie L 12124 16,0 Fs

The proportions of the upper and lower cheek teeth to one another are
similar in P. capensis and P. rovereti, and differ from some extant monachines
(Table 1).

The height of the mandibular corpus is fairly constant between the poste-

A PLIOCENE PHOCID FROM SOUTH AFRICA 83

rior limit of the symphysis and the fifth postcanine (Plate 8). In the mandible
fragment L 12299, the masseteric fossa begins about 20 mm behind the fifth
postcanine, and the ascending ramus begins inclining at about this point. ‘There
is no corresponding upward inflection of the inferior margin of the corpus as
there is in Leptonychotes, and the fossil resembles most other phocid species in this
respect. ‘

The symphysis of the mandible is short relative to that of some monachine
seals, and terminates below the posterior root of the second postcanine. It is
typically monachine, however, and has a strong, oval articular surface over the
entire depth of the jaw. The mandibular condyle is not known.

There are multiple mental foramina towards the anterior part of the cor-
pus.

THE POSTCRANIAL SKELETON

While the skull characters of Prionodelphis capensis show its relationships to
lie with the Monachinae, certain features of the postcranial skeleton are more
commonly found among the Phocinae (see King, 1966). Presumably, the
‘phocine’ characteristics are inherited from the primitive ancestral stock, and
are features which were lost by the Monachinae during their later development.
Characteristics which are typically monachine are also evident.

Vertebrae

Associated with the innominate to be described later (L.15849), were a
number of vertebrae, most of which were badly crushed and incomplete.
However, two lumbar (L 15849A1 and A2) and one caudal vertebra (L 15857)
were reasonably well preserved although still incomplete (Plate 10C, D, F).
Other vertebrae recorded are an axis (L 7563), one other cervical (L 15689),
and a sacral (L 15396) (Plate 10B, E), all of which are damaged, and a number
of other fragmentary specimens.

Only the centrum of the second cervical vertebra is preserved. The odontoid
process is prominent, with a length of 16 mm and a maximum transverse
diameter of 17,7 mm. The total length of the centrum is 53 mm, and the trans-
verse diameter of the anterior articular end is estimated to be about 55 mm. The
other cervical vertebra, probably a fourth, consists of the centrum, parts of the
left transverse processes enclosing the vertebrarterial canal, and part of the left
half of the neural arch. The centrum is 47,6 mm long and the transverse
diameter of the anterior epiphysis is 31,4. mm. Both these specimens have the
reduced transverse processes which characterize the Phocidae (King, 1964:98).

The two lumbar vertebrae, a second or third and a fifth, are similar in size
to those of Pliophoca etrusca (see Ugolini, 1902, and Table 4). The transverse
processes of the fifth lumbar vertebra are very prominent, which is characteris-
tic of all Phocidae (King, 1964:99).

Although the sacral vertebra, a third, is from an adult individual it was not
fused to the second. However, the anterior end of the preserved right transverse
process is markedly rugose, suggesting that there was a strong cartilaginous

84 ANNALS OF THE SOUTH AFRICAN MUSEUM

Tase 4. Dimensions of lumbar vertebrae of Prionodelphis and Pliophoca.

Prionodelphis capensis Pliophoca etrusca
and or 3rd 5th and 3rd 5th
Lenpth ef.centriim 5. sis een 44 60,4 55.4 62,0 61,5 55,5
Transverse diameter of anterior
epiphysis : : : : 41,0 42,2 41,0 - 40,0 44,0

attachment between it and the posterior end of the second sacral transverse
process. This specimen is unusual in that it lacks the left transverse process, and
the anterior end of the centrum and neural arch are, as a result, asymmetrical
with the dorso-ventral median axis directed from right to left at a slight angle to
the normal line. The length of the centrum is estimated to be 38 mm, while oe
of a Pliophoca etrusca specimen is 36 mm (Ugolini, 1902).

In the caudal vertebra, a first, much of the centrum is lost, apparently
having been gnawed away. Crushing, punctures, and gnaw-marks resulting
from grasping and chewing by carnivores are a not uncommon feature of the
fossils from Langebaanweg. Both anterior and posterior zygapophyses of the
caudal vertebra are well developed, and the distance between their anterior and
posterior limits is 47 mm. The transverse diameter of the anterior epiphysis of
the centrum is approximately 25 mm.

Anterior Limb
Scapula

A single incomplete pinniped scapula (L 2160) is known from Langebaan-
weg (Plate 10A). This specimen consists of the articular end, neck and lower
parts of the blade. The acromion and internal margin of the articulation are
damaged.

The glenoid cavity is markedly elongated and concave; the concavity as
well as the breadth/length ratio being comparable only to that of Lobodon and
Monachus (67,9% for L 2160, 72,0% for one Lobodon scapula, and 67,1% for one
Monachus schauinsland: scapula). Other monachine genera have shallower and
more nearly equidimensional scapular glenoid fossae.

Also most comparable to Lobodon and particularly Monachus, the neck of the
fossil scapula is extremely short and has an antero-posterior diameter of 45,8 mm.
The spine is strongly developed for a monachine seal, but might not exceed the
development of that of Lobodon. Too little is preserved to be certain. The sharp
scapular notch, where the anterior margin of the coracoid process turns into the
inferior border of the supraspinous fossa, suggests that the anterior border of the
scapula may have been straight and vertical as in the Antarctic monachines.

The medial surface of the subscapular fossa is divided into two parts by a
prominent ridge most resembling that of Monachus in its prominence and lo-
cation. The ridge is more prominent than in Monachus, however, and is better
developed than in any living phocid. Although the scapula differs greatly in
other respects, the prominence of this subscapular ridge is equalled only by the

A PLIOCENE PHOCID FROM SOUTH AFRICA 85

Miocene Phoca vindobonensis 'Toula (1897: pl. 9, fig. 15a).
Humerus

All fourteen humeri recovered to date in which the distal end is preserved
have an entepicondylar foramen. The supinator ridge is well developed in all
specimens (Plate 11). These features are characteristic of the extant Phocinae
(King, 1966), but in at least one extinct monachine, Monotherium, an entepicon-
dylar foramen was present (see Van Beneden, 1877). This is a characteristic of
particular significance in the interpretation of the relationships of Prionodelphis
capensis.

Because of the rather startling phocine appearance of the humerus of P.
capensis, it is appropriate to examine this bone in greater detail. Plate 12 shows
the right humeri of Monachus schauinslandi, P. capensis and Cystophora cristata.
The humerus of Cystophora was selected because it, of all phocine genera, most
resembles that of the fossil. The humerus of another phocine, Erignathus barbatus,
is shown with those of P. capensis and M. schauinslandi in Figure 2. It is imme-
diately evident that the P. capensis humerus exhibits not only phocine, but also
monachine characteristics.

Apart from the two characters already mentioned, there is also a conside-
rable difference between phocine and monachine humeri in the region of the
deltoid crest. In all pinnipeds the pectoralis muscle is prominent, and its inser-
tion on the humerus is strengthened. In the phocine seals this has been accom-
plished by an anteriorly directed enlargement of the medial edge of the deltoid
crest toward the enlarged lesser tubercle, so that the intertubercular groove
becomes circular in cross-section, coming to within 40° of completely encircling
the bicipital tendon in some species. In the otarioid seals, the pectoral insertion
is similarly strengthened in this area, and the intertubercular groove becomes
trenchant, although the lesser tubercle remains ‘lesser’. Strengthening of the
pectorial insertion on the phocine humerus does not take place by extending the
insertional area distally along the shaft of the humerus but, rather, the insertional
area terminates abruptly at a strong process on the distal end of the deltoid
crest. Beyond this point the anterior margin of the phocine humerus shaft is
concave as it curves to meet the distal articulation, and is devoid of muscle

scars.
The transverse development of the deltoid crest of phocine humeri is also

in evidence laterally, where a lip of bone overhangs the area of insertion of the
deltoid muscle.

In the monachine seals anterior enlargement of the deltoid crest is minimal,
the intertubercular groove remains widely open, and there is no overhanging of
bone on the lateral edge of the crest. The pectoralis insertion on the humerus is
strengthened by extending its area distally down the shaft toward the radial
fossa (much reduced in monachine seals) and the distal articulation. The ante-
rior margin of the monachine humerus is, therefore, straight or even convex,
and muscle scars are prominent where the deltoid crest blends distally into the
shaft. A similar elongation of the pectoral insertion is also present in the otarioid

86 ANNALS OF THE SOUTH AFRICAN MUSEUM

[ 2 3

Fic. 2. Anterior and medial views of the right humeri of Monachus schauinslandi (1), Prionodelphis
capensis (2), and Erignathus barbatus (3). (D — deltoid crest; S — supinator ridge.)

A PLIOCENE PHOCID FROM SOUTH AFRICA 87

seals, consistent with the greater development of the pectoral muscles in that
group.

In the P. capensis humerus, the bicipital or intertubercular groove is widely
open, the deltoid crest blends smoothly into the distal part of the shaft, and the
muscle scars are prominent on the shaft below the deltoid crest. In these respects
it is typically monachine. It does, however, have the lateral overhanging of the
deltoid crest similar to that of the phocines. The deltoid crest is as a result, more
prominently developed than in modern monachines.

It is concluded that the pattern of strengthening of the pectoral muscle
insertion is a more useful character in classifying phocids than the two characters
given by King (1966), which apply to modern species only. In the P. capensis
humerus the loss of the entepicondylar foramen, a reduction in the size of the
supinator ridge and reduction of the lateral development of the deltoid crest,
would reduce it to an almost exact replica of that of M. schauinslandi.

The humerus of P. capensis is stoutly proportioned, and the most complete
specimen known (L 4638) has a total length of 138 mm between the head and
median condyle. The transverse diameter of the distal end is 52,7 mm (mean of
8 specimens). It is stouter than that of living Monachus and Leptonychotes, and some
fossil monachines such as Pliophoca etrusca and Monotherium aberratum; comparable
in stoutness to the humerus of living Hydrurga and fossil Palaeophoca nystiz; it is
less stout than that of living Lobodon and Ommatophoca. Relatively shorter and
stouter humeri suggest more pelagic adaptations in living phocids.

Ulna

According to King (1966:390) there are no consistent differences between
the ulnae of phocines and monachines. In our sample of ulnae from living
species there is a suggestion that the tuberosity for insertion of the internal
anconeal muscle (Howell, 1929:75) is much more produced and somewhat
more posterior in location in the phocine seals. In the monachine seals, if any
anconeal tuberosity can be said to exist, it is continous with the triceps insertion
at the anterodorsal apex of the olecranon (ulnar orientation is here considered
to be with the long axis vertical, as in fissiped carnivores). There also appears to
be slight but persistent sigmoid flexure of the phocine ulnar shaft, in anterior or
humeral aspect, because of a lateral curve distal to the radial notch, whereas
the shaft of modern monachine seals is straight distal to the radial notch.

Of these two suggestive characters, the nature of the anconeal insertion is
not preserved on the ulnae of P. capensis, but there appears to be a slight phocine
outward curvature of the shaft in the vicinity of the interosseous crest, distal to
the radial notch. It thus seems possible that the straight shaft of the living mona-
chine seals has been recently acquired.

Except for the slight curvature of the shaft, the ulna of P. capensis (Plate 13)
greatly resembles that of M. schauinsland:. The proportions are nearly equal and
both are characterized by an extremely elongated posterior process of the
olecranon, giving the bone a very hatchet-like appearance.

88 ANNALS OF THE SOUTH AFRICAN MUSEUM

There is considerable generic variation, and no subfamilial differentiation,
in the configuation and relative location of the humeral and proximal radial
articulation on the phocid ulna. In these articular facets P. capensis also resem-
bles living Monachus. The radial facet has minimal medial offset and distal
separation from the facet for the humerus, and faces anteriorly rather than
anterolaterally as in other phocids. The humeral facet has a distinct medial
curvature, as does that of many other phocids. It also seems possible, therefore,
that the lateral positioning of the head of the radius, as reflected in its articula-
tion with the ulna, may be a rather recent development, at least in the mona-
chine seals.

The total length of the ulna of P. capensis, estimated from two incomplete
specimens, is about 170 mm.

Radius

Reflecting the more anterior orientation of the radial articulation of the
ulna, the radius of Prionodelphis capensis (Plate 14) was orientated more anteriorly
from the ulna than is that of some living monachines and apparently all living
phocines. As a result, the radial tuberosity lies distinctly on the medial side of the
radius, as on the radius of living Monachus, and not on the posteromedial surface
as is the case of the radius of living phocines.

King (1969: fig. 31) has pointed out that the distal articulations of the radii
of Hydrurga and Ommatophoca have convex surfaces that curve on to the medial
(flexor) side of the radius at or near the anterior (preaxial) limit of the articular
surface. The radius of Halichoerus has this medially curving segment of the arti-
culation about midway between the anterior and posterior limits of the arti-
cular surface. The condition in the latter produces a moderate indentation in
the medial margin of the articular facet when viewed distally.

The pattern of distal articulation on the radius of Halichoerus seems to be a
characteristic of all extant Phocinae. The radii of living monachine seals
follow the patterns shown by King (1969) for Ommatophoca and Hydrurga. In
Ommatophoca and Mirounga the distal articular surface is roughly rectangular,
and the anteromedial quarter curves on to the medial surface. In Hydrurga,
Lobodon and Monachus the anteromedial corner of the articulation is extended so
that the part of the surface that curves on to the medial side of the radius is
almost a separate articulation (see King, 1969: fig. 31b).

In Prionodelphis capensis the distal articulation of the radius (L 2935) most
resembles that of Mirounga. In general configuration the radius is markedly
spatulate with a prominent anterior crest for insertion of supinator and pronator
teres muscles, most closely resembling in this respect, Hydrurga and Ommatophoca
of the living monachines.

The total length of the radius of P. capensis, estimated from two incomplete
specimens, is about 145 mm with the greatest anteroposterior diameter of the
shaft being 51,4 mm. The dimensions of the proximal end are 30,8 by 23,4 mm
(mean of seven specimens).

A PLIOCENE PHOCID FROM SOUTH AFRICA 89

Posterior Limb
Innominate

The most complete innominate known (L 15849A) lacks the most anterior
part of the ilium, and the posterior parts of the ischium and pubis (Plate 15).
There is, however, sufficient of this bone remaining to enable a fairly confident
assessment of its characteristics.

The ilium is weakly everted, approximately to the extent of that of Erig-
nathus and most monachine seals. Erignathus is atypical of the Phocinae, the
remainder of which have a strongly everted ilium ‘with a deep lateral exca-
vation’ (King, 1969:392) ; the latter character is absent in P. capensis.

Comparison of the post-acetabular proportions of the entire innominate is
precluded by the incompleteness of the specimens. However, a comparison of
the distance between the centre of the acetabulum and the apex of the ischiatic
spine, to the width of the obturator foramen ventral to the ischiatic spine (not
always maximum width), results in an equally distinct separation between
monachine and phocine seals (Table 5). Interestingly, measurements of photo-
graphs and drawings published by King (1956, 1966, 1969), also conform quite
well.

TABLE 5. Innominate proportions of some phocids.

O = Width of
A = Acetabulum obturator foramen
center to tip of ventral to ischiatic

Species ischiatic spine spine O/A xX 100
Monachusm.. ‘ ! : 37,1 18,3 49,3
Prionodelphis capensis . «Sti 90,8 39,8 43,8
Hydrurga l. . : : . : 132,8 57,8 4355
Hydrurga 1.? . 4 oat) ee 25,6 9,9 38,7
Leptonychotes w.? , : ; 24,0 955 30,6
Leptonychotes w.1 BY MIET “te 4454 15,9 35,8
Monachus s. : i : ‘ 95,6 36,5 38,2
Ommatophoca r.2 . ‘ ; : 22,5 8,2 36,4
CT ae so 95,0 3354 35,2
Cystophora c. or ee, OL ee 132,8 4454 33,4
Pagophilus g. ’ Pad 93,8 30,3 32,3
Halichoerus g.2_. ; ; ‘ 23.3 6,9 20,6
Halichoerus g+ . a 34,9 9,4 26,9
Phocav. . ‘ A ae ie 97;9 28,6 29,2
Eriguathas bee Wk 141,2 39,1 27,7
1D) ACO i 28,5 Tif 27,0

1 From photo in King, 1956.
2 From drawing in either King, 1966, or King, 1960.

As with the living Antarctic monachines (but not Monachus schauinsland.)
and some phocine seals, the innominate of P. capensis appears rather thick
across the acetabulum. It appears similarly thick in available specimens of
Pagophilus and Cystophora. ‘This appearance is caused by a relatively small
acetabulum.

gO0 ANNALS OF THE SOUTH AFRICAN MUSEUM

Femur

Only the distal end of the femur of P. capensis is known (L 10131) (Plate
16A). The patellar facet is somewhat broader than tall, as in the monachine
seals, while a fairly marked pit for the popliteus muscle on the lateral epicondyle
shows resemblance to the phocine seals.

Tibia

The tibia of P. capensis is remarkable for the development of pronounced
fossae on the posterior and antero-lateral surfaces. In the specimen L 10128/9
(Plate 16B, C, D) the thickness of bone between these two fossae is as low as
0,75 mm. This condition most resembles that in the tibia of Halichoerus. King
(1966) states that the post-tibial fossa is more pronounced in the Phocinae than in
the Monachinae.

Distally, the tibia of P. capensis is conspicuously broad and anteroposteriorly
flattened (Plate 16E, F), a condition very similar to Pliophoca etrusca (Tavani,
1942: fig. 18). The fibular contact is sharply angled outward suggesting that the
fibula was rather markedly bowed.

Pes

A calcaneum (L 10118), two astragali (L 10130, L 10993), one navicular
(L 15851), one entocuneiform (L 10124), one metatarsal V (L 10996) and two
phalanges (L 10999, L 10205) of Prionodelphis capensis are known (Plates 17 & 18).
At first glance it seems obvious that P. capensis has long metatarsal bones
relative to the size of the astragalus, and in fact, a sampling of the relative sizes
of these two bones in seven living genera seems to bear this out (Table 6). The
Langebaanweg seal appears to have relatively longer metatarsals than those of
the compared living genera except Monachus.

TABLE 6. Tarsal-metatarsal comparison of some phocids.

(1) (2) (3) (2/1) (3/1)
Greatest length
Greatest length Greatest length of cuboid-MT

Genus of MT V of astragalus IV facet

Phoca . Ze 70,7 mm 59,7 mm 28,2 mm 0,84 0,398
Lovodon * a *. 98,2 75,0 3552 0,76 0,358
Cystophora . 94,6 69,9 32,4, 0,74 0,342
Mirounga . . 129,5 90,5 47,8 0,70 (0,369)
Hydrurga . ; 120,2 82,0 40,8 0,68 0,339
Pagophilus . . 85,6 5737 25,9 0,67 (0,302)
P. capensis . 89,5 575 30,4 0,64 0,339
Monachus .. 98,0 56,3 30,2 0,57 0,308

As the fossil astragali and metatarsal V were not found in association,
they could represent different sized individuals, creating a false impression of
relative metatarsal size. To check this possible error a similar comparison of the
cuboid-metatarsal IV articular facet on metatarsal V, to the total length of the

A PLIOCENE PHOCID FROM SOUTH AFRICA gI

the metatarsal was also made (Table 6). Comparable results were achieved,
except for Mirounga and Pagophilus whose relative proportions were not con-
sistent. It therefore appears probable that P. capensis did indeed have relatively
elongated metatarsal bones and hence had relatively large hind flippers.

King (1966:393-394) has suggested that the more distal articular facet
between the astragalus and the calcaneum is relatively long in the phocine
seals and short in the monachines. Although there is no exception in the mona-
chine specimens available, there seems to be great variation in the form of this
facet on the astragalus of the phocine seals. Our specimen of Cystophora appears
decidedly ‘monachine’ in this character: astragali of Erignathus appear to vary
from distinctly ‘monachine’ to distinctly ‘phocine’. This articular surface on P.
capensis is distinctly elongated and hence ‘phocine’ to the extent that the charac-
ter is valid.

In all respects other than their mutual lower articulation, the astragalus
and calcaneum of P. capensis are extremely similar to those of M. schauinslandi.
If the tibial articulation of the astragalus is arbitrarily taken to be dorsal, so
that the fibular articulation is vertical, these surfaces are low relative to the body
of the astragalus and the fibular articulation extends about to the most ventral
limit of the bone. Among living monachines a similar condition is found in both
Monachus and Mairounga, which differ greatly in this respect from HAydrurga.
Wide variation is found also in the phocine seals: the astragalus of Pagophilus
is perhaps most similar to that of P. capensis, while that of Phoca differs the most.

The tibial articulation is cylindrical, as in Mirounga, Monachus and Lobo-
don, rather than spherical as in Hydrurga. No angular boundary separates the
distal articulation for the navicular from the adjacent articulation for the
calcaneum. The astragalus is rather short-necked and has a short calcanear
process.

The dimensions of the calcaneum L1o0118 (Table 7) are remarkably
similar to that of a M. schauinslandt specimen recorded by Robinette & Stains
(1970: table 1), while its porportions are clearly monachine rather than phocine
(Robinette & Stains, 1970: table 2). The metrical data presented in Table 7
confirms the observations on the similarities between the calcanea of P. capensis,
Monachus (especially M. schauinslandi and M. tropicalis) and Mirounga, and also
illustrates differences from those of Hydrurga, Leptonychotes and Ommatophoca.

The posterior! astragalar articulation of the calcaneum of P. capensis is
narrow relative to its length. This contrasts with that facet of the calcaneum of
Hydrurga and Monachus schauinslandi, in which it is nearly as wide as it is long.
Robinette and Stains (1970:535) state that the facet is narrower on the calca-
neum of M. tropicalis than on that of M. schauinslandi.

The facet of the navicular, for articulation with the entocuneiform, is
notably equidimensional and flat. The articular surface on the entocuneiform,
for contact with metatarsal I, is notably elongated, suggesting a more slotted

1 As used by Robinette & Stains (1970: fig. 1); this is the anterior articulation of King (1966:
393):

Q2 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 7. Dimensions of calcaneum of Prionodelphis capensis from Langebaanweg, compared with
those of some modern monachines.

Species N TL W DVH W/TL DVH/TL DVH/W —
Monachus monachus t ‘ I 63,2 29,8 30,0 AG 47 IOI
Mirounga angusti-

FastEsp OS OT OS BG 790% 40,9* 38,7* = 52 49 95
Monachus

schauinslandit : I 59,1 31,9 30,3 54 51 95
Monachus

tropicalis t I 5459 27,3 28,5 50 51 104
Prionodelphis capensis I 58,4 30,7 30,4. 53 52 99
Hydrurga leptonyxt . I 7455 38,7 42,0 52 56 108
Leptonychotes weddellit I 7155 38,2 41,9 53 58 110
Ommatophoca rossit I 56,0 30,4 34,6 54 61 114

+ From Robinette & Stains, 1970. . * Average figures.

TL = total length. W = width.

DVH = dorsoventral height.

proximal articulation rather than the basined articulation on metatarsal I of
most living seals. Although both the navicular and the entocuneiform appear
large relative to the known astragali and calcaneum, they appear to be un-
diagnostic of subfamily affinities.

As has been mentioned, metatarsal V appears to be relatively elongated.
The two known phalanges also appear to be conspicuously elongated and
slender when compared with those of living phocids. Otherwise they seem to
have no distinctive features.

Discussion

In assigning the Langebaanweg phocid to the genus Prionodelphis, it is
recognized that reassessment may be required when more material of P.
roverett is found. Generic identity is based upon the remarkable similarity of the
few fragments from Argentina to the South African material and on the belief
that the lack of greater knowledge is a stronger argument against the establish-
ment of a new genus than it is against tentative assignment to the same genus.

The similarity of the cheek teeth of P. rovereti and P. capensis is very strong.
The transverse narrowness of the cheek-teeth, more evident in the latter species,
and a posterointernal shelf on the upper cingula only, are features found only in
some of the Antarctic genera of the monachine seals. As mentioned, the tooth
proportions and cingular shelf are most similar to the condition in Hydrurga.
In addition, the greatly reduced and distinctly recurved last upper postcanine,
of both species of Prionodelphis is singularly suggestive of a close relationship, and
is not known in other monachine seals, although there is some resemblance to
the more anterior cheek teeth of Leptonychotes, and to a lesser extent also Omma-
tophoca.
Apart from differences in the dimensions of the teeth of P. capensis and P.
rovereti, some differences in morphology are also evident. Other than the upper

A PLIOCENE PHOCID FROM SOUTH AFRICA 93

fifth, the only known upper postcanine of P. rovereti (Frenguelli, 1922: fig. 1),
differs from the second to fourth upper postcanines of P. capensis in having a
far larger posterior root, which is at least partially divided longitudinally. ‘This
feature of the Argentinian species, as well as its slightly broader cheek-teeth,
can be interpreted as being less advanced characteristics. The anterior part of
the crown of this tooth of P. rovereti lacks’a distinct accessory cusp, which is
present in all the known postcanines of P. capensis. The anterior part of the P.
roverett tooth is markedly convex, and the lingual view (Frenguelli, 1922: fig. 1c)
shows a small step more or less where an accessory cusp might be expected. It is
possible, therefore, that this specimen has the anterior accessory cusp masked by
some individual variation, and that normally such a cusp was present. The
lower postcanines of P. rovereti (Frenguelli, 1922: fig. 2B, C) are illustrated as
having not one, but two anterior accessory cusps, with the anterior and posterior
parts of the teeth being almost mirror images of one another. In none of the
postcanines of P. capensis is a second anterior accessory cusp known, and its
development in the Argentinian species could be a more advanced specializa-
tion.

The age of the Argentinian species is even more uncertain than that from
Langebaanweg. The Entre Rios deposits, from which the specimens of P.
roverett came, appear to be of Pliocene age (Langston, 1965: table 3) and would
seem, therefore, to be roughly the same age as those from Langebaanweg.
However, most often these deposits have been referred to as being late Miocene
or early Pliocene, and the latter age is given by Romer (1966).

Discussion on the relationship of P. capensis to P. rovereti will be more mean-
ingful when more specimens of the latter are known, but the observed differ-
ences between the two sets of specimens, the possibility of a temporal difference
in the deposits from which they come, and their geographical separation sug-
gest a distinction between the South African and Argentinian fossils at least at
the species level.

The relatively poor fossil record of the Phocidae in general renders inter-
pretation of the wider relationships of P. capensis equally problematical. Some
features of the fossil seal from South Africa are found, among the living seals,
only in the Phocinae. However, a number of features are clearly monachine and
these suggest that the dichotomy from the primitive phocid into the two extant
subfamilies was a result of two distinctly different patterns of specialization to
better adapt to pelagic existence. These adaptations relate to, amongst other
things, greater swimming ability and underwater hearing. As has been pointed
out throughout the description, in all of these adaptations P. capensis has clearly
followed the monachine pattern.

To judge from the comparisons between P. capensis and the living phocine
and monachine seals, differences in adaptation toward greater swimming
ability appear most evident in the proximal limb elements. Subfamilial dif-
ferences in the structure of the humerus related to the strengthening of the
pectoralis muscle have been outlined in the discussion of this bone. In the Phoci-

94 ANNALS OF THE SOUTH AFRICAN MUSEUM

nae the pectoralis insertion has been strengthened by exaggeration of the deltoid
crest, and in the Monachinae strengthening of this same muscle has been accom-
panied by a distally extended insertional area on the humerus shaft.

One might infer that the phocid ancestral to the living subfamilies had a
humerus of relatively slender proportions showing a moderate development of
both types of pectoralis insertion, such as seen on the humerus of ‘Phoca’ vindo-
bonensis Toula (1897: pl. 1, fig. 16) or Leptophoca lenis ‘True (1906: pl. 75). Such
fossil phocids as Monotherium aberratum Van Beneden (1877: pl. 17, figs. 1-4)
appear to have the insertional area extended so far distally on the shaft that a
monachine condition seems indisputable, while others such as Phocanella
pumila Van Beneden (1877: pl. 14, figs. 1-4), have clearly evolved the phocine
condition by strengthening the deltoid crest and eliminating all pectoralis
insertion on the shaft distal to the crest. In addition, the presence of an entepi-
condylar foramen on the humerus appears to be a primitive feature. It is present
in all of these fossil seals, including Prionodelphis capensis, and is retained in the
living phocine seals as well. Only at the stage of evolution evident in the living
phocids does the presence of this foramen become diagnostic of subfamily
affinity.

Consideration of the subfamilial differences in swimming adaptations
which might be found in the pelvic limbs has been hampered in this study by the
lack of a complete specimen of the femur of P. capensis. Nevertheless, King
(1966:392) has pointed out that, except for the genus Erignathus, the phocine
seals may be recognized by their extremely everted ilium. Both leverage and
strength of the insertion of the massive muscles of the back, the iliocostalis
system, are benefited by this structure, as well as are most of the gluteus group
which directly transfers the forces of the back to the femur to produce the charac-
teristic phocid swimming motion. The advantages of this structure seem so
obvious that it is puzzling why none of the monachine seals have developed it,
or why it developed so late in the history of the phocid seals. Few fossil seals in
which the pelvis is known, exhibit the phocine everted ilium.

The interpretation of the functions of osteological characters in the phocine
and monachine ear regions is subjective, but according to one interpretation
(Repenning, in press), two of these differences relate to improved underwater
hearing.

The presence of a more or less horizontal crest on the external surface of the
mastoid bone in all phocine seals is correlated with a greater directional
selectivity of sounds in water originating above or below the head; this crest is
not present in the monachine seals, including P. capensis.

The development of an enlarged petrosal apex in all seals is correlated with
a greater sensitivity to sound in water, and this development is conspicuously
less in P. capensis than in the living Antarctic seals. The enlargement of the petro-
sal apex is in the form of a globular mass in the phocine seals, while in the
monachine seals, with the exception of Mirounga, the apex is enlarged as a
rather low and broad structure. In this respect P. capensis is clearly monachine.

A PLIOCENE PHOCID FROM SOUTH AFRICA 95

It should also be noted that broadening is slight in Hydrurga, and that enlarge-
ment of the apex is partly accomplished by thickening; the structure does not
appear globular as in the phocine seals, however. Furthermore, enlargement of a
petrosal apex is minimal in Monachus, less than in P. capensis; in this respect
Monachus might be expected to be most similar to the ancestral phocid from
which the extant subfamilies evolved.

Monachus is the least specialized of the living monachine seals in the enlarge-
ment of the petrosal apex, strengthening of the humerus, distal broadening of
the radius, enlargement of the ilium, strengthening of the femur and modifi-
cation of the dentition. Except for the condition of the femur, which is unknown
in the South African fossil, and possibly the enlargement of the ilium, which is
incompletely preserved, Monachus is also less specialized in these features than
P. capensis.

From the preceding consideration of Prionodelphis capensis and related seals,
the following features appear most likely to be those that would characterize the
ancestral protophocid from which the two modern subfamilies, the Phocinae
and the Monachinae, evolved: dentition with primary cusp flanked by one
accessory cusp anteriorly and one posteriorly, much the same as seen in Praepusa
pannonica Kretzoi (1941: fig. 1), ear region much as in living Monachus, and
postcranial skeleton unspecialized as in ‘Phoca’ vindobonensis 'Toula (1897).

The relationship of P. capensis to the Antarctic monachines is evident in a
broad sense, but it is not clearly ancestral to any of the four living genera. The
highly modified dentitions of the living genera differentiate them most strikingly
from the Pliocene fossil. ‘The great reduction of the last upper postcanine tooth
of P. capensis seems to preclude the possibility of it being ancestral to Hydrurga or
Lobodon, while the great simplification of the teeth of Leptonychotes and Ommatop-
hoca leave little basis for interpretation. Hydrurga, Lobodon and Ommatophoca are
all clearly better adapted to pelagic life in their postcranial specializations than
was P. capensis. All of the Antarctic seals, and Lepionychotes in particular, have a
greater development of the petrosal apex than does P. capensis, which indicates
that the latter had less acute hearing underwater. In all these respects P.
capensis is less advanced than the Antarctic seals, but more advanced than
Monachus.

It seems probable that Prionodelphis, presently known only by the species P.
rovereti and P. capensis, was not the only Pliocene monachine of the southern
seas, and that some or all of the modern Antarctic genera derive from a related
but unknown form.

All previously described fossil phocids are known from incomplete remains
and a good many from a very few, or even one bone. The humeri and femora
are the most commonly described because they are among the more durable
bones of the body, and, presumably, their size lends them to discovery. Excluding
mandibular fragments, Pliophoca etrusca Tavani (1942; a skull), Phoca pontica
Eichwald (1853; a cranium) and Phoca pontica Alekseev (1924; a rostrum) are
the only fossil phocids of which the skulls are even partially known. Despite its

96 ANNALS OF THE SOUTH AFRICAN MUSEUM

fragmentary nature, Prionodelphis capensis is one of the most completely known
fossil phocids, and since the systematic investigation of the Langebaanweg
deposits is still in its early stages, it can be expected that much more material
will become available for study in the future.

SUMMARY

Pinniped remains from the late Pliocene deposits at Langebaanweg in
South Africa are described. The material is referred to Prionodelphis capensis n.
sp. (family Phocidae, subfamily Monachinae). On the basis of this material, the
genus Prionodelphis Frenguelli 1922 is defined. ‘The relationships of the Lange-
baanweg species to extant and fossil monachines are discussed, and morphologi-
cal characters, by which fossil Phocinae and Monachinae can be differentiated
are suggested.

ACKNOWLEDGEMENTS

We wish to thank A. W. Mansfield and D. E. Sergeant of the Fisheries
Research Board of Ganada; C. W. Mack of the Harvard University Museum of
Comparative Zoology; O. P. Pearson and R. L. Jones of the University of
California Museum of Vertebrate Zoology; R. H. Mansville, K. W. Kenyon,
and E. Kridler of the (U.S.) Bureau of Sport Fisheries and Wildlife; C. O.
Handley, Jr. and C. E. Ray of the (U.S.) National Museum of Natural History;
E. H. Bryan, Jr. of the Bernice P. Bishop Museum (Honolulu) ; F. H. Fay of the
(U.S.) Arctic Health Research Center; and L. Giannelli of the Universita di
Pisa Museo di Paleontologia for the loan or gift of material used in this study.

The current investigations at Langebaanweg are being supported by the
South African Council for Scientific and Industrial Research, Chemfos Ltd. (a
subsidiary of the African Metals Corporation) and Shell South Africa (Pty.) Ltd.
The Wenner-Gren Foundation for Anthropological Research, New York,
provided the vehicle used in the field work at Langebaanweg (Grant no.
2752-1834).

We are indebted to Mr. H. Krumm and Mr. G. Benfield of Langebaanweg,
whose co-operation with the South African Museum led to the recovery of most
of the material described in this paper.

Francis H. Fay and Clayton E. Ray kindly offered comments on the manu-
script of this paper, and we are grateful to them for their observations and
criticisms.

REFERENCES

ALEKSEEV, A. K. 1924. [Seals in the Sarmatian deposits of southern Russia.] <h. nauchno-issled.
Kafedr Odesse 1: 26-34. [In Russian. ]

ALLEN, J. A. 1880. History of North American pinnipeds: a monograph of the walruses, sea-
lions, sea-bears and seals of North America. Misc. Publs U.S. geol. geogr. Surv. Territ. 12:
i-xvi, 1-785.

Bonk, E. L. & Sincer, R. 1965. Hipparion from Langebaanweg, Cape Province, and a revision
of the genus in Africa. Ann. S. Afr. Mus. 48: 273-397.

A PLIOCENE PHOCID FROM SOUTH AFRICA 97

Brookes, J. 1828. A catalogue of the anatomical and zoological museum of Joshua Brookes. London.

Burns, J. J. & Fay, F. H. 1970. Comparative morphology of the skull of the Ribbon seal, His-
triophoca fasciata, with remarks on systematics of Phocidae. 7. Zool. 161: 363-394.

CasBrERA, A. 1926. Cetaceos fésiles del Museo de La Plata. Revta Mus. La Plata. 29: 363-411.

EICHWALD, C. E. von. 1853. Lethaea Rossica, ou Paleontologie de la Russie. 3: Stuttgart: Schweizer-
bart.

FRENGUELLI, J. 1922. Prionodelphis rovereti un representante de la familia ‘Squalodontidae’ en al
Paranense Superior de Entre Rios. Boln Acad. nac. Cienc. Cérdoba 25: 491-500.

FRENGUELLI, J. 1926. El Entrerriense de Golfo Nuevo en el Chubut. Boln Acad. nac. Cienc. Cordoba
29: 191-270.

Gray, J. E. 1825. An outline of an attempt at the disposition of Mammalia into tribes and
families, with a list of the genera apparently appertaining to each tribe. Ann. Philosophy (n.s.)
10: 337-344-

Hay, O. P. 1930. Pinnipedia. In Hay, o. P. Second bibliography and catalogue of the fossil Vertebrata
of North America. 2: 555-505. Washington: Carnegie Institution. (Publs Carnegie Instn 390.)

HENDEY, Q.B. 1970a. A review of the geology and palaeontology of the Plio/Pleistocene deposits
at Langebaanweg, Cape Province. Ann. S. Afr. Mus. 56: 75-117.

HeEnpbey, Q.B. 1970). The age of the fossiliferous deposits at Langebaanweg, Cape Province.
Ann. S. Afr. Mus. 56: 119-131.

Howe 1, A. B. 1929. Contribution to the comparative anatomy of the eared and earless seals
(genera <alophus and Phoca). Proc. U.S. natin. Mus. 73: 1-142.

ILuicER, C. 1811. Prodromus systematis mammalium et avium. Berlin: Salfeld.

Ke.LtLoce, R. 1922. Pinnipeds from Miocene and Pleistocene deposits of California. Univ. Calif.
Publs geol. Sct. 13: 23-132.

KeEttoce, R. 1942. Tertiary, Quaternary, and Recent marine mammals of South America and
the West Indies. Proc. 8th Am. Sci. Congr. (Washington) 3: 445-473.

Kine, J. E. 1956. The monk seals (genus Monachus). Bull. Br. Mus. nat. Hist. (Zool.) 3: 201-256.

Kine, J. E. 1964. Seals of the world. London: British Museum (Natural History).

Kine, J. E. 1966. Relationships of the Hooded and Elephant seals (genera Cystophora and Miroun-
ga). F. Kool. 148: 385-398.

Kine, J. E. 1969. Some aspects of the anatomy of the Ross seal, Ommatophoca rossi (Pinnipedia:
Phocidae). Scient. Rep. Br. Antarct. Surv. 63: 1-54.

KRAGLIEVICH, L. 1934. La antigiiedad pliocena de las faunas de Monte Hermoso y Chapadmalal.
Siglo ilustrado, Montevideo 938: 1-136.

Kretzot1, M. 1941. Seehund-Reste aus dem Sarmat von Erd bei Budapest. Féldt. Kézl. 71: 350-
356.

Lancston, W. 1965. Fossil crocodilians from Colombia and the Cenozoic history of the Croco-
dilia in South America. Univ. Calif. Publs geol. Sci. 52% i-vii, 1-157.

Mactio, V. J. & HEeNnpEy, Q.B. 1970. New evidence relating to the supposed stegolophodont
ancestry of the Elephantidae. S. Afr. archaeol. Bull. 25: 85-87.

Mckenna, M. 1969. The origin and early differentiation of therian mammals. Ann. N.Y. Acad.
Sci. 167: 217-240.

REPENNING, C. A. Underwater hearing in seals: functional morphology. Jn Functional anatomy
of marine mammals. London: Academic Press. (In press.)

RosineTTE, H. R. & Strains, H. J. 1970. Comparative study of the calcanea of the Pinnipedia.
J. Mammal. 51: 527-541.

Romer, A. S. 1966. Vertebrate paleontology. 3rd ed. Chicago: University Press.

SavacE, D. E. & Curtis, G. H. 1970. The Villafranchian stage-age and its radiometric dating.
Spec. Pap. geol. Soc. Am. 1242 207-231.

SCHEFFER, V. B. 1958. Seals, sea lions and walruses. Stanford: University Press.

SincER, R. & Hooyer, D. A. 1958. A Stegolophodon from South Africa. Nature, Lond. 182: 101-102.

Smirnov, N. A. 1908. [Review of Russian pinnipeds.] Mém. Acad. Sci. St. Petersb. (8) 23: 1-75.
[In Russian].

Tavani, G. 1942. Revisione dei resti del pinnipede conservato nel Museo di Geologia di Pisa.
Palaeontogr. ital. 40: 97-113.

Tavani, G. 1943. Revisione dei resti di pinnipedi conservati nel Museo geo-paleontologico di
Firenze. Atti. Soc. tosc. Sci. nat. Memorie 61: 34-42.

Touta, F. 1897. Phoca vindobonensis n. sp., von Nussdorf in Wien. Beitr. Paldont. Geol. Ost.-Ung.

IX: 47-70.

98 ANNALS OF THE SOUTH AFRICAN MUSEUM

TROUESSART, E.-L. 1897. Catalogus mammalium tam viventium quam fossilium. Berlin: Friedlander.

True, F. W. 1906. Description of a new genus and species of fossil seal from the Miocene of
Maryland. Proc. U.S. natn. Mus. 30: 835-840.

Uco int, R. 1902. I] Monachus albiventer Bodd. del Pliocene di Orciano. Palaeontogr. ital. 8: 1-20.

VaN BENEDEN, P. J. 1877. Description des ossements fossiles des environs d’Anvers. Annls Mus.
r. Hist. nat. Belg. 1: 1-88.

Ann. S. Afr. Mus., Vol. 59

Dorsal view of skull L 15695. Scale represents 5 cm.

Ann. S. Afr. Mus., Vol. 59 Plate 3

Lateral view of skull L 15695. Scale represents 5 cm.

SENS

E
3)
Yo)
nN
te
i=)
8)
n
oO
i
jor
©
a
©
e—
i)
S)
N
Te)
Dp
©
Yo)
| |
a
=
=)
4
n
et
(eo)
Fs
©
oil os!
>
—
8
i
~~
t=
©
>

Ann. S. Afr. Mus., Vol. 59

Right lateral view of rostrum L 12695.

Ann. S. Afr. Mus., Vol. 59 Plate 6

A, B, G & D. Dorsal, ventral, lateral and internal views of temporal L 15652.

Plate 7

Ann. S. Afr. Mus., Vol. 59

‘poonpes Apysiys aeag

“BSSOF VI[YIOIO[NGNSOA — "fF 9A SUIs1eU [esourenbs — ‘urs ‘xode [esomod— ved {yeuea
[e1owy — “O'F + BSsOF V[Joqor90 — *J-o Syeu¥s proaed — ‘9°90 $;onpanbe varyvoo — -e-o : SUOTIEIADIQGY

-(uojgaT pue (q) sisuadva siyqjapouowg ‘(w)

"(D) sappam sazoys
ipuvjsuimoyas snysvuopy JO sjeroduts, 9y) JO sMaTA [esiog

d

2)

)

Ann. S. Afr. Mus., Vol. 59 Plate 8

A, B & C. Lateral, internal and dorsal views of mandible L 7556.

Ann. S. Afr. Mus., Vol. 59 Plate 9

A. Lateral view of upper incisor L 11689.

B. Lateral view of upper canine L 11686 (aged individual).

C. Buccal and occlusal views of first upper postcanine L 11687.

D. Buccal and occlusal views of second, third of fourth upper postcanine L 15664.
E. Lingual and occlusal views of fifth upper postcanine L 15429.

F. Lateral view of lower canine L 13152 (young individual).

G.

Buccal and occlusal view of lower second, third, fourth or fifth postcanine L 15444B.

Ann. §, Afr. Mus., Vol. 59 Plate 10

. Lateral view of scapula L 2160.
. Lateral and ventral views of second cervical vertebra L 7563.

A
B
C. Ventral view ofa second or third lumbar vertebra L 15849 Ag.
D. Ventral view of fifth lumbar vertebra L 15849 At.

E. Anterior view of third sacral vertebra L 15396.

F. Anterior view of first caudal vertebra L 15857.

Plate 11

Ann. S. Afr. Mus., Vol. 59

i f humerus L 2157.

10r VIEWS O

Medial view of humerus L 4638.

A & B. Anterior and poster

C.

Ann. S. Afr. Mus., Vol. 59 Plate 12

Anterior views of the right humeri of Monachus schauinslandi (A), Prionodelphis capensis (B) and
Cystophora cristata (C). Scale approximately x 4.

Ann. S. Afr. Mus., Vol. 59

Lateral view of left ulna L 15682.
Medial view of right ulna L 2:61.

A.
B.

Ann. S. Afr. Mus., Vol. 59 Plate 14

A. Lateral view of right radius L 12869.
B. View of distal articulation of right radius L 2935 (anterior parts of specimen are lost).

L 15849A

innominate

C=
a
Cet
o}
>
v
=
>
—
fas}
a
8)
~
S&
qe)
S
fa
rs
~
MH
e)
Q
~Q
a)
<

Ann. S. Afr. Mus., Vol. 59

Ann. S. Afr. Mus., Vol. 59

A. Distal view of femur L 10131.

B, C & D. Posterior, anterior and proximal views of right tibia L 10128/9.
E. Posterior view of left tibia L 2138.

F. View of distal articulation of right tibia L 10128/9.

Plate

16

Ann. S. Afr. Mus., Vol. 59 | Plate 17

A. Lateral, ventral and medial views of right calcaneum L 10118.
B. Dorsal and ventral views of left astragalus L 10993.

Plate 18

Ann. S. Afr. Mus., Vol. 59

Dorsal, lateral and ventral views of left metatarsal V, L 10996.

Dorsal view of phalanges L 10205 and L 10999.

A.
B.

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE, 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm X 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, 5, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

Bu.LoucH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les changes respiratoires des littorines.
Archs Kool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGIGAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 80.

a

a

cay

.

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 #Band
March 1972 Maart
Panes. 5 + Weel

THE EVOLUTION AND DISPERSAL OF THE
MONACHINAE (MAMMALIA: PINNIPEDIA)

By
Q. B. HENDEY

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

1, 2(1, 3, 5, 7-8), 3(1-2, 5, t-p-i.), 5(2, 5, 7-9);
6(i5. VtSpat); 7(1-3)s 8, 9(1-2), 10(I),
I1(1-2, 5, 7, t--p-i.), 21, 24(2), 27, 31(1-3), 33

Price of this part/Prys van hierdie deel
R1,40

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 05 4

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

THE EVOLUTION AND DISPERSAL OF THE MONACHINAE
(MAMMALIA: PINNIPEDIA)

By
Q.B. HENDEY

South African Museum, Cape Town
(With 2 figures)

[Ms. accepted 15 November 1971]

CONTENTS

PAGE
Introduction . : é : - : 99
The monk seals. 3 : 3 - 5 100
The Antarctic monachines . . F Shy. sOm
The elephant seals . : : - ok SOF
Discussion : ; F é 3 ety Os
Conclusion : a 5 Q uh 15!
Summary. ° 2 = 2 - BR ih 5
Acknowledgements . : ; : Lapa @ -
References rf : : ; : Pe eas ih

INTRODUCTION

The recent description of a late Pliocene monachine seal from Langebaan-
weg in South Africa (Hendey & Repenning 1972) included a discussion on
the possible relationships of this species (Prionodelphis capensis) to other Mona-
chinae. It is the purpose of the present paper to enlarge upon this topic, and
also to comment on the phylogeny and zoogeography of the Monachinae.

The origin and evolution of the Pinnipedia have been the subject of
numerous publications (e.g. McLaren 1960; Sarich 1969 etc.) and of all the
subfamilies of this order, the Monachinae is probably that with the poorest
fossil record, and consequently the one whose evolution is least well understood.

A number of fossil monachines ranging in age from Miocene to Pleistocene
have been recovered from localities in Europe, while there are Pleistocene
records of monk seals from the south-eastern United States (King 1964). The
Pliocene monachine from South Africa is only the second Southern Hemisphere
record of its kind, the other being the poorly known Prionodelphis rovereti
Frenguelli 1922 from the Pliocene of Argentina. There are later Pleistocene
and Holocene records of Monachinae from the southern continents, but these
are all of species still extant, and are from areas within, or near to, the present
ranges of the species concerned. For example, remains of seals have been
found preserved by icy conditions in Antarctica, although none of these
specimens is more than a few thousand years old (Crane & Griffen 1968). There
are also late Pleistocene and Holocene records of Mirounga leonina from coastal
human occupation sites in South Africa, an area where occasional stray indivi-
duals of this species are still found.

99

Ann. S. Afr. Mus. 59 (5), 1972: 99-113, 2 figs.

100 ANNALS OF THE SOUTH AFRICAN MUSEUM

The sparse fossil record of the Monachinae in the southern continents is
curious, since modern representatives of this subfamily are found mainly in
the higher latitudes of the Southern Hemisphere. The only extant northern
monachines are the comparatively small and widely scattered populations of
monk seals (Monachus monachus, M. tropicalis and M. schauinsland:) and the
northern elephant seal (Mirounga angustirostris).

Modern monachine distribution is one of the more remarkable features
of this subfamily, and in speculating on phyletic relationships within the group
it is convenient to consider the possibilities under three separate headings.

THe Monk SEALS

Modern monk seals occur in the Mediterranean area and adjacent coast
of north-west Africa (Monachus monachus), the Caribbean area (M. tropicalis,
which may now be extinct), and in the Leeward chain of islands north-west
of Hawaii (M. schauinslandi). On the basis of the admittedly poor fossil record,
it seems probable that this group, and the subfamily as a whole, originated
from a generalized phocid in the western Europe and Mediterranean areas.

The earliest record of a monachine seal anywhere is Monotherium aberratum
Van Beneden 1877 from the late Miocene of Belgium. As far as can be seen
from the fragmentary remains known, this species is little different from other
generalized phocids of the Miocene. In the absence of any other monachines
of comparable age, with the possible exception of M. maeoticum Nordmann
1860 from southern Russia, it can conveniently be regarded as ancestral to the
rest of the subfamily. Indeed, there is nothing in the morphology of known
skeletal elements of M. aberratum which would preclude it from being ancestral
to later Monachinae.

Other European fossil monachines include Pliophoca etrusca ‘Tavani 1942
from the later Pliocene of Italy, and Palaeophoca nystui Van Beneden 1877 from
the Pleistocene of Belgium. The status of the Pliocene Pristzphoca occitana
Gervais & Serres 1847 is uncertain, although Kellogg (1922: 78) states that it
‘certainly belongs to the same genus, and possibly to the same species, as the
fossil form from the Orciano in Italy’ (i.e. Pliophoca etrusca). P. etrusca is very
similar to the extant M. monachus, and it was in fact originally described as a
fossil M. albiventer (=monachus) by Ugolini (1902). Whatever the correct
nomenclature of the Italian fossils, they are clearly closely related to @M.
monachus, as is also Monotherium maeoticum (Kellogg, 1922). Several other species
of Monotherium have been described (Kellogg 1922; King 1964).

This sparse and taxonomically confused fossil record clearly does not
allow an unequivocable statement to be made on the phyletic relationships of
the species concerned. However, the Monotherium aberratum — Monachus monachus
lineage is here taken to include Pliophoca etrusca as an intermediate form, and
Palaeophoca nystit as an offshoot.

Elsewhere in the areas of distribution of modern monk seals, the fossil
record is extremely poor or non-existent. Nevertheless, some inferences can be

EVOLUTION AND DISPERSAL OF MONACHINAE IOI

drawn from the few available records of fossil monk seals and related forms.

Although the South African Prionodelphis capensis does exhibit certain
characteristics which are typical of modern Antarctic seals, in totality of
characters there is a more marked resemblance between it and the monk seals
and their ancestors. However, it differs from Pliophoca, Palaeophoca and Monachus
in at least one important anatomical detail, and that is the presence of an
entepicondylar foramen in the humerus. This feature is also present in the
humerus of Monotherium aberratum. It is therefore concluded that the dichotomy
of the European monachine lineage, and the lineage which includes P. capensis,
occurred in the late Miocene or early Pliocene.

Similarly, in view of the late Pliocene date of the fossils from Langebaan-
weg, it is certain that the southern lineage was established in the south Atlantic
sometime earlier in this epoch, or perhaps even during the late Miocene.

Having taken Europe as the centre for monachine evolution, the possible
routes followed by early monachines into the South Atlantic need to be con-
sidered. When intercontinental migrations of seals took place, it is reasonable
to suppose that these would be successful where distance between landfalls was
least, and prevailing ocean currents favourable. The role of ocean currents in
pinniped migrations may well be more complex than appears at first sight.
For example, a factor directly related to ocean currents is that of water
temperatures. King (1964) has noted a connection between pinniped distribu-
tion and water temperatures, and thus also ocean currents. It is to be expected
that pinnipeds undertaking long migrations would follow prevailing ocean
currents, not only because these would facilitate long-term directional move-
ments, but also because water temperatures remain more equable within a
single current system. ‘Temperature is important in determining the ecosystem
of any oceanic environment, and since pinnipeds are simply one element of
such a system, their dispersal in the past is likely to have been influenced by
sea temperature and ocean current patterns.

Pliocene to Recent records of monk seals in the Atlantic Ocean are as
follows:

M. monachus and its ancestors— Europe and north-west Africa
M. tropicalis— Caribbean

P. rovereti— Argentina

P. capensis—South Africa

With Europe and north-west Africa taken as the centre from which the
Monachinae dispersed, the other records can be accounted for in a number of
ways.

However, bearing in mind the factors already mentioned, the pattern of
dispersal outlined below is considered the most probable of the alternatives.

While east to west crossings of the Atlantic could conceivably have taken
place in both hemispheres, it is simpler to suppose that some time in the late
Miocene or early Pliocene a single such crossing took place, starting from the

102 ANNALS OF THE SOUTH AFRICAN MUSEUM

north-west coast of Africa to the most easterly parts of South America. Even
today M. monachus occurs as far south as Cape Blanc in Mauritania, and historic
records indicate its presence even further south in Senegal (Van Wijngaarden
1962). Assuming a similar distribution for an early ancestor, an east to west
crossing following the prevailing ocean currents by way of the Canary Islands
at the shortest distance between Africa and South America would not have
been difficult to achieve.

Once established in South America, these early monachines could have
spread both north and south, the former to give rise eventually to M. tropicalis
in the Caribbean, and the latter to give rise to P. rovereti in Argentina. All the
monachines involved in these movements would have been species adapted to
warm water conditions.

An early monachine population adapting itself to colder waters and moving
into the far south of South America would then have been ideally situated for
further dispersal into Antarctic regions, eventually to give rise to modern Ant-
arctic monachines. This same group could also have undertaken another Atlantic
crossing, this time from west to east, by way of islands near the Antarctic
convergence and the prevailing ocean currents of the southern mid-latitudes,
to South Africa. The presence of P. capensis at Langebaanweg is thus accounted
for, and while Europe is regarded as the centre from which all later Monachinae
were dispersed, the east coast of South America is here regarded as the centre
from which the southern monachines arose (Fig. 1).

Palaeontological evidence relating to the date of arrival of the ancestors
of M. tropicalis in the Caribbean is inconclusive. While there are Pleistocene
records of M. tropicalis in Florida and South Carolina (King 1964), early forms
of this species must have been present in this area before the Pleistocene. ‘This
is a statement which is also based on inference rather than direct evidence.

Since the two records of Prionodelphis both date from the Pliocene, and
only one east to west crossing of the Atlantic has been indicated, it suggests that
ancestors of M. tropicalis were present in the Caribbean during the Pliocene as well.

There is further indirect evidence to substantiate this statement. It is
well known that there is greater morphological correspondence between M.
tropicalis and M. schauinslandi than exists between either of these species and
M. monachus, which indicates that these two species have had a common ancestor
more recently than the genus as a whole. In order that this common ancestor
could give rise to the two modern species, it must have been present in both the
Atlantic and Pacific Oceans.

Bearing in mind the existing distribution of M. tropicalis and M. schauinslandt,
the most probable Atlantic—Pacific link penetrated by the common ancestor,
was that which existed before the North and South American continents were
linked by the isthmus of Panama. Once established on the Pacific coast of
Central America, the dispersal of ancestors of M. schauinslandi to islands in the
central Pacific would have been facilitated by the prevailing currents in the
north equatorial region of this ocean (Fig. 1).

103

oa
<
g
=
2)
<
Zz
2)
=
ca
ce)
4
<
wn
ad
a
=
4
Qa
a)
Z
<
a
=
my
=)
=
2)
>
ca)

‘QU200I[q 9}¥] 24} Sutmp Ayqeqoid ‘eouewy

YON 0} BVOLIOUIY YNOG Jo svOd JsaM Buoye vsunospy Jo peoaidg ‘PF

“QUSI0I[g 2Y} SuLmp swygjapouowg Jo MuepYy yO jo Burssoiy °F
‘Qus00I[g 94} Sutmmp Ayqeqoid ‘suorwe1

OOIEJUY UT SSUTYIVUOU! UIOY}NOS UIBPOUT JO S1O}s90UR JO [VALLIY *S
"QUd90I[ g

Ajrea 10 SUSDOTIAY 2}¥] UT VOLIDUIY YINOG Jo JsvOd yseva 0} BuIssoIy ‘1

:Tesodstp o1oydstuiefFy uoy NOG

“oyloeg [e1}]UI0 0} wuDjsumpyaIs snyIDUOPY JO S10}s29U Jo peaidg °E

*BOLIDUIY [BI]UID Jo
ysvoo OYtIvg 0} Sutsso1o pue ‘asusd0I[g ApIvs 10 BUDDOTF, 9}¥] By
Sulinp uvoqqiey ut siv2dou snysnuopy JO S10\sooue Jo [eAIIIY ‘%

“Ud90I[ J
Ajiva IO DUZDOI 93¥] UI VOTIOUIW YINOG Jo 4svOd jsv9 0} BUISSOIT ‘I
:[esiodstp ssoydstumozy UIOYIION
“SQUIYOVUOUL JO Sp10991 1UIIIY — Y
“SOUTYOVUOUL JO sp10901 DUZ90T[ g — J
*SOUIYOVUOU JO Sp1OIII DUDIOIPI — JI
‘QVUIYOeUOW 9Y} JO saynor [esiodsip pojsossng—smose ye
*sUvIIO JY} JO syUIIIND Vo¥JINs [edroutsd oy} Jo sUIOgG —sMO IL WYSTT

*s[e9s SUTYDeUOUI Jo soynod [esiodstp poysasBng “1 “O17

104 ANNALS OF THE SOUTH AFRICAN MUSEUM

The last time a crossing of the Central American region by pinnipeds was
possible is not known for certain. Olsson (1932) has stated that the first evidence
for a separation between marine invertebrates of northern Peru and the
Caribbean was in the late Miocene. This suggests that the common ancestor
of M. tropicalis and M. schauinslandi must have crossed somewhere near the
present isthmus of Panama by this time. However, Simpson (1950) has referred
to the Pliocene mammals migrating between the two New World continents as
‘island hoppers’, and indicates that the final North and South American land
connection was achieved in the late Pliocene, a view supported by Whitmore
& Stewart (1965). A Pliocene Atlantic to Pacific crossing by early monachines
would seem more reasonable in view of the late Miocene date for the earliest
monachine in Europe. Even though the movement of marine invertebrates
between the Atlantic and Pacific may not have been possible since the late
Miocene, the movement of pinnipeds across the area between Central and
South America during the Pliocene is probable. It is possibly the imprecise
relative dating of the fossils and deposits referred to in the preceding discussion
that causes the discrepancy in the evidence provided by invertebrate and
vertebrate fossils. Gabunia & Rubinstein (1968) give an indication of the
difficulties encountered in the relative dating of deposits at the period of time
which is critical in the present instance.

The preceding interpretation of the evolution and dispersal of the monk
seals can be summed up as follows:

1. Monachus monachus evolved from a late Miocene monachine, largely in the
area in which it is found today.

2. In the late Miocene or early Pliocene an early monachine crossed the
Atlantic from the west coast of Africa to the east coast of South America.

3. One group of these immigrants moved southwards along the east coast of
South America and gave rise to Prionodelphis rovereit.

4. This southward movement of monachines continued into Antarctic regions,
while at least one population crossed the South Atlantic to South Africa
and gave rise to Prionodelphis capensis.

5. A second group of the original immigrants to South America moved
northwards into the Caribbean, and also across into the Pacific before the
final linking of the North and South American continents. They eventually
gave rise to Monachus tropicalis and M. schauinslandi, ancestors of the latter

having moved from the Pacific coast of Central America to the central
Pacific.

The monk seals are a remarkably conservative group. Although the three
modern species have had a history independent of one another, perhaps dating
back to the Miocene, and certainly dating back to the Pliocene, they are still
morphologically very similar to one another. It seems that having once adapted
to their environment they received little, if any, pressure to cause adaptive
change. This conservative, low latitude group survives today, effectively
isolated from other pinnipeds.

~~

EVOLUTION AND DISPERSAL OF MONACHINAE 105

THe ANTARCTIC MONACHINES
(Excluding the southern elephant seal)

Before dealing with the evolution and dispersal of Antarctic monachines,
it is necessary to consider in detail their possible phyletic relationships to the
two species of Prionodelphis, based on considerations of comparative morphology.

A detailed study of the known remains of Prionodelphis capensis revealed
that there is an undoubted connection between this species and the Antarctic
monachines in general (Hendey & Repenning 1972). However, it was also
clear that no definite phyletic connection between the South African species
and any one of the modern Antarctic seals could be demonstrated. In some
respects P. capensis is morphologically intermediate between the genus Monachus
and the Antarctic seals. Such features of P. capensis as the relatively narrow
cheekteeth, the flat and broad petrosal apex of the tympanic, and the relatively
stout humerus and spatulate radius, suggest a stage of development from the
ancestral monachine condition (exhibited by the conservative Monachus)
towards Hydrurga, Lobodon, Leptonychotes and Ommatophoca.

The highly modified dentitions of the Antarctic monachines are the features
which differentiate them from P. capensis most strikingly. On the basis of their
dentitions it is improbable that either Hydrurga or Lobodon could be derived
from a P. capensis-like ancestor, since this would have necessitated the reversal
of at least one evolutionary trend.

Both species of Prionodelphis have a reduced upper fifth postcanine, with
only the vestiges of a single anterior accessory cusp remaining, this apparently
having derived from the basic three-cusped pattern still evident in Monachus.
Both Hydrurga and Lobodon have little reduction of this tooth, which in both
cases has well-developed accessory cusps anteriorly and posteriorly. While it is
not impossible that lost features can be redeveloped (Kurtén 1963), this is
not usual, and the possibility of a phyletic connection between Prionodelphis
and both Hydrurga and Lobodon becomes more improbable.

The multiplication of accessory cusps in Lobodon is yet another complicating
factor. The lower postcanines of P. rovereti have two anterior and two posterior
accessory cusps, but even this does not match the proliferation of posterior
cusps in the second to fifth postcanines of Lobodon. Cusp development in this
genus is variable, and additional accessory cusps appear to develop as buds
from the principal cusp and as projections from the cingulum. There is no
evidence to suggest that the archetype monachine cheektooth had an array of
cusps such as is seen in Lobodon. Consequently, although deviation from the
basic three-cusped pattern may not have been difficult to achieve, in the case
of the Prionodelphis upper fifth postcanine it would require an even more complex
reversal of the trend already mentioned.

In addition, Lobodon differs from other Antarctic monachines in having
teeth which are relatively broad. A comparison of the cheekteeth of P. rovereti,
P. capensis and Hydrurga shows a progressive narrowing of the teeth (Hendey

106 ANNALS OF THE SOUTH AFRICAN MUSEUM

& Repenning 1972: Table 1), and a similar transverse compression is also
evident in the teeth of Leptonychotes and Ommatophoca. ‘The teeth of Lobodon are
more like the ancestral monachine condition in this respect.

Kellogg (1942: 453) has already suggested that P. rovereti might be ancestral
to Leptonychotes. This suggestion is not affected by the one that P. capensis is a
derivative of the Argentinian species. The phyletic connection between the
two species of Prionodelphis is indicated partly by the inferred manner of dis-
persal of southern monachines, the earlier suggestion that in some respects
P. capensis has more advanced characteristics (Hendey & Repenning 1972: 93),
and that the Entre Rios deposits are early Pliocene (Romer 1966), while those
at Langebaanweg are late Pliocene (Hendey 1970).

Prionodelphis capensis could be regarded as intermediate between P. rovereti
and Leptonychotes in view of the reduction in the size of the posterior root in its
upper postcanines, the reduction in the number of cusps in its lower post-
canines, and in the greater transverse compression of its cheekteeth. Much the
same can be said in the case of Ommatophoca, but in this instance the differences
between the cheekteeth of P. capensis and O. rossi are far greater, and changes
would have needed to be at a greatly accelerated rate. Perhaps the feature most
suggestive of a connection between Prionodelphis and Leptonychotes is the recurved
crown and virtual lack of accessory cusps in the upper fifth postcanine of the
former genus, characteristics which are matched in most of the cheekteeth of
Leptonychotes. To a lesser extent these characteristics are evident also in the
cheekteeth of Ommatophoca.

Thus on the basis of teeth alone, it appears that Prionodelphis is more likely
to be ancestral to Leptonychotes and Ommatophoca, although P. capensis itself is
probably not on a direct line to these species, but merely paralleled the trends
which led to them. On the other hand AHydrurga and Lobodon are likely to be
derived from an even more generalized monachine—a proto-Prionodelphis.

Looking once more at the distribution of the southern monachines, it is
clear that, other factors aside, P. capensis was not geographically well situated
to give rise to the Antarctic monachines, and Leptonychotes and Ommatophoca in
particular.

In view of the suggestion that Hydrurga and Lobodon arose from a ‘proto-
Prionodelphis’, it is possible that early representatives of these genera reached
Antarctica before the ancestors of Leptonychotes and Ommatophoca. Since the
latter two genera have a more southerly distribution than Hydrurga and Lobodon
(King 1964), it is likely that their ancestors entered Antarctica from the southern
tip of South America (55°S) via the South Shetland Islands to the Palmer
Peninsula of Antarctica itself. The most likely route into Antarctica from South
Africa (34°S) would, because of prevailing ocean currents, have been by way
of the south Indian Ocean, via islands near the Antarctic convergence and the
pack ice to the Antarctic continent. In addition to being a more difficult route,
it would have necessitated the crossing of an area (the outer fringes of the pack
ice) which may already have been inhabited by the ancestors of Hydrurga and

EVOLUTION AND DISPERSAL OF MONACHINAE 107

Lobodon. While the four modern Antarctic species do not compete for food,
this must become progressively less so further back in time when their respective
ancestors were more generalized.

A final point concerning the possibility of a phyletic connection between
P. capensis and one or more of the modern Antarctic monachines, concerns the
age of P. capensis. Since P. capensis is considered to date back 4—5 million years
(Hendey 1970), it would be required that the marked morphological differences
between it and, say, Leptonychotes be developed in this period of time. If the
early Pliocene date for P. rovereti is correct, the period of time for similar changes
to occur might be two or three times as long. In itself this is meaningless, but
clearly the greater the time involved, the greater the possibilities for radical
morphological change.

In the absence of convincing evidence to the contrary, it is probable that
differentiation and evolution of the Antarctic monachines took place in the
high latitudes in which they occur today. The more extreme climate and climatic
changes in high latitudes during the Pliocene and Pleistocene may well have
provided the mechanism for accelerated changes in the Antarctic monachines,
the nature of the environment acting as a stimulant to adaptive change. Owing
to the lack of deposits of Pliocene and Pleistocene age in Antarctica, except for
those beneath the Antarctic Ocean, it is very improbable that the nature of the
development of Antarctic seals will ever be known from fossil remains. Thus
with this group of pinnipeds, interpretations of their evolutionary history will
probably remain speculative.

It is clear that there was a far more rapid evolution of specialized characters
in the southern monachines than was the case with the monk seals, even
though the independent history of the four southern species was probably no
longer than that of the three monk seals. Thus the Antarctic seals contrast
with the monk seals by being a high latitude and highly specialized group, as
opposed to a low latitude and conservative group.

THe ELEPHANT SEALS

The genus Mirounga has long been the subject of controversy as to its
relationships. It is now included in the Monachinae (King 1966) but anatomi-
cally and in its distribution it does still to some extent stand apart from other
monachines.

While some conclusions on relationships can be drawn from comparisons
between the cheekteeth of other monachines, the peg-like teeth of Mirounga
are singularly uninformative. Burns & Fay (1970: 389), however, concluded
that, ‘craniologically Mirounga is more like other phocids of the Southern
Hemisphere than is Monachus’. The mere fact that Mirounga is a highly specialized
genus makes it probable that it would have closer connections to the diversified
Antarctic monachines, rather than with the conservative monk seals. In other
words, it is suggested that adaptive radiation in the Monachinae is confined
to the southern high latitude group, to which Mirounga must therefore belong.

108 ANNALS OF THE SOUTH AFRICAN MUSEUM

Furthermore, the distribution of the two modern species is more readily
accounted for in assuming a southern origin for the genus. This is already
suggested by the fact that M. angustirostris maintains a pattern of breeding (in
the northern winter) which is typical of Southern Hemisphere phocids (King
1964).

The present distribution of the southern elephant seal includes the extreme
southern end of South America, but within historic times it has been reported
as far north as the Juan Fernandez Islands, which is in fact the type locality.
A wider spread along the west coast of South America during the past, in a
manner comparable to that of modern Arctocephalus australis, is conceivable
owing to prevailing ocean currents and water temperatures of that coast. This
would have brought the southern and northern ranges into far closer proximity,
and indeed have allowed the southern species, or an ancestor, to reach the area
in which the northern species occurs today. A split in this distribution, centred
in low latitudes, of a single population of Mirounga, and a gradual retreat in
its range in the south, would then account for the present pattern of elephant
seal distribution. Mirounga angustirostris can thus be regarded as a relict species,
surviving in isolation far from the origins of the genus.

Since there is an early Pleistocene record of Mzrounga in California (C. A.
Repenning, pers. comm.), it is probable that the genus arose during the Pliocene,
and the present pattern of elephant seal distribution developed during the
Pleistocene.

The suggested phyletic relationships of modern and some fossil Monachinae
are illustrated in Figure 2.

DIscussION

The present pattern of monachine distribution is perhaps the most unusual
of any of the pinniped subfamilies. The Otariinae and Arctocephalinae, which
are taken by some to be a single subfamily (Mitchell 1968), have essentially
similar distributions, with the eastern Pacific shorelines providing the closest
links between the species of the two hemispheres. ‘The Odobeninae and Phocinae
are purely Northern Hemisphere groups, largely confined to high latitudes.
The Monachinae alone can be subdivided into two geographically widely
separated groups, while a third subdivision includes two equally widely sepa-
rated species of a single genus.

King (1964) has discussed pinniped distributions as they relate to ocean
currents and water temperatures, and has remarked upon the correlation
with colder waters. The monk seals are one of the few exceptions to the general
rule that, ‘the 20°C summer isotherm in either hemisphere, where it approaches
continental coasts forms a reasonable pointer to the limits of where one might
expect to find seals’ (King 1964: 89). The critical temperature for the monk
seals is higher. By contrast the Antarctic seals are limited to waters with a
maximum temperature of 3-4°C. The southern elephant seal also occurs in

109

EVOLUTION AND DISPERSAL OF MONACHINAE

NORTHERN HEMISPHERE

SOUTHERN HEMISPHERE

To Northern Hemisphere
M.angustirostris

Monachus _ Monachus Monachus
schauinslandi tropicalis monachus

Palaeophoca
nysti

Pliophoca
etrusca

“Ancestral monachine
(Monathenum aberratum)

a oa
fa | Prionodelphis
: rovereti
ye Prionodelphis capensis
Mirounga Lobodon Hydrurga Ommatophoca Leptonychotes
leonina carcinophagus leptonyx rossi weddelli

Fic. 2. Suggested relationships of some monachine seals.

Pleistocene &
Holocene

Pliocene

Miocene

Pliocene

Pleistocene &
Holocene

IIo ANNALS OF THE SOUTH AFRICAN MUSEUM

the cold waters of the far south, while its northern relative conforms to the
general rule quoted above.

It is reasonable to suppose that at one time the monachines had a less
interrupted distribution, and the records of Prionodelphis give an indication that
this was indeed the case. In order to explain the curious pattern of modern
distributions, it is necessary to refer to the occurrence of other pinnipeds. It is
found that the southern fur seals (genus Arctocephalus) and southern sea lion
(genus Otaria) inhabit much of the intervening area which is suitable for
occupation by seals. Bearing in mind the suggested manner of dispersal of early
monachines, it is probable that much of their range during the Pliocene, and
perhaps part of the Pleistocene as well, included the areas in which at least
some of the species of Arctocephalus and also Otaria occur today. The indications
are, therefore, that there has been replacement of monachines by otariids,
especially in the Southern Hemisphere.

The earliest Southern Hemisphere records of the Otariidae are Otaria
fischer Gervais & Ameghino 1880 from Argentina, a Phocarctos hooker: from
New Zealand (Berry & King 1970), and ‘Arctocephalus’ williams: from Australia
(King 1964). The two Australasian species are Pliocene in age, and while
King (1964) listed the Argentinian record as ‘? Miocene’, it too probably
dates from the Pliocene (Davies 1958). Excluding Pleistocene records, other
recorded fossil otariids are from either the west coast of North America or
Japan. Thus it is probable that while the Phocidae had the Atlantic Ocean as
the centre for their evolution, the early history of the Otariidae was confined
to the Pacific. The indications are, therefore, that the Otariidae only reached
southern Africa after they had become established in Australasia and South
America, i.e. the two southern continents bounding the Pacific Ocean. It must
have been very late in the Pliocene, or perhaps only in the Pleistocene that
the otariids became established in southern Africa. Although the fossil record
is not conclusive on this point, it does seem likely that the extinction of Priono-
delphis in South America and South Africa was more or less synchronous with
the advent of otariids in these regions.

The present distribution of the two species of Mirounga discussed earlier,
may at least be partly determined by the presence of Arctocephalus australis
populations on the South American coast. King (1964) states that even though
M. leonina and A. australis have a common diet they appear to co-exist without
difficulty, but she also notes that when breeding season conflicts occur ‘the fur
seal usually wins’ (King 1964: 25). The fact that M. leonina distribution has
become considerably less in recent times suggests that the sympatric existence
of these two species may be in a state of delicate balance, and, if upset, A.
australis is in a position of dominance.

Unsuccessful competition by monachines with otariids cannot be proven,
but the present distribution of these two groups suggests that there is a mutually
exclusive inter-relationship between them.

The possible influence of climatic changes on monachine distribution

EVOLUTION AND DISPERSAL OF MONACHINAE Pr?!

cannot, of course, be ignored. For example, M. leonina may be more sensitive
to increasing temperatures than Arctocephalus, and the virtual disappearance of
this species from the South African coast in the Holocene may be due to the
general warming of conditions during this epoch. It is unlikely that prehistoric
man played any significant role in the disappearance of the South African
Mirounga, although this possibility cannot be entirely discounted. Late Pleisto-
cene coastal hominid occupation sites are all below present sea-level, and the
relative importance of Mirounga and Arctocephalus in diets of the hominids of the
time is not known.

It is probable that the radiation of the otariids, coupled with changes in
climate, jointly determined the nature of the present distribution of the Mona-
chinae, which survive as isolated populations in low latitudes, and as firmly
entrenched populations only in the southern high latitudes.

In reviewing the evolution of the Monachinae as outlined above, it is
evident that this group does not conform exactly to some widely accepted
principles. Crowson (1970: 133) lists certain ‘rules’ “for determining the areas
of origin of systematic groups from the patterns of distribution of their present-
day members’. He states that it ‘has been suggested that any group should be
considered to have originated in that area where (a) it is represented by the
greatest number of existing species . . . or (d) its most primitive living forms
occur’. Crowson’s conclusion that these ‘rules’ can be misleading is substan-
tiated by the Monachinae, since the greatest number of living species (the
Antarctic seals) are far removed from the most primitive living forms (Monachus).
In this case ‘rule’ (d) applies.

CONCLUSION

There are undoubtedly alternatives to the speculations outlined in this
report. However, since the facts which form its basis are so limited, their
interpretation must to a large extent be subjective. ‘This is inevitable when
dealing with a group of mammals with a poor fossil record which is supple-
mented only at very infrequent intervals. ‘The only two Southern Hemisphere
records which throw any light on the early history of the southern Monachinae
are Prionodelphis rovereti from Argentina and P. capensis from South Africa, the
descriptions of which were published nearly 50 years apart.

Given such a situation it seems preferable to speculate now, rather than
wait for further significant discoveries to be made.

If this paper serves to stimulate better reasoned interpretations of the
limited fossil record of the Monachinae, it will have served a purpose. What-
ever is finally accepted, it is clear that a reassessment of the taxonomy of the
monk seals and all fossil Monachinae is necessary.

SUMMARY

The fossil record of the Monachinae is outlined, and it is concluded that
this pinniped subfamily arose in the western Europe—Mediterranean area. It

II2 ANNALS OF THE SOUTH AFRICAN MUSEUM

is concluded that the monk seals, and the early ancestors of the Monachinae
as a whole, are an essentially low-latitude and conservative group, from which
arose the high-latitude and specialized ‘southern’ monachines, which include
Mirounga angustirosiris. It is suggested that the latter group underwent their
differentiation in the high latitudes in which they are most commonly found
today, having arisen from Prionodelphis and ‘proto-Prionodelphis’ populations
which spread into Antarctic regions from South America. It is suggested that
the decline of the Monachinae in southern mid-latitudes was due at least partly
to the rise of the Otariidae in these regions.

ACKNOWLEDGEMENTS

The present study was a direct outcome of an investigation of the Prionodel-
phis capensis fossils from Langebaanweg, which was undertaken jointly with
Charles A. Repenning of the Pacific Coast Center of the U.S. Geological
Survey. I am greatly indebted to him for the stimulation and assistance he
provided in both these studies. The opinions expressed in the present paper do
not necessarily reflect those of Mr. Repenning.

The current investigations at Langebaanweg are being supported by the
South African Council for Scientific and Industrial Research, Chemfos Ltd. (a
subsidiary of the African Metals Corporation) and Shell South Africa (Pty.)
Ltd. The Wenner-Gren Foundation for Anthropological Research, New York,
provided the vehicle used in the field work at Langebaanweg (Grant no.

2752-1834).
REFERENCES

Berry, J. A. & Kine, J. E. 1970. The identity of the Pliocene seal from Cape Kidnappers,
New Zealand, previously known as Arctocephalus caninus. Tuatara 18: 13-18.

Burns, J. J. & Fay, F. H. 1970. Comparative morphology of the skull of the Ribbon seal,
Histriophoca fasciata, with remarks on the systematics of the Phocidae. 7. Zool. 161: 363-394.

Crane, H. R. & GrirFen, J. B. 1968. University of Michigan radiocarbon dates XII. Radiocarbon
10: 61-114.

Crowson, R. A. 1970. Classification and biology. London: Heinemann.

Davugs, J. L. 1958. The Pinnipedia: an essay in zoogeography. Geogr. Rev. 48: 474-493.

FRENGUELLI, J. 1922. Prionodelphis rovereti un representante de la familia ‘Squalodontidae’ en al
Paranense Superior de Entre Rios. Boln Acad. nac. Cienc. Cordoba 25: 491-500.

GasuntA, L. & RusinsteIn, M. 1968. On the correlation of the Cenozoic deposits of Eurasia
and North America based on the fossil mammals and absolute age data. 13th Int. geol.
Congr. 10: 9-17.

Gervais, H. & AmeEcuHINo, F. 1880. Les mammiféres fossiles de l’Amerique du Sud. Buenos Aires;
Paris: Savy. .

Gervais, P. & SErReEs, M. de. 1847. Nouvelles observations sur les mammiféres dont on a
trouvé les restes fossiles dans les sables marins de Montpellier. Annls Sci. nat. (3) 8: 224-226.

HENDEY, Q.B. 1970. The age of the fossiliferous deposits at Langebaanweg, Cape Province.
Ann. S. Afr. Mus. 56: 119-131. |

Henvey, Q.B. & Repennine, C. A. 1972. A Pliocene phocid from South Africa. Ann. S. Afr.
Mus. 59: 71-98.

KeLitocc, R. 1922. Pinnipeds from Miocene and Pleistocene deposits of California. Univ.
Calif. Publs geol. Sci. 13: 23-132.

EVOLUTION AND DISPERSAL OF MONACHINAE IIlg3

Ke.toce, R. 1942. Tertiary, Quaternary, and Recent marine mammals of South America
and the West Indies. Proc. 8th Am. Sci. Congr. (Washington) 3: 445-473.

Kine, J. E. 1964. Seals of the world. London: British Museum (Natural History).

Kinc, J. E. 1966. Relationships of the Hooded and Elephant seals (genera Cystophora and
Mirounga). F. Kool. 148: 385-308.

KurtéEn, B. 1963. Return of a lost structure in the evolution of the felid dentition. Commentat.
biol. 26: 1-12. .

McLaren, I. A. 1960. Are the Pinnipedia biphyletic? Syst. Zool. g: 18-28.

MirtcHELL, E. 1968. The Mio-Pliocene pinniped Jmagotaria. 7. Fish. Res. Bd Can. 25: 1843-1900.

Norpmann, A. 1860. Palaeontologie Siidrusslands. Helsingfors: Friis.

Oxsson, A. A. 1932. Contributions to the Tertiary paleontology of northern Peru: Part 5, The
Peruvian Miocene. Bull. Am. Paleont. 19: 1-272.

Romer, A. S. 1966. Vertebrate paleontology. Chicago: University Press.
Saricu, V. M. 1969. Pinniped phylogeny. Syst. Zool. 18: 416-422.
Simpson, G. G. 1950. History of the fauna of Latin America. Am. Scient. 38: 361-389. (Reprinted

in Stimpson, G. G. 1965. The geography of evolution; collected essays: 165-208. New York:
Capricorn Books.)

Tavanl, G. 1942. Revisione dei resti del pinnipede conservato nel Museo di Geologia di Pisa.
Palaeontogr. ital. 40: 97-113.

Uco.int1, R. 1902. Il Monachus albiventer Bodd. del Pliocene di Orciano. Palaeontogr. ital. 8: 1-20.

VAN BENEDEN, P. J. 1877. Description des ossements fossiles des environs d’Anvers. Annis Mus. r.
Hist. nat. Belg. 1: 1-88.

VAN WIJNGAARDEN, A. 1962. The Mediterranean monk seal. Oryx 6: 270-273.

Wauitworg, F. C. & STEWART, R. H. 1965. Miocene mammals and Central American seaways.
Science 148: 180-185.

em U MRM iy tig urclav agai diy SSG
Wend Wie aris tah be rae Lig ve as le “ Westy ys ‘eh

‘ z i F i
= _ s i
bel 4 e * SR eee | 2% i: rte aay
-* i ty f
y 9 ; © sms & }
od n ae ' 1 t, TE
, ‘ ‘ ‘ « ( b al
: i ri 7
— M . » mr
- } = ‘
~/% ~ ee : « a4 : 4
‘ Sd, ‘
ad ye 9 ‘
d ye & ‘ 4 . \
: , : 4 . + ren
f - ri fy
te we 4
> ek
a } * ¥
aa) re L 9 f yt
5 f vs “Sy ey ao)
é y cia 8
~ 2 ot
i a oi
r i ¥
\ he
\ } Tee Ve
\ ; eh
7 ' . i
Nl a a \ eS
ca “ A
3s iD i
’ =
io ,
y fi a ; :
4 y
; ae
= ‘ > e
> et. sh
pele oan “)
\) 2 =
i, ir.
' ‘ 5 x
2
A =
; ‘
‘; =
> A } Ue
u 7 = i 4
ve . :
‘ wet
,
/ ’
’ ,
\ dy;
r= Nis ats
=) 3
- Re. ~~ i
in t= 5 i
)
.
= '
4 oe
~/ 2
‘ : t
of cul :
, 4
, oe
i < fos eh tn
3 ‘ =
t
+ ~
4
4 —
’
‘a ¢
a 4 *
ft
7 ‘
: -
. t x
‘3
¥ - Pe t
1 ~ Mi Pe
f if i
p e i *

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE, 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12cm X 18cm (19cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.
For journal articles give title of article, title of journal in italics (abbreviated according to the

World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,

volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHER, P.-H., Duvat, M. & Rarry, A. 1933. Etudes sur les changes respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 19602. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region
of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In scHULTZE. L,
Koologische und anthropologische Ergebnisse einer Forschungsreise 1m westlichen und zentralen Stid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51).
The Harvard system of reference to be used in the synonymy lists, with the full references
incorporated in the list at the end of the article, and not given in contracted form in the synonymy
list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 8o.

©
1 é
~
—
9
' x
£ 5
.
~
«J
is
oS
f
r
~

prints it aes Te aaa hy Pa aed Whig Bie

itd ai es re PO ak aD ade fe ian abe el ee

aN, tee

} ‘ae
; A ;
—, = et ( ayy! : j hat
1 ze } :
= ee
4 Se
= - ra) * J hy
2ae Va

SS) ot ie bi or a

ceotttnto lt, Leiber ah ete ee itpabesaS
“ . “r } ? 7 :

sunita Wola ie Hie uve fice
hd Nagi hes Aes aa ae Re,

ERE CO Ee a eT

d¢ Deets ym sae win: seats balan

70 Feta ap x ee i “a eee ih Ble (Qe nna
; , eet

2 b ; ; , iG : ‘hii
") & ; \ ho F a iT 4 oF cv,
ORE ric’. ele Alaa ng WA i tt. ry ede REE. sat i et wil seo
2 .. as aad, sweet ay i mba Mec eh b K fated at 4 hits Mi eee ‘1h -
aa iow) eke) yeloh ara , eh
eae i Ree AGP Sa Tah ee Ae oo ae ra ba vial, |
FN rea ; er a at Ar M
. eae Me PO iTh LORNA he LSTA Vet POEs i: eign Ri
3 s fea ta ee, ' eo ee PLT: \ RTO) Bee) ead a ey oby 2 a aay
. x tay ‘ th by 24 ‘ . 7 Y M4 ‘ ? 4 i. ;
i ol ati LEA wba) tealiaiek Mair st! Siren ae aa aya
MH 4 ee : 5 Shh gh xt me — y ral gst
cs

- 3 he ae AG ee F o Seb
ra 4 - -
? : ‘ Pe aKa: Wasi
= ie 5 oy, 7 sy
A aes 4 r he } i y és i ‘
2 or, 5,
i ad ¥ a ier os i \ rm
- y a Ge i ; ‘ ay
‘i = :

Z 4 i Pe ;
PS = 7 Dy
a r a os at ne y j sn ne

_ oe "
ial t
v0 NV
J y

c f 3 ait d

e : ee me

a
1 i o / } ri
1
~— =

e ie?

vi se i | ; Loe
; ra va : thi Oe Ut Re) oe te res
fa — , ag 7
Ps S Hh ; i J Ly i. i, a ‘a
i 4 r Lita r ry Ps, AL ee er aA en Dea Bit) 0 ; Z2 ‘ay ne wee
: ’ i oa yi
id
iJ) rf.

5O7,6F

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
March 1972 Maart
Part 6 Deel

A PLIOCENE URSID FROM
SOUTH AFRICA

By
Q. B. HENDEY

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 205 3> 5s 7-8), 3(1=2; 5» e016) 51-2, 5; 7-9)s
6(1, t.=p.1.), 7(1-3); 8, g(1-2), 10(1),
in (2. 5> 7s tpt.) 21, 24(2), 275 Br(t—3)). 33

Price of this part/Prys van hierdie deel
R2,00

Trustees of the South African Museum © Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 06 2

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

A PLIOCENE URSID FROM SOUTH AFRICA

By
Q.B. HENDEY

South African Museum, Cape Town
(With plates 19-20, 2 figs and 3 tables)

[MS. accepted 16 February 1972]

CONTENTS

PAGE

Introduction . ; ; : : fy kaye
The family Tae , <p hIG
Intra-familial categories = Fhe the reac ELS
The Langebaanweg bear. : d : 126
Description 3 . ; , ‘ . 126
Discussion . : ; P F ; ~- 199
Summary . 5 2 : : 2 EST
Pcinosledeemionts: : : : d ra 9
References. " 5 : : ; eign

INTRODUCTION

There are few mammal-bearing deposits of Pliocene age known in Africa
(Kurtén 1971: 134) and consequently occurrences which date from this
epoch are of significance in that they may produce records of species which are
of special phylogenetic and zoogeographic interest. The only Pliocene occur-
rences presently being investigated in southern Africa are those at Langebaan-
weg in the Cape Province (Hendey 1970a, 19706), and one of the more
remarkable records from this locality is that of an agriotheriine ursid (Hendey
1969). It was the first record of an agriotheriine in Africa, and is still the only
ursid known from sub-Saharan Africa.

The Agriotheriinae, which in the most restricted sense may be taken to
include the genera Agriotherium and Indarctos, are known from a number of late
Tertiary and early Pleistocene occurrences in Eurasia and North America,
and the South African record adds a new dimension to concepts of the evolution
and dispersal of this group. In general, the recorded species are represented by
rather fragmentary material and, unfortunately, this is also the case with the
Langebaanweg form. However, the significance of the material is not
diminished, although the description which follows might well require revision
if more specimens are recovered in the future.

The terrestrial mammal fauna with which the Langebaanweg agriotheriine
is associated includes a number of species which are unexpected in an African
context. For example, one of the hyaenids is referred to Percrocuta, a genus
which is otherwise known in Africa only from the Algerian Miocene (‘Thenius

115
Ann. S. Afr. Mus. 59 (6), 1972: 115-132, 2 pls, 2 figs, 3 tables.

116 ANNALS OF THE SOUTH AFRICAN MUSEUM

1966). It was, however, widely distributed in Eurasia during the late Tertiary
(Kurtén 1957a). A second unusual species is an as yet unnamed boselaphine
antelope, which apparently derives from the Miocene Protragocerus labidotus of
Kenya (Gentry 1970). The boselaphines were also common in Eurasia during
the late Tertiary.

Other significant, although not unexpected, records from Langebaanweg
include an early ancestor of Hyaena hyaena, a primitive form of Mammuthus
subplanifrons which is one of the earliest of the true elephants (Maglio & Hendey
1970), and an ancestor of the white rhinoceros, Ceratotherium simum (Hooijer, in
press).

The Langebaanweg deposits are unique amongst the major late Cenozoic
fossiliferous occurrences of sub-Saharan Africa in that a marine fauna is asso-
ciated with the terrestrial vertebrate fossils. Vertebrates linked with the marine
environment include the first recorded fossil penguin from Africa (Simpson
1971) and an unusual monachine seal, Prionodelphis capensis (Hendey & Repen-
ning 1972), which has shed some light on the origins of the Antarctic seals
(Hendey 1972).

Viewed in relation to the fauna as a whole, the agriotheriine is but one of a
series of important additions to the fossil record of the late Cenozoic of Africa.

In order that this species might be placed in taxonomic perspective, its
description is preceded by a brief review of the Ursidae.

THE FAMILY URSIDAE

The Ursidae are a comparatively recently evolved mammalian group,
with a relatively small number of constituent genera, and they have received
a considerable amount of attention from palaeontologists and neontologists
alike. Despite this they have proved an extremely controversial group and it is
only recently that a measure of agreement has been reached on their phyletic
and intra-familial relationships.

The most comprehensive account of the Ursidae is that of Erdbrink (1953),
and this has provided an invaluable basis for more recent work on the family.
Since the appearance of Erdbrink’s monograph much attention has been focused
upon the ursine bears (the genus Ursus and closely related forms). Their phylo-
geny is now one of the best known of all mammalian groups and this successful
study has been largely due to the work of Kurtén (1957), 1958, 1964, etc.) and
Thenius (1959a). Much the same can be said of the tremarctine bears (the
genus Tremarctos and its close relatives) (Kurtén 1966, 1967), although the
origin of this group remains obscure.

The taxonomic position of the giant panda, Azluropoda melanoleuca, has been
a matter of controversy for more than a century owing to its combination of
ursid and procyonid characteristics. Davis (1964) presented what is perhaps the
definitive study of the anatomy of this animal, and he held the view that it is
indeed a bear. While this conclusion is still not universally accepted, Ailuropoda
is regarded as an ursid in the present study.

A PLIOCENE URSID FROM SOUTH AFRICA 117

The Tertiary ursids, excluding the immediate ancestors of the ursine
group, can conveniently be placed in two categories. The first includes those
genera which bridge the evolutionary gap between the Canidae and ‘true’
bears. They are Cephalogale, Hemicyon and Dinocyon. Secondly, there is a more
advanced group in which many specialized ursid characteristics were deve-
loped. This group includes Agriotherium and Indarctos. Not surprisingly these two
groups may be broadly differentiated on temporal as well as morphological
grounds, the former being essentially Miocene in age, and the latter dating
largely from the Pliocene. The Tertiary ursids are not as well known as the
Quaternary forms because they are less well represented in the fossil record.
There are also many uncertainties regarding the relative ages of some members
of the groups.

Frick (1926) separated these Tertiary forms from both the Canidae and
Ursidae, placing them in a single unit which he called ‘Hemicyoninae’. Subse-
quently Pilgrim (1932) referred them all to the Ursidae and this arrangement
is generally followed today.

‘There have been many differences of opinion concerning formal subdivisions
within the family. For example, Kraglievich (1926) proposed separation into
three subfamilies, which Simpson (1945: 225) argued was ‘of very doubtful
theoretical validity and of little or no practical convenience’. Indeed, he even
appeared to be in some doubt as to whether the bears merited family rank.
Erdbrink (1953) was another not in favour of having subfamilies within the
Ursidae. ‘Today it has become fairly widely accepted that subfamilial grouping
is both reasonable and desirable and the arrangement usually followed is that
of ‘Thenius (1959a), who recognized the Hemicyoninae, Agriotheriinae, Tre-
marctinae and Ursinae.

In spite of the formal system of zoological nomenclature, there is a strong
element of personal opinion in the definition of many taxa. The inclusion or
exclusion of the Hemicyoninae from the Ursidae appears to depend largely
upon whether a ‘vertical’ or ‘horizontal’ system of classification (Simpson 1945)
is favoured. An excellent example of this basic difference in approach is afforded
by the controversy which has surrounded the taxonomy of early Pleistocene
Hominidae. Reed (1967) has given a concise summary of the two points of view,
and states that ‘as gaps in the fossil record . . . have been filled, the tendency
has been . . . to shift from a horizontal (grade) type of classification to a vertical
(clade) type’.

While the general pattern of ursid evolution has long been appreciated,
the intra-familial groupings have tended to emphasize morphological rather
than phylogenetic aspects of the family. An attempt has been made here to
classify the ursids according to their phylogeny, although this was hampered by
the fact that there are still certain critical points in ursid evolution which are
not satisfactorily resolved. The intra-familial classification to be given presently
follows that of Thenius (19592), but is modified on the basis of the tentative
phylogeny presented in Figure 1. This phylogeny is based on the work of

ANNALS OF THE SOUTH AFRICAN MUSEUM

118

EOCENE

RECENT

PLEISTOCENE

PLIOCENE

MIOCENE

&
OLIGOCENE

Ursus arctos group

Dinocyon

Hemicyon

Ursus americanus group
i (Boeoata® North America,Arctic) (Asia,

North America)

Cephalogale

Melursus

(Asia)

x
Protursus
Europe )

Helarctos

(Asia)

/ | JL.
7

Indarctos ,

Tremarctos
(South America)

Indarctos

}
/

Ailuropoda
(Asia)

Agriotherium

Fic. 1. Tentative phylogeny of the Ursidae.

A PLIOCENE URSID FROM SOUTH AFRICA 11g

Erdbrink, Kurtén and Thenius, and the morphological characters, temporal
range and geographical distribution of each genus was taken into consideration.

The pattern which emerges is one of three distinct radiations within
the family, each’successive radiation cutting out the one preceding temporally
and/or geographically. In each case the stem genus is recorded in Europe,
and dispersals were largely confined to the northern continents. ‘The intra-
familial classification arrived at is as follows:

Subfamily Tribe Genera included
Hemicyoninae .. Hemicyonini .. Cephalogale, Hemicyon, Dinocyon
Agriotheriinae .. Agriotheriini .. Ursavus, Indarctos, Agriotherium
Ailuropodini .. Ailuropoda
Tremarctini .. Arctodus, Tremarctos
Drsinae ioco. a: Ursin, .. ... Protursus, Ursus; Helarctos, Melursus

The replacement of the Hemicyoninae by the Agriotheriinae, and the
latter in turn by the Ursinae, can be accounted for in general by assuming that
there was competition between better and lesser adapted groups. Such compe-
tition would, of course, have taken place at the species level, but would
ultimately have been manifested in higher taxonomic categories as well.

Kurtén (1957¢: 224) has concluded that the replacement of a species by an
ecologically related species may occur in one of three ways:

(1) “The extinction of the earlier form has no causal connection with the immi-
gration of the later form. Both result from the action of other factors, for
instance climatic.’

(2) ‘The extinction of one species permits the subsequent immigration of
another.’

(3) “The immigrating form is adaptively superior to the local form, and ousts
it through competition.’

While the third alternative is the one concluded to be the most generally
applicable in the present instance, it was almost certainly not the sole factor
involved in the extinction of ursids. Furthermore, competition was probably not
confined only to members of the family, and some ursids probably became
extinct as a result of competition with members of other families. This might
well have been the case with the somewhat aberrant ursid Agriotherium.

INTRA-FAMILIAL CATEGORIES WITHIN THE URSIDAE
Subfamily Hemicyoninae

Diagnosis (adapted from Pilgrim 1931). Ursidae with the upper profile of
the skull almost rectilinear; snout relatively long and narrow; infra-orbital
foramina rather remote from orbits; temporal fossae long and deep; occiput
low; sagittal and lambdoidal crests prominent; zygomatic arches relatively
narrow; P? and P? double-rooted; P* situated behind infra-orbital foramen,

I20 ANNALS OF THE SOUTH AFRICAN MUSEUM

and antero-posterior diameter equal to or slightly exceeding that of M?!; P*
with prominent protocone situated towards the midpoint of the tooth, and
parastyle absent; M! larger than M?, with the transverse diameter of these teeth
exceeding antero-posterior diameter, and with internal cusps crescentic inwards;
M$ always absent; mandible with premasseteric fossa and full complement of
teeth; P, to P, double-rooted; M, large with talonid becoming prominent;
M, smaller than M,, double-rooted with antero-posterior diameter slightly
greater than transverse diameter; M, small, single-rooted and slightly elongated
antero-posteriorly; postcranial skeleton a cursorial type, feet digitigrade;
humerus with entepicondylar foramen.

Discussion. In general there was an increase in body size with time, and
some later Hemicyoninae were as large as the biggest of modern bears. Although
probably carnivorous on the whole, at least one genus (Cephalogale) apparently
became progressively more omnivorous, and it is regarded as the ancestor of all
ursids.

Early in their history the Hemicyoninae were confined to the Old World
and only spread to North America at the peak of their radiation in the late
Miocene. Although they were a restricted group generically, they were wide-
spread and apparently very successful in the northern continents during the
Miocene. ‘They became extinct early in the Pliocene when the radiation of the
Agriotheriinae was beginning.

Subfamily Agriotheriinae
Tribe Agriotheriini

Diagnosis. Ursidae with the upper profile of the skull rather convex; snout
fairly short and broad; infra-orbital foramina close to orbits; occiput moderately
high; sagittal crest not prominent; zygomatic arches becoming broad; P#
situated below or slightly anterior to infra-orbital foramen, with antero-
posterior diameter approximately equal to that of M!; P* protocone becoming
progressively larger and sometimes with accessory cusps developed anterior
to it; P* parastyle develops and becomes progressively larger; M! roughly
square; M? sometimes with talon; M? always absent; mandible sometimes with
premasseteric fossa; M, usually large and sectorial; M, elongated; M, usually
circular or very slightly elongated; postcranial skeleton progressively more
heavily built; humerus (usually) with entepicondylar foramen.

Tribe Aitluropodini

Diagnosis. Ursidae with upper profile of skull convex; snout short and
broad; infra-orbital foramina close to orbits; sagittal crest not prominent; P!
vestigial or absent; P* large with prominent protocone and antero-internal
cusp (protocone lobe), and prominent parastyle; P* situated below infra-
orbital foramen, with antero-posterior diameter slightly greater than that of
M?; M! roughly square with prominent internal cingulum; M? elongated with

A PLIOCENE URSID FROM SOUTH AFRICA I2!I

prominent internal cingulum and talon on which multiple cusplets are deve-
loped; M? always absent; mandible without premasseteric fossa; P, vestigial;
M, large, elongated and non-sectorial; M, large and slightly elongated with
multiple cusplets on occlusal surface; postcranial skeleton robustly propor-
tioned; non-cursorial; humerus with entepicondylar foramen.

Tribe Tremarctini

Diagnosis. Ursidae with the upper profile of the skull rather convex; snout
relatively short and broad; infra-orbital foramina close to orbits; sagittal crest
sometimes prominent; zygomatic arches moderately broad; P* situated anterior
to infra-orbital foramen, with antero-posterior diameter usually less than that
of M!; P* protocone prominent and parastyle absent; M! roughly square or
slightly elongated; M? elongated with prominent talon; M? always absent;
mandible with premasseteric fossa; M, large, elongated and non-sectorial; M,
elongated, antero-posterior diameter approximately equal to that of M,; M,;
moderately large and slightly elongated; postcranial skeleton robustly propor-
tioned; non-cursorial; humerus with entepicondylar foramen.

Discussion. As here defined the Agriotheriinae are the most diverse of the
ursid subfamilies. It includes the genus Ursavus which is the earliest member of
the family which is unmistakably ‘bear-like’, and which is regarded as the stem
genus of the ursids by those who include the Hemicyoninae in the Canidae
(e.g. Kurtén 1966). It is known only in Europe.

An apparent off-shoot from Ursavus was Indarctos, a genus which is first
recorded in the early Pliocene of Europe. It is uncertain which species of
Indarctos is the earliest. One possibility is J. vireti Villalta & Crusafont 1945 from
Spain, while another is the agriotheriine from the lignites of Monte Bamboli
in Italy (Erdbrink 1953). The transition from Ursavus to Indarctos appears to be
principally a matter of an increase in size.

A number of other species of Jndarctos have been recorded in the Pliocene
of the northern continents. European species are J. arctoides Deperet 1895,
I. atticus Dames 1883 and I. ponticus Kormos 1913, while I. lagrelli Zdansky 1924
is from China, J. punjabiensis Lydekker 1884 from India and J. oregonensis
Merriam et al 1916 from North America. In general the recorded specimens of
Indarctos are rather fragmentary, and consequently definitions of the species
are often inadequate. Pilgrim (1931) noted the similarity between J. ponticus
and J. lagrelli, and Kurtén (1957c) regarded them as conspecific. Probably the
recovery of additional material and a review of the genus would result in further
synonomies being recognized.

Much the same can be said of Agriotherium, except that this genus survived
into the Villafranchian of Europe (Kurtén, 1968), and as the new record from
Langebaanweg shows (vide infra), it also became established in Africa. As far
as could be determined there is no early Pliocene record of this genus. Species
include A. insigne Gervais 1853 from Europe, A. maraghanus Mecquenen 1925

I22 ANNALS OF THE SOUTH AFRICAN MUSEUM

from Iran, A. palaeindicus Lydekker 1878 and A. sialensis Falconer & Cautley
1836 from India, and A. gregory: Frick 1921 from North America.

Probably the best published account of the differences between Agriotherium
and Indarctos is that of Pilgrim (1932). He was apparently the first person to
conclude that Agriotherium is the more advanced of the two genera, although
only in some respects and he stated that Indarctos was a development from
Agriotherium. ‘This is a traditional point of view which is still widely held, and
Agriotherium has often been referred to as a link between the Canidae and
Ursidae.

The present study led to the conclusion that the characteristics of
Agriotherium are not primitive, but rather the result of the development of
specializations. Agriothertum is regarded as a derivative of Jndarctos in which there
was a trend towards the development of more carnivorous habits. The alternative
view that Indarctos was derived from Agriotherium is rendered a little unlikely
by the known temporal ranges of the two genera, and the possibility that both
are descended from a hypothetical common ancestor is an unnecessary theory.

The general trend in ursid evolution has been towards the development of
characters suited to an omnivorous or herbivorous diet. There is at least one
well-documented reversal of this trend. The polar bear, Ursus maritimus, is a
purely carnivorous form which still retains many of the characters of U. arctos,
the species from which it is derived (Kurtén 1964). This anomalous develop-
ment may have come about during one of the Pleistocene glaciations when an
U. arctos population adapted to life in a vegetation-less peri-glacial environment.
Since the dichotomy of U. maritimus and U. arctos took place comparatively
recently, the dentitions of the two species are still essentially similar. Given
sufficient time the U. maritimus dentition would become increasingly modified,
with the carnassials developing at the expense of the molars.

Agriotherium was probably just such an exception to the general rule in
ursid evolution, but in this instance the carnivorous habits are reflected by the
nature of the dentition. The numerous references to its ‘primitive’ and ‘canid-
like’ characteristics imply ‘carnivore-like’, which is not necessarily primitive
at all. In fact Erdbrink (1953: 582) referred to the upper carnassial of Agrio-
thertum as being ‘very carnivorous in aspect’.

The most important characters which distinguish Agriotherium from Indarctos
are to be found in the dentitions, and basically the differences are centred on
the emphasis of the carnassials and the reduction of the other cheekteeth in
Agriotherium (Pilgrim 1932: 42).

In Agriotherium the anterior premolars in both maxilla and mandible are
reduced in size and sometimes number. This is not necessarily an indication of a
carnivorous diet, since many ursines also have the anterior premolars reduced
or lost, but in this group the molars are correspondingly enlarged and the
reduction also affects the carnassials. This is not so in Agriotherium.

This genus differs from Indarctos in having the lingual margins of M! and
M? shorter than buccal margins as a result of the paracones and metastyles

A PLIOCENE URSID FROM SOUTH AFRICA 123

being more strongly developed than the protocones and hypocones. Since the
buccal cusps are higher than lingual ones it is possible for them to act as shearing
as well as crushing agents. The crushing function of the M? of Indarctos is further
indicated by the presence of a talon, and in this genus the M? is always longer
than M1?. By contrast the M? of Agriotherium is nearly always smaller than M1
and lacks the talon. Erdbrink’s (1953: 571) conclusion that there is ‘at best a
beginning of a talon... in A. insignis’ is probably incorrect, and the talon in
this species is regarded as vestigial.

The ‘carnivorous aspect’ of the upper carnassial of Agriotherium has already
been mentioned, and this tooth is also the most important in so far as the inferences
on the ancestry of the genus are concerned. It is characterized by the presence
of a prominent parastyle, a cusp which is not found in any of the Canidae,
Hemicyoninae or species of Ursavus. It is, however, present but small in some
species of Indarctos (e.g. I. punjabiensis). ‘This suggests a progressive development
of a P* parastyle in the Agriotheriinae as follows:

Ursavus (absent) —Indarctos (small) —Agriotherium (prominent).

In Azluropoda, which is here regarded as another descendent of Indarctos, it is
also prominent.

The lower carnassial of Agriothertum also exhibits ‘carnivorous’ charac-
teristics. The talonid is reduced relative to the trigonid and the hypoconid is
higher than the entoconid, which makes it a more efficient shearing tooth than
that of Indarctos.

The reasons for the development of an apparently carnivorous lineage
stemming from Jndarctos are not known, but this might have been in response to
competition with early Ursinae. By the late Pliocene when the ursine radiation
was getting under way in Europe, Agriothertum was the only agriotheriine sur-
viving in this area and it only became extinct in the early Pleistocene. It follows
that if it was indeed a purely carnivorous form, its ultimate extinction in
Europe and elsewhere cannot be explained in the same way as the extinction or
limitation of other Agriotheriinae. In this instance competition with other
fissiped carnivores may be the answer, although other undetermined factors
might have been involved.

In the late Pliocene Jndarctos was still present in Asia, an area in which the
Ursinae had not yet become common, and it is only in Asia that Ailuropoda is
recorded. ‘Those authors who have accepted Ailuropoda as an ursid have invari-
ably suggested its descent from the Agriotherium|Indarctos group (see Davis 1964),
and JIndarctos appears to be the only known fossil form from which Azluropoda
can be satisfactorily derived. Since Ailuropoda was well established early in the
Pleistocene, and since it is a good deal more advanced than Indarctos, its differen-
tiation must have taken place during the Pliocene, probably at about the time
that Indarctos itself was nearing extinction.

Much of the controversy about the status of Ailuropoda seems to stem from
the fact that it is almost always compared with modern ursine bears, from
which it does indeed differ quite markedly. However, if it is taken into account

124 ANNALS OF THE SOUTH AFRICAN MUSEUM

that their common ancestor was a Miocene form (Fig. 1), the differences are
hardly surprising. The differences between the European early Pleistocene
Ursus minimus and Agriotherium insigne are as great, or even greater, than those
between modern ursines and Azluropoda, yet the referral of Agriotherium to the
Ursidae is no longer questioned.

In order to illustrate that on dental evidence alone Indarctos could be
ancestral to both Ailuropoda and Agriothertum, a list of some characters of the
upper cheekteeth of these three genera is given in Table 1. On the one hand the
dentition is modified for a herbivorous diet (Azluropoda), and on the other a
carnivorous dentition is developed (Agriotherium).

TABLE I. Some characters of the upper dentitions of Ailuropoda, Indarctos and Agriotherium.

Ailuropoda Indarctos Agriotherium
Herbivorous <———————_ Ancestral genus -————————->_ Carnivorous
lineage lineage

Anterior premolars
P! sometimes absent P! present P! present (?)
P? & P® double-rooted P? & P® double-rooted in P? & P single-rooted
early forms (?)
Carnassial
P* with prominent parastyle P* parastyle absent or P* with prominent
and antero-internal cusp small, antero-internal parastyle, antero-
cusp small internal cusp usually
absent
Molars
M! square with prominent M! square, four main M! narrower lingually,
lingual cingulum, four main cusps only four main cusps only
cusps and smaller cusplets
M? elongated with prominent M? slightly elongated with M? nearly square, talon
talon, four main cusps and small talon and four vestigial or absent,
many cusplets main cusps usually four main cusps
only

The range of Azluropoda diminished considerably during the Quaternary
and it now survives in a natural state only in isolated areas in China. During
the Pleistocene it was widely distributed in China, and is also recorded from
Burma (Smith-Woodward 1915). It was during the Pleistocene that the Asiatic
radiation of the Ursinae took place and this suggests that Azluropoda may have
been an unsuccessful competitor with this group. It is therefore another agrio-
theriine whose decline is attributed to the Ursinae.

The decision to include the tremarctines in the Agriotheriinae is not easily
justified. Superficially at least, there are similarities between the extant Ailuro-
poda melanoleuca and Tremarctos ornatus, and both differ from Ursus. There are
resemblances in general skull morphology, both being ‘short-faced’ forms,
and Davis (1955: 29) states that, ‘Except for the pre-masseteric fossa, the fea-
tures that distinguish the skull of Tremarctos from the skull of Ursus, although
much less exaggerated, are similar to the features that distinguish the skull of the

A PLIOCENE URSID FROM SOUTH AFRICA 125

giant panda (Ailuropoda)’. Kurtén (1967) mentioned the similarity between the

tremarctine Arctodus and Indarctos.
However, the teeth of the tremarctines are much closer to those of Ursus

than any other agriotheriine. The P* lacks a parastyle and the molars are
similar to those of Ursus.

Its ancestry may lie with the agriotherune Ursavus, and the problematical
(?) Ursavus pawniensis Frick 1926 from the North American Miocene may be the
ancestral form. Erdbrink (1953) suggested that the tremarctines are not a
homogeneous group, and he derived Tremarctos from ursine stock, but the
‘arctotheres’ (Arctodus) from Indarctos. However, Kurtén (1966: 7) found that
although the ‘earlier history of Arctodus is poorly documented . . . there can be
little doubt that it is a tremarctine’.

Another significant characteristic of tremarctines is the entepicondylar
foramen of the humerus. This is a feature also present in the humerus of
Indarctos (e.g. I. oregonensis), Ailuropoda and the Hemicyoninae from which the
Agriotheriinae are derived. It is, however, not present in the humerus of the
Ursinae. It is here regarded as a primitive characteristic retained in at least
two agriotheriine lineages (Indarctos—Ailuropoda and ?Ursavus—'Tremarctini),
but lost in the Ursinae and perhaps also the Indarctos— Agriothertum lineage.

The conclusion reached here is that the tremarctines do belong in the
Agriotheriinae, having stemmed from an Ursavus-like ancestor, and having
paralleled the Ursinae in some respects.

As with Aizluropoda, the only surviving tremarctine, Tremarctos ornatus,
occurs isolated from the Ursinae, in this instance in South America. Both
Ailuropoda and the tremarctines co-existed with ursines for much of the Pleisto-
cene, and in the case of the tremarctines for part of the Pliocene as well (Bjork
1970), so their inferred replacement by the ursines was a slow process. However,
the fact remains that they were definitely in decline by the end of the Pleistocene,
whereas the Ursinae were still remarkably successful. But for the advent of
human civilization the Agriotheriinae might well have become extinct while
the Ursinae might have remained a prominent part of the world’s fauna.

‘Subfamily Ursinae
Diagnosis (see Pilgrim 1931).

Discussion. There is an extensive literature on modern and fossil ursids
and a substantial proportion is devoted to the Ursinae. It is the best known and
least controversial of the ursid subfamilies and only in the case of the sun bear,
Helarctos, and the sloth bear, Melursus, are there any real doubts about ancestry.
The subfamily apparently stems from the early Pliocene Protursus (Kurtén
1971), and the genera Ursus and Helarctos are first recorded in the late Pliocene,
while Melursus is known only from the Quaternary.

Four categories may be distinguished within the subfamily. The first two
are the Helarctos and Melursus groups, both of which are represented by a single

126 ANNALS OF THE SOUTH AFRICAN MUSEUM

extant species, and in neither case is there a good fossil record. The genus
Ursus can conveniently be divided into two groups. The first comprises
U. americanus and U. thibetanus, the North American and Asiatic black bears,
and the second is the brown bear group, U. arctos and related forms. The latter
includes the polar and grizzly bears as well as a number of extinct species such
as the giant U. spelaeus. ‘They are an extremely successful group and at one
time or another they have been distributed through much of the Northern
Hemisphere, including the Arctic and North Africa.

THE LANGEBAANWEG BEAR

Agriotherium africanum n.sp.

Holotype. A left maxillary fragment with P* (South African Museum
No. L 2045).

Referred material. A part of an ulna (L 2154) and isolated teeth as follows:
L 1868A—E: I, I,, ?P? and parts of P* and M1}.
1.126972, Ve.
L 1844 & L 3141: I, and I.
L 12561: Ms.

Locality. All the specimens are from ‘E’ Quarry, Langebaanweg.

Diagnosis. A species of Agriothertum of large size, in which the P* has a
prominent parastyle and a well-developed protocone lobe, the latter consisting
of the protocone, an antero-internal cusp and a small intermediately situated
cusp; the protocone lobe projects and is flattened posteriorly where it functions
as a shearing surface additional to that of the paracone and metastyle. The M?
is smaller than M! and is without a talon. The antero-buccal surface of M, is
inflated.

Etymology. The specific name is given in recognition of the fact that this
is the first agriotheriine recorded from Africa.

DESCRIPTION

Only two of the specimens, the ?P? and M, are complete in all respects,
while the referred P*4, M1 and M2? are so poorly preserved that not a single
standard measurement could be taken on them. The latter are important,
however, since they do give an indication of the morphology of the teeth con-
cerned. In general the Agriotheriinae are not well represented in the fossil
record, but the Langebaanweg species can be less adequately defined than most
of the recorded species of the subfamily.

The ?P? (L 1868C) (Plate 19 F, G) is referred to this species since it was
found in association with the other L 1868 specimens, which unquestionably do
belong to Agriotherium. In size (10,8 Xx 7,8) it is comparable to the P? of an
Indarctos atticus specimen described by Thenius (1959)), and it resembles this

a

A PLIOCENE URSID FROM SOUTH AFRICA 127

tooth in being broadest anteriorly. It is a simple, low-crowned tooth with a
barely perceptible principal cusp from which arise keels, one running posteriorly
and the other antero-internally. The crown is supported by a single antero-
posteriorly elongated root.

The carnassial fragment L 1868D is incomplete, but what remains matches
corresponding parts of the holotype P* (Plate 19 A-E). This tooth differs in
some respects from those of previously described specimens of Agriothertum and
Indarctos, although in size (Table 2) and general appearance it is similar to the
P* of these genera.

TABLE 2. Dimensions of the P* of some species of Agriotherium

A. insigne A. sp. A. palaeindicus | A. sivalensis
France Spain India India
(1) (1) (2) (2)
length . : ; 29,1 30,0 28,0 33,0
breadth : 21,0 23,0 21,0 19,8
A, gregoryi A. africanum
N. America South Africa

(r)
UC 24027. UC 24025 AM 18121A L 2045

length sr. 20,5 354 36,5 €32,5

breadth . 4 21,7 25,0 25,0 2555

(1) Frick (1926)
(2) Lydekker (1884)

The crown consists of a parastyle, which is damaged, paracone and meta-
style, which are flanked lingually by a large protocone lobe made up of a
protocone, antero-internal cusp and a small, intermediately situated cusp.
There are two roots on the buccal side of the tooth and another supporting
the protocone lobe. Although the parastyle is damaged, sufficient remains to
indicate that it was large and made up about 25% of the total length of the
tooth. In this respect it is typical of the P* of Agriotherium in which the parastyle
is always large, whereas in Jndarctos it is usually not as well developed. The para-
cone and metastyle are approximately equal in length and make up the
remaining 75% of the total length of the tooth. Shear facets have been worn
on the lingual surfaces of these cusps.

The Langebaanweg P* differs most markedly from previously described
Agriotherium and Indarctos carnassials in the size and morphology of the proto-
cone lobe. Its length (22,5 mm) can be measured accurately since its anterior

128 ANNALS OF THE SOUTH AFRICAN MUSEUM

and posterior limits are clearly defined. In most other Agriotheriinae this is not
the case as the posterior limit of the protocone merges gradually with the lingual
surface of the metastyle, but in any case they all have shorter protocone lobes.
In addition, the protocone lobe of L 2045 differs from other species of Agrio-
therium in that it has a fairly prominent antero-internal cusp, although this cusp
is present in Jndarctos. It is small in J. punjabiensis,* but quite large in I. lagrelli
(Zdansky 1924) and J. atticus (Thenius 1959)).

The protocone itself is unique in that instead of being conical, it has its
apex elongated antero-posteriorly and compressed towards the paracone and
metastyle. The elongation of the protocone lobe as a whole is largely due to
the shape of the protocone. The functional advantage of this elongation is
readily evident, since the posterior part of the protocone has developed on it a
shear facet which is supplementary to that of the paracone and metastyle.
Of all the Indarctos and Agriotherium upper carnassials presently known,
that of the Langebaanweg species seems the best adapted to a shearing
function.

Another unusual feature of the protocone lobe is the small cusp situated
between the antero-internal cusp and the protocone. The cusp itself has been
all but worn away, but its presence is marked by a circular patch of exposed
dentine. It, and the most anterior part of the protocone have almost horizontal
wear facets, indicating that the P* served a crushing function as well.

Parts of the enamel of this tooth show the ‘wrinkling’ or rugosity said to be
characteristic of Agriotherium (Erdbrink 1953).

In its general morphology the P* of the Langebaanweg species is not
dissimilar to that of Azluropoda.

A small part of the maxilla of the holotype is preserved. The most anterior
part of the alveolus of M!, and part of the antero-external root of this tooth are
present. ‘The M! must have had a transverse diameter of at least 30 mm, which
is in keeping with the size of this tooth in Agriotherium and Indarctos. The inferior
margin of the infra-orbital foramen is also present and it is situated above and
slightly posterior to the P*.

Little of the M! (L 1868E) is preserved. Parts of the roots supporting the
paracone and protocone are present, and that root beneath the protocone is
large, antero-posteriorly elongated and inserted at an angle to the plane of the
palate. The other preserved root is smaller, transversely elongated and inserted
vertically into the maxilla. It presumably matched the now missing root which
supported the metastyle. Most of the crown is lost and the only enamel preserved
is near the paracone. Judging from the preserved parts of the crown and the
roots, this tooth appears to have been narrower lingually. The transverse
diameter is estimated to have been 30 mm, which is comparable to the figure
inferred for the missing M? of the holotype. The antero-posterior diameter
must also have been about 30 mm.

* This cusp was shown in Lydekker’s (1884) illustration, but others (e.g. Matthew 1929)
apparently overlooked it.

A PLIOCENE URSID FROM SOUTH AFRICA 129

The M2? (L 12637) is an important specimen, since although it is incom-
plete, its morphology indicates that the affinities of the Langebaanweg agrio-
theriine lie with Agriotherium rather than Jndarctos. It consists of a paracone and
metastyle which are equal in size, situated parallel to a protocone and hypocone
which are also similar in size. The latter cusps are lower than the paracone
and metastyle. The enlarged and posteriorly elongated talon which charac-
terizes the M? of Indarctos is not in evidence. This tooth is appreciably smaller
than the M! and its dimensions are estimated to be 25 X 25 mm. It is thus
smaller than the M? of previously described species of Agriotherium (see Frick
1926: 81).

A reconstruction of the posterior upper dentition of the Langebaanweg
agriotheriine is illustrated in Figure 2.

Fic. 2. A reconstruction of the posterior upper dentition of Agriotherium africanum based on the
specimens L 1868C, L 2045, L 1868E and L 12637 (Natural size).

Little can be said of the lower incisors which are preserved (L 1868A, B,
L 1844, L 3141) (Plate 20 A) other than that they are large and agree in all
morphological respects with the corresponding teeth of Jndarctos lagrelli (Zdan-
sky 1924) and an Jndarctos specimen from Samos (Helbing 1932). No descrip-
tions or illustrations of the lower incisors of Agriotherium could be located, but
presumably they are essentially the same as those of Indarctos. |

The M, (L 12561) (Plate 20 B) is a single-rooted and low-crowned tooth
with an almost circular and flat occlusal surface. The antero-buccal part of the
crown is inflated and there is a wear facet in this region angled from the occlusal
surface across the inflation towards the cingulum. This presumably results
from occlusion with the lingual surface of the paracone or metastyle of M?.
This is an indication that the post-carnassial teeth of this species functioned as
shearing as well as crushing agents. The M, measures 16,5 x 16,9 mm.

The ulna (L 2154) (Plate 20 C, D), which lacks the distal end, lower
part of the shaft and anconeus process, is far too large to be confused with the
ulna of any other carnivore species in the Langebaanweg assemblage. It com-
pares closely in size (Table 3) and morphology with the ulna from Pikermi
referred to Indarctos atticus by Pilgrim (1931).

130 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 3. Dimensions of the Langebaanweg Agriotherium ulna, compared with that of an
Indarctos (?) cf. atticus specimen from Pikermi (Pilgrim, 1931).

L 2154 Pikermi

Dorso-ventral diam. at coronoid process : 87,0 85,0

Transverse diam. at coronoid process . : 59,0 61,0

Transverse diam. at proximal end : : 50,0 52,0
DISCUSSION

The genera Agriotherium and Indarctos share many dental and osteological
characteristics, but it is clear that the Langebaanweg agriotheriine has greater
affinities to the former genus. It is regarded as a species distinct from those
previously recorded since it exhibits certain apparently unique characteristics,
and in addition it is the most geographically isolated record of the genus.

Agriotherium africanum differs from previously described species in the size
of M? and in the nature of its P* and Mg. All the species of Agriotherium are
known from single, or perhaps a few individuals and it has therefore not been
possible to assess the range of variation in any of them. However, R. H. ‘Tedford
(pers. comm.) has found ‘considerable variation [in the P4] within and among
populations of Hemphillian Agnotherium from the United States’, although
none of the North American specimens matched the A. africanum P*. If the
Langebaanweg species is conspecific with a known species, then it is likely that
it would be one of the Eurasian forms, which presumably also had variable
upper carnassials.

In this connection the geographical location of A. africanum is probably
significant. Of the 18 species of fissiped carnivores known from Langebaanweg,
only four have affinities with contemporary Eurasian species. Much the same
applies to the non-carnivorous mammals. Consequently it is probable that
although there is a general similarity between the late Pliocene mammal faunas
of Eurasia and Africa, each area was represented by its own lineages. For
example, although the Langebaanweg Percrocuta is fairly similar to the Eurasian
P. eximia, it is sufficiently different to warrant the status of a separate species.
Similarly the boselaphine from Langebaanweg resembles Tragoportax salmontanus
from the Siwaliks of India, but the two are clearly not conspecific.

Even if larger numbers of individuals of the Eurasian species of Agriotherium
become available in the future, it seems unlikely material matching that from
Langebaanweg will be recorded. It was on this basis that the decision was made
to refer the Langebaanweg Agriotherium to a new species.

Probably it is just a matter of time before more agriotheriine remains are
recovered elsewhere in Africa, especially in view of the attention presently
being focused on Pliocene deposits in East Africa. It is also possible that more
material of A. africanum will be found at Langebaanweg, since some of the
deposits from which present specimens were derived remain unexcavated.
Consequently more adequate definition of A. africanum might still be possible,
and its phyletic relationships might yet be more accurately determined.

A PLIOCENE URSID FROM SOUTH AFRICA 131

SUMMARY

An account of the family Ursidae (Mammalia: Carnivora) is given and a
new ursid species, Agriotherium africanum, is described.

ACKNOWLEDGEMENTS

The first draft of this manuscript was completed in 1967 and at that time
I benefited greatly from correspondence with Dr. Bjérn Kurtén (University of
Helsinki) and Dr. Richard H. Tedford (American Museum of Natural History).
Both were very generous in sharing their knowledge of the Ursidae, and any
merits which this paper may possess are due largely to them. Its shortcomings
are, however, entirely of the author’s own making.

The current investigations at Langebaanweg are being supported by the
South African Council for Scientific and Industrial Research, Chemfos Ltd.
(a subsidiary of the African Metals Corporation) and Shell South Africa (Pty.)
Ltd. ‘The Wenner-Gren Foundation for Anthropological Research, New York,
provided the vehicle used in the field work at Langebaanweg (Grant no.

2752-1834).
I am indebted to the management of Chemfos Ltd, and also Mr. H.
Krumm and Mr. G. Benfield for their unfailing assistance in the recovery of

fossils from the quarries at Langebaanweg.

REFERENCES

Byork, P. R. 1970. The Carnivora of the Hagerman Local Fauna (Late Pliocene) of south-
western Idaho. Trans. Am. phil. Soc. 60: 1-54.

Davis, D. D. 1955. Masticatory apparatus in the spectacled bear Tremarctos ornatus. Fieldiana,
Kool. 37: 25-46.

Davis, D. D. 1964. The giant panda. A study of evolutionary mechanisms. Fieldiana, Zool.
Mem. 3: 1-339.

Erppsrink, D. P. 1953. A review of fossil and recent bears of the Old World. Deventer: Jan de Lange.

Frick, C. 1926. The Hemicyoninae and an American Tertiary bear. Bull. Am. Mus. nat.
Hist. 542 1-119.

Gentry, A. W. 1970. The Bovidae (Mammalia) of the Fort Ternan fossil fauna. Fossil vertebrates
of Afr. 23 243-323.

Hesinc, H. 1932. Uber einen Indarctos-schadel aus dem Pontien der Insel Samos. Abh. schweiz.
paldont. Ges. 52: 1-18.

HENpDEy, Q. B. 1969. Quaternary vertebrate fossil sites in the south-western Cape Province.
S. Afr. archaeol. Bull. 24: 96-105.

HENDEY, Q@.B. 1970a. A review of the geology and palaeontology of the Plio/Pleistocene deposits
at Langebaanweg, Cape Province. Ann. S. Afr. Mus. 56: 75-117.

HeEnbDEy, ©. B. 1970b. The age of the fossiliferous deposits at Langebaanweg, Cape Province.
Ann. S. Afr. Mus. 56: 119-131.

HEnpbEy, Q.B. 1972. The evolution and dispersal of the Monachinae (Mammalia: Pinnipedia).
Ann. S. Afr. Mus. 59: 99-113.

HEnpDeEy, Q.B. & RepEnninc, C. A. 1972. A Pliocene phocid from South Africa. Ann. S. Afr.
Mus. 59: 71-08.

Hooyer, D. A. A late Pliocene rhinoceros from Langebaanweg, Cape Province. Ann. S. Afr.
Mus. 59- (In press.)

Kracuievicu, L. 1926. Los arctoterios norteamericanos (Tremarctotherium n. gen.) en relacién
con los de Sud América. An. Mus. nac. Hist. nat. B. Aires 33: 1-16.

132 ANNALS OF THE SOUTH AFRICAN MUSEUM

KurtTEn, B. 1957a. Percrocuta Kretzoi (Mammalia, Carnivora), a group of Neogene hyenas.
Acta zool. cracov. 2: 375-404.

KurtTEn, B. 1957). The bears and hyenas of the Interglacials. Quaternaria 4: 1-13.

KurtEn, B. 1957c. Mammal migrations, Cenozoic stratigraphy, and the age of Peking Man and
the australopithecines. 7. Paleont. 31: 215-227.

KurtEn, B. 1958. Life and death of the Pleistocene cave bear. Acta zool. fenn. 95: 1-59.

KurtEn, B. 1964. The evolution of the polar bear, Ursus maritimus Phipps. Acta zool. fenn.
108: I-30.

KurtEn, B. 1966. Pleistocene bears of North America. 1. Genus Tremarcios, spectacled bears.
Acta zool. fenn. 115: I-120.

KurtEn, B. 1967. Pleistocene bears of North America. 2. Genus Arctodus, short-faced bears.
Acta zool. fenn. 117: 1-60.

KurtEn, B. 1968. Pleistocene mammals of Europe. London: Weidenfeld & Nicolson.

KurteEn, B. 1971. The age of mammals. London: Weidenfeld & Nicolson.

LyDEKKER, R. 1884. Indian Tertiary and post-Tertiary Vertebrata. Siwalik and Narbada
Carnivora. Palacont. indica (10) 2: 178-354.

Mactuio, V. J. & HENDEy, ©. B. 1970. New evidence relating to the supposed stegolophodont
ancestry of the Elephantidae. S. Afr. archaeol. Bull. 25: 85-87.

MattTHew, W. D. 1929. Critical observations upon Siwalik mammals. Bull. Am. Mus. nat.
Hist. 56: 437-560.

Pitcrm, G. E. 1931. Catalogue of the Pontian Carnivora of Europe. London: British Museum
(Natural History).

Pitcrim, G. E. 1932. The fossil Carnivora of India. Palaeont. indica (n.s.) 18: 1-232.

Reep, C. A. 1967. The generic allocation of the hominid species habilis as a problem in
systematics. §. Afr. F. Sct. 63: 3-5.

Simpson, G. G. 1945. The principles of classification and a classification of mammals. Bull.
Am. Mus. nat. Hist. 85: 1-450.

Simpson, G. G. 1971. Fossil penguin from the Late Cenozoic of South Africa. Science 171:
1144-1145.

SmiTH-Woopwarp, A. 1915. On the skull of an extinct mammal related to Aeluropus from a
cave in the ruby mines at Mogok, Burma. Proc. zool. Soc. Lond. 1915: 425-428.

TuHeENtus, E. 19594. Ursidenphylogenese und Biostratigraphie. <. Sdugetierk. 24: 78-84.

THENtIus, E. 1959). Indarctos arctoides (Carnivora, Mammalia) aus dem Pliozan Osterreichs nebst
einer Revision der Gattung. Neues Fb. Geol. Paldont. Abh. 108: 270-295.

TuEntus, E. 1966. Zur Stammesgeschichte der Hyanen (Carnivora, Mammalia). <. Sdugetierk.
31: 293-300.

ZDANSKY, O. 1924. Jungtertiare Carnivoren Chinas. Paleont. sinica (C) 2: 1-149.

Ann. S. Afr. Mus., Vol. 59

Plate 19

G vent E

A-E Buccal, anterior, occlusal, lingual and oblique views of the Agriotherium africanum holotype
L 2045.

F & G Lingual and occlusal views of the ?P? of Agriotherium africanum L 1868C.

Ann. 8. Afr. Mus., Vol. 59 Plate 20

C D

A Lingual view of the I, of Agriotherium africanum L 3141 and L 1868A.
B Occlusal and posterior views of the M, of Agriotherium africanum L 12561.
C&D Anterior and medial views of the ulna of Agriotherium africanum L 2154.

INST RUCTLIONS,TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm x 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

Bu.iLoucu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHER, P.-H., DuvAu, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 19602. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, b; Liste: 11. Turton, 1932: 80.

Ps
| ors ra rl
ee Pte f

Bo io 8

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
May 1972 Mei
Part 7 Deel

A NEW SPECIES OF PARADOXOSTOMA
(CRUSTACEA, OSTRACODA)
FROM SOUTH AFRICA

By
K. G. McKENZIE

Cape Town Kaapstad

<All HSON;>
\

( JUN 211972 )
\ ‘i f

\ / p
- 7 nT “+i -* FA
S/O RAR it _—- —

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 2(a5 3> 5s 7-8); Stas 5» ile) 51-2, 5, T)s
6(1, tpi), 7(1-3), 8, (1-2), 10(r),
GR Ss 7h eS) I) Si Se), oe

Price of this part/Prys van hierdie deel
Ro,60

Trustees of the South African Museum © Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 07 0

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

A NEW SPECIES OF PARADOXOSTOMA (CRUSTACEA, OSTRACODA)
FROM SOUTH AFRICA

By
K. G. McKenzie

British Museum (Natural History), London
(With 12 figures)

[MS'.. accepted 30 November 1971]

CONTENTS
PAGE
Introduction . : : : : ; : 133
Systematics . . é : 2 : ‘ 133
Summary ; - 4 : : 3 : 137
References ‘ : : F : ‘ : m7
INTRODUCTION

During a working visit to the South African Museum in September 1970,
I was shown a marine sample which had been collected at Sea Point, near
Cape Town, by the late Dr. K. H. Barnard and which included a large popula-
tion of a Paradoxostoma species. This species has proved to be new.

I am very grateful to the Director of the South African Museum, Dr. T. H.
Barry, for making its facilities available; to Mr. B. F. Kensley, Curator of
Crustacea, who drew my attention to the sample; to the National Institute
for Water Research, Pretoria, and to my own Museum for financial support;
and to Mr. D. Goode, the Transvaal Museum, who inked my original drawings.

Types are stored at the South African Museum under register number
SAM A r1or4 and some paratypes are at the British Museum (Natural History)
under register number BM(NH) 1971.10.13.1-25.

SYSTEMATICS
Paradoxostoma kensleyi n. sp.

Figures 1-12
Derwation of name

For Mr. B. F. Kensley, who drew my attention to the sample and who has
contributed several papers on the crustaceans of South Africa.

Diagnosis
In lateral view, carapace ovate-subtriangular; of medium size (length up

to about 0,65 mm); smooth; without conspicuous colour patches in the speci-
men dissected (possibly, such coloration was present in life but disappeared

133
Ann. S. Afr. Mus. 59 (7); 1972: 133—137, 12 figs.

134. ANNALS OF THE SOUTH AFRICAN MUSEUM

Paradoxostoma kensleyi n. sp., ovigerous 9, paratype

Fic. 1. Internal view of right valve, x 300. Fic. 2. Antennule, x 750. Fic. 3. Antenna with
lobate antennal gland, x 750. Fic. 4, Mandible coxale, x 750.

NEW SPECIES OF PARADOXOSTOMA (CRUSTACEA, OSTRACODA) 135

following preservation) ; anterior margin subacuminate anteroventrally; dorsal
margin strongly convex with a weak anterodorsal flexure in the right valve;
posterior margin subacuminate posterodorsally, broadly rounded ventrally;
ventral margin weakly inflexed in the vicinity of the oral cone; greatest height
medial and about 60% of the length. In dorsal view, compressed; evenly
elliptical. Internally, inner margin regular} line of concrescence also regular;
vestibule continuous; radial pore canals few (about 7 anteriorly and 5 pos-
teriorly) unbranched, short and straight; normal pore canals fairly numerous,
scattered, simple; hinge adont or modified adont, with a weak terminal posterior
projection in the right valve and a corresponding accommodation in the left
valve; muscle scars comprising four adductors in a subvertical series, others not
observed (Fig. 1). Garapace sex dimorphism weak.

Antennule (Ar) 6-segmented; length ratios of the last four segments
18:27:13:3 (Fig. 2). Antenna {A2) 5-segmented, the penultimate segment
appears to be sutured in its proximal half; length ratio of the terminal claws is
about 3:2; the flagellum (Spinnborste) extends beyond the tips of the claws
and is jointed at about ¢ its length from the proximal end; the gland to this
Spinnborste is large, and lobate proximally (Fig. 3). Mandible with a styliform
coxale (Fig. 4); palp two-segmented with 5 terminal bristles (Fig. 5). Oral
cone present, with the characteristic suctorial modification (Fig. 6). Maxilla
lacking a palp, trilobate (one lobe hidden in Fig. 7); epipod with about 13
Strahlen and with two downwards-pointing setae. First thoracic leg (P1)
pediform, four-segmented; protopod armed with a powerful dorsodistal claw-
like spine (Fig. 8). Second and third thoracic legs (P2 and P3) also pediform
but with dorsodistal bristles instead of claw-like spines on their protopods (Figs
g, 11). None of the terminal claws on these legs are strongly spinose. Posterior
of the body (2) extended into a caudiform process with a terminal spine
(Fig. 10). Hemipenis of male as illustrated (Fig. 12).

Material

A very large population comprising numerous mature individuals and
juveniles of both sexes.

Locality
Sea Point, near Cape Town, Republic of South Africa.

Collector and date collected
The late Dr. K. H. Barnard; March 1928.

Dimensions
Holotype (3) Length=o,52 mm; Height=o,31 mm; Breadth=o,18 mm.
Allotype (2) Length=o,56 mm; Height=o0,34 mm; Breadth=o,20 mm.

Discussion

Of the previous workers on Recent South African marine Ostracoda
(Brady 1880; Miiller 1908; Klie 1940; Benson & Maddocks 1964) only Klie

136 ANNALS OF THE SOUTH AFRICAN MUSEUM

Paradoxostoma kensleyi n. sp., ovigerous 2, paratype
(same specimen as in Figs 1-4)
Fic. 5. Mandible palp, x 750. Fic. 6. Oral cone with suctorial disc, x 750. Fic. 7. Maxilla,
two lobes plus the downwards-directed setae, X 750. Fic. 8. P1, x 750. Fig. 9. P2, x 750. Fic.
10. Posterior of body, X 750. Fic. 11. P3, x 750.

Paradoxostoma kensleyi n. sp., mature g, paratype
Fic. 12. Hemipenis, muscles only illustrated in the upper region, xX 750.

n= See

NEW SPECIES OF PARADOXOSTOMA (CRUSTACEA, OSTRACODA) 197

described any paradoxostomatids. He keyed six species in a Table (Klie 1940:
447) from which it appears that P. kensleyi is closest to the Klie species P.
auritum and P. reflexum in characters based on the first four paired limbs (anten-
nule, antenna, mandible, maxilla) and because it lacks pilosity on the ventral
margin of the P3 third segment. But P. kensleyi differs from both these species in
maximum size and has a different shape to that of P. reflecum. Apart from the
maximum size difference, P. kensleyi has a different hemipenis to that of P.
auritum and although similar in general carapace shape also appears to have a
different line of concrescence (Klie 1940: 444). Another similar species is
P. hypselum Miiller 1908, which was described from the sub-Antarctic. I have
recently determined a specimen which probably belongs to this species (USNM
137380) and the spines on the distal claws of the thoracic legs, for the P3 in
particular, are distinctive, as pointed out by Miiller (1908: 118, 119). Such
distinct spines do not feature on the distal claws of the thoracic legs in kensleyt.
Further, hypselum (length 9 0,72 mm, ¢ 0,68 mm) is a slightly larger species
than kensleyz.

Summarizing, the known paradoxostomatid fauna of South Africa now
comprises 7 species, namely: Paradoxostoma caeruleum Klie 1940, P. griseum Klie
1940, P. angustissimum Klie 1940, P. auritum Klie 1940, P. reflexum Klie 1940,
P. semilunare Klie 1940 and P. kensleyi n.sp. It is likely that Dr. G. Hartmann,
of the Zoologisches Museum und Staatinstitut, Hamburg, will describe further
species when he monographs his large Recent South African collections.

SUMMARY

Paradoxostoma kensley, a new marine ostracode collected near Cape Town,
South Africa, is described and compared with previously described South
African paradoxostomatids.

REFERENCES

Benson, R. H. & Mappocks, R. F. 1964. Recent ostracodes of Knysna Cape Province, Union
of South Africa. Paleont. Contr. Univ. Kans. 34 (Arthropoda 5): 1-39.

Brapy, G. 8. 1880. Report on the Ostracoda dredged by H.M.S. Challenger during the years
1873-1876. Rep. Voy. Challenger 1873-76. x (Zoology 3): 1-184.

Kuz, W. 1940. Beitrage zur Fauna des Eulitorals von Deutsch-Siidwest-Afrika. II. Ostracoden
von der Kiiste Deutsch-Siidwest-Afrikas. Kieler Meeresforsch. 3: 404-448.

MULLER, G. W. 1908. Die Ostracoden der Deutschen Siidpolar-Expedition 1901-1903. Dé.
Stidpol.-Exped. 10: 52-182.

BY ;

- Par “” 4
bots fica ;

a7 - = iva!

1 on : MAd ‘ ar
ee ee ee
1 i) 2WeED ia ef wah} yt

a f * mes . Pipa en

> <2 )
7
a] 7
e
—— c ile
‘ ay Bow ~~
r "4 tc
AITPN2OKO barsq)
e3 : ‘
_ 2
~ 1209 { —T
Tae PREY aa\b4
f 7 7
Pi. aathe ."
C4 Sani *)
es yeti i} ”
=) ae => —
: ¢ } *
SS See
- . ; eZ aa
{ J Pes sf,
= 4 =?
n ' oo,
i . y ‘
as A

O owy ~
e;
-
err
é 7% e
*
x
5 ~ ty
~ a
,
§ a
¢
a ’ if?
» : hk
’ } rt t.
a bw
{
‘ a
od 7
H Se
Min
F Hs
.
« os
Es ‘
=,
‘Ad
hd a
»
o—
J
, L- 7
- : J
ae
i 5 ve
h Te
* 4 r,
v
q ae |
cp) ‘hh
at
at ,
: » VUE «
. Vw
ii f bs dei 7, 8 ”
4 ‘ee
j i
tale a 7
7 L

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE, 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

x

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm X 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination,

Examples (note capitalization and punctuation)

Bu.LLoucu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, b; Liste: 11. Turton, 1932: 80.

7 ci -.,. © Pe et eee
- y Sout of | ® ; a fo y ath ah ;
Cwat Chi. evi a
os ‘ie es:
; P =. é
: , ‘ ' a ohn 3 A ee
é S . "
_ ! : A "a Rea p
2 nN i x ' i r ~
ot Phar? Whe
tar. ;
Te . } = :
i we : Any
bees 7 £ ei a ' ‘
= ne =)
feneeine heneaes e suotieal astm
‘ i a feet : 3) i S os ge pee Nas

Misha guivelet nd ‘tl boys, anigessn
to (oleetbe [aisodum to (elonmen 4
sbsivis edt i (esictoes eter syne hareraniting oe

: “dst i egeribas A pele Babvy, aoikivib feqionitey atte '
| Mime (h)\ babiovs od ca eganibaod Ye , . Bn

g elope (8) shores to gikkentysl ‘ot = fd oe i

rE as Bee WTR
= x a toe : he ae ye
f a SS Note sai eyey otaey Soe Scie a" sor : lepagi ad oF (uy ‘baw!
; pte NOES sseennel Seed iscibodadgis bogrerte| ploina. old to Dew,
: treat ahs ds qe ote enacts oon wh okey ods GF. 30%, é é written ae
a sete rt tft node rie lo sick eochevsctet seutin pene joule
Oh peteiass st he sell ep): toile a wd laereeray to s1xis spioises do Ghia. OVE.
Sagi ii corte .(ga54 esr ede tobe ut be ie pertain,

| Jotibaned 2sediadmg al tbe vaeqg Tioshasssboi i vine) indexed on é

: - ) (ittiastonig bea nol eee
— sft meaal cimebiera.T be belie sgomlacae shurtaiyesat tapi se

1") ...oui %, MOI Sat Miisiiy of of be] ronnsedeoN al 1h: eranad !
si oes ; ‘ Bic!
Hh 230 Syst et Mpa sal tee a gobitel pit r,t poe ¥ ad's ba
PIF ie vex. co) aM T8S hae orb i
amboanT 989 oi (eboqorine) cacetiht) ua) imo eaioe’ Lay yoloot::
: wy i _ a hse Q »p “es (ei) ‘he 14 Soe age, De ae
34 se bere Teietyt SS: eta) Dre Seamed WS eae retort oy LY wie a Pa
Fee ee tls La ates tet +p joy Mo) pees aa NaS) Gi
ioe a sii <igieass octane), cranelcpe aly lost. 8 : meuitoM, i.
+ . y al i J = ne ‘ SE
Lb SON atts Petal TAS TE Sh LALIT See WA Vie “h ee ole iors ey
CS-pee 1Ar awh a) sormaiedhade sched Seah et pris Osa st
i cite “eaten,
beret vetsionanten \aatyvicas \o shay leieitastetih dental ody ter eguilise galt. bese 9

i AT 4) 6,02 nolan cbaalavois May) sunralonesroyl faotpelogs wt moe b ‘1 colin
2 SR PCY #11 29 meat 1) i a} A af Liye cams (anqoayE mana obs odo ean Baha

t

WOT 6d

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 ~~ Band
May 1972 Mei
Part 8 Deel

DEVELOPMENT OF TRACHURUS TRACHURUS
(CARANGIDAE),
THE SOUTH AFRICAN MAASBANKER

By
E. H. HAIGH

Cape Town Kaapstad

ait! HSON/
VY

JUN 21 1972
S/BRARIED

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

LAGI Gs 3: 7-9) 31-25) 5,0 t-P-1-) 5 ( 2515, 7-9)>
6(1, t.—p.1.), 7(1-3); 8, (1-2), 10(1),
11(1-2, 5, 7, t--p.i.), 24(2), 27, 31(1-3), 33

Price of this part/Prys van hierdie deel
R1,35

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 08 9

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

DEVELOPMENT OF TRACHURUS TRACHURUS (CARANGIDAE),
THE SOUTH AFRICAN MAASBANKER

By
E. H. Haigu
South African Museum, Cape Town
(With 4 figures and 3 tables)
[MS. accepted 29 November 1971]

CONTENTS

PAGE
Introduction : : ) £30
Materials and methods . ee LAT
Description . : : Ar eit
Distribution : : =) 48
Summary . 2 : a 49
Acknowledgements ‘ LAG
References . : ; 6 RAO

INTRODUCTION

The systematics of the genus Trachurus of the family Carangidae seem to
be rather confused and need revision, based on an adequate world-wide
collection. Berry (personal communication) is of the opinion that there are
two species in South African waters. However, his review paper on the genus
Trachurus is only to appear in three or four months. The present larval fish
collection is not geographically wide enough for accurate comparison of the
two possible species and as the distinguishing features are essentially adult
characters, the designation Trachurus trachurus for these specimens is felt to be
the most accurate, at the same time indicating the similarity of these larvae to
European Trachurus trachurus larvae.

Several papers describing the larval development of specimens bearing
either the generic or specific name Trachurus have appeared since late in the
nineteenth century and this seems an opportune time to review what is known
about larvae of this genus.

The description of Caranx trachurus by Holt (1898) does not give details
and the illustrations in Annales du Musée d’ Histoire Naturelle de Marseille (5, 1899:
27-32, figs 53-63) to which he refers were not available. However, Ehrenbaum
(1909: 27-30) reviews Holt’s publications as do Heincke & Ehrenbaum (1900).

All authors seem to be satisfied as to the identity of the egg of Trachurus
trachurus and describe it as having a diameter of 0,7 to 1,09 mm, a totally
segmented yolk and an oil globule of 0,19 to 0,28 mm diameter with yellowish
to brownish pigment around it which remain in the anterior part of the yolk

139

Ann. S. Afr. Mus. 59 (8), 1972: 139-150, 4 figs, 3 tables

I40 ANNALS OF THE SOUTH AFRICAN MUSEUM

until resorbed. South African maasbanker eggs that have been measured are
seldom larger than 1 mm in diameter—usually about 0,9 mm with an oil
globule of 0,2 mm diameter. This falls within the range mentioned above.

Ehrenbaum (1909) gives a short description and six rather inadequate
figures of Trachurus trachurus, but these do serve to confirm the basic similarity
to the larvae described by Schnakenbeck (1931). Unfortunately Schnakenbeck
fails to provide adequate dimensional data. By modern standards Schnakenbeck’s
size groups are rather large, but comparative reworking of measurements in the
present paper produces values which approximate these.

Delsman (1926) described eggs and early larval stages from the Pacific
round the Indonesian islands under the name Caranx kurra. However, both eggs
and larvae are so similar to descriptions by Ehrenbaum (1909) that they could
easily be larvae of the genus Trachurus, probably maccullochi, which Nichols
(1940) put as a race of Trachurus trachurus.

The description of Trachurus symmetricus larvae from the Pacific coast of
America by Ahlstrom & Ball (1954) again shows the small differences between
species of Trachurus. Trachurus symmetricus larvae can perhaps be separated from
Trachurus trachurus larvae by the very slight difference in degree of pigmentation,
the former being less pigmented. However, degree of pigmentation hinges to a
very large extent on the length and method of preservation, and perhaps also
on the time of capture, be it daylight or after dark. Pigmentation is thus an
unreliable characteristic for the distinction of species so closely related.

Aleev (1957) described a good developmental series of Yvrachurus medi-
terraneus ponticus Aleev. Unfortunately his diagrams do not show ossification
details or pigmentation patterns and can thus not be successfully compared
with the figures published by Dechnik & Seniokova in 1964 of Trachurus
mediterraneus (?).

In 1969 Zhudova published three figures of Trachurus larvae described as
T. trachurus and a distribution map giving distribution of Trachurus larvae
between 5°N to 12°S and 10°E to 14°W in the Gulf of Guinea and in 1970
Kiliachenkova published several good figures of eggs and larvae up to 8,9 mm
in length caught along the West African west coast between 24° and 15°N and
19°50’ and 17°50’W.

The adult maasbanker is found in most parts of the Atlantic down to a
depth of 400 metres. It occurs abundantly on the west coast of southern Africa,
where, with Sardinops ocellata (Pappe, 1853), the South African pilchard, it
forms the basis of the fishmeal industry. Like the pilchard, it is a plankton
feeder, its diet consisting mainly of zooplankton including amphipods,
euphausids and fish larvae.

Eggs and larvae are usually found in deep seas (see section on distribution),
while juveniles are found mostly in sheltered bays close inland. Several authors
report the clustering of small and juvenile Trachurus under jellyfish (Ehrenbaum
1909) and floating debris and seaweed (Delsman 1926). Apparently this habit
is common to the Carangidae.

DEVELOPMENT OF TRACHURUS TRACHURUS I41

MATERIALS AND METHODS

Specimens were obtained by research vessels of the Division of Sea
Fisheries, Cape Town, using N100B and NiooH plankton nets, from 1950 to
1967. Samples were fixed and stored in formalin which was replaced by
70% ethyl alcohol. Specimens were stained, using methods of Hollister (1934)
but modified slightly by reducing the clearing time in KOH and reducing the
concentration of the KOH used. ‘This was done in order to preserve pigment in
specimens. As pigment is inclined to fade with time, more than one larva in the
size range was used in order to obtain the most characteristic pigment pattern.
Stained specimens were preserved in glycerin.

Measurements were taken as follows:

Standard length (s.l.): tip of lower jaw to end of caudal peduncle

snout: tip of lower jaw to anterior margin of eye

eye diameter: the eye being essentially round, could be measured
in any direction

head length: tip of snout to cleithrum

trunk length: tip of snout to posterior end of anus, measured
along the midline with verticle to snout and anus

depth: taken at posterior edge of head

All proportions are presented as percentage of standard length.
All lengths cited in text are the standard length.

DESCRIPTION

The adults and juveniles of Trachurus are distinguished from other genera
of the family Carangidae by the absence of separate anal and dorsal finlet, the
presence of laterally expanded scutes over the full length of the lateral line and
a procumbent spine before the spinous dorsal fin. As far as the distinctive
South African species is concerned, Smith (1953) says that Trachurus trachurus
has between 70 and go scutes on the lateral line, a depth of 45°4 and more than
30 dorsal rays.

Among later larval stages, dorsal fin-count of 8 + 31-33 and anal
fin-count of 3 + 27-29 combined with distinctive pigmentation pattern
will serve to separate Trachurus larvae from Decapterus larvae which have a
darkly pigmented area on the head.

Ossification

The smallest specimens show slight ossification of the premaxilla, dentary
and cleithrum. The three horizontal, one corner and one vertically situated
preopercular spines are lightly ossified (Fig. 1A). Between 3,0 mm and 4,5 mm
ossification takes place rapidly. The premaxilla with four fine teeth is well
formed. ‘The maxilla is lightly ossified. The dentary and angular can be clearly
distinguished and there are six ossified branchiostegal rays. The preopercle now

I42 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fic. 1. Trachurus trachurus. Early larval stages showing pigmentation, pattern and position and
degree of bone development. Measurements indicate standard length.
A. 3,30 mm. B. 4,50 mm. C. 5,80 mm. D.)752)\ mame

bears four spines on the horizontal arm, the corner spine has become very long
and there are two spines on the vertical arm. The number of spines laid down
between this stage and 5,4 mm is usually the number that remains for the rest
of the development. There is a spiny ridge on the outer anterior edge of the
preopercle with six to seven small spines on it.

The cleithrum broadens very gradually and by 5,4 mm the supracleithrum
has appeared as have the first traces of the opercle, the quadrate and the
pterygoid. There are now seven branchiostegals. Between 5,4. mm and 6,6 mm
further ossification takes place, all above-mentioned bones becoming heavier,
darker and broader. New bones make their appearance, i.e. the frontal,
sphenotic and parietal as well as the supra-occipital crest over the brain region.
Laterally from this is the first indication of the pterotic and behind the pterotic

DEVELOPMENT OF TRACHURUS TRACHURUS 143
what appear to be the first traces of the exoccipital. The hyomandibular and
symplectic can be clearly distinguished as can the opercle, the subopercle and
interopercle. The epihyal and ceratohyal show the first traces of ossification.
The articular and dentary have become fairly well amalgamated. By 7,8 mm
the pterygoid has become fairly extensive (Fig. 2A) and traces of the lacrimal
and nasal can be seen. The postcleithral bones which first become evident at
about 6 mm are now joined by traces of the scapula and the pelvic girdle is also
evident. Ceratohyal and epihyal can be easily seen.

The cranial ossification proceeds rapidly after 7,8 mm. The whole brain
region becomes ossified and bones join up. The post-temporal appears above
the supracleithrum and there are several centres of ossification in the supra-
temporal region. The circumorbital bones are evident but the nasal region is
still fairly unossified. The maxilla has now become the major bone in the upper
jaw, almost completely obscuring the dentary, when the mouth is closed. The
spines on the preopercle are less prominent. Ceratohyal and epihyal are well
ossified but still separate. The pectoral girdle elements present are scapula and
three pterygials while the pelvic girdle is well ossified.

TABLE I

Average meristic counts during development of Trachurus trachurus.

Average Dorsal Anal Vertebrae | Neural|Haemal|Pectoral| Pelvic | Caudal.
Site spines rays | spines rays trunk tail | spines | spines | fin fin fin
Aj25 | — == = — | o-10+ o-1 } 3-16] 4-6 0-3 — I-5-+1-5
6,03 o- 4] — o-5] o-10+ o-4 | 0-18] o-8 | 0-5 — |3-543-5
6,82 o-6+ 0-17 |O-1 + 0-14] O-10+ 0-12} 2-20] 3-10] 3-8 — 1-g+ 2-8
8,34 o-7-+ 3-24 | 1-2 + 7-22 10-+ 10-14] 20-22 | 11-13 | 6-14] 1-3 |8-9+8
9,16 6-8+ 13-25 | 1-3 +12-19 10+ 11-14] 20-22 | 11-13 | 8-15] 3-4 9+8
9,80 5-8+ 20-30 |3 +16-24 I0o+13-14] 22 |12-13] 9-17] 3-5 9+8
11,61 | 8+1+28-33 | 2+1-+26-28 Io+14 22 13 | 16-19] 1+4-5| 9+8
13,51 |{8-9+1+25-33 |2+1+ 23-28 I0o+14 22 13 7-18] 1+5 9+8
15,37 | 8+1+31-32 |2+1-+26-27 10+ 14 22 13 | 18-21] 1+5 9+8
26,00} 8+1+ 30 |2+1+26 10+14 22 13 | 1+ 2t | 1-5 9+8
29,25| 8+1+ 33 |2+1+29 I0+14 22 13 |1+20] 1+5 9+8
34,54| 8+14+ 34 |2+1+30 10+14 22 rg ale) | [he 9+8

48,50} 8+1+4+ 33 |2+1+30 I0o+14 22 19 | B21) 1-5 9+8

In the postcranial region ossification starts at about 4,5 mm in some
specimens while others show only the first traces of the haemal and neural
spines at about 6,0 mm. The ossification of the vertebral column appears to
start anteriorly with two to three neural spines, then some haemal and neural
spines ossify medially, and after this the neural spines over the intestinal sac
ossify, followed by the posterior haemal and neural spines. Between 5,4 mm
and 6,6 mm the vertebral centra ossify rapidly, and apparently several at once,
from the anterior, with the result that by 6,6 mm there can be as many as

144 ANNALS OF THE SOUTH AFRICAN MUSEUM

ten trunk centra and up to twelve tail centra partly or fully ossified. The urostyle
ossifies before the penultimate vertebra.

By 7,8 mm ten tail centra are ossified, and by 9,9 mm the full complement
of ten trunk and fourteen tail centra, including the urostyle, is ossified. ‘The
neural and haemal spines have broadened by now, especially in the caudal
plate where three haemal and two neural spines support the caudal fin rays.
Schnackenbeck (1931) gives an extensive account of the caudal ossification of
European species and ossification in the specimens described follows the same
pattern.

The median fin rays are first evident between 5,8 and 6,6 mm and those
in the dorsal and anal fins are formed simultaneously. Between 6,5 mm and
7,8 mm the dorsal and anal spines appear. The third anal spine is associated
with the soft anal and only becomes thickened at about 9,1 mm. The middle
dorsal spines appear first then the most posterior and the seventh and eighth
spines last of all, usually between 9,0 and 10,0 mm. The small procumbent
dorsal spine is formed only during juvenile stage. The spine of the soft dorsal
becomes thickened only at about 8,5-9,0 mm. The rays of the anterior and
midsections of both fins are formed first and ossification proceeds posteriorly
gradually until both fins are fully ossified at about 11,0 mm s.l. Between 9 mm
and 12 mm the fin supports of both fins are formed.

The pectoral lobe is present even in smaller specimens and ossification of
the rays starts as early as 5,2 mm in a few specimens but in the majority of
specimens examined general pectoral fin ray ossification commences between
6,0 and 6,50 mm at the most dorsal aspect of the fin and proceeds round the
periphery ventrally in sequence. Ossification takes place rapidly and by 10,0 mm
there are usually about 17 rays ossified. The full complement of 22 rays with the
most anterior ray considerably thickened can be seen only in juveniles of over
20,0 mm.

There are no evident pelvic lobes and the first ray appears only at 7,5 mm
but by 9,8 mm there can be as many as five rays ossified and by 10,0 mm the
outermost ray has become noticeably thickened. The full number of one spine
and five rays is present at 11,0 mm.

The caudal fin ossification starts very early on in development, as soon as
jaws and cleithrum have ossified, and well before the urostyle has turned up.
Rays ossify from the middle outward and by 6,5 mm some specimens show the
full complement of nine dorsal and eight ventral primary caudal rays. Secondary
caudal rays start ossifying almost immediately after this size and proceed
gradually until well into the juvenile stage. The urostyle turns up gradually at
about 6,0 mm s.1.

Juvenile ossification takes place mainly in terms of consolidation of bones
although there is of course a tremendous amount of growth taking place. The
juvenile stage is reached between 10,0 and 12,0 mm although the lateral line
scutes, a major distinguishing feature of the species, appear only between
18 and 19 mm.

DEVELOPMENT OF TRACHURUS TRACHURUS 145

Ossification of Trachurus trachurus larvae does not take place at the same
rate in all specimens and commencement also varies considerably. If position
of capture, time of year and water temperature are considered, it appears that
larvae of similar sizes caught in the same latitudes will show different degrees of
ossification, depending upon whether they were caught early or late in spring
or summer. Larvae of 6,5 mm were caught in water temperatures of 15,79°
and 22,18°C. Those caught in 15,79°C (August) showed far less vertebral
ossification than those caught at 22,18°C (January).

Pigmentation

AWA A

I] : a \ ‘

Fic. 2. Trachurus trachurus. Late larval stages with ossification nearing completion. A. 8,4 mm
standard length. B. 10,5 mm standard length showing ossification. sop. subopercle. iop.
interopercle. C. Same specimen showing pigmentation pattern.

The pigmentation pattern in Trachurus species from the southern African
region is much the same as that described for European specimens by Ehren-
baum (1909) and Schnakenbeck (1931) as well as that of Trachurus symmetricus,
described by Ahlstrom & Ball (1954).

146 ANNALS OF THE SOUTH AFRICAN MUSEUM

In the early larvae (Fig. 1A—B) the pigment is distributed in three main
areas, dorsally along the edge of the body and over the brain area, medio-
laterally along the peritoneal wall and over the medioposterior region of the
chorda and ventrally along the edge of the trunk and tail with a few scattered
melanophores on the upper and lower jaws and in the region of the cleithrum.
The dorsal pigmentation is soon augmented (Fig. 1C) by melanophores
appearing on the dorsolateral sides of the body musculature. The median row
of elongated contracted melanophores in the lateral line region which is so
characteristic of many carangid larvae, becomes darker and more distinct.
A second row of pigment appears directly over the notochord. On the peritoneum
pigment darkens dorsally and scattered melanophores appear laterally.
Chromatophores are still present on the nose, the jaws and underneath the
branchiostegal rays. By 6,6 mm (Fig. 1D) a second pigmentation on the caudal
plate itself has become more clearly defined. Between 7 mm and 10 mm the
body wall thickens and fins are formed obscuring much of the deeper-lying
pigmentation. The balance of pigmentation changes rapidly, the dorsal aspect
of the fish becoming far more heavily pigmented than the ventral (Fig. 2C).
There are large and small melanophores on the dorsolateral aspect. ‘The more
dorsally situated ones seem to be far larger than those on the lateral side.

The dorsal aspect of the head has also become pigmented to uniformity
with the body. There are still chromatophores on the jaws and nose but none is
visible on the trunk region. The ventral aspect of the tail still bears widely
scattered melanophores but the caudal plate and unpaired fins have become
pigmented. Both spiny sections of the median fins are pigmented, while the
rayed parts are clear. The lateral line row of melanophores is still very distinct
and remains so until the juvenile stage is reached.

Changes in body-form

The earliest stages of Trachurus sp. present in these collections have the
yolk-sac absorbed and intestine with one fold developed at 2,45 mm. The eye
has become pigmented but the snout is shorter than eye diameter (7,7% v-s.
10,0% ofs.l.). During the next millimetre increase in length the head increases
proportionately rapidly from 25% to 33,5% of standard length and the depth
remains constant at about 30%. However, this depth is measured at the region
where the head joins the trunk and it is noticeable (Table 2) that this original
proportion is slightly but definitely decreased in the larger stages. ‘This is no
doubt due to rapid increase in the head during early development and the
increase in tail size in later development. If larvae shown in Figures 1 and 2 are
compared, it will be noticed that the younger stages are proportionately much
deeper anteriorly than posteriorly while the later stages are of more even
proportions.

The larval development of Trachurus sp. is thus characterized by a smooth
and gradual development. Fins form in adult positions and body proportions
change gradually and slightly.

DEVELOPMENT OF TRACHURUS TRACHURUS 147

TABLE 2
Mean measurements of Trachurus trachurus in mm.

Average
Size range SUZE No. Snout Eye Head Trunk Depth

2,5- 395 3,17 5 0,25 0,33 0,90 1,70 0,95
355- 455 4,04. 29 0,45 0,40 1,43 2,40 1,30
455- 555 4593 45 0,53 0,50 1,65 2,80 1,55
55- 6,5 5,04. 32 0,70 0,60 2,00 3,40 1,80
6,5- 795 6,92 32 0,75 0,75 2,50 4,00 2,30
75- 8,5 8,15 29 0,95 0,95 3,00 4,90 2,60
8,5- 955 9,04 16 1,05 1,05 320 5,30 2,80
9,5-10,5 10,01 13 1,15 1,10 3,40 5,80 3,05
10,5-11,5 10,89 6 1,20 1,25 3,70 6,20 3,20

11,5-12,5 11,28 17 1,40 1,40 4,10 6,80 3,65
12,5-1355 12,99 3 1,60 1,40 4,60 7320 3,90
13,5-14,5 13,90 7 1,50 1,50 4,80 7,70 4,00
14,5-15,5 15,31 6 1,50 1,70 5,20 8,40 4,50
15,6-16,6 — oO — — — — a

16,6-17,6 16,90 3 2,00 1,80 5,80 9,70 4,80
17,6-18,6 18,20 I 1,80 2,00 5,90 9,80 5,50
18,6—19,6 18,88 4 2,10 1,90 6,40 9,80 5520
19,6—20,6 20,15 2 2,20 2,00 6,50 10,50 6,10
20,6—21,6 21,01 4. 2,10 2,20 6,70 II,10 6,00
23,67 23,67 I 2,27 2,27 7,80 12,02 5,85
24,05 24,05 I 2,27 2,60 7,00. |) 11,30 6,20
25,67 25,607 I 2,60 2,92 8,45 13,65 6,82
29,00 29,00 I 3,00 3,00 9,50 14,60 7:50

TABLE 3
Mean body proportions of Trachurus trachurus larvae as % of standard length.

Size range Average
in mm SUZE No. Snout Eye Head Trunk Depth

in mm

255- 355 3,17 7.5 10,0 25,0 54,0 30,0

3,5- 455 4,04, 10,5 955 3355 56,0 30,5
455- 555 4,93 11,0 10,5 32,0 57,0 3155
5.5- 6,5 5,84, 11,5 10,5 35,0 53,0 31,0
6,575 755 6,92 11,0 11,0 36,0 59,5 32,0
Poros 8,15 13,0 11,5 36,0 5955 31,5
8,5- 955 9,04 12,0 11,5 35,0 58,0 31,0

11,5 11,0 34,0 58,0 31,0
10,5 11,5 34,0 57,0 29,0
11,5 11,5 3455 57,0 30,5
12,0 10,5 35,0 5555 30,0
10,5 11,0 3455 5505 29,0

9,6 11,0 34,0 5455 29,0

Q,5-10,5 ‘10,01
10,5-11,5 10,89
11,5-12,5 11,28
12,5-1335 12,99
13,5-14,5 13,90
14,5-1535 15,31

16,6—17,6 16,90 11,0 11,0 34,0 57,0 28,0
17,6—-18,6 18,20 10,0 10,5 32,0 5395 30,5
18,6—19,6 18,88 11,0 10,0 33,5 52,0 27,5
19,6-20,0 20,15 11,0 10,5 32,5 53,0 30,0
20,0—21,0 21,01 10,0 10,0 32,0 52,5 28,5
23,67 23,67 955 9.5 33.9 51,0 24,5
24,05 24,05 955 11,5 3255 47,0 2555
25,67 25,67 10,0 11,5 33,0 53,0 26,5

= eH we HN) OOOO B WN

29,00 29,00 10,5 10,5 32,5 50,5 26,0

148 ANNALS OF THE SOUTH AFRICAN MUSEUM

The air-bladder starts as a small clear patch behind the cleithrum and
extends rapidly posteriorly until it occupies about 50% of the dorsal longitudinal
distance of the peritoneum.

det!
pt
mx
pmx
d
la
Corp
Fic. 3. Trachurus trachurus juvenile
ang. angular hyp. hypurals ptm. post-temporal
art. articular la. lacrimal pto. pterotic
brstg. branchiostegal rays mx. maxilla ptr. pterygoid
clt. cleithrum na. nasal ptryg. pterygials
cor. coracoid pa. parietal quad. quadrate
c. orb. circumorbitals pelt. 1 & 2 postcleithra scap. scapula
deth. dermethmoid pelv. pelvic girdle soc. supraoccipital
dn. dentary pf. prefrontal supcl. supracleithrum
f. frontal pmx. premaxilla sym. symplectic
hyom. hyomandibular pop. preopercle u. urostyle

DIsTRIBUTION

The area covered by the research vessels of the Division of Sea Fisheries
on the pilchard research programme has varied since the inception of the
programme. During 1951 and 1952 the area worked lay between 32° and
35° 30'S and was delimited by the 200 fathom depthline to the west. Approxi-
mately the same area was worked between 1953 and 1957. In 1958 the eastern
limit of the work area was extended round Cape Point to 19° 30’E. This area
was worked until the end of 1960 when the eastward delimitation was extended
to 21°E. During these years the westward delimitation extended to 16° 31’E.
From July 1963 to December 1965 the area covered by the ships was between
32° 10’ to 36° 10’S and 16° to 21° 30’E. Station lists are obtainable from the
Annual Reports of the Division of Sea Fisheries, for the relevant years. Figure 4
shows positions where larvae and juveniles were captured from 1951 to 1965.
The most productive year was 1964 when Trachurus sp. were caught at 81
stations visited, followed by 1965 when 68 stations yielded Trachurus sp.

Trachurus larvae were caught throughout the year, the smallest number of
stations yielding larvae in July and the largest number during October, that is
late spring in the Southern Hemisphere. The spring months, August, September
and October, yielded the largest collection of larvae while the three other
seasons yielded far fewer and were not markedly different.

DEVELOPMENT OF TRACHURUS TRACHURUS 149

Ee | Bee al Raa eee
14 15 16 17 18 19 20 21 22

Fic. 4. Distribution of Trachurus trachurus larvae in the research area between 1951 and 1965.

SUMMARY

The development of the South African larvae of Trachurus trachurus is
described and a general map of the distribution of the larvae provided.

ACKNOWLEDGEMENTS

The author wishes to thank the Division of Sea Fisheries’ sea-going staff for
the collection of study material, the Fisheries Development Corporation for
financial assistance, and the South African Museum, Cape Town, for housing.
Dr Naomi Millard read and criticized the manuscript.

REFERENCES

Autstrom, E. H. & Batt, O. P. 1954. Description of eggs and larvae of jack mackerel ( Trachurus
symmetricus) and distribution and abundance of larvae in 1950 and 1951. Fishery Bull.
Fish Wildl. Serv. U.S. 56: 209-254.

150 ANNALS OF THE SOUTH AFRICAN MUSEUM

ALEEV, Y. G. 1957. [On ten species of Trachurus in the U.S.S.R. seas.] Trudy sevastopol’. biol. Sta.
g: 167-242. (In Russian.)

Decunik, T. V. & Sentoxova, V. I. 1964. [Distribution of pelagic fish eggs and larvae in the
Mediterranean Sea.] Trudy sevastopol’. biol. Sta. 15: 77-115. (In Russian.)

DetsMAN, H. C. 1926. Fish eggs and larvae from the Java Sea (1). Treubia 8: 199-239.

EHRENBAUM, E. 1909. Eier und Larven von Fischen. 1.Teil. Word. Plankt. Lief. 4: 1-216.

Heincke, F. & EHRENBAUM, E. 1900. Eier und Larven von Fischen der deutschen Bucht.
Wiss. Meeresunters. (Abt. Helgoland) 3: 277-279.

Ho.uistEr, G. 1934. Clearing and dyeing fish for bone study. Zoologica, N.Y. 12: 89-101.

Hott, E. W. L. 1898. Notes on the reproduction of teleostean fishes in the south-western
district. 7. mar. biol. Assoc. U.K. 5: 107-155.

KimiAcHENKOvA, V. A. 1970. Development and distribution of eggs and larvae of Trachurus
trachurus L. Rapp. P.-v. Réun. Cons. perm. int. Explor. Mer 159: 194-108.

Nicuots, J. T. 1940. Notes on carangin fishes. V. Young Trachurus in the Gulf of Mexico.
Am. Mus. Novit. 1067: 1-4.

SCHNAKENBECK, W. 1931. Carangidae. Rep. Dan. oceanogr. Exped. Mediterr. 2(A 14): 3-13.

SmitH, J. L. B. 1953. The sea fishes of southern Africa. 4th ed. Cape Town: Central News Agency.

Zuupova, A. M. 1969. [Materials and study of the eggs and larvae of some species of fish from
the Gulf of Guinea and the adjacent waters of the open ocean.] Trudy AilantNIRO Inst.
ryb. Khoz. Okeanogr. 22: 135-163. (In Russian.)

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS
To be reducible to 12cm X 18cm (19cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.
For journal articles give title of article, title of journal in italics (abbreviated according to the

World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,

volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHer, P.-H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Konan, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region
of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE. L,
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51).
The Harvard system of reference to be used in the synonymy lists, with the full references
incorporated in the list at the end of the article, and not given in contracted form in the synonymy
list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 80.

i pes etree s i

tie bee J
S
u 5 4
~ 4 are
: a
Bi ante YY *, bh fata
RS \ Fae ae se he 4 RAY REN
i Es = é Mr ae if os aN pu eae
el + ean GARE hi 7 pide} eels py ee et bE ad
x = ¢ 2 2 , i Fak
6 © i fe
~ “1 be t 2
es
of =," ait
se ryan abe Boy IS St
. Y mae Se =
‘ D Bier} rae ak es Ay at | a
z oe
7 ia 1
> n Pas a
i
= nS
\s =
aes
ey 2 3 t
& .
5 iss psec) i
Ce ee tee
> _ a i
i E SE, ney a
& ey & Us
a5" es Re
é dS
8
ne
ss :
F Af
1 i
~
= . , fy
~ Lepr ees
i
A P %
i a lee Gy
2 ;
a i}
i
- *
ad = Pe ‘
) *
P|
=
< as 1:4.
-

507, (of

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
August 1972 Augustus “Lion
Part) | 9. Deel

<

|
\

Use r OX
‘J

|
3

A LATE PLIOCENE RHINOCEROS FROM
LANGEBAANWEG, CAPE PROVINCE

By
D. A. HOOIJER

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 2(5 3> D> 7-8), iS. 5> JS ony 5 5(2, 5» 7-9);
G(i;. + pee); 7(1-3); &, ‘o{t=2),.- -s0ft),
ri(1—2; 5575 tp), 24(2), 27; B13). 33

Price of this part/Prys van hierdie deel
R5,30

Trustees of the South African Museum © _ Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 12 7

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

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG,
CAPE PROVINCE

By
D. A. Hooter
Riuksmuseum van Natuurlyke Historie, Levden

(With plates 21-34 and 51 tables)
[MS. accepted 14 February 1972]

CONTENTS
Introduction . ‘ 4 : ; ; f ik
Ceratotherium praecox Hooijer & Patterson : : a 152
Dentition and skull . , ; 5 ‘ : aie 158
Postcranial skeleton . ‘ ’ 4 ; : 2S tO8
Other C. praecox sites in East and South Africa ; oy Oz
Summary 4 i ; : . : : 2 190
Acknowledgements . °. : ; : ; -} KOO
References : : ; ‘ ; - A out OT
Explanation of the plates. ; d : ; LOX
INTRODUCTION

The rhinoceros remains described in the present paper are from the
‘E’ Quarry at Langebaanweg, situated approximately 32°58’S, 18°9’E in the
Sandveld region of the south-western Cape Province, some 105 km N.N.W. of
Cape Town. They are more abundant than those of any other large mammal in
the Langebaanweg fauna. The literature on the geology and palaeontology of
the Langebaanweg deposits is reviewed in Hendey (1970a); the geological age
is discussed in Hendey (19700) and Maglio & Hendey (1970). The ‘E’ Quarry
rhinoceros has been cited as Diceros aff.bicornis, but I found it to be a very early
Ceratotherium, the same as that from Kanapoi, Ekora and Lothagam-1 in
N.W. Kenya described as Ceratothertum praecox Hooijer & Patterson (1972). This
species is still very close to a Diceros like D. bicornis (L.) in some dental characters
which take the eye even at a cursory look, such as the transversely placed
proto- and metaloph, absence of medifossettes, well-developed paracone style,
and angular antero-internal crown corners. In these as well as other characters
the fossil teeth from Langebaanweg and those of D. bicornis differ from those of
Ceratotherium simum (Burchell), which has obliquely placed proto- and metalophs,
medifossettes, no paracone style, and rounded antero-internal crown angles.
We believe that Ceratotherium praecox is directly ancestral to the living C. szmum,
and its occurrence at the Kenya sites, near the 4 million year level (Maglio
1970; Cooke & Maglio 1971; Bishop 1971a: 511) is perfectly in accordance
with the Late Pliocene age that is now becoming accepted for the Langebaanweg
deposits. 7

I5I
Ann. S. Afr. Mus. 59 (9), 1972: 151-191, 14 pls, 51 tables.

152 ANNALS OF THE SOUTH AFRICAN MUSEUM

Abbreviations used in this paper are:

K.N.M. Kenya National Museum

L.M. Leiden Museum

M.C.Z. Museum of Comparative Zoology, Harvard University
S.A.M. South African Museum

CERATOTHERIUM PRAECOX HooiER & PATTERSON
Ceratotherium praecox Hooijer & Patterson 1972: 19.

The present species was based on three incomplete skulls and mandibles
with teeth, some fragments without teeth, an upper molar and an imperfect
humerus from Late Pliocene sites in north-western Kenya. ‘The Langebaanweg
rhino collection comprises four upper dentitions, parts of three skulls and ten
mandibles (mostly with teeth), 100 isolated upper and 50 isolated lower cheek
teeth, 3 upper incisors, 20 deciduous cheek teeth, and 650 postcranial bones.
The cranial and dental characters of the Langebaanweg rhinoceros are the same
as those of the Kenya collection already described, but the Langebaanweg
collection adds to our knowledge of the species information on the upper
incisors and milk teeth which were unknown before, and the postcranial
characters which were virtually unknown until the Langebaanweg material
became available. The data provided in the present paper show the amount of
individual variation within a single species of Pliocene rhinoceros. It is not
saying too much now that C. praecox odontologically and osteologically is better
known than its extant descendant, although, of course, its external characters
are for ever lost to us.

The cranial characters of the present species are as follows: dorsal surface
more concave, posterior portion less extended behind, occiput less posteriorly
inclined, nuchal crest less thickened than in Ceratotherium simum. The premaxillae
bear two incisors each, about 10 mm in diameter. The symphysial part of the
mandible is narrower than in C. stmum, and similar to D. bicornis. The premolars
and molars (upper as well as lower) are more hypsodont than those in D. bicornis,
but decidedly less so than in C. stmum. The flattened ectolophs, marked protocone
folds in the molars, strong internal cingula in the premolars, angular antero-
internal corners of the crowns in premolars and molars alike, the posterior
bulging of the protocones, which make up three-fifths of the internal crown
faces, the medifossettes that rarely occur (mostly in P?-3 and M3, if at all), and
the medisinus and postsinus depths being very nearly equal, all these are
characters shared by the Kanapoi and the Langebaanweg C. praecox.

The species in question is rather Diceros-like in skull and dentition, the
teeth differing in their relatively higher crowns, with a flattened ectoloph on
which the paracone style is almost completely suppressed, the postsinus being
very nearly as deep as the medisinus, and the posterior protocone bulge slightly
more marked. In these points the Kanapoi and Langebaanweg rhinoceros is
evolving toward the Quaternary Ceratothertum simum, in which the crown height

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 15%

is still greater, the paracone style completely suppressed and the parastyle
raised, forming a concave area on the ectoloph where the paracone style had
been, medifossettes are common, formed by the union of crochet and crista,
postsinus and medisinus are equal in depth, the protocone bulge is more
marked, the protoloph is obliquely placed and the antero-internal crown angles
are rounded. In the early subspecies C. stmum germanoafricanum (Hilzheimer),
which is indistinguishable from the extant C. szmum simum cranially, and which
occurs at Laetolil, the basal Olduvai Beds, and Chemeron Formation locality
J.M.90 (=91), the crowns are not quite so hypsodont and the metaloph is still
transverse in its course rather than oblique as in the modern form, although the
rounded antero-internal crown angles and the medifossettes of C. simum
germanoafricanum are as in C. simum simum. In my earlier paper on Pleistocene
East African rhinoceroses (Hooijer 1969), published at a time when I had not
yet studied the material from Kanapoi, Lothagam—1 and Ekora, I referred
specimens from the Chemeron Formation, locality J.M.507, and from the
Mursi Formation of the Omo Basin (=lower level of the Omo collection made
by Mr. R. E. F. Leakey in 1967) to C. simum germanoafricanum which I now
recognize as belonging to Ceratotherium praecox instead; this will be dealt with in
the final section of this paper. The discovery of Ceratotherium praecox vindicates
the view of Thenius (1955) that Ceratotherium split off from the Diceros stock
sometime in the Pliocene.

DENTITION AND SKULL

The individually youngest upper dentition, L13035, comprises P?-M3
from the right side as well as M!~ sin. (Pl. 21). The crowns of P? and M? dext.
only are virtually complete. There was a DM! or P? as there is an anterior facet
on: P?. )

P? is worn down to a height of 28 mm externally. The external enamel
layer is missing for the most part; only the metastyle portions remain. There is
a very marked internal cingulum, rising on the protocone and the hypocone
from its lowest point at the medisinus entrance. The internal portions of
protoloph and metaloph are connected at their bases by a small ridge; there is
a small pit between it and the internal cingulum. The medisinus is slightly
deeper than the postsinus, and there is a very small crochet, hardly more than
a point.

P3, worn to 45 mm from the base externally, has a very prominent internal
cingulum, reaching its lowest point at the entrance to the medisinus, which is
narrow and V-shaped. There is a very weak crochet, and no crista or ante-
crochet. Medisinus and postsinus are equal in depth. The ectoloph is flattened,
with a weak cingulum; there is no parastyle fold or paracone style. The
protoloph is hardly indented anteriorly, but there is a vertical groove in the
metaloph marking off the hypocone.

P4, with an external height as worn of 60 mm, has the flattened ectoloph
detached from the remainder of the crown, which shows the narrow, V-shaped

154 ANNALS OF THE SOUTH AFRICAN MUSEUM

medisinus entrance, the heavy internal cingulum, weak crochet, and the
medisinus depth equal to that of the postsinus, as in P?. An internal view of
P2-4 dext. of L13035 is given in Plate 25, top.

M}, the right of which lacks most of ectoloph and protoloph, and the left
of which is entire but for the antero-external angle, is worn to a height of
52 mm externally. The lingual entrance to the medisinus is V-shaped, and there
is an anterior fold in the metaloph marking off the hypocone. This molar, in
contrast to the premolars, has a deep fold anteriorly in the protoloph marking off
the protocone (the protocone fold), a strong crochet extending all across the
medisinus, not receding near the base, and the internal cingulum hardly -
marked except along the protocone and for a tubercle at the entrance to the
medisinus. ‘The inner portion of the protoloph is recurved backward, forming
three-fifths of the internal surface. The ectoloph is flattened, without styles,
and medisinus and postsinus are of the same depth.

M2?, nearly entire on both sides, is worn to 75 mm from the external base.
This is clearly a hypsodont tooth, the anteroposterior diameter of the crown
being 62 mm externally. A weak paracone style is seen in the upper part of the
crown only, to 60 mm from the base, flattening out further rootward. There is
no groove marking off the hypocone, but the description of M! would otherwise
fit the M?.

M? of dentition L13035, both incomplete behind, are 90 mm high as worn
and the length of the outer surface is about 75 mm. The marked protocone fold,
internal protocone cingulum, and strong, even bifid crochet, are as in the other
molars of this individual. ‘The paracone style is weak but discernible, reaching
from the top of the crown to approximately 50 mm from the crown base.

Another upper dentition, L2519, likewise consists of isolated teeth, which
are P?-M? dext. and P?-M! sin. (Pl. 22). They are rather well preserved
although a number of crown angles are missing.

P? is just 20 mm high as worn externally. The medisinus is still open
internally. It shows a crochet united with a small crista so that a medifossette is
formed. The same feature is seen in both P’, which are worn down externally
to 35 mm from the crown base. The protocone fold, which is preserved only in
P? dext., is more marked than that in P? of dentition L13035. The postsinus is
almost as deep as the medisinus; the ectoloph is just as flattened, with a weak
cingulum, and the internal cingulum is just as prominent as that in P® of
L13035. P*, present on both sides in L2519, is 45 mm high as worn externally,
and the right specimen has an imperfectly formed medifossette, while the left
has a bifid crochet and a small crista that do not join. There is no difference in
depth between the postsinus and the medisinus, and the ectoloph and the
internal cingulum are as in P?®.

M}?, on both sides, has a particularly powerful crochet, nearly twice as
thick as that in M1 of L13035, but no crista. The external crown height is 35 mm,
as worn. In addition to the anterior protocone fold there is an internal indenta-

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 155

tion in the protocone as seen in M! sin. The posterior bulging of the protocone
is such that it forms three-fifths of the internal crown face. There is a distinct
hypocone fold, visible in both the right and the left molar. The internal
cingulum is not preserved in these molars but externally there is a very weak
cingulum, mainly posteriorly, as in all the molars. The ectoloph is flattened, and
the narrow postsinus appears to be slightly less deep than the medisinus.

The M? dext. of L2519, 50 mm high as worn, has a crochet that is not
thickened; it is recurved outward at the apex but the crista is just barely
indicated and no medifossette is formed. The posterior bulging of the protocone
is such that it occupies 27 out of the 45 mm long internal basal anteroposterior
diameter. The lingual entrance to the medisinus is V-shaped, and the protocone
is indented lingually. ‘The portion of the crown that would have shown the
protocone fold is missing; the hypocone fold is weakly developed, and so is the
internal cingulum; the ectoloph is flattened, without styles. The two sinuses are
equally deep.

M? dext. of L2519 lacks most of the outer surface (ecto-metaloph) so that
no measurements can be given. It has a crown height as worn of about 60 mm.
There is a very marked protocone fold, and a weak cingulum on the depressed
internal surface of the protocone. The crochet is well-developed and there is a
crista, too. These projections, however, remain separate down to the bottom of
the medisinus. On this rather worn molar there is no trace of a paracone style
such as we see on less worn specimens; the paracone style is no longer visible
in the basal 50-60 mm of the crown.

The next upper dentition to be described is L13747 (Pl. 23). Of this set of
teeth the small anterior premolar P!, or a persisting DM}, is preserved on the
right side, as the teeth are still zm sztu in the maxillary. It is about 23 mm
anteroposteriorly, and about 20 mm transversely; nothing can be said about its
structure as the crown is worn flat.

P? is nearly entire on both sides. Although the worn crown height is the
same as that in P? of L2519 (20 mm) the valley between protocone and hypocone
is closed as wear has reached the bottom of the sinus in between. There is a
rather strong internal cingulum, and a pit is formed between it and the joint
bases of proto- and hypocone, as in P? of L13035. Postsinus and medisinus are of
equal depth. There is a crochet but no crista.

P? lacks the entire outer surface on the left side, and has only the antero-
external angle on the right. The protocone fold is weakly developed. The
internal cingulum is very marked, continuous with that on the anterior surface,
and it carries a series of tubercles. It extends all along the protocone, reaching
its lowest point at the narrow medisinus entrance, and rises along the hypocone,
i.e. the same development that we noticed in the premolars of the two upper
dentitions dealt with above. There 1s only a crochet, which is not very prominent,
making the central portion of the medisinus rather wide. The depth of this
portion of the medisinus is the same as that of the postsinus.

I 56 ANNALS OF THE SOUTH AFRICAN MUSEUM

P*, the worn ectoloph height of which is 40 mm, is rather damaged on the
left but well preserved on the right side. There is a paracone style, which is
rather more developed than that in the less worn dentitions L13035 and L2519;
in these teeth there is no trace left of the paracone style, but in L13747 it
continues to about 20 mm from the crown base. There is a weak external
cingulum along the posterior moiety of the ectoloph. The posterior bulging of
the protocone occupies three-fifths of the internal surface of the crown. The
internal cingulum is less developed, and the crochet more prominent than that
in P’. The protocone fold is hardly shown. The medisinus is as deep as the
postsinus and has a narrow, V-shaped entrance.

M}, which is between 20 and 25 mm in worn ectoloph height, has the
protocone fold well marked. The protocone takes up 30 out of the 50 mm of
internal anteroposterior crown diameter, and is slightly indented internally.
The bottom of the narrow internal medisinus entrance is almost reached by
wear, but its central portion is still about 15 mm deep, which is also the depth
of the postsinus. The very thick crochet is free from the ectoloph at the level of
wear. In M?! dext. it would have closed off a medifossette with the small crista
if wear had proceeded some 5 mm more, but in M! sin. no medifossette would
have been formed in this way. In this advanced stage of wear no trace remains
of the paracone style; the internal cingulum is so weak as to be practically
absent.

M7? is 40 mm high at the worn ectoloph. M? dext. has a vertical fracture in
the ectoloph, but the external surface of M? sin. is undamaged although
detached from the remainder of the crown. The paracone style is shown as a
weak bulge only along the worn edge of the crown. The protocone fold is very
distinct, and the internal indentation of the protocone shows just as it does in
the M?. The protocone occupies 40 out of the 65 mm of internal anteroposterior
crown diameter. The crochet is narrower than that in M! and does remain free
at its apex so that no medifossette is formed. There is hardly any trace of an
internal cingulum. |

M3, incomplete on both sides, has the external surface worn down to
50 mm; the paracone style is shown only in the apical 15 mm. The anterior
fold, internal indentation, and posterior bulging of the protocone are as in M?.
The crochet extends all across the medisinus but does not close off a medifossette.

Whereas the two dentitions first described (L13035 and L2519) are rather
similar in dimensions (see Table 1) dentition L13747 is larger, but there are no
significant differences in structure. The only point worth making is that the
paracone style is slightly more marked in these large teeth than in those earlier
described.

A crushed skull, L6658, has a good portion of the palate with P4—-M? dext.
and P?-M? sin., and the two last molars detached (Pl. 24). The dental dimen-
sions are more or less intermediate between those of L2519 and L13747

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 157

(Table 1). There is a small portion of the anteriormost premolar, on the left
side. P* sin. is incomplete internally and much worn down: the ectoloph height
is reduced to 15 mm, and the medisinus is cut off from the lingual border. ‘There
is a tiny medifossette, which would have disappeared with a little more wear.
P’ sin. has a medifossette too; its external height as worn is almost 30 mm, and
no paracone style is shown. The medisinus is just closed off lingually. The
internal cingulum, with its lowest point at the junction of protocone and
hypocone, is well developed. It is slightly less marked in P*, present on both
sides, with a worn ectoloph height of 40 mm. The crochet is bifid in P* dext.,
and single in P4 sin. The premolars P? and P* agree in the postsinus being as
deep as the medisinus, the posterior bulging of the protocone forming three-fifths
of the internal surface (21 out of 35 mm in P%, and 27 out of 45 mm in P*), and
in their weak protocone folds.

M1, lacking the antero-external angle on both sides, is some 25 mm high as
worn externally. The protocone fold is very marked, and there is an internal
indentation in the protocone, which occupies three-fifths of the internal border.
The lingual cingulum is weakly developed, the lingual medisinus entrance very
narrow, and the crochet is rather thick, as usual in first molars. M?, the right of
which is partially embedded in the bone, has a worn ectoloph height of 45 mm,
and does not show the paracone style any more. The characters are those of M?;
only the crochet is more slender. The M3, of which the left is virtually complete,
is 60 mm high as worn externally. The paracone style can be traced in the
apical 15-20 mm only. The crochet extends all across the medisinus, and joins
the posterior wall of the protoloph, thus cutting off the external portion of the
medisinus. The protocone fold is strongly marked, the internal cingulum very
weak.

TABLE I
Measurements of upper teeth of Ceratotherium praecox from Langebaanweg (mm)
No. of specimen L13035 L2519 L13747 L6658
P2, ant.post. C. 35 33 36 32
ant.transv. = 40 44 37
post.transv. c. 45 == 50 40+
P3, ant.post. 45 46 — 43
ant.transv. 58 57 66 62
post.transv. 54 — — 58
P*, ant.post. 48 51 53 50
ant.transv. _— 65 75 67
post.transv. — 60 73 63
M}, ant.post. — 58 c. 60 57
ant.transv. 70 70 80 73
post.transv. 64 c.64 7 =
M?, ant.post. 62 64 68 64
ant.transv. 72 72 82 74
- post.transv. 67 65 c.70 69
M35, ant.post.(int.) 65 — 67 66
ant.transv. 72 — 78 73
length outer surface ¢. 75 — 79 83
Length P?-M? c. 300 c. 300 330 305
Length P?—P4 c. 135 130 135 125

Length P*-M? c. 230 C. 225 245 235

158 ANNALS OF THE SOUTH AFRICAN MUSEUM

To dentition L13035 belongs a skull portion, giving a zygomatic width of
390 mm, very near to the maximum, observed by Heller (1913) in modern
Ceratotherium simum, viz., 384 mm. The length from M® to the back of the
postglenoid process is c. 220 mm, slightly less than the length P*-M3 (c. 230 mm).
In subadult skulls of C. simum in which M? has not erupted yet the length
P4_-M®$ exceeds the postdental length from M®? to the back of the postglenoid
process (e.g., S.A.M. 21381: P*-M? c. 225 mm; postdental length c. 190 mm).
In skulls with M$ slightly worn the two lengths are subequal (S.A.M. 21382:
P4_-M®8 215 mm; postdental length 200 mm; S.A.M. 21379: P4-M?3 225 mm;
postdental length 210 mm). In fully adult C. stmum skulls with M? well worn
down the length P*-M® is exceeded by the postdental length (M.C.Z.,
Dept Mamm. 24917 and 34850: P4-M? 190-205 mm; postdental length
270-275 mm). In the holotype skull of Ceratotherium praecox from Kanapoi, which
is quite adult, the postdental length is the larger of the two, though not to the
extent seen in the recent species (Kanapoi P4-M® 205 mm; postdental length
230-250 mm).

The premaxillaries of L13747 are preserved, and they show two alveoli on
each side, one behind the other. The anterior alveolus holds a tooth crown that
is unerupted, about 12 mm anteroposteriorly and 9 mm transversely. The
posterior alveolus is of the same size but empty; its depth is only 7 mm. The
specimens are shown in Plate 28, top. The occurrence of rudimentary upper
incisors in C’.. praecox is interesting, as the recent species of Ceratothertum no longer
shows them. An isolated I’ has a rounded crown and a strong, posteriorly
recurved root. The height of the crown and root combined is 37 mm, while the
crown diameter is 11 mm (PI. 28, top right).

The nasal horn boss of L6658 is crushed, but its width is about 180 mm.
The nasal portion of another skull, L2520, is 180 mm wide at the horn boss;
this width is 170-208 mm in adult males, and 146-173 mm in adult females of
recent Ceratothertum simum (Heller 1913). The frontal region of the skull L2520
shows the second horn boss, on the frontals, but the upper borders of the orbits
are not preserved. Dorsal views of skulls L2520 and L6658 are given in Plate 27.

Skull L13747 is broken in many pieces; the right half of the top of the skull
has been reassembled (Pl. 26). Although the angle between the dorsal and the
occipital planes cannot be exactly measured it is approximately 60°. This is
65° in skull K.N.M. KP30 from Kanapoi, against 65—80° in Diceros bicornis,
and 45—50° in Ceratotherium simum. These figures tend to show that in the fossil
C. praecox the occiput is less posteriorly inclined relative to the dorsal surface
than in C. szmum, and rather resembles D. bicornis in this respect. In keeping
with the less marked posterior inclination of the occiput, the nuchal crest in
C. praecox is not as thickened as it is in modern C. simum, in which it is quite
massive, overhanging the occipital condyles.

The mandible L13035 is nearly entire, lacking only part of the ventral
border of the left horizontal ramus and the right coronoid process (PI. 30).

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 159

P,—-M, dext. and P,—Ms sin. are in situ; an internal view of the right ramus is
given in Plate 31. Mandible L11849 has the symphysis as well as P,-M, dext.,
somewhat more worn than L13035 (PI. 32, right). There is further a symphysial
portion of the mandible, L6058, with the alveoli for P, (Pl. 33). There appear
to be small alveoli for incisors in the symphyses examined, but none of these
elements has been found. The premolars and molars of the Langebaanweg
Ceratotherium do not show the tendency toward obliqueness of the lophids, or
that toward fossettid formation seen in Ceratotherium simum.

The length of the mandible, L13035, is 570 mm; this measurement is
565-635 mm in adult males, and 550-588 mm in adult females of recent
C. stmum (Heller 1913). The length of the symphysis is 125 mm in L13035 and
145 mm in L11849; 129-155 mm in adult males, and 128-147 mm in adult
females of C. stmum. The fossil specimens agree with the recent in both length
measurements. However, the width at the symphysis is 60 mm in L11849 and
65 mm in L6058, which is decidedly less than that in recent males (g6—-125 mm)
and females (gi-111 mm) of C. simum (Heller 1913). It follows from this that in
C.. praecox the symphysis is relatively (and absolutely) narrower than in C. simum.
It is in D. bicornis that we find such a narrow symphysis: $.A.M. 35658 has a
length of symphysis of 105 mm by a width at symphysis of only 45 mm.

In the height at M,, 125 mm, the fossil mandible L13035 equals C. semum
(S.A.M. 21379), whereas in D. bicornis (S.A.M. 35658) this height is only 85 mm.
The distance from the dental foramen to the base of the posteromedial articular
surface is 160 mm in L13035, against 230 mm in C. simum and 135 mm in
D. bicornis; the jaw orientation in the fossil was evidently nearer to that in
D. bicornis than to that in C. stmum. The condyles in L13035 are not entire, but
the condylar area appears to be more massive, and wider below the condyle
than in C. simum. The medial surface below the condyle is more hollowed than
in either of the two living species. These are also the characters of the Kanapoi
C’. praecox.

TABLE 2

Measurements of lower teeth of Ceratothertum praecox (mm)

No. of specimen L13035 L11849 L13035 L11849
P,, ant.post. 30 _ M,, ant.post. (54) =
ant.transv. 15 17 ant.transv. 32 35
post.transv. 17 7, post.transv. 32 37
P,, ant.post. os — M,, ant.post. 63 56
ant.transv. — 24 ant.transv. 36 27
post.transv. 27 = post.transv. 34. 37
P,, ant.post. 48 45 M,, ant.post. c. 64 60
ant.transv. 30 31 ant.transv. 35 35
post.transv. 32 32 post.transv. 32 34
Length P,-M, - 290 290
Length M,—-M, 175 170

Dental measurements of the two mandibles are given in Table 2. Isolated
lower teeth to be recorded further on considerably expand the variation ranges
Ingsize:

160 ANNALS OF THE SOUTH AFRICAN MUSEUM

Among the isolated teeth from Langebaanweg there are a few unworn or
very slightly worn crowns showing the degree of hypsodonty; these will be
mentioned in the following pages.

An unworn P* sin., L13760 (Pl. 25, bottom left) has an ectoloph height of
go mm by a greatest anteroposterior length of the ectoloph, in the apical third of
the crown, of 55 mm, which gives a height/length index of 164. An unworn
recent P* of Diceros bicornis (Leiden Museum, cat.ost.e) has the same greatest
ectoloph length by an ectoloph height of 80 mm, giving a height/length index
of 145. On the other hand, an unworn P* of recent Ceratotherium simum
(S.A.M. 21382) has an ectoloph height of 103 mm by a greatest anteroposterior
ectoloph length of 46 mm, giving a height/length index of 224.

Among the last upper molars in particular there are several nearly unworn
crowns, as follows: an M? dext., L6696 (PI. 25, middle), an M?® sin., L7106, of
the same individual; an M® dext., L6291 (Pl. 25, middle), an M? sin., L6461,
of the same individual as L6291; an unworn M®? dext., L6638, incomplete
basally and a very slightly worn M? sin., L6636. In L6606 the total height of
the outer surface is 94 mm by a length of the outer surface of 78 mm, giving a
height/length index of 121. The paracone style is a narrow ridge, which fades
away in the basal 35 mm of the ectoloph. L6291 has a height of the outer
surface of 85 mm; the length of the outer surface is 70 mm, giving a height/length
index of 121. Finally, L6638 has a height of the outer surface of 96 mm by a
length of this surface of approximately 80 mm; height/length index c.120.
This is just about the height/length index of M® in modern Duceros bicornis
(outer surface height 64 mm, outer surface length 54 mm, height/length index
119: Hooier 1969: 87), but the Pleistocene Diceros bicornis from the Omo Beds
is lower-crowned than the living form (two specimens of M3, height of unworn
outer surface 56-59 mm, length of outer surface 55-58 mm, height/length
index 102: Hooijer 1969: 87). In modern Ceratotherium simum M3 is 120-130 mm
high (Dietrich 1945: 59), and a slightly worn recent M?® (S.A.M. 21379) is
100 mm high at the outer surface, while an unworn recent M? (S.A.M. 21382),
the outer surface of which is not quite fully calcified at base, is just over 100 mm
high at the incompletely formed external surface. In these recent M? there is
no paracone style but a depression behind the parastyle instead.

The M? from Langebaanweg, L6636, slightly worn, has an ectoloph height
at the metaloph origin of 98 mm by a greatest anteroposterior ectoloph length
of 73 mm; its height/length index is 134. The hypsodonty of Ceratothercum
praecox M®? has already been demonstrated in a slightly worn M? from
Lothagam-1 (K.N.M. LT89 in Hooijer & Patterson 1972) that has an ectoloph
height at the metaloph origin of 74 mm by a greatest anteroposterior ectoloph
length of 63 mm, giving a height/length index of 117. In modern Diceros
bicornis M? (two specimens) the ectoloph is not so much higher than wide,
although the difference is small: M.C.Z., Dept Mamm., no. 51479, height at
metaloph origin 56 mm, length 54 mm, height/length index 104, and Leiden
Museum, cat.ost.b, height 74 mm, length 68 mm, height/length index 109.

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 161

Since the Omo M3? of Diceros bicornis (Pleistocene) is less hypsodont than the
modern M?, the same doubtless holds for the M?. In the fossil M? from Lange-
baanweg the paracone style is present on the apical half of the crown only.

There are two very slightly worn P, in the Langebaanweg collection,
_L5356 and L6693, both from the right side. They are rather similar in dimen-
sions (Table 3), and intermediate in height/length indices between recent
Diceros bicornis (first column) and recent Ceratothertum simum (last column of
Table 3). The discrepancy in height/length indices is the same as that
found in P*.

TABLE 3
Measurements of P, in Diceros and Ceratotherium (mm)
D. bicornis Ceratotherium praecox C. simum
No. of specimen L.M.cat.e L5356 L6693-—s« S.A.M..21382
Greatest length of outer surface 44 49 48 47
Height of metalophid 63 74 70 94
Height of hypolophid (64) 55 68 65 88
Height/length index (a) 143 151 146 200
Height/length index (b) 125 139 135 187

Four isolated lower molars, either M, or M,, are unworn or very slightly
worn. These are L6667 and L2526, from the right side, and L6664 and L6680,
from the left. The height of the anterior (meta-) lophid, taken from the external
base of the crown, varies from 70 to 80 mm; the hypolophid height varies
between the same limits. Unworn M,_., of recent D. bicornis are c. 55 to 65 mm
high, and those of C. szmum c. 80 to 100 mm.

The isolated upper premolars and molars from Langebaanweg are
enumerated in the tables that follow.

Of P? we have nine specimens (Table 4) the first three of which are from
the right side, the others from the left. There is a double crochet in L6649, a
medifossette in L6751, L4750, and L6648, while a bifid crochet is shown in
L6623.

TABLE 4

Measurements of P? of Ceratotherium praecox (mm)
No. of specimen L6649 L6751 L4750 L6648 Lo124 L11957
Ant.post. 38 C. 34 35 — 34 c. 40
Ant.transv. 41 40 42 39 44 39
Post.transv. 40 — 46 41 45 40
No. of specimen L6623 L6629 Lg129
Ant.post. C237 C. 34 c. 40
Ant.transv. 38 a7 44
Post.transv. 41 41 46

P? is represented by nineteen specimens (Table 5) the first eight of which
are from the right side, the others (starting with L11801) from the left. In
L6630 there is seen a slender crista extending to the tip of the crochet; the
internal cingulum is rather weak in this specimen as well as in L6627. L6625 and
L5665 have a bifid crochet, L5665 has in addition a very small crista.

162

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 5

Measurements of P® of Ceratothertum praecox (mm)

L6629 ©6©L6630 )«=6—L6631 = 444 )«©=—sL6625, S—- L6295
48 ¢. 50 47 40 43 40
65 c. 67 64 59 60 60
58 =r 56 52 55 57

L13765 Li1801 L6639 Li1996 L6627 L5695
45 45 43 Pe 47 44
58 64 62 6o 68 61
51 57 58 56 61 58

L5671 Lor14 15451 L6640 L5665 Li3g099 Aterir
45 a 50 46 48 44
60 64 66 63 65 . 61
54 58 63 57 59 56

There are twenty specimens of P* (Table 6) the first eleven of which are
from the right side, the others (from L11132 onward) from the left. L2525 has a
crista joining the crochet. L11132 belongs to the same individual as Li1i21,
has a double crochet the lateral part of which is joined to a crista, thus forming
a medifossette (Pl. 29, bottom). L6655, slightly worn, shows a double crochet
and a crista (Pl. 29, bottom). L13760 shows the full height of the ectoloph, with
a height/length index of 164 (Pl. 25, bottom left). Medifossette formation is
very rare in P* and M!~%, one in twenty or three in forty Langebaanweg teeth.

No. of specimen
Ant.post.

Ant. transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

No. of specimen
Ant.post.
Ant.transv.
Post.transv.

TABLE 6
Measurements of P* of Ceratotherium praecox (mm)

L6717. L2525 L6652 L6299 Li167 L6619
5! 54 49 49 5! 50
70 74 70 70 UD 71
63 69 61 61 67 61

L6739 6 L5606)=—L3454) «=Lit12a1 Li1132 L6618

G57 51 47 48 48 c. 52
75 69 67 68 68 71
64 63 56 62 62 66

L6632 L6296 L4612 L6622 113760 L130909
— 48 50 c. 50 47 48
69 70 7a 68 68 64.
63 64 68 63 62 60

TABLE 7
Measurements of M? of Ceratotherium praecox (mm)

L6626 L6624 L6703 5445 Lo113 L6628
58 55 a 57 60 52
73 72 73 72 74 71
68 67 66 65 65 63

L5912 L6293 ~Li2039 L5311 L6647 15418
55 — C.52 56 56 61
68 67 66 69 73 72
65 60 61 65 68 68

L6465 14749 L5g919

c. 53 54 58
69 70 70

66 61 65

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 163

M? is represented by seventeen specimens (‘Table 7) the first eight of which
are from the right side, the left specimens beginning with L6293. Medifossettes
are not formed; L6626 has a double crochet (Pl. 28, bottom right).

There are twenty-four specimens of M? (Table 8) the first ten of which are
from the right side, the remaining specimens (starting with L6636) from the
left. A true medifossette, formed by the union of crochet and crista, is shown
only in Lg116, L5916 (external surface broken off: Pl. 28), and L10983. A small
crista is seen in L5917, L6617 (in which the crochet makes a contact with a
small projection on the posterior face of the ectoloph: Pl. 28), L6746, L6654,
L6641, L6644A, and L12360. Lo118 consists of ectoloph and crochet only; the
crista is in contact with the crochet apically (Pl. 29). L6636 is the specimen with
the ectoloph slightly worn, and a height/length index of 134, already referred
to above.

TABLE 8

Measurements of M2 of Ceratotherium praecox (mm)
ING; of specimen ... L5917 L6617 L66q4 Lo116. L663: L6634 L5911 L6746

Ant.post. ¢. 50 67 c. 62 62 c. 60 61 63 c. 64
Ant.transv. 75 82 82 75 79 77 79 74
Post.transv. 68 75 69 73 ii 72 — —
No. of specimen L6637 L6643 \L6636 L5916 Lo118 L6654 L6645 L10983
Ant.post. — — c. 60 — — 6.55 56 55
Ant.transv. 80 80 73 — — 76 78 71
Post.transv. 72 73 66 — — 68 67 67
No. of specimen L6644B L6654 L6641 L6644A L11898 Li12360 L6653 Lo115
Ant.post. c. 62 63 67 65 65 6.55 ¢.55 —_
Ant.transv. 74 70 75 79 77 73 71 77
Post.transv. 67 65 67 72 68 70 65 68

We have seventeen specimens of M? (Table g) the first ten of which are
from the right side, the remaining (to begin with L7106) from the left. L6606 is
a slightly worn specimen with a height/length index of 121 (Pl. 25); L6291 isa
somewhat smaller specimen likewise slightly worn and with the same index
(Pl. 25). The left M? L7106 belongs to the same individual as L6696, and the
left M? L6461 belongs to the same individual as L6291. The specimen L6638 is
unworn but incomplete at the base of the crown; its height /length index is c. 120.

TABLE 9

Measurements of M3 of Ceratotherium praecox (mm)
No. of specimen L6696 L629: L6638 L6294 L6620 L5666 L10984
Ant.post. (int.) 66 58 c. 65 56 58 61 61
Ant.transv. 71 65 75 68 69 74 66
Length outer surf. 78 70 c. 80 78 75 75 70
No. of specimen L6641 L6290 Litg97 L7106 L6461 L13614
Ant.post. (int.) 64 61 72 — 58 60
Ant.transv. 71 75 75 — 65 7g:
Length outer surf. 76 81 80 78 &. 70 72
No. of specimen L6289 L6642 Litog1 L6466
Ant.post. (int.) 60 57 58 66
Ant.transv. 67 69 69 72

Length outer surf. 75 73 68 74.

164 ANNALS OF THE SOUTH AFRICAN MUSEUM

These specimens have already been referred to above. L6294 has a crochet
extending all across the medisinus; L6641 and Li1997 have a very large
crochet, and L10984 has a small crista and an internal projection at the base of
the crochet.

Some of the remaining lower cheek teeth are zn stu in incomplete mandibles,
as follows: L6615, a right mandibular ramus, has the posterior portion of P, and
the three molars; the lengths are reduced as a result of interproximal wear, their
transverse diameters slightly exceed those in L13035 and L11849 (Table 2),
and the height at M, is 130 mm. P,—M, sin. and P,—P, dext. of one and the same
individual, L6659, are narrower-crowned, as are those recorded in Table 2. A
right and a left mandibular ramus with the much worn M,—M, on either side
(L6612, L6614), give an M,—M, length shorter than that in the less worn
dentitions. The height of the ramus at M, is 125 mm. L6793 isa right mandibular
ramus with P,-M,; L1198q is a right ramus fragment with M,, Mg, and part
of M;. Two parts of right rami, L13759 and L13805, have M,, and Mg,
respectively, in situ. ‘The measurements of these teeth are given in Table to.

TABLE 10

Measurements of lower teeth of Ceratotherium praecox (mm)

No. of specimen L6615 L6659 L6612 L6793 L11989 L13759
L13805
P,, ant.post. = = a cae re Rr.
ant.transv. — 18 — 30 3 7
post.transv. — 20 = = = rie
P,, ant.post. — (47) = = = —
ant.transv. — 24 = — =a =
post.transv. — 28 = = as nee
P,, ant.post. = 52 ms (43) por i;
ant.transv. = 29 = 31 is aa
post.transv. ae 31 3 7a en 7h
M,, ant.post. (40) 57 (44) a 55 ras
ant.transv. 37 32 37 ara 39 aay
post.transv. —= 34 36 a came ae
M,, ant.post. (51) 65 (52) (54) 59 cs
ant.transv. 38 — 37 37) 37 36
post.transv. — = 39 37 38 oy
M,, ant.post. 59 — 58 = a 63
ant.transv. 38 — = = =e a”,
post.transv. 38 —_ 35 = ae 38
Length M,-—M, 160 — 150 — = =

There are seven isolated specimens of P, (Table 11) the first four of which
are from the right side. The first and the last specimen are decidedly larger than
the P, in the two mandibles of Table 2.

TABLE II

Measurements of P, of Ceratotherium praecox (mm)

No. of specimen Li1812 L6684 L6676 Liig99 Li1959 L6665 L11815
Ant.post. 33 (30) a (27) 30 32 34
Ant.transv. 20 18 7 17 18 17 21

Post.transv. 23 23 21 18 20 18 23

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 165

There are ten isolated specimens of P, (Table 12) the first three of which
are from the right side.

TABLE 12

Measurements of P, of Ceratotherium praecox (mm)
No. of specimen L12107 Li1810 L6671 L6681 L6669 Lir811
Ant.post. (40) (44) (41) 48 (45) (38)
Ant.transv. 25 27 24 25 28 24.
Post.transv. 28 30 31 26 28 26
No. of specimen L2525 L5698 L11809 L5697
Ant. post. (42) (42) (43) 50
Ant.transv. 25 24 27 27
Post.transv. 30 28 30 28

P, is represented by eight specimens (Table 13) the first two of which are
from the right side. L6687, L12108, and L11804 are larger, especially wider,
than their homologues in the dentitions of ‘Tables 2 and ro.

TABLE 13

Measurements of P, of Ceratotherium praecox (mm)
No. of specimen L6687 L12108 L11804 L11816 L6762
Ant.post. 49 53 51 (45) 50
Ant.transv. 33 33 33 30 29
Post.transv. 34 34 37 31 31
No. of specimen L6670 L475! L6662
Ant.post. | (47) 55 (49)
Ant.transv. 28 — 31
Post.transv. 35 31 33

Fourteen isolated lower molars represent either M, or M, (Table 14); the
first eight are from the right side, the remaining six (beginning with L6672)
from the left.

TABLE 14
Measurements of M, and M, of Ceratotherium praecox (mm)
No. of specimen L6678 L6679 L2525 L6690 L6302 L5690 L11894
Ant.post. (56) 58 60 (57) 60 65 (50)
Ant.transv. 34 30 31 34 35 34 37
Post.transv. 35 33 33 35 34 34 37
No. of specimen L6677. L6672 14752 Lo126 L5669 §©6L6689 613390
Ant.post. 62 60 — 60 (55) (49) 65
Ant.transv. 38 oF 38 38 38 35 38
Post.transv. 36 39 35 By 37 87 36

The last lower molar, Ms, is easily distinguishable from M, or M, by its
reduced posterior cingulum; in well-worn specimens the absence of a posterior
pressure scar of course is characteristic for M;. There are eight isolated M,
(Table 15) the first two of which are from the right side.

TABLE 15
Measurements of M, of Ceratotherium praecox (mm)
No. of specimen L5667 L11802 L6613 Li11989 Lo609 Lo125 Loi1o Lg120
Ant.post. 68 66 (55) 65 67 65 65 (57)
Ant.transv. 38 40 37 42 41 39 36 36
Post.transv. 34 37 36 oH 30 34 34 34

166 ANNALS OF THE SOUTH AFRICAN MUSEUM

In Tables 10 to 15, inclusive, the anteroposterior diameter is in parentheses
when it is much reduced because of interproximal wear.

There are a number of teeth belonging to the milk dentition of Ceratotherium
praecox. Deciduous teeth were not present among the material of this species
described from Kanapoi, Lothagam-1 and Ekora (Hooijer & Patterson 1972).
Therefore, the Langebaanweg milk teeth are compared below with those of the
two living African species. The differential characters of the milk molars of
Diceros bicornis and Ceratothertum simum are recorded in Hooijer (1959). In a
collection from Late Pleistocene sites near Swartklip, Cape Province, reported
upon by Hendey & Hendey (1968), there are milk molars of C. semum, which
have been used for comparison.

The maxillary milk dentition of Ceratotherium praecox comprises two isolated
and much worn DM}, both from the left side, L6674 and L6675, measuring
22 mm anteroposteriorly and 23 mm transversely. In C. stmum DM! is more
elongated anteroposteriorly than in D. bicornis because of the greater forward
projection of the parastyle in the former, but this character is lost in much worn
specimens like those from Langebaanweg and a distinction cannot be made at
this stage of wear.

Of DM? there are two specimens in the Langebaanweg collection, L4608
(Pl. 29) and L5664, both from the left side. DM? is represented only by a single
specimen, L9105B (Pl. 29), from the left side and lacking most of the ectoloph.
Finally, of the last upper milk molar, DM‘, we have three specimens, one right
lacking the outer surface, L6727, one entire left DM*, L13818, and another left
specimen, much worn down, L6651 (Pl. 29). The upper milk molars in
C. stmum are distinguished from those in D. bicornis by the more prominent
parastyle, suppression of paracone style, greater crown height, absence of inner
cingula, stronger crista joining the crochet and forming a medifossette, and the
postsinus being approximately as deep as the medisinus instead of shallower.
The inner portion of the protoloph is more distinctly curved backward in
C. sumum than in D. bicornis, but this difference is more marked in the posterior
milk molars than in DM2?, in which it is not or hardly evident. Upper milk
dentitions of C. stmum and of D. bicornis have been described from the Early
Pleistocene Makapansgat caves (Hooijer 1959); they tend to be on the large
side but otherwise indistinguishable from their recent homologues. Variation
ranges in dimensions of the milk teeth of the recent species are presented in
Table 16 along with the measurements of the Langebaanweg specimens and
those from Swartklip in the South African Museum; the Swartklip specimens
conform to those of C. simum in every respect (they bear catalogue numbers
preceded by ZW).

The DM? of Ceratotherium praecox, L4608, has a prominent parastyle as in
C. simum but has an internal cingulum along the protocone, as in D. bicornis.
There is a tubercle at the medisinus entrance that is absent in L5664; both
specimens have a well-developed crista joining the crochet and forming a

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 167

medifossette. The postsinus is almost as deep as the medisinus; these are, again,
C’. simum characters.

TABLE 16
Measurements of upper milk molars of C. praecox and recent species (mm)
DM?, no. of specimen L4608 L5664 D.bicornis C.simum ZWi92 ZW2610

Greatest length ectoloph 42 — 38-41 41-51 41 42
Antero-transverse 37 35+ 33-39 36-41 36 34
Postero-transverse 42 38+ 35-40 35-43 — 33
DM%, no. of specimen Lo105B- Dz. bicornis C.simum ZW1842

Greatest length ectoloph — 45-52 53-61 53

Antero-transverse — 40-50 46-48 46

Postero-transverse — 39-47 44-46 45

DM,‘ no. of specimen L13818 L6651 OD. bicornis C.simum

Greatest length ectoloph 60 54+ 50-55 66-68

Antero-transverse 56 53 45-53 = ¢. 54-55

Postero-transverse 53 52 40-51 52-60

L5664 is incomplete internally, but the minimal transverse diameters can
be given. The Langebaanweg DM? tally well in size with those of C. simum. The
two Swartklip specimens of DM?, ZW192 and ZW2610, both from the right
side, lack the internal cingulum, display well-formed medifossettes, and have
the postsinus as deep as the medisinus, as in C. semum to which they belong. The
same holds good for the Swartklip DM?, ZW1842, which is from the left side.
In the Langebaanweg collection there is but one DM?, Lg105B, wanting most
of the ectoloph. There is a slender crista, not joining the crochet, hardly any
trace of an inner cingulum (except at the medisinus entrance), but the postsinus
is less deep than the medisinus, as in D. bicornis. No measurements can be given.
Of DM? we have three Langebaanweg specimens, one right, lacking the outer
surface, and two from the left side, as listed above. The entire specimens show
the absence of the paracone style, the formation of a medifossette, and the
absence of an inner cingulum, as in C. szmum, although the postsinus is decidedly
less deep than the medisinus, as in D. bicornis. Thus, the C. praecox milk molars
combine characters found in C. simum and D. bicornis, whereas in size they are
intermediate between the two.

Of the mandibular milk dentition there are the following specimens:
L6686, DM, dext., slightly worn; Lg105C, DM, dext., unworn (Pl. 32),
metalophid height 41 mm, and hypolophid height 38 mm; L6301, DM, dext.,
slightly worn; Lg105A, DM, dext., unworn, metalophid height 50 mm, and
hypolophid height 46 mm; L6689, DM, dext., much worn down; L6795, left
ramus with incomplete DM,_,; L6660, DM, dext. in ramus fragment, slightly
worn (Pl. 32), crown not fully erupted, anteroposterior diameter 54 mm,
as in Lo1o5A; L2524, DM, sin. in ramus fragment, crown edge broken,
lingual base not exposed; L12870 and L6757, both DM, sin., slightly worn.

As shown in Table 17, the Langebaanweg lower milk molars are larger
than those in D. bicornis, as were the upper milk molars, but they correspond
rather well with those from Swartklip, which represent C. simum. These Swart-
klip specimens are: ZW1837, DM,_, dext. in ramus fragment; ZW2036,

168 ANNALS OF THE SOUTH AFRICAN MUSEUM

DM,_, dext.; ZW1867, DM, dext.; ZW1876, DM, dext., and ZW1966,
DM, sin., unworn, metalophid height 45 mm, hypolophid height 42 mm. The
DM, of C. praecox that is unworn, Lg105C, has the anteroposterior diameter

TABLE 17

Measurements of lower milk molars of C. praecox and recent species (mm)

DM.g, no. of
specimen L6686 ZW ZW ZW D2. bicornis
1837 2036 1867
Ant.post. 41 40 40 39 27-33
Ant.transv. 16 16 — — 13-15
Post.transv. 18 20 — — 15-18
DMs, no. of
specimen Lo1o5 L6301 16795 ZW ZW ZW OD. bicornis
1837 1876 1966
Ant.post. 48 46 CaAg 46 45 44 38-41
Ant.transv. 20 = (228 22 = 22 19-20
Post.transv. 22 QI 25 23 — 23 20-22
DM,, no. of
specimen Lgo105 L668g L12870 16757 ZW D. bicornis
1837
Ant.post. 54 (46) 51 51 48 41-45
Ant.transv. 23 23 25 25 — 22-23
Post.transv. 26 25 27 23+ — 23-25

longer than that in the unworn DM, of C. simum, ZW1966 (48 against 44 mm),
whereas both in metalophid height and in hypolophid height Lg105C is less
than is ZW1966 (41 and 38 mm against 45 and 42 mm). It follows from this
comparison that the milk tooth of C. praecox is less hypsodont than that of
C. simum; we got the same result from the unworn permanent premolars and
molars. }

PoOsTCRANIAL SKELETON

The postcranial material, which is very abundant at the Langebaanweg
‘E’ Quarry, is listed in the tables of measurements that follow (18 through 50).
Measurements of the bones of D. bicornis and C. simum have been given in
previous papers (Hooijer & Singer 1960; Hooijer 1969) from skeletons in the
South African Museum, Cape Town, and in the Osteology Department,
National Museum Centre for Prehistory and Palaeontology, Nairobi, respec-
tively. In both cases the C. simum skeleton is larger than that of D. bicornis, with
more massive metapodials (higher width/length ratios), but other than that no
skeletal differences between the two extant species are apparent. Most of the
Langebaanweg bones are larger than their homologues even in C. simum.

Eleven proximal portions of scapulae are in the Langebaanweg collection
(Table 18) the first five of which are from the right side.

Pont > 7

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 169

TABLE 18
Measurements of scapula (mm)

No. of specimen L8244 L11773 Lir524 L8245-
1. Ant.post. diameter of collum scapulae 135 — — 6.130
2. Ant.post. diameter over tuber scapulae and glenoid

cavity c. 170 155 — ¢.170
3. Ant.post. diameter of glenoid cavity 4 CLL” GAIOO> | CF 210))).. 16; 105
4. Transverse diameter of idem ¢. 95 ¢. gO 95 95
5. Transverse diameter of tuber scapulae — 55 — 60
No. of specimen L8306 113857 L8288 18287 L8290 L8266 113779
Ti — 125 120 130 125 — 135
2 — 165 6.155 165 165 165 175
3. ¢. 105 105 ¢. 100 105 105 105 —
4. 95 c. gO Cc. gO 100 — — _—
5 — 65 65 c. 60 c. 60 6.70 —
No. of specimen D. bicornis C. simum
I 100 130
2 130 160
3. 85 105
4. 80 100
5 45 60

The mid-portion of the shaft of a right humerus, L3421, has a width at the
deltoid tuberosity of 170 mm, and a least width of 85 mm, as in C. simum
(Hooijer 1969: 91). There are further only distal portions of the humerus, nine
in all (Table 19) the first six of which are from the right side.

TABLE 19

Measurements of humerus (mm)
No. of specimen L6886 L6977 L13559 L6947 L6878 L6899
1. Least width of shaft go 80 c. gO 85 — go
2. Greatest distal width —- 190 — — c. 180 c. 190
3. Width of trochlea 135 130 130 c. 130 125 125
No. of specimen L3423 L6965 L13463 =D. bicornis C.simum
I. 80 — 80 60 70-85
a 185 —- = 150-155 180
ee 120 125 — 100 120

A radio-ulna dext., L12818, is slightly damaged proximo-medially; the
radius is longer than any of the fossil radii, four of which are nearly entire
(Table 20); only the last specimen in this table, L4967, is from the left side.

TABLE 20

Measurements of radius (mm)
No. of specimen L12818 L7997 L6375 L8114
1. Median length 400 375 ¢. 370 385
2. Proximal width — 125 130 130
3. Proximal ant.post. diameter (medial side) _ 80 85 85
4. Least width of shaft 75 75 70 70
5. Greatest distal width 120 120 115 —
6. Width distal articular surface 105 100 100 95
No. of specimen L4967 8D. bicornis C.simum
I, 390 345-350 365-380
2. 135 100 120-125
3. go 60 75
4. 75 45-55 65-70
5. — 95 120
6. 105 80 100

170 ANNALS OF THE SOUTH AFRICAN MUSEUM

There are twenty-one proximal portions of the radius (Table 21); the first
seven are from the right side.

TABLE 21
Proximal measurements of radius (mm)

No. of specimen L13175 L6371 L6370 L3425 Lg981 L7983
2. Proximal width 135 130 125 125 120 c. 130
3. Proximal ant.post. diameter

(medial side) go c. 80 80 85 c. 85 go
4. Least width of shaft -- 65 65 — — —
No. of specimen L7934 L7968 L8o017 L8007 L4205 113845 L12888
va 130 — 125 ¢.125 115 125 130
2 go ¢.95 80 — 85 85 go
4. aa — — 65 70 65 70
No. of specimen L7986 L4959 Lo978 =L2229:)«€6LL8015 ~=L2289 §=6—6L7958 Lg988
2. 125 130 13h) | Gal25 125 125 125 C20
3. 80 85 c. 90 85 80 c. 80 c. 85 80
4. 75 70 80 70 — _ — —_

Distal radius portions number thirty-three (Table 22), fifteen from the
right and eighteen from the left side.

TABLE 22
Distal measurements of radius (mm)

No. of specimen L6369 «©=—«L6367 - L7973° «=L6177 «= L4202 ~L13649 ) Egogn
4. Least width of shaft 65 70 70 _- — 65 —
5. Greatest distal width 120 120 120 120 E115, 86, tO 110
6. Width distal articular

surface 100 105 105 110 95 go 95
No. of specimen Lo985 L7911 L8o010 L70961 L6é362\ LGi19g3itgess
4. ioe a i a aare ai rte
5. 120 120) ) Jove 120 120" “e186 120
6. 100 105 105 105 100 105 100
No. of specimen 15174. “L8o01r2 1139842 Le290° L2293 §©6L22qn ages
4. = 70 70 70 65 65 5
5. c. 120 T1Op ues 20 I15 —- 120 -—
6. 05 90 105 105 100 100 go
No. of specimen Lo986 4203 IL9730 ©L6170° L4194 L6372))EeoG7
4. ie aa at = rR ra a}
5. 115 1 US pam on iS LTO 9 0.10); 120 110
6. 105 100 100 95 100 100 100
No. of specimen L7924 L8006 L4957 L7920 L4200
4. a prz a) a ea
5. 120 ¢.105 110 IIO 115
6. 95 100 go 90 100

The ulna of the radio-ulna dext., L12818, is the only entire ulna in the
Langebaanweg collection; it has a maximum length of 530 mm (D. bicornis
450 mm; C. stmum 510 mm), and a length from the processus anconaeus (beak)
to the extremity of the olecranon of 175 mm (D. bicornis 140 mm; C. simum
165 mm). Further measurements are given in Table 23. In this table, twenty
proximal and distal ulna portions are listed; the first twelve are from the right
side, the remaining eight (beginning with L8052) from the left.

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE

171

TABLE 23
Measurements of ulna (mm)

No. of specimen L12818 L8060 113836 L8038 L8071 L8055
1. Width at semilunar notch — — ¢. 95 95 ¢. 105 105
2. Greatest distal diameter go — — — — _
3. Ant.post. diameter distal articular

surface 60 : — — _ —
No. of specimen L4210 L7984 L7959 L8029 L8025 4L7985 L804!
Te — —— — — — _ —
2. 85 c. 80 c. 80 c. 80 C70 85 _
3. 60 Boy 8 50 55 55 55 BD
No. of specimen L8052 113833 Lgg94 L6251 113839 Li2891 13554
it. 105 110 105 105 _- — —
Pls — — — — 80 85 go
2. _ — — —_ 60 60 60
No. of specimen L7927_ =D. bicornis §=C. simum
Ti — go 110
P 80 75 go
3. 55 60 65

There are twenty-six scaphoids (Table 24) the first ten of which are from
the right side.

TABLE 24

Measurements of scaphoid (mm)
No. of specimen L6010 L6o012 L6003 L6o0g 19477
1. Posterior height 63 60 58 57 60
2. Anterior height 58 60 59 59 61
3. Proximal width 55 56 54 55 54
4. Proximal ant.post. diameter 75 73 74 70 78
5. Maximum diameter, distal facets 70 70 69 68 ak
No. of specimen Li1767 L7850 111768 L5284 Lo483 113472 L6218
lis 63 60 65 62 60 66 62
P 62 60 63 61 58 64 59
3- 55 61 57 60 53 55 61
f 74 76 79 77 75 77 77
5: 75 U1. 75 79 75 73 19
No. of specimen L7809 L6014 L6008 L7849 L5986 L3569 §©6—L7803
Tis 60 67 60 63 67 63 65
2. 59 60 63 63 64. 64. 59
3- 55 60 57 58 60 55 57
4. 1 87 73 84 80 74 78
5: 75 78 72 (See 74 76
No. of specimen L7735 L7826 L4290 L13616 L7861 L7738 L5282
I. 66 65 61 64 64 67 57
ois 63 65 59 60 63 68 58
3. 60 60 56 55 59 59 54
4. 76 79 73 81 82 a7 68
5: 73 75 69 75 80 77 68
No. of specimen D. bicornis C. simum
I. 50 62
De 54-60 58-65
3. 55 60
4. 63 75
5.

172

ANNALS OF THE SOUTH AFRICAN MUSEUM

Of the lunar there are thirty-six specimens (Table 25), and the first

eighteen are from the right side.

TABLE 25

Measurements of lunar (mm)

No. of specimen L4253 L12379 L6006 L7853 L7829
1. Anterior height 64 61 65 69 61
2. Proximal width 67 61 64 68 63
3. Greatest ant.post. diameter 79 83 81 87 78
No. of specimen L528: L4270 L304g L4787A Lo475 4287
i 65 57 58 60 59 60
ou, 68 58 66 58 61 60
3- 75-5 ling 76 78 78 Vd
No. of specimen L5290 «639113823 «7737. «=2L7755 +5975 L13824
Ts 62 62 58 64 59 66
2. 64 60 59 62 56 66

3 79 78 76 80 75 78
No. of specimen Li1596 L7885 L7822 17771 113727 .Lo184
I 66 58 61 60 58 63

2 65 60 63 60 66 64

3 86 76 76 78 76 83
No. of specimen L5293 ~4L11598 L7896 L4286 L5972 L780
I 60 58 64 63 61 65

2 60 63 65 63 63 65

3 75 75 80 81 79 78
No. of specimen L3806 L7774 D. bicornis C.simum

I 59 64 44-48 54-60

2 58 65 48 58-62

3 77 80 64-68 75

Fifteen specimens of the cuneiform are in the Langebaanweg collection
(Table 26), ten right and five left; L9465 is presented in anterior view in

Pl. 33 (bottom).

TABLE 26

Measurements of cuneiform (mm)

No. of specimen L12765 L3405 L5218 14265
1. Anterior height 57 51 52 47
2. Distal width 56 48 46 49
3. Proximal ant.post. diameter 51 47 44 43
4. Greatest horizontal diameter 68 63 58 58
No. of specimen L7808 L7833 L7869 L7898 L13821 Lg254
I. 55 53 56 57 56 45
2 a 44 59 51 C. 47 43
3- 48 41 43 47 45 43
4. 61 56 61 64 60 57
No. of specimen L3566 Lo471 Lo9465 OD. bicornis C. simum

I. : 59 52 53 50 56-58

2. 52 50 53 38-40 45-59

3. 51 43 48 38-40 48-51

4. 67 57 62 53 66

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 173

Three pisiforms, one right, L6004, and two left, L7854 and L78g2, are in
the Langebaanweg collection (Table 27); L78g2 is presented in anterior view
in Plate 33 (bottom). The bones have the two facets, for ulna and cuneiform.

TABLE 27

Measurements of pisiform (mm)

No. of specimen L6004 L7854 L7892 Dz. bicornis C.simum
1. Length 71 72 67 61 60
2. Distal height 51 eto) 43 35 36

An exceptional bone is L7823, a cuneiform sin. with the pisiform completely
ankylosed to it. The part representing the cuneiform is normal in shape, but it
forms a solid mass with the pisiform, and the ulnar facets of the two bones are
confluent (Pl. 33, middle). The greatest horizontal diameter of the anomalous
bone is just over 110 mm (the distal extremity of the pisiform is incomplete).
For comparison a cuneiform and a pisiform are figured along with the
cuneipisiform (Pl. 33, bottom).

The trapezium, the radial of the distal row of carpal bones, with facets for
the scaphoid and the trapezoid, is represented in the Langebaanweg collection
by a single specimen, L3497; it is from the right side. In Table 28 the fossil bone
is shown to be larger than its homologue in C. simum, as is usual for
Langebaanweg bones.

TABLE 28

Measurements of trapezium (mm)

No. of specimen L3497 Dz. bicornis C. simum
1. Height 35 31 35
2. Proximal diameters Boe 22025 oT a gs 2) Only

The trapezoid is represented in the Langebaanweg collection by five
specimens, three from the right and two from the left side (Table 29).

TABLE 29

Measurements of trapezoid (mm)

No. of specimen L7798 Li1881 Lr3999 4263 «=9L4267_ ~(COD.:. bicornis C. simum
1. Anterior width 35 a7 37 35 38 30 35
2. Anterior height 38 45 34 39 38 31 32
3. Posterior height 35 50 33 36 37 29 36
4. Ant.post. diameter 51 55 49 48 52 41 49

Of the magnum we have twenty-one specimens (Table 30) the first ten of
which are from the right side.

174

No. of specimen
1. Anterior width
2. Anterior height

3. Proximal ant.post. diameter

4. Greatest diameter

No. of specimen
ie
2.
3.
4.
No. of specimen

b

OO NW

No. of specimen
I .

2.
3.
4.

ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 30
Measurements of magnum (mm)
L7793 «44244 Li12824 L4078 15568
59 59 65 58 56
40 40 38 34 42
83 80 83 68 76
103 104. 115 c. 85 96
Lo9473. Lo460 Lo459 «©6—LO6013 «—L 4264 «6095184 3 L6217
62 61 57 52 56 56 56
38 45 41 38 40 40 41
80 85 84 74. 74 82 78
— — — — — 105 107
L7876 15259 «27743 ~L11592 ~L9476 =9L4283 L745
59 ¢. 60 57 60 6. 55 54 57
38 42 45 43 42 40 ¢. 40
oF] 86 88 — 80 77 78
04 105 112 101 IOI — —
L7797. L7759 =D. bicornis = C. simum
51 cide 55 44-49 57-58
34 40 32 38
Ta 82 63-67 70-71
aa a 77-85 84-85

There are forty specimens of the unciform in the Langebaanweg collection
(Table 31), twenty from the right, and the same number from the left side.

No. of specimen

1. Anterior height
2. Anterior width

3. Greatest diameter

No. of specimen

. of specimen

. of specimen

. of specimen

. of specimen

QO Nes

L11590

TABLE 31

Measurements of unciform (mm)

L7762 L7812 Li2824 Lo193 Li2766 L5262 L7747
58 64 65 57 62 56 61
81 87 83 77 81 75 83

103 112 107 100 108 98 105

L6016 L4240 Lo4g61 Lo2o1 L6005 L7879 L11597
62 55 60 56 55 63 61
81 76 78 72 80 88 85

105 102 100 98 Koen a I 103

L7855 L5260 L7870 Lo184 L4076 Lito97 L7837
64. 55 60 57 58 56 62
87 73 77 75 83 75 83

108 98 102 95 109 97 105

L4285 L12826 Lo468 15973 113829 Li1591 L7840
57 61 60 54 58 63 58
76 79 81 76 83 QI 85

100 105 102 96 108 112 110
L4789A L4256 Li1600 15263 L5258 L7742 Lg469
55 63 53 55 65 60 57
74 85 78 rH] 82 80 80
104 107 95 99 106 102 108

L5257 Lo466 Lo464 OD2z. bicornis C. simum
5 60 57 49-51 55
74 84 79 63-65 74-78

100 106 105 84-90 99-100

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 175

There are twenty entire second metacarpals (Table 32) the first five of
which are from the right side. The ratio middle width/median length in the
fossil series varies from 0,21 to 0,28, which includes the observations on the
recent Mc.II (taken from Hooijer & Singer 1960, and Hooijer 1969).

TABLE 32

Measurements of second metacarpal (mm)

No. of specimen L3066 L4890 L5934 6—Li2819 ©=L5988
1. Median length 173 Gz 167 172 160
2. Proximal width 42 39 43 44 38
3. Proximal ant.post. diameter 52 Cc. 50 55 54 50
4. Middle width 43 40 42 — 38
5. Middle ant.post. diameter 24 23 23 — 21
6. Greatest distal width 50 49 56 56 50
7. Width distal trochlea 44 43 45 46 44
8. Distal ant.post. diameter 51 49 49 50 48
g. Ratio middle width/length 0,25 0,23 0,25 — 0,24
No. of specimen oger, Le7i1r Lyr29 «Li71096)«6LL7072 «©|© L9395: «= L083
ie 175 158 172 167 159 172 176
2 45 37 44 40 37 41 4!
3 55 55 53 51 47 58 51
4 45 37 42 36 35 41 37
5 25 20 23 23 QI 25 24
6 53 52 49 46 47 52 56
7 46 44. 45 42 42 45 47
8 48 48 48 47 45 51 45
’ 0,26 0,23 0,24 0,22 0,22 0,24 0,21
No. of specimen L7093. ~L7o90 L4104 L7064 L7154 L4132 L6064
re 162 176 176 158 183 157 163
2 42 40 42 40 41 40 38
5 ¢. 50 52 by) 49 53 54 5!
4 4! 39 4! 35) 45 39 35
5 19 26 29 22 23 24 24
6 50 53 53 47 57 52 50
7 42 46 42 42 46 45 45
8 45 47 48 43 51 48 45
9 0,25 0,22 0,23 0,22 0,25 0,25 0,21
No. of specimen L7071 D. bacornis C. simum Chemeron
I 166 147 148 160 +160 165
2 — 32 40 44 45 ¢. 45
3 58 46 36 44 49 C. 50
4 47 33 31 40 40 42
5 28 19 18 20 24 23
6 57 39-37 45 50 =
7 48 Cn = gh. gA@ 5
8 50 Ar 38 43 45 =
9 0,28 0,22 0,21 0,25 0,25 0,25

The third metacarpal is represented in the Langebaanweg collection by
twenty-two entire specimens (Table 33), eleven right and eleven left. The bone
L3070 is a diseased specimen, somewhat like the second metatarsal of Dicerorhinus
leakeyi Hooijer (1966, pl. 15) from Rusinga Island.

I 76 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 33

Measurements of third metacarpal (mm)
No. of specimen L5962 =L11356 L7086 L708: L7100
1. Median length 186 206 P20 1192 198 185
2. Proximal width 67 78 65 76 71
3. Proximal ant.post. diameter 58 67 55 65 63 |
4. Middle width 57 66 56 63 57 |
5. Middle ant.post. diameter 27 31 27 28 26 |
6. Greatest distal width 74 78 70 78 73
7. Width distal trochlea 63 WP 62 69 65
8. Distal ant.post. diameter 51 57 52 58 51
g. Ratio middle width/length 0,31 0,32 0,29 0,32 0,31
No. of specimen L7o080 L13750 Lgo070 Lée1, e275 Lbesg a eesge
1: 193 186 195 187 203 194 200
2. 70 64. 69 65 72 67 74
3. 56 57 — 5h a 58 57
4. 58 54 C200 58 61 59 61
5. 30 24 — 27 27 26 31
6. 75 67 — 69 80 70+ 78
ie 63 58 65 62 67 62 68
8. 51 50 = 54 55 50 54
Q. G30 | iasgo — 0,31 0,30 0,305) iene
No. of specimen L7001 L12822 L5937 L593: L414g 113756 Log381
i: 183 195 182 188 188 206 192
2 72 Ti 70 72 78 69 71
3 57 61 ey) 56 58 58 54
4 63 66 55 58 64 63 57
5 25 28 24 27 26 24 24
6 73 = 73 ps 5 71 73
7 61 66 61 58 64 60 62
8 = 56 53 50 54 54 51
9 0,34 0,34 0,30 0,31 0,34 0,31 0,30
No. of specimen Lo408 L7085 L13580 D. bicornis C. simum
i: 186 183 192 162 166 173. 1176
2 72 78 77 59 60 70 68
3 58 61 60 48 51 55 52
4 64 64 58 46 45 50 58
5 27 30 30 22 22 24 28
6 — 82 74 61 52 66 71
i 69 64 60 Be eg ree
8 ar 52 53 44 41 48 48
9 0,34 0,35 0,30 0,28 0,27 0,32 0,33

The variation range in width/length ratio in the Langebaanweg Mc.II]I,
0,29 to 0,35, is such that it includes the observations of C. stmum but the two
D. bicornis metapodials are relatively more slender than the fossil specimens,
although the difference is small.

There are sixteen entire fourth metacarpals in the Langebaanweg collec-
tion (Table 34), the first five of which are from the right side. In this metacarpal,
only one of the two D. bicornis is below the variation range in width/length ratio
in the fossil specimens.

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 177

TABLE 34

Measurements of fourth metacarpal (mm)
No. of specimen L6631 L7084 L5936 L5949 L7098 L12820
1. Median length 148 147 155 151 149 157
2. Proximal width 53 50 54 55 53 58
3. Proximal ant.post. diameter 47 52 51 48 50 50
4. Middle width ' 39 42 40 37 40 44
5. Middle ant.post. diameter 24 24 24 23 27 25
6. Greatest distal width 50 58 — 48 52 55
7. Width distal trochlea 46 49 46 42 42 45
8. Distal ant.post. diameter 46 45 — 42 44 46
g. Ratio middle width/length 0,26 0,29 0,26 0,25 0,27 0,28
No. of specimen L2285 Lo4g1r L7078 L413: Lo9246 L7089 L7102
Te 150 153 147 160 161 163 155
2. 54 50 51 57 57 58 51
3: 50 re 49 53 52 55 5!
4. 39 39 40 36 38 43 37
5: 23 21 24 22 25 24 26
6. 54 48 50 51 51 52 51
7: 46 43 42 43 45 42 42
8. 44 41 42 43 46 44 45
Q. 0,26 0,25 0,27 0,23 0,24 0,26 0,24
No. of specimen L7095 Lo4go1 Ly7101 D. bicornis C. simum
I. 156 145 157 136 = 135 145 143
2 64 57 59 43 38 on Ahad
3 53 49 51 43 44 51 59
4 42 38 38 33. 30 40 4!
5 25 25 26 18 19 23 23
6 58 47 52 43 35 48 52
| 47 46 46 BG ta is ra:
8 48 43 44 38 34 43 45
9 0527 0,26 0,24 0,24 0,22 0,28 0,29

The fifth metacarpal of Ceratotherium praecox is reduced, mammiform, as it
is in the recent species. ‘There is one specimen in the Langebaanweg collection,
L11606, with the two facets for the unciform and Mc.IV. It is 46 mm in length,
and 35 by 29 mm in proximal diameters. In D. bicornis these diameters are
35 mm, and 27 by 26 mm;; in C. simum the bone is larger, as usual, viz., length .
45 mm, and 33 by 26 mm proximally.

Of the femora in the Langebaanweg collection there is only one that is
nearly entire, L12292, from the left side, lacking portions of the caput and of
the trochanter major, and most of the medial part of the trochlea (first column
in Table 35). In length from caput to medial condyle it exceeds the femur of
C. simum, but in diameter of the caput it is just as large as the larger of the two
C.. semum femora. There are two isolated femur heads, L12632 and L12676, with
the same diameter as L12292. The width across the third trochanter, 175 mm,
is also found in a mid-shaft portion of a left femur, L13254. There are several
juvenile shaft portions showing the third trochanter, viz., L13831, L13867—
13869, and L3409. Three distal portions of femora, L8118 and L12681 from the
right side, and L11758 from the left, complete the list of femora in the
Langebaanweg collection (Table 35).

I 78 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 35

Measurements of femur (mm)
No. of specimen L12292 D. bicornis C. simum
1. Greatest length 590 440 460 510 530
2. Diameter of caput 110 80 85 110 05
3. Width across third trochanter 175 — 140 155 —
4. Greatest distal width ¢. 175 120 125 155 150
5. Distal ant.post. diameter, medial side — 160 ~=-:165 190 190
6. Distal ant.post. diameter, lateral side 155 — 125 155 —
No. of specimen L8118 Lr12681 L11758
4. Greatest distal width 175 180 165
5. Distal ant.post. diameter, medial side 210 2Ora 6. 200
6. Distal ant.post. diameter, lateral side €300\.— 6. ey 16, AAG

Twenty-one entire patellae are in the Langebaanweg collection (Table 36),
nine from the right, and eleven from the left side. All of them are larger than
the recent bones even of C. semum.

TABLE 36

Measurements of patella (mm)
No. of specimen L14035 113725 L11589 Lo2z10 L4250 111387 L3o069
1. Length 130 115 105 110 115 L390) sy Gala
2. Width 120 110 105 105 105 120 125
No. of specimen L6226 L6o60 L4268 L7766 L7895 L7739 © L4o61
I. 130 T15 120 125 120 115 130
2s 110 110 110 110 105 105 110
No. of specimen L12833 Li13968 L7787 L4246 L5817 L5927
i 135 120 120 115 110 115
2. 125 110 115 105 105 100
No. of specimen L5926 D. bicornis C. stmum
ci 115 95 100 105 105
2; 105 85 90 go 95

Of the tibia there are no entire specimens in the Langebaanweg collection;
the most complete specimen, L1805, has only the medial portion of the proximal
articular surface, and distally the lateral portion is damaged. The length,
measured along the medial surface, is 355 mm, and the greatest length was
probably 380 mm (D. bicornis 335 mm, C. simum 350-380 mm). There are five
proximal portions of the tibia (Table 37), the first three of which are from the
right side.

TABLE 37

Proximal measurements of tibia (mm)
No. of specimen Lo7o2 Li2619g L7934 13174 L7944
1. Proximal width — 150 140 150 155
2. Proximal ant.post. diameter 115 145 120-++ ==" 65135
3. Least width of shaft 70 — — 75 80
No. of specimen . D. bicornis C. stmum
ie TAO) sata, 135 140
2. — 120 — 145

3- sees Stes

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 179

There are no less than forty-one distal portions of the tibia (Table 38) ; the
first nineteen from the right side, and the remaining twenty-two (starting with
L13477) from the left.

TABLE 38

Distal measurements of tibia (mm)

No. of specimen L7947. L7908 L6171 L7953 L1806 L4968
3. Least width of shaft 80 75 70 55 70 65
4. Distal width 105 100 95 go 95 100
5. Distal ant.post. diameter 100 go 85 85 80 go
No. of specimen L11770 L7909 L4963 14742 Log80 L3073 L6167
3- mas aa ae ae 7 ar

4. 95 100 105 100 ¢. 95 95 100
5: go 95 95 go 85 85 95
No. of specimen Lyo10 L7930 666174 L6373 L6374 Ly7948 L13477
3: a aaa F aa Pil er We:
4. go 100 go 100 105 100 100
5: 85 95 80+ go 95 =’ 95
No. of specimen L4965 113858 L7941 L7940 L4969 ) §6—L2262 =6L6165
3. 65 as 70 85 75 70 =
4. 95 95 95 100 95 95 100
5. go go 85 100 — 85 95
No. of specimen L7946 L7914 L7950 L7931 L7947 L7951 Lyg12
3: 75 ma aa an rah a oa:
As go 95 100 105 100 95 100
5. 90 go 95 100 85 85 go
No. of specimen L2264 L7926)§ 66366) = Lit529 64186. =L4187

3: mS 7g ae a Nes Te

4. 100 95 go 95 100 go

5: go 95 85 go =a =

No. of specimen L7921 D., bicornis C. simum

3. ri nt 155 a) G5

+ go O51 7695 95 115

5: a TOL 5195 8085

The astragalus is represented in the Langebaanweg collection by sixty-
seven entire specimens (Table 39), thirty-six from the right side, and thirty-one
(beginning with L4166) from the left. The astragalus is the numerically best
represented bone in the Langebaanweg collection, to which its solid build
undoubtedly contributed.

Like the other bones from Langebaanweg, the astragali are on the large
side when compared with their homologues in the living African species.
Twenty-six out of the sixty-seven Langebaanweg astragali exceed the larger of
the two C’. simum astragali in all dimensions taken. The ratio medial height /total
width varies between much wider limits in the Langebaanweg series (0,74—0,91)
than it does in the few recent bones of D. bicornis and C. szmum, as may be
expected. However, the variation range in this ratio in the Ceratotherium praecox
series does not overlap with that in the Miocene brachypotheres of Africa and
Europe (Brachypotherium heinzelin and B. brachypus: 0,64—0,73; cf. Hooijer
1966: 148). In nearly all of the Langebaanweg astragali the trochlea width is

180

No

be tah ac A

No.

Se Ree A a casagpee ee uctte an peane er

SE eee Ne aaa zy ee lite OS Nea

Cie Ga OS Uae OS) Ni eis COU Go Nee

oO.

oO.

oO.

oO.

oO.

O.

Measurements of astragalus (mm)

. of specimen

Lateral height

Medial height

Total width

Ratio medial height/total width
Trochlea width

Width of distal facets

of specimen

of specimen

of specimen

of specimen

of specimen

of specimen

of specimen

TABLE 39
L5886 L58o1
93 96
go 94
110 110
0,82 0,85
103 95
86 go
1.7222 1.7230
88 84
85 83
100 103
0,85 0,81
96 93
86 83
L7212 L7195
go 94
89 94
108 119
0,82 0,79
99 103
84 95
L11577 L11903
88 gI
92 94
116 112
0,79 0,84
95 98
94 93
15427 14874
gI 87
84 84
113 107
0,74 0,79
99 96
86 85
L12515 L4166
88 88
88 89
103 109
0,85 0,82
90 102
79 93
L7213 | L7225
86 94.
gI gi
115 120
0,79 0,76
105 105
96 95
L7488 19495
87 80
82 81
105 98
0,78 0,83
go 87
83 80

L5929
89
88

104
0,85

85
107
0,79

95
84

ANNALS OF THE SOUTH AFRICAN MUSEUM

105
0,76
94
83

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 181

No. of specimen L7210 L7203 L7206 Li1578 111583 L4164
Ms gi 88 85 87 g2 84
2. 87 gI 87 93 92 92
Be 107 100 110 120 106 103
4. 0,81 0,91 0,79 0,78 0,87 0,89
5. 92 gi 98 105 100 g2
6. 86 79 86 101 86 84.
No. of specimen L11582 L5889 L5869 L4870 Logg: 14863
ii 85 94 87 87 85 83
2. 93 97 93 go 82 84
om 114 108 108 104 106 100
4. 0,82 0,89 0,86 0,87 0,77 0,84.
5: 94 97 96 93 go 92
6. QI 88 85 83 84 86
No. of specimen L9486 D. bicornis C. simum

I 92 Goin Ue! 74 76

2 94 68 = 70 75 84

3 114 86 83 95 104

4 0,82 0,79 0,84 0,79 0,81

5 99 78 78 83. 87

6 87 73 72 85 88

greater than the medial height, although in some by a narrow margin only; in
three specimens (L7203, L4164, and L588Q) the trochlea width equals the
medial height, and in one (L12655) the trochlea width is just a little less than
the medial height. This evidently exceptional condition in C. praecox is the rule
in Aceratherium and Dicerorhinus (Hooijer 1966: 173); in Brachypotherium trochlea
width exceeds medial height, as it does also in Paradiceros (Hooijer 1968: 89) and
Chilotherrdium (Hooijer 1971: 377).

The calcaneum is represented in the Langebaanweg collection by fifty
specimens, twenty-four right and twenty-six left (Table 40). In greatest height
all of these exceed the recent bones used for comparison; in anteroposterior
diameter thirty-six fossil calcanea exceed the recent.

TABLE 40

Measurements of calcaneum (mm)

No. of specimen L11584 L11771 5867 14174 L7186 L5980
1. Greatest height 149 153 143 140 141 145
2. Greatest width 85 go a == se =
3. Ant.post. diameter —- 83 76 72 73 76
No. of specimen L5893 )S—-L5855 + L5982 L598: L4177 15851 L3052
Ls 142 144. 134 146 143 148 140
25 83 81 81 _- -- 82 =
3. 70 78 69 75 76 76 73
No. of specimen Lgo52 L488: L7180 L7169 L7184 15856 L7190
Te 140 143 146 142 149 152 146
2. ae os om re 95 97 88
3. 73 74 76 73 77 88 79
No. of specimen E7629 E7198 “VL7181 YY Lyro1 7166 6348 (Lrs8e4
i 145 157 153 150 152 146 148
2. go 95 er 94 ai =e a

oe Th 81 vig 81 80 76 76

182

No. of specimen
I.
2:
3.
No. of specimen
i.
2.
Ze
No. of specimen
i
2.
eB
No. of specimen
I.
2.

3.

ANNALS OF THE SOUTH AFRICAN MUSEUM

L5892 3536 «63149 =L7194 L6055
144 145 140 156 153
83 — 87 — —
78 76 70 84 80
L4175 15461 17175 L3790 = L8654
150 145 152 142 143
73 77 79: (Megs 80
Lai77) Lyrg7r) | tegr7e” Eggo i565
143 145 140 146 161
ss 79 — == ¢. 95
75 70 73 77 83
Lo501 L7192 L13825 D. bicornis
152 153 138 LIONS hLO
82 83 — 65 70
75 82 79 60 65

15853 L9503
151 145
84 83
79 72
L7188 17187
I4I 147
ze 85
80 77
L6054 L1802
I41 T41
85 83
72 74
C. simum
125 125
80 82
75 66

The naviculars in the Langebaanweg collection number twenty-seven
(Table 41), the first fifteen of which are from the right side.

No. of specimen

1. Anterior height

2. Total width

3. Ant.post. diameter

No. of specimen L7852
ie 32
2. 56
3. 76
No. of specimen L7757
I. 33
2. 66
3- 70
No. of specimen L12627
ii 33
2, 60
3- 73

TABLE 41

Measurements of navicular (mm)

Lg516 17775 315567 13675 19515 Lg512 Lor81
35 32 34. 37 31 31 34
56 62 58 54. 57 63 60
71 72 82 75 72 76 80

L7854 L7888 L11623 L6064 L4242A L6067 14251
31 32 30 33 33 34 33
By 61 53 60 55 59 54
72 73 69 76 69 72 74

Lo507. ~L6065 Lo510 L5241 L4242B L4257 15264
32 32 30 37 30 32 34
58 58 56 60 52 53 54
72 69 72 78 70 76 75

L7889 L7841 L6066 D. bicornis C. simum
33 33 35) 24 29
58 55 63 45 55
76 66 78 56 62

There are twenty-nine cuboids (Table 42), the first eleven of which are

from the right side.

No. of specimen

1. Anterior height

2. Anterior width

3. Greatest ant.post.
diameter

No. of specimen

I.

ay

3.

TABLE 42

Measurements of cuboid (mm)

L6221 13804 L4262 13676
53 49 49 48
53 48 54 49
77 75 76 69

L4269 L7796 Li11750 L4o069
5! 55 51 5!
49 54 50 49
76 79 73 78

Lo482 = L12823
50 54
49 58
73 86

L4260 17871
49 46
44 48
73 70

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 183

No. of specimen L9458 Lo9474 L6620 L4068 L7785 L12008 L7803
I. 48 52 52 52 46 50 53
2. 5! ras o! 52 49 54 54
3. 72 A 81 83 76 78 87
No. of specimen Lo9472 L7770 «5273 «=94L4289 )6=6sL5287 = L288 )3=—s L280
I. 49 52 55 4+ 50 55 47
2. 44 50 48 43 44 47 53
3. 77 79 78 73 71 76 78
No. of specimen L5294 OD. bicornis C. simum

I. 48 37 43

2 46 44 52

3- 73 65 80

The cuboid of Ceratotherium praecox is higher than wide anteriorly in sixteen
specimens, and wider than high in nine. We find the same variation in
Acerathertum and Dicerorhinus (Hooijer 1966: 176); it is in Brachypotherium and
Chilothertum that the width is distinctly greater than the height, and this is true
to a lesser extent in Chilotherrdium (Hooijer 1971: 380).

Eight ectocuneiforms, six right and two left (Table 43), have the anterior
width about two times the anterior height, as in the recent African species,
Aceratherium and Dicerorhinus, and Chilotheridium; in Chilotherium the width is
three times the height (Hooier 1966: 177; 1971: 380-381).

TABLE 43

Measurements of ectocuneiform (mm)
No. of specimen L4075 Lo514 Lo517 L7820 L4053 L4070
1. Anterior height 30 32 28 30 27 27
2. Anterior width 60 56 54 57 53 55
3. Ant.post. diameter il 59 57 57 51 56
No. of specimen L7749 L7773. Dz. bicorns C. simum
I. 33 33 24 27
2. 57 58 45 57
3- 58 59 53 54

One mesocuneiform, from the left side, is the remaining tarsal bone in the
collection (Table 44).

TABLE 44
Measurements of mesocuneiform (mm)
No. of specimen L12663 Dz. bicornis C. simum
1. Height 24. 14 19
2. Width 24. 24 22
3. Ant.post. diameter 45 34. 43

A set of right metatarsals, L13548-13550, belong to one and the same
individual (Pl. 34). Their measurements are given in the first columns of
Tables 45-47.

Of the second metatarsal there are fifteen entire specimens (Table 45),
the first seven of which are from the right side. The variation range in width/

184 ANNALS OF THE SOUTH AFRICAN MUSEUM

length ratio is rather small, 0,18—-0,22 only, and one of the D. bicornis meta-
podials remains below these limits, that is, it is more slender in build.

TABLE 45
Measurements of second metatarsal (mm)
No. of specimen L13550 Lea27q L4118 L4886 L13802 -
1. Median length 176 166 161 164 162
2. Proximal width 41 35 37 35 37
3. Proximal ant.post. diameter — 49 55 51 54
4. Middle width 32 32 30 33 29
5. Middle ant.post. diameter 32 30 29 26 20
6. Greatest distal width 49 45 — 44 —
7. Width distal trochlea 44 4I 39 39 —
8. Distal ant.post. diameter 52 43 44 45 43
g. Ratio middle width/length 0,18 0,19 0,19 0,20. 0,18
No. of specimen L5943 L7075 L11772 L6o052 L410g L4127 Lg380
ia 174 158 162 153 160 168 162
2 43 33 36 33 33 33 39
3 56 51 58 50 50 54 56
4 35 29 30 28 32 30 35
3) 33 27 32 24 27 30 30
6 48 43 49 41 41 44 48
7 43 38 42 39 38 38 44
8 48 43 45 42 44 48 45
0,20 0,18 0,19 0,18 0,20 0,18 0,22
No. of specimen L7097 Li1g04 L4142 D. bicornis C. simum
ie 157 153 160 126.5) 185 148 151
: 38 35 39 25. 24 ~~ SSO eae
3 56 51 55 42 33 49 47
4 32 33 32 25 22 30 28
5 30 30 31 19 20 22 24
6 46 45 44 23 eae 40 39
7 45 40 41 2024 lies =e
8 47 44 48 36 35 42 40
9 0,20 0,22 0,20 0,20 0,16 0,20 0,19

There are twenty entire third metatarsals (Table 46), ten from the right
and ten from the left side. The range of variation in width/length ratio of the
fossil bones (0,26—0,33) is very nearly the same as that in the four recent bones.

TABLE 46
Measurements of third metatarsal (mm)

No. of specimen L13548 L6048 14138 113752 113754
1. Median length 198 192 179 181 171
2. Proximal width 70 62 62 61 55
3. Proximal ant.post. diameter = 57 57 50 pci aS
4. Middle width 60 53 55 54 47
5. Middle ant.post. diameter 35 31 30 30 25
6. Greatest distal width 82 68 — 71 58
7. Width distal trochlea 69 57 54 58 54
8. Distal ant.post. diameter 55 50 50 — 49
g. Ratio middle width/length 0,30 0,28 0,31 0,30 0,27

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 185

No. of specimen L7068 L7065 L7o062 Li2615 L11855 L5960 L6043
ii Ig! 187 182 180 188 198 186
2 63 64 57 61 62 67 58
3 61 55 a 57 55 c. 60 51
4 54 52 53 52 57 62 49
5 29 29 26 31 27 33 28
6 67 67 — 68 70 80 —
7 59 56 56 56 60 66 53
8 52 51 rE 51 51 58 49
Q. 0,28 0,28 0,29 0,29 0,30 0,31 0,26
No. of specimen L7000 L5932 L4148 L13801 L13749 Lo37q-)6§6L7152
Mie 180 171 178 190 183 177 182
2 59 58 58 59 59 61 60
3 53 52 55 ar 54 51 53
4 48 53 49 55 53 54 60
5 26 29 28 29 31 31 29
6 66 64 66 ap —- 68 74
7 55 54 55 59 55 59 62
8 50 46 48 54 51 49 51
9 0,27 0,31 0,28 0,29 0,29 0,31 0,33
No. of specimen L7092 D. bicornis C. simum Aterir

I 183 148 152 160 169 180

2 59 48 50 59 55 58

3 7 43 45 47 49 50

4 50 40 40 5r 48 49

5 28 QI 19 22 25 24

6 65 54 45 56 66 68

7 57 A Nie oly 45: =

8 50 | 42)" 40 46 47 45

9 0,27 0,27 0,26 ©:92),4:0;20 127

There are sixteen entire fourth metatarsals in the Langebaanweg collection
(Table 47), eight right and eight left. The width/length ratio does not vary a

TABLE 47
Measurements of fourth metatarsal (mm)

No. of specimen L13549 13555 L4888 L7073 13785
1. Median length 170 166 153 166 155
2. Proximal width 65 54 52 53 48
3. Proximal ant.post. diameter 52 44 46 48 43
4. Middle width 33 32 33 31 27
5. Middle ant.post. diameter 48 40 38 38 32
6. Greatest distal width == 43 42 41 37
7. Width distal trochlea _ 39 39 Al 35
8. Distal ant.post. diameter 49 44 Al 45 41
g. Ratio middle width/length 0,19 0,19 0,22 0,19 0,17
No. of specimen L7114 Ly7o70 Lye63 Lari1 L4armr L4158 113748

; 162 157 156 160 163 158 174

© Or Moh ON
ee)
©
ee)
re)
ee)
N
ise)
ou
ee)
Oo
ey)
(ee)
aN
()

186 ANNALS OF THE SOUTH AFRICAN MUSEUM

No. of specimen L5942 Log90 Log241 L7099 D. bicornis C. simum
I 152 157 155 154 125 127 138 146
2 49 57 45 49 42 39 44 49
3 46 55 56 46 40 40 47 45
4 29 30 30 27 26 =. 26 35. 29
5 35 40 34 33 24 23 26 28
6 38 42 40 41 36") Si 44 39
7 34 36 36 38 33) == =e
8 39 39 42 40 Borat Al” haa
9 0,19 0,19 0,19 0,18 0,21 0,20 0,25 0,20

great deal (0,17—-0,22). One of the recent bones (the first under the head
C. simum) is not within these limits but above them; it is more massively built
than the other recent, and the fossil fourth metatarsals.

There are nineteen first phalanges of median digits (Table 48), whether
from the manus or from the pes I am unable to tell.

TABLE 48
Measurements of phalanx I, median digit (mm)

No. of specimen L3046 L8418 L6o099 L8416 L4214 Lg25o0
1. Median length 38 41 40 43 Gy) 42
2. Proximal width 66 67 67 66 64 60
No. of specimen L8417. L5276 Lo520 L8415 L5326 L8420 =©6L6216
I. 43 44 37 37 37 40 42
2. 61 64 58 61 — 60 57
No. of specimen L8419 Lo251 L5993 L13767 L5275 L7252
I. 39 ZY 40 44 39 7a
2. 65 57 61 64 57 Fi
No. of specimen D. bicornis C. stmum

manus pes manus pes
3 31 33 42 4I
2. 51 49 58 63

Four second phalanges of median digits are available (Table 40).

TABLE 49
Measurements of phalanx II, median digit (mm)
D. bicornis C. simum
No. of specimen L8426 Lo9253 Lo518 Li1607 manus pes manus pes
1. Median length 30 33 34 26 26 28 30 30
2. Proximal width 64 62 53 64 55 56 65 73

There is one third phalanx of a median digit (Table 50).

TABLE 50
Measurements of phalanx III, median digit (mm)
D. bicornis C. simum
No. of specimen L8427 manus pes manus pes

1. Median length 32 26 28 — 34.
2. Greatest width 93 84 80 — 107

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 187

Five bones (L8421, L8422, Lo256, Lo519 and L11879) represent first
phalanges of lateral digits; they vary in median length from 35 to 37 mm, and
in proximal width from 42 to 50 mm. A third phalanx of a lateral digit, L9257,
has a median length of c. 30 mm, and a greatest diameter of 51 mm. ‘Iwo
proximal sesamoids remain to be recorded; the larger bone, L7364, length
41 mm, width 21 mm, presumably belonged to a median digit, while the
smaller, L4074, length 28 mm, width 17 mm, may have belonged to a lateral
digit.

OTHER C’. PRAECOX SITES IN EAST AND SOUTH AFRICA

We have evidence of the occurrence of Ceratotherium praecox at sites other
than Kanapoi, Ekora and Lothagam-1 in Kenya, and Langebaanweg in the
Cape Province. Fragmentary teeth from the Mursi Formation of the Omo
Basin in southern Ethiopia and from the Chemeron Formation in Kenya,
previously referred to Ceratotherium simum germanoafricanum (Hooijer 1969: 86, 77),
in the light of the discovery of Ceratotherium praecox at Kanapoi and Lange-
baanweg, should be identified as C’. praecox. The teeth from the ‘lower level’
(Mursi Formation), which had been collected by R. Leakey in 1967, were
re-examined by me in July 1971 at the Centre for Prehistory and Palaeontology,
National Museum, Nairobi. There are a P* sin. and a M2?-% sin. in palatal
portions (Hooijer 1969, pl. 5, figs 4-5) displaying, as far as preserved, an
angular antero-internal corner. In M? there is a true medifossette, whereas in
P4 and M? the crochet extends across the medisinus without uniting with a
crista to form a medifossette. P* shows the internal indentation of the protocone
also seen in M?. The internal face of M? is 50 mm anteroposteriorly, and 30 mm
of this are taken up by the protocone. Although all the teeth are incomplete
externally the basal external crown outline is preserved, and the transverse
diameters can be approximately given (Table 51). They are within the limits
of their homologues in the Langebaanweg collection. Although the ectolophs of
the Mursi Formation specimens cannot be studied, in all observable characters
these teeth agree with those of Ceratotherium praecox; the medifossette is not
normally formed in this species, and its presence in the Mursi M? is exceptional.
The Chemeron maxilla with M1-3 (Hooijer 1969, pl. 2, fig. 1), from locality
J.M.507, do not have medifossettes, and M? has a distinct antero-internal crown
angle. The teeth are very much worn down, and M! and M? are so fragmentary
that the width cannot be determined, but those of M? are approximately the
same as those in the Mursi specimen (Table 51). The skull from J.M.g1,
Chemeron Formation (Hooijer 1969: 76, pl. 1) is more advanced in its dentition
and shows the rounded antero-internal crown angles, the medifossettes, and the
posterior extension of the protocone characteristic of the modern species; this
specimen moreover has the backwardly inclined occiput, extending beyond the
occipital condyles, characteristic of C. stmum germanoafricanum, and as such it was
identified in my earlier paper. The presence of both Ceratotherium praecox and
Ceratotherium simum germanoafricanum in the Chemeron Formation is puzzling,

188 ANNALS OF THE SOUTH AFRICAN MUSEUM

for the mammalian fossils in the Chemeron Formation were found so closely
together (Dr. W. W. Bishop, pers. comm.) as to make it unlikely that they were
not of the same age. The Chemeron locality J.M.go (=J.M.g1) is placed by
Cooke & Maglio (1971, fig. 2) at the 2 million year level, whereas the remainder
of the Chemeron Formation is left at the 4 million year level. This arrangement
is in accordance with the evidence provided by the rhinoceroses. Bishop (1971),
with a faunal list, gives the age of the GChemeron Formation as greater than
2,0 m.y. and less than 5,4 m.y.

A metapodial of a rhinoceros from the Chemeron Formation, locality
J.M.511, is a left second metacarpal. Whether it represents C. praecox or
C. simum I am unable to tell; the measurements have been added to Table 32
and agree with those of either of the two species.

From locality J.M.511 of the Chemeron Formation there is a P* dext. of a
large chalicothere, a new element to the Chemeron Formation fauna (cf. Bishop
19710). It was collected on 5 August 1967; I found it in the Chemeron collection
of the Department of Geology at Bedford College, London, on 18 November
1971, and it was given to me for study by Dr. W. W. Bishop. The specimen is of
considerable interest as it adds to the younger elements of the Chemeron
Formation fauna, and chalicothere teeth are rare anyway. The specimen is
referable to Ancylothertum hennigi (Dietrich), a species recorded before from
Laetolil and Bed I at Olduvai (Dietrich 1942: 105; Butler 1965: 226). It is very
well preserved and not much worn; the lingual cusp is only just touched by
wear, and the height of the worn ectoloph is 33 mm. The crown measures
28 mm anteroposteriorly and 31 mm transversely, and has all the characters of
Ancylotherium (Thenius 1953: 98 and fig. 1). The Olduvai material consists of a
few carpals, metacarpais and phalanges only, but among the Laetolil collection
there is an M? (Dietrich 1942, pl. IV, fig. 37; pl. XII, fig. 79), measuring
55,0 mm anteroposteriorly and 40,0 mm transversely. The newly found P* and
the Laetolil M?, when compared with their homologues in an upper dentition
of Ancylotherium pentelicum (Gaudry & Lartet) as figured by Thenius, prove to be
on a par for size. In the A. pentelicum dentition P* measures 33,3 by 37,5 mm,
and M? 67,2 by 50,5 mm (Thenius 1953: 105); the Chemeron P* and the
Laetolil M? are both one-sixth smaller in dimensions than the corresponding
teeth in A. pfentelicum. Laetolil and Olduvai Bed I are around the 2 million year
level (Maglio 1970; Cooke & Maglio 1971), and that is where part of the
Chemeron Formation (locality J.M.go and 91) was placed by Cooke & Maglio.
However, as stated above, in the opinion of geologist Dr. Bishop, the geological
evidence does not support a time gap of 2 million years between some Chemeron
sites (J.M.go, 91) and others. The tooth of Ancylotherium hennigi (locality
J.-M.511) as well as the skull of Ceratothertum simum germanoafricanum (locality
J.M.g1) and stage 2 or 3 of Elephas recki (Cooke & Maglio 1971) suggest an age
for the Chemeron Formation closer to 2 million years than to 4 million years.
On the other hand we have elements like the maxillary of Ceratotherium praecox
(locality J.M.507) in addition to Loxodonta adaurora Maglio, Mammuthus

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE 189

subplanifrons (Osborn), Anancus cf. kenyensis (Cooke & Maglio 1971) or Anancus sp.
(Bishop 19716), and Nyanzachoerus species ‘A’ of Cooke & Ewer, which are
suggestive of an age around 4 million years. If the Chemeron Formation fauna
is really unified as to age, it may tentatively be placed around the 3 million year
level, as suggested to me by Dr. W. W. Bishop. However this may be, further
faunal studies are needed, and the record of Ancylotherium hennigi from locality
J.M.511 of the Chemeron Formation is here given as a contribution for that end.

TABLE 51
Measurements of upper teeth of Ceratotherium praecox (mm)
Swartlintjes
Langebaan- MursiFm. Chemeron Farm,
weg Fm. Namaqua-
J-M.507 land
P4, ant.post. 47-C. 57 — — —
ant.transv. 65-76 c. 68 — —
post.transv. 56-73 c. 60 — —_
M?, ant.post. c. 50-68 --- — _
ant.transv. 70-82 ¢. 75 Ca 75 74
post.transv. 65-75 c. 65 c. 65 70
M®, ant.post. (int.) 56-72 c. 60 _— —
ant.transv. 65-78 6x67 — —
length outer surface 68-83 = — ==

The Aterir Beds in the Baringo area, Kenya, which are placed by Maglio
(1970), Cooke & Maglio (1971) and Bishop (19715) near the 4 million year
level (as is the Mursi Formation = Yellow Sands), contain material of
C. praecox. The inner portion of an upper right molar, marked 5/B4/6, has the
marked protocone fold, internal indentation of the protocone, slight internal
cingulum, a crochet but no medifossette, and the marked antero-internal crown
corner characteristic of the present species. No measurements can be given, but
in its internal anteroposterior diameter, nearly 50 mm (of which 28 mm for the
protocone), and the massive crochet it is nearest to M?. Another Aterir specimen,
an upper left premolar, marked 5/B1, again with the angular antero-internal
corner, no medifossette, and a very weak paracone style, has the dimensions of
P? in dentition L13035. Its measurements have been added to Table 5. There
is further in the Aterir collection a right third metatarsal (marked 1/18 and 1/23)
indistinguishable from its Langebaanweg homologue; its measurements have
been added to Table 46: Finally, there is a proximal sesamoid, presumably of a
median digit, marked 5/B1. This Aterir specimen is 41 mm long and 20 mm
wide, just about as large as the Langebaanweg sesamoid L7364.

An isolated, rolled M? dext., lacking the antero-external and postero-
external angles, and originating from Swartlintjes Farm, Hondeklipbaai,
Namaqualand, C.P. (about 160 km north of Langebaanweg), represents the
same species of Ceratotherium as that from Langebaanweg. According to the
geologist who presented the specimen to the South African Museum,
Mr. A. J. Carrington, the fossil molar came from ill-sorted angular felspathic
fluviatile gravels at an elevation of c. 18 m. The gravels overlie what are taken

Igo ANNALS OF THE SOUTH AFRICAN MUSEUM

as Lower Pleistocene marine sands, and would be Upper Pleistocene. However,
it is difficult to reconcile this view with the characters of the rhinoceros molar,
which are those of the Late Pliocene Ceratotherium praecox. Its rolled condition
suggests that it was derived from an earlier deposit. The specimen is figured in
Plate 25 (bottom right), and bears the South African Museum number Q1771.
The ectoloph is 77 mm high as worn. There is a well-marked protocone fold and
internal indentation of the protocone, an angular antero-internal crown
corner, and further there are a weak cingulum internally at the protocone, a
strong but relatively slender crochet, no crista, and a postsinus very nearly as
deep as the medisinus. The antero-transverse diameter is 74 mm, the postero-
transverse 70 mm, very close to those in L6658.

The Namaqualand site is the only one in the Cape Province other than
Langebaanweg from which Ceratothertum praecox is recorded, and this species is
further known only from north-western Kenya and southern Ethiopia. It is
already proving useful in African correlations, and may become more so if and
when found in other parts of Africa.

SUMMARY

Numerous remains of an extinct species of rhinoceros have been obtained
by parties of the South African Museum at the ‘E’ Quarry of the Langebaanweg
site, 104. km north-northwest of Cape Town, C.P. They are more abundant
than those of any other large mammal in the Langebaanweg fauna; there are
remains of seven skulls, ten mandibles (most of them with teeth zm sztu),
170 isolated teeth, and 650 postcranial bones. This material is referred to
Ceratotherium praecox Hooijer & Patterson described from the Late Pliocene of
Lothagam-1, Kanapoi, and Ekora in north-western Kenya. Ceratotherium
praecox is little removed from the point of divergence of the genus Ceratotherium
and the genus Diceros, and is held to represent the immediate ancestor of the
modern white rhinoceros, Ceratothertum simum. The species is further recorded in
the Cape Province from Swartlintjes Farm, Hondeklipbaai, Namaqualand
(approximately 160 km north of Langebaanweg). It is also known from the
Mursi Formation in southern Ethiopia, and the Chemeron Formation and the
Aterir Beds in the Baringo area, Kenya, all deposits dated around the 4 million
year level. The discovery of this species is proving most useful in inter-African
correlation and adds to the evidence already available that the ‘E’ Quarry
Langebaanweg site is Late Pliocene in age.

ACKNOWLEDGEMENTS

It is a great pleasure to thank Dr. T. H. Barry, Director, Mr. and Mrs. Q. B.
Hendey, and Mrs. D. Hirschon, Palaeomammalogy Department, South
African Museum, for facilitating my work at the museum in May and June
1971, and for courtesies extended. I am indebted to Dr. W. W. Bishop for
permission to include C. praecox remains from other East African sites, to
Mr. R. E. F. Leakey who let me study Mursi Formation, Ethiopia, and

A LATE PLIOCENE RHINOCEROS FROM LANGEBAANWEG, CAPE PROVINCE IQI

East Rudolf, Kenya, rhinoceroses, and to Mr. Neville Eden who took the
photographs. My journey to South Africa has been made possible by a grant
from the Wenner-Gren Foundation for Anthropological Research, Inc.,
New York.

REFERENCES

BisHop, W. W. 19714. The Late Cenozoic history of East Africa in relation to hominoid
evolution. Jn TurEKIAN, K. K., ed. Late Cenozoic glacial ages: 493-527. Cambridge (Mass.):
Yale University Press.

Bisoop, W. W. 19710. Stratigraphic succession ‘versus’ calibration in East Africa. In
BISHOP, W. W. & MILLER, J. A., eds. Calibration of hominoid evolution. Edinburgh: Scottish
Academic Press.

But er, P. M. 1965. East African Miocene and Pleistocene chalicotheres. Bull. Br. Mus. nat. Hist.
(Geol.) 10: 163-237.

Cooke, H. B.S. & Mac tio, V. J. 1971. Plio-Pleistocene stratigraphy in East Africa in relation to
proboscidean and suid evolution. Jn BIsHop, w. w. & MILLER, J. A., eds. Calibration of
hominoid evolution. Edinburgh: Scottish Academic Press.

Dietricu, W. O. 1942. Altestquartare Saugetiere aus der siidlichen Serengeti; Deutsch-
Ostafrika. Palaeontographica 94(A): 43-133.

DietricH, W. O. 1945. Nashornreste aus dem Quartar Deutsch-Ostafrikas. Palaeontographica
96(A): 46-90.

HELLER, E. 1913. The White Rhinoceros. Smithson. misc. Collns 61: 1-77.

HENDEY, @.B. 1969. Quaternary vertebrate fossil sites in the south-western Cape Province.
S. Afr. archaeol. Bull. 24: 96-105.

HENDEY, © .B. 1970a. A review of the geology and palaeontology of the Plio/Pleistocene deposits
at Langebaanweg, Cape Province. With an Appendix: The Langebaanweg Bovidae by
A. W. Gentry. Ann. S. Afr. Mus. 56: 75-117.

HENDEY, Q.B. 1970). The age of the fossiliferous deposits at Langebaanweg, Cape Province.
Ann. S. Afr. Mus. 56: 119-131.

HEnpDEY, Q.B. & HENDEy, H. 1968. New Quaternary fossil sites near Swartklip, Cape Province.
Ann. S. Afr. Mus. 52: 43-73.

Hooyer, D. A. 1959. Fossil rhinoceroses from the Limeworks Cave, Makapansgat. Palaeont. afr.

6: I-13.

Hoorer, D. A. 1966. Miocene rhinoceroses of East Africa. Bull. Br. Mus. nat. Hist. (Geol.)
13: 117-190.

Hooyer, D. A. 1968. A rhinoceros from the Late Miocene of Fort Ternan, Kenya. Zool. Meded.,
Leiden 43: 77-92.

Hooyer, D. A. 1969. Pleistocene East African rhinoceroses. Fossil Vertebr. Afr. 1: 71-98.

Hooyer, D. A. 1971. A new rhinoceros from the Late Miocene of Loperot, Turkana District,
Kenya. Bull. Mus. comp. Zool. Harv. 142: 339-392.

Hooyer, D. A. & PATTERSON, B. 1972. Rhinoceroses from the Pliocene of north-western Kenya.
Bull. Mus. comp. Zool. Harv. 144: 1-26.

Hooyer, D. A. & SincER, R. 1960. Fossil rhinoceroses from Hopefield, South Africa. ool.
Meded., Leiden 37: 113-128. :

Mactio, V. J. 1970. Early Elephantidae of Africa and a tentative correlation of African Plio-
Pleistocene deposits. Nature, Lond. 225: 328-332.

Mactio, V. J. & HENDEy, Q.B. 1970. New evidence relating to the supposed stegolophodont
ancestry of the Elephantidae. S. Afr. archaeol. Bull. 25: 85-87.

THENtIus, E. 1953. Studien tiber fossile Vertebraten Griechenlands. III. Das Maxillargebiss von
Ancylotherium pentelicum Gaudry und Lartet. Annls géol. Pays hell. 5: 97-106.

Thenius, E. 1955. Zur Kenntniss der unterpliozdnen Diceros-Arten (Mammalia, Rhino-
cerotidae). Annin naturh. Mus. Wien 60: 202-211.

EXPLANATION OF THE PLATES

Note. All specimens are Ceratotherium praecox Hooijer & Patterson from Langebaanweg, except
Plate 25, bottom right, which is from Swartlintjes Farm, Namaqualand.

‘9
we?
*
|
(f
¥
"
Om

yn

4

TE

ew

ny

Ann. S. Afr. Mus., Vol. 59 Plate 21

Upper dentition, L13035, crown view, X 0,44.

Ann. S. Afr. Mus., Vol. 59 Plate 22

Upper dentition, L2519, crown view, X 0,52.

Ann. S. Afr. Mus., Vol. 59 Plate 23

Upper dentition, L13747, crown view, X 0,35.

Ann. S. Afr. Mus., Vol. 59 Plate 24

Skull, L6658, palatal view, x 0,37.

Ann. S. Afr. Mus., Vol. 59 Plate 25

Top, P?~* dext., L13035, internal view, x 0,78.

Middle, outer surfaces of M® dext., L6291 and L6606, external views, x 0,57.

Bottom left, ectoloph of P* sin., L13760, external view, x 0,56.

Bottom right, M? dext., Swartlintjes Farm, Namaqualand, S.A.M. Q1771, crown view, x 0,82.

Ann. S. Afr. Mus., Vol. 59 Plate 26

Top of skull, L13747, right lateral view, x 0,67.

Ann. S. Afr. Mus., Vol. 59 Plate 27

Left, top of skull, L2520, dorsal view, x 0,33.
Right, skull, L6658, dorsal view, x 0,35.

*oS°O & “SMOTA
UMOID “QzQQT “IXEP AT pur ‘Lr1ggT “1x9p zW ‘g16S7 “UIS ,IA “JYSII 0} yo] wo’ ‘mor WIOWOg
"0o'r x “Mota [esaqe] “LPLE ry y[nys ym “J s9ddn payejost qysta doy,
“GSO x ‘mara [eqeyed ‘LPL E17 [mys Jo sorerprxeutoad “yo doy,

Plate 28

Ann. S. Afr. Mus., Vol. 59

*L9‘0 X ‘SMOTA UMOIO

‘91167 “us ,Jy JO ydojo}9o pure ‘SSgoq “urs ,g ‘sS111'T “urs yg “VYst1 07 YJ9T WOIZ ‘MOI WOW0g
*@L°O X ‘SMOTA

uMOID ‘1SgQT “UTS ,JNC pue ‘gogrT “urs ZING ‘qSor6T] “urs Aq ystr 07 YoT wos ‘Mos doy,

Plate 29

Ann. S. Afr. Mus., Vol. 59

Ann. 8. Afr. Mus., Vol. 59 Plate 30

Mandible, L13035, top view, x 0,28.

Ann. S. Afr. Mus., Vol. 59 Plate 31

Right mandibular ramus, L13035, internal view, x 0,27.

Ann. S. Afr. Mus., Vol. 59 Plate 32

Top left, DM, dext., Lg105C, crown view, X 0,99.
Bottom left, DM, dext. in ramus fragment, L6660, crown view, X 0,73.
Right, mandible, L11849, top view, x 0,36.

Ann. S. Afr. Mus., Vol. 59 Plate 33

Top, symphysis of the mandible, L6058, top view, x 0,67.
Middle, ankylosed cuneiform and pisiform sin., L7823, anterior view, Xx 0,61.
Bottom, cuneiform sin., L9465, and pisiform sin., L7892, as they articulate, anterior views, X 0,61.

Ann. S. Afr. Mus., Vol. 59 Plate 34

\

OES

Metatarsals II, III and IV dext., L13548-13550, articulated, front view, x 0,61.

INSTRUCTIONS TO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS
To be reducible to 12cm xX 18cm (19cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.
For journal articles give title of article, title of journal in italics (abbreviated according to the

World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,

volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

BULLOUGH, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHErR, P.-H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Konn, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region
of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In scHULTZE. L,
Koologische und anthropologische Ergebnisse einer Forschungsreise 1m westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med.-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51).
The Harvard system of reference to be used in the synonymy lists, with the full references
incorporated in the list at the end of the article, and not given in contracted form in the synonymy
list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, b; Liste: 11. Turton, 1932: 80.

rs

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
September 1972 September
Part 10 Deel

A REVIEW OF THE SEPIIDAE (CEPHALOPODA)
OF SOUTHERN AFRICA

By
MARTINA A. ROELEVELD

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, 2a. 3, 5s 7-8); 3 (1-2, 5» US) 3 51-2, 5, 7-9),
6(1, t.—p.i.), 7(1-3), 8, g(1-2), 10(1),
(i 2, 5» 7s t-piela)s 24(2), 27; Z1(1—3)5 33

Price of this part/Prys van hierdie deel
R10,80

Trustees of the South African Museum © Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 13 5

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

A REVIEW OF THE SEPIIDAE (CEPHALOPODA)
OF SOUTHERN AFRICA

By

Martina A. ROELEVELD
South African Museum, Cape Town

(With plates 35-45, 20 figures and 53 tables)
[MS. accepted 1 March 1972]

CONTENTS
Introduction . d ‘ ‘ F ; ; : . 194
Methods . : E : , , , . 194
Abbreviations and Plast : ; A ; + 196
Key to the Sepiidae of southern Attics : ; ‘ 2 REND
Key to shells only . - : : 4 : é * Teg
Systematic discussion
Diagnoses of family and genera . i : ; sv 200
Description of southern African species
Sepia zanzibarica Pfeffer, 1884 . : - 202
Sepia officinalis vermiculata Quoy & (rai art 1832 Peer)
Sepia acuminata Smith, 1916 ; , ‘ . 208
Sepia confusa Smith, 1916 : : : : | 2to
Sepia incerta Smith, 1916 5 } : : : hong
Sepia burnupt Hoyle, 1904 : : : : st) 2ET
Sepia joubint Massy, 1927 : : : ‘ PN 2-1
Sepia adami n. sp. : ; yee
Sepia australis Ghee & ane ee : : Jy 2a
Sepia tuberculata Lamarck, 1798 : : ‘ - 231
Sepia papillata Quoy & Gaimard, 1832. : Lhiagr
Sepia simoniana Thiele, 1920 . : : ‘ . 240
Sepia angulata n. sp. : : : : a eae
Sepia hieronis (Robson, mater 3 : ‘ i 245
Sepia insignis Smith, 1916 : : : : “Fy 2Ao
Sepia robsoni (Massy, 1927) : : : «. 250
Sepia faurei n. sp. . . : | 25%
Sepia (Hemisepius) typica peel ee oh : <i) 257
Sepia (Hemisepius) dubia Adam & Rees, 1966. . 264
Sepiella cyanea Robson, 1924 . : . : :. 206
Discussion
Relationships . : : : : : . 268
Geographical diemsbaietoa f : . ; JN 208
Vertical distribution ¢ ‘ ‘ : ‘ . 276
Growth . ‘ : : NONE i: ‘ ‘ sy LOD
Summary. : i : é : : . . 280
Acknowledgements - ‘ ; . 2 : e 2ae
References . : : : : : : MPO
Appendix: Tables — : F é : ; ’ , 1285
193

_ Ann. S. Afr. Mus. 59 (10), 1972: 193-313, 11 pls, 20 figs, 53 tables

194 ANNALS OF THE SOUTH AFRICAN MUSEUM

INTRODUCTION

The aim of this paper is to collate and extend, where possible, the know-
ledge of the Sepiidae of southern Africa (defined by Day 1967: vu, as Africa
south of the twentieth parallel of latitude). Hitherto, the sepiids have been
described in various papers dealing with South African cephalopods (e.g.
Robson 1924a, b; Massy 1925, 1927, 1928; Voss 1962b, 1967) and in some
detail by Adam & Rees (1966) in their excellent review of the family. The
collection in the South African Museum, however, comprises some 664 animals
and 411 shells, belonging to 18 species, and these specimens provide a number
of interesting additions to our knowledge of the southern African Sepiidae.
Three new species and the female animal of Sepza insignis (previously known
only by its shell) are described.

All but two of the species have been redescribed from the specimens
available and compared with previous descriptions. Unfortunately the collec-
tion is poor in species occurring off the Natal coast (called here the ‘dorato-
sepion’ group). This gap has been filled to some extent by the donation of a few
Sepiidae caught off the Mocambique coast by the Oceanographic Research
Institute, Durban, and by the loan of some Sepiidae from the Natal Museum,
Pietermaritzburg. Most of the latter specimens have previously been described
by Miss A. L. Massy (1925, 1928). No specimens were available for Sepia
robsoni and S. (Hemisepius) dubia. Only one specimen, the type, is known of each
of these two species; these types are deposited in the British Museum (Natural
History), and are not sent out on loan.

Before his death Dr. K. H. Barnard compiled notes on the South African
Sepiidae and constructed a rough key to the species. But since several specimens
have been added to the collection in the interim, and the notes include a
number of errors, the present work was not based on his notes, but was started
afresh. Some points of interest found in Barnard’s notes are discussed under the
relevant species.

This paper was submitted in partial fulfilment of the requirements of the
Degree of Master of Science in Zoology at the University of Stellenbosch.

METHODS

All animals in reasonable condition were measured with dividers and
millimetre rule. Each dimension was calculated as a percentage of dorsal
mantle length for animals and of shell length for shells. For animals of mantle
length less than 25 mm, only mantle dimensions were recorded, as the animals
are too small to handle without damaging them. Measurements of southern
African Sepiidae previously published by other authors were included where
possible to increase the numbers to a significant level. Tables of all relative
dimensions are given at the end of the paper. For some of the more significant
dimensions (mantle width, head width, fin width, length of tentacular club and
shell dimensions) ranges and arithmetic means were calculated. Calculations

ae

ee ts 8 ge wee. A ae

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA — 195

for male and female animals of each species were carried out separately, but
shell dimensions of both sexes were combined, since in the case of a shell found
on the beach, the sex of the animal by which it was secreted cannot be
determined (except perhaps in the case of Sepzella cyanea).

The dimensions of preserved specimens are affected by a number of factors.
The relative mantle width is generally greater in dead animals than in live ones,
as the mantle collapses and flattens out. However, since all animals measured
were dead and fixed in formalin, one may assume the error to be fairly constant.
Relative arm lengths are not considered to be of much significance, since they
vary considerably, depending on the extent of contraction. However their
ranges and means were calculated for comparative purposes, where only large
differences may be considered significant. ‘The length of the tentacles is similarly
affected, but more so, as the tentacles seem to be more contractile than the
arms. In addition, some animals were preserved with the tentacles retracted,
and these tentacles were sometimes difficult to straighten out for measuring.
Similarly the presence or absence of keels on the arms frequently appears to
depend on the condition of the animal at the time of preservation.

In most cases coloration is an unreliable guide within the Sepiidae, as these
animals are masters of the art of camouflage, and can show a wide range of
colour patterns. Generally the background is a pale cream, overlaid by brown-
black and orange-red chromatophores. By progressive expansion of these
chromatophores, the animal can assume a range of colours from black, through
purple to reddish-brown, over all or part of the body. With complete contraction
of the chromatophores only the pale background is seen.

Five of the species described below, namely Sepia confusa, S. incerta,
S. burnupi, S. joubint and S. adami, have been called the ‘doratosepion’ group for
convenience, since these species seem to be closely related and are frequently
discussed as a group or compared with each other in the text. Most of these
species have been included at some time in Rochebrune’s genus Doratosepion.
This genus was created for Sepiidae with an elongated body, short arms with
biserial suckers, short tentacular clubs with unequal suckers, a very elongated
shell with two posterior wings and a spine. Whilst Rochebrune’s classification
of the Sepiidae has been rejected by Adam (1944) and Adam & Rees (1966)
with good reason, ‘doratosepion’ is used here as a collective name for the
abovementioned five species. Its use, however, in no way implies the retention
of the genus Doratosepion.

Distribution ranges have been constructed from localities of specimens in
the South African Museum collection and from published records, and have in
some cases been extended by locality records from the University of Cape Town
Ecological Survey. In the distribution figure (Fig. 18) only localities of animals
are recorded. Shell localities cannot be considered to extend the distribution
range of a species, since sepiid shells are known to drift over long distances, and
their place of origin is thus unknown.

Finally, it should be noted that the Natal Museum has listed the shells of

196 ANNALS OF THE SOUTH AFRICAN MUSEUM

animals in the collection under separate numbers, e.g.:

N.M.956: S. incerta, 13 from Natal coast, in stomach of Ground Shark.

N.M.957: 5S. incerta, 15 from the same locality.

N.M.g958: shells of the above two specimens (and one shell of S. burnupz,
presumably added later).

ABBREVIATIONS AND GLOSSARY

MlLd —dorsal mantle length along midline

MLv —ventral mantle length along midline

MW —maximum mantle width (excluding fins)

HL —head length dorsally (from anterior tip of nuchal cartilage to edge
of dorsal interbrachial membrane)

HW —maximum head width (usually across the eyes)

BE —length of single fin along curve of mantle

FW —width of single fin, from lateral edge of mantle to free edge of fin

AL I-IV—arm length measured from inner base of most proximal sucker to
tip of arm

ni —length of tentacle, from point of emergence from tentacular sac to
tip of club

ict —length of tentacular club, from basal sucker to tip of club

Shell:

L —total length of shell, excluding posterior spine (where present)

W —maximum width of shell

‘Eh —maximum thickness of shell, along midline

Str z —length of striated zone

km —kilometres

m —metres

mm —millimetres

N —number of specimens measured

N.M. —Natal Museum

O.R.I. —Oceanographic Research Institute

Pr. — Cape Fisheries survey vessel Pieter Faure

S.A.M. —South African Museum
U.C.T. —University of Cape Town Ecological Survey

acuminate —forming an acute angle
arm length formula —comparative lengths of the arms in decreasing order
attenuated —suddenly becomes very slender distally (usually referring

to arm tips)

‘doratosepion’ group—includes Sepia confusa, S. incerta, S. burnupi, S. joubini and
S. adami. See page 195 for definition.

emarginate _ —with a broad semicircular or rectangular notch (usually
referring to anterior mantle margin ventrally)

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 197

KEY TO THE SEPIIDAE OF SOUTHERN AFRICA

Figures 1 and 2 illustrate the external morphology of Sepia and its shell,

and most of the terms used in the keys. Sepza angulata is not included in the first
key as this species is known only by its shell.

I

Ne)

ee)

10

II

12

13

La

Posterior gland present, opening via a pore situated between the posterior extremities of

the fins (genus Sepziella) : y : : : ; , Sepiella cyanea
Posterior gland absent (genus Sepia) : : : g : : : ; roe
Tentacular club with numerous subequal suckers (Figs 14c, 17b) _.. F ven a
Tentacular club with a few median suckers enlarged (Fig. 11) “ : 7 iene
Mantle produced dorsally . : : ; 3 : ; ; ; , ae
Mantle convex dorsally : 5 : 4 } : ‘ ; , - ms
Mantle ventrally entire : : ; , ; ; : ; ; ; p45
Mantle ventrally emarginate : ; ; ; : ’ ; : . oS. insignis
Dorsal arms with biserial suckers ; j ; : : : ‘ . 9. hieronis
Dorsal arms with quadriserial suckers . J : , : P : J ie iG
Buccal membrane with a few small suckers . ; : : : . §. zanzibarica
Buccal membrane without suckers : : ; : : 3 S. acuminata
Mantle ventrally entire : ‘ : : ‘ : f ‘ , iS. stmoniana
Mantle ventrally emarginate : : : , , Pei:
Tips of dorsal arms finger-like, devoid of suckers ae teh : ; : ; See:
Tips of dorsal arms normal, with suckers to the tips : ; : : : ee 2
Few or no papillae dorsally along outline of shell and on head ; ; . §. robsoni
Densely papillose dorsally (Fig. 15a) . : F ‘ ‘ : . . §. faurer
About 12 pairs of pores in the ventral mantle surface (Fig. 17b) . §. (Hemisepius) typica
No pores in the ventral mantle surface : é F ; . §. (Hemisepius) dubia
Skin tuberculate dorsally; ventrally with two a wrinkled patches on the mantle
(Fig. 12) . : : P j : ee:
Skin smooth dlongal ln no ene veninicled eens F ; : : ; i ag
Diameter of large tentacular suckers approximately ee to width of club
(Fig. 11b) ; ; S. papillata
Diameter of lanes Bueealee eee less than width so Efe (Fig. 11a) . d S. tuberculata
Shell ovate, width 32-46% length; inner cone well developed, reflexed and completely
fused to outer cone (Pls 35c, d, 36a, b) ‘ A : . S. officinalis vermiculata
Shell elongate; inner cone wea uivecl with narrow Panis ‘ ; , ‘ : ape!
Shell broadly elongate, width 29-37% length; no posterior wings on outer cone

(Pl. 39a, b). ‘Light organ’ in mantle cavity . : s : S. australis
Shell narrow elongate, width 14-26% length; outer cone aes eerion wings (Fig. 2).

No ‘light organ’ in mantle cavity. (“doratosepion’ group) ; ‘ : ‘ . 15
Males (male of S. adami not known) 16
Females. The females of the ‘doratosepion’ seccite are difficult to squannis. ned tas eee

of the key is very tentative . ; f : : F F : : : «- FO
Dorsal arms normal . : : f : : ‘ : : em ia |
Dorsal arms modified (Figs Ge. 8e) . ‘ : : < : : : = 1G
Fins rounded posteriorly. : : ; é : . S. joubini
Fins extended posteriorly to form ‘tail’ (He 4) : ‘ ‘ ‘ : . S. confusa
Ventral arms with hectocotylized region and distal cirri (Fig. 8c, d); fins extended into

points posteriorly 4 : . 8. burnupi
Ventral arms without ectacerrlined region ¢ or Cirri; fae rounded aseriody . oS. incerta

198 ANNALS OF THE SOUTH AFRICAN MUSEUM

VENTRAL DORSAL

dorsal arm
dorsolateral arm

ventrolateral arm

ventral arm
outer lip

:

position of head

retracted tentacle } SN eee produced
Fee et FID . eye dorsally

protective
membrane

position of shell

fin transverse
mantle rows
posterior point
of mantle C
a
arm |
arm Il
arm Ill
armIV

tentacular club

funnel opening
funnel

funnel cartilage

mantle cartilage

funnel retractor light organ”

nidamental gland
ink sac

position of stomach
ovary

Fic. 1. General morphology of the Sepiidae, illustrated by Sepia australis. a. External features;

b. the mantle cavity (the ‘light organ’ is characteristic of S. australis and does not occur in other

Sepiidae); c. dorsal arm, to show the quadriserial arrangement of the suckers. Longitudinal

series numbered 1-4. The suckers of series 1 and 4 are obscured distally by the protective
membranes.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 199

19 Lateral arms attenuated distally . ; ; , ; ; ; , - a/ 20
Lateral arms not attenuated distally . : : ‘ ’ ‘ . S. adami
20 Lateral arms attenuated over their distal half; distal suckers biserial ; ‘ pm iY

Lateral arms attenuated over less than half the arm length; distal suckers quadriserial . 22

21 Protective membranes on distal part of dorsal arms expanded . ‘ ; . 9. joubini
Protective membranes on distal part of dorsal arms not expanded . : . S. confusa

22 Shell with inner cone raised posteriorly; striated zone convex; striae convex
(Fig. 6d) . : : : : ; : : : , ; ; . oS. incerta
Shell with inner cone low posteriorly; striated zone //\-shaped; striae angular
(Fig. 8a, Pl. 4d) , : : ; : ; f ‘ ; : . S. burnupi

smooth
zone

or
last
loculus

phragmocone

striated
zone

median groove

limb of
inner cone

wing of
outer cone

ye

inner cone
spine

Fic. 2. Ventral view of the shell of Sepia joubini (A30141)
showing some of the features mentioned in the descriptions.
Length 37 mm.

KEry TO SHELLS ONLY

Sepia robsoni is not included in this key, as its shell is insufficiently known.

1 Posterior spine present : . ; : : : : ‘ , : ae
Posterior spine absent 9
2 Ventral part of inner cone well developed. ; . ; : : ; "3
Ventral part of inner cone reduced 4

200 ANNALS OF THE SOUTH AFRICAN MUSEUM

3 Ventral part of inner cone not reflexed (Pl. 35b) . i . S. zanzibarica
Ventral part of inner cone reflexed and completely fused ‘a outer cone 35d, 36b)
S. officinalis vermiculata

4 Outer cone with posterior wings (Fig. 2) : : : : ; " : 2.15
Outer cone without posterior wings . é ; 2 3 : : : ee
5 Inner cone raised posteriorly, forming a deep pocket over the end of the phragmocone
(Pls 37d, 39b) . : ; : , : : ‘ ; ae
Inner cone low pesisceels (PI. 38d) : : 4 5 ‘ é : : He, wy
6 Posterior part of inner cone with a deep longitudinal groove (Fig. 6d) ‘ . S. incerta
Posterior part of inner cone rounded, without longitudinal groove (Fig. 5) . S. confusa
7 Limbs of inner cone raised, lying on the phragmocone; striae /\-shaped (Pl. 38d,
Fig. 8a) . ‘ . 8. burnupi
Limbs of inner cone non aed lying at fos of sihsawine 23 striae convex
(Figs 2, 10a). : : : ; : . *S. joubini, S. adami
8 Shell rhomboidal, posterior spine not keeled eb 37a, b) 4 : : S. acuminata
Shell elongate oval, posterior spine keeled (Pl. 39a, b) . : ‘ : S. australis
9 Phragmocone considerably shorter than dorsal shield 3 : : : - 20
Phragmocone almost as long as dorsal shield : : ; ; : ‘ 1 MES
10 Anterior margin of phragmocone transverse, not oe with corresponding margin of
dorsal lamella (Fig. 1 7G) : 2 II
Anterior margin of phragmocone convex, more or ies sense! “ih isis ce
margin of dorsal lamella (Fig. 16c) : f ‘ 3 ‘ : : . 8. faurer
11 Shell dorsally chitinous, inner cone indistinct (Fig. 17c, d) : . S. (Hemisepius) typica
Shell dorsally calcified, inner cone distinct . ' ‘ , . §. (Hemisepius) dubia
12 Outer cone expanded posteriorly, inner cone reduced (Pl. 42d) (genus Sepiella) Sepiella cyanea
Outer cone narrow posteriorly, inner cone well developed and completely reflexed . 13
13 Shell rounded anteriorly . : : : : A : s ; : Gb
Shell acuminate anteriorly . 3 4 : : : ‘ ; F : yea
14 Ventral surface flat or concave (Pls 39d, ae : ‘ : : : S. tuberculata
Ventral surface convex i i : é : : : : : ae x
15 Limbs of inner cone fairly broad, narrowing etetually anteriorly . 16

Limbs of inner cone broad posteriorly, narrowing suddenly anteriorly (Pl. 4b) S. simoniana

16 Shell very broad (width 50-60% length) (Pl. 45); Sa between striated zone and

smooth zone pronounced in lateral view (Pl. 10d) . S. angulata
Shell elongate oval (width 37-55 /o length) (Pl. 41); no pronounced baci between
striated zone and smooth zone in lateral view : : 5 S. papillata
17 Limbs of inner cone lie at sides of phragmocone (PI. ie) : ; : . S. Ateronis
Limbs of inner cone raised, lying on phragmocone : : 4 5 . S. insignis

SYSTEMATIC DiscussION
DIAGNOSES OF FAMILY AND GENERA
Family Sepiidae
Mantle short, oval or rounded, dorso-ventrally flattened; fins lateral,

occupying almost entire lateral margin of mantle; eye covered by a continuous
membranous lid; arm suckers usually quadriserial, occasionally biserial; left

* On the available specimens no constant differences could be found between the shells
of these two species.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 20!

ventral arm of male usually hectocotylized; tentacles completely retractable,
tentacular club distinct from stalk, with subequal or unequal suckers. Shell
internal, usually calcareous; phragmocone retained; conotheca reduced
ventrally, represented by inner cone; posterior spine present or absent.

Genus SEPIA Linnaeus, 1758
Diagnosis as for family. Posterior gland absent; tentacular club with
subequal or unequal suckers; locking apparatus simple, oval (Fig. 3a). Outer
cone of shell not expanded posteriorly.

—— ==,

Fic. 3. Comparison of the mantle locking apparatus of
a. Sepia (S. australis, 9, A3go0154) and b. Sepiella (S. cyanea,
( 36, A6526). Left, mantle component; right, funnel component.
Scale = 1 mm.

202 ANNALS OF THE SOUTH AFRICAN MUSEUM

Subgenus Sepia s.s.

Mantle slender to moderately broad; arm suckers biserial or quadriserial;
tentacular suckers subequal or unequal. Shell slender to broadly oval, length
approximately equal to dorsal mantle length; phragmocone occupies almost all
dorsal shield; posterior spine present or absent.

Subgenus Hemisepius Steenstrup, 1875

Mantle very broad; arm suckers biserial; tentacular suckers subequal. Shell
very thin, phragmocone shorter than dorsal shield; posterior spine absent.

Subgenus Metasepia Hoyle, 1885

Mantle broadly oval; arm suckers quadriserial; tentacular club with few
unequal suckers. Shell rhomboidal, much shorter than mantle; posterior spine
absent. (No southern African representatives. )

Genus SEPIELLA Gray, 1849

Posterior gland present; tentacular club with numerous subequal suckers;
locking apparatus with tubercle on mantle component and corresponding
depression in funnel component (Fig. 3b). Shell without posterior spine; outer
cone expanded posteriorly.

DESCRIPTION OF SOUTHERN AFRICAN SPECIES
Sepia zanzibarica Pfeffer, 1884
(Pl. 35a, b. ‘Tables 9, 10)

Sepia zanzibarica Pfeffer, 1884: 9, fig. 11, 11a. Hoyle, 1886: 22, 217. Smith, 1916: 21. Tomlin,
1923: 40. Massy, 1925: 211. Voss, 19626: 248. Adam & Rees, 1966: 7, pl. 2, figs 9-11, pl. 41,
fig. 247.

Type locality
Zanzibar.

Distribution

Animals: Zanzibar (Pfeffer 1884: 9), German East Africa (Massy 1925: 211).
Depth not known.

Shells: | Mombasa (Adam & Rees 1966: 7) to Tongaat, Natal (Smith 1916:
21); also Malagasy (Adam & Rees 1966: 7).

Material

N.M. 959, 960, German East Africa (det. A. L. Massy); 1 3

S.A.M. A2141, Chinde, mouth of Zambezi River; 2 shells

Locality unknown; 1 shell

Description
Male (N.M.959) originally described by Massy (1925: 211), now in rather
poor condition.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 203

Mantle broadly oval, anterior mantle margin produced dorsally, ventrally
entire. Fins narrow, rounded, separate posteriorly. “The buccal membrane has
a single sucker on five of its tips’ (Adam & Rees 1966: 8). These not clearly
visible in male, due to distorted state of buccal membrane.

Skin smooth, except for few tiny papillae mid-dorsally on mantle. Colour
dark purple dorsally, with paler fins. Narrow dark purple line along fin bases.
Colour ventrally lighter purple laterally on mantle, fading to mottled buff-
purple mid-ventrally.

Arms unequal in length; shortest dorsally, longest ventrally, of formula
4.3.2.1. In female, arms I to III subequal, arms IV considerably longer
(Pfeffer 1884: 9). Arms joined by shallow interbrachial web, except between
ventral pair. All arms keeled, provided with well-developed protective mem-
branes folding over inner surface; arm tips attenuated.

Suckers on all arms quadriserially arranged to tips. Chitinous rings of
suckers smooth.

Left ventral arm hectocotylized. Basal two-thirds of arm bearing quadri-
serial suckers; suckers minute over six rows on distal third (as described by
Massy 1925: 211). Adam & Rees (1966: 8) were mistaken in presuming that
the hectocotylus was situated on the proximal third of the arm. Three dorsal
series of minute suckers in normal position, ventral series displaced, leaving
naked, grooved region on arm; tip of arm bearing normal quadriserial suckers.

Tentacular club long, bearing numerous subequal suckers in rows of
about six, ‘but probably form oblique transverse rows of eight’ (Adam & Rees
1966: 8). Of these, three median series of suckers a little larger than others.
In addition, three larger distal suckers partially concealed by reflexed tip of
club. Chitinous rings of tentacular suckers finely dentate. Protective mem-
branes well developed, remaining separate proximally and continuing along
tentacular stalk for some distance. Natatory membrane a little longer than
club.

Shell (N.M.960) of male specimen badly damaged. Three other shells
(Pl. 35a, b) in fairly good condition, but somewhat worn.

Shell broadly oval, tapering anteriorly and posteriorly. Dorsal surface
roughly granular posteriorly, more finely so anteriorly. Median ridge ill-defined.
Two fairly well defined dorsal grooves diverging from posterior end correspond
with position of ventral limbs of inner cone. Posterior spine short but strong,
directed dorsally and coloured blue; spine not keeled. Striated zone long
ventrally, occupying two-thirds to three-quarters shell length. Striae broady
/\-shaped, becoming somewhat more rounded anteriorly. Median groove broad,
with phragmocone raised on either side. Inner cone very well developed
posteriorly, free, curving over posterior part of striated zone to form a pocket.
Limbs of inner cone broad, curving over lateral edges of phragmocone. Outer
cone of present shells damaged, but according to Pfeffer’s figure (fig. 11a) it
broadens somewhat posteriorly without actually forming wings. Shell thickest
on either side of midline, near anterior end of striated zone.

204 ANNALS OF THE SOUTH AFRICAN MUSEUM

Remarks

It is not certain if this may be considered a southern African species, since
only shells have been found on our coasts.*

Tomlin (1923: 40) mentions a specimen, presumably a shell, from
Isipingo, but gives no details.

Sepia officinalis vermiculata Quoy & Gaimard, 1832
(Pls 35c, d, 36a, b. ‘Tables 1, 2, 11-13)

Sepia vermiculata Quoy & Gaimard, 1832: 64, pl. I, figs 1-5. Férussac & d’Orbigny, 1835-1848:
2709, pl. I11dis. Gibbons, 1888: 202. Bartsch, 1915: 250. Smith, 1916: 20. Robson, 19242: 12.
Massy, 1925: 209; 1928: QI.

Acanthosepion vermiculatum Rochebrune, 1884: 113. Adam, 1944: 234.

Acanthosepion vermiculata: Robson, 19246: 639. Massy, 1927: 156.

Sepia officinalis vermiculata Adam, 1940: 130; 1941: 99, 102, 106, pl. IV, fig. 1; 1962: 11. Voss,
1962b: 248, 249. Adam & Rees, 1966: 30, pl. 10, figs 55, 56, pl. 45, fig. 271.

? Sepia jousseaumi Rochebrune, 1884: 117. Smith, 1916: 22. Adam, 1941: 108, pl. IV, fig. 3;
1944: 235-

? Sepia jousseaumei: Bartsch, 1915: 250.

? Sepia hierredda (non Rang) Turton, 1932: 2.

Type localities
Cape of Good Hope (S. vermiculata and S. jousseaumt).

Distribution

Animals: 30° 42'S, 15° 59’E (Voss 19626: 250) to Delagoa Bay, Mocambique
(Massy 1927: 156; Adam 1962: 11). Depth 0-248 m.

Shells: | Saldanha Bay to Chinde (S.A.M.).

Material

S.A.M. A2143, Chinde, mouth of Zambezi River; 1 shell
A2144, Durban; 1 shell
A30125, S 2° W of Cape Point, 23 km, 156 m; 4 juveniles
A30128, SE of Cape St. Blaize, 9 km, 62 m; 1 juvenile
A30129, Swartkops; 1 juvenile
A30130, Algoa Bay fishing grounds; 1 3
A30131, Table Bay; 1 9
A30182, locality unknown; 1 ¢
A30183, locality unknown; 1 9
A30483, Still Bay; 1 shell
A30487, locality unknown (det. A. L. Massy); 1 shell
A30496, locality unknown; 1 shell
A31238, off Hartenbos, near Mossel Bay, 18-24 m; 5 9
A31292, 33° o1’S, 17° 58’E (Saldanha Bay); 1 9

* Voss (personal communication) has two females caught off Beira in 1964 by the U.S. Indian

Ocean Expedition. The localities are 19° 51’S, 36° 21’E, 62 m, and 20° 30'S, 35° 49’E, 32 m.
The latter locality is within the limits of southern Africa as defined on page 194.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 205

Locality unknown; 1 9

Knysna estuary; 30 g, 18 9

Durban Bay; 1 3, 69

Ysterfontein beach; 1 shell

Nature’s Valley; 1 shell, discarded

Breede River mouth; 1 shell

Krom River mouth, Cape St. Francis; 12 shells
Saldanha Bay; 2 shells

Description

Mantle broadly oval, anterior margin somewhat produced dorsally,
ventrally entire. Head short and broad; fins broad and rounded but separate
posteriorly. Mantle slightly more slender and fins slightly wider in males than
in females.

Skin smooth, except in one male (A3o0130), which is sparsely papillose
dorsally, mainly on head. No indication in any specimens of long ridge-shaped
tubercles near fins, as mentioned by Massy (1925: 210), although some have
pale pink round spots in this region. Colour dark dorsally, pale ventrally and
some specimens show the well-known transverse zebra-like stripes dorsally on
mantle and on fourth arms, or at least at fin bases. Three specimens (males
A30130 and Ago182 and female, locality unknown) with pale stripes on dark
background, but one female (A30183) with dark stripes on pale background.

Arms longest ventrally, shortest dorsally, with arm length formula 4.3.2.1.
Arms joined by shallow web; arms III and IV keeled, arms II sometimes keeled,
arms I usually not—depending on manner of preservation. Arm tips somewhat
attenuated. Suckers quadriserially arranged on all arms to tips, decreasing
regularly in size from arm base. All sucker rings finely dentate, distal teeth
being longer than proximal ones. Protective membranes well developed.

Left ventral arm of male hectocotylized basally. About six normal suckers
at base of arm followed by 9-12 rows of modified suckers. The latter much
smaller and separated by transverse ridges on arm. Arm normal distally.

Tentacular club bearing small distal suckers in oblique rows of eight.
Suckers on proximal part of club variously enlarged: from ventral side, first
series of suckers of normal size, second series 1,5—2 times as big as first series,
third series 2,5—3 times as big as first series, fourth series 1-1,5 times as big as
first series, fifth series same size as first series. Rings of large club suckers smooth,
those of smaller suckers toothed distally. One sucker at tip of club concealed by
reflexed tip; immediately below this, two suckers, about twice as big as their
proximal neighbours, partly concealed by tip. Protective membranes of club
not meeting proximally; natatory membrane a little shorter than club.

Shell (Pls 35c, d, 36a, b) broadly oval, tapering somewhat anteriorly and
posteriorly; posterior spine present. Posterior end of shell and base of spine
covered with horny covering. Dorsal surface of shell tuberculate, with fairly
broad chitinous margin. No marked dorsal ridge. Ventral striated zone fairly

206 ANNALS OF THE SOUTH AFRICAN MUSEUM

long (about half total length), with median longitudinal ridge. Anterior border
of striated zone convex on either side of median ridge. Inner cone broad,
reflexed and fused to broad outer cone.

Seven shells from Durban Bay have two deep lateral grooves in ventral
surface (Pl. 35d). These grooves do not occur in any other shells, and it is
strange that all seven shells should have them, as these animals, though from
the same locality, were not all caught at the same time (two were caught in
February 1970 and the other five in April 1970).

Remarks

Adam (1941: 104) has shown that Sepza officinalis in the eastern Atlantic
Ocean is represented by four geographic races:
S. 0. officinalis Linnaeus, 1758: from the Atlantic coast of France to Rio de Oro
(Cap Blanc).
S. 0. filliouxi Lafont, 1868: from the Atlantic coast of France to the southern coast
of Scandinavia.
S. 0. hierredda Rang, 1837: from south of the Baie du Lévrier (Mauritania) to
the coast of Angola.
S. 0. vermiculata Quoy & Gaimard, 1832: southern Africa.
In addition a fifth race, S. 0. mediterranea Ninni, 1884, lives in the Mediterranean
Sea (Adam & Rees 1966: 32).

TABLE 1. A comparison of the relative dimensions (as °% MLd) of the animals of

the races Sepia officinalis vermiculata and Sepia officinalis hierredda. Only animals with

dorsal mantle length greater than 100 mm are included. The figures for S. officinalis
hierredda were calculated from relative dimensions given by Adam (1941).

Females Males
Range Mean Range Mean
MW S. 0. vermiculata 43,0-63,1 55,0 38,4-57,4 49,9
S. 0. hierredda 41-51 46,0 40-51 44,1
HW S. 0. vermiculata 34,9-50,0 4354 36,9-46,3 41,5
S. 0. hierredda 31-42,5 3553 28-36 31,9
AL I S. 0. vermiculata 27,7-55,0 38,9 28,8-46,3 37,3
S. 0. hierredda 265-3355 30,2 26,5-40 34,1
AL II S. 0. vermiculata 29,8-57,9 42,7 34,1-49,3 39,4
S. 0. hierredda 28-36,5 31,3 30-40 36,1
AL III S. 0. vermiculata 31,9-60,3 44,0 3557-54 4353
S. 0. hierredda 28,5-3755 33,8 335-43 39,2
AL IV S. 0. vermiculata 37,2-79,3 53,6 40,5-63,2 48,9
S. 0. hierredda 32,5-48 3955 40-63 50,2
Tcl S. 0. vermiculata 25,7—38,2 30,4 23,8-34 28,2
S. 0. hierredda 20,5-24,5 23,0 19-23 22-2

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 207

Adam (1941: 103) was able to show some differences in the shells of the
Atlantic races, but could find no marked differences in the forms of the animals.
He observed that S. 0. vermiculata differs from S. 0. hierredda in that the shell is
relatively wider and thicker and is more strongly tuberculate dorsally. Adam &
Rees (1966: 32) added that the posterior part of the shell of S. 0. vermiculata
seems to be more broadly rounded than that of S. 0. hierredda.

A comparison of the measurements of the S. 0. vermiculata specimens
described above, with those given for S. 0. hierredda by Adam (1941: 157-159),
shows that there are indeed some differences, but only in specimens having a
dorsal mantle length exceeding 100 mm. The most marked differences between
the two races is observable in the tentacular clubs (Table 1). The clubs of
S. o. vermiculata are relatively longer than those of S. 0. hierredda, and there is
apparently no overlap in the ranges of this dimension for the two races. The
differences in the other relative dimensions are less marked, but S. 0. vermiculata
has a relatively wider mantle and head, and longer arms (except the fourth arm
in the males) (Table 1). The relative dimensions of the shells also show some
differences between the two races (Table 2). The shells of S. 0. vermiculata are
relatively wider and thicker and have a slightly shorter striated zone. These
differences become more marked, the larger the shells.

TABLE 2. Comparison of relative shell dimensions (as % L) of
S. officinalis vermiculata and S. 0. hierredda. Only shells of length
greater than 100 mm are considered. The figures for S. 0. hierredda
were calculated from relative dimensions given by Adam (1941).

Range Mean

o. vermiculata 32,8-41,7 38,5

Width S.
S. 0. hierredda 30-36,5 34,0
Thickness S. 0. vermiculata Q,1-15,2 12,5
S. 0. hierredda 9,4-12,6 II,I
Length of striated zone S. 0. vermiculata 40,5-74,8 50,3
S. 0. hierredda 4355-7055 5755

The males of S. 0. vermiculata and S. 0. hierredda cannot be distinguished by
their hectocotyli. The hectocotylus of S. 0. hierredda has 8-13 transverse rows of
modified suckers (Adam 1941: 106). In the present specimens of S. 0. vermiculata
the hectocotylus has 9-12 rows of modified suckers. The extent of sucker
modification seems to increase with the size of the animal. Massy (1925: 210)
reported a large male of S. 0. vermiculata (MLd 64 inches = 165 mm) having
17 rows of modified suckers on the hectocotylus.

The southernmost record of S. officinalis hierredda is the Baia dos Tigres,
Angola (about 15° 20'S); S. officinalis vermiculata is known to occur as far north
as 30° 42’S (north of the Olifants River) off the west coast of southern Africa

208 ANNALS OF THE SOUTH AFRICAN MUSEUM

(Fig. 18). Dr. M.-L. Penrith (personal communication) points out that no
sepiid shells were found on any but the northernmost beaches of South West
Africa, although much other debris was washed up. In Angola, where S. o. hier-
redda is known to occur, sepiid shells were present on the beaches. The apparent
paucity of Sepiidae off the South West African coast suggests that there may be
no region of overlap between S. 0. Aierredda and S. 0. vermiculata.

On the east coast, animals of S. 0. vermiculata have been found as far north
as Delagoa Bay (Massy 1927: 156; Adam 1962: 11). Barnard collected a shell
(A2143) from Chinde, near the mouth of the Zambezi River; but this cannot be
considered a reliable locality, since sepiid shells have been known to drift over
long distances.

It is of interest to note that Sepza officinalis is the only species of Sepia known
to occur in estuaries in southern Africa. This, its wide geographic distribution
and its division into several geographic races, indicate that it is a very adaptable
species.

Sepia acuminata Smith, 1916

(Pl. 37a, b. ‘Tables 14-16)

Sepia acuminata Smith, 1916: 21, pl. II, fig. 3 (partim). Tomlin, 1923: 40. Robson, 1924a: 12
Massy, 1928: g1, pl. VIII, figs 1-7. Turton, 1932: 1. Voss, 1962b: 248. Adam & Rees
1966: 53, pl. 16, figs 91, 92, pl. 43, fig. 261.

non Sepia acuminata Smith, 1916: non pl. II, fig. 4 (= S. hieronis).

Rhombosepion acuminata Robson, 19245: 643.

Sepia sp. Adam, 1941: 121, pl. IV, fig. 6.

Type localities
Port Elizabeth; Tongaat beach, Natal (shells only).

Distribution

Animals: 29° 54’S, 31° 15’E (Robson 19245: 643, Sta. 103) to Punta Zavora,
Mocambique (S.A.M.). Depth 64-369 m.
Shells: | Cape St. Francis (S.A.M.) to ? Mombasa (Adam & Rees 1966: 53).

Material

S.A.M. A30147, locality unknown; 1 3
A31398, Corner, Mocambique, 25° 15'S, 35° 10’E, + 266m;1 g,19
A31399, Corner, Mocambique, 25° 15'S, 35° 10’E, 248 m; I
A31400, Corner, Mocambique, 25° 15/8, 35° 10’E, 257 m; 1 juvenile
A31401, Punta Zavora, Mocambique, 24° 35'S, 35° 25’E, 220-257 m;

Ig

N.M. 964, 965, off Tugela River, 48 km out, 64 m (det. A. L. Massy); 1 g, 12

Krom River mouth, Cape St. Francis; 1 shell (broken)

Umngazana River mouth, west Pondoland; 1 shell

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 209

Description

Mantle broadly oval, anterior margin produced to fairly sharp point
dorsally, halfway along eyes or more; ventrally entire. Fins narrow, rounded
and separate posteriorly.

Skin smooth. Coloration (described from recently caught specimens from
Mocambique; other specimens considerably faded): dense concentration of
purple chromatophores dorsally on mantle, head and arms; fins pale at edges.
Chromatophores less dense ventrally, resulting in paler purple colour near fins
and brownish colour medially on mantle. Ventral surface of head and arms
almost white.

Arms short (less than half MLd), but somewhat longer in females than in
males. Arms I to III subequal in length, arms IV somewhat longer. Shallow
interbrachial web present, except between arms IV. All arms keeled to some
extent and bear moderately well developed protective membranes.

Arm suckers globose, quadriserially arranged to tips of arms. Chitinous
rings of suckers finely toothed distally, with nodular surface; adjacent skin of
suckers wrinkled.

Left ventral arm of male hectocotylized distally. Six or seven rows of
normal suckers on basal third of arm, followed by nine to ten modified rows,
in which all suckers greatly reduced in size. Two dorsal series of modified
suckers in normal position, but the two ventral series have moved close together
on ventral edge of sucker-bearing surface, almost forming single line. Thus two
dorsal series widely separated from two ventral series by naked, wrinkled
region. Distal third of arm with normal suckers, decreasing in size to arm tip.

Tentacular club recurved, bearing numerous small subequal suckers in
transverse rows of eight. Reflexed tip of club obscures two somewhat larger
suckers. Sucker rings toothed distally; protective membranes separate
proximally; natatory membrane somewhat longer than club.

Shell (Pl. 37a, b) broadly elongate, sharply pointed anteriorly, rounded
posteriorly, bearing spine with tip directed dorsally; spine not keeled. Dorsal
surface of shell usually pink in colour, with broad chitinous margins laterally
and with median ridge sometimes sunken below level of rest of dorsal sur-
face. Striated zone long ventrally (about two-thirds total length); median
groove very faint or absent; striae regularly convex. Inner cone not much
thickened posteriorly, with narrow limbs curving over lateral edge of striated
zone. Outer cone only slightly calcified and mostly chitinous posteriorly,
where it curves sharply in ventral direction. Shell thickest at anterior end of
striated zone.

Remarks

In Smith’s (1916: 21) original description, he mentioned and figured
(pl. II, fig. 4) one specimen from Tongaat Beach which differed from the other
examples of S. acuminata. From the figure, this specimen is clearly S. hieronis (cf.).

210 ANNALS OF THE SOUTH AFRICAN MUSEUM

The shell from the ‘Ph. Dautzenberg’ collection, illustrated by Adam
(1941, pl. IV, fig. 6) looks very much like that of S. acuminata.*

Adam & Rees (1966: 53) reported some broken shells, almost certainly
pertaining to $. acuminata, from Mombasa. This identification is the more
probable since S$. acuminata is now known to occur as far north as Mocambique,
and is apparently a subtropical species. The male specimen reported by Adam
& Rees from the U.C.T. Ecological Survey, AFR 1051 K, comes from 29° 54'S,
31° 13’E, 369 m (ex U.C.T. catalogue).

Sepia confusa Smith, 1916
(Pl. 37c, d. Figs 4, 5. Tables 17-19)

Sepia burnupi Hoyle, 1904: 27, pl. I, fig. 192 (partim).

Sepia confusa Smith, 1916: 24, pl. II, figs 7, 8. Tomlin, 1923: 41. Robson, 1924a: 12. Turton,
1932: 1. Voss, 1962b: 248. Adam & Rees, 1966: 65, pl. 18, figs 112, 113, pl. 42, fig. 248.

Doratosepion confusa Massy & Robson, 1923: 435, figs 1-3. Carleton & Robson, 1924: 259,
pl. 3, figs 1-6.

Doratosepion confusum: Robson, 1924): 647.

Sepia (Doratosepion) confusa Massy, 1925: 221, pl. XIII, figs 20, 21, 24-28, pl. XIV, fig. 38;
1928: 93.

Type localities
Port Elizabeth; Tongaat beach, Natal (shells only).

Distribution
Animals: 29° 52’S, 31° 17’E (off Durban) (Robson 1g24a: 12, 19246: 647,
Sta. 95) to 5° 39'S, 39° 16’E (Zanzibar area) (Adam & Rees
1966: 65). Depth 64-352 m.
Shells: Port Elizabeth (Hoyle 1904: 27) to Chinde, mouth of Zambezi
River (S.A.M.).
Material
S.A.M. A2140, Chinde, mouth of Zambezi River; 1 shell
A6516, off Tugela River, 116-134 m;1 3
A30292, 35 km S of Tugela River, 116-134 m; 2 3, in poor condition
A31402, Corner, Mocambique, 25° 15'S, 35° 10’E, + 266m; 4 3
A31403, Corner, Mocambique, 25° 15'S, 35° 10’E, 257 m; 1 2
A31404, Punta Zavora, Mocambique, 24° 35'S, 35° 25'E, 220-257 m;
2g
N.M. 961, 962, 48 km off Tugela River, 64 m (det. A. L. Massy); 2 3
N.M. 963, Natal coast (det. A. L. Massy); 2 shells

Description

Mantle elongate, anterior margin strongly produced dorsally, ventrally
slightly emarginate. Fins of male very broad, especially towards posterior end,
and extended to form ‘tail’ beyond posterior end of mantle (Fig. 4). “Tail’

* This identification has been confirmed by Adam (personal communication).

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 2II

Fic. 4. Ventral view of Sepia confusa male, to show ‘tail’-like
extension of fins. Modified after Massy & Robson (1923).

212 ANNALS OF THE SOUTH AFRICAN MUSEUM

length approximately equal to or greater than MLd. Fins of female rounded
and separate posteriorly, not forming ‘tail’.

Skin smooth, except in one male (A6516) which has few small papillae
mid-dorsally on mantle. Colour dark purple mid-dorsally on mantle and above
eyes. Rest of body paler pinkish-brown, where chromatophores less expanded.
Ventral surface of head almost white.

Arms relatively short (not more than half MLd). Dorsal arms of males
slightly longer than others, which are subequal in length. Arms of female
subequal. Interbrachial web deepest between arms III and IV, absent between
fourth pair. Arms I to III slightly keeled, arms IV more strongly so. Protective
membranes fairly well developed, especially along attenuated part of arms,
where membrane curves over, partially covering suckers. Arms attenuated at
tips in males, over about half arm length in female. No sign of hectocotylization
in males.

Arm suckers globose, with smooth rings; quadriserially arranged to tips of
arms in males. Proximally, two median series of suckers much larger than
lateral series, but on attenuated part of arm all suckers minute and of same
size. In female, suckers as those of males proximally, but becoming biserial and
widely spaced over flattened attenuated part of arm.

Tentacular club slightly recurved, with distal suckers in oblique rows of
eight, and with nine median suckers variously enlarged. Of these, five much
larger, other four grading to normal size. Three slightly larger suckers partially
concealed by reflexed tip of club. Sucker rings broad and nodular, with finely
dentate inner edge. Protective membranes of club moderately well developed,
separate proximally. Natatory membrane a little longer than club.

Shell (Pl. 37c, d, Fig. 5) very narrow, pointed anteriorly, with posterior
spine directed dorsally. In male specimens from Mocambique, delicate keel
runs from base of spine and for some distance along dorsal face of shell; dry
shells show no indication of keel on spine. Dorsal surface of shell with narrow,
heavily calcified region medially, coloured pink; shell chitinous laterally.
Distinct broad median ridge, narrowing posteriorly, limited by lateral grooves.
Whole dorsal surface shows concentric striae, /-shaped. Ventral striated zone
just over half total length. Striae wavy, anterior border of striated zone angular
on either side of midline (//\-shaped). Median longitudinal groove distinct.
Inner cone forms free ledge posteriorly (Fig. 5) and has long narrow limbs.
Outer cone broad and deep posteriorly, forming chitinous wings. Shell thick,
ventral surface strongly convex; maximum thickness occurring immediately
anterior to striated zone.

Remarks

Detailed structure and possible function of the ‘tail’ are discussed by
Massy & Robson (1923) and by Carleton & Robson (1924).

The male of S. confusa is distinguished from the other ‘doratosepion’ species
by its ‘tail’-like extension of the fins. The female may be separated from all

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 213

except S. joubini by the biserial sucker arrangement on the attenuated part of
the arms. It is distinct from the latter species in that the protective membranes
on the tips of the dorsal arms are not expanded as they are in S. joubini.

Fic. 5. Sepia confusa male (A31402). Detail of posterior end of
shell, ventral view.

Sepia incerta Smith, 1916
(Pls 38a, b, 39a, b. Fig. 6. Tables 20-22)

Sepia burnupi Hoyle, 1904: 27, pl. I, figs 190, 191 (partim).

Sepia incerta Smith, 1916: 23, pl. II, fig. 6. Tomlin, 1923: ah Voss, 19625: 248. Adam & Rees,
1966: 67, pl. 19, figs 114, 115, pl. 41, fig. 241.

Sepia (Doratosepion) incerta Massy, 1925: 219, pl. XIII, figs 22, 23, 29-36, pl. XIV, figs 40, 43.

Sepia (Doratosepion) burnupi (non Hoyle) Massy, 1925: 215, pl. XII, figs 12-19, pl. XIV, figs 39,
41, 42; 1928: 94. Barnard, 1962: 252, fig. 4.

? Sepia incerta: Turton, 1932: 1.

Type localities
Port Elizabeth; Tongaat beach, Natal (shells only).
Distribution

Animals: 33° 07'S, 27° 56’E* (East London area) (Barnard 1962: 252) to
Durban (Adam & Rees 1966: 67). Depth 70-79 m.

* Not 33° 04’S, 27° 54’E, 27 fm., as stated by Barnard (1962: 252).

214 ANNALS OF THE SOUTH AFRICAN MUSEUM

Shells: Port Elizabeth (Hoyle 1904: 27) to Tongaat beach, Natal (Smith
1916: 23).

Material
S.A.M. A30143, 33° 07'S, 27° 560°E,* 79m; 5 g, 19
Ag30480, locality unknown; 2 shells
S.A.M. S1, Punta Zavora, Mocambique; 9g shells
N.M. 956, 957, 958 A, B, Natal coast, in stomach of Ground Shark (det.
A. L. Massy); 2 3, 2 shells
N.M.g969, 970, Cape Henderson (det. A. L. Massy); 1 9, 3 shells

Description

Mature males: mantle elongate, about three times as long as wide.
Anterior mantle margin produced dorsally, emarginate ventrally. Fins narrow,
fused posteriorly over tip of mantle.

Chromatophores densest mid-dorsally, fewer towards fins and sides of
head; ventrally sparse, forming spots on mantle and, in one large specimen,
larger spots on funnel. Series of large spots present along dorsal midline of
each fin.

Arm lengths unequal (formula 1.4.2 = 3). Dorsal arms longest, modified
(Fig. 6c): proximal quarter bearing 10-18 normal suckers basally, then
8-10 suckers on thickened transverse ridges on arm. Above this, protective
membranes expanded over about half arm length, supported by transverse
thickenings in the membrane. Membranes joined over inner surface, distal to
sucker bearing portion of arm; ventral membrane more expanded than dorsal
one and forming lamella with maximum width of 10-20 mm, about three-fifths
from arm base. Distal quarter of arm attenuated, with protective membranes
rapidly reduced towards arm tip.

Lateral and ventral arms bearing suckers quadriserially arranged almost
to tips; median suckers larger than lateral suckers, situated on protective
membranes. Suckers on these arms decrease gradually in size towards tip—
none enlarged. No sign of hectocotylization on either ventral arm in any males.

Suckers globose; rings very finely serrated, almost smooth. Skin wrinkled
immediately adjacent to chitinous rings of suckers.

Arms joined by shallow web, deepest between arms III and IV; no web
between ventral pair.

Tentacular club recurved, bearing numerous small suckers distally in
obliquely transverse rows of eight. Four or five suckers at proximal end of one
of the median rows much enlarged. Large club suckers with finely serrated
rings, small ones with toothed rings. Natatory membrane extends a little
beyond proximal limit of club; protective membranes separate distally.

Young male (A30143, MLd 93 mm): dorsal arms shorter than ventral
arms, but modified to some extent. Protective membranes well developed,

* See footnote on p. 213.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA =. 215

Fic. 6. Sepia incerta. a. Ventral and b. dorsal views of female

(Ago143). c. Right dorsal arm of male; d. detail of posterior

end of shell (g, A30143), ventral view. c. Modified after
Barnard (1962); a. b. and d. original.

216 ANNALS OF THE SOUTH AFRICAN MUSEUM

increasing gradually in width towards distal end, but not forming marked
lamellate expansions found in mature males. Protective membranes only 3 mm
wide at maximum width, but already fused over inner surface of arm. On
proximal two-fifths, 14 normal suckers and 10-12 suckers on transverse ridges.
Ventral arms relatively shorter than in mature males.

Females (Fig. 6a, b): small in comparison with large males, but mature.
Ovary well developed, containing numerous oval eggs. Well developed
nidamental glands present. Females differ from males in that dorsal arms not
modified; arm formula variable. Lateral arm pairs not equal in length, unlike
those of males. Dorsal arms of female with well developed protective membranes,
but these not expanded into lamellae. All arms of female bear quadriserial
suckers. Arm tips (less than half arm length) attenuated, bearing quadriserial
suckers and having well-developed protective membranes folding over inner
surface of arm.

Shell (Pls 38a, b, 39a, b) narrow, elongate, acuminate anteriorly and
posteriorly. Dorsally with wide chitinous margins and three median longitudinal
ridges, pink in colour, separated by two longitudinal grooves. Median ridge
broad anteriorly, narrowing posteriorly; two lateral ridges narrow and
indistinct anteriorly, but broader and more marked posteriorly. Ventrally with
shallow longitudinal median groove. Striated zone long and anteriorly convex.
Striae /\-shaped posteriorly, becoming more convex anteriorly. Inner cone with
long narrow limbs lying at sides of striated zone; inner cone raised posteriorly
to form ledge over end of striated zone. Centre of this posterior ledge bisected
by deep longitudinal groove (Fig. 6d). Outer cone forming posterior, chitinous
wings, and thickened over base of spine. Spine not keeled; directed dorsally.

Remarks

The soft parts of Sepza incerta were first described by Massy (1925: 215, 219).
She described two males as S. burnupi and two females as S. incerta, but a
re-examination of the shells of these specimens has indicated that in fact all
four specimens are referable to S. incerta. Barnard (1962: 252) described another
six specimens of S. incerta under the name S. burnupi.

S. incerta is distinguished from the other species of the ‘doratosepion’ group
as follows: The male of S. incerta is characterized by its modified dorsal and
unmodified ventral arms. The female is distinct from those of S. confusa,
S. joubim and §. adam: in that less than half the arm length is attenuated distally,
and the suckers on the attenuated part are quadriserial. From the female of
S. burnupi it can only be separated by the differences in the shells (and the larger
size of the animals).

The shell of S. incerta differs from those of S. burnupi, S. joubini and S. adami
in that the inner cone is raised posteriorly, forming a deep pocket over the end
of the striated zone, as in S. confusa. It differs from the latter species in having a
deep longitudinal groove in the centre of the raised posterior portion of the
inner cone.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 217

Sepia burnupi Hoyle, 1904
(Pl. 38c, d. Figs 7, 8. Tables 23, 24)
Sepia burnupi Hoyle, 1904: 27, pl. I, figs 188, 189 (partim). Bartsch, 1915: 250. Smith, 1916: 23,
pl. II, fig. 5. Voss, 19626: 248. Adam & Rees, 1966: 81, pl. 20, figs 127, 128.
non Sepia burnupi Hoyle, 1904: non pl. I, figs 190, 191 (= S. incerta), non pl. I, fig. 192 (= S. confusa).
Sepia exsignata Barnard, 1962: 250, fig. 3.
non Sepia burnupi: Barnard, 1962: 252, fig. 4 (= S. incerta).
? Sepia burnupi: Turton, 1932: 1.

Type localities
Umkomaas, Natal (S. burnupi, shells only); off Umhlanga River, Natal,
40-48 m (S. exsignata).
Distribution
Animals: Off Umhlanga River, Natal (Barnard 1962: 252). Depth 40-48 m.
Shells: ? Port Alfred (Turton 1932: 1) to Tongaat beach, Natal (Smith
1910229).
Material
S.A.M. A2147, Scottburgh, Natal; 3 shells
A6525, off Umhlanga River, Natal, 40-48 m; 1 3, 1 9 (holotype and
allotype of S. exsignata)
N.M.958 C, ? Natal coast; 1 shell
N.M.4073, off Umhlanga River, Natal, 40-48 m; 1 ¢ (paratype of S. exsignata)

Description

Mantle elongate oval, anterior margin somewhat produced dorsally,
emarginate ventrally. Mantle pointed posteriorly. Fins beginning at anterior
mantle edge, fairly wide. Fins of male drawn out into overlapping points
posteriorly (Fig. 7b); rounded ventrally and meeting in midline in female
(Fig. 7a).

Skin sparsely papillose dorsally over head and mantle, with a series of
elongate tubercles along fin bases dorsally. Skin smooth ventrally, except for
longitudinal dermal fold on each side of mantle, between midline and fin bases.

Colour dark dorsally, particularly between eyes and on that part of mantle
covering shell; fins pale. Ventral surface pale, but with scattered chromato-
phores between dermal folds and fins. In addition, female has four oval silvery-
blue spots dorsally between eyes.

Arms of female subequal in length, except ventral pair, which a little
longer. Arm tips somewhat attenuated; arms III and IV keeled. Suckers on all
arms quadriserial, decreasing in size towards tips. Median series of suckers
somewhat larger than lateral series.

In male, dorsal and ventral arms modified. Dorsal arms (Fig. 8e) bearing
long cirriform processes laterally, and shorter ones on dorsal margin, supporting
broad protective membranes. Basally, suckers normal and quadriserial, but
rapidly decreasing in size distally, where cirriform processes extend further

218 ANNALS OF THE SOUTH AFRICAN MUSEUM

b

Fic. 7. Sepia burnupi, A6525. a. Female (allotype of S. exsignata),
b. male (holotype of S. exsignata). Left: dorsal and right:
ventral views.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 219

towards midline of arm, until at distal end they alternate medially with minute,
biserial suckers. Tips of lateral cirri swollen. Distal tip of arm (about 2 mm)
bare, without suckers or cirri.

Right ventral arm of male with normal quadriserial suckers over most of
its length, decreasing markedly in size distally, followed by short bare portion
of the arm. Tip bearing short cirri, of which dorsal cirri better developed than
ventral ones (Fig. 8d). There are a few small suckers between the cirri. Left
ventral arm (Fig. 8c), the hectocotylus, basically like right ventral arm; but in
second quarter of arm, ventral protective membrane very well developed,
thrown into folds, and ventral series of suckers absent.

Lateral arms like those of female, with quadriserial suckers decreasing in
size towards attenuated arm tips. In both sexes, sucker rings without teeth,
nodular on upper surface.

Tentacular club of both sexes slightly recurved (Fig. 8b), bearing small
suckers in oblique rows of eight, and five suckers greatly enlarged. Two suckers
at tip larger than their neighbours, partly covered by reflexed tip of club. All
tentacular sucker rings finely dentate. Protective membranes well developed,
approximating very closely at base of club, without fusing. Natatory membrane
only a little longer than club.

Shell (Pl. 38c, d, Fig. 8a) narrowly elongate, pointed anteriorly and
posteriorly. Dorsally, shell calcified only along narrow median strip; laterally
with broad chitinous margins. Calcified region finely granular, coloured pink,
with fairly well defined median ridge. Posterior spine directed dorsally, not
keeled. Deep median ventral groove present over whole length of shell. In
region of striated zone, ventral surface highest at limbs of inner cone, shelving to
lowest point at median groove, giving ventral surface //\-shape in cross-
section. Striated zone long, with acuminate anterior border. Striae /\-shaped.
Inner cone has curved around lateral edges of striated zone, covering its sides
with very thin layer (through which striae can be seen) and having, like
S. insignis, its limbs lying on the phragmocone and not at its sides. Inner cone
only very slightly raised posteriorly, not forming ledge. Outer cone fairly
narrow, but forming wings posteriorly. Shell thickest just anterior to striated
zone.

Remarks

Hoyle (1904: 27) originally described S. burnupi on the basis of five shells.
Smith (1916: 23), on re-examining these shells, and with additional material,
separated them into three species, S. burnupi, S. incerta and S. confusa.

The first animals referred to S. burnupi were two male specimens described
by Massy (1925: 215), but these in fact pertain to S. incerta. Barnard (1962: 250)
described two males and one female (S.A.M. A6525, N.M. 4073) as a new
species, Sepia exsignata. Adam & Rees (1966: 83), going only by Barnard’s
description, suggested that S. exsignata is probably synonymous with S. burnupi.
The decalcified shell of S. exsignata was examined and, despite its rather poor

220 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fic. 8. Sepia burnupi, A6525. a. Ventral view of shell of female;
b. left tentacular club of male; c. hectocotylized left ventral arm of
male; d. tip of right ventral arm of male; e. modified dorsal arm of
male. c. d. e. Modified after Barnard (1962); a. and b. original.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 221

condition, was found to be identical with that of S. burnupi. The two species are
here formally synonymized.

The shell of S. burnupi differs from those of S$. confusa and 5S. incerta in that
the inner cone is low posteriorly and does not form a deep pocket over the end
of the phragmocone. It differs from S. joubini and S. adami in the /\\-shape of
the striated zone, with the raised limbs of the inner cone lying near the peaks
of the (\ and by the angular shape of the striae.

The male of S. burnupz is distinguishable from those of the other species of
the ‘doratosepion’ group by the modified tips of the ventral arms. The female is
distinct from S. confusa, S. joubint and S. adami in that the arms are attenuated
over less than half their length, and the suckers on the attenuated portion of the
arms are quadriserial; from S. zncerta it can be separated only by the shell.

Sepia joubin Massy, 1927
(Figs 2, 9. Tables 3, 25, 26)

Sepia (Doratosepion) joubint Massy, 1927: 161, pl. XVIII, figs 1-10.
Sepia joubini: Voss, 1962b: 248. Adam & Rees, 1966: 70, pl. 43, fig. 257.

Type localities
Tugela River mouth, NW by N 2 N, 25 km, 66-77 m; Cape Natal,
W by N, 10,5 km, 99 m.

Distribution
Animals: Off Tugela River mouth to Cape Natal (Massy 1927: 161, and
S.A.M.). Depth 66-134 m.

Material

S.A.M. Ag3o141, off Cape Natal, S 79°E 10,5 km, 99 m; 1 3, 15 2
A30142, 35 km S of Tugela River mouth, 116-134 m; 1 3, 2 9
Ago172, S 11°E of Tugela River mouth, 29 km, 84-101 m; 4 3g, 1 9
A31393, off Cape Natal, S 79°E 10,5 km, 99 m; 11 ¢

Description

Animals small. Mantle elongate, anterior margin produced dorsally,
slightly emarginate ventrally. Posterior end pointed. Fins narrow, widening
somewhat posteriorly, where rounded and separate.

Skin smooth in most specimens, but some (A30172) have a few dark
papillae dorsally. Colour pale, with sparse chromatophores. Dark area dorsally
over shell, and oval orange tubercles at fin bases. Pale silvery-blue areas present
dorsally in front of and behind eyes, and sometimes also between eyes. These
silvery regions not visible in all specimens. Ventral mantle surface pale, with
somewhat darker region between midline and fin base. Also a few tubercles
present on lateral margin of this area. Head translucent under and on either
side of funnel, but beige at arm bases. Males with distinct red spot on lateral
side of each dorsal arm, near base. Smaller red spot present on lateral side of

222 ANNALS OF THE SOUTH AFRICAN MUSEUM

ventral arm, and sometimes small spots also present on other arms. These red
spots absent in females.

Arms short (about one-third MLd or less). In males arms subequal in
length, attenuated only at tips. Suckers on dorsal arms arranged in about
three pairs basally, then quadriserial to arm tip. Suckers on lateral arms
arranged in oblique quadriserial rows, except for few irregularly arranged
suckers basally. Minute suckers on extreme tip of arm biserial for a few rows.
Right ventral arm bearing large irregularly arranged suckers on basal half;
suckers much smaller distally, arranged quadriserially. Left ventral arm
hectocotylized, somewhat longer than its right counterpart. Suckers large
basally, variously arranged. Distal hectocotylized region obscure and not
always recognizable as such. Protective membranes expanded and small
suckers arranged quadriserially but widely spaced over about eight rows.
Distal tip of arm bearing minute quadriserial suckers.

In female, lateral arms longer than dorsal and ventral ones. Dorsal arms
not markedly attenuated, bearing quadriserial suckers separated distally by
deep median longitudinal groove. Protective membranes very well developed
near arm tip and folded over inner surface of arm. Lateral arms attenuated
over about half their length. Proximal part of arm bearing suckers arranged
quadriserially, except a few irregular ones basally. Distally, suckers become
minute and biserial, separated by longitudinal groove. Ventral arms not
attenuated, bearing quadriserial suckers to tips.

Suckers globose in both sexes, having smooth rings. Dorsal and ventral

Fic. 9. Sepia joubini, A30172. Detail of posterior
part of shell, ventral view. See also Figure 2.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 223

arms keeled; interbrachial web highest between lateral arms, absent between
fourth pair.

Tentacular club with many small subequal suckers distally in obliquely
transverse rows of eight, and four median basal suckers greatly enlarged. All
tentacular suckers with finely toothed rings. Tip of club reflexed, partly
concealing two somewhat larger suckers. Natatory membrane extending
beyond club.

Shells all decalcified, and damaged when removed. Shape of shell (Fig. 2)
narrow elongate, pointed anteriorly. Indication of median longitudinal ridge
dorsally; posterior spine present, not keeled, directed dorsally (Fig. 9). Median
longitudinal groove present ventrally. Striated zone long (about two-thirds
shell length), convex on either side of median groove, becoming flattened
posteriorly. Striae convex and wavy, becoming more angular anteriorly. Inner
cone with long narrow limbs, forming low ledge posteriorly. Outer cone
forming posterior wings. Shell thickest in region of anterior end of striated zone.
Not known if peculiar structure at base of posterior spine (Fig. 9) occurs in all
specimens.

Remarks
The specimens A30141 are topotypes, since they were from the same haul
as Massy’s syntypes (P.F.10715). These topotypes were not seen by Massy.
Barnard stated in his notes: “the whip-like tips (of the arms, in the female)

TABLE 3. Sepia joubini females: increase in arm length and
extent of attenuation of the dorsolateral arm with increase
in size.

MLd Arm II Attenuated part
Length Length 8

(mm) (mm) (mm) Arm II
A3oI4i 26 8 3 3755
A3or4i 28 8 3 3755
Ago14i 29 8 2 25,0
Agoi4I 30 II 5 45.5
A3oI4i 32 II 4 36,4
A30142 32 13 7 53,8
A3oI4i 34 10 4 40,0
A30142 35 14 7 50,0
Ago14I 35 13 7 53,8
A30141 36 13 7 53,8
A3ZorI4i 36 15 10 66,7
A3o141 37 14 8 5751
A3O141 37 17 9 52,9
A3o172 38 20 12 60,0
A3o141 39 16 9 56,3
Ago14I 39 17 9 52,9
A3O14I 40 13 7 53,8
A30141 43 17 9 52,9

224 ANNALS OF THE SOUTH AFRICAN MUSEUM

begin to develop from 30 mm mantle length; below this length the whip is not
clearly demarcated from the proximal portion and specimens may be difficult
to separate from australis. The latter in fact occurred in the same haul. The whip
becomes longer relatively to the proximal portion, and may become a little
longer than the latter.’

Table 3 shows the extent of attenuation of the dorsolateral arm in females
of 26 to 43 mm dorsal mantle length. In the smallest females, the tips of the
lateral arms already show some degree of attenuation, but the suckers on the
attenuated part are quadriserial. ‘The degree of attenuation of the arm increases
with growth of the animal, to a mantle length of about 36 mm, then levels off,
although the increase in arm length with growth appears to be linear. In the
larger specimens the suckers on the attenuated part of the arm are arranged
biserially, although this probably corresponds to a crowding of the quadriserial
condition.

The ‘australis’ specimens caught in the same haul as S. joubint (A30141) are
in fact S. adam: (A30149, A31394). In any case, young S. joubint without
attenuated lateral arms may be distinguished from S. australis by the absence of
a ‘light organ’ in the mantle cavity, and by the differences in the shells.
S. australis shells are less narrow and elongate; they have a more marked dorsal
rib and have a spine with a strong dorsal keel, whereas S. joubini shells have an
unkeeled spine. The outer cone forms posterior wings in S. joubint but not in
S. australis. |

The male of S. joubint may be distinguished from the other ‘doratosepion’
species by the red spots on the arms, the unmodified dorsal arms and rounded
fins. The female is characterized by the expanded protective membranes
distally on the dorsal arms.

Sepia adami n. sp.
(Fig. 10. Table 27)
Type locality
S 79°E of Cape Natal 10,5 km, 99 m.

Material
S.A.M. A31394, S 79°E of Cape Natal 10,5 km, 99 m (P.F.10717, 14 December

1900); 1 2 (holotype)
A30149, same locality as holotype; 5 2 (paratypes)

Description

Only females of this species known.

Mantle elongate oval, sharply pointed posteriorly (Fig. 1oc, d). Anterior
mantle margin produced dorsally to about anterior level of eyes in large
specimens (a little less in smaller ones) ; ventrally entire or slightly emarginate.
Head short and broad. Fins beginning a few mm behind mantle margin, narrow
and rounded posteriorly, but fused over posterior tip of mantle.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 225

Fic. 10. Sepia adami female, A31394 (holotype). a. Ventral view of shell; b. dorsal view of
anterior part of shell; c. ventral view of holotype; d. dorsal view of holotype; e. right tentacular
club.

226 ANNALS OF THE SOUTH AFRICAN MUSEUM

Skin sparsely papillose dorsally on head and mantle. Raised lunate
tubercles along fin bases dorsally. Colour (of preserved specimens) pale cream,
with somewhat darker region mid-dorsally over shell and over eyes. Scattered
reddish-brown spots present on arms. Ventral surface pale. Sometimes also a
line of pale lunate tubercles in this region (not in type specimen).

Arms subequal in length, joined by shallow web except between ventral
arms. All arms keeled. Suckers globose, quadriserial on all arms from base to
tip. The two median series of suckers much larger than the lateral ones; but on
tips of arms all suckers the same size and markedly smaller than basal suckers.
Sucker rings smooth.

Tentacular club (Fig. 10e) short, slightly recurved, with about five median
suckers enlarged and smaller distal suckers in five to six longitudinal series.
Two suckers at tip of club slightly larger than their neighbours and partially
concealed by reflexed tip of club. Rings of large club suckers finely toothed
distally, those of small suckers toothed right round. Natatory membrane
broad, extending a little beyond club proximally. Protective membranes
separate at base of club.

Shells of all specimens very soft and difficult to remove undamaged. Only
shell of holotype extracted in reasonably complete state (shell length 59 mm,
including spine). Unfortunately, this shell is somewhat abnormal, having
apparently been damaged and repaired, since outgrowth present mid-dorsally,
corresponding with deep cleft in mid-ventral region. ‘These do not occur in other
shell examined by dissection.

Shell (Fig. toa, b) narrow, elongate, broadest in anterior third, tapering
sharply towards posterior end. Shell calcified only in median region dorsally;
calcified part consisting of flat median ridge, broadest anteriorly, narrowing
posteriorly, and two lateral ridges, narrow and indistinct anteriorly, broadening
posteriorly to about twice width of median ridge. Posterior part of shell
somewhat damaged dorsally, but ridges apparently fuse and become less
distinct in this region. Calcified part of shell pale pink in colour, partially
rugose. Rest of shell chitinous dorsally. Posterior spine directed dorsally, not
keeled.

Striated zone occupies just over half ventral surface of shell. Anterior
border of striated zone convex, with invagination at median groove. But this
apparently due to repair outgrowth mentioned above, since anterior border
/\-shaped in other shell. Inner cone forming shallow pocket posteriorly over
striated zone, but not markedly raised. Limbs of inner cone long and narrow,
but it is not certain exactly how far they extend. Outer cone widens posteriorly,
forming small wings. Shell thickest anterior to striated zone.

Remarks

These specimens were found in the collection of the South African Museum,
labelled Sepia australis, and indeed they do superficially resemble this species.
S. adami differs from S. australis in the following ways:

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 227

1. The shell is markedly different in that it is more elongate, thicker, has

posterior wings and a spine without keels.

2. There is no ‘light organ’ in the mantle cavity.

The arm tips are more slender than in S. australis.

4. The colour is much paler than that of S. australis and there is no indi-

cation of an orange band in the region of the fin bases.

With the six female specimens of S. adam: described above were caught
eleven males which almost certainly pertain to S. joubini, and have been given a
separate catalogue number (A31393). The females of $. adami certainly differ
from those of S. joubini. In S. adami the suckers on the dorsal arms are clearly
quadriserial to the tips, whereas in S. joubinz females the suckers on the dorsal
arms appear to be biserial, although in fact they are in very oblique rows of
quadriserial suckers. Neither are the tips of the dorsal arms provided with
expanded protective membranes, as in S. joubini. In the latter species the
lateral arms are very attenuated over at least half their length, and the suckers
on the attenuated portion of these arms are biserial. In S. adami the lateral arms
are not attenuated and bear quadriserial suckers to the tips of the arms.

Distinctive characters

1. Tentacular club with unequal suckers

2. Shell with unkeeled spine and posterior wings

3. Arms subequal, with quadriserial suckers

This species apparently falls into the “‘doratosepion’ group, and is most
closely related to the other species of this group, viz. S. confusa, S. incerta,
S. burnupi and S. joubini. It differs from the females of all these species in that the
arms are not attenuated distally.

Sepia adami has been named after Professor Dr. W. Adam of the Institut
Royal des Sciences Naturelles de Belgique, in recognition of his extensive work
on the Sepiidae.

Sepia australis Quoy & Gaimard, 1832
(Pl. 40a, b. Figs 1, 3a, 21. Tables 28-30)

Sepia australis Quoy & Gaimard, 1832: 70, pl. 5, figs 3-7. Hoyle, 1912: 281, fig. 8. Bartsch,
1915: 250. Smith, 1916: 24, pl. II, fig. 9. Tomlin, 1923: 41. Robson, 1924a: 11. Turton,
1932: 2. Adam, 1941: 117, pl. IV, fig. 5; 1942: 10; 1959: 149, fig. 10. Voss, 19625: 248, 252;
1967: 64. Adam & Rees, 1966: 89, pl. 21, figs 138-142, pl. 45, fig. 270.

Sepia capensis d’Orbigny, 1845a: 283. Férussac & d’Orbigny, 1835-1848: 278, pl. VII, figs 1-3,
pl. XII, figs 7-11, pl. XVII, figs 18, 19. Tryon, 1879: 198, pl. 94, figs 440-442. Hoyle,
1886: 23, 217. Gibbons, 1888: 202. Bartsch, 1915: 250. Thiele, 1920: 438, pl. LII, fig. 14,
pl. LIII, figs 1-5.

Sepia sinope Gray, 1849: 106.

Rhombosepion australe Rochebrune, 1884: 85. Adam, 1944: 223.

Rhombosepion capense Rochebrune, 1884: 85. Robson, 19246: 641, fig. 24. Adam, 1944: 222.

Sepia (Doratosepion) australis Massy, 1925: 214.

Rhombosepion australis: Massy, 1927: 156.

non Sepia australis: d’Orbigny, 1845a: 294. Férussac & d’Orbigny, 1848: 285, pl. VII, fig. 4.
Hoyle, 1886: 22, 220. (= S. novaehollandiae).

non Sepia capensis d’Orbigny, 1826: Gray, 1849: 110 (= S. cultrata).

228

ANNALS OF THE SOUTH AFRICAN MUSEUM

Type localities
Agulhas Bank (S. australis), Cape of Good Hope (S. capensis), ? China

(S.

sinope) .

Distribution
Animals: South African coast, from 31° 43'S, 16° 13’E (off the Olifants River)

Shells:

Material
S.A.M.

(Adam & Rees 1966: 89) to S 73°E of Rame Head, 5 km (near
Port St. Johns) (S.A.M.).

Red Sea (Rochebrune 1884: 85; Adam 1942: 10, 1944: 222,
1959: 149).

Depth 2-459 m.

South African coast, from Braak River (Namaqualand coast)
(S.A.M.) to Port Alfred (Turton 1932: 2).

? China (Gray 1849: 106).

2727, Kalk Bay; 17 shells

A8982, Hout Bay, 17-37 m (det. A. L. Massy); 2 g, 3 9, in poor
condition

A8984, S 78°W of Lions Head 19 km, 110 m (det. A. L. Massy); 2 ¢

A8986, south of Knysna Heads 16 km, 95 m; 25 3, 34 9, 51 juveniles,
all in poor condition

A8988, N 48°W of Lions Head 80 km, 422 m; 4 specimens in poor
condition

A8g92, S 42°W of Cape St. Blaize 17,5 km, 2-6 m (det. A. L. Massy) ;
1 2 in poor condition

A8994, N 87°E of Cape Point lighthouse 15 km, 59 m (det. A. L.
Massy) ; II specimens in poor condition

A8g997, mouth of Hout Bay, 73-92 m (det. A. L. Massy); 1 9 in poor
condition

A29627, 11 km W of Slangkop, 128 m (det. G. L. Voss); 1 g, 1

A29734, west of Slangkop, 100 m (det. G. L. Voss); 2 9

Azo8o1, False Bay, 59 m; 1 9

A29802, False Bay, 59 m; 1 ¢

Ago148, S 57°W of Cape Point lighthouse 11 km, 147 m; 10 g, 9 9

Ago150, S 54°W of Gericke Point 21,5 km, 79 m; 7 3, 35 9

A30151, S 11°W of Great Fish Point lighthouse 5 km, 73 m; 4 4,6 2

A30152, S 6°E of Cape Infanta 21,5 km, 77m;1 4,1? 6

A30153, 33° 04'S, 27° 54'E, 50 m; 4 9, 2 PY

A30154, 34° 19'S, 18° 32’E (Buffels Bay), 59-64 m; 6 J, 12 9

A30155, SE of Cape St. Blaize 9 km, 62 m; 1 Q

A30156, Mossel Bay, 34° 14’S, 22° 23’E, 66 m (det. A. L. Massy); 1

A30157, S 37°W of Cape Hangklip 45,5 km, 183 m; 2 g, 49

A30158, S 70°W of Cape Infanta 8 km, 64 m (det. A. L. Massy); 1 @

A30159, S of Knysna Heads 16 km, 95 m; 3 g, 1 2

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 229

A30160, S 79°W of Table Mt. 64 km, 459 m; 1 3,69
A30161, S 16°W of Cape Point light 16 km, 156m ; 2 3, 5 9
A30162, S 34°W of Cape Infanta 30 km, 84.m;1? 3, 1? 9, 2 juveniles,
all in poor condition
A30163, S 73°E of Rame Head 5 km, 79 m; 2 3
A30164, S 62°E of Bird Island light 13,5 km, 72 m; 7 3, 3 2
A30165, S 70°W of Cape Infanta 8 km, 64 m (det. A. L. Massy);
13,29
A30166, S 20°E of Sebastian Bluff 15,5 km, 73 m; 1 9
A30167, S 48°W of Cape St. Blaize 43 km, 81-84 m; 2 3
A30168, off Buffels Bay (False Bay), 55 m; 2 3, 4 9
Ago169, S of Cape St. Blaize 57,5 km, 99 m (det. A. L. Massy);
5 specimens in poor condition
Ago170, S 37°W of Cape Infanta 7 km, 68 m; 1 g, 1 9
Ago171, 8S 14°W of Gericke Point 8 km, 64 m; 3 9
A30173, S 6°E of Cape Infanta 21,5 km, 77 m; 2 g, 1 ? 9 in poor
condition
A30175, S 42°E of Cape St. Blaize 9,5 km, 66 m; 2 J, 2 9
A30190, Salt River power station (Cape Town); 1 9
A30263, N 87°E of Cape Point lighthouse, 59 m (det. A. L. Massy) ;
5 specimens in poor condition
A30264, S 14°W of Gericke Point 8 km, 64 m (det. A. L. Massy);
I 4g, 1 9, in poor condition
A30266, S 50°E of Sebastian Bluff 13,5 km, 62 m (det. A. L. Massy); 1 3
A30267, S 34°W of Cape Infanta 30 km, 84 m (det. A. L. Massy);
I g, I 9, in poor condition
A30334, 5 11°W of Cape Point 15 km, 149-160 m; 1 9
A30504, Bloubergstrand; 1 shell
A30555, Bloubergstrand; 2 shells
A30564, Between Strandfontein and Muizenberg; 1 ¢
A30604, 32° 24'S, 17° 28’E, 193 m; 13 3g, 53 Q, all in poor condition
Ysterfontein beach; 1 shell
Mossel Bay (south of Slangkop on Cape Peninsula); 22 shells
Olifantsbosbaai (Cape Point Reserve); 7 shells
Millers Point, Simonstown; 6 shells
48 km N of Olifants River; 2 shells
Strandfontein (False Bay); 1 shell
Bloubergstrand; 1 shell
Between Strandfontein and Muizenberg; 11 shells
Namaqualand coast, between Sout and Braak Rivers; 1 shell
Simonstown; 2 shells
Arniston; 2 shells
Still Bay; 40 shells
Krom River mouth, Cape St. Francis; 1 shell (broken)

230 ANNALS OF THE SOUTH AFRICAN MUSEUM

Description

Mantle elongate oval, anterior margin dorsally produced, ventrally entire
or slightly emarginate. Head short and broad, fins narrow and rounded but
separate posteriorly.

Skin smooth. Colour dark purple dorsally on head and mantle, with
slightly paler area medially over shell, and narrow reddish-brown to orange
line at bases of fins. This line broadens posteriorly, but not meeting in the
midline. Fins pale yellow, with small purple spots. Mantle as dark ventrally as
dorsally, with purple chromatophores more concentrated near fin bases and
posteriorly on mantle. Fins pale ventrally, without chromatophores; head pale
ventrally except for some chromatophores on keeled edges of ventral arms.

Arms subequal in length, joined by shallow interbrachial web, deepest
between arms II and III, absent between ventral pair. Protective membranes
fairly well developed, especially at arm bases. Arms III moderately, arms IV
well keeled.

Suckers globose, quadriserially arranged, with the two median series much
larger than the lateral ones. Enlargement of median sucker series more marked
in male than in female. Arm tips relatively blunt (definitely not attenuated),
bearing minute suckers still arranged quadriserially. In male, distal suckers
with long teeth on distal part of rings; rings on proximal suckers smooth on
arms I and II, finely dentate proximally on arms III and IV. In female, rings
finely dentate in distal suckers, smooth in proximal suckers. Sucker rings
nodular on upper surface in both sexes, and adjacent skin of suckers wrinkled.

Left ventral arm of male hectocotylized over proximal two-thirds: five or
six normal suckers at base, followed by six to seven rows of minute suckers.
The two dorsal series of modified suckers arranged in normal positions, but the
two ventral series merged to form a single series of 12-14. minute suckers
situated on edge of ventral protective membrane. Distal third of arm normal.
Modified suckers described by Adam & Rees (1966: go) as consisting of
‘13 slightly alternate pairs of smaller ones’. This in fact corresponds to
6—7 quadriserial rows of suckers, of which two ventral series have merged to
form a single series, as mentioned above.

Tentacular club somewhat recurved, with small distal suckers in rows of
about five. Four median suckers at base enlarged, and two suckers at tip a little
larger than their neighbours, partially concealed by reflexed tip of club. Large
suckers with smooth rings, smaller ones finely dentate. All rings broad and
nodular. Natatory membrane very broad, a little longer than club. Protective
membranes not meeting proximally.

Shell (Pl. 40a, b) broadly elongate, somewhat pointed anteriorly, more
sharply so posteriorly. Posterior spine with dorsal keel, continuing along dorsal
surface of shell for some distance. Shell with broad median rib dorsally, limited
on either side by lateral groove. Striated zone long ventrally (about three-
quarters shell length) ; striae wavy. Anterior border of striated zone also wavy
and may be sharply pointed or rounded medially. Ventral surface raised in

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 231

middle to form broad ridge on either side of distinct median longitudinal
groove. Lateral to these ridges, shell concave. Inner cone not well developed
posteriorly but raised to form deep pocket; limbs of inner cone narrow. Outer
cone narrow, not forming posterior wings. Shell thickest and widest near
anterior border of striated zone.

Remarks

In the past there has been considerable confusion in the naming of this
species. Sepia australis was first described from a specimen from the Cape of
Good Hope by Quoy & Gaimard in 1832. Later d’Orbigny claimed that he had
described the same species as Sepia capensis in 1826, and used the name S. australis
for an Australian species. Hoyle (1909: 266) could find no confirmation of the
use of S. capensis prior to 1832, and upheld Quoy & Gaimard’s name S. australis,
renaming S. australis d’Orbigny as Sepia novaehollandiae.

Regarding Sepia sinope, Smith (1916: 24) remarked: “The name S. sinope
was substituted by Gray for the S. australis, Q. & G. (non d’Orb.), and he
quoted a single imperfect shell in the British Museum collection which was
labelled “‘China’’. No information concerning its acquirement is attached to the
specimen, and consequently in all probability the locality cannot be relied
upon. It certainly belongs to the present species’ (i.e. S. australis).

S. australis is one of the commonest Sepiidae along the coast of the western
Cape and its shells are found in abundance on the beaches. Turton (1932: 2)
mentioned that this species is rare at Port Alfred, and indeed not many records
are known east of Port Elizabeth.

The University of Cape Town Ecological Survey has a record of S. australis
from Durban. Unfortunately the animals were discarded and their identity
cannot be checked. Since S$. australis has been confused with S. joubini and
S. adami in the past, this record remains doubtful.

Adam & Rees (1966: 89) report a shell from Grahamstown, Natal. But the
only Grahamstown in South Africa known to the author is in the eastern Cape,
and is 64 km from the sea.

Voss (19625: 252, 1967: 64) is mistaken in stating that S. australis is known
only from southern Africa, since Rochebrune (1884: 85) and Adam (1942: 10,
1944: 222, 1959: 149) have reported this species from the Red Sea. The
occurrence of S. australis in China (Gray 1849: 106) is doubtful (see above).

Sepia tuberculata Lamarck, 1798
(Pls 39c, d, 40c, d. Fig. 11a. Tables 4, 31-33)

Sepia tuberculata Lamarck, 1798: 130; 1799: 9, pl. I, figs 14-8; 1822: 668; 1845: 372. Bosc,
1802: 45. Montfort, 1805: 274, pl. vii, figs 1-6. Blainville, 1825: 368, pl. I, fig. 2; 1827a:
figs 2-6; 18275: pl. 1, figs 2-6. Deshayes, 1832: 945. Férussac & d’Orbigny, 1835-1848:
277, pl. VI, figs 1-4. d’Orbigny, 18454: 281; 1845[-47]): pl. 3, fig. 11; 1845[-47]c: pl. 3,
fig. 11. Hoyle, 1886: 24, 217. Gibbons, 1888: 202. Smith, 1903: 356. Adam, 1941: 113,
pl. III, fig. 8. Voss, 1962b: 248. Adam & Rees, 1966: 106, pl. 26, figs 169, 170, pl. 27,
figs 171, 172, pl. 28, figs 173, 174, pl. 44, figs 265, 268.

232 ANNALS OF THE SOUTH AFRICAN MUSEUM

? Sepia mammilata Leach, MS. Feérussac & d’Orbigny, 1835-1848: 277 (S. mamillata), pl. [Vbis
(S. mammilata).

Spathidosepion tuberculatum Rochebrune, 1884: 93, pl. IV, fig. 3. Adam, 1944: 226.

Hemisepius (?) tuberculatus Smith, 1916: 25.

non Sepia tuberculata: Férussac & d’Orbigny, 1835-1848: non pl. XVII, figs 13-15. Steenstrup,
1875: IV, pl. I, figs 20, 21, pl. II, fig. 6. Hoyle, 1910: 265, figs 9, 10, pl. Va, figs 4-14.
(= S. papillata).

Type locality
?

Distribution
Animals: Melkbosstrand (S.A.M.) to Knysna (U.C.T.). Depth o-3 m.
Shells: | Kommetjie, Cape Peninsula to Nature’s Valley (S.A.M.) and
Malagasy (Adam, 1941: 114).
Material
S.A.M. A29781, Mossel Bay; 1 2
A29867, Strandfontein; 1 9
Agor121, David’s Kraal (near Cape Hangklip); 1 3
A30123, Simonstown Harbour, 3 m; 1 9
Ag30139, locality unknown; 1 9
Ago180, Melkbosstrand; 1 9
Ag30279, locality unknown; 1 ¢
A30485, Simonstown; 1 shell (of juvenile)
A30493, Simonstown; 1 shell
Ago500, Gordon’s Bay; 2 shells
Ag30511, Dalebrook; 2 9, 1 juvenile ? 9
A30559, Dalebrook; 1 9
Ago6o00, Millers Point, Simonstown; 2 3
A31235, locality unknown; 1 ¢
Nature’s Valley; 6 shells, discarded
Between Strandfontein and Muizenberg; 3 shells
Millers Point, Simonstown; 3 shells
Betty’s Bay; 1 shell
Strandfontein (False Bay); 1 shell
Kommetijie; 1 shell
Die Kelders; 1 shell, discarded
? Cape Agulhas; 4 shells
? Milnerton; 2 shells
? Pearly Beach, S of Gansbaai; 4 shells
? Die Kelders; 1 shell
? 11 km NW of Cape Agulhas; 7 shells

Description

Mantle short and broad, anterior mantle margin convex dorsally (not
produced), ventrally entire; mantle rounded posteriorly. Fins wide, beginning
a few mm behind anterior mantle margin, rounded but separate posteriorly.

|

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 233

Skin densely tuberculate on dorsal surface of head, arms, mantle and fin
bases. Ventral surface smooth except for large oval wrinkled area on each side
of mantle and on ventral surfaces of fourth arms. Colour dark purple dorsally,
pale buff ventrally.

Arms subequal in length, about half dorsal mantle length. Tips of all arms
except ventral pair attenuated. Interbrachial membrane present, except
between ventral arms. Arms III and IV keeled.

Suckers not globose; with finely dentate horny rings. Protective membranes
well developed. Suckers quadriserial on all arms to tips.

Left ventral arm of male hectocotylized basally. Dorsal series of suckers
normal, but the two ventral series are widely separated from these by a broad
naked area with transverse ridges. Distal half of arm normal.

Tentacular club long, slightly recurved, bearing small distal suckers
arranged in very oblique rows of eight. Four or five median suckers enlarged.
Suckers at tip of club no larger than their neighbours. Horny rings of large
suckers smooth, those of small suckers dentate. Protective membranes meeting
proximally. Natatory membrane extending a little beyond club.

Shell (Pl. 40c, d) oval but somewhat angular anteriorly. Dorsal surface
finely granular. Median ridge faint or absent. No posterior spine or knob, but
only a small hump present in this position. No distinct median groove ventrally.
Striated zone long, striae wavy with overall convex shape. Inner cone com-
pletely reflexed and fused to outer cone, but not well developed posteriorly;

Fic. 11. Comparison of tentacular clubs of a. Sepia tuberculata
(Agor139) and b. S. papillata (A30124).

234 ANNALS OF THE SOUTH AFRICAN MUSEUM

limbs of inner cone long and narrow. Outer cone broad laterally, narrowing
sharply posteriorly so that inner cone almost reaches posterior margin of shell.
Shell generally very thin, ventral surface flat or slightly concave.

Several shells (indicated with query in list of material) (Pl. 39c, d), found
on beaches, with markedly longer striated zone, and sometimes also with
somewhat wider outer cone.

Remarks

This species was synonymized with S. papillata by Férussac & d’Orbigny
(1835-1848: 277), but the two were rightly separated by Rochebrune (1884:
93-95). Hoyle (1910: 267), following Férussac & d’Orbigny’s synonymy,
described a specimen of S. papillata under the name S. tuberculata. In fact the
differences between the two species are quite clear (Table 4), the most marked
being the relative sizes of the large tentacular suckers. The two specimens
mentioned by Steenstrup (1875: IV) almost certainly also pertain to
S. papillata (cf.).

Adam (1941: 114) has mentioned a specimen of S. tuberculata from Port
Dorey, New Guinea. The correctness of this locality is doubtful (Adam,
personal communication), and has not been included in the distribution lists.

The wrinkled areas on the ventral surfaces on the mantle and fourth arms
are apparently used to hold on to a hard substratum such as rocks. S. tuberculata
kept in aquaria have been observed to use these wrinkled areas to cling to the
glass walls.

TABLE 4. Comparison of Sepia tuberculata with S. papillata.

S. tuberculata S. papillata

Dorsal mantle margin Convex Produced

Tips of arms I-III Attenuated Not attenuated

Tentacular Large Diameter less than width Diameter approximately equal

club suckers of tentacular club to width of tentacular club
Protective Joined proximally Separate proximally
membranes

Shell Dorsal No median ridge; no Median ridge and posterior
surface posterior knob knob present
Ventral Flat or concave; no dis- Convex; median groove
surface tinct median groove present
Inner Not well developed pos- Well developed posteriorly;
cone teriorly; limbs narrow limbs usually broad (but see

description)

Thickness Mean 6,6% shell length Mean 10,1% shell length

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 235

The shells with an exceptionally long striated zone (70,0-88,5°% shell
length) show no other differences from those of S. tuberculata (striated zone
54,1-73,9% shell length) and may constitute an extension of the known range
of striated zone length for this species, or may be distinct. No decision can be
made on their status until the soft parts can be studied.

Sepia papillata Quoy & Gaimard, 1832
(Pl. 41a—d. Figs 11b, 12. Tables 4, 34-37)

Sepia papillata Quoy & Gaimard, 1832: 61, pl. I, figs 6-14. Férussac & d’Orbigny, 1835-1848:
pl. IlIte', figs 1-5. Bartsch, 1915: 250. Tomlin, 1923: 40 (partim). Massy, 1925: 211;
1928: 92. Turton, 1932: 1. Adam, 19394: 55, pl. III, fig. 6; 1941: 112. Voss, 19626: 248, 251.
Adam & Rees, 1966: 108, pl. 28, figs 175-178.

Spathidosepion papillatum Rochebrune, 1884: 94. Adam, 1944: 226.

Sepia tuberculata (non Lamarck) Férussac & d’Orbigny, 1835-1848: pl. XVII, figs 13-15.
Steenstrup, 1875: IV, pl. I, figs 20, 21, pl. II, fig. 6. Gibbons, 1888: 202. Hoyle, 1910:
265, figs 9, 10, pl. Va, figs 4-14.

non Sepia papillata: Smith, 1916: 22, pl. II, figs 1, 2 (= S. simoniana).

Type locality
Cape of Good Hope.

Distribution
Animals: Liideritzbucht (Angra Pequena) (Hoyle 1910: 265) to Natal coast, off
Tugela River and Umvoti River (Massy 1928: 92). Depth 26-127 m.
Shells: Orange River mouth (S.A.M.) to ? Tongaat, Natal (Tomlin
1923: 40).
Material
S.A.M. Agor18, Woodstock power station, Cape Town; 2 ¢ (one shell missing)
A30119, S 62—79°E of Bird Island lighthouse 15 km, 73-48 m; 1 Q
(shell missing)
Ago120, Woodstock power station, Cape Town; 1 3
A30124, locality unknown; 1 Q (shell missing)
A30136, S 87°E of Cape St. Blaize 9 km, 51 m; 1 2
A30137, Hout Bay, 37-73 m; 1 3
A30138, N 3°E of Green Point lighthouse 4 km, 40 m;1 ¢
A30140, S 8-17°E of Maalgaten River mouth 11-14 km, 61-64 m; 1 9
A304.76, Bloubergstrand; 1 shell
A30482, S 62—79°E of Bird Island lighthouse 15 km, 73-48 m; 1 shell
A30497, locality unknown; 1 shell
Ag30507, Elandsbaai beach, after red tide; 2 3, 2 9
A30509, Elandsbaai beach, after red tide; 1 3
A30553, Orange River mouth; 2 shells
A30554, Bloubergstrand; 1 shell
A31250, Castle Rock, False Bay; 1 3
Ysterfontein; 1 shell

236 ANNALS OF THE SOUTH AFRICAN MUSEUM

Nature’s Valley; 1 shell, discarded

Between Strandfontein and Muizenberg; 14 shells
Mossel Bay (S of Slangkop on Cape Peninsula); 1 shell
Olifantsbosbaai (Cape Point Reserve); 1 shell

Millers Point, Simonstown; 2 shells

48 km N of Olifants River; 4 shells

Strandfontein (False Bay); 6 shells

Betty’s Bay; 1 shell

Umngazana River mouth, west Pondoland; 2 shells
Muizenberg; 2 shells

Kommetjie; 4 shells

Milnerton beach, Cape Town; 2 shells

Namaqualand coast, between Sout River and Braak River; 1 shell
Still Bay; 8 shells

Simonstown; 2 shells

Arniston; 2 shells

Die Kelders; 3 shells, discarded

Krom River mouth, Cape St. Francis; 32 shells

Description

Mantle broadly oval, anterior margin somewhat produced dorsally,
slightly emarginate ventrally. Fins broad, beginning a few mm behind mantle
margin; posteriorly rounded and separate.

Skin tuberculate dorsally and laterally on head, arms and mantle, but
smooth ventrally except for two large wrinkled patches on mantle, and smaller
patches along ventral surfaces of fourth arms, as in S. tuberculata. Colour dark
reddish-brown to purple dorsally on tuberculate surfaces, pale ventrally with
scattered chromatophores, which are somewhat more dense on fins.

Arms subequal in length, about 50% MLd; arms I to III attenuated over
about distal quarter. All arms keeled, and joined by fairly deep interbrachial
web, which may attain half arm length.

Suckers not globose; with finely toothed chitinous rings. In the female,
suckers quadriserially arranged on all arms. Suckers on attenuated arm tips
minute, but still quadriserial. In the males, suckers quadriserial over most of
arm, but attenuated tip broader and flatter than in female, and minute suckers
arranged in eight series. Ventral arms not much attenuated at tips, but suckers
also show multiplicity, although less extensively than on other arms.

Left ventral arm of male hectocotylized proximally. Basally about
13 suckers arranged normally, followed by modified region, consisting of eight
quadriserial rows of suckers. In this region the two dorsal series are separated
from the two ventral series by a naked region with transverse ridges. Modified
suckers described by Massy (1925: 212) as ‘15 rows of diminutive suckers placed
2 in a row, in zigzag order, with shallow grooves between them’. This in fact
corresponds with a quadriserial condition in which two series come close

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 237

oae.

=—S
=
=

=

=

SS
— =

SS
<<

>>
Ba
—— SF
=

=>

—S
= —~S
SSSs SASS
PRS
= SSNS

—=>
SS SN
SS >>

So
=<

—
—S

—~ =

yy

YIME
Hi
)

Retwera SOLER

SS Sa)
——
Ss

Fic. 12. Sepia papillata male, A30137.
Ventral view to show wrinkled areas on
mantle and ventral arms.

together (on either side of median grooved section of arm) until apparently
forming single longitudinal zigzag series on each side. Suckers normally
quadriserial distally, but at arm tip (which is slightly attenuated), minute
suckers arranged in oblique rows of eight.

Tentacular club long, bearing small distal suckers in oblique rows of eight.
Four suckers enlarged medially; of these, the middle two extremely large, with
diameters equal to width of sucker-bearing surface of club. Chitinous rings of
large suckers smooth, those of small suckers dentate. Protective and natatory
membranes very well developed; protective membranes not meeting basally,
natatory membrane extending a little beyond club.

Barnard (unpublished notes) discovered that there are two forms of shells

238 ANNALS OF THE SOUTH AFRICAN MUSEUM

belonging to S. papillata. Both shell forms (Pl. 40a—d) broadly oval, tapering
somewhat anteriorly. Dorsal surface rugose, with faint, broad median ridge,
broadening anteriorly. Usually no spine posteriorly, but a broad rounded knob
present. In a few shells, however, this knob continues as small spine lying close
to shell surface and thus directed ventrally. Spine not exceeding posterior
margin of shell. Striated zone long ventrally, with narrow smooth area on either
side of it anteriorly. Striae wavy, with overall /-shape. Median longitudinal
groove distinct. Shape of inner cone different in the two forms, but completely
reflexed and fused to outer cone in both. Outer cone broad laterally. Inner cone
frequently (but more often in form A than B) drawn out posteriorly, almost
reaching edge of shell, as in S. stmoniana.

Shell form A (Pl. 40a, b) with broad inner cone; difficult to distinguish
from S. stmoniana in extreme cases. Phragmocone of this form generally thicker
at anterior region of striated zone, on either side of midline. Shell form B
(Pl. 40c, d) differs in that inner cone narrow and thicker, with limbs forming
narrow raised ridges, as in S. tuberculata. Phragmocone generally thinner than in
form A, but always convex to some extent.

Remarks

As mentioned above, the animals of this species have frequently been
confused with those of S. tuberculata, but they are in fact distinct (see Table 4).

Of the two shell forms, form A (having a broad inner cone) markedly
resembles the shell of S. stmoniana, but the animals differ primarily in that
S. stmoniana has a tentacular club with numerous subequal suckers, and has a
smooth skin, whereas S. papillata has a tentacular club with very unequal
suckers, and has a tuberculate skin. Shell form B of S. papillata closely resembles
the shell of S. tuberculata, but the differences between the two species are as
listed in Table 4. A comparison of the relative dimensions of shell forms A and B
of S$. papillata revealed no statistically significant differences. The present
collection includes six male and four female animals having shells of form A,
and only two males with shells of form B. No differences could be found
between the animals having either shell form, and all undoubtedly belonged to
S. papillata. The significance of these shell forms is at present unknown, but they
may reflect a relationship with S. simoniana on the one hand and S. tuberculata
on the other.

Steenstrup (1875: IV) mentioned two specimens (male and female) of
Sepia which he thought to be S. tuberculata. He illustrated the buccal view of the
female (pl. II, fig. 6), suckers (pl. I, fig. 21) and an arm with minute suckers at
the tip arranged in eight series (pl. I, fig. 20). As S. tuberculata and S. papillata
have so often been confused, we may assume that Steenstrup’s specimens could
belong to either species, since neither tentacular club nor shell were illustrated.
But of these two species only the male of S. papillata has the minute suckers at the
tips arranged in eight series. Hence the male specimen was almost certainly
S. papillata, and presumably the female belonged to the same species.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 239

As mentioned above, Hoyle’s (1910: 265) three female specimens, described
as S. tuberculata, were in fact S. papillata. ‘This can clearly be seen from his figures
(pl. [Va, figs 4-6). In the description, Hoyle said that the arms are about as
long as the mantle, but according to his measurements, they are only half as
long, as also in the present specimens.

The description of S. papillata shells given by Smith (i916: 22) could cover
both S. papillata and S. stmoniana, and indeed he stated that ‘It should be noted
also that the limbs of the inner cones are rather variable. Sometimes, as in the
Astrolabe figure, they do not expand much posteriorly. On the contrary, in some
specimens they spread considerably, and become rather pointed posteriorly.’
The shell in Smith’s figure (pl. II, fig. 1) is that of S. semoniana. Unfortunately,
he does not give a list or the number of specimens examined, but mentions only
that there were shells from Port Elizabeth (Ponsonby, Spencer) and Tongaat
beach, Natal (Burnup). One of the specimens presented to the British Museum
by Spencer was 134 mm long and 60 mm wide.

Adam (1941: 113) lists one of Smith’s specimens (length 134 mm, width
45%* length) from Port Elizabeth in a table of measurements of S. papillata
shells, but does not mention a shell from Port Elizabeth in his list of material
examined.

Adam & Rees (1966) do not include Smith (1916) in the synonymy for
S. papillata, but included in the list of material for S. st¢moniana are three shells
from Port Elizabeth (Ponsonby) and three shells (of which one is doubtful)
from Port Elizabeth (Spencer). These are presumably the same shells as
originally examined by Smith. Of the specimens donated to the British Museum
by Spencer, one shell is recorded by Adam & Rees as being 137 mm long, with
a width of 44% of the length. This must be the shell measured by Smith.
Presumably Adam reconsidered his opinion of 1941 of Smith’s specimens, and
now considers them all to pertain to S. stmoniana. Adam & Rees (1966: 109)
in fact state that the shell figured by Smith (1916, pl. II, figs 1, 2) as S. papillata
belongs to S. s¢moniana.

Tomlin’s (1923: 40) localities listed for S$. papillata were obviously quoted
from Smith (1916), and thus refer to S. stmoniana (with the exception of Quoy &
Gaimard’s locality of the type of S. papillata).

Massy (1925: 211) is the only author to have previously described a male of
S. papillata, but she does not mention the presence of multiplicity of the suckers
on the arm tips. Since this condition occurs in all the present male specimens,
Massy probably overlooked it in hers.

Voss (19625: 251) was mistaken in saying that S. papillata ‘was not reported
from South Africa by either Massy or Robson and may be an uncommon
species’. S. papillata was described by Massy in 1925, and mentioned again in
1928. In fact the species appears to be very common, at least around the
western Cape, since the shells are commonly washed up on the beaches.

* Not 57% (Adam, personal communication).

240 ANNALS OF THE SOUTH AFRICAN MUSEUM

Sepia stmoniana ‘Thiele, 1920
(Pl. 42a, b. Tables 38-40)
Sepia simoniana Thiele, 1920: 436, pl. LII, figs 5-13. Odhner, 1923: 7. Tomlin, 1926: 285.
Voss, 1962): 248, 250. Adam & Rees, 1966: 109, pl. 29, figs 179-182, pl. 42, fig. 254.
Sepia natalensis Massy, 1925: 212, pl. XI, figs 1-11, pl. XIV, fig. 37.

Sepia tuberculata (non Lamarck) Gray, 1849: 101, 102.
Sepia papillata (non Quoy & Gaimard) Smith, 1916: 2a, pl. II, figs 1, 2. Tomlin, 1923: 40 (partim).

Type localities
Simons Bay (S$. semoniana); 25,5 km NE of Bird Island (S. natalensis).

Distribution
Animals: Simons Bay (Thiele 1920: 436) to off Tugela River (S.A.M.).
Depth 14-134 m. |
Shells: | 48 km N of Olifants River mouth (S.A.M.) to Tongaat beach, Natal
(Smith 1916: 22).
Material
S.A.M. Ago127, locality unknown; 1 3
AZOl32, 94 125, To 9O ee 40 msn ©
A30133, Simons Bay; 1 9
A30134, locality unknown; 1 Q
A30135, 35 km S of Tugela River mouth, 116-134 m; 1 9
Ago179, Millers Point, Simonstown; 1 juvenile
Ag30498, ‘Table Bay; 1 shell
A30499, Still Bay; 1 shell
A31239, Sunny Cove, False Bay, 14 m; 2 juveniles
A31251, Castle Rock, False Bay; 2 3
Between Strandfontein and Muizenberg; 4 shells
Simonstown; 2 shells
48 km N of Olifants River mouth; 1 shell
Betty’s Bay; 2 shells
Strandfontein (False Bay); 2 shells
Umngazana River mouth, west Pondoland; 4 shells
Still Bay; 15 shells
Arniston; 4 shells
Die Kelders; 1 shell, discarded
Krom River mouth, Cape St. Francis; 46 shells

Description

Mantle broadly oval, anterior mantle margin somewhat produced dorsally,
entire ventrally. Fins fairly narrow, beginning a few mm behind anterior mantle
margin, rounded and separate posteriorly.

Skin very finely papillose dorsally on head and mantle. Skin somewhat
wrinkled ventrally on mantle in some specimens, but this is possibly due to
preservation. One juvenile (A30179), however, has distinct wrinkled oval

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 241

patches ventrally on mantle, as in S. papillata and SS. tuberculata. Wrinkled areas
usually also present on ventral surfaces of fourth arms. Colour pinkish-brown
dorsally, pale ventrally, with sparse chromatophores, except for two slightly
darker regions on either side of midline, beginning at anterior mantle margin
and fading towards posterior.

Arms subequal in female, about half dorsal mantle length, with third arms
a little longer than others. In males, arms unequal in length, of formula 3.2.1.4.
All arms except ventral pair somewhat attenuated over about distal quarter.
Depth of interbrachial web approximately half arm length, except between
ventral pair, where web absent. Third and fourth arms keeled.

Suckers on all arms quadriserial to tips; suckers largest about one-third
from arm base, then decrease in size to tips, where they suddenly become
minute. On fourth arms, although tips not attenuated, distal suckers also
minute. Rings of large suckers smooth, those of smaller ones very finely dentate.
Suckers of male more globose than those of female.

Left ventral arm of male hectocotylized proximally. About five to nine
normal suckers at base of arm, followed by modified region, where two dorsal
series of suckers smaller than usual and separated from the two ventral series by
naked region with transverse ridges. Suckers in the two ventral series minute.
Distal half of arm normal. According to Thiele’s (1920: 438) description of the
hectocotylus, second longitudinal series of suckers almost lacking in modified
region, although in one specimen this absence was less extensive. But in
the present specimens, second series of suckers complete, though smaller,
and it seems probable that in Thiele’s specimens some suckers were lost after
capture.

Tentacular club very long, occupying about half length of tentacle. Club
bears very many minute subequal suckers, and four (not two, as stated by Voss
19626: 250) somewhat larger suckers, partly concealed by reflexed tip of club.
Protective membranes very well developed, also bearing suckers, except at
outer edge. Rings of small suckers with well defined teeth; those of four larger
suckers at tip of club smooth. Protective membranes joined basally. Natatory
membrane well developed, somewhat shorter than club.

Shell (Pl. 42a, b) broadly oval, tapering a little anteriorly. Dorsal surface
finely granular. At most a faint indication of median longitudinal ridge dorsally,
broadening anteriorly. Longitudinal grooves on either side of median ridge
ill defined. Spine absent posteriorly, but broad knob present, sometimes
tapering into small point, but not attaining posterior margin of shell. Striated
zone long, about two-thirds shell length. A narrow smooth area present on
either side of striated zone. Striae wavy, with overall /\-shape. Deep median
longitudinal groove present ventrally, with convex striated zone on either side.
Striated zone most strongly convex near its anterior end. Inner cone well
developed, reflexed and fused to outer cone. Limbs of inner cone very broad,
usually narrowing suddenly anteriorly, giving characteristic shape. Inner cone
drawn out posteriorly, almost reaching posterior margin of shell. In some shells,

242 ANNALS OF THE SOUTH AFRICAN MUSEUM

however, anterior narrowing of limbs not as sharp, and these shells are difficult
to separate from those of S. papillata form A.

Remarks
As has been pointed out above, shell form A of S. papillata is very similar

to that of S. stmoniana, but the animals differ markedly.

A shell from the Cape of Good Hope, identified by Gray (1849: 102) as
S. tuberculata, belongs in fact to S. simoniana, according to Adam & Rees (1966:
109), who re-examined the shell.

As mentioned above, the specimens described by Smith (1916: 22) as
S. papillata are probably S. szmoniana.

S. simoniana Thiele and S. natalensis Massy are indubitably synonymous, as
has already been remarked by Voss (19625: 251) and by Adam & Rees
(1966: 109).

Sepia angulata n. sp.

(Pls 44d, 45a—-d. Table 41)

Type locality
Bloubergstrand (shells only).

Distribution
Shells: Bloubergstrand to Still Bay (S.A.M.).

Material
S.A.M. A31317, Bloubergstrand (coll. Roeleveld, g November 1969); 1 shell
(holotype)
A31318, Milnerton beach (coll. Kensley, 29 January 1969); 3 shells
A31319, Still Bay (coll. Du Preez, April 1969); 2 shells
A31320, Still Bay (coll. Du Preez, October 1969); 3 shells
A31395, Bloubergstrand (coll. Roeleveld, 9 November 1969); 3 shells

Description

Only shells of this species known.

Shell (Pls 44d, 45a—-d) broadly oval, rounded anteriorly and posteriorly.
Dorsally, part of dorsal shield covering phragmocone clearly distinguishable
from outer cone: the former brown in colour; the outer cone white. Dorsal
surface finely granular, becoming somewhat more coarsely so posteriorly;
entire dorsal surface with iridescent sheen. Faint indication of dorsal longitudinal
rib, widening anteriorly. No posterior spine or knob.

Ventrally, striated zone fairly short, flat from side to side,* except at
extreme lateral edges, which are slightly rounded. Striated zone rises rapidly
from posterior to anterior. Then smooth zone, also flat, descends sharply
towards anterior, forming distinct obtuse angle between striated and smooth
zones (Pl. 44d). A number of shallow grooves radiate over striated zone from

* In one specimen (A31319, length 63 mm) the striated zone is concave.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 243

posterior end in holotype; median longitudinal groove no more distinct than
others. Striae wavy over the grooves, with overall convex shape. In some shells,
however, radiating grooves less distinct than median groove, and in these
shells, striae more regularly convex.

Inner cone completely reflexed, fused to outer cone; posteriorly well
developed, forming transverse ridge. Limbs of inner cone broad, as in S. papillata
form A. Outer cone very broad and deep laterally and posteriorly, giving shell
its characteristic tubby shape. Slight fold present in outer cone on each side,
near posterior end, from inner cone to margin.

Remarks

This species somewhat resembles S. papillata and S. simoniana, but the shell
of S. angulata is somewhat shorter and broader, differing also in that the median
ventral groove is indistinct, and the striated zone and last loculus are remarkably
flat and form an angle at the point of meeting, about halfway along the shell.
S. angulata differs from S. tuberculata in that it is relatively broader, the phragmo-
cone is much thicker, particularly at the anterior end of the striated zone, and
the outer cone is much broader posteriorly. The inner cone is also more strongly
developed in S. angulata.

S. angulata has been so named after the angle formed between the striated
and smooth zones in lateral view (angulatus, L.—angular).

Distinctive characters
I. Shell very broad (50-60% length)
2. Obtuse angle between striated and smooth zones, in side view
3. Both striated zone and smooth zone very flat from side to side

Sepia hieronis (Robson, 1924)
(Pl. 43a-d. Fig. 13. Tables 5, 42-44)
Sepia acuminata Smith, 1916: 21, pl. II, fig. 4 (partim).
Sepia sp. A Robson, 1924a: 13.
Rhombosepion hieronis Robson, 19246: 645, pl. II, figs 9, 11. Massy, 1927: 158.
Sepia hieronis Voss, 19626: 248, 251; 1967: 64. Adam & Rees, 1966: 112, pl. 30, figs 187, 188,
pl. 43, fig. 262.
Type localities
Soo eke, Sta. 2) 99° 03's, 17 49 Kh, Thm
Stay 7 og SoS, Tye aos ZOO I
S15 gel Sih sud OMS ies ean Nd ees i gal
Sta so: Soyo ly Ahi, 2 75 el
Distribution
Animals: West coast: 30° 13'S, 15° 18’E* (Adam & Rees 1966: 112, g) to
west of Slangkop (Voss 1967: 64). Depth 43-457 m.
East coast: Monte Belo, Mocambique (S.A.M.). Depth 431-459 m.
Shells: | Bloubergstrand (S.A.M.) to Tongaat beach, Natal (Smith 1916: 21).

* Not 30° 09’S, 19° 02’E, as stated by Adam & Rees.

244. ANNALS OF THE SOUTH AFRICAN MUSEUM

Material |
S.A.M. Agooo, 80 km N 42°W of Lions Head, 422 m (det. A. L. Massy);
I g, 1 juvenile, in poor condition
A29728, west of Slangkop, 250 m (det. G. L. Voss); 1
Ago145, locality unknown; 1 2
Ago146, S 76°W of Lions Head 45 km, 257 m; 1 g
A30260, N 48°W of Lions Head 80 km, 422 m (det. A. L. Massy);
2 specimens in poor condition
A30268, N 51°W of Lions Head 75 km, 321 m (det. A. L. Massy);
I g in poor condition
A30563, 34° 04'S, 17° 45 E, 275 m; 1 g
AS12439) 29° 15 9, 17 108 1263 mia G
A31405, Monte Belo, Mocambique, 25° 35'S, 33° 30’E, 431-455 m; 1 4
A31406, Monte Belo, Mocambique, 25° 35'S, 33° 30’E, 440-459 m; 1 9
A31407, Monte Belo, Mocambique, 25° 35'S, 33° 30’E, 459 m; 1 ¢
Mossel Bay (S of Slangkop on Cape Peninsula); 19 shells, broken
Olifantsbosbaai (Cape Point Reserve); 1 shell, broken
Millers Point, Simonstown; 1 shell, broken
Bloubergstrand; 2 shells, slightly damaged

Description

The specimens previously described by Massy (1927: 158) now very
shrivelled and cannot be measured or properly examined. Remaining specimens
(seven males and one small female) in good condition.

Mantle ovate, anterior mantle margin produced to fairly sharp point
dorsally between eyes in males, less strongly produced in small female (about
half as much as in males). Ventrally, anterior mantle margin entire in males,
emarginate in female. Fins narrow, beginning a few mm behind mantle margin,
rounded and separate posteriorly.

Colour generally mottled reddish-brown dorsally (but in two males and in
female chromatophores contracted to small spots on pale background). Mantle
ventrally pale in middle, somewhat darker laterally towards fin bases, where
chromatophores more dense. Skin sparsely papillose dorsally on head and
mantle in female; smooth in males.

Arms generally shortest dorsally, longest ventrally, having formula 4.3.2.1,
but varying somewhat in different specimens (female measured by Adam &
Rees 1966: 113, had subequal arms). Interbrachial web present, except
between ventral arm pair, and is deepest between arms II and III.

In males, dorsal arms attenuated over about distal third, bearing biserial
suckers gradually diminishing in size towards tips, where suckers minute.
Sucker arrangement on lateral arms very characteristic. On dorsolateral arms,
suckers of basal two-thirds biserial, sometimes becoming somewhat irregularly
arranged or even quadriserial for a few rows on middle third of arm, then three
to five pairs of greatly enlarged suckers. Of these, middle ones have diameter

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 245

ID SOGS>..
>

(9

Fic. 13. Sepia hieronis male, A30563. a. Ventral view; b. dorsal view; c. hectocotylus; d. sucker
from the base of a lateral arm, diameter about 1 mm; e. one of the suckers distal to the enlarged
suckers on arms II to IV, diameter about 1 mm; f. one of the enlarged suckers on the distal
part of the hectocotylus, diameter about 1 mm; g. enlarged sucker from one of the lateral arms,
diameter about 2 mm;; h. oral view of sucker from the tentacular club, diameter about 0,2 mm.

about twice that of basal suckers. The more distal enlarged suckers of peculiar
shape (Fig. 13e), being elongated and attached by long stalks. Distal suckers
minute, arranged quadriserially. Arrangement of suckers on ventrolateral arms
generally the same as that on dorsolateral arms, but a number of quadriserial
rows of suckers present on middle third of arm. Enlarged suckers not as large
(relative to those at arm base) as on dorsolateral arms. On right ventral arm,
suckers on basal third arranged as on lateral arms. Quadriserial suckers present
on middle third, followed by three to eight biserial rows; then about five
biserial rows of suckers larger than those immediately proximal to them, but
not as large as basal suckers. Minute suckers at tip of arm also biserial, unlike
those of lateral arms.

Left ventral arm of male hectocotylized. One to three normal suckers at
base followed by modified region, extending approximately half way along arm,
which is transversely wrinkled and bears minute suckers laterally: seven pairs of

246 ANNALS OF THE SOUTH AFRICAN MUSEUM

minute suckers on dorsal border; ventrally, the two series of suckers have moved
together to form a single series of 14 suckers situated on extreme ventral edge of
arm. Suckers biserial distally; the more proximal of these suckers (9-10 pairs)
large, those on the tip minute.

In the female, all arms attenuated distally. Skin wrinkled basally on arms,
next to sucker-bearing surface. This apparently not due merely to preservation,
and has not been observed in any other Sepia species. Suckers on all arms
biserial, and diminish gradually in size from base to tip. No enlargement of
suckers on middle third of arms as found in males (Adam & Rees 1966: 113,
however, found that females may also have enlarged suckers on the arms.
See below).

Tentacular club small and recurved, with numerous subequal suckers in
transverse rows of about eight. Rings of suckers without teeth, but covered with
numerous small knobs. Protective membranes separate proximally. Natatory
membrane extends along tentacular stalk for a distance a little less than half
club length.

Shell (Pl. 43a-d) ovate, acuminate anteriorly, narrowing and rounded
posteriorly. Dorsal surface usually pink in colour, finely granular posteriorly.
Chitinous margins broad in smaller shells. Distinct rounded median rib present
dorsally, broadening somewhat anteriorly. Rib limited by distinct lateral
grooves. Spine absent posteriorly, but pronounced rounded knob present.
Behind this, shell bends sharply towards ventral. Striated zone long, with
narrow concave region laterally, raised and flattened medially. Striae convex to
angular (/\-shaped), but with notch on either side, corresponding with point of
meeting of raised middle region of striated zone with concave lateral areas.
Median groove usually very faint, sometimes almost indistinguishable, but in a
few shells is quite marked. Outer cone broad and deep.

TABLE 5. Sepia hieronis: a comparison of the relative dimensions (as % shell
length) of the shells from the east and west coasts of southern Africa.

West coast East coast
N Mean Range N Mean Range

Width
Thickness
Striated zone

39,1 38,3-40,4
11,0 9,6-11,7

72,2 79,0-73, I

42,4 49,7—-45,7
14,3 13,0—-15,2
65,6 64,8—-66,7

op
oo 6 OO

Shells from east and west coasts show some marked differences, although
having an overall similar appearance. In shells from west coast (Pl. 43a, b),
inner cone not as well developed posteriorly, and with narrow limbs. In eastern
shells (Pl. 43c, d), on the contrary, inner cone well developed posteriorly,
reflexed and fused to outer cone, and with broad limbs. Eastern shells wider and
thicker, have shorter striated zone, and median region of striated zone is more
markedly raised than in western shells (Table 5). Yet differences between the

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 247

two shell forms are those of degree, and do not suggest different species (except
perhaps differences of inner cones). No differences could be found between
animals from the two coasts. Differences in relative dimensions of shells from
east and west coasts possibly due to size, since shells from east coast generally
smaller than those from west coast. Alternatively, differences may be due to
contrasting environmental conditions on east and west coasts.

Remarks

Smith (1916: 21), in the original description of S. acuminata, mentioned
and figured (pl. II, fig. 4) a shell from Tongaat beach, Natal, which differs
from the other examples of S. acuminata. From the description and figure, this
specimen clearly belongs to S. heronis, and differs from S. acuminata in that there
is no spine, but only a posterior knob (just visible in Smith’s figure), the outer
cone is well developed posteriorly, the striated zone is more raised, and is
divided into three distinct regions: one flat median and two concave lateral
areas.

The arrangement of the suckers seems to vary considerably in S. /eronis,
since every author describes them somewhat differently. But all are agreed that
there are a number of pairs of enlarged suckers on the lateral arms of the males.

Robson (19245: 646) described the hectocotylus as having one series of
suckers ventrally in the modified region. This in fact corresponds to two series
which have moved together to form a single series, as mentioned above. In one
of the present specimens the hectocotylized arm is contracted, and the ventral
series has become a zig-zag line, illustrating its origin from two longitudinal
series of suckers.

According to Massy (1927: 159) the distal suckers on the lateral arms are
biserial, but in the present male specimens they are quadriserial.

Adam & Rees (1966: 113), in the description of the male, mention that the
suckers of all the arms are proximally quadriserial, and imply that those of the
dorsal arms are quadriserial throughout. In the present males, the suckers of
the dorsal arms are biserial from base to tip, and on the lateral arms are
biserial over at least part of the proximal half, although these conditions are
sometimes rendered less clear due to contraction of the arms.

In one female (MLd 61 mm) described by Adam & Rees (1966: 113) the
dorsolateral arms have four pairs of enlarged suckers in the middle, minute
suckers quadriserially arranged distally; ‘the tips of all the arms, except the
dorsal ones, seem to have the quadriserial arrangement of the suckers, but we
are not sure that this is not due to contraction and that, in fact, all the suckers
are biserial’. ‘The female syntype has no enlarged suckers on the dorsolateral
arms. In another female the enlarged suckers are well developed on the lateral
and even on the ventral arms. In the present female (A30145, MLd 33 mm)
all the arm suckers are biserial, and none is enlarged. It is possible that the
sucker enlargement only develops as the animals grow to maturity.

Previously, ‘animals of S. hieronis were known only from the west coast of

248 ANNALS OF THE SOUTH AFRICAN MUSEUM

South Africa, from 30° 13'S, 15° 16’E to west of Slangkop (Robson 1924a: 13,
1924): 645; Massy 1927: 158, 159; Voss 19625: 251, 1967: 64; Adam & Rees
1966: 112). Adam & Rees (1966: 53, 112) found one shell of S. Aieronis with
specimens of S. acuminata from Stations 95 and 103 (Robson 1924a: 12, 13,
1924b: 643). ‘These localities are approximately 30°S, 31°E, which is off the
Natal coast, near Durban. Smith’s (1916: 21) shell from Tongaat beach, Natal,
was the only other record of S. hieronis from the east coast.

The present specimens from Monte Belo, Mocambique, are the first
records of S. /ieronis animals from the Indian Ocean. These specimens
undoubtedly pertain to S. fzeronis, although the shells show some differences
from those found off the west coast. It is strange that neither the shell nor the
animal of S. jieronis has so far been recorded between Cape Point and Durban.

Sepia insignis Smith, 1916
(Pl. 44a-c. Fig. 14. Tables 45, 46)
Sepia insignis Smith, 1916: 25, pl. II, fig. 10. Tomlin, 1923: 41. Voss, 1962b: 248. Adam & Rees,
1966: 114, pl. 31, figs 189-191.
Type locality
Tongaat beach, Natal (shells only).

Distribution
Animal: 34° 15'S, 18° 47’E (False Bay) (S.A.M.). Depth 42 m.
Shells: | Bloubergstrand (S.A.M.) to Tongaat beach, Natal ies 1916: 2 5).

Material
S.A.M. A30486, locality unknown; 1 shell
A31236, Kommetjie; 1 shell
A31241, Simonstown; 3 shells (1 broken)
A31247, 34° 15'S, 18° 47’E, 42 m; 1 9
A31248, between Strandfontein and Muizenberg; 1 shell
Still Bay; 1 shell
Bloubergstrand; 1 shell, broken

Description

Female specimen rather badly preserved, with mantle compressed
laterally.

Mantle elongate oval, anterior margin produced dorsally, emarginate
ventrally. Head short and broad. Fins rather wide, beginning a few mm behind
anterior margin of mantle, rounded and separate posteriorly.

Skin sparsely papillose dorsally with concentric chromatophores on head
and mantle. Mantle very wrinkled ventrally, but this may be due to preserva-
tion. Colour dark purple dorsally on head and mantle; chromatophores on
mantle concentrated mid-dorsally, less dense laterally towards fins. Fins pale,
with a few scattered chromatophores both dorsally and ventrally. Mantle pale
mid-ventrally, with band of chromatophores on each side along bases of fins.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 249

Fic. 14. Sepia insignis female, A31247. a. Dorsal view; b. ventral view; c. right tentacular club.

Arms I and II subequal in length, about one-third MLd, arms III a little
shorter, arms IV longer, very well developed, and wrinkled ventrally. Shallow
interbrachial web present, except between ventral arms. Dorsal arms attenuated
distally over about half arm length. Arms III and IV strongly keeled. Protective
membranes well developed on all arms, and wide enough to meet over inner
surface of arm.

Arms unfortunately rather distorted, and some suckers lost, especially from
delicate dorsal arms. Suckers on dorsal arms biserially arranged from base to
tip. On arm II, five pairs of suckers basally followed by very oblique quadri-
serial rows to arm tip. On arm III, two pairs of suckers basally, then one or two
to fill in gap before obliquely quadriserial suckers, which continue to tip.
Suckers on ventral arms very obliquely quadriserial, with one row of three
suckers basally. Quadriserial rows so oblique that suckers appear to be almost
biserial. None of the suckers enlarged; size of suckers decreases gradually from
base to tip on all arms. Suckers globose, with nodular rings which have
irregular but not toothed edges.

Tentacular club fairly small, bearing a number of subequal suckers in
transverse rows of about eight. Suckers have large rings with nodular surface
and smooth edge. Natatory membrane well developed, continuing a little
beyond club. Protective membranes well developed; ventral membrane curves
around base of club, but not meeting dorsal membrane, which ends just before
base of club. Upper surface of club with a number of transverse rows of
chromatophores.

Shell (Pl. 44a-c) elongate, fairly narrow, sharply acuminate anteriorly,
then about same width over approximately two-thirds of its length; rounded

250 ANNALS OF THE SOUTH AFRICAN MUSEUM

posteriorly. Posterior spine absent. Shell a pale pink colour dorsally. Mid-dorsal
longitudinal ridge present, but not limited by lateral grooves. Ridge ends
posteriorly in blunt knob, whereafter shell curves sharply to ventral. Striated
zone long, narrowing markedly posteriorly; striae /\-shaped between limbs of
inner cone. Well defined median longitudinal groove present over striated zone,
continuing along last loculus, where it is less distinct. Inner cone well developed
posteriorly, completely reflexed and fused to outer cone. Inner cone curved
over sides of striated zone laterally, as in S. burnupi, and limbs of inner cone lie
on phragmocone. In smallest shell, limbs of inner cone much nearer edges of
phragmocone than in larger shells. It would appear that limbs of inner cone
move towards middle of shell in older shells. Outer cone broad laterally.

Remarks

Until now only shells of S. znsignis were known. The present female was
found in the collection of the University of Cape Town Ecological Survey, and
is now lodged in the collection of the South African Museum.

The shell from Bloubergstrand (PI. 44c) must have measured about 55 mm
when whole, and is the largest shell known of S. insignis.

This species seems to be related to S. hieronis. The shells of the two species
show some resemblance, and the animals are similar in that both species have
subequal tentacular suckers and biserial suckers on the dorsal arms.

Sepia robsoni (Massy, 1927)
(Tables 6, 50)

Rhombosepion robsoni Massy, 1927: 159, pl. XVII, figs 1-8.
Sepia robsoni Voss, 1962b: 248. Adam & Rees, 1966: 120, pl. 46, fig. 279.

Type locality
Hout Bay, 17-37 m.

Description

Only one specimen (male) ever caught, and was not available for examina-
tion. Following description after Massy (1927: 159) and Adam & Rees
(1966: 120).

Animal small. Mantle broadly oval, anterior mantle margin dorsally
slightly convex (Adam & Rees 1966: 120), very slightly produced in the centre
(Massy 1927: 160), ventrally deeply emarginate. Fins wide, beginning three mm
behind mantle margin, separate posteriorly.

A few tubercles present dorsally along outline of shell and on head. Fleshy
ridge present on either side on mantle ventrally, near fin bases, as in S. typica,
but pores absent. Flesh-colour to pale brown, with minute dark chromatophores
dorsally, extending partly on to fins. Ventrally with a few chromatophores
along fin base.

Arms subequal in length, with ventral arms slightly longer than dorsal
ones. Interbrachial web well developed between arms I to III, attaining half

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 251

arm length, lower between arms III and IV, absent between ventral pair.
Suckers globose, with smooth rings; skin adjacent to rings grooved; suckers
biserially arranged, none enlarged.

Dorsal arms with finger-like tips, devoid of suckers, and with eight pairs of
suckers proximal to naked tips. Dorsolateral arms with eight pairs of large
subequal suckers and a few small suckers distally. Ventrolateral arms with nine
pairs of large subequal suckers. Right ventral arm with seven pairs of large
suckers, of which suckers of ventral series larger than those of dorsal series, and
much larger than those of other arms. Left ventral arm hectocotylized over
basal three-quarters of its length, bearing ten pairs of minute suckers on modified
region.

Tentacular club crescent shaped, bearing about 53 more or less subequal
suckers, in transverse rows of four to six, median suckers being slightly larger
than others. Rings of tentacular suckers papillose (? nodular), dentate on at
least part of the ring (Massy 1927: 160). Natatory membrane very broad,
extending beyond base of club for a distance equal to about half club length;
dorsal protective membrane very broad at base of club.

Shell completely chitinized (? decalcified), in..poor condition. Anterior
part strongly acuminate, resembling that of S. Azeronis.

Remarks

According to Massy (1927: 159), Sepza robsoni resembles S. hieronis, but
differs from the latter in having grooved suckers (smooth rings) biserially
arranged on all the arms, and none is enlarged in the male. “The shell somewhat
resembles that of S. insignis, but the latter is even more acuminate anteriorly
and seems to be much wider in its posterior part?’ (Adam & Rees 1966: 121).

S. robsoni also resembles S. dubia (for comparison, see Table 6) in that all
the arm suckers are biserially arranged, the mantle is very broad, and the
ventral mantle surface has fleshy keels without pores. The hectocotylus is like
that of S. typzca, and the interbrachial web is well developed. The tentacular
club is exactly like that of S. typica (Adam & Rees 1966: 121). Sepia robsom
differs, however, from both S. typica and S. dubia in its shell, which presumably
has the phragmocone covering almost the entire dorsal lamella (since it is
compared with those of S. hzeronis and S. insignis) and in the dorsal arms with
the tips devoid of suckers. In this latter feature S. robsont resembles S. fauret.

Sepia fauret n. sp.
(Figs 15, 16. ‘Tables 6, 50)
Type locality
S 14°E of Cape Seal 88 km, 168 m.
Material
S.A.M. A30144, S 14°E of Cape Seal 88 km, 168 m; 1 9 (holotype)

252 ANNALS OF THE SOUTH AFRICAN MUSEUM

Description

Animal small. Mantle broadly oval, anterior mantle margin convex
dorsally (not markedly produced), emarginate ventrally. Fins narrow, beginning
a few mm behind mantle margin, separate posteriorly.

Colour uniform brown on head and mantle dorsally, with small darker
spots on arms. Skin densely papillose dorsally. Fleshy ridge present on either
side of mantle ventrally, near fin bases, as in S. typica, but pores absent.

Fig. 15. Sepia faurei female, A30144 (holotype). a. Dorsal view and b. ventral view.

= =e

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA =—_.253

Arms subequal in length; interbrachial web high dorsally, attaining about
half arm length, but low between ventrolateral and ventral arms, and absent
between ventral pair. Suckers globose, without teeth on chitinous rings, but
these covered with numerous raised knobs. Suckers on all arms biserially
arranged. Distally all arms except ventral pair attenuated over about half arm
length; suckers on attenuated portion (about six pairs on arm I, 17 pairs on
arm II, and 16 pairs on arm III) minute, and protective membranes very well
developed. Tips of attenuated distal half of dorsal arms finger-like, devoid of
suckers and protective membranes (Fig. 16a). Six to nine pairs of subequal
suckers proximally (nine pairs on arm I, six pairs on arm II, and seven pairs on
arm III), of which last pair somewhat smaller. Ventral arms not markedly
attenuated distally, bearing eight pairs of subequal suckers proximally, and
about 12 pairs of minute suckers on tips.

Fig. 16. Sepia faurei female, A30144 (holotype). a. Detail of distal part of dorsal arm. Position of
missing suckers indicated by dotted rings. b. Right tentacular club. c. Ventral view of shell
(incomplete).

Tentacular club broad, slightly recurved, bearing 33 subequal suckers in
transverse rows of four to six (Fig. 16b). Median suckers a little larger than
lateral suckers. Sucker rings not toothed, with nodular surface. Dorsal protective
membrane well developed, separate from ventral membrane proximally.
Natatory membrane broad, continuing along tentacular stalk for a distance
equal to club length.

Shell (Fig. 16c) not calcified, very thin, as in S. typica, broadly ovate,

ANNALS OF THE SOUTH AFRICAN MUSEUM

254

SUIIe [[@ UO [eLIOsSIq fasoqoT+ [eliosiq pue |[euls 19y)e yy

AJ sure u99M30q

qUo}SIXo-UOU PU AT puUe [I] sulIe
U29M}9q JIMOT YONUE ST 7I fsuLIe oy}
dn Aemyjyey Suryovoi ‘TTT pue [I ‘T
sulre UsaMjoq Ysty AIOA SI qoM OUT,

AI suite us9Mj0q JUISqe

SI pue SA] pue JJ] swe us9MjJ0q JOMOT
UDAD SI HI ST] pue [] ‘J swe usaMmj0q
Yjsus] Use Jey UTe}7e JOU s2Op qa

yisuz] ut yenbaqns suy yisug] ut penbaqns sunny

sarod ynoyiiM ynq ‘sn71dqq "Fy
UI SB ‘UISIVUI 19]NO 9} 0} JoT[e1ed
pay AYsay YI? & YIM “yIOOUTg

sazod ou ynq ‘soseq uy
24} IvaU apis Joyj}10 UO aBpis AYsoy VW

UISIVUL 9TJULUT 94} IeoU ATIOLIO}Ue
ou0 pily} & pue ‘ou, ULIpPsUT oY} Jo
apts yove uo ouo ‘oeTIded poayoe1}U09
Jo soyojzed [eao OM} oe J19y} aU
oy} UO AT[esIOp-prpy ‘souvireadde
asosni AIDA & SuNVoID ‘oviIded
punos pooeds-]jam YIM paisA09 sure
pue peoy ‘opUeU Jo s0RyINs [esI0g

oeyided
[[ews yIIM posoAod Ajasuap sure
pue suru ‘peoy uo ATesi1op uty

ojyeursieuls A[dsop

o}euUIsi1eUS AjjerUIA ‘xaAUOD AT[TesIOG Ajjeaquaa ‘xaauoo ApYsys ATTes1I0q

WI 1% wu LI
Jew} o[eWo}
woanvf’“s Diqnp *S

SUIIe OU} Te
uo posuesse Ajjeliosiq fasoqoys sisyong

AI suze

us0Mj0q Suryovy] Ajo}o,du09 st

pue Ia9MO] YONUI sI I AT pue T]] suse
u99M}0q SJ] pue I] ‘[ sulre usaM10q
sue oy} dn Avmyyey soyovor qom ou T,
yisus] ur penbaqns sway

soiod Aur jo 2081}

OU SI 9104} nq ‘osprs AYsop oy} YIM
sng1gQ] snigasiwmapy JO yey} a¥I] YOn|T

pesy 943 UO pur TOYS
94} JO 9UTTINO oY} SuoTe sopo19qn} Moy VW

oyeurs1euls A[dsop
Ayjeajuaa ‘xoAuo0d ApYSYs ATTes10q

wut L1
oTeul

1u0sqos *S'

SIOWVONS WAY

qom
[erpoesq19jyUy

sy}sus] WAY

ooeyins
opueW [enUusA

aejjided jesi0og

UISIVUL
opueU IOI0Uy

PIW

x9g

‘ds -u iainvf‘s pure ssoy 29 Wepy vignp *s ‘(AsseyAq) 2Uosgos vIdag Jo s1d}DeIeYI UTeUT ay} Jo UOSIIedUIOD *g ATAV],

255

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

foa}4a00 panuyuor

Sue] qnyd sy} 0} enba

QOUL}SIP B OJ [VIS OY} Suoye sonunuos
pue ‘proiq oueiquiour A10}e}eNy *ATTeU
-Ixo1d 9U0 [v1]UDA 94} WOIy o}eIedas
‘padoyaaap [Jam sueIqusUT 9AN99}01d
[esioq “souUO [eUTSIeUI oY} ULY}

I9ZIL] IIT BV squo sppprur sy} ‘g—P jo
SMOI ISIDASULI} UT sIayons [enboqns &&

souvIquioul

2A1]}99}01d pue silayons Jo ploagp

pue oyI]-1osuy o1e sure [esIop 9y} Jo
Jyey [eastp poyenuaye oy} jo sdy ayy,

YASUI] sit Jjey ynoqe 10x qnyd jo oseq
puodsq Sutpus}xo ‘podoyaaap [jam
ouviquiow A10}v}JeNT “qnyo Jo oseq
7@ QUO [eIJUIA OY} WOT poyesredas

pue opm suviquiout 39AN909}01d
[esioq *(uoT}eorUNUAWIOD [euOsiod
‘9961 s0j;Av yz) soBre] APYSITs o1e

S9UO ULIPSUT 9Y} YITYM Jo ‘s1ayons HS
Joey UL O1¥ IBY, *(QQ61 sody

® wWiepy) sores jeurpnytsuoy S 10 F
UI posuvsiie sIoyons penbaqns oynutlyy

(uonvoruNUIUIOS

jeuosiod ‘g961 s0,Ae 7)

sd oy} Jo ww Po uIYyIIM 0} snUTUOD
SUIIe [esIOp IY} UO sIay¥ONs YJ,

qnyo
34} JO yysUZ] oy J[ey ynoge I0j qnyjo jo
aseq puodoq Surpuaj}xo pue peoiq AI9A
oueviquioul A10}2JeNJ “Qnyjo Jo aseq 3e
peoiq AroA DUBIqUIOUL 9A1}909}01d [esIOC
"}SO1 SY} ULY} SIOYONS JOSIe] DAVY SMO
uvipou & ynoqy ‘Q—P JO SMOI 9SIOA
-SuUeI] UI posuere sioyons 6S ynoqy

Sutuoyory} Aue uous

jou op Aoy J, *.s19yx9Ns JO prloaop

‘sd oyl]-1osuy savy sure [esiop

oui, ‘uowIoeds 9y} pourTuIexo-31 OYM
*(9961) sooy 29 Wepy 0} SuIpPIODDy
‘sdry 94} 3 19puajs reodde

sue [esIop oy} (% ‘I sBy ‘TTAX ‘[d)
soinsy s.Asseyy ut yng *(L61 Asseyy)
qjuepriooe 0} onp 2q s10JoI19y) AvU d0Ue
-1eodde oy], ‘UOT}eULIOJ JO 9sINOD 9Y} UT
OQ JYSIUI s1oyONs JI se popjoj 9q 0} Uses
SI 30BJIMs 9} AT]V191"] PoAIOsqoO Udy AA,
*sIgyYONs JNOYIIM sqouy popunos ur pus
pue pousyory} o1e suonsod [e}stp sy,

qnyo repnoejUa T,

SULI® [eS1OC

ANNALS OF THE SOUTH AFRICAN MUSEUM

256

a a a a ee ee re

SUIVUIII 9UOD JOUUT 94} JO
UONVOIPUL OU pue 4so] UI9q sey T[IYs
3y} Jo y1ed s01193s0d ayy AjayeuN,I10;UA,
‘pepunos Apysiis pue ‘osioA

-SUBI}] DIV JLII}S SY], *eore [eUIsSIeUI
YOOUIS & SI 9194} DUOZ PoqerI}s

94} JO apis 19y}10 UG ‘sUODOWIseIYd

94} JO UOTdad [e1]UI9 9Yy} sotdnd00

QUuoz pr}elys oY], ‘uoooWwse1Yd

24} JO YISus] OY} J[ey ynoqe paizA00
Ajqeqoud ‘1api0q 10119}Ue xaAUO0D

@ sey YIM ‘sn[nsoy yse'T *e]jaurey
[estop 94} Jo y1ed s0119}ue 9Y} 19A00
jou soop pur ‘odeys ur re~nsuers}y
Ajysnor suocousesyd [eUsA oy J,
“A[IOLIO}U 9JeUTUINIe yeEYMOUIOs ‘[eAo
A[pvoaq st 3] *snouryrys Aya}a;duN09 {ays

vainof *s!

au0d J9UUT 94} AT2}0;duI09

SpUNOJINSs 9UOD J9jNO prog 9UT,
‘uolssoidap MoTeys AOA & spunosins
yxed r10110}S0d s}1 {aud 193n0

24} 0} pasny AjJa}07dur109 oie 1eY}
‘sqUII] peorqg JoYyyVeI YIM ‘Joyooye

UI INO[OO UI YsTUMOIq ‘ouUOd JOUUT
JOUT}sIp & st atoyy, “AAVM ATYSITS
ATuO Sutoq Yysre1js Apreou ‘astaAsueIy
‘pooeds Ajapim ore “roquinu ut G1
ynoqe ‘ovII}s oY], ‘svore [eUIs1eUI
“‘yjoouIs ‘peoiq om} Aq poxuey st pue
QUOZ IOUUL 3JOYM 9y} Jo uoTI0d paryy
[e4jus0 943 Ajuo sotdnd90 varie poyerys
9} ‘sn[NIo] IseT VY} puTyog ‘vose
PPIs SY} JO YSU] oy} J[ey ynoqe
sjUssoidor sn{nooy Ise] SITY, ‘UIs1eUr
JOII9}Ue 78 }SOpPIM ‘prozodes} sn{ns07
sey “snoid@] FZ ut se ‘y)8u9q [JOYS

oy} Jjey uey) s10u 97331] & sotdn990 71
‘yied Surpunosms oy} uey) s9ys1y
Apysys st pue odvys [eou09 passsAor
e sey yred [e1,U09 oy], ‘:podeys-uoods
pue savou0d Ajsuods d0VjJANS [eI]UI A,
‘guids ON ‘usoyyed aye~nNoyOI & YIM
‘shOoTeO[¥9 SI JOVJANS [esIOp ITOYM 9}
A[IVONJ *SUISIeUL 9Y} IeoU 19972] SI
ynq s[pprur oy} ur odevys-ow0p Moy e
SI }] ‘purysq popuno. Ajpeorq st pue
snngG *F] Ui uey} Os ssoy ynq ‘Ay1o1I9}Ue
oyeurumnoe ‘feao Ajpeoiq T[IYS

pignp “s)

siuosary *S' JO

Jey} So[quisso1 puv o}yeuUTUINOe AjSUO0.1}s
st yazed JO110}Uv s}J “UONTpUOD 100d ut
‘(payroyesap ¢) pezrunryo Ayo}o7du0,+)

1UO0sqoL *S'

IPs

(panunuos) 9 ATaV J,

> gE ae

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 257

somewhat acuminate anteriorly. Posterior end damaged. As in S. typica,
phragmocone not covering anterior part of dorsal lamella. Anterior border of
last loculus convex. Striae fewer and last loculus longer than in S. typica.
Striated zone occupying central region of phragmocone; smooth marginal area
present on either side of striated zone. Striae transverse, slightly convex in
shape. No indication of inner cone remains.

Remarks

Sepia faurec resembles S. robsoni in the absence of suckers distally on the
dorsal arms. It resembles both S. robsont and S. dubia in that the mantle is very
broad and has fleshy keels, without pores ventrally, and the arm suckers are
biserially arranged. ‘The three species are compared in Table 6.

The shell of S. faurez is similar to those of S. typica and S. dubia, being very
thin, with the phragmocone not covering the entire dorsal lamella. In S. typica
and §. dubia, however, the anterior border of the phragmocone is not parallel to
the corresponding sides of the dorsal lamella, whereas in S. faurei it is more
nearly so, and the phragmocone is somewhat longer in the latter species.

Sepia fauret, showing relationships with both S. robsoni and S. dubia, seems
to represent an intermediate link in the transition from Sepia to Hemisepius.
On the one hand it is related (by virtue of the dorsal arms) to S. robsoni, which
apparently has a Sepza-like shell with the phragmocone covering almost the
entire dorsal lamella anteriorly, and on the other hand S. faurei is related to
S. (Hemisepius) dubia and S. (Hemisepius) typica with a Hemisepius shell, in which
the phragmocone is much shorter than the dorsal lamella.

Distinctive characters

1. Tips of dorsal arms finger-like, devoid of suckers and protective
membranes
Suckers biserially arranged on all arms
Mantle very broad, with fleshy keels ventrally
Skin densely papillose dorsally on head, mantle and arms
Shell with phragmocone considerably shorter than dorsal lamella, but
with anterior margin of phragmocone convex in shape

The holotype of Sepia faurei was collected during one of the cruises of the
Cape Fisheries survey vessel Pieter Faure (P.F.14290, 19 oo 1902), after
which this species has been named.

2 Soe

Sepia (Hemisepius) typica (Steenstrup, 1875)
(Fig. 17. Tables 7, 47-49)

Hemisepius typicus Steenstrup, 1875: 468, pl. I, figs 1-10, pl. II, fig. 1. Hoyle, 1886: 26, 217;
1912: 281. Gibbons, 1888: 202. Smith, 1903: 356; 1916: 25. Chun, 1915: 411, figs 33, 34.
Massy, 1927: 164. Thore, 1945: 50, fig. 1. Voss, 19625: 248, 252; 1967: 64.

Hemisepion typicum: Rochebrune, 1884: 78, pl. 3, fig. 1.

Rhombosepion sp. A Massy, 1927: 161.

Hemisepius typicus var. chuni Thore, 1945: 50.

Sepia (Hemisepius) typica Adam & Rees, 1966: 117, pl. 32, figs 192-195, pl. 33, figs 196, 197.

258 ANNALS OF THE SOUTH AFRICAN MUSEUM

Type localities
Table Bay; St. Francis Bay (var. chun).

Distribution
Animals: Saldanha Bay (Hoyle, 1912: 281) to Cape Natal, W by N, 10 km
(Massy 1927: 161). Depth 2-156 m.
Material
S.A.M. A889, locality unknown; 1 g in poor condition
A29608, south side of Schaapen Island, Saldanha Bay, 4 m; 3 9
A29717, SSE of Ystervarkpunt, 92 m (det. G. L. Voss); 7 g, 6 2
A29783, Saldanha Bay, 7m; 15 dg, 31 9
A30176, S 34°W of Cape Infanta 30 km, 84 m; 5 ¢ (of which two in
poor condition), 2 2
A30177, 34° 14'S, 22° 23’E (Mossel Bay), 60 m; 5 g, 4 2
A30269, S 16°W of Cape Point lighthouse 16 km, 156 m; 1 4, 4 9, all
in poor condition |
Ag30484, locality unknown; 1 shell in poor condition

Description

Animals small; largest specimen in present collection being a female of
MLd 25 mm; largest male has MLd 21 mm.

Mantle very broadly oval, almost as wide as long. Anterior mantle margin
convex dorsally, emarginate ventrally. Ventral surface of mantle bearing two
fleshy ridges near fin bases; each ridge with a number of pores anteriorly
(most commonly 10-12 on each side, but the number may vary between five
and 15. Number of pores not always the same on both sides). Ridges becoming
less distinctive posteriorly. No evidence of longitudinal groove linking pores, as
described by Steenstrup (1875: II).

Head short and broad. Fins narrow, beginning a few mm behind anterior
mantle margin, fused posteriorly.

Colour dark reddish-purple dorsally on head, arms and mantle, with
darker diamond-shaped region mid-dorsally over shell. Two pale round
tubercles with concentric chromatophores present in middle of darker region.
Fins pale dorsally with scattered dark chromatophores, most dense near fin
bases. In some cases short transverse orange bands present on fins. Colour pale
ventrally, except for darker colour of ridges bearing pores, and sparse chromato-
phores between ridges and fin bases; a few chromatophores present laterally on
funnel.

Skin very sparsely papillose dorsally, a few papillae around eyes being the
most marked. Mid-dorsal tubercles on mantle very flattened.

Arms subequal in length, fairly short (about 40-5094 MLd). Interbrachial
web deep, attaining half arm length, but absent between ventral arms. All arms
triangular in cross section, but only well developed ventral arms keeled.

Suckers on all arms biserial, globose, not flattened and disc-shaped, as
described by Steenstrup (1875: II). Chitinous rings smooth edged, with nodular

;

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 259

surface. In female, size of suckers decreases gradually from base, but suckers
suddenly become much smaller at tips of arms; about six to eight pairs of
minute biserial suckers present. In male, one to four pairs of suckers near tips
of arms I to III enlarged. No enlargement occurs on ventral arms, but in some
cases suckers of dorsal series of right ventral arm larger than those of ventral
series. Thore (1945: 52) found, however, that suckers of 12th to 18th rows on
right ventral arm were enlarged in his specimens from Table Bay and Oukraal
(Oude Kraal). One male from Saldanha Bay (A29783) shows abnormal
arrangement of suckers on third arms, perhaps due to contraction. Middle of
right arm III bears three rows of suckers in irregular quadriserial arrangement,
and one row of three suckers. Left arm III has one row of three suckers, four
very oblique quadriserial rows, then a single dorsal sucker medially. These arms
normal distally.

Fic. 17. Sepia (Hemisepius) typica. a. Dorsal and b. ventral view of female, A29783. c. Dorsal and
d. ventral view of shell of female, A29608.

Left ventral arm of male hectocotylized. Basal half or more of arm modified,
bearing nine to 13 pairs of minute suckers arranged in two widely spaced series
separated by fleshy transverse ridges on arm. Distal half of arm normal, bearing
five to six pairs of normal suckers, then about eight pairs of minute suckers on
arm tip. A few basal modified suckers sometimes larger than the rest.

Tentacular club small and straight, bearing numerous subequal suckers in
oblique transverse rows of six. Rings of suckers broad and nodular, without
teeth. Protective membranes separate proximally. Natatory membrane very

260 ANNALS OF THE SOUTH AFRICAN MUSEUM

well developed, continuing along stalk for a distance from one-half to once
club length.

Shell (Fig. 17c, d) not calcified, very thin and fragile. One shell, success-
fully dissected out (A29608), has dimensions: length 19,5 mm, width 10,5 mm,
length of striated zone 10 mm. It is broad, pointed anteriorly, rounded
posteriorly. No posterior spine or knob, and no median dorsal ridge present.
Phragmocone very short, triangular in shape, with last loculus constituting base
of triangle anteriorly. Last loculus not covering anterior part of dorsal shield.
No median longitudinal groove ventrally. Striated zone occupies most of
phragmocone; striae wavy. Inner cone barely discernible, completely fused
with outer cone; forming, together with its limbs, a circle about posterior point
of striated zone. Outer cone broad.

Remarks

When first describing this species, Steenstrup (1875: 468) created a new
genus for it, on the basis of the following characters:

1. The ventral mantle surface has deep pores, which in Hemisepius typicus

are arranged in two lines of 12 pores each.

2. The shell is poorly developed, with very rudimentary, calcareous loculi
not covering the anterior part of the dorsal lamella, and their anterior
border is not parallel to the corresponding sides of the extremely thin
lamella.

3. All arms with biserial suckers, which are very flattened, almost smooth,
disc-like.

Until recently, these characters clearly separated H. typicus from all the
remaining Sepiidae (although the number of pores was found to vary somewhat,
and the suckers are not flattened, but globose) and the genus was valid. But a
recently discovered species, Sepia dubia, first described by Adam & Rees (1966:
119), shows a number of characters which bridge the gap between the genera
Hemisepius and Sepia, and Adam & Rees relegated Hemisepius to subgeneric
Status.

Chun (1915: 412) first described the male of S. typica, and illustrated the
hectocotylus. He found that most of the suckers in the ventral series of the
modified region were absent, but these had probably been lost, since all the
males in the present collection have a complete ventral series on the hectocotylus.

Smith (1916: 26) suggested the possibility that S. tuberculata be identical
with S. typica. It is difficult to understand how Smith came to this supposition,
since both the shells and the animals of these two species are very different. ‘The
only possible similarities in the shells are the general outline and the thinness
(although the shell of S. typica is much broader and very much thinner than that
of S. tuberculata). As remarked by Adam (1941: 116), the suggestion is clearly
untenable. |

Thore (1945: 50) found that his specimens of S.. typica from Table Bay and
Oukraal (MLd 22-27 mm) were larger than that described by Chun (1915: 412)

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 261

(MLd 17 mm) from a more easterly locality, St. Francis Bay, but comparable
with those of Steenstrup (from Table Bay) and of Massy (1927: 164) (from
False Bay and Hout Bay). On the basis of this, he stated: ‘I think we have to
postulate a constant difference in size between the eastern and western form of
Hemisepius, the latter being the largest’ (Thore 1945: 50), and proposed that the
eastern forms be named H. typicus var. chuni.

Thore found that his western male specimens also differed from that
described by Chun in the number and size of the suckers on the ventral arms.
The hectocotylus of Thore’s specimens differed from that figured by Chun in
that the dorsal row of suckers was separated from the edge of the arm by a
distinct longitudinal groove; the proximal suckers in the ventral row were not
enlarged; the second-last sucker of the dorsal row (in the modified region) was
not enlarged; the enlarged suckers distal to the modified region were about
12 in number (five in Chun’s specimen) and there were about 20 minute suckers
at the tip of the arm (13 in Chun’s specimen).

In all, only three specimens of S. typica are known from localities east of
24°EK: one male (MLd 17 mm) from St. Francis Bay (Chun 1915: 412), one
female (MLd 18 mm) from east of Port Elizabeth (Voss 19625: 252) and one
male (MLd 13 mm) from Cape Natal (Rhombosepion sp. A, Massy 1927: 161,
determined as S. typica by Adam & Rees 1966: 117). In the male from Cape
Natal, the suckers distal to the modified region of the hectocotylus were absent,
presumably lost. This leaves one eastern male with which to compare the western
specimens. A comparison of the specimens of Thore and Chun is given in
Table 7, together with the specimens in the collection of the South African
Museum.

From observations on the specimens and the data in the table, the
following points become apparent:

1. Chun’s specimen from St. Francis Bay is not unusually small, as it falls

into the size ranges of mature specimens from Saldanha Bay and
Mossel Bay, and is larger than those from Cape Infanta and Yster-
varkpunt.

2. The number and position of enlarged suckers on arms I to III
apparently varies randomly; there is no correlation with mantle length
or the locality of the specimens. The distal suckers on arms III are not
always markedly enlarged.

3. Right arm IV: the total number of suckers increases with the size of
the animal, but the specimens from Ystervarkpunt have fewer suckers
than those of the same size from other localities; Chun’s specimen has
an unusually low number of suckers for its size—a specimen of MLd
17 mm from Mossel Bay has 21 pairs of suckers on the fourth right arm.
The position of the distal enlarged suckers is not closely related to the
total number of suckers, but some enlarged distal suckers are almost
always present; in this, Chun’s specimen again differs. There are
usually three pairs of enlarged suckers basally, but occasionally two or

ANNALS OF THE SOUTH AFRICAN MUSEUM

262

6-3

61-Z1

sued € 0} & [eseg

sired y}VI 0} WGI

sued G3-1%

oI

0% ynoqy

sited & jeseg

sited YIQI 0} WISI

sired of ynoqy

OI-9

OZ-OI
sited % 10 € jesegq

SMOJI Y}LI-11
UIOIF JUIWIOZIE[US JIQeIIT A,

sated Lz-61

siayons [e}sIp

oynuIu pue uoIse1 poyipour

us9M}oq posie[US JEqUINNy
A]yeistp oynuTUL JoquIn}yy

Ayjeutxoid posiepuy

Aj[eistp posiepuy

Joquinu [e310],

(AT wre aq)
snjA}090}99y4
uo sioyong

AI wae 14811
uo siayoNG

| ee
_————————mmwuoei

SMOJI Y}II—-O1

jo 9soy} SurIpnjour Ajjensn
‘sMOI YJGI—OI WOIF s1axONsS
jo JUSUIOSIV[US JIqQeIIE A

Mos 416
jo 9soy} SurIpnjout Ajjensn
‘smOI Y}I1I—L WOI S19yONS

jo JUOWOSIE[US sIqQuLIE A

Mod 436
jo 9soy} surpnjout Ajjensn
‘sMOI Y}II-Q UOJ] s1ayONs

jo JUDUIOSIV[US oIqQeIIETA

(poyroods
JOU SMO YOTYM) sioxoONs
jo JUOWOSIL[UD [eISIC

J wie sy

SoTI9S DUO JO
sIayONs IWOS Jo 10 ‘si1ayons
[e3stp Jo sured —z jo juouI

-a81e[U9 oqeImeA pue ‘ured
Jayons YOI Jo UIWIOSIe[UT

sMOI UVI-E1

jo 9soy} surpnypour Ayjensn
*‘sMOI YJQI—GI UOJ] sioaxyONs
jo JUSUIOSIV[US oIqQeIIE A

SMOJ Y}1I-O1
jo 9soy} Surpnjour Ayyensn
‘sMOI YIGI-—Q VOI} sIOyONS

jo JUOWIOSIV[US oIqQeIIeE A

SMOI 401-6

JO 9soy} Surpnjout Ajjensn
‘sMOI Y}GI—Q ULOI] SsIdyONS
jo JUOWIOSIV[US J[qQuiIe A

TET Gy,

IT Uy

siayons
posiepuy

ee | eet | A |

Su] sULU [esIOq

(gL10€vy)
ejurjuy eden HO

(SP61 a104,,,)
Aeg 2192,

(6g46sy)
Aeg eyuepyes

‘papnpour oie snj4y000}99y sTqQeztuUSooeI & YIM susuIOIds AJUG ‘stoyONS WIE Jo JUIWIOZIe[US pue JoquINU jo UOSIIedUIO_D ‘safeuUr DIGG] vidag *L ATA J,

263

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

(ég qo) S

S1OyONs [e}SIP oyNUTUI
OI-9 pue uolse1 poyipour
usaMIJoq posIe[US Joquinyy

Z1-Q Aj[eystp oynuTUL Jaquinyy

(AI wire yeq)
snj4}090}994
uo s1oxING

dn 0} oseq WO azIs Ut
Sutsvoioop Ajjenpeis ‘sired 1

sued $-€ |eseg
sired yIQI-11

sited gZ-1%

su1ed €-2% jeseg Ayjewtxoid posiepuq

sired yiv1-6 Ajjeistp posiejuq

si1ed 13-91 Jaquinu [ej0 J,

AI Wue 7Y8I1
uo siayong

___ ) |) |

poqodar jusWIaZIe;Ua ON

pesiejue sed yj01

posivjua sed y}01

SMOI UVI-6 Woy siayONs
JO JUDWOZIV[US JTQeIIeE A

SMOI 401-6
jo soy} Surpnyzour Ayjensn
‘sMOI Y}E1—Q UIOIF sIOyONS

JO JUDWOSIV[UD JIqQuIIT A

MoI Y16

JO 2soy} Surpnjour Ajjensn
‘smo Y}11—L Woy siayons
Jo JUOWOSIV[US s[qQuIIVA

SMOJI Y}1 I-O1
JO osoy} Surpnyour ATjensn
*‘sMOI YISI—OI WOIT s1oyons
jO JUSUIOSIV[US sIqeLUueA

Il] SV

MOI YQ
jo 9soy} Surpnyour Ayjensn
‘sMOI YJOI—Q WIOJJ s1ayONS

jo JUDUIOSIe[US JIqQuLueA

ty

smo yIg—L

jo 9soy} Surpnjour Ajjensn
‘smMOI 4}6—Q UIOJJ s1oyons
jo JUDWIOSIV[UD oIqQeLIeA

sioyons
posiepuq

ee” | Fe

(S161 unyy)
Aeg stoueiy 1g

(LL10€&y)
Aeg [assoyy

(L1L6zy)
yundyivAsi9isx JO

SUI] WU TesIOGg

(panuyuos) L alav J,

264

ANNALS OF THE SOUTH AFRICAN MUSEUM

four pairs; Chun’s description does not preclude the possibility that
these were also present in the specimen from St. Francis Bay.
Hectocotylus: the number of minute distal suckers of Chun’s specimen
is within the normal range for its size. The number of enlarged distal
suckers in Chun’s specimen is unusually small; from the figure (Chun
1915, fig. 34), ‘hore deduced that the penultimate sucker of the dorsal
series on the modified region is enlarged, and that there are five enlarged
suckers distal to the modified region of the hectocotylus. An alternative
interpretation is that this single enlarged dorsal sucker is part of the
group of distal enlarged suckers, of which the second pair has failed to
become enlarged. A similar case was observed in a specimen (MLd
16 mm) from Cape Infanta, in which one sucker of the second pair of
enlarged suckers has remained minute. The longitudinal groove
separating the dorsal series of suckers from the edge of the modified
surface, as reported by Thore, was observed in some specimens but not
in others, and apparently depends on the state of preservation of the
specimens. The three enlarged proximal suckers in the ventral row on
the modified region, as illustrated by Chun, are absent in Thore’s
specimens. In the present specimens the proximal suckers of the dorsal
and/or ventral row were sometimes found to be somewhat larger than
the more distal suckers on the modified region of the hectocotylus. But
the proximal enlargement is not marked, and the size of the suckers
of the modified region gradually diminishes distally. The number of
large suckers, and their position, apparently varies randomly.

Thus Chun’s specimen differs from those from more westerly localities only

in that it has fewer suckers on the right arm IV, of which none are enlarged
distally. Whether or not this constitutes a valid character for separating eastern
and western forms of S. typica cannot be decided on the basis of the presently
known specimens, and a decision must await the collection of further specimens
from the eastern coast of South Africa. In any case, the ‘variety’ is no longer
recognized as a valid taxon within the system of nomenclature, and should be
replaced by the term subspecies, provided the eastern and western forms of
S. typica are found to differ sufficiently.

Sepia (Hemisepius) dubia Adam & Rees, 1966
(Tables 6, 50)

Sepia (Hemisepius) dubia Adam & Rees, 1966: 1109, pl. 34, figs 198-201, pl. 46, fig. 272, text fig. 1.

Type locality

False Bay, 34° 11'S, 18° 27'E, 25 m.

Description

The only known specimen (female) was not available for examination.

Following description after Adam & Rees (1966: 119) and Taylor (personal

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 265

communication), who kindly supplied some additional information about the
type specimen.

Animal small. Mantle broadly oval, anterior mantle margin slightly
convex dorsally, deeply emarginate ventrally.

Dorsal surface of mantle, head and arms covered with well spaced round
papillae; in addition, three oval patches of contracted papillae present medially:
one on either side of median line, and third one anteriorly near mantle margin.
Ventral mantle surface smooth, with thick fleshy keel, parallel to outer margin,
as in §. typica, but without pores.

Arms subequal in length, laterally compressed, keeled on outer sides;
protective membranes narrow. Web very high between dorsal and lateral arms,
attaining half arm length; lower between arms III and IV, absent between
ventral pair. Arm suckers rather small; biserially arranged.

Tentacular club small, crescent shaped, bearing 54 minute subequal
suckers arranged in four or five longitudinal series (according to Taylor,
personal communication, the median suckers are slightly larger than the others).
Natatory membrane well developed, extending beyond base of club for a
distance equal to about half club length. Dorsal protective membrane wide,
separated from ventral membrane at base of club.

Shell broadly oval, somewhat acuminate anteriorly, broadly rounded
posteriorly. Almost whole dorsal surface calcareous, with reticulate pattern.
Posterior spine absent. Ventral surface strongly concave, spoon shaped.
Phragmocone has reversed conical shape, occupying a little more than half
shell length, as in S. typzca. Last loculus trapezoid, widest at anterior margin.
Striated zone about twice as long as last loculus, but occupying only central
third of width of phragmocone; broad smooth marginal area present on
either side of striated zone. About 15 widely spaced, transverse, slightly
wavy striae present. Inner cone distinct, brownish in colour, with rather
broad limbs, completely fused to outer cone. Posterior part of inner cone
surrounds shallow depression. Outer cone broad, completely surrounding
inner cone.

Remarks

Sepia dubia resembles S. typica in that it has biserial suckers on the arms, the
tentacular clubs are the same, and the shell is very similar, with a very short
phragmocone, whose anterior border is not parallel to the corresponding sides
of the thin dorsal lamella. The shell differs however, in that it has a calcareous
covering of the dorsal surface, and a distinct inner cone. Sepia dubia has fleshy
keels on the ventral mantle surface, but there are no pores. It also differs from
S. typica in that the skin is not smooth, but is covered with well spaced papillae
dorsally.

Whilst $. dubia is obviously closely related to S. typica, the differences
between the two species include a character (pores in the ventral mantle surface)
listed by Steenstrup (1875: II) as defining the genus Hemisepius, and led Adam &

266 ANNALS OF THE SOUTH AFRICAN MUSEUM

Rees (1966: 143) to state that ‘the genus or subgenus Hemisepius may be main-
tained for H. typicus and H. dubius, but its separation from other Sepiidae has
become less distinct’. In the taxonomic section of their review (Adam & Rees
1966: 117, 119) they have relegated Hemisepius to subgeneric status for S. typica
and S$. dubia.

Septella cyanea Robson, 1924
(Pl. 42c, d. Fig. 3b. Tables 51-53)

Sepiella cyanea Robson, 1924a: 13; 1924): 648, figs 25-27, pl. II, fig. 6. Adam, 19398: 109,
figs 14A-B, pl. IV, figs 3, 4. Voss, 19625: 248. Adam & Rees, 1966: 121, pl. 36, figs 208-215.

Sepia sp. a Voss, 1962a: 3.

? Sepiella obtusata (non Pfeffer) Massy, 1928: 95.

Type locality
S.S. Pickle, Sta. 476: 29° 17'S, 31° 33’E, 51 m (lectotype, designated
Adam & Rees 1966: 121).

Distribution

Animals: Port Elizabeth (Adam & Rees 1966: 121) to 29° 17’S, 31° 33’E (off
Tugela River) (Robson 1924a: 14, Sta. 476) and Nosy N’Tangam
(Malagasy) (Adam & Rees 1966: 121). Depth 51-73 m.

Shells: Port Elizabeth (Adam & Rees 1966: 121) to ? Tongaat (Massy
1928: 95) and Ambavanibé, Malagasy (Adam 1939): 109).

Material
S.A.M. A6526, Port Alfred; 1 3, 5 shells

Description

Mantle elongate oval, anterior mantle margin produced dorsally to eye
level, emarginate ventrally. Mantle bluntly rounded posteriorly and posterior
gland (characteristic of genus) opens via pore situated between and just below
posterior extremities of fins. Fins fairly wide, beginning a few mm from mantle
margin, rounded and closely approximated posteriorly. Posterior region of fins
somewhat damaged in present specimen, but according to Adam & Rees
(1966: 122), fins fused at base.

Skin smooth; colour of head and mantle dark blue-purple mid-dorsally,
paler towards fins, on each of which a series of dark wedge-shaped patches
present in male. Ventral surface pale, with sparse chromatophores medially,
more concentrated near fins.

Arms fairly short, with arm length formula 4.3.1.2, and joined by low
interbrachial web, also present between ventral arms (but Adam & Rees 1966:
123, report that web absent between ventral arms). All arms keeled. Arm tips
attenuated. Suckers quadriserially arranged on all arms; chitinous rings of
suckers toothed distally with about 12 long teeth.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 267

Left ventral arm hectocotylized over basal half. Minute suckers in modified
region arranged in one dorsal, one medio-dorsal and two ventral series. Arm
surface between dorsal and ventral series is transversely ridged, and suckers of
medio-dorsal and dorsal series are situated on the ridges. Middle and distal
parts of arm somewhat mutilated, but distal part apparently normal, with
quadriserial suckers. :

Tentacles of present specimen missing. According to Adam & Rees (1966:
123) ‘the tentacular stem is triangular in cross-section with a rounded keel on
the outer side and a flat, transversely-striated, inner surface, which is limited by
two membraneous ridges, these being the continuation of the protective
membranes of the club. The swimming-membrane of the latter is not very
broad and barely reaches the base of the club. The protective membranes are
narrow. The minute, subequal suckers are arranged in about 12 longitudinal
series in both sexes (pl. 36, fig. 215). Their chitinous rings are each armed with
a few, blunt, spaced teeth.” Robson (1924b: 648) also observed two enlarged
suckers at the extremity of the club.

Shell (Pl. 42c, d) elongate oval, somewhat narrower anteriorly, but not
sharply acuminate; posteriorly rather broadly rounded. Dorsal surface
calcareous, with median longitudinal ridge and lateral grooves. Chitinous
margin fairly broad. No posterior spine, but slight hump present over posterior
extremity of striated zone. Broad shallow groove present, running from hump
to posterior margin. Striated zone long, fairly broad posteriorly. Anterior
margin of striated zone broadly convex, somewhat wavy. Last loculus continues
along sides of striated zone to meet limbs of inner cone. Anterior part of smooth
zone shows some compression in most shells. Faint indication of median longi-
tudinal groove over striated zone only. Inner cone forms knob posteriorly, and
has very short narrow limbs. Outer cone very broad.

Females differ from males in that arms are relatively much shorter, protec-
tive membranes on arms better developed, covering distal suckers (Adam &
Rees 1966: 123) and sucker rings almost smooth (Robson 19246: 648). Shell of
female broader, ventral surface thinner, inner cone more developed and outer
cone broader (Adam & Rees 1966: 123) and striated zone less pointed (Robson

19245: 649).

Remarks

The genus Sepiella is represented by only one known species, §. cyanea, in
southern African waters. Together with S. ornata and S. weberi, it differs from the
other species of Sepzella in the number of sucker series (10-14) on the tentacular
club (S. melwardi from Australia is known only by its shell).

Sepiella ornata (from West Africa) differs from S$. cyanea in that the shell is
narrower, the striated zone is shorter, and the posterior part of the shell is less
broad; in S. weberi (from Timor and Soemba) the posterior part of the striated
area is less acuminate than in S. cyanea (Adam & Rees 1966: 123).

268 ANNALS OF THE SOUTH AFRICAN MUSEUM

Discussion

RELATIONSHIPS

The genus Sepza was created by Linnaeus (1758) to include all cephalopods
without an external shell. Of these, the only true Sepza included in the genus was
S. officinalis. The genus was restricted by Lamarck (1799: 4) to include only
those cephalopods with an internal calcareous shell. At that time, the only
known species were S. officinalis and S. tuberculata.

Gray (1849: 106) first used the name Sefzella for a group of shells which are
‘oblong, posterior end expanded, produced, cartilaginous, not beaked, convex
beneath’. Steenstrup (1875: 468) created the genus Hemisepius, with H. typicus
as the type species. This author (1880: 347) redescribed the group Sepiella and
on the basis of the characters of both the shell and the soft parts raised Sepiella
to generic status.

With the discovery of numerous species of Sepia in the course of time,
several attempts were made to divide the genus into subgeneric groups, mainly
on the basis of sucker arrangement on the sessile arms and tentacular clubs,
and on the structure of the shell (d’Orbigny 1845a: 261-298; Gray 18409:
g6—112). Attempts have also been made to split up the genus Sepza into several
genera, mostly without success. Rochebrune (1884: 74) divided the Sepiidae
into ten genera, mainly on shell characters. This classification has been shown
to be extremely contradictory and unnatural (Adam 1944). More recently,
Iredale (1954: 81) divided the Australian Sepiidae into three families, four
subfamilies and 13 genera! Criticisms of this classification are given by Adam
(1964: 265) and Adam & Rees (1966: 132). These authors have suggested a
more reasonable classification,* retaining only the genera Sepia and Sepiella,
which is outlined below in slightly modified form, together with the southern
African representatives of each group:

I. Shell with well-developed posterior spine; ventral part of inner cone
strongly developed. Tentacular suckers subequal, in 8-20 series.
Sepia zanzibarica

II. Shell with well-developed posterior spine; inner cone well

developed, with wide limbs, but completely reflexed on to and fused

with the outer cone. Tentacular suckers unequal, usually in eight
longitudinal series. Sepia officinalis vermiculata

III. Shell with posterior spine which is generally not keeled, but may be

keeled dorsally, ventrally, or on both sides; inner cone more

reduced, with narrow limbs; in most of these species the outer cone

has two posterior wings which form, in the narrower shells, a

typical cup-like expansion. Tentacular suckers nearly always

arranged in eight longitudinal series.
* Adam (pers. comm.) says that he and Rees do not attach any systematic value to these

groups. Where, however, a genus contains as many species as does Sepia, the use of some such
grouping of similar species greatly facilitates the study of interspecific relationships.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 269

PV.

VI.

VII.

VIII.

a. Outer cone without wings; posterior spine with or without
keels. ‘Tentacular suckers minute and subequal or slightly
unequal. Sepia acuminata

b. Outer cone with wings; posterior spine not keeled. Tentacular
suckers subequal. No southern African representatives

c. Outer cone with wings; posterior spine without keels. Tentacular
suckers unequal. In some species the ventral part of the inner
cone forms a short, rounded ledge. Sepia confusa, S. incerta,
S. burnupi, S. joubint and S. adami

d. Outer cone with or without wings; posterior spine keeled.
Tentacular suckers unequal or subequal. Sepia australis

e. Outer cone with wings; posterior spine absent. Tentacular
suckers subequal or unequal. No southern African repre-
sentatives

Shell relatively broad, with a more or less developed inner cone and

without posterior wings at the outer cone; posterior spine absent.

Tentacular suckers unequal or subequal. Sepza tuberculata, S. papillata,

S. semoniana, S. angulata, S. hieronis and S. insignis

Form of shell not well known. Tentacular suckers subequal; dorsal

arms with finger-like tips, devoid of suckers. Sepia robsoni and

S. faurer

Subgenus Hemisepius: Shell very thin, without posterior spine;

phragmocone considerably shorter than dorsal shield; inner cone

reduced. Sepia typica and S. dubia

Subgenus Metasepia: Shell rhomboidal, much shorter than the

mantle, with a completely chitinous dorsal surface. Tentacular

suckers very few in number, unequal. Inner cone very narrow;
posterior spine absent. No southern African representatives

Genus Sepiella: Mantle with a posterior gland and characteristic

locking apparatus (Fig. 3b). Tentacular suckers subequal, in

8-32 longitudinal series. Shell with outer cone expanded, inner cone

reduced; posterior spine absent. Sepzella cyanea

The degree of affinity within the different groups varies. The southern
African species falling into group IIIc, the ‘doratosepion’ group (Sepia confusa,
S. incerta, S. burnupi, S. joubini and S. adamz) show close interrelationships and a
marked resemblance in the characters mentioned above (p. 195) as defining
Rochebrune’s genus. Not all the suckers on the sessile arms are biserial, but in
most of these species there is a biserial arrangement of suckers on some part of
the arms in one or both sexes. These species are also remarkable in that nearly
all show sexual dimorphism, such as the ‘tail’ in males of S. confusa, the modified
dorsal arms in the male of S. incerta, and the modified dorsal and ventral arms
in the male of S. burnupi. The males of S. jowbini are less remarkable, being distin-
guished from the females (apart from the hectocotylization of the left ventral
arm) mainly by the red spots on the arms. The male of S. adami is not known.

270 ANNALS OF THE SOUTH AFRICAN MUSEUM

The southern African ‘doratosepion’ species fall into two orders of size.
S. confusa and S. incerta attain dorsal mantle lengths of 85-90 mm in the females
and about 150 mm in the males. S. burnupi, S. joubint and S. adami attain dorsal
mantle lengths of 36-59 mm in the females and 41-45 mm in the males (the
male of $. adami is not known). All these species occur off the east coast of
southern Africa.

Sepia australis (group IIId) shows a superficial resemblance to the smaller
‘doratosepion’ species, and particularly to S. joubini and S. adami. But the shell is
wider, has a keeled spine and no posterior wings on the outer cone. Its distri-
bution also differs, and this species has been assigned to the Cape faunistic
province (see below).

Sepia acuminata (group IIIa) differs markedly from the other species of
group III. The animal is generally broader and the shell is not narrow elongate
but broad, almost rhomboidal, with no posterior wings on the outer cone.

The affinities between the southern African representatives of group IV
(Sepia tuberculata, S. papillata, S. simoniana, S. angulata, S. hieronis and S. insignis)
are much less clear than those of group III. Whilst there are similarities
between some of these species, the only characters common to all are the
absence of the posterior spine and wings on the outer cone of the shell, and the
well developed inner cone which is completely reflexed and fused to the outer
cone. The shell of S. papillata (Pl. 41a—d) is similar to that of S. tuberculata
(Pls 39c, d, 40c, d) on the one hand, and on the other hand to that of S. stmoniana
(Pl. 42a, b), from which it is sometimes almost indistinguishable, and to that of
S. angulata (Pls 44d, 45a—d). The shells of S. hzeronis (Pl. 43a—-d) and S. insignis
(Pl. 44a-c) are very different to these and to each other in general shape. A
consideration of the tentacular suckers, however, divides these species
differently: S. tuberculata and S. papillata have unequal tentacular suckers,
whereas in S. simoniana, S. hieronis and S. insignis they are subequal. The soft
parts of S. angulata are not known. The distribution of these species also varies
(Fig. 18, Table 8), but all occur in the Cape—South West African province.

Only two species, S. typica and S. dubia, fall into group VI, and both are
southern African. These small animals show a close affinity in general
appearance and in the shells. S. typica has previously been separated from the
other Sepiidae in the genus Hemisepius, but the discovery of S. dubia has rendered
the separation of Hemisepius from Sepia less clear. The distribution of these
species, and those of group V, falls under the Cape-South West African
province.

The systematic position of Sepia robsoni and S. faurei (group V) seems to be
intermediate between those of groups IV and VI. Both species show a relation-
ship with the species of group VI in general shape (animal very small, mantle
very broad, with fleshy keels ventrally) ; other similarities between the species of
groups V and VI include the similarity of hectocotylus of S. robsont and S. typica
(those of S. fauret and S. dubia not known), the biserial arrangement of the
suckers on all the arms, and the similarities of the tentacular clubs, bearing

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 27

subequal suckers. The shell of S. robsoni, of which little is known, apparently
resembles those of S. Aieronis and S. insignis (group IV). The shell of S. faurez,
on the other hand, approaches the Hemisepius-like shell: it is very thin and
chitinous, and the phragmocone is considerably shorter than the dorsal lamella,
but is not as short as those of S. typica and S. dubia. The shell of S. faurei differs
from those of group VI in that the anterior margin is more nearly parallel to the
corresponding margins of the dorsal lamella.

An evolutionary series leading to the Hemisepius condition can be traced
as follows: Sepia robsoni (shell Sepia-like, dorsal arms with bare tips, fleshy keels
on mantle ventrally, pores absent) to S. faure: (shell approaching Hemisepius
condition, dorsal arms with bare tips, fleshy keels on mantle ventrally, pores
absent) to S. dubia (Hemisepius-like shell, dorsal arms normal, fleshy keels on
mantle, no pores) to S. typica (Hemisepius-like shell, dorsal arms normal, fleshy
keels with pores).

It is impossible at this stage to draw any further conclusions regarding the
phylogenetic relationships between the sepiids. Adam (19394: 92) concluded
that the shell and tentacular club present the best features for distinguishing the
species, and Adam (1964: 268) and Adam & Rees (1966: 135) have suggested
that the following characters are probably primitive: a well-developed ventral
part of the inner cone; the presence of a posterior spine; arm suckers of equal
size and quadriserially arranged; subequal tentacular suckers, arranged in
eight longitudinal series; and the presence of minute suckers on the buccal
membrane (of the species described here, only S. zanzibarica has buccal suckers).
But an arrangement of species according to shell structure does not agree with
an arrangement of a series according to the structure of the tentacular club,
e.g. a comparison between S. papillata and S. simoniana, whose shells are very
similar, but whose tentacular clubs are very different.

GEOGRAPHICAL DISTRIBUTION

Sepiids are cephalopods inhabiting the continental shelf and slope, and in
some cases the intertidal zone. Although they are capable of active swimming
they apparently spend most of their time on the bottom and do not move over
very wide areas or go far beyond the continental shelf (sepiids are not generally
found deeper than about 500 metres*). This is borne out by the distribution of
the individual species; it is found that although sepiids occur around the coasts
of Europe, Africa, Asia, the Indo-Pacific islands and Australia, there are no
known cosmopolitan species. On the other hand, the number of endemic species
is high.

The earliest fossil record of sepiids is that of Voltzia palmeri Schevill, from
Upper Jurassic deposits. Apart from this, sepiids are known almost exclusively
through the Tertiary to the present. Five genera have been recorded from the
Eocene (Bilow-Trummer 1920; Roger 1952): Archaeosepia, Belosepia, Pseudosepia,

* Three known exceptions are Sepia elliptica, S. hedleyi and S. pharaonis, which have been
collected from depths to 1 000 metres.

272 ANNALS OF THE SOUTH AFRICAN MUSEUM

Sepia and Stenosepia, but of these only Sepza has been recorded from later Tertiary
deposits (Oligocene, Miocene, Pliocene) and still occurs in recent times. Thus it
would appear that Voltzia became extinct in the Jurassic, and Archaeosepia,
Belosepia, Pseudosepia and Stenosepia in the Eocene. All recent sepiids were
presumably derived from the fossil genus Sepza.

All but three of the fossil sepiids were found in European deposits. The
three exceptions are Voltzia palmeri Schevill from Cuba (Upper Jurassic),
Belosepia incurvata Cossmann & Pissaro from West Pakistan (Sind Region;
Eocene) and Belosepia ungula Gabb from North America (Texas, Missouri,
Alabama; Eocene).

The absence of sepiids from New Zealand and both coasts of the American
continent today is apparently due to the separation of these land masses by
extensive, deep oceans and/or very cold water in the regions where migration
could otherwise occur. The ocean between Australia (where numerous species
of Sepiidae occur) and New Zealand is wide and deep, and the currents are
adverse to a crossing in this direction.

Tas_eE 8. Distributional categories of the Sepiidae of southern Africa
(in the same order as in figure 18).

Mocambique—Malagasy Natal Cape—South West African
province’ province province

Sepia confusa Sepia acuminata Sepia faurer
Sepia joubini Sepia stmoniana
Sepia burnupi Sepia insignis
Sepia adami Sepia dubia
Sepia incerta Sepia robsoni
Sepiella cyanea Sepia tuberculata

Sepia typica

Sepia officinalis vermiculata
Sepia papillata

Sepia australis

Sepia hieronis

? Sepia angulata

At least two species of fossil sepiids have been recorded from America, the
most recent from the Eocene. The absence of later fossil evidence suggests that
these species became extinct, and that recolonization was prevented, perhaps
by the low temperatures prevailing in the only relatively shallow areas via which
Sepiidae could migrate from Asia or Europe, viz. the Bering Straits or via the
Faeroe Islands, Iceland and Greenland.

In the waters around southern Africa, 19 species of Sepiidae are known to
occur. Of these, 16 are endemic, one is tropical, one is an Atlantic species and
the remaining species has an interrupted distribution.

The tropical species, Sepia confusa, enters southern African waters at the
southern end of its range. It is known to occur from Zanzibar to Durban, and is

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 273

20°S
ty WALVIS BAY
xe)
S
ay
sa) ata
n LUDERITZBUCHT
[o)
” 5
= =
ome = ORANGE R.
Sm|c
one |2
ey ees
fat)
B ® wy ”
es ss
a39Y%2Z 5d
aici eno: SALDANHA BAY
= r= 2
@eo")
°° CAPE POINT
ne
ake)
os
o®
22 KNYSNA
<
S
(42)
el @ 9)
> PORT ELIZABETH
@
@
ay
EAST LONDON
nN
DH BASHEE R.
NP AL o
820238
S2cag
ec pe DURBAN
at

DELAGOA BAY

PUNTA ZAVORA

259°S

Fic. 18. Distribution ranges of the Sepiidae of southern Africa.

274 ANNALS OF THE SOUTH AFRICAN MUSEUM

related to other sepiids of the ‘doratosepion’ group occurring off the east coast
of southern Africa.

One Atlantic species, Sepia officinalis, enters southern African waters. It
occurs from Scandinavia to Delagoa Bay (Mocgambique), but the subspecies
S. officinalis vermiculata is endemic to southern Africa, occurring from off the
Groene River mouth on the west coast to Delagoa Bay on the east coast.

Sepia australis has an interrupted distribution, occurring in southern
African waters from the Olifants River on the west coast to Rame Head, near
Port St. Johns, on the east coast, but is also found in the Red Sea.

The endemic species are divided into two groups: the subtropical species
restricted to the east coast of southern Africa, and the temperate species,
occurring from the west coast round to the south coast and gradually diminishing
along the east coast.

Stephenson (1948: 228), on the basis of extensive surveys of the intertidal
fauna, divided southern Africa into three faunistic provinces, namely the
subtropical population of Natal, the warm temperate fauna of the south coast
and the cold temperate fauna of the west coast. Day (1967: 11) found, however,
that these divisions do not apply to shelf fauna, since the change in temperature
at, say, 100 m is not as marked as it is in the intertidal zone. Thus the surface
temperature of the south coast (Bashee River to Cape Point) ranges from 15° to
20°C, whereas on the west coast (Cape Point to South West Africa) the surface
temperature range is about 12° to 15°C. At 100 m, however, the bottom
temperature is far more constant, the range being about 12° to 14°C from
Port Elizabeth to Liideritzbucht (Day 1967: 12).

Day (1967: 12) suggests the following faunistic provinces to include both
intertidal and shelf fauna:

1. The Mocambique—Malagasy province, reaching Delagoa Bay;
dominated by tropical species.

2. The Natal province, from Delagoa Bay to Bashee River; many tropical
species, but also fair numbers of endemics and Atlantic species.

3. The Cape-South West African province, from Bashee River to about
Cape Frio; dominated by endemics but with a few tropical species and
several other components. The intertidal fauna of this province differs
on the Indian and Atlantic coasts, i.e. are separated into warm
temperate and cold temperate forms.

The distribution of the Sepiidae of southern Africa agrees with Day’s
faunistic provinces, except that the boundary between the Natal and the
Cape-South West African species seems to lie a little further south, between
Port Elizabeth and East London (Fig. 18). The species of Sepiidae arbitrarily
assigned to the various provinces are listed in Table 8.

1. The Mogambique-Malagasy province: only one tropical sepiid species, Sepia
confusa, is known to occur here. It also occurs further south, as far as Durban.
Sepia hieronis has been caught in this region, but as it has also been recorded off

A REVIEW OF THE SEPITDAE (CEPHALOPODA) OF SOUTHERN AFRICA 275

the west coast, this is not a tropical species but an endemic one with a peculiar
interrupted distribution. Sepia acuminata, a subtropical species, enters this
province at the northern end of its range.

2. The Natal province: all the species of this category are endemic, and are
restricted to the east coast of southern Africa; they do not extend further south
than Port Elizabeth. The main component of this group of subtropical species
is the ‘doratosepion’ group, including Sepza incerta, S. burnupi, S. joubint and
S. adami. These species are very similar anatomically and are obviously closely
related. S. confusa (see above) also belongs to the ‘doratosepion’ group, but it is
not endemic to southern Africa, as it has been recorded as far north as Zanzibar.
Since, however, the waters off the east African coast have not been extensively
sampled, further collection may well show that the sepiids of the Natal province
occur further north than is known at present.

Two other species, Sepiella cyanea and Sepia acuminata, are found in the
Natal province. The latter species extends into the Mocambique province, and
has been recorded as far north as Zavora. Sepiella cyanea has also been recorded
from Nosy N’Tangam, Malagasy (Adam & Rees 1966: 121).

Several species from the Cape—South West African province extend into
the Natal province (see below).

3. The Cape—South West African province: Sepiids do not seem to enter South
West African waters to any marked degree. Only one Cape species, S. papillata,
has been recorded from South West Africa. The apparent absence of Cape
Sepiidae from South West African waters does not seem to be due to tempera-
ture, since this varies little along the west coast of southern Africa. It may
however be due to the difference in coastal conditions, since the South West
African shores consist largely of long sandy beaches, with pounding surf and
very little kelp. Rocky outcrops are few and far apart. According to Penrith &
Penrith (1969: 100), fishes of the genus Lizthognathus show a similar break in
distribution off the South West African coast. L. lithognathus occurs from Natal
to the Orange River mouth, but has not been recorded off South West Africa,
where the species is replaced by L. aureti and L. oliviert off the northern half of
South West Africa. No species of Lithognathus have been recorded off the
southern half of South West Africa, between Sandwich Harbour (on the tropic
of Capricorn) and the Orange River mouth.

Of the Cape species of Sepiidae, Sepza tuberculata occurs from Melkbosstrand
to Knysna. The relatively short range of this species may be linked with its
shallow-water habitat (depth range o-3 m).

S. papillata and S. typica have wider ranges of distribution, extending into
the Natal province (to Durban). On the west coast, S. papillata has been
recorded as far north as Liideritzbucht—one of the few rocky areas on the
South West African shore; S. typica does not seem to occur north of Saldanha
Bay.

As mentioned above, S. officinalis vermiculata is endemic, and occurs in the

276 ANNALS OF THE SOUTH AFRICAN MUSEUM

Cape faunistic province. It has been recorded from Groene River mouth to
Delagoa Bay, the latter being the northernmost record for temperate Cape
species of Sepiidae along the east coast (except perhaps S. hieronis).

Sepia simoniana is the only Cape species which does not occur off the west
coast. It has been recorded from False Bay to the Tugela River mouth.

Other endemic species of the Cape province are Sepia insignis, S. robsoni,
S. faurei and S. dubia, each known from only one locality, and S. angulata. ‘The
soft parts of the latter species are as yet unknown; it is not included in Figure 18.
S. hieronis has provisionally been allocated to the Cape species, although it has
not so far been recorded from the south coast; it is known to occur off the west
coast (from Hondeklip Bay to Slangkop) and off Monte Belo, Mocambique.

VERTICAL DISTRIBUTION

The vertical distributions of the southern African sepiids are not well
known. Many records give no indication of the depths at which the specimens
were caught, and others, where open trawls were used, are unreliable. In the
latter case the depth at which the trawl was hauled is given, but the specimens
could as easily have been caught while the trawl was raised or lowered.
However, since sepiids tend to live on or near the bottom, the depth over which
the trawl was fished probably gives a fairly good indication of the depth at
which the sepiids were living. The depth records available (Fig. 19) show some
interesting features.

Off the south coast, where the continental shelf is very wide (about 220 km
at the widest point), few sepiids have been recorded below 100 m. Off the east
and west coasts the continental shelf is much narrower (maximum about
40-50 km) and sepiids have been recorded as deep as 460 m, that is, some way
out beyond the edge of the shelf.

Sepia officinalis vermiculata is unusual in that it is found in shallow water in
estuaries and sheltered bays (Saldanha Bay, Breede River mouth, Knysna
lagoon, Bushmans River mouth and Durban Bay). Strangely it has not so far
been found in False Bay. This subspecies is not limited to shallow water,
however, and has also been recorded from depths to 249 m off the Natal coast
(Massy 1925: 200).

Sepia tuberculata is apparently more closely restricted to the inshore waters,
and has been collected mainly from rock pools. There is also one record from
Simonstown harbour, at a depth of 3 m. Many of the records do not give any
reference to depth, and this species is not included in Figure 19.

Sepia typica is common in fairly shallow water in Saldanha Bay, but else-
where it is found somewhat deeper. In Table Bay and False Bay this species is
found below 17-18 m, off the south coast below 40 m, and off Durban at 99 m.
Similarly the upper depth limit of S. szmoniana is deeper off the east coast
(116-134 m) than off the south coast (below about 10 m in False Bay).
S. papillata and S. australis show a similar though less marked trend at the
eastern end of their distribution ranges.

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 277

oe
3 ee: Z
<td z
3S 5 = 2 °o Ee)
ae < z a ™N [=] 3 —_
N a iis zy: par] ze as
[es vm = 42 = 4 wm oO w z 3° at
= oO = = “2 = =z Ow wo 2
uw z oO uw w Www - Se = is) at 29 N
8 ¢ A ialibres VeOk Sorcok 28 §<
3 o ” Oo $62 a. w wo a (=) ON
0 U U feed U J
S

® S. hieronis

0 S. confusa

4 S, adami

4 S. incerta

® S. papillata

© S. officinalis vermiculata

w
w
=
w Oo
=
=
=
Ke
a
Ww
a
200
O S. joubini
@ S. acuminata
~¥ S. burnupi
@S. typica
OS, simoniana
VSepiella cyanea
400

Fig. 19. Vertical distribution of the Sepiidae of southern Africa. Linked symbols indicate
continuous hauls between the indicated depths.

278 ANNALS OF THE SOUTH AFRICAN MUSEUM

The east coast sepiids also show a tendency to occupy deeper water
further northwards, as the climate becomes hotter, though the absence of
records from shallow water off the Mocambique coast may well be due to
insufficient sampling. The east coast species are generally caught below 60 m
off the Natal coast and below 200 m off the Mocambique coast. Two exceptions
are §. burnupi, caught at 40-48 m off the Umhlanga River, and Sepzella cyanea,
recorded from 51 and 73 m off the Tugela and Umvoti Rivers. For both these
species, these are the only available depth records.

GROWTH

For each species the relative dimensions were calculated for as many
specimens as possible, and the ranges and means were calculated for those
dimensions thought to be of some significance. Tests for correlation between
various relative dimensions and the standard length (MLd or shell L) were
performed for species where ten or more specimens had been measured (males
and females being considered separately), to test if there is any change
in relative body (or shell) dimensions with growth. In general the results of
the correlation tests were disappointing, due to the great variation in measure-
ments.

The only species giving significant correlations for most of the body
measurements was Sepia australis, and even here the scatter about the regression
lines is wide, though 59 males and 77 females were measured. Figure 20 shows
the scatter diagrams for most of the dimensions of S. australis males and shells.
Where correlation was found to be statistically significant, the regression lines
were calculated by the least squares method. These lines are included in
Figure 20 for interest, but they clearly do not fit the scatter diagrams well, since
the variation is very wide. Trends can, however, be observed. With growth, the
mantle becomes much narrower relative to its length; the head becomes
relatively shorter and narrower, and the fins also become relatively narrower.
The change in relative length of the tentacular club is small, but the arms
become relatively longer. In fact there seems to be a general trend to elongation
with growth, producing a more streamlined body, perhaps enabling faster
locomotion.

Growth of the S. australis females is very similar to that of the males,
except that the arms are relatively shorter than those of the males, although
also becoming longer in larger animals.

The shell dimensions also show a wide scatter, despite the fact that the
shell is a rigid structure and is not subject of contractility as are the soft parts.
The shell is however secreted by the animal and the variation in its dimensions
suggests that the wide scatter of the relative dimensions of the soft parts is
not solely due to their contractility. The shell becomes relatively narrower
and slightly thinner with growth, and the striated zone becomes relatively
longer. |

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 279

Fic. 20. Sepia australis males and shells. Change in dimensions (expressed as % MLd or % shell

length) with growth (increase in MLd or shell length). The straight lines through the scatter

diagrams indicate the calculated regression lines, where a statistically significant correlation was
found to obtain.

280 ANNALS OF THE SOUTH AFRICAN MUSEUM

SUMMARY

A complete synonymy, distribution and depth ranges, and descriptions of
external morphology are given for each of 20 species. Three species, Sepia adami,
S. angulata and S. faurei, are new. The soft parts of S. znszgnis (previously known
only by its shell) are described for the first time. The animals and shells
described by Massy (1925) as S. zncerta and S. burnuft have been re-examined,
and all are found to pertain to S. incerta. The first known animals of S. burnupi
(formerly the types of S. exsegnata) have been redescribed, and the first recorded
specimens of §. hieronis from the east coast are described. Keys to the soft parts
and to the shells are provided. The relationships of the southern African
Sepiidae, their geographical and vertical distribution, and their growth, are
discussed.

ACKNOWLEDGEMENTS

I am grateful to Professor W. Adam of the Institut Royal des Sciences
Naturelles de Belgique, Dr. D. van Z. Engelbrecht of the University of Stellen-
bosch, Dr. N. A. H. Millard of the South African Museum, Dr. P. B. Best of the
Division of Sea Fisheries, Cape ‘Town and Dr. M.-L. Penrith, formerly of the
South African Museum, for valuable advice on the manuscript. Also to
Mr. S. X. Kannemeyer, for the photography.

For the donation and/or loan of specimens, I should like to thank Mr. J. Bass
of the Oceanographic Research Institute, Durban, Dr. J. A. Pringle of the
Natal Museum, Dr. J. D. Taylor of the British Museum (Natural History),
Dr. J. G. Field of the University of Cape Town, and the Division of Sea
Fisheries. I would also like to thank the numerous individuals who have
brought me sepiid shells.

REFERENCES

Apam, W. 19394. Cephalopoda. II— Révision des espéces Indo-Malaises du genre Sepia Linn
1758. Siboga Exped. Monogr. 55b: 35-02.

Apa, W. 1939). Cephalopoda. II1I—Révision du genre Sepiella (Gray) Steenstrup, 1880.
Siboga Exped. Monogr. 55b: 93-122.

Apam, W. 1940. Les races de la seiche commune (Sepia officinalis Linné). Bull. Soc. zool. Fr.
65: 125-131.

Apam, W. 1941. Résultats scientifiques des croisiéres du navire-école belge ‘Mercator’. III (4).
Cephalopoda. Mém. Mus. r. Hist. nat. Belg. (2) 21: 83-161.

Apam, W. 1942. Les céphalopodes de la mer Rouge. Bull. Inst. océanogr. Monaco 822: 1-20.

Apam, W. 1944. Révision de I’ ‘Etude monographique de la famille des Sepiadae’ d’A. T.
de Rochebrune (1884). Mém. Mus. natn. Hist. nat., Paris (n.s.) 18: 219-242.

Apam, W. 1959. Les céphalopodes de la mer Rouge. Jn Mission Robert Ph. Dollfus en Egypte
(Décembre 1927—Mars 1929). S.S. ‘Al Sayad’. Résultats scientifiques 3 (28): 125-193.
Paris: Centre National de la Recherche Scientifique.

Apam, W. 1962. Céphalopodes de l’Archipel du Cap-Vert, de l’Angola et du Mozambique.
Mems Fta Invest. Ultramar (2) 33: 7-64.

Apam, W. 1964. Considérations sur la systématique des Sepiidae (Cephalopoda). Zodl. Meded.,
Leiden 39: 263-278. ;

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 281

Apam, W. & Rezzs, W. J. 1966. A review of the cephalopod family Sepiidae. Scient. Rep. John
Murray Exped. 11 (1): 1-165.

BARNARD, K. H. 1962. New species and records of South African marine Mollusca from Natal,
Zululand and Mocgambique. Ann. Natal Mus. 15: 247-254.

BartscH, P. 1915. Report on the Turton collection of South African marine mollusks, with
additional notes on other South African shells contained in the United States National
Museum. Bull. U.S. natn. Mus. 91: 1-305. '

BLAINVILLE, H. M. D. de. 1825. Manuel de malacologie et conchologie. Paris: Levrault.

BLAINVILLE, H. M. D. de. 1827a. In cuvisEr, F., ed. Dictionnaire des sciences naturelles. Paris.

BLAINVILLE, H. M. D. de. 18275. Atlas to Manuel de malacologie et conchologie. Paris: Levrault.

Bosc, L. A. G. 1802. Histoire naturelle des vers. 1. Jn BUFFON, G. L. L. de. Histoire naturelle de
Buffon. Nouv. éd. (Suite). Paris.

BuLow-Trumme_r, E. v. 1920. Cephalopoda dibranchiata. Fossilium Cat. (1: Animalia) 11: 1-313.

Car etTon, H. M. & Rosson, G. C. 1924. On the histology and function of certain secondary
sexual organs in the cuttlefish Doratosepion confusa. Proc. R. Soc. (B) 96: 259-271.

Cuun, C. 1915. Die Cephalopoden. II. Teil: Myopsida, Octopoda. Wiss. Ergebn. dt. Tiefsee-
Exped. ‘Valdivia’ 18: 405-552.

Day, J. H. 1967. A monograph on the Polychaeta of southern Africa. Part 1. Errantia. London: British
Museum (Natural History).

DesHayeEs, G. P. 1830-32. Histoire naturelle des vers. Par M. Brugiére (continuée par M. G. P.
Deshayes). 2-3. In Encyclopédie méthodique. Paris. 5

Férussac, A. de & Orzicny, A. d’. 1835-48. Histoire naturelle générale et particuliére des céphalopodes
acttabuliferes vivants et fossiles. Paris: Bertrand.

GisBons, J. S. 1888. Partial list of the South African Mollusca. Trans. S. Afr. phil. Soc. 4: 201-219.

Gray, J. E. 1849. Catalogue of the Mollusca in the collection of the British Museum. 1. Cephalopoda
Antepedia. London: British Museum (Natural History).

Hoyvte, W. E. 1885. Preliminary report on the Cephalopoda collected during the cruise of
H.M.S. ‘Challenger’. 2. The Decapoda. Proc. R. phys. Soc. Edinb. 13: 281-310.

Hoye, W. E. 1886. Report on the Cephalopoda collected by H.M.S. Challenger during the

years 1873-1876. Rep. Voy. Challenger 1873-76 16: 1-245.

Hoye, W. E. 1904. Sepia burnupi n. sp. from Natal. 7. Conch., Lond. 11: 27-28.

Hoye, W. E. 1909. A catalogue of recent Cephalopoda. Second supplement, 1897-1906.
Proc. R. phys. Soc. Edinb. 17: 254-299.

Hoye, W. E. 1910. Mollusca: Cephalopoda. Jn scHULTZE, L. Koologische und anthropologische
Ergebnisse einer Forschungsreise im westlichen und zentralen Stidafrika 4: 261-268. Jena: Fischer.
Denkschr. med-naturw. Ges. Jena 16: 261-268.

Hove, W. E. 1912. The Cephalopoda of the Scottish National Antarctic Expedition. Trans.
R. Soc. Edinb. 48: 273-283.

IREDALE, T. 1954. Cuttle-fish ‘bones’ again. Aust. Zool. 12: 63-82.

Lamarck, J. B. 1798. Extrait d’un mémoire sur le genre de la séche, du calmar et du poulpe,
vulgairement nommés polypes de mer. Bull. Soc. philomath. Paris 2: 129-131.

Lamarck, J. B. 1799. Sur les genres de la séche, du calmar et du poulpe, vulgairement nommés
polypes de mer. Mém. Soc. Hist. nat. Paris 1: 1-24.

Lamarck, J. B. 1822. Histoire naturelle des animaux sans vertébres. 1st ed. 7. Paris: Bailliére.

Lamarck, J. B. 1835-45. Histoire naturelle des animaux sans vertébres. 2nd ed. 11. Paris.

Leacu. [Unpublished manuscript, referred to in Férussac & d’Orbigny, 1835-48. ]

Linnaeus, C. 1758. Systema naturae. 1oth ed. 1. Holmiae: Laurentii Salvii.

Massy, A. L. 1925. On the Cephalopoda of the Natal Museum. Ann. Natal Mus. 5: 201-229.

Massy, A. L. 1927. The Cephalopoda of the South African Museum. Ann. S. Afr. Mus. 25:
151-167.

Massy, A. L. 1928. On the Cephalopoda of the Natal Museum. Part II. Ann. Natal Mus. 6:
89-96.

Massy, A. L. & Rosson, G. C. 1923. On a remarkable case of sex-dimorphism in the genus
Sepia. Ann. Mag. nat. Hist. (9) 12: 435-442.

MontrortT, P. D. de. 1805. Histoire naturelle des mollusques. Jn BUFFON, G. L. L. de. Histoire
naturelle. Nouv. éd. (suite). Paris.

Opune_r, N. Hj. 1923. Contribution to the marine molluscan faunas of South and West Africa.
Géteborgs K. Vetensk.—o. VitterhSamh. Handl. (4) 26 (7): 1-40.

282 ANNALS OF THE SOUTH AFRICAN MUSEUM

Orsicny, A. D. d’. 1826. Tableau méthodique de la classe des céphalopodes. Annis Sci. nat.
7: 96-169.

Orsicny, A. D. d’. 1845[-47]a. Mollusques vivants et fossiles. 1: 1-605, with Atlas of 36 plates.
Paris.

Orsicny, A. D. d’. 1845[-47]b. Paléontologie universelle des coquilles et des mollusques: 1-392,
plates 11-104. Paris.

Orsicny, A. D. d’. 1845[-47]c. Paléontologie des coquilles et des mollusques étrangéres a la France.
Plates 9-60. Paris.

PenrITH, M. J. & PENRITH, M.-L. 1969. A new species of Lithognathus (Pisces: Sparidae) from
the northern coast of South West Africa. Cimbebasia (A) 1: 99-111.

PFEFFER, G. 1884. Die Cephalopoden des Hamburger Naturhistorischen Museums. ADA. naturw.
Ver. Hamburg 8: 63-90.

Quoy, J. R. & Gamarp, J. P. 1832. Mollusques. In Zoologie du voyage de l’ Astrolabe, pendant les
années 1826-1829. 2. Paris: Tastu.

Rosson, G. C. 1924a. Preliminary report on the Cephalopoda (Decapoda) procured by the
S. S. “Pickle”. Rep. Fish. mar. biol. Surv. Un. S. Afr. 3 (Spec. Rep. 9): 1-14.

Rosson, G. C. 19246. On the Cephalopoda obtained in South African waters by Dr. J. D. F.
Gilchrist in 1920-21. Proc. zool. Soc. Lond. 1924: 589-686.

RocuesrunE, A. T. de. 1884. Etude monographique de la famille des Sepiadae. Bull. Soc.
philomath. Paris (7) 8: 74-122.

RocErR, J. 1952. Sous-classe des Dibranchiata Owen 1836 (Coleoidea Waagen, Endocochlia
Schwartz). In PIVETEAU, J., ed. Traité de Paléontologie. 2: 689-755. Paris: Masson.

SmiTH, E. A. 1903. A list of species of Mollusca from South Africa, forming an appendix to
G. B. Sowerby’s “Marine shells of South Africa’’. Proc. malac. Soc. Lond. 5: 354-402.

SmiTH, E. A. 1916. On the shells of the South African species of Sepiidae. Proc. malac. Soc. Lond.
12: 20-26.

STEENSTRUP, J. 1875. Hemisepius, en ny slaegt af Sepia-blaeksprutternes familie med bemaerk-
ninger om Sepia-formerne i almindelighed. K. danske Vidensk. Selsk. Skr. (5) 10: 465-482.

STEENSTRUP, J. 1880. Sepiella Gray. Stp. Vidensk. Meddr dansk naturh. Foren. 1879/80: 347-356.

STEPHENSON, T. A. 1948. The constitution of the intertidal fauna and flora of South Africa.
Part III. Ann. Natal Mus. 11: 207-324.

THIELE, J. 1920. Die Cephalopoden der deutschen Siidpolar-Expedition 1901-1905. Dt. Siidpol.-
Exped. 16: 433-465.

THORE, S. 1945. On the Cephalopoda of Prof. O. Carlgren’s expedition to South Africa in 1935.
K. fysiogr. Sallsk. Lund Férh. 15: 49-57.

Tomuin, J. R. le B. 1923. On South African marine Mollusca with descriptions of several new
species. 7. Conch., Lond. 17: 40-52.

Tomuin, J. R. le B. 1926. On the South African marine Mollusca, with descriptions of new
species. Ann. Natal Mus. 5: 283-301.

Tryon, G. W. 1879. Cephalopoda. Man. Conch. 1: 1-316.

Turton, W. H. 1932. The marine shells of Port Alfred, S. Africa. London: Oxford University Press.

Voss, G. L. 1962a. List of the types and species of cephalopods in the collections of the Academy
of Natural Sciences of Philadelphia. Notul. Nat. 356: 1-7.

Voss, G. L. 19626. South African cephalopods. Trans. R. Soc. S. Afr. 36: 245-272.

Voss, G. L. 1967. Some bathypelagic cephalopods from South African waters. Ann. S. Afr. Mus.
50: 61-88.

APPENDIX

TABLE 9. Sepia zanzibarica. Relative dimensions as % MLd.

N.M.959 Holotype
» (Adam & Rees 1966)
MLdinmm . . 163 168
13/11) 0 Oa A 71,8 89
MOV, eee es 52,1 42
Pe Ade . kid te | 33,1 21
PAV Viet huit Ge erst 31,9 32
FL 103,1 92
FW 10,4 9
AL I 39,9 30
AL Th. 42,3 30
AL, IT. 43,6 33
Al EVoy 4 46,0 40
oT) ee 128,2 92
Tcl 26,4 21

Table 10. Sepia zanzibarica shells.
Relative dimensions as % shell length.

A2I4i Mean
Linmm . 189 164
i aa 31,2 335 32,4
Th 8,5 8,5 8,5
Str z 68,8 75,0 71,9

283

ANNALS OF THE SOUTH AFRICAN MUSEUM

284

o°GE-9'Es
1g1—3°&6

z‘E9—-S‘ov

VS-L‘GE
£6b-1‘VE
&“gh-g‘gz
1‘91—L‘11
VE11— 96
2°oS—6‘96
1°EE-G‘ao
Z°19-V'Q6
G‘z6-— gl

98ur yy

1°6% €1 UE

1°Qa

C‘or .

6‘0&
o‘VE
Vos
CSI
Q°So1
VI1v
oLa
G‘6P
3°69

II!

9‘0&

a‘SV
966
696
Gos
o‘QI
6601
o‘GP
8°gz
U'1G
016
III

9s
L°Sv
L‘GE
1°VS
L°1S
6°G1
1‘LOI
SIP
BOS
U°aG
‘06
gol

9‘La G‘6a ee PL
I‘t1r | 356 SNS aoa 1 |

901 | SG6 | « * + «ay DE
otS eet . e . Al IV
= 9‘g& ee en:
6°6§ Cbs ° . ° S neny
6°gé | Poe an I'IV
L‘G1 6°G1 * ° ° Ma
1‘9O1 g‘gol . . . . “Ta
696 o‘ov a NET
L‘ao rebey . . . . TH
akeke) 6‘0G ee NAAT
Pg 606 ° 5 : - ATIN

g61 03s * + WU UI pT

9961
uesWy | N || soo
29
uepyV

Qc 6a SVS OoS&6 of9% V'9z Ez
renee | Moy Ponti a= — — - a
1gt | VSr1 — —- — — _—
€G a‘E9 SOM. O6o Gly 1°Cr Cor
oS 6°S Ove tags 8 aly ao lv. eo OV
LV £60 9°96 ofov V6E E96 1°GE
bY £‘oh Sor, G2o Sieg VEL& “foc
1 Lv Ot SOO eal eZ Ole =. abr
96 SOM | oir vor Crt 26 Sor ap TOT
ov &‘QV Coie OOF a Orci mec lve ay uy |
Qz 1°G& 6‘qoz «GLa GHG CGasS CL
1G VLG o19 96S G&S ofr =S°0S
Lg L‘69 G26 0.06 §=6S16 =a ‘z6 69
Vol g&I L9 08 v6 SOD ze aie

zg61 | Avg

wepy |ueqing Azenjsq eushuyy

sg1osy | os 10by

. oO . .
PIW % S¥ SUOISUDIP DANLIOY ‘sopeu Yyv7NIIUULAAA SyDUInf{O Diday ‘11 ATAV,],

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 285

TABLE 12. Sepia officinalis vermiculata females and juveniles. Relative dimensions as % MLd.

Locality
A30131} 430183} A31292}] unknown Knysna Estuary Durban Bay
MLdinmm. 208 287 188 242 103 105 100 102 98 96 156 7p
Muley ee 88,5 87,5 88,8 86,4 99,0 93,3 90,0 87,3 92,9 89,6 91,7 92,4
MUNN gk bs 43,8 50,9 44,7 55,8 63,1 50,5 50,0 56,3 53,1 54,2 59,6 59,9
HL — 24,4 23,9 34,3 25,2 DO, F270 . §25'5 29,6 26,0 33,3 33,1
HW — 37,6 38,8 39,3 47,6 43,8 42,0 44,1 44, 47,9 48,7 50,0
FL 101,0 | 107,3 | 109,0 110,7 119,4 106,7 104,0 OS) SUID 4a 110.4 ts 4
FW Kf y. 16,0 16, 19,4 12,4 16,0 — 12: VIET 13,5 16,3
ALI 31,7 29,6 Diet 55,0 40,8 34,3 32,0 42,2 40,8 35,4 40,4 42,4
AL II 33,7 32,8 29,8 57,9 40,8 34,3 31:0). f 4351 40,8 38,5 50;0 ... 47,7
AL Ill — 36,6 31,9 60,3 42,7 36,2 36,0 44,1 38,8 42,7 44,2 48,8
ALIV = 44,3 37,2 79,3 53,4 44,8 42,0 52,0 52,0 51,0 56,4 50,0
TL Rt 101,0 | 158,5 85,1 203,3 — — — — — — 105,8 116,3
Lt 117,8 | 163,1 74,5 159,1 — — — —_ — — 115,4 104,7
Tcl 26,4 27,9 26,6 36,0 29,1 25,7 31,0 38,2 34,7 30,2 31,4 32,0
Adam
; &
Durban Bay (cont.) Adam | Rees N Mean Range A30128 | A30129
1962 | 1966 Q Q fe) juvenile | juvenile
MLdinmm . 162 149 172 168 147 120 50 . 39
Mily yer. 3 91,4 89,9 92,4 89,9 90 85 18 90,3 85 -99,0 90,0 87,2
INEW 9 fa. 4k 58,0 62,4 56,4 58,3 52 54 18 54,6 | 43,8-63,1 52,0 61,5
HL 31,5 32,9 29,7 SOat 31 33 it9/ 29,0 | 20,0-34,3 28,0 30,8
HW 49,4 48,3 34,9 36,9 46 42 1 43,7 34,9-50,0 48,0 48,7
FL 113,6 117,4 104,7 114,3 100 98 17 108,5 98 -119,4 106,0 97,4
FW 4,8 5,4 17,4 14,9 10 13,5 17 14,7 10 -19,4 2,0 10,3
EEL 38,9 41,6 39,5 44,0 42 33 18 38:4) |) 2it—-55.0 22,0 Dil
ALII 46,3 S23) 46,5 49,4 — 33 17 41,6 29,8-57,9 28,0 30,8
AL Ill 48,1 49:7 47,1 47,2 46 33 17 43,1 31,9-60,3 26,0 30,8
AL yy fae ey eo ieee 58 46 ha S21 37,2-79,3 ine 38,5
t ‘ i 73, Si — hk 117,9
TL yy Ho. seto7a’ ass 13812] || 15° 21 | 129,3 | 74,5-203,3° 1) j000 | 125.6
Tcl . 28,4 28,9 29,1 33,3 — 32 17 30,6 | 25,7-38,2 20,0 20,5

286

TABLE 13. Sepia officinalis vermiculata shells. Relative dimensions as % shell length.

Ago129

A30130
Ago182
A30183

Breede
River

Knysna
Estuary

in mm

ANNALS OF THE SOUTH AFRICAN MUSEUM

W

Th

Str z

43,2
49,5
63,2

69,1

46,7

Saldanha
Bay

Krom
River

Durban
Bay

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 287

TABLE 14. Sepia acuminata males. Relative dimensions as % MLd.

Adam
&
Massy | Rees
A30147 | A31398 | A31399 |N.M.964| 1928 1966 || N| Mean Range
MLd 89 67 89 78 80 67
in mm
MLv 83,1 80,6 80,9 82,1 80,0 78 6 80,8 78 -83,1
MW 44,9 68,7 60,7 64,1 49 || 5 5755 | 44,9-68,7
HL 27,0 3453 36,0 3752 25 || 5 31,9 | 25 —37,2
HW 40,4 52,2 48,3 42,3 41,3 | 36 || 6 43.4 | 36 52,2
FL 95,5 86,6 82,0 9459 87 115 89,2 | 82,0-95,5
FW 11,2 6,0 7,9 9,0 9 5 8,6 6,0-11,2
AL I 25,8 38,8 39,3 3353 42,5 | 25 || 6 34,1 | 25 —42,5
AL II 24,7 35,8 36,0 29,5 41,3 | 27 | 6 32,4 | 24,7-41,3
AL Ill 24,7 38,8 42,7 359 40,0 28 6 35,0 24,7-42,7
AL IV 27,0 50,7 44,9 43,6 51,3 | 33 || 6 41,8 | 27,0-51,3
se ae ree — aS se 105 6 112,3 44,9-182,1
Te 11,2 22,4 — 15,4 15,0 13 5 15,4 11,2—22,4

TABLE 15. Sepia acuminata females and juveniles. Relative dimensions as % MLd.

Adam &
A31398 | N.M.964| Rees N Mean Range A31400
1966 2 2 2 juv.
&. 2 2

MLd in mm 92 92 79 39
MLv . 7359 82,6 83 3 79,8 739-83 87,2
MW 65,2 5453 53 3 5755 53 —65,2 59,0
HL 38,0 31,5 29 3 32,8 29 -38,0 38,5
HW 51,1 40,2 40 3 43,8 40 -5I,1 48,7
FL 76,1 90,2 89 3 85,1 76,1-90,2 82,1
FW 6,5 12,0 II B 9,8 6,5 —12,0 77
ALI 40,2 29,3 32 3 33,8 29,3-40,2 3559
AL II 44,6 38,0 32 3 38,2 32 —44,6 3539
AL Ill 40,2 37,0 32 3 36,4 32 —40,2 33,3
ALIV 50,0 41,3 35 3 42,1 35 —50,0 48,7
Rt — 137,0 —
TLy, eae: oa 95 4 123,2 95 —155.4 Be
Tcl 22,8 18,5 13 3 18,1 13 —22,8 23,1

288 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 16. Sepia acuminata shells. Relative dimensions as % shell length.

West
A31398 | A31399 | A31400 | Pondo- N.M.965 N| Mean Range
land
Linmm 72 92 +39 +108 iol +77
W 38,9 38,0 | +48,7 | +43,5 | £35,6 +39 6 | 40,6 | 35,6-48,7
Th. : 9,7 9,2 = 10,3 + 10,2 9,9 Sis 10,4 6 10,0 9,2—-10,4
Str z 59.7 68,5 | +£64,1 | +64,8 64)4 = Ot 6 | 63,8 | 59,7-68,5

See opposite for Table 17

TABLE 18. Sepia confusa females. Relative dimensions as % MLd.

A31403 Adam & Rees 1966 N Mean Range
MLdinmm . +53 77 85 80 84
NE oe eee 73,6 80 79 75 81
Mw 49,1 36 33 35 33 5 3752 33 —49,1
1EGL 26,4 22 23 19 18
HW 32,1 31 33 30 33 5 31,8 30 —33
Pi 79,2 87 79 81 80
BW 557 13 13 12,5 8,5 5 10,5 5»7-13
ALI 45:3-|--38 33 34 30 5 36,1 30 —45,3
ALII 453 | 34 32 35 32 5 3557 32 —45,3
AL Ill 49,1 34 32 32 32 5 35,8 32 —49,1
ALIV 50,9 | 36 33 35 34 5 37,8 33 —50,9
Th — a 82 pa pee
Tel om 13,5 14 aa cas 2 13,8 13,5-14.

TABLE 19. Sepia confusa shells. Relative dimensions as % shell length.

Adam
&
A2140 |N.M.961 A31402 A31403 nae N Mean Range
1
Linmm 3. . 88 +141 106 +4105 +93 +104 +53 84
Wear er 19,3 1556 17,9 18,1 19,4 19,2 24,5 19 8 19,1 15,6—24,5
5 1 o Sea ae ee 9,1 6,4 8,5 7,6 8,6 Wi) 11,3 9 8 8,5 6,4-11,3
Str z Pig Be Lee app 65,2 54,7 58,1 59,1 — — +62 6 39; 54,7—-65,2

289

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

— +1°GE1 | &96 — +009 +6%%g +V‘LL | +8&‘Q& G‘°Go1 | o'S6 Ges ken
€*g1-66 ZI | 6 OOr sr‘o1 V1 G1 -- €‘g1 a — 1°C1 G‘o1 6°6 L‘or ey one HOWE
ofoo1r oGaI 06 Sg =e io" i: == fas ay Oa ae a e "IL
bgt gz | 6'9f | E1 |] ofoF EGP o‘ob | of 1& gz | oGb eb 66h OS GP Q‘L& ob 4 L‘o& TN aL
o‘Sb-1‘Le | PGE | €1 |} ofg& SE give | gz 6z 92 |*S°e6 = Otr Gah oor &‘gS 9‘a& 1°LG Ley
1‘gh—P‘gz 1°96 | 1 || g‘g& ofzP G96 | of 6z Qo | TOV 4 OGcy- Veo | {O0V &°GE 696 Vga * “ESTAR.
6sS— EE | ssh | ar |} OLY a‘LE 6‘9h =| SE cE 6& | 62S — g&P 1‘6P 9‘Qv 1°1P L‘GE oN ay
6°L1-9‘L 11 | O1 GI GSI G1 | Sor £6. G8 9‘ E11 9S 6°L1 ANT
— — — Lia 29°09) O19) S557 S‘9g9 1°06 L‘Ge eee |
G‘gE-6'0z | 9'gz | aI 6‘0z oGo | gz Lz o& | Gof = PSHE COEHE OES 1°1G Qvz L‘°Gs im MH
Gz Va 66 | Sis “ote 9a. Sire 66% E1% 00% es he ae
o6F-1°t& | 9°6E | 11 — — o‘ov | SE 9& OF | OGP “TOV bol § a266 1°66 6°96 1°3& = OE cues NAIA
6*zL =6°g9 6°99 | o£ cL Sla| “GOO OGL, GGL 1°G9 6°GL o‘GL 6°LL "Se ATTA
obi ght GE1 | oo1rF SorF _ 06 Vol £6 Gor gol 6&1 IVI ovi | ww ul py

osuey Ss | ueOP | N gz61 Cz61 gg961 sooy 29 Wepy covisy ZOO WEN II96''N | 91S9V

Asseyy Asseyy

‘PIW % Se suotsuowIp sanepoy ‘soyeur vsnfuos nidag *L1 ATaV J,

ANNALS OF THE SOUTH AFRICAN MUSEUM

290

1°SS—G‘SI

xo SOI— 6G
9 °SQ—-6 ‘QE
%9°19-3‘SE
40°L31-S‘99
I1-2‘9

V‘Q9z-V‘0z

LE—-L‘o€

GI

vr GI oo Lal aor 1 ae
O'VL .-&°Sq TeOGT op

98 96 cay ya al ace 1081
69 6S LV OVE VVL Zz °sO!1
66 g& TOW Ov gSb S‘Zo
66 LE Gra Le. OL. Orig
bg 19 UNO meno O olor 6*La1
Teta

Gz ve ViOGs oL°66

Lig iGo

bE LE

VIE Jh0) Dus oo ean ciel ®)

ov SEI Lvr oG1 Cot ..g61
9961

SOY Qc61 Asseyy Cz61 Asseyy
IEP,

‘PIW % SB SUOISUSUNIP ATILIDY ‘sopeur vj1a9u2 vidas *Os AIAV J,

*popnypour jou (wut 66 pyPY) seu Sunox x,

Ver — Gr
= — G9
G00" = = :6'Zor
vEQ 1°99 ‘og
ELE LE V6
oGE V6E g'9&
ogg S‘gg Pgh
ZO 2 Ol o OM
iMate) teri) VERS)
cs ok I oY chk To
CFG ae Ol | olor
Lote ay SG LsOe
Cfo a on =a
aiae ellie wera
Ehiosy

9°18
6&9
o‘oV
9‘or
oS
&°9

0‘S6
HN
261
oSGE
0‘G9

Oa!

EVI
o SQ
S‘99
£°9&
oL&
9‘S6
3°

Q°b6
6°E%
6°6z
EVE
Z‘SQ

gv

GSI . . . eae
G19 . . . 1

eile Weare LE
O09 le he AT.
CLG lo) TT, Ti
S198 la aang
chek en Ut Sia ee | 57
C6 e 5 . Ma
Of6— | a ME
Caz ° . ° TH
PPS . ° ° MI
Cbg A . . ATW

QI * WU Ul pT

LS6-TUN 9S6° WN

a

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 291

TABLE 21. Sepia incerta females. Relative dimensions as 94 MLd.

A30143 Massy 1925 N Mean Range

MLd in mm 82 go 80
Miivy a. 85,4 ,
MW . . 36,6 48,9 53,8 3 46,4 36,6-53,8
HL ». 18,3
HW 26,8 3353 30,0 3 30,0 26,8-33,3
Fis. 92,7
FW . 8,5
ALI 26,8 5454 45,0 3 42,1 26,8-54,4
AL II 22,0 57,8 61,3 3 47,0 22,0-61,3
AL Ill 31,7 64,4. 56,3 3 50,8 31,7-64,4
AL IV 3554 60,0 62,5 3 52,6 35,4-62,5

Rt 87,8
TL i ae 14.2,2 —
Tcl 14,6 1G. 7 -—— 2 15,7 14,6-16,7

TABLE 22. Sepia incerta shells. Relative dimensions (approximate) as % shell length.
Many of the shells are broken.

Massy 1925 N.M.958 N.M.970 Punta Zavora N Mean Range
No. 8 No. 16 A B B C
L in mm 135 125* 148* 146* 81 65 106 a 66 49
Wee es 17 16,8 14,9 13,7 19,8 20,0 17,9 20,8 19,7 22,4 10 18,3 13,7-22,4
Tieti* 8 6,1 5,5 7,4 EY 7,5 7,8 7,6 8,2 9 13 5,5-8,2
Str Zia . 270 266 57,4 47,9t| 67,9 56,9 69,8 62,3 56,1 63,3 8 64,0 56,1—? 70

* Length of shell if complete
+ Part of striated zone missing; percentage should be higher. These dimensions excluded from the calculation of the mean
relative length of striated zone for this reason.

TABLE 23. Sepia burnupi. Relative dimensions as % MLd.

A6525 N.M.4073 Mean A6525
3 3 3 2
Mild inimm’... .. 44 45 36
IMM erie BN 90,9 QI,1 g1,0 88,9
MW ain =) Bids 40,9 40,0 40,5 47,2
Jeg Lye ei * ae s 25,0 24,4. 24,7 27,8
EW yey 3 Beg 36,4 35,6 36,0 38,9
ABNEY ge dy Sy ae ae 102,3 97,8 100,I 100,0
OW caheoeecss oh. Bema ts 11,4 Id yl 1.3 10,1
ABT nga 2) ee: 4352 3353 38,3 30,6
7 big er er 31,8 24,4 28,1 33,3
AT Vil Ot Oa 38,6 28,9 33,8 30,6
fod OA gee ee ee 545 46,7 50,6 38,9
Pet 8 ke —- —
Peay ik ae 68,2 ee 69,7 83,3

Gl). seit ME giade ce 11,4 baa 11,3 sh

292

ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 24. Sepia burnupi shells. Relative dimensions (approximate) as % shell length.

A2147
Linmm AT* 41*
he 2555 24,4
Th 74 6,1
Str z 85,1 78,0

* Anterior tip of shell missing.

N.M.958
€ N
50*
24 4
7 4
74 4

TABLE 25. Sepia joubini males. Relative dimensions as % MLd.

Mean Range
24,4 23,6-25,5
6,7 61-734
7957 74 —85,1
A30141} A30142 A30172 Massy 1927 A31393
MLd in mm 29 34. 36 36 Al 39 = 39 40 33 35. At 4k
MLyv 79:3 | 85,3 | 75,0 80,6 80,5 79,5 84,6 77,1 82,9 75,6
MW 37.9 | 41,2 | 38,9 4157 31,7 - 38,5 3855 40,0 36,6 36,6
jeu 24,1 26,5 | 22,2 22,2 24,4 20,5 15,4 22,9 19,5 22,0
HW 34,5 | 35:3° | 3353 3353 31,7 28,2" 30,8, | 130,0°°°'36,4. {eaten 7 eegieg
79:3 | 82,4 | 83,3 86,1 85,4 87,2 84,6 82,9 80,5 80,5
FW. 552 BB, 558 he Osu Oe a> dw ealoth 8,6. 7,3 733
ALI 24,1 324m) 22)2530;00 2454 025,00620,2 10,0 ti 25,7 24,4 24,4
AL II 24,1 20,4. 1. 2232°90,.6°994 4 0129.7 630.6) 87 Fe 99,9 2930 220
AL III 20,7 20,4" |"25,0°° 27,8" 19)5 23,1 “25,6 "|"40;0° 3034. ("22407 = 22-ameern
ATTY, 17,2 92,4 | 22,2 25.0 22,0 23,1 25,6 | 45,0 9 45,500) 2h 2aeeeee
LCi 110,3 — 91,7 — 117,9 125,60 | — ; MS 91,4 100,0 78,0
Hie 89,7 Tata ay rat 7352 102,6 one j 94,3 104,9 82,9
10,3 — — 13,9 12,2 10,3 7.7 | — 12,1 10,0 9,8 9,0
A31393 (cont.) N Mean Range
MLd in mm 39 36 34 Pep oP igor) van 20
MLyv 7454 177;8°-70,5. 95,0 78,9 70,7. 76520 70,0
MW 35:9 38,9 38,2 42,4 42,3 40,0 47,6 35,0 18 39,0 31,7-47,6
Ai 17,9 22,2. 20,6 21,2 “23:1 \ 20,0
HW 30,8 30,6 32,4 36,4 42,3 36,7 18 3304 28,2-42,3
: 82,1 77,6 © 82,4 78,8 "80,8 So,0
FW 5,1 5,6 539 6,1 11,5 6,7 16 733 5,1-11,5
ALI 25,6 27,8 20,6 21,2 26,9 26,7 17 26,5 20,6—40,0
AL II 20,5 25,0 20,6 21,2 26,9 23,3 17 25,1 20,5-3755
AL III 20,5 25,0 °-20,6 21,2 30,0 2953 18 2555 19,5-40,0
ALIV 23,1 27,8 23,5 24,2 34,6 23,3 18 27,0 17,2-45,5
Rt. 94,9 100,0+ 105,9 75,8 — 86,7
Lr. 102,6 102,8 111,8 — — _ 106,7
Tcl T0,3- (13,9 11,8 [2,1 =" 1353 14 1,3 757-139

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 293
TABLE 26. Sepia joubini females. Relative dimensions as % MLd.
A3Zoi41
MLd in mm Se Mie geen, mG & afSt 95. — Bon » B56. 85m. gag -~ga B89
276 Ol, 5) 62,00 82,1 , 80,6 ~82;9~ $1,868 80,0 ‘82,9 “81;6) 81,6 82,1
35,0 40,7 41,4 42,9 41,9 40,0 39,4 40,0 37,1 36,8 36,8 35,9
29.5 25,9 20,7 25,0 25,0 25,7 24,2 29,9 25,7 23.7 21,1 20,5
32,5 40,7 37:9 39,3 38,7 343 36.4 343 343 28,9 31,6 33,3
82,5 81,5 79,3 82,1 83,9 82,9 84,8 77,1 82,9 81,6 78,9 76,9
5:0 74 10,3 Tot 6,5 751 9,1 597 597 ar 593 6,4
Shas 8998 a 27. 220.6 Gago 1314.0 27a. eo OSI k. 6 eG. 2B.97 Ba.9
35,0 25,9 31,0 28,6 25,8 34,3 30,3 31,4 45,7 36,8 39,5 41,0
37,5 25,9 31,0 28,6 29,0 37,1 30,3 28,6 42,9 39,5 39,5 43,6
30,0 29,6 31,0 32,1 29,0 31,4 33,3 28,6 34,3 31,6 34,2 30,8
90,0 140,7 sa 107,1 E129 (949-7 -67,9nie 100,09 a 70,0) =F
92,5 107,4 96,6 110,7 — 122,9 87,9 105,7 100,0 84,2 84,2 100,0
O90 14,0 —igG 14.9 j12i9 14,39. 12,1 “14,3 S18,9> JO)5 13,2 ay
Massy 1927
(Adam
& Rees
A3or141 (cont.) Ago142 |Agor72| 1966) N | Mean Range
PF11741 PF10715
MLd in mm | 40 40 43 ah) 32 39 47 36
89-5 Goo" 76,7 66,1) $755 82,1 83
35,0 32,5 32,6 | 38,9 40,6 | 38,5 | 32 19 | 37,8 | 32 —42,9
20,0 22,5 20,9 | 30,6 40,6 20,5 25
30,0 27,5, — 33.3 3454 3353 30 30,6 19 | 33,8 27,5-40,7
80,0 80,0 76,7 | 80,6 90,7 87,2 85
By BS 0,3 5.0: G53 Tal 8,5 18 |) 753 5,0-10,3
30,0 30,0 2759 27,8 43,8 30,8 30 38,9 20 30,8 25,0-43,8
3755 37.5 3752 | 38,9 43,8 | 43,6 | 38 47,2 20 | 36,5 | 25,8-47,2
27 Gior 87520389 3453 1| 43,0 | 38 47,2 20 | 36,2 | 25,9-47,2
32,5. 30,0 2759 | 30,6 43;8)| |28,2 | — 38,9 19 | 32,0 | 27,9-43,8
90,0 95,0 81,4 ae . 7 94,9 72 127,8
G75, PORS+ 14 ls 93:8 | 1 92,3
15,0 15,0 11,6 | — 15,6 12,8 935 13,9 19 | 13,0 9,5-15,6
TABLE 27. Sepia adami females. Relative dimensions as % MLd.
A31394 A30149 N | Mean Range
MLd in mm 59 47 44 28 24 25
MLyv 76,3 74.5 72,7 7830 75,0 = 80,0
MW 40,7 38,3 38,6 39,3 45,8 44,0 || 6 | 41,1 | 38,3-45,8
EVES SE 2% Ach 20,5 25,0
HW 35,6 36,2 36,4 39,3 4 | 36,9 | 35,6-39,3
BL, < 86,4 SOS! 77.3" 92,1
FW . 6,8 6,4 6,8 omit 4 6,8 6,4-7,1
AL I Be Pe 290,85 7937.5“) Bro 4 28,6 25,0-32,2
ALI 30,5 25,5 25,0 21,4 4 | 25,6 | 21,4-30,5
AL III 28,8 29,6 .25,0 25,0 4 27,2 25,0—-20,8
AL IV 32.2 29,8 31,8 25,0 4 20,7 25,0—-32,2
TRE Rt 98,3 63,8 Te 85,7
Lt 78,0 78,7 88,6 92,9
Tcl 13,6 P2Cobet2.5 1458 4 13,3 | 12,5-14,3

oS Ser St gol L‘gI 6°E1 GCI EVI oie ay,
9‘19 L‘9g ~—s« ‘0g solr 68611 1.98 8 9'V6 == 2 a
ofoo1 | SEQ 694 Lisor =r1Stt1 § §=gsso1_—- 1 “GOI L‘orr ae S| LL
oVS o.0& ~=—s-_- 6‘ga 1GE = gf EOor -40Le 9‘gz ~ ALIV
g‘1& of0& 90S 1°GE 6&6 90% g°Lé 9‘gz = Tit ow
g‘1f Liga = 6g GE go gQ'0& gQLE V1z I) ea Pf
6°gz 0‘o& 692 Vos GEE 9‘Lz 1°GE €‘V1 : l'IV
GE 6°S1 SGI 9‘O1 Tap £°9 gol L‘or * Ma
G‘69 0.06 =—s« Gg 9g «=: QS s«HQQSCéOG “16 L‘Ge sme Tete Z|
s 1°SP Lig9f = a‘ SOy = Ly = OfeG 1S*or 6°zP MH
2 69% L‘Qz 1°SS L‘6% oGs oGs Eve V1z pms |) "|
S L°bv SSb sob of9h of0G 1ZS =o TLy, STG Goe = Vir S Vr aay Veg =GOh = GOP " MI
= s9g_ | 4og = Sag. tess EEG LGQo*bg | Vag SCOgg «| gg r'9g Ss ‘og. S‘9g | sgh = gg ‘gd = ASHI
A ge o& gz Gz va VI Gz Lt gl L& 9& 9& LE 1G Le Qo WU UI PTI
s Rapper | cee ee ee ee a ee Se a 2 dk ee ee ee
x gS10by bGi0ty ZG1o&y 1S1ofy (-yu02) oG10&y
<
se
5
2 QF Tt Gols Oral EF — Sv = Vor or 26°21 a G‘gI eds ; PL
a) low Ti S= g‘zo1 | 96g — ofoor gor gfs11r = L‘Lo1 — Q VII eat ete
a @OO1; 1°11 Scot, £66 EYE — vizor 1°96 866546 ~— ‘a0 — 111 ; gin goehs ut
a 9°6z 6Ga o0‘0& 6S L‘o& — octg Las ove Gene 13S 962 a ee ’ AL IV
© SOG | 6.66% Gu Fags L‘vS = ogo Ss «Lis&— HIE — SOE 9°92 ~=—s_-o 6 > aie ie eee ee ~ TIT TV
i. gcc | Zee oof = 9G eee — G0 mag Oo = 19'0o8 89.06 9°92 ~—- 96 Sao aac" Il IV
S eee | GGes SSC ofSa si Wah So 2S Oe Os ee 6 0e BOGe 99a —s-«o6z Biot ang tailed l'IV
vd I‘rt oe oO1 gal o‘OI oOO1 G‘6 HONE, G‘9 €‘o1 ‘V1 QV : : : " MA
a Sens QL. 0°09) Onn cp Ochs OTO = aIcGe 1GQ o‘OOI ABI OPE IE : ee Ta
VAY WU = OSG es oor Gaye oor Sor Soe ocr ofS? vov 1‘9V ; : MH
Ger 6% See Sow kt a hic OD OD | 261 61% 1°&% Miz - 6Gz ce age ; : TH
UV 4) el OV a aS Glen! a Gea bicva (Oo%sns GOY — Got oar OG oS Vor 1°9v = 009 ie em 2 ; MW
SC 19.t- plo OneG 1°Q gle ogg Sg 8 Lag 1GQg — o‘¥Q—Ssé HQ 1%g 026 CG‘1g—Ss« OL ss rid : ATW
Le Lo ov 6§ 6P oS oP oS LY 6& Fa Qa Lz 0% ees : WU UT PTI
oS10fy grioty

294

‘PIN % S¥ SUOISUDWIP DATVLPOY ‘soTeu synajsnv vidas °Qz ATAV J,

295

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

gfoz-g‘or| L‘&1
9‘LE—-0'Sa | 1E
Q‘LE-z'ss| SSS
9°LE-F'1za| 062
1°GE-E*F1| 6'9z
bG1-1‘9 | Sor
of0S-£'EE | S‘ob
1‘1g-0°LE| L‘SP
osuey | uray
6°E1 61
arog. Ores,
— oGL
g‘o& 6°SE
SSE EEE
go& = 0%
glo = SEE
rr rit
g‘LL = SEQ
SSE = =LS1h
o'Gs 991
6°96 = UP
gill = 90g
g& gf
Footy

fest
ag "IL
Si NU en
III "IV
II ‘IV
ie cera |
» + Yong
MH
’ "TH
MW
ATTIN
CE OT ALVNL

. .

uaa bl bey

. .
. ° .

"ATTN
ui UT PTI,

Q‘II P11 6‘o1 ote Sol Gor 6-1 1°SI een Cer

€*S6 VLG 1‘g oSO0I L‘s6 — $96 o’Sor1 | 9‘GL =

€°C6 L‘Ge bog o%sO1 3°06 906 L‘G9 ofoo1 | gfSl L*sol

6‘La Pons 69% 9°Qz DEG o0'Gs 9‘Qz bof akeys) 1°GE

9‘a& P11 1°Qz G‘gz €‘6% SAG 992 alee) Gee Vos

a‘oS PE 1°Qz VESS LST S oGs 9°Qz SSS BLS V‘st

6‘Lz PE bo 99a €‘6% oGz 9‘gz 6‘oS G‘oS bss

€‘6 9‘9 G‘9 1°9 SSI v6 oy rap reo a oe

b°99 6°s9 6°SL bl z‘06 b‘b9 L‘Gg I‘901 | 9‘bg 769

C6E ISLS oL& L°vS 6ST G‘LE LSGE V‘sV bof Q‘LE

6‘Le 6°ss ay a &‘QI C61 9°Q1 oGs ove 1S 691

9‘eh L‘GP CEP 6‘bP 6ST OEP bot G1G CGP a‘SP

L‘E9 TLL 6°91 GCL 9‘Le 1‘9L 9‘gl 6‘06 9‘9L G‘9g
Pv cg gv 67 Iv aS re && So LE

boSofSy |ggzoby GLiog’y oL1osy ggl0by Lo10&y |Sgioty

g‘b1 Q‘II 9°05. Via O61 gi OSI Gar

HOO.) 34°94 Gort | & VOL o-Git} cVOY =2-coL | G00

G06 = be LTO I-on1 9°26 7) o'o0r “Sse 9°66

kor S96 Ebr — “VO = =-O°GG > GIS 6G S1E

Liza = V6% c6e 4 86—Vo& 8 = & ‘gz 1Le 96% 61g

Lo = 6 26a = V0& 1°Qz 26a =. 9 6% 1°Qz

Slo = 6% 0°Gz gz. Giger ei6s. - gids 1‘Qz

16 Q‘II £9 L°9 G‘9 &°g Vee v6

91g = gg OGL iol Sgt [ene Gro 6°19

o*0S 1‘LY OGY Keli 70Mls | GIG ~" O2e8 G‘LE

gis 6a omic Hos iGfSc b ioe —~ SiG1 6°12

GS of0G eros — {Sig& iOS | GOS —“Or0G QEP gSb =: Oz

19 ZaE mol, jolie - op | Se ~~ Gare 61g 66L tae
oS VE tz &% ov QV FG oS Va Li

Egqg10by IQ1OSYV ogioty 6S10by

(panuyjuor) 9% ATAV I,

SV Q‘1I Q°o1
— Gy, PULLS)
o‘og)=—s«sW‘zg—StséL: ‘650
O06 = s6ee = 0'05
&°VE 6% GSE
V16 > —S‘Qz g‘o0&
P16 Sees OS 24006
V1 Q‘II 6‘OI
6g agg hg
‘VS GE G‘9E
SV O'0a © = Son
o‘ov Ain OcIP
00g £°G9 1°ZQ
cE ES 6§
(yu02) booty
6°s1 (AG oom
6‘06 6EOI sal
1‘L9 — I‘11I
(Ser Pane WIS
6s GLa ene
06a ‘Se HOG
0°6z GLa EEE
L‘6 36 6‘9
PLL Li9g9 ~=—s OGL
L‘9S Ate 9‘GSE
9°Ga O.1s 00%
617 SLE o‘oV
9‘0g bsQ b‘b9
1& 1S Cy
LSi10&y

v1 ot ZG — L‘g1 aL Gal wt ol aoe 1‘QI at ob EVI L‘g1I 1S Pe Rae 8
G6S oGL &‘L9 = — gE11 | 66 69h 1601 oOf001 — gvor | a6s1 a1 eta
—_ o‘Gg 9°19 — oot, TOs! 26g O'6S Goo. | 4*601 ~OfSGar 3°S6 gosr = aL 11 Sanaa, IL
‘go G‘Lz Geo) O25 = 7g 6 O26 Bo a OG =050G- a ca 95" = O° 9s os 1Vs regen) lel Ng
gogo GLa C‘gz ge O66 ~off¢ OGe-. @06 | G'Gq “Sots Z'oct- o'1s 0Gz 1FS 1 Spay.
ogo 86G°Le 9‘ga CVS SEE gfLe 9G6z ss 0G Gz o6¢ L’go orf 90% L‘o% ~-<* PRay
zw‘go Ss G‘L& Oige- | Off EGE xiaG ogo Qo SIE ESSE L6gz == 1 gos ~—s L ‘0% eee (ag
6°11 GSI oS Sor o‘O1 6‘o1 6‘o1 Lat 9‘9 G‘9 6‘9 G6 Ga Q‘S1 = SNAG
S06 Ss S‘z6 6°66 Lg ofo01 32‘g9Q alg 66 #98 TWAOs el Loe = 26G0 SiGor. - Meee tS ag
s 1°96 = oGE fo hte} CS A Gt A ht 1VZ Giotn aes Oia LOY ots Fa LiGs gSb gb TS NEE
a o61 002 Os @Gi4c . O66 +, Iho nes |) ogo | Kas eG, the | pine \belog = ade Soh, REY
Z och 86G‘sP gsb oSb 6b 60GGOL‘gb = ‘bP Ger) ocr | Goh iver orby | ofaP zvG = & ‘9h * 2™ SAAN
a nie | OR | “icp (veg © gibg | gieg «= GSggtsp | rag 6 6t¥g = gttg «i icdg «6 Log «= LGge | gfol- 6°G2 on &, SATIN
a ov ov 66 66 60 6z o§ 62 66 66 a 16 cP ov Vz 6z UIU Ul PTI
a eee ee | eee eee ee ee | ee on Pe se Pn ee | ee oe ee de
ow PS108y ES10fy 1S10fy (-7u09) oG106y
a
se
5
o ORG ieee Te 2 OTe AGG or 6for o%%1 SOE YEON OAH IE Ste ae OIG At eee se OT
ca Orr Gli is cor 209 — EVS ool 6Gz1 &%9 ofol 1-o@ 2 Goon WSe ISO) 2 Gro1d 1e09 ; VT
2 o‘gz1 1%%g 8 8=6—L’gor =S6S) = =6—P‘gitr =~ 00g g‘6z1 S‘Lg ofoL S*Le — 6°26 1611 Vgor 9°19 : tbs | AE
fy O'0G "VIG Sita 061 [SE ome | IN ie = OSS OCGse score OO Oo GSoe One hao TTY.
C) Duce Oc tc Eyam =) oge tiTmEe oven |toCsn fGen ofese eosbe 16s re Lilo Ye O1S ab oGc Se Ge rere EET
fs Dou Ofc |Gec Bite 1 oom ion, ober |16Gze Gee “ogo ofgs00 PGs, Lilo ao 618) cn 618 20 ole Il IV
s Oca REO Gass. OfOG) y Vara) AGce = FoG1 0°9z GGice i SGe OO OOcts Voate Lilet) ovis Oro, 62a I IV
4 OSI SV 6S 6°11 oS O61 001 III £°9 08 6‘g QI 9‘O1 G‘9 Q‘aI L°st eee |G
0109, 76695) Log * StoG’ of1g (otlee o0'06 IPL COAe Ooger oOOLRh SoZ Lian e600 2 Os ee o'og : esa it |
o9b = «V‘gh SEP Eger -| SOc “OG. ~ Ofgb Temas ree O'Gvae -S° LG Vie Viorel Lh 2 Von OOK Pe ee ay VJ
0.92 V1G-—s«OB L‘ol 6°Qz o‘G 00% 9‘6z 64a i OSBB_—s—“‘éi@* SG G‘E% G1z LAL Gabe GSGa, tL ror : * sa
09h = o0G Loh «=86'ch 8=60—SLD CGE oP 61G2 “Lv ober Sle LV ee OCR. VO 1 Owen OoU - MN
OGoe.tsG Ghee Ligls “ise, “| oer iocae o9R SiGe “boom ‘ofoO= Ocore T2978 1co.8e Pog. oleuee ocg St = SAGTTAT
sf Qo o§ av g& ov oS Lo 9b oS 9S Ss LY Lv LY c¢ WU UT PTI

296

gbioty

‘PIW % SB SUOISUDLUIP DATILDY ‘sopeuoy sapvsjsnv vidag *6% ATAV.T,

5 g‘Il O11 P11 o'G1 CF O91 o‘61 L‘g! Q‘II g‘V1 g‘V1 gial Qa oe WONT;
o Gk , L‘S9 Big §=6o'0g’—Ss« QQ OOM w'gb- ign. “Gg! Vt GeO se FOL oun a
0°66 1°99 — G‘zg — 0.96) = « L‘Lor = =a‘vor = 999 6vor Sg S‘osr 6°F6 * Sr LL
< 6°Lz Q‘Qa ELS of0& 6‘gz 0'Qz g'o& 36s &°EE 9‘9z 0'GG 1°&% 9‘o& : AL IV
ic 9‘Sa 9‘gz G6z = o0& £‘gz ove §g‘ok 56% GEE 9.95 («OGG Ss 9Sz III IV
pe 9‘Sa B‘Qz Lo = o0& g‘1f ove gf06 gos G‘La 8.9% ~—s_ «Qa G‘oz BQ mee, (ia in
< 6Le = Gz 6‘La G‘Lz 69% 09a ~—s gS cl oY al OY VVZ 02% G‘oz 1% eas (a 4
S Qg‘II 611 P11 oS) oS O%GI Gri £9 96 96 GL 6‘O1 gal = KK
= of6 = L “Gg 99g S‘z6 8 S‘6g ofl it oy em to HOE &‘bg G‘og. Sq — offe “=: “TW
> C*6E 1°g& 1v& ofoF 90% oer. sor Sr “SLE go g'9S G66 6°CE NET
Ss 9‘gI Q°bs Lizz OG 1°1S 09a ~—s 6‘gz 56% GSS G61 Oco p OUI 1°E% ee nel
a Gob 6aP ete pShics hie L’ogsrs oc 00S "a'0S 1°EP OOpS OOS “SOY wagner. Sint ~~ Ginn © = RATN
5 Gf _..0°T9 9°98 of0g_—Ss a ‘FQ ea COR I hG rg/ S oe OO. OSL Oe. OF ote aire - * ATIN
= &P oP bb ob gs oI Gs Qs Vs 1S ot 0% 1v Iv 6& 66 WUE UL PTY
6 Ggioty Po10ty I910fyV ogioty
q
<
=
oy
a 6‘orI o‘gI 661 611 BSI — gb L‘gl 1‘QI PCr 9G 6°61 PGI “eee
= 9‘tg OfOOI OCOII Q*QzI a‘gor — o9g of001 Ofoo1 og 99 )6=6QftS ~~ = (‘tg sae at ||
a bog — ofo1r S601 ooo! — IGOR IG EGR Egor. 169. 790 [Tope ole oe ORT WL
Q 6°Ez Lo Ligt §=9Qz—s 08 6Qz Olae WOR "06s ces" “60e"4 io'feee OSs ° . NOW
e 1‘gz Slo EE gs 992 L°Ss OLoe, 14 00m O62 OS%>% Gao) Oeare Ge " . Til Tv
na 1‘Qz oS 6s o°Q6 = ‘Qa 6‘gz Q‘Lz Louw 0 6c 16% =: 6 ‘Qa Q‘La 9‘GSz ee | aN
= 6‘Sz VE = OC0G”—“(iéiB'QU—s«Q'QU L‘Ss glee Liget ec t9Ge Gos oGeer ofG6 6 AW
H L‘g 16 L‘9 6°11 2‘Q EG QS 6G 6%a1 Q6aI oS 6°61 Qs fe te hi
a ge Lisg, «=o — RPTL &‘L9 G69 E69 GS6" G06) E'cby, GSi6g = 76 2°69 eee 4) (15
2 eae v6 ofob 1°96 L°VS G‘6E ive) Of0V= Glp SSO “Ooh = eb.aG G‘9& ; MH
is 9‘61 avo 4 o0f08)=—_ QEB—ts«US'CG] 1°1G Oa. Bel Ont Soca s -Si0Ge 7 Ver V‘61 1°S% $ TH
> G‘EP Sve for -°Gfav-< 6'or L‘by TLV Ooo Soe Lis eo 0S = Oy. “coy Ly ohh Oey MN
= g‘rg gig Sg Sg 964 avg Vege S69" 6°06) (9'cg ea Loo. Gta ica sve i997 -o Le ATI
< or && of ov 67 g& Ea 61 ZS 6z of 1 66 g& 9& 66 WU Ul PTI
6S108y LS1ofy GS1o$yv (quos) bS10&y

(panuyuor) 6% ATAV J,

ANNALS OF THE SOUTH AFRICAN MUSEUM

co

29

2‘61-G ‘6 VT i) VO) |eor 061 Tee Gor Q‘11 QI etn ZS ov1 fg ry ley Sam
— 6‘ro1 | 9'GG G‘L9 BLO » Lok 6°66 g‘LL 1°oh 9°69 pe ae
0°69 Gmon (Recs | 216 L‘ol 16L =‘sQ L‘&Q 1°9G o1G Ds | TL

626-061 | G*Ze 199 || OfLe 6°Gz oGe =: 6 ‘zz 6Ee Ge Loz G‘gz 6‘La 6‘La “So LAUD Y.

v‘GE—-0'0% | 6.9% | 99 || 9Lz 1‘VZ ZS SOG 9Gz = ‘0G ~— 6 ‘gz 9‘Qz 9‘GSz 9‘Gz roe 6) ee 1g

G*vE—o'g1 | L‘gz | 99 || &'6za 6Gz oGz = s «6a OGG Gila 240 G‘gz 9‘Gz 9‘Gz "ss. A Ue

o‘VE—-9‘61 | 99% | 99 || S‘6z 9‘6z oGe = 0G gGe Ss SEs LL SQ Cbs 9‘Gz 6‘Lz \~ *) Lp ay

o6G1-£'G L‘or 99 || 1G £6 9S &‘9 66 66 I‘TI o‘OI Q‘I1 QI eA
Gg Q‘LL REL e sOrOL 0.98 ~=—s-: 09 1°16 9‘19 b‘99 i dgk (2 yl edie eee ey =| |
1°9b-1°va | SOE | gg || BLE o°L& fouy svc IG IS MOIS L‘g& 6°VE G‘6E oe a INET
061 1‘V% BSS 981 9‘gI 6.03 QSL I Dos 6g 9Gz eg eT ea

L‘gg-S‘L& | ofgh | ZL || L‘6E L‘ov bbb 966 A a hd a aE Hd 9‘ov GcVS ofv...glb abr Gora |) > ANIA

2‘93 6°99 6‘E9 Whe Ba OL ee Ole LOO L‘&Q hy Ay A © 01° © bi <1 ° NR Za 1d os | ERY Let PN)

9S vS 9f ob Sh SP GP 6h Zz Gz Gz Sh Sb | wut ur pq

o8uryy uv. | N ||PEEoby job10fy GLio’y IL10ty oL10ty 9g10bY gglo$y

(panurjuod) 66 AITAV J,

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

TABLE 30. Sepia australis shells. Relative dimensions as °% shell length.

2727 .

A30504

Mossel Bay (Cape Peninsula) .

North of Olifants River

Strandfontein to Muizenberg .

Millers Point

Namaqualand coast
Simonstown .

Arniston .

299

Lin mm W
33 36,4
47 34,0
48 3504
47 34,0
46 34,8
49 32,7
51 3353
43 32,6
48 3353
51 vr
54 35,2
41 36,6
52 B27
58 29,3
57 3353
55 29,1
45 3353
46 32,6
55 30,9
52 B27
43 34,9
59 32,2
40 32,5
53 R250
56 32,1

24
33,2
29, 1-36,6

300 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 31. Sepia tuberculata males. Relative dimensions as % MLd.

Adam
&
Ago121}| A30279 A30600 A31235| Rees || N | Mean Range
1966
MlLd in mm 25 42 27 33 51 47
Misy 2 92,0 85,8 96,3 90,9 90,2 92
NEW eet oS 64,0 69,0 63,0 66,7 58,8 66 6 | 64,6 58,869
HL 48,0 | 50,0 | 59,3 57,6 | 37,3 | 60
HW 60,0 | 52.4 | 70,4 63,6 | 47,1 | 53 6 | 57,8 | 47,1-70,4
FL 104,0 85,8 96,3 103,0- | P109,9 96
FW 16,0 16,7 18,5 15,2 19,6 Usa iS 17,2 15,2—109,6
ALI 52,0 a FO:A V12:7 We Ag) tl a7 5 | 59,8 |) 47a gee
AL II 48,0 = 70,4 66,7 | 43,1 | 60 5 | 57.6 | 43,1—70,4
AL Ill 56,0 | 45,3 | 66,7 66,7 | 49,0 -| 57 6 | 56,8 | 45,3-66,7
AL IV 52,004) 24 155.0) 915455) | eet Sa G | 5351 || aay ae
TL 76,0 — 140;7 127.3 49,0 | 106
Tcl 32,0 = 37:0 /30,3 | 21,6 |. 23,5, ]| 5 | 289 1) BiG oaga

See page 302 for Table 32

TABLE 33. Sepia tuberculata shells. Relative dimensions as % shell length.

Lin mm W Th Str z

AsoeG7 -.. <Bo SSR ae aaa Sales 69 A335 8,7 73,9
VNC tc 47(0 Re eee ee RDS STR Bs 39 46,2 5,1 7150
ABHOR G totes ep See ae 45 4454 6,7 6454
AGORIE ae Eo ae ee Satie ae 55 4555 555 63,6
45 46,7 6,7 62,2

ASOGEO: Mis i eo ae oe ee wk 32 46,9 6,3 62,5
AGOGO OD) ie Pea. See 27 48,1 5,6 55,6
22 50,0 6,8 63,6

Strandfontein to Muizenberg .. 62 _ 5,6 71,0
38 47,4 553 7397

39 48,7 757 56,4

Betty’s Bay > so). 7a ae be eae 37 48,6 8,1 54,1

Koommeie: (3.95 ds ee 40 52,5 755 62,5

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA = 301

Adam 1941

Adam & Rees 1966 .

Cape Agulhas

Milnerton

Pearly Beach .

Die Kelders

Beach 11 km NW of Cape Agulhas .

TABLE 33 (continued)

L in mm

5555
59,5
41,5
46
+68
61
39
51

.

39
48,9

43,5-55

Th

6,1

4,4-8,7

Str z

72,2

54, 1-88,5

ANNALS OF THE SOUTH AFRICAN MUSEUM

302

g‘gbv— Ga
b‘E9-S‘EP
9‘L9-S‘SP
3°99-S‘6E
9‘Lg—-&‘EV
3‘ES-Z‘Q

1‘19-S‘1P

ofog— 19

osuey

9°65

PG
CpG
b‘GG
G‘2G
en

Gs

0‘69

oI

oI
II
GI
ol
II

GI

oI

uvsyy | N

16 G‘9z of
Lai — Vor
6S Ch 8 GS*Eh
ZG [ip een
aS €S S‘6E
6S 1G GCP
61 o‘8 =
ZOl z6 —
GG 1S LV
Lv 6S G96
ol 19 oL
36 88 96
V9 6P 6S
sl P “l
gQ61 s99 XJ 1v61 wepy
Ste OM

1°96 | g‘9h
S°go01 | og&1
QaG Q1G
69S G‘EG
6°9S o‘LS
1°19 Q1S
V‘61 L‘0%
g‘Gor | L‘gor
1°19 o‘GG
gicS | gob
6°69 &‘L9
o‘L6 b‘96
ot 9S

VES
CVI
PQS
‘v9
‘v9
o‘oV
g‘0%
G‘o11
b°QS
00S
L‘99
0‘96
9v

6SSo0&V 11S06y

ole Gof L°&Y L‘o& &°EE ; 10
o‘o61 | 06g Von i Le Aoyoie If Me Aoys, ee See
o*6S vot v‘E9 L‘ov 00S =" SATAY
a‘LG 00S 9‘L9 E‘LG L‘ov oe TAN.
0°69 G‘zG @‘Q9 L‘9S 00S = TN
Q‘GG o6GG 9‘L9 0‘9t ESP : LIV
raat rob ok 6°91 SLI L‘gt " Ma
Gyrehen tovtaxoye | toil | Oral || tir po W
06S G17, 9‘&S v‘6P 0°09 ; MH
06S o‘Gh Goh Gh L‘o¥ ; WH
9°89 o‘EL SEE L‘99 00g MW
006 0°69 a‘L6 b‘69 009 > “ ATIN

oL ZB TV GL of * WUT UT pT

oglosy |6f10by |Ez10fy |Log6zy |19L6sVv

‘PIW % S¥ SUOISUSUIIP DATLIDY ‘sopeUMloy vyvINIVAgn] DIdasy *s& ATAV J,

ae

ete -

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA

A30118 A30120 | A30137| A30138} A30507 A30509 | A31250|| N | Mean
MLdinmm.| 100 115 142 1. |, t0Ose | 110) | |S. 01 105 rT aa ee |
mire. . | 95,0, 87,0 | +69,.6 | 87,0 | 86, 78,3 81,0 90,5 | 87,6
MW 68,0 60,9 62,5 | 60,0°| 59,1 | 59.1 76,2 64.8 | 58.4 9 63,2
HL 50,0 38,3 35,7 | 44,0 | 40,9 | 26,1 41,0 33,3 | 46,9
HW 50.0 46.1 42.9 .| AzO | “45.5 \- 34,8. 52,4 43.8 | 46.0 9 44,8
FL 113,0 117.4 89,3 | 105,0 | 100.0 | 95.7 104.8 | +100 104.4
FW 26,1 14:3.) 15, 1G 13Oed 28: ee in 8 18,5
AL Bs0) 62.6 Qo | 540 |} 52.7 1 478° 56,2 ak 64.6 8 54,1
ALI 57,0 56,5 42.0 | 55,0 | 57.9 1 47:8 61.9 x 64.6 8 55.3
AL lll 64,0 61,7 49:9.) 550.) 65,5-| 47,80 ) 64,8 a 58.4 8 57,5
ALIV 65,0 55,7 56,3 |. 480°! 509.| 43.5- 54,3 ST 1h) 52.2 9 53,7
TH 105:0P 64.Sx\) 160,7 | 750T 77:3 fie: tha a 167.3
Tcl 8.0 30.4 30,41 30,0 |. 32.7 = Be = 30,1 6 31,9

TABLE 35. Sepia papillata females. Relative dimensions as % MLd.
Hoyle
Agorig} Ago124| A3g0136} Ago140 A30507 IQIO N | Mean

MLd in mm 130 115 95 140 120 135 115
MLv. 92,3 | 91,3 S854 S258 91,7. 88,9
MW . 63,8 82,6 61,1 60,7 75,0 66,7 62,6 7 67,5
HL 4253 Goo) 46,3 | 4453) 5853. $44.4
HW AD Se i0T50,54 Agee 37,0 | 493 1444) 40, 1.7 | 45,0
FL 115.4 | 124,3 106,3 103,6 116,7 88,9
FW 1554 21,7 19,7 17,9 20,8 18,5 1553 7 T7750
ALI FOO -1-O3554 | 54.7 1, 42,0 1 5053 9 940.7 | 475851) 7 | 5st
ALII 50,001 -O453m ie 58,9 || 404 uh 50.9 9 AAA) Nebi4e,7: (|| Telents0
AL II oA pales 20h ea Oim | u 4 Os4iae OO gn oad 56,5 || 7| 56,6
ALIV 5O;0na, (68,75 i) 4955 Hau45,0 [5255 8° 45:9 | 4758) 1 7 | 5153
TL 10155, | 23450 sl94,7 | tO7,1 | Wr40,0 9.139, || 113,0
Vel AAO || 55.7) Weed. 0 93,0 | 20,2 § 25,9 6 | 39,4

TABLE 34. Sepia papillata males. Relative dimensions as % MLd.

B93

Range

58,4-76,2
34,8-52,4
13,0-28,6
42,0-64,6
42,0-64,6
42,9-65,5
43,5-65,0

30,0-38,0

60,7-82,6
3739-59,5

I 1,3-2 I 97
40,7-63,5
44,4-64,3
46,4-66,7
45,0-08,7

2539-557

304 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 36. Sepia papillata shells (form A). Relative dimensions as % shell length.

Linmm W Th Str z
Agonis | 4.) Bape bee 110 46,4 8,6 7257
AGOASS fol) fo hey) Sa 134 —_ 9,7 68,7
AgGEOg SO.) Buk meee, eee te 131 45,8 935 61,1
94 47:9 9,0 61,7
RAO5OQ Ss eee ree 104 42,3 8,7 65,4
PA CT ete fees. We: 100 48,0 10,0 67,0
117 43,6 9,0 68,4
Strandfontein to Muizenberg . . 124 — 10,5 69,4
100 42,0 10,0 69,0
103 -- 9,2 71,0
100 46,0 11,5 55,0
Beliy Ss Bay -. ea eee ee 103 48,5 8,3 70,9.
Strandiontem., .°-"230.) 2 4-4 95 49,5 12,1 5739
92 52,2 11,4 60,9
Whinerton beach: =a 122 46,7 12,3 59,0
123 4535 9,8 69,1
Asniston .cieelé 7. 7] 2iGi oe |. 92 4537 10,3 64,1
Hoyle 1910-43) -.4b A art 103 47,6 12,4 —
Nat 28 Feet oc CNP ee oe 15 18 17
Meany f. 5.2... Tak oh = & 46,5 10,1 65,4

Range, to O03. Toe eee ee 42,0—52,2 8,3-12,4 550-7257

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 305

TABLE 37. Sepia papillata shells (form B). Relative dimensions as % shell length.

Lin mm W Th Str z
Si ae haa lg ee; |S ime | ae ee 57 4754 9,6 61,4
Ark | 48,8 9,8 53:7
Strandfontein to Muizenberg .. 109 4539 Q,2 7552
88 43,2 9,7 68,2
Siiiautsbosbaal. |. . ... 2s 97 4594 10,3 76,3
Siramenpnteia |) i.e a PP. 76 48,7 9,9 64,5
Namaqualand coast . .. . 92 44,6 8,2 68,5
PEMA carols Ba ap. sk al fe 70 50,0 7:9 7239
Rrerorme ie } 5 i.e.) ae. 7 4551 8,5 74,6
PPM ew 65,5 48 — —
AGMMMAGEMREES TQGHweirs 2. 108 4355 10 +66
EP ee a WR eg II 10 10
PM a gh fe 46,4 9,3 68,1
Jodi Pay VS a a a 43,2—50,0 7,9-10,3 53,7-76,3

TABLE 38. Sepia simoniana males. Relative dimensions as 9 MLd.

A30127 A3I251 N | Mean Range

MLd in mm. 116 141 144
Mikyic: 20]. 92,2 91,5 87,5
NEN 1S 62,9 63,1 57,6 3 61,2 57,0-63,1
ell se ao ie 49,1 4353 444
BN fas en te 3751 49,6 42,4 3 43,0 37,1-49,6
FL ay, 94,8 113,5 109,0
FW — 15,6 16,0 2 15,8 15,6-16,0
ALI 62,9 59,6 55,6 3 59,4 55,6-62,9
AL II 63,8 66,0 63,2 3 64,3 63,2-66,0
AL Ill 7353 7939 68,8 3 71,0 68,8-73,3
ADV 60,3 5352 5459 3 56,1 53,2-60,3

Rt = 177;3 172,2
une Lt — 173,8 160,4

66,7 63,2—70,2

|
oO
—
~J
&
NO
nH
Ge
N
NO

306

ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 39. Sepia simoniana females. Relative dimensions as % MLd.

A30132| Ago133} Ago134| Ago135| Adam & Rees 1966 | 1925

Nilay 0 87,4
MW 5354
HL 46,0
HW 40,2
FL. 100,0
FW 10,9
ALI 4357
2.0 WAG Ug 43,1
AL Til 44,3
AL IV 41,4

jae 103,4
TL Ht 100,6
Tcl 48,9

172
92,4
5756
43,6
4757

116,3
15, I
52,3
535
55,8
54,1

142,4

127,9
55,2

See opposite for Table 40

139
87,8
57,6
40,3
4157

105,8
10,8
39,6
41,0
4593
43,2
89,9

734
51,8

RII
87,4
60,4
3758
45,0

100,0

9,0
38,7
41,4
46,8
41,4

101,8

TiT37
51,4

Massy

140 147

84

56

49

41 3955
+93

+8,5

39 45,6

43 40,8

46 5454

43 51,0

me 187,1

— 60,5

N | Mean

©
Oo
qm
O

locech och echo) ©
aN
ee
Ie)

~I
on
=
i)

TABLE 41. Sepia angulata shells. Relative dimensions as % shell length.

A31317 (Holotype) .

A31395

A31318

A31319

A31320

55,8

50,8-60,0

13,3
14,0
19.1
13,0
15,2

14,0
16,0

11,9
11,9

12,2

14,3

11,8
12

1304

11,8-16,0

Str z

51,7
47,4
4755
52,2
50,0
5935
60,0

60,3
61,2

58,5

5453

58,8
12

5454

47,4-61,2

Range

53 60,4
39 —4757

9,0-15,1
38,7-52,3
40 —53,5
43 —55,8
41,4-54,1

48,9-60,5

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 307

TABLE 40. Sepia simoniana shells. Relative dimensions as % shell length.

Lin mm W Th Str z
oe Os ir A a ee 113 42,5 10,6 67,3
og oo UE Bs a oe 169 ; 45;6 10,1 71,6
ees, LEPC. She LR 166 47,6 9,6 78,3
Pee = OW A TL MO. Pt. 23 52,2 8,7 56,5
Strandfontein to Muizenberg .. 185 43,8 9,2 7955
14! = 10,6 71,6
PEMenstewh:. | 1) d.6e¥! & Ba 127 4353 9,4 7357
152 46,1 8,2 7551
MEWESPAY Se elk ELLs 181 44,8 11,0 79,6
DreamGnentein kk ts 98 46,9 10,2 60,2
Umngazana River mouth. . . 73 46,6 — —
Pea Sk Pe eA 100 44,0 10,0 65,0
104 42,3 8,7 62,5
80 46,3 10,0 70,0
PREMISE GP PLD ED eh hae’ 38 47,4 9,2 63,2
meam € Rees r9o66;5 . . . . 137 44 11,5 +59
131 38 9 +64
121 43 10 +65
120 45 12,5 Sa 7)
114 49 12,5 z= 7O
112 44 10 +62
PUY 48 14,5 —
107 47 12 S25) 0
105 4I II +62
go 43 13 63
N 24 24 15
Rice OO 2 eee ies cf ae 4551 10,5 68,7

Range: Gata ..) « Reh. 38 —52,2 8,2-14,5 56,5-79,6

308

ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 42. Sepia hieronis males. Relative dimensions as 9% MLd.

A29728} A30146} A30563} A31243| A31405| A31406] A31407|| N | Mean

62 61 47
75,0 82,0 83,0
56,5 65,6 70,2
3751 50,8 | 44,7
43,5 | 50,8 | 46,8
62,3 || 90,29) 78,7
6,5 9,8 8,5
38,7 | 55.7 | 34,0
35,5 | 62,3 | 42,6
38,7 60,7 46,8
45.2 | 689 | 48,9
90,3 114,8 142,6
14,5 13,1 12,8

48
93,8
62,5 || 7
45,8
5452) 47
83,3
SS lez
EIN) i
47,9 || 7
58,3 || 7
Be 7
185,4
14,6 || 7

TABLE 43. Sepia hieronis females. Relative dimensions as % MLd.

No NNN WN

Adam &

A30145 Rees 1966
33 61
78,8 74
60,6 54
48,5 38
51,5 51
84,8 85
Q,1 8
18,2 38
24,2 38
30,3 38
4555 38
157,6 ri ek
TOOK 10

Relative dimensions as % shell length.

MLd in mm 59 54
MLy . 69,5 | 79,4
MW . 49,2 | 55,6
HL 37,3 | 37,0
HW 44,1 | 42,6
FL 78,0 79,6
FW 10,2 5,6
ALI 30,5 26,0
AL II 33,9 26,0
AL Il 37:3 | 31,5
AL IV 49,2 | 46,3
TE 127,10 glatOL.g
Tcl 11,9 14,8
MLd in mm
MLv
MW
HL.
HW
|
FW .
AL I
AL ll
AL Ill
AL IV
INE
Fel
West coast A29728
A30146
A30563
A31243
East coast A31405
A31406
A31407
N
Mean

40,5

38,3-45,7

12,4

9,6-1 5,2

Str z
70,0
731
+72,6
7331
65,2
66,7
64,8
7
69,4

64,8-73,1

Range

49; 2-—70,2
3735-5452

5,0-10,2
26,0-5557
26,0-62,3
31,5-60,7
35,7-68,9

I 1,9-14,8

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 309

TABLE 45. Sepia insignis female.
Relative dimensions as % MLd.

A31247

Midaininimi. 1h, 44
WEN, Ee ibe ay = 90,9
ON ee + 52,3
Lo ae ae a ee 27.8
: | on 4595
ee bo ho Ba Se 9535
WE Ba eh 6,8
yl Clg Pata meet 36,4
ALAR tS 36,4
BbolIE 4. 29,5
Pe Ny RN ee, 4535

intl co Pets 68,2
— Le 8 75,0
VL Soaliee CER Pee 13,6

TABLE 46. Sepia insignis shells. Relative dimensions as % shell length.

Linmm W Th Str z
IGOARG 2 3, os 29 3455 10,3 552
OTIC MS rs ii 29 3759 10,3 5592
NOTOA Wee ye 34 3553 8,8 61,8
1735 3751 a 60,0
PROT OAR Tie i n>’, 44 31,8 Ty. 63,6
Sull Bays 5 +32 aie 10,9 65,6
Adam & Rees 1966 . 26 33 16,5 65
Me bcoe | Lie Pearse Lo
INiCarm Bet pce: 34,4 11,4 60,9

Rangers <4). 31,3-37,9 8,8-16,5 55,2-65,6

ANNALS OF THE SOUTH AFRICAN MUSEUM

310

L‘g1—L‘g [org 9 9G —— mor 5 i et == 69 pee sen ae een aS Pisce eo ° . . Ma
Gorn Vir 6 Sor “evor |) ‘606 O‘OOL Of001 | Of00I Of001 Of001 Of001 Of0O1 | °° Rete aie
g‘LL—o'0G | L‘og | gz |} 6°24S RHEE ore G, 1°LS GS 5590 7 8°99 o‘09 )6=s« S&‘gG—Ss«éO0Q HG C070) ves 2 NEE
og 81h 99S Tiga AST HAMOYS)* (Sef L990 «= S‘wQg Ss & EG TAG SEES a a |) |
DCO 90) 0-00 7) Lolly Qo. 8 ia § Sof ara afl, CoG, “oy setae TOKE, 7 OMe} Copteyl, — (ohoVe) - MINN
L446 66 a99 S06 6.06) = « & EQCCé#G “1 [Eo \o vam ASIG/L/ oy same (0)(0Y9) sr w/e toe co Z10}0) : SS ATTIAT
61 V1 LA 1G II oI gI Gi gI GI V1 GI UWIUT UI PTIN

suey ueayy | N LLioty gLioty (uo) E9L6ey
eS PGr ees wee Ss ee Seach aes ~~ aes ae Coy L‘9 ee L‘gl . . . . Ma
gor LLor of001 Gig L‘go1 ofoo1 &'g01 9Gor — Zor i111 L‘gor of001 o0f06 of001 Sac : aie
595 = Ging. "oftg-2 Gi1g™ “0.09% “ofcS= =E‘gS 836 o ‘0G — SO Ole 0-09" 2 0°09) | OL0Lm 8409 : oF SNA
LOO) pe Qres me oILS ae iO Coe Soot erly | Gcon — ofGG — Ines 2 OQ) oud FOG, O02 ou : ees (|
boo) fea cO) ae ONG tO OL sO ON ELL eG) COREL OG) Via O99 = GL Ligg 4 0'Oc | Se : ’ MI
MO foyikey (OK(loy Coiide) = /Eovoe Geyoym ISG Zroy aa hi49) pSilfey = LA oy {otoray CopKoje ye | YAfovo\n Fa Kooyols = fot eho) : * ATIN
oI I II 1 G1 61 gl gI 8 V1 6 GI oy Ol oI ees og, TRUE TLE PEPTAT

EgLb6sy Lil6syv

‘PIW % Se suotsusulIp saT}EOY ‘soyeu vI91dq vidas ‘LV atav,

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 311

TABLE 48. Sepia typica females. Relative dimensions as °%% MLd.

A29608 A29717 A29783
MLdinmm | 22 21 18 9 10 12 8 10 II 23 25 19
Mlv . .| 93,2 100,0 94,4 | 88,9 90,0 83,3 75,0 80,0 72,7 | 91,3 92,0 89,5
ie. | 72,7 70,2 77,0-| 869 60,0 75,0 75,0 70,0 72,7 | 82,6 64,0 89,5
Pies 62,6, Gi5o-! ‘61,1 66,7" “7o,0 ° Gb,7 — 60,0 63,6 56,5 48,0 57,9
PW 2) 5455.5 721.—.55,0..|. 66,7. 70,0) 66,7... —._,60,0.. .63,6 |. 52,2. 48,0 57,9
few | 113,0 10,5. 105,6° | 100,0 | 100,06 FrO0,0 — 110,0 90,9 | 113,0 104,0 110,5
PV ea Q,1 9.5° 9D,1 _ — 12,5 — — — 10,9 12,0 10,5

A29783 (cont.)

MlLdinmm . 15 18 20 14 It 14 10 10 II 10 13 14

WEE ss fos; °G339 + .90,07-2G2,9' 61,8 85y7eRh 90,0 90,0 90,9 .80,07. 284,65 35,7
DEW s. daio 677,8.' $0,0°.085,7 72,7 78;60¢ 80,0 ‘90,0 .81,8 90,0 //69,2 85,7
lc | et Goro 55.0" 50,00,050,0 63,6 . 579192 Go,o. 70,0 .63,6 -80,0. J6%,5. 5751
HW . . .| 60,0 55,6 55,0 .64,3 63,6 50,0 70,0 70,0 63,6 70,0 53,8 57,1
FL ie. | LOMO ¥LCO,ONIIO5 Oni is4e 100,0 LO 7AleTT00;0 1 10;0° 10Q;3 \4.9030 100,0 107,1
PW —— fd = Ek Vl Rome re Maar Weer ay I ar es Te OY

A29783 (cont.)

MLdinmm . 14 15 15 15 16 15 16 18 17 14 18 19
Mlv . . .| 85,7 86,7 93,3 86,7 87,5 93,3 93,8 88,9 88,2 92,9 83,3 94,7
MW. . .. | 78,6 73,3 80,0 86,7 81,3 80,0. 81,3)) 77:8.) 82:4)-85,:7)1,77,8 8452
HL. - .. | 57,1 53,3 -60,0 66,7 62,5 66,7 62,5 66,7 64,7 64,3 66,7 68,4
EUV Miah 57504 60,0 ...60,0...Go,0)) 56:3 ..G0:0) 56,3... 5050 56,8) 967,11 9°55;56 52,6
Pe 1G7.1) 63.9 106,7 [106.7 112.5 10G-G 106.9, 160j0,511,8 114,9).705,6 100,0
LON et AS a Fak — _- -- _— — 1838 83 11,8 grt 53,9 . =
A29783 (cont.) A3o0176 A30177 N | Mean Range
MLd in mm 18 17 18 15 19 20 15 15
Miv . .| 889 94,1 | 94,4 86,7 | 100,0 78,9 93,3 9353
MW . .| 77,8 82,4 | 72,2 73:3 | 78,9 75,0 80,0 86,7 || 44 | 79,3 | 64,0-90,0
HE, YOO ee GE,r 164;7 55,0 66,7 63.9% 55,0" 66:79 6657
HW “Arid 55,6 58,8 55,6 60,0 52,6 60,0 60,0 66,7 || 43 | 59,0 | 48,0-70,0
Pee. = htO5.6 105,9 1, 100;0° | 66,7 | Ti0,5 110,07 1139/9 no

PW 2". see — 539 — 10,0 — — — — 17 | 40,9 5;9-18,8

312 ANNALS OF THE SOUTH AFRICAN MUSEUM

TABLE 49. Sepia typica shells. Relative dimensions as % shell length.

Linmm W Th Str z
A2Q608, 4 he am 19,5 53,8 — 51,3
Adam & Rees 1966 . 21 5751 — —
Mean, ao.) | Samy 5535

TABLE 50. Sepia robsoni, S. faureti and S. dubia. Relative dimensions as % MLd
(or shell length).

S. robsoni S. fauret S. dubia
(Massy 1927 A30144 (Adam & Rees
& Taylor*) 1966
& Taylor*)
3 2 2
Mlidinmm. . . 19 21 17
Miva: Goes, & 87,1 85,7 91,7
NEW 0.80 3 LAP) +76,5 66,7 68,5
TUL, 0.98.9 ACEO: |. 33:5 42,9 53,6
SAW OE) SES 5751 42,9 52,4
FL ne: 7259 90,5 gI,I
FW 8,8 oi 955
ALI 64,7 38,1 —
AL II 70,6 38,1 —
AL II 70,6 38,1 —
ALIV 76,5 42,9 Re i
i; 60,6 81,0 68,5
Tcl 17,6 +9,5 16,1
Shell Linmm ... 14,5
W sbos 62,1

* Taylor, personal communication.

TABLE 51. Sepiella cyanea males. Relative dimensions as % MLd.

A6526 Adam & Rees 1966 N | Mean} Range
MLd inmm | 62 7475 56 55 53:5 50 48
NEL i oe 75,0 74 81 88 87 6 86
MW .. 7%: 40,3 45 41 57 56 60 58 60 8 52,2 40,3—60
BE ee 19,4 23 24 25 25 26 26 27
HW . ./°- 419 | 41 39 52 49 49 52 50 8 | 46,7 | 39 -52
FL : 80,6 93 85 100 98 98 98 92
FW 8,1 i5 9,5, 19,5. 18,20 20 12,5 8 15,3 8, 1-20
ALI 355 | 41 41 36 35 36 38 40 8 | 37,8 | 35 -41
AL II 33.9 | 41 41 36 36 36 38 40 8 | 37:7 | 339-41
AL Il 419 | 47 45 46 44 43 48 50 8 | 45,6 | 41,9-50
AL IV 46,8 | 61 55 54 49 50 54 56 8 | 53,2 | 46,8-61
+E: — | 81 106 = —_—- — —_—- —

ne
eh
_
ie}
_
x
oOo
i)
1S)
LS)
iS)
iS)
jo)
iS)
nN
i)
_
~—I
iS)
2
©

17,5-24

A REVIEW OF THE SEPIIDAE (CEPHALOPODA) OF SOUTHERN AFRICA 313

TABLE 52. Sepiella cyanea females. Relative dimensions as % MLd.

Adam & Rees 1966 N | Mean Range
MLd in mm 75 71 69 a7 WaT aa 29
MLv 80 84 86 Sy «87 64 89

40 42 49 ae oe on OD i 48,4 | 40 -55

13,5 14 19 m (1 16 12 7 16,1 12 -I9
ALI 36 38 29 By Bao igs. 128 7 30,3 | 27 -38
AL II 33 39 29 27] 427 ay. 26 7 30,0 | 27 -39
AL III 36 39 30 a7 2h), ae 6° 3t 2 31,0 | 27 -39
AL IV 47 48 38 95 38 38 35 7 39:9 | 35 —48

14 —_ — —- —- —_- —

CUE gee eV dees 2 aaa ae a 4 23,4 | 22,5-24,5

TABLE 53. Sepiella cyanea shells. Relative dimensions as % shell length.

Sex Linmm W Th Str z
A6526 3 61 31,1 11,5 50,8
Adam & Rees 1966 3 69 29 10 58
3 54 33 14 56
2 74 39,5 II 5!
g 70 31,5 12 53
2 69 3555 13 58
A6526 ? 77 33,8 Biey 54,5
? 76 32,9 11,8 52,6
? 62 32,2 11,3 54,8
? +75 30,7 12,0 61,3
Adam & Rees 1966 80 32,5 12 60
? 76 31,5 12,5 58
INEST re 12 12 12
Meantotal > oe Ts 32,0 I1,9 55,7
Rangetotal - - - 29 35,5 10 -14 50,8-61,3
Ng . 3 5 3
Meang 31,0 11,8 54:9
Rangeg 29 -33 IO -I4 50,8-58
No pe ie Hei 2s 3 3 3
HG AI LS aa. Ok 32,5 12 54

Rangeg PT 30,5-3555 Il -I3 51 58

a
an i) yee
‘ Bi tat 4 ae
1 je = in eli
¥ f ’
i rs 3
x m
- — vs »
Ae al ‘ 14
— ‘
; a ;
~ ; ee
bn = son < ae
< i
P e r =
ss wig’ f 2 e
. = 4
oo eee a rng ete 2A vines «am
x = 1",
= x sh ¥
2 = ve 3
4 Ae nts b . - . wer nd ic eerie ps chitin eal mira tiasian ich abs (age ahi ete pera etielnmatne!
; a . ‘ . . a) oa): nas
: i= a aah). To ate f Ye ne aid.
ie 3 ie
= e iy ‘ ‘
: s 5 = ‘
- x
ey, = ¢ E
r zi 08 “ 4
=f
otek:
z = =
an HS
; §
3 -
=|
s oe Bin x
= =
. ; ]
wu = A - S
‘
> :
= s
4 “ ase 2 . es
= ! 4
4 5
S =
Le
- 4
" %, % te
—
z
5 :
se {
= e me)
4
’
7 ory.
F, ;
_ af - P
w
J
ee .
- i . P
> ca

Plate 35

Ann. S. Afr. Mus., Vol. 59

f shell
c. Dorsal and d. ventral

1E€WS O

ventral v

Dorsal and b.

a.
Sepia officinalis vermiculata Quoy & Gaimard, ? from Durban

A2i4i:

Pfeffer,

ica

ibar

Sepia zanz

f

VIEWS O

.
e

shell.
Scale = 20 mm.

Ann. 8S. Afr. Mus., Vol. 59 ) Plate 36

Sepia officinalis vermiculata Quoy & Gaimard, 2, A3o0183: a. Dorsal and b. ventral views of shell
(posterior spine broken).
Sepia officinalis hierredda Rang, 3, A31291, from Baia de Cabo Negro, Angola: c. Dorsal and
d. ventral views of shell of comparable size to the above.
Scale = 100 mm.

Ann. S. Afr. Mus., Vol. 59

Sepia acuminata Smith, 3, A31398: a. Dorsal and b. ventral views of shell.
Sepia confusa Smith, A2140: c. Dorsal and d. ventral views of shell.
. Scale = 10 mm.

Ann. S. Afr. Mus., Vol. 59 Plate 38

Sepia incerta Smith, N.M.g70: a. Ventral view of shells described by Massy (1925: 219) as
S. incerta.
Sepia incerta Smith, N.M.g58: b. Ventral view of shells described by Massy (1925: 215) as
S. burnupt.
Sepia burnupi Hoyle, A2147: c. Dorsal and d. ventral views of shell (anterior tip broken).
Scale = 10 mm.

Plate 39

Ann. S. Afr. Mus., Vol. 59

SS

CDDC—R?Ana CUGCGG ON . \ \

AQ nny

OOKD> WS

SS

Sy

—

DG?)RK_E ~
SSS S

f shell

1€WS O

ventral vi

Dorsal and b

Sepia tuberculata Lamarck, from Cape Agulhas: c. Dorsal and d. ventral views of shell with

a.

from Punta Zavora

b

ith

ta Sm

incer

Sepia

lly long striated zone.

Scale = 10 mm.

.

exceptiona

Ann. S. Afr. Mus., Vol. 59

Plate 40

Sepia australis Quoy & Gaimard, from Still Bay: a. Dorsal and b. ventral views of shell.
Sepia tuberculata Lamarck, 2, A30511: c. Dorsal and d. ventral views of shell with normal striated
zone.
Scale = 10 mm.

N

\ <

a. Dorsal and b. ventral

f shell.

from Milnerton beach
Sepia papillata Quoy & Gaimard, shell form B, 3, A30120

VIEWS O

Sepia papillata Quoy & Gaimard, shell form A,

f shell.

VIEWS O

c. Dorsal and d. ventral

Scale = 10 mm.

Ann. S. Afr. Mus., Vol. 59 Plate 42

Sepia simoniana Thiele, from Still Bay: a. Dorsal and b. ventral views of shell.
Sepiella cyanea Robson, g, A6526: c. Dorsal and d. ventral views of shell.
Scale = 10 mm.

Ann. S. Afr. Mus., Vol. 59 Plate 43

Sepia hieronis (Robson), 3, A29728 (west coast): a. Dorsal and b. ventral views of shell.
Sepia hieronis (Robson), §, A31407 (east coast): c. Dorsal and d. ventral views of shell.
Scale = 10 mm.

Ann. S. Afr. Mus., Vol. 59 Plate 44

Sepia insignis Smith, A31241: a. Dorsal and b. ventral views of shell. c. Ventral view of part of
large shell from Bloubergstrand.
Sepia angulata fi. sp.: d. Median view of half shell, cut longitudinally to show angle between
striated zone and smooth zone.
Scale = 10 mm.

iews of shell.
iews of shell

>'S
5 §
S 8
>.>
Qo
solae
Ss
a &
3
m
Oo O°
AA
ad

ta angulata. n. sp., paratype, A31320:

Sepia angulata n. sp., holotype, A31317

Sep

Scale = 10 mm.

INSTRUCTIONS: TO VAUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE, 1960.

Style manual for biological journals. Washington: American Institute of Biological Sciences.

.

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm x 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

Bu.LLoucu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Paris
88: 100-140.

FiscHErR, P.-H., Duvat, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19600. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, b; Liste: 11. Turton, 1932: 80.

ho (es )oanththes leptin. tes F
: shuirte. al ni. (rkelnd. Where phan | Sees
. show Sepetibesd wiper Biv _ervigived age ca

bhi sity ; YTRMAO fey | ohiden. a “i. Gb dW | Diy rs
v3 . ~accaleny (8) eae yritsrzel en et (cy. ated om

HER uri ia Wi f teal |
‘ i, - a wes "t
i ’
ay Re _tlibw | mae ‘aabie vito A stunitgss goibulsai. 9 ey rei Sax
oEeN, © o a . we tlie
4 _ yy ‘, ; ‘ x 75 + =a) : * ‘e r
; : ig ri if ms a 4 i , ae th b 7 A aM x
ovine noiasistld sas’ bo hie av bres STA £1 abun. sats SO) | Caacry 7 oes Sis 1
ais Roses fr aston vd

(HSS inte | US8GL" 1F ebiria Pre Yo bry: al) 3
ran . Sefte ade od THIAS UNS pnd syeaT TO). Oy Meat ihe a. sh ebaitisre: ce

; f > i ene
cid cou oildug ta saela ax hie ae: ee B3 iviritia ol
= ae SE AIGS 5 st y, ¢ at de > ‘lexi fi iy trun ( it) sbi gat 12 - 1): shila
ewmdireredg atsorme 4 Pie Ueltioe tare in pha nl Dep sity, Spaltaeireng tay
Sa ig vr at 5 » bp
odehrtirreraie VOHAS Was Cboged yO ian bel Yh iis ay cdisue” rise va
ro : aye =e ¥ ee > ‘| , H Aoi eutaagy f, ites aolies sake ;
- a ew Se Feat ee) ek h eo
om EMTS RE i snob Bet, hai x ek) oh; vayiramn lasityas’ Ve Sat
ae m erik Oe ea OU). bastie i £3 Sain al rshs 7. Pots zk 7" a wile cabana
SOHO IIL >: ‘ MAG tuna eporiidys & 4 tke eps POOL As IAS ae a ae 4
3 AEE
Eel es Lee 20 1et nl
mmo wnd ES ed Catiofousn lescullett) wangd to shor ligongents 2 “sola
— oa = = . ~ » \ - +; - ‘
es ! Se hee 2m Ler) walk toe woh ay ae ol ¥
HOY wi ri iraserhioved ievisl boa essaney gos cain ibe? eg
= 2 ee =. Li at c ‘ft a | j } oe it tO) BIRR gs 4 C. wih
c aie tne << y 7 iz vo 40 ‘
z 4 OAISTIGH@ SL aviivid noneor dbooneeD jsanlgeadeneiott ab beseiied
> ~ = r Pal? B . Reet NS od ;
ins. Le wnat an Salevia WA hie eine eae aha Ata
<0 OR roe tHe hee 200) od el taaes laa bunt, Aorted + anol: or
oy; ky ; - see yes
3 ; ’ ; . ms ' if MUTASE
Weal erkrnne looiyelos” Ye shag Incakiciorsstel feagal oat 2 op psicelbayi cal cake m
re ' A 2 ‘ -?
re CPE Bia 89 old cx viata vintalssain. Lericlno’® sor mutT Ins
Se! nti sonia Hit ole likey | eed yiape mops af ii ai bive od G1 s30ne pales Ue
; = _ seh wongeoanre sely sat ore bet TIO Hi move 2087 boa oloittg odi'te
of a eT Bo
; : ed
AR oper nornch AS: topeils ‘ Per en Tap ley i Eo
7 ‘ 3 gk, rod e f= i. a. ? ees ' ay
— > v ws => aes
: y x -
rays “2 weet
2 ai
4 | y 4 one
- = <4 ‘ ‘ ay
i} +
> -¢ = é
; rs , .
= a — —- a? '
a 4
e ‘7 « i La
Rie ‘ay

a LS

ANNALS OF THE SOUTH AFRICAN MUSEUM
ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

Volume 59 Band
September 1972 September
Part «uu Deel

DISCARD THE NAMES
THERIODONTIA AND ANOMODONTIA:
A NEW CLASSIFICATION OF THE THERAPSIDA

By
L. D. BOONSTRA

Cape Town Kaapstad

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material
becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid
van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad

OUT OF PRINT/UIT DRUK

I, Cir, 35 5s 7-8), Z(—2, 5» $= |0515))p 51-2, 5; 7-9)s
6(1, t.—p.1.), 7(1-3), 8, g(1-2), 10(1),
II ( I-2, 5; 7» tpl.) 24(2), 27, 3! (3) 318)

Price of this part/Prys van hierdie deel
R2,10

Trustees of the South African Museum © _ ‘Trustees van die Suid-Afrikaanse Museum

1972

ISBN 0 949940 14 3

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

DISCARD THE NAMES THERIODONTIA AND ANOMODONTIA:
A NEW CLASSIFICATION OF THE THERAPSIDA
By

L. D. Boonstra
South African Museum, Cape Town

(With 2 figures)

[MS. accepted 25 Fanuary 1972]

CONTENTS

PAGE
Introduction. : . ; : : Phe 5
Diagnoses , : ; 318
Diagnoses of the higher taxa . : ‘ A ha22
Summary . 5 ; d é : 2 9h334
Ineferences ~~ - 3 , ; : «)) 894

INTRODUCTION

Since the discovery of certain Permian reptiles in Russia and South Africa
in the early thirties of the 19th century showing some characters of a mammalian
nature and their first descriptions by Kutorga in 1838 and Owen in 1844
numerous attempts have been made by various authors to fit them into the
taxonomic system in such a way as to indicate in what manner these reptiles
could be considered related and ancestral to the mammals.

I have found the historical study of the various classifications proposed by
authors most interesting and illuminating and was tempted to publish a de-
tailed historical review, but on second thoughts have decided that confining
myself to the essentials of the ever increasing precision of the phylogenetic
views would be more important and valuable and satisfying.

At the present time we know over 300 recorded genera of ‘reptiles’ which
possess to varying degrees characters indicating a development in mosaic
pattern in a general mammalian direction.

For this assemblage of ‘reptiles’ we have the name Therapsida coined by
Broom in 1905.

In the present state of our knowledge this group as a whole appears to have
evolved from captorhinomorph and sphenacodontid precursors.

The oldest known therapsids, from low down in the Permian, consist of an
assemblage in which discrete lines of development are clearly evident.

For these lines of development we have the following denominations
available:

1. Anomodontia (Owen 1859)
2. Dicynodontia (Owen 1860)

315
Ann. S. Afr. Mus. 59 (11), 1972: 315-338, 2 figs.

316 ANNALS OF THE SOUTH AFRICAN MUSEUM

- Cynodontia (Owen 1860)
Theriodontia (Owen 1876)
Dinocephalia (Seeley 1895)
Gorgonopsia (Seeley 1895)
Therocephalia (Broom 1903)
Scaloposauria (Boonstra 1953)
g. Phthinosuchia (Romer 1961)

10. Eotitanosuchia (Boonstra 1963)

Cope ee

There exists strong evidence that the Dinocephalia and Gorgonopsia
evolved through the Eotitanosuchia from common sphenacodontid ancestors
of the morphological habit such as that of the genus Haptodus. Taxonomically
there may thus be some reason to coin a higher denomination to include the
Eotitanosuchia, Gorgonopsia and Dinocephalia in order to indicate their
consanguinity. Both these sublines of development, each showing the develop-
ment of certain mammalian characters, became extinct—the Dinocephalia at
the end of the Middle Permian and the Gorgonopsia at the end of the Upper
Permian.

The dinocephalian line shows certain stages of development. The most
primitive stage is represented by a group of animals for which the name
Eotitanosuchidae has been used by Tchudinov (1960).

Ascending directly from the Eotitanosuchidae is the group Brthenaaee
(Efremov 1954). From the Brithopodidae two higher groups arose dichoto-
mously, viz. the Anteosauridae (Boonstra 1954) and the Titanosuchidae
(Broom 1903). From the Titanosuchidae three higher stages diverged, viz. the
Tapinocephalidae (Owen 1876), the Styracocephalidae (Haughton 1929) and
the Estemmenosuchidae (Tchudinov 1960).

The Gorgonopsia line shows the following stages of development repre-
sented by groups with the following denominations:

Phthinosuchidae (Efremov 1954)
Hipposauridae (Watson & Romer 1956)
Gorgonopsidae (Lydekker 1890)
Burnetiidae (Broom 1923)

The dicynodontian line arising from unknown but probably sphenacodont
ancestors has as its oldest known representative the genus Otsheria from which I
have proposed the group name Otsheriidae (Boonstra 1963). Diverging from
this base there are the short-lived groups Venyukoviidae (Efremov 1940) and
Galeopsidae (Broom 1912) and the longer lived but also sterile line of the
Dicynodontia (Owen 1860) which split up into the Endothiodontidae
(Lydekker 1890) and Dicynodontidae (Owen 1876) from which arose the
Kistecephalidae (Seeley 1895), Lystrosauridae (Broom 1903) and the Kan-
nemeyeriidae (Von Huene 1948).

The third line, first encountered in the Middle Permian is that’ of the
Therocephalia (Broom 1903). Arising from as yet unknown. but. probably

A NEW CLASSIFICATION OF THE THERAPSIDA 317

sphenacodont ancestors they formed an important group of carnivores during
Tapinocephalus zone times, when they had already developed diverging branches
with the following denominations:

Pristerognathidae (Broom 1906)

Lycosuchidae (Broom 1910)

Alopecodontidae (Broom 1932)

This line, strongly developed in the Middle Permian, became extinct at the
end of the Upper Permian with the last off-shoots represented by the Whait-
siidae (Haughton 1918) and the Euchambersiidae (Broom 1931).

The last line, also beginning in the Middle Permian, with a few inade-
quately known forms, is that of the Scaloposauria (Boonstra 1953). This line
may have arisen from the therocephalian line during the Lower Permian and
is certainly closely related. Both the therocephalian and scaloposaurian lines
apparently arose from some earlier sphenacodonts, but we have no certain
indication of this as we have in the first line where a Haptodus-like form is
indicated.

Commencing in the Cistecephalus zone and continuing to the top of the
Cynognathus zone we find the Bauriamorpha (Watson 1917) which are generally
considered to have arisen from the earlier ictidosuchian Scaloposauria.

Also commencing in the Cistecephalus zone and continuing into the Red
Beds we have a final branch—the Cynodontia (Owen 1860) culminating in the
near-mammals—the Tritylodontia (Simpson 1925). If the Cynodontia are not
a parallel branch to the Scaloposauria, both to be derived from primitive
therocephalians, then one must postulate a direct and separate derivation from
some earlier sphenacodont.

From the foregoing it is clear that I think that we have three main branches
of Therapsida.

An older view was that there were two main branches which have been
labelled:

Anomodontia (Owen 1859) and

Theriodontia (Owen 1860)

Anomodontia

Although Owen initially in 1859-60 clearly intended the terms Anomo-
dontia and Dicynodontia to have as type, Dicynodon, he later included some
theriodonts.

Since then the term Anomodontia has had a chequered career, being used
by authors to include a variety of other forms manifestly un-Dicynodon-like.
The term Anomodontia being thus misused it would be pragmatic to drop it
altogether and rather retain the name Dicynodontia solely for those forms
showing a Dicynodon-like structure as originally intended by Owen.

Theriodontia —
In 1860 Owen coined the term Cynodontia with Galesaurus as the type and
included it as a ‘family’ of his Anomodontia. In 1876 Owen introduced the

318 ANNALS OF THE SOUTH AFRICAN MUSEUM

term Theriodontia for the same genera included in his former Cynodontia,
apparently to supersede the latter name. As Owen included genera now con-
sidered as Gorgonopsia and others as Cynodontia in his Theriodontia, and
Watson recently also the Titanosuchia, which three groups are now known to
lie on different lines of development, the term Theriodontia unites incom-
patible groups and should be dropped.

Deciding to discard the names Anomodontia and Theriodontia because
they each bracket together lines of development which, as I have indicated,
are not nearly related, it appears necessary to coin three new names for the
three main branches of the Therapsida, and to include as subdivisions of each
of these three new denominations those groups which are in fact closely related.

For the first of these branches I propose the name Alphatherapsida to
include the subdivisions Eotitanosuchia, Dinocephalia and Gorgonopsia.

For the second branch—Betatherapsida—to include only those forms
related to Dicynodon.

For the third branch the name—Gammatherapsida—to include those
fertile groups directly related and finally leading to the first mammals, viz.
Therocephalia, Scaloposauria and Cynodontia.

DIAGNOSES
ALPHATHERAPSIDA

Dentition primitively carnivorous with pointed incisors, canines and
postcanines adapted for snatching and tearing out flesh without cutting or
chewing. Short-lived side branches with dentition transformed to herbivorous
talon and heel teeth adapted for piercing and crushing without cutting and
chewing.

Choanae anteriorly situated with air passage without bony ventral floor
to partition it off during feeding.

The temporalis primitively originating from under surface of skull roof and
inserted in the adductor fossa and on the upper edge of the dentary.

In the gorgonopsian branch the insertion was improved by the precocious
development of a coronoid process on the dentary. Here we thus have a
primitive origin coupled with an advanced insertion.

In the dinocephalian branch the origin of the temporalis shifts away from
the under surface of the skull roof to the lateral surface of the intertemporal
bones, but no coronoid process is developed. Here we thus have an advanced
origin coupled with a primitive insertion. The later pachyostosis bedevils this
aspect. The postdentary bones of the mandible are persistently well developed
due to their retaining the primitive insertion of the adductors which exerted a
greater horizontal than vertical pull, with the joint a simple hinge. But in the
herbivorous forms a fore-and-aft motion allows for a crushing bite.

Concomitant with the primitive carnivorous jaw-mechanism the loco-
motory apparatus is of a crawling habit with sprawling limbs and little
upraising of the body and only slightly reduced digital segments.

A NEW CLASSIFICATION OF THE THERAPSIDA 319

In the braincase the sphenoidal complex is very well ossified, but the
prootic is feebly ossified, thus leaving a wide gap in the lateral wall with a loose
standing unwidened epipterygoid. In the pachyostotic Dinocephalia the gap in
the lateral wall is greatly reduced, but the narrow epipterygoid remains
uninvolved. r

In the Alphatherapsida the gorgonopsian branch has its characteristic
structures developed early and these are retained, with only insignificant
variations, throughout its span of life, notwithstanding that they survived to the
end of the Upper Permian.

The dinocephalian branch commencing as a primitive carnivorous group
early in its history, develops herbivorous twigs but the whole branch is short
lived and unprogressive and is soon cut short by the pathological pachyostosis.

BETATHERAPSIDA

Dentition herbivorous, primitively with a series of marginal teeth, later
with marginal teeth in part or wholly replaced by horny sheaths. Choanae
shifted moderately posteriorly, with part of air passage with bony partition,
separating it from buccal cavity, formed by plates of the premaxilla.

The jaw adductors highly specialized, particularly in their origins, with
concomitant great lengthening of the temporal fossa and the development of a
unique triradiate squamosal, and everted zygoma accompanied by a lengthened
sliding articular facet allowing fore-and-aft sectorial movement of the jaw when
feeding, insertions tending to shift on to the outer face of the dentary.

The postdentary bones unreduced.

Feeding on upland plants (except lystrosaurs and Kistecephalidae) the
locomotor apparatus is adapted for a more upright walking gait, with an
acromion process, greatly enlarged anterior iliac process, obturator foramen and
reduced digital segments. In the braincase the sphenoidal complex is very well
developed and situated far anteriorly; the prootic short, thus leaving a very long
gap in the lateral wall, with a loose standing slender epipterygoid.

Notwithstanding its long span of life this group remained stationary on its
early achieved developmental niveau. What variations arose were quite
insignificant, initiating nothing phylogenetically fertile.

GAMMATHERAPSIDA

Dentition primitively carnivorous, but variations commence quite early,
viz. reduction of postcanines (in lycosuchids and whaitsiids); development of
additional precanines (in alopecodontids and Scaloposauria); tricuspid
sectorial postcanines arose (in some Scaloposauria and some Cynodontia) ;
grinding surfaces developed on postcanines (in some Scaloposauria and Cyno-
dontia); differentiation into ‘premolars’ and ‘molars’ (in Cynodontia). ‘Thus
the primitive snatching and tearing dentition became adapted to cutting and
grinding with a process of chewing.

Primitively the choanae were anteriorly situated, but concomitant with

320 ANNALS OF THE SOUTH AFRICAN MUSEUM

the developing of a chewing habit the choanae shifted backwards with the
development of a secondary bony palate partitioning off the air passage from
the buccal cavity during the process of mastication.

The temporalis originated from the lateral face of the intertemporal bones
and inserted on the coronoid process.

Primitively, where the posterior mandibular bones are still well developed,
the superficial masseter inserted on the reflected lamina of the angular and the
internal pterygoid wrapped round the ventral edge of the angular. But in the
cynodonts part of the superficial masseter inserted on the postero-ventral corner
of the dentary and the internal pterygoid also partly moved on to the inner face
of the corner of the dentary, which resulted in a reduction of the function and
thus the size of the angular.

With the pull of the jaw muscles having strong horizontal components in
the early forms the posterior mandibular bones remained strong to withstand
the strain on the jaw joint. But with the jaw-closing muscles developing less
horizontal and greater vertical pull the strain on the jaw joint decreased with a
resulting decrease in size of the posterior mandibular bones.

With the increased strain on the dentary, due to its capturing some of the
muscle insertions from the posterior bones, it developed a very large coronoid
process and a prominent angle and, extending further and further posteriorly,
in the final stages made contact with the squamosal to form a double jaw joint.

In the locomotor apparatus a more upright walking gait is developed with a
reduction of the phalanges to 2,3,3,3,3 in all but the early cynodonts where the
4th and 5th digits have 4 segments.

Primitively with pubic foramen, advanced with obturator foramen.
Primitively without, advanced forms with, infra-spinatus fossa.

In the braincase the sphenoidal complex is primitively not well ossi-
fied, but moderately so in some advanced forms. Primitively the prootic
is feebly developed, but in advanced forms extends anteriorly to meet the
epipterygoid.

Primitively the epipterygoid is usually slender but broadened in the
primitive lycosuchids. In some advanced forms it is greatly broadened and
meeting the prootic enters into the sidewall of the braincase (cynodonts).

Postorbital bar primitively well developed as also in some advanced forms
but in some others it becomes weak and even incomplete.

Occipital condyle primitively single (tripartite), in advanced forms double
and formed by the exoccipitals. Primitively with suborbital fenestra, but some-
times reduced, large in Scaloposauria, absent in Cynodontia.

In contrast to the Alphatherapsida and Betatherapsida the Gamma-
therapsida were a very versatile group in which developments, besides leading
into a number of early as well as later blind alleys, produced very progressive
parallel branches all in a general mammalian direction, with the procynosuchid
—galesaurid—tritylodont branch most probably including the actual ancestors
of the first mammals.

A NEW CLASSIFICATION OF THE THERAPSIDA 321

SYNAPSIDA

The term Synapsida (Osborn 1903) has by all recent students been used
with the taxonomic rank of Subclass to include those vertebrates popularly
known as the mammal-like reptiles. These animals possess a mosaic of characters
some of which pertain to the Class Mammialia and others to the Class Reptilia.

They are thus neither true mammals nor true reptiles and do thus not fit
into the Class Mammalia or into the Class Reptilia.

I thus support those recent authors who have proposed that the name
Synapsida should have the rank of a separate Class.

The Class Synapsida would then include the two Subclasses Pelycosauria
and Therapsida.

For the Subclass Therapsida I propose the following classification :

Subclass Superorder Order Suborder Family

Therapsida Alphatherapsida | Eotitanosuchia Eotitanosuchidae
Phthinosuchidae
Rubidginidae

Dinocephalia Brithopia Brithopodidae
Anteosauridae

Titanosuchia Titanosuchidae
Tapinocephalidae
Styracocephalidae
Estemmeno-
suchidae

Gorgonopsia Hipposauridae
Gorgonopsidae
Burnetiidae

Betatherapsida Venyukovioidea Otsheriidae
Venyukoviidae

? Dromasauridae

Dicynodontia Endothiodontidae
Dicynodontidae
Kistecephalidae
Lystrosauridae
Kannemeyeriidae

Gammatherapsida | Therocephalia Pristerognathidae
Lycosuchidae
Whaitsiidae

Scaloposauria Ictidosuchia Alopecodontidae
Ictidosuchidae
Scaloposauridae

322 ANNALS OF THE SOUTH AFRICAN MUSEUM:

Subclass Superorder Order Suborder Family
Therapsida | Gammatherapsida Scaloposauria Bauriamorpha Bauriidae
(continued)} (continued) (continued) Ericiolacertidae
Cynodontia Procynosuchia Procynosuchidae
Galesauridae

? Silphedestidae

Cynognathia | Cynognathidae
Diademodontidae
Chiniquodontidae —
Traversodontidae

Tritylodontia Tritylodontidae
Trithelodontidae
Diarthrognathidae

DIAGNOSES OF THE HIGHER THERAPSID TAXA
SUBCLASS THERAPSIDA

Advanced synapsids of the Permian and Triassic. There is strong evidence
that one therapsid superorder, at least, was directly derived from sphenacodont
pelycosaurs, but the derivation of the other two superorders from sphenaco-
donts, although very probable, is less certain. The therapsids include the direct
ancestors of the mammals.

Further advanced than the pelycosaurs in that: the pterygo-basicranial
joint is no longer freely movable; a longitudinal girder is developed, the inter-
pterygoid vacuity is never widely open but partly or completely closed; the
squamosal is outflaring with a posterior face; there is no supratemporal; the
lacrimal never reaches the nostril and the maxilla is deep.

At the beginning of the Middle Permian the therapsids had already
developed in diverse directions each showing a lesser or greater acquisition of
certain mammalian characters.

Of the three main branches one became successfully adapted and domina-
ted the scene during the Middle and Upper Permian, comprising herbivores
and their predatory carnivores, but proved sterile; a second branch of herbi-
vores became adapted to their special niche, waxed exceedingly and very
successfully maintained themselves to near the end of the Triassic when they
died out without issue; the third branch, already well established at the begin-
ning of the Middle Permian, firstly as predators and later developing herbi-
vorous side branches, developed more and more in the mammalian direction,
with one or more twigs producing the first mammals late in the Triassic.

A NEW CLASSIFICATION OF THE THERAPSIDA 323

SUPERORDER ALPHATHERAPSIDA

Permian therapsids a stage further developed than the early Permian
sphenacodonts from which they arose, not leading to mammals.

The intertemporal skull table is primitively broad and flat, with the
posterior flange of the postorbital lying horizontally in the dorsal skull roof
(but modified in some Dinocephalia) and reaching the squamosal; the postor-
bital bar is always complete.

The lower jaw primitively without a prominent coronoid process (but
present in Gorgonopsia), the dentary always strong, but without a definite
postero-ventral angle; the postdentary bones always well developed.

The quadrate is primitively robust with the quadrate ramus of the ptery-
goid strong (except in Gorgonopsia).

Primitively with simple conical incisors, canine and postcanines (but
modified in some Dinocephalia) and palatal teeth on the pterygoid and palatine.

There is no secondary palate and no suborbital fenestra.

The epipterygoid is slender and does not enter the sidewall of the braincase;
the prootic is weakly developed with a free anterior edge; the sphenoidal com-
plex is well ossified.

The postfrontal is well developed; the dorsal premaxillary process is long
(but short in Gorgonopsia).

_ A pineal foramen is always present.

The occipital condyle is single.

ORDER EOTITANOSUCHIA

The most primitive therapsids descending from sphenacodonts, with all
the primitive characters listed in the diagnosis of the superorder Alphatherap-
sida of which they are the morphological ancestors.

Family Eotitanosuchidae
Eotitanosuchidae with the squamosal not extending into the intertemporal
skull roof and the primary palate closed.

Family Phthinosuchidae

Eotitanosuchians with the squamosal developing a lappet entering the
intertemporal skull roof and the primary palate with a median cleft.
Family Rubidginidae

Younger relict eotitanosuchians with a short series of serrated postcanines
and a small temporal fenestra.

ORDER GORGONOPSIA

Middle and Upper Permian alphatherapsids descending from eotitano-
suchians, which have developed a prominent coronoid process; the quadrate
posteriorly situated is reduced in size and the quadrate ramus of the pterygoid
is lightly built; the dorsal process of the premaxillary is shortened; a preparietal
is developed. Extinct.at the end of the Permian leaving no descendants.

324 ANNALS OF THE SOUTH AFRICAN MUSEUM

Family Hipposauridae

Primitive gorgonopsians with a very broad intertemporal skull roof; small
temporal fenestra, deep suspensorium, fairly long postcanine series, with the
dorsal skull contour strongly curved.

Family Gorgonopsidae

Intertemporal skull roof somewhat reduced in both earlier and later forms,
but in the latter sometimes secondarily greatly widened; large temporal fenestra,
fairly shallow suspensorium, postcanine series reduced, gape of jaws in some
Upper Permian forms enormous with very strong canines.

Family Burnetiidae

Later aberrant gorgonopsians with very wide intertemporal region,
reduced temporal fenestra, dentition reduced with weak teeth; with pachyosto-
tic thickening of roofbones of skull in the form of bosses and ridges.

ORDER DINOCEPHALIA

Early alphatherapsids derivable from an eotitanosuchian niveau and die
out without issue at the end of the Middle Permian; with basically primitive
structure obscured in some branches by pachyostosis and some abortive
specializations.

Width of intertemporal skull table reduced, sometimes greatly so, but
secondarily greatly widened where the pachyostosis is great, temporal fenestra
moderate to large except where secondarily reduced by the pachyostosis.

No coronoid process on the dentary; primitively with a carnivorous
dentition of simple conical teeth, later specialized carnivorous with an early
development of a herbivorous branch and intermeshing of some or all the upper
and lower batteries.

The quadrate robust, as also the quadrate ramus of the pterygoid;
quadrate shifting anteriorly.

Dorsal process of the premaxillary long to very long.

SUBORDER BRITHOPIA

Primitive dinocephalians linked to the eotitanosuchians and morpho-
logically ancestral to the other dinocephalian groups.

The intertemporal skull roof reduced in width with the posterior process
of the postorbital lying at a slant down from the horizontal; the temporal
fenestra large.

The dentition carnivorous, with the incisors tending to lengthen and the
postcanines becoming reduced; the lower and upper incisors and canines
intermesh.

Palatal teeth primitively well developed, later practically confined to the
palatine.

Quadratojugal never a surface bone.

With no or little general pachyostosis.

Dorsal process of the premaxillary moderately long.

A NEW CLASSIFICATION OF THE THERAPSIDA 325

Family Brithopodidae

Primitive brithopians with a fairly long postcanine series; incisors not
greatly lengthened; postfrontal not bulbously swollen and no other pachyo-
stosis; quadrate with little anterior shift; moderate outflaring of squamosals.
Palatal teeth well developed.

x

Family Anteosauridae

A stage further advanced than the brithopids. Postcanine series reduced,
incisors greatly lengthened. Postfrontals becoming greatly bulbously swollen
and the skull roof moderately pachyostosed. Strong outflaring of squamosals,
especially posteriorly.

Palatal teeth reduced, practically confined to the palatine.

SUBORDER TITANOSUCHIA

Advanced dinocephalians derived from a brithopid niveau.

The intertemporal region reduced in width, sometimes to a narrow sagittal
crest, but secondarily greatly to enormously widened where the pachyostosis is
great; the temporal fenestra large to very large but secondarily greatly reduced.

The dentition is herbivorous, initially with a large conical canine and with
only the incisors developing a talon and heel; later the canine is not distin-
guishable as such and all the marginal teeth develop a talon and heel; the
postcanine series always very long. Palatal teeth practically absent.

Quadratojugal sometimes a surface bone.

The pachyostosis is moderate to enormous.

Family Titanosuchidae

Primitive titanosuchians developed from a brithopid level and indicating
the morphological level from which the tapinocephalids arose.

The intertemporal width reduced with a low thick sagittal crest, posterior
process of the postorbital reduced, temporal fenestra fairly large, but squamo-
sals not outflaring.

Strong incisors with piercing talon and crushing heel and large conical
canines, a long series of spatulate postcanines, which do not intermesh as do the
incisors and canines of the two jaws. No palatal teeth.

Moderate pachyostosis.

Family Tapinocephalidae

Specialized titanosuchians derived from a titanosuchid level.

The intertemporal region very variable, mostly of moderate width, some-
times with a sharp sagittal crest, in one subfamily enormously broadened as a
result of the excessive pachyostosis; temporal fenestra large to greatly reduced.

All the marginal teeth with talon and heel, upper and lower battery
intermeshing.

Pachyostosis light to great.

Family Styracocephalidae
Middle Permian aberrant titanosuchians. Intertemporal region very

326 ANNALS OF THE SOUTH AFRICAN MUSEUM

broad, but temporal fossa roomy with posteriorly flaring squamosal; pachy-
ostosis in the form of ‘horns’ and bosses.

Weak conical incisors and canine and a long series of postcanines; palatal
teeth very well developed, even on the vomer.

Family Estemmenosuchidae

The ‘horns’ situated on the frontals, and directed dorsally, whereas in the
Styracocephalidae the ‘horns’ are formed by the tabular and directed pos-
teriorly. Otherwise with features very similar to those of the Styracocephalidae.

SUPERORDER BETATHERAPSIDA

Permian and Triassic therapsids on a developmental niveau far above that
of the early Permian sphenacodonts, not leading to mammals.

The intertemporal skull table is primitively reduced in width, but flat,
becoming narrow and later developing a sagittal crest, but secondarily widened
in the Kistecephalidae; the posterior process of the postorbital inclined down-
wards from the horizontal and reaching the squamosal, later reduced; the
postorbital bar is always complete.

The lower jaw without a coronoid process of the dentary and without a
coronoid bone; the dentary always strong without a postero-ventral angle; the
postdentary bones well developed. The quadrate is robust lying low down on a
pedicel of the uniquely triradiate squamosal, and the quadrate ramus of the
pterygoid is weak. Primitively with a modified set of marginal teeth in both
jaws, which very early are radically reduced and sometimes wholly lost and
replaced by horny sheaths; there are no palatal teeth.

The premaxillaries, primitively paired but later fused, develop plates to
form a unique type of secondary palate and the choanae are shifted posteriorly;
primitively the maxilla and palatine have no inward palatal growth but later
extend palatally but never meet below the air passage. The epipterygoid is
slender and does not enter the sidewall of the braincase; the prootic is weakly
developed with a free anterior edge; the sphenoidal complex is well ossified and
lies far anteriorly.

The postfrontal is primitively well developed but reduced later; the dorsal
premaxillary processes are primitively paired and long, later fused and short.

A pineal foramen present; preparietal primitively absent, later present.
The occipital condyle is single. ‘There is a fenestra in the mandible between the
dentary and angular.

ORDER VENYUKOVIOIDEA

Primitive betatherapsids not directly linked to the sphenacodonts, leading
to the higher Dicynodontia.

The width of the intertemporal region is reduced, without sagittal crest
and the temporal fenestra is short, the jugal has a large entry into the zygoma
which is not strongly everted.

A NEW CLASSIFICATION OF THE THERAPSIDA 327

The dentition consists of a well-developed series of bluntly conical marginal
teeth on the premaxilla, maxilla and dentary.

No inward growth of palatine and maxilla and the posterior part of the
palate is thus primitive, except that the lateral ramus of the pterygoid is
somewhat or much reduced. ,

The premaxillaries are not fused and have a long dorsal process; the
septomaxilla is largely superficial, the postfrontal is well developed and there is
no preparietal.

Family Otsheriidae

The incisors are enlarged, the choana is short, the palatine does not meet
the premaxilla; the lateral ramus of the pterygoid is still prominent, the
lacrimal is short and there is no pachyostosis.

Family Venyukoviidae

The incisors are enlarged, the choana is long; the palatine meets the
premaxilla; the lateral ramus of the pterygoid is much reduced; the lacrimal
is long; there is some pachyostosis.

Family Dromasauridae
The dentition consists of a series of isodont marginal teeth or the jaws are
edentulous; the temporal fenestra is very short and deep.

ORDER DICYNODONTIA

Advanced betatherapsids morphologially derivable from Otsheria; a
long-lived order, rich in species varying in minor characters with a single main
theme and phylogenetically sterile.

The intertemporal region reduced in width, sometimes very much so, but
is secondarily widened in one aberrant family; the sagittal crest feeble to very
high or wholly absent. The temporal fenestra is very long and the jugal is
practically ousted from the zygoma by the squamosal which is uniquely
everted.

The marginal teeth are greatly modified, there are never any incisors;
an upper conical canine present or absent; postcanines present or absent and
when present shifted medially and variously disposed.

Palatal flanges of the palatine and maxilla tending to grow inwards to form
a variable open trough for the air passage; the lateral ramus of the pterygoid
is lost.

The premaxillaries are fused and the dorsal process is short; the septo-
maxilla tending to shift interiorally; the postfrontal is primitively present but
is lost in later forms; a preparietal is developed.

Family Endothiodontidae
With postcanine teeth, number and disposition very variable, canines
present or absent, with a postfrontal.

328 ANNALS OF THE SOUTH AFRICAN MUSEUM

Family Dicynodontidae
Without postcanine teeth, canines present or absent, the postfrontal is
frequently absent.

Family Kistecephalidae

The intertemporal region is secondarily greatly widened, without sagittal
crest, edentulous, without canines, without pre- and postfrontals and pre-
parietal.

Family Lystrosauridae
Without postcanines and canines usually present, postfrontal present,
nares shifted posteriorly and premaxilla lengthened.

Family Kannemeyeriidae
Very high sagittal crest; depression leading into pineal foramen.

SUPERORDER GAMMATHERAPSIDA

Permian and Triassic therapsids probably derived from Early Permian
sphenacodonts and including the immediate ancestors of the mammals.

The intertemporal skull table narrow, usually with a sagittal crest, but
secondarily widened in one late family; the postorbital is reduced and never
reaches the squamosal; the postorbital bar primitively and usually complete,
but incomplete and even wholly absent in some advanced forms.

The dentary primitively and usually with a prominant coronoid process,
strong and finally greatly enlarged with a strong postero-ventral angle and
making contact with the squamosal; the postdentary bones primitively well
developed but greatly reduced in some advanced groups.

The quadrate small, with weak to incomplete quadrate ramus of the
pterygoid.

Primitively with carnivorous dentition of simple conical teeth, later very
variable, often with accessory cusps in the postcanines and in some advanced
forms with highly elaborated crowns of a mammalian nature.

Primitively without secondary palate, in later groups incipient in various
ways, to incomplete, and finally fully developed in mammalian fashion.

The epipterygoid is primitively slender and remains so in many forms, but
is widened in some early forms and in advanced forms very broad and
incorporated into the sidewall of the braincase; the prootic is primitively weakly
developed, but later growing forwards meets the epipterygoid (alisphenoid)
suturally; the sphenoidal complex is usually not well ossified.

The postfrontal small or absent.

The occipital condyle single in earlier forms, later becomes notched and
finally with double condyles formed by the exoccipitals.

ORDER THEROCEPHALIA

Middle to Upper Permian gammatherapsids; initially primitive but with
a wide gap between them and their sphenacodont precursors; with a degenerate
family in the Upper Permian; relation to higher gammatherapsids is uncertain.

A NEW CLASSIFICATION OF THE THERAPSIDA 329

The intertemporal region is narrow with a reduced postorbital; the
postorbital bar always complete; pineal foramen always present; dentary with
a well-developed coronoid process and the postdentary bones well developed;
no prominent postero-ventral angle to dentary.

Primitively without secondary palate but aberrantly incipient in the
Upper Permian family.

Epipterygoid slender or widened but never incorporated in the sidewall
of the braincase.

Dentition primitively carnivorous with a long postcanine series, later
greatly reduced to lost, always uncusped. Postfrontal small or absent, suborbital
fenestra large in earlier forms but reduced to absent in later forms.

The occipital condyle is always single.

Family Pristerognathidae
Middle Permian therocephalians with a well-developed carnivorous

dentition with a single canine; no secondary palate, epipterygoid slender,
postfrontal small, suborbital fenestra large.

Family Lycosuchidae

Middle Permian therocephalians with a well-developed carnivorous
dentition, with double canines; no secondary palate, epipterygoid widened;
postfrontal small, suborbital fenestra large.

Family Whaitsiidae (including Lycedeopsidae and Euchambersiidae as
subfamilies)

Upper Permian therocephalians with reduced dentition, postcanines
feeble, few or wholly absent, lower incisors sometimes absent, aberrant develop-
ment of an incipient secondary palate, epipterygoid widened, suborbital
fenestra large to small or absent, dentary scimitar-shaped with postdentary
bones not robust.

ORDER SCALOPOSAURIA

Middle Permian to Lower ‘Triassic gammatherapsids, probably independ-
ently derived from sphenacodonts with a considerable gap; fairly primitive in
the Middle Permian, but advanced in the Upper Permian and Lower Triassic;
relations to cynodonts uncertain.

The intertemporal region is usually narrow with a sagittal crest, but
widened in one family with loss of crest; the postorbital bar is slender and
complete or incomplete; the postorbital is sometimes greatly reduced or even
absent; the jugal spur of the postorbital bar is usually present but absent in a
few forms; the pineal foramen is sometimes absent.

The coronoid process of the dentary is strong, feeble or absent; there is no
prominant postero-ventral angle to the dentary; the postdentary bones are well
developed or weakened. A secondary palate is primitively absent, later
incipient to well developed. The epipterygoid is mostly slender but sometimes
widened and partially included in the sidewall of the braincase in one form.

330 ANNALS OF THE SOUTH AFRICAN MUSEUM

A prominent canine is usually retained but sometimes not recognizable as
such, accessory small canines are usually present, maxillary teeth usually
numerous and the postcanines variable, being simple, cusped or with trans-
versely widened crowns.

The suborbital fenestrae are always well developed, the postfrontal is
reduced or absent.

The occipital condyle is initially single but later sometimes incipiently
double.

SUBORDER ICTIDOSUCHIA

Mostly Permian scaloposaurians just extending into the Triassic; the
intertemporal region is usually narrow but later widened in one family, the
postdentary bones are weakened and the dentary lightly built; the secondary
palate is primitively absent, sometimes incipient but never complete, the
epipterygoid is slender but in one case partially enters the sidewall of the
braincase.

Family Alopecodontidae

Primitive Middle Permian ictidosuchians close to the contemporary
therocephalians.

The intertemporal region is narrow with a sagittal crest; the postorbital
bar is complete, the coronoid process is strong, the dentary robust and the
postdentary bones well developed; there is no secondary palate.

Dentition carnivorous, always with a prominent canine and two small
accessories, the postcanines are simple conical teeth, the postfrontal is small and
the occipital condyle single.

Family Ictidosuchidae

Upper Permian ictidosuchians linked to the alopecodontids.

The intertemporal region is narrow, usually with a sagittal crest and a
pineal foramen present; primitively with a complete postorbital bar, but later
incomplete; the coronoid process is prominent, the secondary palate absent,
incipient to weakly developed; with an enlarged canine behind smaller acces-
sories and the postcanines simple conical teeth, a single occipital condyle.

Family Scaloposauridae

Permian to Lower Triassic ictidosuchians.

Primitively with a narrow intertemporal region, but in some later forms
this is widened and the pineal foramen is often absent.

Primitively with a complete postorbital bar, but later incomplete and
sometimes even without a jugal spur.

The coronoid is weak or absent.

The secondary palate is incipient to weakly developed.

Primitively with an enlarged main canine, but in advanced forms not
distinguishable, the postcanines are primitively conical but sometimes cusped.

A NEW CLASSIFICATION OF THE THERAPSIDA 331

SUBORDER BAURIAMORPHA

Triassic scaloposaurians further advanced than the ictidosuchians; the
intertemporal region is narrow, usually with a sagittal crest and pineal foramen.

Postorbital bar, complete or incomplete.

Dentary with a prominent coronoid process; weak or robust with post-
dentary bones well developed or weak; with a well-developed closed secondary
palate; epipterygoid moderately widened, lying lateral to braincase.

Incisors and canine conical or peglike, postcanines with cusps and trans-
versely expanded. Occipital condyle notched or double.

Family Bauriidae

Intertemporal region narrow with sagittal crest; pineal foramen absent
or present; the postorbital bar complete or incomplete, sometimes without
jugal spur.

Dentary strong with well-developed postdentary bones. The vomer does
not enter the secondary palate. Incisors and prominent canine conical and
postcanines expanded.

Family Ericiolacertidae

Intertemporal region broadened, without sagittal crest, no pineal foramen;
postorbital bar incomplete, without a jugal spur.

Dentary and postdentary bones lightly built.

The vomer enters the secondary palate.

No outstanding canine, incisors modified and postcanines peglike with
cusps and expanded transversely.

ORDER CYNODONTIA

Advanced Upper Permian and Triassic gammatherapsids, derived from
sphenacodonts probably through an intermediate stage at a morphological
level near that of the Middle Permian therocephalians and scaloposaurians;
including the ancestors of the mammals.

The intertemporal region is narrow with a sagittal crest, pineal foramen
primitively present but later lost; postorbital bar complete in earlier forms but
later incomplete. Dentary with weak to very strong coronoid process, the
postero-ventral angle to the dentary is primitively weakly developed but very
prominent in later forms; primitively without a posterior process but this is
developed in later forms and in some forms reaches the squamosal to form an
accessory articulation; the postdentary bones well developed in earlier forms
but later much reduced. Primitively with a cleft secondary palate, but later
closed.

The epipterygoid is widened and enters the sidewall of the braincase and
becomes suturally joined to the prootic.

The dentition primitively with conical incisors and canines sometimes with
accessory small canines; the postcanines developing cusps and later with
widened variously elaborated crowns; primitively polyphyodont later diphyo-

332 ANNALS OF THE SOUTH AFRICAN MUSEUM

dont with ‘premolars’ and ‘molars’ distinguishable.
Postfrontal lost; no suborbital fenestra; occipital condyle notched and later
double.

SUBORDER PROCYNOSUCHIA

Primitive Upper Permian and Triassic cynodonts, related to the two older
gammatherapsid orders and linked to the first mammals.

Pineal foramen present and postorbital bar complete, zygoma lightly built.

In the dentary the postero-ventral angle is absent or only moderately
developed; a masseteric fossa on the coronoid process is incipient to fairly well
developed but there is still no masseteric process on the jugal; the posterior
process of the dentary is still undeveloped; the postdentary bones are still well
developed, but the reflected lamina of the angular is reduced. Initially the
secondary palate is still cleft but is later closed.

The incisors and canine conical with accessory small canines sometimes
present; accessory cusps on the postcanines.

Family Procynosuchidae

Upper Permian primitive procynosuchians still with a cleft palate, fairly
weak coronoid process, with accessory precanine maxillary teeth; the occipital
condyle is incipiently double.

Family Galesauridae

Upper Permian and Lower Triassic procynosuchians with a closed
secondary palate, strong coronoid process without precanine maxillary teeth,
and a double occipital condyle.

SUBORDER CYNOGNATHIA

Advanced specialized Triassic cynodonts, with a carnivorous and herbi-
vorous branch, derived from Upper Permian procynosuchians, becoming
extinct in the Upper Triassic.

Pineal foramen present, postorbital bar complete and a very strong
zygoma.

Dentary greatly enlarged, prominent and strong coronoid process with
masseteric fossa fairly to very well developed; strong to very strong postero-
ventral angle; posterior process of dentary moderately to well developed and
in some advanced forms making contact with the squamosal in an accessory
articulation; masseteric process on jugal present or absent with a step between
maxilla and jugal; postdentary bones greatly reduced with all but loss of
reflected lamina of the angular. Secondary palate well developed and closed.
Incisors and canines conical, without accessory anterior canines, postcanines
with fore and aft accessory cusps or with crowns transversely expanded and
further elaborated.

Family Cynognathidae
Earlier carnivorous cynognathians with the dentary not making contact

A NEW CLASSIFICATION OF THE THERAPSIDA 333

with the squamosal, and with small angular process and jugal process; maxillary
teeth divided into premolars with crenulated crowns and molars with a series
of sectorial cusps in a longitudinal row.

Family Chiniquodontidae

Later carnivorous cynognathians with the dentary making contact with
the squamosal in some advanced forms, usually no angular process to the
dentary.

Family Diademodontidae

Earlier herbivorous cynognathians with very strong masseteric process on
the jugal; long series of maxillary teeth with peg-like premolars and transversely
widened crushing molars.

Family Traversodontidae

Later herbivorous cynognathians with a step between the maxilla and jugal
and no masseteric process on the jugal.

SUBORDER TRITYLODONTIA

Advanced Upper Triassic cynodonts, derived from Upper Permian
procynosuchians.

Without pineal foramen, postorbital bar incomplete without postorbital
and postfrontal; zygoma very strong or fairly weak.

Dentary greatly enlarged with strong coronoid process and well developed
masseteric fossa; postero-ventral angle very prominent, posterior process of
dentary well developed and making contact with the squamosal in advanced
forms; postdentary bones greatly reduced; no jugal process.

Secondary palate closed but greatly reduced in width with median shift of
postcanines. Transverse ramus of pterygoid reduced.

Incisors primitively conical or specialized and recumbent, conical canine
present or absent, postcanines cusped and further elaborated with crushing
crowns.

Mononarial or binarial.

Family Trithelodontidae

Zygoma moderately strong; posterior process of the dentary not reaching
the squamosal; incisors primitively conical, long diastema, long series of
widened molars; mononarial.

Family Tritylodontidae

Zygoma strong; posterior process of the dentary not meeting the squamosal ;
incisors reduced, one enlarged, recumbent in dentary, no canine, long diastema,
molars quadrangular with elaborate crushing crowns; mononarial.

Family Diarthrognathidae

Zygoma fairly weak; posterior process of the dentary making contact with
the squamosal in an accessory articulation; incisors and canine primitively
conical, no diastema, molars transversely widened with cusps; binarial.

334 ANNALS OF THE SOUTH AFRICAN MUSEUM

SUMMARY

The classification of the Therapsida is re-evaluated and the older view of
two main branches, Anomodontia and Theriodontia, discarded in favour of
three main branches for which the names Alphatherapsida, Betatherapsida and
Gammatherapsida are proposed.

REFERENCES

Boonstra, L. D. 1953. A new scaloposaurian genus. Ann. Mag. nat. Hist. (12) 6: 601-605.

Boonstra, L. D. 1954. The cranial structure of the titanosuchian: Anteosaurus. Ann. S. Afr. Mus.
42: 108-148.

Boonstra, L. D. 1963. Early dichotomies in the therapsids. S$. Afr. 7. Sci. 5g: 176-195.

Brink, A. S. 1963. The taxonomic position of the Synapsida. S. Afr. F. Sci. 5g: 153-159.

Broom, R. 1903. On the classification of the theriodonts and their allies. Rep. S. Afr. Ass. Adumt
Sci. 13 286-204.

Broom, R. 1905. On the use of the term Anomodontia. Rec. Albany Mus. 1: 266-269.

Broom, R. 1912. On some new fossil reptiles from the Permian and Triassic beds of South Africa.
Proc. zool. Soc. Lond. 1912: 859-876.

Broom, R. 1923. On the structure of the skull in the carnivorous dinocephalian reptiles. Proc.
zool. Soc. Lond. 1923: 661-684.

Broom, R. 1931. Notices of some new genera and species of Karroo fossil reptiles. Rec. Albany
Mus. 4: 161-166.

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

Erremovy, I. A. 1940. Preliminary description of the new Permian and Triassic Tetrapoda from
USSR. Trudy paleont. Inst. 10 (2): 1-140.

Erremov, I. A. 1954. [A fauna of terrestrial vertebrates in the Permian cupriferous sandstones
of the western Ural region.] Trudy paleont. Inst. 54: 1-416. (In Russian.)

Haucnuton, S. H. 1918. Investigations in South African fossil reptiles and Amphibia. (Part 11.)
Some new carnivorous therapsids, with notes upon the brain-case in certain species. Ann.
S. Afr. Mus. 12: 175-216.

Haucurton, S. H. 1924. A bibliographic list of pre-Stormberg Karroo fossils. Trans. R. Soc. S. Afr.
12: 51-104.

HaucutTon, S. H. 1929. On some new therapsid genera. Ann. S. Afr. Mus. 28: 55-78.

HuEng, F. von. 1948. Short review of the lower tetrapods. Jn ROYAL SOCIETY OF SOUTH AFRICA.
Robert Broom commemorative volume: 65-106. Cape Town: Royal Society of South Africa.
(Special publication.)

Kutoroea, S. 1838. Beitrag zur Kenntniss der organischen Ueberreste des Kupfersandsteins am westlichen
Abhange des Urals. St. Petersburg: Eggers.

LyDEKKER, R. 1890. Catalogue of the fossil Reptilia and Amphibia in the British Museum. Part IV.
London: British Museum.

Oszorn, H. F. 1903. The reptilian subclasses Diapsida and Synapsida and the early history of the
Diaptosauria. Mem. Am. Mus. nat. Hist. 1: 449-507.

Owen, R. 1844. Description of certain fossil crania, discovered by A. G. Bain, Esq., in sandstone
rocks at the south-eastern extremity of Africa, referable to different species of an extinct
genus of Reptilia (Dicynodon) and indicative of a new tribe or sub-order of Sauria. Proc.
geol. Soc. Lond. 4: 500-504.

Owen, 1859 (1860). On the orders of fossil and recent Reptilia, and their distribution in time.
Rep. Br. Ass. Adumt Sci. (Aberdeen, 1859) 29: 153-166.

Owen, R. 1860. Palaeontology ; or, A systematic summary of extinct animals and their relations. Edinburgh:
Longman.

A NEW CLASSIFICATION OF THE THERAPSIDA 335

Owen, R. 1876. Descriptive and illustrated catalogue of the fossil Reptilia of South Africa in the collection
of the British Museum. London: British Museum.

REED, C. 1960. Polyphyletic or monophyletic ancestry of mammals, or: what is a class? Evolution
14: 314-322.

Romer, A. S. 1961. Synapsid evolution and dentition. In International colloquium on the evolution of
lower and non specialized mammals . . . September 1960. 1: 9-56. Brussels: Koninklijke Vlaamse
Academie voor Wetenschappen, Letteren en Schone Kunsten van Belgie.

SEELEY, H. G. 1895. Researches on the structure, organization, and classification of the fossil
Reptilia. Part IX, section 1. On the Therosuchia. Phil. Trans. R. Soc. (B) 183: 311-370.

Smpson, G. G. 1925. A Mesozoic mammal skull from Mongolia. Am. Mus. Novit. 201: 1-11.

Tcuupinov, P. K. 1960. [Upper Permian therapsids of the Ezhovo locality.] Paleont. Zh. 1960
(4): 81-94. (In Russian.)

VAN VALEN, L. 1960. Therapsids as mammals. Evolution 14: 304-313.

Watson, D. M. S. 1917. Sketch classification of the pre-Jurassic tetrapod vertebrates. Proc. zool.
Soc. Lond. 1917: 167-186.

Watson, D. M.S. & Romer, A. S. 1956. A classification of therapsid reptiles. Bull. Mus. comp.
Kool. Harv. 114: 37-89.

336 ANNALS OF THE SOUTH AFRICAN MUSEUM

DEVONIAN CARBONIFEROUS PERMIAN | TRIASSIC JURASSIC CRETACEOUS | TERTIARY

uelpeuuuieu-a4q

uel sipidiyy

pue plurysoydes
pisdesoyy,

SNOJIUIOJOQUIT
prauoposeuayds

/
/

S, _eridoy

BS erquyduiy

ust
7
PD endo

2
6
Qa.
7
x
z
--
fa

Fic. 1. Schematic representation of the evolutionary story of the mammals. The diagram is
based on the number of described genera.

We commence the story nearly 400 million years ago.

At that time (the Devonian) there lived a group of freshwater fish, known as the Rhipidistia,
whose paired fins had become adapted to propelling the body forwards with a purchase on the
muddy floor of shallow pools.

These rhipidistians were succeeded (during the Carboniferous) by a group of amphibians,
known as the Embolomeri, with two pairs of extremities capable of ungainly locomotion on
dry land and with the ability of utilizing atmospheric oxygen but returning to the water for
reproduction.

Just before the next period (the Permian) we encounter two groups of reptiles that had
become completely adapted to life on land. These were the primitive Captorhinidae, with a
simple adductor muscular mass for closing the jaw when feeding and the Sphenacodontia, where
the adductor muscles had developed into a more efficient mechanism for feeding and able to
raise the body from the ground for better locomotion.

Arising from these early reptiles (during the later part of the Permian) we have the
Therapsida, which evolved in various directions, but with this in common viz. a great improve-
ment in the jaw mechanism and locomotor ability. The majority of the therapsids, developing
along differing but nearly parallel lines, became extinct, some in the Permian and others
successfully competing to the end of the Trias.

Of the more successful therapsids we indicate in the scheme a group consisting of the
familiar Scaloposauridae, Bauriidae, Tritylodontidae and Diarthrognathidae, brigading them

A NEW CLASSIFICATION OF THE THERAPSIDA 337

together as the Premammalian Therapsida. In this group the locomotor ability is greatly
improved and the jaw mechanism adapted more and more for chewing their food. For the latter
the lower jaw became more and more dominated by a single bone— the dentary —and the teeth
developed shearing, cutting and crushing cusps and the respiration during the chewing process
was facilitated by the development of a bony secondary palate, separating the air passage from
the buccal cavity.

Up to the end of the Triassic period all the vertebrates were poikilothermic or ‘cold-blooded’,
i.e. they had no built in mechanism for temperature control. About this time certain of the higher
therapsids, with their higher rate of metabolism, made possible by the improved locomotor and
masticatory ability, developed mechanisms to dissipate excess body heat or to conserve it.
For the former a skin with glands for sweating and a diaphragm for panting became imperative.
To conserve heat the development of an insulating cover of hair or fur took place. These features,
together with the dentary-squamosal jaw hinge made these small rat-like creatures mammals.

From their beginning late in the Triassic (about 150 million years ago) these first mammals
were small rat-like animals forming a very inconspicuous part of the vertebrate fauna. This
“continued throughout the Jurassic and Cretaceous, when vertebrate life was dominated by the
sauropsid reptiles which included the dinosaurs during their heyday.

‘But from the Tertiary the mammals waxed exceedingly to fill every possible ecological

niche including besides terrestrial conditions varying from arctic to tropical climates, excursions
into fresh and salt water and into the air. The culminating event, less than a million years ago,
is the emergence of Man. SS

Since then this single genus has attained a dominant position in the living world, which it is
ravaging at an alarming rate.

338 ANNALS OF THE SOUTH AFRICAN MUSEUM

LOWER PERMIAN MIDDLE PERMIAN PPER PERMIANLOWER TRIAS |MIDTRIASUPPERTRIAS| —_|

+ } .
VS nos
2 ene
S---------- >_~— ee ce
p Ti

VISCAONODUOD
VITVHdd DONIC
VIHONSONV.LILOF

(SNGOLdV Hi). LNOGOOVNAHdS

2)

VILNOGONA
VIYNVSOdOTVOS
VITVHdd SOUTH

(QLNOGODVNdHdS

VILNOGONADIC

Otsheriidae
nes

(Q.LNOGODVNaAHdS

— ae Farag regs Molteno | Red (G3
| i eagle, -witicl eruleed a \auiipue ereonae Erect ceisias SRST Geta | Malem | Red PREY |

Fic. 2. Schematic phylogeny of the Therapsida.

INSTRUCTIONS FO AUTHORS

Based on

CONFERENCE OF BIOLOGICAL EDITORS, COMMITTEE ON FORM AND STYLE. 1960.
Style manual for biological journals. Washington: American Institute of Biological Sciences.

‘

MANUSCRIPT

To be typewritten, double spaced, with good margins, arranged in the following order:
(1) Heading, consisting of informative but brief title, name(s) of author(s), address(es) of
author(s), number of illustrations (plates, figures, enumerated maps and tables) in the article.
(2) Contents. (3) The main text, divided into principal divisions with major headings; sub-
headings to be used sparingly and enumeration of headings to be avoided. (4) Summary.
(5) Acknowledgements. (6) References, as below. (7) Key to lettering of figures. (8) Explana-
tion to plates.

ILLUSTRATIONS

To be reducible to 12 cm Xx 18 cm (19 cm including caption). A metric scale to appear with
all photographs.

REFERENCES

Harvard system (name and year) to be used: author’s name and year of publication given
in text; full references at the end of the article, arranged alphabetically by names, chronologi-
cally within each name, with suffixes a, b, etc. to the year for more than one paper by the same
author in that year.

For books give title in italics, edition, volume number, place of publication, publisher.

For journal articles give title of article, title of journal in italics (abbreviated according to
the World list of scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses,
volume number, part number (only if independently paged) in parentheses, pagination.

Examples (note capitalization and punctuation)

Butitoucu, W. S. 1960. Practical invertebrate anatomy. 2nd ed. London: Macmillan.

FiscHER, P.-H. 1948. Données sur la résistance et de le vitalité des mollusques. 7. Conch., Parts
88: 100-140.

FiscHER, P.-H., Duvau, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines.
Archs Zool. exp. gén. 74: 627-634.

Koun, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee
region of Ceylon. Ann. Mag. nat. Hist. (13) 2: 309-320.

Koun, A. J. 19605. Spawning behaviour, egg masses and larval development in Conus from the
Indian Ocean. Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. Jn scHULTZE, L.
Koologische und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-
Afrika. 4: 269-270. Jena: Fischer. Denkschr. med-naturw. Ges. Jena 16: 269-270.

ZOOLOGICAL NOMENCLATURE

To be governed by the rulings of the latest International code of zoological nomenclature issued
by the International Trust for Zoological Nomenclature (particularly articles 22 and 51). The
Harvard system of reference to be used in the synonymy lists, with the full references incorporated
in the list at the end of the article, and not given in contracted form in the synonymy list.

Example
Scalaria coronata Lamarck, 1816: pl. 451, figs 5 a, 6; Liste: 11. Turton, 1932: 80.

ce

v

£
TS
f Pee RES
= ee P fl ¥ ay
i < 2 " = eg ‘
at) y- aK. P pate “jae
% ° nny ~ he) a
~ oa >
“ 2 7 v .
" ie oF ‘a
pe,
_ + 6 oft 12 oe ‘ ,
Ps . J ae .
—~— 5 3
— id 4 7 is
oo
= ‘
;
» -
3 a : i
Fi “ i
r af oc eek
7 =
*
<a *
ul, “ -
we ‘
“4
’ " we 5
/
= “4
» >
y ,
i ‘ : er)
®
= " :
‘ = . ‘
» ~ b
x !
, ‘ =
a .
= 2 —
=) oes P g ,
£ P 3
be i
r] 3 =
1
ee
k so ” ;
J 7 > om ©
oe
‘ ~Niyg
aes ay
an S * .
>
= 3
oy 2
' ¢
= f
j
a
- :
5 3 he
4 ;
. m
S& ,
“ .
. « an
adie i
* ‘ n
!
wT
~ ?
te ee,
J
_> / f
- ., x y
+ Py me ; Li

Si ey COS = 2G 4 pit =) aes mae |=. oO} §— WSS * > Fe {9 mi
; jx] DL £5 app me yy “aad 4 SF hes es 2 \ — \e | 7
muse” om ord ov = “if? Ay LON sy 2 habe a WS G Yofab <i y |
eo , rani ” ; ° oe ‘ | :
ITUTION NOILNLILSNI_ NVINOSHLINS S33 luvud Mul BRARI ES SMITHSONIAN INSTITUTION
YX = Wes = <= = ee K |
YX or <8 |
EE 2 B g e 2 NR
AN z = Z = S = y |
YUVUGIT_ LIBRAR!I ES SMITHSONIAN INSTITUTION NOLLAIILSNt ae fyVvy |
a a 4 a = & = ams Oe
= a. = o aa ao a3\
‘s “4 =)
= a c = a < :
SS a Ye: O <= oO fa
. Fa — = at = © eae
ITUTION | NOLLALILSNI_NVINOSHLINS _ $3 huvuyd tl BRARI ES_ SMITHSONIERS
“Ss
2. ee 2 oo = yy,
= me \y = a S a) oes
= = SW i > e y >
i- a S\ \Y r me = . Oe = 4 ‘SS
LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S31iyuvyaT}T
z = z= = = =
S NX re) = Ae a
Ss ASS o aie: a S ere © Oo
& : 2 = ie = z
= “SS > = hie >" = > j
79) = ” 23 z ; 7p) . =
NOLLNLILSNI NVINOSHLINS S23 lYVYEIT LIBRARIES SMITHSONIAN INSTITUTION
” z = w” De =
TT) Th tad iW ,
c =a e 2 = Uy?
a = ee a. < Gf fea
a = = of pe yn
e O ae, Pat ca - rat
wl = ond = a = i
uvudia LIBRARI ES_ SMITHSONIAN INSTITUTION NOLO TITSN! iNG-) 3 iuvud
tg LY 0 “9 ie
EM = = > >
i i: Yes Ly om - ie -
= “ijl m — rm 72)
= on = uw | Ro
SMITHSONIAN

NOILNLILSNI NVINOSHLINS S3IYVYdEIT LIBRARIES

NVINOSHLINS S3!1Y¥Vudil

w m w
Z SN ae z =<
= SS Sy tee a z
: = ay E a :
= > GS = > Ss ;.
w ma Vv? . 75 w)
LIBRARI ES SMITHSONIAN _ pNOILALILSNI NVINOSHLIWS S32 1yYvug
7 ss ‘ a us = uw Sona yy
= = a = E :
=. < i i < =f < |
Cc oe sar oe Ss o
me om — [0] om oO
wee AG Oo —— — f
2 ow a ol S . i F
NOILNLILSNI_NVINOSHLINS $3 fuvud Dit LIBRARI ES_ SMITHSONIAN INSTITU
Fu oa! ‘ ey et , ‘
= ye = o = wo PE |
5 z —~ a a j =
- 3 = res = “oy
~ — se ft , ane
MYYGIT LI B RARI ES SMITHSONIAN INSTITUTION NOLLALILSNI NVINOSHIIWS (23 1yvugl
: Sar 7: “* ip j
; <x . = <Gi> x ‘ z= . < — z dl
fll, = &. & YL = RO VE ee 2S eee flu'"”,

|] = “GyyG nxn \, J = i Ja WO = J] 2
= UG > Sur = Nay 2 SN fe
y D Viv m Non oc ” sy m o x li my m

= 7 = . £ 7)
JLSNI NVINOSHLINS S3IYVYdIT LIBRARIES SMITHSONIAN

%
m
7)
” 7)
z 4 2 & & z G& 5
y © 2 i ‘See F ee
= = , = > = ae
n = no ofp 77) ae
RIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVYUsII_
w > “ny 3 - a Zz
a : «”? a9 wn Moss _ i!
om | Ay, =« =-{ x
< ivf Lp > < = ON =
¢ Wass Cc axed ‘ G
«Gy 4 = a ~. ps 4
ea) —= © _ ran) a
pital O _ oO ~ Oo
inal =z a = ahi =
LIBRARIES SMITHSONIAN INSTITUTION
a z c a ie z
o = w = Ly OD =
> Ys = 2 =, GY ‘= S
‘ae NSS ce > i Le ti +s i=
ee th Ne sae: 2
RIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SSIYVYUSIT
on Se wn z 7) z
= < = V4 < = <
\. on D w Bip a oH n
SSW CO a ves Vd fb :, XL. Oo =
S = é = 7 | = Z =
¥i 2 e S 3 = 77

ILSNI NVINOSHLINS S3IYVYdIT LIBRARIES SMITHSONIAN

Mi 5 ni
2 ul z ul
o nf at ae
Ss co 5 mm
GC WN ST Cc =
cS ‘!, c
OBS ata 5 =
a | a a |
RIES SMITHSONIAN INSTITUTION NOILNLILSNI
x E ie z oy a
7 ny 4 2 a ow SS . :
> = a. 7. BS
= bys La ca = a ‘SN
= ’GF = = =
P 8 «ee »% 7 €
LILSNI_ NVINOSHLINS S31uVvus aot BRARI ES SMITHSONIAN
Ses = eae Ks =
oe Y* = cam | aS , <= SAY YK a
G: = i 2 NS in
= > : = >
”) za Ww za

\RIES SMITHSONIAN INSTITUTION NOJLOLILSNI

RIES

NOILNLILSNI
LIBRARIES

S3i¥vVudIT LIBRARIES

INSTITUTION
Saluvyagit

5
ro |
ne
> Ss We
a
=
=

NIAN
LIWS

=
= "a NO

STITU
«> "%

s:
LAP wr

y

SMITHSONIAN

lUVUSIT LIBR

SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS

INSTITUTION

Ssaiuvudil

INSTITUTION

INSTITUTION

ARIES SMITHSONIAN

Saiuvugiy_
=
wm
=f
any
Cc
=
oO
Fo

INSTITUTION |
Es
eK
=,
a

wo z
a ahd de
= Wy, =

NOILn

NVINOSHLINS S3SIYVUGIT LIBRARIES

LIBRARIES

NOILALILSNI

LIBRARIES

LILSNI

LIBRARIES

LIBRARIES

NOILNLILSNI

LIBRARIE©

NVINOSHLINS S3I1YVYSIT LIBRARIES

INSTITUTION NOILNLILSNI

INST

x ~ SMITHSONIAN _

2am

NOILNLILSNI

SMITHSONIAN

LIWS SZIYWVYHUSIT LIBRARIES

SMITHSONIAN INSTITUTION LIBRARIES
“TAMA
| 3 9088 01206 6197