ISSN 0078-8554

PALAEONTOLOGIA
AFRICANA

ANNALS OF THE

BERNARD PRICE INSTITUTE

FOR

PALAEONTOLOGICAL RESEARCH

UNIVERSITY OF THE WITWATERSRAND

JOHANNESBURG

VOLUME 37, 2001

ACKNOWLEDGEMENTS
The Bernard Price Institute for Palaeontological Research gratefully acknowledges financial support for its

programmes and for publication of this journal by

THE COUNCIL'S RESEARCH COMMITTEE, UNIVERSITY OF THE WITWATERSRAND.
NATIONAL RESEARCH FOUNDATION (NRF)

PALAEO-ANTHROPOLOGY SCIENTIFIC TRUST (PAST)

for
PALAEONTOLOGICAL RESEARCH

University of the Witwatersrand
Johannesburg

2001

,lt~

P·A·S·T

DTP by the Central Graphics Service of the University of the Witwatersrand.
Printed in the Republic of South Africa by THE NATAL WITNESS (PTY) LTD. , Pietermaritzburg, KwaZulu Natal

BERNARD PRICE INSTITUTE FOR PALAEONTOLOGICAL RESEARCH

2001

STAFF
Academic Staff

Director and Chair of Palaeontology
B.S. Rubidge BSc (Hons), MSc (Stell), PhD (UPE)

Deputy Director
M.K. Bamford BSc (Hons), MSc, PhD (Witwatersrand)

Research Officer
A Renaut BSc (Hons), PhD (Witwatersrand)

Research Fellow
S.E. de Villiers BSc (Hons) (Lond), PhD (Witwatersrand), PRSM

Collections Curator
M.A. Raath BSc (Hons), PhD, UED (Rhodes)

Post Doctoral Fellows
R.J. Damiani BSc (Hons), PhD (La Trobe)
B. Gomez DEA (Lyon, Paris), PhD (Lyon)

Editorial panel
B.S. Rubidge: Editor
M.A Raath: Editor

M.K. Bamford: Associate Editor
L.R. Berger: Associate Editor

Consulting editors
Dr. lA Clack (Museum of Zoology, University of Cambridge, Cambridge, UK)

Dr. H.C. Klinger (South African Museum, Cape Town)
Dr. K. Padian (University of California, Berkeley, California, USA)

Dr. R.M.H. Smith (South African Museum, Cape Town)
Prof. L. Scott (University of O.F.S., Bloemfontein)
Dr. J.F. Thackeray (Transvaal Museum, Pretoria)

Dr K.B. Pigg (Arizona State University, Arizona, USA)

Technical and Supporting Staff
Principal Technician

W. Costopoulos

Senior Administrative Secretary
H. Kempson

Assistant Research Technician
G.M. Modise

C.B. Dube

Technician/Fossil Preparator
B.M. Mbense
M.Monyane

P.R. Mukanela
H. Naiker

AN. Ntheri
IN. Sithole

S. Tshabalala

Custodian, Makapansgat Sites
J. Maluleke

Honorary Staff
Honorary Professor of Palaeoanthropology

P.V. Tobias FRS, FRCP, MBBCh, PhD, DSc (Witwatersrand), Hon. ScD (Cantab, Pennsylvania), Hon. DSc (Natal, U West., Ont.,
Alberta, Cape Town, Guelph, UNISA, Durban-Westville, Pennsylvania, Wits, Mus d' Hist Naturelle-Paris, Barcelona, Turin, Charles U,

Prague, Stellenbosh), For. Assoc. NAS, Hon. FRSSA, Hon. FCMSA

Honorary Research Professorial Fellow
J.W. Kitching PhD (Witwatersrand), Hon DSc (Port Elizabeth, Witwatersrand), FMLS, FRSSA

Honorary Research Associates
c.K. Brain BSc, PhD (UCT), DSc (Witwatersrand), Hon. DSc (UCT, Natal, Pret, Witwatersrand), FZS, FRSSA

F.E. Grine BA (Hons) (Washington & Jefferson College), PhD (Witwatersrand)
J.M. Maguire BSc (Natal), BSc (Hons), PhD (Witwatersrand)

I.R. McLachlan BSc (Hons) (Witwatersrand)

Hon. Warden, Makapansgat Sites
W. Murzl

PALAEOANTHROPOLOGICAL UNIT FOR RESEARCH AND EXPLORATION

STAFF
Academic Staff

Director of Palaeoanthropology Unit for Research and Exploration
L.R. Berger BA (Hons) (GA Southern), PhD (Witwatersrand)

Senior Research Officer
J. Moggi-Cecchi DR (University of Florence)

Honorary Research Associates
S.E. Churchill PhD (New Mexico)

A. Keyser BSc (Pret), MSc (Pret), PhD (Witwatersrand)
R.S . Kidd Dip Pod Med (Salford), BA (Hons) (Open University), PhD (Western Australia)

P. Schmid PhD (ZUrich)
P.S. Ungar BA (SUNY Binghamton), MA, PhD (SUNY Stony Brook)

Tecbnical and Supporting Staff
Senior Technician

P. Chaushev MD (Varna)

Assistant Collection Curator
B. van Rensburg

Administrative Secretary
A. Adsetts

Financial Officer
N. McCormick

Palaeont. afr., 37, (2001)

CONTENTS

1. Interpretive problems in a search for micro-invertebrate fossils from a Neoproterozoic
Limestone in Namibia

Page No.

by C. K. Brain, K.-H. Hoffmann, A. R. Prave, A. E. Fallick, J. Coetzee and A. J. Botha ................ 1-12

2. Carboniferous pycnoxylic woods from the Dwyka Group of Southern Namibia
by Berthold Bangert and Marion Bamford ... .. .. .................................. ........... ... ... ... ... ... ... ... .. .... ....... 13-23

3. A new actinopterygian fish species from the Late Permian Beaufort Group, South Africa
by Patrick Bender .. ..... ....... ....... .................................... .... ... ... ... ... .. .. ... .... .. ... .... ................................ 25-40

4. Cranial anatomy of the giant Middle Triassic ternnospondyl Cherninia megarhina and a
review of feeding in mastodonsaurids
by Ross J. Damiani .. .. .. .... ... .... .... ....... ... ................ .............. ... ... ... ... ... ...... ... ....... ..... ....... ......... .......... 41-52

5. Preliminary phylogenetic analysis and stratigraphic congruence of the dicynodont
anomodonts (Synapsida: Therapsida)
by Kenneth D. Angielczyk ......................... ... ....... ... ....... ... ... .... ...... ... ... ... ... ............ .................. .. ....... 53-79

6. Cranial description and taxonomic re-evaluation of Kannemeyeria argentinensis (Therapsida:
Dicynodontia)
by A. J. Renaut and P J. Hancox ...................................... .. .... .. ....................................................... 81-91

7. A partial skeleton of the Tritheledontid Pachygenelus (Therapsida: Cynodontia)
by C. E. Gow ..... ... ..... .. .................................................................................. .. ................................. 93-97

8. New Viverrinae (Carnivora: Mammalia) from the basal Middle Miocene of Arrisdrift,
Namibia
by Jorge Morales, Martin Pickford, Dolores Soria and Susana Fraile ......................................... 99-102

9. New evidence of the giant hyaena, Pachycrocuta brevirostris (Carnivora: Hyaenidae),
from the Gladysvale Cave Deposit (Plio-Pleistocene, John Nash Nature Reserve, Gauteng,
South Africa) .
by Raoul J. Mutter, Lee R. Berger and Peter Schmid .......... .. .............. .. .......... .. .......................... 103-113

10. Locomotor and habitat classifications of cercopithecoid postcranial material from
Sterkfontein Member 4, Bolt's Farm and Swartkrans Members 1 and 2, South Africa
by Sarah Elton ........... .. ......................................................... ... .. ........................................... .. ...... 115-126

Pa/aeont. afr., 37,1-12 (2001)

INTERPRETIVE PROBLEMS IN A SEARCH FOR MICRO-INVERTEBRATE FOSSILS FROM A
NEOPROTEROZOIC LIMESTONE IN NAMIBIA

C. K. Brain,! K.-H. HotTmann,2 A. R. Prave,3 A. E. Fallick,4 J. Coetzee5 and A. J. Botha5

ITransvaal Museum, P. 0. Box 413, Pretoria 0001, South Aj'rica
2Geological Survey oj' Namibia, P. 0. Box 2168, Windhoek, NamIbia

3School oj' Geography and Geosciences, University oj'St. Andrews, Fife KY 16 9AL, Scotland
4Scottish Universities Environmental Research Centre, East Kilbride, Glasgow G75 OQF, Scotland

5 Electron Microscopy Unit, University oj' Pretoria, 0002, Pretoria, South Aj'rica

ABSTRACT
Attention is focussed on a carbonate sequence in the Auros Formation of the Otavi Group in

northern Namibia, where several limestone layers are found to have been phosphatised. These contain
an abundance of unusual objects, some of which suggest sponge-like microfossils, whereas others
superficially resemble bivalved shells. Alternatively they may be pseudofossils - the deceptive
products-of a phosphatisation process and subsequent diagenetic effects in the limestone. Since this
deposit antedates the ca. 590 million-year-old Ghaub or Marinoan glaciation, the presence of any
potential metazoan fossils is worth investigating. The objects in question are described and alternative
interpretations are discussed.

KEYWORDS: Neoproterozoic, Otavi Group, microfossils, pseudofossils

INTRODUCTION
For the last few years, one of the authors (C.KB.)

has been conducting a search for micro-invertebrate
fossils in Proterozoic limestones of Namibia. Many
samples oflimestone from the Otavi Group in the Otavi
Tsumeb-Grootfontein area were examined in thin
section and as acetic acid-treated residues, but
recrystallisation of the limestones in the folded Otavi
Mountainland appeared to militate against the finding of
well preserved microfossils there. In fact, as Martin
Pickford (1995) has pointed out, virtually no fossils other
than stromatolites have thus far been found in Otavi
Group sediments. But, north of the Otavi Mountainland
lies the flat, calcrete-covered plain of the Etosha Basin,
which apparently represented a Bahamas-type
carbonate platform on the Congo Craton in Late
Proterozoic times. About 50 km beyond the northern
border of the Otavi Mountainland outcrops, and well
away from the metamorphic folded belt, some isolated
hills make their appearance, through the surrounding
calcrete cover, close to the Halali restcamp in the Etosha
National Park. Two km east of Halali are the twin peaks
of Helio Hill and when C.KB. examined acetic acid
residues of samples of a black grainstone from the base
of Helio Hill, he encountered microfossil-like objects
that appeared to have been phosphatised. Such objects,
referred to here as 'otavias', were later found in similar
limestones from the Halali Hill itself and the Halali
Quarry, 2 km further west. They are considered in detail
in this paper, together with shell-like objects
encountered in the Halali Quarry grainstones. Although
the Halali outcrops appear on the 1 : 1 000 000 Geological
Map of Namibia (1980) as 'Otavi Group -

undifferentiated', it was clearly important to establish
just where within the Group's stratigraphy the sequence
is positioned. The outcrops were therefore examined in
detail in the field by K-H.H. and C.KB. and samples
were taken at close intervals for carbon isotope analysis
and interpretation by AR.P. and AE.F. Comparable
studies were made on the Abenab Subgroup sequence
in the Kaokoveld by K-H.H. and A.R.P., while
confirmation of the presence of calcium phosphate in the
microfossil-like objects from the Halali outcrops was
provided by J.C. and AJ.B.

As has been pointed out by several specialists in the
field of Proterozoic palaeontology who commented on
the first draft of this paper, the interpretation of
phosphatised objects in limestones of this age is by no
means straightforward and is fraught with uncertainties.
The purpose of the present paper therefore is to draw
attention to, and to discuss such uncertainties in the the
Halali context, with the hope that greater interpretive
confidence will be possible in the future.

GEOLOGICAL CONTEXT AND CARBON
ISOTOPE VALUES

As pointed out earlier (Hoffmann & Prave 1996,
p. 77), "the Otavi Group is a thick succession of
Neoproterozoic carbonates exposed within the Otavi
Foreland fold belt of northern Namibia. It overlies
predominantly coarse-grained terrigenous siliciclastic
and local volcanic rocks of the Nosib Group and is
overlain by fine- to coarse-grained Mulden Group
siliciclastic sediments." The extent of Otavi Group
outcrops in northern Namibia is shown in Figure 1, while
a lithostratigraphic subdivision and correlation of the

main
figure

...... Angola
. -_. -- . -- . -- . -- . -- .. - . -- . -- . -- .. - . -- .

Namibia

100 km

,--------------,
.--- __ .1 • ,

Etosha Park • Halali :
I

Figure 1. Map of northern Namibia showing the extent of the Otavi Group outcrops (shaded) in
the Otavi Mountainland and Kaokoveld. The position of the Halali exposures, on the
Etosha plain is also shown.

Eastern Kaokoveld Otavi Mountainland Group in the Otavi Mountainland and the eastern
Kaokoveld, as was proposed by Hoffmann & Prave
(1996), is given in Figure 2. Just prior to the publication
of this subdivision, field studies along the Fransfontein
Ridge to the east of the Otavi Mountainland, had
demonstrated the presence of two stratigraphically and
lithologically distinct glacial diamictite intervals, each
succeeded by a unique cap-carbonate (Hoffmann 1994;
Prave & Hoffmann 1995; Hoffmann & Prave 1996).
Prior to this, a single glacial interval only, within the
Otavi Group had been known for a considerable time
(e.g. Le Roex 1941), but the presence of the two
diamictites is now generally accepted.

Gp sGp Formation Gp sGp Formation

c
Upper (!)

"0
"S
~ Lower

c
Kombat (!)

:2
::J

:2: Tschudi

Huttenberg Huttenberg

Elandshoek Elandshoek
.0 '.0
(!) Q)

E Maieberg
::J

E Maieberg
::J en en ..... Keilberg ..... Keilberg

6. Ghaub 6.
"~

6. Ghaub 6. -0
Ombaatjie Auros

">
ctI .0 - ctI Gruis
0 c

(!)

.0
ctI Gauss c
Q)

.0 Rasthof « .0 Berg Aukas «
6. Chuos 6. 6. Varianto 6.

0
.0

no formal E
ctI formation
.0 subdivisions
E exist
0

;Q (transitional)

en
0 Nabis Z

I ~ Nabis

Although evidence for very severe glacial conditions
during Neoproterozoic time has been recognised for 37
years (Harland 1964) and the concept of a 'snowball
Earth' was proposed by Kirschvink (1992) almost a
decade ago, the application of the 'snowball Earth'
scenario to the two glacial episodes reflected by Otavi
Group diamictites is much more recent (Hoffman et al.
1998a, b). It postulates extreme glacial conditions, even
in the equatorial regions, a cessation of continental
runoff and virtual shutdown of biological activity. Such
conditions would have had a dramatic effect on the
evolution of early animals, whose lineages apparently go
back considerably further in time on the basis of
molecular evidence (Doolittle et al. 1996; Wray et al.
1996). It is therefore of interest to note that the Halali
carbonate sequence with which we are concerned here
is found to fall within the Auros Formation of the Abenab
Subgroup, below the upper of the two glacial episodes,
the Ghaub, which is thought to have occurred about 590
million years ago.

Figure 2. Simplified lithostratigraphic subdivision of
Neoproterozoic rocks in northern Namibia (modified
slightly from Hoffmann & Prave 1996). Triangles denote
the two glaciogenic units . Gp = Group; sGp = sub
Group.

The continuity of the Halali carbonate sequence has
been traced in four separate exposures. Where exposed,
close to the Halali restcamp, these rocks are found to dip
at a shallow angle towards the east, with the result that
the oldest part of the profile outcrops in a low ridge at the
Halali Quarry, 2 km west of the Halali camp, and this

'Chert' Hill

Helio Hill
1""/<111 1

1
1

1 /
/

""11f!I 1
1

I{/JI "" / / /
/ If!I / <QJ/

/ / !
/ / /

i"l ./
• 1 "I'

• I' • . /
. 1

~/ . ! -;
. I·
'1

! . f . /.

.l.....-
h;o;.

.L -
f-- -Z!'II!!-

Halali Hill
-= . -.,....

1 , '11
T /

I / /I

I
T •

cherty grey
dolostone

med.·light grey
dolo·grainstones
and thrombolitic
dolostones wi
chert nodules

med. grey
dol<>-grainstones
& locally oolitic
dolostones

Otavia
da rk ribbon-rack
lim estone

Otavia
dark grey & buff
limestone

brown weathered
dolomicrite

laminated
dolomicrite wi
minor chert

nodular·bedded
limestone wi
minor chert

dark rhythmite
limestone

Halali Quarry

f--

I-- I

'JIIIIIII _*

II I 1
I J

-I

dark grey
ribbon-rock
& rhythmrte
limestone

Otavia

dark grey &
buff limestone

• • • • • •
• • • •

·2 0 2 4 6 8

composite stratigraphic section 013C (%0 V·PDB)

Figure 3. Composite stratigraphic section of the Otavi Group,
Halali area, Etosha National Park. Each locality occurs
in isolation but stratigraphic continuity is assured because
the area is structurally simple and marked by shallow
dips, thus the amount of stratigraphic omission between
sections is minimal (no more than several metres between
each section).

sequence is continued in the 40 m-thick layers of the
Halali Hill, situated in the restcamp enclosure. Two
kilometres to the east are the prominent twin peaks of
Tweekoppies, or Helio Hill, where 60 metres of grey
dolostones continue the sequence in this geographically
isolated exposure of Otavi carbonates. Finally, about
one km farther east is a low ridge of exposures referred
to in this paper as "chert hill", which represents the
stratigraphically youngest rocks reported upon here.

The lower parts of the Halali profile contain highly
distinctive ribbon rocks, or nodular limestones,
consisting of interfingered grey and buff layers. These
outcrop again 45 km to the south, on the northern fringe
of the Otavi Mountainland, just within the boundary
fence of the Etosha National Park, opposite the farm
Olifantslaagte. Here the ribbon rocks can be seen in
stratigraphic context some distance below the Maieberg
carbonate, that lies above the Ghaub diamictite
elsewhere in the Otavi succession.

Lithological details of the composite Halali profile are
shown in Figure 3, together with results of the carbon
isotope analyses, the implications of which will now be
considered .

The carbonate rocks expos~d in the Halali area were
measured and sampled by K.-H.H. and C.K.B .
Subsequently, A.R.P. , accompanied by C.K.B., field
checked the sections and sample localities. A total of77
samples was analysed for C-isotope signatures .
Samples were processed following standard techniques
and screening process . Samples were rnicrodrilled (in
order to minimise lithological differences, only micritic
constituents were drilled) and the powders sent to the
Scottish Universities Environmental Research Centre
for stable isotopic analysis under the supervision of A. E.
F. Approximately 1 mg of powder is loaded into a 4 ml
glass tube that is sealed and heated at 70° C for 30
minutes. The sample is then placed in the Carbonate
Acid Injector and a 3-way needle inserted which pumps
He into the sample to purge all atmospheric gases. The
He flow and vent are closed, phosphoric acid added and
the sample is then allowed to react (at least 8 hours for
calcite and at least 24 hours for dolomite). Once the
reaction time is completed, the samples are transferred
to the Gas Prep Interface Analyser linked to an AP 2003
triple-collector mass spectrometer in which the vented
CO2 is analysed. Calibrated standards are included at
every run to check machine precision and
reproducibility.

The resulting l3C trend is shown in Figure 3 and,
combined with the lithofacies character of the rocks,
provides convincing evidence for correlating the isolated
Halali exposures to the Auros-Ombaatjie Formations
(see Figure. 2) of the Otavi Group . As discussed above,
the dark grey limestones that are exposed at Halali can
be observed to occur subjacent to the Keilberg cap
carbonate near the Etosha border fence south of Halali.
In addition, the dark grey, foetid ribbon-rocks and
rhythmite limestones of the Halali successions are
lithologically identical to the pre-Keilberg rocks of the
Ombaatjie Formation in the eastern Kaokoveld. In the

Otavi Mountainland, the rocks immediately beneath the
Keilberg cap carbonate are laminated dolomicrites and
dolograinstones, facies in every respect similar to those
of the Ombaatjie Formation and Halali rocks except that
they are dolomitised and recrystalised. Thus,
lithostratigraphic ally there is good evidence to conclude
that the Ombaatjie, Auros and Halali rocks are
correlati ves.

This inference is corroborated by the l3C data. The
limestones exposed in the Halali Quarry and Halali Hill
localities display a sharp rise from initially slightly
negative values to mostly +2 to +4 values; several
singular excursions to around 0 punctuate this trend. In
the overlying Helio Hill and Chert Hill sections the l3C
values rise to +8 and remain uniform through the entirety
of the exposed succession. Note that this rise is not
attributable to dolomitisation because it initially occurs in
the dark grey ribbon-rock limestones in the lower 15 m
ofthe Helio Hill section (these limestones are identical
to the underlying ones in the Halali Hill and Quarry
sections). This overall trend (sharp rise from slightly
negative to moderately positive and then uniformly +8)
mirrors that known from the Ombaatjie Formation in the
eastern Kaokoveld (e.g. Hoffman et al. 1988; Prave,
Hoffmann & Fallick unpub. data) and the Auros
Formation in the Otavi Mountainland (Prave, Hoffmann
& Fallick unpub . data). Thus, the combined
lithostratigraphic and chemostratigraphic data for the
Ha1ali rocks indicates that these, and any microfossils
that they might contain, are pre-Ghaub glaciation in age.

EVIDENCE OF PHOSPHATISATION
As mentioned above, acetic acid dissolution of dark

grey grainstone from the base of the Helio Hill revealed
objects that superficially appeared to have been
phosphatised. Examples of these objects, in the form of
sponge-like 'Otavias' and shell-like 'Ha1alias', were
mounted on SEM stubs and subjected to an EDS process
(Energy-dispersive Spectometry), linked to a JEOL
5800L V Scanning Electon Microscope at the University
of Pretoria's Electron Microscopy Unit, by two of the
authors, J.e. and AJ.B. As a control, the same
procedure was carried out on an Early Cambrian 'small
shelly' tubular fossil, dissolved from a sample of Ajax
Limestone from the Aroona area of South Australia.
The phosphatisation of such fossils has been
documented by Bengtson et al. (1990). The elemental
composition of the three samples mentioned here is
shown in Figure. 4 a - c, from which it can be seen that
the peaks for Calcium and Phosphorus are virtually
identical for the three samples. These results are taken
to indicate that the the two kinds of Halali objects, be
they micro- or pseudo-fossils, have been preserved as
calcium phosphate. At the same time it should be
emphasised that the heights of peaks for any particular
element are not a reliable indicator of its actual
abundance in the sample being investigated, though the
presence of the particular element is documented
beyond doubt.

It should be mentioned that this is not the first time that
calcium phosphate has been demonstrated in Otavi
Group carbonates. In his study of carbonates hosting the
Tsumeb ore-body, Hughes (1987) remarked on the
presence of calcium phosphate being present both as
cryptocrystalline collophane and as crystalline apatite.

A
700

STUB187 OTAVIA NR16 (EM778)
650 ·

C,
600

550

500

450

400

350

300

250

200

150

100

JL \A 50 ~ r,

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Energy (keV)

B
1100

NR 10

C,
1000

900

800

700

600

500

400

300

200

100
'" r, r,

7 7.'

C STUB191 CONTROL SOUTH AUSTRALIA (EM722)
1200

1100

1000

900

800

700

600

500

400

300

200

r. r,

0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5
Energy (keV)

Figure 4. Energy-dispersive spectometry diagrams documenting
the occurrence of elements in samples of : A. The outer
envelope of an 'Otavia' specimen from the Halali Quarry
section; B. A 'Halalia' shell from the same horizon and
C. A control sample, consisting of a 'small shelly'
phosphatised tube fossil from the Early Cambrian Ajax
Limestone, Aroona, South Australia. It will be seen that
Calcium and Phosphorus are represented in a comparable
manner in each of the samples by this technique.

PHOSPHATIC "OTA VIA" OBJECTS WITH A
SPONGE-LIKE APPEARANCE

For convenience, the objects discussed here are
referred to informally as 'otavias', although no formal
taxonomic designation is provided at this stage. In the
Halali sequence, otavias have been found in three layers
of black/grey foetid grainstones - in the Halali Quarry
section, at the top of the Halali Hill and within the
lowermost few metres of the Helio Hill sequence. They
are extremely common in the Quarry section, less so on
the Halali Hill and rare at Helio Hill, where the enclosing
matrix has been partially dolomitised. In size, otavias
vary from 300 micrometers to 3.5 mm in maximum
dimension, with the largest specimens occurring in the
Quarry section. From here a sample of200 otavias were
individually measured, as well as 90 from the Halali Hill.
The greatest dimensions of these specimens were
recorded after the individual objects had been
manipulated under water over a micrometer scale with
the aid of a stereo microscope. Most otavias from the

Quarry section are 500-600 micrometers in their
greatest dimension, while those from Halali Hill are
typically 300-400 micrometers in length.

As shown in Figure 5, there is a good deal of variation
in shape of otavia specimens, though they all share some
features in common. When viewed as isolated objects
sorted from acetic acid residues under a stereo
microscope and then imaged with a scanning electron
microscope (in this case, a Hitachi 510 at the Transvaal
Museum), each otavia appears as a hollow bag, with one
to five comparatively large openings, typically on raised,
volcano-like mounds . These penetrate the phosphatic
outer wall and lead directly into the internal cavity. The
wall is also pierced by numerous smaller openings that
generally lead into a "peripheral labyrinth' (for want of
a better term), shown in Figure 6. This system of
irregularly interlinked spaces is present beneath some
parts of the outer wall and is variable in its extent; it also
has connections with the internal cavity.

Figure 5. SEM images of 'otavia' objects recovered from acetic acid residues ofOtavi limestone; A and B are from the base of Helio Hill,
the others come from the Halali Quarry section. All scale bars represent 200 micrometres .

In thin sections, as shown in Figure 7, the outer
envelope and peripheral labyrinth typically appear dark,
or almost opaque; the walls of these structures are
composed of calcium phosphate in an amorphous or
cryptocrystalline form. Thin sections show the interior
of the otavia structure to be filled with crystalline
calcium carbonate, very similar to that of the
surrounding matrix, which can perhaps best be
described as a sparite.

In view of the superficial similarity in appearance of
these objects to small sponges, special attention has
been given to the outer walls for the possible presence

of spicules, characteristic of many later sponges.
Elongate crystals do occur occasionally in the walls, but
they are not convincing as spicules and are more likely
to be inorganic with a diagenetic origin.

In an earlier draft of this paper, one of the authors
(C.K.B.) attempted an interpretation of the otavia
structures, suggesting that they were phosphatised
fossils of small calcareous sponges (Alternative A
below). The draft was sent for comment to various
specialists on Proterozoic palaeontology and valuable
comments were received from Stefan Bengtson,
Andrew Knoll and Bruce Runnegar, necessitating a

Figure 6. SEM images of 'otavia' obJects from the Halali sequence. A & B show detail of openings through the outer envelope on raised
eminences; scale bars each represent 50 micrometres. C & D show detail of the 'peripheral labyrinth " where the outer envelope
has been removed; the scale bar for C represents 200 micrometres; that for D is 20 micrometres.

revision of some of the proposals made in the draft.
These are taken into account in Alternative B, to be
presented shortly, while a third possibility is expressed in
Alternative C. Studies will continue on the otavias from
the Halali sequence and a more firmly based
interpretation may be possible in the future.

Alternative A
This interpretation assumes that an otavia existed

originally as a multicellular sponge-like organism in the
form of a hollow bag, the organic walls of which were
penetrated by one or several larger openings, typically
25 - 200 micrometers in diameter and usually situated on
raised eminences. These connected directly with the

Figure 7. Thin sections of three 'otavia' objects. The black areas
represent the extent of the 'peripheral labyrinths',
which may fill depressions among composite
microphytolite grains, according to one interpretive
alternative. See text for details . Scale bars represent the
following . A: 1 mm; B: 200 micrometres; C: 500
micrometres.

internal cavity and would have been the equivalent of the
exhalent oscula of later sponges. Numerous smaller
holes in the outer wall, representing incurrent pores or
ostia, typically led into an irregular 'peripheral labyrinth'
of interlinked chambers of variable extent, within some
parts of the outer wall. These chambers in turn had
connections with the internal cavity and could be
interpreted as the paragastric cavities of sponges, in
which flagellated collar cells were originally housed.

In his comment on this interpretation, Bruce
Runnegar has reservations about the 'peripheral
labyrinth' being of organic origin. He writes (pers.
comm.): "The meshwork of phosphate is sponge-like but
it is also the sort of structure one might expect to find in
phosphate nodules formed inorganically". This is an
observation that clearly needs to be taken seriously.

Another reservation, expressed by Stefan Bengtson
(pers. comm.), is that the overall size of the otavias is too
small for these to have functioned properly as sponges.
In particular, he points out that there are critical lower
limits for the size of incurrent pores or ostia, if these are
to function effectively as the entry points of water
streams carrying potential food items. He observes
(pers. comm): "The force necessary to pump water
through a tube is inversely proportional to the 4th(!)
Power of its diameter (the Hagen-Poiseuille equation).
In other words, resistance increases dramatically with
diminishing size, and a 2 micrometer tube needs 54=625
times as much pressure as a 10 micrometer one to
maintain the same flow, and 58=-400,00 times as much
as a 50 micrometer one. Sponges are basically
suspension feeders, and ostia this small (ie. about 10
microns across) would exclude most suspended cellular
material from entering the system. Inhalent canals are
typically lined with choanocytes, which are about 5-10
microns apart in diameter. Ostia down to 2 microns in
diameter would be out of proportion with even the
smallest inhalent canals."

Another reservation, expressed both by Bengtson and
Knoll (pers. comm.), concerns the conclusion inherent in
this interpretation, that the interior of the otavia "bag"
was, in fact, empty prior to fossilisation. Bengtson
suggests that the phosphatic 'skin' might have
surrounded a solid structure. He writes (pers. comm.):
"Looking at your thin sections and SEMs, I rather have
the impression of irregular galleries of calcium
carbonate that have been enveloped in calcium
phosphate, and that the protruding necks with the
"oscula" are places where the enveloped calcium
carbonate has been sticking out".

This comment leads naturally to the consideration of
'microphytolites', central to Alternative B, a matter
raised by Knoll and clearly critical to the understanding
of Proterozoic objects of this kind.

Alternative B
In their paper on Late Proterozoic sediments from

Spitsbergen, Swett & Knoll (1985 p. 341) provide the
following information about microphytolites: "In the
Soviet palaeontological liter:ature the term

"microphytolite"is used to embrace a wide variety of
carbonate grain morphologies ranging from fine sand to
gravel in size and exhibiting highly variable shapes and
internal structures. Morphologies that fall within this
group include ooids, small oncolites, grapestones, algal
lumps, and other, more problematic intraclasts; some
have been assigned Linnean bionomials and used in
biostratigraphy (e.g., Raaben and Zabrodin 1972)".
They point out that microphytolites are known in various
upper Proterozoic carbonates from many parts of the
world and, although they do occur occasionally in
modern carbonate environments, they are not common
there. The reason for this is that, in a microphytolite, a
variety of irregular grains is held together by a bacterial
and/or algal film and, in the contemporary context, such
a film tends to be consumed by grazing metazoans,
causing the individual grains to separate. The overall
shape of a composite microphytolite aggregate can be
very irregular with projections and deep embayments in
connection with which Swett & Knoll (1985) remarked:
"As viewed in thin section, external shapes may be
extremely elongate (L:W > 20:1) or compact (L:W -
1:1)".

It is of interest to note that microphytolite grains
provide some information of the turbulence of the water
in which they formed. Where water movement is
sufficient to regularly roll carbonate grains around on the
sea floor, oolitic or pisolitic grainstones can develop. In
conditions of intermediate turbulence, microphytolites
might well form, where the composite grains are
occasionally moved, but not with sufficient regularity to
acquire oolitic coatings. Finally, in tranquil aqueous
environments, continuous microbial mats can develop to
cover the sea floor, incorporating any microphytolites
that may happen to be there. Adolf Seilacher (eg. 1999)
has stressed the importance of biomats to the existence
of the unusual 'Ediacaran' fauna that dominated shallow
marine environments during terminal Neoproterozoic
times . Such biomats disappeared with the advent of
abundant surface-feeding and burrowing metazoans in
Cambrian times.

Alternative B postulates that an otavia is essentially
no more than a microphytolite that has acquired a
phosphatic coating. Deep embayments between
individual grains were originally filled with adhering
bacteria, cyanobacteria or other algae and their
substance was phosphatised at the time that the
phosphatic envelope developed. The entire otavia
structure is therefore biological only in the sense that
calcium phosphate replaced the slime which served as
the 'glue' for the component microphytolitic grains .
Large openings in the phosphatic envelope were places
where parts of grains protruded; small openings
represented nothing more than defects in the envelope's
continuity.

Alternative C
This interpretation takes into account the very

considerable role that microphytolites played in
Neoproterozoic grainstones and acknowledges that
these were very likely the nuclei upon which otavia

structures were built. However, in Alternative C it is
speculated that the deep embayments, present in some
of the microphytolite complex structures, could well
have been the ideal microhabitats that promoted the
evolution of the first encrusting proto-sponges. These
rather loosely coordinated 'multicellular' organisms
must have started as social choanoflagellates, such as
still may be found in contemporary habitats, and it is
suggested that these could have thrived in the deep
recesses among microphytolite grains. Selective
pressures may have favoured those colonies that
increased their 'multicellular' tendency, promoting a
coordinated flow of water, with its food particles, past
the food-gathering collar cells. This would have involved
the development of an outer envelope with incurrent and
excurrent openings and a series of interlinked chambers
in which flagellated cells were housed.

Close examination of otavia structure suggests that
some of the large openings, on their raised mounds, are
not simply places where microphytolite grains protruded,
or where the phosphatic envelope was defective, but
were biological structures. This possibility will be
considered further in the future as the research project
proceeds.

OOIDS WITH PHOSPHATIC RIMS AND THE
GENESIS OF 'HALALIA' SHELLS

In the otavia-rich grainstone of the Halali Quarry
section, complete ooids occasionally occur, typically 3 -
4 mm in diameter. These have laminated outer shells of
calcium phosphate, a situation that has been described
elsewhere previously, as, for instance, in Lower
Cambrian sediments of Spitsbergen (Swett & Crowder
1982). As complete spheroids, however, they are rare in
the Halali grainstones, when compared to the relatively
common pieces of broken ooids such as those shown in
Figure 8A, B. These pieces consist of an outer, often
laminated shell of calcium phosphate, around the
remains of the micritic 'ball' that the shell enclosed.
Very frequently the pieces of outer shell have separated
completely from their original cores and these
phosphatic shell pieces are seen in abundance in some
layers of the grainstone. In thin section, the micritic
interiors of the ooids typically appear dark or opaque.
What caused the ooids to break up is debatable, but there
are indications that the pellet-like micritic cores of some
ooids have been partially removed by diagenetic
dissolution causing pieces of the outer shells to collapse
inwards upon themselves. Results of this process can be
seen in thin sections of oolites (Figure 9A) from the
Ombaatjie Formation in the eastern Kaokoveld, of
approximately equivalent age to the Halali carbonates.
As shown in Figure 9B, pieces of the outer shells remain
in various degrees of apposition to one another, even
though the fine-grained filling has been partially
removed. In the thin sections and acetic acid residues
from the Halali Quarry grainstone it is not unusual to see
shell pieces in bivalve-like apposition, as shown in Figure
8C, D and these are referred to (by C.K.B.) as halalias.
Initially it was thought that some of these could have

Figure 8.

A & B: Pieces of broken ooid from the Halali Quarry section. Notice the laminated external shell of calcium phosphate and
the dense micritic interior. Scale bars each represent 500 micro metres. C & D : collapsed ooids from the Halali Quarry section,
showing a deceptive bivalve-like condition of 'halalia' shells. C is a SEM image showing the apposition oftwo shell pieces
with remnants oftheir adhering interiors, while D shows two apposed shell pieces in thin section. Scale bars each represent
500 micrometres.

Figure 9 A& B: Thin sections of oolitic limestone from the Ombaatjie Formation in the eastern Kaokoveld, showing complete ooids
with resistant rims in A, while B shows the effects of diagenetic dissolution of the interior of ooids and resulting collapse
of the exterior shells. Each scale bar represents Imm.

been true bivalved shells of biological origin, but in view
of subsequent evidence, this interpretation seems
unlikely. They are more probably highly deceptive
pseudofossils.

CONCLUDING COMMENTS
As mentioned above, there is a strong possibility,

based on molecular evidence, that metazoan lineages go
back in time beyond the Ghaub glaciation. Soft-bodied
micro-invertebrates were therefore presumably present
during deposition of the Auros/Ombaatjie Formations
and the best chance for their preservation is probably in
a phosphatised limestone such as occurs at Halali. The
superb preservation potential for embryos and algae has
been amply demonstrated by Proterozoic and Early
Cambrian phosphorites in China (Bengtson & Zhao
1997; Xiao et al. 1998; Zhang et a!. 1998), so any
phosphatic limestone from this time period is well worth
investigating.

The possibility, remote as it may seem, of proto
sponge fossils being present in a pre-Ghaub limestone
would imply that at least one of the sponge lineages
survived the Ghaub or Marinoan 'snowball Earth'
episode, to continue on a wider scale once more
favourable conditions returned. Remains interpreted as
being those of sponges have been well documented in
the 'Ediacaran' period, following the glacial (e.g.
Brasier eta!' 1997; Collins 1999; Gehling & Rigby 1996;

Li eta!' 1998; Dong & Knoll 1996). There is even some
evidence of Ediacaran-type organisms preserved in
intertillite beds of Canada (Hofmann eta!' 1990). It has
been suggested that survival during the snowball Earth
glacial would have been possible in locally warmed
microhabitats such as close to hydrothermal vents
(Erwin 1999, Farmer 2000) . But new computer
simulations of the snowball Earth scenario suggest that
there may, in fact, have been ice-free equatorial oceans,
the presence of which would have been critical for the
survival oflong-standing metazoan lineages (Hyde eta!'
2000; Runnegar 2000). Studies of the palaeontological
potential of the Halali phosphatic limestones will
continue in the future, in the hope that they could throw
some light on the antiquity of metazoan lineages in the
Namibian region.

ACKNOWLEDGEMENTS
Fieldwork on this project by C.K.B . was supported by a grant

from the PAST fund in Johannesburg; this help is gratefully
acknowledged. Permission to undertake fieldwork in the Etosha
National Park was provided by the Park authorities. Comments
received on the first draft of this paper from Stefan Bengtson, Andy
Knoll and Bruce Runnegar have proved of great value. The assistance
in the field of Conrad Brain and Arno Glinzel is much appreciated .
Laura Brain's contribution in making numerous thin sections was
invaluable. A.R.P. was supported by NERC grants and he also
gratefully acknowledges the Geological Survey of Namibia for
additional support. We would also like to thank Bruce Rubidge for
his encouragement and editorial help.

REFERENCES

BENGTSON, S. & ZHAO, Y. 1997. Fossilized metazoan embryos from the earliest Cambrian. Science 277, 1645-1648.
BENGTSON, S., CONWAY MORRIS, S., COOPER, B. J., JELL, P. A. & RUNNEGAR, B. N. 1990. Early Cambrian fossils from South

Australia. Association of Australasian Palaeontologists, Memoir9, 1-364.
BRASIER, M. , GREEN, O. & SHIELDS, G. 1997. Ediacaran sponge spiCUle clusters from southwestern Mongolia and the origins of the

Cambrian fauna. Geology24(4), 303-306.
COLLINS, A. G. 1999. The early fossil record of sponge-grade metazoans . Abstract 52302, Geological Society of America AnnualMeeting,

Denver, Colorado.
DONG, X. & KNOLL, A. H. 1996. Middle and Late Cambrian sponge spicules from Hunan, China. JournalofPalaeontology70(2), 173-184.
DOOLITTLE, R. F.,FENG, D. F., TSANG, S., CHO, G. & LITTLE, E. 1996. Determining divergence times ofthe major Kingdoms ofliving

organisms with a protein clock. Science 271, 470-477.
ERWIN, D. H. 1999. Biotic survi val and recovery during the Neoproterozoic 'Snowball Earth' . Abstract 50039, Geological Society of America

Annual Meeting, Denver, Colorado.
FARMER, J. D. 2000. Hydrothermal systems: doorways to early biosphere evolution. GSA Today 10(7), 1-9.
GEHLING, J. G. & RIGBY, 1. K 1996. Long awaited sponges from the Neoproterozoic Ediacaran fauna of South Australia. Journal of

Palaeontology70(2), 185-195.
HARLAND, B.W. 1965. Evidence for late Precambrian glaciation and its significance. In: Nairn, A.E.M. E(ed.), Problems inPalaeocltinatology,

119-149. London, Interscience.
HOFMANN, H. J., NARBONNE, G. M. & AITKEN, J. D. 1990. Ediacaran remains from intertillite beds in northwestern Canada. Geology

18, 1199-1202.
HOFFMAN, P. F ., KAUFMAN, A. J . & HALVERSON, G. P. 1998a. Comings and goings of global glaciation on a Neoproterozoic tropical

platform in Namibia. GSA Today 8(5) , 1-9.
HOFFMAN, P. F. , KAUFMAN, A. J. , HALVERSON, G. P. & SCHRAG, D. P. 1998b. A Proterozoic snowball Earth. Science 281,

1342-1346.
HOFFMANN, K-H. 1989. New aspects oflithostratigraphic subdivision and correlation of late Proterozoic to early Cambrain rocks of the

southern Damara Belt and their correlation with the central and northern Damara Belt and Gariep Belt. Communications of the Geological
Survey of Namibia 5,59-67.

HOFFMANN, K-H, 1994. New constraints on the timing of continental breakup and collision in the Damara belt. Abstracts volume,
Proterozoic Crustal and Metallogenic Evolution. Geological Society & Geological Survey of Namibia, Windhoek, p. 30.

HOFFMANN, K-H. & PRAVE, A. R. 1996. A preliminary note on the revised subdivision and regional correlation of the Otavi Group
based on glacigenic diamictite and associated cap dolostones. Communications of the Geological Survey of Namibia 11, 77-82.

HYDE, W. T., CROWLEY, T. J ., BAUM, S. K & PELTIER, W. R. 2000. Neoproterozoic 'snowball Earth' simulations with a coupled
climate/ice-sheet model. Nature 405, 425-429.

HUGHES, M . J. 1987. The Tsumeb ore body, Namibia, and related dolostone-hosted base metal ore deposits of central Namibia. PhD thesis ,
Faculty of Science, University of the Witwatersrand, Johannesburg.

PALAEONTOLOGIA AFRICANA VOG7 2001 _ B

KENNEDY, M. J ., RUNNEGAR, B., PRA YE, A. R., HOFFMANN, K-H & ARTHUR, M. A. 1998. Two or four Neoproterozoic
glaciations? Geology26(12), 1059-1063 .

KIRSCHYINK, lL. 1992. Late Proterozoic low-latitude global glaciation: the Snowball Earth. In: Schopf, lW. and Klein, C. (eds), The
Proterozoic Biosphere. A multi-disciplinary study, 51-52. Cambridge, Cambridge University Press .

LI, c.-W., CHEN, J.-Y. & HUA, T.-E. 1998. Precambrian sponges with cellular structures. Science 279, 879-882.
LE ROEX, H. D. 1941. A tillite in the Otavi mountains, S. W. A .. Transactions of the Geological Society of South Africa 44,207-218.
PICKFORD, M. H. L. 1995. Review of the Riphean , Vendian and early Cambrian palaeontology of the Otavi and Nama Groups, Namibia.

Communications of the Geological Survey of Namibia 10,57 -81.
PRA YE, A. R. 1996. Tale of three cratons: tectonostratigraphic anatomy ofthe Damara orogen in northwestern Namibia and the assembly

of Gondwana. Geology 24(2), 1115-1118.
PRA YE, A. R. & HOFFMANN, K.-H. 1995. U nequi vocal evidence for two Neoproterozoic glaciations in the Damara succession of Namibia.

Geological Society of America, Abstracts with Program 27, 380.
RAABEN, M. Ye . & ZABRODIN, Y. Ye 1972. Upper Riphean problematic algae: stromatolites, oncolites. Academiya Nauk SSSR,

Geologicheskij Institute, Trudy, Moscow.
RUNNEGAR, B. 2000. Loophole for snowball Earth . Nature 405,403-404.
SEILACHER, A. 1999. Biomat-related lifestyles in the Precambrian. PAL410S14, 86-93.
SWETT, K & CROWDER, K 1982. Primary phosphatic oolites from the Lower Cambrian of Spitsbergen. JournalofSedimentaryPetrology

52(2), 587-593.
SWETT, K. & KNOLL, A. H. 1985. Stromatolitic bioherms and microphytolites from the Late Proterozoic Draken Conglomerate Formation,

Spitsbergen. Precambrian,Research 28, 327 -34 7.
WRA Y, G. A. , LEYENTON, J. S. & SHAPIRO, L. H. 1996. Molecular evidence for deep Precambrian divergences among metazoan phyla.

Science 274, 568-573 .
XIAO, S. , ZHANG, Y. & KNOLL, A. H. 1998. Three dimensional preservation of algae and animal embryos in a Neoproterozoic phosphorite.

Nature 391, 553-558.
ZHANG, Y., YIN, L. , XIAO, S. & KNOLL, A. H. 1998. Permineralised fossils from the terminal Proterozoic Doushantuo Formation, South

China. Journal of Palaeontology 72, supplement to No.4, Memoir 50, 1-52.

Pal aeon!. a/r., 37,13-23 (2001)

CARBONIFEROUS PYCNOXYLIC WOODS FROM THE DWYKA GROUP OF
SOUTHERN NAMIBIA

Berthold Bangertl and Marion Bamford2

/Institut fur Geologie, Universitiit Wur.zburg, Pleicherwall I, 97070 Wur.zburg, Germany.
2 Bernard Price Institute for Palaeontological Research, University of the Witwatersrand,

Private Bag J, WITS 2050, Johannesburg, South Africa.

ABSTRACT
Glacial deposits of the Dwyka Group between Keetmanshoop and Mariental in southern Namibia

have been reinvestigated for palaeontological remains and associated tuff horizons in an attempt to
accurately date the deposits. SHRIMP-based dating of juvenile zircons from these tuff horizons
provide ages which cumulate in the latest Carboniferous (Gzelian) . The pycnoxylic woods
Megaporoxylon scherziKrausel and Megaporoxylon kaokense Krausel are 'described in detail for the
first time and are compared with similar permineralised woods from Gondwana. Based on previous
fossil wood studies covering the rocks of the main Karoo Basin, these species occur onl y in the Dwyka
and lower Ecca Groups in southern Africa and do not extend to the upper Ecca Group.

KEYWORDS: Dwyka, fossil wood, Megaporoxy/on, Namibia

INTRODUCTION
Geological setting and general stratigraphy of

the Dwyka Group
During the Carboniferous most of the southern part of

Gondwana was glaciated. In the Late Carboniferous
and Early Permian glaciers receded, depositing the
glaciogenic sediments of the Dwyka Group (Karoo
Supergroup). In southern Namibia sections of the
Carboniferous, glaciogenic Dwyka Group ofthe Aranos
Basin have been subdivided into four deglaciation
sequences (DS I-IV, Bangert et al., in press), which
were initially recognised by Theron & Blignault (1975)
and Visser (1997) in the Western Cape Province, South
Africa. In southern Namibia DS I and II began with
tillites and debris-rainout diamictites which are overlain
by minor siltstones, sandstones and conglomerates
representing sediment-gravity flow deposits and fluvio
glacial outwash. The tops of the sequences are formed
by lacustrine or offshore marine shale units which are
essentially drops tone-free. Of particular importance is
an offshore marine mudstone-unit, up to 45 m thick,
named the Ganigobis Shale Member (Martin &
Wilczewski 1970), which represents the top ofDS II and
which is especially well exposed in the area between
Tses and Ganigobis (Figure 1). This shale member
contains well preserved pieces and trunks of
permineralised wood some of which are described in this
contribution, ashfall tuffs (Bangert et ai. 1999,
Stollhofen et al. 2000), concretionary nodules bearing
remains of paleoniscoid fishes (Gtirich 1923, Gardiner
1962, Bangert et al. in press) and spiral coprolites
(McLachlan & Anderson, 1973, Bangert et al. in press),
bivalves (Bangert et al. in press), gastropods (Dickins
1961), Conularia sp. (Schroeder 1908), sponge spicules
and crinoid columns (Bangert et al. in press), as well as
microbial bioherms (Grill 1997).

Juvenile, magmatic zircon grains were separated
from ashfall tuff horizons of the Ganigobis Shale
Member and provided 206Pbf238U ages of 302.0±3.0 Ma
and 299.5±3 Ma (Bangert et al. 1999). Within the
framework of available numerical time-scales the
Ganigobis Shale Member relates to the Late
Carboniferous and more specifically to the Kasimovian
(cf. Harland et al. 1990) or Gzelian (Menning 1995).

Deglaciation sequence III comprises mainly
proglacial tunnel mouth and debris rain-out diamictites
and gravity flow sandstones which terminate with the 75
m thick Hardap Shale Member (SACS 1980). The latter
is characterised by occurrences of the marine bivalve
Eurydesma mytiloides (Heath 1972), Bryozoa and
Asteroidea (Martin & Wilczewski 1970), and has been
related to the Gondwana-wide Eurydesma
transgression (Dickins 1984). The fourth deglaciation
sequence, DS IV, begins with gravity flow sandstones
and dropstone-bearing mudstones which are capped by
greenish, dropstone-free mudstones immediately below
the base of the Ecca Group.

Lithology and fossil content of the fossil wood
sites

The fragment of permineralised wood (BP/16/935)
described here was found at the Brukkaros River about
400 km S of Windhoek and about 23 km NNW ofTses
(No.1 in Figure 1). The outcrop is located downstream
along the Brukkaros River about 2 km SW of the bridge
of the B 1 road which crosses the Brukkaros River
(Figure 1). Stratigraphically the outcrop lies within the
upper part of the Ganigobis Shale Member (DS II,
Dwyka Group). The up to 5 mhigh outcrops oftheriver
cuttings have exposed black massive mudstones with
interbedded sandy and calcareous horizons which are
intruded by five carbonatite dykes. Abruptly lenticular

o
li)CI)
N

40'

o
I.{)
o
Ii)
N

55'

NAMIBIA

WINDHOEK

•
Mariental

•
area of-+O
insert map •

Keetmanshoop
>-----<

200 km

•

main
map -~'"

Ganigobis -t---\-{

Owyka Group
(OS I-IV)

Aretitis

• Soekmekaar

G) Fossil wood sites
Outcrop of the Ganigobis Shale
Member (OS II, Dwyka Group)

• Settlement N
Major road A
Major railway ~
Scale in km

o 1 2 3 4 5

•
GANIGOBIS

Figure 1. Detailed location map of the type area of the Ganigobis Shale Member in the vicinity of Ganigobis and Tses, southern Namibia',
showing the location of the described outcrops.

calcareous sandstone bodies up to 3 m long and 60cm
thick are commonly exposed along the ri ver cuttings. A
bentonitic tuff horizon is exposed at the top of the ri ver
cuttings of the Brukkaros River (VIlla, cf. Bangert et at.
in press). The outcrop extends from the river cuttings
(25°41.24' S, 18°01.54' E) to a few tens of metres to the
east (25°41.26' S, 18° 01.88' E) where flat mounds of
nodular siltstones which have diameters of 4-5 m,
heights of up to 1.5 m, are exposed. The mounds are
covered by round to elongate nodules of brownish-grey
massive micrite, which probably represent microbial
bioherms (cf. Bangert et al. in press).

The following fossils and ichnofossils were detected
within the outcrop (cf. Bangert et at. in press):
- a single example of Conularia sp. (cf. Schroeder

1908),
- numerous shells of Peruvispira vipersdorfensis

Dickins (1961),
- disrupted boundstone samples with column ossicles of

crinoids and small sponge spicules,
- spiral and anvil-like permineralised coprolites (cf.

McLachlan & Anderson 1973)
- trace fossils such as Planolites montanus, Planolites

sp. and other indeterminable traces.

Permineralised fossil wood was found in situ on the
tops of the mounds. The wood was not preserved as
complete trunks but as isolated fragments of less than 10
cm in diameter. One especially well preserved sample of
completely silicified wood displaying 1-7 mm thick
annual growth rings is described in this contribution (BPI
16/935, Figures 2-5).

The second described fragment of permineralised
wood (BPI161746) was found at the Asab River about 22
km NW ofTses and 5.5 krn SW (25°43.04' S, 1 r59.08'
E) of the previously described outcrop (No.2, Figure 1).
Stratigraphically, the outcrop lies within the lower part of
the Ganigobis Shale Member (DS II, Dwyka Group)
which contains abundant concretionary mudstone
nodules bearing remains of paleoniscoid fishes and spiral
coprolites (cf. Bangert et at. in press). The outcrop is
also a ri ver cutting which extends a few hundred metres
along the western side of the Asab River. The outcrop
is dominated by monotonous mudstones sporadically
including irregular mudstone concretions (some
horizontally aligned) lacking fossils. Four bentonitic tuff
horizons occur in the lower part of the outcrop. The
described sample of wood was found in situ within the
lower part of the outcrop.

A third piece of permineralised wood (BPI16/547)
was discovered at the Wasser River, near the farm
Tsaraxa, about 27.5 krn SW of Tses (26°02.55' S,
1r57.20' E, No. 3 in Figure 1). This site is the
southernmost outcrop of the lower part of the Ganigobis
Shale Member with its characteristic drops tone-free
facies changing southwards into shales containing
varying amounts of dropstones (cf. Bangert et at. in
press). Outcrops of the black, massive and drops tone
free mudstones occur in the river cutting of the Wasser

River. In this outcrop area three bentonitic tuff horizons
were detected, which can be correlated to those located
to the north but no other fossils were discovered. The
described sample of permineralised wood was collected
by Hermann Grill in the outcrop area.

Previously described fossil woods from Namibia
Fossil wood is fairly abundant within deposits ofthe

Karoo Supergroup in Namibia but has seldom been
described in detail (Pickford 1995). Descriptions were
given by Krausel (Krausel & Range 1928, Krausel
1956a, b), who erected several new genera of
pycnoxylic woods with well-preserved primary
vasculature and piths. It is the secondary xylem amongst
gymnosperm woods that is most frequently preserved.
This has led to the development of a dual nomenclatural
system. There is a large number of genera for woods
with only secondary xylem (woods without a pith =
homoxylous, or woods where the pith has simply not
been preserved) . A separate group of genera
incorporates wood with intact primary xylem, pith and
secondary xylem (pycnoxylic woods = dense and
massive with secondary xylem more abundant, as in
tree-like conifers; manoxylic = loose, soft and scanty
with the pith more abundant, as in pteridosperms). Pith
is generally considered to be the most important feature
and woods preserving this tissue are segregated into two
groups. Homocellular piths have only parenchymatous
tissue whereas heterocellular piths have sclereids,
secretory cells, or secretory canals within the
parenchymatous tissue. The shape of the pith,
cylindrical or variously lobed, is another feature of
taxonomic significance.

Lower Gondwanan woods can be divided into five
types based on the tracheid pit and spiral thickening
arrangements in the secondary wood (Lepekhina1972,
Pant & Singh 1987). Thus two specimens with the same
secondary xylem but only one having a pith, are placed
in different taxa. This is essential because they could
belong to different plant groups. Affinities of the woods
have been discussed by various authors, but there is little
consensus (cf. Krausel et at. 1961, Lepekhina 1972,
Pant & Singh 1987). Homoxylous woods have recently
been described from the younger formations of the
Karoo Basin (Bamford 1999).

The pycnoxylic woods from Namibia and their
characteristics are summarised in Table 1. Table 2
shows a comparison of Megaporoxylon-type woods
described in the literature. In this paper a newly
discovered specimen of Megaporoxylon scherzi is
described because it is particularly well preserved and
well dated. Two less well preserved specimens of
Megaporoxylon kaokense are also described.
Polished thin sections were made of all of the silicified
woods in the following orientations: transverse section,
radial longitudinal section and tangential longitudinal
section. The sections were mounted on petrographic
slides, ground and polished to a thickness of
approximately 25-30 /-1m, studied and photographed
under a Zeiss Axioskop microscope.

DESCRIPTION OF WOOD
Megoporoxy/on scnerzi Krausel 1956b.

Specimen No: BB27 and Slide No: BPI16/935
Locality: Brukkaros River, about 23 km NNW ofTses,
25°41.26' S, 18°01.88'E (No.1: Figure 1)
Stratigraphy: Ganigobis Shale Member, DS II, Dwyka
Group, Karoo Supergroup
Collector: Berthold Bangert
Figures: 2-8

Description
The wood is silicified, yellow and brown, and

measures 9x5x6 cm. The pith is only a few centimetres
in diameter and heterocellular with cells containing a
dark substance (secretory cells?) scattered within the
parenchymatous pith. The primary xylem lobes are very
small and end arch (Figures 2, 3). In longitudinal section
the protoxylem tracheids exhibit typical annular or
helical thickening (Figure 4). The growth rings are very
clearly seen, varying in width from 1-7 mm, with an
average of 4,6 mm (Figure 5). Latewood comprises one
tenth to one quarter of each ring and ends abruptly at the
beginning of the earlywood. The transition from
earlywood to latewood is gradual. There are usually 20-
30 rows of latewood cells. In transverse section the
tracheids of the secondary xylem are square to
polygonal and thin walled (wall between two adjacent
cells is 5)..lm wide) in the earlywood, and only slightly
thicker in the latewood (7,5)..lm). The earlywood mean
tangential diameter is 35)..lm (range 25-42 )..lm) and mean
radial diameter is 36)..lm (range 25-47)..lm). The latewood
mean tangential diameter is 29)..lm (range 25-35)..lm) and
mean radial diameter is 12)..lm (range 7-22)..lm).
Bordered pitting on the radial walls of the tracheids is
araucarian, predominantly uniseriate and contiguous
(90%), slightly flattened, but also biseriate and alternate,
and rarely uniseriate and separate (Figure 6) . In the
earlywood the mean diameter of the pits is 12,5)..lm, and
1 O)..lm in the latewood. The pit apertures are mostly 5)..lm
wide and round but some areas show the cross-like
structure of elongated pits of adjacent cells overlapping
at right angles, particularly in the narrow latewood
tracheids (Figure 7).

The rays are uniseriate and low, 2-10-15 (minimum,
mean, maximum) cells high, with thin, unpitted walls.
The cross-field pits are oopores: large, simple, oval to
fusiform and obliquely orientated, the single pit (very
rarely two) filling most of the field (Figure 8). In the
earlywood they are on average 37 )..lm long and 15 )..lm
wide, orientated in the same direction. The latewood
cross-field pits are orientated in the opposite direction
and are smaller, but because the field is also smaller, they
occupy most of the field. These pits are 20 )..lm long and
5 )..lm wide, and also without a border. No resin canals,
resin or axial parenchyma were seen within the
secondary xylem.

Identification
This wood is the same as Megaporoxylon scherzi

described by Krausel (l956b) from the upper Dwyka
beds (now considered to be the Lower Ecca GroupYof
the Karoo Supergroup near Mariental. He did not give
measurements of the cells or pits but said that the cross
field pits of M. kaokense were up to three times as high
as the tracheid pits. In M. scherzi (Krausel 1956b) the
cross-field pits were very similar to those of M.
kaokense (Krausel 1956a) but were oval and slanting,
not round. M. zellei has slightly shorter oopores
(Krausel 1956b).

Megoporoxy/on kookense Krausel 19568
(1) Specimen No: BB28, and slide No: BP1161746
Locality: Asab River, about 22 km NW of Tses,
25°43.04' S, 17°59.08' E (No.2: Figure 1)
Stratigraphy: Ganigobis Shale Member, DS II, Dwyka
Group, Karoo Supergroup
Collector: Berthold Bangert
Figures: 9-12

Description
The specimen BP1l61746 is grey, laterally

compressed with longitudinal grooves, but has
surprisingly well preserved secondary xylem. The pith is
heterocellular with secretory cells scattered between
the parenchyma cells. Diameter of the pith is unknown.
The primary xylem is in lobes extending into the pith and
is endarch (Figure 9). In longitudinal section these
protoxylem tracheids have close spiral thickening. The
metaxylem tracheids have alternate biseriate pitting.

In the secondary xylem the growth rings are on
average 1 mm wide and have less latewood, only 2-3
rows, than the specimen described above (BPI16/935).
Tracheids are squarish in transverse section and have a
dark substance in the lumina which is likely to be an
artefact of preservation (Figure 9). The earlywood
tracheid mean tangential diameter is 35 )..lm (range
25-45)..lm) and mean radial diameter is 33 )..lm (range
27 -45)..lm). The growth rings are narrow (<1 mm) near
the pith. The latewood is made up of 2-3 rows of radially
compressed tracheids. The late wood tracheid mean
tangential diameter is 33)..lm (range 20-40)..lm) and mean
radial diameter is 12)..lm (range 1O-18)..lm). Adjacent cell
walls are 7-1 O)..lm thick. Tracheid bordered pitting is
clear in some areas (Figure 11). These pits are uniseriate
or biseriate and alternate, contiguous and slightly
flattened. Their diameter is 12-15)..lm. There are
approximately equal proportions of uniseriate and
biseriate pitting.

The rays are low, 1-5-12 cells high, exclusively
uniseriate and relatively rare (Figure 10). The cross
field pits are large, completely filling the field, simple and
almost round, 17.5)..lmx 17.5)..lm(Figure 12). No resin or
resin canals occur in the secondary xylem.

Figures 2-5. Megaporoxylon scherzi Krausel, BP116/935; 2: Diagram of specimen showing growth rings in secondary xylem and endarch
primary xylem: metaxylem (m, large thick-walled cells), and protoxylem (p, cluster of small thick-walled cells) which is nearer
the pith (large thin-walled cells, some with resiniferous contents); 3: Transverse section (TS) of specimen BPI16/935 showing
the primary xylem lobe: (secondary xylem above photograph), metaxylem (large central cells), small protoxy1em cells just below
and large pith cells. Scale bar = 60 Ilm; 4: Radial longitudinal section (RLS). Protoxylem tracheids have helical and annular
thickenings (left), pith cells on the right. Scale bar = 40 Ilm; 5: (TS). Earlywood tracheids with thin walls in upper part and thicker
walled latewood tracheids below. Scale bar = 350 Ilm.

Figures 6-8: Megaporoxylon scherzt~ BP116/935; 6: (RLS). Uni- and biseriate bordered pitting on radial walls of early wood tracheids. Scale
bar = 20 /-lm; 7: (RLS). Uniseriate pits on radial walls of latewood tracheids. Scale bar = 40 /-lm; 8: (RLS). Crossfield pitting.
Oopores in the earlywood are oval oblique. Scale bar = 40 /-lm. 9: Megaporoxylon kaokense Krausel BPI161746. TS showing
secondary xylem (top), primary xylem lobe and pith with dark cells below. Scale bar = 175 /-lm.

Figures 10-13: Megaporoxylon kaokense BPI161746; 10: Tangential longitudinal section (TLS) showing low rays, uniseriate and 3-5 cells
high. Scale bar = 40 /-lm. 11 : (RLS). Bordered pits on the radial walls of the tracheids are uniseriate and contiguous. Scale bar
= 60 !-lm. 12: (RLS). Cross-field pits are round oopores, filling the field. Scale bar = 60/-lm. 13 : BP116/547, Megaporoxylon
kaokense (Ganigobis Shale Member). (RLS). Cross-field pits are large round oopores, filling the field . Scale = 20 /-lID .

(2) Specimen and slide number: BPI16/547
Locality: Wasser River, near the farm Tsaraxa, about
27.5 km SW ofTses, 26°02.55' S, 17°57.20' E (Figure 1
insert).
Stratigraphy: Ganigobis Shale Member, DS II, Dwyka
Group, Karoo Supergroup
Collector: Hermann Grill
Figure: 13

Description
This specimen is grey-black, measured 9 x 9 x 4,5 cm,

has growth rings 1-2 mm wide and was from a trunk with
a diameter greater than 50 cm. No pith was preserved,
only secondary xylem. The earlywood tracheid mean
tangential diameter is 261lm (range 20-30llm) and the
radial diameter is 241lm (range 20-30Ilm). For the
late wood the mean tangential diameter is 251lm (range
23-27Ilm) and mean radial diameter is 13llm (range 10-
15Ilm). Bordered pitting occurs only on the radial walls
of the tracheids and is 1-2 seriate, alternate, contiguous
or rarely separate, with a diameter of 1Ollm. The rays
are low and uniseriate, 2-5-8 cells high and rare. Cross
field pits are large, simple, round to oval and fill the cross
field, 17.51lm by 17.51lm (Figure 13).

Identification
These specimens are conspecific with woods

assigned to Megaporoxylon kaokense Krausel
(1956a) from the Tsarabis Formation (Lower Ecca
Group, formerly Upper Dwyka) in the southern
Kaokoveld. They are very similar to the specimen of M.
scherzi, described above, except that the cross-field pits
are more rounded, rather than elliptic and oblique. Only
single pits were seen but there were very few cross-field
pits visible altogether.

Comparison with other fossil woods
Krausel (1956a, b; Krausel & Range 1928) described

17 taxa of fossil woods from Dwyka and Ecca Group
deposits in Namibia which had pith preserved (Table 1).
Mostly the piths have small diameters, less than 5 cm, the
primary xylem was endarch in the Dwyka woods (one
exception), and mesarch in the Ecca woods, and all the
secondary woods have araucarian tracheid pitting on the
radial walls. Generic differences lie in the pith types and
the cross-field pitting. Typical araucarian cross-field
pitting is cupressoid (or narrowly bordered), but one
specimen, Dadoxylon arberi, has small, simple pits of
the Zalesskioxylon secondary xylem type. The other
woods have secondary xylem of the
Protophyllocladoxylon-type and are shown here in
Table 1 as having oopores in the cross-field pit column.
In his original description of Protophyllocladoxylon
Krausel (1939) described the secondary xylem as
having purely araucarian tracheid pitting (slightly
compressed pits, usually 2 or more seriate and
alternate), but MUller-Stoll & Schultze-Motel (1989)
included woods with mixed tracheid pitting (araucarian
and abietinian) in this genus. The type of pitting rather
than the secondary wood genera are used here to avoid

confusion. The secondary xylem of Megaporoxylon is
the same as that of Protophyllocladoxylon Krausel.

Maheshwari (1966) described a fourth species of
Megaporoxylon, M. krauseli from the Raniganj
(Upper Permian) of India, and two more species from
the Permian of Antarctica (Maheshwari 1972). All six
species are very similar (Table 2). The species from
India and Antarctica are a little younger than the
Namibian species. To the best of the authors'
knowledge Megaporoxylon has not been described
from South America but it does occur there (Da Rosa
Alves, pers. comm). Krausel & Dolianiti (1958)
described several other genera shared between Brazil
and Namibia (Lobatoxylon, Taxopitys), although
different species characterised each area.
Protophyllocladoxylon is common to both continents
(Guerra-Sommer, 1977) but without a pith preserved it
is difficult to determine whether the same plant is being
considered. The Protophyllocladoxylon woods, i.e.
those without a pith, occur in the Mesozoic rather than
the Palaeozoic (MUller-Stoll & Schultze-Motel 1989), so
Megaporoxylon and Protophyllocladoxylon are
unlikely to belong to the same gymnospermous groups.

DISCUSSION
Megaporoxylon scherzi (Figure 5) has very clear

growth rings and abundant latewood whereas
Megaporoxylon kaokense has narrower latewood. In
both cases the wood nearest to the pith has been studied,
which is the wood of the very young tree. The specimens
originate from different localities: M. scherzi was found
in the upper part and M. kaokense was collected from
the lower part of the Ganigobis Shale Member and so
they are very unlikely to be exactly contemporaneous.
The authors cannot assume that their immediate
environments were the same, nor do we know for
certain if the two species of wood represent different
plant taxa because woods tend to be more conservative
than other parts or organs of the 'plant. The growth rings,
however, imply a seasonal environment which is to be
expected due to the high latitude positiqn of southern
Gondwana during the Carboniferous and Early Permian.

The stratigraphically restricted distribution
(Carboniferous, - Lower Permian) of Megaporoxylon
makes it a potentially useful biostratigraphic indicator.
Gondwanan woods with piths seem to be confined to the
Carboniferous and Lower Permian deposits which could
be a preservational artefact but of the many intact
specimens collected from the Upper Permian and
Triassic of southern Africa (Bamford 1999,2000) none
has a pith. The dominant Gondwanan gymnosperms at
this time were the Cordaitales, Glossopteridales and
early conifers. The Cordaitales have been recorded
from the Upper Carboniferous to the Upper Permian
deposits with many organ genera having been assigned
to this group. The cordaitalean woods Mesoxylon and
Cordaioxylon have large piths with secondary xylem of
the Araucarioxylon-type (Trivett & Rothwell
1991).The glossopterids are predominantly a Permian
group; leaves have also been found in rocks of the

TABLE 1.
Fossil woods described from Namibia by Kriiusel (in Kriiusel & Range 1928, first eight taxa; KriiuseI1956a,b, the rest of the
taxa). All the woods have alternate pitting on the radial tracheid walls of the secondary xylem, variously described as
''Amucorioxylon-type, ZoIesskioxylon-type or PllyUoclot/oxylon-type". The main differences are in the cross-field pits
which for the three pnwdingtaxa are cupressoid, simple and ooporoid respectively. The Kaoko Formation here is equivalent

to the Dwyka and lower Ecca Groups in this region. (Bp = bordered pits on tracheid walls; R = ray width).

FOSSIL WOOD PITH TYPE PRIMARY 2 XYLEM: LOCALITY
XYLEM CROSS-FIELD STRATIGRAPHY

Medullopitys heterogeneous mesarch oopores Keetmanshoop
sclerotica sclerenchyma Ecca Group

strands

Abietopitys homogeneous mesarch cupressoid Keetmanhoop
perforata Ecca Group

Phyllocladopitys homogeneous mesarch oopores Ganigobis
capensis Dwyka Group

Phyllocladoxylon - - oopores Ganigobis
capense Dwyka Group

Dadoxylon rangei homogeneous mesarch cupressoid Various
(?) Ecca Group

Dadoxylon homogeneous mesarch cupressoid Keetmanshoop
porosum (?) Ecca Group

Dadoxylon arberi homogeneous mesarch Simple, small, Doros crater
(?) numerous Tsarabis Fm

(Lower Ecca Group)

Taxopitys heterogeneous mesarch cupressoid Doros crater
africana secretory cells spiral Tsarabis Fm

thickening (Lower Ecca Group)

Solenoxylon wissi discoid, large endarch cupressoid Kaokoveld
bp: 1-3 Tsarabis Fm

JLower Ecca Group)

Solenoxylon kurzi diSCOid , large endarch cupressoid Kaokoveld
bp: 1 Tsarabis Fm

(Lower Ecca Group)

Solenoxylon discoid , large endarch cupressoid Kaokoveld
oberholzeri bp: 2-4; R: 2 Tsarabis Fm

ser. 1Lower Ecca Group)

Lobatoxylon lobed endarch cupressoid Kaokoveld
kaokense Tsarabis Fm

(Lower Ecca Group)

Megaporoxylon heterogeneous endarch oopores Kaokoveld
kaokense secretory cells (round) Tsarabis Fm

~Lower Ecca Group)

Megaporoxylon heterogeneous endarch oopores j Mariental
scherzi secretory cells (oval, slanting) Lower Ecca (?) Group

Megaporoxylon heterogeneous endarch oopores Amalia
zellei secretory cells shorter Dwyka Group

Kaokoxylon heterogeneous endarch cupressoid Kaokoveld
reuningi sclerenchyma single, small Tsarabis Fm

(Lower Ecca Group)

Kaokoxylon heterogeneous endarch cupressoid Kaokoveld
durum sclerenchyma several , smaller Tsarabis Fm

(Lower Ecca Group)

Phyllocladopitys homogeneous mesarch oopores Mariental
martini bp: flattened Lower Ecca (?) Group

TABLE 2.
Comparison of woods of MegopoTO.1J'/on described in the literature and in this paper. Growth rings are distinct in all of

the species.

SPECIES LOCALITY PITH TYPE
STRATIGRAPHY Primary xylem
AUTHOR

Megaporoxylon Kaokoveld , Namibia heterogeneous
kaokense Tsarabis Fm. , Lower secretory cells

Ecca Group endarch
Krausel1956a

M. scherzi Mariental, Namibia heterogeneous
Lower Ecca (?) secretory cells
Group endarch
Krausel 1956b

M. zellei Amalia, Namibia heterogeneous
Dwyka Group secretory cells
Krausel 1956b endarch

M. krauseli Raniganj, India heterogeneous
Upper Permian secretory cells
Maheshwari 1 966 endarch?

M. antarcticum MtWeaver, homogeneous
Antarctica
Permian endarch
Maheshwari 1972

M. cana/osum Mercer Ridge heterogeneous
Antarctica secretory cells
Permian endarch
Maheshwari 1 972

BP/16/935 Brukkaros, Namibia heterogeneous
M. scherzi Ganigobis Shale secretory cells

Member endarch
Dwyka Group
Bangert & Bamford

BP/161746 Ganigobis, Namibia heterogeneous
M. kaokense Ganigobis Shale secretory cells

Member endarch
Dwyka Group
Bangert & Bamford

BP/16/547 near Tses. Namibia unknown

M. kaokense ? Ganigobis Shale
Member

Dwyka Group

Bangert & Bamford

Dwyka Group in the south-western main Karoo Basin
(Anderson & McLachlan 1976). Araucarioxylon has
been considered the wood type of the glossopterids
(Gould & Delevoryas 1977; Pigg & Taylor 1993) but this
wood type may have been produced by more than one
plant group, especially as this wood type occurs after the
extinction of the Glossopteris flora. The lack of
phylogenetically informative characters currently
prohibits assignment of Megaporoxylon woods to any
gymnosperm order.

TRACHEID PITS CROSS-FIELD RAYS
number Number of pits width
arrangement shape height
size size (cells)

mostly 1 seriate, 1 large or 2 smaller, 1 seriate
araucarian simple, round 1-14
-- 3 x size tracheid

Jl.its

mostly 1 seriate 1-2 simple, round to 1 seriate
araucarian oval and slanting 1-25

1-3 seriate 1 large, simple, 1 seriate
araucarian broader and 1-12-18
- not as high as in

M. kaokense

1-3 seriate 1-2 (-3) large, 1 seriate
simple, round 1-6-23

2-3 seriate 1 large, simple, 1 seriate
araucarian elliptical 2-6-17
8-11 flm 18x 11 ,5f.1m

1-2 (-3) seriate 1 large, simple, 1 seriate
araucarian oval-elliptic, oblique 1-4-8
- 16x22 - 5x12flm

1 (-2) seriate 1 large or 2 smaller, 1 seriate
araucarian simple, oval and 2-10-15
10-12,5f.1m oblique pits

20 x 5 - 37x15 flm

1-2 seriate 1 large, round, 1 seriate
araucarian simple pit , 17,5 flm 1-5-12
12-15flm completely fills the

field

1-2 seriate 1 large, round 1 seriate

araucarian simple pit, 30 x 2-5-8

10flm
25flm, completely
fills the field

ACKNOWLEDGEMENTS
Volker Lorenz and Markus Geiger (WUrzburg) helped with the

field work and collecting samples of permineralised wood. Richard
Lewis (BPI) and Rupert Wassermann (WUrzburg) are thanked for
preparing the slides. Research was funded by the German Research
Foundation (DFG) and the Postgraduate Research Program
"Interdisciplinary Geoscience Research in Africa". The National
Monuments Council of Namibia is thanked for their permission to
study the wood samples outside Namibia. Logistic support by the
Geological Survey of Namibia is gratefully acknowledged. Kathleen
Pigg and an anonymous reviewer are thanked for their comments.

REFERENCES

ANDERSON, A.M. & McLACHLAN, I.R. 1976. The plant record in the Dwyka and Ecca Series (Permian) of the southwestern half of
the Great Karoo Basin. Palaeontologia a.fricana 19,31-42.

BAMFORD, M.K. 1999. Permo-Triassic fossil woods from the South African Karoo Basin. Palaeontologia a.fricana 35,25-40.
BAMFORD, M.K. 2000. Fossil woods of Karoo-aged deposits from South Africa and Namibia as an aid to biostratigraphic correlations.

Journal 0.1 A.frican Earth Sciences 31,119-132.
BANGERT, B., STOLLHOFEN, H., LORENZ, V. & ARMSTRONG, R. 1999. The geochronology and significance of ash-fallout tuffs in

the glaciogenic Carboniferous-Permian Dwyka Group of Namibia and South Africa. Journal 0.1 A.frican Earth Sciences 29 (1), 33-49.
BANGERT, B., STOLLHOFEN, H., GEIGER, M. & LORENZ, V. (in press). High resolution tephrostratigraphy, fossil record and age

of Carboniferous glaciomarine mudstones, Dwyka Group, southern Namibia. submitted for the Henno Martin Commemorative Volume,
Communications 0.1 the Geological Survey 0.1 Namibia, Windhoek, Namibia.

DICKINS, J.M. 1961. Eurydesma and Peruvispira from the Dwyka beds of South Africa. Palaeontology 4, 138-148.
DICKINS, J.M. 1984. Late Palaeozoic glaciation. BMR Journal Australian Geology and Geophysics 9, 163-169.
GARDINER, B.G. 1962. Namaichthys schroederiGiirich and other Palaeozoic fishes from South Africa. Palaeontology 5,9-21.
GOULD, R.E. & DELEVORYAS, T. 1977. The biology of Glossopteris: evidence from petrified seed-bearing and pollen-bearing organs.

Alcheringal,387-399.
GRILL, H. 1997. The Permo-Carboniferous glacial to marine Karoo record in southern Namibia: sedimentary facies and sequence stratigraphy.

Beringeria 19, 98 p., Wiirzburg, Germany.
GUERRA-SOMMER, M. 1977. Damudoxylon (Maheshwari) Maheshwari 1972, urn genero ocurrente no Gondwana do Brazil. Pesquisas,

PortoAllegre7,131-144.
GORICH, G. 1923. Acrolepis lotziund andere Ganoiden aus den Dwyka Schichten von Ganikobis, Siidwestafrika. Beitdge zur geologischen

Eiforschung der Deutschen Schutzgebiete 19, 26-74.
HARLAND, W.B., ARMSTRONG, R.L., COX, A.V., CRAIG, L.E., SMITH, A.G. & SMITH, D.G. 1990.Ageologictimescale. Cambridge,

Cambridge University Press, 263 p.
HEATH, D.C. 1972. Die geologie van die Sisteem Karoo in die gebied Mariental-Asab, Suidwes-Afrika. Memoirs 0.1 the Geological Survey

0.1 South A.frica 61, 36 pp.
KRAuSEL, R. 1939. Ergebnisse der Forschungsreisen Prof. E. Stromers in den Wiisten Agyptens, pt. 4, Die fossilen Floren Agyptens.

Bayerische. Akademie der Wissenschaflen Abhandlungen Mathematisch Naturwissschafllichen Abteiling, n.f. 47, 1-140.
KRAUSEL, R. 1956a. Der "Versteinerte Wald" im Kaokoveld, Siidwest-Afrika. Senckenbergiana lethaea 37,411-445.
KRAUSEL, R. 1956b. Holzer aus dem siidlichen Gebiet der Karru-Schichten Siidwest -Afrikas. Senckenbergiana Lethaea 37,447-453 .
KRAUSEL, R. & DOLIANITI, E. 1958.Gyrnnospermenholzer aus dem PaHiozoikum Brasiliens. Palaeontographica I04B, 115-137.
KRAuSEL, R. , MAITHY, P.K. & MAHESHWARI, H.K. 1961. Gymnospermous woods with primary structures from Gondwana Rocks

- A review. Palaeobotanist 10, 97 -107.
KRAUSEL, R. & RANGE, P. 1928. Beitage zur Kenntnis der Karruformation Deutsch-Siidwest-Afrikas. Beitdge zur geologischen

Eiforschung der Deutschen Schutzgebiete 20, 1-54.
LEPEKHINA, V.G. 1972. Woods of Palaeozoic pycnoxylic gymnosperms with special reference to North Eurasia representatives.

Palaeontographica 138B, 44-106.
MAHESHW ARI, H.K. 1966. On some fossil woods from the Raniganj Stage of the Raniganj Coalfield, Bengal, pt. 28, in Studies in the

Glossopteris flora ofIndia. Palaeobotantstl5, 243-257.
MAHESHW ARI, H.K. 1972. Permian wood from Antarctica and revision of some Lower Gondwana wood taxa. Palaeontographica 138B,

1-43.
MARTIN, H. & WILCZEWSKI, N. 1970. Palaeoecology, conditions of deposition and the palaeogeography of the marine Dwyka Beds

of South West Africa. In: Haughton, S .H. (ed.) International Gondwana Symposium 2, 225-232: Proceeding Papers lUGS 2nd Gondwana
Symposium (South Africa), Pretoria.

McLACHLAN, I.R. & ANDERSON, A. 1973. A review of the evidence for marine conditions in southern Africa during Dwyka times.
Palaeontologia a.fricana 15, 37 -64.

MENNING, M. 1995. A numerical time scale for the Permian and Triassic Periods: an integrated time analysis. In: Scholle, P.A., Peryt, T.M.
& Ulmer-Scholle, D.S. (eds) : The Permian 0.1 Northern Pangea, Vol. 1,77-97. Berlin, Springer.

MULLER-STOLL, W.R. & SCHULTZE-MOTEL, J. 1989. Gymnospermen-Ho1zer des deutschen Jura. Teil2: Die Protopinoiden Holzer.
Zeitschriji derdeutschen geologIschen Gesellschafl l40, 53-71.

PANT, D.D. & SINGH, V.K.1987. Xylotomy of some woods from Raniganj Formation (Permian), Raniganj Coalfield, India. Palaeontographica
203B, 1-82.

PICKFORD, M. 1995. Karoo Supergroup palaeontology of Namibia and brief description of a thecodont from Omingonde. Palaeontologia
a.fricana 32, 51-66.

PIGG, K.P. & TAYLOR, T.N. 1993. Anatomically preserved Glossopteris stems with attached leaves from the central Transantarctic
Mountains, Antarctica. American Journal 0.1 Botany 80, 500-516.

SCHROEDER, H. 1908. Marine Fossilien in Verbindung mit permischen Glazialkonglomeratin Deutsch-Siidwestafrika. Jahrbuchkiinigliche
Preussische geologlsche Landesanstalt29, 694-697.

SACS (South African Committee for Stratigraphy) 1980. Stratigraphy of South Africa. Part 1: Lithostratigraphy of the Republic of South
Africa, South West AfricalNamibia and the Republics of Bophuthatswana, Transkei and Venda. (L.E. Kent, Compiler), Handbook
Geological Survey 0.1 South A.frica, 8, 690 pp.

STOLLHOFEN, H., STANISTREET, I.G., BANGERT, B. & GRILL, H. 2000. Tuffs, tectonism and glacially related sea-level changes,
Carboniferous-Permian, southern Namibia. Palaeogeography, Palaeoclimatology, Palaeoecology. 161, 127-150

THERON, J.N. & BLIGNAULT, H.J. 1975. A model for the sedimentation of the Dwyka glacials in the Southwestern Cape. In: Campbell,
E.S .W. Ed, International Gondwana Symposium 3, 347-356. Canberra, University Press.

TRIVETT, M.L. & ROTHWELL, G.W. 1991. Diversity among Palaeozoic Cordaitales. Neues Jahrbuch.fur Geologie und Paldontologie,
Abhandlungen 183, 289-305.

VISSER, J.N.J. 1997. Deglaciation sequences in the Permo-Carboniferous Karoo and Kalahari Basins of southern Africa: a tool in the analysis
of cyclic glaciomarine basin fills. Sedimentology44, 507 -521.

Palaeont. a.fr. , 37, 25-40 (2001)

A NEW ACTINOPTERYGIAN FISH SPECIES FROM THE LATE PERMIAN BEAUFORT
GROUP, SOUTH AFRICA

Patrick Bender

Council for Geoscience, Private Bag Xl12, Pretoria, South Africa.
e-mail·bender@.?li:co.za

ABSTRACT
A new genus and species of actinopterygian (ray-finned) fish, Bethesdaichthys kitchingi, is

described from the Tatarian, Late Permian, Lower Beaufort Group of South Africa. Bethesdaichthys
is presently known from three localities, two in the New Bethesda and one in the Victoria West districts
of the Karoo region respectively. The fossils were recovered from within the Abrahamskraal
Formation Tapinocephalus Assemblage Zone at the Victoria West locality, and from an uncertain
Formation possibly closely equivalent to the Balfour Formation, within the Dicynodon Assemblage
Zone at the New Bethesda sites. Bethesdaichthys kitchingi is a fusiform fish, up to approximately
300mm in total length, with the skull displaying a moderately oblique suspensorium, and a maxilla
with a large sub-rectangular postorbital blade. Furthermore there is a complex offour suborbital bones
adjacent to the orbit. The pectoral fin is large relative to body size and the tail is heterocercal with
an elongate tapered dorsal body lobe. The anterior midflank scales in particular exhibit a distinctive
dermal ornamentation consisting of numerous ganoineridges. The phylogenetics and interrelationships
of Bethesdaichthys kitchingiare examined. It appears to exhibit a relatively conservative morphology
similar to that found in possibly related Carboniferous taxa such as the South African taxa
Australichthysand Willomorichthys. Bethesdaichthys kitchingiis derived relative to stem-actinopterans
such as the Howqualepis and Mimia, and also derived relative to southern African Palaeozoic
actinoptyerygians such as Mentzichthys jubbl; and Namaichthys schroeden; but basal to stem
neopterygians such as Australosomus, Perleldus and Saurichthys.

KEYWORDS: Bethesdaichthys, palaeonjscid, Late Permian, Tatarian, Beaufort Group, Actinopterygii.

INTRODUCTION
A new genus and species of Late Permian

actinopterygian fish is described here from the Lower
Beaufort Group of South Africa, based essentially on
well preserved and diagnostic skeletal elements.
Taxonomically relevant actinopterygian fossil remains,
in particular diagnostic skull remains, have up to now not
been described from the Lower Beaufort Group,
although incompletely preserved skeletal remains and
isolated body scales of fossil fish have been recorded
from much or most of the biostratigraphic range (see
Broom 1913a, 1913b;lubb & Gardiner 1975; Woodward
1888,1889,1893). Recently, Bender (2000) documented
for the first time well preserved and relatively complete
actinopterygian remains from the Lower Beaufort
Group, tentatively describing several new species,
including Bethesdaichthys kitchingi.

Bethesdaichthys kitchingi is an actinopterygian fish
which belongs to a group of early actinopterygian taxa
collectively referred to as "palaeoniscids" (Traquair
1877-1914; Gardiner 1967) or "Palaeoniscomorpha"
(Lund et al 1995). It is generally accepted that the
palaeoniscids constitute a paraphyletic group of mostly
Palaeozoic actinopterygians (Coates 1993), with a global
distribution. These palaeoniscids or lower
actinopterygians represent the "primitive" or basal
members of the Subclass Actinopterygii (Gardiner
1973).

The sedimentary rocks of the Beaufort Group have
yielded diverse and important fossils, including macro
and micro-palaeobotanical remains, vertebrate and
invertebrate body fossils and traces (Hancox & Rubidge
1997). Analysis of the fossils provides information on the
evolution of life in the Permo-Triassic, and has proved
significant in unravelling the geological development of
the Karoo Basin (Hancox & Rubidge 1997). On the
basis of its uniquely large and relatively complete
continental Permo-Triassic sedimentary sequence, the
Beaufort Group is considered almost as a 'world
stratotype' for continental Permo-Triassic age
geological and palaeontological research (Smith 1990).
The Beaufort Group is particularly renowned for its
diversity and range of therapsid fossils, which elucidate
the evolutionary transition to mammals (Broom 1932;
SACS 1980). The therapsids have been utilized as a
basis for an eightfold biostratigraphic subdivision of the
Group (Rubidge 1995), with the Lower Beaufort Group
comprising six of the eight biozones (Figure 1).

MATERIALS AND METHODS
Three laterally compressed Bethesdaichthys

specimens were recovered from a Tapinocephalus
Assemblage Zone locality on the farm Blourug, Victoria
West district, Abrahamskraal Formation, Adelaide
Subgroup, Lower Beaufort Group. These specimens
were contained within a single, thin, buff-coloured, fine

STRATIGRAPHY

WEST 0 F 24°E EAST OF 24°E NORTHERN OFS
ASSEMBLAGE

ZONE

- 7:Z 7777:?<::::

E ..•..•.. ~~~~~~~ .. ~ .•...•.••..........•..

~ CYllogllathus
;:J

U 0 .... ~ BURGERSDORP F. DRIEKOPPEN F .
00 '-'
~ CQ

~
;:J
(/J

~
<

~ E-<

I ••• • •• • •••••••••••• :·:·

> .. :.:.:: .. : .... ::::: .. :.:::. )< 1\<) .: ... :: ....... : .......... U ... ·:.:.C·. 8GG 07 (/J

~ Iii ~G~. ? 1·<% ~R~ ~B~ f.
Lystrosaurus

~
~dt(tn <

E-<

I ••••• •·•· •••••••• · ?? ••••• : ••• ·:.:U •• ··:.·:·.:.< >? I> ..... ::.:: .. :: .. <. :.::::::::;::.::::::: ::::::: -

Palingkloof M. Harrismith M.

f--- ::::.:. <

It : ;~:;:;:,: .;;

j%i~;
•••••••••••••

~ W;rJ ~ Elandsberg M.
0
~ ~
c.!i
E-- ~

• ••••• • ••• ·2 •• • ••• ·,.·. ,112111 0111
:z
~

~
~

...
0 ~ jSjtj"J u

Q DicYIIOdoll
~

.· ........... ·.·.·.?i· ..
z I· ,Sill;;;1 ~]mi

••••••••••

~ 0
~ -< ~

~ ....:l
~ I) ):.: ••• U): ••• )C)C)) .. H 1=0 -<

1=0 0
~ :z
;:J
0 Steenkal llpsvlakte M. Daggaboersnek M. I'C:UU'}:tt>: ~

'-' CQ ::::::::,,:;:;:>:::>:: ::·:.:·:::·:·:·::: ...•• : .... · •. :.i.·:· •.••. \ •. ;:J

~ I;!);;:I~!
(/J

J~bi~ .· ......
....
~

1 •••••••••• • •••• ••• •••• g~Ji#igj) •• c ••••• :;;: Cistecephalus

'Z
...l

•••••••••••••••••••••••••••••••••••••••••••••••••

....
< ~ .... < ~

~ ~ Hoed emaker M.
E-- MIDDLETON F. Tropidostoma

~
:::::. ::::::

~ 'ii::,II: 1~ ·1· ••••••••• • •••••
Pristerognathus

VOLKSRUST F.
ABRAHAM SKRAAL

KOONAPF.
F.

Tapillocephalus

Eodicynodoll

~
~
0
~ KOEDOES BERG F./ WATERFORD F./
c.!i WATERF ORDF. FORT BROWN F. -<
U
U
~

Sandstone-rich Unit

Figure 1. Lithostratigraphic units and Vertebrate Assemblage Zones of the Beaufort Group (after Rubidge et af 1995).

to medium grained sandstone unit, which outcrops over
a lateral distance of approximately 70m, and contains
numerous specimens of other early actinopterygian
species. Until now the exact biostratigraphic zone of the
site has been uncertain, but as skull elements of a
dinocephalian therapsid were found approximately 15m
below the fish site, it appears that the site falls within the
Tapinocephalus Assemblage Zone.

A total of 15 Bethesdaichthys specimens were
recovered from a Dicynodon Assemblage Zone
roadside locality on the farm Wilgerbosch, New
Bethesda district, Adelaide Subgroup, Lower Beaufort
Group. These specimens were derived from a blue
green to green, ripple cross-laminated, silty mudstone
with a mudstone veneer on the upper surface; and also
from an interbedded mudstone/siltstone/fine-grained
sandstone sequence, up to 13 cm thick with fossil fish
found throughout the sequence. A single specimen was
recovered from a site located at a reservoir also on the
farm Wilgerbosch and approximately 35m
stratigraphically above the roadside site; preserved in a
blue-green siltstone horizon which is situated below a
laminated sequence similar in form to that at the roadside
site. The formational designation of the Wilgerbosch
sites is uncertain (Cole et al. in press), but they appear
to be situated in roughly the stratigraphic equivalent of
the Balfour Formation (see Figure 1).

Most specimens required mechanical and chemical
preparation before analysis was possible. Air scribes
and an assortment of needles and probes were used
initially to remove covering rock matrix, and in certain
cases to finely prepare the specimen surface prior to
analysis. A dilute 10% solution of acetic acid was on
occasions used to soften surrounding matrix. In the case
of a number of specimens, analysis was problematical
because of the weathered nature of the surface bone
and ganoine. In this case preparation of latex rubber
casts or peels was found to reveal excellent underlying
morphological detail; study and illustration of the casts
was facilitated by whitening with ammonium chloride.
Thin sections of scales were prepared, and studied for
histological analysis using a Zeiss standard petrographic
microscope with polarised light. Interpretive drawings
were made using a Leica MZ6 microscope with drawing
tube. Photographs were taken using a Nikon FM camera
mounted on a copy stand, and for the thin sections, a
Zeiss polaroid camera was used . The scanning electron
micrographs were made on a Leica Stereoscan 440 at
the Council for Geoscience, Pretoria. The phylogenetic
analysis was carried out by using the Gardiner and
Schaeffer 1989 (cladogram III) as a basis, since this is
the most recent comprehensive early actinopterygian
phylogenetic analysis. The relevant Bethesdaichthys
characters were compared to those of the constituent
taxa in the Gardiner and Schaeffer (1989) cladogram,
thus determining the phylogenetic position of
Bethesdaichthys. The results of the phylogenetic
investigation of Bethesdaichthys were illustrated
together with a revision of the Gardiner and Schaeffer
cladogram III, in which certain nodes and taxa within the
cladogram were updated.

SYSTEMATIC PALAEONTOLOGY
Class Actinopterygii Woodward 1891

Infraclass Actinopteri Cope 1871
Genus Bethesdaichthys gen. nov.

Derivation of name: Named after the Karoo mountain
hamlet, Nieu Bethesda, which is close to the
Wilgerbosch site where the first specimen of this taxon
was found . Bethesda means "place of flowing waters"
(Bible: John, Chap. 5: 2-4), probably appropriate to Late
Permian fluvial conditions in the region of the
Wilgerbosch fossil fish site. The suffix - 'ichthys' is
derived from the ancient Greek word for fish.

Diagnosis: A fusiform fish , approximately 300cm in
total length. Skull relatively broad, with a moderately
oblique suspensorium. Dermopterotic broadens
anteriorly and does not suture with the nasal.
Dermosphenotic is a crescent-shaped bone. Snout
region consists of a small premaxilla, antorbital, narrow
rostral and fairly elongate nasal. Jugal is a broad wedge
shaped bone. Maxilla has a large subrectangular
postorbital blade. Dentition consists of a median row of
large pointed conical teeth and an outer row of numerous
smaller pointed teeth. Preopercular with almost right
angled inflexion between the wedge-shaped dorsal, and
ventral limbs. There is a complex of four suborbital
bones including a large triangular suborbital between the
maxilla and the jugal. Opercular is broad and
rectangular, subopercular approximately 2/3 of the
height of the opercular with an obliquely angled ventral
margin. Branchiostegal rays number nine. Distal
bifurcation of the fin-rays is visible on the dorsal and
caudal fins . Pectoral fin is relatively large and consists
of fin-rays which are proximally jointed. Caudal fin is
heterocercal with an elongate dorsal body lobe. The
anterior rows of flank scales exhibit a distinctive dermal
ornament, with up to 14 curved and steeply inclined
dorsal ganoine ridges distinct from a series of up to 10
horizontally inclined ventral ridges. A series of enlarged
ridge scales is present along most of the dorsal margin.
Scale histology shows a laterally continuous,
multilayered ganoine layer.

Remarks: Bethesdaichthys is clearly different from
any of the other Beaufort Group actinopterygian taxa,
based on: fewer than 12-13 branchiostegal rays, the
shape of the dermosphenotic, maxilla, preopercular, and
opercular in particular, and in the morphology of the
suborbitals. Bethesdaichthys can be compared to the
South African Carboniferous genus Australichthys on
the basis of maxilla, preopercular and opercular shape
and form (see Gardiner 1969), but differs with regard to
the dorsal fin shape, size and form.

Type species: Bethesdaichthys kitchingi nov.
Derivation of the name: In honour of Mr 'Croonie'
Kitching, Nieu Bethesda resident and road builder, who
first discovered the Wilgerbosch fossil fish site in about
1928 while constructing a new road over part of the site.

sci

pmx

ant

Figure 2. Bethesdaichthys kitchingi holotype BP/1/4373/3. A.
Lateral view showing the dermal skull region. B. Camera
lucid a interpretation. (See p. 38 for abbreviations)

Holotype: BP/1/4373/3 , in the Bernard Price Institute
for Palaeontological Research (BPI), University of the
Witwatersrand, Johannesburg. From the Wilgerbosch
roadside locality, Dicynodon Assemblage Zone, Lower
Beaufort Group.

Rejerredspecimens: BP/lI4373/2, 3, 19, 110, 119, 120,
121, 122, 123, 124, 134, 138; BP/1/116, housed at the
BPI Palaeontology, Johannesburg. PB/95/6; PB/96/15,
housed at the Council For Geoscience, Pretoria; VIOl ,
housed at the Victoria West Museum, Victoria West;
TM 20, housed at the Transvaal Museum, Pretoria.

Horizon and locality: V 10 1 is from the Blourug
locality, Tapinocephalus Assemblage Zone, Lower
Beaufort Group. TM 20 is from the Wilgerbosch
reservoir site, and the rest of the specimens from the
Wilgerbosch roadside locality, Dicynodon Assemblage
Zone, Lower Beaufort Group.

Diagnosis: As for genus.

Remarks: A total of three specimens are recorded from
Victoria West, one from the Wilgerbosch reservoir site,
and fifteen from the Wilgerbosch roadside locality.

DESCRIPTION
Skull Roof

The skull roof is made up of paired parietals, frontals
and extrascapulars, a large dermopterotic and small
dermosphenotic located on either side of the frontals.
The dermal ornament over the whole of the head region
is fairly robust, in the form of mixed short ridges and
denticles, which are similar in shape and form to that
seen in the cheek region, but more robust. Bones of the
skull roof region were studied mainly in specimen BP/lI
4373/3 (Figure 2).

Parietals: (Figures 2, 3). The median section of the right
parietal is rather poorly preserved in BP/1/4373/3. It is
rectangular and approximately one third of the frontal
length. Anteriorly it sutures with the frontal and
posteriorly with the extrascapulars.

Frontals: (Figures 2, 3). The right frontal is preserved
in a somewhat distorted state in BP/1/4373/3. It is
relatively long, and narrow although this could be an
artifact of preservation. Anteriorly it sutures with the
rostral and posteriorly with the parietals.

Dermopterotic: (Figures 2, 3). Most of the right
dermopterotic is preserved in BP/1/4373/3. It appears
broadest anteriorly and tapers posteriorly. The anterior
overlap with the dermosphenotic appears smooth and
almost straight vertical. It does not contact the nasal, and
is fairly far removed from