OE ay ier = =< => = = = y

hae ee = s = _ = Se

EF NY 2 Ff 2) a 2 = K

sy . ~~ = my >’ .

= xB = = = 2 a |

IN NOILALILSNI NVINOSHLINS LIBRARIES SMITHSONIAN INSTITUTION NOILf

o aa . z a ey eal uk

= = = “ - Uy,” =

on =, ANN er 4 oO ng fre — fe

e Fe a a Zz a = ze

17 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVUdIT LIBR

Z 5 = S x 5 =

ae = O exe: e)

we IK ee : \S OE =

C a =) a: =D) : URS S =

Jiu ae a i N ee a

Hs Z z Z Z ie

ON NOILALILSNI SAIYVYEIT LIBRARIES

SMITHSONIAN INSTITUTION NOIL!

NVINOSHLINS S3SIYVUEII

NVINOSHLIWS

SMITHSONIAN

NVINOSHLIWS

SMITHSONIAN

NVINOSHLIWS

« .

SMITHSONIAN

I7 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS Saiuvugia LIBR

v x a 4 ¥

2 "i a us a : o

ee o. si oc a. oc =

Le => e - = 2 =

a mn a oO A ae =

2 = 2 a = ee S

z |

IN NOILALILSNI ~°4 ryYVvuUua bts LIBRARI ES_ SMITHSONIAN INST

: : : : : : :

= — D> : =

E a = oa = c= =

; : b - ae : ;

Z o Z 2 Z a zZ

17 NVINOSHLINS SSHIYVYGIT LIBR

Yo,

INSTITUTION NOILI

eat

NYINOSHLIWS

SMITHSONIAN

Nv NOSHLINS

NVINOSHLIWS

SMITHSONIAN

Pian

ON NOILNLILSNI NVINOSHLINS S3luvagiy

LIBRARIES SMITHSONIAN

NOILNLILSNI

NOILOLILSNI

LIBRARIES

tel da Sg

LIBRARIES

17 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3INVYEIT LIBR

S3IYVYSIT LIBRARIES

i z= = ra oe re

) = =e = a =) >

fe) op. - > E

<} is _ Pe) — Pa)

ce 5 F = F :

o z Z a qe C

ON _NOILNLILSNI SSIYVYUGIT LIBRARIES SMITHSONIAN INSTITUTION. NOIL

” = on z tees ” s wee 22)

ij me & = Pe aa Sarit < Gs =

= ra = r= \ < = = : S = Ls

ai O =. 7 oO NW YS 6 Oo NS S\ ie : 44

| wo ” AL Tae \ ” ” IN O

Oo ac Ca. Ee Oo Tr \N O Lip

= _ = & z ‘ z Gj

> > > = > = > 4

ma ” Fae wn Pas Ww z=

117 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS SaluvudiT LIBF

a <clpNeee pe): RESON | I era aN LU SYN OG <t

‘= = —_ pat = _—

z = Zz = Zz 4

O Be © a & O ae

: S aps 2 SN 3

E ae = Z E WO" 2

= >" = > = AS >

7) Zz 7) z ” s z

INSTITUTION NOILALILSNI_ NVINOSHLIV

YW) = YW) cs (ep) > *

P @ ul i uw >

0 = - za a -

‘ —{ ee aN

& S : "S : Sc OW

— (®) rae @) Zs @) SS

ae) z - Zz a} 2

INSTITUTION NOILNLILSNI NVINOSHLINS S3IMYVYGIT LIBRARIES SMITHSON

re Ze i < z i S

— 2) mee NP oO om

19.8] — SENN — 109) ae

- E SQ = 2 E

Wes =

oa = = Ces = oe I

2 - 2 SOQ = : 7

C = B 2 » zZ

S SJIYVYEIT_ LIBRARIES SMITHSONIAN NOILNLILSNI NVINOSHLI}

w z= OH WY Fa Ww z=

= pas =i < = <<

JNP2 a = ay : z

= ‘ ©: *s = = i SS hi =

Qs - S ae 6 i 2

= = 2 FE z Uy =

> = > S > =

ze ”) . ae ” a ze O

_INSTITUTION NOILNLILSNI NVINOSHLINS SZIYVYGIT LIBRARIES SMITriso.

= = Y)

z us 2 ees a ‘

=I ox sal te za Dae yy

a < = = z = PL

S o S fod S oe “Wily,

5 é 5 : 5 0 Gt

z 4 Fa ; ai PA ol

SJINVYGIT LIBRARIES SMITHSONIAN INSTITUTION NOILALILSNI NVINOSHLI!

Be i\2 a = yy Be é eS

: : & Gry = : 2 We

E = E VM = = = SN

= = a fc a = a NS

ASN = m a ma 3 i SNS

— Ww xed eres

MITHSONIAN INSTITUTION NOILNLILSNI NVINOSHUING 2S rau ain eee ARTES Ome

Fe = 4 = z . aA

: z 2 ; 3 NW 3

. SONS

: Ele SW SKHy

Z g, = g SN SE

>" >" = x >

i 2 rs ke Ae oe

i LIBRARIES SMITHSONIAN INSTITUTION NOILOLILSNI_ NVINOSHL

— Ww — WY ~

aj zZ in Zz i Zz

UY, ”

- = = Yy,* zi z=

é Lt at < Yi foi a < at

ss Y fea | c& ow G/ GLSp es oc &

STRESS _ (S) =aty O a e)

| - Zz ay Zz 4 2

SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3IYVUAIT LIBRARIES SMITHSON

| tg z a Z Gi =

| ay ee = “ S a S)

19. Np ue ~x es wD i

Ok: nee = > = > =

: “iy = = a = 2 a

| UP m 2 m S m 2

NVINOSHLINS SAINYVYGIT LIBRARIES) SMITHSONIAN _ INSTITUTION NOILALILSNI NVINOSHI

: w me BAe on z a Y =

= ist = OV S = =

> hits = = \ . A lf Sk Z ae

Tre Te W Bie Waa os we ‘bf a

ong Be ) r NN Oo Y% =e

= es z a : 2 “iy =

> = > = > i

Wy Fas (op) . FZ ” Ls oA é Ww

SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS S3I¥Wusl) LIBRARIES SMITHSON

sm — ‘yy sss (op)

ISSN 0303 2515

VOLUME = JUNE 1976

OF THE SOUTH AFRICA sAN

~ CAPE TOWN >

INSTRUCTIONS TO AUTHORS

1. MATERIAL should be original and not published elsewhere, in whole or in part. When

accepted, copyright becomes the property of the Trustees of the South African Museum.

2. LAYOUT should be as follows:

(a) Centred masthead to consist of

Title: informative but concise, without abbreviations and not including the names of new genera or species

Author’s(s’) name(s)

Address(es) of author(s) Gnstitution where work was carried out)

Number of illustrations (figures, enumerated maps and tables, in this order)

(b) Abstract of not more than 200 words, intelligible to the reader without reference to the text

(d) Introduction

(e) Subject-matter of the paper, divided into sections to correspond with those given in table of contents

(f) Summary, if paper is lengthy

(g) Acknowledgements

(h) References

(i) Abbreviations, where these are numerous

3. MANUSCRIPT, to be submitted in triplicate, should be typewritten and neat, double spaced

with 2,5 cm margins all round. First lines of paragraph should be indented. Tables and a list of

legends for illustrations should be typed separately, their positions indicated in the text. All

pages should be numbered consecutively.

Major headings of the paper are centred capitals; first subheadings are shouldered small

capitals; second subheadings are shouldered italics; third subheadings are indented, shouldered

italics. Further subdivisions should be avoided, as also enumeration (never roman numerals)

of headings and abbreviations.

Footnotes should be avoided unless they are short and essential.

Only generic and specific names should be underlined to indicate italics; all other marking

up should be left to editor and publisher.

4. ILLUSTRATIONS should be reducible to a size not exceeding 12 x 18 cm (19 cm including

legend); the reduction or enlargment required should be indicated; originals larger than

35 x 47 cm should not be submitted; photographs should be rectangular in shape and final

size. A metric scale should appear with all illustrations, otherwise magnification or reduction

should be given in the legend; if the latter, then the final reduction or enlargement should be

taken into consideration.

All illustrations, whether line drawings or photographs, should be termed figures (plates

are not printed; half-tones will appear in their proper place in the text) and numbered in a

single series. Items of composite figures should be designated by capital letters; lettering of

figures is not set in type and should be in lower-case letters.

The number of the figure should be lightly marked in pencil on the back of each illustration.

5. REFERENCES cited in text and synonymies should all be included in the list at the end of

the paper, using the Harvard System (ibid., idem, loc. cit., op. cit. are not acceptable):

(a) Author’s name and year of publication given in text, e.g.:

‘Smith (1969) describes...’

‘Smith (1969: 36, fig. 16) describes...’

‘As described (Smith 1969a, 1969b; Jones 1971)...”

‘As described (Haughton & Broom 1927)...’

‘As described (Haughton et al. 1927)...’

Note: no comma separating name and year

pagination indicated by colon, not p.

names of joint authors connected by ampersand

et al. in text for more than two joint authors, but names of all authors given in list of references.

(b) Full references at the end of the paper, arranged alphabetically by names, chronologically

within each name, with suffixes a, b, etc. to the year for more than one paper by the same

author in that year, e.g. Smith (1969a, 1969b) and not Smith (1969, 1969a).

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

For journal article give title of article, title of journal in italics (abbreviated according to the World list of

scientific periodicals. 4th ed. London: Butterworths, 1963), series in parentheses, volume number, part

number (only if independently paged) in parentheses, pagination (first and last pages of article).

Examples (note capitalization and punctuation)

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

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

FIscHER, P.-H., DuvAL, M. & Rarry, A. 1933. Etudes sur les échanges respiratoires des littorines. — Archs

Zool. exp. gén. 74: 627-634.

Konn, A. J. 1960a. Ecological notes on Conus (Mollusca: Gastropoda) in the Trincomalee region of Ceylon. —

Ann. Mag. nat. Hist. (13) 2: 309-320.

Konn, A. J. 19606. Spawning behaviour, egg masses and larval development in Conus from the Indian Ocean. —

Bull. Bingham oceanogr. Coll. 17 (4): 1-51.

THIELE, J. 1910. Mollusca: B. Polyplacophora, Gastropoda marina, Bivalvia. In: SCHULTZE, L. Zoologische

und anthropologische Ergebnisse einer Forschungsreise im westlichen und zentralen Siid-Afrika 4: 269-270.

Jena: Fischer. — Denkschr. med.-naturw. Ges. Jena 16: 269-270.

(continued inside back cover)

ANNALS OF THE SOUTH AFRICAN MUSEUM

ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

VOLUME 71 BAND

oy, ¢

5 S

rang

THE TRUSTEES OF THE DIE TRUSTEES VAN DIE

SOUTH AFRICAN MUSEUM SUID-AFRIKAANSE MUSEUM

CAPE TOWN KAAPSTAD

1976

The ANNALS OF THE SOUTH AFRICAN MUSEUM

are issued in parts at irregular intervals as material

becomes available

Obtainable from the South African Museum, P.O. Box 61, Cape Town 8000

Die ANNALE VAN DIE SUID-AFRIKAANSE MUSEUM

word uitgegee in dele op ongereelde tye na beskikbaarheid

van stof

Verkrygbaar van die Suid-Afrikaanse Museum, Posbus 61, Kaapstad 8000

OUT OF PRINT/ UIT DRUK

1, 2(1, 3, 5-8), 3(1-2, 4-5, t.-p.i.), 51-3, 5, 7-9),

6(1, t.—p.i.), 70-4), 8, 9-2), 101),

11(1-2, 5, 7, t.—-p.i.), 15(5), 24(2), 27, 31(1-3), 33

Price of this part/Prys van hierdie deel

R12,20

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

1976

ISBN 0 949940 92 5

Printed in South Africa by In Suid-Afrika gedruk deur

The Rustica Press, Pty., Ltd., Die Rustica-pers, Edms., Bpk.,

Court Road, Wynberg, Cape Courtweg, Wynberg, Kaap

PROCEEDINGS

OF THE

SOUTHERN AFRICAN SOCIETY

FOR

QUATERNARY RESEARCH

Conference on recent progress in later Cenozoic Studies

in southern Africa, 31 May — 3 June 1975

PREFACE

The most recent biennial conference of the Southern African Society for

Quaternary Research (SASQUA) was held at the South African Museum in

Cape Town from 31 May to 3 June 1975. This volume contains a series of

papers presented at the conference which itself was part of the one hundred

and fiftieth anniversary celebrations of the South African Museum. SASQUA

is grateful to the Board of Trustees, the Director, Dr T. H. Barry, and to

the staff of the museum for acting as hosts to the conference and for making

the more than eighty delegates most welcome. A further acknowledgement

is due for the invitation to publish the proceedings in the museum Annals. A

Cenozoic Study Group was established at the museum some years ago to

promote Palaeontological, Geological and Archaeological research in the

Quaternary and later Tertiary in the southern Cape and a number of papers —

in this volume reflect the direct or indirect stimulus given to Quaternary

research by the museum.

The conference was held over four days from a Saturday to a Tuesday.

To cope with the unpredictable Cape winter weather, meetings and excursions

were held on alternate days on the advice of the consulting geographer on

the organizing committee, Professor Barnard, but happily there was fine

weather throughout. The Saturday morning sessions were divided between

Geology A (sea-levels), chaired by Dr O. Davies with a lead address given by

Mr A. J. Tankard, and Geology B (terrestrial) chaired by Professor G. Soéhnge

with the lead address given by Professor A. O. Fuller. The Saturday after-

noon session was devoted to dating problems in the Cenozoic, the chairman

for the session being Dr J. C. Vogel and an address was given by Professor

A. Brock. The Sunday was given over to a field trip to the Walker Bay coast

near Gansbaai with visits to the Die Kelders and Byeneskrans archaeological

excavations, led by Mr F. R. Schweitzer.

On the Monday, Palaeoecology was the subject-matter of the first session

with the chairman Dr C. K. Brain, and the lead address given by Professor

E. M. van Zinderen Bakker. The second morning session was devoted to Anthro-

pology sensu lato with Dr H. J. Deacon as chairman and the lead address

given by Professor R. J. Mason. In the final meeting of the day, keynote

addresses were given by the distinguished guests of the Society, Professor

H. de Lumley of the University of Provence and Professor W. W. Bishop of

the University of London. The fourth and final day was spent visiting the

palaeontological site of Langebaanweg and geological exposures in the Saldanha

Bay area, the excursion being led by Dr Q. B. Hendey and Mr A. J. Tankard.

Thanks are due to the chairman of the sessions for guiding the discussions

which followed on the lead addresses, and on the papers that were allocated

to their sessions. As the papers were circulated before the meeting, individual

papers were not formally read and the chairman’s role in ensuring adequate

discussion was an important one. Thanks are also due to the leaders of the

Vi PREFACE

excursions which allowed for useful informal discussions and a glimpse of

on-going research. The conference itself was organized by a local committee

that included Professor W. S. Barnard, Dr H. J. Deacon, Dr Q. B. Hendey,

Mrs D. M. Leakey, Mr J. Rogers, Mr F. R. Schweitzer and Dr W. G. Siesser.

As secretary of the committee and as assistant editor of the proceedings, much

of the success of the conference is due to the untiring efficiency and hard work

of Mrs Leakey.

This volume is a fair reflection of the substance of the conference and the

papers are grouped in the same order as in the conference programme. Contri-

butions in French have been translated into English by Mrs Leakey for publi-

cation for which further acknowledgement is gratefully made. Mrs Ione Rudner

a delegate and member of the South African Museum staff and editor of the

Annals, has given invaluable guidance in editorial matters and has effectively

seen the papers to press. All delegates to the meeting will agree that if we are

indeed the product of the union between Paranthropus and Australopithecus

as was so well argued by Bill Bishop in his best alto to the acclaim of all present

then much good came of that union. SASQUA, as a body devoted to the study

of the setting and results of this lowly hominid love-match, has an important

contribution to make to our knowledge. These proceedings, the third in the

history of the Society, are part of this contribution.

H. J. Deacon

University of Stellenbosch

Honorary Editor of Proceedings

LIST OF CONTENTS

PREFACE

SECTION 1. GEOLOGY A. SEA-LEVELS

DAVIES, O.

The older coastal dunes in Natal and Zululand and their relation to former shore-

lines

FLEMMING, B. W.

Rocky Bank —evidence for a relict wave-cut platform

GIRESSE, P. see KOUYOUMONTZAKIS, G.

KOUYOUMONTZAKIS, G. & GIRESSE, P.

L’evolution a la fin du Pleistocene et a Il’Holocene du littoral Angolais de Lobito-

Benguela et Mossamedes

MONTAGGIONI, L.

Holocene submergence on Réunion Island (Indian Ocean) ..

Orme, A. R.

Late Pleistocene channels and Flandrian sediments beneath Natal estuaries

TANKARD, A. J.

Lead address. Cenozoic sea-level changes: a discussion

SECTION 2. GEOLOGY B. TERRESTRIAL

FuLtier, A. O.

Lead address. Correlated environmental parameters and their bearing on palaeo-

environmental reconstruction in incomplete stratigraphic sections

Hopspay, D. K.

Quaternary sedimentation and development of the lagoonal complex, Lake

St Lucia, Zululand .

MARKER, M. E.

Aeolianite: Australian and southern African deposits compared

VERHOEF, P.

Late quaternary colluvial deposits (progress report) ..

WEINERT, H. H.

Past climate derived from calcrete and N-value

SECTION 3. PALAEOECOLOGY

CoLLincs, G. E., CRUICKSHANK, A. R. I., MACGUIRE, J. M. & RANDALL, R. M.

Recent faunal studies at Makapansgat Limeworks, Transvaal, South Africa

CRUICKSHANK, A. R. I. see COLLINGS, G. E. ef al.

Hoower, D. A.

Phylogeny of the rhinocerotids of Africa

KLEIN, R. G.

The fossil history of Raphicerus H. Smith, 1827 Sogiirias meee in the

Cape biotic zone . ae ae

MaAcGurrE, J. M. see COLLINGS, G. E. et Bie

RANDALL, R. M. see COLLINGS, G. E. et al.

SCOTT, L.

Preliminary palynological results from the Alexandersfontein Basin near Kimberley

_ ZINDEREN BAKKER S8r., E. M. VAN

Lead address. Laté Quaternary environmental changes in southern Africa

Vil

Page

UW)

115

125

113)3)

167

169

193

141

Vill LIST OF CONTENTS

Page

SECTION 4. ARCHAEOLOGY

Brooker, M. see DEACON, H. J.

DEACON, H. J. & BROOKER, M.

The Holocene and Upper Pleistocene sequence in the southern Cape 3, a2 ~208

Mason, R. J.

Lead address. Analogy and archaeology ae ae at ae he oo» 201

Mason, R. J.

Exploration archaeology of the Kaokoveld and southern Angola and potential

australopithecene sites me oe as - oe aps a 2 eS

KEYNOTE ADDRESS

BisHop, W. W.

Comparison of australopithecine-bearing deposits in eastern and southern Africa

—a new look at a sixteen-year-old problem a ae a os pe 225

SECTION 1

GEOLOGY A. SEA-LEVELS

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION

ANTHONY J. TANKARD

South African Museum, Cape Town

(With 2 figures)

ABSTRACT

Recent studies of the history of existing ice sheets, sea-floor spreading, deep-sea cores,

isotopic composition changes of ocean waters and isostatic responses to varying loads, are

reviewed with regard to their bearing on sea-level changes. Only sea-levels of the past 125 000

years are discussed in any detail.

CONTENTS

mtroduction= i 2 Fe es Os 1

Glacio-eustatic control of sea-level . ce ks! D,

Sea-floor spreading and tectono-eustatic changes . 5}

Pleistocene high-level shorelines. . . . . . 6

Last interglacial shorelines . . 7

Sea-levels of the last Glacial Age

Inatenstadialiisea=levic lane nnen nnn se 8

Glacial maximum sea-level . . . . . . 9

ost-clacialisea=levelims an ee ee Se el

Conclusionsar sae, wee ee te Lee Uk 13

IRCICREMCCSaheaaeeucanuen as ee mali, Fok aha

INTRODUCTION

Palaeogeographic studies show that all the major coastlands of the world

have been subjected to alternating periods of submergence and emergence.

Old high-level shorelines and submerged terraces show that sea-level has

oscillated considerably in the past. In 1842 Maclaren introduced his concept

of glacial control to explain an oscillating sea-level during the Quaternary

Period. According to ‘classical’ theory a series of stepped marine terraces are

produced by a sea-level which fluctuates in response to waxing and waning of

continental ice sheets, and each successive high sea-level was lower than the

previous (Trowbridge 1954). To cite an example, transgression complexes on

the Namaqualand coast have been recognized at 75-90 m, 45-50 m, 29-34 m,

17-21 m, 7-8 m, 5 m, and 2 m, and these range in age from basal Pleistocene

to Recent (Carrington & Kensley 1969).

The assumption that the order of decreasing altitude of the elevated

shorelines corresponds with a decreasing age was confirmed by Flemming

(1968). He showed that an oscillating sea-level will not necessarily obliterate

previous erosion features, but that there is a general absence of random positions

of high-level shorelines. A descending chronological order implies that the

Shorelines are the result of either a series of sea-level oscillations of decreasing

Proc. sth. Afr. Soc. Quat. Res. 1975.

2 ANNALS OF THE SOUTH AFRICAN MUSEUM

magnitude, or the result of a series of oscillations of approximately equal

magnitude superimposed on an overall regression (Flemming 1968). The fact

that the Pleistocene glaciations were of nearly equal magnitude (Flint 1971)

would support the second option.

There have been several attempts in recent years to synthesize data on

eustatic sea-level changes (e.g. Fairbridge 1961; Guilcher 1969; Hey 1971).

As Hey points out, any attempt to interpret the field-evidence becomes an

exercise in correlation. South African literature on the subject of high-level

shorelines is generally of doubtful quality as there are no means of dating these

shorelines and reliance is placed on long-distance correlation with the ‘firmly

established’ Moroccan and other sequences. The horizontality of a shoreline

over considerable distances does not disprove the possibility of tectonism

(Cotton 1963; Bloom 1967), while the southern African coastlands, in par-

ticular, have a poor record of Cenozoic stability. Not only can we not separate

the climatic from the tectonic influences on sea-level, but it is also common

practice to ascribe high-level shorelines to named Late Cenozoic stages on the

mistakert assumption that the northern hemisphere climatic history is well

known and the stages adequately defined.

The recent publication of new evidence bearing on eustatic sea-level

studies makes a discussion opportune. The purpose of this paper will be to

discuss some of the more important factors and data dealing with eustatic

sea-level changes. Current research on the existing ice sheets, deep-sea cores

and oxygen isotope studies, sea floor spreading, and the bearing these have on

eustatic sea-level studies will be discussed. Particular attention will be paid to

glacio-eustatic sea-level oscillations of the last 125 ka (125 000 years) because

only for that period are there adequate and reliable data, and the effect of

ocean floor spreading would be expected to be minimal. In this paper the term

‘eustatic’ will apply to changes of sea-level relative to a fixed datum, say the

centre of the earth, and which are synchronous over the whole globe.

GLACIO-EUSTATIC CONTROL OF SEA-LEVEL

If high-level shorelines, higher than 30 m above present mean sea-level

(a.s.l.) for instance, were of glacio-eustatic origin it would be implied that the

existing continental ice sheets had melted and again built up several times in

the Quaternary. The history of these ice sheets show that this has not

happened.

Although there is little evidence for the existence of large continental

ice sheets prior to the Middle Miocene, some evidence suggests the presence of

calving glaciers in the Eocene (Denton ef al. 1971). The first major cooling is

Closely associated with the Eocene—Oligocene boundary (Devereux 1967;

Kennett et al. 1974), and although Oligocene temperatures on Antarctica were

close to freezing there was no development of an extensive ice sheet (Shackleton

& Kennett 1974). But Margolis & Kennett (1970) document evidence of some

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 3

glaciation during the Oligocene. Evidence of glaciation prior to 22 m.y. B.P.

is provided by radiometrically dated basalts overlying a Tertiary tillite (Craddock

et al. 1964; Rutford et al. 1968). It seems highly likely that the glaciers were

only mountain or valley glaciers. A rich and varied fossil flora suggests not

only that extensive ice sheets had not developed in the Lower Tertiary, but

also that until at least the Early Miocene the climate was generally temperate

(Denton et al. 1971).

Analysis of cores from the floor of the Ross Sea (DSDP leg 28) suggests

extensive glaciation on eastern Antarctica at least by the Early Miocene (Hayes

et al. 1973). Shackleton & Kennett (1974) (DSDP leg 29) have found that the

Antarctic continental ice sheet developed to present thickness between the

early Middle Miocene and early Late Miocene, and that by the Late Miocene—

Early Pliocene the Antarctic ice sheet was much more extensive than at present.

This development was accompanied by a major regression. The maximum of

ice accumulation 4 to 5 m.y. ago was followed by an abrupt melting and ice

retreat to the present position (Hayes et al. 1973). Subsequent fluctuations in

the Antarctic ice cover have been minor. Studies of deep-sea cores show that

Antarctic glaciation has been continuous for the last 3 to 5 m.y. (Koster 1966;

Goodell et al. 1968).

It has been suggested by Savin et al. (1975) that until the beginning of the

Middle Miocene high- and low-latitude temperatures changed in similar fashion.

They record a lowering of temperature from the Late Eocene until the Late

Oligocene, followed by a rise in temperature through the Early Miocene. Onset

of major glaciation and development of the circum-Antarctic oceanic circula-

tion in the Middle Miocene led to a sudden divergence of high- and low-latitude

marine temperatures. —

On radiolarian evidence Fillon (1973) has shown that for the last 3 m.y.

surface water temperatures south of 65°S were never warmer by more than

about 3°C than they are today. Mercer (1968a) has found that at some time

duting the Pleistocene Antarctic temperatures could have been 7 to 10°C

higher than today, but he notes that the present —10°C January isotherm lies

close to the coast. A low amplitude temperature fluctuation could not lead to

significant ice melting.

In the northern hemisphere full glacial conditions have prevailed in Alaska

since the Late Miocene (Miller 1953; Bandy et al. 1969; Denton & Armstrong

1969). In a study of several cores from the Arctic Ocean Clark (1971, 1974)

showed that since at least the Early Pliocene the Arctic Ocean has been con-

tinuously ice covered and that the present thickness of the ice is a minimum.

He found that the Arctic Ocean was ice-free until the Eocene, but that the

ice sheet must have formed some time between the Eocene and Pliocene.

Shackleton & Kennett (1974) suggest a Late Pliocene development of northern

hemisphere glaciation.

Melting of the existing ice sheets (east and west Antarctica, Greenland,

and others) would cause sea-level to rise 65 m (Flint 1971). But there is over-

4 ANNALS OF THE SOUTH AFRICAN MUSEUM

whelming evidence that the land-based eastern Antarctic ice sheet was stable

throughout the Pleistocene and could have contributed little to sea-level move-

ment. The Antarctic ice sheet possibly underwent periodic surges into the

Southern Ocean. but there is little evidence to support this (Denton et al. 1971).

Data from deep-sea cores argue against surges, but do not eliminate the possi-

bility of smaller surges.

Mercer (19685, 1973) has found that the marine ice sheet in western Ant-

arctica is more vulnerable to climatic change and its melting would have caused

a eustatic rise in sea-level of about 5 m. Isotopic studies of the Greenland ice

(Dansgaard et al. 1969) suggest that only 15 per cent of that ice is residual from

the last glacial and preceding interglacial age. This led Emiliani (1969) to suggest

that high Late Pleistocene sea-levels resulted from melting of this Greenland

ice Sheet and that melting would cause a eustatic rise in sea-level of about 10 m.

Taking a conservative view, that both the Greenland and western Antarctic

ice sheets contributed to Pleistocene eustatic sea-level movement while the

large eastern Antarctic ice sheet remained stable, the maximum glacio-eustatic

rise of sea-level could not have exceeded 15 m at any stage in the Pleistocene.

These conclusions are broadly supported by oxygen isotope studies of

deep-sea cores. Accumulation of the vast isotopically negative Pleistocene ice

sheets resulted in a lowering of sea-level. At the same time the oceans became

isotopically positive and more saline. The original aim of oxygen isotope

studies was to measure palaeotemperature (e.g. Emiliani 1955, 1966). Shackleton

& Opdyke (1973) have published a record of ocean isotopic composition changes

for the last 800 ka. They note that such a record is of greater stratigraphic

value than a record of temperature change.

Based on a rough equivalent of 0,1%, isotopic deviation to 10 m sea-level

change, Shackleton & Opdyke (1973) have drawn a glacio-eustatic sea-level

curve for the past 130 ka. The remarkable agreement between positions of sea-

level shown on this curve and those measured on Barbados (Broecker ef al. 1968)

and New Guinea (Veeh & Chappell 1970) proves the correctness of the method.

It thus becomes possible to use the record of oxygen isotopic composition in core

V 28-238 (Shackleton & Opdyke 1973: fig. 9) as a sea-level curve for the past

800 ka. This record shows that sea-level during isotopic substage Se (+6 m)

was the highest in the last 800 ka, and that only twice in this interval did sea-

level exceed present sea-level, at about 400 ka and again at about 320 ka.

In conclusion, it must be emphasized that a sea-level curve derived from

oxygen isotope measurements gives a reasonably clear record of glacio-eustatic

sea-level changes without the complicating effects of tectono-eustasism or

isostatic adjustments of the globe to changing loads. The history of the existing

ice sheets shows that the glacio-eustatic component could not have caused a

sea-level rise much above present sea-level in the Pleistocene. The major Ant-

arctic ice sheet appears to have existed in its present form since the Pliocene.

It is unlikely that the history of these ice sheets could account for the major

Late Cretaceous and Tertiary (Eocene and Miocene) transgressions.

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 5)

SEA-FLOOR SPREADING AND TECTONO-EUSTATIC CHANGES

The new theory of global tectonics has transformed numerous branches

of geological research, and the field of eustatic sea-level change has not emerged

unscathed. By this concept crustal plates grow along the mid-ocean ridges and

sink or override each other where they again come into contact. Changes in

the rate of accretion at mid-ocean ridges will lead to changes in the depth of

the oceans, since elevation or subsidence of sea-floor is a function of ridge

activity. The idea that ridge activity may affect sea-level has been postulated

by Hallam (1963, 1971), Russell (1968), Menard (1969), Frerichs & Shive

(1971), Flemming & Roberts (1973) and Vine (1973). Flemming & Roberts

have attempted to correlate eustatic changes with contemporaneous global

spreading discontinuities, and can explain the major Late Cretaceous, Eocene

and Miocene transgressions in this way. Hallam (1971) and Hays & Pitman

(1973) believe that the Late Cretaceous transgression was due to oceanic-floor

uplift consequent upon accelerated sea-floor spreading. The Miocene eustatic

rise due to spreading rate changes could be of the order of hundreds of metres

(Flemming & Roberts 1973). If there had been no ridge activity in the Cenozoic,

and allowing for isostatic adjustment, sea-level would today be 350 m lower

than in fact it is (Vine 1973).

According to Jacoby (1972) the effect of plate movement is to create an

environment of changing density in which continental blocks of constant

density will float up or down. For instance, in an environment of greater density

caused by cooling of the upper mantle the continental blocks would float

upward, causing a regression. Conversely, decreasing density would result in

transgression.

Rona (1973) has recorded average rates of sediment accumulation from

wells on the continental shelves and slopes. He found maxima of sediment

accumulation, corresponding to major transgressions, in the Middle Eocene

and Miocene. Each maximum is separated by a minimum which corresponds to

regressive phases. Rona equates the transgressions with a volume increase of

the mid-ocean ridges resulting from fast-spreading and orogenic quiescence of

continents. Regression corresponds to decrease of mid-ocean ridge volume

with slower spreading and orogenic activity of the continents.

Increased ridge activity in the Eocene and Miocene would adequately

account for marine transgressive complexes of those ages around the southern

African coast. Tertiary marine sediments are preserved in depressions in the

Precambrian basement south of Liideritz. Haughton (1963) suggested that

most of the fossil fauna was Miocene, and that Eocene deposits were preserved

at higher elevations. In situ Middle Eocene marine sediments also occur in

Mozambique (Du Toit 1954). Along the South African coast the only evidence

for an Eocene transgression consists of reworked deposits at Uloa (Frankel

1968) and Birbury (Bourdon & Magnier 1969). Otherwise the major trans-

gressive complex deposits in South Africa are of Miocene age (King 1953;

Frankel 1968; Ruddock 1968; Tankard 1975).

6 ANNALS OF THE SOUTH AFRICAN MUSEUM

These authors also describe Pliocene transgressive complexes. The west

coast Pliocene deposits can be attributed entirely to tilting and there is no need

to invoke a eustatic sea-level rise. Furthermore, Frankel (1968) disputes that

there are Pliocene marine sediments along the Zululand coastal plain.

Hallam (1973) suggests that due to continued subsidence of the ocean

floors, sea-level was generally regressive through the Pliocene and Pleistocene

following the Miocene peak. A change of elevation of the ocean floors implies

a subcrustal transfer of mantle material. Depression of the ocean floors would

transfer mantle material towards the continental blocks and would cause them

to float upwards. But there would be a transitional zone of flexure where the

rising continental block was coupled to the downwarping ocean floor. Mention

has already been made of the maximum of ice accumulation 4 to 5 million

years ago in the Antarctic which was followed by abrupt melting and ice retreat

to the present position. The time lag between this melting and attainment of

hydro-isostatic equilibrium could have been marked by a Late Pliocene trans-

gression, albeit minor compared with the Miocene transgression.

PLEISTOCENE HIGH-LEVEL SHORELINES

Assuming a uniform regression from the Miocene eustatic high, +300 m

according to Flemming & Roberts (1973), the basal Pleistocene sea surface

would have been 50-60 m above present. This general regression, the result of

downwarping of the ocean floors, would probably be accompanied by uplift

of the continental blocks and continued flexuring of the margins. The increased

Late Pliocene melt water suggests a basal Pleistocene sea-level even higher than

50-60 m before hydro-isostatic equilibrium was achieved. But this is all specula-

tion since little is known about the course of events following a Late Pliocene

melting of Antarctic ice and since our knowledge of sea-floor spreading is far

from complete. Bloom (1971) warns that it ‘would be foolhardy to infer any-

thing about glacial-eustatic control of sea level during one of the early Pleisto-

cene glaciations, for instance, in the face of evidence that the ocean basins are

widening at rates of up to 16 cm per year...’.

To obtain sets of discrete high-level shorelines such as have been described

for the South African coast (Carrington & Kensley 1969; Davies 1970-1973,

and others) one could speculate that it would be possible to superimpose a

glacio-eustatic curve derived from oxygen isotope composition changes in the

Pleistocene on a sea-level regressing due to subsidence of the ocean floors.

It is doubtful, however, whether this would produce meaningful results. Shackle-

ton & Opdyke (1973) record a sea-level at stage 9 (approximately 560 ka) near

present sea-level. To raise this shoreline to +30 m either by uniform rate of

uplift of the coastal area, or tectono-eustatic effects, would ensure that the

120 ka shoreline (+6 m) would today be recorded at +15 m. Davies (1970)

suggests that the Natal 60 m shoreline is of Cromerian age. The ‘Cromerian

Complex’ probably extends from 350 ka back into the Matuyama Epoch

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 7

(Shackleton & Opdyke 1973). Furthermore, Davies (1970, 1973) claims that

the 30 m and 60 m beaches contain Acheulian artefacts, but the lower limit of

Acheulian time is about 700 ka according to Klein (1974). The oxygen isotopic

composition record shows a peak at about that time, stage 19, close to the

present sea level. At a uniform rate of uplift to elevate such a shoreline to

+60 m, the 120 ka shoreline would be found today at about +16 m. What is

believed to be the 120 ka shoreline has been identified in Zululand at +8 m

(Hobday 1976), and in the Saldanha area at +6,3 m (Tankard 1976). This

type of reasoning suggests that the 30 m, 60 m, and higher shorelines, could

be no younger than the Early Pleistocene on a ‘stable’ coast.

LAST INTERGLACIAL SHORELINES

In the Late Quaternary a more realistic appraisal of sea-level fluctuation is

possible because:

1. The effect of sea-floor subsidence due to spreading changes would be

minimal.

2. Elevations of the shorelines can be explained by changes in the ice

sheets.

3. Adequate data is available. In this discussion the last interglacial will

be taken as equivalent to isotope stage 5, a time range 128 to 73 ka (Suggate

1974).

Emiliani (1961), Shackleton (1969), Shackleton & Opdyke (1973), and

Emiliani & Shackleton (1974) have analysed deep-sea cores spanning the last

500 to 800 ka. In all of these studies substage Se of the last interglacial registers

a higher palaeotemperature than the maximum post-glacial (stage 1) or any

earlier stage. Substage 5e corresponds to a sea-level maximum at 120 ka.

Shackleton & Opdyke have derived, from oxygen isotope measurements, a

glacio-eustatic sea-level curve for the past 130 ka. The curve shows four maxima

which have been compared with estimated sea-levels on Barbados (Broecker

et al. 1968) and New Guinea (Veeh & Chappell 1970). Only the substage 5e

sea-level (Barbados III) at 120 ka is higher than the present sea-level. There are

also two other maxima in the last interglacial, both 10 to 20 m lower than

present sea-level: BII at 100 ka and BI at 80 ka.

Uranium series dating of corals from shorelines between 1,5 and 9 m above

present datum from various parts of the Indian and Pacific Oceans yield dates

of the order of 120 ka (Veeh 1966). Undoubtedly the most detailed and most

reliable results in recent years have come from Barbados and New Guinea.

Both these islands have been uplifted at a uniform rate. On Barbados Broecker

et al. (1968) have recognized three last interglacial sea levels at elevations

+6 m (BIT), —13 m (BIJ), and —13 m (BI), which have been dated at 122 ka,

103 ka, 82 ka respectively. A similar sea-level curve based on a series of trans-

gressive-regressive cycles identified from a series of coral reefs on New Guinea

independently confirms each high sea-level stand on Barbados (Veeh & Chappell

8 ANNALS OF THE SOUTH AFRICAN MUSEUM

1970). The evidence from Barbados and New Guinea for two —13 m peaks at

80 ka and 100 ka explains why these levels are seldom recognized from conti-

nental coastlands.

On Mallorca Butzer & Cuerda (1962) recorded Tyrrhenian II shorelines at

12,5 m and 7,2 m with a 120 ka age, and a Tyrrhenian III shoreline at 2,2 m

with an age of 80 ka. Molluscs from cemented terrace deposits from the western

Mediterranean and Moroccan coasts are suggestive of sea-level stillstands in

those areas at 120 ka and 80 ka.

The shoreline between 2 and 7 m above present sea-level which is widely

recognized in places remote from plate boundaries, and which formed at

120-130 ka, is becoming commonly used as a Late Pleistocene sea-level datum

(Chappell 1974). If this shoreline is situated higher than --7 m it would infer

tectonic uplift (Chappell 1974). Hobday (1976) has recognized three last inter-

glacial shorelines from the St. Lucia area of Zululand: 8 m, 3,4-5,3 m, and

4,5 m. Three last interglacial shorelines have also been recognized in the Sal-

danha area of the south-western Cape: 6,3 m, 2—3,5 m, and 0 m (Tankard 1976).

In both of these cases the highest beaches would agree with the so-called Late

Pleistocene datum and should therefore be unaffected by displacements such

as hydroisostatic adjustments. It therefore becomes difficult to reconcile’ the

two low levels with the BII and BI levels, although they agree well with the

Mallorca series described by Butzer & Cuerda (1962). Perhaps all three of

these last interglacial shorelines from South Africa should be equated with BIII.

SEA-LEVELS OF THE LAST GLACIAL AGE

Interstadial sea-level

In North America Milliman & Emery (1968) cited 15 radiocarbon dates on

carbonate samples to predict an interstadial sea-level as high as the present at

35 ka. Numerous other authors have followed them (see Thom 1973 for a full

discussion). However, a considerable amount of evidence has been published

which shows that such a high interstadial sea-level is unlikely. This subject has

been treated in great detail by Thom (1973), and only a few comments will be

given here.

A glacio-eustatic sea-level curve derived from oxygen isotope measurements

on deep-sea core V28—238 shows sea-level to have been considerably lower than

the present level during the interstadial (Shackleton & Opdyke 1973: fig. 7).

This curve agrees with the New Guinea data (Veeh & Chappell 1970) although

it was derived from independent lines of reasoning, and must therefore be

basically correct. The NG III data shows a corrected shoreline at —20 m at

35 ka. According to Broecker & Van Donk (1970) sea-level could not have

been closer than 18 m from present sea-level between 35 and 45 ka.

Glacio-climatic evidence proves the impossibility of sea-level being close

to present sea-level during the interstadial (M6rner 1971). The last glaciation

did not possess a warm interval of comparable intensity or duration to that

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 9

which exists at present, or prior to 60 ka and oxygen isotope measurements on

Greenland ice cores failed to show values equivalent to an interglacial or to

present conditions (Thom 1973). Continental ice fronts during the interstadial

lay in south-western Sweden and at the line of the Great Lakes in North America

(Fairbridge 1971).

Fairbridge has recorded discussions at the eighth INQUA congress. It was

stated that a high interstadial sea-level to +3 to 5 m relies on dated carbonate

material. Studies carried out by the Columbia University radiocarbon labora-

tory show that all shell dates older than 20 ka are subject to miniscule con-

tamination which gives meaningless ‘dates’ and which should be reported as

‘greater than. . .’. Fairbridge concludes that any ‘postulated glacio-eustatic

level for this epoch anywhere near present sea-level is absolutely out of the

question’.

Figure 1 is a South African time-depth plot of sea-level over the last 47 ka

based on 20 radiocarbon dates derived from wood, peat, or carbonate samples

from below sea-level, or very deeply buried, so that the chance of contamination

by atmospheric CO, is thus minimized. These samples are all identified with

shallow nearshore environments, but are unfortunately representative of

widely scattered localities from west of the Cape Peninsula to Zululand. A line

is drawn through the points to draw the eye, and should be regarded only as a

crude approximation to sea-level history over the past 47 ka.

The curve suggests that between 47 ka and 25 ka there was a sea-level

maximum at approximately —20 m. A lengthy stillstand at this level would

be expected to leave topographic evidence. Rocky Bank, south of the Cape

Peninsula, was most likely formed by a sea-level at this elevation (Flemming

1976) as was a prominent submerged cliff at —20 m along parts of the west

coast of the Cape Province and South West Africa (Wright 1964; Murray et al.

1970).

Glacial maximum sea-level

Evidence for low sea-levels between 15 and 20 ka, during the glacial maxi-

mum, have recently been discussed by Chappell (1974). Dated shallow marine

deposits vary in depth from —65 m to —150 m. The most consistent results

are from the Texas continental shelf. There a sea-level lowering of 130 m at

15 ka compares with a lowering of 90 m at 17 ka for the eastern continental

shelf and 130-170 m for Australia (Chappell 1974). Using Walcott’s (1972)

corrections for elastic warping, Chappell calculates a minimum sea-level of

—135 m. As he points out this agrees very favourably with Flint’s (1971)

estimate of —130 m calculated from ice volumes. Oxygen isotope results

Suggest a lowering to —120 m (Shackleton & Opdyke 1973).

The South African time-depth curve (Fig. 1) suggests a rapid fall of sea-

level with advance of the final Wiirm glaciation and that a minimum sea-level

of —130 m was reached at 17-18 ka.

ANNALS OF THE SOUTH AFRICAN MUSEUM

"POLY YINOG JOJ VY Lp ISL] IOJ SAIN JOAS]-V9S OANL[OY *] “SIA

(ey) aby

OS G7 07 GE O€ GZ 02 Gl

(91914) EZ61 PHTEYIS O

(siouel4 3S adeg JOMS) [16] SIBJeW +]2D0\ V

(TEN) ~BO9BL PNeW x

(SSeUJaPIIM) ZO9G| UlMeW O

OS|

yjdeq

(Ww)

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 1]

POST-GLACIAL SEA-LEVEL

In this section the author proposes to discuss the changes in sea-level

during the Holocene, and mainly the past 6 000 years. Widely divergent opinions

about the course of sea-level rise persist because little account is taken of the

isostatic responses of the earth crust to changing ice and water loads. The

hydro-isostatic effects will vary with location according to continental shelf

geometry and structure of the underlying mantle. A prerequisite in any Holo-

cene shoreline study is the accurate identification of the 125 ka datum

(+5 +3 m) in that area (Thom & Chappell 1975).

Following the maximum glacial advance, the rate of retreat of the ice was

nearly constant, and the present extent was reached at about 6,5 ka (Bloom

1971: fig. 4). Bloom’s curve is reproduced in Figure 2 along with several esti-

Age (x 1 000 years)

oe F~ee

=10

2 20

ree

gy c

wo =|]

H- 40 20 =

5 =

@ 8

= 80 Ss

40 w

a 2

= 60 ©

= &

(@)

seeeeee Godwin et al 1958 5 ee

----- Fairbridge 1961

-.-..—- Shepard 1961

—— Emery 1969

------- Bloom 1971 (ice sheet area) \ 60

—— Scholl 8 Stuiver (1967)

= 8}0)

Fig. 2. Holocene sea-level rise, showing different concepts.

UD ANNALS OF THE SOUTH AFRICAN MUSEUM

mates of sea-level rise. Generally there appears to have been a rapid rise of

‘sea-level, but the rate has decreased with diminishing age.

Ignoring the isostatic responses of the crust to changing loads, there is a

threefold division of opinion on the course of sea-level rise (discussed in detail

by Jelgersma 1971). These are (Fig. 2):

1. The oscillating sea-level concept (Fairbridge 1961). Fairbridge finds

evidence, mainly from Australia, for postglacial sea-levels at 3-5 m a.s.l. (5 ka),

1,5-2 m a.s.l. (3,7 ka), and 0,6-1,0 m a.s.l. (2,3 ka).

2. The steady sea-level concept (Godwin et al. 1958) suggests sea-level

rose rapidly until reaching the present position at 5,5 ka. Sea-level supposedly

remained constant since.

3. Finally, there is the continuously rising sea-level concept (Shepard

1961, 1963). According to this concept the postglacial rise of sea-level has been

asymptotic and for the last 5 ka sea-level has been rising continuously. Shepard’s

data indicates a rapid rise from 17 to 6 ka when sea-level stood at —6 m. This

was followed by a slow rise to —1,8 m between 3,5 and 2 ka.

A very detailed work is that of Scholl & Stuiver (1967) on the stable Ever-

glades coast of southern Florida. They see a rapid rise of sea-level to —1 m at

3 ka, since when sea-level has risen slowly.

Scholl & Stuiver could find no evidence in the Florida region in support

of high Holocene sea-levels as postulated by Fairbridge (1961). They pointed

out that a slight rise in sea-level in the Everglades region would have left a very

substantial record. Furthermore, they question the significance of radiocarbon

dates from the Everglades that were used by Fairbridge. Shepard (1963,

1964), Russell (1963), Shepard & Curray (1965), Hails (1965) and Thom et al.

(1969) are rather sceptical of the significance of Australian radiocarbon dates

used by Fairbridge (1961), Gill (1961) and Ward (1965). Thom e¢ al. (1969)

have found no morphological or stratigraphical evidence for sea-levels higher

than the present in eastern Australia, which supposedly has a stable coastline,

between 2 985 and 9 000 B.P. Although there was a post-glacial climatic opti-

mum between 4 000 and 2 000 B.C., Lamb (1966) is doubtful that there could

have been a corresponding rise in sea-level because continental ice-sheets were

still too extensive. The Brenner Pass, for example, may not have been open

before 1 800 B.C.

Recent studies (e.g. Bloom 1967, 1971; Walcott 1972; Chappell 1974)

demonstrate that the continental margins are not necessarily stationary, but

that changing ice and water loads induce displacement of the earth’s surface

which may take the form of rapid elastic adjustments and slow viscous mantle

flow (Chappell 1974). The course of sea-level rise relative to the continental

margins varies with location, and is affected by shelf geometry and physical

structure of the underlying mantle (Thom & Chappell 1975). Bloom (1971)

argues that, because the shelf off the Florida Everglades is very shallow, down-

warping due to hydro-isostatic compensation would be minimal, and that the

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 13

curve drawn by Scholl & Stuiver (1967) would be a realistic glacio-eustatic

curve. Walcott’s (1972, fig. 1) map of elastic warping strengthens this argument

since the Florida Everglades lie on his 100 per cent contour, i.e. no vertical

displacement due to elastic deformation of the earth. Florida is also remote

from plate boundaries. According to Walcott (1972) no substantial change of

sea-level is necessary to explain sea-level of the last 6 000 years.

Deglaciation and the consequent sea-level rise following the —130 m low

at about 17 ka involved on average depression of the ocean floors of about 8 m,

and an average upward movement of the continents of about 16 m (area of the

continents half that of the oceans) in the last 7 ka (Chappell 1974). Furthermore,

there would be a zone of flexuring where ocean basins and continents meet.

These effects differ with locality. If there is a vertical movement of the continent

and flexuring of their margins relative to the ocean floors, then the most desirable

place to measure glacio-eustatic change would be on mid-oceanic islands which

follow the movement of the ocean floors (the ‘dip-stick’ effect).

Data from eastern Australia and Christchurch, New Zealand, both stable

areas, allowed Chappell (1974) to draw a continental coast sea-level curve.

This curve shows present sea-level to have been attained by 6 000 B.P. and to

have remained static relative to the continent since then. Recalculation of the

sea-level data from the continental curve to give movement of sea-level relative

to ocean basins shows that sea-level has been rising continuously throughout

that period, and that the course of sea-level rise relative to the ocean basins

agrees with the results from midoceanic islands and the Florida Everglades.

The South African sea-level curve (Fig. 2) agrees with the Queensland—

New South Wales results (Thom & Chappell 1975), although many more radio-

carbon dates are required to substantiate its correctness. After the eustatic low

at 17 ka, sea-level initially rose very rapidly with deglaciation (167 cm per

100 years). At 9 000 B.P. sea-level, relative to the continental coast, stood at

—25 m, and was within a metre of present sea-level between 5 000 and 6 000

B.P. Detailed work in Zululand (Hobday 1976) and the Saldanha area (Tankard

1976) has not revealed much evidence of Holocene shorelines higher than

present sea-level. Furthermore, identification of the Late Quaternary datum

in these areas at 8 m and 6 m respectively shows comparative stability during

this period.

CONCLUSIONS

Pleistocene sea-level research is probably on the verge of a major revolution

and studies such as that of Shackleton & Opdyke (1973) on isotopic composi-

tional changes in the oceans should soon give us a clear account of glacio-

eustasism through the Neogene and Pleistocene. Already this record, and the

history of the existing ice sheets, show that glacio-eustatic sea-level could not

have been much higher than present datum at any time in the Pleistocene. It

will require a basic change in philosophy if the record of ocean isotopic com-

14 ANNALS OF THE SOUTH AFRICAN MUSEUM

positions is to be accepted, and a reaction against these ideas similar to that

which greeted the radiocarbon dating technique can be expected. In addition,

it is clear that information on the effect of sea-floor spreading and tectono-

eustasism in general is required.

Recent quantitative studies (e.g. Walcott 1972; Chappell 1974) show that

the earth’s crust responds isostatically to changing ice and water loads, and that

the best place to observe the movement of sea-level relative to the ocean basins

is on the midoceanic islands, which are remote from plate boundaries, and move

with the ocean floor in its response to changing water loads.

Studies on Holocene shorelines along the continental margins take on a

new dimension because of these recent trends. No longer is there need for

absolute change in sea-level to explain shorelines of the last 6000 years. By

measuring differences in the Holocene record with locality on the continental

margins, and comparing these with data from the midoceanic islands, for

instance, we should have a means of examining the structure of the underlying

mantle. Holocene shoreline studies will, in future, demand a more sophisticated

understanding of sedimentary, marine and crustal processes than has been the

case in the past in South Africa and elsewhere.

REFERENCES

BANpDy, O. L., BUTLER, E. A. & WRIGHT, R. C. 1969. Alaskan upper Miocene marine glacial

deposits and the Turborotalia pachyderma datum plane.— Science 166: 607-609.

BLoom, A. L. 1967. Pleistocene shorelines: a new test of isostasy.— Bull. geol. Soc. Am. 78:

1477-1494.

Bioom. A. L. 1971. Glacial-eustatic and isostatic controls of sea level since the last glaciation.

In: TUREKIAN, K. K. ed. Late Cenozoic glacial ages: 355-379. New Haven: Yale University

Press.

BourDON, M. & MAGNIER, P. 1969. Notes on the Tertiary fossils at Birbury, Cape Province.

—Trans. geol. Soc. S. Afr. 72: 123-125.

BROECKER, W. S., THURBER, D. L., GODDARD, J., Ku, T-L., MATTHEWS, R. K. & MESOLELLA,

K. J. 1968. Milankovitch hypothesis supported by precise dating of coral reefs and deep

sea sediments.— Science 159: 297-300.

BROECKER, W. S. & VAN Donk, J. 1970. Insolation changes, ice volumes, and the 01° record

in deep-sea cores.— Rev. Geophys. Space Phys. 8: 169-198.

BuTzer, K. W. & CUERDA, J. 1962. Coastal stratigraphy of southern Mallorca and its implica-

tions for the Pleistocene chronology of the Mediterranean sea.—J. Geol. 70: 398-416.

CARRINGTON, A. J. & KENSLEY, B. F. 1969. Pleistocene molluscs from the Namaqualand coast.

—Ann. S. Afr. Mus. 52: 189-223.

CHAPPELL, J. 1974. Late Quaternary glacio- and hydro-isostasy, on a layered earth. — Quatern.

Res. 4: 405-428.

CriarK, D. L. 1971. Arctic Ocean ice cover and its late Cenozoic history. — Bull. geol. Soc. Am.

82: 3313-3324.

CLARK, D. L. 1974. Late Mesozoic and early Cenozoic sediment cores from the Arctic Ocean.

—Geology 2: 41-44.

Cotton, C. A. 1963. The question of high Pleistocene shorelines.— Trans. R. Soc. N.Z. 2:

51-62.

CRADDOCK, C., BASTIEN, T. W. & RUTFORD, R. H. 1964. Geology of the Jones Mountains

area. In: ApiE, R. J. ed. Antarctic Geology: 171-187. Amsterdam: Inter-Science.

DANSGAARD, W., JOHNSEN, S. J., MOLLER, J. & LANGWay, C. C. Jr. 1969. One thousand

centuries of climatic record from Camp Century on the Greenland ice sheet.— Science

166: 377-381.

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 15

Davies, O. 1970. Pleistocene beaches of Natal—Ann. Natal Mus. 20: 403-442.

Davies, O. 1971. Pleistocene shorelines in the southern and south-eastern Cape Province

(Part 1).—Ann. Natal Mus. 21: 183-223.

Davies, O. 1972. Pleistocene shorelines in the southern and south-eastern Cape Province

(Part 2).—Ann. Natal Mus. 21: 225-279.

Davies, O. 1973. Pleistocene shorelines in the western Cape and South-West Africa.— Ann.

Natal Mus. 21: 719-765.

DENTON, G. H. & ARMSTRONG, R. L. 1969. Miocene—Pliocene glaciations in Southern Alaska.

—Am. J. Sci. 267: 1121-1142.

DENTON, G. H., ARMSTRONG, R. L. & STUIVER, M. 1971. The Late Cenozoic glacial history

of Antarctica. In: TUREKIAN, K. K. ed. Late Cenozoic glacial ages: 267-306. New Haven:

Yale University Press.

DEVEREUX, I. 1967. Oxygen isotope paleotemperature measurements on New Zealand Tertiary

fossils. —N.Z. Jl. Sci. 10: 988-1011.

Du Toit, A. L. 1954. Geology of South Africa. 3rd ed. S. H. HAUGHTON, ed. Edinburgh:

Oliver & Boyd.

EMILIANI, C. 1955. Pleistocene temperatures. —J. Geol. 63: 149-158.

EMILIANI, C. 1961. Cenozoic climatic changes as indicated by the stratigraphy and chronology

of deep-sea cores of Globigerina facies.— Ann. N.Y. Acad. Sci. 95: 521-536.

EMILIANI, C. 1966. Palaeotemperature analysis of Caribbean cores P 6304-8 and P 6304-9

and a generalised temperature curve for the last 425,000 years.—J. Geol. 74: 109-126.

EMILIANI, C. 1969. Interglacial high sea levels and the control of Greenland ice by the pre-

cession of the equinoxes.— Science 166: 1503-1504.

EMILIANI, C. & SHACKLETON, N. J. 1974. The Brunhes Epoch: isotopic paleotemperatures and

geochronology. — Science 183: 511-514.

FAIRBRIDGE, R. W. 1961. Eustatic changes in sea-level. Jn: AHRENS, L. H. et al. eds. Physics

and chemistry of the earth 4: 99-185. London: Pergamon Press.

FAIRBRIDGE, R. W. 1971. Quaternary shoreline problems at Inqua, 1969.—Quaternaria 15:

1-17.

FILLon, R. H. 1973. Radiolarian evidence of late Cenozoic oceanic paleotemperatures, Ross

Sea, Antarctica. —Palaeogeogr. Palaeoclim. Palaeoecol. 14: 171-185.

FLEMMING, B. W. 1976. Rocky Bank—evidence for a relict wave-cut platform.—Ann. S.

Afr. Mus. 71: 33-48.

FLEMMING, N. C. 1968. Derivation of Pleistocene marine chronology from morphometry of

erosion profiles.—J. Geol. 76: 280-296.

FLEMMING, N. C. & Roperts, D. G. 1973. Tectono-eustatic changes in sea level and seafloor

spreading. — Nature, Lond. 243: 19-22.

FLINT, R. F. 1971. Glacial and Quaternary geology. New York: Wiley & Sons.

FRANKEL, J. J. 1968. Tertiary sediments in the lower Umfolozi River Valley, Zululand. —

Trans. geol. Soc. S. Afr. 71: 135-145.

FRERICHS, W. E. & SHIVE, P. N. 1971. Tectonic implications of variations in sea floor spreading

rates.— Earth Planet. Sci. Lett. 12: 406-410.

GILL, E. D. 1961. Changes in the level of the sea relative to the land in Australia during the

Quaternary Era.—Z. Geomorph. Suppl. 3: 73-79.

GopwiIn, H., SuGGATE, R. P. & WILLIS, E. H. 1958. Radiocarbon dating of the eustatic rise

in ocean-level.— Nature, Lond. 181: 1518-1519.

GOODELL, H. G., WATKINS, N. D., MATHER, T. T. & Koster, S. 1968. The Antarctic glacial

history recorded in sediments of the Southern Ocean.— Palaeogeogr. Palaeoclim. Palaeo-

ecol. 5: 41-62.

GUILCHER, A. 1969. Pleistocene and Holocene sea-level changes.— Earth Sci. Rey. 5: 69-97.

HAIzs, J. R. 1965. A critical review of sea-level changes in eastern Australia.— Geogr. Studies 3:

63-79.

HALLAM, A. 1963. Major epeirogenic and eustatic changes since the Cretaceous, and their

possible relationships to crustal structure.—Am. J. Sci. 261: 397-423.

HALLAM, A. 1971. Mesozoic geology and the opening of the North Atlantic.—J. Geol. 79:

129-157.

HAUvuGHTON, S. H. 1963. Stratigraphic history of Africa south of the Sahara. London: Oliver &

Boyd.

16 ANNALS OF THE SOUTH AFRICAN MUSEUM

Hayes, D. E., FRAKES, L. A., BARRETT, P., BURNS, D. A., CHEN, P-H., Forp, A. S., KANEPS,

A. G., Kemp, E. A., McCCoLLum, D. W., PIPER, D. J. W., WALL, R. E. & Wess, P. N.

1973. Leg 28 deep-sea drilling in the Southern Ocean.—Geotimes 18 (6): 19-24.

Hays, J. D. & PirMan, W. C. III. 1973. Lithospheric plate motion, sea level changes and

climatic and ecological consequences. — Nature, Lond. 246: 18-22.

Hey, R. W. 1971. Quaternary shorelines of the Mediterranean and Black Seas.— Quaternaria

15: 273-284.

Hospay, D. K. 1976. Quaternary sedimentation and development of the lagoonal complex,

Lake St Lucia, Zululand.— Ann. S. Afr. Mus. 71: 93-113.

JAcoBYy, W. R. 1972. Plate theory, epeirogenesis and eustatic sea-level changes. — Tectonophys.

15: 187-196.

JELGERSMA, S. 1971. Sea-level changes during the last 10,000 years. In: STEERS, J. P. ed. Intro-

duction to coastline development: 25-48. London: Macmillan.

KENNETT, J. P., Houtz, R. E., ANDREWS, P. B., EDwArps, A. R., GosTIN, V. A., HaAsos, M.,

Hampton, M., JENKINS, D. G., MARGOLIS, S. V., OVENSHINE, A. T. & PERCH-NIELSEN, K.

1974. Cenozoic palaeooceanography in the Southwest Pacific Ocean, Antarctic glaciation,

and the development of the circum-Antarctic Current.—Jnitial Reports of the Deep Sea

Drilling Project: 1155-1169. Washington: U.S. Govt. Printer.

KinG, L. C. 1953. A marine Miocene fauna from Zululand.—Trans. geol. Soc. S. Afr. 56:

59-91.

KLEIN, R. G. 1974. Environment and subsistence of prehistoric man in the southern Cape

Province, South Africa.— World Archaeology 5: 249-283.

Koster, S. 1966. Recent sediments and sedimentary history across the Pacific-Antarctic

Ridge.— Contr. Dep. Geol. Fla St. Univ. 17: 1-83.

Lamps, H. H. 1966. The changing climate: selected papers. London: Methuen.

McCLaren, C. 1842. The glacial theory of Professor Agassiz.— Am. J. Sci. 42: 346-365.

MARGOLIS, S. V. & KENNETT, J. P. 1970. Antarctic glaciated during the Tertiary recorded in

sub-Antarctic cores. — Science 170: 1085-1087.

MENARD, H. W. 1969. Elevation and subsidence of oceanic crust.— Earth Planet. Sci. Lett. 6:

275-284.

MERCER, J. H. 1968a. Antarctic ice and Sangamon sea level. — Publ. int. Ass. scient. Hydrol. 79:

217-225.

Mercer, J. H. 1968b. The discontinuous glacio-eustatic fall in Tertiary sea level.—Palaeo-

geogr. Palaeoclim. Palaeoecol. 5: 77-85.

Mercer, J. H. 1973. Cainozoic temperature trends in the southern hemisphere: Antarctic

and Andean glacial evidence.— Palaeoecol. afr. 8: 85-114.

MILter, D. J. 1953. Late Cenozoic marine glacial sediments and marine terraces of Middleton

Island, Alaska.—J. Geol. 61: 17-40.

MILLIMAN, J. L. & Emery, K. O. 1968. Sea-levels during the past 35,000 years. — Science 162:

1121-1123.

Morner, N.-A. 1971. The position of the ocean level during the interstadial at about 30,000 BP

—a discussion from a climatic-glaciologic point of view.—Canadian J. Earth Sci. 8:

132-143.

Murray, L. G., JoyNT, R. H., O’SHEA, D. O. C., Foster, R. W. & KLEINJAN, L. 1970. The

geological environment of some diamond deposits off the coast of South West Africa. —

Rep. Inst. geol. Sci. 70 (13): 119-141.

Rona, P. A., 1973. Relations between rates of sediment accumulation on continental shelves,

sea-floor spreading, and eustasy inferred from central north Atlantic.— Bull. geol. Soc.

Am. 84: 2851-2872.

Ruppock, A. 1968. Cainozoic sea-levels and diastrophism in a region bordering Algoa Bay. —

Trans. geol. Soc. S. Afr. 71: 209-233.

RUSSELL, K. L. 1968. Oceanic ridges and eustatic changes in sea level.— Nature, Lond. 218:

861-862.

RUSSELL, R. J. 1963. Recent recession of tropical cliffy coasts. — Science 139: 9-15.

RUTFORD, R. H., CRADOCK, C. & BASTIEN, T. W. 1968. Late Tertiary glaciation and sea level

changes in Antarctica. —Palaeogeogr. Palaeoclim. Palaeoecol. 5: 15-39.

SAVIN, S. M., DouGLas, R. G., & STEHLI, F. G. 1975. Tertiary marine palaeotemperatures. —

Bull. geol. Soc. Am. 86: 1499-1510.

CENOZOIC SEA-LEVEL CHANGES: A DISCUSSION 7

ScHOLL, D. W. & STUIVER, M. 1967. Recent submergence of Southern Florida. A comparison

with adjacent coasts and other eustatic data.— Bull. geol. Soc. Am. 78: 437-454.

SHACKLETON, N. J. 1969. The last interglacial in the marine and terrestrial records.—Proc.

R. Soc. Lond. B. 174: 135-154.

SHACKLETON, N. J. & KENNETT, J. P. 1974. Palaeotemperature history of the Cenozoic and

the initiation of Antarctic glaciation: oxygen and carbon isotope analyses in D.S.D.P.

sites 279, 277, and 281.—Initial Reports of the Deep Sea Drilling Project: 743-755.

Washington: U.S. Govt. Printer.

SHACKLETON, N. J. & OppykE, N. D. 1973. Oxygen isotope and palaeomagnetic stratigraphy

of equatorial Pacific core V28—238: Oxygen isotope temperatures and ice volumes on a

10° year and 10° year scale.— Quatern. Res. 3: 39-55.

SHEPARD, F. P. 1961. Sea-level rise during the past 20,000 years.—Z. Geomorph. Suppl. 3:

30-35.

SHEPARD, F. P. 1963. Thirty-five thousand years of sea level. Jn: CLEMENTS, T., ed. Essays in

marine geology: 1-10. Los Angeles: Univ. Calif.

SHEPARD, F. P. 1964. Sea level changes in the past 6 000 years: possible archaeological sig-

nificance.— Science 143: 574-576.

SHEPARD, F. P. & CurRAY, J. R. 1965. Carbon-14 determination of sea level changes in stable

areas. — Progr. Oceanogr. 4: 283-291.

SUGGATE, R. P. 1974. When did the last interglacial end ?— Quatern. Res. 4: 246-252.

TANKARD, A. J. 1975. The marine Neogene Saldanha Formation.—Trans. geol. Soc. S. Afr.

78: 257-264.

TANKARD, A. J. 1976. Pleistocene history and coastal morphology of the Ysterfontein—Elands

Bay area, Cape Province.— Ann. S. Afr. Mus. 69: 73-119.

THom, B. G. 1973. The dilemma of high interstadial sea levels during the last glaciation. —

Progr. Geogr. 5: 167-246.

THomM, B. G. & CHAPPELL, J. 1975. Holocene sea levels relative to Australia. — Search 6: 90-93.

THoM, B. G., Hairs, J. R. & Martin, A. R. H. 1969. Radiocarbon evidence against higher

postglacial sea levels in eastern Australia— Mar. Geol. 7: 161-168.

TROWBRIDGE, A. C. 1954. Mississippi River and Gulf Coast terraces and sediments as related

to Pleistocene history—a problem.— Bull. geol. Soc. Am. 65: 793-812.

VEEH, H. H. 1966. Th?*°/U?*8 and U*4/U?8 ages of Pleistocene high sea level stand. —J. Geophys.

Res. 71: 3379-3386.

VEEH, H. H. & CHAPPELL, J. 1970. Astronomical theory of climatic change: support from

New Guinea.— Science 167: 862-865.

VINE, F. J. 1973. Continental fragmentation and ocean floor evolution during the past 200 m.y.

In: TARLING, D. H. & RUNCORN, S. K. eds.— Implications of continental drift to the Earth

Sciences: 831-839. London: Academic Press.

WALCOTT, R. I. 1972. Past sea levels, eustasy and deformation of the earth.— Quatern. Res. 2:

1 -14.

Warp, W. T. 1965. Eustatic and climatic history of the Adelaide area, South Australia. —

J. Geol. 73: 592-602.

WRIGHT, J. A. 1964. Gully pattern and development in wave-cut bedrock shelves north of

the Orange River mouth.—Trans. geol. Soc. S. Afr. 67: 163-171.

a (ita a Oe

=n - r ro I

,e )

Call

S . 7

' i

J it

‘

] <

+¥ =

| {

i :

f ;

} I

i 2 ‘

18)

~ r

i ' ¥

i -

¥ ‘ ;

ic w

i : ;

a :

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND

AND THEIR RELATION TO FORMER SHORELINES

OLIVER DAVIES

Natal Museum, Pietermaritzburg

(With 5 maps)

ABSTRACT

In Natal and southern Zululand there extends along the coast a single dune-cordon, or

in a very few places two fossil dune-cordons, in the rest of Zululand at least six. The older

cordons are formed either of bright red sand or, close to the coast, of white cemented aeolianite,

and are easily distinguishable from relatively unconsolidated low dunes of white, yellowish

or light brown sand which have been formed recently, often by redistribution of the older

sands. The main cordons were constructed during the earlier parts of glacio-eustatic regres-

sions, the latest after the end of the Eem interglacial. They are not amenable to radiometric

dating and are hardly ever fossiliferous; an attempt is made to date them by means of artefacts

on or beneath them, by considering the marine terraces on which they rest, especially the

lowest. It thus appears that the two cordons of Natal belong to the beginning of the last

glaciation, with cores probably dating to the Saale glaciation; while the Zululand multiple

cordons can be correlated with at least three glacio-eustatic regressions, as far back as the

Elster or perhaps to an earlier glaciation. There is no reason to think that any of the extant

dunes are older than the Pleistocene.

This paper has been condensed from three reports by the author written for the Tertiary/

Quaternary Group of the South African Committee for Stratigraphy. The map was compiled

for one of these reports, and contains more detail than is included in this text.

CONTENTS

PAGE

Natal and southern Zululand . ... . 19

Central and northern Zululand See MS)

Conclusion: suggested Pleistocene chronology 30

NATAL AND SOUTHERN ZULULAND

Despite much variation in relief, the coast of Natal is remarkably

unindented. All along, from the Msikaba River in northern Transkei to the

Mlalazi River in southern Zululand, there extends a dune-cordon, nowhere

more than 5 km from the present shore, in places rising directly from it. Almost

everywhere it is composed of bright red quartz-sand, in places containing

ilmenite and magnetite. In depth the sand tends to yellowish or grey, but is not

less consolidated than close to the surface. It seems to be mostly wind-

transported, but in places there may be inclusions of clay from temporary

vleis. In a very few places within Durban, such as Burman’s Bush (King L.

1962; Frankel 1965) and Montclair (King L. C. 1966), there is hard white

aeolianite in place of red sand. Tests of foraminifera in this rock indicate that

the sand is at least partially of littoral origin. The Tertiary-Quaternary Working

Group of the South African Committee for Stratigraphy has provisionally

called this cordon the Berea Red Sands member of the Bluff formation.

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 19-32, 5 maps.

20 ANNALS OF THE SOUTH AFRICAN MUSEUM

28°00

28°30-

em CORDONS ABCOEF

SW Small white dunes

wut Red sand,believed rede-

posited

N Thin soil,no dune

In Inaccessible

MAP 1

edie e if |

3T 30’

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND 21

The top of this cordon is irregular, and it has undergone considerable

erosion. In a few places it rises to > 150 ms.l. Its vertical thickness is variable;

at the University of Natal not less than 70 m.* It is interrupted even by small

rivers, nearly always at their lagoons which mark gorges grading to low glacio-

eustatic sea-level; many of them had probably been incised before the last

glaciation (Weichsel) and reactivated at that time. Landwards of the cordon

are shallow valleys parallel to the coast, presumably cut after the cordon was

first formed.

Except where incised by rivers, the cordon takes no account of under-

lying topography. It rests both on rock-slopes and on fossil beach-gravels;

some of these gravels may be of Tertiary age (e.g. Umgeni Quarry, Clarewood

Quarry; King L. C. 1966), but most contain rolled artefacts and were deposited

during the Quaternary regression. The highest beach-gravel which has been

noticed overlain by red sand is at > 101 m at Frasers; there are sites nearly

as high at Malakazi and Ilfracombe (Davies 1970). The lowest site is at +6 m

at Ngane Mouth, where a rock-platform at +5,5 m, perhaps a river-channel,

is overlain by 70-80 cm of marine sand beneath dune-sand; in the gravel have

been found rolled artefacts of Late Acheulian (Fauresmith) type. Red dune-

sands rest on beach-gravel at +8 m at Tugela Mouth, +9 m at Sheffield Beach

and Willard Beach, +8,5 m at Umdoni Park. There is no evidence at these

sites that the sand has slumped. The writer assumes that the present cordon

(discounting a possible older core) was formed as a single stage, after the

+8-9 metre sea-level. Though we have no radiometric dates, the writer is

inclined to equate this sea-level with the second peak of the last interglacial

(Eem).?

As one would expect, there are no clean sections through the red dune.

Two very deep gullies near Stanger (at Hyde Park and Kijabe), which cut

right through it, are thickly overgrown and inaccessible. Nor have uncon-

formities been observed normally within the dune. Excavations to 15-20 m

1 This figure is calculated on the assumption that the Berea Cordon at this point rests on

the 60-70 metre marine platform, as it does at the Mgeni and at the Tollgate; both east and

west of the latter this platform was exposed some years ago. McCarthy (1967: 148, 197)

describes an excavation at Pigeon Valley Reservoir just west of the University, but this did

not reach rock.

2 The evidence is set out in detail in Davies (1970) and Davies (1972) and cannot be repeated

here. It is assumed on a world-wide basis that the peak of Eem I was at about +7 mand Eem IIT

just above or just below 0 m; but the emergence throughout the Quaternary of the coastal

regions of south-eastern Africa would produce higher levels for the Eem. There are two

levels along these coasts, at about +18 and +9 m. The gravels on these terraces contain

archaeological material, which can be roughly dated: on the 18-metre terrace both rolled and

unrolled pieces which it is difficult to assign to a later stage than the Acheulian; on the 9-metre

terrace rolled pieces which typologically are best assigned to the End-Acheulian. It is now

being recognized that the Acheulian was giving place to more developed industries of Middle

Stone Age type before the end of the Eem interglacial—not as sometimes claimed at the

beginning of the interglacial. The archaeological evidence is confirmed by faunal evidence

from the south-eastern Cape. Members of the warm-water Swartkops fauna occur in estuarine

deposits at about +15 mand +6 m; these beds almost certainly belong to the 18 and 9-metre

sea-levels. The fauna could not have survived a large drop of ocean-level for which in Italy

and elsewhere there is evidence before Eem I; so these two sea-levels must be regarded as

stages of a single interglacial.

Dy ANNALS OF THE SOUTH AFRICAN MUSEUM

depth at Overport and elsewhere in Durban reveal nothing but compacted

sand. At Pigeon Valley Reservoir, behind the University of Natal, McCarthy

(1967: 148, 197) observed traces of calcification about 14 m below surface,

but no unconformity. At Hyde Park a terrace seems to be cut in red sand at

+7,5 m. Though there is no clean section, a terrace at this altitude suggests a

marine-cut platform, perhaps rather eroded, separating the main dune from

an older dune, perhaps of Saale age.

At Umwabi Road, Isipingo, a sand-pit on the flank of a small valley which

cuts through the red dune exposed marine terraces at +33 and +45m, separated

by a cliff formed of a dyke of rotten dolerite which must have emerged as a

stack above the upper terrace, which is cut in shale (Maud 1968: 167; Davies

1970: 406, 419, 425-6; section no longer cleanly exposed). Banked against the

cliff was yellowish sand with layers of ferricrete and ilmenite. This was overlain

by a surface of shale chips and a few unrolled artefacts, one probably a hand-

axe. Both terraces and this surface were covered by dune-sand, red passing into

yellow in depth, to an altitude of +82 m, with a line of quartzite pebbles on it

such as frequently occur on the main Berea Cordon and are probably manuports.

It appears that the lower sand against the cliff was the remains of an older

cordon which had been almost entirely removed, perhaps in the last interglacial.

Most of the exposures on top of the main Berea Cordon have suffered

erosion in the past, and reveal a mixture of cultural material which may date

from any periods after the cordon was formed. But on a few two surfaces are

visible. The lower is 1-3 m below the surface of the red sand; it carries picks

and other artefacts of the Tugela Industry. The upper line is on the surface of

the red sand, and carries artefacts of Pietersburg affinity (probably Sampson’s

(1972) M.S.A. Phase 3). It therefore appears that since the last interglacial

there were two stages of dune-construction, separated by an unconformity.

In a large exposure at Umgababa ilmenite-quarry there is evidence for

stratification of the dune which can be reconstructed as follows:

1. On the quarry-floor at about +45 m bright red sand, perhaps older

dune, with a sloping surface carrying stone chips, nodules and Late Acheulian

artefacts. The base of the sand is not exposed, so it is not known if it rests on

a beach or on a rock-slope.*

2. Above the lower red sand, up to 6 m of bright red sand capped by a

surface with terrestrial molluscs, root-tubes and artefacts of the Tugela Industry.

3. On this surface, up to 18 m of whitish calcified dune, capped by a layer

of ferricrete nodules containing flakes, points and other artefacts of Pietersburg

type.

4. On this surface, soft red sand passing conformably into grey humified

sand.

3 Formerly Sangoan; but the writer has come to the conclusion that this name is not justified

for the Natal coast (see Davies 1976)

_ * McCarthy (1967: 196) gives a schematic section, clearly not based on direct observation,

eieanee that in his view the red sand rests on pebbles on the seaward-slope of the 18-metre

shoreline.

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND 23

The calcified sand which here separates the Tugela from the Pietersburg

surface is thicker than the upper red sand found elsewhere between the two

surfaces; but the contemporaneity of each surface at different places is guaran-

teed by the artefacts. McCarthy (1967: 192-6) misinterpreted the Umgababa

site because he relied on dubiously datable foraminifera and took no account

of the archaeological evidence (see Davies 1976).

In some places, for instance at the Upper and Lower Red Deserts, Port

Edward, there are two discrete areas of red dune; but the writer considers that,

apart perhaps from an older core, the Berea member is a single cordon which

has in places been subdivided by erosion. There are, however, remains in Natal

of an outer cordon, for which has been recommended the name Calcarenite

member of the Bluff formation.

The Bluff Cordon is in places submerged as an offshore reef (e.g. Aliwal

Shoal; McCarthy 1967: 136-7). Only at Durban Bluff are all three beds

preserved :

Bed 3. Lightly calcified aeolianite with pipes and nodules, nearly 100 m

thick and rising to a maximum of +106 m. It apparently contained vlei-beds

and lenses of boulders, perhaps pieces of cemented sandstone, but no con-

tinuous unconformity. Near the surface are unknown thicknesses of decalcified

red sand (see King & Maud 1964: 21-2).

Bed 2. A beach less than 1 m thick, composed of boulders, shells and sand.

At Cave Rock it rises to a maximum of +8,1 m and dips steeply seaward.

At Mbokadweni Mouth and Umbhloti Beach (see Maud 1968: 177-9) there

were apparently fluctuations of sea-level soon after the start of the regression,

but the sequence of these events has not been fully elucidated.

Bed 1. A very hard aeolianite down to —90 m, resting in places on Karoo

rocks, elsewhere on planed Cretaceous beds.

Bed 2 occurs only on the seaward face of the Bluff, at Umhloti Beach and

Tiger Rocks. Where it is absent, e.g. at the Point, the contact of beds 1 and 3

is not clear and there is no obvious unconformity. Bed | probably rose higher

than +8 m where not marine-planed, perhaps considerably higher, as on its

face has been found slumped sand, equally calcified and containing stone chips

and mammal-bones.

Except along the Bluff, bed 3 is missing and bed 2 is truncated. On the

offshore reef presumably bed 1 also is truncated.

It is likely that at Durban, as elsewhere, the 8-metre beach correlates with

the second peak of the last interglacial. There is no sign in Durban of the first

peak, perhaps to +18 m, except far up the Mgeni Estuary. The fluctuations

below the 8-metre beach may cover the third peak (cp. Suggate 1974). If this

dating is correct, the lower aeolianite or bed 1 of the Bluff formation was

formed during the eustatic low of the penultimate (Saale) glaciation, the upper

aeolianite or bed 3 during the Weichsel glaciation.

Practically no artefacts have been found in the Bluff beds. There was one

pick of Tugela-type in the beach, perhaps one or two on a river-flat behind the

24 ANNALS OF THE SOUTH AFRICAN MUSEUM

beach; but on the surface of the Bluff Ridge nothing older than Wilton.

The highest beds of both Berea and Bluff Cordons rest on a beach at

+8-9 m, and probably belong to the same age; perhaps the seaward cordon

is a little later than the landward. A pause in dune-construction could have

been due to restriction in the supply of sand caused by a marine transgression

which flooded much of the continental shelf. It is unlikely that this transgression

was the third peak of the last interglacial, because between the second and third

peaks the sea probably did not regress far enough to supply sand for the whole

of the Berea Cordon; so the transgression may have been that of the Brérup

interstadial.

Therefore, with great hesitation owing to the absence of typical artefacts

on and in the Bluff Ridge, the writer would propose the following chronology

for the Natal dunes: the most active period of dune-formation would be the

earlier part of a marine regression, when supplies of sand on the continental

shelf were being continually exposed but only slowly anchored by pioneer

vegetation. Changes of prevailing wind and rainfall during the onset of a

glacial period may also have contributed:

After the mid-Weichsel transgression (Dennekamp—Hengelo). Partial decalcifi-

cation of Bluff bed 3 and of bed 3 at Umgababa.

Roughly at the mid-Weichsel interstade. Upper dune-surface with Pietersburg-

type industry.

Between the Brérup and mid-Weichsel interstades (about 60000 and 35 000

B.P.). Construction of the Bluff upper aeolianite, of Umgababa bed 3,

and of the upper red sand on the Berea Cordon.

Before the Brorup interstade (before 60 000 and perhaps as early as 80 000 B.P.).

Tugela Industry on the lower surface of the Berea Cordon and on the still

exposed 8-metre beach of the Bluff (bed 2). Construction of the principal

body of the Berea Cordon.

Third peak of the last interglacial (c. 80000 B.P.). Sea-level near modern.

Between second and third peaks of the last interglacial (c. 107 000-80 000 B.P.).

Sea-level a little below modern.

Second peak of the last interglacial (c. 107 000 B.P.). Formation of the 8—9 metre

beach.

L. C. King’s suggestion (1966) that Uloa (Map 3 013) and other beds in

the Mfolozi Estuary, the Bluff beds and the outcrops of calcareous sandstone

in the Berea Cordon in Durban are all Pliocene is unacceptable. The detailed

foraminiferal reports which he quotes in this but not in subsequent papers

say for the Mfolozi series ‘Age; Pliocene or younger’, for the Durban series

‘Age: None given’. To argue from these reports that all these occurrences are

Pliocene is illogical. As the Miocene beds at Uloa and other Zululand sites

are close to modern sea-level, the writer would agree that there has been a post-

Miocene transgression to >150 m, the highest known level of the Pliopleistocene

shorelines in Natal (Davies 1970); some of the calcified beach-deposits in

Durban, especially on terraces which on altitude do not fit into the estab-

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND DS

lished Pleistocene regressive sequence, may belong to the transgression. But

the presence of Middle Pleistocene artefacts within beach-gravels beneath the

calcified patches of dune excludes the possibility that any of the calcified dunes

belong to the Miopliocene stage.

CENTRAL AND NORTHERN ZULULAND

North of the Mlalazi River the Zululand coast trends from north-north-

east to north-east to form a wide coastal plain bounded landwards by the

Lebombo Mountains. The northern part of this plain is composed of Cretaceous

rocks with perhaps a Tertiary veneer, marine-planed and then lightly dissected;

the planation has extended a short way into the basalts. In the southern part

the Cretaceous has been largely submerged beneath younger sediments, and

there are lagoons and swamps which have been isolated fairly recently from the

sea. The Pliopleistocene shoreline is a long way inland, nearly 100 km near the

Mozambique border.

At least six cordons of red dune-sand can be identified in the coastal plain,

besides a large number of lower and shorter dunes of brown, yellow or white

sand, perhaps formed by redeposition; there may be more than six cordons,

but heavy afforestation in the south makes it impossible to trace some of the

cordons over long distances, and much of the northern area is trackless grass-

land and inaccessible. It is presumed that each cordon was constructed not far

from the contemporary shore. Lack of exposure makes it difficult to determine

their bases. They do not rest each on a single shoreline, but there has probably

been slumping into valleys as they graded to fluctuating sea-levels.

Cordon A, the coastal cordon, follows the present coast closely. Both

faces are steep. It rises to a maximum altitude of +145 m. It is interrupted only

at four river-mouths (Mhlatuze and Nseleni, Mfolozi and Msunduze, Sordwana

Bay, Kosi Bay). At Mission Rocks (just south of Bats Cave) and Lake Sibayi

it overlies fossil mouths, marked by up-country pebbles in beach-rock which

was formed during or perhaps before the last interglacial. It may rest on well-

cemented beach-sands which incorporate older aeolianite pebbles, in some

places at + 1-2 m, at Bats Cave at +8 m, at Black Rock (Map 5 05) at +12 m;

also on uncemented marine or lagoonal sands at +18 m at Mbonambi Beach

(Map 2 05) and Port Durnford (Map 2 06; Hobday 1975). At Kosi Bay cemented

aeolianite rested on calcified beach-sand at —3 m. While Cordon A may

incorporate fragments of an older cordon, in its present form it seems to post-

date the 18-metre sea-level. It may well be younger than the 8-metre level; but

nowhere is the cliff of the underlying beach exposed.

An entrenched channel north of Cape Vidal (Hill 1975), which apparently

drained Lake St Lucia to the north-east, must be older than the fossil mouth

at Mission Rocks and was probably blocked by a cordon older than the 18-metre

sea-level.

Cordon A seems to be formed of red sand; but in places there is evidence

26 ANNALS OF THE SOUTH AFRICAN MUSEUM

KEY to MAPS 2-5

of e2 etc. Place-names,

see text; for coastal and

river names see Map 1.

‘<A Betc. Approx. border of dune, by exposure

or contour, name of cordon.

70 >70 <70 Dune-base exposed or inferred, m

above sea-level.

4(70) Summit of dune (m).

a (a) Archaeological material beneath

(on) dune.

poe Dune-sand believed redeposited.

Dannie ~ Rock-cliff backing dune.

Sand-colour: R red; W white; Ca calcified;

WR white turning red below.

f Forest probably on sand, no exposures.

ey f ?

= ses) yo = e7 °75

73 NB ax oe eae 6P

’ n)a Water-laid beds

) 2

A %

of calcification, at Meersig Quarry (Map 3 010) up to +65 ms.l. and at no

great depth from the summit. Its composition therefore resembles Durban

Bluff. But though no artefacts have been found beneath Cordon A and fossils

very rarely, a stone pick of the Tugela Industry was found by D. K. Hobday

on the dune at Port Durnford. Behind the village of Richards Bay there seemed

to be evidence, at the back of the cordon, for two stages of slumped red sand

separated by a ferricrete layer; on the lower sand were two rolled pebbles which

must be manuports, on the upper a pick. Though exposures are rare on Durban

Bluff and no watch was kept on the widespread building twenty years ago, in

fact no piece of the Tugela Industry has been found on the ridge; so Cordon A

in Zululand may in part be slightly older than the Bluff Ridge.

Cordon B can be traced along the west side of Lake St Lucia. At Charters

Creek (Map 3 07) and Hell’s Gate (Map 3 08) it rests on a platform at +8-10 m,

at Mpathe F.R. (Map 3 012) and Fanies (Map 3 09) on water-laid beds at

+30 m. South of Malakatana (Map 3 06) it apparently diverges southwest-

wards, follows the eastern side of the Mpate valley, is interrupted by the Mfolozi

Estuary, and resumes at Sokulu (Map 3 011) and Red Hill (Map 2 04), where

it diverges from Cordon A. Its base is at about +25 m on both banks of the

Mkuze River near Lower Mkuze Store (Map 4 03). It cannot be traced north

of 27.30 S, but may continue as one of the inaccessible dunes west of Maputa

(26.59 S 32.45E).

Cordon B is formed of red sand. No artefacts have been found on it, other

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND v4 |

27°45'

we Oe

~~.

ono 5

oo"

men

eed

<--->

=== Ee

Ca,

ofde R on surface

/? olderdune 28°15’

tobelow SLL.

28 ANNALS OF THE SOUTH AFRICAN MUSEUM

than fairly recent material. Beneath it the writer found at Mpathe F.R., on a

surface formed partly of water-laid grit and partly apparently of older dune,

broken pebbles of nondescript type. As Cordon B rests in places on the 8-10

metre terrace, it must be younger than one peak of the last interglacial.

Cordon D can be traced fairly continuously from the Mlalazi River

probably to the Mozambique border. It is interrupted by several rivers, and

in the far north is mostly inaccessible. It is formed of red sand. Its sides are not

steep, and it rarely rises to a great height above the adjacent plain. Its maximum

altitude is +112 m south of Mtubatuba and at Kuleni (Map 3 ol), +130 m

north of the Mkuze River (see Map 4). Its base is generally between +60 and

+75 m, nowhere below +30 m save where it seems to have slumped. In places

it rests on one of the higher beach-gravels (-+-49-+70 m).

Cordon D appears to be aligned with a cordon behind Mtunzini, which

rests on rock at +61 m. South-west from Mtunzini there is a single cordon,

which appears to belong to the Berea member. It is, however, difficult to associ-

ate Cordon D with the Berea Cordon, which the writer has shown to date

from the last interglacial. Cordon D is 22 km from the present coast near

Mtubatuba, >35 km behind Kosi Bay. Its profile is much less steep than the

Berea Cordon. On it have been found a very few artefacts—in addition to

pieces of the Tugela Industry, a hand-axe at the University of Zululand (Map 2

ol) and a small cleaver near the Mlalazi River; these are likely to be Late

Acheulian and older than the last interglacial. Beneath the dune have been

== —=---"”

wee eo eK =

— eww ene ~~ ee

—

Sa ear td

---—.

(146)! a\>100

<130*!

-~e

PS)

SS ee) aad

--—

oor

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND 29

found rare pieces of some stage of the Acheulian. Near the Mhlatuze River

there is a surface 3 m below the top of the red sand, indicating reworking or

renewed dune-building during the last glaciation.

It appears therefore that Cordon D is continuous with the older core of

the Berea Cordon, revealed at Umwabi Road, Durban; whereas Cordons A

and B correlate with the present Natal cordons, and were formed after the last

interglacial.

Cordon C is a discontinuous line from south of the Mhlatuze River to

the promontory on the south side of False Bay. Possibly the outer dune at

Tshongwe (Map 4 05) is part of it. North-east of Mtubatuba there may be

included in it more than one cordon. It rests at >21 m on an estuarine gravel

of the Nyalazi River and farther south on a platform at +27 m.

Cordon C seems to be older than the 18-metre shoreline. Both it and

Cordon D may have been constructed just after the 30-metre shoreline in the

Mfolozi basin; but farther north, where the Cretaceous shelf is higher, there

is only one cordon. There is no satisfactory archaeological evidence for dating

Cordon C. There have been found on it in Dukuduku F.R. a few indeterminate

artefacts possibly of Tugela affinity; at Hluhluwe Bridge (Map 3 04) and

Nyalazi Dip (Map 3 05) it was impossible to tell if the artefacts had come from

beneath or above the red sand.

Cordons E and F are fragmentary. It is often difficult to decide which

cordon is represented, and there may be remains of more than two cordons

west of D. Hesitatingly the writer has plotted as Cordon E a prominent ridge

extending from the Mozambique border and Ndumu Hill, west of the Pongola

River past Nkononde (Map 4 04) into Mkuze Game Reserve (Map 4 02);

it has not been identified south of the Msunduze River. It rises to +189 m near

Makhane’s Pont (Map 5 04) and +175 m at Ndumu, and in places >70 m

above the adjacent plain. In the extreme north its base is at about +100 m.

West of the Pongola River and in Mkuze Game Reserve it rests on the inner

edge of the 60-metre beach. Between the Nsumu and Msunduze Rivers its

base seems to be at <45 m; but the area is very difficult of access, so it has not

been possible to control information obtained in 1950.

Nor can Cordon E be archaeologically dated save within wide limits. The

60-metre beach-gravel has yielded rolled choppers of adiagnostic type at

Mzinyeni Dip (Map 5 03) and Mlambongwenye (Map 5 02) (Davies 1970).

On the dune south of the Nsumu River there were found lightly abraded rough

hand-axes.

Although Cordon E is well preserved, it appears to date to an earlier

interglacial-glacial cycle than Cordon D, the two being up to 15 km apart.

If D with fragments of C belong to the Holstein-Saale, E may date to the

Cromer-Elster cycle or to an unrecognized interglacial during the Elster.

Cordon F is almost certainly older than E. In Mkuze Game Reserve,

where it can be distinguished from E, it is 30 m high and its base is at +95 m.

Elsewhere it is fragmentary, and in many places all that is preserved is a low

30 ANNALS OF THE SOUTH AFRICAN MUSEUM

body of red sand banked against a cliff. The writer has assigned to it fragments

with base as follows:

Empangeni. ce ee 3= Oa sion

Homeleigh against a cliff Glew 2 03) tr ee eee eee +147m

Ncemane against a cliff (Map 3 a errs heer es +90 m

Hluhluwe (Map302) . . SO Ne eee +73 m

Mkuze Game Reserve (Map 4 Oo). Se ie Tt ih Fone =-95 mi

east ol Lubombu Driiti(Viap > olin) a ee +90 m

It seems therefore to rest on surfaces at +73 m and +90-+100 m. The cliff

at Ncemane is the most probable indicator of its altitude—i.e. it is associated

with a marine platform at about +90 m; but it is possible that really two

cordons are represented. There is no archaeological evidence for dating Cordon

F; but there is no reason to suppose that any part of it is pre-Pleistocene.

15 MOZAM32230B1QUE 32145

ae e

MAP 5 fo fBI PRI jay

SAIND / :

ve | ia(90)

7 mR} inc

2700'

pe) Te

ee BLA

a) g

meee) Se

=5-s--—A eee ees

CONCLUSION: SUGGESTED PLEISTOCENE CHRONOLOGY

All the cordons in Natal and Zululand were probably constructed during

marine regressions, which uncovered abundant sand on the continental shelf.

Their formation was probably rapid, as they do not include well-marked land-

surfaces. So each dune was for a time mobile, but owing to changing conditions

became fixed by vegetation. Fixation may have been due to continued regression,

which removed the shore so far that beach-sand no longer reached the dune;

but partial transgression, change of prevailing winds or aridity could reactivate

it. So far as we know, the deepest regressions round the emerging sub-continent

were in the Saale and Weichsel glaciation-cycles. To each of these cycles there

are assigned two cordons, and there may be others offshore.

Lack of informative exposures makes it difficult to study these dune-

cordons. Natural drainage has incised the dunes but left slumped and heavily

THE OLDER COASTAL DUNES IN NATAL AND ZULULAND 31

vegetated scars. Industrial development has seldom cut straight through a

cordon to reveal a clean section. We can reasonably identify events of the last

cycle (Eem—Weichsel) from the lowest terraces on which the dunes rest and

from artefacts which by then had become more differentiated; during older

cycles, artefacts are much less diagnostic.

Provisionally the author would propose the following table:

Last cycle, Eem/Weichsel, Natal, Berea Cordon; On dunes, Tugela

with 3 interglacial Natal, Bluff bed 3; and later

peaks of sea-level Zululand, Cordons A industries ;

followed by deep and B beneath dunes,

regression Late Acheulian

and earlier

Penultimate cycle, Natal, core of Berea On dunes,

Holstein/Saale Cordon; Acheulian

Natal, Bluff bed 1;

Zululand, Cordons C

and D; aeolianite

pebbles incorporated

in beach-rock of last

interglacial

Antepenultimate cycle, Zululand, Cordon E Beneath dunes,

Elster or Cromer/Elster pre-Acheulian

APPENDIX

Maps are included only of the complicated areas of Zululand, where

there are several cordons. South of the Mlalazi River at most places only a

single cordon occurs. To enable the reader to find localities mentioned in the

text, a list of their latitudes is given:

MitwaleSivoah =. ya. We ea te | SORTS!

BiurmamspBuShl 5. va Nel bya. 3 29°49"

ClarewoodsOuary. 2). 45 e2) ie 67, 29°55"

RASCESie en en ae roe ctr tet a DOT SST

Fibberdene soad-cutuimes)) | 44 4. 30-32)

eG ew leakcnns hae ar et to ce ee DOLD

WitAcOMmbe Te Oe oe, is. SOA

isipingo, Umwabideoad) 9,9. 4. =. = 29.59

KG bommnates i PE a ek es DOO DS!

WMalaKaZIo Re Ui a Fal detent) gle! oa gy 30°00!

Mbokadweni Mouth ... . . . 30°01’

Montclair Mee Re soa Peer ae) )

32 ANNALS OF THE SOUTH AFRICAN MUSEUM

Msikaba Mouth ak as Deere peel esl 9;

NeaneMouthrieawte 262.0 4. Bee ee Oalily

PortEdwardined deserts) See eee Os

Shetiield: Beach cr" psoas - es) ae a oo:

Tiger ROCKS: 2: cn. s.c. i; a Se ee SO ROY

‘Fugela, Mouth: 22> 072 (1p: See Se Oley

Umdont Park. 2--..) OS Sk 2 8024

Umgababa ilmenite quarry. . . . . 30°08’

Winigeni‘Quarry: 5) 2 eo

WimblotimBeach 9 2) So ee a ee eee Oe

Willard Beach) ~.0 % “322 2) 2 29338

REFERENCES

Davies, O. 1970. Pleistocene beaches of Natal.— Ann. Natal Mus. 20: 403-442.

Davies, O. 1972. Pleistocene shorelines in the southern and south-eastern Cape Province

(part 2).—Ann. Natal Mus. 21: 225-279.

Davies, O. 1976. The ‘Sangoan’ Industries.— Ann. Natal Mus. In press.

FRANKEL, J. J. 1965. On the Pleistocene rocks at Burman Bush.—S. Afr. J. Sci. 61: 183-185.

HILL, B. J. 1975. Origin of Southern African coastal lakes.—Trans. R. Soc. S. Afr. 41:

225-240.

Hospay, D. K. 1975. The Port Durnford Formation. — Trans. geol. Soc. S. Afr. 77: 141-149.

KING, L. 1962. The Post-Karroo Stratigraphy of Durban.— Trans. geol. Soc. S. Afr. 65: 85-99.

Kinc, L. C. 1966. An extensive marine Pliocene formation in Natal and Mocgambique. —

Trans. geol. Soc. S. Afr. 69: 201-210.

Kina, L. C. & MaAupb, R. R. 1964. The geology of Durban and environs.— Bull. geol. Surv.

S. Afr. 42:1-54.

McCartTny, M. J. 1967. Stratigraphical and sedimentological evidence from the Durban

region of major sea-level movements since the Late Tertiary.— Trans. geol. Soc. S. Afr.

70: 135-65, 191-198.

Maup, R. R. 1968. Quaternary geomorphology and soil formation in coastal Natal.—

Z. Geomorph. N. F. Supplementband 7: 155-199.

SAMPSON, C. G. 1972. The stone age industries of the Orange River Scheme and South Africa.

— Mem. natn. Mus. Bloemfontein 6: 1-288.

SUGGATE, R. P. 1974. When did the Last Interglacial end?— Quat. Res. 4: 246-252.

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT

PLATFORM

BURGHARD W. FLEMMING

N.RJI.O.(C.S.I.R.), Marine Geoscience Unit, University of Cape Town

(With 12 figures and 1 table)

ABSTRACT

Rocky Bank is an isolated rocky outcrop 10 km south-east of Cape Point. Between 30

and 40 m below sea-level, a wave-cut platform is sparsely covered with well-rounded pebbles,

cobbles and boulders of Table Mountain Sandstone. The extremely low quartz content in the

sediments on the platform indicates that there is no marine abrasion active today. This supports

evidence from other parts of the world that the abrasive energy of ocean waves is restricted

to depths shoaler than 12-15 m. Calculation of orbital velocities from local wave conditions

demonstrate the concentration of wave energy near the sea surface. To account for the wave-

cut platform and its features a sea-level 20-25 m below the present must be assumed. This

agrees well with other observations along the west coast of southern Africa.

CONTENTS

PAGE

Introduction . : - 2 33

Methods : ; é : 35

Morphology . ; : : 35

Geology ; : ; : 36

Sediments : ; : : 3i/

Discussion : : é : 39

Summary ; 3 . : 47

Acknowledgements . : : 47

References : : ’ ; 47

INTRODUCTION

Rocky Bank is an isolated rock outcrop about 10 km (6 nautical miles)

south-east of Cape Point at the mouth of False Bay, Western Cape, South

Africa. The geographic position of its shallowest part is 18.35,5S, 24.24,8E.

The outcrop has been traversed on several research cruises in the past

but no samples were recovered to establish its lithology or the nature of its

sediment cover. To fill this gap an in situ investigation was arranged during the

winter months of 1973. Besides obtaining this initial geological information,

clear evidence was found that Rocky Bank represents a wave-cut platform

not exposed to submarine abrasion under the present hydrological conditions.

To explain this feature a lower sea-level must be assumed.

This study presents the evidence that characterizes the wave-cut platform

and proves its present non-activity. The implications of such evidence with

respect to the methodological difficulties experienced when attempting to infer

former sea-levels from raised marine terraces are discussed.

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 33-48, 12 figs, 1 table.

ANNALS OF THE SOUTH AFRICAN MUSEUM

uolje}s Hulaip — K

sidjaw ul Yyydap

WwW Z b 0

MNVaG AWDOY

‘AWow Aye pue Ajl[e0] —yueg AYOY “[ ‘3l{

(SH

Zr ve

OC"

fo)

S) q

Vv ° §

LNIOd AdvVO

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM 35

First results of this investigation were reported in the annual technical

report of the Marine Geology Unit at the University of Cape Town (Flemming

1973). Since then new information has been obtained by side-scan sonar and

shallow seismic profiling, clarifying some of the problems that had remained

unsolved.

METHODS

Observations, documentation and sample collection was carried out in

situ by the application of SCUBA diving. Notes and sketches were made on

plastic sheets. These included strike directions measured by compass, and

dip angles measured with a spirit-level clinometer. The photographs were

taken with a Nikonos II underwater camera. Sediment samples were recovered

by short hand-operated corers in order to minimize sampling errors. Finally,

pebbles and cobbles as well as bed-rock chips were collected for closer

inspection.

Laboratory work (viz. Miiller 1967) included dialysis, drying and sieving

of the sediment at half-phi intervals. Each fraction was then inspected under

a binocular microscope. Seven different component groups were identified,

separated, counted and weighed. Very small numbers of coralline algae and

worm tubes, too few to allow accurate weighing, are grouped together with

bryozoans. Similarly, even smaller numbers of unidentifiable fragments are

grouped with cirripeds which constitute 61 per cent of the total sample, thus

avoiding distortions in the final results.

In this way the whole sample was processed down to the 1,5—2,0 phi fraction.

The remaining four fractions total only 0,23 per cent by weight and the indi-

vidual. components were too small to weigh accurately. All presentations are

therefore true weight percentages of the total sample. The grain size parameters

were determined according to the graphic measures proposed by Folk & Ward

(1957) and Friedman (1962).

MORPHOLOGY

Rocky Bank rises rather steeply from the surrounding sea-bed reaching

a maximum slope gradient of 1 : 7. At about 40 m below sea-level the slope

begins to level off into a rugged plateau which gives the outcrop a guyot-like

appearance interrupted only by a small central peak which reaches 22 m below

surface (Fig 1).

The surface area of the plateau, enclosed by the 40 m depth contour,

amounts to 7,65 km? or 32 per cent of the total surface area projected by the

65 m contour amounting to 24 km?. When plotting the relative frequencies of

projected surface areas at 5 m intervals (Fig. 2) a first sharply defined maximum

is reached between 35 and 40 m (A). A second less conspicuous maximum lies

between 55 and 60 m (B). Two platforms are therefore clearly defined.

36 ANNALS OF THE SOUTH AFRICAN MUSEUM

2 2Sts«80 40 50 60 70

WATER DEPTH [m]

Fig. 2. Frequency histogram of projected surface areas at 5 m intervals.

Maximum A— 35-40 m platform

Maximum B— 55-60 m platform

GEOLOGY

The in situ investigation confirmed the previous assumption that Rocky

Bank consisted of Table Mountain Sandstone (Gentle 1970). This was by no

means obvious as it would not have been surprising if it were actually found

to be a granite outcrop (Fig. 12).

At —35 m the submerged ‘landscape’ gave the impression of a rugged

yet relatively level plateau dissected by shallow irregular gullies (Fig. 3). The

sandstone formation dips gradually to the south-west with a general north-

westerly to south-easterly strike. The gullies do not reflect the strike of the

bedding but run more or less at right angles to it. It seems hardly coincidental

that the orientation of these gullies is parallel to the direction of the dominant

south-westerly swell which is refracted by Rocky Bank (Shipley 1964). It is

possible that at some stage wave action has selectively eroded a major joint

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM Si

Fig. 3. The rugged yet relatively level plateau at —35 m.

direction. The plateau is loosely covered by well-rounded pebbles, cobbles

and boulders most of which are overgrown by calcareous algae and other

marine organisms (Fig. 4). Biogenic sediments, often forming patches of several

Square metres, are usually confined to the gullies (Fig. 5) but occasionally

occur between boulders also (Fig. 4).

Side-scan sonar recordings show that all of Rocky Bank consists of sand-

stone and it remains uncertain at what depth the contact to the granite takes

place. From 60 m downward the whole outcrop is mantled by a belt of sediment

attaining a fan-like shape especially along the eastern flank. Boomer profiles

record this sediment fan as a transparent wedge. However, they also reveal

an abrupt change in the slope of the bedrock at 90 m below sea-level, possibly

indicating a third wave-cut platform.

SEDIMENTS

The sediments of Rocky Bank are almost entirely of biogenic origin; the

terrigenous content being merely 0,37 per cent by weight. The distribution of

this sediment cover, consisting predominantly of coarse sands and fine gravels,

is thin and patchy with a tendency to sheet accumulations in low depressions

and gullies. This paucity of sediment can only be explained by low production

rates and probably gradual transport across the platform into deeper water

by the combined forces of wave action, gravity and possibly currents. Sediment

movement is in fact clearly illustrated by well-defined ripple marks that reach

heights up to 15 cm and crest-to-crest distances well in excess of 60 cm (Fig. 5).

ANNALS OF THE SOUTH AFRICAN MUSEUM

tlie, # iii = a i

Fig. 4. Well-rounded pebbles and cobbles with coarse biogenic sediment patches

over bedrock.

Fig. 5. Gully in bedrock with coarse biogenic sediment.

The ripple marks indicate water movement.

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM a0

The Recent sediments along the lower slopes of Rocky Bank should

therefore consist essentially of fragmented carbonate skeletons produced on

the upper reaches of the bank, becoming increasingly diluted by pelagic sedi-

ments. Further down in the sedimentary column of the fan coarse sands,

pebbles, cobbles and boulders, produced in the high energy environments of

a lower sea-level, can be expected.

Pertinent grain size parameters are listed in Table 1 below. The lognormal

grain size distribution is demonstrated by the identical values for median,

mean and mode.

TABLE 1.

Grain size parameters

Median . , : : : : : — 0,90 phi (very coarse sand)

Mean (Folk & Ward 1957) : : ; : — 0,90 phi ( - )

Mode . : : ; ; E : ; — 0,90 phi ( zd )

Std. Deviation (Folk & Ward 1957) . : .+ 1,02 phi (poorly sorted)

Skewness (Friedman 1962) ; : : ‘ — 0,03 (nearly symmetrical)

Kurtosos (Friedman 1962) : : : : + 0,69 (platykurtic)

The corresponding cumulative curve and frequency histogram are presented

in Figure 6A and B respectively. Poor sorting is a common feature of most.

biogenic sands when processed mechanically by sieving. Settling results show

that the same sediment is hydraulically equivalent to well-sorted coarse quartz.

The weight distribution of the sediment components is threefold, being

dominated by cirriped fragments (61 per cent). A second group is formed by

bryozoans (20 per cent) and molluscs (15 per cent) and a third group is made up

of echinoderms (2 per cent), corals (1,2 per cent), quartz (0,37 per cent) and

benthic forams (0,2 per cent). Figure 7 illustrates this relative frequency very

clearly. The size distribution of each individual component also follows the

lognormal pattern, indicating their selective response to the hydraulic sorting

process. Only corals deviate from this pattern. They seem to reflect only the

finer tail of an otherwise much coarser population not sampled representatively

(Fig. 8). The contribution of these individual components to the sediment is

summarized in a composite frequency polygon (Fig. 9).

The composition of the sediment allows interesting conclusions regarding

the ecological conditions on Rocky Bank. This, however, is not the aim of

this paper. On the other hand it also provides powerful evidence for a non-

active abrasion platform, an aspect to be discussed in the following section.

DISCUSSION

Although pronounced submerged terraces always suggest wave-cut plat-

forms, such features alone do not justify this conclusion. Tectonic activity

cannot always be ruled out. In the case of Rocky Bank, however, the presence

of well-rounded pebbles, cobbles and boulders provides the necessary evidence

for a wave-cut origin; at least for the upper platform. Too little is known about

ANNALS OF THE SOUTH AFRICAN MUSEUM

"speAIOjUI TYd-% 7e wIeIsO\sTy AOUONbely *g ‘1oded AzIqeqoid UO 9AIND OANLINUIND “VY 9 ‘314

gq Vv

pe 20 bey ser r- pt ee ce” WD 30> = = =

0”

% YyBiam

“wind

iyd sad

% WYyBIem

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM 41

FORAMS.

QUARTZ

CORALS

ECHINODERMS

MOLLUSCS

BRYOZOANS

CIRRIPEDS

S 2 2 ) & 2 2

weight %

Fig. 7. Relative frequency of individual component groups.

the two deeper platforms to allow the same interpretation. However, they do

show a striking coincidence with Pleistocene sea-levels discussed elsewhere in

the literature (e.g. Emiliani 1955, 1966; Donn, Farrand & Ewing 1962; Broecker

1971; Shackleton & Opdyke 1973; Evans 1974). Furthermore, at similar levels

beach deposits and extensive terraces have been recorded off the east and west

coast of southern Africa on several occasions (Bremner 1973; Flemming 1975).

Having established the wave-cut origin of the upper platform, criteria

have to be sought to establish whether this platform is being actively abraded

today or whether we are dealing with a relict feature. As early as 1949 Berthoise

suggested that marine abrasion is restricted to the upper few metres of the sea.

From work carried out along the Pacific coast of the U.S.A., Longwell & Flint

(1955) as well as Bradley (1958) reach the conclusion that submarine abrasion

is restricted to the upper 10-12 m. Flemming (1965, 1968) presents geomorphic

evidence for an abrasion base at —10 m and the author of this article has found

sedimentological evidence in the Western Baltic which distinctly defines this

limit at —13 m (Flemming & Wefer 1973). Recent underwater investigations

in Saldanha Bay, South Africa, have revealed similar evidence at —15 m

(Flemming, unpublished). In both cases side-scan sonar showed this to be a

continuous feature. Hence, although the wave base, calculated roughly at

half the wave length, may well reach down to 100 m or more, the above authors

agree that submarine abrasion is restricted to the upper 10-15 m. The unexpected

agreement of abrasion levels observed in extremely diverse marine environ-

ments in various parts of the world indeed suggests a common cause.

The hypothesis that the abrasion platform is essentially a function of the

energy generated by the most frequently occurring waves would seem a plausible

explanation. To test this hypothesis theoretically, the attenuation of orbital

velocities with increasing water depth was calculated for a variety of waves,

42 ANNALS OF THE SOUTH AFRICAN MUSEUM

| 1 D

SAND

aa ECHINODERMS

/ je

, AP : =3 22) 24) omen

CIRRIPEDS \

i a QUARTZ -

e_, x

BRYOZOANS ~~

| c FORAMS. <

TAC

OLLUSCS e

-3 -2 -1 QO +1

phi—units

Fig. 8. Size distribution of individual component groups.

= se -1 0 +1

utilizing the method of Rankine (Zenkovich 1967). Figure 10 presents a

theoretical velocity vs. depth diagram for deep water waves with periods between

5 and 15 seconds and heights from 1 to 10 metres. Not shown in the diagram

are the various frequencies at which these individual waves occur since they

might change from place to place. For South African waters frequencies for

wave heights and zero-crossing periods were calculated for the depth ranges

0-15 m and > 15 m using data compiled by Darbyshire & Darbyshire (1964),

Darbyshire & Pritchard (1966) and Shillington (1974). The 15 m division line

was chosen because it demarcates the observed abrasion base in this area.

Two threshold velocities were arbitrarily selected from the empirical data of

Hjulstrém (1935), modified by Sundborg (1956), and modified by Allen (1965).

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM 43

20 % (weight)

CIRRIPEDS

BRYOZOANS

MOLLUSCS

ECHINODERMS

CORALS

QUARTZ

FORAMS

phi -units

Fig. 9. Sediment components displayed in a composite frequency polygon.

The first, 40 cm/sec, would be sufficient to move residual sediments composed

of coarse sands and fine gravels, a size range commonly encountered on abrasion

platforms. The second, 100 cm/sec, was chosen following the assumption that

higher velocities would move this sediment with increasing violence, thereby

Sharply increasing the abrasion rates.

44 ANNALS OF THE SOUTH AFRICAN MUSEUM

The resulting figures, displayed in Figure 10 (inset), strongly support

the suggested hypothesis. Velocities above 100 cm/sec are practically non-

existent below 15 m and even the threshold velocity of 40 cm/sec occurs at a

frequency of only 15 per cent below this depth. The conclusion that the boundary

between abrasion and sedimentation reflects a delicate equilibrium maintained —

only by the constancy of the wave spectrum over a long period of time, seems

justified. Furthermore, the evidence for very similar depth levels of abrasion

bases reported from various parts of the world suggests that wave spectra are

very similar on a global scale. A recent compilation of long-term wind and

wave data on a world-wide scale by Davies (1972) indicates that this approach

certainly points in the right direction. From the above considerations we can

conclude that Rocky Bank cannot represent an active abrasion platform under

the present hydrological regime. This is also reflected by the sediments of

Rocky Bank. Quartz, which would certainly form a major sedimentary com-

ponent if appreciable abrasion took place, contributes a mere 0,37 per cent to

the total mass of the sediment (Fig. 9). Furthermore, all cobbles with diameters

CRITICAL FREQUENCY BELOW

PICK-UP VELOCITY OF ABRASION THRESHOLD

(CM/SEC) 0-15m (%)| >15m(%) (%)

(=)

(—)

ORBITAL VELOCITY [cm/sec]

(=)

(—)

IIS

10 SS

0 20 70

[m]

Fig. 10. Theoretical velocity vs. depth diagram of deep water waves with periods between

5 and 15 seconds and heights from 1 to 10 metres.

Inset: percentages of wave frequencies for 40 cm/sec and 100 cm/sec threshold velocities.

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM 45

> 10 cm are entirely coated by encrusting algae that show no sign of wear,

indicating that pick-up velocities for this grain size are rarely if ever achieved.

Yet they are very well rounded (Fig. 11).

To explain the wave-cut origin of Rocky Bank, a sea-level at least 20-25 m

below the present must be assumed. Wright (1964) and Murray et al. (1970)

have traced a submerged cliff at —20 m along parts of the South African west

coast and ascribe this feature to the post-Pleistocene transgression. The previous

Pleistocene event with sea-levels permitting abrasion at this depth dates between

c. 110000 and 80000 B.P. (Shackleton & Opdyke 1973). However, it seems

doubtful whether 70 000 years of sub-aerial erosion would have left intact the

features discussed above. The platform could well have been pre-formed at

that time but its present appearance would nevertheless be a modification due

to the Flandrian transgression. To develop a pronounced abrasion platform

the sea-level must have been reasonably stable for some time around —20 to

—25 m or at least risen at a reduced rate. Inflection points in the rising sea-level

curve have in fact been suggested by a number of authors at about 9 000 years

B.P. (e.g. Morner 1971). Whether 1 000-2 000 years would be sufficient to

render a platform of this dimension, however, remains a matter of argument.

Recent C™ dates from South Africa suggest that during the Wirm I/II Inter-

stadial the sea-level actually rose to c. — 25 m (Tankard, pers. comm. 1975)

and not merely to — 50 m as suggested by Shackleton & Opdyke (1973). If

this could be substantiated then the observed wave-cut platform (Fig. 12)

would certainly be adequately explained.

Fig. 11. Rock samples.

Top—well-rounded pebbles and cobbles.

Bottom— bedrock chips.

ANNALS OF THE SOUTH AFRICAN MUSEUM

"Ul GE — Je aSeq UOIseIqe dAljdsadsoI oY} PU LU OZ — J¥ [9Ad]-BOS JOJO OY} BUIMOYsS ‘;y/W SUIT SuoTe yueg Kyooy YSNOIY} UOI9S-SSOID “Z| “31

+ + +

JALINVYD AdVD + +

+ + 4+

ANOLSONVS “LW 31798V1L

+ + + + +

eseq uoiseiqe juaseaid

JAX] eas jJUasaid

MNVG ANION

ROCKY BANK—EVIDENCE FOR A RELICT WAVE-CUT PLATFORM 47

A final point in this discussion deals with the problem of inferring former

sea-levels from raised marine deposits. First of all, raised beaches must be

distinguished from raised terraces. A raised beach always allows the definition

of the related sea-level, but it will not indicate whether it was stable or not.

A raised terrace on the other hand always implies a relative stable sea-level

but it does not automatically define the related sea-level unless it is known with

certainty which part of the abrasion platform it represents. As we have seen,

the sea-level can be anywhere up to 15 m above a wave-cut platform.

SUMMARY

An underwater investigation, carried out in 1973, revealed that Rocky

Bank, a sandstone outcrop at the mouth of False Bay, South Africa, represents

a wave-cut platform. This is demonstrated by the relatively large surface area

taken up by the 35 m to 40 m depth contours as well as the loose cover by well-

rounded pebbles, cobbles and boulders. The platform is dissected by numerous

shallow gullies running parallel to the direction of the swell. They were probably

eroded by wave action. The sediments are almost entirely of biogenic origin,

indicating that submarine abrasion of the sandstone platform is virtually

non-existent under present hydrological conditions.

This is further supported by the theoretical calculation of orbital velocities

from various wave parameters. Wave frequencies in South African waters

clearly demonstrate the concentration of wave energy in the upper 15 m of

the water column. Similar depth levels of abrasion bases (10-15 m) reported

from various parts of the world suggest that the constancy of wave spectra

over long periods of time are similar on a global scale. The wave-cut platform

is thought to be essentially a feature of the Flandrian transgression, although

recent C™ dates would not exclude the possibility of a Wiirm I/II Interstadial

event.

The problem of inferring former sea-levels from raised terraces is pointed

out. It is essential that the position of the exposed terrace relative to its position

on the former abrasion platform is known before a definite sea-level can be

inferred.

ACKNOWLEDGEMENTS

The author wishes to express his appreciation to all those members of

the University of Cape Town Underwater Club who assisted in the diving

venture. Special thanks are due to Dr T. F. W. Harris, Department of Oceano-

graphy, University of Cape Town, for his advice in matters concerning physical

oceanography.

REFERENCES

ALLEN, J. R. L. 1965. A Review of the origin and characteristics of Recent alluvial sediments.

— Sedimentology 5: 89-191.

BERTHOISE, L. 1949. L’erosion marine et la formation des galets.—C. r. hebd. Séanc. Acad.

Sci., Paris 229: 841-843.

48 ANNALS OF THE SOUTH AFRICAN MUSEUM

BRADLEY, W. C. 1958. Submarine abrasion and wave-cut platforms.—J. Geol. 69: 967-974.

BREMNER, J. M. 1973. Cruise Report—R. V. Thomas B. Davie: Cruise 283,13.03.—09.04.73.

—Jt. geol. Surv. Univ. Cape Town Mar. Geol. Progm.

BROECKER, W. A. 1971. Calcite accumulation rates and glacial to interglacial changes in

oceanic mixing.— Jn: TUREKIAN, K. K. ed. Late Cenozoic glacial ages. New Haven:

Yale University Press.

DARBYSHIRE, J. & DARBYSHIRE, M. 1964. Wave observations in South African waters.—S. Afr.

J. Sci. 60: 183-189.

DARBYSHIRE, M. & PRITCHARD, E. 1966. Sea waves near the coasts of South Africa.— Dt.

hydrogr. Z. 15: 218-225.

Davies, J. L. 1972. Geographical Variation in Coastal Development. Edinburgh: Oliver & Boyd.

Donn, W. L., FARRAND, W. R. & Ewinc, M. 1962. Pleistocene ice volumes and sea-level

lowering.—J. Geol. 70: 206-214.

EMILIANI, C. 1955. Pleistocene Temperatures.—J. Geol. 63: 538-578.

EMILIANI, C. 1966. Paleotemperature analysis of Caribbean cores P 6304-8 and P 6304—9

and a generalized temperature curve for the last 425 000 years.—J. Geol. 74: 109-126.

Evans, P. 1974. Late Pleistocene chronology and the glacial-interglacial cycle.— Geol. Mag.

111: 421-430.

FLEMMING, B. W. 1973. The Geology of Rocky Bank.—Tech. Rep. jt. geol. Surv. Univ.

Cape Town Mar. Geol. Progm. No. 6: 78-82.

FLEMMING, B. W. 1975. Cruise Report—R. V. Thomas B. Davie: Cruise 311, 20.01.—02.02.75.

—Jt. geol Sury. Univ. Cape Town Mar. Geol. Progm.

FLEMMING, B. W. & WEFER, G. 1973. Tauchbeobachtungen an Wellenrippeln und Abrasions-

erscheinungen in der westlichen Ostsee.— Meyniana 23: 9-18.

FLEMMING, N. C. 1965. Form and relation to present sea level of Pleistocene marine erosion

features.—J. Geol. 73: 799-811.

FLEMMING, N. C. 1968. Derivation of Pleistocene marine chronology from morphometry of

erosion profiles.—J. Geol. 76: 280-296.

Foik, R. L. & Warp, W. C. 1957. Brazos River Bar: a study in the significance of grain size

parameters.—J. sedim. Petrol. 27: 3-26.

FRIEDMAN, G. 1962. On sorting, sorting coefficients, and the lognormality of grain size distri-

bution of sandstones.—J. Geol. 70: 737-753.

GENTLE, R. I. 1970. Pre-Quaternary Geology of the Continental Margin between Cape Infanta

and Cape Town.— Tech. Rep. SANCOR Mar. Geol. Progm. (3).

HJULstTROM, F. 1935. Studies of the morphological activity of rivers illustrated by the river

Fyris.— Bull. geol. Instn. Univ. Upsala 25: 221-527.

LONGWELL, C. R. & FLINT, R. F. 1955. Physical Geology. New York: Wiley & Sons.

Morner, N.-A. 1971. The Holocene Eustatic Sea Level Problem.—Geologie Mijnb. 50:

699-702.

MULLER, G. 1967. Methods in Sedimentary Petrology. Stuttgart: Schweizerbart.

Murray, L. G., JoyNT, R. H., O’SHEA, D. O. C., Foster, R. W. & KLEINJAN, L. 1970. The

geological environment of some diamond deposits off the coast of South West Africa.

In: DELANY, F. M. ed. The Geology of the East Atlantic Continental Margin 1. General

and Economic Papers.— Rep. Inst. geol. Sci. 70 (13): 119-141.

SHACKLETON, N. J. & OppykE, N. D. 1973. Oxygen Isotope and Palaeomagnetic Stratigraphy

of Equatorial Pacific Core V 28-238: Oxygen Isotope Temperatures and Ice Volumes

on a 10° year and 10° year scale.— Quatern. Res. 3: 39-55.

SHILLINGTON, F. A. 1974. Surface Waves near Cape Town. 1. Measurements and Statistics.

Unpubl. M.Sc. thesis, University of Cape Town.

SHIPLEY, A. M. 1964. Some aspects of wave refraction in False Bay.— S. Afr. J. Sci. 60: 115-120

SUNDBORG, A. 1956. The River Klaralven: a study of fluvial processes.— Geogr. Annl. 38:

127-316.

WRIGHT, J. A. 1964. Gully pattern and development in wave-cut bedrock shelves north of the

Orange River mouth.— Trans. geol. Soc. S. Afr. 67: 163-171.

ZENKOVICH, V. P. 1967. Processes of Coastal Development. New York: Wiley & Sons.

‘ a, so

L°7EVOLUTION A LA FIN DU PLEISTOCENE ET A L?’HOLOCENE DU

LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES

Par

GEORGES KOUYOUMONTZAKIS & PIERRE GIRESSE

Laboratoire de Géologie, Université de Brazzaville

(Avec 7 figures et 1 tableau)

RESUME

L’étude entreprise ici devra étre considérée comme une mise a jour du travail effectué par

les auteurs portugais sur les cétes de l’Angola.

On essayera de dégager les chronologies anciennes basées sur les sequences méditerranéen-

nes et de les actualiser en les parallélisant avec les séquences utilisées en Afrique de l’?Ouest

(Mauritanie).

Les arguments sédimentologiques permettent une reconstitution paléogéographique

précise, et, plusiers mesures au C* servent de base a la chronologie des lignes de rivage

surélevées.

Les faunes et les microfaunes nous permettent pour chacun de ces niveaux de nous rendre

compte que la température de l’océan semblait plus élevée que de nos jours, et ainsi d’émettre

des hypotheses sur la pérennité du courant froid de Benguela au cours de l’époque envisagée.

ABSTRACT

THE EVOLUTION AT THE END OF THE PLEISTOCENE AND HOLOCENE OF THE ANGOLAN COAST OF

LOBITO-BENGUELA AND MOSSAMEDES

The present study should be considered as a survey of research carried out by Portuguese

workers on the Angolan coast.

An attempt is made to depart from former chronologies based upon Mediterranean

sequences and to bring them up to date by comparing them with those used in West Africa

(Mauritania).

Sedimentological evidence allows detailed reconstruction of the palaeogeography and

several C" assays form a framework for the chronology of the raised beaches.

In each level the fauna and the microfauna suggest that sea temperature was higher than

it is today and thus allow the formulation of hypotheses concerning the permanance of the cold

Benguela current during the period under discussion.

TABLE DES MATIERES

PAGE

Introduction . . es 50

ieenCadnre séographique actuel SEA hie So) ge oe 50

2. Cadre tectonique et géologique . . Si

3. Aspects sedimentologiques des dépdts des terrasses

NC HOEES 220) mises 5 5 5 6 Sy

INIORPIOLOSICs sae eens ee el eee 52

Mingralogicumn m+ -e er et ee 52

Les carbonates . . ; 55

4. Etude paléontologique et micropaléontologique ; 56

Resionide Mossamedes*s= =. )s)- 56

ILES (STRSSES ITAUIIES «on 6 5 5 56

ILES (GUMISSES DERSES 6 5 6 5 5 4 co 3)//

Régionde Lobito-Benguela. . . . . . 60

Conclusionspy es oe hee se ce RS 63

RESIN CES i) ee a Ie Mae ag a a 66

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 49-67, 7 figs, 1 table.

50 ANNALS OF THE SOUTH AFRICAN MUSEUM

INTRODUCTION

Notre étude considérera tout particuli¢rement les deux secteurs de Lobito-

Benguela d’une part et de Mossamedes d’autre part, mais dans le cadre plus

général de la cote atlantique.

Nos observations, en effet, ont pour objet examen de l’€volution eustatique

au Pléistocéne final et a l’Holocéne sur la cOte africaine sud atlantique et

linterprétation en particulier de la céte angolaise a la lumiére des phénoménes

des secteurs voisins.

Les études micropaléontologiques et sédimentologiques se sont souvent

appuyées sur la connaissance des facteurs actuels, notre postulat a priori admet-

tant dans cette région, une certaine chronicité des conditions physiques depuis

au moins la fin du Pléistocene. Les modifications n’intervenant qu’a l’échelle

dune évolution progressive.

1. CADRE GEOGRAPHIQUE ACTUEL

Le littoral angolais procede de différentes zones climatiques fonction de la

latitude:

—le secteur au nord de Luanda est régi par un climat tropical peu humide

(type soudanien) que l’on retrouve a quelques nuances pres sur la céte

angolaise.

—de Luanda a Mossamédeés, des vents secs alizés interviennent et condition-

nent un climat semi-aride de type sénégalien.

—a la limite du Sud-Ouest Africain et de l’Angola, l’aridification s’intensifie,

la température s’abaisse, un climat désertique de type saharien est en

place.

La dynamique des eaux littorales est dirigée par le courant froid de Benguela

orienté du Cap vers l’Equateur. Le courant a une influence directe sur la zonation

climatique que nous venons de décrire, mais encore sur la température des eaux

qui, sur une verticale donnée, présente des variations saisonniéres de plus en plus

contrastées en allant vers le Nord.

Pendant 1|’été austral, les eaux chaudes guinéennes peuvent descendre

jusqu’a Luanda. Pendant l’hiver, les eaux froides et salées du courant de

Benguela remontent jusqu’au Gabon et sont donc prépondérantes dans la zone

considérée.

L’installation des eaux froides est synchrone de la mise en place de vents de

terre qui poussent les eaux de surface vers le large, facilitant ainsi les arrivées

d’eaux profondes, riches en matiéres nutritives.

On concoit donc l’importance considérable de la dynamique de ce courant

de Benguela quant a4 la composition de faunes benthiques et pélagiques de la

plate-forme.

Des variations de la progression vers le Nord de ce courant, peuvent

découler de trés importants changements de population lors de histoire

quaternaire récente.

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 51

Les vents cOtiers dominants sont d’origine SW et déterminent une houle

progressant vers le Nord. Comme pour les cotes congolaises, on observe de trés

fréquentes fléches littorales orientées du Sud au Nord. Les plus remarquables

par leur extension sont celles de Baia dos Tigres, de Porto Alexandre, et celles

qui abritent les eaux du port de Lobito et surtout de Luanda ow les bancs

sableux servent de fondations a certaines installations portuaires.

Du point de vue morphologie, les affleurements tertiaires et secondaires

généralement assez meubles, forment des falaises souvent érodées ot l’escarpe-

ment est assez fréquent. Parfois, les couches basaltiques post-turoniennes

couronnent les reliefs et forment cuestas. Les formations ont pu étre aisément

entaillées par l’érosion des mers quaternaires qui ont laissé des surfaces

d’abrasion assez remarquablement étagées jusqu’a 145 m (Soares de Carvalho

19615), celles de 20 m et 40 m étant assez réguliérement visibles.

2. CADRE TECTONIQUE ET GEOLOGIQUE

Les deux bassins sédimentaires étudiées ici sont situés sur une cOte quasi

rectiligne de direction générale NNE-SSW depuis le 11° de latitude Sud

jusqu’au 17° de latitude Sud soit entre Lobito et Port Alexandre sur prés de

800 km.

Les deux unités sont séparées par un affleurement de socle précambrien au

niveau du Cabo Santa Maria, nous nommerons celui du Nord bassin de

Benguela—Lobito, et celui du Sud, bassin de Mossamédes.

Au nord du Précambrien de Cabo Santa Maria, les séries sédimentaires

crétacés et cénozoiques affleurent avec un pendage de 20° vers le NW et forment

de légéres structures plissées dont les axes anticlinaux et synclinaux sont NE-SW,

ces plissements sont accompagneés de failles de méme direction ou de direction

NS (Mascarenhas Neto 1960); ces failles ont joués durant le Miocéne, et a des

époques bien plus récentes car les auteurs portugais signalent le ‘Pléistocéne’

sensu lato a des altitudes variant entre quelques metres et plus de 150 m; ainsi

Soares de Carvalho (1960) cite un conglomérat a Arca senilis L. et industrie

paléolithique de facture acheuléenne au Sud de la Ponta de Sombreiro: ‘a topo

dos depositos tem cotas da ordem dos 155 metros.’ Encore que ce dép6t trés

€levé doive sans doute étre daté du Pléistocéne (Clark 1963).

On peut donc penser a un systéme de flancs de synclinaux ou d’anticlinaux

faillés en marches d’escaliers qui portent ‘le Pléistocéne sensu lato’ a des altitudes

aussi hautes; ce Pléistocéne doit correspondre a un Tafaritien ou a la rigueur a

un Aioujien car des datations au C"* dans le région de Baia Farta donnent des

ages supérieurs ou égaux a 35 000 ans BP.

Au Sud du Cabo Santa Maria, les affleurements quaternaires existent dans

la région de San Nicolau et dans celles de Mossamédés et Port Alexandre, la

seule tectonique qui leur soit associée est un faillage presque Nord Sud bordant

le socle. Cependant, nous relevons le fait que des faunes 4 Ostréides, Arca

senilis L. et Anomia ephippium L. aient été trouvées dans le conglomérat de

52 ANNALS OF THE SOUTH AFRICAN MUSEUM

Chalunga a des altitudes comprises entre 170 m et 180 m (Soares de Carvalho

19615: 156).

Dans la région de Port Alexandre un beach rock a été trouvé a une altitude

comprise entre 40 m et 50 m et contient des faunes a Ostrea sp. L. et Rotula cf.

augusti Klein, cet affleurement pourrait correspondre a une terrasse “Tyr-

rhénienne’ au sens que l’emploient les auteurs portugais c’est-a-dire a l’ Aioujien,

ou, a un Inchirien.

3. ASPECTS SEDIMENTOLOGIQUES DES DEPOTS DES TERRASSES

INFERIEURES A 20 METRES

Les analyses ont portés successivement sur les plages actuelles et celles plus

anciennes surélevées et en vis a vis. Comme indiqué en introduction, les condi-

tions physiques de sédimentation seront mises en parallele et supposées peu

variables. :

Morphologie

La quasi totalité des grains de quartz observés aussi bien, dans les plages

actuelles (jusqu’a Luanda) que sur les terrasses de Lobito-Benguela et Mossa-

médeés sont de type non usé et l’émoussé faconné par l’abrasion marine y est a

peine décelable.

Egalement les caries d’origine pédogénétique ne sont pas remarquées, méme

a Luanda pourtant sous climat sub-tropical. Ces observations conduisent a deux

interpretations d’ailleurs complémentaires:

—lLa permanence d’un climat 4a dominante aride depuis le Pleistocene permet

une érosion importante par thermoclastisme dont G. Soares de Carvalho

souligne d’ailleurs le modelé et les figures d’érosion caractéristiques. Un tel

climat a pu présenter quelques variantes, mais son caractére essentiel est demeuré

assez constant.

—Le transport de la roche-mére a la bordure du bassin a été bref: les relais

détritiques sont courts et peu fréquents. A partir d’une érosion mécanique

toujours active, l’épirogénie de la céte et le débit irrégulier des fleuves ont permis

une sédimentation importante des sables. G. Soares de Carvalho avait déja

souligné la réunion de conditions favorables 4 une accumulation de considérables

réserves détritiques, notamment lors de la régression préflandrienne.

En fait, il semble que l’accumulation rapide soit intervenue aussi bien avant,

qu’apres cette régression.

Minéralogie

La plupart des plages de l’Angola sont caractérisés par la dominance des

minéraux verts (€pidote, hornblende et parfois hypersthéne et olivine). Les pour-

centages relatifs des deux espéces cardinales (épidote et hornblende) permettent

d’établir des divisions locales sur lesquelles nous n’insisterons pas ici.

LE LITTORAL ANGOLAIS DE LOBITO—-BENGUELA ET MOSSAMEDES

"BIOBUY UD SOgSI[I]N SOgIIVA SOISO[OUOIYS sep UOSTeIedUIOD “| IIqQeL,

NAllLlIdVsVi

pBunjpy) ap Josawojbuo7 3a

JDWDJIW ap sajgos

valuayssJAl 9}UV

IN, SHEL NY (al Dd \/

O01

VEEEZELEDENYV

Sap auiDsso|

Nae MO Oly NSINSHYYAL L NSINSHYSAL

WANMO IF

L- 0s

¢ NSINSHYYAL

O€

WadNM OAN

NSIHEEINMEEINR)

auualljng aig voissasbay

Zv NV

LSSSVSESNV

|ZS1SOS67 87 LYNV

oo0se < 9yNV

LZNV

oo0se < 22NV

aJPUOXa}y 40g

S\N se) OO 18S Sh die

N sli Gill ial DN] NaIF INO

04140) ojanbuag

N SHO DD ©

auUuaIp/DWIID Volssasbay

NMG NW fof Dt

+O0V0E:9LNV

SLNV

OOL +0€VE-VLNV

SUJepPOW “LENV

JIS0d04d JIDOTOWO YN)

SDYDUO) Sap DIDI

N3AIlLLOHOYVNON NOISSSYOSNVASL

N3AIYGNV 14

0001 * a4

AISOIONINY SL snosay Sady

a/DUI 4 @AISSaJBeYy aSDYd

IN SII IO) sl 7 Ht

dag Low JISSSYNSNVYL IANLIV

{S3N0-1 Jd INDI s¥-1 000! * 49 SIVDNLYOd SUNILNY $31

Jd JDO TONOY) SNVOSAY SIV GE UVd TWS3d FDO TONOYN)

SDYIU0) Sap DIDI

: 4IN39010H

54 ANNALS OF THE SOUTH AFRICAN MUSEUM

Dans les plages actuelles des poids exceptionnels de minéraux lourds sont

obtenus: les pourcentages par rapport a la fraction légére atteignent 20 a 30 et

méme plus de 50 pour cent dans plusiers cas (fractions fines). Par comparaison,

les sables des plages surélevées jusqu’a 20 métres sont moins riches (1 a 10 pour

cent). Ce sera un des rares aspects permettant une différenciation entre les deux

séries de plages, les plus anciennes sont plus riches en débris coquilliers ce qui

abaisse d’autant le poids relatif des espéces lourdes. De plus, comme les miné-

raux verts sont extrémement altérables, peut-étre qu’un temps de sédimentation

et une diagénése plus longs et un climat plus contrasté ont abouti a l’élimination

de certains d’entre eux.

L’origine de ces minéraux verts est trés voisine de la cote: le trés vaste massif

granitique précambrien longe le littoral depuis Mossamédeées jusqu’a Novo

Redondo. Ce massif présente de fréquentes enclaves de dolérites pigeonitiques.

Enfin sur la falaise marine affleure une couche volcanique souvent basaltique qui

est interstratifiée entre le Turonien et le Campanien; des petits cristaux d’olivine

sont parfois décrits dans les basaltes (Armado) par G. Soares de Carvalho. Les

matériaux d’érosion se concentrent localement en fonction de l’apport irrégulier

des oueds cotiers. |

Les plages constituent une vaste province pétrographique allant vers le

Nord, au moins jusqu’a Luanda malgré l’éloignement du massif éruptif; les

apports fluviaux et surtout importante dérive littorale rendent compte de cette

unité dont la limite nord ne peut étre précisée; en tout état de cause, les plages du

Zaire, du Cabinda et du Congo ont des sables trés différents par l’?émoussé des

grains et les cortéges minéralogiques. Vers le Sud, la limite est située au milieu

de la Baie de Mossamédes dont les plages septentrionales sont tres riches en

minéraux verts alors que celles du Sud (Praia Amelia) voient apparaitre deux

minéraux de métamorphisme peu fréquents jusqu’alors: le grenat et landa-

lousite. Cette limite correspond a l’embouchure d’un oued important, le Rio

Bero qui descend des reliefs granitiques, traverse les porphyres, et dont les

alluvions sont ensuite entrainées vers le Nord de la Baie.

Il est tres remarquable de constater que cette division de la Baie au niveau

des plages actuelles se retrouve trés fidélement au niveau des plages anciennes

surélévées jusqu’a 15 métres. On démontre ainsi une pérennité pendant la période

correspondante des vents dominants qui orientent la dérive littorale vers le Nord

et de importance des apports du Rio Bero a travers d’éventuelles petites fluctua-

tions climatiques.

_ Les terrasses de Lobito et de Benguela présentent des dépdts eux aussi,

minéralogiquement semblables a ceux des plages actuelles sous-jacentes méme

quand une rubéfaction est intervenue (AN 51).

Les minéraux accessoires soulignent souvent cette distribution générale:

—la staurotide souligne l’importance du métamorphisme au Sud du Rio

Bero, mais encore 4 Luanda éloignée des apports directs du massif éruptif.

Il en est de méme du disthéne.

—la tourmaline, le zircon et le rutile sont aussi plus fréquents 4 Luanda et

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 55

au Sud du Mossamédeés. Leurs teneurs ne s’élevent pas dans les dépéts

anciens et n’indiquent pas un vieillissement du cortege.

L’olivine est présenté parfois dans le secteur Benguela—Lobito tant sur les

plages actuelles que fossiles; elle semble procéder des intrusions basaltiques

crétacées de la céte.

L’hypersthéne n’accompagne que rarement les amphiboles et les épidotes

sauf dans la Baie de Mossamédés (5 a 10 pour cent en fraction moyenne).

L’augite, trés rare, lui semble associée.

La barytine est trouvée dans la terrasse de 20 métres de Benguela (AN 54,

55) ou elle est remaniée a partir des géodes de séries calcaires de |’Albien

inférieur. Ce minéral fragile a été peu dispersé.

Enfin, dans les minéraux opaques ow les altérations ferrugineuses sont

absentes, signalons des ilménites irréguli¢rement concentrées en fonction des

oueds locaux. Une forte teneur en ilménite correspond a une forte teneur

générale en grains lourds.

Les carbonates

Les plages actuelles, mises a part les bandes localisées ot les coquilles sont

concentrées, sont relativement peu carbonatées pas plus de 5 pour cent. sauf

exception.

Les plages anciennes sont beaucoup plus riches: toujours plus de 30 pour

cent. Donc en dépit des analogies physiques constatées plus haut dans les deux

séries de dépdt, un certain refroidissement des eaux est envisageable en allant vers

Pactuel.

Les montages de fraction lourde ont permis d’observer l’aragonite dans la

plupart des sédiments. I] en est de méme pour I’ankérite qui n’existait pas sur les

cotes du Congo: un des niveaux de terrasses (AN 46) prés de Lobito est trés

riche en ankérite. La précipitation des carbonates authigénes (calcites le plus

souvent) est nette dans beaucoup de niveaux de terrasses de Lobito a Mossa-

médés ou les silts quartzeux sont cimentés en petites boules friables (AN 40,

AN 46, AN 52, AN 53, AN 54, AN 57); certains quartz sont méme encroidtés par

les carbonates, phénoméne peut-étre diagénétique que l’on ne voit pas sur les

plages modernes.

Les carbonates d’origine organique sont prépondérants: ce sont des

débris de test de mollusques sous forme de lamelles corrodées et de petite

taille, de tubes de ver, mais aussi des ostracodes et des foraminiféres assez

fréquents.

Des plaques minces ont été effectuées dans les gros tests de mollusques

conserves. Aucune recristallisation de la calcite n’est observée dans les coquilles

des niveaux inférieurs 4 20 m qui nous intéressent surtout ici. Par contre, quel-

ques tests récoltés vers 40 m (niveaux tyrrhéniens 4 Strombes de G. Soares de

Carvalho) laissent voir quelques minéralisations secondaires qui pourrait fausser

des mesures radiochronologiques. Par contre, les mesures au C™ des témoins de

bas niveaux ne semblent présenter aucune source de rajeunissement.

56 ANNALS OF THE SOUTH AFRICAN MUSEUM

En résumé, quels arguments paléogéographiques relatifs aux plages suréle-

vées, peut-on retirer de cette esquisse sédimentologique?

1. Les études morphoscopiques et minéralogiques se rejoignent dans |’essen-

tiel de leur conclusion: les apports sédimentaires ont été rapides, actifs car les

quartz sont frais et les minéraux fragiles surabondants, les oueds compétents

étaient les mémes qu’aujourd’hui et la dérive littorale de méme orientation et de

méme efficacité.

2. Un climat de type aride a controlé la sédimentation et a permis |’érosion

caractéristique, les apports alluviaux irréguliers et locaux et une précipitation de

certains carbonates authigénes pendant la diagénese. Une légeére altération a pu

se traduire par Il’élimination chimique de quelques minéraux fragiles présents

dans le sable actuel.

3. L’abondance de tests calcaires grossiers par rapport a la plage moderne

présume, avant l’inventaire faunistique détaillée de conditions écologiques

favorables lors de périodes ot le courant de Benguela était moins présent

qu’aujourd’hui. Cette limite variable de l’influence du courant de Benguela ayant

d’évidentes conséquences climatiques et donc géodynamiques comme nous

Pavons remarqué en (2). .

4. ETUDE PALEONTOLOGIQUE ET MICROPALEONTOLOGIQUE

Nous décrirons tout d’abord les faunes observeées et leurs habitats dans la

région de Mossamédes puis dans celle de Benguela—Lobito.

REGION DE MOSSAMEDES

Sur une cote NS, la Baie de Mossamédeés est une échancrure semicirculaire

ouverte a l’Ouest et fermée par deux caps, la Ponta de Giraul au Nord et la

Ponta de Noronha au Sud.

Les terrasses hautes

Les sédiments les plus anciens que nous y trouverons ont été décrits par les

auteurs portugais (Soares de Carvalho 19615) et les auteurs belges (Dartevelle

1952) qui avaient défini une terrasse marine haute contenant des Strombus

bubonius Lamk., Arca senilis L., Thais haemastoma L., et des restes d’échino-

dermes sous forme de digitations de Rotulinae; on y a méme trouvé (Faber

1926) des restes de madréporaires (AN 42).

Nous avons observé ces affleurements sur les escarpements de Torre de

Tombo au Sud de la ville et sur les falaises surplombant la Praia Amelia. On y

trouve la faune décrite plus haut accompagnée de restes d’ostréides et des

gastéropodes tels que Bulla aff. adansoni Philippi et Pugilina aff. morio L.

(AN 32, 33, 34).

G. Soares de Carvalho attribue a4 ces coquilles un 4ge Tyrrhénien auquel

nous donnerons le nom d’Aioujien en liaison avec la nouvelle stratigraphie de

Quaternaire définie au Sénégal et en Mauritanie (Elouard & Faure 1967).

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES ii

Nous avons vu plus haut que pres de Port Alexandre, donc 80 km au Sud un

affleurement avec un beach rock aréno-calcaire contenait une faune a Ostrea sp.,

Ostrea aff. folium L., des fragments de madréporaires et des digitations d’échino-

dermes qui pourrait appartenir a Radiorotula orbiculus L. et a Rotula augusti

Klein, nous avons hésité sur l’age de ce dép6t, mais vu l’état de conservation des

animaux, il semble inchirien.

Les terrasses basses

La région de Mossamédeés nous fournira encore de beaux affleurements,

vers Baia des Pipas, sur la route y conduisant ou sur le plateau calcaire de la

Ponta de Giraul et de Praia des Conchas. Mais nous sommes 1a en présence

d’affleurements beaucoup plus récents et qui ont pu étre datés.

Au Sud W de la ville 4 Praia Amelia, au-dessous de la falaise portant

l’Aioujien et 4 +2 m, nous trouvons une plage dont la faune a été datée AN 10/2

(GIF 3220) 1620 + 80 BP, mais cet Age est certainement trop faible car nous

pensons cet affleurement Nouakchottien, les géologues portugais (Soares de

Carvalho) lui donnaient l’age Ouljien, cet Age Ouljien englobant le Nouak-

chottien et l’Inchirien.

A la Ponta do Giraul, Davies (1959) signale un stade Ouljien (Plage V) et un

stade Monastirien (Plage IV) daté par un outillage Acheuléen ancien (30 m).

Le site de Praia des Conchas est supporté par un entablement calcaire a

pendage faible vers le NW et sa surface est parsemée de débris coquilliers, on y a

méme trouvé des fragments d’os de cétacés (5 m).

Nous sommes en présence de trois plages échelonnées suivant leurs altitudes.

A lEst du plateau, une faune typique dont |’état d’usure fait penser a celle des

Fig. 1. Praia des Conchas, Plage nouakchottiene (5 m) montrant des rochers corrodés

par des Echinides (l’echelle mesure 25 cm).

58 ANNALS OF THE SOUTH AFRICAN MUSEUM

falaises de Praia Amelia donc a un Aioujien, puis en descendant en altitude des

Strombus bubonius Lamk. épais, avec Thais haemastoma L. d’age inchirien car

non encrotités commes plus haut et, en bordure de falaise vers 3 4 5 m des rochers

corrodés par des oursins au pied desquels se trouvent les débris coquilliers de la

‘plage 4 Vermetus ouljienne’ de Soares de Carvalho (19615).

La, des datations au C' ont donné sur ces faunes des ages de 3420 + 100

BP (AN 14, GIF 2945) et 3040 + 100 BP (AN 16, GIF 2946), c’est-a-dire

Nouakchottien pour des cotes respectives de +5 et +3 m.

L’échantillon AN 36 prélevé sur le méme gisement contient la faune sui-

vante:

Patella granularis Krauss, Conus aff. ventricosus Gmelin, Thais haemastoma L.,

Cantharus viverratus Kiener, Nerita sp. L., Ostrea folium L., Vermetus adansoni

Daudin.

Fig. 2. Praia des Conchas, dépdts de la plage nouakchottienne (3 a 5 m) a faune de

gastéropodes Patella granularis, Thais haemastoma, et des ostréides (Echelle 25 cm).

En se rapprochant de la ligne de rivage actuelle et se mélangeant avec elle,

nous irouverons une plage appelée ‘plage a Perna perna L.’ dont l’age est Post

Nouakchottien a actuel; on y trouve une faune 4a cirripédes, a gastéropodes:

Patella granularis L., Fissurella sp. Brug., Thais haemastoma L., Monodonta

sagittifera Lamk., Calyptrea trochiformis Gmelin, Opalia crenata L., Cantharus

viverratus Kiener, et a lamellibranches Perna perna L.

Sur la route conduisant 4 Baia des Pipas, nous trouvons un trés riche

affleurement dans des sables argileux (AN 39/40) (15 m) auquel nous donnerons

eee a

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 59

Fig. 3. Plage nouakchottienne a Ostrea folium, Calyptrea chinensis, Thais haemastoma

(3 a5 m) G de la grandeur naturelle).

un age Nouakchottien contenant une faune trés bien conservée de lamelli-

branches et de gastéropodes:

Spisula nivea Gmelin, Mactra largillierti Philippi, Cardium ringens Gmelin, Solen

marginatus Pennant, Natica aff. collaria Lamk., Natica collaria Lamk., Terebra

senegalensis Lamk., Mesalia mesal.

D’autres espéces sont en cours de détermination; de méme ce sédiment a

fourni une microfaune de foraminiféres et d’ostracodes assez abondante:

Nonion asterizans Fichtel et Moll, Florilus boueanus Brady, Elphidium crispum

L., Elphidium complanatum W@Orb., Ammonia beccarii L., Quinqueloculina

disparilis d’Orb., Quinqueloculina lamarckiana d’Orb., Eponides sp. Montfort,

Eponides repandus Fichtel et Moll, Bolivina spathulata Williamson, Globigerina

aff. bulloides d’Orb.

Cette plage semble montrer un milieu infra littoral 4 apport d’eau douce de

maniére intermittente, et la faune que 1’on trouve correspond bien 4 un Nouak-

chottien typique tel qu’il est décrit en Afrique de l’Ouest (Elouard 1968). Ainsi la

présence de plages présentant des faunes identiques de maniére symétrique par

rapport a l’Equateur, Dakar 15° Nord, et Mossamédés 15° Sud, semblerait

montrer l’existence d’une sédimentation et de conditions biologiques constantes

au cours du Quaternaire sur les cdtes d’Afrique. Cependant, se pose le probleme

de la présence au Sud de l’Equateur du courant froid de Benguela dont l’existence

au Quaternaire ancien peut étre discutée.

60 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 4. Route de Baia des Pipas (15 m), faune littorale nouakchottienne milieu infra-

littoral de fond de baie (le marteau mesure 45 cm).

REGION DE LOBITO—BENGUELA

Dans cette région la succession des plages fossiles sera moins bien définie

que dans la région précédente, on retrouvera la plage aioujienne “Tyrrhénienne’ a

120 m (AN 43) sous forme d’un conglomérat concrétionné de graviers et de

coquilles indéterminables a |’Est du phare de Lobito.

Par contre nous n’avons pas rencontré les affleurements signalés par Soares

de Carvalho (1960) et Mascarenhas Neto (1960) aux alentours de la Ponta de

Sombreiro.

Ensuite aux cétes d’altitude comprises entre 8 et 20 m trouvera souvent des

affleurements a faunes diverses, auxquels nous donnerons un Age constant

inchirien AN 46 (GIF 3232) > 35 000.

A la Ponta das Vacas, Davies (1959) signale trois terrasses entre 0 et 25 m, la

plus récente qu’il date du Flandrien (Plage VI) contiendrait un outillage d’Age

Sangoen récent, et est précédé par deux plages plus anciennes, une plage d’age

Ouljien (Plage V) et une autre d’4ge Monastirien (Plage IV).

Dans les alentours immédiats de Lobito au Bairro de Santa Cruz au-dessous

de sables trés fins certainement ogoliens, nous trouvons des fragments coquilliers

d’huitres et de Spisula nivea Gmelin, et de Mactridées (8 4 10 m).

Puis dans la région de Catumbela, une sabliére montre deux passées a Arca

senilis L., Chlamys varius L., Glycimeris sp. da Costa et Ostrea sp. L. séparées

par des sables plus ou moins rubefiés, suivant la succession suivante de haut en

bas:

a: Sable ocre argileux fin, 3,00 m

b: Passée a Arca senilis L. (AN 46 > 35 000) 0,50 m

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 61

c: Sable ocre non argileux, 1,00 m

d: Passée a Arca senilis L., Ostrea sp., Glycimeris sp. da Costa, et Chlamys

varius L. (AN 47, 48, 49), 0,50 m

e: Passée de sable grossier blanc non argileux, 1,00 m

C’est-a-dire un Ogolien ocre dunaire régressif recouvrant un Inchirien a

Arca senilis L. qui présente deux épisodes, I’épisode supérieur étant plus lagunaire

que l’inférieur. Cette hypothése se confirme au SE de la ville de Benguela ot l’on

trouve une passée d’age inchirien présentant une faune a Arca senilis L., Pugilina

morio L. et Ostrea sp. L. prises en masse avec des sédiments de type lagunaire

franc a gypse (AN 54).

La passée supérieure de Il’Inchirien contient suivant les affleurements des

coquilles de Tympanotonus fuscatus L.

La microfaune de la passée inchirien a Arca senilis L. est assez réduite et

contient des animaux de faciés saumatre tel Ammonia beccarii L., assez abon-

dante, et Reophax sp. Montfort.

La microfaune de l’affleurement AN 54 est cependant plus riche en espéces:

Lenticulina iota Cushman, Lenticulina aff. calcar L., Bulimina inflata Sequenza,

Bulimina striata d@’Orb., Nonionella sp. Cushman, Elphidium incertum William-

son, Ammonia beccarii L., Eponides sp. Montfort, Bolivinoides sp. Cushman,

Globigerina aff. bulloides d’Orb.

Entre les deux gisements existe un affleurement (AN 53) dont la faune

semble moins lagunaire et contenant:

Cardium ringens Gmelin, Mactra largillierti Philippi, Natica sp. Scopoli, Mesalia

mesal.

La microfaune est composée de nombreux ostracodes et de foraminiféres

parmi lesquels:

Nonion asterizans Fichtel et Moll, Nonionella atlantica Cushman, Ammonia

beccarii L., Globigerina aff. bulloides d’Orb.

A l’Est de ces affleurements, aux alentours du terrain d’aviation de Benguela

une plage surélevée (20 a 25 m) montre Ostrea sp., Arca senilis L., Cardium

ringens Gmelin, Conus pulcher Lightfoot, ces animaux étant trés faiblement

recristallisés.

Plus au Sud, vers le petit ville de Baia Farta, les affleurements sont portés en

altitude par une néotectonique cassante, on trouve deux affleurements inchiriens,

un situé 4 2 ou 3 m (AN 22) (GIF 2947) > 35 000 ans a Arca senilis L., Glyci-

meris sp. da Costa, Mactra largillierti Philippi, et 4 2 km de 1a, mais 20 m

au-dessous un affleurement (AN 56) dont laltération est de méme aspect

contenant:

Ostrea folium L., Arca senilis L., Smaragdia viridis L., Tympanotonus fuscatus L.

Prés de Baia Farta, a 80 m d’altitude, O. Davies (1959) signale une passée

graveleuse qui pourrait étre une plage intermédiaire entre les bas niveaux de la

Porta das Vacas et les hauts niveaux de la Ponta do Sombreiro.

Le dernier affleurement que nous verrons sera peuplé essentiellement

62 ANNALS OF THE SOUTH AFRICAN MUSEUM

Figs 5-6. Cdte de Baia Azul (prés de Baia Farta secteur de Benguela). Accumulation

d’Arca sensilis L. et de Tympanotonus fuscatus L. et Tympanotonus fuscatus var. radula

(15 m). Cette accumulation est certainement d’age inchirien (4 de la grandeur naturelle).

d’Arca senilis L. et de Tympanotonus fuscatus L., la microfaune a Ammonia

beccarii L. et Ammonia beccarii L. var. tepida, Quinqueloculina sp. d’Orb. et de

nombreux ostracodes a test lisse peut montrer un milieu a début de dessalure;

quant a son age, nous ne pouvons pas le définir exactement comme inchirien

ou nouakchottien (AN 57 BAIA FARTA + 12 m).

La céte de la région de Lobito 4 Benguela, semble donc ne montrer que des

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 63

affleurements inchiriens, peut-étre des dragages effectués dans la baie de Benguela

permettraient de démontrer la présence d’un Nouakchottien.

En résumé, nous avons donc vu entre Mossamédeés et Lobito une con-

cordance entre les faunes. Le Tyrrhénien—Aioujien et l’Inchirien sont toujours

caractérisés par les mémes animaux du Sud au Nord.

Les dép6ts aioujiens appartiennent souvent a des facies littoraux de milieux

chauds de fond de baie, mais sans apport d’eau douce, alors que les dépéts

inchiriens semblent caractérisés par des animaux a tendance lagunaire Arca

senilis L., Tympanotonus fuscatus L., seul ?’Inchirien de Port Alexandre, montre

des échinides et des madréporaires.

Quant au Nouakchottien, dans la région de Mossamédeés ow nous l’obser-

vons, nous lui verrons deux faciés:

—Un facies infralittoral vaseux de fond de baie calme, a apports intermit-

tents d’eau douce (AN 39/40).

—Un facies littoral battu avec des rochers perforés par les échinides, des

patelles et des huitres (AN 38/36).

CONCLUSIONS

L’étagement des diverses lignes de rivages surélevées observees sur les cétes

d’Angola est a peu prés constant; cependant, il est 4 noter que les niveaux les

plus récents, Nouakchottien et Inchirien, disparaissent de l’affleurement en allant

vers le Nord.

L’Inchirien trouvé a Port Alexandre vers 50 m, existera sous un autre

facies entre Baia Farta et Lobito a des altitudes plus basses (+15, -+-25 m).

Quant au Nouakchottien, il n’existera plus au niveau du secteur de Lobito-

Benguela.

Vers. Luanda, Ambriz, seules affleureront les plages aioujiennes, puis

au-dela du fleuve Congo, il n’existera plus que quelques plages holocénes qui

disparaissent vers le Nord.

Les témoins de niveaux transgressifs trouvés au Nord du fleuve Congo

présentant une faune assez bien conservé, l’ont été dans des carottages ou dans

des dragages du plateau et du littoral congologabonais.

Dans les carottages de géotechnique effectués dans le port de Pointe-Noire,

certaines passées entre —7 et —8 m ont montré des faunes de mactridés et

d’ostréides, semblant étre donc Nouakchottien, dont l’Age absolu est de

4920 + 140 ans BP (éch. PG 79, GIF 2209) (Giresse, Kouyoumontzakis 1971).

Au Gabon des carottages et des sondages effectués 4 Port-Gentil ont permis

a Nicklés (1952) de définir une faune de Mollusques quaternaires en deux passées

entre —15 et —20 m et —27 et —32 m qui sont des sédiments de niveaux trans-

gressifs certainement nouakchottien.

Sur le plateau continental, nous possédons une datation pour une carotte de

80 cm prélevée 4 —110 m de fond, la base de celle-ci montre un Age supérieur ou

égal a 35 000 ans (PN 1198 g, GIF 3233), mais les espéces pélagiques contenues

ANNALS OF THE SOUTH AFRICAN MUSEUM

"91X9] S| SUPP SAIOPISUOD S}Qdep So] 39 SIN9}90S SOT *1 “SI

JIIDUJ9}ONH 32 PIIDIZJa] sSjzUaWa Nay FEY 4! /

Wy OL G 0 | S,9I a

J AYONVX31V LYOu gs:

+ 9g NV :

SS NV, NYS E777 NV

LZ‘9Z NV+ 7 VLYV4

“9g VS NV - viva

OJIBIQUIOS

3p D}UOY

V1SNONZAE

EG NV \

S 0€,2; —— ye

ZS‘LG NV+ oy

0s D Gp NV+ a SAG AWVIOW

Y, ) ) DUUOJON

NE y ” DP DIUOd

) A

p 9

J 0€ él OV 6E€ NV/+ nee 9S

EV NY, O1I8S01

SLNV#+/SLNV VISNON 8g g

——

WS 0

4 8E NV+ OLIGO YY Gg

SG!

LE LITTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 65

dans la microfaune de ce sédiment sont en cours de détermination et ne permet-

tent pas de donner la profondeur de dépét de cette association thanatocoeno-

tique. Il peut donc s’agir d’un dépdt formé durant la période transgressive

inchirenne (—25000 a —40000) ou la période régressive aguerguerienne

(—40 000 4 —70 000).

Les seuls dépots de surface que nous avons trouvés et qui semblent corre-

spondre a un niveau récent Tafolien (—4 000 4 —2 000) sont des sables blancs

qui résultent d’un épisode é€olien, de méme la présence de tourbes de mangrove

dans le bassin des fleuves cdtiers et la largeur de leur plaine estuarienne

démontrerait une transgression récente.

Il est donc important de constater que les dépdts quaternaires trouvés en

altitude au Sud de l’Angola, voient leur altitude diminuer et celle-ci devient

négative au Nord du fleuve Congo. Les derniers niveaux quaternaires cités a

terre le sont au Zaire et au Cabinda (Dartevelle 1950; Rémy 1954). Plus au

Nord et jusqu’a l’Afrique Occidentale le ‘quaternaire sl’ n’affleure plus.

Peut-étre des dragages systématiques sur tout le plateau continental entre ie

Gabon et la Céte d'Ivoire permettrait de trouver l’explication de ces phénoménes

et de relier les affleurements sous-marins du Congo avec ceux de la Céte d’Ivoire

(Martin 1973) et de la Guinée et de la Sierra Léone (MacMaster et al. 1970).

Des observations d’ordre faunistique vont nous permettre ensuite de définir

les aires d’extension géographique et chronologique du courant froid de

Benguela.

Les auteurs portugais parlent au fond de la Baia do Chapeu Armado d’un

site Ouljien 4 madréporaires a une altitude de 8 a 20 m.

Cet animal déja cité par Faber en 1926, est rattaché par Carvalho (19615) a

Pespéce Siderastrea siderea Ellis et Solander; Chevalier & Hebrard (1967) citent

dans Il’Inchirien de Mauritanie un Siderastrea radians Pallas daté vers 31050

ans BP.

Ces auteurs pensent que Siderastrea siderea cité par Carvalho pourrait étre

rattaché a l’espéce Siderastrea radians, cette espéce pouvant vivre en milieu a

apport terrigéne, exposé a I’air, au soleil et pouvant subir des abaissements de

salinité et de température dans des conditions exceptionelles. Dans le cas nous

intéressant, la faune collectée sur le méme site est une faune chaude de milieu

infralittoral qui n’existe plus actuellement que dans les eaux guinéennes, ou les

eaux sursalées et surchauffées de baies fermées dans les régions actuellement

baignées par le courant froid de Benguela.

La faune qui accompagne ces madréporaires est de méme composition

qualitative que la faune des affleurements GKAN 14, 16, 36, 39, 40 et paraitrait

donc avoir une age nouakchottien, cependant, en l’absence de datations absolues,

nous sommes amenés a faire deux hypothéses.

(a) Les gisements a Siderastrea radians sont inchiriens.

(b\ Ils sont nouakchottiens.

Si les Siderastrea radians ou siderea sont inchiriens la présence du courant de

Benguela sur les cdtes d’Afrique peut étre infirmée jusqu’a moins 25 000 ans BP,

66 ANNALS OF THE SOUTH AFRICAN MUSEUM

par contre si ils sont d’Age nouakchottien, la présence du courant de Benguela ne

remonterait qu’a 4 ou 5 000 ans BP.

En outre, au cours des récoltes de faune sur les plages du Congo entre le 4°

et le 5° de latitude Sud, nous avons ramassé des Astrangia sp. Quel que soit leur

lieu de récolte, ces animaux n’ont jamais été observés vivants et la plupart du.

temps, sont ou brisés ou recouverts d’une épifaune. De méme au large du littoral

du Congo, au cours de dragages, nous avons récolté, sur un replat a 110-115 mde

profondeurs, des polypiers solitaires semblant appartenir a l’espéce Caryophyllia

clavus Sc. Ces animaux présentent le méme degré d’abrasion que les animaux

trouvés sur la céte; on ne les a jamais trouvés vivants.

Cette “plage fossile’ de —110—-115 m nous a fourni une faune de mollusques

d’eaux plus chaudes qu’elles ne le sont de nos jours, actuellement on trouve ces

animaux sur le bord Nord du Golfe de Guinée, en Cote d’ Ivoire.

Au point de vue microfaunistique, on trouve dans cette ‘plage’ de trés

nombreuses Amphistegina lessonii d’ Orb. qui vivent en général dans des eaux

chaudes et associées a des faciés coralligénes. Ces foraminiféres ont été datés de

l’Holocéne PN 596 (GIF 2564) 11980 + 250 BP, PN 611 (Fontes) 12620 + 240

BP.

Il a donc fallu pour qu’ils vivent 1a que le courant de Benguela n’ait pas

existé a cette €poque, soit entre 12860 et 11730 BP, car la zone datée est comprise

dans sa zone actuelle d’extension.

REFERENCES

BraGA, J. M. 1960. Foraminiferos da costa da Mocambique.— Mems Jta Invest. Ultramar,

Estud. ens. et docum. 67: 1-211.

CHEVALIER, J. P. & HEBRARD, L. 1967. Découverte de Madréporaires dans le Pléistocéne

supérieur de Mauritanie.— Actes VI Congrés Panafricain de Préhistoire: 453-456.

CLARK, J. D. 1963. The distribution of prehistoric culture in Angola.— Proceedings V Pan-

african Congress on Prehistory: 225-309.

DARTEVELLE, E. 1950. La céte et l’estuaire du Congo. — Mém. Inst. r. colon. belge Sect. Sci. nat.

meéd. 19 (2): 1-S8.

DARTEVELLE, E. 1952. Echinides fossiles du Congo et de I’Angola. 1° Partie: Introduction

distorique et stratigraphique. — Annls. Mus. r. Congo belge Ser. 8vo (Sci. géol.) 12: 1-71.

DARTEVELLE, E. 1953. Echinides fossiles du Congo et de l’Angola. 2° Partie: Description

systématique des Echinides fossiles du Congo et de l’Angola.— Annls. Mus. r. Congo

belge Ser. 8vo (Sci. géol.) 13: 1-240.

DAVEAU, S. & RIBEIRO, O. 1967. Bilan des recherches concernant I’étude du Quaternaire en

Angola et au Mozambique.— Actes VI Congrés Panaftricain de Préhistoire: 353-354.

Davies, O. 1959. The raised beaches of Angola and South West Africa.— Proceedings IV

Panafrican Congress on Prehistory: 289-294.

Daviess, O. 1971. Pleistocene shorelines in southern Africa.— Quaternaria 15: 317-323.

DELIBRIAS, G., GIRESSE, P. & KOUYOUMONTZAKIS, G. 1973. Géochronologie des divers stades

de la transgression holocéne au large du Congo.—C. r. hebd. Séanc. Acad. Sci., Paris 276:

1389-1391.

DESBROsSES, J. 1966. Les associations de Foraminiféres sur le rebord du plateau continental au

large de la Céte d’Ivoire.— DES 55 Lab. Géol. Dijon.

ELOUARD, P. 1968. Le Nouakchottien, étage du Quaternaire de Mauritanie.— Annls. Fac. Sci.

Univ. Dakar 22: 121-131.

ELOuUARD, P. & Faure, H. 1967. Quaternaire de l’Inchiri, du Taffoli et des environs de

Nouakchott.— Actes VI Congrés Panafricain de Préhistoire: 446-492.

LE LiTTORAL ANGOLAIS DE LOBITO-BENGUELA ET MOSSAMEDES 67

FABER, F. J. 1926. Concerning the occurrence of Quaternary corals in Angola.— Proc. K. ned.

Akad. Wet. 29: 843-845.

GirRESSE, P. & KOUYOUMONTZAKIS, G. 1971. Géologie du sous-sol du Port de Pointe-Noire et

des fonds sous-marins voisins.— Annls. Univ. Brazzaville: 97-114.

GIRESSE, P. & KOUYOUMONTZAKIS, G. 1973. Cartographie sédimentologique des plateaux

continentaux du Sud du Gabon, du Congo, du Cabinda et du Zaire. Cah. ORSTOM,

Sér. Géol. 5: 235-257.

GONCALVES, F. & ROMAN, J. 1974. Une sous sepéce nouvelle de Rotula orbiculus. L. dans les

formations plio-quaternaires de l’Angola.— Biol .Mus. Lab. min. geol. Fac. Cienc. Lisboa.

5: 99-106.

LoeEBLicH, A. R. & TAPPAN, H. 1964. Protista 2; Sarcodina, chiefly “Thecamoebians’ and

Foraminiferida. In: Moore, R. C., ed. Treatise on Invertebrate Palaeontology: Part C.

Lawrence, Kansas: Geological Society of America and the University of Kansas.

MACMASTER,: R. L., LACHANCE, T. P. & ASHRAF, A. 1970. Continental Shelf, geomorphic

features off Portuguese Guinea, Guinea and Sierra Leone.— Mar. Geol. 9: 205-215.

Martin, L. 1973. Morphologie, sédimentologie et paléogéographie au Quaternaire récent du

plateaut continental invoirien.— These Doct. és Sci. Nat. Paris.

MASCARENHAS NETO, M. G. 1960. Géologie de la région Benguela Cuio (Bande sédimentaire). —

Bol. Serv. geol. min. Angola 1: 89-99.

MASCARENHAS NETO, M. G. 1961. As Bacias sedimentares de Benguela e Mocamedes.— Bol.

Serv. geol. min. Angola 3: 63-93.

MASCARENHAS NETO, M. G. & GRACA DE CRUZ, A. 1960. Geologia da faixa sedimentar entre

a Baia dos Elefantes e o Cabo Santa Maria.— Bol. Sery. geol. min. Angola 1: 9-36.

NickLes, M. 1950. Mollusques testacés marins de la Cote Occidentale d’ Afrique. Paris: Le

Chevalier.

Nick ies, M. 1952. Mollusques du Quaternaire marin de Port-Gentil AEF.— Bull. dir. mines

Géol. Afr. equat. fr. 5: 73-101.

PHLEGER, B. F. 1965. Ecology and distribution of recent Foraminifera. Baltimore: Johns Hopkins

University Press.

Remy, J. M. 1954. Contribution a l’étude de la terrasse marine de la Pointe Kudevele (Congo).

— Rey. Zool. Bot. afr. 48: 24-26.

SOARES DE CARVALHO, G. 1960. Sobre los depositos cretacicos do litoral de Angola.— Bol. Serv.

geol. min. Angola 1: 37-48.

SOARES DE CARVALHO, G. 196la. Alguns problemas dos terracos quaternarias de litoral de

Angola.— Bol. Serv. geol. min. Angola 2: 5-15.

SOARES DE CARVALHO, G. 1961b. Geologia do deserto de Mocamedes.— Mem. Jta Invest.

Ultramar 26: 1-227.

TAVARES Rocua, A. & UBALDO, M. L. 1964. Foraminiferos do terciario superior e do quater-

nario da provincia Portuguesa de Timor.— Mems Jta Invest. Ultramar 5: 1-180.

HOLOCENE SUBMERGENCE ON REUNION ISLAND

(INDIAN OCEAN)

L. MONTAGGIONI

Centre Universitaire de la Réunion—France D.O.M. and

Centre d’Océanographie d’ Endoume— Marseille

(With 5 figures and 1 table)

ABSTRACT

A series of dates obtained from samples of coral taken from a core drilled into the fringing

reef of Réunion Island shows that during the last 7 300 years sea-level has risen continuously,

although at a progressively decreasing rate, until it reached its present position. Between

7 300 and 5 600 years B.P. sea-level rose at the rate of 0,59 cm/year; from 5 600 to 3 500

years B.P. the rate of rise was about 0,29 cm/year. Since 3 500 years B.P. the sea has trans-

gressed at a rate of 0,10 cm/year.

The clear increase in the rate of transgression relative to the rates suggested in other

areas could be the result of compensatory eustatic processes which tended to offset the delayed

beginning of the Holocene transgression in this part of the Indian Ocean. Astronomic and/or

oceanographic factors would account for the obvious differences observed in the eustatic

movements better than unusually rapid subsidence of the island.

CONTENTS

PAGE

IimtroductiOne “.o.cy ee 1 aha Ss 69

Location and chronostratigraphic data of the core 69

Rates and mechanics of sea-level rise . . . . 72

GOnclusiOnste: == eee ee ee ae sk Oe Eee G4

ANcknowledgementSs se uses 2 eee 7S

IRGLETEM CESS. ucch tec) aati varie cer ometece el ghee raeskoinisneucye

INTRODUCTION

The coral reefs of Réunion were long neglected by research workers. The first

morphological studies were undertaken in 1970 (Faure & Montaggioni 1970;

Montaggioni 1970). The present study concerns the mechanics of the Holocene

transgression in the Mascarene Archipelago. It is based on eight radiocarbon

dates obtained from core samples from the fringing reef of La Saline.

LOCATION AND CHRONOSTRATIGRAPHIC DATA OF THE CORE

Situated in the western Indian Ocean, at 55.32E, 21.07S, the island of

Réunion consists of two shield volcanoes dating from between the end of the

Pliocene and the present time (McDougall 1971). The mountainous and very

rugged topography of the hinterland (maximum height 3 069 m) is extended

into a very narrow submarine platform (maximum width 7 km). In consequence

the total surface of the coral reefs (12 km?) is negligible compared with that

of the island (2 500 km?). (See Fig. 1.)

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 69-75, 5 figs, 1 table.

ANNALS OF THE SOUTH AFRICAN MUSEUM

SSIGSVOVVVVVVV

SNANAASS

$,SS Loz

SSSAS~~va07.Q

SIN

st sien 4

DAGASCAR

older volc anoes(>2.1 — 0.07 m.y.)

to present)

younger /active volcanoes (0.35m.y.

LOCATION & PHYSIOGRAPHY of REUNION ISLAND

Fig. 1. Location and physiography of Réunion Island.

HOLOCENE SUBMERGENCE ON REUNION ISLAND 71

In 1972 core-drilling was conducted on the fringing reef of La Saline-Trou

d’Eau on the landward side of the reef flat (Acropora pharaonis flat) at a distance

of about 150 m from the seaward side of the reef. The core, which did not

reach the volcanic bedrock, was about 18 m long. The sedimentary material

which was obtained consisted mostly of coral fragments of which more than

80 per cent comprised the remains of branching Acropora forms (Fig. 2).

reef flat sai bottoms

Sores ea mae |

low tide

a sn an se SS eS

= aaa SSS Ss SS SS

SS SS SSS SS

Fig. 2. Schematic cross-section of the fringing reef of La Saline.

The stratigraphic position and radiometric age determinations of the

coral samples are given in Table 1. Dates were obtained by Kigoshi, Gaku-

shuin University, Tokyo.

Table 1. Depth and age of coral samples

Depth in metres of

Field Laboratory sample relative to

Sample Sample present sea-level

No. No. (low tide) Cage BP:

FRE 1 Gak 5067 2 1) 168) 3 110 + 90

FRE 3 Gak 5069 4,2 to 4,3 4950 + 120

FRE 4 Gak 5070 5,3 4 840 + 100

FRE 5 Gak 5071 6,8 5 390 + 120

FRE 6 Gak 5072 9,6 5 860 + 130

FRE 7 Gak 5073 les} Shh) se ills)

FRE 8 Gak 5074 13,8 6 870 + 150

FRE 9 Gak 5075 18,0 7 280 + 120

These radiometric dates make it possible to calculate the average rate of

reef accretion which was about 0,4 cm/year with maximum and minimum

values of 1,02 and 0,11 cm/year respectively. This average value is clearly less

than that suggested by Vaughan (1919) for branching forms with a rapid rate

of growth like the genus Acropora (2,55 cm/year). The average speed of reef

accretion is, however, closely comparable to the 0,4-0,8 cm/year suggested by

Vaughan (1919) and Hoffmeister & Multer (1964) for massive coral forms.

Moreover, it is clear that there has been a general and progressive decrease in

the rate of reef accumulation during the last 7 000 years.

The sum of the foregoing evidence suggests the influence of a major eco-

logical factor such as emergence which would inhibit the growth of the coral;

in this way the upper members of the living coral colonies would be closely

controlled by the level of spring low tides throughout the Holocene marine

transgression. As a result, the position of the corals in the reef structure may

TP ANNALS OF THE SOUTH AFRICAN MUSEUM

be considered as an excellent indicator of the level of spring low tide during

the Holocene transgression (Easton & Olson 1974).

RATES AND MECHANICS OF SEA-LEVEL RISE

Figure 3 shows the curve of median eustatic sea-level rise during the last

7 300 years.

On Réunion Island the sea-level rose at a rate of 0,59 cm/year between

7 300 and 5 600 years B.P. (Fig. 4). This rate seems very high by comparison,

for example, with those obtained for the same time period on New Caledonia

(0,29 cm/year: Baltzer 1970) along the west coast of America (0,33 cm/year:

Redfield 1967) and in the Netherlands (0,32 cm/year: Jelgersma 1966).

years Mee x 1000

— —— higher tide

O— lower tide

depth in metres

@ stratigraphic position of dated samples(see table)

Fig. 3. Holocene submergence of Réunion.

Between 5600 and 3500 years B.P. the rate of sea-level rise declined

markedly to 0,28 cm/year (Fig. 4). Easton & Olson (1974) obtained a comparable

result (0,29 cm/year) on Oahu. On the other hand the world eustatic curve

shows a rate of 0,17 cm/year (Shepard 1963) during the same period. Bloom

(1969) maintains that this rate of 0,17 cm/year is of purely eustatic origin in

Micronesia. If this latter assertion is correct the obviously much greater velocity

of transgression on Réunion must have been due to the addition of isostatic

effects to account for the increase of approximately 0,10 cm/year in the rate

of subsidence of the island. Such a rate of subsidence would, however,

— + = Saw Sel wed.

HOLOCENE SUBMERGENCE ON REUNION ISLAND 73

dated samples (with margin of error

P----4

on the dates of samples)

rates of sea-level rise in cm/year

years B.P. x 1000

Fig. 4. Rates of rise of sea-level on Réunion.

appear too high if one considers those suggested for other subsiding volcanic

areas: Micronesia (0,03 cm/year during the last 10000 years: Bloom 1969),

Mururoa (0,006 to 0,012 cm/year during the last 8 m.y.: Labeyrie et al. 1969)

and atolls in the Pacific Ocean (average rates of submergence 0,002 to 0,004

cm/year: Hess 1965). The rates generally accepted for the degree of subsidence

appear ta be negligible by comparison with average speeds of transgression,

being only about as great as the errors allowed on the calculation of these

velocities of transgression. This being the case, the phenomenon of subsidence

could not account for the large real differences in the rates of transgression on

Réunion.

Furthermore, prior to 5 500 years B.P. sea-level seems to have been 2 to

6 m lower on Réunion than eustatic level elswhere, according to evidence

recently discovered in various parts of the world (Fig. 5). Conversely, the

suggested curve seems to provide evidence that sea-levels were slightly higher on

Réunion than elsewhere from 5 500 to 3 000 years B.P. This lack of correlation

in the height of sea-levels during the period under discussion appears to be

part of a general world-wide trend. Schofield (1967) and Morner (1971) postu-

late respectively that this was due to the effect of astronomic and/or oceano-

graphic factors such as the uneven distribution of ocean waters consequent

upon the local addition of large amounts of glacial melt-water, variation in

the density of ocean water or else fluctuations in the rate of rotation of the

earth. Schofield (1967: 116) wrote that ‘the level in the higher latitudes must

have been initially higher than those in the equatorial regions’.

74 ANNALS OF THE SOUTH AFRICAN MUSEUM

years B.P. x 1008

— ean,

==.

New caledonia _ Baltzer 1970

depth in metres

. Scandinavia , average curve —Morner 1971

Micronesia __ Bloom 1969

eccccccces Florida _ Scholl et al. 1969

Reunion

world average Curve __ Shepard 1963

Hawaii Easton & Olson 1974

Fig. 5. Comparison between sea-level curves of Réunion and other countries.

One might therefore suppose that around 7000 years B.P. there was an

increase in the rate of sea-level rise to compensate for the delay in the beginning

of the first movements of the Holocene transgression in the Mascarenes. After

5 500 years B.P. the rate slowed progressively in sympathy with world-wide

reduction in the velocity of marine transgression (Scholl et al. 1969) and the

partial attainment of eustatic equilibrium in this part of the Indian Ocean.

However, the abatement of this compensatory increase took place gradually

(on Réunion the velocity of transgression remained higher than the average

generally accepted rates) and ceased completely only when the sea reached its

present position which it has never surpassed (Montaggioni 1973).

CONCLUSIONS

1. The eustatic curve obtained indicates that sea-level has risen con-

tinuously, although at a generally decreasing rate, during the last 7 300 years.

This phenomenon seems to have been general throughout the world.

2. In the Mascarenes the Holocene transgression began later than in

other geographic regions. This delay could have been due to oceanographic

and/or astronomic factors.

3. The relatively high rates of increase in sea-level height could have been

due to a process of eustatic compensation which would tend to make good the

differences in sea-level. An isostatic origin (subsidence) for such rates must be

discounted.

4. The sea reached its present level without ever surpassing it.

HOLOCENE SUBMERGENCE ON REUNION ISLAND We

ACKNOWLEDGEMENTS

I wish to thank the Conseil général du Département de la Réunion for its

financial support, thus making possible the project of C™“ datings and core-

drilling. I am grateful to Mrs Margaret Leakey of the South African Museum

for her help in translating this paper.

REFERENCES

BALTZER, F. 1970. Datation absolue de la transgression holocéne sur la c6te ouest de Nouvelle-

Calédonie sur des échantillons de tourbes a Palétuviers. Interprétation néotectonique. —

C. r. Acad. Sc. Paris (D) 271: 2251-2254.

Bioom, A. L. 1969. Holocene submergence in Micronesia as the standard for eustatic sea level

changes. — Quaternaria 12: 145-154.

Easton, W. H. & OLSON, E. C. 1973. Carbon-14 profile of Hanauma reef, Oahu, Hawaii. —

Second Int. Coral Reef Symp., Australia, Abstracts: 91.

EASTON, W. H. & OLSON, E. C. 1974. Late recent rise of sea level in Hawaii.—Jnt. Symp. on

Indo-Pacific Tropical Reef Biol., Guam, Abstracts.

Faure, G. & MONTAGGIONI, L. 1970. Le Récif corallien de St-Pierre de la Réunion (Océan

Indien): géomorphologie et répartition des peuplements.— Rec. Trav. Sta. Mar. Endoume

(h.s. suppl.) 10: 271-284.

Hess, H. H. 1965. Mid-oceanic ridges and tectonics of the sea floor.—Proc. 17th Symp.

Colston Res. Soc., Butterworths edit.: 317-334.

HOFFMEISTER, J. E. & MULTER, H. 1964. Growth rates estimates of a Pleistocene coral reef

of Florida.— Bull. geol. Soc. Am. 75: 353-358.

JELGERSMA, S. 1966. Sea-levels changes during the last 10,000 years. In: Proceedings inter-

national symposium world climate from 8 000 to 0 B.C., 1966: 54-71. London: Royal

Meteorological Society.

LABEYRIE, J., LALOU, C. and DELIBRIAS, G. 1969. Etude des transgressions marines sur I’atoll

de Mururoa par la datation des différents niveaux de corail.— Cah. Pacif. 13: 59-68.

McDouGaLL, I. 1971 The geochronology and evolution of the young volcanic island of

Réunion, Indian Ocean.— Geochim. Cosmochim. Acta 35: 261-288.

MOonrTAGGIONI, L. 1970. Répartition et zonation géomorphologique des structures récifales de

V’Ile de la Réunion (Océan Indien).—C. r. Acad. Sc. Paris (D) 270: 663-665.

MONTAGGIONI, L. 1973. Coral reefs and quaternary shorelines in the Mascarene Archipelago

(Indian Ocean).— Second Int. Coral Reef Symp., Australia, Abstracts: 89-90.

Morner, N. A. 1971. The Holocene eustatic sea level problem.— Geol. Mijnbouw 50: 699-702.

REDFIELD, A. C. 1967. Post-glacial change in sea level in the western North Atlantic Ocean. —

Science 157: 687-692.

SCHOFIELD, J. C. 1967. Post-glacial sea level maxima: a function of salinity?—J. Geosc.,

Osaka City Univ. 10: 115-118.

SCHOLL, D. W., CRAIGHEAD, F. C. & STUIVER, M. 1969. Florida submergence curve revised :

its relation to coastal sedimentation rates.— Science 163: 562-564.

SHEPARD, F. P. 1963. Thirty-five thousand years of sea level. In: CLEMENTS, T. Essays in

marine geology in honor of K. O. Emery. Univ. S. Calif. Press: 1-10.

VAUGHAN, T. W. 1919. Corals and formation of coral reefs. — Ren. Smithson. Instn 17: 189-238.

LATE PLEISTOCENE CHANNELS AND FLANDRIAN SEDIMENTS

BENEATH NATAL ESTUARIES: A SYNTHESIS

ANTONY R. ORME

University of California, Los Angeles, California 90024, U.S.A.

(With 2 figures)

ABSTRACT

Borehole investigations in Natal estuaries reveal late Pleistocene bed-rock channels

buried beneath complex sequences of marine, lagoonal and fluvial sediments deposited during

and since the Flandrian transgression. The dimensions of these channels are conditioned

primarily by the size and gradient of the contributing watersheds and, for the larger rivers,

indicate that sea-level fell to more than —55 m below the present during the Weichselian

Stage. Sandy marine sediments accumulated as wedges in the larger estuaries during the

Flandrian transgression, but are now found only as barrier deposits along the open coast.

Lagoonal sediments accumulated as black to grey organic-rich clays, silts and sands during

and after the transgression, but in recent times have been largely masked by continuing

fluvial deposition. Owing to fluvial scour and fill, which may range over 20 m vertically during

major floods, the radiometric dating of organic debris from these deposits reveals little useful

data concerning the progress of the Flandrian transgression. The Mgeni Estuary, north

Durban, is discussed to illustrate the kinds of evidence upon which this study is based.

CONTENTS

PAGE

Introduction : : : : ah

Buried bed-rock channels : ; 78

Flandrian marine sediments . : 80

Flandrian lagoonal sediments 4 81

The Mgeni Estuary: a case study . 82

Acknowledgements : ; ; 85

References . : 4 p , 85

INTRODUCTION

The following study is based on a detailed analysis of buried bed-rock

channels and sedimentation in about two-thirds of the 60 estuaries and lagoons

that characterize the 570 km of Natal coast between the Mozambique and

Transkei borders (Orme 1974). Sites examined included the Greater St Lucia

Lagoon and its nine principal contributing drainages, the Richards Bay Lagoon

and its two main contributing streams, and 28 independent rivers that discharge

directly into the Indian Ocean without passing through an intermediate lagoonal

filtering system other than their own estuaries. Of these, cross-sections for

20 estuaries were constructed by this investigator from raw borehole data

supplied by various agencies. The remaining estuaries are generally very small

and largely repetitive in detail. From this analysis, a number of observations

may be synthesized and conclusions drawn concerning the late Pleistocene

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 77-85, 2 figs.

78 ANNALS OF THE SOUTH AFRICAN MUSEUM

marine regression and the subsequent Flandrian transgression along the coast

of south-east Africa. Figure 1 shows the location of estuaries and contributory

drainages in Natal and KwaZulu.

BURIED BED-ROCK CHANNELS

The late Pleistocene bed-rock channels described in this paper were

examined as part of a detailed investigation into the nature and magnitude of

Holocene sedimentation in the estuaries along the Natal coast. These bed-rock

channels are important because the ability of modern estuaries and lagoons

to absorb and store sediments is a function of the size of the sedimentary basin

below the level of erosion to be anticipated as a result of flood scour and tidal

effects. The size of the sedimentary basin is in turn related to the amount of

erosion accomplished by Pleistocene rivers attuned to base levels below present

sea-level. In general terms, the larger the watershed and steeper the gradient

of these rivers, the greater the discharge and velocity of the water, and the

greater the erosion and transporting power within their channels. It is thus

to be anticipated that the larger rivers of the present landscape may well be

flowing seaward over wide and deep bed-rock channels cut during later Pleisto-

cene times. Although the pattern of sea-level changes for south-east Africa

during late Pleistocene times has yet to be determined in detail, eolianites

and near-shore deposits at depths of up to —100 m off the Natal coast (Anderson

1906; McCarthy 1967) provide some clues to the magnitude of the drawdown

to which neighbouring rivers sought to adjust. Data revealed by this present

study are consistent with the above assumptions.

The largest river of Natal, the Tugela with a watershed of 28 000 km,

reveals a bed-rock channel at approximately —50 m below present mean sea-

level at a point 6 km above its mouth. The Mfolozi, the second largest river of

Natal with a watershed of 10 700 km?, flows over a bed-rock channel whose

floor was not reached by borings at —40 m below mean sea-level at a distance of

34 river km from the sea. Borehole investigations in and around Richards Bay

indicate a well-defined buried valley for the late Pleistocene Mhlatuze which

descends at least to —55 m below the present surface of the lagoon, and is

clogged with Holocene estuarine and barrier-beach sediments. The Mgeni, the

fourth largest river of Natal with a 4 400 km? watershed, flows seaward over

a bed-rock channel at least —52 m below mean sea-level.

Rivers of intermediate size have bed-rock channels of lesser depth—the

Tongaat at —30 m, the Mhloti at —31 m, the Mtwalume at —27 m, the Mtam-

vuna at least —30 m below mean sea-level, and so on. In some instances, such

as the Mkomazi Estuary, boreholes have revealed only the edge of what must

be a wider and deeper channel. In other instances, such as beneath the Isipingo

Flats, no bed-rock channel has been revealed although coarse sands and gravels

typical of basal channel deposits suggest the existence of a channel at not much

greater depth. Small rivers commonly have small bed-rock channels at com-

NATAL ESTUARIES: LATE PLEISTOCENE CHANNELS AND FLANDRIAN SEDIMENTS 79

MOZAMBIQUE

TRANSVAAL

SIBAY!

ST. LUCIA ESTUARY

& ei Mom BY : ;

ZR ae

S } NSELe M5 SJ CAPE ST. LUCIA

= Ae A ee A 7 = Es

Ss LAG

os ( ye ‘

f ) Lg Ww

( EMPANGENI

CES:

( = ) 7S RICHARDS BAY

(/ / | i

= x { , LALAZI

j as = ( ran 5

/ i

—~ 7 /

eos ie 2) <\C/ pwaGeNDRIFT eee

} pei f WG y Bosc)

( B

\\ MOO! RIVERS (

wes ) Side Ae

LESOTHO J | aes = INDIAN

LA , i { *AAHLALI re) Cc E A N

wo MGENI Z

ue ss DAM 4, TONGAAT

i arr Ws, MAHLOTI

~MHLANGA

}

\% oy

p> CM DURBAN

AAS

Z IsiPINGO

OT!

LITTLE AMANZIMTOTI

~ILLOVO

T AGABABA MS IMBAZ/

SGN “NGANE

\ A MAHLONGWANA

- MM PAMBINYONI

~AZINTO

~IFAFA

(KAT WALUME

UNAMFU

FAZAZAAN

ZUMBE

NJAMBILI

ROY MTENTWENI

PORT SHEPSTONE

HLANGENI

IBILANHLOLO

ITTLE IBILANHLOLO

TRANSKE! MPENJATI

Fig. 1. Location of estuaries and contributory drainages in Natal and KwaZulu.

paratively shallow depth, such as the channel at —12 m beneath the Little

Ibilanhlolo.

From these data it is clear that the depth of the buried channels is to be

positively correlated with the size of the contributing watershed and that,

whereas no definitive conclusions may be reached about the maximum extent of

the late Pleistocene marine regression, sea-level must have fallen to at least

—55 m below the present, presumably during the Weichselian Stage. The sea-

ward gradient of these buried channels is presumed to continue, though with

80 ANNALS OF THE SOUTH AFRICAN MUSEUM

diminishing slope, across the continental shelf toward a former sea-level as

much as —100 m below the present. Although masked by subsequent marine

sedimentation, the evidence for such channels might be revealed by geophysical

investigations.

The precise dimensions of the buried channels vary considerably, but in

many instances a relatively broad bed-rock channel just below mean sea-level,

perhaps the Eemian channel from the Last Interglacial Stage, is incised by a

deeper, narrower trough, thought to be of Weichselian Age. This relationship

is well seen in the Tongaat, Mgeni, Mtwalume, and Mhlangeni Estuaries. The

Mtwalume channel is particularly interesting in that two rock terraces are

preserved above the deepest trough, stranded perhaps as the river shifted against

the south bank during the Weichselian Stage. In general terms, channels exca-

vated in the more resistant sandstones, granites, and shales reinforced with

dolerite have steeper slopes than those found in basalts, unreinforced shales,

limestones, and saprolites derived from all these rocks.

FLANDRIAN MARINE SEDIMENTS

The Flandrian transgression, which commenced some time before 15 000

B.P. and culminated around 5 000 B.P., flooded the bed-rock channels described

above and produced the slack-water estuarine environments in which deposition

of marine, lagoonal, and fluvial sediments subsequently occurred. In general

terms, fluvial deposits characterize the inner portions of these estuaries, giving

way progressively seaward to lagoonal and marine sediments. Such facies

changes are particularly well illustrated in the sedimentary records beneath

the larger estuaries. Such records are rarely simple, however, primarily because

flood scour and fill have frequently disturbed the sequence of deposits. Today,

for example, major summer floods may significantly disturb estuarine sediments

to as much as 10 m below and 10 m above the winter channel bed, and there is

abundant evidence to indicate that similar fluvial regimes existed during the

Flandrian transgression. It is primarily because of this repeated disturbance

of the sedimentary sequence by scour and fill that attempts to define the progress

of the Flandrian transgression by reference to radiometrically dated organic

debris contained within individual strata have proved unsatisfactory.

Marine deposits are found in Natal estuaries in two principal sedimentary

situations: either as materials carried into the estuary by wave and tidal action

during the Flandrian transgression, or as sands washed and blown over the

coastal barriers and spits during and since the close of the transgression.

The first category of deposits is not found in the shallower estuaries that

are subject to total scour at the present time, nor are such deposits found

toward the inner regions of major estuaries where fluvial scour and fill pre-

dominated during the transgression and persist today. Toward the sea in the

larger estuaries and lagoons, however, such deposits may form substantial

wedges in the sequence. In the Mgeni Estuary near its mouth, for example, a

NATAL ESTUARIES: LATE PLEISTOCENE CHANNELS AND FLANDRIAN SEDIMENTS 81

10 m thick wedge of fine sand with abundant shell fragments at a depth of —5

to —15 m suggests a marine incursion toward the close of the Flandrian trans-

gression before the coastal barrier had accumulated (Fig. 2). The light grey,

medium sand with shell debris beneath the Isipingo Flats at —2 to —8 m

suggests the last marine incursion into the estuary prior to the development

of the present swamp facies. The marine materials are essentially sands ranging

from fine to very coarse in texture.

The second category of marine deposits, namely sands forming the coastal

barriers and spits, and washed or carried over into the estuaries by waves and

wind action, are found across the mouths of all estuaries. Sections through

the Little Ibilanhlolo and Mtamvuna Estuaries along the Natal South Coast

reveal the position and character of the modern barrier spits particularly well.

In the former, the sand-spit forms a wedge which thickens to 4 m across the

estuary, lying upon earlier flood deposits. In the latter locality, fluvial deposits

are overlain by up to 13 m of medium to coarse sand that extends as a spit for

over 300 m across the estuary. The Mkomazi and Tongaat Estuaries also

illustrate this characteristic, though the Tongaat sand-spit merges without

marked discontinuity into estuarine muds and sands. The nature and magnitude

of recent barrier growth across Natal estuaries have been examined in detail

elsewhere (Orme 1973, 1974) and will not be discussed further in the present

context.

FLANDRIAN LAGOONAL SEDIMENTS

As Natal’s estuaries began to assume their present shape during the Flan-

drian transgression, lagoonal silts and fine sands accumulated on the margins

of fluvial deposition, while finer silts and clays accumulated in deeper water

and also in backswamps to the side of the main drainage channels. These

deposits included much organic material, derived in part from trees and other

vegetal fragments washed into the estuary or lagoon by inflowing rivers, and

in part from the vegetation growing within the estuary.

In the shallower estuaries, in which earlier Flandrian sediments have been

mostly removed from the bed-rock channel by subsequent flood scour, lagoonal

deposits are found mainly toward the top of the refill sequence and are of

comparatively recent origin. In the deeper estuaries, however, lagoonal deposits

may occur at intervals throughout the sequence beneath the depth of present

flood scour. Such strata represent lagoonal facies that accumulated during the

Flandrian transgression, both during temporary stillstands during the rise of

sea level and as backswamp deposits to the side of the main estuary.

The thick lagoonal sequences found in the Mfolozi and Mgeni Estuaries

are excellent illustrations of this phenomenon. Black to grey organic-rich clays

and silts occur in several lenses beneath the Mfolozi valley and the 10 m thick

swamp facies toward the top of the sequence must represent the last stages in

the infilling of the Greater St. Lucia Lagoon, subsequently overwhelmed by

82 ANNALS OF THE SOUTH AFRICAN MUSEUM

more recent flood deposits. A similar sequence occurs beneath the Mgeni, and

again, a black organic-rich clay merging into a black silt must indicate lagoonal

conditions just before the estuary became clogged with sediments and the river

began to braid with difficulty through its recent fluvial deposits. This lagoonal

deposit lies over a marine sand in which a stream channel was carved and

filled with brown sand. In some of the broader embayments where little recent

fluvial scour has taken place, as at Isipingo, Amanzimtoti, and Msimbazi,

lagoonal muds of recent origin are relatively extensive, ranging from black

clays to fine dirty silts and clays.

Since the close of the Flandrian transgression, most Natal estuaries have

further diminished in size as a result of fluvial deposition. This will pot be

discussed here. Lagoonal sedimentation is now confined to comparatively

small areas while marine deposits are restricted to barriers at the mouths of

the estuaries.

THE MGENI ESTUARY: A CASE STUDY

The Mgeni River drains the fourth largest watershed in Natal (4 400 km?)

and reaches the Indian Ocean through the northern suburbs of Durban. Because

of its watershed’s importance to the water supply of Durban and Pietermaritz-

burg, and because its lower reaches were formerly a major obstacle to road and

rail communications north of Durban, the Mgeni is the most thoroughly studied

of Natal rivers. Between 1958 and 1961, its mean annual flow was gauged at

12,5 m/sec and its suspended sediment load was estimated to be 0,07 x 108

tons annually, or less than 0,7 per cent of the load carried by the Tugela. Its mean

discharge varies seasonally from 18,4 m?/sec during the summer rainy season

to only 6,5 m3/sec during the winter.

The Mgeni rises in open hill country at an altitude of 1 829 m some 32 km

east of the Drakensberg, in an area that experiences more than 1 000 mm of pre-

cipitation annually. Unlike many Natal rivers, the Mgeni’s upper course is

relatively open whereas the last 40 km above the floodplain around Durban

cross deeply dissected terrain in The Valley of A Thousand Hills where deeply

decomposed Archaean granites contribute abundant debris to the river. The

last 16 km toward the ocean are typified by a wide sandy river-bed composed

of abundant granitic debris, with a mean gradient of 2,6 m/km. Because of

the seasonal variability of its flow, the Mgeni may also be temporarily blocked

at the ocean by a sandy barrier, its waters spreading both north and south into

a 3 km-long lagoon parallel with the shore. Historical records from the period

of the early British colony in the mid-nineteenth century suggest that the

Mgeni, on some occasions at least, was diverted by this barrier southward

into Natal Bay (Durban Harbour). The Blue Lagoon, now little more than a

small swamp immediately south of the Mgeni Estuary, is a remnant of this

former diversion. Borehole data are forthcoming from several localities along

the lower reaches of the Mgeni and these will now be discussed in sequence

downstream toward the ocean.

NATAL ESTUARIES: LATE PLEISTOCENE CHANNELS AND FLANDRIAN SEDIMENTS 83

Boreholes for the bridge on the new Durban By-Pass, at the confluence of

the Mgeni and Palmiet Rivers 7 km above the ocean, reveal a bed-rock channel

excavated in Dwyka tillite and Ecca shale to a depth of —25 m below present

sea-level. This channel is filled in succession with up to 2 m of basal gravels,

up to 15 m of sand with some clay and silt lenses, and an uppermost complex of

sands, silts and clays. These deposits are all fluvial in origin and the sequence

displays numerous scour and fill episodes. The present river-bed lies only 1 m

above mean sea-level and below this point occupies a broad alluvial plain

2 km wide.

In the vicinity of the Connaught Bridge, 2,5 km above the ocean, the flood-

plain narrows to 300 m between steep bluffs in a locality that early offered

an important bridging point across the Mgeni. The original Connaught Bridge

was built at the beginning of this century, and the excavations for this bridge

revealed an essentially three-tier sequence of deposits: a basal stratum of

‘white sea sand’, an intermediate stratum of “silty clay’ with thin layers of sand,

and an upper stratum of “coarse sand’. Boreholes for a new Connaught Bridge

obtained in 1958 confirm this general sequence and reveal the basal sand to

a depth of —28 m below mean sea-level without reaching bed-rock. This sand

contains abundant shells and is in places mixed with small amounts of clay,

seemingly confirming a marine incursion into an estuarine environment. The

intermediate deposit is more clayey than silty clay and is up to 13 m thick,

reaching a base of —24 m below mean sea-level. This stratum is a reasonably

typical estuarine deposit. The upper sand is both coarse and rich in gravels,

presumably a typical Mgeni flood deposit filling the contemporary scour trough

of the river which extends down to 11 m below the present river-bed.

Borings for the new Athlone Bridge over the Mgeni approximately 1 000 m

inland from the ocean were conducted in 1964 for the Durban City Engineer,

and these have been interpreted by this investigator (Fig. 2). The sequence of

events revealed is exceedingly complex, reflecting appreciable lateral and

MEDIUM

BROWN SAND

MSL FINE BROWN SAND =< BLACK CLAY Arpepeoe

COARSE GREY iss == MEDIUM SAND COARSE BROWN TO PINK SAND =

: SSE Gee AEST EEE

3s D ass eg oncaras H

a= BLACK SILT = ——— =>

5 —— = Y i

ERS a ACK CLAY = : = / a

Y ) = — ERROR a :

oo eS BON

FINE SAND

WITH SHELL FRAGMENTS ©

“~ ® 6

IUM TO COA

BROWN SAND

eee MEDIUM

Za =

YELLOW TO BROWN —@ = 00,2

CLAY 5 068%

MGENI

ATHLONE BRIDGE, DURBAN

R AND =.9'

¢ RN “S<

Fig. 2. Cross-section through the Mgeni Estuary along the line of the new Athlone Bridge,

north Durban.

84 ANNALS OF THE SOUTH AFRICAN MUSEUM

vertical variations in the sedimentary record. The bed-rock channel is cut

into black Ecca shales which become well fractured and weathered toward their

surface, suggesting that the basal clays found locally across the profile may be

in situ weathering deposits. This bed-rock channel at first slopes gently from

the north bank and then plunges more steeply to at least —5S2 m below mean

sea-level. Though the coarse sand and boulders in the southernmost profile

were not bottomed during boring operations, they are thought to lie near the

floor of this channel. Among the overlying deposits, the dark-coloured, brown

to black sands are interpreted as fluvial deposits accumulating in an organic-

rich estuarine environment, while the yellow to brown clays are interpreted

as lagoonal and backswamp materials. At depths of between —5 and —15 m,

a pronounced fine sand occurs with abundant shell fragments, suggesting a

marine incursion into the estuary towards the close of the Flandrian trans-

gression, at a time when the coastal barrier was not in position. This fine sand

is overlain by lagoonal black clays and silts, and by fluvial brown and pink

sands. The base of the coarse brown to pink sand beneath the more southerly

arm of the Mgeni is thought to represent the lower limit of normal scour at

the present time, though the top of the fine grey sand with shells may have been

scoured out by an extreme flood event related to present base level. Pink sands

are fairly typical of the Mgeni, having been derived from Archaean granites

in The Valley of A Thousand Hills.

Exploratory borings for the Ellis Brown Viaduct across the Mgeni Estuary

200 m inland from the ocean reveal complexity similar to that farther upstream:

an irregular bed-rock channel cut into Ecca shale and Karoo dolerite intrusions;

lenses of marine sand with shell fragments up to 10 m thick, frequent channel

fills of typical Mgeni brown to pink sands and grits, interspersed with clay

and silt lagoonal deposits. As is to be anticipated so close to the ocean, incursions

of marine sand are more frequent and the deposits much thicker than farther

upstream, while fluvial sands are proportionately thinner and finer in texture.

The present surface deposits in the estuary—fluvial sands, swamp clays and

silts, and washover marine sands —provide a useful reference key to the materials

at greater depth.

A sample of wood, Spirostachys africanus, from a depth of —29 m in the

Mgeni channel has yielded a C“ date of 8 420 + 140 years B.P. (Maud 1968).

Assuming that this terrigenous material was then buried by the Flandrian

transgression, it lies somewhat deeper than most estimates of sea-level for the

time. Whereas extreme views place sea-level around 8500 B.P. as high as

—15 mand as low as —35 m, most investigators favour a —20 to —25 m depth.

However, the scour-and-fill character of the Mgeni sequence suggests that the

material may have been emplaced at its comparatively low level during the

alluviation of a flood channel scoured several metres below the sea-level of

the time. Less easily explained is the date of 24 950 + 950 years B.P. obtained

for a peat sample from a black clay stratum in the Harbour Beds at a depth

of —22 m below mean sea-level beneath central Durban. These two examples

NATAL ESTUARIES: LATE PLEISTOCENE CHANNELS AND FLANDRIAN SEDIMENTS 85

illustrate the difficulties in defining the Flandrian transgression in terms of

radiometrically dated debris from estuarine environments.

ACKNOWLEDGEMENTS

The author gratefully acknowledges the assistance of the Natal Roads

Department, the Durban City Engineer’s Department and the South African

Railways and Harbours in providing access to borehole records.

REFERENCES

ANDERSON, W. 1906. On the geology of the Bluff bore, Durban, Natal.— Trans. geol. Soc.

S. Afr. 9: 111-116.

Mavp, R. R. 1968. Quaternary geomorphology and soil formation in coastal Natal.—Zeitsch.

f. Geomorph. (n.f.) 7: 155-199.

McCartnHy, M. J. 1967. Stratigraphical and sedimentological evidence from the Durban

region of major sea-level movements since the late Tertiary.— Trans. geol. Soc. S. Afr.

70: 135-165.

Orme, A. R. 1973. Barrier and lagoon systems along the Zululand coast, South Africa. In:

Coates, D. R. ed. Coastal Geomorphology: 181-217. New York: State University of New

York Press.

Orme, A. R. 1974. Estuarine sedimentation along the Natal Coast, South Africa. Arlington,

Va.: Office of Naval Research. Tech. Rep. 5.

\ =

, t

— ¥

if it B

“

. ¥

' : i

; i

: »

i |

i / |

- }

= eS Se : ; = =

~~ ~ a = - eS et a = od ra

SECTION 2

GEOLOGY B. TERRESTRIAL

CORRELATED ENVIRONMENTAL PARAMETERS AND THEIR

BEARING ON PALAEOENVIRONMENTAL RECONSTRUCTION

IN INCOMPLETE STRATIGRAPHIC SECTIONS

A. O. FULLER

Department of Geology, University of Cape Town

(With 1| figure)

ABSTRACT

In circumstances under which the probability of sedimentation at a particular locus is

correlated with the state of an environmental parameter, systematic loss of information con-

cerning that parameter will occur during subsequent palaeoenvironmental reconstruction

based on the preserved sequence. All other parameters whose states vary sympathetically

will be similarly affected, whether or not they themselves directly influence sedimentation.

Stratigraphic sections with poor time control may not contain clear evidence of hiatus. How-

ever, if two or more environmental parameteres are correlated, but with lag, their imprint

on the stratigraphic record will be marked by discontinuities.

CONTENTS

PAGE

Introduction . : : : 5 : : i : : 87

Fluctuating environmental parameters and preservation potential . 88

Correlated environmental parameters . : : 5 ; : 88

The effect of lag ; ‘ 3 ; : ‘ ; : 3 89

Limitations ‘ : : ; : : ; : : ; 91

INTRODUCTION

Palaeoenvironmental reconstruction, when based on aspects of a sedi-

mentary sequence representing the time interval in question, will be potentially

most successful under circumstances where sedimentation at the study locus

was more or less continuous, that is, where breaks in the record were short

compared to the time intervals over which it is desired that definition of serial

changes is to be achieved. Such circumstances are rare in nature, especially in

non-marine and intermediate environments, and reconstruction of past con-

ditions must often be based on stratigraphic sections which are incomplete.

The time intervals represented by the breaks may not be known, and will not be

if palaeontological evidence is lacking, or absolute ages are absent or too few.

This is particularly true for short breaks, which will not normally be recognized,

especially if evidence of intraformational erosion is lacking.

A lack of preserved sedimentary responses in a section under study may

be due partly to their low preservation potential under a specific depositional

regime, or to circumstances which precluded any form of sedimentation at a

site during a particular time interval.

Proc. sth. Afr. Soc. Quat. Res. 1975

Ann. S. Afr. Mus. 71, 1976: 87-91, 1 fig.

88 ANNALS OF THE SOUTH AFRICAN MUSEUM

The palaeoenvironmentalist dependent on a preserved vertical succession

is obviously familiar with the above factors which may reduce the data-yielding

potential of his section. As stated above, if the gaps are small compared to

the time-scale of the phenomena to be resolved, then no problems exist. But

if they are not, as will often be the case in Quaternary investigations, then

serious limitations are impressed on the study.

FLUCTUATING ENVIRONMENTAL PARAMETERS AND

PRESERVATION POTENTIAL

There is in addition another circumstance, which the author wishes to

explore, which imposes further limitations on reconstruction of the palaeo-

environment where incomplete vertical sections are under study. This relates

to situations in which the probability of a gap in the record occurring is closely

correlated with a particular state of an environmental parameter.

Gaps in a stratigraphic section represent intervals during which a condition

of either non-deposition or erosion existed. Such conditions may be influenced

by a number of environmental parameters operating singly or jointly. These

include, amongst others, temperature, especially as it affects weathering and

erosion, precipitation (amount and distribution of), vegetation, eustacy, and

tectonism (local or regional). Palaeoenvironmental reconstruction is concerned

with these and related elements of the past, and successful identification of

previous conditions will depend firstly on the extent to which their fluctuations

are impressed on the physical and biological record and secondly, and more

obviously, on the degree to which this record is preserved.

Clearly there are circumstances in which the probability of preservation of

evidence of a specific environmental state is very low or zero. Thus temporary

lowering of base level may cause a halt of sedimentation at some locus near

enough to be influenced by the event. Resumption of aggradation would

normally follow a recovery of base level, so that the resulting stratigraphic

section, containing a diastem or paraconformity, would reveal no information

with regard to the nature of the event.

The situation is aggravated if cyclic repetition of the event occurs, since

in this circumstance the systematic loss of information will vitiate any attempt

to interpret serial aspects of environmental parameters.

CORRELATED ENVIRONMENTAL PARAMETERS

An important corollary is that if the states of two or more environmental

parameters are highly correlated in time, then if any one of them strongly

influences the probability of stratigraphic preservation, the others will be

similarly affected, with the result that systematic loss of data concerning all

correlated parameter states will occur. It is necessary therefore, in any particular

study, firstly to consider which of the many environmental parameters subject

CORRELATED ENVIRONMENTAL PARAMETERS 89

to fluctuation is likely to influence the probability of stratigraphic preservation

under the appropriate depositional setting, and secondly the degree of correlation

between such parameters and others which do not directly influence the preser-

vation potential. Thus, following the example cited above, in which a change

of base level led to a loss of stratigraphic record, a number of other environ-

mental parameters such as temperature and precipitation, whose fluctuations

may reasonably be expected to correlate through time with base level, would

similarly be poorly registered in the preserved section.

The limitations that these considerations impose on the interpretation of

ancient environments will apply especially to situations in which only poor

palaeontological control is available. Regional studies, especially with good

stratigraphic control, are more likely to provide continuous records in the

form of composite sections, where gaps in one record may be filled by obser-

vations from another.

In situations where considerable systematic loss of information with regard

to certain parameter states has occurred, due to factors such as those outlined

above, reconstruction of the palaeoenvironment at times specified by radio-

metric, palacomagnetic, or palaeontological criteria are, of course, of

undiminished value, so that should the investigator suspect loss of record

along the lines suggested, emphasis should be given to those datable palaeoen-

vironmental conditions that are impressed on the preserved record. However,

any attempt to define serial phenomena should be undertaken with caution,

especially if one objective is to match parameter fluctuations through time with

those recorded at other sites.

THE EFFECT OF LAG

Figure | has been prepared to illustrate some of the points raised, as well

as to show the result of introducing lag in the correlation of two parameters,

one of which directly influences deposition. It is interesting to note that, where

no lag exists, no obvious discontinuity in the stratigraphic record occurs

(profile II]) whereas hiatuses are very sharply defined in situations involving lag.

It will be noted that profile [V was obtained as follows: an environmental

parameter A (e.g. base level, rainfall) directly influenced sedimentation, such

that certain intervals (t,—ts, t,-t;) were non-depositional. Assuming that fluctua-

tions in parameter A could be perfectly interpreted through a study of physical

or palaeontological elements of the stratigraphic record, curves II and Ill

could be drawn and these curves would also apply to other parameters whose

states fluctuated sympathetically with those of A, without lag. If however the

stratigraphic section contained physical or palaeontological evidence of an

environmental parameter C which fluctuated sympathetically with parameter A

but with lag At, then the profile [TV would result, bearing a clear imprint of

hiatuses which were absent in III. Both profiles III and IV might be obtained

from this same section.

90 ANNALS OF THE SOUTH AFRICAN MUSEUM

™

Figure 1.

| Parameter A

i! + II] Parameter B

IV Parameter C

(For explanation see text)

—

Non-

iti

deposition Deposition

Profile I. A fluctuating environmental parameter A is portrayed through time t,—t,. For

purposes of illustration it is assumed that this parameter directly influences the probability

of stratigraphic preservation, so that a threshold axis has been drawn which separates states

associated with depositional and non-depositional episodes.

Profile II and III. These curves illustrate fluctuations in parameter A corresponding to

depositional episodes in I which may be recorded in the stratigraphic section. Times corre-

sponding to those in I are indicated. The curves can also be taken to represent fluctuations

in any other parameter (B) which is perfectly correlated with A. The two curves are identical

except with regard to the non-depositional interval.

Profile IV. If the states of a parameter C are perfectly correlated with those of A, but are out

of step by lag /A\t, then the profile illustrated by IV results. The state of parameter C at the

times indicated have been read from I.

CORRELATED ENVIRONMENTAL PARAMETERS 91

LIMITATIONS

Discussion has assumed that there was perfect positive correlation between

the fluctuating states of various parameters, and further that the preserved

stratigraphic sections contained elements of one kind or another which could

be uniquely matched with a variety of palaeoenvironmental states. It is clear

that under natural conditions the former assumption is unlikely to be fully

justified, and the present state of the art of palaeoenvironmental reconstruction

makes the latter somewhat unrealistic. Nevertheless the author considers that

in spite of these shortcomings there are broad implications of the arguments

presented above which might be of assistance in interpreting stratigraphic

sequences.

QUATERNARY SEDIMENTATION AND DEVELOPMENT OF THE

LAGOONAL COMPLEX, LAKE ST LUCIA, ZULULAND

David K. HOBDAY

Department of Geology, University of Natal, Pietermaritzburg

(With 10 figures)

ABSTRACT

Lake St Lucia is a saline lagoon on the Zululand coastal plain. It has a surface area of

380 km? and an average depth of less than 1,5 m. A compound barrier and a series of parallel

relict beach-dune ridges record a succession of depositional events related to Pleistocene

sea-level fluctuations. Aeolian sandstones of the penultimate glaciation were planed by the

Eem transgression, during which three distinct beach deposits accumulated a few metres

above the present shoreline. This high stand flooded the back-barrier area and coral reefs

developed in False Bay, the western arm of the existing lagoon system. Sea-level lowering

concomitant with the last glaciation initiated renewed valley incision. It was then that the high

dunes of the modern barrier originated. The smaller river mouths were sealed by littoral

processes during the post-glacial transgression and the lagoonal complex became established.

This has been substantially reduced in depth and extent. Coring reveals a maximum of 33,5 m

of late Pleistocene and Holocene sediment reflecting a change from marine to estuarine and

lagoonal conditions.

CONTENTS

PAGE

Introductioniea, .) hoc. ews Ss ce ee ee we ke 98

TENS Golanyorowinyel WeVERKIe 5 6 5 6 6 6 6 6 «6 o 86

Winitelie mre nee — ey ee ete es ca SOG

OG eS em ey ke Us ue a he eT

LON Gui ae nian 0) Rn SM. << ame arn Bes. 2

Oni er es a A ee Be, ese MOO

NOT a ag oe amet mete rte oh eae ns oe olay ace SLO

Coastaledunesr ee Ge ee OL

Possible relationship of barrier development to

Pleistocene sea-level fluctuations . . . . 102

Fossiliferous limestones of False Bay. . . . . . 103

Sands of the coastal plain Be eee ahah aed Bee OS)

HEDKerSteleUclarCOres;| | tteae oe te Oe , AOF

Discussion andrconclusions) = es ee ee Ld

Acknowledsementstane. ae cyte. Shee oe) Re 2. 2

INGER CNCCS teria ME aie i on ee ee we tL)

INTRODUCTION

Lake St Lucia is a large saline lagoon on the Zululand coastal plain (Fig. 1).

It is separated from the Indian Ocean by a compound barrier, the basal portions

of which consist of calcareous sandstones bearing raised marine-cut terraces.

These are surmounted by semi-consolidated sandstones followed by loose

sands of the high coastal dunes which attain elevations of 180 m.

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 93-113, 10 figs.

94 ANNALS OF THE SOUTH AFRICAN MUSEUM

IN,N7 y

San)

MLN

No N

Cretaceous mantled

ae moe) by Quaternary

ST LUCIA |f°354 Pre-Cretaceous

ESTUARY eI

M folozi @) Core locations

O 5) 10 km

—————— a a

Fig. 1. Generalized distribution of the main geological and geomorphological elements of the

Lake St Lucia area. Core sites and other important localities are indicated.

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 95

The lagoon comprises two north-south trending bodies of water, False

Bay and St Lucia proper, with a connecting strait known as Hell’s Gate. The

total surface area fluctuates between 420 and 225 km? depending upon lake-

level. This is in turn determined by the relative intensity of inflow and evapora-

tion, varying from a position some 1,5 m above mean sea-level following floods

to below sea-level during times of drought. Fresh water is contributed largely

by the Nyalazi, Hluhluwe, Mzinene and Mkuze Rivers, with a combined

annual discharge of 45,2 <x 10° m?. The Narrows, a shallow winding channel,

is the only outlet, extending 21 km to St Lucia Estuary. The occasional negative

gradient accompanying low lake-levels creates reverse flow of sea water up

The Narrows into Lake St Lucia. Prior to its artificial diversion in 1952 the

Mfolozi River, with a discharge greater than the total freshwater inflow into

the Lake, flowed into the Estuary. It now enters the sea 1,5 km south of its

former outlet. Tidal amplitudes of up to 1,6 m in the Estuary are totally absorbed

within The Narrows. Rapid wind induced changes in lake level are common,

however.

False Bay and the western shores of Lake St Lucia are incised into Creta-

ceous strata. The gentle seaward dip exposes progressively younger stages from

the Neocomian of the upstream Mzinene to the easternmost cliff exposures

which probably span the Mesozoic-Cenozoic boundary into the Paleocene.

These rocks are sporadically overlain at an elevation of 10 m by Middle to

Upper Tertiary coquina (‘pecten bed’) and sandstone. Above these are the

unconsolidated red, brown and grey sand deposits which mantle the coastal

plain. The most conspicuous topographic features of the coastal plain are

north-south oriented dune ridges, the larger of which are composed of red

sand. To the west of Lake St Lucia these commonly have a boulder bed at

the base.

In False Bay there are fossiliferous late Pleistocene deposits with a fauna

indicative of unrestricted marine conditions. These must have accumulated

prior to the development of a continuous seaward barrier. At a later stage the

barrier was more substantial and had a profound influence on the nature and

rate of sedimentation.

It is the purpose of this paper to outline the salient sedimentary and

geomorphic characteristics of the barrier and lagoon and to relate these to the

processes involved in their development. It is apparent from the outset that a

dominant factor has been eustatic changes in sea-level. Most of the data were

obtained from the study of surface exposures and shallow auger samples.

Important information concerning the sediments beneath the Lake bed was

provided by seven continuous cores obtained during the latter half of 1973

through the efforts of Mr T. Blok in charge of St Lucia Reclamation. The

cores extend from the sediment-water interface to Cretaceous bedrock at a

maximum depth of over 30 m. Foraminifera, ostracods, diatoms and

a variety of macrofossils were extracted and set aside for specialist

examination.

96 ANNALS OF THE SOUTH AFRICAN MUSEUM

THE COMPOUND BARRIER

Several distinct stratigraphic units are exposed along the seaward side

of the coastal barrier, but the sandstones do not outcrop along the landward

margin where they are mantled by unconsolidated deposits. Certain units are

correlative with outctops in the Durban area described by a number of authors

including Krige (1932), King & Maud (1964), McCarthy (1967) and Maud

(1968).

UNIT |

The oldest deposits of the barrier consist of fine to medium-grained sand-

stone which is best exposed at First Rocks and Bats Cave 2 km to the north of

Mission Rocks (Fig. 1). This is probably equivalent to the ‘Bluff Beds’ of

Krige (1932) or the ‘first aeolianite’ of Maud (1968). It extends from an unknown

depth below sea-level to a maximum observed elevation of 8 m. The higher

parts have been subjected to considerable karst weathering.

Large-scale cross-bedding is a characteristic feature (Fig. 2). Dips range

from 8 to 36 degrees and are variable in azimuth, although a strong westerly

component is evident in most outcrops. Foreset laminae commonly display

evidence of minor slumping prior to lithification. In thin section the rock is

seen to consist predominantly of well-sorted sub-angular to rounded quartz

grains with diagenetically altered shell fragments and a sparite cement.

Fig. 2. Aeolian sandstones of unit 1 displaying large-scale cross-bedding at First Rocks.

eS ee ee

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 97

These attributes together support the interpretation by Belderson (1961)

and McCarthy (1967) of an aeolian origin for the equivalent formation in Natal.

Sand was presumably piled up as coastal dunes during a stage of eustatically

lowered sea-level. Ample sand would have been available from a belt of emergent

continental shelf several km wide. Present-day prevailing wind directions are

from the north-east and south-west, blowing roughly parallel to the shoreline.

During glaciations however, temperature contrasts between land and sea would

have been increased, thus leading to a higher frequency of strong onshore winds.

This probably accounts for the high proportion of landward inclined foresets

in unit |.

At certain localities, such as Bats Cave, these aeolian sandstones are

truncated by a flat, horizontal erosion surface at between 4 and 5 m above

sea-level. Elsewhere the surface is less regular, ranging from a height of 8 m

to below present sea-level.

UNIT 2

The depressions in this undulating surface eroded into unit 1 are overlain

between present spring low tide level and an elevation of up to 5 m by coarser-

grained, conglomeratic sandstone. These rocks, here referred to as unit 2, are

subdivisible into two distinct facies on the basis of differences in sedimentary

structures, and to a lesser extent, in lithology.

The first type consists of low-angle, eastward-inclined planar foresets of

medium to coarse-grained sandstone with small discoidal and blade-shaped

pebbles composed largely of lydianite. Well-defined parallel lamination dips

seaward at between 4 and 10 degrees (Fig. 3). Microscopic examination reveals

that these sandstones are less perfectly sorted than the aeolian deposits. They

contain in addition to quartz up to 10 per cent feldspar and a variety of mollusc

and echinoderm fragments, foraminifera and algae. These deposits strongly

resemble the upper foreshore of modern beaches. The preponderance of flattened

pebbles of small size (less than 3 cm maximum diameter) is suggestive of the

size-shape sorting processes that operate in the swash zone, and the eastward

dipping stratification records the successive profiles of a prograding beach.

Local discontinuities mark changes in the beach profile.

The second facies occurs at a lower level (generally below 2 m a.s.].) and

is gradational into the upper foreshore beds. It consists of trough cross-bedded

sandstone, very similar in lithology to the beach deposits, but differing in the

shape of the pebbles. Both flat and equant shapes are represented, and many are

almost perfect spheres. Furthermore, the maximum size is slightly larger, some

pebbles attaining diameters of 15 cm. Whereas discoidal and blade-shaped

pebbles are selectively cast up on a beach, shape sorting results in a larger

proportion of equant pebbles being concentrated below low tide level (Bluck

1967). Thus it appears that this trough cross-bedded sandstone facies origi-

nated in a nearshore shallow marine environment.

Trough axis azimuths (Fig. 4) reveal that the currents which deposited

98 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 3. Upper foreshore deposits of unit 2. The upper surface is planed and overlain by a

younger conglomerate and the coastal dunes to the right. Mission Rocks.

these sands flowed in directions which were either sub-parallel or perpendicular

to the existing shoreline. Since the late Pleistocene shoreline trend was

undoubtedly very similar, these currents were of a semi-permanent longshore

variety. Modern-day patterns along certain sections of the Natal and Zululand

coasts are very similar. Both northward and southward flowing currents are

observed at different times of the year. The offshore directed cross-beds are

characteristic components of the nearshore inner rough facies immediately

seaward of the swash zone (Clifton et al. 1971).

The upper foreshore facies of unit 2 extend to a maximum observed

elevation of between 3,5 m and 5 m where they are bevelled by the same seaward-

dipping surface that locally truncates the older aeolian deposits of unit 1. Thus

the high water level corresponding to the deposition of these beach sands must

have been a little more than 5 m above present sea-level and is possibly approxi-

mated by a poorly preserved 8 m surface at Bats Cave. Deposition of unit 2

must therefore have coincided with an inter-glacial higher sea-level which

succeeded the low stand during which the aeolian sandstones (unit 1)

accumulated.

During the ensuing marine retreat, probably related to an interstadial of

short duration, these shoreline deposits were lithified, and together with the

older sandstones, subjected to karst weathering.

<< lO ee

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 99

Fig. 4. Distribution of cross-bed azimuths in the lower nearshore

facies of unit 2 (34 readings)

UNIT 3

Renewed advance of the sea above its present level planed both aeolian

and beach deposits along a seaward-sloping surface at present preserved between

5,3 and 3,4 m. This wave planed surface is mantled by a coarse-grained con-

glomerate (Fig. 5) which fills in karst depressions and potholes produced by

marine abrasion. The thickness of this conglomerate, designated unit 3, is

therefore highly variable, ranging from a few cm to 2 m. The inclusions consist

of locally eroded clasts of calcareous sandstone together with a variety of

extrabasinal pebbles. Of the latter quartzite pebbles are most numerous followed

by lydianite and gneiss. Some of the inclusions are of boulder dimensions, the

largest encountered measuring 65 cm in diameter. The mean diameter is approxi-

mately four times greater than that of the pebbles in unit 2. A discoidal shape

predominates. Large numbers of oysters and other abraded mollusc fragments

are present locally. Some oysters are preserved in living position in potholes.

Unit 3 is ascribed to transgressive shoreline processes which charac-

teristically deposit coarse shingle. The fact that the coarsest material is preserved

in the potholes suggests that these pebbles and boulders were responsible for

their abrasion. The sides of some potholes are distinctly undercut. Vigorous

100 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 5. Elongate depressions and potholes in truncated surface of unit 2 overlain by oyster-

bearing conglomerates. South of Mission Rocks.

surf action broke off blocks of sandstone which were later incorporated in the

pebbly matrix.

The superabundance of pebbles derived from the Precambrian hinterland

suggests that, in addition to the Mfolozi, other large rivers with outlets further

north may have been providing this material. There is, however, a northward

decrease in the maximum dimension of these inclusions suggesting that the

competence of the Mfolozi considerably exceeded other rivers in the area.

Larger pebbles are encountered to the north of the Mfolozi than to the south,

indicating that beach drift patterns were probably the same as they are today.

Maximum high water mark corresponding to this episode was a little

more than 5 m above present sea-level, but could not be determined precisely.

It was nevertheless distinctly lower than the previous transgression which had

been accompanied by the deposition of unit 2.

UNIT 4

An aeolian sandstone very similar to unit 1 in terms of its composition,

texture and sedimentary structures overlies a 4,4 m pebble veneered (unit 3),

eroded surface at Bats Cave. The foresets are inclined westward at angles of

35 degrees. This sandstone was apparently not deposited to any great thickness,

extending to an elevation of only 10 m at this locality. Nevertheless, the karst

and decalcified upper portions suggest that some of the overlying sands were

derived by the leaching of this unit.

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 101

The limited distribution and thickness of this deposit suggests accumula-

tion during a brief marine regression. A narrow belt of exposed sand would

thus have been subjected to deflation, the sand accumulating against the older

deposits of the barrier.

UNIT 5

Evidence for at least one further rise in sea-level above the present shore-

line comes from a second generation of potholes incised through unit 3 and

commonly extending into unit 2 (Fig. 6). These potholes differ from the older

variety associated with unit 3 in the preservation of outflow channels produced

by seaward runoff. Oysters occur most abundantly in these potholes within

1,5 m of modern spring high water mark. Many adhere to the sides or to one

another in living position and there are few reworked fragments. Potholes

related to this episode at elevations of up to 4,5 m contain reworked oysters

and discoidal pebbles.

Fig. 6. The three raised beach deposit of the barrier. Unit 5 beneath the scale occupies a

cavity in unit 3 which overlies the eroded surface of unit 2. Mission Rocks.

COASTAL DUNES

Practically continuous high coastal dunes extend from Mtunzini north-

ward into Mozambique. They consist of light coloured, almost pure quartz

sand with local concentrations of heavy minerals such as ilemnite, rutile,

zircon and magnetite. Grain size analysis reveals a normal distribution with a

median diameter in the medium sand range.

102 ANNALS OF THE SOUTH AFRICAN MUSEUM

In places the coastal dunes overlap landward on to an older ridge of red

sand. Dune peaks attain an elevation of 188 m to the south of St Lucia Estuary.

Between the Estuary and First Rocks to the north they are low and insignificant.

Beyond First Rocks two lines of high dunes are separated by a narrow

depression.

Deep augering revealed an increasing percentage of heavy minerals, but

the absence of a core of older consolidated deposits. The sands have been

weakly calcified in places. There is no evidence of the lignite-bearing Port

Durnford Formation except to the south of St Lucia Estuary. Excavation of

the seaward dune flanks reveals multidirectional cross-bedding. Occasional

cross-bedding on a large scale, in sets 10 m thick, is reflected in places by

patterns in the vegetation growing on the flanks (A. J. Tankard, pers. comm.).

This, together with the topography of the dune crests, suggests that winds from

the north-east and south-west have been equally effective.

An origin related to a lowered sea-level is indicated by the present narrow

beach, and the dominance of destructive processes which are effectively reducing

the average elevation of the dunes. North-easterly winds in particular have

produced parabolic blowouts which cause local destruction of vegetation and

thus accelerate deflation. Certain dune crests are actually growing in elevation

(C. J. Ward, pers. comm.) but this appears to be exceptional.

As with older aeolian deposits sand would have been carried landward

across the exposed sandflat. The pre-existing ridge of older cemented sands

coincident with the modern shoreline would have localized accumulation as

an almost continuous dune belt, broken only by river outlets. Stabilization by

vegetation would have occurred during the subsequent climatic amelioration.

Further limited accretion of wind-blown sand along the seaward margins has

been accompanied by local erosion of the dune base elsewhere depending

upon differences in shoreline trend and offshore topography.

POSSIBLE RELATIONSHIP OF BARRIER DEVELOPMENT TO PLEISTOCENE SEA-LEVEL

FLUCTUATIONS

Changes in the relative level of land and sea, such as are evident in the

Lake St Lucia barrier, are attributable to two possible basic mechanisms,

tectonic or eustatic. Monoclinal warping (King & King 1959), although demon-

strable on a large scale involving seaward tilting of land surfaces, does not

account for the wide extent of low terraces at comparable elevations along

much of the east coast of southern Africa (Krige 1927; Davies 1970, 1971).

A eustatic origin appears to be the more likely.

There are marked differences of opinion as to world-wide sea-level heights

during early and middle Pleistocene times, and also with regard to the validity

of Holocene levels higher than the present (Curray 1969). Nevertheless, there

appears to be general agreement that during the late Pleistocene, between

approximately 75 000 and 130000 years ago, there were times when sea-level

stood several metres higher than at present. It is also accepted by a number of

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 103

investigators that this Eemian (Sangamon) high was not a single event, but

consisted of three peaks separated by slightly lower levels (Fairbridge 1960;

Mesolella et al. 1969; Emiliani 1970, among others).

It is possible that the stratigraphic units 2, 3 and 5 of the St Lucia barrier

correspond with these three Eemian peaks. Unit 1, which is widely developed

along the Mozambique, Natal and probably much of the south-eastern Cape

coast, is therefore interpreted as originating during the preceding Saalian

glaciation.

FOSSILIFEROUS LIMESTONES OF FALSE BAY

Highly fossiliferous limestones containing a variety of corals and molluscs,

together with local concentrations of brachiopods and cirripeds, are exposed

in very small outcrops at Lister’s Point and Picnic Point on the western shores

of False Bay (Fig. 1). These rocks occur intermittently along the shallowly sub-

merged shelf bordering these outcrops, and are present in at least two places

beneath the surface of False Bay near to the opening into Hell’s Gate. An

additional small pinnacle protrudes above the mud floor at the south-eastern

end of Hell’s Gate, and at low lake-level is within a few centimetres of the

surface.

The faunal characteristics of the limestones are similar at all localities,

but there are differences in the relative abundance of various coral types. At

Lister’s Point several large colonial corals are preserved in growth position

together with scattered solitary forms (Fig. 7), all of which await specialist

identification. Casts of branching Acroporidae are the common form in the

upper part of the Picnic Point exposure. Large bivalves such as Hyotissa hyotis,

with a maximum observed length of 23 cm have been obtained along with

large specimens of the gastropod Jonna. Other commonly occurring molluscs

include Circe scripta, Decatopecten sp., Polinices sebae, Rapana rapiformis,

and Ostrea. The barnacle Balanus is abundant, along with a small brachiopod

of the family Kraussinidae which resembles Megerlina (T. Mason, pers. comm.).

Also present are sponge and echinoderm fragments and possible crab remains.

Trace fossils are mainly large (1-4 cm diameter) vertical tubes with horizontal

offshoots and probably include both Ophiomorpha and Thalassinoides.

Petrographic examination reveals that the bulk of the allochems are

biogenic, consisting of mollusc fragments, foraminifera and red algae, with a

variable proportion of lithoclasts, mainly quartz. According to Folk’s (1959)

classification this rock is a biosparite, a well-sorted, moderately high energy

deposit, with a sparry calcite cement.

During times of very low lake-level the lower parts of the formation are

seen to contain concentrations of spherical to discoidal pebbles of rhyolite,

basalt, jasper, chert and lydianite along with many oysters, mainly broken.

This merges upward into fossiliferous biosparite with a great variety of fossils,

followed by a coral rich zone with a matrix of broken shells. There are con-

104 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 7. Coral at Lister’s Point on the western shore of False Bay.

spicuous differences between one outcrop and another in the proportion of

articulated to disarticulated bivalves and also in the degree of alignment of

the fossils. South of Lister’s Point up to 50 per cent of the bivalves are articu-

lated and there is a tendency toward upward convexity of individual valves.

No preferred disposition is apparent in the predominantly disarticulated valves

at Picnic Point.

Along the shores of False Bay the limestones overlie an eastward dipping

surface eroded into Cretaceous siltstones. The basal contact is undulatory in a

direction parallel to the lake shore. In places there is evidence of intense boring

of the Cretaceous surface by marine organisms prior to deposition of the pebble

and shell debris. The upper surface extends to a maximum observed elevation

of 3,4 m a.s.l. Probing beneath the submerged shelf suggests that the base

extends to a maximum of | m below sea-level.

These deposits have been radiometrically dated as older than 50 000 years

(Orme 1973). This, together with their elevation and geomorphic situation,

suggests deposition during the Eemian high stand concomitant with plantation

of the barrier terraces. The lithology and faunal content point to an unrestricted,

high energy shallow marine situation in close proximity to the shore. Sea-level

was at least 3 m above the present, and the water temperature possibly a few

degrees warmer. It appears that False Bay was at this time a marine embayment

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 105

with Hell’s Gate opening directly to the sea. This would have been contem-

poraneous with one of the stages of bevelling of the barrier sandstones to

seaward. These at this stage would have resembled some of the shallowly

submerged ‘reefs’ off the present-day Natal coast, e.g. Aliwal Shoal and Glenton

Reef. Living corals were restricted to areas of free circulation in the vicinity

of Hell’s Gate, and have not been encountered as fossils in those areas of False

Bay which would have been very sheltered or periodically diluted by fresh

water inflow.

The upper surface of the limestone slopes gently eastward (Fig. 8) and

was possibly wave-planed. This was very likely a consequence of the Eem sea-

level oscillations postulated above. On the other hand the surface may have

been eroded during a Holocene transgression above the present sea-level.

This is believed to be unlikely, however, in view of the apparent antiquity of

some of the sands which overlap the landward margins of the limestone surface.

The approximate position of the Hluhluwe River outlet at this stage was

probably in the vicinity of a patch of fluviatile or estuarine conglomerate some

2 m a.s.l. which is situated 2 km south-west of the present mouth.

Fig. 8. Fossiliferous limestone at Lister’s Point displaying eastward inclined surface.

SANDS OF THE COASTAL PLAIN

Several distinct varieties of unconsolidated sand mantle the coastal plain.

Prominent north-south trending dune ridges consist largely of homogeneous

red sand with pebbles and boulders towards the base. Six dune cordons,

106 ANNALS OF THE SOUTH AFRICAN MUSEUM

decreasing in age from west to east, have been mapped by Davies (pers. comm.).

A thick basal boulder bed occurs in the lowermost parts of the western

cordons, e.g. at Bushlands Halt inland of the present study area. This is replaced

eastward by scattered pebble bands at all elevations. A kilometre inland from

Fanies Island Rest Camp pebbly coarse-grained sand attains a maximum

thickness of 8 m. Heavy mineral laminae, containing up to 60 per cent zircon,

rutile, magnetite, garnet and ilmenite, occur interlayered with quartz sand,

which itself contains prominent textural banding. Pebbly coarse-grained beds

1-15 cm thick alternate with finer material in approximately the same pro-

portion. These layers commonly display opposed (herringbone) cross-

lamination. The pebbles are discoidal and of physically resistant lithologies,

mainly vein quartz and quartzite. Further south near Charters Creek and in

the Makakatana plantation area are similar deposits displaying low angle

cross-bedding of variable azimuth.

These deposits are interpreted as of estuarine or shore zone origin, with

indications of tidal reversal in flow. The presence of numerous westward

inclined cross-beds precludes a normal fluviatile origin. The moderately mature

composition and texture and the predominantly discoidal pebble shape suggest

energetic processes with shape-sorting such as occurs on beaches or estuarine

shoals.

The homogeneous red sand, which in most dune ridges overlies the boulder

and pebble bearing deposits, has a median diameter in the fine-grained sand

range and displays little regional variation in texture. Subangular to rounded

quartz grains make up between 70 and 95 per cent. The heavy mineral content

decreases eastward from a maximum of 8 per cent near Bushlands. In most

samples clays constitute between 6 and 18 per cent, occasionally more. These

appear to have originated by the kaolonization of feldspar, which is almost

entirely absent. The red sand has a maximum thickness of 50 m immediately

to the west of False Bay.

The innermost dune ridge, which passes through Hluhluwe village, is

probably related to a very high early Pleistocene sea-level. The dune bases

become progressively lower eastward concomitant with an increase in topo-

graphic relief and dune continuity. This is suggestive of decreasing antiquity,

a conclusion which is supported by east-west differences in soil profiles. Indu-

rated ferruginous zones, a product of advanced podsolisation, are well developed

in the west. Although present 4 km inland from St Lucia Estuary this soil

horizon has not been recognized seaward of Lake St Lucia.

The red sands of the dune ridges are attributed to aeolian processes which

accompanied marine regressions. Together with the coarser estuarine or beach

deposits at the base each probably constitutes a transgressive-regressive couplet,

but the details are masked by advanced diagenesis, which has destroyed most

primary structures, plus the obscuring effects of younger sands and dense

vegetation.

Very fine-grained light coloured sand is intermittently exposed in a dune

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 107

ridge adjacent to the western shores of Lake St Lucia and The Narrows. It rests

on the Cretaceous at an elevation of 7-12 m and varies between 3 and 9 m in

thickness. It is abruptly overlain by pebbly coarse-grained sands. Over 90 per

cent of the sand is quartz with about 8 per cent calcite. It is well sorted, and

unlike other sands in the area, displays marked positive skewness. This, in

combination with other textural and compositional attributes, is suggestive

of deposition by wind. An absence of internal structures and the weathering

characteristics both resemble loess (Hobday 1965).

The coastal plain is veneered with coversands which are generally between

0,5 and 3 m thick. Some have undoubtedly been derived by leaching of red

sands, as evidenced by white sand tongues which extend down into the red

sand. There is nevertheless textural evidence of local redistribution, probably

during Holocene times. Ventifacts have been recovered from this deposit at

the top of a quarry to the west of Fanies Island Rest Camp.

Near Charters Creek a discontinuous thin layer of reworked hardpan is

encountered between the red sand and overlying light grey sand. This gravel

layer confirms that the upper sands are not simply the leached counterpart

of the red sands but are substantially younger. This is further supported by

marked textural contrasts. The median grain size of the coversands coincides

closely with the homogeneous red sands, but some samples are slightly better

sorted (Hobday 1965).

Low dunes and hummocky topography are present over much of the

eastern half of the coastal plain. These are quite distinct from the larger red

sand ridges from which they have been derived by local deflation. A north-

south alignment is once more evident but is less consistent. Augering reveals

that these dunes consist entirely of Holocene coversands.

WAKE SI eUCIA CORES

A coring programme conducted by the Provincial Roads Department at

the request of Mr T. Blok, engineer in charge of St Lucia Reclamation, estab-

lished that sediment thicknesses are considerably greater than had previously

been suspected (Kriel 1965). Seven continuous cores, extending from the

sediment-water interface to Cretaceous bedrock, were obtained from a floating

rig (Fig. 9). Their locations are indicated in Figure 1. The depth to bedrock and

the salient lithological and faunal characteristics are depicted in Figure 10.

This diagram illustrates the variability in thickness of the Quaternary fill,

ranging from a minimum of 7,8 m to a maximum of 33,5 m. Figure 10 also

demonstrates a lack of stratigraphic similarity between the various cores apart

from a general tendency for the coarser sediment grades to occur in the lower

parts. Carbonaceous mud and silt, such as dominate much of the present bed

of the lake, constitute the bulk of the upper part of the cores.

The foraminiferal content of core samples taken at 50-100 cm intervals

has been analysed by Phleger (in press). A number of carbonaceous samples

108 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 9. Floating rig utilized to obtain deep sediment cores in operation in False Bay.

are being processed by J. Vogel of the Council for Scientific and Industrial

Research for radiocarbon age determination. The preliminary findings, together

with those reported by Kriel (1965), indicate a Holocene age apart from the

deeper parts of some cores which extend into the late Pleistocene.

Core 1 from northern False Bay extended to a depth of 33,5 m. Below a

depth of 19 m silt and shelly sand predominate and contain a variety of molluscs,

some of which are of marine affinity (R. N. Kilburn, pers. comm.). This accords

with the findings of Phleger (in press) who detected foraminiferal evidence of

marine conditions, with unrestricted access to oceanic water at depths of

21,5 m and at the bottom of the core. Sedimentological evidence for high

energy processes comes from 50 cm of reworked Cretaceous at the base, but

this is overlain by deposits of sand and mud suggesting rapid reduction in

energy as a result of deepening and the sheltering effect of a barrier.

Bioturbated mud containing broken, thin-shelled molluscs makes up all

but the basal 3 m of Core 2. Phleger (in press) reports the presence of gypsum

crystals at depths of between 18,5 and 25,5 m. This indicates that even during

the early Holocene there were periods of intense desiccation initiated by effec-

tive lagoonal confinement and lack of appreciable inflow of fresh water. Phleger

notes foraminiferal indications of marine conditions just above transgressive

white sands at the base of the core, and at depths of 4,5-5 m.

109

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA

[eCUNPJOIOVU Pd1OJUNODUD ATUOWUUOS JO UOIINGLISIP PoZI[V1oUuSs PUL OZIS UILIS JUVUILIOP SUICOIPUI SOIOO OYLT JO UONe}UNSOIdSI SNPWIOYIS “OT

BUlT[oL

SNIICSSPN

Ud[OS

eliejOoulnsues

XBUOPOIIIIH

BPUOIR ||

*e1oUuds

BUOORIA|

vided

snueleg

BI11ISO

BIUISOG

BIue|o-4

SNIIBSSEN

BSUIT

snueleg

sAWe]y)

BlosIewNn a

“SI

O€

110 ANNALS OF THE SOUTH AFRICAN MUSEUM

Core 3 struck Cretaceous bed-rock at a depth of 13,7 m. Between 10 m

and the base are poorly sorted sand and oxidized silt containing calcareous and

ferruginous concretions. These are interpreted as fluviatile deposits. They are

overlain by 2,5 m of sand containing broken shells, apparently of estuarine

origin. Above these are typical lagoonal muds with abundant Eumarscia

pauperscula giving way upward to Solen corneus. Scattered sand-size quartz

grains with a red oxide coating were probably blown in from the Pleistocene

red sand cordons.

The basal sediments of Core 4, situated in the north-east, consist of 20 cm

of aeolian sand. This is followed by 1,5 m of very shelly sand with shallow

water estuarine or lagoonal macrofauna such as Solen and Balanus. Phleger Gn

press) finds the upper parts of the core to be characterized by high production

of foraminifera, indicating a plentiful nutrient supply by tidal incursions or

river inflow.

Core 5 to the south of Bird Island indicates a thin basal layer of aeolian

sand. This is followed by 40 cm of relatively poorly sorted sand containing

articulated bivalves suggestive of a sheltered environment. The remainder of

the core consists of rapid alternations of silt and mud, a textural contrast which

characterizes the floor of the Lake in this area at the present day. The included

fauna are essentially estuarine or lagoonal forms.

Core 6 in the southern portion of the Lake struck bed-rock at a compara-

tively shallow depth of 7,8 m. White sand at the base is overlain by 50 cm of

high oxidized medium-grained red sand identical to the wind-blown red dune

deposits of the modern coastal plain. A metre of estuarine mud which overlies

the red sand is followed by fine-grained brown sand with Solen and mud clasts.

The latter are indicative of periods of subaerial exposure followed by inundation

such as occur in a tidal mudflat, or in a non-tidal situation such as Lake St Lucia

today where the level is subject to fluctuation. The uppermost deposits consist

of 4,5 m of black mud.

The Narrows (Core 7) display a sequence which is conspicuously different

from any other. Whereas shells are generally absent finely disseminated vegetal

detritus is common throughout. Phleger (in press) records evidence in the

presence of Trochammina sp. of a marine marsh environment above a depth of

15 m. Mud at a depth of 15 m coarsens upward through silt with ferruginous

concretions into silty sand, a typical lacustrine or ‘bay fill’ pattern produced

by lake shore encroachment over lake floor muds. Below a level of 15 m Phleger

finds a complete absence of foraminifera, which is suggestive of a freshwater

marsh deposit.

A second core was only partially retrieved from a position very near to

Core 7 but slightly further seaward. The succession was generally similar apart

from a well-defined layer of beach sand at a depth 7,7-12 m. Its position within

a marine marsh setting is suggestive of washover processes whereby beach-

barrier sand is transported’ landward over back-barrier salt marsh during

storms.

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA let

DISCUSSION AND CONCLUSIONS

Despite controversy over the degree and timing of Quaternary sea-level

changes it is generally accepted that early Pleistocene levels were considerably

higher than at present. Because glacio-eustatic oscillations were superimposed

on this non-glacial regression, retreat of the sea from the Zululand coastal

plain was undoubtedly erratic and accompanied by several reversals. North-

south aligned dune ridges, whose positions are related to former shorelines, are

conspicuously younger eastward. Many of these ridges comprise a basal trans-

gressive beach deposit overlain by aeolian sands. Unfortunately the details

of these early to middle Pleistocene events have for the most part been masked

by reworking, vegetation and advanced diagenesis.

The drainage pattern was profoundly influenced by these dune ridges.

The Mzinene River was deflected first northward and then southward to unite

with the Nyalazi and Hluhluwe Rivers, which had been confined behind the

same ridge. River valleys became deeply entrenched during stages of sea-level

lowering. Northern False Bay and Hell’s Gate were both incised to depths

greater than 30 m. Shorelines corresponding to these low stands were situated

near the outer edge of the continental shelf. The combined discharge through

Hell’s Gate probably joined with the Mkuze River to extend across the emergent

shelf and erode the deep canyon which is present off Leven Point. Hopefully

further coring in the northern lake will document this hypothetical confluence.

Unlike most barrier-lagoon systems which were initiated in response to a

deceleration in the rate of sea-level rise during the period approximately 7 000

to 5000 years ago (Curray 1969; Phleger 1969), the St Lucia barrier had an

earlier origin. A spine of aeolian sandstones, thought to be related to the pen-

ultimate or Saalian glaciation, was truncated during the Eem transgression.

There were probably three distinct peaks of sea-level advance during which

beach and nearshore sand and gravel accumulated. The two intervening stadials

are marked by a karst erosion surface and a younger aeolian sandstone

respectively.

Planation and inundation of the ancestral barrier sandstones during the

sea-level maxima would have permitted free marine circulation in the area

now occupied by Lake St Lucia. Hell’s Gate opened directly to the sea, with

small islands protruding above the barrier, which otherwise was shallowly

submergent and breached by a number of channels. The western shores of

Lake St Lucia were sea-cliffs at this time, thus possibly accounting for their

present linearity. Unimpeded oceanic incursion in those areas of False Bay

adjacent to Hell’s Gate permitted the growth of coral reefs. A variety of molluscs

flourished and accumulated in situ as oyster beds and as bioclastic deposits.

Other commonly occurring organisms were brachiopods, barnacles, sponges

and echinoderms.

During the last glacial retreat the bulk of the earlier valley-fill deposits

were eroded. Onshore winds crossing the emergent shelf blew sand against the

sandstone barrier core. A prominent submarine ridge near the outer margin

LD ANNALS OF THE SOUTH AFRICAN MUSEUM

of the continental shelf may be a relict barrier marking the maximum extent

of the Weichsel regression. Lower ridges to landward possibly represent beach

ridges developed and subsequently flooded during the Flandrian transgression.

Sedimentological and faunal evidence from the Lake St Lucia cores indi-

cates that lagoonal confinement was established at an early stage during the

post-glacial transgression. Nevertheless, marine conditions were periodically

reintroduced by the breaching of outlets. Progressive reduction of river gradients

resulted in the blocking of river mouths in the north, presumably by littoral

drift followed by dune formation. Only the Mfolozi River was sufficiently

large to maintain an exit.

The combined waters of the northern rivers became impounded behind

the barrier thus formed. As the lagoon level rose the margins overflowed the

adjacent depressions between the coastal barrier and older ridges inland. This

ancestral Lake St Lucia system occupied an area of 1 165 km? (Orme 1973),

covering the existing Mkuze swamps. Between St Lucia and the drowned

Mfolozi valley to the south was a relatively shallow spillway where the waters

were deflected over older alluvium to the only outlet at the mouth of the Mfolozi.

Subsequent changes in the configuration of the Lake St Lucia system

were accomplished by delta infilling, segmentation, swamp encroachment and

local shoreline erosion (Hobday 1965; Orme 1973). Pronounced accumulation

of mud in the immediate vicinity of rivers entering the Lake is probably due to

flocculation of clays in saline water. As noted by Phleger (1969) a marked

reduction in salinity during flood stages allows clay to become more widely

distributed over a lagoon. At the present day the entire floor of the Lake is

above effective wave base. As a consequence wind-induced waves and associ-

ated currents are the most effective sediment transport mechanisms.

ACKNOWLEDGEMENTS

This study began in 1964 as part of an M.Sc. thesis supervised by Professor

L. C. King. Further investigations during the past three years have been sup-

ported by the Natal Parks, Game and Fish Preservation Board. I am grateful

to Mr R. S. Crass, Principal Scientific Officer, and other officers of the Parks

Board including Gordon Forrest, Frik Joubert, W. Jacobs and A. Mitchell

for their help. Messrs T. Blok and Yeld of St Lucia Reclamation are thanked

for their assistance and practical advice. I. van Heerden helped with the field-

work. Certain aspects of the study, including most of the sediment analyses,

were financed by the Council for Scientific and Industrial Research. Mr R. N.

Kilburn kindly identified many of the molluscs.

REFERENCES

BELDERSON, R. H. 1961. The size distribution characteristics of the Recent shallow marine

sediments off Durban, South Africa. Unpublished M.Sc. thesis, University of Natal.

Biuck, B. J. 1967. Sedimentation of gravel beaches: examples from south Wales.—J. sedim.

Petrol. 37: 128-156.

SEDIMENTATION AND DEVELOPMENT OF LAKE ST LUCIA 113

CurrTon, H. E., HUNTER, R. E. & PHILLips, R. L. 1971. Depositional structures and processes

in the non-barred high-energy nearshore.—J. sedim. Petrol. 41: 651-670.

CurrAy, J. R. 1969. History of continental shelves. In: STANLEY, D. J. The new concepts of

continental margin sedimentation. Washington D.C.: Am. Geol. Inst. JC61—JC618.

Davies, O. 1970. Pleistocene beaches of Natal.— Ann. Natal Mus. 20: 403-442.

Davies, O. 1971. Pleistocene shorelines in the southern and south-eastern Cape Province

(Part 1).—Ann. Natal Mus, 21: 183-223.

EMILIANI, C. 1970. Pleistocene paleotemperatures.— Science 168: 822-825.

FAIRBRIDGE, R. W. 1960. The changing level of the sea.— Scient. Am. 202: 7079.

Fok, R. L. 1959. Practical petrographic classification of limestones.— Bull. Am. Ass. Petrol.

Geol. 43: 1-38.

Hospay, D. K. 1965. The geomorphology of the Lake St Lucia area. Unpublished M.Sc.

thesis, University of Natal.

Kine, L. C. & Kina, L. A. 1959. A reappraisal of the Natal Monocline.—S. Afr. geog. J. 41:

15-30.

Kina, L. C. & MaAup, R. R. 1964. The geology of Durban and environs.— Bull. geol. Surv.

S. Afr. 42.

KRIEL, J. P. 1965. Report on the hydrology of the St Lucia Lake system: Appendix No. 5.

In: Report of the Commission of Inquiry into the alleged threat to animal and plant life

in St Lucia Lake: 228-303. Pretoria.

Krice, A. V. 1927. An examination of the Tertiary and Quaternary changes of sea-level in

South Africa with special stress on evidence in favour of Recent world-wide sinking of

ocean level.— Ann. Univ. Stellenbosch. 5: 1-18.

KriGE, L. J. 1932. The geology of Durban.— Trans. geol. Soc. S. Afr. 35: 37-67.

McCarthy, M. J. 1967. Stratigraphical and sedimentological evidence from the Durban

region of major sea-level movements since the late Tertiary.— Trans. geol. Soc. S. Afr. 70:

135-165.

Mavp, R. R. 1968. Quaternary geomorphology and soil formation in coastal Natal.— Z. Geo-

morph. 7: 155-199.

MESOLELLA, K. J., SEALY, H. A. & MATTHEWS, R. K. 1970. Facies geometries within Pleisto-

cene reefs of Barbados, West Indies.— Bull. Am. Ass. Pet. Geol. 54: 1899-1917.

Orme, A. R. 1973. Barrier and lagoon systems along the Zululand coast, South Africa. In:

Coates, D. R. Coastal Geomorphology: 181-217. New York: State University of New

York Press.

PHLEGER, F. B. 1969. Some general features of coastal lagoons. Laguanas Costeras, Un

Simposio, UNAM-UNESCO, Mexico: 5-26.

PHLEGER, F. B. (In press). Holocene ecology of St Lucia Lagoon, Zululand, based on foramini-

fera. In: Report of ihe St Lucia Scientific Advisory Council Meeting, Charters Creek, 1976.

Pietermaritzburg: Natal Parks, Game and Fish Preservation Board.

AEOLIANITE: AUSTRALIAN AND SOUTH AFRICAN

DEPOSITS COMPARED

MARGARET E. MARKER

University of the Witwatersrand

(With 5 figures)

ABSTRACT

A comparison of Australian and South African Pleistocene consolidated beach and dune

deposits is presented. Their distribution, largely south of 30°S on former marine platforms,

seawards of deeply weathered erosion bevels and extending below present sea-level, is discussed.

Their spatial relationships both with underlying bedrock and with other similar deposits are

considered. It is shown that whereas on a coast of submergence aeolianite deposits of different

ages will be superimposed, in areas of emergence the strandlines will be spatially separated.

In the Lower Southeast of South Australia where this is the case, the ridge complexes have

been shown to conceal breaks of slope in the underlying limestone bedrock. The latter are

attributed to transgressive marine erosion. It is postulated that similar breaks of slope will be

found to underlie South African aeolianite deposits. Furthermore it is postulated that erosion

of the underlying limestone provided a source of calcium carbonate for subsequent lithification

of the backshore beach and dune ridges.

CONTENTS

PAGE

Introduction . : ; Z 2 115

Aeolianite distribution Z 3 i 116

Spatial relationships ; , : 120

Conclusions . < , 2 : 122

Acknowledgement . : , : 123

References E ‘ : : : 123

INTRODUCTION

Consolidated calcarenite dune or beach material, generally termed

aeolianite* in South Africa, is a widespread Cainozoic legacy along subtropical

coasts (Fairbridge 1968). As relict strandline material it offers considerable

potential for the elucidation of Cainozoic coastal fluctuations. In South Africa,

however, specific studies of these deposits have, until recently, been neglected. By

contrast, in Australia aeolianite deposits have attracted considerable attention

as a potential basis for Quaternary chronologies (Bird 1960; Gill 1967; Hossfeld

1950; Sprigg 1952; et al.). Nevertheless interpretation of the results is fraught

with difficulty.

Through a comparison of the better documented Australian occurrences

with their South African counterparts, this paper focuses attention on aspects

of aeolianite distribution and spatial relationships.

*Although cemented calcareous beach sandstones, properly termed calcarenites, are domi-

nantly of aeolian origin, the presence of cross-bedded sands incorporating shell bands suggests

that the term aeolianite is not strictly correct. For the purpose of this paper the term will,

however, be retained.

IBIS

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann S. Afr. Mus. 71, 1976: 115-124, 5 figs.

116 ANNALS OF THE SOUTH AFRICAN MUSEUM

Aeolianite consists of consolidated beach and dune sands distinct from

currently forming beach rock. Most aeolianite deposits are of considerable

thickness, exhibit pronounced cross-bedding and are Pleistocence to Holocene

in age. Being largely backshore deposits, they mark former strandlines. The

beach and dune sands have been consolidated to varying degrees by calcium

carbonate. Lithification therefore implies a considerable amount of calcium

carbonate grains or comminuted shells in the original sands, a long period of

leaching to cement the core material with concomitant formation of overlying

dominantly quartzose sands. The current absence of the latter in association

with the aeolianite further implies that the aeolianite deposits must now be

considerably diminished in volume. Where several phases of aeolianite deposi-

tion have occurred, it is likely that the leached overlying sands have been

incorporated into the later deposits. The degree of lithification depends partly

on age but also apparently on exposure. Thus resistant aeolianite frequently

outcrops along eroding cliffs and on the present beach, lithification being asso-

ciated with spray wetting.

AEOLIANITE DISTRIBUTION

Australian aeolianite deposits are widely distributed along the south-

eastern coasts from Westernport Bay, on the Bass Strait islands and northern

Tasmania to Eyre Peninsula. The red silicious dunes of the east coast are,

however, very different and they contain no aeolianite core (Bird, E. C. F. &

Thom, B. pers. comm.). Aeolianite deposits are also found along the western

coast between Cape Leewin and North West Cape where they rest on the

Coastal Limestones, in origin also aeolianite (Fig. 1).

Aeolianite is also widespread along the South African coastline between

Saldanha Bay and East London. More recently the red sands of the coastal

fringe between the Transkei and Zululand have been recognized as similar in

origin but now deeply weathered (McCarthy 1967; Maud 1968). A distinction

- YEAST

= LONDON

ee PORT

SS —— ELIZABETH

~ 700 0 200 400 600km

w = Deep Duncrust and weathering

= Tertiary Limestone es Baee tn sen ® (aki ates (6 ons ees

Bee \ i

“7, Reolianite x !

——— 200m tsobath Ny)

Fig. 1. The distribution of Australia and South African aeolianite deposits in relation to

Tertiary limestones and deep weathering profiles.

AEOLIANITE: AUSTRALIAN AND SOUTH AFRICAN DEPOSITS COMPARED 117

must be drawn between the Berea Red Sands, a weathering product of the

aeolianite Pliocene Bluff Beds, and the high red dunes which have a core of

aeolianite material (Maud 1968). Although at least three distinct phases of

aeolianite formation have been reported from widely separated localities

(Martin 1962; Maud 1968; McCarthy 1967), detailed analyses of specific

localities have been largely neglected.

Comparison of the distribution of aeolianite deposits in Australia and

South Africa highlights certain similarities (Fig. 1). In both Australia and South

Africa, aeolianite deposits are most common south of latitude 30°S. The north-

ward extension towards the Tropic of Capricorn of the South African red

weathered aeolianite dunes has no counterpart in Australia although recently

weathered aeolianite has been reported from a rainshadow coast near Towns-

ville (Hopley 1971). The absence of relict aeolianite in the tropics generally

may therefore be a function of leaching efficiency rather than of original deposi-

tion patterns.

In both countries aeolianite deposits rest on marine benches and extend

below sea-level. The present coastline generally truncates the alignment of the

aeolianite ridges, indicating their emplacement under conditions of low sea-

level. Such truncation is clearly seen between George and Knysna (Fig. 2),

at Kenton-on-Sea (Fig. 3) and in the high dunes of Zululand. The Aliwal

Shoal, and probably the Protea Bank also, are seaward extensions of the Bluff

Beds which have been proved to a depth of —100 m on The Bluff itself

(McCarthy 1967). The Wilderness aeolianite ridges have been mapped to a

CLV |r >

Tuan

Le IE

pee

qwtere ey CRonde\ swartvie,\ }

: 2 ylei Sy

—— SUBMARINE RIDGES

qenecee FORMER CLIFF LINE

FBZ cALcARENITE RIDGES MIDNA OC IL IN

22] COASTAL PLATFORM

Fig. 2. Aeolianite ridges in the George-—Knysna embayment (after Tyson 1971).

118 ANNALS OF THE SOUTH AFRICAN MUSEUM

GRAHAMSTOWN Ge ee Silcrete Profile

= Tertiary Limestone

-:-. Aeolianite

~”

BATHURST = :

<a i)

Ty |

A G Mug

vm"

? PORT ALFRED

Fig. 3. Aeolianite deposits of Eastern Cape Province.

depth of —55 m and the reefs of Plettenberg Bay are also aeolianite in origin.

Such discordance of Pleistocene and modern shorelines is equally true of

Australia. In Western Australia Rottnest Island off Perth is formed of aeolianite.

Soil profiles developed on aeolianite and buried by later aeolianite deposition

occur below mean sea-level at Sorrento in Victoria and the base of the youngest

ridge in the Lower Southeast of South Australia lies at —11 m on the seaward

side (Fig. 4).

A second feature is the close juxtaposition of aeolianite deposits and mid-

Tertiary erosion bevels carrying relict deeply weathered soils (Fig. 1). The

emergent beach ridge sequence of the Lower Southeast of South Australia

commences immediately shorewards of the laterite capped Dundas Tablelands

(Fig. 4). The West Australian limestones abut on to the laterite-capped Darling

Ranges peneplain. In South Africa the George-Knysna ridges lie seawards of

the laterite-capped Coastal Platform (Fig. 2) and the Kenton-on-Sea aeolianites

have the Grahamstown silcrete profiles inland of them (Fig. 3).

AEOLIANITE: AUSTRALIAN AND SOUTH AFRICAN DEPOSITS COMPARED

wot

2. %.e?

ee wue

ere °

ed ® ®

eee @ e

5 OOO DO

1A EAST NARACOORTE

1B WEST NARACOORTE

2 HARPERS

3 STEWART-CAVE

4 WOOLUMBOOL

oO PEACOCK

6 BAKERS

7 ARDUNE

8 EAST AVENUE

9 WEST AVENUE-CAVETON

10 REEDY CREEK-BURLEIGH

11 DAIRY-KONGARONG-MACDONNELL

12 WOAKWINE

13 ROBE

119

Ss Recent sediments Including blown sands d

VV,

Volcanic ash and blown sand SEI

Volcanic rocks (Pleistocene) LI™~

Calcified aeoliantte beach ridges (Pleistocene)

Pa Oligo-Miocene Limestone

|@s| Granitic Basement complex

Fig. 4. Aeolianite beach ridges in the Lower Southeast of South Australia.

120 ANNALS OF THE SOUTH AFRICAN MUSEUM

In every case aeolianite covers marine benches immediately below the

level of the weathered Tertiary bevel. Frequently, it rests on marine limestones,

deposited contemporaneously seawards of this planated land surface. In the

eastern Cape these marine deposits are believed to overlie three marine-cut

benches (Dingle 1971; Ruddock 1968; Siesser 1972), and are typical regressive

marine deposition facies. However, the degree of diagenesis and in particular

of recrystallization indicates that three transgressions and regressions were

involved and that reworking of the marine limestones and superimposed

aeolianite has destroyed much of the earlier material (Siesser 1972). The Coastal

Limestones of Western Australia have also been so modified that little trace of

cross-bedding remains and the recrystallized limestons are capable of hosting

extensive cave systems. This alteration is attributed to marine transgression

which has bevelled and eroded the deposit. In many respects the much modified

West Australian Coastal Limestones bear a close resemblance to the Cainozoic

Limestones of South Africa. Recently the term Coastal Limestone has been

proposed to include the Dorcasia Limestones of Saldanha Bay, the Bredasdorp

Beds and the Alexandria Formation (Siesser 1972). All these limestones consist

of marine beds overlain by redistributed and altered aeolianite. In places

younger, recognizable, cross-bedded, calcareous sandstone aeolianite rests on

top of the altered Cape Coastal Limestones, just as it does in Western Australia.

The close association of cross-bedded aeolianite and underlying limestone,

whether of aeolian or marine provenance, suggests that their juxtaposition

is not fortuitous. Destruction of the limestone by bevelling during a transgression

would provide a source of calcareous sand to add to beach quartzose material

and thus enable lithification of the aeolianite during the subsequent regression.

In south-eastern South Australia the bedrock is Miocene bryozoal limestone and

even the present beach sands are rich in calcium carbonate. However, in many

other aeolianite localities present-day beach sands are so low in calcium car-

bonate that lithification is hard to envisage. It is suggested that in such localities

the Tertiary limestones have now been totally eroded and that the aeolianite

is a relict deposit determined by the former presence of calcareous bedrock.

The patchy distribution of Tertiary limestones in South Africa lends credence

to this suggestion.

SPATIAL RELATIONSHIPS

Elucidation of the relationships of different aeolianite deposits to each

other and to the underlying bedrock is dependent on precise and detailed

mapping. The tendency for aeolianite occurrences to be superimposed, juxta-

posed and elided and their association with recent blown sands of the present

coastline have hampered the collection of comprehensive data.

Spatially distinct aeolianite ridges representing former strandlines have

been preserved only in areas of Pleistocene emergence. The Lower Southeast

of South Australia provides one of the best sequences, but the series associated

AEOLIANITE: AUSTRALIAN AND SOUTH AFRICAN DEPOSITS COMPARED i

with the George-Knysna embayment is similar. Both areas have attracted

considerable attention (Sprigg 1952; Hossfeld 1950; Martin 1962; Tyson

1971).

Where subsidence was current, aeolianite deposits are superimposed with

intercalated soil profiles. Such sequences are well seen at Sorrento, Victoria

(Bird 1972) and along the coast north of Perth, Western Australia. The Zulu-

land red dunes with an aeolianite core are superimposed on Berea Red Sands,

but the supposed superposition at The Bluff, Durban, has been shown to be a

raised beach deposit (McCarthy 1967). Subsiding coasts have caused sub-

mergence and partial destruction of aeolianite surface morphology and its

burial under more recent deposits. It seems probable that this has been the case

in the Eastern Cape.

The ubiquity of these inter-ridge recent deposits conceals in most cases

the contact of aeolianite and bedrock, so that the ridges have been presumed

to rest on gently shelving marine benches (Davies 1971). Recent studies have,

however, indicated that this may not be true. In the Durban area aeolianite

deposits have been shown to abut on three distinct breaks of slope (Maud 1968).

Analysis of the surface topography of the Miocene Limestone, a karst host

rock, in the Lower Southeast of South Australia using a close borehole data

grid has produced information of relevance in this respect (Marker 1975).

There aeolianite beach ridges abut on and cover cliffs in the underlying bed-

rock (Fig. 5). Each distinct aeolianite ridge series therefore records a marine

transgression when the step was cut and a regression when the sands were

blown from the exposed shore platform against the cliff. The borehole plot

suggests that certain strandlines overlie pronounced cliffs and these can be

interpreted as representing major periods of stillstands. The East Naracoorte

beach abuts against the early Pleistocene or late Pliocene Kanawinka cliff.

The Mingbool, Compton, Caveton, Allendale and Robe coastlines, the latter

partly buried by more recent sands at C. Northumberland, are also dominant.

In some cases raised beach pebble deposits occur at the same altitudes to confirm

the borehole evidence. The Lower Southeast of South Australia has a sequence

of strandlines extending in time from at least early Pleistocene to + 8 000

years B.P. at Robe. The central Caveton complex has yielded C™ dates exceeding

35 000 years B.P. (Blackburn 1966) but on geomorphological grounds is con-

sidered more likely to date from about 135 000 years B.P. (Chappell, J., pers.

comm.).

In the George-Knysna embayment a series of four strandlines have been

identified (Martin 1962). However, it appears likely that at least twelve such

strandlines exist. The earliest aeolianite material has recently been related to

the 100 m bevel, the second to the 60 m bevel, the third to the 18 m bevel and

the fourth to the 8 m bevel (Butzer & Helgren 1972). However, these workers

did not recognize the greater number of ridge complexes. It appears, therefore,

well established that aeolianite ridges are associated with marine cliffing. The

implication is that whenever contacts between aeolianite and bedrock are

DD ANNALS OF THE SOUTH AFRICAN MUSEUM

REGION 6 REGION 5 REGION3 NO KARST REGION 2 REGION |

ML_Burr

frowned :v\ doline-uvala Leake: done fine

SW drowne: ahd oline—u Leake swamp doline—uvela NE

Tantanoolay v ~ v! 1

cave

deep uvalas ridge with terra rossa swamp. dolines ACS Ese

ag poe eb aE. oe “ E SE pi00

Coonawarra

REGION 6 REGION 5 REGION 4 REGION 7 NO KARST N

BEACH DEPOSITS ingboo!

LOXTON SANDS Bureigh “7p” Yipee

Ss lan yes y) Ts ce eeemmanannnas (2

BRIDGEWATER FM. (beach ridges}

EES | MALANGANEE FM

Vv] WOLCAWIC ASH & LAVA

BRIDGEWATER FM. [clay & sift)

EE: |] reworkeo GAMBIER LIMESTONE

| =Se5

Sx]

GAMBIER LIMESTONE

EOCENE & OLDER ROCKS

[E===] Plezomeraic sunrAace

Fig. 5. Sections across the Lower Southeast of South Australia to show the association of

aeolianite beach ridges with bedrock, breaks of slope (compiled from borehole data).

recorded it is essential to establish whether the site lies landwards or seawards

of the deposit.

On the Eastern Cape coast between Kwaaihoek and Kleinemonde aeolianite

outcrops in cliffs, on the beach and extends seawards in a north-easterly direction

as reefs (Fig. 3). The base of the deposit is only rarely exposed. In the Bushman

and Kariega estuaries aeolianite rests on and incorporates beach cobbles at

—1 to —2 m below mean sea-level. Since this aeolianite extends seawards, this

altitude can be taken to be the landward altitude of the marine bench that has

been incised. At Port Alfred another aeolianite rests on Cape System bedrock

at an altitude of approximately 4 m (Mountain 1962). Inland from the coastal

aeolianite at Kenton-on-Sea a marine cliff is cut into the Coastal Limestone

with a summit bevel at 60 m, which is overlain by another, presumably older

aeolianite. Further aeolianite deposits mapped with the Coastal Limestones —

overlie other breaks of slope inland. However, even if the coastal outcrops

alone are considered, it is apparent that remnants of a series of ridges exist

(Fig. 3).

CONCLUSIONS

Comparison of Australian and South African deposits has emphasized

certain salient facts. Aeolianite deposits record former, probably Pleistocene,

AEOLIANITE: AUSTRALIAN AND SOUTH AFRICAN DEPOSITS COMPARED 123

strandline positions. The deposits appear to be located over marine-cut breaks

of slope. They are regression deposits, subsequent to the transgressions that

eroded underlying calcareous bedrock, thus providing both calcium carbonate

in sufficient quantity for lithification and cliff loci for sand deposition. The

marine bedrock seems to have been laid down contemporaneously with deep

weathering of the mid-Tertiary planation surfaces. The degree of diagenesis

and alteration of early aeolianite deposits appears to be the consequence of

marine submergence and planation. Such altered deposits retain nothing of

their original surface morphology and little of their structural morphology,

therefore their value for elucidation of a Quaternary chronology is minimized.

The degree of lithification of later aeolianite deposits is a function of age and/or

exposure to sea-spray wetting. Although hard aeolianite occurs in shore loca-

tions it may be contemporaneous with less consolidated inland deposits. In

general the older the aeolianite, the greater the degree of lithification. Aeolianite

deposits have a neglected potential for the elucidation of Cainozoic coastal

chronologies.

ACKNOWLEDGEMENT

Much of the research embodied in this paper was funded by travel grants

from the South African Council for Scientific and Industrial Research.

REFERENCES

BirpD, E. C. F. 1960. Formation of sand beach ridges.— Aust. J. Sci. 22: 349-50.

Birpb, E. C. F. 1972.—Ancient soils of Diamond Bay, Victoria.— Vic. Nat. 89: 349-53.

BLACKBURN, G. 1966. Radiocarbon dates relating to soil development, coastline changes and

volcanic ash deposition in southeast, South Australia. — Aust. J. Sci. 29: 50-2.

BUTZER, K. W. & HELGREN, D. M. 1972. Late Cenozoic evolution of the Cape coast between

Knysna and Cape St. Francis.— Quat. Res. 2: 143-169.

Davies, O. 1971. Pleistocene shorelines in southern Africa. — Quaternaria 15: 317-23.

DINGLE, R. V. 1971. Tertiary sedimentary history of the continental shelf off southern Cape

Province.—Trans. geol. Soc. S. Afr. 74: 173-86.

FAIRBRIDGE, R. W., ed. 1968. Encyclopedia of geomorphology. New York: Rheinhold Book Co.

GILL, E. D. 1967. Evolution of the Warrnambool—Port Fairy coast and the Tower Hill erup-

tions. In: Landform Studies. JENNINGS, J. N. & MassutTt, J. A. eds: 341-364. Cambridge:

University Press.

Hop ey, D. 1971. The origin and significance of north Queensland island spits. —Z. Geomorph.

15: 371-89.

HossFELp, P. S. 1950. Late Cainozoic history of the Southeast of South Australia.—Trans.

R. Soc. S. Aust. 73: 232-79.

Marker, M. E. 1975. The Lower Southeast of South Australia: A karst province. Occ. Pap.

Dept. geogr. envir. Stud. Witwatersrand Univ.

MartTIn, A. R. 1962. Evidence relating to the Quaternary history of the Wilderness Lakes. —

Trans. geol. Soc. S. Afr. 65: 19-39.

MAup, R. R. 1968. Quaternary geomorphology and soil formation in coastal Natal.—Z.

Geomorph. Suppl. 7: 155-99.

McCarthy, M. J. 1967. Stratigraphical and sedimentological evidence from the Durban area

of major sea level changes since the Late Tertiary.— Trans. geol. Soc. S. Afr. 52: 135-65.

Mountain, E. D. 1962. The geology of the country round Port Alfred, C.P.— Explanation

of sheets 3326D & 3327C Geol. Surv. S. Afr.

124 ANNALS OF THE SOUTH AFRICAN MUSEUM

Ruppock, A. 1968. Cainozoic sea levels and diastrophism in a region bordering Algoa Bay.

—Trans. geol. Soc. S. Afr. 71: 209-33.

SIESSER, W. G. 1972. Petrology of the Cainozoic Coastal Limestones of Cape Province. S.A.

—Trans. geol. Soc. S. Afr. 75: 178-85.

SPRIGG, R. C. 1952. Geology of southeastern province of South Australia with special reference

to Quaternary coastline migrations and modern beach development.— Bull. geol. Surv.

S. Aust. 29.

Tyson, P. D., ed. 1971. Outeniqualand: the George-Knysna area.—S. Afr. Landscape Series

2. 8S. Afr. Geogr. Soc.

LATE QUATERNARY COLLUVIAL DEPOSITS (PROGRESS REPORT)

P. VERHOEF

Department of Geography, University of South Africa, Pretoria

(With 3 figures and 2 tables)

ABSTRACT

At the northern entrance of the Daspoort tunnel in Pretoria a massive gravel deposit is

exposed at the junction of the pediment and hillslope. On the upper section of the pediment

the gravel is discordantly overlain by red sand beds, and the top of the gravel layer

bears a ferricrete horizon. Lower down on the pediment, excavations reveal colluvial

sand beds, representing two major stages of deposition. The lower sand beds, bearing

iron concretions, apparently constitute the downslope facies of the gravel layer. It is proposed

that the gravel deposit and its extension is a sheet-flow deposit dating from the last Pleistocene

pluvial period, and the upper, red sand a lower energy sheet-flow deposit dating from an

Holocene pluvial phase of lesser intensity.

CONTENTS

PAGE

Graveldeposita 5. ee Sa ene

IRedisand beds: 2- 4 2) 3: = =] «129

The Sandfontein sand deposits . . . 129

DISCUSSION Pee ete ee ee = 130

Conclusionsae 7 ts ee ee ee dt

IRCEKeNnCeSie te ee eae, ee ted an LSD

GRAVEL DEPOSIT

At the northern entrance of the Daspoort tunnel in Pretoria a massive, red

(SYR4/6) colluvial gravel deposit, 6-7 m thick, is revealed at the foot of the

Daspoort hillslope (1330 m; 25.44S). The deposit is composed of an assemblage

of sub-angular quartzite and some vein-quartz fragments set in a matrix of

finer gravel and fine-earth. The rock fragments and matrix are heavily iron-

stained and iron-impregnated. The gravel deposit obliterated the original

sharp knickpoint between pediment and hillslope and the present junction

is a concave break of slope from 20° on the hillside to 3° on the upper pediment.

Upwards the deposit pinches out against the hillside and downslope it also

diminishes in thickness. The geological relationships are shown in Figure 1.

The dip of the strata is about 28° and the deposit rests discordantly upon

partly decomposed shale.

Coarse layering can be detected and five sub-phases of deposition may be

distinguished, as borne out by analyses of samples. (See Fig. 2.)

All quantitative determinations quoted in this paper have been performed

by J. H. Reynhardt (1975).

Layer A: Thickness 1 m; about one-third by volume composed of rock

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 125-132, 3 figs, 2 tables.

125

ANNALS OF THE SOUTH AFRICAN MUSEUM

126

"BIIOJOIg “JUOWOIRID “osuvsI 1oodseq sy} Jo jusUIpaed pue sdojs UJOY}IOU oY) SSOIDe UOTIOES OIVeUIAYDS “[ “314

UOI}EWIO4 BIOOdSeG :3}3!1Z7}1ENH Opel

UOI}JeWA0Y Hiaqsaijebew :ajeus SEL

DULS 2

SUO!I}91DUOD UOAI 40 DJaADIAIaY +44

JPARID oo,

€29USI0}SI3}/q) S9l2ey pues + J@ABID -WNIAN|/OD 7

({29uUa00/0H) pues pay -wnianiyog 7

LATE QUATERNARY COLLUVIAL DEPOSITS 127)

1 . 4 eS * Bx

an ote BAS os Sac

=* So . es —

are “S

Fig. 2. Colluvial gravel deposit at the northern entrance of the

Daspoort tunnel, Pretoria.

(quartzite) fragments, average long-axis 10 cm; loamy sand matrix; gradual

transition to B.

Layer B: Thickness 2,5 m; about one-half by volume composed of rock

fragments, average long-axis 12 cm; sandy loam matrix; clear transition

to. C

Layers C and D: Thickness 2,5 m; about three-quarters by volume com-

posed of rock fragments, average long-axis 15-16 cm; sandy clay loam matrix;

quartzite fragments of up to 30 cm are quite frequent in these layers; abrupt

boundary with E.

Layer E: Thickness 0,5 m; distinctive well-rounded, heavily patinated

(some black) pebbles and cobbles, gravel size smaller than 10 cm.

Thus there is an increase in the coarseness of the gravel from top to bottom,

layer E being the exception, and also an increase in gravel content.

The roundness of the gravel was determined according to the method

described by Sames (1966: 127). Looking at the outline of the pebble the convex-

shaped parts are measured and the total convex length expressed as a per-

centage of the whole circumference. According to the wear six roundness

grades are distinguished by Sames; angular 0-10% rho; subangular: 15-25%

rho; subrounded: 30-40% rho; rounded: 45-60% rho; well-rounded: 65-80%

rho; extremely well-rounded: 85-100°% rho. For layer A the average roundness

is 33% rho and Standard Deviation 20% rho; for layer B the respective figures

anes0 /Arho and 177, rho. tor layer ©, 26.7, tho and 112,74 rho; for layer,

128 ANNALS OF THE SOUTH AFRICAN MUSEUM

20% rho and 8 % rho; for layer E, 56% rho and 33% rho. The gravel in layer E

is significantly more rounded than the gravel from all other layers.

Mechanical analysis of the matrix material (< 2 mm) from the various

layers shows a significant increase in clay content from top to bottom and a

corresponding decrease in the sand fractions (see Table 1). This is probably due

to a greater measure of weathered shale admixture during the earlier phases of

deposition, the area of shale exposure having diminished as the deposit accumu-

lated, or to better sorting as the slope on the pediment increased as a result of

the progressive accumulation of sediment at the junction of the hillslope and

pediment.

TABLE 1

Mechanical analysis of matrix material (< 2 mm) of the gravel deposit

Layer Sand (%) Silt (%) Clay (%)

A 86 4 10

B qT 5 20

C 70 6 24

D 48 12 40

39 14 47

Heavy residue analysis shows a significant difference between layers D and

E (see Table 2), the latter showing a greater degree of weathering.

To differentiate further between the various layers in the gravel deposit

a cluster analysis was performed with the aid of a computer programme (see

Davis 1973), using sixteen variables based on particle size distribution, gravel

content, morphometric characteristics, and heavy mineral composition. The

most significant boundary was found to be that between layers D and E (differ-

ence of 19 units), while the smallest difference was that between layers A and B

(6 units). The difference between layers B and C was 9 units.

TABLE 2

Heavy residue analysis of gravel deposit

Layer Zircon/ Tourmaline Zircon|(Epidote + Pyroxene)

Ratio Ratio

A 5,0 0,6

B 4,7 0,6

C 4,0 0,4

D 5,0 0,4

E 4,0 0,9

From the combined analyses the conclusion may be drawn that a major

lithologic discontinuity is present between layers D and E and subordinate

discontinuities between layers A and B and between B and C.

The well-rounded, patinated gravel of layer E probably represents reworked

residual rubble of more resistant components lying on the pre-existing surface

and therefore reflects climatic conditions prevailing before the onset of collu-

viation. The residual character of this layer is further borne out by the higher

LATE QUATERNARY COLLUVIAL DEPOSITS 129

degree of weathering of the matrix material as exemplified by the higher zircon/

(epidote + pyroxene) ratio. The pyroxene most probably derives from the

diabase intrusions in the quartzite.

In layer D there is a sudden increase in coarse quartzite fragments, and

of much greater angularity, which would be indicative of high energy weathering

and mass transport over a short distance so that little sorting could take place,

as the relatively high percentage clay in the matrix seems to indicate. Upwards

there had been a gradual decline in transport energy conditions, the deposition

having taken place in phases, as minor discontinuities in the profile indicate.

RED SAND BEDS

Just below the break of slope a red sand layer commences, discordantly

overlying the gravel deposit. From this and other exposures it can be observed

that the red sand deposit, with increasing thickness, extends across the pediment

surface, sloping about 2°, to a line about halfway the distance between the

foot of the hillside and the lowest point (Modderspruit) in the longitudunal

valley between the Daspoort and Magaliesberg ranges, where the sand deposit

pinches out and weathered shale bedrock is exposed.

About one kilometre west of the tunnel entrance, on the upper section of

the pediment, an excavation shows the same sequence of gravel and discordant

red sand overburden, but the contact between gravel and underlying bedrock,

presumably shale, is not exposed. In the top of the gravel bed, directly under-

neath the red sand, a ferricrete horizon is present. The abrupt discordant boun-

dary between the sand and gravel most probably represents a denudation hiatus

and the ferricrete horizon, part of a former soil horizon, has been truncated

and buried by deposition of the colluvial red sand and is now fossil.

THE SANDFONTEIN SAND DEPOSITS

Lower down on the pediment over 6 m of sand is exposed in excavations.

No gravel layer is present in the profile and as yet it is not clear whether the

gravel from the upper part of the pediment extends underneath the sand to this

point.

A section through the sand deposit reveals three major zones, an upper, red

(SYR 5/8) sandy loam (81% sand, 5% silt, 14% clay), about 2,5 m thick,

followed by a yellow (SY 7/3) loamy sand (87% sand, 4% silt, 9°% clay), about

2 m thick, and finally a grey (2,5Y 7/2) sandy loam (82% sand, 4% silt, 14%

clay), 1,5 m thickness being exposed. The differences in colour are most probably

due to differences in groundwater conditions.

The most striking characteristic of the yellow-stained layer is the presence

of a high concentration of iron concretions (see Fig. 3). In the author’s opinion

this zone of iron accumulation is also fossil and the continuation of the ferri-

crete horizon exposed in the gravel layer at the upper section of the pediment

130 ANNALS OF THE SOUTH AFRICAN MUSEUM

il 2

: ee *«

=e se

i ee 1

¢ > ad os f 8)

as = “ Set

wes we ¢ i

Te |

ees ¥ AS se

oe ~ oo = SP ae = “

ok ¥

Fig. 3. Colluvial sand deposit showing the yellow-stained zone

bearing iron concretions (dark specks), overlain by red sand.

described above. This would imply that there is a gradual facies change from

gravel to sand in the downslope direction across the pediment. However,

further investigations would be necessary to determine the precise relationship

between the gravelly and sandy colluvial deposits at various points on the

pediment slope.

The upper, red zone of this sand deposit appears to correlate with the red

sand overlying the gravel bed at the upper end of the pediment. Support of this

inference is furnished by evidence of fluviatile sorting across the pediment as

shown by textural analysis of the red sand forming the smooth surface of the

pediment. Samples taken at intervals, and at the same depth of 10 cm, show a

downslope increase of the clay fraction from 8 per cent to 26 per cent and a

corresponding decline of the sand fraction from 87 per cent to 66 per cent.

DISCUSSION

Concerning the gravel deposit described above, Linton (1969: 85) quotes

a comparable instance from the northern slopes of the Magaliesberg, west of

Hartbeespoort dam, and attributes its origin to gelifluxion sludge, ‘of the type

termed ‘‘head’’ or “‘rubble drift’ in England’. According to Linton this and

other similar surficial deposits in South Africa should be considered distinctive

cold-climate forms and would serve ‘to indicate the reality of Pleistocene

cryonival and geliflual activity in South Africa’.

LATE QUATERNARY COLLUVIAL DEPOSITS 131

The present author cannot subscribe to the view of a distinct periglacial

origin for the deposits concerned, first of all, as cryergic activity of such magni-

tude would only be generated at temperatures fluctuating around 0°C, such

a drastic depression of temperature during the Pleistocene can hardly be assumed

for the Pretoria area, the present annual mean temperature being 17°C. Secondly,

the considerable proportion of fines contained in the matrix of the gravel, and

especially the vast sand deposits on the pediment, the lower beds probably

representing the downslope facies of the gravel deposit, attest to a comminu-

tion of quartzite bedrock, from which they have been derived, that could not

have been accomplished by mechanical frost-shattering, but rather by deep

chemical weathering under more humid conditions than prevailing at the

present time. The iron-staining and impregnation giving the sediments their

typical dark red colour, also indicate that the material has been released by

intense chemical weathering. The well-developed joint system of the quartzite

bedrock would permit of deep penetration of moisture, and hence weathering,

producing not only fines but also coarse subangular fragments as seen in the

gravel deposit

CONCLUSIONS

The major discontinuity in the colluvial deposits is that between the gravel

bed and the overlying red sand (Colluvia II and I in Fig. 1). The nature of the

deposits appears to indicate that the material originated from chemical weather-

ing of the quartzite hillslope and was translocated by running water (sheet and

rill wash). The two deposits represent climatological episodes during which

more humid conditions obtained than at present, the older of the two, the

gravel deposit, being attributable to fluviatile action and weathering of a greater

intensity than the upper, red sand beds. Presumably, therefore, the gravel

colluvium may be ascribed most reasonably to the last Pleistocene pluvial

period, whereas the red sand beds may date from an Holocene pluvial episode

of lesser intensity. The temperature depressions during pluvial periods would

certainly not have inhibited chemical weathering, but rather, the increased

availability of moisture, as a result of higher rainfall, higher frequency of

winter snowfall, and lower rate of evaporation, must have more than offset

the retarding effect of the lowering of temperature by, say, 5°.

It is also suggested by the present author that the sequence of two colluvia

described above corresponds to the generalized section of typical superficial

deposits as inferred by Cooke (1941: 26) for areas underlain by Karoo beds,

the gravel layer being the equivalent of Cooke’s pebble band, and the dis-

continuity between the gravel and the red sand overburden representing a

former land surface on which the soil horizon with ferricrete had developed.

Cooke terms this discontinuity the ‘Middle Stone Age horizon’, since artefacts

of that age occur at this level. In the Pretoria area Middle Stone Age and older

artefacts occur throughout the two colluvial layers.

132 ANNALS OF THE SOUTH AFRICAN MUSEUM

REFERENCES

Cooke, H. B. S. 1941. A preliminary survey of the quaternary period in southern Africa.—

Archaeol., Ser., S. Afr. 4. .

Davis, J. C. D. 1973. Statistics and data analysis in geology. Wiley: New York. i

Linton, D. L. 1969. Evidence of Pleistocene cryonival phenomena in South Africa. In: VAN :

ZINDEREN BAKKER, E. M. ed.—Palaeoecol. Afr. 5. k

REYNHARDT, J. H. 1975. ’n Bodemgeografiese studie van die Moot en aangrensende Bosveld

ten noordweste van Pretoria. Unpublished MSc thesis, University of South Africa.

SaMEs, C. W. 1966. Morphometric data of some recent pebble associations and their appli-

cation to ancient deposits. —J. sedim. Petrol. 36: 126-142.

PAST CLIMATE DERIVED FROM CALCRETE AND N-VALUE

H. H. WEINERT

National Institute for Road Research, CSIR, Pretoria

(With 2 figures and 2 tables)

ABSTRACT

Since many pedogenic processes are climate-sensitive, it should be possible to use certain

older pedogenic materials as indicators of the palaeoclimate. This statement is discussed on

the basis of outcrops of hardpan calcrete in the northern and central Transvaal where the

present climatic conditions do not appear to favour the development of this type of calcrete.

It is attempted to show that the interaction of certain climatic factors as expressed by the

N-value may provide a means of understanding the palaeoclimatic conditions during the

later Cenozoic. It is shown in particular that changes of individual climatic factors need not

be very great to cause such a change in the overall conditions that the formation of pedogenic

materials may be initiated or terminated.

CONTENTS

PAGE

Introduction ; : 2 ; ; : 133

Calcrete and climate . , ’ u ; 134

‘Fossil’ calcrete and past climate . : ue AS

Rainfall : : : 3 i 2 137

N-value ‘ ; 5 : ‘ : 138

Conclusions : x : : . . 140

Acknowledgement : 5 i ‘ . 140

References : : : y ; 5 140

INTRODUCTION

Although the road-building properties of weathered dolerite and other

basic crystalling rocks as revealed by research do not appear to tie up with our

present knowledge of the later Cenozoic of southern Africa, this research has

been responsible for a closer look being taken as to how the present climate

influences the weathering of these rocks.

That there is a relationship between the geological process of weathering,

the performance of weathered natural construction materials and the present

climate has always been assumed and as a result of the research mentioned a

better understanding has been obtained of the interaction of these processes.

Indeed, this research has provided the means from which the engineering

performance of weathered rocks can be predicted from a mathematical expres-

sion which defines the climatic equilibrium within which the rocks have been

weathering.

133

Proc. sth. Afr. Soc. Quat. Res. 1975

Ann. S. Afr. Mus. 71, 1976: 133-140, 2 figs, 2 tables

134 ANNALS OF THE SOUTH AFRICAN MUSEUM

This expression is

12E,;

Nees

where N is a numerical value

E, is the computed evaporation during January, the warmest month,

and

P, is the annual precipitation

The significant values of N are the integers 1, 2, 5 and 10 (Fig. 1), of which

N = 5 is the most important. The precise meaning of these integers as well

as the deduction of the expression (Weinert 1965, 1974) and the actual method

of calculation (Weinert 1974) have been described elsewhere.

CALCRETE AND CLIMATE

Certain pedogenic materials, particularly calcrete and ferricrete, are

sensitive to climate, a fact which has often been used to explain their distribu-

tion. Very frequently, probably because of the numerous stations in the world

where rainfall is recorded and which allow the compilation of relatively reliable

rainfall maps, the total annual rainfall has been the considered indicator.

Rainfall is, however, only one climatic factor and its influence on soils and

rocks does not depend only on the absolute quantity of rainwater that has

fallen whether monthly or annually, but rather on the time available to the

rainwater to act on soils and rocks before it is lost by run-off and evaporation.

In his work on calcrete, Netterberg (1969a, 1971) also considered that

there is a connection between rainfall and the formation and distribution of

these pedogenic materials. He suggested 550 mm and 800 mm of annual

precipitation as critical quantities: in areas receiving less than 550 mm rain per

year hardpan calcrete is common; where the rainfall is between 550 mm and

800 mm nodular calcrete can still be found and where the rainfall is more than

800 mm no calcrete occurs. This applies to most parts of southern Africa

although occasional deviations from these observations do of course occur.

Apart from this relationship to rainfall, Netterberg also observed that

there is a close relationship between the occurrence of certain types of calcrete

and the N-value in which the integers mentioned previously play a particular

role:

(i) where N is greater than 5 hardpan calcrete most commonly occurs,

(ii) where N is between 2 and 5 nodular calcrete is the most advanced stage of

development,

(iii) where N is less than 2 no calcrete is found.

It may be added that ferricrete occurs predominantly where N is less than 5

and that the areas of use of both pedogenic materials in road construction are

separated remarkably well by the contour of N = 5 (Fig. 2).

“COLIFY UIOYINOS IOJ SON[VA-N ONCUII[D Jo dew In0jUOD “| “34

ove oct off

135

"OIJY NYBHLNOS YO4 S3NTWA-N

DILVWITD JO dVW YNOLNOD | iy

IMVA -N it

HLIM SNOLWIS § § Nve"uNnde |

S |

»O& v

SSITTIVA-N DILVAWN ¢

WwNOS JO S3ININ 3 |

0% — |

AYVONNOG

JINVWYOSY3d AS

ONV ONIN3SHLV3M| ——— Sy

092 TVNISIVO

QN3931

ove

PAST CLIMATE DERIVED FROM CALCRETE AND N-VALUE

136 ANNALS OF THE SOUTH AFRICAN MUSEUM

22 24

CALCRETE Ea ZB 7

ee © \

e PIETERSBURG g

\ o +24

Rss» FERRICRETE

Ve. 7

SOUTH WEST AFRICA S \

ORANGE

28> lax vy fore s\\ A\SV\ y alee

Yy yy i. , | “\"

\. Uy Y Yj ZaCOEMPONTEIN va

ss SPRINGBOK — — fla . meas oy DURBAN

FIGURE 2: Occurrence

“N MSM}

WAS of roadworthy pedogenic

co ae tS ai Lb BRIBE TM Materials in relation

no WS So

Fig. 2. Occurrence of roadworthy pedogenic materials in relation to N = 5.

There are of course also exceptions to the latter observation, e.g. the

isolated occurrence of hardpan calcrete in the Springbok Flats of the central

Transvaal, where the present climatic environment has an N-value of less

than 5. An exception such as this may, however, be of particular interest, as

it may be indicative of the climatic conditions that prevailed in the area during

some part of the later Cenozoic and may provide the information necessary

for estimating palaeoclimatic conditions of certain localities.

The isolated occurrence of large quantities of hardpan calcrete in the

Springbok Flats appears to be an outlier of the tongue of hardpan calcrete

which stretches south of the Soutpansberg towards Pietersburg. The southern

boundary of this hardpan calcrete is about 15 to 20 km north of Pietersburg

while in the vicinity of Pietersburg and south to about Potgietersrus only

scattered calcrete nodules occur. There is a narrow zone of hardpan calcrete

near Potgietersrus, particularly on outcrops of the Transvaal dolomite. That

there has ever been a connection between the hardpan calcrete north of Pieters-

burg and that in the Springbok Flats can only be assumed but not be proved

at this stage.

If, however, we consider present-day climate in connection with these

calcrete occurrences, two apparent inconsistencies emerge.

PAST CLIMATE DERIVED FROM CALCRETE AND N-VALUE 137

1. Annual rainfall. Pietersburg receives 521 mm rain annually which is

less than the limit suggested by Netterberg (1969a, 1971). The southern

boundary of the hardpan calcrete, north of Pietersburg, is not reached

until the annual rainfall has dropped far below 500 mm. On the other hand, most

of those parts of the Springbok Flats where hardpan calcrete occurs, as well as

most of such occurrences near Potgietersrus, receive more than 550 mm and even

more than 600 mm rain per year. The only station in the Springbok Flats area

with a reasonably comprehensive weather record since 1920 is Kopje Alleen

(latitude 24.51S, longitude 28.41E), which has an annual mean rainfall of 621 mm.

2. N-value. The N-value of Pietersburg is 3,3 and, if N =—5 is considered

as the limit of the occurrence of hardpan calcrete, the absence of this most

developed type of calcrete at Pietersburg appears to be in line with this obser-

vation. In contrast, the N-value of the country between Pietersburg and the

Soutpansberg, where there is much outcropping hardpan calcrete, is also

generally less than 5 although the presence of smaller isolated areas where N

is more than 5 cannot be entirely excluded. The N-values of the Springbok

Flats are less than 5, that of Kopje Alleen being 3,9, and yet much hardpan

calcrete is present.

“FOSSIL? CALCRETE AND PAST CLIMATE

Netterberg’s (1969a, b, c) work has revealed that the hardpan calcretes

north of Pietersburg as well as those in the Springbok Flats are ‘fossil’, which

means that they are remnants of past conditions which were more favourable

for their formation than those of today. It is obvious that calcrete, once it

has been formed, can be preserved in a climatic environment not favourable

to its formation. Both the stage of development attained by the calcrete and

the degree of difference between the original and any later climatic conditions

will affect the duration of this ‘survival’.

A hardpan calcrete is the most advanced stage of calcrete development

and it will be preserved longer and under more severe long-term changes of

climate than any other stage. ‘Fossil’ calcrete may therefore be taken as an

indicator of climatic conditions of the past and this will be particularly the

case if the hardpan stage had been reached.

The occurrences of hardpan calcrete north of Pietersburg and in the

Springbok Flats will now be considered in the light of climatic changes, and

in the discussion which follows it will be assumed that the presence of the most

advanced stage of calcrete development really is an indicator of the climate

that prevailed during the period of its formation, i.e. a N-value of 5 and more

or a mean annual rainfall of less than 550 mm.

Rainfall

It has been said that the country between Pietersburg and the Soutpans-

berg, in fact the whole area between the Strydpoortberge and the Soutpansberg,

receives less than 550 mm rain per year. This means that, if rainfall were the

138 ANNALS OF THE SOUTH AFRICAN MUSEUM

sole climatic factor which favours or disfavours the formation of hardpan

calcrete, this material should be present in the whole area. The present climate

or rather the present rainfall would then be that which also caused the occurrence

of the hardpan calcrete whose southern boundary is now about 15 to 20 km

north of Pietersburg and which coincides better with the 450 mm isohyet of

the present annual rainfall. It should be noticed that the other pedogenic factors,

parent material, topography, biology and time are not considered in this paper.

The annual rainfall in most of the Springbok Flats where hard,an calcrete

occurs 1s now more than 550 mm, and in parts it is even more than 600 mm;

for instance, the annual rainfall at Kopje Alleen is 621 mm. The annual averages

in this area may vary from decade to decade by more than 100 mm. The calcrete

in this area appears to be ‘fossil’ and a long lasting decrease in the annual rain-

fall of the order of 70 to more than 100 mm, i.e. toward the lower end of the

present ‘normal’ variation, would be required if the formation of hardpan

calcrete were governed by rainfall alone. This would be a considerable decrease

which would also have a noticeable effect on other climatic factors. Presuming

such a decrease the rainfall in the Springbok Flats would then be of the same

order as it is now about one degree further north, and the general climatic

conditions would be rather different from those pertaining at present in the

country between the Strydpoortberge and the Soutpansberg.

N-value

All the areas in the northern and central Transvaal which have been dis-

cussed possess outcrops of hardpan calcrete and in all of them the N-value is

obviously less than 5. Most hardpan calcrete in southern Africa, however,

occurs where N is more than 5 (Netterberg 1969a, 1971).

The following climatic data are required for the determination of the

N-value:

(1) the mean annual rainfall;

(2) the average air temperature of the warmest month in southern Africa;

this is usually January;

(3) the average relative humidity of the warmest month; and

(4) the mean wind speed of the warmest month.

Applying Olivier’s (1964) expression for daily evaporation and a wind

correction, the N-value is calculated from:

372 c,(0,58 + 0,14 V;)

IN an er eee

Pa wos

Where c, is the wet bulb depression in °C for January;

V, is the mean wind speed of January in m/s;

P, 1s the mean annual rainfall in mm, and

wos is the correction for the monthly field-water requirements

depending on latitude (Olivier 1964)—this is 0,78 for the latitude

of Pietersburg and 0,79 for that of the Springbok Flats.

PAST CLIMATE DERIVED FROM CALCRETE AND N-VALUE 139

It has already been said that the N-value of Pietersburg is 3,3 and that of

the Springbok Flats is between 3 and 4, that of Kopje Alleen being 3,9. Any

individual climatic factor would have to change considerably, leaving all others

unaffected, for the N-value to change to 5; for instance, if the N-value of the

two stations were to increase to 5, e.g. the annual rainfall at Kopje Alleen

would have to drop from 621 mm to about 480 mm, and the relative humidity

at Pietersburg, where the relative humidity during January is remarkably high,

would have to decrease from the present 69 per cent to 52 per cent.

In any climate, the change of any one factor is influenced by the overall

situation and may affect all the other climatic factors. A decrease in the rainfall

will thus mostly be associated with a decrease in the relative humidity, possibly

an increase in the air temperature and perhaps also in the average wind speed.

Considering therefore the climatic changes required to increase the N-value

at Pietersburg and in the Springbok Flats to 5, an entirely different picture

emerges from that in which rainfall alone was being considered as the only

cause of the formation of hardpan calcrete.

The average relative humidity during January at Pietersburg is 69 per cent

nowadays and it would have to decrease to about 60 per cent to agree with

the present conditions north and south of the town. Such a decrease would

most likely be associated with a slight rise in the average January temperature

and the wind speed, and a minimal decrease in the total annual rainfall might

also occur.

The following situation would then be one of many which would change

the N-value of Pietersburg from the present 3,3 to 5:

Table 1 Possible change in conditions required at Pietersburg to bring N-value to 5

Climatic factors Present condition New condition

Mean annual rainfall ; ; : : ; 5 521 mm 500 mm

Mean temperature in January . : ‘ : : ilo (C Dane

Mean relatieve humidity in January . ; : : DY, 60%

Mean wind speed of January . ; ; 5 ; 2,6 m/s 2,9 m/s

N-value : ; . , ; ‘ 5 ; 3558) B)

At most localities of the Springbok Flats, the major change would have to

occur in the annual rainfall which is now 621 mm at Kopje Alleen. A com-

parison of the relevant relations at all major weather stations of the Republic

reveals that, when the annual rainfall is of the order of 500 to 600 mm, a decrease

of 4 to 5 per cent in the relative humidity of the warmest and most rainy month

would be a reasonable estimate if the total annual precipitation decreased by

50 mm. If the average wind speed of the warmest month remains unchanged,

the average temperature of the warmest month might rise by one degree

Celsius.

To bring the N-value of Kopje Alleen to 5 the following changes in the

climatic conditions represent one of many possibilities:

140 ANNALS OF THE SOUTH AFRICAN MUSEUM

Table 2

Possible change in conditions required at Kopje Alleen to bring N-value to 5

Climatic factors Present condition New condition

Mean annual rainfall i : A : : : 621 mm 570 mm

Mean temperature in January . : ‘ 2 ; 23.5 € 24,5°C

Mean relative humidity in January . s ’ : 62% 58%

Mean wind speed of January . : é : ; 3,1 m/s 3,1 m/s

N-value . ‘ : ; , : : t : 39 5

CONCLUSIONS

The above two tables show that no radical changes in the present climatic

conditions would be required if the value of N = 5 were considered a require-

ment of the development of hardpan calcrete in the Pietersburg, the Springbok

Flats and possibly all other relevant areas. It is very unlikely that a natural

event, such as the pedogenic process of calcrete formation, depends only on

one single climatic factor. With numerous interacting climatic factors, however,

each individual factor need not change much to bring about such an alteration

in the general conditions that calcrete development may begin or be brought

to an end. A mathematical expression which is a ratio of the interaction of

various climatic parameters offers a means of estimating such past conditions,

particularly if the former climate can be connected to the formation of datable,

climatic-sensitive materials such as calcrete. The N-value has been used to

demonstrate this and it shows that all required variations are small and well

within those fluctuations which ‘normally’ occur today. This may indicate

even that no radical changes of climate may have occurred in the areas concerned

at least during the later Pleistocene.

ACKNOWLEDGEMENT

This paper is published by permission of the Director, National Institute

for Road Research, Pretoria, Republic of South Africa.

REFERENCES

NETTERBERG, F. 1969a. The geology and engineering properties of South African calcretes,

Ph.D. thesis, University of the Witwatersrand, 4 volumes.

NETTERBERG, F. 19695. The interpretation of some basic calcrete types.—S. Afr. archaeol.

Bull. 24: 117-122.

NETTERBERG, F. 1969c. Ages of calcrete in Southern Africa.— S. Afr. archaeol. Bull. 24: 88-92.

NETTERBERG, F. 1971. Calcrete in road construction.— CSIR Res. Rep. 286, NIRR Bull. 10.

Ouivier, H. 1964. Irrigation and climate. 2nd ed. London: Edward Arnold.

WEINERT, H. H. 1965. Climatic factors affecting the weathering of igneous rocks.— Agric.

Meteorol. 2: 27-48.

WEINERT, H. H. 1974. A climatic index of weathering and its application in road construction. —

Géotechnique 24: 475-488.

eee

SECTION 3

PALAEOQECOLOGY

ON A ee — 7 i oe. = 7 a) ae . ata

4) Le ,ow a: 1 ae e a"

Wf = =

’ Poss a ; n¢@

oF > Ve

- ~ - f

4 iy La

= e

« . 7 e

- ae

i r

ii |

) aa

iF ~

| hi

{

| a

iy i

Mi)

=—

Bi ‘I

i) hi

rit i

p 4

| yr

it '

“hh

i ra

)

~ = <—

4 .

,) |

yi Ly

}

\ sf “

} 4 _—*

ita ‘ 'P-

Ms ti) : - ,

“ | ra - af

hy, 1) j

[ }

Gh ee =

LATE QUATERNARY ENVIRONMENTAL CHANGES IN

SOUTHERN AFRICA

E. M. VAN ZINDEREN BAKKER SR.

Institute for Environmental Sciences, University of the Orange Free State,

Bloemfontein

(With 5 figures)

ABSTRACT

The climatic regime of southern Africa is strongly influenced by the evolution of the

Antarctic ice sheet which dominates the atmospheric and oceanic circulation. This climatic

system is of mid-Tertiary age. During the Quaternary glaciations world-wide lowerings in

temperature caused northward shifts of the climatic belts in the southern hemisphere. As a

consequence of these shifts southern Africa was much more exposed to the zone of the Wester-

lies with their strong winds, the invasion of many depressions, regular influxes of cold polar

air and much more rain than at present.

The drastic changes which occurred in the major ecological regions during the last

glacial cycle are described in broad outline. From archaeological and radiometric datings the

inference can be made that during full-glacial and interglacial periods climatic conditions in

several regions were so adverse to human life that long time gaps exist in the occupation of

prehistoric sites in these regions.

CONTENTS

PAGE

Introduction : : : 5 : oe ai

The glacial climate : : i ; , ay

Environmental changes 4 F . 144

The austro-afro alpine region . 145

The plateau of the Orange Free State 5 2) F45

The Cape coastal region : . 146

The west coast and the Kalahari Basin . 148

References . : : ; : : 5 ISO

INTRODUCTION

The present general systems of atmospheric and oceanic circulation of the

southern hemisphere are of mid-Tertiary age. These systems could develop

only after the greater part of Antarctica was ice-covered and the South Polar

heat sink had originated. The production of cold Antarctic Bottom Water

(Shackleton & Kennett 1975) and the formation of the Circum-Antarctic

Current (Kennett et al. 1974) in Oligocene times caused the present general

distribution of the different water masses in the Antarctic Ocean.

The atmospheric circulation was strongly affected by the great difference

in temperature which had developed between Antarctica and the tropics and

which had steepened the pressure gradient between these regions. The result

of this development was a strong zonal circulation in the upper troposphere

which had again a great influence on the position of the high and low pressure

141

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 141-152, 5 figs.

142 ANNALS OF THE SOUTH AFRICAN MUSEUM

cells of our surface climate and therefore on the steering of the atmospheric

and oceanic circulations.

The present climatic system of southern Africa is dominated by the following

three climatic zones:

1. The Inter-Tropical Convergence Zone (1.T.C.Z.) which moves northward

and southward with the sun. This tropical system brings precipitation to

meridional zones of Africa about as far south as the parallel of 20°. Further

southward the interference of the next system disturbs the regular rainfall

distribution because of a strong aridifying influence coming from the west.

2. The subtropical zone, which in the region of southern Africa consists of

three anticyclones, one over the oceans on either side of the subcontinent

and one which persists over the plateau of the interior at 2 000 m altitude

(Jackson 1952; Trewartha 1966). The drying influence of these anticyclones,

which are subject to small seasonal meridional displacements round 30°S.

latitude, is at present the most important component of the climate of

southern Africa. The more stable South Atlantic high pressure system has

the greatest influence on this climate as it is situated much nearer to the

coast than its Indian Ocean counterpart. This extremely aridifying influence

along the west coast which reaches far inland is aggravated by the Benguela

Current and its associated cold upwellings, which are responsible for the

existence of the hyper-arid Namib Desert. The anticyclone over the Indian

Ocean is situated further away from the continent and humid maritime

and tropical air can as a consequence often enter the eastern half of southern

Africa.

3. The zone of mid-latitude westerlies moves slightly northward in winter and

then brings cyclonic rain to the south-western Cape coastal area.

This climatic system is ancient, but the topographical distribution of

its ocean and wind currents changed significantly in the course of time. The

very short description given above applies to the conditions prevailing

during the second half of the interglacial in which we are living at present.

Variations of considerable magnitude have occurred during the Quaternary

glacial and interglacial maxima and during stadials and interstadials which

were of different intensity and duration.

THE GLACIAL CLIMATE

During the last glacial important changes took place in the oceans

surrounding southern Africa. Oceanographic research by Hays et al. (1976)

has shown that during the coldest maximum of this glaciation of Wurm-

Wisconsin age, 18000 years B.P., the Antarctic Polar Front was displaced

considerably northward, especially in the Atlantic and Indian Ocean sectors

but only slightly south of South Africa. The cold Antarctic Upper Water

penetrated from 6 to 10° latitude northward and pack ice covered the Antarctic

Ocean even in summer as far north as 55°S. The cooling effect of these changes

LATE QUATERNARY ENVIRONMENTAL CHANGES IN SOUTHERN AFRICA 143

greatly steepened the atmospheric pressure gradient and as a consequence

strongly activated the atmospheric and oceanic circulations. Cold air and water

masses could penetrate much further northward and must have had a profound

influence on the mid-latitudes and the subtropical regions of the southern

hemisphere. The general world-wide lowering in temperature during the glacial

maximum will consequently have been greater in those parts which were exposed

to these polar influences. A typical example of this is the sub-antarctic island

Marion which was glaciated during the last glacial period (Van Zinderen

Bakker Sr 1969).

The northward displacement of cold water along the west coast of southern

Africa during the last glaciation has been studied by Bornhold (1973). The

results of his investigations of marine sediments off the coast of Angola can

only be explained by the assumption that a northward shift of the Atlantic

Ocean anticyclone forced the Benguela Current with its high productivity

water along the well-aligned coast north of Mossamedes. The drying influence

of this system must have affected western Angola, Zaire and Congo as has been

anticipated by the author (Van Zinderen Bakker Sr 1967).

It is possible that cold water also penetrated further north along the

south-east coast of South Africa so that the warm Agulhas Current could not

reach the south coast of the continent as is the case at present. Vincent (1972)

has shown that the temperature of the surface water in the southern end of the

Mocambique Channel before 10 000 years B.P. was 5°C colder than at present.

The consequences of these considerable changes in the atmospheric and

oceanic circulation, which aggravated the general lowering in temperature,

were dramatic for the palaeoenvironment of southern Africa. The northward

movement of the climatic zones extended the influence of the westerlies with

their accompanying winter rainfall, low temperatures and great windforce.

The cold fronts, which caused regular influxes of very cold polar air deep into

southern Africa, could, by comparison with present conditions at higher

latitudes, have penetrated the continent a hundred times a year. It is therefore

possible that in exposed areas temperature drops of 10°C occurred regularly,

especially in winter. Cryoclastic phenomena as described by Butzer (1973),

‘periglacial’ features as studied by a number of authors (Alexandre 1962;

Sparrow 1967, 1974; Harper 1969; Hastenrath 1972) and the production of

roof spall in caves of the Cape coastal region (Butzer 1973) could well be

explained by recurrent drastic lowering in temperature in winter. Decreases in

temperatures of 8 to 10°C have been suggested by Harper (1969) for the high

mountains of Lesotho, by Butzer, (1973) for the phenomena at the southern

coast and by Talma et al. (1974) even for the central Transvaal. More palaeo-

temperature determinations will, however, be necessary to come to an accurate

assessment of the average temperature during the last glacial maximum.

The cyclonic winter rain may not have penetrated southern Africa as far

as the influxes of polar air. Indications for the northern limit of these cyclonic

rains can perhaps be obtained from biogeographic evidence in the southern

144 ANNALS OF THE SOUTH AFRICAN MUSEUM

25 var

SOUTH AFRICA

35° 350

Fig. 1. Locality map of the interior plateau and the Kalahari region.

1 = Okavango delta; 2 = Mababe depression; 3 = Lake Ngami; 4 = Makarikari;

5 = Florisbad; 6 = Alexandersfonteinspan; 7 = Aliwal North; K = Kimberley.

Namib (Haacke in Van Zinderen Bakker Sr 1975) from the high lake levels of

Alexandersfonteinpan (Butzer, Fock et al. 1973), the pollen analytical results

of Aliwal North (Coetzee 1967) and Florisbad (Van Zinderen Bakker Sr 1957)

and the sequence of the Riverton Formation of the Vaal Basin (Butzer, Helgren

et al. 1973). It is well possible that the cyclonic trajectories were displaced north-

ward in winter by at least 10° latitude, so that they could in certain localities

reach the 24th parallel of latitude. During interglacial maxima the reverse

development will have taken place and tropical rain could freely penetrate the

Transvaal while the Orange Free State and the Cape Province would generally

have had a fairly dry to very dry climate as the winter rain did not touch the

south-western Cape area regularly.

ENVIRONMENTAL CHANGES

The impact of the drastic climatic changes on the vegetation will be treated

in more detail elsewhere (Van Zinderen Bakker Sr 1976). On purely theoretical

grounds certain vegetation reconstructions can be implied, but is not yet possible

LATE QUATERNARY ENVIRONMENTAL CHANGES IN SOUTHERN AFRICA 145

to draw any maps of former vegetation patterns as pollen analytical records

are so far only available for the Cape coastal region and the inland plateau.

These reconstructions can therefore be attempted only in very broad outline

and will apply primarily to the rather extreme conditions of the warmest and

coldest periods. It should, however, be realized that during the long duration

of the Quaternary an endless variation of ecological conditions occurred. These

various stages, which could have been of long duration, all form part of the

full glacial-interglacial cycle. For some regions the important stages will be

described using the terminology proposed by Von Post (1946), Firbas (1949: 110)

and Iversen (1954, 1958) in which the stage of the maximum cold is called the

cryocratic stage. The protocratic stage with rising, though oscillating, tempe-

rature leads to the mesocratic ecological stage, during which the temperature

reaches optimum conditions. The cycle is closed by the telocratic stage during

which temperature is declining.

The austro-afro alpine region

The ‘periglacial’ phenomena described from the higher altitudes of the eastern

escarpment suggest that during the last glacial maximum, depending on the

accepted temperature gradient, a drop in temperature of 5,5 to 9,0°C could

have occurred (Harper 1969). Under such conditions the upper limit of the

vegetation will have been lowered by some 1 000 m. This drastic change will

have driven all the trees and ravine forests away from altitudes above 1 100

to 1 200 m and will have robbed the upper reaches above about 2 000 m of

all their vegetation. Radiocarbon dates of the age of the swamps, which occur

at about 3 000 m altitude in Lesotho, show that vegetation only settled here

again after the beginning of Hypsithermal warming of the climate (Van Zinderen

Bakker Sr & Werger 1974). It would certainly have been impossible for Middle

Stone Age and Later Stone Age people to live permanently in the mountains

at higher altitudes during full-glacial or late-glacial times. Willcox (1974) offers

this possibility as one of his explanations for the lack of prehistoric settlements

of such age at higher altitudes.

The plateau of the Orange Free State (Fig. 2)

The present-day grassveld plateau of the Orange Free State and adjacent

areas has been subjected to radical ecological changes. During full-glacial

times the alpine grassland, which at present covers the eastern escarpment

above the tree limit, will have invaded this plateau under severely cold and

rainy conditions. Climatic amelioration will have led to the protocratic condi-

tions of the second half of our present interglacial which are characterized by

Sweet grassveld (Cymbopogon-Themeda) with summer rain and limited night

frost in winter. The warmest mesocratic period of the interglacial must,

according to pollen analytical results, have seen the invasion of the semi-arid

Karoo (Van Zinderen Bakker Sr 1957; Coetzee 1967) (Fig. 3). According to

146 ANNALS OF THE SOUTH AFRICAN MUSEUM

interglacial

DESERT ky

of ee

protocratic

ALPINE GRASSLAND

SAVANNA

WNNWAWS

GNWISSWHD ALWeddWaL

TEMPERATE GRASSLAND

glacial

Fig. 2. The vegetation types of the full glacial-interglacial cycle of the Highveld and Cape

Plateau.

geomorphological information (Butzer, Helgren et al. 1973) even drier conditions

in the western part promoted the spread of windblown sanddunes.

The valuable study by J. Deacon (1974) indicates that this inland plateau

was inhospitable to prehistoric hunter-gatherers between 9500 and 4 600

years B.P. During this period no occupation sites are known on this plateau,

which was then covered by a very arid vegetation.

The Cape coastal region

This coastal region received the full impact of the strong winds; the influxes

of polar air and heavy rainfall were not limited to winter only. The cryoclastic

evidence and the fossil assemblages found in the caves suggest that during full

glacial times the coastal plain was mainly covered by grassland (Butzer &

Helgren 1972; Klein 1974).

LATE QUATERNARY ENVIRONMENTAL CHANGES IN SOUTHERN AFRICA 147

European Chronostratigraphy

(B.P.)

Stratigraphy] Vegetation

probably 4400

Atlanticum

probably z

8 7200

3°

dry very dry

Karoo warm Boreal

PS eS

occupation

sites

Basal Peat

a ea ee =

9650 Zone 2 Karoid very dry, warm Preboreal

dry grassveld 10100

S Zone Y grassveld wetter, cooler Upper Dryas

=e 11250 2 10900

Bs Zone X Karoo dry, warm < Aller6éd Int.

Zo 11650 3 11850

2) Zone W grassveld wet, C) Older Dryas

fo)

=O ca 12200 & 12100

<P Zone V Karoid | Bélling Int.

nm |ca 12600 12400

ca 13185

Zone U grassveld Oldest Dryas

12600

el ee ee ee ee St ee eee ae

= 19350 Peat III aieine wet, cold Upper

2 Jca 25000 eee grassveld M.S.A Pleniglacial]

A oscillations: 4

ax 28.450 Peat II dry M.S.A wet - cold 5

Cs Karoo eee to 4 29000

“ee dry - warm o Denekamp Int.

£ |> 48900 Peat I

~~“

wet oscillation

Karoo very dry, warm

Fig. 3. The interior plateau during the Late Quaternary. Chronostratigraphy and environ-

mental changes.

These climatic conditions will have been most adverse to the life of pre-

historic man. In this connection Klein (1974) gives an interesting summary

of the hiatus in occupation which is known for the caves in the coastal region

and further inland. This gap, which is in the order of tens of thousands of years,

falls between the occupation by MSA and LSA people. It has been suggested

that the very inhospitable climate during full-glacial times may have been one

of the factors which had driven man away from those caves which were very

exposed (Van Zinderen Bakker Sr 1976). Other factors, such as scarcity of food

resources or the great distance from the sea, may well have contributed to the

non-occupance of the caves during cryocratic times.

The different stages of the glacial-interglacial ecological cycle are not known

for the Cape coastal region, but the pollen analytical results obtained by Schalke

(1973) and Martin (1968) provide some valuable information which can be

combined with palaeontological evidence (Klein 1974). The following general

explanation has been offered (Van Zinderen Bakker Sr 1976) (Fig. 4). In the

Cape coastal region temperature and humidity are negatively correlated as

can be concluded from the winter-rainfall climatic pattern. Evergreen forest

could spread widely only during protocratic and telocratic times when tempe-

rature and humidity reached medium values. These conditions prevailed at the

148 ANNALS OF THE SOUTH AFRICAN MUSEUM

interglacial

SEMI-ARID KAROO

protocratic S

GRASSLAND

glacial

MACCHIA-FOREST

LSAYOd-WIHOOWW

Fig. 4. The vegetation types of the full glacial-interglacial cycle of the southern Cape coastal

region.

beginning and end of the Holocene thermal optimum. During the optinium

the climate was too dry for the evergreen forest, and Karoo-like vegetation

and dunes spread along the coastal plain. Macchia and Podocarpus forest,

which are not so sensitive to frost, will have been dominant during protocratic

and telocratic times respectively before conditions were favourable for the

coastal evergreen forest. These considerable changes in climate and vegetation

may not have been of the same nature in the entire extended coastal region.

Much more research is needed, especially pollen analytical studies, to elucidate

the complicated evolution of the environment in this region.

The west coast and the Kalahari Basin (Figs 1, 5)

As previously discussed, the aridifying influence of the South Atlantic

anticyclone and the Benguela Current moved in a meridional direction north-

ward during colder periods and poleward during warmer episodes. Hyper-

LATE QUATERNARY ENVIRONMENTAL CHANGES IN SOUTHERN AFRICA 149

Pee ieee Mts.

Bettas Sata

Ba es

By > Sai: b R.

ah Ra Ya

A SossuB Vlei

: om i.

ae ie suns

ete os Bg Bit poses ES

ard et ose)

26 26

Orange R

Fig. 5. Locality map of the Namib Desert.

150 ANNALS OF THE SOUTH AFRICAN MUSEUM

desertic conditions existing in the coastal region and Kalahari sand movements

are strongly correlated with this system (Van Zinderen Bakker Sr 1975). Some

radiocarbon dates from northern Angola support the inference that the rede-

posited Kalahari Sand members were active during colder periods.

Too little information is at present available for the assessment of former

changes in rainfall in this vast region. On theoretical grounds it can be assumed

that at certain warmer stages tropical rainfall invaded the northern part, while

winter rainfall may have penetrated the southern Namib and perhaps even

Botswana during full-glacial periods. A study of the history of the large pans

in Botswana, especially of those which are not connected with the Okavango—

Botletle river system, could throw much light on former climatic conditions.

Similar studies of the isolated pans which occur in the barchan dunes of the

southern Namib are of great importance as is already shown by the description

of the Tsondab vlei by Seeley & Sandelowsky (1974). The vlei deposits contain

much Early Stone Age material and it appears as if the Tsondab River, which

originates in the Naukluft Mountains to the east, was able to reach the ocean

during a former period with higher rainfall. The Sossus vlei further south

could also yield important palaeoclimatic information. The Mirabib Rock Shelter

north of the Kuiseb River, which is being studied by Sandelowsky (1974),

contains a wealth of information covering a period of nearly 10000 years.

From a biological viewpoint it seems as if former higher rainfall in the

southern Namib Desert could only have been of a southern origin and must

consequently have been winter rain which penetrated far northward during

full-glacial times. Summer rain of I.T.C. origin could only reach the southern

Namib during abnormal mesocratic conditions when the South Atlantic anti-

cyclone was situated extremely far south. A climatic assumption of the latter

kind cannot offer a valid explanation as it would mean that the whole Namib

received summer rainfall over its full length. Such ‘pluvial’ conditions are not

congenial with the rich hyper-desertic adaptations which are found in the

barchan biota. These speciations point to great antiquity and an undisturbed

environment. It can therefore be accepted that a rainfall maximum of a northern

origin can never have affected the core of the Namib Desert (Koch 1962).

An attempt will not be made to elaborate on the Late-Quaternary history

of the vast woodland regions which cover extensive areas of southern and East

Africa. Detailed information on this region is still lacking, although it is known

that these woodlands received less rainfall during the last glacial (Hamilton

1972) and were favoured with higher humidity during warmer periods. It is

also not possible to venture into the unknown climatic and vegetational history

of the east coastal region of southern Africa.

REFERENCES

ALEXANDRE, J. 1962. Phénoménes périglaciaires dans le Basutoland et le Drakensberg du

Natal.— Biuletyn Peryglacjalny 11: 11-13.

BORNHOLD, B. D. 1973. Late Quaternary sedimentation in the eastern Angola Basin.— Woods

Hole Oceanogr. Inst. WHOI-73-80. Unpublished manuscript.

LATE QUATERNARY ENVIRONMENTAL CHANGES IN SOUTHERN AFRICA 151

BuTzer, K. W. 1973. Pleistocene ‘periglacial’ phenomena in southern Africa.— Boreas 2: 1-11.

BuTZER, K. W. & HELGREN, D. M. 1972. Late Cenozoic evolution of the Cape coast between

Knysna and Cape St. Francis, South Africa.— Quatern. Res. 2: 143-169.

BuTzer, K. W., Fock, G. J., STUCKENRATH, R. & ZILCH, A. 1973. Palaeohydrology of Late

Pleistocene Lake, Alexandersfontein, Kimberley, South Africa.— Nature, Lond. 243

(5406): 328-330.

BuTzer, K. W., HELGREN, D. M., Fock, G. J. & STUCKENRATH, R. 1973. Alluvial terraces

of the lower Vaal River, South Africa: a re-appraisal and re-investigation.—J. Geol. 81:

341-362.

CoETZEE, J. A. 1967. Pollen Analytical Studies in East and Southern Africa.— Palaeoecology

of Africa Til: 1-146.

DEACON, J. 1974. Patterning in the radiocarbon dates for the Wilton/Smithfield complex in

Southern Africa.— S. Afr. archaeol. Bull. 29: 3-18.

FirBas, Fr. 1949. Waldgeschichte Mitteleuropas 1. Jena: Gustay Fischer.

HamMILTon, A. C. 1972. The interpretation of pollen diagrams from highland Uganda.—

Palaeoecology of Africa VIL: 45-149.

HARPER, G. 1969. Periglacial evidence in southern Africa during the Pleistocene epoch.—

Palaeoecology of Africa 1V: 71-101.

HASTENRATH, S. 1972. A note on recent and Pleistocene altitudinal zonation in Southern

Africa.— S. Afr. J. Sci. 68: 96-102.

Hays, J. D., LoZANo, J. A., SHACKLETON, N. J. & IRVING, G. 1976. An 18000 year B.P.

reconstruction of the Atlantic and Indian sectors of the Antarctic Ocean. In: CLINE, R. M.

& Hays, J. D. eds. (In press.)

IVERSEN, J. 1954. The late-glacial flora of Denmark and its relation to climate and soil.—

Geol. Survey Denmark, ser. 2, 80.

IVERSEN, J. 1958. The bearing of glacial and interglacial epochs on the formation and extinction

of plant taxa. Uppsala Univ. Arsskrift (b): 210-215.

JACKSON, S. P. 1952. Atmospheric Circulation over South Africa.—S. Afr. Geograph. J. 34:

48-59. -

KENNETT, J. P., Houtz, R. E., ANDREws, P. B., EDwarps, A. R., GOSsTIN, V. A., HAJos,

M., Hampton, M. A., JENKINS, D. G., MARGOLIS, S. V., OVENSHINE, A. T. & PERCH-

NIELSEN, P. 1974. Development of the Circum-Antarctic Current.— Science 186 (4159):

144-147.

KLEIN, G. R. 1974. Environment and subsistence of prehistoric man in the Southern Cape

Province, South Africa.— World Archaeol. 5: 249-284.

Kocn, C. 1962. The Tenebrionidae of Southern Africa. XXXI. Comprehensive notes on the

tenebrionid fauna of the Namib desert. Ann. Tvl. Mus. 24: 61-106.

Martin, A. R. H. 1968. Pollen analysis of Groenvlei Lake sediments, Knysna (South Africa).—

Rev. Palaeobotan. Palynol. 7: 107-144.

Post. L. von. 1946. The prospect for pollen analysis in the study of the earth’s climatic

history.— New Phytol. 45: 193-217.

SANDELOWSKY, B. H. 1974. Archaeological investigations at Mirabib Hill Rock Shelter.—

S. Afr. archaeol. Soc., Goodwin Ser. 2: 65-72.

SCHALKE, H. J. W. G. 1973. The Upper Quaternary of the Cape Flats Area (Cape Province,

South Africa).— Scripta Geol. 15: 1-57.

SEELEY, M. K. & SANDELOWSKY, B. H. 1974. Dating the regression of a river’s end point.—

S. Afr. archaeol. Soc., Goodwin Ser. 2: 61-64.

SHACKLETON, N. J. & KENNETT, J. P. 1975. Paleotemperature history of the Cenozoic and

the initiation of Antarctic glaciation: oxygen and carbon isotope analyses in DSDP

sites 277, 279, and 281. In: KENNETT, J. P., Houtz, R. E., et al. Initial Reports of the

Deep Sea Drilling Project 29. Washington: U.S. Government Printing Office.

SPARROW, G. W. A. 1967. Pleistocene periglacial topography in South Africa.—J. Glaciology

6: 551-59.

SPARROW, G. W. A. 1974. Periglacial cirque formation in southern Africa.— S$. Afr. archaeol.

Soc., Goodwin Ser. 2: 25-28.

TALMA, A. S., VoGEL, J. C. & PARTRIDGE, T. C. 1974. Isotopic Contents of some Transvaal

Speleothems and their Palaeoclimatic Significance.— S. Afr. J. Sci. 70: 135-140.

TREWARTHA, G. T. 1966. The Earth’s Problem Climates. London: Methuen.

152 ANNALS OF THE SOUTH AFRICAN MUSEUM

VINCENT, E. 1972. Climatic change at the Pleistocene-Holocene boundary in the southwestern

Indian Ocean. — Palaeoecology of Africa V1: 45—54.

WiLLcox, A. R. 1974. Reasons for the non-occurrence of Middle Stone Age material in the

Natal Drakensberg.—S. Afr. J. Sci. 70: 273-274.

ZINDEREN BAKKER Sr, E. M. VAN. 1957. A pollen analytical investigation of the Florisbad

deposits (South Africa).— Proc. 3rd Pan-Afr. Congr. Prehist. Livingstone 1955: 56-67. |

ZINDEREN BAKKER SR, E. M. VAN. 1967. Upper Pleistocene and Holocene stratigraphy and

ecology on the basis of vegetation changes in sub-Saharan Africa. In: W. W. BisHop

et al. eds. Background to evolution in Africa: 125-147. Chicago: University of Chicago

Press.

ZINDEREN BAKKER SR, E. M. VAN. 1969. Quarternary pollen analytical studies in the southern

hemisphere with special reference to the sub-Antarctic.—Palaeoecology of Africa V:

175-212.

ZINDEREN BAKKER SR, E. M. VAN. 1975. The origin and palaeoenvironment of the Namib

Desert biome.—J. Biogeography 2: 65-73.

ZINDEREN BAKKER Spr, E. M. VAN. 1976. The Evolution of Late Quaternary Palaeoclimates

of Southern Africa, Palaeoecology of Africa IX (in press).

ZINDEREN BAKKER SR, E. M. VAN & WERGER, M. J. A. 1974. Environment, vegetation and

phytogeography of the high-altitude bogs of Lesotho.— Vegetatio 29: 37-49.

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS,

TRANSVAAL, SOUTH AFRICA

G. E. CoLiincs, A. R. I. CRUICKSHANK, J. M. MAGUIRE & R. M. RANDALL

Bernard Price Institute for Palaeontological Research,

University of the Witwatersrand, Johannesburg

(With 5 figures and 1 table)

ABSTRACT

A brief review is given of recent work done on the Makapansgat Limeworks Hyaenidae,

machaerodont and feline sabre-tooth cats and the Hystricidae. The smallest Hyaena present

is confirmed as H. h. makapani. H. brevirostris is noted as possibly occurring. Crocuta is

recorded for the first time in this deposit. A form of sabre-tooth, recently wrongly described

as Megantereon problematicus, is reassessed as Homotherium cf. nestianus. Details of several

specimens of Dinofelis are given, including a complete skull of D. barlowi and of the large

part of a semi-articulated skeleton belonging to another individual. Three species of hystricid

rodent are present at Limeworks. The largest, Xenohystrix crassidens Greenwood, has ather-

urine features in its teeth. Hystrix makapanensis and representatives of the modern species

H. africaeaustralis are very similar, except for size and minor differences in tooth morphology.

Comments are made on the palaeoecology of the area based on the described groups and it

is postulated that the immediate vicinity of the cavern supported subtropical riverine bush

and forest up to the end of Lower Phase I times. From comparisons with radiometrically

dated deposits in East Africa, it is suggested that the Phase II Breccia was laid down at the

most 2,75 m.y. ago, which would make the vast majority of the deposit technically Pliocene.

CONTENTS

PAGE

Introductionwintea. a tee hoc) a 1538

Material studied

Eiyaemidacwen, et ee ee Bo yA:

Relidacwe apres 157

Fy StHiCiddees. 2 Pgs ee 1 - a, Sek SS

Generale bona to eee te eee JSD

Gonclusionso= «4-2... 2. = & «& %. “lot

SUMMaAnyAs Alen pele ee See iae Gl

/OTVOMMECIISINSNIIS . so 1 6 oo o o IJG6ll

IREIGRINCSS gg cn US CES

INTRODUCTION

This paper serves to summarize research on three groups of fossil mammals

from the Makapansgat Limeworks which has been completed recently, or

which is about to be completed. The animal groups concerned are the Hyaenidae

(R.M.R.), the machaerodontine and feline sabre-tooths (G.E.C.) and the

Hystricidae (J.M.M.). Collation and some interpretation, as was the original

supervision of the projects, are by A.R.I.C. Full accounts of each aspect of

this report will appear elsewhere.

ss)

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 153-165, 5 figs, 1 table.

154 ANNALS OF THE SOUTH AFRICAN MUSEUM

MATERIAL STUDIED (Table 1)

HYAENIDAE

Hyaena hyaena makapani (Toerien) (Figs 1-2). This is the smallest hyaena

occurring at the Limeworks and on the basis of statistical analyses of the

molars and premolars is confirmed as being a subspecies of H. hyaena, as was

originally suggested by Ewer (1967). A supposed ancestor of H. h. makapani

from the Cape has recently been described by Hendey (1974). The horizon of

his species (H. abronia) clearly antedates the Limeworks Grey Breccia.

The technique used here for comparing the dentitions of hyaenas is an

analogue of that devised by Simpson ef al. (1960: 357). This is a graphical

method using percentages (Fig. 1), in preference to logarithmic scales, as it is

more convenient to use.

An interesting discovery was a lower jaw and a maxilla with the deciduous

dentitions in situ. From these specimens, the order of tooth eruption and replace-

ment could be deduced (Fig. 2).

S242]

The adult formula of H. h. makapani is ee = 36 while the juvenile

32° 123:0

formula is : = = 28 (Figs. 2A-B). These discoveries are important in

confirming Toerien’s (1952) description of isolated deciduous upper cheek

teeth, which were not accepted by Ewer (1967: 114) as being from Hyaena.

The deciduous dentition of hyaenas is so unlike the permanent dentition that

isolated deciduous teeth are more likely than not to be wrongly identified

(Kitching pers. comm.). All age classes are represented, from the very young

to the almost senile.

About 37 per cent of the Hyaena h. makapani material was estimated to be

one year old or less at death, based on tooth wear and eruption features. This

compares with 33 per cent for hyaenas observed by Kitching (1963: 40) at

Pin Hole Cave in Derbyshire, England. The implication from this is that

juvenile hyenas are at risk when their teeth are replacing and tend to die in

their dens from malnutrition and/or infection. It is therefore possible that

parts of the Limeworks cavern may have been used as a hyaena lair during

Grey Breccia times.

Hyaena cf. brevirostris.. All the specimens assigned to this large taxon

are very fragmentary, with heavily worn teeth and this identification is there-

fore uncertain. This size class of scavenger (if scavenger it was), with its charac-

teristic tooth size and structure, is clearly adapted to feed on the remains of

animals possessing thick-walled bones, and hence by implication of moderate

to large size. Speculations on the imbalance of specimens when compared with

the foregoing is not profitable, when so little is known of the conditions under

which the deposit accumulated. However, in the lower levels of the Limeworks,

i.e. Grey Breccia, there are two species of hyaena, one a small generalized type

135

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL

s[enpraA

~Ipur Jo

SIDAISNDADIIAfD

xMISA

S[enprA

-Ipul Jo

Sisuauadoyou

x1ISCH

“CIT, Pue OF, ‘76

SUaPISSDAD

xX1AjsSdyouaX

s]enpiA

purl Jo

DINIOID

x JO DINIOAD

‘ds syafouiq 1MO]40q

syafoulq

snuv1jsau ‘yo

WIN1ddYyJOUWO

“Ayyeorpdeisiyeijs [elloyeul JO UOTINGINsIG “[ IQVyL

SIAJSOAIA IAQ

‘JO Duavdy]

‘BY BS6] UleIg) UCD pasde]]oD, oy) WOIZ,

1uvdpyou

puavdy nuavd yy]

s[eq0 J,

SOUTJIOARIL,

poyeurwieju0Z Jo

POW pow [eseq

J 9seyd JOMOT

eInd01g ASTID

J 9seyd IOMOT

eINNIIg

Proseyidoso19—g)

Jo ‘ul

J eseyg Joddy

II eseud

(€6 “Sy

866] UleIg J0}je)

Aydeis1ye.1S

quosoid exe]

156 ANNALS OF THE SOUTH AFRICAN MUSEUM

P \

. H.brunnea

2 N H.hyaena

P ————

ih H.h.makapani

ny eee ee ees

P ny

M 4 ae

Pig

| 4

Pa \.

‘XN

‘N

‘4

Bs \

p \

80 85 90 95 100 K@}s)

Fig. 1. Ratio diagram comparing mean tooth crown lengths of three hyaenas. Data from

Kurtén, Toerien and this study.

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL 157

closely related to the present-day striped hyaena of North Africa and Asia,

and the other much better adapted to large prey. Therefore they might be seen

to be compatible with the known bovids of that time. Crocuta makes its appear-

ance for the first time in the Phase II Breccia in the form of one isolated left

mandibular ramus belonging to a young individual (Fig. 3).

FELIDAE

Machaerodontinae

Homotherium cf. nestianus (Fig. 4). This skull and lower jaw was originally

described as Megantereon problematicus Collings (1972) on the basis of its

peculiarities as preserved. The specimen itself is laterally flattened and most

of the top of the skull was removed during blasting operations by the lime-

workers. In addition, and for reasons which are not entirely clear, different

degrees of wear had occurred on each tooth row. Finally, the individual must

have been quite old at death because of the total amount of tooth wear seen.

Assignation to the species nestianus is tentative because of the excessive and

unequal wear on the teeth and the lack of information about the missing

portions of the skull. The upper canines are in fact serrated on both their

front and rear edges, with the former being less distinctly serrated than the

latter. The postcanine dentition comprises long, narrow P4, and much reduced

P3. The lower canine is relatively small. The isolated single upper and lower

canines from the Grey Breccia agree with those of the skull.

Felinae

Dinofelis barlowi (Broom). This animal shows some machaerodont ten-

dencies, but is clearly much more closely allied to the true cats than is Homo-

therium. D. barlowi is one of the least specialized of the Dinofelis species, and

supports the idea of the lower levels at Makapansgat being very old in terms

of the South African Neogene. The following material is in the new collections:

(i) A large crushed and distorted skull in two blocks with some scattered

metacarpals and other elements. Basal Red Mud.

(ii) Snout and lower jaw symphyseal region containing canines and incisors.

Grey Breccia.

(iii) A detached naso-frontal region with part of right jugal arch and orbit

plus part of the palate. No teeth. Grey Breccia.

(iv) A nearly complete right mandibular ramus containing all teeth except the

incisors. Grey Breccia.

(v) A complete left ankle joint found in the vicinity of the skull ((i), above).

Basal Red Mud.

Dinofelis sp. The proportions of this partial skeleton are very close to

those of the modern leopard (Panthera pardus), but because positively identified

skeletons of D. barlowi are not known it is realistic to leave this material

158 ANNALS OF THE SOUTH AFRICAN MUSEUM

unassigned as to species for the present. The degree of specialization of this

skeleton seems less than in D. diastemata from Langebaanweg. The following

material is in the new collections:

(a) Accrushed and incomplete mandibular ramus without teeth. Grey Breccia.

(ii) A partial right mandibular ramus with only the roots of the canine and

premolars preserved. Upper Phase I Breccia.

(iii) A large selection of postcranial bones belonging to one individual, in

partial articulation. Basal Red Mud.

(iv) An isolated distal end of femur. Grey Breccia.

Representatives of each of the Makapansgat sabre-tooth cats occur in

the Basal Red Mud. D. barlowi extends into the Grey Breccia and only the one

specimen of Dinofelis is known in the Upper Phase I Breccia.

The association of a large hyena (H. cf. brevirostris) with a true sabre-tooth

cat (Homotherium cf. nestianus) is probably an indication of the size of the prey

species available to them. On these grounds it seems likely that the apparent

disappearance of the large ungulates at the end of lower Phase I times caused

the extinction of these carnivores as well. Quite clearly detailed analyses of

the bovid fauna will tell us a great deal more than is presently known of their

biology and may serve to dispute this subjective view of the happening of that

time.

It is also believed that the more typical sabre-tooth cats had relatively

weak hind limbs and short tails and because of this tended not to be open-

country hunters. Thus they may have lain in wait for their prey in close cover

and did not run it down in a chase. Certainly sabre-tooth cats were unlikely

to have jumped on to the back of their prey, because their long thin canines

would have been extremely vulnerable to breakage against the neural spines

and other bony excrescences of the prey’s skeleton. It is much more likely that

sabre-tooth cats attacked the soft thinner-skinned underparts, or the throat,

of their herbivore prey and this could be better effected by a short rush from

behind cover (Cruickshank 1973).

HYSTRICIDAE

Xenohystrix crassidens is unusual in its extremely large size, and is a rare

component of Plio-Pleistocene deposits. It is, however, the best represented

porcupine at the Limeworks and all specimens are from mature adults. It has

relatively brachydont dentition when compared with other porcupines, and

both upper and lower cheek teeth appear to be rooted, as opposed to the semi-

rooted condition seen in Hystrix africaeaustralis. The root pattern in Hystrix

makapanensis is not seen clearly in any of the cheek teeth available for study,

but the dentition is less brachydont than in Xenohystrix crassidens but also less

hypsodont than in Hystrix africaeaustralis. In X. crassidens the diastema is

relatively much shorter and the curvature of the incisors much more abrupt

than in either of the others present in this deposit, which leads to the conclusion

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL 159

that this species had a shorter snout in comparison. With its more vertical

orientation of the jaw adductor musculature and its brachydont dentition its

diet could have been of softer vegetation than is usual in modern porcupines.

This is supported by the fact that the roots of the cheek teeth close earlier in

life and thus its teeth did not continue erupting for as long as is required in

H. africaeaustralis. The rooted teeth of X. crassidens have their closest parallel

in the atherurine porcupines of the tropical forests in south east Asia and in

west and central Africa.

Hystrix makapanensis (Greenwood, 1958) is intermediate in size between

X. crassidens and H. africaeaustralis but shows a strong tendency towards

the specializations of the latter.

Hystrix africaeaustralis, some specimens of which were originally classified

as Hystrix cf. africaeaustralis by Greenwood (1955), is identical in most respects

to the living species and can thus be regarded as being an ‘early form’ of the

South African porcupine. The presence of three species of porcupine living

sympatrically in the one area is difficult to explain. Xenohystrix was probably

a non-burrowing forest dweller similar to modern atherurines in this respect

and the other two smaller species must have had contrasting ecological pre-

ferences. What they were is subject for speculation, but it can be expected that

the H. africaeaustralis form living them must have had a mode of life very

similar to its present-day descendants, tending to favour open-country savannah

environments. All these species were likely to have been attractive prey. Today

porcupines have few natural enemies barring man, leopard and other carnivores

living under stress.

GENERAL

There is evidence from other sources (e.g. Maier 1970) that the area near

to the Makapansgat cavern supported a body of permanent woodland, and this

would, on analogy with modern times, be a strip of subtropical riverine bush

and forest following the bed of the local river and even extending up into the

mountains behind. This river runs very close to the Limeworks site today and

if it were associated with a body of permanent forest then that forest must

have approached very closely to the cavern entrance. Notwithstanding the

presence of several forest-loving forms in the deposit, it is also clear that open

plains were very close to hand and therefore the ecology of the Makapansgat

region during Phase I times must have been very complex indeed. The present-

day topography in the Makapansgat Valley is broken and varied and probably

similar to the condition existing in the past. This may be one good explanation

for the ecological complexity of the region as reflected in the fossils entombed

in the deposits of the Limeworks.

It is also significant that while the Basal Red Mud has yielded a number

of semi-articulated skeletons, no association of skeletal parts from single

individuals is known from the Grey Breccia. This is understood to mean that

160 ANNALS OF THE SOUTH AFRICAN MUSEUM

during Basal Red Mud times the animals fell in through shafts or avens and

were thus preserved.

The conditions under which accumulation of the Grey Breccia took place

have been the subject of much speculation. It is probable that the bones in

this level were accumulated through the activities of several agents, e.g. leopards,

as suggested by Brain several times for the Swartkrans deposit, or largely by

hominid activities, as the authors prefer. It is clear from a detailed study by

J. M. M. that porcupines played a very minor part in assembling this vast

collection, as the characters of modern Hystrix africaeaustralis accumulations

are vastly different from this one.

The age of the deposit on the above groups

Hyaenidae

Three sites in Africa which have yielded hyaenas have also had radiometric

dates assigned to them. They are Laetolil, Ileret and White Sands in the Omo

Basin. The White Sands site has yielded the earliest dated record of Crocuta

crocuta which is assessed at 2,75 m.y. (Howell et al. 1969: 81), and is the most

important record because it refers to the latest form of hyaena at Makapansgat

and means that the Phase II Breccia could have a maximum age of 2,75 m.y.

b.p. The Grey and Upper Phase I Breccias are older, but by how much cannot

be estimated on this evidence. We should like to stress the importance of the

first appearance of Crocuta cf. crocuta in the Phase II breccia, as it is known

that carnivores have much wider ecological tolerances than do herbivores

It is therefore probable that because of their lesser sensitivity to environmental

changes, the first appearance of a new carnivore species may well be almost

simultaneous throughout a wide area, even in areas as widespread as east and

southern Africa.

Felidae

Deposits at Koobi Fora and East Rudolph have been radiometrically

dated and four felids are associated with these deposits. All are from zones 1

and 2 which have a time span of 1,3—-2,6 m.y. (Maglio 1972). From these horizons

the genera Panthera, Dinofelis and Homotherium are reported in association

with Megantereon eurynodon. Dinofelis and Panthera are long-ranging genera.

The E. Rudolf Homotherium has affinities with both H. crenatidens and H.

nestianus (Maglio 1972; Meave Leakey pers. comm.), while Megantereon

eurynodon is known also from Kromdraai (Faunal Site: Brain 1958) and the

Upper Pliocene of China (Nihowan: Collings 1973). Therefore the levels

at Makapansgat which have Homotherium, i.e. the Basal Red Mud and Grey

Breccia, would seem to fall within the limits set by zones | and 2 at the E.

Rudolf localities, with clearly a trend beyond the older date being probable,

bearing in mind the postulated dates for the Phase II Breccias.

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL 161

Hystricidae

The two species of Hystrix preserved at Makapansgat tell us little of the

age of the deposit. From an examination of a cast of the only available specimen

referred to cf. Xenohystrix crassidens reported by Bishop (1972) from the

Chemeron Formation in East Africa it appears that the material is not assign-

able to this taxon and therefore adds little to our knowledge. The smaller

porcupines from the datable East African deposits have not yet been identified

and therefore cannot help.

CONCLUSIONS

On the basis of the foregoing, it would seem that the evidence would

imply a maximum age of about 2,75 m.y. for the Phase II Breccias at Makapans-

gat Limeworks and if a limit of 1,8 m.y. is accepted as the lower boundary

of the Pleistocene then most, if not all, of the Limeworks Deposit may be

regarded, technically, as belonging to the Pliocene.

SUMMARY

Two hundred and twenty-five specimens of hyaenas ascribable to 3 species,

12 sabre-tooth and ‘false sabre-tooth’ specimens ascribable to 2 species and

71 porcupine specimens belonging to 3 species are briefly reviewed and their

ecology and geological age are discussed.

From these speculations it is thought that the country immediately adjacent

to the Makapansgat Limeworks at the time that the Lower Phase I levels

were being deposited must have supported a much denser vegetation than at

present, approximating to a subtropical riverine bush and forest.

On the basis of the meagre evidence presented, the only possible age

estimate reached is for the Phase II Breccia which is not more than 2,75 m.y.

old and therefore technically Pliocene.

ACKNOWLEDGEMENTS

None of this work could have been done by us without the deep personal

knowledge of the Limeworks site shown by Dr J. W. Kitching, and so freely

imparted. We should also like to acknowledge the help, criticism and assis-

tance of Mr Brian Maguire, Dr C. K. Brain, Dr Q. B. Hendey, Prof. H. B. S.

Cooke and Dr C. E. Gow. Dr J. M. Harris of the National Museums of Kenya

made available information about the Chermeron Formation porcupines. The

Council for Scientific and Industrial Research has provided funds to help the

research programmes summarized here, in the form of personal grants to

G.E.C., J.M.M. and R.M.R. and in the form of running grants to the Bernard

Price Institute which sponsored them.

162 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 2. Stereo photographs of maxillary (right) and mandibular (left) fragments of juvenile

Hi. h. makapani to show milk dentition being replaced by permanent dentition. Specimens

M603 and M2284 of the Bernard Price Institute for Palaeontological Research. Grey Breccia.

(Metric scale).

wok &

Centimetre

Ce cea »

Fig. 3. Partial mandible of Crocuta cf. crocuta. Specimen M2567 of the Bernard Price Institute

for Palaeontological Research. Phase II Breccia.

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL 163

Fig. 4. Skull and mandible of Homotherium cf. nestianus. Specimen M8280 in the Bernard

Price Institute. Basal Red Mud. (Metric scale.)

—~

164 ANNALS OF THE SOUTH AFRICAN MUSEUM

Fig. 5. Maxillary and mandibular fragments of Xenohystrix crassidens Greenwood (upper)

compared with a lower jaw ramus of the extant porcupine, Hystrix africaeaustralis (lower).

Dentitions of both mandibles at approximately the same stage of eruption. Specimens M.1006

and M.1007, in the Bernard Price Institute for Palaeontological Research, Xenohystrix cras-

sidens from the Grey Breccia. Specimen PPM.117, Aystrix africaeaustralis in the Bernard

Price Institute for Palaeontological Research. No locality data. (Metric scale.)

RECENT FAUNAL STUDIES AT MAKAPANSGAT LIMEWORKS, TRANSVAAL 165

REFERENCES

BisHop, W. W. 1972. Stratigraphic succession ‘versus’ calibration in East Africa. In: BisHop,

W. W. & MILLER, J. A. eds. Calibration of hominoid evolution: recent advances in isotope

and other methods applicable to the origin of man. Edinburgh: Scottish Academic Press.

BRAIN, C. K. 1958. The Transvaal Ape-man bearing cave deposits.— Trans. Mus. Mem. 11:

1-131.

CoLLinGs, G. E. 1972. A new species of machaerodont from Makapansgat.—Palaeont. afr.

14: 87-92.

COLLINGS, G. E. 1973. Some new machaerodonts from Makapansgat Limeworks. Unpublished

M.Sc. thesis, University of the Witwatersrand.

CRUICKSHANK, A. R. I. 1973. The mode of life of gorgonopsians.—Palaeont. afr. 15: 65-67.

Ewer, R. F. 1967. The fossil hyaenids of Africa—a re-appraisal. In: BIsHop, W. W. & CLARK,

J. D. eds. Background to Evolution in Africa. Chicago: University Press.

GREENWOOD, M. 1955. Fossil Hystricoidea from the Makapan Valley, Transvaal.— Palaeont.

afr. 3: 77-85.

GREENWOOD, M. 1958. Fossil Hystricoidea from the Makapan Valley, Transvaal. Hystrix

makapanensis nom. nov. for Hystrix major Greenwood.— Ann. Mag. nat. Hist. (13) 1: 36S.

HENDEY, Q. B. 1974. The Late Cenozoic Carnivora of the South-Western Cape Province. —

Ann. S. Afr. Mus. 63: 1-369.

HowELL, F. C., FICHTER, L. S. & Eck, G. 1969. Vertebrate assemblages from the Usno

Formation, White Sands and Brown Sands Localities, Lower Omo Basin, Ethiopia.

— Quarternaria 11: 65-88.

KITCHING, J. W. 1963. Bone, tooth and horn tools of Palaeolithic man. Manchester: University

Press.

Maac1uio, V. J. 1972. Vertebrate faunas and chronology of hominid-bearing sediments East

of Lake Rudolf, Kenya.— Nature, Lond. 239: 379-384.

Mazer, W. 1970. New fossil Cercopithecoidea from the Lower Pleistocene cave deposits of

the Makapansgat Limeworks, South Africa.—Palaeont. afr. 13: 69-108.

SIMPSON, C. G., Roc, A. & LEWONTIN, R. C. 1960. Quantative Zoology. New York: Harcourt

Brace.

TOERIEN, M. J. 1952. The fossil hyaenas of the Makapansgat Valley.—S. Afr. J. Sci. 48:

293-300.

PHYLOGENY OF THE RHINOCEROTIDS OF AFRICA

D. A. HOOWER

Rijksmuseum van Natuurlijke Historie, Leiden

(With 1 table)

CONTENTS

PAGE

exten : 3 A eee LGy/

Selected bibliography . . 168

TEXT

The rhinocerotids as here understood comprise the Epiaceratherium—

Trigonias group and its descendants. There are seven genera in Africa, as

listed in Table 1. Four of them (Brachypotherium, Aceratherium, Dicerorhinus

and Chilotheridium) emerge in the Early Miocene. As they represent full-fledged

TABLE 1. The rhinocerotids of Africa, distribution in space and time

S S

oO S <= a i 5 2

> S D = S = o

= = 3 S Ss S 3 aS

& = S 5 RS 5 S iS

= S 8 aS = aS 5 5

E Sa) < Q Ss) Q S) x

. [V—Up. II — — — — x x —

SORE To 0 ge = = = sf 2 x 5

Chemeron Fm. 42 — — — — — x —

Shungura Fm. 42 -- = — — x x —-

Mursi Fm. 4 — — — — < SK —

Aterir Fm. 4— — — — — — x —

Kanapoi 4 — = = = = Xx —

Langebaanweg — — — — — x —

Lothagam Hill : 6 x _ _ — _ x —

Sahabi 6 x -- — _ — — —

Mpesida Beds Ui x — — — - x —

Ngorora Fm. 12- x <a Oni aX x _ — —

Kirimun — < On xX x — _ —

Douaria 12 - - — — x — —

Alengerr Beds 14-12 — x x — — — —

Fort Ternan 14 — — — — — — x

Loperot 18 — — — 4 — — —

Sinda x< x — — — _ —

Rusinga 18,5 x x x< x — — —

Ombo — — x x - _ ~—

Napak 19 x < x _- _ _ —

Bukwa II 23 x -- u < = — —

Moghara x ? — — — — —

167

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 167-168, 1 table.

168 ANNALS OF THE SOUTH AFRICAN MUSEUM

species different from their Eurasiatic counterparts they must have invaded

Africa already in the Oligocene. Brachypotherium sprang from the Oligocene

aceratheres. The African Brachypotherium snowi (Moghara, Egypt), Brachy-

potherium heinzelini (Bukwa II to Sinda, Zaire), Aceratherium acutirostratum

and Dicerorhinus leakeyi (both Napak to Alengerr) combine characters found

in different Oligocene to Pliocene European species of the same genera, showing

that the phylogeny in Africa was different from that in the rest of the Old

World since the Oligocene. Brachypotherium lewisi (Ngorora to Lothagam

Hill) is the latest representative of its genus in the world. Chilotheridium

pattersoni is the African representative of the Eurasiatic Chilotherium div. spec.,

and differs in a combination of primitive (presence of metacarpal V, more

slender metapodials) as well as progressive (small nasal horn, larger frontal

air sinuses) characters. Diceros appears first in the Late Pliocene of North

Africa (Douaria) and Eurasia (Pikermi, Samos, Maragha), and may be derived

from the stock represented by Paradiceros mukirii (Fort Ternan). Diceros

bicornis of Mursi, Ethiopia, has lower-crowned molars than the modern species

(Shungura to Recent). Ceratotherium praecox (Mpesida to Mursi) is the direct

ancestor of Ceratotherium simum (Shungura to Recent), and split off from the

Diceros stock in the Late Miocene. |

SELECTED BIBLIOGRAPHY

Hoower, D. A. 1963. Miocene Mammalia of Congo.— Ann. Mus. Royal de I’ Afrique Centrale,

Sci. Géol. 46: 1-77.

Hoover, D. A. 1966. Miocene rhinoceroses of East Africa.— Bull. Br. Mus. (Nat. Hist.),

Geol. 13: 117-190.

Hoower, D. A. 1968. A rhinoceros from the Late Miocene of Fort Ternan, Kenya.— Zool.

Med. Leiden 43: 77-92.

Hooer, D. A. 1969. Pleistocene East African rhinoceroses.— Fossil Vertebr. Afr. 1: 71-98.

Hoouer, D. A. 1971. A new rhinoceros from the Late Miocene of Loperot, Turkana District,

Kenya.— Bull. Mus. Comp. Zool. Harvard 142: 339-392.

Hoowwer, D. A. 1972. A Late Pliocene rhinoceros from Langebaanweg, Cape Province. —

Ann. S. Afr. Mus. 59: 151-191.

Hoover, D. A. 1973. Additional Miocene to Pleistocene rhinoceroses of Africa.— Zool.

Med. Leiden 46: 149-178.

Hoover, D. A. & PATTERSON, B. 1972. Rhinoceroses from the Pliocene of Northwestern

Kenya.— Bull. Mus. Comp. Zool. Harvard 144: 1-26.

THE FOSSIL HISTORY OF RAPHICERUS H. SMITH, 1827

(BOVIDAE, MAMMALIA) IN THE CAPE BIOTIC ZONE

RICHARD G. KLEIN

Department of Anthropology, University of Chicago, Chicago, IL 60637

(With 5 figures, 1 map and 2 tables)

ABSTRACT

The present distribution and to some extent the fossil record of R. melanotis suggest it

may have evolved in the Cape Biotic Zone, perhaps in response to the substantially different

environmental conditions which characterized much of the later (and ?mid) Pleistocene and

which supported large populations of now extinct endemic or near endemic grazing species.

R. campestris appeared in the Zone no later than the earlier part of the Upper Pleistocene,

but clear evidence for its essentially modern distribution and abundance is so far restricted to

very recent Holocene contexts. Perhaps it was similarly distributed and abundant in climati-

cally comparable portions of earlier interglacials. There is some evidence from dental measure-

ments to suggest that Raphicerus spp. may have been significantly larger during especially

cold intervals in the Zone, but this requires further investigation. There is also some evidence

for selective predation on Raphicerus individuals of different ages at various sites, but this

too requires further research for secure demonstration. Limited evidence suggests that at

least Later Stone Age peoples did not prey selectively on animals of either sex. The chances

of obtaining additional information both from and about Raphicerus in the Cape Biotic Zone

are excellent since it occurs in high frequency in virtually all local palaeontological

and archaeological sites.

CONTENTS

PAGE

NCROGUCHIONS ie ee ee bee eee OG ake hes bre fe 169

ithe modern species of Raphicerus’ - - >: =. =. = 7

OSSilvRaphicerus™ “are. Ps Sole Ss ee As

Changespmsize throuch) times 4) nie) ee Sl

Population¢sthucturem san ioe en eee ee 182

Acknowledgements. 5 ase 4) oe en ee SG

IRGIEKENCES em ey en ee eke Oe ba ee ee SG

IND pPenGixeliig Solments Ciera is ea tl owes J TOS

AP DENdIxe2. Hala act ene Oka eet See 5 See 19

INTRODUCTION

The far south-western corner of the African continent is distinguished

from the rest of sub-Saharan Africa by its warm temperate climate, unique

‘fynbos’ (“Cape Macchia’) vegetation, and endemic fauna. Its biological dis-

tinctiveness has been recognized by several authors, for example, by Meester

(1965) who suggested the name ‘South-West Cape Biotic Zone’, shortened by

Hendey (1974: 16) to ‘Cape Biotic Zone’. As defined by Hendey and used here,

the ‘Cape Biotic Zone’ includes three subdivisions (Map 1): (1) the south-

western Cape sensu stricto; (2) the southern Cape; and (3) the Cape Folded

Mountains.

169

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 169-191, 5 figs, 1 map, 2 tables.

170 ANNALS OF THE SOUTH AFRICAN MUSEUM

9 ser

EASTERN CAPE

Fave “%e | \ CAPE FOLDED

Langebaanweg pean MOUNTAINS

Sea Harvest_\\ > il

‘° ey “>

= lendetomeint |

—_- —_

= — —

—_— >= —

e y

Fay

\5 7

(eo)

2 i

m bee 2

(@)

m Ve

Q0,4

Fg |

joo)

fo}

17 2| 25 29

Map 1. The schematic extent of the major subdivisions of the Cape Biotic Zone (after Hendey

1974:7) and the approximate locations of the principal sites providing Raphicerus samples

relevant to this paper.

Historically, the south-western Cape has been differentiated from the southern

Cape by more seasonally restricted (winter) rainfall and by a relatively

impoverished mammalian fauna (see Klein 1974a, table 1). Evidence from sites

such as Sea Harvest (Hendey 1974, table 6), Melkbos (Hendey 1968; Klein,

unpublished), and Swartklip (Hendey & Hendey 1968; Klein 1975a) in the south-

western Cape and Die Kelders I (Klein 19755), Nelson Bay Cave (Klein

1972a), and Klasies River Mouth (Klein 19745) in the southern Cape suggests

the two areas were more alike faunistically in the late Pleistocene, perhaps in

part because of greater climatic similarity and in part because faunal inter-

change was facilitated by exposure of the continental shelf during periods of

lowered sea-level. The late Pleistocene fauna of the two areas is further notable

for a number of extinct taxa (Klein 1974a), including at least three forms

which at that time were almost certainly Cape Biotic Zone endemics: a small

springbok (Antidorcas australis), a wildebeest at least subspecifically distinct

from the extant Connochaetes gnou, and a reedbuck at least subspecifically

distinct from the extant Redunca arundinum. Historically recorded endemics

include (after Meester 1965: 91) four rodents (Bathyergus suillus, Praomys

verreauxi, Acomys subspinosus, and Tatera afra) and two large mammals—the

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 171

blue antelope (Hippotragus leucophaeus) and the bontebok (Damaliscus dorcas

dorcas). In the late Pleistocene, the blue antelope occurred somewhat outside

the area of the Cape Biotic Zone as defined here (Klein 1974c), suggesting the

possibility that the boundaries of the Zone have varied over time, presumably

in response to climatic change.

In addition to the endemic mammals listed above, the grysbok (Raphicerus

melanotis) is largely restricted to the Cape Biotic Zone, where it overlaps with

its much more widespread relative, the steenbok (Raphicerus campestris). The

purpose of this paper is to summarize the fossil history of these two species in

the Cape Biotic Zone, with particular emphasis on the later Pleistocene and

Holocene.

THE MODERN SPECIES OF RAPHICERUS

In addition to R. melanotis and R. campestris, most authorities recognize

yet a third extant species of Raphicerus, R. sharpei, known popularly as Sharpe’s

erysbok. In contexts where R. sharpei and R. melanotis are both discussed, the

latter is known as the Cape grysbok. Some authors, for example, Haltenorth

(1963: 78) and Smithers (1971: 221) consider the two grysboks to be distinct

only at the subspecies level (R. melanotis melanotis and R. m. sharpei), but the

more conventional specific distinction has been retained here.

Individuals of all three species of Raphicerus are small (7-14 kg), incon-

spicuous, non-gregarious antelopes (Dorst & Dandelot 1970: 264, 266; Bigalke

1974: 833; Smithers 1971: 220, 222; Astley Maberley 1967: 129-138). Unlike

many larger, gregarious species, they are still found in good numbers over

large parts of their historic ranges. As already indicated, R. melanotis is largely

confined to the Cape Biotic Zone, although it is also distributed in adjacent

areas of the south-eastern Cape (Bigalke & Bateman 1962, map 4). R. campestris

is far more widespread, occurring with minor discontinuities over nearly all of

southern Africa and also in parts of Kenya and Tanzania in East Africa (Ansell

1971: 66; see also Dorst & Dandelot 1970: 269). In the Cape Biotic Zone, it is

most numerous in the south-western Cape where it probably outnumbers

R. melanotis. In the southern Cape, it is generally less numerous than R. mela-

notis, and in fact is probably absent altogether from some areas with especially

heavy cover (for example, portions of the Knysna Forest) where R. melanotis

is fairly common (pers. obs.). In southern Africa, R. sharpei occurs no further

south than Swaziland, extending from there into southern Tanzania (Ansell

1971: 67; see also Dorst & Dandelot 1970: 269). It therefore nowhere overlaps

R. melanotis and is sympatric with R. campestris only in the north-eastern

portions of the latter’s range.

Survival for all three species of Raphicerus requires the presence of adequate

cover, but the steenbok exhibits a clear tendency to inhabit more open country

than the grysboks (Bigalke 1972: 176). All three may be broadly characterized

as selective browsers (Jarman 1974: 223), but perhaps in keeping with its

preference for more open environments, the steenbok apparently incorporates

172 ANNALS OF THE SOUTH AFRICAN MUSEUM

a higher proportion of grass into its diet (Smithers 1971: 221, 222: Hoffmann

& Stewart 1972: 228).

The three species are fairly easy to tell apart on the basis of pelage and

other external characters (Dorst & Dandelot 1970: 264, 266). There are no

published criteria for distinguishing them on hard parts, but A. W. Gentry

(pers. comm.) has suggested several cranial characters which together allow

R. melanotis 2

(S.A.M. 36002)

R. melanotis 2

(S.A.M. 35109)

R. melanotis 6

(S.A.M. 36204)

R. melanotis &

(S.A.M. 36205)

FR. campestris 9

R. campestris 6° S

(S.A.M. 35120)

(S.A.M. MI68)

R. campestris Oo"

(SAM 15285)

R. campestris %

(S.A.M. 13915)

eee eee es

Fig. 1. Outlines of left mandibles of Raphicerus melanotis and R. campestris, illustrating the

differences between the two species in the configuration of the inferior margin of the hori-

zontal ramus. An arrow indicates the approximate position of the canine in each case (after

originals by K. Scott).

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 173

complete separation of R. campestris from R. melanotis. These include a more

upright insertion of the horncores in R. campestris as well as significantly

broader ethmoidal fissures, smaller preorbital fossae, and larger auditory bullae

than in R. melanotis. Although it may be impossible to distinguish them on

dental morphology, K. Scott (pers. comm.) has observed that the configuration

of the inferior margin of the mandible differs significantly between the two

species (Fig. 1): in R. campestris it tends to be fairly straight, while in R. mela-

notis it is markedly convex. This difference is well marked in animals of all

ages, but seems especially clear in juveniles. R. sharpei probably resembles

R. melanotis in mandibular shape, but the author has seen too few R. sharpei

specimens to be sure. This mandibular difference is particularly useful in

dealing with fossil samples because the relevant portion of the mandible is

generally preserved in high frequency versus other cranial elements on which

distinctions may be made. Mandible configuration was the principal and in

most cases the only feature which could be used to calculate the relative fre-

quencies of grysbok and steenbok in the fossil samples discussed below.

In an attempt to establish further cranial characters which might allow

separation of R. campestris and R. melanotis, measurements were made on

dental rows and on the diastemata of a series of comparative specimens of each

species housed in the South African Museum. The means, standard deviations,

and numbers of specimens involved are recorded in Appendix 1 (see also

Figs 2-3). The results of ‘Student’s t’ tests for the significance of differences

between means, where such differences were significant at the 0,05 level or

below, are recorded in Appendix 2.

All the specimens in the Museum’s collection were either of unknown

provenience or were derived from the Cape Biotic Zone so that it was not

possible 10 test for geographic variability in either species. Student’s ‘t’ tests

run to compare the sexes within each species revealed no significant differences

in mean dental dimensions, perhaps because the samples involved were too

small. It is thought that the samples were large enough, however, to conclude

that sex-related dental size differences, if any, are probably small in magnitude.

This is the justification for ignoring sex in comparing the R. melanotis and

R. campestris comparative samples to the fossil ones (below) in which sex

cannot be established in any case.

The results suggest that the mean length of the diastema (measured from

the posterior margin of the canine alveolus to the anterior margin of the

alveolus for P,) may be used as a nearly discrete feature to separate R. campestris

from R. melanotis (at least as adults). The relevant Dice-Leraas diagram in

Figure 2 (see also Appendices 1-2) shows that the mean diastema length in R.

campestris is significantly greater than the mean length in R. melanotis, while

in the comparative samples and in fossil samples from the site of Die Kelders I

(where R. campestris mandibles were sorted from those of R. melanotis on

the degree of curvature of the inferior margin of the horizontal ramus), there

is little or no overlap in observed ranges. The observed ranges do overlap

174 ANNALS OF THE SOUTH AFRICAN MUSEUM

M,—Mz

OK! EBC EBC

Comparative LSA IO 11-15

Rm. Re. Re. Rc. R.c.

lo—"4

(8) aa

(1)

(5) (0)

DK! EBC EBC

Comparative LSA 1-10 II-I5

Rm Re Re. Re Re.

(7) okt EBC EBC

Comparative LSA I-10 II-I5

Rm. Re Re Re Re.

wl—m3

(12)

Comparative

Rm.

(8)

Comparative

Rim.

(10)

Comparative

R.m.

Fig. 2.

(18)

R.c.

(10)

Re.

(9)

R.c.

Olastema

(19)

(2)

(6)

(8)

OK! OKI

Comparative LSA LSA EBC EBC

Rm. Re Rm Re. Rm Re.

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 175

substantially in a second set of fossil samples from Elands Bay Cave (Fig. Dy

but the Elands Bay samples involved are small and contrast with the compara-

tive and Die Kelders samples to such an extent that it seems possible a recording

or measuring error is responsible.

The premolar and molar row lengths of Raphicerus specimens were

measured always at the level of the occlusal surface. A complicating factor

here is that the mean measurements may change somewhat as the teeth wear,

and the samples involved were too small to test for this directly. However, the

author believes that changes with wear are probably not substantial, since as

one tooth within a row shortens with wear, another tends to lengthen corre-

spondingly, preventing row lengths from changing significantly during the life

of an individual. Additionally, the samples on which measurements were

made were roughly comparable in their age frequency composition. As a

consequence, the author thinks that any statistically significant differences

observed in mean dental row lengths between comparative R. campestris and

R. melanotis or between either and the fossil samples discussed below are

probably open to non-trivial interpretation.

The measurements indicate that the mean length of the molar row is

probably greater in R. campestris, while the mean length of the premolar row

is probably greater in R. melanotis, though for both dental rows the species’

ranges overlap substantially, and with the present data, a statistically significant

mean difference is demonstrable only for dP,-dP, and P,—P, (longer in R.

melanotis) (see Fig. 2 and Appendices 1-2). The mean premolar row/molar

row length ratios are significantly smaller in R. campestris than in R. melanotis

(Appendices 1-2), in keeping with the hypothesis that the steenbok has a shorter

mean premolar row and a longer mean molar row. This finding further makes

sense in view of the somewhat greater propensity of the steenbok to incorporate

Fig. 2. Dice-Leraas diagrams (modified following procedures discussed by Simpson ef al.

1960: 355) for the comparison of mean dental row and diastema lengths between Raphicerus

melanotis and R. campestris and among R. campestris samples from various Cape Biotic

Zone sites. (R. m. = R. melanotis; R.c. = R. campestris; DK 1 = Die Kelders 1; EBC =

Elands Bay Cave). Since the two species cannot be securely separated on maxillary characters,

only mandibular dental rows are considered in comparisons involving fossil samples. The

various samples are arranged so that older ones are generally to the right (see Table 1 for

the geologic ages of the samples). Vertical lines show observed ranges; open rectangles mark

standard deviations, with solid black indicating the 95 per cent confidence intervals for the

mean. The means are indicated by horizontal lines, and the number of specimens in each

sample is shown in parentheses.

176

ANNALS OF THE SOUTH AFRICAN MUSEUM

M, -M3

(37)

(77)

24 DK!| EBC NBC NBC EBC NBC KRM |

Rm. LSA I-lO W A U5 R ZWI msa_ EFT

2 pF 1)

23 Po— Pq P —P

22 na’

(1)

2! (3) =

(2)

20 ()

SS (6)

19 (3) (8)

(8)

18 ——_

(1)

(3) (12) (2)

17 (0) (0)

DK! EBC NBC NBC EBC NBC

Rm. LSA I-10 W A z

(2)

(10) (13)

(49)

(0) (QO)

DK! EBC NBC NBC EBC NBC

R.m. LSA I-lIO W A II-I5 R

NBC NBC NBC

Rm w "A R

KRM 1

ZWIl wsa EFT

Fig. 3.

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 177

grass in its diet. It is an easily demonstrable generalization for bovids that

grazing species usually have longer molar rows relative to premolar rows than

browsing species.

FOSSIL RAPHICERUS

Although as yet undescribed specimens from the Pliocene locality of

Langebaanweg in the south-western Cape suggest that Raphicerus may have

been in existence as long as 4 to 5 million years ago (Hendey 1973), the genus

has a remarkably poor fossil record (Gentry, in press). Among other things,

it is not possible to say when the various species of Raphicerus diverged. Material

from the Krugersdorp australopithecine site of Swartkrans includes mandibles

which are very similar in horizontal ramus configuration to those of R. cam-

pestris (pers. obs. on material kindly shown to the author by E. Vrba), while

the (?somewhat later) mid-Pleistocene locality of Elandsfontein in the south-

western Cape has provided remains of a very large species of Raphicerus,

which in the relatively great inclination of the horncores and the convex shape

of the inferior margin of the mandible (Fig. 4) is more reminiscent of R. mela-

notis. This may indicate that the lines leading to R. campestris and R. melanotis

had already diverged by mid-Pleistocene times, with the R. melanotis lineage

perhaps evolving in the Cape Biotic Zone. Whatever the case, the great size

of the Elandsfontein species relative not only to modern R. melanotis and

R. campestris, but also to samples from such earlier Upper Pleistocene sites

as Klasies River Mouth Cave I and Swartklip I (Fig. 3), in combination with

the tendency of its horncores to display an irregular (vs. relatively rounded)

cross-section, is another indication of the fact that the basic Elandsfontein

Fig. 3. Dice-Leraas diagrams (modified following procedures discussed by Simpson ef al.

1960: 355) for the comparison of mean dental row and diastema lengths among samples of

Raphicerus melanotis from various Cape Biotic Zone sites. (R. m. = comparative R. melanotis;

DK 1 = Die Kelders 1; EBC = Elands Bay Cave; NBC W, A, R = Nelson Bay Cave Wilton,

Albany, and Robberg; ZW 1 = Swartklip 1; KRM = Klasies River Mouth Cave 1; EFT =

Elandsfontein. The Elandsfontein sample is actually derived from an as yet undescribed

species of Raphicerus, but is included here since the species may be ancestral to R. melanotis).

The sites are arranged so that older ones are generally to the right (see Table 1 for the geological

ages of the samples). Since criteria have not yet been developed to separate R. melanotis and

R. campestris on maxillary characters, for sites where both are clearly present (see Table 1),

only mandibular parameters are presented in the diagrams. Vertical lines show observed

ranges; open rectangles mark standard deviations, with solid black indicating the 95 per cent

confidence intervals for the mean. The means are indicated by horizontal lines, and the number

of specimens in each sample is shown in parentheses.

178 ANNALS OF THE SOUTH AFRICAN MUSEUM

(SAM ZW .3426)

(SAM EFT ISOl)

Fig. 4. Right mandibles of Raphicerus melanotis from Swartklip 1 (top)

and Raphicerus sp. from Elandsfontein (bottom) (drawings by

K. Scott).

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 179

fauna, and perhaps also the “Saldanha Skull’, antedates the Upper Pleistocene

by a significant time interval (Klein 1973).

Fossil remains, especially mandibles, indistinguishable from those of

modern R. melanotis are well known in the Cape Biotic Zone from a series of

earlier Upper Pleistocene sites in both the south-western and southern Cape

(Table 1, Map 1). However, only one of the sites—the Sea Harvest locality at

Saldanha Bay—has provided remains attributable to R. campestris and then

in circumstance that suggest the species may have been less numerous near

Saldanha than R. melanotis. At present, R. campestris is numerically dominant

in the area. The absence of steenbok at Klasies I was perhaps to be expected

since the-site is in an area today (near the eastern margin of the Knysna Forest)

where steenbok are very rare, and both the Klasies sediments and the fauna

taken as a whole suggest an environment similar to the modern one at the

time the fauna accumulated (the Last Interglacial). The absence of steenbok

at the remaining sites (Die Kelders, Swartklip and Duinefontein) is more

surprising since the species occurs in fair numbers near each today. It is unlikely

that sampling error is involved—if, for example, the ratio of grysbok to steenbok

in the earlier Upper Pleistocene near Die Kelders were assumed to be the same

as in the late Holocene sample from the site, that is, roughly 3 : 1, the binomial

probability of encountering 26 identifiable grysbok and no steenbok would

be considerably less than 0,001. Even if the ratio were 10 : | in favour of grysbok,

the probability of finding 26 grysbok and no steenbok in the earlier Upper

Pleistocene sample would be only 0,065. The full faunal lists of Die Kelders,

Swartklip, Duinefontein and Sea Harvest contain taxa (especially Antidorcas

and Connochaetes in fair numbers) which suggest a substantially different

vegetation cover than the modern one. More particularly, a much larger-

than-modern grass component is implied, and while there is nothing a priori

in a grassier environment to exclude R. campestris (quite the reverse), it seems

possible that the particular conditions which prevailed placed it at a distinct

competitive disadvantage relative to R. melanotis. Such peculiar (vs. any modern)

environmental conditions may in fact have occupied a far larger proportion

of Quaternary time than have modern ones and may even be the ones in which

R. melanotis evolved, perhaps as part of a faunal complex involving the now

extinct, larger Cape Biotic Zone endemics mentioned earlier.

Mid-Upper Pleistocene faunas in the Cape Biotic Zone are essentially

unknown at present, but on-going excavations directed by H. J. Deacon at

Boomplaas Cave near Oudtshoorn promise to fill the gap in the near future.

Terminal Pleistocene faunal samples containing specimens of Raphicerus are

known from Elands Bay Cave in the south-western Cape (technically, just

beyond its northern boundary as defined here) and Nelson Bay Cave in the

southern Cape. The Nelson Bay samples do not allow estimates of the relative

frequencies of R. melanotis and R. campestris, but the Elands Bay samples

indicate that R. melanotis and R. campestris may have been about equally

common near that site (Table 1). In recent times. R. campestris has been far

ANNALS OF THE SOUTH AFRICAN MUSEUM

180

g[QeyNUSp! JOU OOM YOIYM syUSWseIZ o[qIpueU puUe SUOTUSp

Jo Sioquinu oY], ‘S}USWIseIy sIqIPUeL 9IQeyUEpPI Woy poyepNoye

Areyrxew snyjd ‘syuowsesy O[QIPULU OUIeS OY} SUISN payepNoyeo o1oM “dds snsao1ydvy

9 910M S[ENPIAIPU! Sz4Zsadups “y pue s1jounjaw snaaaiydoy JO SIaquanuU sy, q

‘9U900)SI29[q Jaddq oy} estiduros 19y}030} [eIDeIH \sey] pure [elov|s19jUy yseT SOUL, “d'A 000 OF Aloyeunxosdde ye soyeurui9) YOryA

[EIORTD Ise OY) Aq AloYEIPOUILUT POMOT[OJ SI IT “dA 000 SZ 0} 000 SZI ATYsNOI WoIZ [eAJOJUT oY} Suds [eIOR[SIOJUT \SVT OY} “oIOY Posn SV e

(poysijqndun “ure, y

pue dooysuy) [eo1so[oovyo1Vy

(B10qqoy Aeq UWOSTON,

IOJS B SOOUDIOJOI) [LIISO[OSRYIIV

(s10qqoy Avg UOSTON,

IOJ SV SOOUDIOJOI) [BDISOTOOBYIIY

“(EL6I SunoX 2 [eyez

“€L61 1OZING *9PL6l “IZLOI

‘OTL6I Ule[ A) [eoIsofoseyo1y

(SI-IT Avg spurl

IOJ SB SOOUDIOJOI) [BIISOTOIeYIIV

(poysiqndun

Ulop yy ‘UoNeredoid Ul pue 7/6]

UOWUIYIV) [LOISOTOSeYI1IV

(VL6I ‘OL6I

IOZHOMYIS) [VdISO[OseyIIV

(4SL6I “ULI

‘ssoid Ul pUe pL6[ JOZIIOMYOS

® pleyueL) [eosolooeyo1y

(poysijqndun

Jozing ‘poysijqndun

pue GyLol UTA +ZL6l Jesuls

ww IOWA) [PdISO[OSRYOIV

(poyslqndun 19zjng

CCL6T UIST - 8961 AepusH

2 AapuoH) (UoTe;nuINSOe

SIOAIUIED) [OISO[OJUOSP[eg

(poysiqndun

JOzINg pue UlO,y) (UONe;NUINdS.e

JIOAIUIVD {,) [edISO[OJUOSR[ ed

(poysyqndun sezing

‘poysyjqndun ulspy : uissed

[pL6[ AopusH) (uoljefnumnsoe

SIOATUICD {,) [BOISO[OJUONL[ eg

(SHONTUaIAA GNV) ALIS JO AdAL

0OT

9ST

Scl

6¢

AUBUL

‘dds snsaoiydoy

9 ‘q:SV AIAVIAILNAGI SIVNGIAIGNI 40 SYdH#NON WAWININ

ouoU JO Mof

SIAJSOQWUDI “YT

Aueul

SIJOUDIAUL “yy

(O-v1) ‘d’& 00S I-000 € *2

(D-r1) ‘dA 000 S-000 8 *2

(O-v1) ‘d’& 000 8-000 ZI 2

(O-V1) ‘d’& 000 vI-000 81 *9

(O-v1) ‘d’@ 0-005 € 2

(O-r1) ‘d’A 009 6-000 TI *2

(0-41) ‘d’& 008 1-000 Z °2

(QdUdIOJUT [VIISO[OOS)

[eIoe[H 4sey] Jolie

(QOUSIOJUT [VIISOTOSS)

[eloe[s19}UT SVT

[SdUdIOFUI [LOISOT[ONS

29 (08 9-1) 000 OF < O-FI]

({.) [elov[s19}UyT IseT 91eT

(QOUDdIOJUT [VdIBO[OSS)

[elov[s19}UT ISVT 91eT

[ooudIOJUT [VdISO[OSS

® (Z8Z-MN) 000 OF < O-F1]

({) [elov[s19}Uy sey 9187

(AONAGIAD JO AUNLVN

CNV) efDV TVOIDOTOUD

Aeg UOSs[on,

S[OAO'T

UOTTAA

<xeul[) pue

xe)

-WIq, SAD

Aeg uOs[oN,

S[OADT

Aueqiy

OAR)

Aeg UOsfon,

S[OAI] 3.10q

-qOxyY 9AeD

Aeg UOS[oN,

OI-T

S[OAST SABD

Aeqspur|q

SI-IT

S[OAIT CARD

Aeqspur|q

cVSTI,

I SI9P[9y{ VI

VSI,

I SIOP[Pw IC

<VSW,

] 9AeD YINO|W

JOATY SOISePS]

I dippzems

(Soqy[e)

U9} UOJOUING

JSOAIVF POS

ALIS

({ dep] 208 ‘sous 9Y} JO SUOTIO] 94} IOJ) BUOZ IHOIg ode_ oY} UI SoqIs SUSDOTOH] pu ousd0jslo[q Joddy ur “dds smuaoiydvy Jo soouonbely *‘T FIAVL

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 181

more common, a fact which is apparently reflected in the late Holocene sample

from the same locality (Table 1). As at Sea Harvest, Swartklip, and Duine-

fontein, the species accompanying Raphicerus in the very late Pleistocene

levels of Elands Bay suggest a vegetation cover with a greater-than-present

grass component. Once again, however, the particular conditions that prevailed

were ones in which R. melanotis was able to compete more effectively with

R. campestris than it can in the area today.

The various Holocene samples—from Elands Bay, Die Kelders and Nelson

Bay—suggest that at least by the late Holocene the essentially modern distribu-

tions and relative frequencies of Raphicerus spp. had become established in the

Cape Biotic Zone. More particularly, the late Holocene fossil samples imply

that as at present, R. campestris became less common and R. melanotis more

common along the coast from north-west to south-east. Supplementary faunal

data from the nearly complete Holocene sequence at Nelson Bay indicate that

the Holocene environment there was variable, but consistently high frequencies

of such creatures as bushbuck (Tragelaphus scriptus) and bushpig (Potamo-

choerus porcus) in the Nelson Bay Holocene deposits indicate the more or less

continuous occurrence of the relatively closed vegetational surroundings

known to be more favourable at present to the grysbok than to the steenbok.

CHANGES IN SIZE THROUGH TIME

Figures 2 and 3 and Appendices | and 2 show that the various Raphicerus

samples sometimes differ significantly in the mean lengths of molar and premolar

rows (significance in this context means that the probability of a chance differ-

ence between two means is less than 0,05 as assessed by ‘Student’s t’). Par-

ticularly interesting are the statistically significant, larger mean dental row

lengths observed in the Nelson Bay Robberg and Elands Bay (levels) 11-15 late

Pleistocene samples. (The Nelson Bay Albany sample does not exhibit the

same tendency towards larger specimen size, but the Albany Industry in fact

extends into the early Holocene, and most of the measurable Nelson Bay

Albany specimens are between c. 9 000 and 8 000 radiocarbon years old. The

Elands Bay 11-15 specimens are all older than 9 600 radiocarbon years, while

the Nelson Bay Robberg ones are all older than 14 000 years.) The Elands Bay

11-15 sample and especially the Nelson Bay Robberg one accumulated under

very much cooler-than-present conditions (see Van Zinderen Bakker & Butzer

1973 for a general statement on late Pleistocene climates in southern Africa;

Butzer 1973 for a specific statement on Nelson Bay). It is therefore tempting

to hypothesize the operation of Bergmann’s Rule in Raphicerus. (This ‘Rule’

states that, all other things equal, individuals belonging to a single mammalian

species will tend to be larger in colder climates. This is because as an animal

grows, its volume, which produces heat, increases more rapidly than its skin

area, which dissipates it. Bergmann’s Rule apparently operates on many extant

mammal species with broad latitudinal distributions, and has been convin-

182 ANNALS OF THE SOUTH AFRICAN MUSEUM

cingly used to explain, for example, the relatively larger mean size of several

species of small mammals in Last Glacial deposits in Pennsylvania. See Guilday

1971: 251-252.) One important difficulty in using Bergmann’s Rule to explain

the relatively long dental rows in the terminal Pleistocene Raphicerus samples

from Nelson Bay and Elands Bay is that there are smaller, but still significant

mean differences among other Cape Biotic Zone samples (for example, between

the average molar row lengths in the Die Kelders I LSA R, melanotis and

comparative R. melanotis samples) which it would be difficult to attribute to

temperature on present evidence. Additionally, it remains to be shown that

relatively small differences in dental row length, even though statistically

significant, would be reflected in overall body size. In particular, it is possible

that a slightly (but significantly) increased mean dental row length might occur

in response to a change in vegetation (and diet), without any corresponding

change in mean body size. It should be possible to gain some contol over this

problem by measurements planned to be made soon on various post-cranial

elements (which should reflect mean body size directly). If the material is

available in museums, it would also be relevant to obtain measurements on

Raphicerus specimens from present-day environments characterized by different

mean annual temperatures. |

POPULATION STRUCTURE

Figure 5 presents the minimum numbers of Raphicerus individuals in

each of seven successive dental eruption/wear categories for each of the sites

listed in Table 1 (excepting Duinefontein). The definitions of the eruption/wear

states are given in the caption of the figure. Unfortunately, there are virtually

no published data on the schedule (in months or years) of dental eruption and

wear in Raphicerus spp. so that it is impossible to place chronological limits

on the dental states in the figure. Since some of the dental states almost cer-

tainly have much broader chronological limits than others, it is therefore impos-

sible to compute life tables or survivorship curves from the data in the figure

(compare Kurtén 1953: 42ff.; Voorhies 1969: 24ff.; or Reher 1974: 117-122).

The figure does show, however, that dentitions assignable to Category VII,

in which the crowns of the teeth are worn down to near the level of the alveolus,

are rare at all sites. This suggests that individuals of Raphicerus do not ordi-

narily survive into very old age, however that may be defined chronologically.

Data presented by Mentis (1972: 74-75) indicate that it is in dental state III

(M2 erupting to erupted, but essentially unworn) that females of Raphicerus

spp. become capable of bearing their first young. As a first step towards deter-

mining if there is meaningful patterning in the data presented in Figure 4,

this fact has been used to regroup the data for each site into two broad sub-

divisions: (1) Categories I and II (with M2 unerupted and presumably including

no sexually mature individuals) and (2) Categories IIJ-VII (with M2 erupting

to erupted and presumably including mostly sexually mature individuals).

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 183

SEA HARVEST

ELANOSBAY /-/0

2 NELSON BAY ROBBGERG

| |

==) O Re Sa O

6 ELANOSBAY 11-15

Sit

DIE KELDERS /LSA

FR. campestris

FR. melanotis

NELSON BAY

WILTON

eee We NAN i 600 i IY Ye YT a OO I AY AIL

Fig. 5. Histograms showing the frequencies of Raphicerus individuals in different dental states

in samples from the sites listed in Table 1 (excluding Duinefontein). The dental states are

defined to include individuals in which (1) dP4 was erupting to erupted, but essentially unworn;

(II) M1 was erupting to erupted, but essentially unworn; (III) M2 was erupting to erupted,

but essentially unworn; ([V) M3 was erupting to erupted, but essentially unworn; (V) P4 was

erupting to erupted, but essentially unworn; (VI) P4 was in early to mid wear; and (VII) P4

was in late wear.

The regrouped data are presented in Table 2. As the caption of Table 2 shows,

application of the chi-square statistic to this data reveals several differences

significant at the 0,05 level or below. The differences are very difficult to interpret,

however, For example, among the archaeological sites, the two earlier Upper

Pleistocene ones (with “Middle Stone Age’ materials)—Klasies I (MSA) and

Die Kelders I (MSA)—display roughly the lowest and highest percentages of

combined categories I and H, while the remaining sites—Elands Bay, Nelson

Bay, and Die Kelders I (LSA) (all with ‘Later Stone Age’ materials) exhibit a

wide range of intermediate values.

ANNALS OF THE SOUTH AFRICAN MUSEUM

184

06'0-s6'0 = d ‘oon = LU

ia ee 060-s6'0 = 4 ‘00'0 = 94

0S‘0-SL'0 = 4 “170 = fe s7%9 = 4 ‘O'T = oH

100’ > d ‘zp'er = “FA

0s‘0-SL‘0 = 4 ‘970 = a olo-szv0 =d eT = a

sco'o-so'o = doe'p = BS sto-oso = IZ = SH

100° > d‘9c%6 = BA szo'0-so'0 = d ‘tee = VN

Too. > d‘Iz'g= ON soo-ol'o = 4 ‘zee = Go

:sonjea orenbs-1yD po}dejes

(%9L) 91 (%0S) § (%L9) 07 (Pail) Gl TIA /TI

ms | AY Z A

(% VT) § (%0S) § (% EE) OL (% LT) 97 i/I

x NM A a

S14JSadwuvo ‘M S1JOUD]AUU “YM S1AJSAdUIDI ‘Y SIJOUDaU “NM SALV.LS

AVd SANV Td I SYACTAM AIG TIVLNAG

Icl OOT [Lie

(%L8) 9b | (ZBL) L (% 06) 97 (%S8) TI (%S9) 17 (%SL) 16 (%9S) 9S (%69) TIE | (E9) ET | (%98) 7 TA /

L S U O d O N W TI > |

(“ED L CAGONG (%OD) € (Aol) C (ASS) Gi ZSows | CAidian | GAs) @s \) CAS) ial (“vl v U/I

fe I H D A aq d S| Vv

VSI LSAAUVH I dI14 B1NQqOry Aueqiy UOTYIM VSIN C= OI-I SALV.LS

I SHISV 14 Vas =~LUVMS aAVO AVA NOSTAN I SYadTay aId AVd SANV Td TIVLNAG

(‘¢ INS UI e}ep UO poseg) ‘sojduies UOZ SNoIg odeD UI [[A-]]] PUe II-] SoIeIS [e}UOP UI S[eNpIAIpUI sn4aoz1ydDyY JO SIOQUINN “7 IIQeL

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 185

It seems unlikely that the differences among sites in the proportions of

combined categories I and II derive from differences among them in the relative

frequencies of R. melanotis and R. campestris since differences in the proportion

of I/II within R. melanotis and R. campestris at the two principal sites that

contain both—Die Kelders I (LSA) and Elands Bay—are far from significant

(Table 2). It is probable that no single factor is responsible, but rather that the

differences in the age group proportions are brought about by multiple factors,

for example, the precise nature of the vegetation near each site, the distance

from the site at which animals were generally killed, the particular hunting and

butchering methods used (in the case of archeological sites), whether the bones

derive from animals that were killed or scavenged (in the case of the apparent

carnivore accumulations at Swartklip and Sea Harvest), and so forth. One

factor that is probably not involved is the season(s) at which the sites were

occupied since both the grysbok and the steenbok apparently breed more or

less continuously throughout the year, with only minor birth peaks, if any

(Mentis 1972: 74-75; Smithers 1971: 221, 222). In this context, it is important

to point out that the data in Figure 5 are presented in such a way that no discrete

age clusters could be apparent, but in fact I observed no such clusters as I

worked through the material.

Finally, in addition to age data, the nearly universal absence of horns in

females of Raphicerus spp. makes it possible to obtain data on the proportions

of the sexes in the various samples. Unfortunately, however, pre- and post-

depositional factors leading to fragmentation of Raphicerus remains provide a

clear source of bias in favour of the males since the hornless female skulls are

far more fragile and are less likely to survive in pieces large enough to identify.

In the Die Kelders I MSA sample, for example, in which fragmentation is

especially pronounced, not a single identifiable female skull fragment was

encountered, though males are represented by a minimum of 19 individuals.

Even the male skulls have obviously suffered substantially from the extreme

fragmentation, since the total of 19 individuals identifiable as to sex compares

very poorly with the total of 81 animals (in dental categories [V-VII) which

would have more or less full-sized horns, if they were males. Swartklip I reveals

a similar, if not so extreme situation with 23 adult or near adult (category

IV—VII) individuals identifiable on dental material, and only 11 on sexually

relevant skull parts. All 11 are males. Here the author believes the explanation

is predepositional—the Swartklip bone collector was probably a carnivore

(Klein 1975a) whose jaws reduced the small hornless female skulls to pieces too

small to identify (Swartklip contains virtually no identifiable braincase frag-

ments of any bovid, except pieces with horncores attached). In any case, among

all the samples, there is only one—the Die Kelders Later Stone Age assemblage

—in which it seems likely to the author that pre- and post-depositional factors

have not seriously biased sex ratios. Even here, the identifiable female skull frag-

ments are generally not adequate for species identification, forcing the author

to present a sex ratio for R. melanotis and R. campestris lumped: males—41,

186 ANNALS OF THE SOUTH AFRICAN MUSEUM

females—50. This does not differ significantly from 1 : 1 (binomial probability

of a difference from parity roughly 0,25). Although there is some disagreement

on the sex ratio in free-ranging Raphicerus spp. (Kerr & Wilson 1967; Wilson

& Kerr 1969; Rowe-Rowe 1971; Penzhorn 1971), an overview of the data

(Mentis 1972: 75-76), in combination with theoretical considerations on the

population structure of small, non-gregarious, territorial antelopes, suggests

a sex ratio of roughly 1 : 1 in living populations (Jarman 1974: 261).

ACKNOWLEDGEMENTS

I thank Dr Q. B. Hendey for the facilities to undertake this study in his

department at the South African Museum. Dr Hendey, J. E. Parkington,

F. R. Schweitzer, Prof. R. Singer, and J. Wymer kindly made relevant fossil

material available to me. National Science Foundation grant GS-39625 provided

financial support. Dr Hendey provided helpful criticisms of a preliminary draft

of the manuscript.

REFERENCES

ANSELL, W. F. H. 1971. Order Artiodactyla. Jn: MEESTER, J. & SETZER, H. W., eds. The

mammals of Africa: an identification manual. Part 15. Washington, D.C.: Smithsonian

Institution Press.

BAKKER, E. M. VAN ZINDEREN & BUTZER, K. W. 1973. Quaternary environmental changes in

southern Africa.— Soil Sci. 116: 236-248.

BIGALKE, R. C. 1972. The contemporary mammal fauna of Africa. In: Keast, A., Erk, F. C.

& GLass, B., eds. Evolution, mammals and southern continents: 141-194. Albany: State

University of New York Press.

BIGALKE, R. C. 1974. Ungulate behaviour and management, with special reference to hus-

bandry of wild ungulates on South African ranches. In: GEIST, V. & WALTHER, F. eds.

The behaviour of ungulates and its relation to management: 830-852. Morges: International

Union for Conservation of Nature and Natural Resources.

BIGALKE, R. C. & BATEMAN, J. A. 1962. On the status and distribution of ungulate mammals

in the Cape Province, South Africa.— Ann. Cape prov. Mus. 2: 85-109.

BuTzer, K. W. 1973. Geology of Nelson Bay Cave, Robberg, South Africa.—S. Afr. archaeol.

Bull. 28: 97-110.

Dorst, J. & DANDELOT, P. 1970. A field guide to the larger mammals of Africa. London:

Collins.

FAIRHALL, A. W. & YOUNG, A. W. 1973. Methodology of radiocarbon dating and radio-

carbon dates from Nelson Bay Cave.—S. Afr. archaeol. Bull. 28: 90-93.

GENTRY, A. W. In press. Order Artiodactyla. In: MAGLIO, V. ed. Fossil Vertebrates of Africa.

Vol. IV.

GUILDAY, J. E. 1971. The Pleistocene history of the Appalachian Mammal Fauna. Jn: HOLT,

P. C. ed. The distributional history of the biota of the southern Appalachians: 233-262.

Blacksburg: Virginia Polytechnic Institute and State University.

HALTENORTH, T. 1963. Klassifikation der Saugetiere: Artiodactyla I (18).—Handbuch der

Zoologie 8 (32): 1-167.

HENDEY, Q. B. 1968. The Melkbos site: an Upper Pleistocene fossil occurrence in the south-

western Cape Province.— Ann. S. Afr. Mus. 52: 89-119.

HENDEY, Q. B. 1973. Fossil occurrences at Langebaanweg, Cape Province.— Nature, Lond.,

244: 13-14.

HENDEY, Q. B. 1974. The Late Cenozoic Carnivora of the South-Western Cape Province. —

Ann. S. Afr. Mus. 63: 1-369.

HENDEY, Q. B. & HENDEY, H. 1968. New Quaternary fossil sites near Swartklip, Cape Province.

—Ann. S. Afr. Mus. 52: 43-73.

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 187

HOFFMANN, R. R. & STEWART, D. R. M. 1972. Grazer or browser: a classification based on

the stomach-structure and feeding habits of East African ruminants.—Mammalia 36:

226-240.

JARMAN, P. J. 1974. The social organization of antelope in relation to their ecology. — Behaviour

48; 215-267.

Kerr, M. A. & WILSON, V. J. (1967). Notes on reproduction in Sharpe’s grysbok.— Arnoldia

(Rhod.) 3(17): 1-4.

KLEIN, R. G. 1972a. The late Quaternary mammalian fauna of Nelson Bay Cave (Cape

Province, South Africa): Its implications for megafaunal extinctions and environmental

and cultural change.— Quatern. Res. 2: 135-142.

KLEIN, R. G. 1972b. Preliminary report on the July through September 1970 excavations at

Nelson Bay Cave, Plettenberg Bay (Cape Province, South Africa).—Palaeoecology Afr.

6: 177-208.

KLEIN, R. G. 1973. Geological antiquity of Rhodesian Man.— Nature, Lond. 244: 311-312.

KeIn, R. G. 1974a. A provisional statement on terminal Pleistocene mammalian extinctions

in the Cape Biotic Zone, southern Cape Province, South Africa.—S. Afr. archaeol.

Soc. Goodwin Ser. 2: 39-45.

KLEIN, R. G. 1974b. Environment and subsistence of prehistoric man in the southern Cape

Province, South Africa.— World Archaeol. 5: 249-284.

KLEIN, R. G. 1974c. On the taxonomic status, distribution and ecology of the blue antelope,

Hippotragus leucophaeus (Pallas, 1766).—Ann. S. Afr. Mus. 65: 99-143.

KLEIN, R. G. 1975a. Paleoanthropological implications of the non-archaeological bone assem-

blage from Swartklip I, South-Western Cape Province, South Africa.— Quatern. Res. 5:

275-288.

KLEIN, R. G. 1975b. Middle Stone Age man—animal relationships in Southern Africa: evidence

from Die Kelders and Klasies River Mouth.— Science 190: 265-267.

KurTEN, B. 1953. On the variation and population dynamics of fossil and recent mammal

populations.— Acta zool. fenn. 76: 1-122.

MABERLEY, C. T. ASTLEY 1967. The game animals of Southern Africa. Johannesburg: Nelson.

MEESTER, J. 1965. The origins of the southern African mammal fauna. Zoologica afr. 1: 87-93.

MENTIS, M. T. 1972. A review of some life history features of the large herbivores of Africa.—

The Lammergeyer 16: 1-89.

PARKINGTON, J. E. 1972. Seasonal mobility in the Late Stone Age.— Afr. Stud. 31: 223-243.

PENZHORN, B. L. 1971. A note on the sex ratio of steenbok Raphicerus campestris in the

Kalahari Gemsbok National Park.— Koedoe 14: 61-64.

ReHeER, C. A. 1974. Population study of the Casper site bison. Jn: FRIsoN, G. C. ed. The

Casper Site: 113-124. New York: Academic Press.

RoweE-RoweE, D. T. 1971. Sex ratios of steenbok Raphicerus campestris Thunberg seen in

two southern African national parks.— Koedoe 14: 55-59.

SCHWEITZER, F. R. 1970. A preliminary report of excavations of a cave at Die Kelders.—

S. Afr. archaeol. Bull. 25: 136-8.

SCHWEITZER, F. R. 1974. Archaeological evidence for sheep at the Cape.—S. Afr. archaeol.

Bull. 29: 75-82.

Simpson, G. G., Roe, A. & LEWONTIN, R. C. 1960. Quantitative Zoology. New York: Har-

court, Brace & World.

SMITHERS, R. 1971. The mammals of Botswana. Salisbury: National Museums of Rhodesia.

TANKARD, A. J. & SCHWEITZER, F. R. 1974. The geology of Die Kelders Cave and environs:

a paleoenvironmental study.—S. Afr. J. Sci. 70: 365-369.

TANKARD, A. J. & SCHWEITZER, F. R. in press. Textural analysis of cave sediments: Die

Kelders, Cape Province, South Africa. In: DAvipson, D. A. & SHACKLEY, M. L. eds.

Geoarchaeology: Earth Science and the Past. London: Duckworth.

Vooruies, M. R. 1969. Taphonomy and population dynamics of an early Pliocene vertebrate

fauna, Knox County, Nebraska. University of Wyoming Contributions to Geology Special

Paper No. 1: 1-69.

WILSON, V. J. & Kerr, M. A. 1969. Brief notes on reproduction in steenbok Raphicerus

campestris, Thunberg.—Arnoldia (Rhod.) 4 (23): 1-5.

Wyner, J. J. & SINGER, R. 1972. Middle Stone Age occupational settlements on the Tzitzikama

coast, eastern Cape Province, South Africa. In: Ucko, P. J., TRINGHAM, R. & DIMBLEBY,

G. W., eds. Man, settlement and urbanism: 207-210. London: Duckworth.

ANNALS OF THE SOUTH AFRICAN MUSEUM

188

T Wao

Cl I aoe

9780 aaa aa

O16I 00°72 are

ral V

T MZ OUdN JOdNn

VSIN

IT NU

090'T

CO'V~

L4H

‘ds snuaaydoy

I I C

=e as 090°T

08°07 XG CG SLNG

ud V

T MZ OdNn OdaN

07'SZ

LAA

‘ds snaaaiydovy

LLE‘O

SLIT

OdN

v VN

SL70 9060 8s

€0°61 » WOUG 2

Giel WN

sAlyeIedwo7d

SIJOUKJaU SnAaz1Iydoy

TaA—“d

8 LN

99€T wae

OE TZ OL‘0T X

dH WW

dAlyeIedwod

S1dJSadupod snaaa1ydoy

6 IN

vr6-0 Foes

86 IZ OS*IZ x

A WW

sAlyereduiod

SIJOUBJaU SnAgz1ydoy

rd P—2dP

L == N

9660 elm tS

LL‘61 = ek

dA WW

sAlyeIedwiog

SIAJSadUIDI SnAgaIYdDY

6 IN

06L‘0 ee

66 02 Or'0~ X

dH WW

sAleIedu0Od

SIJOUBjau SnABIIYAdDY

"ap—‘*dp

“SOIJOUIT[IUL Ul ore s}UOWOINSeveW [[Y (UloJUojspuelyq = L4q ‘] yop JOATY sorsepyy = | WWM {1

CIP[UVMS = T MZ ‘SIoqqoy pur ‘Aueqyy “uoyiM coved Aeg UosPaN = Y “W ‘A OMN ‘eAeO Avg spurlg = Ogg ‘1 sIOploy

Id = 1 AC -seyeuley pue sopew = FAWN ‘Soyewoy = 4A ‘soyeul = PY) *}X0} OY UI paropisuoo sajduies snsao1ydoy

SNOLIVA 94} UI SY}SUS] VUTD\seIP pUk MOI [eJUOp Joy (N) ozIS o[dures pue “(s) UoNeIANp prepur}s ‘(x) uvoU snowyIYy

[ xIpuoddy

189

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE

100°T

9L"67

VSINN

OO*TS

8L°67

I MZ

L09‘T

OO°EE

LAA

“ds snsaoiydvy

Cc €

es 6L0°7

09°TE 99°0€

u Vv

OdN OaN

IZO'T

UG Ce

"ds snaaoiydoy

G 9

r100 IPE T

OL‘TE ETOE

CcI-IT OI-T

Jd Jd

I i

O€‘ZE O€ OE

cI-Il OI-T

Jd Jd

09°€7

‘ds snaaoiydoy

C G

SEST Ivl‘O

00°61 06°07

ud V

JEN JdNn

aoe I

ae Ors

cI-II OI-T

Jd Jd

(‘7u02) {| xipuoddy

Z6L‘T

LTOE

VSI

I 1d

(Le

OL7'T

L9'OE

VST

I Ad

y i? IN

Iv10 leg ©

Tr'67 Evoe X

dA WIN

sATyeIedWIOd

SIAJSAdUIDI SnAanIydoy

Vv VN

8£8°0 volo Ss

86°87 SL'6C X

fel WW

sAl}eIeduIODd

S1joubjaul snsaaiydvy

WIA

¢ ¢N

IIZ0 8760 Ss

O£ ‘07 976 X

dH WN

dATIVIVAUIOD

S1AJSAdWUDI Sniaaydoy

174 vVN

6670 veoT Ss

€0°07@ 8EOT X

ra fa WW

sAT}eIeduIOZ

SIJOUDJaU Sndaa1ydoy

rd — ed

3 6 INI

LSL‘O 9690S

ELT SILI X

fa fal WW

sAleIedu07D

S1AJSAdUIDI Snsavalydoy

ANNALS OF THE SOUTH AFRICAN MUSEUM

190

077'T

Cr'vZ

VST

I Ad

e ¢

1100 L700

IE€L‘O 9€L‘0

Ha WW

oATyeIeduro*y

SIAJSAAWUIDI *Y

14 v

v100 €Z0°0

¢19°0 L9S‘0

Aa WW

sAlyereduo;d

SIAJSACWUDI “yy

S1ASadub) snsaaydoy

€ ey

90€°0 08L‘0

LVSC EL°97

VSI

IT Naat T MZ

L760

09°€Z

sAlyereduio*d

0L8°0

10°62

LAA

‘ds snaaoiydoy

g 8

LES‘0 S8L'0

98°97 TL'SZ

a Vv

OdNn JdNn

(juo2) | xipuoddy

O€ cl

9160 PIL‘0

6£°97 ELST

MM HAWN

JOAN

¢ dl

€£0°0 6£00 Ss

99L'0 T1830 x

44 WW

sATVeIVdWIOZ

SIJOUDJAU SnAanIydoy

eW —1W/sd —2d

€ c¢ N

7700 LEe0'0 S$

LS9°0 SL9'0

dA WW

dAlyeredulo*d

SIOUDJAU Snszz1ydoy

"W—W/"d—-"d

8 N

Cay |S

80°6I x

sATyeIeduI0Zd

SHOUDJAU SnAzaiydoy

euloSvIG

L ITN

C9ET lor'T Ss

95°97 LO Gaex

44H WN

sATyeIeduo0,d

S1Saduvs snsaaiydoy

¢ [5 IN

76-0 Isso. Ss

8L°ST 69°ST X

dS WIN

sAneIedwuodg

SIJOUDJAU SnAad1YdDy

aN xz IW

FOSSIL HISTORY OF RAPHICERUS H. SMITH IN THE CAPE BIOTIC ZONE 191

Appendix 2

List of samples of Raphicerus melanotis and R. campestris between which

statistically significant mean differences in dental row or diastema length may

be demonstrated. “‘t’ = value of Student’s t; df = degrees of freedom; p = sig-

nificance level (taken as significant if < 0,05) [R. m. = Raphicerus melanotis

(comparative unless otherwise indicated); R. c. = R. campestris (comparative

unless otherwise indicated); NBC W, A, R = Nelson Bay Cave Wilton, Albany

and Robberg; EBC — Elands Bay Cave; ZW 1 = Swartklip 1; DK 1 = Die

Kelders 1; KRM | = Klasies River Mouth Cave 1).

t df )

dP,—dP,

Rem CRC crn oe A ie a We) Gy ee Md a es 2D 15 0,05—0,02

P,—P,

Rema IRC... Ce WG ern ea eae, ean oe Seu gu eA 11 <a OOO!

INEROeR@ Resi PROM. a on My) se Be we DELO i 0,05—0,02

INE @aRGhen. > 2Wil Roms. so os ea ww Ew 3523 11 0,010,001

p2— P4

PaVmleReme = NBC WeRom: 9. 95-15. 0. « . . e252 7 0,05-0,02

M,—M3

EBC 11-15 R. m. > NBC W R. m. a ea 5 alae AO, 31 0,02-0,01

INB@ RR. im. > R.m... eee te eee RM co mics a, TORO 8 0,02-0,01

NBC R R. m. > NBC W re m. PE Rea e Pee eee en 4 OO 32 0,02-0,01

NBG ROR. im. = KRMIMSA Rim . . . . « . . 2,36 8 0,05—0,02

DK 1 LSA R. m. > NBC W R. m. SOE peed ET Meats ae DSS 107 0,01—-0,001

DKGIESAUR im. > ZW iR.m. 2.) 2. . . = 2 2,88 100 0,01-0,001

DKGIGESA Rom: => KRM i MSA R.:m.. . . . . . 2,06 83 0,05—0,02

DKGIBIESAGREm: = Ram, . . 2. we ee ek ww. 282 83 0,01-0,001

EBC 11-15 R.c. > R.c 3,49 8 0,02-0,01

M!— M2

NBC WR.m.>R.m. . ere et alin ae ae ene PS) 40 0,05—0,02

NBC W R. m. > KRM 1 MSA R. Meter ee Se ee ih oe ae) TO) Sil 0,05—0,02

INB@sReRS im: = NBC AUR. fie: 64-5 Seba Ae ew 6 273 11 0,05—0,02

INB@ ROR im. RM TMSA Rom: .- . « + -. :.. 3,09 23 0,01-0,001

XV miphkvonm: = KRM 1 MSA Rom. 3. So. a a 3539 18 0,01-0,001

MURS ROMM sa eek See) I eh sl hw BUSA 2H, < 0,001

=P /iMe—

ee eC y Wee ea ec ee a ee ee DE S6 9 0,02-0,01

eep ine pC) ese ee hoo ee ke os, eae a 8 AOS ita 0,01-0,001

Diastema

Rec. S Rem, SRS dee RM ae Peat ohh ee lle Se SO) 10 < 0,001

DK 1R.c.>DK1R.m. Se earache Bera une et ict ey 2 OOS Poff < 0,001

PRELIMINARY PALYNOLOGICAL RESULTS FROM THE

ALEXANDERSFONTEIN BASIN NEAR KIMBERLEY

Louis SCOTT

Institute for Environmental Sciences, University of the Orange Free State,

Bloemfontein

(With 3 figures)

ABSTRACT

An extensive sampling programme for pollen analysis was carried out around Kimberley,

Republic of South Africa, as part of the multi-disciplinary Alexandersfontein Project. Pre-

liminary results from late to middle Holocene spring and alluvial deposits at Uitzigt,

in the Alexandersfontein Basin, suggest a wetter environment during deposition of part of

the spring sediments, and conditions not unlike the present for the alluvial deposits.

CONTENTS

PAGE

introductioness Ge 2.) 2.8% (2 2” 198

The sampling programme. . . . . 194

Werctationvat Uitziet 8 ess re95

Pollen diagrams from Uitzigt. . . . 195

The alluvial terrace near Uitzigt . . . 197

Conclusions 6. 2 os se Be 2 es 2 198

Neknowledzemems =~. 2 fs 22. = 198

INGIGTENGES, 5G SE! 6) ee Ee, ee

INTRODUCTION

Pleistocene lake deposits in the Alexandersfontein Basin near Kimberley,

were first described by Butzer et al. (1973) and Van Zinderen Bakker & Butzer

(1973). In order to study this region in terms of its environmental history and

prehistoric ecology, a multi-disciplinary programme, the Alexandersfontein

Project, was undertaken in 1974 under the direction of K. W. Butzer of the

University of Chicago. A complex succession of mid-Pleistocene to Holocene

lakes, with related spring deposits, was identified and archaeological surveying,

mapping, collecting, testing and excavating was initiated. In order to assist

in developing a model of prehistoric settlement patterning related to spring

and shoreline micro-habitats in time and space, a sampling programme for

chronometric dating and palynological study was also undertaken. The isotopic

studies, by R. Stuckenrath of the Smithsonian Institution, are still in progress,

and no final chronometric dates for the pollen profiles are reported here. The

palynological investigations were carried out at the Institute for Environmental

Sciences at the University of the Orange Free State, under the supervision of

E. M. van Zinderen Bakker Sr. The slides were mounted in glycerine jelly and

are kept in the slide collections of the Institute and identifications were made

by means of the pollen reference collection, built up by Van Zinderen Bakker.

193

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 193-199, 3 figs.

194 ANNALS OF THE SOUTH AFRICAN MUSEUM

THE SAMPLING PROGRAMME

A total of 87 samples was collected for pollen analyses from the Alex-

andersfontein area. These cover the stratigraphic and lithologic range of

sediments at the fossil spring sites Uitzigt, Mauritsfontein and Benfontein.

Good pollen yield was limited to organic spring deposits at Uitzigt and a

closely linked alluvial terrace a little downstream; these units have been dated

by C'* as late Holocene, i.e. within the last 5000 years. Pollen diagrams

of these two sites are presented in this paper. Exploratory samples from this

area also demonstrated an absence of pollen from weathered Karoo shale

exposures, mid-Pleistocene lake chalks and Upper Pleistocene spring marls.

Poor, but in some cases suitable, pollen or spores were recovered from Upper

Pleistocene spring chalks, underlying the organic spring deposits at Uitzigt

(sample no. 5884), and from the B-horizon of the paleosol developed in late

Pleistocene littoral deposits at the Mauritsfontein site (5689, 5690 and 5762).

Pollen analyses of the samples are in progress.

In order to obtain a better understanding of the fossil pollen assemblages

of the whole region (see Fig. 1), 35 additional samples were taken from the

late Holocene spring mucks and lake marls at the rock-engraving site complex

of Klipfontein (60 km west of Kimberley), from Holocene alluvia on the Vaal

D 9 =

«Del port :

sate: = 7 es

~ Riy Barkly eZ e Ly

ce mS est a’ Te

Riverton | 1, / ms

eo profiles on Vaal Bg S

described h a a

aie Klipfontein |S. =

A Other sites ve

___Doornlaagte Kimber! 2

“NA Rivers

——Roads

__Provincial

me Boarden

any

Benfontein

Alexandersfon- -S/@Uitzigt

bene tein Sella. Ce

/ “+ Maurits

/ nf fontein

Riet River

0 lO

Fig. 1. Alexandersfontein in its regional setting.

‘Modder River

PRELIMINARY PALYNOLOGICAL RESULTS FROM ALEXANDERSFONTEIN BASIN 195

River, and from mid-Pleistocene, freshwater limestones at Doornlaagte and

Rooidam (+ 50 km west of Kimberley). Pollen was found in most of these,

with the exception of those from Rooidam, and analyses on them are in progress.

VEGETATION AT UITZIGT

During December 1974, a survey of the modern vegetation around the

Uitzigt spring was carried out for comparison with data on the present pollen

rain and fossil spectra. The site lies near the north-eastern margin of the treeless

60 km? Alexandersfontein Basin. The regional vegetation is Kalahari Thorn-

veld invaded by Karoo and around Kimberley this veld type consists of Acacia

giraffae-savannah with grasses of the Dry Cymbopogon—Themeda veld and

some of those of the Bankenveld (Acocks 1953). The spring site is surrounded

by an isolated group of trees comprising the following species: Diospyros

lycioides, Rhus pyroides, Ziziphus mucronata, Asparagus laricinus and Lycium

sp., while the near-by savannah is composed mainly of Acacia giraffae, A.

karoo, A. tortilis, Ziziphus mucronata, Ehretia rigida and Lycium sp. In the

veld around the spring many species of Gramineae occur, e.g. Eragrostis spp.,

Cynodon dactylon, Sporobolus tenellus etc., and Compositae like Pentzia spp.,

Gazania sp., and Gnaphalium sp., as well as other herbs such as Lasiosiphon

sp., Thesium hystrix, Stachys sp., Salvia sp., Portulaca sp., Walafrida densiflora,

Sutera sp., Limosella capensis, Alectra orobanchoides, Psilocaulon sp., Hypertelis

sp., Limeum aethiopicum, Asclepias sp., Cordylogyne sp., Convolvulus sp., Sal-

sola sp., Polygala sp., Indigofera alternans, Zygophyllum sp., Apium graveolens,

Umbellifera centella, Amsinckia sp., Datura sp., Lycium sp., Albuca sp., Trachy-

andra sp., etc. Cyperaceae occur in moist patches around the spring.

POLLEN DIAGRAMS FROM UITZIGT

THE SPRING-EYE DEPOSITS

Samples were removed from the calcareous organic spring-eye deposits at

Uitzigt, which are resting on calcrete, by means of a Hiller auger (sample

nos. 5677-82 and 5715-39), as well as a surface sample for the assessment of

the modern pollen rain. All samples proved to contain countable pollen.

Detailed pollen analyses were performed on four samples from the profile

(1974, cores 3, 4) and also on the surface sample. The species composition (see

Fig. 2) was calculated by examination of 250, or in a few cases 100, pollen

grains, depending on the yield of the sample. Cyperaceae, Chenopodiaceae,

Gramineae, Compositae, Aizoaceae, Diospyros and Rhus are the main pollen

components. The oldest sample shows an increase in percentage of Cyperaceae

which are probably related to spring activity and a decrease in pollen of the

isolated group of trees now found growing round the spring (Diospyros and

Rhus). The surface sample from the eye shows a sharp increase in the per-

centage of grasses, probably a result of the high precipitation during the previous

196 ANNALS OF THE SOUTH AFRICAN MUSEUM

EXPLANATION : [J =10%

COMPOSITAE ne '

( without <x re] = a 2)

TARCHONANTHUS) —— = = a

pw 2 = Ww > (=) = wd

oz WwW a (=

wD S a = > _ rv) 4 qa <x wn vn = =< Ss Ss =

a S = i] Q x nu et HO uw o fw a =

i= a a _ _ i>) uJ Se —) =) ce oO re) e = o [4 a z=

S = = = - [4 (=) <x >a 8D 434“ HE & lu a = 2) Sion =

< = = = e uJ Ww uw z [) ao > uw wi "=" "Nn PEO HOD Ez =

a =< wi a > = Sr a Sa = pone, e222: 3 2 S275 ne

> ec =x :

aes —- Fe 1p a7 3

i Dee ee seen oF a ee i: |

cere ae ee ae ae

—Boiiddg i. cue a:

Says a a yale hk aR! :

a i 8 Jes eo :

= S S41) 13 ee Dye oe : )

eS et a ey ee == ee |

ao S Sars 1. ileal ne

ae SSjsi1 | Ua eae a! : 1

SJ sl fd = ]

Fig. 2. Pollen diagram of the spring deposits at Uitzigt.

season. There is also a higher percentage of Diospyros and Rhus in this sample.

The presence of rare pollen of the introduced plants, Pinus, Eucalyptus and

Alternanthera, as well as the presence of a seed of Datura, a naturalized plant,

at 11 cm, suggests that the deposits are disturbed to some extent.

DEPOSITS ADJACENT TO THE SPRING-EYE

Further samples from a different generation of calcareous organic spring

sediment were collected from a pit (1975, no. 1) about 5 metres from the eye

and situated about | metre higher on the slope. The material is similar in appear-

ance to those from the eye, but are more consolidated and undisturbed. The

deposits have a thickness of about 90 cm. A few in situ artefacts and pieces of

ostrich eggshell are present. The pollen of ten samples, which are not as rich

as the eye-deposits, suggests wetter conditions for these deposits than those

from the eye, for the following reasons (see Fig. 2):

I. They show a marked reduction in the percentage of pollen from the

halophytic Chenopodiaceae.

2. They show a reduction in the percentage of Aizoaceae pollen, which

is taken to represent a karoid type of vegetation. It is believed that this group

of pollen grains may also include some specimens from the closely related

Phytolaccaceae and Portulacaceae families.

3. There is an increase in the percentage of Gramineae (between 30 and

PRELIMINARY PALYNOLOGICAL RESULTS FROM ALEXANDERSFONTEIN BASIN 197

64%), as the highest percentage obtained in the spring-eye samples was 19,2 per

cent. The former values compare favourably with the surface sample (59,6 %),

which as noted possibly reflects a very wet year.

Although the Compositae are also more abundant in these deposits, it is

believed that the grass : Compositae ratio used by Van Zinderen Bakker (1957)

and Coetzee (1967) to measure dryness is not applicable here, because the

Compositae present are not the usual types for this area. There are few of the

normal spiny forms which make up the majority of Karoo and South African

species and which were found in the spring-eye deposits. Compositae types II

(Stoebe-like morphology) and HI (Artemisia-like morphology) probably form

part of the spring vegetation, firstly because they show the same distribution

as the Cyperaceae, which are probably related to spring activity, and, secondly,

because these types are often found together in clusters, showing that they do

not represent the pollen rain, but probably grew on the site. These forms

therefore cannot suggest dryness of the surrounding veld. On the other hand,

the pollen of Compositae type I (Lactuca-like morphology) which is absent

in the surface and younger samples of the eye, is probably blown in from the

surrounding veld because it shows a similar distribution as the Gramineae

which mainly represent the veld. The samples contain very few tree pollen

grains. Rhus is only rarely recorded at 80 cm, while Tarchonanthus and winged

pollen grains of Podocarpus which could have been transported for hundreds

of kilometres by the wind, are found in small numbers throughout the profile.

In contrast to the present situation the pollen spectra suggest the absence of

trees growing directly around the spring at the time of deposition.

THE ALLUVIAL TERRACE NEAR UITZIGT

During early February 1975, a series of samples (nos. 5799-5816) was

collected from the side of a gully in an alluvial terrace of Holocene age near

Uitzigt. The gully is situated about 100 metres from the spring at a lower alti-

tude in the Alexandersfontein Basin. The terrace is composed of grey organic

calcareous deposits with calcrete pebbles. Three samples were analysed from

depths of 10, 60 and 110 cm in the wall of the gully as well as a surface sample.

The terrace sediments, which naturally represent a much drier environment

than the spring, show a very high percentage (> 60%) of Chenopodiaceae,

with smaller numbers of Compositae, grasses and trilete spores (Pteridophyta

and ?Bryophyta) (see Fig. 3). The pollen spectra at the three levels are similar

to the surface sample from the gully except in the higher percentage of grasses,

and correspondingly lower percentage of Chenopodiaceae in the latter sample.

As in the surface sample from the spring-eye, this is likely to result from the

unusually high precipitation of the last two years. The drier environment as

suggested by the pollen spectra from the terrace is therefore probably not much

different from the average conditions of recent times. Unlike the spring deposits

adjacent to the eye, the Compositae types in the alluvial terrace were the same as

those in the surface samples and spectra from the spring-eye.

198 ANNALS OF THE SOUTH AFRICAN MUSEUM

EXPLANATION: a) =10%

COMPOSITAE a

tad

oe ~~ = = ! =

ta Ow w a

= = i pew) ta —! a ct Za cn wn =

S = ie < < wn 3 qgow qtw wu (=) a

be = ky. = tad tJ ==) = — @&W Keo oO w @& [> 4

= = = an @ wo oO ee q€odcat oc ce Ke > < <x

x = = oe ed oS < <x > a Snr oazraeuww a — oO ”n <x

SHS 2£ FE 28. ,B ef seSSules 2 See

[==} (4) — oo So oOo a C-¢ — MN OQoorzt od oO K- 2 [= 4 = 42 a =

———— eee

Fig. 3. Pollen diagram of the alluvial terrace near Uitzigt.

CONCLUSION

As was expected, the palynological work in the northern Cape region

produced better pollen assemblages from the younger organic spring deposits

and alluvial terraces than from the lake chalks and calcretes. However, sporo-

morphae obtained from deposits of the latter kind at Doornlaagte, Klipfontein

and Uitzigt show that these deposits are not entirely devoid of potential for

palynology. When comparing the three pollen profiles of the organic spring

and alluvial deposits from the Alexandersfontein Basin, as described here,

the spectra from the deposits adjacent to the spring-eye differ considerably

from the assemblages of the spring-eye deposits and the alluvial terrace, which

show greater similarity to assemblages of present pollen rain as obtained from

the surface samples. The former appear to represent a more humid environ-

ment. The final interpretation of these results can best be made when the

isotopic and sedimentological studies are completed and the remaining pollen

results available.

ACKNOWLEDGEMENTS

I wish to express my thanks to Professor K. W. Butzer for initiating this

research project and for his encouragement and suggestions. Professor E. M.

van Zinderen Bakker Sr of the Institute for Environmental Sciences supervised

PRELIMINARY PALYNOLOGICAL RESULTS FROM ALEXANDERSFONTEIN BASIN 199

the project and kindly supported the fieldwork. I am also grateful to Mr D. B.

Miiller and Dr H. J. T. Venter for assisting me with the identification of plants

and to Mr C. E. Boshoff and Professor A. Horowitz for their assistance with

the sampling.

REFERENCES

Acocks, J. P. H. 1953. Veld types of South Africa. — Bot. Surv. Mem. 28: Pretoria. Government

Printer.

BuTZzER, K. W., Fock, G. J., STUCKENRATH, R. & ZILCH, A. 1973. Palaeohydrology of Late

Pleistocene Lake Alexandersfontein, Kimberley, South Africa.— Nature, Lond. 243:

328-330.

CoETZEE, J. A. 1967. Pollen analytical studies in East and Southern Africa. In: ZINDEREN

BAKKER, E. M. VAN Sr. Palaeoecology of Africa 3. Cape Town: Balkema.

ZINDEREN BAKKER, E. M. VAN Sr. 1957. A pollen analytical investigation of the Florisbad

deposits (South Africa). In: CLarK, J. D. Third Pan-African Congress on Prehistory,

Livingstone 1955: 56-67. London: Chatto and Windus.

ZINDEREN BAKKER, E. M. VAN Sr. & BUTZER, K. W. 1973. Quaternary Environmental Changes

in Southern Africa.— Soil Science. 116 (3): 236-248.

SECTION 4

ARCHAEOLOGY

' es

(

x |

{

= C

. : ‘

| aa

ANALOGY AND ARCHAEOLOGY

R. J. MASON

Archaeological Research Unit, University of the Witwatersrand, Johannesburg

Analogy, the relationship of things to their models, is the lifeblood of the

historical sciences. Archaeology, as one of the historical sciences, would not

exist without analogy. We can know about the prehistoric past only in terms of

analogy with models taken from the present. It is extraordinary, therefore, to

observe that intensive systematic teaching in the philosophical mechanisms of

analogy is neglected in our schools of prehistory; few of our teachers of pre-

history are able to present a theoretical account of analogy and its practical

application in archaeology. (Sapire 1972.)

The practical implications of analogy for the research archaeologist are very

clear. Each and every research archaeologist should have his own personal log-

book of analogies and models for every aspect of his work from his reconstruc-

tion of prehistoric technological behaviour to the analysis of geological processes

responsible for the mass deposition agencies preserving his site. Knowing

archaeologists as I do, they may accept this exhortation, but do little about it.

It is sad to observe the Palaeolithic or Stone Age archaeologist who has never

lived with living representatives of his specialized knowledge, hunter—gatherers

who have managed to survive into the latter twentieth century. Examples of

ivory-tower prehistorians, entirely lacking in personal experience of hunter—

gatherers or other living peoples related to archaeology, are only too common.

An engineer specialized in the fabrication of metals need not do the work

himself, but he keeps close to it in the work of his factory managers; archae-

ologists do not have factory managers, so they should be obliged to do the

dirty work for themselves and get in on the hunter-gatherer scene personally.

What essentially practical science has the richest store of models for the

1975 field archaeologist? In my opinion the science of mineral exploration is a

prime source of models for modern archaeology. Both archaeology and mineral

exploration search for obscure materials located as minute fractions in vast

bodies of enclosing but sterile deposit.

The science of mineral exploration combines three sister sciences: explora-

tion geophysics, exploration geochemistry and exploration geology.

Somehow or other the archaeologist has got to get the practical help of

scientists in all three exploration disciplines. Of course, for many years archae-

ologists have recognized the relevance of some exploration techniques to their

subject, but no full-scale attempt to use mineral exploration techniques for

archaeological purposes has ever been made.

To quote a single example: in 1959 Dr Tony Brink and I were able to use a

30 inch (76 cm) boring machine to drill a line of seven holes in the Chuniespoort

Valley slope deposits, some up to 12 metres deep. We were able to probe the

201

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 201-202.

202 ANNALS OF THE SOUTH AFRICAN MUSEUM

Chuniespoort soils in our search for archaeological data. Alas, we were not

able to expand our 1959 experiment to other sites, but substantial areas of

South African landscape cry out for the 1959 Chuniespoort treatment.

If we were able to acquire mineral exploration teams to help in fieldwork

we might be able to locate entirely new sources of archaeological data. The

history of South African mineral exploration at once suggests a valuable model

for the location of new archaeological deposits. South Africa’s new frontier in

mineral exploration lies along the south bank of the Orange River in the northern

Cape. Here we have a region known as Bushmanland which was ignored as a

potential source of base minerals because it was a granite region and granites were

not recognized as a source of base minerals. However, things have changed.

Exploration geophysics, geochemistry and geology have now discovered the

history of granite development revealing that base metal ore bodies exist in the

earliest pre-granite rocks subsequently formed into granites and gneisses. Vast

wealth awaits the mineral exploration team capable of locating the eo kind

of pre-granite rocks in a granite assemblage.

A new wealth of archaeological information relating to man or environment

may await the archaeologist if we search new deposits with geophysical, geo-

chemical and geological techniques borrowed from mineral exploration sciences.

The traditional sources of our data are caves and slow-fill areas: what about all

other forms of superficial deposit? The evidence we want may be present only

in geological or geochemical form, but it may surprise us all.

Finally, I should like to remark that the pace of archaeological research

is too slow. For example, the slow and painful reassessment of European

Cenozoic environments now taking place could have been completed in a

fraction of the time by a single mineral exploration company. The same com-

ment applies to the unduly prolonged exploration of some archaeological sites.

REFERENCE

SAPIRE, D. 1972. Models and analogies in archaeological interpretation.—S. Afr. archaeol.

Soc. Goodwin Series 1: 6-7.

THE HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE

SOUTHERN CAPE

H. J. DEACON & MARY BROOKER

Department of Archaeology, University of Stellenbosch

(With 3 figures)

ABSTRACT

In 1971 a research project aimed at the study of the late Quaternary environment and

culture change in the southern Cape region was initiated and observations have been made

at several archaeological sites. The most promising is Boomplaas Cave where some five metres

of deposits range in age from Holocene to early Upper Pleistocene. The well-stratified sequence

includes cultural material as well as plant and animal remains. Most recently the cave was

used by prehistoric herders, ancestral Hottentots, as a stock-kraal and prior to that the cave

was used by hunters and gatherers as a locality for the storage of Pappea capensis fruits.

Occupation levels of the earlier Holocene and Upper Pleistocene are still under study. The

larger mammalian faunal remains show shifts from closed to more open and back to closed

habitat species through the whole sequence that may reflect major trends in the pattern of

environmental change. Confirmation of this patterning is being sought in the current study

of the rich microfaunal samples accumulated from owl pellets and unrelated to human activi-

ties. The major industrial change in the sequence is prior to 21 000 years B.P. and while this

does not directly correlate with environmental change, there are lower order technological

changes which may reflect adaptive responses to changing environment in the late Upper

Pleistocene and Holocene.

CONTENTS

PAGE

Introduction . : ; x 1203

Project progress ; : . 204

Boomplaas Cave : ; 5 Ady

Conclusions . : : 5 6 AS}

Acknowledgements . . . 214

References ; 5 ; = 214

INTRODUCTION

The Southern Cape which is defined here as equivalent to the eastern zone

of the Cape Folded Mountain Belt (Wellington 1955: 108) comprises a series

of east-west trending mountain ranges and a narrow coastal plain. It has been

described as part of the Cape biotic zone (Hendey 1974: 16) and is part of the

landscape of the Cape Flora, the ‘fynbos’. A major ecological boundary separates

the Southern Cape from the more arid Karoo of the immediate interior of

South Africa and the boundary lies along the inland margin of the fold moun-

tains. In 1971 a research project, “The Langkloof Archaeological Research

Project: Late Quarternary environment and culture change in the Southern

Cape’ was initiated at the University of Stellenbosch. This is conceived as

primarily an archaeological study with a focus on tracing changing human

adaptations in the Upper Pleistocene and Holocene. Because the study of

203

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 203-214, 3 figs.

204 ANNALS OF THE SOUTH AFRICAN MUSEUM

adaptation is the study of goodness of fit between human behaviour and the

total environmental context in which it operates, the research project is of

necessity multi-disciplinary and indeed archaeological fieldwork where under-

taken on an adequate scale and with due attention to research design can

yield basic data that contribute directly to palaeoecological and palaeoclimatic

studies.

The southern Cape is archaeologically one of the more intensively studied

regions of southern Africa and, in the main, cave occurrences along the coast

have attracted attention. Much of the work has, however, been done outside

of reference to any specific design and has been in this sense opportunistic,

but it has shown the establishment and continuation of stable densities of

human populations in the region through the last hundred thousand years.

That Upper Pleistocene and Holocene settlement patterns included the occupa-

tion of caves allows sampling to be directed at this type of occurrence with

some advantages in preservation of materials and their context. A bias in

previous studies in the southern Cape lies not so much in the emphasis on cave

excavation as in the selection of cave sites mainly in the coastal zone. At coastal

caves there is the maximum effect of wide-ranging changes due to eustatic

lowering of sea-level during the Upper Pleistocene which are difficult to factor

out from more direct effects of changes in climate, flora and fauna on past

human habitats. The goal of the Langkloof archaeological project has been

to build on and extend previous observations on culture change and changes

in past environment.

PROJECT PROGRESS

The initial observations were made in the Langkloof and were to follow

up on preliminary excavations undertaken by Jolly in 1969. Of several sites

sampled by Jolly only Rautenbach’s Cave (33.318, 23.42E, 1 : 50000 sheet

reference 3323 DA Voorkloof) would seem to warrant further investigation.

It occupies a small anticlinal structure in Witteberg quartzite perched on a

hillside in the entrance to a small kloof draining into the Baviaans River. The

floor is 12 metres by 6 metres and the deposit is some 2 metres deep in Jolly’s

cutting. The upper metre of deposit is a grey ashy unit with some preserved

organic remains. Surface disturbances made by rodents and the cutting in this

upper unit show the occurrence of pottery, Themeda triandra grass bases,

Boophone disticha bulbar leaves, Pappea capensis fruits and Freezia sp. corm

scales. The middle unit of the deposit is a red-brown loam with spall material

and oxidation here precludes organic preservation. The basal part of the deposit

from 15 cm above bedrock is a black spall-free iron humate stained horizon.

The upper unit is Holocene in age and it is evident on the typology of cultural

material included in the lower units that the sequence extends into the Upper

Pleistocene. A single radiocarbon determination gave a result of 12 560 + 100

(Pta. —251) apparently related to the lower portion of the sequence. The precise

205

HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE SOUTHERN CAPE

“SOUS DALD POIVAPOXS JULIIOAUI 1OY}O O} UONLIOI UI DARD SeL[dWOOg Jo UONedO] 94} SuIMoYs sdeD UsoYyINOs oy} Jo dey] *] “SIF

Y SCO} WeD

SISN3

SNMOL @

SAIS G3LWAVOXa 4

Io py RESO

4 3

now

JOAIY Salse|>

EY ySaIIDUYy,

Id ON! 4

Lt oN) NM

13NI3U- JVVuD

oy OS Ov Of O« Ol 0) vl

LOld

NOIDSY AdVD NYSHLNOS

SVV 1d WOO

¥ Ss\unoD

Ava 13SSO

Aeg uos|an\< 9ZIe|g IS ade)

JOny sahei., VNSANY@L LS :

vy Q)

Wo» 390I9@

y ISINY|EO 4 nO

ae Ps ; a \

ey) / « weybde vonoD

ems Sj »

E] O

“ane 3 } : ain i 6 U o~ |

§ v 9 n oO »

én _ GNHOOHS1GNO

Upequainey 4 ct 44002 LI1Wo

A 2 @ wy POOP S|PHNG a s

seed (008 i ” Vv NA

IYOWMOTIIMe aa

INIEIV JONIUd@

OBNESONIVT

iS 8

_N330Y38V @ Ss

1sam LYOsnv38 - N

206 ANNALS OF THE SOUTH AFRICAN MUSEUM

provenance of the sample submitted by Jolly is unknown. Rautenbach’s Cave

draws attention to the possibility of extending observations on the subsistence

ecology of Holocene hunter—gatherers made for example in the far eastern

limit of the Cape Folded Belt at Melkhoutboom Cave (Deacon H. J. 1972).

The same potential, however, exists at other sites in the Langkloof. The cul-

tural materials from the Upper Pleistocene units are of some typological

interest as they include blades within the size range of the Howieson’s Poort

industry. Jolly’s samples, however, are inadequate for any detailed study.

In excavations carried out in 1972 by H. J. Deacon and R. G. Klein in

collaboration with J. Deacon, M. Brooker and with the assistance of students

from the University of Stellenbosch, two cave sites in the Langkloof were test

excavated. The Kangkara Cave (33.47S, 23.05E, sheet 3323 CC Kruisvallei)

is an open rock shelter cave in Table Mountain sandstone in the Upper Keur-

booms River valley. Excavation here showed 2 metres depth of deposit without

reaching bed-rock. The lowest 30 cm of the sequence sampled is an iron humate

stained unit that includes faceted platformed flakes, prepared cores and pre-

eminent selection of quartzite raw materials. This would appear to relate to

the early Upper Pleistocene and correlates with the basal units at caves on the

adjacent coast. A disconformity marked by a rock tumble separates the earlier

culture-stratigraphic unit from the overlying terminal Pleistocene and Holocene

sequence. The latter covers more than 12 500 years (Pta.-782) and artefactual

as well as organic remains (bone and shell) are preserved. Within the culture-

stratigraphic terminology in use in the eastern zone of the Cape Folded Belt,

the rock tumble and overlying layers include materials related to the Robberg,

Albany and Wilton industries.

The second test excavated site was Paardeberg Cave in the Kougaberge,

north of Joubertina (33.44S, 24.01E, sheet 3324 CA Sandvlakte). This is a

large cave which in spite of its remoteness had suffered some disturbance by

casual digging in 1966. Although there is an estimated thickness of 1 metre

of later Holocene deposit in a part of the cave, the test excavation was located

towards the mouth where the deposits are thinner to establish the sequence

and preservation in pre-Holocene levels. The sequence intersected was 80 cm

deep and overlying an oxidized stony yellow-brown loam layer (10 cm) con-

taining a few large quartzite flakes was a unit less oxidized and of greyer hue

that showed a high density of silcrete artefacts. Typologically the industry is

of some interest as it includes unifacial and bifacial points and ‘nosed’ scrapers

associated with a strong blade element. The top of this unit is dated to greater

than 44 000 years (Pta.-773). A disconformity separates the above unit from

cultural material and organic remains of Holocene age. The Holocene remains

are of interest as at Rautenbach’s Cave through the good condition of preserva-

tion of plant and bone remains. These initial investigations within the project

showed that there were a number of cave sites within the mountain zone with

sequence extending into the Upper Pleistocene. They also pointed to the problem

of obtaining observations on associated organic remains for much of the Upper

HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE SOUTHERN CAPE 207

Pleistocene in areas underlain by the acid rocks of the Cape System. A further

aspect made more prominent by these observations was the discontinuities

in the sequences exposed. At these sites inland and again at sites on the coast

it appeared that the time range between some 40000 and 15000 years B.P.

was not represented by any marked build up in deposits through occupation.

Partial explanation for this may lie in eustatic changes in sea-level and settle-

ment taking place on the now submerged coastal plain beyond the reach of

investigators.

Further progress has depended on overcoming some of the limitations of

the sites initially investigated. Following on Goodwin’s (1930: 565) results

from a 1930 excavation in the Cango Caves, a conscious search was made for

sites in the Cango Beds, an upfaulted block of pre-Cape rocks that includes

bands of limestone that would be more favourable for bone preservation. In

the more durable sandstone—quartzite caves and shelters the build-up of deposits

is slow, giving minimal separation of culture-stratigraphic units and making

for maximal physical fragmentation of organic remains such as bone through

simple trampling. The limestone area offered the prospect of giving a more

expanded sequence. As the Cango Caves themselves serve as a tourist facility

there are some added problems of excavation there and Boomplaas Cave

4 kilometres away was chosen for excavation. Subsequent work within the

project which has amounted to some four months’ excavation time during

three separate field seasons has been concentrated on this site. As a long-

sequence site, covering perhaps the last 80 000 years with human occupation

throughout and reasonable preservation of associated faunal materials, Boom-

plaas Cave offers the potential of providing an excellent record of changing

human adaptation in the southern Cape in the Upper Pleistocene and Holocene.

While this investigation has become central to the Langkloof project in terms

of the capabilities of the research team from Stellenbosch, another site, a large

rock shelter cave on the farm Buffelskloof 15 km from Calitzdorp and some

30 km from Boomplaas, is currently being excavated by H. Opperman of the

University of Fort Hare.

BOOMPLAAS CAVE

Boomplaas Cave occupies part of a fissure system in Drupkelderkop

(33.238, 22.11E, sheet 3322 AC Kangogrotte) opening above the floor of the

valley at the same elevation as the Cango Caves. The opening to the fissure

forming the cave has a floor area of some 250 metres. Thus far some 100 square

metres have been excavated to a depth of approximately 1 metre and a sounding

of 1 square metre (square P12) excavated to bed-rock. The latter has provided

a section through the complete sequence. The deposits are some 5 metres

thick and well stratified with discrete depositional units representing occupation

and non-occupation horizons. A suite of seven radiocarbon age determinations

currently gives a time scale for the sequence, the lower half of which is older

208 ANNALS OF THE SOUTH AFRICAN MUSEUM

BOOMPLAAS CAVE

SECTION

= ROCK PAINTINGS

ie EXCAVATED AREA

Fig. 2. Plan and section of Boomplaas Cave showing the grid and excavation area.

HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE SOUTHERN CAPE 209

than 40 000 years and beyond standard radiocarbon age determination. Major

disconformities are not apparent and the site affords a more complete Holocene-

Upper Pleistocene sequence than is yet known from the southern Cape and it

includes occupation horizons in the time range between 40000 and 15 000

years. A minor stratigraphic disconformity has been noted in this time range

at Boomplaas but may be a depositional break of only some 2000 to 3 000

years. It is, however, important in relation to the cultural sequence.

The sequence at Boomplaas can be viewed in different terms, from the

aspect of the depositional history, the litho-stratigraphic units, from the aspect

of the organic remains, the biostratigraphic units, and from the aspect of the

cultural traces of preserved features and artefacts, the culture-stratigraphic

units. A detailed repo1t on investigation is in preparation for publication

elsewhere and only soine of the major features are discussed here. As might

be expected in any long-sequence cave-fill the lithological units are varied and

complex. There is, however, a constant tendency for the sand fraction column

samples through the sequence to show a trimodal particle size distribution.

This may reflect a balance in the contribution to the deposits from different

sources. In the build-up of the deposits a fissure opening to the surface in the

rear of the cave and material transported by human and natural agencies

through the mouth of the cave are of importance. There is, in addition, contri-

bution of a coarse element, roof spalls and rock tumble, to the deposit. The

most notable effect of human occupation is the increase in organic material

and ash. Red-brown loams form with lower intensity of occupation and charac-

teristically have a relatively high micro-fauna content. Three layers with high

angular roof spall content mark changes in the stresses within the domed roof

and are explicable as the result of frost fracturing. These can be placed in time

and may have palaeoclimatic significance. An initial series of sediment samples

have been studied by J. J. N. Lambrechts.

The sequence includes well-preserved fauna from all levels, but plant

remains are preserved in humified form from the Holocene horizons only.

The mammalian fauna is under study by R. G. Klein and additional observa-

tions on the micromammalian remains have been made by V. Scott and

T. Pocock. The expectation in the Holocene-Upper Pleistocene time range is

that the mammalian fauna will be largely composed of extant forms. Interest

in the fauna lies in the main in its potential as a reflection of environmental

change and changing hunting preferences. In the occupation horizons it can

be postulated that the main source of larger mammalian faunal remains is

the return of man-made kills to the site. Non-occupation horizons also include

larger mammal remains, the result of carnivore predation and natural deaths,

and are of principal importance for the study of micro-fauna and changes in

patterns of owl predation.

Following out of Klein’s study of the samples of mammalian fauna from

the current stage of excavation progress it is apparent that there is cyclic change

through time between assemblages dominated by grazing and browsing antelope.

210 ANNALS OF THE SOUTH AFRICAN MUSEUM

It is reasonably well established that the main adaptation of Holocene hunter—

gatherers was to the hunting of small browsers and it can be argued that this

reflects the disappearance of a niche for grazers through increased shrub and

bush cover in the Holocene in the fynbos areas (Deacon H. J. 1972: 34; Klein

1974: 273). In the Boomplaas sequence the Holocene is marked by the occur-

rence of marine shell, an indication of the eustatic recovery of sea-level to a

point that brought the coast within range of foraging activities of the cave users,

and again in the dominance of grysbok in the faunal remains. In the later and

middle Upper Pleistocene, in a time range extending back to more than 40 000

years, changed environmental conditions are indicated by the absence of grys-

bok and the presence of wildebeest. Still earlier in the Upper Pleistocene

grysbok are again the dominant antelope. Although human selection may be

a factor contributing to this cyclic pattern, real changes in the environmental

mosaic are suggested. A test is afforded by the micro-fauna remains from owl

pellets and preliminary studies indicate that the micro-fauna will be a useful

source of independent information on past environment. Rich accumulations

of micro-fauna are a feature of some stratigraphic units in the middle part of

the sequence and there are modern owl roosts in the same limestone outcrop

for comparative study. 7

A control study of the modern vegetation cover of the area has been initiated

by R. O. Moffett both for the identification of plant remains recovered and to

aid in understanding the direction and amplitude of past vegetation changes.

Thus far this part of the study has contributed a collection of over 400 plant

species to the district herbarium in Stellenbosch. Limited fragmentary humified

macroscopic plant remains have been recovered from the upper metre of

~ deposit. Little success has thus far been achieved in the study of pollens from

cave sites anywhere in South Africa and at Boomplaas an initial series of

samples analysed showed negative results. At present the prospects for the

study of micro-plant fossils must be seen as low.

The cultural sequence has been best studied in the topmost units and the

upper 80 cm of deposit (CBM, DGL and BLD 1-2A, Fig. 3) represent the use

of the cave as a kraal. This accumulation of banded calcined dung and sandy

loam layers dates to the last 1 700 years and is related to the appearance of

pastoralists in the region. Associated with the herding way of life is the intro-

duction of pottery and the construction of relatively elaborate hearth features

(BLD2), some of which are packed with stone. Kraal sites of early herders

in the Cape have low archaeological visibility, making this occurrence signifi-

cant. There are very limited associated artefacts in the upper unit.

Below a uniform calcined dung band (2A) forming a floor to the unit is

a carbonized and humified dark brown marker horizon (BLD3). This would

appear to have been formed through the burning of dung accumulation over-

lying the former natural surface layer of the cave floor which was littered with

woody and leafy wind drift. The horizon includes a range of artefactual materials

typical of the microlithic Holocene industry that is found along the length of

HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE SOUTHERN CAPE PAL

LITHOLGG)

BIOSTRATIGRAPHIC

UNTTS

LITHOLOGICAL

WILTON INDUSTRY

HOLOCENE

UNDIF FERENTIATED

PLEISTOCENE

Fig. 3. Section through the Boomplaas Cave deposits showing the lithology, dating, cultural

and biostratigraphic units.

PD) ANNALS OF THE SOUTH AFRICAN MUSEUM

the eastern zone of the Cape Folded Mountains. Numbers of pits lined with

Boophone disticha leaf and grass are cut from the base of this layer into the

underlying deposits in the central area of the cave. These are storage pits and

the few that have not been emptied contain fruits of Pappea capensis. This late

Holocene practice of storing fruits can be suggested as related to the harvesting

of oil rich fruits rather than the collection and storage of edible foodstuffs. At

present the only comparable observations are from Melkhoutboom Cave

where similarly oil rich but less palatable fruits were stored in grass lined pits.

Although the plant remains are humified, the excavation technique was designed

specifically to maximize the recovery of plant materials and it was apparent

that few edible plant species are represented. A single corm of Watsonia and

two Hypoxis villosa corms are amongst the limited sample of food remains

thus far recorded from the pits. At this time level probably between 2 000 and

2 500 years B.P. the cave would seem to have served as a base for a specialized

and seasonally restricted activity. Ashy material (BLD3AM) underlies the

marker horizon and at least two distinct surfaces representing short term

occupations could be distinguished on the occurrence of humified woody litter

in the unit. Below this upper Stone Age occupation complex is a coarse sand

loam horizon with abundant micro-faunal remains (BLDBL). This would

seem to represent a break in occupation. At the base of the micro-fauna rich

layer is a floor (FBL) exposed thus far over some 30 sq. metres that shows a

low density scatter of fauna, artefacts and humified plant remains. Below this

floor is a series of discrete occupations associated with extensive lenticular ash

bodies one of which (AF1 BRL) is dated to 6 400 + 75 (UW 306). Area stripping

of the interdigitating ash lenses is the stage reached by larger scale excavation

at the present and further comments on the culture-stratigraphic succession

are restricted to information gained from the excavation of the P12 sounding.

A red brown sandy loam (BRL) underlies the ash lenses developed in the

deposits detailed above. While not devoid of traces of occupation it represents

a period of low intensity utilization of the cave by man. The Pleistocene—

Holocene boundary is represented in this layer by the disappearance of marine

shell manuports. A complex of ash lenses and intercalated dark humic loams

(CL) developed to a thickness of 30 cms underlies the end Pleistocene red brown

loam. The base of this unit is dated to 14 200 + 240 (UW —301) and the whole

is by contrast a phase of more intensive site utilization. The cultural materials

from this unit are referred to the Albany Industry in the local culture-

stratigraphic terminology. A further late Upper Pleistocene occupation (Rob-

berg Industry) is represented in the underlying depositional units (GWA-

BWA) that in turn bottom on a disconformable surface. The disconformity

as noted correlates with a major cultural discontinuity and below the dis-

conformity the artefact industry is macrolithic in character. The latter industry

for obvious sampling limitations has been designated as an undifferentiated

Upper Pleistocene blade industry but in all probability represents a complex of

industries.

HOLOCENE AND UPPER PLEISTOCENE SEQUENCE IN THE SOUTHERN CAPE 213

The Holocene and late Upper Pleistocene Stone Age sequence at Boom-

plaas and other occurrences in the Southern Cape appears to represent a

pattern of three stable adaptations each with a 6000-8 000 year duration

during which set norms of technological and subsistence behaviour were

established. It is a time range of marked environmental change and each

adaptation appears to be an adjustment to gradually changing environmental

conditions within a single level of human behavioural potential (Deacon H. J.

1974).

The upper Pleistocene cultural sequence is not well understood in the

Southern Cape. Goodwin (1930: 565, 1933: 517) described it as comprising

three content units, the Mossel Bay, Still Bay and Howieson’s Poort cultures,

but he noted difficulties in precise definition of these variants or variations.

The excavation of new sequences and the re-examination of previously exca-

vated sequences is under way but many problems remain (Klein 1969: 134).

The Boomplaas sequence offers the potential for the study and possible resolu-

tion of some of the problems of the Upper Pleistocene cultural sequence and

this indeed is one of the more important aims of the research programme.

An indication of the value of the site in this regard is given by the recognition

of a circular stone hearth feature in a.dark ashy loam excavated as AF1/LP

and dated to 21 100 + 420 (UW-300). This represents a short term occupation

horizon, one of several surfaces on which cultural and faunal remains have

accumulated and which are separated by non-cultural deposits. The implication

for the excavator is that a very fine distinction between occupation horizons

can be made. This has not been possible at some other sites, as for example

in the lower part of Kangkara where the slow rate of accumulation of deposits

and intense weathering has resulted in a much condensed sequence.

Interest in the Boomplaas sequence is then from a number of aspects.

What was sought in the initial reconnaissance excavation in the Langkloof

was minimally a cultural sequence extending into the Upper Pleistocene with

good preservation of fauna. There are few known occurrences in the region

that fulfil this condition. The Boomplaas Cave has well preserved faunal

remains throughout and in addition the sequence encompasses much of the

Upper Pleistocene. The stratigraphic separation of different occupation levels

is seen as important in the study of how cultural residues accumulate on and

are distributed over temporary stable surfaces exposed for varying periods

during the build up of a series of cave deposits. Generally described as ‘floors’,

these do not always have the same unit value in terms of occupation.

CONCLUSIONS

This is a preliminary statement on the Langkloof project which is seen

as a long term study. It is estimated that the excavation of the site which is

the present major focus, Boomplaas, Cave will involve fieldwork over the next

15 years. However, concurrently with this excavation, studies will be made of

214 ANNALS OF THE SOUTH AFRICAN MUSEUM

other Holocene and Upper Pleistocene sequences and efforts devoted to the

study of particular problems. An example of problem orientated research —

forming part of the overall project is that directed at the study of Watsonia

and other geophytes which were dietry staples of Holocene gatherers. On the

evidence from Boomplaas Cave this work needs to be extended to include oil

rich seed producing species. The object is to gain an understanding of the

adaptive advantages of the exploitation of certain specific resources and ulti-

mately a more deterministic explanation of aspects of behaviour. While Holo-

cene studies are more advanced and problem orientated research is possible,

within the lesser known Upper Pleistocene time range much more basic descrip-

tive observations are needed.

ACKNOWLEDGEMENTS

The Langkloof archaeological research project has been supported by the

Fund for Scientific Research (University of Stellenbosch) and the Human

Sciences Research Council. The authors acknowledge the interest and

co-operation of A. Botha, J. Deacon, R. G. Klein (University of Chicago),

J. J. N. Lambrechts (Soil Science Department, University of Stellenbosch),

T. O’'C. Maggs (Natal Museum), R. O. Moffett (Botany Department, University

of Stellenbosch), B. Adams, P. P. H. Hamel and F. B. Silberbauer (Archaeology

Department, University of Stellenbosch), V. Scott, M. Wilson (South African

Museum) and the Municipality of Oudtshoorn and their staff at the Cango

Caves.

REFERENCES

Deacon, H. J. 1972. A review of the post-Pleistocene in South Africa.—S. Afr. archaeol. Soc.

Goodwin Series 1: 26-45.

Deacon, H. J. 1974. An archaeological study of the Eastern Cape in the post-Pleistocene

Period. Unpublished Ph.D. thesis, University of Cape Town.

Goopwin, A. J. H. 1930. Chronology of the Mossel Bay Industry.—S. Afr. J. Sci. 27: 562-572.

Goopwin, A. J. H. 1933. The Cape Flats complex.— S. Afr. J. Sci. 30: 515-523.

HENDEY, Q. B. 1974. The late Cenozoic Carnivora of the south-western Cape Province.—

Ann. S. Afr. Mus. 63: 1-369.

KLEIN, R. G. 1969. Problems in the study of the Middle Stone Age of South Africa.—S. Afr.

archaeol. Bull. 25: 127-135.

KLEIN, R. G. 1974. Environment and subsistence of prehistoric man in the Southern Cape

Province, South Africa.— World Archaeol. 5(3): 249-284.

WELLINGTON, J. H. 1955. Southern Africa: A geographical study. 1. Physical Geography.

Cambridge: Cambridge University Press.

EXPLORATION ARCHAEOLOGY OF THE KAOKOVELD AND

SOUTHERN ANGOLA AND POTENTIAL AUSTRALOPITHECINE

SITES IN THE SERRA DA CHELA MASSIF, SOUTHERN ANGOLA

R. J. MASON

Archaeological Research Unit, University of the Witwatersrand, Johannesburg

(With 5 figures, and a note by R. G. Welbourne)

ABSTRACT

The main characteristics of surface archaeology in and near the Kunene River on the

border of Angola and South West Africa are described. No living-floors or organic materials

were observed but stone artefact sites are abundant. The pre-Cenozoic history of the Kunene

may be inferred from tillite deposits which were located.

A potential australopithecine site, mapped by F. Mouta, on the Serra da Chela was

visited and is described.

CONTENTS

PAGE

Mntroductomar met ee a 1S

GiravelSey-waete pee ee ee

StoneyAgelassemblagesi ne ee ld

Dwikayclllites. va ee a ce ts = 218

Potential australopithecine sites on the

Serraida ChelayMassit 4 =. 4 219

INCleREnCCST ae tae g We ee D3

Remarks on the fossil fauna of the Serra

da Chela by R.G. Welbourne . . 223

References SL mee we Be OA

INTRODUCTION

Archaeological exploration is progressively revealing the south-western

side of southern Africa as a potential Early Hominid discovery region. The

discovery of Tertiary land mammals at Elisabethvelder on the South West

African coast was followed by P. V. Tobias’s announcement of the discovery

of a possible Hominid femur fragment in a dolomite fissure at Berg Aukas in

South West Africa. In January 1975 the author was able to investigate a poten-

tial Hominid site at Leba on the Serra da Chela Massif in Angola whose dis-

covery was announced by R. A. Dart in 1950 (Dart 1950).

In 1958 H. Martin kindly offered the author a seat in a vehicle bound for

the Kaokoveld in South West Africa, where he was to accompany an expedition

from the University of Cape Town by E. Simpson. This was accepted with

pleasure, in the knowledge that in 1954 C. K. Brain had discovered an Earlier

Stone Age site near the Kunene River gorge, west of the Marianfluss tributary,

in the arid western portion of the river’s course (Brain 1959) (Fig. 1 Site 1).

The intention was to search the eastern portion of the Kunene Valley

between the Marianfluss and the Ruacana Falls in the hopes of discovering

215

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 215-224, 5 figs.

ANNALS OF THE SOUTH AFRICAN MUSEUM

216

<= +

ake

ST1v4 VNVOVNy

S11V4 NHLOYO.LNO 6

‘ejOsUY UJOYINos ‘P[sAoyory ‘vore suouny jo dew °] ‘317

/\ (F

Cal ‘ b . >| senewonod a5 onde 0 \

Al e A le 2 iw ) \ (uoiyeuejdxa 104 3x9} as)

j Se eae PN IN li—1 Salis

¥ — Ca Se CGKe si \

yi. ea a ee, — 2 ES

* OG 7] ay 3 |

= d1aAOxOVy

® * mM =) fh |

\ aN (& [LOSE ae oa s |

¥ \ ee ac Nee \ \

“ — { —) <i } :

t Tey \ (

®\ She =) - ;

Ns 4"

ns eC vwasvo ol AR an"

Si sy 49 ae

oe Nv390

STWwa Winds vv ‘SLW andouaa f

‘SLW IGNITVWIHO3L Oa A iY SILNV LV

a c /

a bed /

; |

~m 7

VIOODNV NYSHLNOS

LD) %

SNES

EXPLORATION ARCHAEOLOGY OF THE KAOKOVELD AND SOUTHERN ANGOLA 217

potential Early Hominid localities of the living-floor variety, preserving both

organic materials and artefacts. The author was able to study sites on the

Kaokoveld bank between C. K. Brain’s Kunene Gorge site and the Ruacana

Falls, and for a month was accompanied mainly by D. Toerien.

In January 1975, through the kind offices of D. J. Malan, Prospecting

Manager, Geology Department of the Johannesburg Consolidated Investment

Company, J. de Villiers and P. Cochran and accompanied by N. Rosendal,

the author walked from Iona to Otjimborongbongo on the Kunene River on

the Angola bank and was able to confirm the essential features of the alluvial

sequence of the Kunene River previously observed in 1958 by identifying three

separate profiles of the river at Sites 3, 4 and 5, indicated on Figure 1. B. Huntley,

Chief Game Warden of Angola, kindly drove the party to sites in western

Iona where Iron Age observations were recorded (Fig. 1 Site 16).

At no site in the Kunene Valley or in the territories visited north or south

of the river did the author observe a single Stone Age living-floor or Stone Age

site with organic materials, although a lens of Earlier Stone Age artefacts near

Cambino (Site 15 on Fig. 1) may be the surface of an intact Earlier Stone Age

camp. At Site 16 on the flats at the eastern edge of the Berorue Mountains

Iron Age floors with bone and charcoal were observed, possibly ancestral to

the present-day Himba now concentrated well to the east of the Berorue. The

author was, however, able to make the following observations in the Stone Age

context (see Fig. 2).

GRAVELS

Well-rounded gravels, presumably of alluvial origin, accumulated in lenses

nearly 2 metres thick at localities tributary to the Kunene, such as Site 17

(north of Ehomba Mountain) which are not forming today. Gravel deposits

at localities such as Site 17 apparently accumulated during the Acheulean and

Sangoan occupation of the area. Site 17 may represent a period of well-rounded

gravel accumulated in the environmental history of south-western Africa

similar to the deposition period represented at the Nakop River on the southern

edge of South West Africa discovered by Brain in 1954 (Brain & Mason 1955).

The Kunene River today is transporting gravels in the arid environment

investigated, but these gravels may well be partly derived from presumably

Dwyka tillites abundantly represented in the Kunene Valley. The discharge

for the transport of such materials is derived from the Highlands of Angola

where present-day rainfall is of the order of 50 cm annually.

STONE AGE ASSEMBLAGES

The presence of substantial quantities of Acheulean artefact assemblages,

possibly representing both Earlier and Later stages of the Acheulean develop-

ment are located at Sites 1, 7A and 7B, 15 and 17, showing that the length of

the Kunene Valley and adjacent territory was extensively used as a source of

218 ANNALS OF THE SOUTH AFRICAN MUSEUM

POTENTIAL AUSTRALOPITHECINE to tegr

BRECCIAS - SERRA DA CHELA

ANGOLA

PZZASERRA DA CHELA

Ze SCARPHENT

1 UPPER DOLOMITE

MIDDLE QUARTZITE | FORMATION

AND OTHER ROCKS

SCALE

5000

METRES

Traced from provisional sheet Folha N° 335 SUL _D-33

ean Geologica De Angola—Humpata Cainde T1

Fig. 2. Map of Serra da Chela potential australopithecine sites.

food and raw materials by Acheulean Man. Site 17 is the only Sangoan locality

that was observed.

In addition to Earlier Stone Age data, Middle Stone Age surface assem-

blages, ranging from assemblages resembling the Cave of Hearths Bed 4 assem-

blage, through to assemblages resembling the Cave of Hearths Bed 6-9

assemblage, especially at Sites 3 and 6, suggesting the occupation or use of

the Kunene Valley by hunter-gatherer communities technologically related

to the Cave of Hearths Middle Stone Age people. Later Stone Age assemblages

resembling both Orange Free State Earlier and Later Smithfield assemblages

are present, but not in substantial quantities. Later Hominid prehistory may

be recorded in the numerous possible cave sites observed from aircraft north

of Espiniera, near Iona (Fig. 1).

DWYKA TILLITES

The most spectacular feature of the history of the Kunene Valley is demon-

strated by horizontal, or nearly horizontal, surfaces of Dwyka tillite, frequently

clad by calcrete beds, exposed as vast sheets surfacing the tillite which plugs

EXPLORATION ARCHAEOLOGY OF THE KAOKOVELD AND SOUTHERN ANGOLA 219

the Kunene Valley from at least the Ruacana Falls downstream to Otjimbo-

rongbongo. The Kunene tillites have been discussed by H. Martin (1953) (Sites

3 to 14). Accurate measurements of the heights of these surfaces above the

present Kunene could not be made, but the author’s estimates suggest that

substantial horizontal surfaces are preserved from heights approximately

50 metres above the Kunene. The finest exposures of these surfaces may be

seen in the approximate 30 km length of the Kunene downstream from the

Ruacana Falls.

The clearest relationship of the Dwyka tillite to the earlier rocks of the

Kunene Valley was observed at Otjimborongbongo where the contact between

the Dwyka tillite filling and the older rocks of the Kunene Valley was clearly

presented.

The Kunene Dwyka tillite is capped by immense calcrete sheets up to

4 metres thick, which also overlie other deposits in or near the Kunene Valley.

The calcrete sheets may yield environmental data if a suitable chronometric

dating technique could be developed.

The Kunene River therefore provides a parallel to the Vaal River as a

stream course owing its origin in part to Dwyka Glaciation, or even earlier

erosion (King 1951: 233). The glacial origin of both the Vaal and the Kunene

valleys is clearly evident in the fillings of tillite observable in both valleys,

but best preserved in the more arid environment of the Kunene River.

The Kunene tillites provide the main evidence for Dwyka glaciation in

the valley, but, in addition, landscape morphology bears traces of glacial erosion.

From the north summit of Otjihipa (approximately 1 500 metres) (Fig. 1 south-

east of Site 2), the author was able to observe remarkable mature hanging valleys

about 700 metres above the Kunene, on the south slopes of Cafema Mountain

presumably representing the truncation of pre-Dwyka valleys by the main

thrust of Dwyka movement supposedly in the direction of the Kunene flow

today.

POTENTIAL AUSTRALOPITHECINE SITES ON THE

SERRA DA CHELA MASSIF

In 1950 Raymond Dart published a short note on breccia deposits near

Humpata in Angola on the Serra da Chela Plateau reported to him by Merlin

W. Ennis (Dart 1950). Late in January 1975 the author was able to visit Leba,

one of the breccia sites on the Serra da Chela Escarpment published by F. Mouta

(1954) (Section and Plan, Figs 4—5) and is particularly grateful to P. Cochran

for making arrangements for this visit.

In 1954 F. Mouta, quoting C. Arambourg, noted the presence of the

following mammals at the Serra da Chela dolomite quarries located at Leba.

Nandinhba and Kangalongue (Fig. 3):

1. Machairodont

2. Taurotragus cf. oryx.

220 ANNALS OF THE SOUTH AFRICAN MUSEUM

LEBA BRECCIA QUARRY

oO Mined breccia

: dump

+> January 1975 reserve

of breccia

Edge of

ancient fissure

Mined breccia dump

approx.

metres

Fig. 3. Plan of Leba Quarry.

EXPLORATION ARCHAEOLOGY OF THE KAOKOVELD AND SOUTHERN ANGOLA 22]

3. Cricetomys cf. gambianus gambianus.

4. Dinopithecus ingens

5. Papio sp.

Arambourg identified the fauna as ‘pertaining to Early Quaternary’

(Mouta 1954). Very similar forms are now known to occur with australopi-

thecine remains at Sterkfontein, Swartkrans, Kromdraai, Makapan Limeworks,

Taungs and Olduvai Bed I. (See note by R. Welbourne.)

The position of the breccias in the Serra da Chela dolomites was first

established by F. Mouta in the late 1940s. Mouta located breccia deposits in the

dolomites which are exposed resting directly on quartzite sandstones on the

Serra da Chela Plateau analogous to the superposition of Dolomite on Black

Reef Quartzite in the Transvaal. In January 1975 the writer visited the Leba

quarry and drew the sketch plan and section illustrated in Figures 4 and 5.

The Leba fissure is simply the remains of a deep crack developed along

the joints of the dolomite, possibly early in the Cenozoic. Tertiary erosion

removed the surface dolomite to expose narrow cracks at sites such as Leba.

Deep cracks in a game-rich area may have acted as animal traps, which may

explain the presence of the bones of a possible large elephant reported to the

writer by the quarry manager in January. The Leba quarry could therefore

have functioned for many hundreds of thousands of years as a natural game

trap. The presence of a fauna elsewhere in south and east Africa associated

LEBA NORTH QUARRY

Profile AB

———_ EDGE OF ANCIENT FISSURE

IN SITU DOLOMITE

COLLAPSED DOLOMITE

SURFACE RUBBLE

1;8

METRES BRECCIA WITH HORIZONTAL

BEDDING

TRAVERTINE

ZAUH ae

Fig. 4. Profile AB, Leba North quarry.

DO? ANNALS OF THE SOUTH AFRICAN MUSEUM

with australopithecine hominids suggests that a thorough examination of the

Serra da Chela dolomite breccias would be eminently worth while.

Plans and sections of the Leba dolomitic fissures are easily observed as a

result of extensive travertine quarrying which is continuing in 1975. The general

appearance of the Leba dolomitic fissure environment is similar to the lime-

stone caves environment in the Makapan Valley. The Leba travertines have

been mined along a crack in the dolomite several hundred metres in length.

At least several hundred tons of breccia are available to scientific research in

dumps (Fig. 5). The breccia has been dumped on the side of the quarry in

exactly the same way as the breccia was dumped outside the Makapansgat

Limeworks quarry. A surface examination of the breccia disclosed a few frag-

ments of bone but it is clear that the proportion of bone to breccia is very low.

No substantial accumulations of bone were observed in the breccias, both

in situ and in the rubble heaps.

KUNENE PROFILES

SITE 3

SITE 10

SITE 7A

SITE 12

SITE 7B oy

————

SITE 8

WS river Ss

ES tint

we SITE 14

ed calcrete om

gravels

approx. vertical scale

0 30 60

metres

Fig. 5. Kunene River profiles.

EXPLORATION ARCHAEOLOGY OF THE KAOKOVELD AND SOUTHERN ANGOLA 223

Time did not permit a visit to the other potential australopithecine breccias

at Nandinhba and Kangalongue or Techinvinguiro. Nevertheless sufficient

information is available both in the literature and from informants in near-by

Sa da Bandera to indicate that the dolomitic caverns and fissures of the Serra

da Chela plateau offer a rich potential hunting-ground both for early and late

hominid prehistory.

REFERENCES

BRAIN, C. K. 1959. The Kunene Gorge.— Bull. Tvl. Mus. 3: 425.

BRAIN, C. K. & MAson, R. J. 1955. A Later African Chelles-Acheul site near Nakop, South

Kalahari.—S. Afr. archaeol. Bull. 10: 22-25.

Dart, R. 1950. A note on the Limestone caverns of Leba, near Humpata, Angola.—S. Afr.

archaeol. Bull. 5: 149-151.

Kina, L. C. 1951. South African Scenery. Edinburgh: Oliver & Boyd.

MartIn, H. 1953. Notes on the Dwyka succession and on some pre-Dwyka valleys in South

West Africa.—Trans. Proc. geol. Soc. S. Afr. 51: 37-43.

Moura, F. 1954. Noticia explicativa do Escobo Geologico de Angola 1:2000 000.

REMARKS ON THE FOSSIL FAUNA OF THE SERRA DA CHELA

R. G. WELBOURNE

Department of Archaeology, University of the Witwatersrand

The fauna from Serra da Chela, identified by Arambourg, has been listed

by Mason as follows (see Mason report):

1. Machairodont (extinct sabre-tooth cat).

2. Taurotragus cf. oryx (the modern eland).

3. Cricetomys cf. gambianus gambianus (a living species of rodent widely

distributed in warmer parts of southern Africa).

4. Dinopithecus ingens (an extinct giant baboon).

5. Papio sp. (the modern baboon).

The last identification, of Papio sp., may be in error. Perhaps the correct

genus is the extinct one, Parapapio, which occurs at Sterkfontein and Taungs.

Freedman (1957: 160) remarks that teeth of these two genera are

indistinguishable.

Machairodont fossils have been recovered from Makapan Limeworks

(Ewer 1956), Sterkfontein, Swartkrans and Kromdraai (Ewer 1955), and from

Olduvai Beds I and II (Leakey 1965).

Dinopithecus ingens was described from the type site, Schurweberg, Trans-

vaal, by Broom in 1937. Swartkrans has yielded 37 specimens (Freedman 1957).

224 ANNALS OF THE SOUTH AFRICAN MUSEUM

REFERENCES

Ewer, R. F. 1955. The fossil carnivores of the Transvaal caves: machairodontinae. — Proc.

Zool. Soc. Lond. 125: 587-615.

Ewer, R. F. 1956. Some fossil carnivores of the Makapansgat Valley. —Palaeont. afr. 4: 57-67.

FREEDMAN, L. 1957. The fossil cercopithecoidea of South Africa.—Ann. Trans. Mus. 23:

121-262.

Leakey, L. S. B. 1965. Olduvai Gorge 1951-61 1. Cambridge: Cambridge University Press.

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS IN

EASTERN AND SOUTHERN AFRICA—A NEW LOOK AT A

SIXTEEN-YEAR-OLD PROBLEM

WALTER WILLIAM BISHOP

Queen Mary College, University of London

ABSTRACT

The historical background to hominid studies in eastern and southern Africa is outlined

together with the contrasting environmental/palaeoenvironmental setting of the two areas.

Parameters operating in favour of hominid studies and those operating in a contrary direction

are analysed for the two areas. Geomorphology is rejected as a basis for dating. Although

the deposits and their contained faunas in the two areas represent very different conditions

of preservation, the contrasting bodies of evidence are complementary. Both must be com-

bined to obtain maximum value from the hominid record of either area. The last word lies

with the fossil hominids.

CONTENTS

PAGE

AMtrOoductiones tye 5 et oh tee oe 225

Elistonicalebackeroundy a) a eee

Chimaticicontusionae seen eee 226

East Africa and South Africa. . . . 227

Conclusione os 6 ee Si Ase, Geo 2 233

Acknowledgements. sara een o>

IRCLEnCNCeSHet ees set ase een 5 8

INTRODUCTION

I am not sure what is implied by the term “Keynote Address’. It appears

to be something between a clarion call to rally the troops and the last trump!

However, there is only need to rally troops when they are retreating in con-

fusion and SASQUA seems to be in good heart. Clearly the last trump on

Quaternary or hominid matters is still far off.

I think it is appropriate and timely to sound a ‘half-time’ keynote by

reviewing critically the position reached in the investigation of Australopithecine-

bearing deposits in eastern and southern Africa. The review will attempt to

highlight differences of setting and of approach and hence in the nature of the

results obtained in the two areas.

HISTORICAL BACKGROUND

The African continent is unique in the field of studies of man’s antiquity.

It has yielded the longest and most complete hominid fossil record and the

longest sequence of archaeological traces of man’s early technological develop-

ment. There is good evidence of occupancy of Africa south of the Sahara by

Proc. sth. Afr. Soc. Quat. Res. 1975.

Ann. S. Afr. Mus. 71, 1976: 225-237.

226 ANNALS OF THE SOUTH AFRICAN MUSEUM

hominoids since early Miocene times, some 18 to 20 million years ago. A

hominid line is generally thought as being separately discernible from about

14 million years ago. There are many long gaps in the early part of the record

but the later phases of the hominid story are remarkably well covered for the

period from about 3,5 million years ago to the present day. Artefact occurrences

ranging back through some 2,5 million years are known and still earlier traces

seem likely to be discovered in the near future. _

It is only 50 years since Raymond Dart’s prophetic publication in February

1925 announced the finding of the Taung child and the birth of the first of the

Australopithecine line. Pioneer investigations during the 1930s at the Transvaal

sites and the more sophisticated post-war phase which continues up to the

present day need no further documentation here. Publications by Dart, Broom,

Robinson, Brain, Tobias and many others tell their own story.

It is important to note that brief though 50 years may seem in scientific

investigation, indeed my own life span is only a few years less than that of the

taxon Australopithecus, the East African research has a much shorter time-base.

It is barely 16 years since Dr Louis and Dr Mary Leakey found the skull of

Australopithecus (Zinjanthropus) boisei at Olduvai Gorge in Tanzania. Only a

single milk molar had been found prior to the discovery of ‘Zinj’ in 1959. Since

his discovery the whole tempo of Quaternary research has quickened percep-

tibly. The boost to funds for research resulted in greater numbers of workers

and in the development of more refined techniques. The mushroom-like growth

can be judged from the numbers of hominid specimens recovered from the

various East African rift valley localities during the last few years:

TANZANIA Olduvai 50 since 1959

Laetolil 1D since 1974

KENYA Baringo 6 since 1966

East Rudolf 125 since 1969

ETHIOPIA Omo 182 since 1968

Afar 50-++ since 1973

A total of about 425 hominid specimens has been recovered in 16 years.

The research has been both interdisciplinary and international and hence is an

ideal subject for review at a SASQUA meeting.

CLIMATIC CONFUSION

By a strange coincidence it is also 16 years since the first moves were

made towards the abandonment of the widespread use in Africa of the notional

system of climatic-stratigraphic pigeon-holes based, in theory, upon four

pluvial and three interpluvial periods. The worst aspect of this ill-founded

sequence had been its misuse for ‘dating’ as a correlative of the familiar but

equally precarious four glacials, with three interglacials of north temperate

areas. Critical studies were carried out by Cooke (1958) from a base in South

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS DOF

Africa and Flint (1959) from a North American north temperate viewpoint.

They showed that convenient though these pigeon-holes might seem, the

evidence for their existence rested on insecure foundations (one could also

say ‘formations’), even in their alleged type areas.

Since that time African Quaternary stratigraphers have restricted them-

selves to the use of normal lithostratigraphic, biostratigraphic and geophysical/

geochemical dating methods for correlation and classification rather than

relying on climatic inference. This has involved a shift away from what has

become accepted practice for Quaternary studies in temperate areas of other

continents. However, it marks a return to what is the standard approach for

most of the pre-Pleistocene geological column.

In addition to turning away from Quaternary pluvials as a panacea for

stratigraphic correlation, recent research has shown that the Pliocene in Africa

was not a time of universal and extreme aridity. As more evidence of hominids

and of their activities becomes available for the Plio-Pleistocene it is important

to emphasize that: ‘Attractive though the idea may have seemed, human

strength of character was neither tempered in the heat of a Pliocene drought

nor quenched in the Pleistocene pluvials’ (Bishop 1976).

EAST AFRICA AND SOUTH AFRICA

Following the general background considerations of the preceding two

sections it remains to compare, and where necessary contrast, the geological

setting of the principal Australopithecine-bearing deposits in eastern and

southern Africa.

EAST AFRICA

The Gregory rift from the Omo Valley of Ethiopia southwards to the

vicinity of Olduvai Gorge in Tanzania contains unique sequences of fossil

mammal and hominid-bearing strata in the Pliocene and Pleistocene time range.

Work has recently been extended further northwards along the Ethiopian

section of the rift to the Afar region. Several aspects of the region may be

isolated as favouring, or acting ‘for’, the preservation and study of the hominid

material.

‘For’

Law of superposition. By application of the basic law of superposition, the

relative positions of hominid fossils within rock sequences, and by inference

through time, have been established within numerous local, well-exposed

‘layer cake’ successions in the rift valley. Even if isotopic dates are not available,

relative dating is possible using stratigraphic mapping combined with careful

collecting.

Geological mapping and strato-geometry. From a basis in geological mapping,

correlations have been made, aided by air photography and satellite imagery

228 ANNALS OF THE SOUTH AFRICAN MUSEUM

and controlled by ground survey, between adjacent stratigraphic sections.

The mapping has been aided by the presence of good marker horizons, in the

form of time synchronous tephra or deposits derived from tephra without

appreciable time lapse. (For definition of ‘tephra’ see Thorarinsson 1954, 1969;

Westgate 1974). Throughout the Rudolf Basin, but especially in the Shungura

Formation of the Omo Valley, at Olduvai Gorge and to a lesser extent in the

Baringo Basin, mapping has been assisted by the presence of such ‘tuffaceous’

marker beds even when the source volcano lies outside the area of study.

Potassium—argon dating. In East Africa many lavas, together with sediments

from predominantly volcanic sources and including some tephra, are potassium

rich and have been dated by potassium*—argon*® decay methods. Olduvai

Gorge witnessed the first application of *“K—*°Ar to the dating of rocks as

young as Pleistocene by Evernden & Curtis (1961), in the Berkeley dating

laboratory of the University of California. These dates were fortunately keyed

into the strato-geometry of Olduvai mapped by Hay. Similar work has now

been carried out in other areas of the rift. For example in the Baringo district

detailed mapping was carried out for the Kenya Government by research teams

directed by Bishop and King, based at Bedford College, University of London,

for which dating was undertaken at the Cambridge University laboratory by

Miller and the Institute of Geological Sciences by Snelling. When the combined

results of such mapping throughout the Gregory Rift are integrated with

numerous isotopic dates obtained from several laboratories a most formidable

calibrated Plio-Pleistocene succession will be established.

Palaeomagnetic polarity studies and the development of a Plio-Pleistocene

‘palaeomagnetic polarity time-scale’ (Cox 1969; Cox et al. 1963; Watkins 1972).

The work of Brock at East Rudolf, of Brock and Cox at Olduvai Gorge and

Brown at Omo has already yielded consistent results (Brock & Isaac 1974;

Brown & Shuey 1976; Brown 1976). These provide cross-checks for the

potassium—argon data. In the Baringo area a thorough sampling for palaeo-

magnetic polarity determination has recently been carried out by a Liverpool

University team of the long section of over 150 lava flow units from circa 14 m.y.

to late Pleistocene (Dagley 1976).

Fission track dates. These have been little used in the Gregory Rift area although

early work at Olduvai was undertaken by Fleischer (Fleischer et al. 1965).

There is a tremendous potential for cross-checking potassium—argon dates

(particularly on sanidines from pumices) against glass shards (representing

fractured bubble walls) and on zircon crystals from the same tephra source.

Such material has yielded consistent sequences of fission-track ages from New

Zealand particularly in the range from 2,0 to 0,2 m.y. (Seward 1973).

Comparative studies of recent environments and processes. The patch-work of

rift environments is worthy of mention as a further advantage. The whole rift

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS 229

valley graben has been for 16 or 17 m.y. a ‘sump’ region into which sediments

have been decanted, including such odd-shaped sedimentary particles as fossils

and, towards the end of the period, artefacts, as an integral part of the sediment.

Phases of recurrent fault movement or the initiation of grid faults have exposed

the sediments to erosion and scientific observation. Volcanic activity occurred

intermittently but persistently—sealing sedimentary basins, diverting drainage

lines and forming fresh catchments for water and sediment. Volcanicity con-

tributed burial materials in the form of tuffs (both primary fall-out and secon-

dary derivatives) that provided environments with a chemistry favouring the

preservation and mineralization of bone. Similar deposits are still being formed,

as for example the calcium carbonate rich tephra from the volcano Ol Doinyo

Lengai in Tanzania.

This similarity between present-day and past volcanic products underlines

the important function of the rift as an open air ‘laboratory’. No study illus-

trates this better than taphonomy, although much remains to be done in this

developing field. However, some “burial laws’ are already beginning to emerge,

thanks to the studies of Behrensmeyer (1976) and Hill (Hill & Walker 1972),

in both ancient and modern rift environments.

‘Against’

There are, however, some factors that are unfavourable to hominid studies:

Limited sample size. The long sequences of superimposed fossil-bearing units

in the rift are advantageous for the dating of hominid material but they have

one major drawback. No large sample of hominid remains referable to one

genus or one species has been found in, or on, a single stratum. Thus inferred

variations within a taxon must be traced through time without the palaeontolo-

gist being able to study variability in a sample even approximating to a popula-

tion from a single time-plane. Frequently only scanty and disarticulated remains

are found, although they can sometimes be assigned to one individual.

Facies faunas and life and death assemblages. Fossil-bearing formations, members

and beds can be traced over 50 to 60 kilometres (as for instance in the Omo

Valley and at East Rudolf) with the aid of reliable marker horizons. However,

change of facies may result in the sampling of ‘facies faunas’ representing

different local environments of broadly equivalent age. Taphonomy (study of

Burial Laws) plays an important role in helping to interpret such deposits.

It is also valuable in showing the extent to which mammalian fossils, including

hominids, have become ‘bedfellows’ only after death by being transported into

one sedimentary environment although representing animals favouring different

ecologies during life.

The differences in composition between the assemblages of fossil hominids

from Omo and East Rudolf suggest different taphonomic selection in the

two areas:

230 ANNALS OF THE SOUTH AFRICAN MUSEUM

Numbers of Hominid fossil specimens

Crania, maxilla,

Post mandible Isolated

Cranial (whole or part) teeth Total

OMO

(since 1967). . 7 13 162 182

EAST RUDOLF

Gince 1969) ae 50 SS) 16 125

Some of the contrasts may reflect different methods of collection, but

there seems little doubt that the principal differences result from different

conditions of preservation.

SOUTH AFRICA

The South African ape-man bearing cave breccias also show factors that

may be considered as acting ‘for’ or ‘against’ the scientific value of the hominid

material.

‘For’

Accumulation. Once caves in the dolomitic limestone are in communication

with the surface they provide repositories where under the influence of gravity,

water, mammalian and human agency, bones of animals that lived and died

in the vicinity are accumulated.

Chemical environment. A limestone cave favours the preservation of bone

structure. Some distortion may occur owing to compaction or through roof

falls, but infiltration of calcium carbonate into the breccia matrix, and the

formation of travertine layers assist rapid consolidation favouring the survival

of bone and tooth.

Comparative studies. The existence in the same area of similar modern caves

permits studies to be undertaken of the various processes involved in sediment

or bone accumulation and mineralization.

Sites. The caves represent localized accumulations of bone from which large

numbers of hominid remains have been recovered, e.g. over 600 specimens

from Swartkrans represent at least 85 individuals (C. K. Brain, pers. comm.).

Unfortunately, the time over which this number of specimens accumulated is

not known. Nevertheless, it may be argued that once a cave is open to the

surface and particularly if the cave mouth or shaft opening is attractive as a

lair or habitation, infill is probably rapid. The fill includes biological and other

material which falls or is washed in, roof fall, travertine and interstitial car-

bonate, together with animal remains and other items brought in. Some phases

of infill have been recognized from changes in the sedimentation pattern.

However, such accumulations are probably the nearest that a palaeontologist

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS 23

studying Plio-Pleistocene hominids approaches to having a large sample from

a ‘population’ as they possibly span only limited periods of time. It is necessary

to add a further caveat that the possibility of secondary solution and later

infill cannot be ignored in a limestone area.

‘Against’

Factors that are unfavourable to hominid studies are:

Cave bias. A cave fill may sample only a biased representation of the total

biomass that lived in the cave area. Bias may be in the direction of selective

eating and collecting habits or of preferences for a ‘life style’ in or near caves.

However, ‘pit fall’ mechanisms probably serve to redress some of this otherwise

selective sampling—even here, however, there is probably a trend towards

natural selection of a sample biased toward the halt, sick, lame or just plain

unwaty.

Stratigraphic isolation. The former caves are now represented by masses of

bone-bearing breccia that are stratigraphically isolated entities. They are

volumes, rather than layers, of rock separated by virgin dolomite or by erosion

from any surviving remnants of an original system of communicating caves or

related deposits. The infills of individual caves have been analysed most elegantly

by sedimentological and stratigraphic investigations (Brain 1958, 1976; Butzer

1974) and I am happy to note that Formations are being mapped within the

caves as recommended in Bishop & Clark (1967), as outlined in papers read by

Brain and Partridge at this conference.

Temporal isolation. The caves are isolated in time as well as in space. As yet

no isotopic dating technique or other dating method independent of the fossils

themselves has been successful despite persistent efforts to explore every

possibility.

Correlation. Correlation between the individual cave sequences and faunas is

therefore dependent upon comparison of fossil assemblages typical of the cave

environments of accumulation. This involves elimination of purely local

ecological differences and use of ‘stage-of-evolution’ as the criterion for estab-

lishing relative age. Some ‘second-hand’ dates have been derived by correlating

the cave assemblages with other mammalian sequences already calibrated

isotopically. This becomes particularly difficult when it involves extrapolation

across distances of 2 500 kilometres or more (e.g. from Sterkfontein to Olduvai

Gorge), from the southern part of a continental mass (latitude 26°S) to a more

central situation almost on the equator (latitude 3°S). There are additionally

the problems referred to above of relating the South African cave assemblages

to East African open plain, lake shore or riverside ‘mixed’ assemblages.

Dating and correlation. It is important to differentiate with care between these

two terms. Dating should be retained for the establishment of dates, measured

D3? ANNALS OF THE SOUTH AFRICAN MUSEUM

in potassium—argon decay years or radiocarbon years. Correlation may be

affected by use of lithological sequence, biological content, palaeomagnetic

characteristics or by comparing isotopic dates established by similar methods.

Correlation may yield ‘second-hand dates’ but these should not then be re-used

for further dating.

An alleged geomorphological dating method. Recently Partridge (1973) attempted

to use a geomorphological approach to derive dates for the opening of four

South African former caves that have yielded hominids. This has not been

referred to above as it is not a viable ‘dating’ method. The three major assump-

tions on which the method is based all have inherent problems:

ASSUMPTION I requires that a geomorphological datum, the African erosion

cycle of L. C. King (1962), is a well-established morphological feature through-

out southern Africa, occurs and is recognizable near each of the hominid-

bearing localities, can be dated by extrapolation from its time of inception at

the coast (in this case about 100 m.y. ago), was developed at a relatively constant

rate, and has remained virtually unmodified since its initial planation.

ASSUMPTION II concerns migration of nickpoints related to L. C. King’s

Post African I cycle (dated as leaving the coast during Burdigalian times—

about 20 m.y. ago). De Swardt (1974) criticizes the validity of this approach.

Partridge assumes that headward erosion along three river systems (two major

tributaries of the Limpopo flowing to an Indian Ocean base level and a tribu-

tary of the Orange River flowing to the South Atlantic) progressed at a suffi-

ciently constant rate during the Post African I cycle to permit comparisons

between the different rivers. The length of each section of river downstream

from a cave situation is compared with total stream length in order to derive a

measure of time elapsed.

ASSUMPTION III involves the establishment of time lapsed in widening the

valley to produce cavern opening in the period since the nickpoint was opposite

the cave site. Partridge lists further assumptions involving rectilinear valley

flanks, down-cutting of the channel and regular retreat of rectilinear valley

flanks. However, it is not certain that all the caves opened as a result of valley

wall retreat as opposed to widening of joints communicating with the surface

of an interfluve.

Assumption I calls for a number of subjective judgements but, apart from

noting the critical comments of B. C. King (1958), Holmes (1965) De Swardt

(1974) and Bishop (1966) these are not considered here. The problems raised

by Assumptions II and III are sufficient to invalidate the method. Recent

studies of modern erosion rates for rivers and slopes (e.g. Douglas (1969),

Gibbs (1967), Young (1969), and recorded in various textbooks) make clear

the variables involved. Widely different rates are obtained, depending upon

the interplay of climate, aspect, slope and substrate.

Past erosion rates as recorded in the geological column are more difficult

to quantify. However, attempts have been made which provide “guestimates’

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS 233

ranging from 0,1 to 100 m?/km?/year (e.g. Clark & Jager (1969), Ruxton &

McDougall (1967) and Dury (1972) cite further references to relevant work).

Other variables include change in ground cover while a slight change in

climate in a semi-arid area grossly changes the sediment yield (Douglas 1967).

Similar effects may be induced through removal of vegetation, however this is

brought about. Variations in lithology, even those which are very local and

not recorded in Partridge’s assignment of his river profiles to four broad divisions

of substrate, can greatly alter rates of headward erosion.

In brief, the variation observed in the parameters controlling the migration

of nickpoints and valley flanks is so great that the geomorphologically derived

dates cited by Partridge have no firm basis. One would welcome a dated sequence

of South African hominid-bearing caves to set alongside the East African

hominid chronostratigraphy. However, this attempt at geomorphological

dating seems to be a case of the wish becoming father to the thought.

It is particularly unfortunate that in two recent papers P. V. Tobias accepts

Partridge’s dates as having ‘implications’ for the study of South African hominids

(1973a, 19736). In his Nature paper (1973a) he refers to the dates as only esti-

mates. However, in an abstract prepared for the LIX INQUA Congress (Tobias

1973b), they are quoted as if they are firm and reliable ages. Various other

anthropologists are also beginning to make use of the geomorphic ‘dates’ as

if they were reliably based.

However, it is important to note that although Partridge has described an

interesting theoretical geomorphological model it has not found acceptance

from geomorphologists as a working model that can be applied in the field.

Stage-of-evolution data. Correlations based on stage-of-evolution of the fossil

mammal faunas may help to assign what Tobias (1973a) calls ‘purely relative

faunal dates’. This has been the basis of correlations suggested by Maglio

(1972) and Cooke & Maglio (1972). Although hominids are considered as part

of the total assemblages used in such correlations, the suggestion by Tobias

that the hominid material could, or should, be used to derive even a rough age

seems to involve a circular argument. The fossil hominids are placed in the

untenable (for living hominids) joint role of ‘judge’ of their own antiquity and

‘accused’ whose age is subject of investigation.

CONCLUSION

The hominid collections from East and South African localities come

from very different palaeoenvironmental settings. The East African area is

one of extensive fossiliferous formations, whereas in South Africa one has

localized fossil ‘sites’. Nevertheless both regions make valuable contributions

to hominid studies in different ways.

The South African caves yield large samples of hominids from deposits

having a comparatively small vertical thickness. It seems probable that the

main hominid-bearing strata within these cave-fills span comparatively short

234 ANNALS OF THE SOUTH AFRICAN MUSEUM

periods of time. The fossils certainly cluster in particular lithological sub-units

of the breccias. The fossiliferous strata are frequently separated from the

original opening of the cave by deposits, suggesting variable and in some cases

possibly long periods, represented by sediments barren of fossils.

In contrast the East African horizons provide only limited numbers of

hominids from each locality although these are firmly localized in respect of

environment, stratigraphic setting and age of deposition. The South African

accumulations are probably the nearest that a palaeontologist studying Plio—

Pleistocene hominids approaches to having a large sample from a ‘population’

spanning only a limited period of time.

The East African hominids are related to a well-established, stratigraphi-

cally controlled chronology and yield unique information concerning the time

range of possible evolutionary lineages. Conversely the South African cave

fills and their contained hominids must still be considered as undated.

Workers in each area need to use evidence from the other if the full poten-

tial of both is to be realized. With this in mind, it seems appropriate to close

this review by allowing an early hominid representative from each area to express

his, or is it her, views on the subject and to have the last word. To obtain the

best effect a robust tenor and a gracile soprano should render the following

duet to Offenbach’s music for “The Bold Gendarmes’.

AUSTRALOPITHECINE ALLIANCE

We are two primates bold yet wary

Who hide a crest beneath our hair.

To go out hunting we are chary

For we haven’t a thing to wear.

Then to our living floor returning

Upon a transient Olduvai shore

Zinjanthropus and Mrs Ples

We lived together side by side

Zinjanthropus and Mrs Ples

And were together stratified.

If some days may lack attraction

We build a pile of stones and earth.

When arranged to our satisfaction

We sit and shake with silent mirth.

For they are sure to prove a problem

To archaeologists in years to come.

Zinjanthropus and Mrs Ples

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS 35

Although our brains are small in size

Zinjanthropus and Mrs Ples

We like to think and theorize.

My maiden name was transvaalensis

But love for ‘Zinj’ just could not fail.

I linked my neat, meat-modified molars

To that seed & root, nutcracking male.

Despite the distance now between us

No rift or pitfall mars life’s whirl

Zinjanthropus and Mrs Ples

I’m her robust East African male.

Zinjanthropus and Mrs Ples

And I’m his gracile South African girl.

And now the toil and chase are quiet

We both have drawn a last long breath.

After a fierce brief life of riot

Nearly two million years of death.

But if you ask us why we’re sleeping

In one enormous Olduvai Bed

Zinjanthropus and Mrs Ples

We're linked in death as well as life

Zinjanthropus and Mrs Ples

For now we're truly man and wife.

ACKNOWLEDGEMENTS

I am grateful to SASQUA for the invitation to attend the Cape Town

meeting and to the British Council who provided funds for travel to South Africa

and made possible the lecture tour to the various universities and cities and the

associated field trips arranged by numerous friends. The South African Museum

kindly provided accommodation and hospitality during my stay in Cape Town

while Margaret Leakey ably supervised all the details of my visit in addition to

acting as organizing secretary for a most successful meeting.

I acknowledge my indebtedness to Bob Brain, Karl Butzer, Basil Cooke,

Oliver Davies, Adam de Swardt, Brett Hendey, David Hobday, Tim Partridge,

Tony Tankard, Phillip Tobias and Martin Williams and many others for helpful

discussion in field and laboratory. Without their help the South African section

236 ANNALS OF THE SOUTH AFRICAN MUSEUM

of the paper would have been very sparse. I benefited from their wide local

knowledge. Any errors that remain arise from my own misinterpretations

following a most exciting but all too brief visit.

REFERENCES

BEHRENSMEYER, A. K. 1976. The habitat of Plio-Pleistocene Hominids in East Africa; tapho-

nomic and micro-stratigraphic evidence. Jn: JoLLy, C. ed. African Hominidae of the Plio-

Pleistocene. London: Butterworths. (In press.)

BisHop, W. W. 1966. Stratigraphical geomorphology. In: Dury, G. H. ed. Essays in geo-

morphology: 139-176. London: Heinemann.

BisHop, W. W. 1976. Pliocene problems relating to human evolution. Jn: IsAac, G. Ll. &

McCown, E. R. eds. Human Origins: Louis Leakey and the East African Evidence. Menlo

Park, California: W. A. Benjamin. (In press.)

BisHop, W. W. & CLARK, J. D. 1967. Background to evolution in Africa. Chicago: Chicago

University Press.

BRAIN, C. K. 1958. The Transvaal Ape-man-bearing cave deposits.—Transv. Mus. Mem. 11:

1-131.

BRAIN, C. K. 1976. Some aspects of the South African Australopithecine sites and their

bone accumulations. In: JoLLy, C. ed. African Hominidae of the Plio-Pleistocene. London:

Butterworths. (In press.)

Brock, A. & Isaac, G. LI. 1974. Palaeomagnetic stratigraphy and chronology of hominid-

bearing sediments east of Lake Rudolf. Kenya.— Nature, Lond. 247: 344-348.

Brown, F. H. 1976. Radiometric dating and palaeomagnetic studies of Omo Group deposits.

In: CoppPENs, Y. et al. eds. Stratigraphy, palaeoecology and evolution in the Lake Rudolf

Basin. Chicago: University of Chicago Press. (In press.)

Brown, F. H. & SHUEY, R. T. 1976. Preliminary magnetostratigraphy of the Lower Omo

Valley, Ethiopia. Jn: Coppens, Y. et al. eds. Stratigraphy, palaeoecology and evolution in

the Lake Rudolf Basin. Chicago: Chicago University Press. (In press.)

BuTzer, K. W. 1974. Palaeoecology of South African Australopithecines: Taung Revisited. —

Curr. Anthrop. 15: 367-426.

CLARK, S. P. & JAGER, E. 1969. Denudation rate on the Alps from geochronological and

heat flow data.—Am. J. Sci. 267: 1153-1160.

Cooke, H. B. S. 1958. Observations relating to Quaternary environments in east and southern

Africa. — Bull. geol. Soc. S. Afr. 20: Annexure 73 pp.

CookE, H. B. S. & MAGLio, V. J. 1972. Plio-Pleistocene stratigraphy in East Africa in relation

to proboscidean and suid evolution. Jn: BisHop, W. W. & MILLER, J. eds. Calibration

of Hominoid Evolution: 303-329. Edinburgh: Scottish Academic Press.

Cox, A. 1969. Geomagnetic reversals. — Science 163: 237-245.

Cox, A., DoELL, R. & DALRYMPLE, G. B. 1963. Geomagnetic polarity epochs and Pleistocene

Geochronometry.— Nature, Lond. 198: 1049-1051.

DaGcLey, P. 1976. Magnetostratigraphy of the area west of Lake Baringo. Jn: BisHop, W. W.

ed. Geological Background to Fossil Man. London: Geological Society. (In press.)

DE Swarpt, A. M. J. 1974. Comments upon Partridge’s paper on Geomorphological dating.

— Nature, Lond. 250: 683.

Douctas, I. 1967. Man, vegetation and the sediment yields of rivers.— Nature, Lond. 215:

925-928.

Douc as, I. 1969. The efficiency of humid tropical denudation systems.—Trans. Inst. Br.

Geogr. 46: 1-16.

Dury, G. H. 1972. Some current trends in geomorphology.— Earth Sci. Rev. 8: 45-72.

EVERNDEN, J. F. & Curtis, G. H. 1961. Age of Bed I, Olduvai Gorge, Tanganyika. — Nature,

Lond. 191: 478-479.

FLEISCHER, R. L., Price, P. B., WALKER, R. M. & LEAKEY, L. S. B. 1965. Fission track dating

of Bed I, Olduvai Gorge. — Science 148: 72-74.

FLINT, R. F. 1959. On the basis of Pleistocene correlation in East Africa.—Geol. Mag. 96:

265-285.

COMPARISON OF AUSTRALOPITHECINE-BEARING DEPOSITS Di

Gisss, R. J. 1967. The geochemistry of the Amazon river system. Part I. The factors that

control the salinity and the composition and concentration of the suspended solids. —

Bull. geol. Soc. Am. 78: 1203-1232.

Hitt, A. P. & WALKER, A. 1972. Procedures in vertebrate taphonomy; notes on a Uganda

Miocene fossil locality.—J. geol. Soc. Lond. 128: 399-406.

Homes, A. 1965. Principles of physical geology. London: Nelson.

Kina, B. C. 1958. The geomorphology of Africa.— Sci. Prog. Lond. 15: 672-681. 16: 97-107.

Kina, L. C. 1962. The morphology of the earth. London: Oliver & Boyd.

MacLtio, V. J. 1972. Vertebrate faunas and chronology of hominid bearing sediments east

of Lake Rudolf, Kenya.— Nature, Lond. 239: 379-385.

PARTRIDGE, T. C. 1973. Geomorphological dating of cave openings at Makapansgat, Sterk-

fontein, Swartkrans and Taung.— Nature, Lond. 246: 75-79.

RuxTON, B. P. & MCDOuUGALL, I. 1967. Denudation rates in northeast Papua from potassium—

argon dating of lavas.—Am. J. Sci. 265: 545-561.

SEWARD, D. 1973. Some aspects of the sedimentology of the Wanganui Basin. Unpublished

Ph.D. thesis, Victoria University of Wellington, New Zealand.

THORARINSSON, S. 1954. The eruption of Hekla 1947-48. Part 2. The tephra fall from Hekla

on March 29th 1947. Reykjavik: University Press.

THORARINSSON, S. 1969. A Pleistocene Ignimbrite in Thorsmork.— Natturufraedingurinn 39:

135-55.

TostAs, P. V. 1973a. Implications of the new age estimates of the early South African homi-

nids.— Nature, Lond. 246: 79-83.

Tosias, P. V. 1973b. New African evidence on hominid phylogeny.— Abstracts, LX INQUA

Congress, Christchurch, New Zealand: 369-370.

WATKINS, N. D. 1972. Review of the development of the Geomagnetic polarity time-scale

and discussion of prospects for its finer definition.— Bull. geol. Soc. Am. 83: 551-574.

WESTGATE, J. 1974. A world bibliography and index of Quaternary tephrochronology.—

LIX INQUA Congress, Christchurch, New Zealand. Printing services, University of Alberta.

Youna, A. 1969. Present rate of land erosion.— Nature, Lond. 224: 851-852.

6. SYSTEMATIC papers must conform with the International code of zoological nomenclature

(particularly Articles 22 and 51).

Names of new taxa, combinations, synonyms, etc., when used for the first time, must be

followed by the appropriate Latin (not English) abbreviation, e.g. gen. nov., sp. nov., comb.

nov., Syn. nov., etc.

An author’s name when cited must follow the name of the taxon without intervening

punctuation and not be abbreviated; if the year is added, a comma must separate author’s

name and year. The author’s name (and date, if cited) must be placed in parentheses if a

species or subspecies is transferred from its original genus. The name of a subsequent user of

a scientific name must be separated from the scientific name by a colon.

Synonymy arrangement should be according to chronology of names, i.e. all published

scientific names by which the species previously has been designated are listed in chronological

order, with all references to that name following in chronological order, e.g.:

Family Nuculanidae

Nuculana (Lembulus) bicuspidata (Gould, 1845)

Figs 14-15A

Nucula (Leda) bicuspidata would 1845: 37.

Leda plicifera A. Adams, 50.

Laeda bicuspidata Hanley, ee. 118, pl. 228 (fig. 73). Sowerby, 1871: pl. 2 (figs 8a—b).

Nucula largillierti Philippi, 1861: 87.

Leda bicuspidata: Nicklés, 1950: 163, fig. 301; 1955: 110. Barnard, 1964: 234, figs 8-9

Note punctuation in the above example:

comma separates author’s name and year

semicolon separates more than one reference by the same author

full stop separates references by different authors

figures of plates are enclosed in parentheses to distinguish them from text-figures

dash, not comma separates consecutive numbers

Synonymy arrangement according to chronology of bibliographic references, whereby

the year is placed in front of each entry, and the synonym repeated in full for each entry, is

not acceptable.

In describing new species, one specimen must be designated as the holotype; other speci-

mens mentioned in the original description are to be designated paratypes; additional material

not regarded as paratypes should be listed separately. The complete data (registration number,

depository, description of specimen, locality, collector, date) of the holotype and paratypes

must be recorded, e.g.:

Holotype

SAM-—A13535 in the South African Museum, Cape Town. Adult female from mid-tide region, King’s Beach,

Port Elizabeth (33.51S, 25.39E), collected by A. Smith, 15 January 1973.

_ Note standard form of writing South African Museum registration numbers, date and geographical positions.

7. SPECIAL HOUSE RULES

Capital initial letters

(a) The Figures, Maps and Tables of the paper when referred to in the text __

e.g. ‘... the Figure depicting C. namacolus...’; *...in C. namacolus(Fig. 10)...’

(b) The prefixes of prefixed surnames in all languages, when used in the text, if not preceded

by initials or full names

e.g. Du Toit but A. L.du Toit; Von Huene but F. von Huene

e.g. Therocephalia, but therocephalian

Punctuation should be loose, omitting all not strictly necessary

Reference to the author should be expressed in the third person

Roman numerals should be converted to arabic, except when forming part of the title of a

book or article, such as

‘Revision of the Crustacea. Part VIII. The Amphipoda.’

Specific name must not stand alone, but be preceded by the generic name or its abbreviation

to initial capital letter, provided the same generic name is used consecutively.

Name of new genus or species is not to be included in the title: it should be included in the

abstract, counter to Recommendation 23 of the Code, to meet the requirements of Bio-

logical Abstracts.

PROCEEDINGS

OF THE

SOUTHERN AFRICAN SOCIETY

FOR

QUATERNARY RESEARCH

i: a ae Tp) ~ eM WY

a wn 2 ae

. ud e) a

— — = FG As, Big

ow itty [0 ane ;

< a <5 | Y,

re We 4

co = an yw

ie = a |

— INSTITUTION NOILNIILSNI NVINOSHLIWS

{ >

ty, oO = moe

Gy, x0 = Pe) Ne

oe gyn - > We

h/ hep : = ENS

7 Z m Ww

wn as ep)

NNOSHIING Saiavad ove BRARI ES SMITHSONIAN

(op) x

=< , < =

= Wy, z N ine

aE YY: O \. oa

o bet SoS EN o

eae ae fe) = WSS 2

> = ays pa

STHSONIAN INSTITUTION NOILALILSNI NVINOSHLIWS

— Ww “ees

z ee

Oy, a Ww

Wily, “4 = ~

We Wh a < ~y

fe cas 4 =I

) = S)

= a 2

INOSHLIINS S3IYVYdE!IT LIBRARIES SMITHSONIAN

Saiuvddl

LIBRARIES

Pg ( Fe

= co

= Be)

~ >

zt xy

i set

wire Hen

1ITHSONIAN INSTITUTION NOILOLILSNI

za XN. (9)

< \ . z

z NA: =

z Se i

e O

= & za

= ana ee

(ap) : a

NNOSHLINS S3JIYVUgIT LIBRARIES

A es

Ww =

ow =

ie 4

= =

ss Oo

‘ J Pag

AITHSONIAN_ INSTITUTION NOILALILSNI

5 - é

Rae)

se a

= -

= =

A z

JINOSHLIWS

Fas

AITHSONIAN

NOILNLILSNI

SMITHSONIAN

INSTITUTION NOILN

LIBRARIES

LILSNI

SMITHSONIAN

NVINOSHLINS S3IYVYEITLIBRARIE

INSTITUTION

SMITHSONIAN

NVINOSHLINS

SMITHSONIAN

NVINOSHLIWS

NOILALILSNI

LIBRARIES SMITHSONIAN

S SaIuvygIt

LIBRARIES SMITHSONIAN

INSTITUTION

INSTITUTION NOILNLILSNI

INSTITUTION

Saiuvyuaiy

NVINOSHLIW

LIBRARIES

SS!YVYUEIT LIBRARIES

sidlauvagi

NOILNLILSNI

INSTITUTION NOILALILSNI

INSTITUTION

SJIYVYGIT LIBR

SMITHSONIAN

NVINOSHLIWS

NOILNLILSNI

NOILALILSNI

LIBRARIES

ARIES

“AS

INSTITUTION NOILONLILSNI

pose ”

= we

= ac

= pipe

S) a

z i

S3'1yvygiy

Zz rc

oO —s

ea w

= AO

im >

E- > 0)

i pis

= Mm

— w

INSTITUTION

* w

Ny F

\ es

=a =

S3JIYVNAG Ete

INSTITUTION NOILNLILSNI

ITHSONIAN

NVINOSHLINS S31YvVUYdIT LIBRARIES

NVIN

Os H. iT

MAS

VAS

X %

SMITHSON

NVINOSHII

SMITHSONI

—_

NVINOSHLIV

SMITHSONI

‘a

VINOSHLIL

COO

=a LANE

aj % . ae

Cc NN |

2 F

LIBRARIES SMITHSO NI

= Yj Jy

= Yin “~Y

ath

= a

NOILALILSNI _ NVINOSH

YY, = Serr

fi, O ‘

‘ =

LIBRARIES SMIT

“)

jem

fe 2 ies

< Us <= Us ot era YW) J eee ae

i7 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLINS SJ3I1YvVYgIN

cored Ww = any 7p)

rr x . & a - Wd

4 =), J a o A fava

) z 2G@iy : : : 3

je) a Yi So) = fe) ze

Zz at z J Fa Sg

IN NOILONLILSNI NVINOSHLINS S3I1YVYgIT LIBRARIES SMITHSONIAN INSTITUTION

= wae z An z a

° w = = © oo

5 = = 2 = >

) 2 29 = a = 5

F = b : me =

Z Do ume a iy es D z D

}1 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWNS S3INVUSIT

= 2 Zz Ks w Zz ” 3

< < = , = =

= = z 4 ie 4

ee aE Oo ae Pee: 28

S fe) Sm fe) Te fe)

a 6 = 3 a =

N NOILALILSNI NVINOSHLIWS LIBRARIES INSTITUTION

ae af = 7) ca wn

ac = a = 2 ae

< e ee e = “eS

fe —_ ce + ia a a

a Oo wes Oo a. =

a Zz at Zz , =

1 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI SAIYVUEIT

- z si = EE Zz

w = a 8.e, = ow °

Be > 2D > Ps) =

| eS i ¥ = E an =

= 5 = b < 5

= = 7) = a . £

N NOILNLILSNI S3SIYVYEIT LIBRARIES SMITHSONIAN INSTITUTION

o z n z ae ” 2

= < = c= @& <= | eo

4 z Aa, he (Z NS =. z KE

r fo) i Dt 9 QR =z SD We

B = SG 2 We 8 2 VQ

3 = zZ “yy Fr wz = .

= = | Ps as

2 a 2 = <i 7

7 LIBRARIES SMITHSONIAN INSTITUTION NOILNLILSNI NVINOSHLIWS Saiuvugity

ane ap) pees Y = wo

a s 6 2 a ud

= ce: = oc = a

ie < = < a <

= a =I = — ==

= co = = rea

Oo zs fe O — oO —

= at = - _ oe eed

Pe NOILNLILSNI_NVINOSHLIWS ~°4 lyYVvug ie LIBRARI ES_ SMITHSONIAN _INSTITUTION

2 : = Sli ales

3 2 = a 5 , o

= = =o as A

w” w ae w os

s o = O

Vu BRA! HSONIAN INSTITUTION NOILOLILSNI NVINOSHLIWS S3 lyYvuyudlt

=e: 2 lhe z wt: < = ,

S 5 = 4 ys S = |

g g z WI ps 3 z

Ee NX = = inher Wie Z.

Be Zee ED 7: 5 :

N NOHLALILSNI_NVINOSHLINS S31YVUSIT LIBRARIES _.INSTITUTION

: £ = @ “ % :

at e we

es feel zou ag RED

= zs 2. < aes

oc 5 ac = a a

= S) a. O a O

= J a at Zz I =

SMITHSONIAN INSTITUTION LIBRARIES

il ll

3 9088 01206 6429 ae

FRE Phd

: Beis ta Fe \ wie, ene . ‘ Aon Ee Rr OR int | Tats : , Fi a

CRETE ee Seetuttag ay Ut : ieee AE. thts coun las : : ’ .

bh ilah >, wldt shay trek, S08 ba are URAL RS ton Har an r OW K ' ‘

. ui Smasees gee De, ’ ae . * . .

ee see het Mr te thes . Hea ER Bea a x (eediee a PAG, ; r i ‘ ;

Spa nseor ¢ eepleia & fede Aen be ar) a THN : Poorest met init:

nom tes tao Shee , ‘ ° ; Be

ie ¢ : FOO EERE RIGS GION ton We bey «obo. bn oh Sl, i nak 1 y FRa

iia 8507) tort lay eet pos vee PVR A, a8 ‘ perce < 4 , « is

ann Sate LCP RTI ACL REAR EFC ar Tove Reet ar date Ane . :

Ver he pestuta ate poe! kisah ; : : : ; = 3

auiapotbler oY . pots eee) BC haan lg oh " pear : : eee

Pow ber lous SN a useless eerent copes trae ' Al : a . '

gies tobias at Cees Nase Ks fe iteea ae hhs n ES a real ‘ : fat > 4

we RSG ACTORS TCE AER Se a MICS nO YD al EU ae ery, Tes aE :

ar VR Nk eG Evy os rit} ‘ Lift ti eT A 3 iy f a f ; i

4 fries eR a ; Fhe e ats bt Li 1 Te aVASN thie Lan aoptes ve ~

; oa? . Pay 5 rah vk OMS ti Coes 5 tat * ‘ ‘ ‘ 4

Petia hoy AURA ceed ety Subs Sta ig edad gi US et) *) . i yA Ce 4 .

PMP See ey, Y Me : 4 ; a

Paty : z . ES ate

: ay ley a ' PS brs

SPE SES OUST ‘ ers te

ee ECR a ‘

€r% ‘ : ; : ' ‘ ¥ ;

ee ta 7a ek : 3 : Dy

Weer eke aly : ve 4 2 ‘

a EA Ae us : . , .

Py sts Wat SRW ; é ait :

Neviaan: Sey eye prery ‘ 3

We kate adm 4: s ca . . : A ‘

ue hie. 3 5 foe ;

nee bine te eis

bop a ea yi 3 : 3 : i

REM Su Le i, Pm :

Ade : ' q . . 7

Bad ble, ins ! i ¥.

B22 ey % ;

Pus eve, t 5 ‘ x es fi