Lo os é

1 iets AROS O NM Mee, hs Bea NG eA Pe BB EM A a a AeA St At A BS PES tay) OB an Se or Rs TS ESAS A UE ee A ae aT a SAY Ans
POR AT IO A | UW ee " 4 ae wa ak 3 : HY aed
yh “es

Athy UE OF ia be
be hopierhy
he *

i hs

£3. eh

ve ree

esa ai
if

ES eetets St Paes Fig

a
> ¥ a
ee ee

y ny fF “
tah)

ae
me nie
yo a
Me

ooh puny

sgh
ae

ese
=

-

Oe pen
woh

See

ea

ao
: Py Yet 755.

rile

Aes van
Pee AT gue x Ba
Pee FN tae vi $

iy aes

*
if

fi
ath 94

:
#

ee

He i . %
Wald y
te
: Pee tue “ eo AA ‘

mA PAN HN, f Cua aaa Mu laek Reis
Cn BAU ORS a) HY, YAR, ee ” 195i ” ae
WE Bek Rt : : MWh: ih a
8, Weed We { SAE BRO. ohoth | ibe ae : ; ‘i Lt
SA Agata aed Z E ; Ok way eh
He WMS

ih LL ERI mem ae OR =

CTA ER Diy
ways

Rue

; ia ‘

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XXXVII

e
a
_s
m
’
iy
=
i
y
)
al
(i
~
‘
'

“)

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

OIG ODI TIA DCR OGE ITE

PRINTED FOR THE
TRUSTEES OF THE SOUTH AFRICAN MUSEUM, CAPE TOWN
BY NEIL AND CO. LTD., 212 CAUSEWAYSIDE, EDINBURGH

1947-1952.

TRUSTEES OF THE SOUTH AFRICAN MUSEUM.

Professor R. 8S. ADAMSON, F.R.S.S.Afr.
The Rev. S. J. Fort.

Professor D. L. ScuHoirz, D.Sc.

C. J. SIBBETT, J.P.

S. H. Sxatre, Ph.D., F.R.S.S8.Afr., J.P.

SCIENTIFIC STAFF OF THE SOUTH AFRICAN
MUSEUM.

KEPPEL Harcourt BaRnaArD, M.A., D.Sc., F.L.S., Director.

ALBERT JOHN HeEsse#, B.Sc., Ph.D., Assistant in Charge of the Entomological
Department.

LizuwE Dirk Boonstra, D.Sc., Assistant in Charge of the Palaeontological
Department.

Miss EK. Marcaret Suaw, B.A., Assistant in Charge of the Ethnological and
Numismatic Department.

Miss G. Joycr Lewis, B.A., Assistant in Charge of the Botanical Department.

A. J. H. Goopwin, M.A., Honorary Keeper of the Archaeological Collections.

STN, cai BME. tet we:

‘
ay
ow
'
4
“a
‘
. ¥
'
'
\
‘
uv

: i ee,
Ae es i et ee

<

on mle, Ne PIA eee mini,
ir : th age. SL

Mo penta glia union neg
— mM ie ‘a

sie Sen ARR aaa

ie . : i a

Ld .
4 :
. Uy
1 7 J
ue
j
pal
Lee
eid
7
'
t
'
j
es
’
UJ y
-

CONTENTS.

A. J. H. Goopwin.
The Bored Stones of South Africa

J. A. KEEN.
Craniometric Survey of the South African Museum Collection of Bushman,
Hottentot and Bush-Hottentot Hybrid Skulls

Plates I-IV.

DATES OF PUBLICATION.

Part 1 AUGUST 1947
Part 2 AUGUST 1952

211

os

Ea z NE iy Live Ww,

nr

i : 4
7
r =
‘
4 5 :
*
i Tae
1
+ x
: = 4
nd 3
s
i
.
i
'
; f
‘ sh
a
¥ :
f f ai
% ay é a
d *
J f ane Bay lex
- hr
‘ 5
/
x t :
; ‘
;
= ara
1
)
}
,
|
i
j
=) s
\
. Pi
1
)
i
‘
1
re 4
‘
i ry
ny q
4
> 6

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XXXVI.

PART I, containing :-—

1. The Bored Stones of South Africa. By A. J. H. Goopwin, M.A.,
F.R.S.S8.Afr., University of Cape Town, and Hon. Keeper of
Ethnology in South African Museum. (With 9 Text-figures
and 16 maps.)

ISSUED AUGUST 1947. PRICE £1, 6s.

PRINTED FOR THE
TRUSTEES OF THE SOUTH AFRICAN MUSEUM, CAPE TOWN
BY NEILL AND CO., LTD.,
212 CAUSEWAYSIDE, EDINBURGH

pee \
Eu) we wel

ANNALS 8. AFRICAN MUSEUM, Vol. xxxvi, Part 2.
Addendum to Barbus holubi: see p. 262.

Corrigenda: Pp. 113, etc., and fig. 6 (map), wherever “holubi”’ occurs,
read aeneus.

Pp. 120, 121, 153, for “‘Holub’s”’ Yellow-fish, read
Burchell’s Yellow-fish.

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XXXVJII.

1.—The Bored Stones of South Africa. By A. J. H. Goopwin, M.A.,
F.R.S.S.Afr., University of Cape Town, and Hon. Keeper of
Ethnology in South African Museum.

(With 9 Text-figures and 16 Maps.)
(MS. received June 1944.)

PART I—GENERAL PRINCIPLES AND METHOD.

Ir is our intention to discuss the prehistoric and protohistoric examples
of the bored stones of the Union, area by area, and type by type. In
undertaking such a task it is essential to remember that the material
of such a survey is useless in itself: it has value only as a source of
legitimate deduction. It is obvious that we are dealing with a great
mass of undigested evidence, of a series of superposed and, in some
cases, completely independent cultural movements, into whose history
we are attempting to gain an insight.

Our evidence is bad. The very abundance of our material has
made us careless in the collection of data. Where stratigraphy has
been present it has been ignored, and in many instances no attempt
has been made to give or to record the exact localities of our finds.

Whatever results we may achieve here can only be regarded as
tentative and incomplete: a framework into which new material
can be built, permitting clearer, more precise deduction. Data are
here limited to the Union, though I had hoped to extend the field
further. The chances of discovery outside the Union are slight, and
only isolated areas have been searched. The distances between
specimen and specimen are too great to permit of legitimate deduction,
and the numbers of examples are too small to be of use at present.
It is most sincerely to be hoped that a similar study will be extended
to the Rhodesias, while the raw material for such a survey in the
Belgian Congo has already been collected and published.

VOL. XXXVII, PART I. 1

SEP 2 0 1947

2 Annals of the South African Museum.

INTRODUCTORY.

Of all the stone implements so abundantly found in South Africa,
the bored stone or “‘!kwe”* has attracted the most widespread
interest. The reasons for this popular interest are not far to seek.
They are the most easily recognizable implements of the Later Stone
Age, and the most widespread. Their symmetry, the trouble taken
in boring the central hole, its obvious artificial origin; the surprise
that so simple a people could have created so artistically simple a
form; and the abundance of material on superficial sites, have all
caused examples to be collected and kept by the finders as door-stops,
manger-weights, or paper-weights. They are frequently sent to
museums with the minimum of information, either singly or in bulk.
No knowledge of provenance, associations or the type of deposit is
available for over half the specimens. In some cases this is the
fault of the collector, in others the museum assistants are to blame,
more especially in the early days when cataloguing and marking were
done by minor assistants with no knowledge of the subject.

The bored stone is sufficiently widespread to need some sort of
explanation as a phenomenon, and numbers of legends have accumu-
lated regarding their use. Many of these are complete fabrications,
and it is essential to note that the Huropean is by no means the worst
offender. Native informants have gladly provided farmers and
archaeologists with invented theories, and quite as happily support
these by personal reminiscences. Some of these legends will be dealt
with as they arise.

As a result of all this, and owing to casual generalisations on the
part of those who should have known better, numerous inaccuracies
have been published. There is scarcely a single adequate description
in any scientific paper, and a thorough resurvey is therefore essential.
It was partly to have some factual basis for future generalisations
concerning size, shape, and distribution that I started collecting data
some twenty years ago. Other intentions were to provide a frame-
work into which local discoveries could be fitted, to discover if any
certain relationship existed between the bored stones and Later
Stone Age cultures, either locally or generally, to discover the limits
of tribal hunting areas, and to plot migrations. It was intended to
provide a general survey of distribution and association. This failed
at the time through lack of exact data accompanying the bored
stones.

* See Peringuey, Annals of the 8. Afr. Mus., vol. viii, 1911, p. 107.

The Bored Stones of South Africa. 3

The present work consists of the digestion of a great quantity of
data, made available to me through the kindness of various museums
and private collectors. The work began in 1923, as a small part of
the vast amount of checking, measuring, comparison, and typological
and distributional study which led me to formulate the existing stone
implement classification. The work entailed has already been
cursorily dealt with elsewhere,* and need hardly be laboured here.

Independent field work by Col. Hardy, Dr. Hewitt, Dr. Heese,
Professor van Riet Lowe, myself and others, has borne out that the
so-called !kwe or !kibi (terms applied by Stow and other writers to
all forms of bored stone) are variously associable in southern Africa
with Smithfield, Wilton, coastal midden, and Bantu sites. The
insistent Bantu tradition in the Transvaal shows that they were used
in earlier days as make-weights for digging sticks (subsequently stated
to have been hoes, on the statement of an old male) by the women
of several tribes, notably the BaTlokwa under the chieftainess
MaNtatatisi. There the matter rested, except for a general accumula-
tion of proof of a somewhat similar nature which added to the list
of names of Transvaal Bantu tribes using bored stones. Engelbrecht
in his study of the Korana states that bored stones were used for
digging by that Hottentot tribe, the only satisfactory reference we |
have to their use by this people,t though the Bushman use of the
bored stone is well and consistently authenticated by early travellers
and explorers.

In 1926 this survey was dropped, after a considerable number of
measurements had been accumulated. When the war broke out, and
military duties hmited my ability to continue work in the field, I was
inspired by Major Oliver H. Myers (a name well known in Egyptian
archaeology) to take up the survey once again. My results are here
published completely for the first time, though a cyclostyled {
Preliminary Survey, in four parts, was published during 1942. In
the Preliminary Survey the known figures were listed, and an initial
attempt was made to formulate some of the theories dealt with more

* A. J. H. Goodwin, Man, 1927, Art. 14; also in Africa, vol. ii, 1929, pp. 174
et seq.

t Sir Langham Dale refers in an early publication to the use of bored stones
as digging-stick stones and as weapons of war by the Hottentots at Wupperthal,
and regrets the absence of further evidence.

t Goodwin, A. J. H., A Preliminary Survey of the Bored Stones of South Africa.
Published by the School of African Studies, the University of Cape Town. Cape
Town, 1942.

4. Annals of the South African Museum.

fully in the present paper. At that time no attempt was made to
discuss the migrations of the prehistoric types. The present publica-
tion therefore includes much material not previously published, and
in addition gives deductions and theories not given elsewhere.

GENERAL PRINCIPLES.

As this is the first full chorological survey attempted in South
Africa, it would be well to outline the general principles of our methods
as an aid to future workers. I here employ a fundamentally statistical
approach, which provides an ethnological tool as certain as that of
geology in the earlier prehistoric field, and of far greater importance
in the later phases of man’s history. Work of this nature should be
limited by certain strict rules. It depends fundamentally upon
deductions or logical judgments, made from available data, while
speculation (probability suggested by an insufficiency of available
data) must not be permitted to affect the main issues of the survey.
It may be cited as a fruitful field for future study.*

1. A survey should confine itself to strictly comparable data, in |
this case the bored stones of South Africa.

2. All available material and exact data should be obtained,
analysed, and employed without suppression. In South Africa only
one important source of material has been debarred from me, and all
available sources have been employed.

3. A fundamental basis for comparison should be sought, and in
the present work a few simple measurements are used.

4. A simple technique for analysis should be employed. Here I
have used the map and the centimetre graph paper. These represent
‘the study of dispersal and the analysis into localized types.

5. While the cultural uses of an element of culture should be
defined, these are not essentials to the validity of the survey. The
obvious failure in the present paper to differentiate absolutely and
precisely between such uses as digging-stones, knob-kerrie heads, etc.,
is admitted, but this does not affect the fundamental facts of
distribution.

6. Extraneous factors should be eliminated. Such points as the
exact petrology and attribution of types of stone are regarded as
unimportant to the survey. '

7. Local environment must be constantly studied, and the deduc-
tions made from distributional statistics measured against this highly

* See A. J. H. Goodwin, ‘“‘ Method in Prehistory,” S. Afr. Arch. Soc.
Handbook I, 1946, chap. xiii.

The Bored Stones of South Africa. 4)

important ethnological factor. The intention here is to deduce the
geographical areas of attraction, migrational routes, and the natural
barriers confining them.

8. The study of attracters, barriers, and migration routes has a
local and a general application. It is local in that it defines our
specific routes and areas for future exact ethnological study, and
general in so far as it augments our knowledge of the fundamental
laws governing man’s movements and concentrations.

9. All reasonable deductions regarding the race, culture, mode of
life, etc., of the users should be employed. Where deduction is not
possible, probabilities may be stated as speculations worthy of
consideration, but must not be permitted to affect the major issues of
the survey. A clear distinction between deduction and speculation
must be observed.

10. Any general theories based upon the material must be
sufficiently inclusive to cover about nine-tenths of the available data,
about one-tenth being permitted for anomalous types and personal
peculiarities.

11. Speculations based upon casual study, even of large numbers
of single instances over a period of time, are extremely dangerous,
misleading, andincorrect. They belong to the realm of popular fancy.

Types and Ethnology.

Though the figures given in this study are divided into three main
groups—Prehistoric, Bantu, and Tlokwa—there are in reality four
series. The first two may be regarded as acceptable. The third is
slightly more speculative; the fourth is highly speculative at present,
but further evidence may provide convincing proof.

1. Prehistoric Series: Going back into the Later Stone Age, and
apparently not preceding that general era. These stones can be
associated with certainty at various sites with one or more of the
following elements: Smithfield B, C, N; Wilton; Coastal Middens;
Pottery; Rock-paintings; Petroglyphs. Apart from definite associa-
tion, they are generally associated, in that they appear with high
frequency from surface sites where those elements occur, or in the
immediate neighbourhood. Association is only specifically mentioned
in cases where it is certain. General association is mentioned in the
prefatory note preceding each area. Hxact association has proved to
be disappointingly rare.

2. Bantu Series: Belonging to the nineteenth century, and perhaps
the eighteenth, and demonstrably in use by a cattle-keeping people

6 Annals of the South African Museum.

(by reason of distribution) who practised agriculture and employed
the typically Mediterranean technique of terracing. There may or
may not be a direct association between the bored stone and proto-
historic gold working, but terracing provides an indubitable link
between the two complexes.

3. The Mantatees * drive into the Upper Molopo-Groenwater area:
This is historical, and the association of bored stones with the Tlokwa
of the period has similar historical backing. A slight element of
speculation enters here, not sufficient to condemn all the specimens,
but perhaps enough to raise doubts as to certain individual allocations
(p. 175).

4, The Mantatees refugee period: After their disastrous battle
with the Griquas near Litako and Kuruman, we know from various
sources that the Tlokwa refugees were scattered throughout the Cape
of Good Hope. We also know that bored stones of the large type
associable elsewhere with the Bantu are scattered over the same area.
Association is highly speculative, and for that reason the examples
are left in their places within the cadre of prehistoric series, and —
are not separated out.

General Typology.

In discussing the distribution of types it behoves us to consider
the value of the distributional method as applicable specifically to
the field of prehistory, employing the present material as an example.
The evidence at our disposal goes to show that the position is a
complicated one. In examining the “Bored Stones” we are not
dealing with a single cultural element, nor are we dealing with a
technique employed by a single cultural group. We are not even
dealing with a single period. To elucidate the position let me augment
the categories given earlier (Prehistoric, Bantu, Tlokwa) by discussing
the bored stone from the viewpoint of probable usage. The
arrangement is by size. “

i. The Bantu digging-stones, used by the women agriculturists
of the Eastern Transvaal, and carried by them through Kurechane,
Litako, and Kuruman, and, presumably, down the Groenwater valley
and across the Cape Province.

il. The prehistoric digging-stone, unassociated with agriculture,

* BaTlokwa and other tribes under the chieftainess MaNtatisi. The form
Makatees is also commonly used at the Cape for these people. “‘Tlokwa” would
only represent the nuclear tribe in this great migration which will be discussed in
detail later (p. 202).

The Bored Stones of South Africa. 7

and used by women food-collectors, of the general “‘Bushman”’
peoples of South Africa. It is vouched for as a Bushman tool, is
frequently depicted in paintings, and was in use until late in the
last century. The terms /kwe and /kibi are often applied to it in
local literature of the period.

iii. The stone-headed club represented in paintings. The Hotten-
tot (7) term kirt (presumably cognate with /kibi) seems to have been
generally applied to this weapon and its wooden derivative by the
early settlers, whence the Afrikaans knopkierte and the English
knobkerry or knobkirt. This weapon may well have been adopted by
the men of any of the South African tribes—Bantu, Hottentot, or
Bushman. It is obviously a man’s weapon. In this group we can
tentatively include the “‘macehead,” generally pyriform and carefully
tooled, deeply bored from one face only.

iv. Stone dagga-pipes for smoking hemp (Cannabis sativa). These
are historically late, having been introduced with hemp by the East
Coast Arabs probably from the Sofala region through Bantu tribes
on the Zambesi, from whence it spread northwards (e.g. to the Ba-
Shilange on the Lulua-Kasai) and southwards to the Sotho-Tswana
people in the Union. We may regard these as masculine, for amongst
most peoples the use of strong narcotics and stimulants is permitted
exclusively (or more generously) to the male.

v. The pebble disc, common only in certain areas, usually within
easy reach of the coast, where flat beach pebbles are easily available.
This is presumably a variant of the digging-stone, a woman’s tool.
The inland discs in contrast are generally of shale or some analogous
material.

vi. The edged disc and stone “‘arm-ring,”’ the former patently a
stage in the making of the latter. The term ‘“‘arm-ring” has been
questioned elsewhere.* If it is an ornament it is presumably male
(see Appendix II).

vu. A further type that may be watched for is the “ring”; a
small stone, with an index between 40 per cent. and 60 per cent., in
which the mouth of the bore exceeds the height of the stone. These
are products of the technique of boring, and obviously reflect the use
of a coarse drill or pecker. Even this slight variation may be cultural.
Mainly confined to the South Coast and the Cape.

vill. ““Stone beads,” bored stones less than 3 cm. in diameter, of

* Goodwin, A. J. H., ‘‘Edged Discs and Armrings,”’ S. Afr. J. Sci., 1943, vol. xl,
pp. 296-302.

8 Annals of the South African Museum.

which relatively few are represented here. These are certainly Bantu
in most instances.

ix. Elongated types. In general we have to presume that the
elongated examples are variants of the circular. There is, as yet,
no means by which we can link the two series with certainty, and here
again we must not lose sight of the possibility that we are dealing
with a separate cultural spread, quite unrelated to some or all of the
above.

Finally it must be remembered that we are dealing with implements
in three stages of manufacture: incompletely bored, completely bored,
and with the finished aperture enlarged by intensive usage. This
last will be dealt with further under the Use of Metal Tools.

Manufacture.

It is first essential to discuss briefly the method by which these
stones were bored and shaped. In general the technique seems to
have been everywhere the same, though variants will be noted.

A suitable pebble was taken, and the piercing process was begun by
pecking a hollow in the two opposing faces. This was done with a
stone pecker, an instrument shaped by use, and found on most rich
sites. In form it consists merely of an obtuse-angled natural or
shaped stone drill, blunted and battered by usage. Numbers of
discards occur at this stage. Some partly bored stones seem to have
been cast aside for no particular reason, others have broken as the
result of too hard a blow.

Once these two cup-shaped hollows met in the centre of the stone,
the resulting aperture would be enlarged by means of a stone reamer
or rimer. A long natural or shaped stone would be used, and would
be inserted and rotated by hand until the hole was large enough.
This rimer eventually developed a phallus-like form as a natural result
of usage (fig. 1).

These two stages (pecking and reaming) provided the simplest form
of bored stone, but in many cases further action was taken to shape
the stone more nearly to the maker’s fancy. In perhaps the majority
of cases the outer surface of the stone has been battered to shape, by
careful pitting over the whole surface by a laborious crumbling process.
This is apparently true only in the Prehistoric Series, and the process
is absent or rare among Bantu examples. It is also possible that
finished stones were sometimes polished over the whole surface, and
specimens with a black smoky polish are fairly frequent, though this

The Bored Stones of ‘South Africa. 9

==> =e

fae
i=

)

Fic. 1.—Rimers, from: 1, a shelter near Wellington (Univ. Stell. Col.); 2, Noord-
hoek kitchen-midden (U.C.T. Col.); 3, Alexandra Hospital, Pinelands;
Stippling shows points of wear. (Hardy Col. U.C.T.)

10 Annals of the South African Museum.

may be the result of grass fires. In most cases I have noted that the
stone has been shaped or pecked, but these are not to be regarded as
unusual. In most cases evidence has not been submitted to me, and
in a number of specimens natural agencies may have produced a
similar effect.
In certain instances (notably, though not exclusively, in the Vaal
area, and related to the rare “‘pyriform”’ or “‘macehead”’ examples)
there is evidence of “‘tooling”’ over the whole surface. Pitting on the
pecked examples is regularly distributed, but on the tooled specimens
there appear to be distinct “chatter marks” or regular lines of pitting,
as though a metal tool had been employed. These lines of chatter
are generally to be observed more or less radially to the mouth of the
bore, covering the whole surface of the stone in much the same way
that chatter marks cover the face of a modern tooled foundation stone.

The Use of Metal Borers.

While there is some reason for the belief that metal borers of some
sort were employed in certain instances by the Bantu—and even
perhaps by late prehistoric peoples—there is no certain proof avail-
able. Two types of bore, the cylindrical and the conical (in contrast
to the usual hour-glass or biconical bore), have been attributed by
various writers to the use of metal borers. In all the specimens I have
examined the “cylindrical bore” is far more likely to have been the
result of hard wear in use, and cannot reasonably be attributed to the
use either of metal, or to the “‘reed and sand’’ method employed in
Neolithic Europe. Here and there the term “‘choke bored” is added
after the description of specimens; this is not my comment, but
refers to stones measured by my correspondents and helpers.

There would seem to be a fairly prevalent belief among Huropeans
that the use of a metal bit should produce a cylindrical hole. In
practice the conical and biconical bores can be more easily obtained
with a triangular metal borer. The cylindrical hole needs a borer of
such a type that it will permit of the escape of the powdered stone as
the hole is drilled. A European “jumper drill” with a cross-shaped
section would serve the purpose. It is far more likely that the Bantu
would employ a common native hoe-tang or axe-tang for the purpose,
which would yield a conical or biconical bore. This tool would consist
of a flat piece of iron, about a quarter of an inch thick, and tapering to
a rough point from a diameter of 2inches(5cm.). This would produce
just such types of bore as we find. There is seldom proof that iron
was employed on a particular specimen.

The Bored Stones of South Africa. 11

The narrow conical bores are in a somewhat different category.
Oddly enough none of these have been attributed to the Bantu. In
size and proportions they appear to fall into the Prehistoric Series.

Diam. Ht. Mouths of Bore. Angle.
Leydsdorp. : a i:8 5:8 3°6 2-2 14°
Welbedacht : 10-6 7-4 3-0 1-4 1
Zwartruggens . 12-0 8-2 3-0 1-4 10°
Klerksdorp 11-0 6-8 4-0 1-6 20°
-Bloemhof A . 5 : 11-4 7:0 4-0 3:6) | lng
Tygerfontein, Hoopstad . 9-6 7:3 2-9 05 | 20°
Douglas . ‘ : : 10-2 9-5 3-4 Thos, 13°
Blaauwheuvel. 8-7 4-6 3°5 2-2 14°
Smithfield 9-0 5-1 4-0 2:5 212
Knysna . 9-0 4-9 4-6 2°8 i

The distribution is mainly confined to the Western Transvaal and
the Free State areas, where Bantu peoples have long existed. The
fact that all specimens fit perfectly into local Prehistoric Series, and in
addition their generally small size (compared with known Bantu
material) suggest that these are protohistoric and that iron was
obtainable from the neighbouring early Bantu. (See later, under
V.M.R. area.) The similarity in size of the last three examples is
suggestive.

An additional specimen (M.M.K. 1535), a large bead from Newlands,
Barkly West (4:9 x 3-7; bore about 0-8, tapering to 0-3 cm.), has quite
certainly been made with fencing wire or some similar thin metal.
This example is almost small enough to fall into the “bead” group.
It has split longitudinally.

The use of a tube for boring was known in Europe in Neolithic times.
A bamboo or reed was rotated in the stone, constantly fed with sand
and water as an abrasive. This action cut an ever-deepening circle
into the stone, producing a cylindrical bore in the stone and in addition
yielding a cylindrical core of stone of a type unknown here. In the
case of unfinished Huropean specimens this core stands isolated in the
incomplete bore; an unmistakable object. This again is absent in
South African sites.

The relationship of Bantu specimens to prehistoric is hardly dis-
cernible in the study of the bore, and we can presume that much the
same boring methods were used. The histograms of all Bantu
specimens, compared with 100 random Bushman examples, show a
slight difference only (pp. 185-191).

12 Annals of the South African Museum.

Hole Mouth. Hole Centre.
Range Max. Range Max
8 Frequency 8 Frequency.
BAaNntTU ; ; 1:9-8-7 5:0 cm. 1:1-7:7 2°45 cm
PREHISTORIC ; n 1:9-7:°5 3:9 cm. 0:5—4:6 2-2 em,

The largest of the Bantu hole-centres probably represents an old
well-used specimen in soft stone. The smallest Bushman figure
represents a small unfinished example. Further histograms may be
worked out from the figures given, and may have some regional value;
but owing to the unknown factors of material in both the bored stone
and the stones used for boring, the histograms would have little
value. The whole question of boring with metals will be dealt with
again.

““Sacredness”’ of the Bored Stone.

It is evident that these stones were considered by their makers as
of sufficient value to be shaped, smoothed, and finished off carefully.
This suggests that shape and size were of some importance to the pre-
historic users. Yet we can hardly attach any directly religious
symbolism or significance to these stones. That they may have
symbolized the female aspect of food-getting (agriculture or collecting)
is perfectly possible, but a directly phallic representation is unlikely
and without any factual justification. The use of the hoe among Bantu
agriculturalists as a symbol of womanhood depends on a double
foundation—the natural association of the hoe with the woman’s
life-work, and the use of the hoe in bride-price payments. This
second basis is lacking in the Prehistoric Series of bored stones, so
far as any available evidence goes.

To disprove the sacredness of the stones per se we may point out
the number of instances in which these bored stones have been used
as hammers or as upper grindstones; so frequent that we can only
attribute these uses to the makers themselves. This does not suggest
a “sacred object,” but rather one of the ordinary everyday tools of a
woman’s life, or a man’s weapon. It is of course quite possible that
these stones had a secondary phallic significance, as they were gener-
ally used by women for digging, but speculation of this sort is not
our affair.

The frequency with which completed stones are found split, so that

The Bored Stones of South Africa. 13

the two halves lie together, has been adduced as a proof of this sacred-
ness. It has been suggested that they were deliberately broken,
either on the death of the user, or to prevent the implement from
falling into enemy hands. The obvious implication is simpler than
this. A wooden digging-stick, if tight fitting, would be liable to
swell and to burst the stone if allowed to get wet. This suggests
chance fracture rather than “ritual breaking.”

S1zE AND SHAPE.*

In studying many hundreds of specimens, it has been found neces-
sary to reduce description to a minimum. Much of the material
has been measured at other centres, and a simple, uniform set of
measurements has had to be developed. Four direct measurements
are given for each complete circular specimen; and five for each of
the elongated examples. From these certain indices are developed.
Simplification of this sort can only be regarded as a conventional
expression of the stone. It does not present the shape of the stone
completely. Thus a sub-cylindrical form is not differentiated from
the true prolate, sphere, or oblate, though this may be noted in the
right-hand column. Evidence is not sufficient to yield results as to
variations in shape, and I cannot make any deductions from the
bulk of material that I have measured personally.

All these measurements have been taken in such a way as to
represent the stone in its original entirety. Thus if a stone has been
cleanly fractured, the shape is presumed to have been symmetrical,
and an allowance made. Such figures are approximate and are given
in brackets in all instances. They are accurate within a few milli-
metres. Worn specimens are similarly measured to allow for the
abrasion of wear.

For circular specimens the following measurements are given in
this order: maximum diameter and maximum height. These are
followed by two figures, index and cube. The index is the percentage
ratio of height to diameter, and depends upon the formula:

100 x ht.
diam.

The “‘cube” represents a square box into which the stone would

fit exactly, and is the product of the formula d? x h.

* For some remarks on the use of the bored stone as a pipe-bowl for hemp-
smoking, see Appendix I at the end of this work, also Laidler, P. W., in Trans.
Roy. Soc. 8. Afr., vol. xxvi, 1938, pp. 1-24.

14 Annals of the South African Museum.

These figures are followed by the external diameter of the larger
mouth of the bore, and finally the minimum diameter towards the
centre of the bore.

In the elongated examples the following figures are given: length,
width, height, cube, mouth of hole, and minimum bore. It has not
been found advisable to develop an index for this group. Cube is
the product of |. x w. x ht.

By circular specimens is meant those which appear to the eye to be
circular in plan; generally length does not exceed width by more
than 8 per cent. of the length.

In the case of the prehistoric circular examples we can divide our
specimens into rough terminological groups. The shapes depend
upon the vertical section of specimens (7.e. the index), while sizes
depend upon the cube. Where terms are used to designate a whole
group, the average index and the average cube should be used to
supply the term. I have here substituted “prolate” for the more
cumbersome “‘hypsisphere,”’ and give “oblate” as an alternative to
“oval” of the Preliminary Survey.

Prolate . over 110 per cent.; oversized . over 2000 cu. cm.
Spherical . 90-110 a large . . 1350-2000 ,,
Subspherical 70— 90 Ry large medium 800-1350 ,,
Oval (Oblate) 50-— 70 ae medium . 450-800 ,,
Suboval . 35-— 50 Hh small . : 180— 450 ss,
Oval disc . 20- 35 3 very small . under 180 _,,
Dise . . below 20 RY

The use of such terms cannot be hard and fast, and in discussing
groups the term “‘disc’”’ may be found to cover a group ranging from
15 per cent. to 40 per cent., while “sphere”’ may include specimens
with indices of 85 per cent. to 105 per cent. Similarly “large” might
apply to a group with a cube ranging from 1450 to 2200 cubic centi-
metres. These terms are given to help in the description of bored
stones in future scientific descriptions of implement types, though it is
most earnestly hoped that exact measurements will be given for each
individual specimen as well.

Cube, Material and Werght.

The term “‘cube” employed here is misleading unless certain
fundamentals are remembered. Owing to a strong mathematical
tradition in our own society, we are prone to believe that the weight
and volume of two spheres, one of 1000 cu. cm. and the other of
500 cu. cm., is in the relation of two to one. The “‘untutored savage,”
on the other hand, would regard a stone measuring 5-3 cm. in diameter

The Bored Stones of South Africa. 15

as only “‘slightly larger” than a sphere measuring 4-2 cm. in diameter,
and, of course, the difference in diameter becomes of greater or lesser
importance as the diameter of the spheres decreases or increases. The
idea of “‘size”’ (apart from the mathematical concepts of mass, etc.)
is probably a mental assessment of a combination of diameter and
expected weight. In practice this should mean that the larger the
type the greater the acceptable deviation in size (or the scatter when
the type is represented on a graph paper), though I have not found
this to be true in most instances given here.

It will be observed that neither the material nor the weight are
here regarded as of primary importance. Where they are known
they are given. The material can seldom be accurately analysed
without the partial destruction of a specimen, and even then chemical
or microscopic analysis would have to be carried out at suitable
centres for some two thousand specimens. Vague terms, such as
sandstone, shale, granite, soapstone, dolerite, slate, marble, quartzitic
sandstone, have little real significance, save that they suggest the
degree of hardness of the materials it was possible to work. Enough
evidence is produced to show that all these materials were used, and
were capable of being efficiently worked.

In a few instances, where a stone has distinctive local peculiarities
and might with certainty be attributed to a particular outcrop, exact
definition would be of great value; but even here it is to be observed
that the material used need not bear any direct relationship to local
petrology. Much of the material comes from a gravel source, and
shows evidence of river pebble origin. It may have been carried
by stream action for many miles from its source, or it may have
been carried by man. There is no means of telling. If we were
dealing with quarried stones (such as the bluestones at Stonehenge)
the source of material would be of immense value in showing the
direction of migration.

The exact assessment of weight is similarly of little value, as it is
part of our own mathematical tradition. While the original pebble
would probably have been chosen with an eye to weight, the boring
subsequently removed a variable and uncertain quantity of the
original stone. Weight has a broad value in dealing with unbroken
specimens, but in many instances the exact weight could not be
obtained from either the private or museum sources of material.
Where it is known it has been given.

An approximation to weight can be obtained from the external
dimensions, and “‘cube”’ in cubic centimetres can be roughly converted

16 Annals of the South African Museum.

into grammes by multiplying the cube by a factor. If the average
diameter of the stone (or the cube root of the “‘cube,”’ which is similar)
is greater than 13-0 cm., the factor willbe 1-6. Ifthe average diameter
lies between 6-0 cm. and 13-0 cm., use 1-55. Ifit is less than 6-0 cm.,
1-5 gives the approximation. Thus:

100 cu. cm. (root si 65) =150 grammes (about 5:5 oz. )
200 29 ( 29 5°85 ) =300 29 ( 99 ll 99 )
300) oe (ae GB) e265 3 Cee 1 Wb: )
500 39 ( 99 7: 94 ) =775 99 ( 99 1 99 Il OZ )
1000 _=SCé,, (,, 10:0 ) =1:55 kilos (ee 3) 14; ORS)
2000 29 ( 29 12-6 ) =3:1 29 ( 29 7 Panes. ry) )
4000 99 ( 99 15: 9 ) =6-4 99 ( 99 14 39 9 99 )

Weight has little relationship to shape, though it is obvious that,
with the biconical bore, the higher a stone in relationship to diameter,
the greater will be the quantity removed by boring. The material
of which a stone is composed has some effect, but only in extreme
cases (such as specular iron ore), and the figures above can be taken
as reasonable approximations in most instances.

Copying.

While in the Bantu Series there seems to have been no real attempt
to copy size or shape from traditional examples, the vast bulk of the
prehistoric examples conforms to one of a variety of local patterns.
These may all be the varied progeny of about seven prototypes
described later from the Transvaal and the Vet Marico area.

As the question of copying is important, for the validity of my
deductions depends upon it, we can study the subject shortly. It is
to be understood that I here use the terms “copied” and “‘selected”’
to carry the same general significance.

So far as one can judge it would seem easier for the untrained eye to
assess the relative proportions of two subspherical objects, and their
relative diameters, than to estimate relative volume. In actual
practice the assessment of relative weight can be very accurate too,
provided that the specific gravity of two objects is reasonably similar.
However this may be, two facts are strikingly brought out by the
plotting of our figures on graph paper.

i. The Bantu Series shows a wide and apparently haphazard
scatter, suggesting that (provided there was sufficient weight for the
purpose) neither proportions nor dimensions were very important.

ii. The Prehistoric Series, in contrast, shows well-defined grouping,
sufficiently distinct in most areas to prove that certain proportions
and sizes were achieved by processes of selection of natural stones,

The Bored Stones of South Africa. 17

or of copying an acceptable size by reducing a selected natural stone
by pecking.

The Bantu Series therefore raises no new problem, but the Pre-
historic material does. The first is the problem of copying. How
were the stones selected or copied? I have seen a Transvaal Bantu
select stones of equal size from a river gravel by comparing the size
of one stone directly with the cast left by another in the mud. The
result was the choice of a series of large pebbles or boulders none of
which varied more than an inch in diameter, about the range of error
found in the Prehistoric Series. It is therefore obviously possible
for that degree of accuracy to be obtained by simple means. In
smaller examples judgment by eye would be more exact, but peculiarly
enough the range of error in small examples appears to be much the
same as that in the larger.

It is obvious that there are two types of copying:

1. Parallel copying, where several copies are made from one original.
In this type the average dimensions of all copies approximate to those
of the original.

2. Serial copying, where a copy is made from an original, a second
copy made from the first, and so on. There may be a cumulative
error here, though this would in part be cancelled and corrected by the
chances of subsequent copying.

If we combine these two types of selection or copying, and make
two first copies from a single original, then make copies from these
two independently, we may get two graphical groups, each slightly
different in their graphical scatter. We can regard these as linked
groups. Curiously, this third graphical pattern is the most usual,
and can be observed in almost any area in South Africa. This may
be called “divergent copying.”

Deviations from the original may be stopped or controlled by
tradition, and tend to approach more closely to the original once
divergence is recognized. This stabilizes a “fashion” in shape and
proportions, probably as exact as those that govern the proportion
of brim to crown in our own choice of a hat in any year, and analogous
with our estimations of sizes in footballs and cricket balls, etc.

In some cases the doubled or linked products of divergent copying
are sufficiently close for us to ignore the difference, and to unite two
clusters under a single average. In other cases, there is such a distinct
line of demarcation that it has been thought more satisfactory to
keep the two clusters separate and treat them as linked. This

necessarily leads to anomalies in comparisons between types. It
VOW. XXXVI PART J. 2

18 Annals of the South African Museum.

therefore becomes essential in such cases to draw deductions in one
area, and re-apply them in the next.

These linked or twinned clusters may vary in diameter, and remain
constant in height; but this is not always true, and in the Vet-Marico
River area we get averages (11-1 8-7 cm. and 10-9 x7-3 cm.) that
show a similar diameter and variable height.

Peculiarities of this sort appear to be the rule rather than the
exception in the middle ranges of types, and it is difficult to know
just how they should be dealt with. The problem becomes more
difficult in focal areas where types are abundant. It becomes essential,
before any area is understood, to plot the position of each specimen
in that area on a piece of centimetre graph paper. The next essential
is to plot the specimens from each of the adjacent areas on separate
graph papers. Only the two dimensions, diameter and height, need
be used. Suitable adjustments of averages should then be made for
the local area, so that direct comparison becomes more easy and more
exact. Unhappily the publication of a series of graphs of this nature
would be too expensive, and only a few are published as examples. ©
Should any worker desire to develop this work further, either locally
or generally, the figures are given, and graphs can quickly be built
up from these. Thisis essential before local applications of the present
work can be completely understood.

In the section containing the accumulated material, the local
graphical clusters are represented by averages. Where only one or
two examples represent a cluster, the average is given in brackets.
Where more examples are known, there is no bracket.

Aberrant Forms.
Aberrant forms are here divided into two groups:

Erratics that have wandered in from surrounding areas; and
Anomalous, those that fit in with no local distribution.

These latter may be the product of individual variation on the part
of the maker, as, for instance, the prolate examples that constantly
crop up, presumably the result of local competition and emulation.
The ever-present factor of wrong location must be remembered.
For instance, a specimen shown to me at Cape Town eventually proved
to have come from the Klerksdorp district, almost 1000 miles away.
Such extreme instances are rare, but they exist. J. M. Bain and Dr.
Kannemeyer appear to be sinners in this respect. Both early
collectors presented batches of implements collected over a series of

The Bored Stones of South Africa. 19

years to the South African Museum, and in many instances a wrong
allocation is almost certain. Kannemeyer, for instance, was known
to have collected numbers of stones, from sites near Smithfield,
Burghersdorp, and elsewhere, to have secreted these in the houses of
friends, and finally to have sent them in bulk to the South African
Museum. Museums are not without blame in this respect. —

How to Use the Present Survey.

Results achieved here can only be tentative. Our knowledge of
any subject only leads us to a better digestion of further material.
By supplying some attempt at a framework on which local workers
may build, I hope to evoke additional knowledge from them. This
publication is out-of-date immediately the manuscript leaves my hands.
It should be the intention of local workers, now that this raw material
has been published, to get a more accurate knowledge of local types.
There are various problems that can only be dealt with adequately
by local workers, and we can discuss a few of these here.

1. The isolation of types of bored stone. This is here dealt with
by fitting the main dimensions of each stone to a simple graph and
observing the scatter. One or two examples are given to illustrate
method.

2. The isolation of areas of distribution of types. This has been
attempted here so far as my present knowledge permits, but much
remains to be done in the future. The term “area” is here used to
cover the approximation of a graphical dispersal to a natural or
environmental region, e.g. a river bed, a type of veld, an altitude, etc.
Finer adjustment of type to natural region should be possible locally,
and an extension and more exact definition of an area should result.
Remeasurement of available material may help us in this.

3. A full analysis of associable finds. The material should be
related to implement types, paintings, engravings, presence or absence
of pottery, shell midden, fishbones, stratification, skeletons, and any
other suitable elements of culture. Implement types should be
related to bored stone types wherever possible. Negative association
is Important. :

4, Extension of knowledge from the particular to the more general.
Once types have been isolated, local areas defined and associations
studied, this whole complex can be extended into surrounding areas,
from simple areas to areas of overlap (focal areas). Work should be
from the simple to the more complicated.

20 Annals of the South African Museum.

This principle of extending our knowledge from the simpler to the
more complicated is fundamental. So far the tendency has been all
in the other direction. Areas of abundance, such as the Free State,
Knysna, and the Cape, have been blindly worked, and various collec-
tions made by inexperienced workers, all seeking the same types of
evidence: bored stones, grind-stones, human skeletal remains,
pottery, and one or two easily recognizable stone implement types.
Collections of this haphazard type are often marked “Free State’,
“Cape Colony”, “Natal”, or “Northern Transvaal”. Even beads
and microliths are missed in many instances, so that our most in-
defatigable workers have in the past done by far the most harm. The
minimum of exact evidence that should be given with a collection is
the registered farm name (or map reference) and the district or nearest
town. If possible museums should ask for every available detail
of the discovery of each specimen, and record it in writing.

In certain instances, in spite of this lack of very exact evidence, it
can be shown that the number of types of bored stone in a definable
area 1s strictly hmited. This is observable in the Vet-Marico area,
for instance, and can be abundantly shown elsewhere. In other cases
a single type will cut across two or three defined areas and show an
independent distribution; an example of this is seen in the Natal
coastal belt and along our southern coast, where it is difficult to find
very precise limits to areas without subdividing them further than
our knowledge of exact provenance permits. This subdivision is
suggested for the Knysna-Tzitzikamma area, where forest, mountain,
and midden types may be discerned. Only further more detailed
evidence of the exact source of each example will prove or disprove
this.

If we show that certain types have a limited and definable distribu-
tion, then we have shown that the types exist, and that we are not
dealing with a chance distribution. It is important to recognize
the validity of this hypothesis, as, if we plot the size and shape of all
specimens from all areas in South Africa, we find that we are dealing
with what appears to be a purely chance distribution. Once the
examples are broken down into areas, the picture is very different.
Granted that we have “pure areas,’ such as the Swartland and
Grootrivier, it is obvious that deductions made in such areas should be
reapplied where a large variety of types exists. Logically we cannot
proceed in the opposite direction.

Those areas in which a great variety occurs we can call “‘focal”’
areas or areas of overlap. These are areas particularly attractive to

The Bored Stones of South Africa. 21

hunting and collecting peoples, and highways of slow migration.
They tend to be secondary cultural centres of settlement and consolida-
tion from whence new migrations will originate. We are therefore
dealing with a more complex problem here; we are studying numerous
overlapping types, the products of successive migrations or periods of
sojourn, or even of serial copying. Some of these can be related to the
products of people outside the area, and are described as erratic.

For these reasons comparison should constantly be made to parallels
from adjacent areas. In actual fact we might divide South Africa
into three general types of area: focal areas, migrational routes, and
marginal (or pure) areas where few people came. This underlying
division is basic. No one is directly interested in the distribution of
types; it is the deduction concerning man’s story and environment
that is of interest.

It might be argued that in those areas where large numbers of
specimens happen to have been collected, large numbers of types
occur, and that the number of types is in direct proportion to the
number of specimens that present-day man has collected. In fact
thisis not true. The number of specimens collected has every appear-
ance of being generally proportional to the original population, and
to the numbers of specimens available. The number of types may
therefore be regarded as related to the number of uses, and to the
number of cultural groups represented in the area over a period of
time, which would partly be a function of environment.

I refer elsewhere to the varied needs that this “bored stone”’ would
meet. No single cultural group felt all those needs. One group
might need only a digging-stone, another might add the need for a
knobkerry, another might use a dagga-pipe, and so on. We may
therefore deduce that the number of types represented is a product
of the number of cultural groups present in an area over a period of
time, multiplied by the number of uses to which this general form
could be put by each group. In other words, where prehistoric
density was highest, and migration most frequent, the greatest number
of types will occur, and be available for collection.

Region and Area.

In the Preliminary Survey the Union was divided into regions and
subdivided into areas. This arrangement will be generally adhered to
here, but with certain changes. Regions will receive Roman capital
numbers; areas will be referred to by two, three, or four initials;
types within an area by the use of small Roman figures.

22 Annals of the South African Museum.

One further difficulty arises: region and area in the Elongated
Series are coincident. Presumably the distribution of types is different,
but the number of examples known at present does not permit us to
make even this deduction with complete certainty. Here regions are
given initial letters, differing from those used for the Circular Series
unless the regions themselves coincide.

The region and the area are concepts based upon both distribution
and environment. In dealing with distribution it has seldom been
found possible to confine a single type to a single area. In some
instances types overlap areas, in others they may be confined to
smaller distributions within an area. This concept of a static area is
therefore a compromise. The second element in this compromise
consists of the various environmental factors which controlled the
movements and sojourns of early man. Here and there small areas
have been singled out, showing perhaps three sites. These necessarily
need further investigation, but the suggestion here is that they may
represent pure types with a limited and local distribution.

The environmental factors entailed will be dealt with later under the
headings “‘ Attracters”’ and “Barriers”’ (p. 117).

PART II.—PREHISTORIC CIRCULAR SPECIMENS.
| MATERIAL AND COMMENT.

The present part consists of a complete list of circular specimens
made available to me by various individuals and institutions. The
material is listed under sites, partly to reduce space, and partly to
permit of the wholesale removal of a single site into another area,
should future material make this necessary.

The term “‘site’? must be thought of in a very wide sense. In
almost all cases no site proper is given. In over half the specimens
only the district is known. The district may either be taken from the
name of the nearest town, or may agree with the administrative
districts of the Union. There is, for instance, no Kofhefontein district
in official publications. It will therefore be realised that a dozen
different collections of implements submitted to an institution as
from “‘Koffiefontein”’ might come from a dozen true sites within a
ten- or twenty-mile radius of that village. Where administrative
districts are cited the position is no better.

I have therefore made it a general rule to refrain from splitting
collections of implements presented at one time as having come from

The Bored Stones of South Africa. 23

a single provenance. Where two collections, generally by different
donors, given at different times or to different institutions, are marked
only by provenance I have had no compunction in dividing them and,
where necessary, placing material into two different areas or even
regions. The rule is not absolute, but has been infringed in two or
three instances at most.

Wherever the letters A, B, etc., are employed after a place-name
material has been split into presumed sites. The result will best be
seen by turning to Koffiefontein A and B series. It may be seen also
at Rustenburg, Lichtenburg, Bloemhof, etc. The administrative
districts covered by these last three names include 9020, 4295, and
840 square miles respectively. The area of Wales is 7360 square miles.
Other figures are given under the Free State Focal Region.

Overleaf is a schematic geographical representation, summarising
all the regions and areas discussed. The regions are numbered in
Roman figures in the order of their treatment. (See next page.)

List oF SOURCES.

In the following list I have added the names of persons who have
helped me in the collection or measurement of material, or by allowing
me access to the material at their disposal. I wish to record my most
grateful thanks for the very great help that these persons and institu-
tions have given me. I am no less grateful to those who have sent
material to me, either at the South African Museum, or at the Univer-
sity of Cape Town. Those who have been of most help have been
mentioned in the text under their own areas.

AMG =Albany Museum, Grahamstown, and Mr. F. Midgley.

ASJ =Archaeological Survey, Johannesburg, Mr. B. D. Malan and Miss J.
Haughton.

DC =Prof. M. R. Drennan, Department of Anatomy, University of Cape Town.

ELM =East London Museum, and Miss Courtney-Latimer.

HAC =Harger Archaeological Collection, including material from the old Chamber
of Mines Geological Collection, Dr. H. 8S. Harger and acting curator.

HC =Dr.C.H.T. D. Heese’s collection, Alt Stadt, Riversdale.

MFH =Dr. Mossop’s collection at Fish Hoek.

MMK =McGregor Museum, Kimberley, and Mr. B. D. Malan.

MW =Mr. Frans Malan’s collection, Groenberg, Wellington.

NMP =Natal Museum, Pietermaritzburg, Dr. R. Lawrence and assistant.

PEM =Port Elizabeth Museum, and Miss Hoekstra.

SAM =South African Museum, Cape Town.

SI =Stone Implement Collection, Durban Museum and Art Gallery, and Mr.
K. C. Chubb.

SR =Southern Rhodesia Museum, Bulawayo, and Mr. Neville Jones.

TMP =Transvaal Museum, Pretoria, and Miss J. Haughton.

UCT =University of Cape Town, Archaeological Collection.

WUA = Witwatersrand University, Department of Anatomy, and Dr. L. H. Wells.

24 Annals of the South African Museum.

SCHEMA.*
I. TRANSVAAL.
Eastern. ~~
Central. \
Southern. Ne
\
~
Il. Nata.
Highlands.
x Lowlands.
III. Vaat. NX —

Schoonspruit- Vaal. Ne
Vet-Marico Rivers. N
Bloemfontein-Sepani. \

/ VNaal-Harts. IV. FREE State.

JT. Koffiefontein A and B.
/ Ula. Vaat (LowER). Central Free State.
ye Middle Vaal. <-——> Alexandersfontein.
Vis Lower Vaal. Modder-Fauresmith.
Va South-East Free State.
Vi x
x \

V. WESTERN Bypass. VII. SourH-East.
Molopo-Groenwater. Upper Thornveld.
Upington-Hartbees. Lower Thornveld.

Bokkeveld. North and South Transkei.
SS VA

1 VIII. Lowianps. ve
VI. THIRSTLAND. Kast London. wee

Kaaienveld. Bushman’s River. “
Namaqualand. Port Elizabeth.
Winterveld. Zuney-Maitland Mines.
Trekveld. fo

Nieuveld-Goup.\, IX. SourH Coast. Pe
SS Sundays River.
Grootrivier.
‘. Tzitzikamma Caves.
Knysna-Tzitzikamma.
Swartberg.
Belvedere- Heidelberg.
Gouritz-Karoo.
XI. Care oF Goop Horr. X. AGULHAS. i
Hinterland. West Langeberg. /
Cape Flats. / Breederivier. “
Cape Middens. / Bot River.
Sandveld. Wa
Swartland.

* Detailed maps follow this section (pp. 109 et seq.).

I. THE TRANSVAAL REGION.

The northern boundary of the known distribution of circular bored
stones runs from the Dwarsberg and Witfonteinberg along the
Hoekberg. From here it passes northward towards the Blaauwberg,
eastward to Klein Letaba, then southward to include Barberton and
Swaziland. The Transvaal region includes most of the province itself
and a few sites in the north-eastern Free State. The northern

The Bored Stones of South Africa. 25

boundary coincides with the limits given above. From Swaziland
the southern boundary runs south-west to Harrismith, and then
includes the Wilge River basin and the Upper Vaalas far as Klerksdorp.
In the Wilge River section it overlaps the Natal Highlands region,
and part of the eastward extension of the Vet-Marico area.

The distribution of known material is not continuous. It starts
at 23° 50’ south, but no examples are known from the Limpopo and
few from the Olifants River basin. Prehistoric specimens are more
limited in this direction than the later Bantu examples, to be dealt
with in a subsequent part (Part V, p. 170).

There seems little reason to believe that the bored stone was Bantu
in origin, or that it was taken over and spread farther south by the
Bushmen. The evidence at Djebel Redeyef* and other North
African sites points to an immediately post-Capsian dating there, and
shows “‘Intercapsioneolithic” associations, which would here be
regarded as broadly Wilton. With similar associations at the two
extremes of the continent, it is probable that the Wilton peoples were
everywhere familiar with the bored stone, at least in certain phases
of that culture. The important question for the future to decide is
whether the term Wilton is more closely associable with the people
we call Hottentots, or with the Bushmen. Evidence on this point
is not yet convincing, though the association of Schofield’s “‘ Hottentot”’
pottery with the Wilton is important.

I regard the Transvaal as. the primary area of distribution of all
our South African prehistoric types, and this might well be called the
Prototype Area, so farasthe Unionisconcerned. The Free State region,
immediately north of the Orange River, provides a secondary area of
distribution to the Cape Province, a Deuterotype or Focal region.
This concept will be developed further in dealing with the migrations.

The literature on the Transvaal is extremely meagre, and more
has been written on the Bantu material than on the Prehistoric.
The notes I have on associations are of little value except for the
Vaal. It is hoped that future material (especially that found in the
vast districts of the northern and western Transvaal) will be more
accurately documented.

ET. EASTERN TRANSVAAL AREA.

This area is marked by few types; an area where absences are more
important than diversity. Our evidence so far is meagre, and the

* EK. Gobert, L’Anthropologie, xxiii, 1912, pp. 151 et seq.

26 Annals of the South African Museum.

first and second types listed below are separated in the light of
analogous material in other parts of the Transvaal region. )

* Type Diam. Ht. Index Cube Affinities

i (123 86 70:0 1070) OT. ii, NH. iii.
i. (105 80 762 880) CT. iv, VMR. iii.
Fie MOS AON NEVES 615
ive lO: 7 Min 4260) 64320 525) CT. vii.

There seems to be no relationship with the Southern Transvaal.
Ermelo and Lake Chrissie suggest a close relationship with the CT. 1
and i1. The area shows only specimens of 10-0 cm. and over in
diameter. The final specimen listed below (Hastern Transvaal)
suggests, by its size and type, the Ermelo area.

* Source Diam. Ht. Index Cube Bore Notes and refs.
BELFAST. —“—
TMP 1861 . 10-2 6:3 60:8 645 2-7 1:3
MALELANE.
TMP 8488 . 1ll-l 8:3.° 74-8) ) 1025) 75-3) 3-1
BARBERTON.

SAM 3022 . 10:3 57:2 625 ‘0
1

5:9 3 1-6
TMP 1912/12 10-2 4:3 42-1 450 4: 2-4 Found in chance

association with
Bantu specimen,
in Kaap R. wash.

J AKSLUST.

TMP 8040 . 10:2 5:6 54-9 585 38 3:3

LAKE CHRISSIE.

SAM 1733 . 10-0 76 76:0 760 43 3:5

ScHaGEN, NELSPRUIT.

WUA 1085 G. 11-2 5:1 45:5 740 4:3 1:8 Indeterminate basic
igneous rock.

ERMELO.
SAM 3183 . 12:3 80 65:0 1210 30 1:6

EZULWENI.
SP6w i: beeen 4:9 43-7 615 3-1 2:8

“EASTERN TRANSVAAL.”
SAM 3. eles 92 74:8 1405 3:1 3-0 Cylindrical bore.

CT. CenTraL TRANSVAAL (PROTOTYPE AREA).

I deal with the Hastern Transvaal first, as it appears to stand apart,
and to represent the eastward distribution of a limited number of
types (fig. 2). The types from the Central Transvaal agree with the
Kastern Transvaal series, but additional types are present, thus
showing the ET. area to be an area of absences, derived from OT.

* See pp. 13 and 14 for definitions of terms used above columns.

HEIGHT.

The Bored Stones of South Africa. 27

DIAMETER.

Fia. 2.—Central Transvaal.

On the other side the CT. shows some affinity with the Molopo-
Groenwater stretch of the Western Bypass. The term “Prototype
Area”’ will explain itself when the part on Migrations is reached.

i 141 10-8 766 2150. MG.i.

ii. (12-2 G20 75-4 3870) | EL. a:

i) 11-7 78 66-7 970 VMB. i, MG. ii.
iv. (10-2 8:2 80-4 865) ET. ii.

v. (11-5 5-4 47-0 715)

vi. 10-0 5:4 54:0 540 NH.v.

vii. (10-8 42 38-9 490) ET. iv, MG. iv.
viii. 7-7 5:6 72-7 330 MG. v.

ix. (5-7 3:6 63-2 117) ST. iii, VH. vi.

As might be expected the Leydsdorp and New Agatha specimens,
from sites tucked away behind the Olifants River and the Drakensberg
and Murchison ranges, diverge slightly from the usual scatter. The
Button’s Kop example (CT. ix) is obviously an erratic, probably
related to the Southern Transvaal.

28
LEYDSDORP.
SAM 2824 12:7 7-7 606 £1240
ASJ 2/38/7 7:8 5:8 74:3 365
New AGATHA.
TMP 1447 11-4 53 =—8-2 690
‘“NORTHERN TRANSVAAL’”’ (?PIETERSBURG).
WOUA 250 10-8 4:2 38-9 490
Butrton’s Kop.
TMP 5437 aw 3:6 63:2 117
MOo.LsGat.
TMP 1496 9-8 5-7 58:2 550
POTGIETERSRUST.
SAM 4706 12-4 93 75:0 1430
WARMBATHS.
TMP 9162 7:6 4:9 65:4 285
PANKOP.
TMP 14:0 11-1 79:3 2180
WITFONTEIN.
SAM 7529 10:8 4:3 40-0 500
RUSTENBURG: A.
WUA 670 12-1 90 74:4 1320
TMP 8491 10-2 4:7 46:0 490
GRASMERE, VENTERSDORP.
ASJ 14/39 11:6 76 65:5 1025
KROMDRAAI, POTCHEFSTROOM.
Univ. Stell. 11:5 5:4 47-0 715
LICHTENBURG: A.
HHC 1042 11-5 8:0 69:6 1060
TMP 8487 10:2 78 76:5 810
SANNIESHOF, LICHTENBURG.
ASJ 29/42 9-8 5:6 57:2 530
VAALBOSPUTTE, LICHTENBURG.
WUA 1240 7:7 59 76:9 350
HEBRON.
TMP 7548 14:0 10:2 72-9 2000
WONDERBOOM.
TMP 8509 10-2 5-4 52-9 560
PRETORIA.
TMP 7560 146 10:5 714 £2240
3670 14-0) 4 81-4) 2235
1879 10-2 8:5 83-3 885
1469 10-2 5-1 50:0 530

2-9
3°6

3:6

4-8

38

3°6

4-4

2-9

4-2

hae A
= vo

1-6
2-2

1-4

2-4

17)

2-9

1-6

2-0

1-6

Cee
oe

bo 62
—

2-2

1-7

1-9

1-7

Sindcslrs
Qwwa

Annals of the South African Museum.

Blue
schist.

micaceous

Conical bore.

Anorthosite.

Erratic.

Broken.

Si

6 feet deep.

2-6 kg.

Woncerstone.

‘45 ke.

The Bored Stones of South Africa.

KENSINGTON, JOHANNESBURG.

WUA 67/1 10-2 6:0
EERSTEFABRIEKEN.

TMP 1876 11-4 7:8
BETHAL.

Univ. Stell. 13-7 10:6

TMP 5187 10:2 6-2
HaARRISMITH: A.

ASJ 14/40 14:3 10-9

ST. SouTHERN TRANSVAAL AREA.

58°8

610

1015

1990
645

2230

bo on
aT Or

2:7

4-]

29

‘88 kg., 12 feet deep.
Serpentine.

While there are a number of specimens from here, there are relatively

few types (fig. 3).

1

Be ae ACD
& bo ko vo

Type iv is an erratic.

1645
405
178
145)

NH. ii.
INE ve

INES vis CT ix.

An example included here in group ili (Johannesburg, 7-6 x 3-9) is an

erratic.

RUSTENBURG-BRITS RoapD.

ASJ 56/40 12:7 10-0
RiIEeTGAT, BRITS.

UCT 42/ 13:05. 10-5
CYFERBULT.

TMP 8045 8-9 5:9

VENTERSDORP Town LANDS.
HHC 1044 12-2 9-6

FREAN’S MILL, VENTERSDORP.

HHC 1058 6:5 4:5
_PoTCHEFSTROOM.

TMP 8461 12-7 9-4
VEREENIGING.

Univ. Stell. 6:9 2-4
SNYDERSDRIFT.

MMK 44 13-2) 1e4
ASVOELSKOP, JOHANNESBURG.

ASJ 47/37 8:5 4-7

nage] TIN LS) 3°9

MAGALIESBERG.

TMP 1848 5:7 4-2

78-7

80-1

66:3

78-7

69-2

74:0

34:8

1615

1735

470

1530

4-6

3°8

3°2

4-0

1-5

4-]

2-2

Ou

3-1

1-2

2-0

3°5

2-2

i-9

em bo
CO Or

1-3

The average, without this specimen, would read 6-0 x 4:2 cm.

Hammer.

Diabase or andesite.
Smooth.

Diabase. Hammer.

Pyriform.

Soapstone.
Erratic. SSV.

HEIGHT.

30

Il-

10-

A=

=

Annals of the South African Museum.

DIAMETER.

Fic. 3.—Southern Transvaal.

ELorr’s CutrtTine, PRETORIA.

TMP 8038 13:3 9-7
BRONKHORSTSPRUIT.

TMP 1469 12:7 10-8
GERMISTON.

HHC 1253 6:3 4-0
SPRINGS.

TMP 2333 8:9 5-1
BaRNARD’S Kop, HEIDELBERG.

WUA 776 8-1 2-7
Grap, Petrus STEYN.

SAM 5025 6:0 3:7
STANDERTON: A.

MMK 1157 13:3 (9-4)

73-0

85-0

63-5

07:3

33°3

61-7

67-7

1705

1740

159

405

177

135

1665

5:3

4-2

2-0

3°5

2:7

2:8

4:2

a1

2-7

1-1

2-3

1-7

2°5

2-8

L.S.A. Associations.

16 feet deep in al-
luvium, in ashes.

Found in spruit.

‘24 kg. Serpentine.

Smooth polished ser-
pentine. Ploughed

up.
SS. Damaged.

The Bored Stones of South Africa. 31

Those interested in the Free State should turn to the Vaal Region
from here (p. 35). Ifthe Western Province of the Cape is of interest,
turn to the Western Bypass (p. 60) and Thirstland Regions (p. 64).

Il. NATAL REGION.

This appears to be a cul-de-sac with little relationship to areas to the
south or west. The main affinities are with the Central, Kastern, and
Southern Transvaal areas. Only one affinity exists with the Free
State (Bloemfontein-Sepani 1.), and that is cut off from Natal by the
Basutoland mountains and by the Vet-Marico area.

The Natal Region starts in the south-eastern Transvaal at
Standerton, then follows the Klip River to Vrede. From here the
distribution so far known coincides with Natal proper, Southern
Zululand, and a very small part of Griqualand East.

The region is divided into two areas—the Natal Highlands, and the
Natal Lowlands. The latter is the more limited, lying generally
below 2800 feet above sea-level (900 m.), running inland with the low
Mooi and Tugela valleys until that altitude is reached. The Highland
area seems generally restricted to levels above 2500 feet (800 m.),
but a stone from Tinley Manor, an anomalous example from Durban,
and three normal specimens from there suggest that the limit is not
everywhere related to altitude.

The Lowland types conform to a limited extent to the Strandveld
or coastal strip, and may prove to be related to a coastal people.
This would suggest that both areas were held by different cultural
groups at the same time, and that each was sufficiently strong to
hold back the other except where the Umgeni and Tugela valleys
provided easy access to the coast.

Certain of the Highland types are associable with Smithfield N.
elements, and a later survey of migrational routes will show that those
feeding Natal diverge from those leading to the Free State, north of
the Vaal River. There is thus reason to suppose that my original
hypothesis, relating the Smithfield N. series to Rhodesian origins, is
remarkably supported (though not proved) by this new evidence.*

NH. Natat HIGHLANDS.

i. (11-0 12:2 111-0 1475) Anomalous.
Hs) be 10-5 86:8 1535 Sita:
iii: 12-6 8-2 65-1 1300 ET. i, BS. i.
Thyo. Bell 6-9 75°8 570 Scattered group. SSV. ii and iii.
Ws), Beis) 5-3 56-9 480 ST. ii, CT. vi.
Yash C0) 3:6 51-4 176 ST. iii.
* Goodwin, Proc. Rhod. Sci. Ass., xxxiv, 1934.

32 Annals of the South African Museum.

STANDERTON B.

WUA 1037 11-6 810-4
VREDE.
VILE 7/34 7:0 3°8
Moo1rHorEk, UTRECHT.
WUA 1266 12:0 8:0
Acton HomEs.
SI 10 13-4 8-4
BERGVILLE.
PEM — E 6:5 4:5
EstTcouRT.
SI 11 12:9 10:9
BERGVLIET, EKENNERSDALE.
Q. E. Carter . 10-2 7:0
12:7 8-2
YORK.
NMP 2515 10-0 7:0
BALGOWAN.
ASJ 32/35 8:8 (6:5)
9-5 (5:5)
7:9 7:2
TINLEY MANoR, STANGER.
A. G. Wills 5 13-2 7:6
DURBAN.
ASJ 24/139 11:0 12-2
Sie one 13-1 7:3
Bg . 12:0 8:2
UR . 10-4 5:6
UNDERBERG.
SI9 . : 6:5 3:2
KOKSTAD.
SAM 4756. 9-6 5:7
CEDARVILLE.
SP227. ; 7:7 3:9
MATATIELE.
UCT 34/20 . 8:9 4:7
MACLEAR.
ELM — 12:5 9:0
** NATAL.”
NMP 99 11-7 10-2

90-0

54:3

66-7

62-7

69-2

84:3

68-6
64-5

70-0

57-6

111-0
95°7
68-3
53°8

49-2

59-4

50:7

52°8

72:0

87-1

1400

186

1150

1510

190

1815

730
1325

700

505
495
450

1325

1475
1255
1180

585

136

525

231

370

1405

1400

4-5

3°5

4-3

4-2

3°5

5:6

4-2

3:1

38

4-2

4-8

747

2°5

3'0

3°5

2:0

bo =
> v9

2°8
2:3
2-1

3:0

2:0

2-9

Pel

2:8

19 kg.
shale.

Indurated

Ss.

Paintings.

1-06 kg.

All associated with
Smithfield N. im-
plements.

Gritty Ss.

Anomalous.

Roughly triangular.

Smithfield N. associ-
ation.

2-65 kg.

The route to the Natal Highlands is discussed fully under Migrational

Routes.

The Bored Stones of South Africa. 33

References.

Not much has been written on the Natal Highlands. P. J.
Kloppers * describes and illustrates a bored stone from the farm
Kalkkrans, Utrecht. It is irregularly quadrilateral in form, and
obviously has a low index. No indication whatsoever is given of the
size. Kloppers suggests that it may have been used as an axe (? adze)
ora hoe. This is highly possible, but the illustration suggests a type
very like Trevor’s Bantu examples from Pietersburg (p. 183).

NL. Natat LowLanpbs.

This series shows little relationship to any upcountry material.
The only similarity is the chance analogy between shale discs from
the South-Western Transvaal and discs from here. In the case of the
Transvaal specimens we are obviously dealing with the anomalous
local products of shale, and little value can reasonably be attached
to the similarity. Only one cluster from the Lowlands area is satis-
factory. The low index is the most regular feature, usually below
40 per cent.

(26m 5 e370 ae) BBP tie
itis (1 diy ba Dey a
fh, 1 OD IG oe
ee (Sh 32, 86:88). )
ve ED EO I a)
vi. (56 (2:7 48-2 | 95)

All are below 5:5 cm. in height, lower than the East London series,
and below the flattest of the Bushman’s River material. The two
comparisons given above probably have little value.

NKWALENI.
NMP — 5 Lass 15 12-0 ae 50 3-9 “34 ke.

Havetock, TucEeLta Movurts.
NIMS iene) 15-6 2:7 48-2 Goa i) -19hke:;

Umutotr Bracu, DuRBAN NorRTH.
A.G. Wills . 11-2 3:8 33-9 475
12-2 1:8 14:8 270
10-4 2:0 19-2 215 46
All from same pottery midden site.

4-3 ‘8 ~~ Pebble.
4-8

3°6 Hard black shale.

99 99

WILLow GRANGE, WEENEN.

SEL... . 14:6 5-4 37:0 ee 40 3:0 10 feet below surface.
PIETERMARITZBURG.
TMP 3199... 7:9 10 12-0 5 4.9 40 Anomalous.

* «'n Vierkantige Deurboorde Klipwerktuig,” S. Afr. J. Sci., 1935, xxxii, pp.
474-475.
VOL. XXXVII, PART 1. 3

34 Annals of the South African Museum.

RICHMOND.
NMP 2516 . 12:2 4-7 39:3 a. A eee 1-09 kg.
KARRIDENE.
Sli2Zo: Bed fs EV 3:3 24-1 aie 5:2 19 Described by J. G.
Cramb.
(10-2 29 28-5 = 5:6 .-) From Cramb’s illus-
tration.
Uvonco.
Se ue lO: 9 1:8.) 16:5 oe 34 1:8 10 feet deep.
28s . 12:9 2:0 15:5 ee 4:2 2-6
HARDING.
SHSi) : 8:7 3:2 36:8 ihe Oi eoril

Considerably more exact evidence from this Lowlands area should
yield better graphical grouping, and some clearer knowledge of the
associable implement types. Pottery, microlithic implements, bone
tools, the presence or absence of seashells should all be carefully
noted, and any new evidence sent to the writer or to the nearest
museum, or given reasonable publicity in some acceptable form.
The Transkei is similarly very badly represented, and evidence from
there would be welcomed. |

The relationship suggested above, between this area and other
coastal areas farther south, is interesting. There would seem to be
a flat coastal or lowland type from Natal to East London. The
Bushman’s River series consists only of specimens over 6:0 cm.
thick, and only one specimen has an index below 50 per cent. No
example from NL. reaches either figure. This refutes any idea of a
“Coastal Type” of bored stone, found exclusively at midden sites
along the coast from Natal to the Cape. The truth is far more
complicated than that, as the South Coast Region bears out.

Bibliography and References.

Literature on the Natal and Transvaal bored stones is scanty. One
or two references may be found in descriptions of sites, but with little
interest. Gooch * divided the bored stones he discovered between
the Cape and Natal into three groups: (1) the full sized, 5-7 lb., from
the Cape to Natal, globular in form, and of any material from quartzite
to shale. (2) Small form, 2-3 lb., “angular” like a thick perforated
disc. (3) A small form weighing about 2 lb., globular and occurring
in many materials, less abundant. He mentions the finding of a ring
near Camperdown; it was of indurated shale, and was discovered by

* W. D. Gooch, “‘Stone Age of South Africa,” J. Anthrop. Inst., vol. xi, 1881,
pp. 784-788, illus.

The Bored Stones of South Africa. 30

Mr. Thresh. He describes the finding of what we here call a bladed
disc (apparently his type 2) at a “‘considerable depth in blown sand,
in the Red Hill railway cutting, Victoria County, Natal.” George
Leith * notes the finding of a bored stone near Pietermaritzburg.
Other instances might be mentioned.

A further note is given by J. G. Cramb + in a somewhat unsatisfying
paper. He describes a midden site, with decomposed shell-heaps,
typical midden chopping tools, etc., without a single suggestion of
Smithfield types described or illustrated in the text. He admits that
“it is obvious that there is little or no correlation between these
implements and those of the known phases of the Smithfield culture.”
He suggests that the variation described might have originated in a
coastal movement, and stresses the dissimilarity between this and the
numerous Strandloper shell middens of the Natal and Zululand coasts,
where pottery occurs. Differences between Natal midden sites were
noted by Gooch in 1881. Cramb’s second example, taken from his
drawing, is not included in the list of types above, but is listed
tentatively under Karridene.

Further Reading.

If primarily interested in the coastal belt, turn on to the South-
Hastern and South-Hast Lowlands Regions, and continue reading from
there (pp. 72 and 77).

Ill. VAAL REGION.

One of the great highways southward from the Transvaal region
very naturally passed along the Vaal River valley. As tributary
streams entered the main valley they provided lateral lines of move-
ment, either directly across the Vaal, or as feeding routes along which
(as in the case of the Harts River) new types were brought to the
lower reaches of the great highway. These types joined the slow
migrations southward.

I have thought fit to separate the Vaal region from both the
Transvaal and the Free State, but all three are intimately inter-
related, feeding one another and making the two provinces difficult
to understand. Perhaps the key lies in the National Museum,
Bloemfontein, but I am unable in the present circumstances to obtain

* J. Roy. Anthrop. Inst., vol. i, 1898.

{t “Smithfield Implements from a Natal Coastal Site,” Trans. Roy. Soc. S. Afr.,
1934, vol. xxii, pp. 205-223.

36 Annals of the South African Museum.

access to that collection. It is to be hoped that Dr. H. C. N. van
Hoepen’s forthcoming work on the bored stones will provide data for
a clearer view of the position.

The Schoonspruit appears to provide a feeder to the Vaal, with
little attempt to transgress the river southward. In contrast the
Vet-Marico River area has a double function. While it feeds the
Vaal River valley from the headwaters of the Marico, it also cuts
directly across the Vaal valley and seems to have peopled the northern
half of the Free State. This area is divided into a northern and a
southern stretch, but the division has no value apart from the
geographical distribution.

The Middle and Lower Vaal areas might be paired as another region,
as influences cut across from the Free State focus on the east. I have
therefore made them into a subregion, [IIa Vaal Region (Lower), to
permit division with further evidence.

SCV. ScHOONSPRUIT-VAAL AREA.

This area shows little real relationship with surrounding areas, and
many of the types seem to be anomalous. I do not regard it as
important.

i, (10:8 8:7. 180-5: 1015) | VMReainyenae
th) OS Fa SOG 645 OFS. ii, LV. v.
Ff, (Gel eI) eG 465) MG. v.

iv. 10:0 3:3 33-0 330 VMR. v.
ee GRA 270 VH. v, MV. xii.
vine WON, 1-2 sen 10-0 as

RUSTENBURG: B.

ASJ 49/36/5 . 7-9 3:8 481 230 eo. ole S
9:8 29 29:6 (280) 1-9) 1-8) ise:

LIcHTENBURG: B.

AMG 1760 7:3 4:4 60:3 23505 boo eS

ASJ 7/38 10:8 8 (7-4) e: 5-0 4:8 Flat disc.

13:0 Sz 0) ~ 50 39 Flat disc.

KLERKSDORP: A.

WUA\752 . 98 1:3 (13-3) 7:0 6:2 Slate disc.

894 9-8 AST (PATE) 3:0 2-4 Ironstone disc.

SLURRY, MAFEKING.

MMK 1894 . 7:8 73 93:6 AAD 32D OE 2D

SHEPPARD ISLAND.

ASJ 7/35/67 . 10-9 8-8 80-7 1050 40 3-0 Under 2 feet gravel.
One face ground.
7/35/7 . 106 8:5 80:2 955 54 2-4 Under 4 feet gravel.
7/35/68 . 9-1 75 = 82-4 620 32 3:0
7/35/48 . 10-5 35 33:3 385 2:9 2-4
7/35/47. 7-8 50 64-1 305 3:0 2-0

The Bored Stones of South Africa. 37

SEVENFONTEIN, KINGSWOOD.

MMK 1425 . 8:3 6-9 83-1 475 44 3-2 Ss. grinder.
TYGERFONTEIN 189, HoopsTaD.

ASJ 8/41 9-6 7-3) 16:0 675 (2:9 0:5) - Conical bore.
FouRTEEN STREAMS.

AMG 1352. 89 76 85:4 600 34 18

MIP 7528 . 10:5 16 15-2 of 3:8 3:6 Disc.

Slurry and Sevenfontein show some affinity with the Molopo-
Groenwater area (Western Bypass).

VMR. Vet-Marico RIvERs.

This provides a most important highway from the sources of the
Marico River in the Western Transvaal, across the Highveld, to the
Vaal. There is some movement along the Vaal, but far more im-
portant are the migrations along the Vet, the Valsch, and probably
the Rhenoster Rivers, to supply the dominant types in the northern
Free State. This area shows some affinities with the Transvaal,
and has been divided into a northern and a southern section. The
northern section includes the Marico sources, the Highveld, and the
upper reaches of the Vaal River as far as the Vaal Harts confluence:
the southern is listed from east to west, across the Free State to the
junction of the Vet with the Vaal.

Ie Gy 8 eT A TAO) | OAD

10-6 9-2 86-8 1035 BS. ii, SCV. 1.
loa? 7-6 71-0 870 CES. i.

iv. 11-3 6-6 58-4 845 BS. ii.

vs (@A Sey Ba 330) SCV. iv.

1
ll.

The most striking fact about the area is the paucity of types in
relation to the number of specimens (fig. 4). But just as striking
when the examples are plotted on a graph paper, is the very slight
variation between groups. Except for the last example (presumably
an erratic from the Schoonspruit-Vaal area), diameters range from
10-5 to 12:2 cm., and heights vary between 9-5 and 5:-8cm. The only
other area where an analogous type of distribution is discernible is in
the Central Free State, to the South, but only one type (VMR. iii and
CFS i) agrees. It would seem that we have here the logical outcome
of the theory of copying outlined above in Part I (p. 16).

The extraordinarily high proportion of conical bores in the northern
part of the area is extremely interesting and may prove important.
This will be touched on again when dealing with the Bantu material
in Part V. There is reason to believe that most (perhaps all) were

38 Annals of the South African Museum.

bored with metal tools by Bushmen living in some sort of association
with Sotho tribes, possibly as serfs.

WELBEDACHT.
SAM 4216

ZWARTRUGGENS.
SAM ZW

BULTFONTEIN.
HHC 1046

KLERKSDORP: B.
L. M. Earle

BLoEMHOF: A.

MMK 902
ASJ 5/35/BS.

WUA 1038

HEIGHT.

10-6

12-0

10:5

11-0

12-1
11-5
11-4
10-7
11-7

DIAMETER.

Trext-Fic. 4.—Vet Marico Rivers.

NortHERN (Marico-VAAL).

7-4

8-2

9-5

6:8

MII 1.
CNIOwWwW Pe

70-0

68-3

90-4

61-9

65:1
63°5
61-4
10-9
90:0

830

1180

1050

825

1230
965
910
860
795

(3:0

(3:0

4-5

(4-0

Lo
WOOoonr

1-4)

1-4)

2-6

1:6)

Fee
IR ASOS

Conical bore. Signs
of striation part
way down.

Conical bore.

2-1 kg. Specular iron
ore. 20 feet deep

Ss. conical bore.

Amyed. diabase.
Soapstone.
Conical bore.
Soapstone.
13kg. Ss.

The Bored Stones of South Africa. 39

CHRISTIANIA Town LANDS.
HHC 1045 . 10-9 8:3 76:1 985 49 2-9 Diabase or andesite.
Top of flood-plain
alluvium. Outside
smoothed. Used
as grinder,
Van AASWEGENSHOEK, CHRISTIANIA.
MMK 1845 . 11-2 77 3868-7 965 4:7 2-6 Dolerite.

This last specimen was found with an Archidiskodon. See Dart,
Sain Je isci., vol. xxvi, 1929; p: 71d.

SOUTHERN (VET-VAAL).
BAVIAANSKRANZ.
PEM 614.V. 10:5 7:8 74-2 860 59 3:5

HARRISMITH: B.

ASJ 7/41 - 10-5 SoZ OL On LOLa I Ae 222
10-7 8-7 813 1000 45 3:8
10-5 75 71-4 830 52 4:0

KRANSFONTEIN, HARRISMITH.
Univ. Stell. . 11-7 8:6) a: o) Sea feta 2. 2h

PaRYS.
MMK 2147 . 11-6 Ci 61-2 955 5:3 46 Cyclindrical bore.

KROONSTAD.
MMK 1299 . 11-2 8-4 75:0 1055 5:5 2:9 Ss.
Univ. Stell. . 11-2 63 56-2 790 49 3-7
9-4 3°7 39-4 330 2:8 2-6 Shale. Erratic from
SSV.
WINBURG.

TMP 6468 . 10-5 TO oe 870 41 2:7

In Parapiso, BRANDFORT.
Univ. Stell. . 11-1 66 59:5 815 4:5 3:3 Shale. Pecked sur-

face.
NAPIER, BosHOF.
MMK 279 . 11:8 83 70:0 1155 4:5 2-9 Ventersdorp traver-
tine.
BuTHABUTHA.
SAM 1758 . 10:5 75 71-4 825 49 2-5

NORTHERN BASUTOLAND.

SAM 466 . 108 9- 84:2 1060
7

1 2:6 Quartzite Ss.
11:0 ‘6 69-0 920 3°3

Ss. Nodule.

He oo
Hs Oo

BS. BLOEMFONTEIN-SEPANI AREA.
This area is somewhat detached from the southern Free State,

and shows affinities with the Vaal. It is possible that it has affinities
with the south-eastern Free State, and that it was the route by which

40 Annals of the South African Museum.

material passed from the Vaal, along the western side of Basutoland,
then up the Orange to the Transkei. More evidence is essential to get
this clear. On the other hand the similarity with some of the
Winterveld types is abundantly clear.

1 13-0 8-4 64-6 1420 VH. ii, KA. i, NH. iii, TK. ii.
ti. (10-5 9-2 87-2 1035) VMR. ii, UTV. ii.

rhtig | (UL loe/ 6-5 55:1 900) VMR. iv.

ive (8:3 8:5 102:4 585) KA. iii, WV. vi.

v. (88 6-2 70-4 480) KA. ii, WV. vii, SEFS. vii, UTV. v.
vi. + (5° 3°8 69-4 108) KA. v, WV. ix, SEFS. xii, TK. v.

It will be seen that most of the comparisons for the first three types
lie to the north, while those from the latter three lie to the south.
The Transkei shows affinities with types i and vi. This last is also
recognizable in the south-eastern Free State. ;

Apart from the last type, all specimens exceed 8 cm. in diameter,
while most exceed 10-3cm. It becomes obvious, if we plot this series
against the Vet-Marico series, that individual examples from here
coincide more nearly with VMR. specimens than the type averages
would suggest. One last question needs comment. What is the
relationship between type BS. i and the Natal Highlands type in?
The fact that the bulk of the southern distribution of the Vet-Marico
River area and the Basutoland massives lie across the road between ~
suggests very strongly that the present area merely represents a
divergent branch, originating in a common source to the north,
probably ET. 1, or CT. ii, both of which show similarities to the Natal
Highlands, but less with the present area.

BLOEMFONTEIN.
TMP 1873 10:8 94 87:9 1095 3:3 2:5
1875 12-4 6-7 54:0 1030 69 3:3
ASJ — 8:3 8:°5 102-4 585 4-4 ye
TMP 1850 5:7 4:7 78-9 146 29 2:0
SEPANI.
SAM 2474 . 13:5 85 62:9 1550 75 £4:3
13:5 Wess © lazy ADA ab) Bo
12:1 86 71-1 1260 5:3 3:2
10:4 90 86-4 Cae) 425 32
11-1 6:3 56:3 UH, xO) Bez
8-8 6-2 70-4 480 41 3:2
4-9 2:9 59-2 70 2:0 :

The extreme paucity of material from this area is very much
regretted. It is possible that the averages given above have some
relation to fact, but little real value can be placed upon them until
Dr. E. C. N. van Hoepen publishes his work on the bored stones in
the National Museum, Bloemfontein.

The Bored Stones of South Africa. 41

VH. Vaat-Harts ARBA.

We are now getting towards the middle reaches of the Vaal. The
most striking fact is the preponderance of examples over 11-0 cm. in
diameter, with a marked step between these and smaller types—
presumably representing a division between the digging-stick stone
_ andthe knobkerrie head. The larger elements recall, in a very general
way, the VMR. and ST. areas, implying that the types were intended
for the same function in all these areas.

ie ss) 11-0 79°8 2125 MG. i, NWT. i, MV. ii.
i 13-0 9-2 70-8 1555 IBS sure, MV Ectvi
ii. 11-4 9-8 85-9 1275 MF. u.
iv. 133 6-8 ol-1 1190
Vee (i-D 3°8 50-6 215) MV. xiii, SWT. ii, NES. iv.
vi. (5-0 3:5 70-0 87) NWT. vii, BS. vi.

Type 1 might reasonably be redivided into two types 13-4 x 8-7 cm.
and 12-8 x 9-8 cm. respectively. The former shows similarities with
BS. 1, the latter with ST.iand MV. iv.

SCHWEIZER REINEKE.

HHC 1043 . 13:0 100 77:0 1690 5:0 3-6 Diabase. 3 feet deep
on gravels.
PNP Si22 127 6-7 52:9 1080 4:2 2-2 Harts River gravels.
MMEG'993... (7-3) 41 56-2 218 3:1 2:5 Amygd. diabase.
Rolled.
BLoEMHOF B.
ASJ 13/41 . 14:5 11:4 78:6 2420 49 1:6
138 1l-l 80-4 2115 65 3:5
S/3o/80 . ) 12:0, 10:2. (85:0 1470 4-2 ae
MINT. 12-9 8-2 63°6 13865 5:9 3:3 Ss.
Asad aaojso). 11-0 10:0 90-9 1210 46 2:7 ‘Ss.

LEEUWKRAAL, HOopstTaD.
ASJ 7/35/6 . 13:4 75 56:0 1345 51 2:7 #£Surface.

BABERSLAGTE, HoopstaD.
ASJ 7/35/1 . 13:3 6-9 51:9 1220 47 2:9 6 feet deep.

HOopstTaD.

MMK 999. 14:0 G:2n 20 em 43 2-5 Ss)

CHRISTIANIA.

HHC 1045 . 13-6 95 70:00 1755 45 3-8

MMK 1292 5-1 3-9) 16-5 101 29 2:0 Amygd. diabase.
Grinder.

BETHELPELLA.

ASJ 5/35/BPS 48 3-1 64:6 7 2-7 1-2) Surface.

KOOPMANSFONTEIN.
MMK 418. 12:8 100 78:1 1640 46 2-9

42 Annals of the South African Museum.

NEWLANDS.
MMK 389. 134 105 784 1860 54 3-1 Ss.
Ms cha 4: 86 642 1520 50 3:8. Ss.
1249 . 12-5 9-277 73°61" 1440" 95-1) 3-5) Sst
1132 12 922) 82-1 5085-0) 0 92-30 Sse

VaAL-Harts CONFLUENCE.
ASJ 4/40/22 . 7:7 3:7 48-0 2G 3-2.) 2-5

Illa. VAAL REGION (LOWER).

The Griqualand stretch of the Vall River is marked by a variety
of types, some migrants down the Vaal, others cross movements to or
from the Free State Focal Area. A more exact knowledge of proven-
ance is essential, especially as to the relation of specimens to the river,
left or right bank, etc. At present the picture is somewhat kaleido-
scopic.

The region is a subdivision of the Vaal Region, and common types
hold the two areas (here called the Middle and the Lower Vaal) into
one whole, difficult to disentangle into two true areas. It is obvious
too that sites such as Rietpan and Botterhoek present types that show
affinity with Smithfield, Postmasburg, and Phillipstown. These
widely separated sites will be dealt with under Migrations. The
column of comparisons will suggest the interweaving of movements
to and from this stretch of the Vaal and surrounding areas.

MV. MippuLe VAAL.

i. (14:3 12-2 85:5 2500) SEFS.i, MG. i, WV. i.
rile Vo 11-4 80:7 1760) ) VER, SE Vesa
iii. 11-4 11-6 101-7 1505 LV. i.
iv. 12-5 9-6 76-9 1500 VH. ii, KB. ii, SEFS. iii, CT. ii.
v. (12:5 8-0 64-4 1260) BS.i, KA.i, NWT. ii.
vi. 11-5 9-0 78:3 1200 MF. iii.
vil. 10-5 7-9 75:2 870 KB. iv, ET. ii, VMR. iii.
vill. (12:3 4-6 37-4 700) Anomalous.
ix. 11-0 5:5 50-0 665 LV. iv, CT. v.
xe ERB} 5:0 53:7 430 AG. i, ST. ii.
ale) | THs) 6-3 87-2 385 LV. vi, KB. vi, MG. v, CES. iii.
MA) de? 4-] 56-9 213 SCV. v, ST. iii, VH. v.
xiii. 6-5 5-0 76-9 210 KA. iv. °

While it is obvious that sites from Barkly West to Gong-Gong and at
Kimberley show an overlap of MV. and LV. types, these have been
included here. Material from Katlani, Rietpan, etc., which would fit
better geographically into the Lower Vaal area, have also been
included here, as their graphical scatter conforms better. As our
exact knowledge of this whole province increases it may be better
to work from the geographical distribution of types and develop
new areas.

The Bored Stones of South Africa. 43

RIVERVIEW ESTATE.
ASJ 5/35/55 . 11-0 8:2 74:5

16/40 . 10-0 8-1 81:0

WINDSORTON.
ASJ 15/41 . 116 11:3 97-4
11-6 8:9 78:8
SAM 1839 . 12:3 4:6 37-4
ASJ 15/41/19 W.10-8 5-7, 1) o2'8
SAM 1839 7:0 Don oak
SLYPKLIP, WINDSORTON.
ASJ 20/43 . 9:6 8:3 86-4
RIVERTON.
ASJ 36/32/16 12-5 (9-5) 79-6
MMK 1294 11:8 9-2 82-1
8:0 6:20 1-5

Goop Horr, PNIEL.
MMK 1488 . 11:2 108 96:4

PNIEL.

ASJ 61/88 . 11:5 9-0 78-2

5/85. 6-2 5:7 91-9

KANTEEN Kop.

ASD KK, ¥.. 6-5 4:4. 67-7
WINTER’S RUSH.

SAM 791 Basel! 66 81-5
BarkKLty WEST.

ASJ 22/41 13-2 116 87-8

WC 297 2 12-5 82 65-6

MMK 1492 9:8 4:4 44-9

WOUA 247 7:5 (4:5) 60-0

BaRKLY WEST TO GONG-GONG.
Van Alphen . 13-0 9:8 75-4
11-5 12:3 107-9
17s 10-2887 -2

12-4 9-2 74-3
12-5 79 63-2
10-9 74 67-9
8-0 6-5 81-2
7-9 64 81-0
6-9 6-4 92-8
6-9 5-1} 73-9
RIETPAN.
MMK 607. 145 12-4 85-5
2608 a ls-On e112: 86-2
26 van 1 93 5-2 55-9
489 9-1] 50 55:4

990
810

1520
1135
700

665
260

765

1485
1280
400

1355

1190
220

186

435

2020
1280
425
255

1655
1625
1415
1415
1240
880
415
400
305
245

2610

1900
450
415

4-0
4-5

CoH G2 OD
bo oO bo CO C1

bo Or

2°8

G2 HR G9 09 GO OT OVE O9 O09
OS — CO © CO HP ©C CO 00

7.

He O2ct
Sye5ien Er

ee WNW he wea
DaAODROARD: >

—

Ree ee
00 00 bo

Sohnge, Visser and
Lowe. Site VI.

Cylindrical bore.
Anomalous.

Ss.
Ss.

Ss.

Erratic. Cf. UH. v,

Radial cuts at
mouth.

Ss.

ad. Annals of the South African Museum.

KIMBERLEY.
MMK 1879 . 13-1 9:6 73-3) G50 o23 ole SS
1898 . 12:9 9:3) ik. La OOU eA) yer Ona S se
ASJ 4/40/21 . 10:5 7:2 68-5 TIONS 8 75) is
ASJ 8 . Bey Mlle, 5:5 49-1 690 3:3 2:7—
a j 9-4 D2 OOO 460 3:9. 2:7
MMK 1296 8:1 6-1 75:3 400 3:5 19 Ss. Grinder.
ce eee 5:9 177-6 340 3-7 1-9
ASJ 4/40/21. 7-4 5:8 78:4 5) Ro es 9 a
MMK 1296 6:4 4-7 73-4 wO2, Sl | 18885 Sse
ASJ 8. 6-9 Beis! | Oey 167. 30 2:5
AMG 1732 8 4:8 82:8 161 2:8 ‘9
SKANSKOP, KIMBERLEY.
MMK 1850 .— 8-8 5:1 58:0 400 39 3-0

Riet River, KIMBERLEY.
SAM — . 10:0 85 85:0 850 42 3:1

BoTTERHOEK, JAKOBSDAL.

MMI 997 92 14:2) 12:1 85:2" 2440" 4-7 © 2-8
1180 . 11:0 5:6 51-0 680 5:2) 2:6 7 oSs:
a - 68 4:7 69-1 Jie 3-2) 2:0) Sse

ROorroort.

ASJ 36/35/15 = 11-7 SI it 8 er 24 6-0 ee
10-8 LO rion 920 54 2-5
6-2 2°5

8-4 73°8 440 4-0
SCHMIDTSDRIF.
MMK 733 12-3 110-0 (81-3. 15154-39124 esse
SAM 4957 . 10-6 8:4 79-2 945 5-1 2-9

7-2 4:1 57:3 210 42 2-6
MIDDELPLAATS.

MMK 1488 . 10:8 5:2 48-2 605 3:9 2:3) iss:
KatTLAnt, Dovuguas.
PEM 614/0 . 11:3 90 796 1150 53 3:4

MAZELSFONTEIN.
MMK 1803 . 12-0 98 81-7 1410 45 3-1

The fact that certain of these Middle Vaal types recur at Douglas
and other sites below the Modder River junction suggests that these
are true Vaal types (MV. ili, ix, xi). They conform to LV. 1, iv, and vi
respectively, confirming the suggestion. The remainder probably
represent lateral movements into this valley, generally from the Free
State region, but sometimes from the Vaal-Harts, Southern Transvaal,

etc., suggesting that these last are also intimately related to the
Vaal valley.

LV. Lower VAAL.

Careful comparison of this area with the Middle Vaal, either geo-
graphically or graphically, will at once show that the division is

The Bored Stones of South Africa. 45

hardly justifiable at present. Sites in both series occur at Douglas,
etc., while types often correspond. There is a difference present,
but the MV. series has been used largely as a receptacle for the sites
showing mixed types, the number of averages is therefore greater
in that series. I suspect that we are dealing with material not from
two stretches of the river, but from two banks of the river. The
picture is not so simple as that, for at times, especially along the
stretch west of Barkly West village, the same types occur on both
banks. These two areas must yield, with further and more exact
knowledge, to more reasonable adjustment.

(IEG 10-@ 9551) 455) MVE aii:
ii. (13-4 60 448 1080) VH. iv.

iii. 10-3 9:0 87:3 955 BS. ii, SCV. i.

iv. (10-7 60 56-1 685) MV. ix.

v. (91 75 82-4 620) KB. v, SCV. iii.
vi. 8-2 5-7 70-0 385 MV. xi, KA. ii.
vii. (5-9 ley USS Ne NC hIKEA ye

While the MV. series shows the more general distribution, the LV.
material is more closely related to the Kaaienveld (p. 65).

Gone-Gone.
MMK 991. 6-4 16 25-0 66 -* 2-6 ~2:2 Ss.
DovUGLAS.
MMK 929 . 134 60 448 1080 55 33 Ss.
MSs 90 63 700 510) 44° 24
999°. 8:0 62 775 400 38 2:8
7-4. 5-6 75:7 305 34 1-9
1666 5:3 1:8 34:0 Y= Sap 1 Dips
St. Ciatr, DovuGLas.
SAM 702 . 11:6 11:0 94:8 1480 40 2-9 Ss.
10-0 8:5 85:0 S00), 3°59) 3:3
11-0 6-4 58-2 TO a2 ;
9-6 7-7 80-1 710 as ; Started.
MMK 258 . 10-4 5-5 52-9 By ZS Bll ISI,
SAM 702 i 8-6 74 86:0 SYED | Fats) | pI
8-2 6-1 74-4 ANO: 73:7. 2-2
8-0 5:3 66-2 390 84:5 ots

- BROADWATERS, DOUGLAS.
MMK 2105 . 10-2

Ne)
Or
Ne)
oY
i=)

990 3:4 1:5 Soapstone. Partly
rebored later.

WITBOOM, GRIQUATOWN.
MMK 1291 . 10:8 90 833 1050 3:8 2-2

GRANGE.

PVE ye eS 10-7 9027") 149076:0 3-3
iS 8:0 5:3 62-4 385 3:8 2-5

In the McGregor Museum collection at Kimberley, a few outstanding
specimens occur. Some years ago Mr. H. G. Bryant of Prieska

46 Annals of the South African Museum.

presented the collection with a complete digging-stick. It consists of
a large bored stone mounted on a stick, perhaps an inch in diameter
and a yard in length; the end of the stick is shod with a buckhorn.
Three specimens from Griqualand West appear to be mace-heads,
possibly of a ceremonial nature. They are of hard igneous rock, tooled
over the whole surface. One has been tooled to a pyriform shape,
markedly pointed at one end; at the opposite pole there is a flat
platform, obviously prepared for boring. The remainder are ovals,
bored into one face only, presumably to take a suitable stick.
(Compare UTV. Aliwal North; type vii, SAM 1361, where the
example measuring 6-3 x 6-3 cm. is of analogous pyriform type.)

Illustrations of further Griquatown material can be found in EH. J./

Dunn’s book,* but the material itself is unhappily in Australia and
the figures are not available.

IV. FREE STATE FOCAL REGION.

This region has had to be rather awkwardly divided. A short
study of the Koffiefontein material will explain quite clearly why this
subdivision is necessary. It is due to the paucity of evidence of
exact provenance. All material should be referred to the smallest
unit of land, viz. a farm. SBoshof district covers about 5000 square
miles of territory, while Fauresmith district covers a little less.
Material marked by districts is therefore as precisely placed as material
from Britain marked by counties. Yorkshire with its 6000 square
miles compares with many districts in South Africa. Such terms as
Smithfield mean little more than this. Compare the areas given in
the beginning of this part (Part II, p. 23).

What makes the position even more difficult is that this region is
focal. It provides the area of attraction that seems to have con-
centrated and mingled every movement that passed between the
Basutoland massives and the Kaap Plateau, a belt 300 miles wide,
representing a bottle-neck of migration between the arid Kalahari
and the impassable mountains.

This region consists of the southern half of the Free State, and the
basins of the Modder, Riet, Kaffir, and Caledon Rivers, together with
the northern basin of the Orange and the junction of the Vaal with
the Orange River. At this last point the types seem to coincide with
the northernmost distribution of the Trekveld-Winterveld area.

* K.J. Dunn, The Bushmen. London, 1931. |

The Bored Stones of South Africa. 47

Immediately west of this region lie the Middle and Lower Vaal areas,
while to the north are the northern Free State and the Bloemfontein-
Sepani areas. To the east the Maluti range of the Drakensberg
confines the area absolutely.

As one would expect of a region served by a network of rivers,
divided only by low plateau country, and lying on the arid side of the
Drakensbergen, these perennial streams (though considerably reduced
by the end of the dry winter) have attracted man from all directions.
Cultural traits converge on the streams and cross and recross from a
number of sources, meeting and mingling, and at times combining to
form new cultural patterns. They create a jigsaw of distribution
difficult to deal with on a map.

My task has been rendered more difficult in this region, as war
conditions and service have prevented me from visiting the McGregor
Museum, Kimberley, for the purpose of measuring specimens. The
museum itself has been shorthanded and incapable of undertaking the
work. For the great quantity of additional material I have received
from this source I am indebted to an old student, B. D. Malan (now
attached to the Archaeological Survey at Johannesburg), who took
advantage of his presence in Kimberley to obtain this valuable
information. My earlier notes did not represent a tithe of the material
thus made available. I have been dissuaded from making use of
material from the National Museum, Bloemfontein, as Dr. E. C. N.
van Hoepen tells me he intends to publish this at some future
date.

Professor C. van Riet Lowe has shown this region to be the heart
of the Smithfield culture, especially of phase B. Indeed to under-
stand this region at all clearly the reader is referred to papers by that
author,* while a site list is given later. Several of the sites listed
here are described in these papers, and some knowledge of the relation-
ships existing between the various phases of the Smithfield culture,
the Wilton, and the bored stone can be gathered from those sources.
Other literature will be mentioned as we proceed. This is the focal
area of the Smithfield B culture, and most of our examples are
associated with this phase.

* Lowe and Cable, *“‘The lithicultural horizon . . . of the bored stone in the
Orange Free State,” S. Afr. J. Sci., 1927, vol. xxiv, pp. 506-513.

C. van Riet Lowe, “The Smithfield Industry in the Orange Free State,’”’ Ann. S.
Afr. Mus., 1929, vol. xxvii, pp. 151-234.

Sohnge, Visser, and Lowe, “The Archaeology of the Vaal River Basin,” Geological
Survey, Mem. 35. Pretoria, 1937.

48 Annals of the South African Museum.

K.A., K.B. KoFFIEFONTEIN AREA.

The material from this ill-defined area, including parts of the
Jakobsdal and Fauresmith districts, presents a very peculiar problem.
If all examples are taken together they cover the range of the Free

DIAMETER.

HEIGHT.

Fic. 5.—Koffiefontein A.

State types fairly comprehensively. If, however, they are divided
into two groups, related to the collections made by independent
institutions, the picture is very different. The series is therefore
arranged in two series, each with its own averages. The first series
(K.A.) is from the Archaeological Survey, Johannesburg, and from
the University of Cape Town collection only. These are from the
same sources, as I collected the latter example with Professor van
Riet Lowe from his sites. These are mainly from the Fauresmith

The Bored Stones of South Africa. 49

district, on the Koffiefontein town lands and neighbouring farms
(fig. 5). The other series (K.B.) consists of material sent in to
the McGregor Museum (apparently from the Jakobsdal side of
Koffiefontein) and to the Anatomy Department of the University of
the Witwatersrand, from comparable sites (fig. 6).

DIAMETER.

Fic. 6.—Koffiefontein B.

The result of this separation of material yields a picture of alternat-
ing types. There are two explanations possible. Hither we are
dealing with the function of two different distributions at Koffie-
fontein, or alternatively we are dealing with the same series at different
stages in the process of serial copying. If the former is true, the
parallel between the two series is the result of a remarkable coincidence.
If the latter, then we are presented with the products of a slow change
induced by copying over a period of time. We can tentatively regard
the two series as being chronologically and spatially separate, though
it is impossible to suggest which may be the earlier.

VOL. XXXVII, PART 1. 4,

50 Annals of the South African Museum.

It is at once evident that in a congested focal area, such as the
Free State, more exact evidence of provenance is essential. It may
even be necessary to mark specimens from a single apparently homo-
geneous site with such additional data as North, South, Central, etc.
A survey of a large area by the methods laid down in Method in
Prehistory (South African Archaeological Society, Handbook I,
1945, pp. 96 et seg.) may considerably increase our knowledge
of a local series. The extension of this ethnological technique
to a series of test sites, each a hundred yards square, would
provide evidence that could be checked by collation. The collection
of data in the field should not be difficult, the real work consists
of collation of material later, which would provide a major task.
Such evidence is not available at present, and is not likely to be
obtainable for some time after the war. The average figures for
the two groups are therefore given, and the data is separated
accordingly.

K.A. KCB.
in 12:3) (98:3) 967-5) 9 1255 i. (14:7 10-1 68:7 2185)
ii} 18'S.) 05:07) 167-0) Hos ii. 126 10:1 80:0 1605
iii. (7-5 7-5 1000 420) iii. (12-8 73 57-0 1195)
iva 6:6 TOs eos iv. 10-0 7-8 78:0 780
v. 56 39 700 122 v. 85 76 89-4 550
vi 65 1:7 261 72 vi. 7-5 59 78:9 330
vii, 7-5 49 653 295
vii, 5-5 45 81-9 136
ix. 4:8 31 64-6 71
x. (3-5 2-3 65-7 28)

While no comparisons are given here, other areas and groups are
compared to these two series, and a cross reference should quickly
show the relationship of the various surrounding averages. Koffie-
fontein B provides the material for the Winterveld migration.

KoFFIEFONTEIN A (ASJ anp UCT onty).

ASJ 10/41 . 13-5 90 66:7 1640 58 2:6
64/38 . 12:0 °(8:0) 66:7 1150 5-5 a
LOA ey 2 8:0 714 1005 54 3:5
G4/38 0 eee OG) aed 525 3:0 2:8
8-9 6-3 70:8 500 3:9 2:9
9-1 6-0 65-9 495 46 2-7
Ter 8-2 106-5 485 47 3-6 Hammer.
10/41 9-4 5:5) ~ 60-1 485 4:2 2:5
64/38 8-6 6-1 70-9 450 4:2 2-7
(9-0) (5:5) 61-1 445 30 1-5 Fragmentary.
8-2 5:8 70°77 390 45 3:4
8:5 5:3 62:3 385 40 2:2 Grinder.
Te 6-9 94:5 370) 424 O°7

The Bored Stones of South Africa. 51

KOFFIEFONTEIN A (ASJ anp UCT onLy)—continued.

ASJ 64/38 6:7 5:2. 776 935 (8-1 2-4
68 45 662 205 35 24

64. 4671-9, 9 190 | 2-9) 1-7

29/35 63 46 73:0 18 39 23
10/41 6:0) 863.3) 1 137 282-8) F1-2
64/38 DoS LOG 7 | 123.5 3-0 18

By aoa Gao TT 3-0. 18

50 44 880 110 28 23

WOMI5197. 61 2-0 32-8 Tle OD ES
ASJ 64/38 70 13 186 64-3308

I suspect that the first specimen listed above represents a group of
its own, and that it really belongs, in spite of its accession number, to
series B. I make this deduction from material listed later under the
Winterveld area, that suggests closer affinities with Koffiefontein B
than with the A series, though types represented in both are present
in WV.

KOFFIEFONTEIN B (MMK anp WUA ownty).

MMK 1310 . 14:7 10-1 68-7 2185 42 2:0
Wi 226). 12-8) 10:3 80-5 - 1690 Gl 3-7
1100a. 12-2 99° 81-47) 4-5 5 Ss. 1-84 kg.
MMK 1432 . 11:7 105 89-7 1440 43 2-6
WUA 235 . 11:8 96) "8i-3 ) 1335), 5:4 3:4 1-8 ke:
MMK 1311 . 12:8 73 57:0 1200 4:8 3:2 Mudstone.
1310 . 10:5 Onn LO 872 3-6 ae
WUA 232 . 10:0 8-1 81-0 810 43 3-0 Irreg. Ss.
236 a7 7:2 74:2 680 43 2-4
MMIKGWSI3 >. 8:7 Sosa G15, 36 oo:
WUA 11008. 8-8 7:8 88-6 605 4-1 : Ss.
HOOE 8-7 76 87:3 GD) AR Se
789A. 8-4 76 90-0 5385 45 3-1 Subcylindrical.
231 8-3 7-1 85-5 490 4-1 2-2 Mudstone.
MMK 398 7:8 6-2 79-5 SoO Mach 2.0
PSLON 7822 5-4 65-9 365 3:8 2-4
WUA1100F. 7:4 6-3 85-1 345 24 .. 0-45 kg.
HOOD. 1-7 5-7 74-0 340 24 41:8 #£Ss. 0-54 kg.
230) 1D 6-0 80-0 S30) 2-400) 3.
MMK 1791 8-1 4-9 60-5 320 46 2:6
2638 7:8 4:9 62:8 300 3:0 2:3
WUA 227 74 5:0 67:6 21 POS) Goro: SS
234 7-0 56 80-0 Zion wie 02 8 223
MMK 1432 7:0 4-6 65:7 259) 3-99 2-7
WUA 7898 55 5:2 94-5 155 30 1:7
MMK 1745 5-5 4:2 76:4 127 2-6 1:5 Four faces ground.
WUA 237 5-4 40 74-0 Phe oF Olea! ta OSs
MMK 1536 5:5 2-7 48-7 83) 3:6) 2:0
1637 4-7 Be ther 82 2:5 1-4
1717 4-7 3:1 66-0 68 2:6 1:2
1537 4-] 36 8678-0 Ol) 1-5: 70:6
2227 4-] 3-1 75-6 Bye) apres © Le
1891 3-5 PAB (CSET) 282-1) 0:9

52 Annals of the South African Museum.

OTHER FREE STATE SITES.

The extraordinary richness of this area, and the vagaries of collec-
tion make it impossible to give any real picture. What additional
evidence is available should be collected by the various institutions
concerned before it is utterly lost. All that I can hope to do here is to
discuss those patches that appear to provide culturally “pure” series,
in order to provide data for the isolation of types by distribution when
further more exact evidence is available. Sites are therefore grouped
by typology rather than on geographical grounds. With nomadic
peoples wandering in relation to the varying availability of water in
an area such as this, that is all we can at present hope to do.

Many types can be linked with the Middle and Lower Vaal areas.
Some types are very “dominant” or persistent, others appear to be
recessive, and to occur in one or two areas only. Perhaps this is
important, more especially when types persist over great areas or
cover long distances, such types should repay watching.

The main reason why the Free State is such a difficult area lies in
the plateau structure of almost the entire area. This plain is cut, ©
never very deeply, by a number of rivers flowing west to the Vaal,
to yield a down-like countryside, never impeding migrations and
minor movements in any direction. The reasons for analogies on the
Middle and Lower Vaal should therefore be obvious as well.

CFS. CENTRAL FREE STATE GROUP.

This group is mainly characterised by the relative closeness of the
averages. All diameters lie between 7-5 and 11-3 cm. and all heights
are below 8-0 cm. This gives the impression that a distribution such as
that given above for the Vaal-Harts area, and a scatter such as this,
would cover most of the Free State types. Alternatively the NWT.
area added to this group would yield much the same result. The CFS
material thus appears to be a fundamental component of the region.

i. (11-0 7:0 63-6 845) VMR. iii.
iis © 3925 75 79-0 680 MF. v, KB. iv, MFG. v, LV. iii, SCV. ii.
ii. 8-4 6-5 77-4 460 SEES. vii.
iv. 8-3 4:7 56-6 325 SEFS. viii.
EAGueE’s Nest (PETRUSBERG-BosHoF RD.).

WA 669) LIEl 77-00 163-10e) SG5n 46 29

ASJ 11/40 . 99 73 73-4 680 49 3-6 Grinder.
94 75 800 665 44 2-7

SAM 4601 . (85) 63 741 455 (3-8) (2:7) Broken.
8-5 AuTi 63:b0) vol 5eeea6 (ea)

ASJ 11/40 8:2 16:6 180-5 445) 2-0) | 12-6

SAM 4601 2 7-6 sola S408 vag Flog

M. C. Burkitt’s example from this site (South Africa’s Past in Stone and Paint,
p. 94) is much smaller, measuring approximately 5-7 x 4-7 em. Compare KB. viii.

The Bored Stones of South Africa. 53

BLAAUWHEUVEL.
ASJ 25/35 8:9 6:5 73:0 515 45 3-0
8-7 4-6 652-9 350 (3-5 2-3) Conical bore.
8:4 4:9 58-3 3450) oa 2:6
8:2 4:9 60:5 320 44 2-9
DRIEKOPSEILAND.
ASJ 14/42 . 9:5 (eS Lo 695 3:9
GOEIEMANSBERG, LUCKHOFF.
WIWIACA/39 . 9°77 7:3 75:3 685 34 1-1 Ss. -82 kg.
LUCKHOFF.
SAM 4684 . 9-2 74 80-4 625 40 2:3 Shale or Ss.
PLOOYSBURG.
MMK 2636 . 8-4 4:4 52-4 310 73:0) 225
EmMavs.

MMK 1294 . 11-0 71 64:5 860 5:8 3-4 Shale.

AG. ALEXANDERSFONTEIN GROUP.

As would be expected this group shows some relationship to
Kimberley. Only two types of any value are represented. All
specimens are small, the largest having a diameter of 9-0 cm. This
suggests a general affinity with CES. iv, with the addition of even
smaller elements.
(9-0 5:2 62-8 425) MV. xi.

ms ((7e33 2-8 38:3 150) Anomalous.
(46 2:2 47-5 48)

Banx’s DRIFT.

MMK 1897 . 9:0 4:8 53:3 400 4:2 3-0 Epidote diabase.
1184. 4:5 2-4 53-3 DO 2:2 2S SS:
ALEXANDERSFONTEIN.
ASJ 36/35/1 . 9-0 5:6 62-2 450), “4022 4s
. 7:3 2°38 38-3 150 2:3 1:9 Anomalous.
4-8 2:0 41-7 46 2:0 1:6

MF. MoppER-FAURESMITH GROUP.

These sites may be taken somewhat uncomfortably together. I
can see no reason why Fauresmith district should associate itself
with the Modder River, and yet show no real relationship with the
Riet River which lies nearer. The Modder River site is immediately
below the junction of the Riet and Modder, where the railway crosses
from Kimberley southward. The nearest point in the Fauresmith
district is perhaps 30 miles, and part of Jakobsdal district intervenes.

ig ood 11-2 85-5 1920 MV. ii.

ii. 11-4 Sz 85-1 1260 als sits
iii. (11-5 8:5 74-0 1125) KA.i, VMR.i, SWT. i.
iv. (9-2 10-5 114-1 890) Anomalous.

We oX0) 74 82-2 600 KB. v, SSV. ii, CFS. i.
vi. 6-0 3:8 63-3 137 KA. v.

54 Annals of the South African Museum.

Mopper RIVER.

MMK 1297 12h LL - Gs 2 S02 18704: Gi 2-0) seis:
A049 3 11-9) 10:26 28): 7) AAD onl oul Ss.
Dain | les 9°62 5 S2:08 Ost de 5:G ose Ss.
SAM 815 . 11-5 86 746 1140 49 3:6
MMK 756 11-4 S47 iat 10999) Sea 973-0
WUA 607 . 9-6 6:9 72:0 G25 me D a as
MMK 192 . 6-2 36 38-1 lipishy, oe iiss}
FavRESMITH DISTRICT.
ASJ 26/35/5 . 13:7 11:0 803 2065 5-7 3-3
66/38 = 13-0. 11-0), (84-65) 1860) 5-8 93-7
23/39 . 10d (9:6) 91-4) 1060) 3-2) 02-7
92 10:5 114-1 890 3:4 2:6
28/36, -% { (9:0). (7-5)? 283°3 (G05): 02:7: 231 Fragmentary.
28/35. 9-] 7-1 78:0 590 42 2-6
6:0 Ae Sema miler t55)) 3422-0
6:5 3°55 86-38 VAT ssn
KLIPPLAAT, FAURESMITH.
WUA A/81 . 6-1 4:4 72-1 165 2-7 16 8feet deep. Diabase.
AVALON.
ASJ 9/38/11 . 8-9 8:0 90:0 635 42 1:7 Surface pecked. One
face flattened.
RamMauH MANorR.
MMK 185. 5-2 3:4 65:4 92 28 1:4

SEFS. SoutH-EASteERN FREE STATE.

Held together by the Smithfield series, these sites show some
homogeneity. The Smithfield material appears to be of less real
value than was at first thought. I have been able to discover that
Dr. Kannemeyer was in the habit of leaving collections of stone
implements at the homes of his country patients, from where they
would be collected at a later date, put together and sent to Dr.
Péringuey at the South African Museum. It is perfectly possible
that the bulk of the collection is from the neighbourhood of Smith-
field, but it is probable that individual examples came from neigh-
bouring districts—Bethulie, Rouxville, and perhaps even from
Burghersdorp and Aliwal North, where he was a fairly regular visitor.
These last two districts lie on the southern side of the Orange River,
outside this region.

i. (14:2 12-4 87:3 2500) Bethulie dist. MV. i, WV. i.
li. 12-4 12-] 96-8 1860 MV. iii.
iii. 12-5 10-0 80-0 1565 KB. ii.
Tie WARD 7:5 62-5 1080 KA. i.
we (GE) 8:8 88-1 860)

was) ORL 7-0 69-8 715 KB. iv, WV. v.
Vil SO 6-3 70-1 500 KA. ii, WV. vii, CES. iii.
Vili, 87 5:3 60-9 400 KA. ii, AG. i, CFS. iv.
1x.) 74 5:7 77-0 310 KB. vi.
xea” 0:0 4-4 13°3 158 KB. viii, WV. ix.
oath (F/G) 2-0 26:7 -- ) Herschel only.
XTi eo-O 2-9 58:0 67 KGB ix Vie xe

The Bored Stones of South Africa. 55

This area shows a fairly considerable resemblance to the two series,
Koffiefontein A and B, taken together. Types vii, vill above are a
twinned cluster. Type v shows a general resemblance to individual
specimens in NWT. i, MV. vu, LV, i, etc. It is an unconvincing
sroup in itself, and no comparisons have therefore been made.

SMITHFIELD. :
SAM T1177’ . 13:0. 11-5 885 1945 4-4
Kann. . 12-4 12:4 1000 1905 60 3:5 Granite.
125 11:33 90:0 1770 2:5 Started. Hole 3-2
cm. deep.
1850 . 13:0 10-4 80:0 1760 56 2:5
13-2 99 75:0 1725 5:4 2-9 Ironstone nodule.
12-0 95 79:1 1870 5:0 3:0. Ironstone nodule.
12-0 8:0 66:7 1150 4:2 2-8 Coarse ss.
Kann. 10:5 8:5 80-4 925 34 3:0 Subcylindrical.
822 11-1 70 63-9 8659) 3°2) 7 2-9 Ss.
796 9-2 9-2 100-0 780 83:3 os
822 QOS) Os 3-7 805 (4:2) (2:6) Shale fragment.
1117 10-0 75 75:0 150) 3°3 aw. Ss.
1830 10:3 70 68:0 745 4-1 a Coarse Ss. Pecked.
796 10:1 7-1 70-0 715 4:3 2:0 Ss. Pecked.
796 10-2 6:5 63:5 675 40 2-1 Ss. Pecked.
822 9-5 10) isa 630 40 2-5 Sandy shale.
ILI 9:6 6:7 70-0 Gb e422 oe Os.
1830 9-0 6-4 71-1 520 40 2:3 Ss. Pecked.
Kann 92 (6:0) 65-2 510 3:9 (2:5) Hollow in centre.
Nodule.
1830 9-0 5:1 56-7 415 (4:0 2-5) Shale. Conical bore.
822 8-9 66 74-2 525 3:0 ie
1830 8-9 6:0 67-4 475 3:5 2:2 Quartzitic Ss.
1117 8:6 6:0 70:0 445 3°5 a Ss.
1830 (8-4) 5:4 64-2 380 45 3:0. Ss.
Wer DS oro 345 3:5 (2:6) Nodule.e Hollow
centre.
71 5:7 80:3 285 30 18 Nodulee Hollow
centre.
6:6 5:2 78:8 226° 3-08) Lea SS:
1880 . 60 5:0 = 83-3 180 3:0 1:0 #Nodule
ETA, eo TO 3°38 54:3 S54 2 One ass-
796 6-4 4:4 68-7 180 2:9 1-0 Ss. pebble. Pecked.
Kann. . 6:5 4:2 64-6 Ti 2:85), AOS ecked.
i 6:0 4:0 66-7 144 2-8 At Ss.
5:8 4:3 74-1 145 2-2 1:5 Fine grained nodule.
1830 . 4:8 29 60-4 67 2:9 1:3. Ss. Pecked. Fine
example.
iss, 46 3:0 65:2 63 2:0 0:8

SMITHFIELD Poort.
SAM 797 29 4:6 77-9 160 2:0 1:0 MHollowcentre. Pos-
sibly metal bored.
Smithfield C.

PHILIPPOLIS.
ASJ 9/38/6 . 12:2 72 59:0 1070 4:7

by
(or)

56 Annals of the South African Museum.

BETHULIE.

MMK 2237 Ae Ae Sion DOONT cts iaed

LO 10:25 78527 Aa 83:5) 2-4:

TMP 1859 7:6 5:6 73-7 Byaay 1) Bis) 4 2

MMK 2237 5:8 3:0 51:7 101 2:8 1-4
Custozza, ZASTRON.

WUAA/I.. 4-9 26 53-1 G20 92-2) ales ‘07 kg.
HERSCHEL.

SAM 496 ; 7-6 2:0 26:3 115” 631

WELTEVREDE, FAURESMITH.

WUA A/41 . 12-7 6-6 52:0 1065
A/40 . 10-2 6-2 60-7 645

iro)

4-9 Mudstone. Cyl. bore.
2-8 -82 ke.

1

ASSOCIATIONS IN THE FREE STATE.

The question of the association of the bored stone in the Free State
has been cleared up once and for all by Professor C. van Riet Lowe
in his field research in that area. Much of the material was destroyed,

or labels were burnt, in a fire at the University of the Witwatersrand

some years ago, but his notes still remain, and the following is a list
of the sites on which he could definitely associate bored stones with
Smithfield B implements. Even in cases where the material has
survived the implements are often too broken to be reasonably
measured. The fact of association is sufficient. Farm names and
numbers are given.

Avalon 554. Mokari 50.

Bankfontein 417.
Blaauwbanksdrift 195.
Brakfontein 231.

De Kiel Oost 101.

Die Aar 232.

HKagle’s Nest 550.

I] Paradiso.

Koffiefontein Town Lands.

Klein Philippolis.
Koningfontein.
Kopjeskraal.
Lockshoek 192.

Paardenvallei 668.
Pramberg 15.
Rooibrak 106.
Smithfield.

Spitzkop 155.
Slagtkraal Drift.
Telegraaffontein 644.
Tygerfontein.
Uitdraai 357.
Wagenmakersdrift 24.
Weenkop 434.
Wepener Town Lands.

The Pramberg specimen is pyriform, the Tygerfontein example is an

unfinished conical-bored example.

The list is taken from van Riet

Lowe’s field-notes and is probably not complete, as many sites listed
there merely state, “full range of implements.” These have not been
included, nor have the sites where borers only are mentioned.

y
—_

The Bored Stones of South Africa. 57

LITERATURE ON THE FREE STATE REGION.

In addition to the papers written by Professor C. van Riet Lowe,
and noted above, there is a certain amount of literature on this region.
Rey. 8. 8. Dornan writing in 1909 * notes that “‘ The first immigrants
came principally from the Cape Colony, from where they were hunted
out by Dutch commandoes. These bands lived in what is now known
as the conquered territory, and their centre was a cave in the neigh-
bourhood of Hermon, to the west of this country near Wepener. The
Bushman name for this part was Qibing—a locative case from Qibi, a
digging stick.”” We may note that the letter “q”’ is here used as a
click, and that, if the derivation is correct, the locative follows the
Sotho, not the Bushman form. He continues (p. 443): “The roots
were dug up with a digging-stick called the Qibi, a straight, tough
piece of wood about three feet long, pointed at the end. It was
inserted into a circular stone and wedged tight. The labour taken to
round and bore these stones with no tools but their hands (s2c) must
have been immense.”

Perhaps the most interesting account, historically if for no other
reason, comes from the pen of Dr. D. R. Kannemeyer, to whom the
South African Museum is indebted for the large collection of stones
listed here from Smithfield, Albert District, and some from Burghers-
dorp. His paper is almost impossible to obtain to-day, and extracts
may therefore be quoted from it. It is to be remembered that
Kannemeyer uses his own orthography in writing Bushman, using an
apostrophe followed by a consonant to designate the appropriate
clicks. In addition this was a popular article, and he makes various
statements without much support.

“Lying scattered over the country, but most frequently found in
the beds of the springs, or in their neighbourhood, these round per-
forated ‘ka ‘ka ‘kouwie, or digging-stick stones, still occur in fairly
abundant numbers. They are often unearthed by the plough. They
are becoming scarcer than they used to be... . . Yet I have managed
to collect about two hundred during the last few years, the largest
weighing close upon seven pounds, the smallest one ounce.

“These perforated stones were used by the women to weight the

* Dornan, ‘‘Notes on the Bushmen of Basutoland,” Trans. 8. Afr. Phil. Soc.,
vol. xviii, pp. 437-450.

7 D. R. Kannemeyer, ‘Stone Implements,’ Cape Illustrated Magazine, vol. i,
1890, pp. 120-130.

58 Annals of the South African Museum.

‘cbr or sharp-pointed fire-hardened stakes employed in digging
taauwing or bulbs, and ‘taau ‘tke ‘tkaubitse, or termite larvae. They
were secured by one or two wedges, or rested on a bulge left in the
middle or upper part of the stick. The pitfalls used for entrapping
game were also dug by means of these weighted digging-sticks.

“There can be no doubt that these digging-sticks were intended for
and used as pick-weights, not as war weapons, nor are they relics
(war-weapons), as Col. Bowker once thought, of an earlier pre-historic
race preceding the Bushman, who, finding them ready to hand,
adapted them to the uses they put them to.

a . what a tale of maternal love and care, what a fond desire to
gratify and amuse their tender offspring do these toy ‘ka ‘ka ‘kouwtes
reveal! In weight an ounce, and only fit to be used by very young
children, they were not very efficient instruments.

“Some of these ‘stick-stones’ were made of traprock. These were
invariably more carefully finished, rounded and smoothened than
those made from sandstone, and must have required a great deal of
expenditure of time and trouble to complete, especially when we
consider the imperfect appliances at command. The older series
shows the most careful workmanship. They are scarcer than the
less highly finished ones, to which I shall presently allude, for they
are found more deeply buried. They were perforated by means of
wedge-shaped stone borers or rimers. . . . Many of these have their
boring end worn down to great smoothness, and show a complete
abrasion of the salient ridges, which gave the necessary grip... .
Several are only partially used, or not at all.

“When contact with other nations gave the Bushmen iron, the
stone drill was discarded. As a rule these later metal-drilled stones
were not so carefully finished. They have a roughish and uniform
bore, and have their shortest diameter in the line of perforation. .. .
In the older stone-drilled ones the aperture is highly polished and
double bell-mouthed, and is made along the longest diameter. This
latter peculiarity is so constant that it may almost be considered as a
diagnostic mark of age. In all cases the stone was first rounded and
then perforated, as these numerous examples, ranging through all
stages of the process, show. . .

“That the cruder and less carefully finished stones are the most
recent in time is proved by various considerations—for instance their
fresh look, their superficiality, and the partially perforated state in
which so many of them are found.

“Occasionally they were used for other than digging purposes. I
have found one on which one side shows evident signs of having been
used as a muller, on the other it has one of the grooves for polishing
the stone (? bone) arrowheads.”

With reference to Kannemeyer’s suggestion that the small bored
stones were made for children, mention may be made of the evidence
of Mr. Turner of British Bechuanaland, who told Miss M. Wilman

The Bored Stones of South Africa. 59

of the McGregor Museum, Kimberley, that he had once seen a family
of Bushmen, all of whom carried digging-sticks. The children had
small specimens. Whether these were made for the use of children
or were picked up by them and used in imitation of their elders it is
not known. This statement, made by Mr. Robert Turner, of
Lukasdam, Postmasburg, is the more surprising when we realise that
the only specimen that comes from that farm is an exceptionally large
soapstone example (MMK 191), and none of the Postmasburg material
is less than 10-8 cm. in diameter. While it is perfectly possible that
small examples were used by children, it is as likely that these were
originally knobkerrie heads. The point has little importance in the
present paper, as in any case we are dealing with examples by size
and form, and have to discard any attempts to say how individual
specimens were used.

An illuminating extract from a letter by Dr. Kannemeyer to Dr.
Péringuey of the South African Museum, dated 3rd May 1909, gives
us some information on the source of the above article, and may well
be published here:

“My informant who gave me the names of the different implements,
etc., was an intelligent Cape-boy, whose father was a child of Landrost
Vollenheim of Swellendam during the Dutch régime. He was brought
up on a mission station in the interior where mostly Bushmen were
located, and he spoke their language fluently and had seen their
handicrafts. I had only one interview with him: he died before I
could pump him further. He called them ‘Ka ‘Ka 'Kouwvie (the
digging-stick) and the—(here follows a sketch of a typical Smithfield
B endscraper)—skinning flake, ’kuon.”’

Further reading on the Bushmen in the Free State can be found
in Ellenberger and MacGregor,* though we need hardly subscribe
to their theories on the origins of the Bushmen.

SUGGESTIONS FOR FURTHER READING.

We have now reached the end of our first major migrational move-
ment, and have discussed the end-products in a heavily ‘‘documented”’
focal area. The Thirstland Region, south of the Orange River, seems
to have held back migration for a considerable time. From the
Orange River southward we follow a second series of migrations,
distinct from those that fed the Free State, but fed from that region.

* Ellenberger and MacGregor, History of the Basuto. London, 1912.

60 Annals of the South African Museum.

The reader is asked to turn on to the South-Eastern Region if his
interest lies in that direction (p. 72). If the Namaqualand route is
to be followed, he should turn to the Thirstland Region, and study
the Kaaienveld and Namaqualand areas (p. 64), ignoring the Trekveld
and Winterveld. Finally, if the southern end of the continent is his
field, he should turn to the Trekveld, Winterveld, and Roggeveld,
or follow the Karoo route to the South Coast, and finally to the Cape
Region (pp. 81, 97). We have reached the parting of the ways.

We now turn back to Region I, to study the Western Bypass, by
which types have passed down the eastern side of the Kalahari, deep
into the Thirstland Region, to the Bokkeveld.

V. WESTERN BYPASS.

We return now to Region I, the Transvaal, and from its western
end we trace the pathway of a major migration, five hundred miles in
length. The route started in the Vryburg district and touched the
sources of the Molopo River; from here the path can be traced south-
ward, parallel to the Vaal but 80 miles farther west. The Groenwater
River is followed to the Orange, where there is some lateral movement
between Kakamas and Prieska, then the highway seems to follow
the Hartbees stream, deep into the flank of the Trekveld and Winter-
veld areas, to Groetvloer, van Wyksvlei, and on to the Bokkeveld area.

The Molopo-Groenwater section coincides generally with the route
taken by the MaNtatisi movement a century ago, and it is possible
that some individual examples may belong to that Bantu movement,
but I would suggest that, whoever may have followed this route in
recent years, it represents the eastern border of a great triangle of
roads, related to the Kalahari border. We are dealing now with the
north-south route and with movements east and west along the
Orange River bed above the Aughrabies Falls, but the Molopo river-
bed itself must have provided another route in good years, probably
feeding the Namaqualand area. Further knowledge of the unknown
desert should give us this information.

The crossroads, where this great highway leaves the Groenwater
valley to follow the Brak and Hartbees Rivers, south of the Orange,
would have been difficult to recognise, had it not been for energetic
help provided by Mr. C. Wilmot, postmaster of Prieska, who has
collected a number of important specimens, through the co-operation
of farmers in that general area, and has sent them to the South African
Museum. All the examples from about here, with numbers SAM

The Bored Stones of South Africa. 61

5014 and over, have been submitted through him, each with the most
exact possible provenance. I would therefore like to thank him for
his keen co-operation and help. Specimens with numbers higher
than SAM 5042 arrived too late for inclusion in type averages.

The Bypass has been divided into three areas: the Molopo-
Groenwater, the central Upington-Hartbees area, and the Southern
or Bokkeveld area. The first shows affinities with the Transvaal
generally, and with the Vaal-Harts area. There seems to have been
some lateral exchange with the Middle Vaal as well, but farther south
contacts do not seem to be so evident.

Typical of the two northern areas on the Bypass, is a gap between
types with a thickness of 10-5 cm. and 7-5cm. The thicker types are
lost in the southernmost area, but the thickness of the remainder
increases slightly to a limit of 8-0 cm., here the upward limit.

We may epitomise the changes that occur by giving the measure-
ments of each of the averages in columns, with their parallels opposite
them. Types which do not recur are left out.

MG. UH. BY.
te Llssey/ 11-2 Ie 225 11:3
TMi 114383 7:2 li. 12:0 7:3 rit, 4 1peel 7:7
ii. 10:8 6:5 iv. 10-7 6:8 li. 10:3 7:2
Vile e025 4:2 Vv 7-1 5:5

I think the picture given provides an unassailable argument in
favour of the approach to the whole subject, here used for the first
time (see map 8).

MG. Motopo-GROENWATER AREA.

ie | 1s 27/ 11-2 81-8 2100 Uli Vine Ve
n. 12-3 7-2 58°8 1090 UPS: BY, 11:
gute, -LO:8 6-5 60-0 760 UH. iv, BV. iii, NES. ii, VMB. iv.
aye (L1-O 4-4 40-0 530) ET. iv, CT. vii.
Wo) toi 6-0 70-6 435 MV. xi, LV. vi, SCV. iii.
vi. 6:5 4-2 64-6 177 Wilals amg Sule sinks

Specimens marked MMK 2078-9 from the Kuruman area yield a
single group with the Wolhaarkop example (MG. v). These might
just as well have been included with the Middle or Lower Vaal areas,
and show a lateral movement from that river. This movement will
be discussed later in dealing with Migrations.

RooIBERG, VRYBURG.

SAM 1104 . 12-6 68 53:0 1080 4:0 2:5 Ss. Fine radial cuts
at both mouths.
VRYBURG A.
MMK 488 . 143 10-8 75:5 2210
679 . 12:7 10-8 850 1740

out
km OO
or 00

hoe

62 Annals of the South African Museum.

KURUMAN.
MMK 920 12-2 75 60:2 1200 5:0 3:6
ASJ 23/39 11:3 At Bylo Bayi Backes.) 94583
MMK 2078 9-0 6:1 67-8 500 .3:8 1-9
2079 8°5 6:7 78-8 485 34 1-7
2078 8-8 5:8 66-0 450 4-1 1:7
717 6:0 AloRy 1 Pflloy/ 15a 3: 4e 4!
POSTMASBURG.
MIMKe 252)" 27) 14°37. 12-09) 83-9" 942455) 5-3 See
2709 . 10:8 6-2 57-4 125) aS) ane O
W OLHAARKOP.
MMK 1285 . 7:9 5-5 70:0 3400) 14: Ores
MAtTZApP.

ASJ 36/35/19 10-7 45 41:0 S60) io:O uae

KLIPFONTEIN, NIEKERK’S HOPE.

SAM 5015 . 12:3 11:1 £490:0 1680 58 3:5

GoUupP.

MMK 1487 . 6-9 4:1 60:0 195 2:9 1-7

BULFONTEIN, KOEGAS.

SAM 5032 . 11-9 70 59:0 1000 41 43:6 #£Hard tale. Pecked.

NUWEVLEI, Koxcas.

SAM 5047 . 9-0 5:8 64:4 470 44 1:6 Granitic rock.

ERFrust, Groot NAUTE.

SAM 5034 . 10-5 6:8 64:8 750 3:3 3:2 Hard tale. Smooth.
Beautiful —_ speci-
men.

The Nuwevlei example arrived too late to be included in the type
averages, but it agrees well with type v.

UH. Urincton-HARTBEES AREA.

This distribution cuts down to the centre of the dry country between
the Orange River and the south-western Cape, with its winter rains.
From here on it is probably lost in the surrounding distributions.
Type iii appears again in the Roggeveld, and type iv at Calvinia.

te l2eo 11:3 90-3 1770 MG. i.
ii. (16-2 5-2 32-1 1365) Possibly Bantu.
lii. 12-0 1-3 61-0 1050 MG. ii, BV. ii.
iv. 10-7 6:8 63°5 780 MG. iii, BV. iii.
Vial 5:5 77-4 275 MG. vi, TV. iii.
VROUPAN, WILDEBEESPAN.
SAM 5017 . 7:8 5:0 64:0 305 31° 2:4

KLIPDRIFT.
PEM 614p . 12:0 6-8 56-7 980 42 2-6

The Bored Stones of South Africa. 63

ERENSPAN, REDLANDS.
SAM 5036 . 12:6 75 60°00 1190 54 4:0 2aS8:

WITVLEI, FRANSENHOF, PRIESKA.
SAM 5043. . 140 12:0 85:7 2350 5:2 3:8 Greyss. Erratic.

MIDDELWATER, PRIESKA.
SAM 5030 . 11:0 6-8 61:8 830 3:5 1:2 Pecked.

UITDRAAI, PRIESKA.

SAM 50385. «13-2 (11:0) 83:3 1915 4-5 x 3-0) Ovalhole. Signs of
sharpening at one
mouth. Soft taley

sandstone.

UPINGTON.

SAM 412 . 11:8 deo (63:5) 1045)" 4-1) 2-8

MMK 697. 10:3 6-8 66-0 720 48 2-4

ErRFsE PLaas, HoPETOWN.

SAM 5042 . 10:1 65 64:3 665 4-4 as Dolerite. Battered.

KENHARDT.

Univ. Stell. . 12-0 73 60:8 1050 40 2:3

MMK 265 . 16:2 52 32:1 13865 3:9 1-9 ? MaNtatisi.

GROOTVLOER.

SAM ’94 elt Go ica 89-4 PIG D: | 329) 9 22-5

KEURFONTEIN, VOSBURG.

UCT 36/— . 12:8 116 90-6 1900 4:2 2:6 Smithfield B. Assoc.

UCT 36/— . 6:3 5:4 885-7 2A Oe eal] Do. do.

GROOTFOURIESKOLK, SPRINGBOKPOORTJE.

Sabogal: 2 11-8 (11-0) 9372 1530 49 .. Soapstone; very
irregular. Worn,
unfinished.

PooRTJE, OMDRAAIVLEI.

SAM 5044 . 13:5 68 504 1240 53 3:7 Ss.

VANWYKSVLEI.

SANEI25 y=) 7-3 6:0 82-2 320 2:8 1:9

The Witvlei and Poortje specimens arrived too late to be included
under the type averages. Both fail to agree well with local types.
The first is an erratic that merely brings the MG. 1 type a few miles
farther south, to the southern bank of the Orange River. The Poortje
example is somewhat anomalous, though it may be a variant of
group UH. iu.

It may be noted here that the Trekveld Area (p. 69) seems to have
some affinities with these first two stages in the Western Bypass, but
not with the further extension to the Bokkeveld area. These will be
discussed later under Migrations. Probably we are dealing with a
divergence of types into two routes—the Trekveld and the Bokkeveld.
If this is so it may in time permit us to differentiate between types

64 Annals of the South African Museum.

that passed to the Bokkeveld, and types that entered the Trekveld,
thus providing the foundation for some more exact definition of the
two groups of people who took the diverging roads.

BV. BoKKEVELD AREA.

This represents the final coherent distribution of this series. Larger
types, with a thickness greater than 8-0 cm., have disappeared, but
except for the anomalous example from Calvinia, the various types are
all represented along the Western Bypass.

i. (14-0 8-9 63-6 1745) Anomalous.
i. 12-4 Corl 63:8 1150 MG. ii, UH. iii.
ii. (10-3 7-2 69-8 740) MG. iii, UH. iv.

HoNDEKLIP Bay.
SAM 4930 . 12:6 78 61:9 1240 4-7 2:6

KLAVER.
Univ. Stell. . 12:0 79 659 1140 29 2-5

CALVINIA.
SAM 410 . 14:0 89 636 1745 40 2-7 Anomalous.
AQ30, ° 2 32-5 4 15922) LEGON 4-0; eee 2
410 . 10-4 TO) “57-3 150 3:95 92:3
Univ. Stell. . 10-1 io daa 145° “39 ee

VI. THIRSTLAND REGION.

This wide region could never have been thickly or regularly
inhabited. If recent climatic conditions are reflected in the present
day, then this is an area of long droughts, sometimes lasting seven
years, then a short spell of wet years, when there is high fertility,
plentiful water and reasonable cloud cover. This has always been an
area of game, plains-loving springbok being the most typical animal.
Possibly giraffe, buffalo, zebra, and elephant lived here, though
sporadic bones of these animals suggest that they were migratory.
Game migration has been stopped by essential jackal-proof fencing
in this region, and most of the game has died as a result of drought
in this typical dry-scrub country.

The region lies south of the Orange River, from about 25° east to
the west coast. In the south it is bounded by the Roggeveld moun-
tains, the Nieuwveld mountains and the Sneeuwberg, roughly along
the parallel 32° soufh. The whole area is arid, increasingly so from
east to west. It is divided into the Kaaienveld, near Prieska; the
Namaqualand area, farther west, together forming the western
highway; the Trekveld and Winterveld areas which probably provide
a focus for all the remainder. From here we may turn on to the

The Bored Stones of South Africa. 65

Nieuwveld-Goup area that links with the Little Karoo and Sundays
River to the south-east, to form the southward highway, and with
the South Coast Region (p. 81).

KV. KAAIENVELD.

This area partly coincides with the Upington-Hartbees area, and
marks the crossing of two roads.

ros kL 13:3. 101-6 2250)
me LL 8:8 80-0 1085 LV. iti, NQL. i, MV. vi.
it.» Vib 5:4 47-0 715 MV. ix.
iv. 9:2 7:4 80-4 625 LV. v, NQL. iii.
wa (0-9 6-9 100-0 330) MV. xi.

(
vier (8:5 4-0 47-0 290) WY. viii.

Group i might well be divided into two, 12-0 x 8-8 and 10-9 x 8-8 cm.
The former related to Namaqualand 1, the latter to Namaqualand u.
The single example from Prieska, forming type v, agrees well with
a single specimen from Barkly West to Gonggong (6-9x6-4). In
general there is considerable agreement with the Middle Vaal.

** BUSHMANLAND.”’
SAM 1587 . 12:0 13:2 110:°0 1900 3-9 2-0 Anomalous.
te a 12-2 5:5 45-0 820 40 2:3
MARYDALE.

SAM 5022... 10-9 87 79:8 10385 40 2:2 #£z¥Fine ss. Bore and
stone deeply pati-
nated.

WITFONTEIN.
SAM 5033 < 12:0 89 742 1280 44 $19 #£Pecked.
PRIESKA.
MNIKers49 2 14-2 «13-5 )=— «95-1 2720 «49 28 Ss Ss.
12-0 So ed p40) 9 5:4 SI Ss.
10-9 SO esa-a 1080" 63-0" 2-5” Ss:

SAM 4607 . 10:6 8-8 83-0 990. 39. 23

MMK 1849 . 12-4 Dro aed 815 4:6 2:2 #£Ss. Grinder.

HHC 1047 9-5 7-5 "79-0 GiD a2" 2-5

MMK 2108 9-4 76 80:9 670 3:2 2:2 Ss. Smoothed.

SAM 4868 . 9-2 74 80-4 625, 4-0 2-7

TMP 8934 . 9-2 7-0: ~ 76-1 590 3:8 21

SAM 1099 6-9 6-9 100-0 BaUMeow aD

MMK 2108 8:3 4:2 50°6 290 35 1:8 Erratic.

BAaNGHOEK, HoOPETOWN.
SAM 5037. 118 50 42-4 695 45 2-3. Bad shape. Sandy
shale.
STOFBAKKIES, PRIESKA.
SAM 5014 . 11-4 Son d4-by 105) 4-6) 3-02 | Ss:
VIGILANTSPUT, PRIESKA.

SAM 5020 . 11-5 5:4 47:0 RS era4e 2-2) Ss:

10-4 (5-1) 49:0 58be ~5:1) 2-6 ~ ‘Hard tale. Worn.

VOL. XXXVII, PART l. 5

66 Annals of the South African Museum.
KEIKAMSPOORT, PRIESKA.

SAM 5045 . 11:9 84 706 1190 46 2-9 Shale.
‘*KAATENBULT.”’

ASJ 14/41 . 9-0 75 83°3 605 4:8 3-5
MaRITZDAM.

SAM 5040. 11-9 91 765 1290 41 23 Ss. Pecked.
ONGERS SIDING.

SAM 5046 . 10-5 8:5 81:0 9403-6) 2-7 se

JOHNSTON'S PAN, GROOTDOORN.
SAM 5019 . 10-9 90 82:6 1070 46 2-6 Hard talcy shale.

Victoria West A.
Sharples 10 . 8-9 7-1 80-0 560 3-0

4. 86 38 442 280 28 2.0
The specimens from Keikamspoort and Ongers Siding arrived too

late for inclusion in the type averages. Both conform to the average
KV. u.

NQL. NaMAQuaLAND AREA.

This is the end of a westward movement along the Orange River
valley from the Kaaienveld, Trekveld, and Winterveld areas. Though
there is no evidence from the north bank of the Orange, I suspect
that the Molopo River also brought types directly across the southern
Kalahari to this area; probably types iv and v from MV.

ry 1122 9-4 ie 1375 KV. i, MV. va; DiVoae
li. 10-6 8-5 81-9 955 KV. ii, LV. iti, MV. vii.
iii. 9-0 7-4 82-2 600 KV. iv, LV. v.
hye |, ate 5-1 57-9 385 LV. vi, MV. x.

v. (4:2 3:3 78:5 58) WY. x, MV. xiv.

Types 1 and 1 above represent the results of divergent copying from
KV.u. These two, with type iii, can be traced from the Lower Vaal,
through the Kaaienveld. The types present in the Trekveld and
Winterveld areas (over 12:5 cm. in diameter and 10-0 cm. in height)
are absent here, and indeed no examples over that height occur in
this area. They probably represent an independent spread which
never reached Namaqualand.

‘*NAMAQUALAND.”’
ASJ 49/36/6 . 12-3 SO) l-o 134) 4 Dea
STEINKOPF.

NMP 2080 12:5 99 792 1545 54 2-2
2592 11-0 88 800 1065 56 2:5
2342/p. 10-7 8:4 78:5 960 59 2-8
2577/a . 10-1 8:4 83-1 855 48 2-1

The Bored Stones of South Africa. 67

NMP 2342/F. 8-8 7:5 85-2 590 5:3 2:8
2342/B . 8-8 7:3 81:8 565 = 4-1 2-0
2342/6 < . 8:9 bel bT-3 405 80 3:0
2577 8-3 4-9 59-0 340 3°8 Pepe
2048 4:2 So OLD yoy Poa an bol Erratic.
CoNCORDIA.

SAM 1394 . 12-0 96 800 1380 41 2-0
9-5 73 76-8 660 40 2-2

OOKIEP.

MMP 2517 . 93 5:3 57-0 630 48 2-9
KAMIESKROON (BOWESDORP).

SAM 1763 . 11-6 95 81:9 1280 54 3-0
GaRIES.

Dreyer Coll. . 11-0 90 81:9 1090 2-4 2-3 Soapstone.

We now return once again to Prieska and the junction of the Vaal
with the Orange, and trace a separate movement southward from
there. This is related to the beds of the Brak and Ongers Rivers.

WV. WINTERVELD.

The Trekveld and Winterveld areas were originally taken together,
but apart from a few overlapping examples, the two show very distinct
distributions of types. The Winterveld appears to provide a great
highway, from Prieska and the junction of the Vaal and Orange Rivers
to the headwaters of the Sundays River in the Koudeveld, making
use of the Brak and Ongers Rivers.

i. 14-6 13-0 89-9 2770 MV. i.

ii. 14-6 10-1 69-9 2150 NG. i.

tit) 12-9 9-0 69-8 1495 NQL. i, KA. i.

ive 10:9 10-1 92-7 1200 MF. ii.

v. 10-0 7-5 75-0 750 KV. iv, KB. iv, CFS. ii, SWB. iv.
vay celal 8-5 104-9 560 KA. iii, BS. iv, KB. v.

viable ath 5:8 63-7 480 NQL. iv, KA. ii, AG. i, BS. v.
Vili. (8-5 4-3 50-6 310) CES. iv.

pe (977, 4:3 75:5 140) MICAS Vy, (BSivi-

Ke oe 2-5 67-6 35 Probably all beads KB. x.

The comparisons with Koffiefontein A and B suggest that this area
is the continuation of a migration from the south-western Free State.
The parallels, in the smaller types, with the Bloemfontein-Sepani area _
seem to carry the migration nearer to its origin.

BELMONT.

MMI 178. 14-2 12-2 85:9. 2460" 4-8 32-9
LE58: 3°8 3:3 86-9 48 1:3 0:5

HoPEeTowN.

SAM 757 . 4-0 2-1 + 52-5 34 2-1]

68 Annals of the South African Museum.

BaNGHOEK, HoPETOWN.

SAM 5039 . = 9-2 6-0
PAARDEKLOOF, WELGEVONDEN.
SAM 5027 . 12:8 9-1
VREDE, UPINGTON.
SAM 5041 . 12:1 (9-0)
GOODHOPE, PRIESKA.
SAM 5029 . 14:8 9-9
ROoIvLOER, OMDRAAISVLEI.
SAM 5088 . 14:5 (12:5)
PLAATSMABOK.
SAM 5021 . 144 # 10-0
FarIrvIEW, PooRTJE.
SAM 5023 . 12:9 (9-2)
KALKAAR, CARNARVON.
SAM 5048 . 13:0 8:6
CARNARVON.
SAM 1098 9-6 8:0
PEM 615 V.. 8-1 8:5
BRITSTOWN.
Heese Coll.1. 15:0 13°4
2. 14-7 10:4
Ba le) 9-1
SAM 2469 . 108 10-2
Heese Coll.4. 10:7 = (7:3)
SAM 1753 . 10:2 7:3
Heese Coll. 5 9-7 8:0
17 9:8 7:4.
Pall ets) 5:7
Ch G7) 4:3
Il 48 (2:7

MIDDELWATER, BRITSTOWN.

Heese Coll. A 3-1

B 2-4.
PHILIPSTOWN.
MMK 1111 . 14:2
HANOVER.
Sis: - 10-1
Ge a, 85d
OH as cil
GRAAFF REINET.
SAM 3206. 13-0
F. Malan Coll. 8-7
J ANSENVILLE.
PEM 6l4w . 12°8
11:0

2°8

2-3

510

1490

1545

2170

2630

2075

1530

1455

740
560

3015
2245
1535
1190
835
760
750
710
515
140
62

27
13

2500

735
310
34

1510
445

1460
1370

a
Ne)

He i
oo

OTD TO) nes Tee les) ON he lye
HE RAADANWOSSNWH

oe
@ ©

2-7

4-3

3:6

2-4

2-4

em bo
~I bo

ae ee RO Pe a COOL
Drmowok KR? TOR

0-7
0-7

2-9

Just started. Ham-
mer.

“Cantaloupe quar-
tering’ at one
end. Ss.

Oval cylindrical
bore. Split. Ss.

Soapstone. Oval

hole.

Quartzitic Ss.
Started.

Fine granitic rock.

Hard tale. Oval
bore.

Sandy shale.

Anomalous.

Anomalous.

Bead. Metal bored.
5 Bead. Metal bored.

The Bored Stones of South Africa. 69

TV. TREKVELD.

The Trekveld and Winterveld were originally taken together as one
great area. While the Winterveld shows a distinct highway, the
Trekveld does not; it appears, so far, to be a backwater, with little
real relationship to surrounding areas.

i.. 13-8 10-4 76-1 1980 MG. i.

i. (11-4 9-5 83:3 1235) KV. ii, NQL. i.
iii. 7:3 4-2 57°5 225 MV. xu.

THREE SISTERS.
MMK 1302 . 13-4 99 73:9 1780 42 2-2

Victoria WssT B.

ELM 1 4-02 ie? 9 80:3 2305 7 4:0) 73-3
SAM 1494 . 13:8 100 725 1905 55 3:3
11-4 95 83:3 1235 48 2-6
1662. —«. 7:5 4:0 53:3 225 29 2-0
Sharples 7 . 71 4:3 60:6 220-5 2-5 :
ELIASLAAGTE.
MMK 39 : 7-4. 4:4 59-5 240 30 41:7
BRANDVLEI.
SAM 4998. 7:5 4:1 54:6 2308) 3:8) 1 2:3
KALABASPUT, VANWYKSVLEI.

SAM 5049 . 13:55 10:0 74:0 1825 54 3:1 Sandy shale.

BLAAUWKRANS, CARNARVON.
SAM 5050 . 134 100 736 1795 61 2:4 Sandy shale.

We have already noted the analogies between this and the Upington
Hartbees area, and have suggested the possibility that the Trekveld
may be in part related to the Western Bypass.

BIBLIOGRAPHY AND COMMENTARY.

Not very much has been written on the bored stones of this region,
but what has been written has some interest. E. J. Dunn * describes
bored stones he collected from various sites in and outside this region,
and adds: “The knowledge of the manner in which they were made,
however, remains, for the same old Bushwoman that showed me the
manner of affixing stone flakes to arrow-shafts described the method,
and particularly mentioned the long-pointed stones of hard material,
which, she said, were obtained from the Kiljan Veldt.” This last is
certainly a misinterpretation of Kaaien Bult (or Kaaien Veld). In
much the same way he gives Zak River as ‘‘Leek River” in giving an

* “Stone Implements of South Africa,” in Trans. §.A. Phil. Soc., vol. ii, 1879-80,
pp. 6-22.

70 Annals of the South African Museum.

account of this old woman’s methods of attaching stone tips to arrows.
This error he partly rectifies in his later work * written thirty years
after he had left “the Colony,’’ and here makes Kiljan Veldt into
Kijan Veldt. In this latter book he gives a somewhat fuller account
of the making of the bored stone, on p. 77 onwards. His illustrations
are excellent, but unhappily, as only one view of each stone is given,
little can be deduced of a type helpful to us here. In one or two
instances it has been found possible to make use of his illustrations,
but such evidence is suspect. The collection itself is now in Australia,
providing that continent with a relatively useless hoard, and at the
same time depriving this continent of valuable material.

W. D. Gooch + comments rather bitterly on Dunn’s work and says
that he himself, Bleek, and Bowker all failed to find a Bushman capable
of making a bored stone, while Dunn had succeeded.

Dr. Bleek seems to have had better luck later, as a letter from his
daughter, Miss D. F. Bleek, shows:

“With regard to the question of boring I may mention what an old
Bushman woman told me at Kenhardt in 1911. She was the same
woman who had been photographed with a digging-stick at Salt River
in 1884; the picture forms the decoration on the cover of ‘Specimens
of Bushman Folklore.’ She told me that women bored them, not men;
that she had bored one herself; that it had not taken her very long;
that it had been done by iron. She was speaking Afrikaans and said
‘met e eister’, and made the motion of twirling with her hands. Of
course the earlier ones must have been made with stone borers, but
in the years 1840 to 1850 when she was young, iron must have been
obtainable.

“The !Xam Bushman have a verb /kiyu, to knock down by a throw.
Their word for a kiriis /kéken. Their kiris were made by nature, just
a stick that happened to have a knob on it. The Nama for a kiri
is /hoas.”

NG. NIEUVELD-GOUP.

Contrary to my usual custom, I have found it advisable to split
a single museum-accession group: the Aberdeen series, PEM 514 F,
is divided into Aberdeen A, considered here, and Aberdeen B, to be
considered under Sundays River area. It seems clear that the ten
specimens sent to the Port Elizabeth Museum were not from a single
site. Some link with the Thirstland Region, others are related to the
Gamtoos, Klein, and Sundays River basin. It is interesting to note

* EK. J. Dunn, The Bushman. London, 1931.
+ “Stone Age of South Africa,” J. Roy. Anthrop. Inst., vol. xi, 1881, pp. 124-82.

The Bored Stones of South Africa. 71

that this series shows no relationship with sites at Jansenville and
Graaff Reinet.

Averages from here are not convincing, but littie is really known of
this area. It is related to the Zwartberg and to the Gouritz basin
to the south, and Winterveld and Kaaienveld to the north.

i. (146 10-1 69:5 2160) WV. ii.

ii. (121 11:0 90-5 1600)

iii, 11-4 81 705 ° 1050 # XK. vii, ZW. iii.
v. (10-5 61 58-1 675) GK. iv.

SUTHERLAND.
Drennan Coll. 11-6 8-1 69:8 940 4:8 2:3
10:6 8:1 76-4 O10 4-27) 2-1
BEAUFORT WEST.
SAM 803 4 AO.) 0-8 215, 4-2 us
WViWANIO4O 2 «12-2 )«=«6 11-0 0S 9071 )3=ss«d63 <e HSCOD
MURRAYSBURG.
SAM 767 . 12:0 10:9 90:8 1570 4-1] 2-5
ABERDEEN A.
PRE GI4n Ss 14-8 10:1 68:2- 2210. 3-7 41:8
11-9 82 689 1160 49 2-2
11-2 80 714 1005 46 2-7
10-5 6-1 58-1 675 4-1 rae Anomalous.

The last specimen listed may belong to Aberdeen B, Sundays
River area.

SUGGESTIONS FOR FURTHER READING.

Those readers who are primarily interested in the Cape Region
should now turn to the Sundays River, Gouritz Karoo (pp. 82, 92),
then continue with the South Coast Region (p. 81), finishing with
the Sandveld and Swartland areas (p. 106).

For those who are interested in what lies to the east, we now return
once again to the Orange River, this time to the eastern end. From

here we take up the south-western route where we last left it, at
Smithfield (p. 60).

EASTERN PROVINCE OF THE CAPE.

We now turn from the Free State, the area of highest concentration
of Smithfield types, towards the south-east, to the area of the greatest
efflorescence of Bushman paintings.

Many of the sites mentioned below, both in the South-Hastern
Region and in the Lowlands, have been described as paintings sites by
Burkitt, Hewitt, myself, and others. The fourfold association of
paintings, bored stones, implement types, and skeletal material is
never very clearly expressed, but the main cultures in this general

72 Annals of the South African Museum.

area during the Later Stone Age were Wilton and Smithfield C, both
fundamentally cave cultures, and a certain quantity of Smithfield B
material on surface sites. The two former are associable with paint-
ings, though no attempt has yet been made to analyse stylistic groups
in relation to deposits. The last culture mentioned is, so far as I
know, not associable with any art forms in the Eastern Province of
the Cape, but it can be associated with certainty with the bored stone
in several instances.

The Clarke’s Siding site, on Victory Farm, is referred to by Burkitt,*
and by Miss Tongue f who covers much of this area in her portfolio
of paintings. Additional references to the art of this part of the Cape
Province can be obtained from the works of Miss D. F. Bleek, in
collaboration with others.{

I would suggest that the areas below the mountain ranges have been
fed at various periods by highways which have at times crossed one
another, often at right angles. One important route, from the Free
State along (or more probably, over) the southern slopes of the
Basutoland mountains, which fed the Northern Transkei and Hast
London areas, is so far unknown (see pp. 75, 77).

Various types can be traced along the coast as far as the Knysna-
Tzitzikamma and the Belvedere-Heidelberg (Cape) areas.

VII. SOUTH-EASTERN REGION.

This region starts at the confluence of the Seacow and Orange
Rivers, at 25° E. long. It includes the stretch between the Orange
and the Witteberg and Elandsberg, then continues to include the
Transkei, so far asitis known. The region is dealt with in three areas:
Upper Thornveld, Lower Thornveld, and Transkei. The whole
Thornveld is so named as it coincides well with the southern portion
of Regional Division 5: The Cape Thornveld, in the Union Year Book
22, 1941, p. 992. So far as the Upper Thornveld area is concerned
the term has little real environmental value. Possibly a better term
for the Upper Thornveld would have been New Hantamveld, according
to the publication of Dr. KE. A. Nobbs,§ which came to hand after
these terms had been chosen.

* M. C. Burkitt, South Africa’s Past in Stone and Paint. Cambridge, 1928.

t Helen Tongue, Bushman Paintings. Oxford, 1909.

t Stow and Bleek, Rock Paintings in South Africa. Methuen, 1930. Van der
Riet and Bleek, More Rock Paintings in South Africa, Methuen, 1940.

§ E. A. Nobbs, Veld. Privately published. Lynedoch, Cape, 1941. Museum
copy received 9th January 1942.

The Bored Stones of South Africa. 73

UTYV. Uprer THORNVELD.

This area covers the open plains between the Orange River and the
first great mountain step towards the sea, formed by the Kikvorsch-
berg, Zuurberg, Bamboesberg, and Stormberg ranges. The country
here is dry, with a summer rainfall of 10 inches to the west, increasing
to 20 inches to the east, but the Orange River is easily accessible in
times of drought, and is fed by winter snow and rain from the Basuto-
land mountains. This appears to have been the route by which the
rivers to the south-east were approached, especially the Kei and Fish
Rivers. It can be regarded as a fairly general “‘feeder”’ area.

mio |6M-0 | °84:60«©— («1860 ~ SEES. ii, ME. i, KTZ. i, GR. STK. i.

ii. (14-0 8-4 60:0 1650) PE. ii, STK. ii.

iii. (10-7 91 85:0 1040) KV. ii, SEFS. v.

iv. (11-0 6-6 60-0 800) PE. iv, BDV. iii.

v. 86 6-4 75-0 470 WV. vii, SEFS. vii, CFS. iii, KA. ii,

KTZ. xiii, BH. vii.

vi. (63 6:3 100-0 255)

sain 7 3 4-7 64-0 250 LTV. v, BDV. vi.
viii, 5-4 38 70-4 113 +Britstown. SEFS. xii, MF. vi, KA. v,

BH. xiv, STK. v.

It is worthy of note that type i is confined to the immediate neigh-
bourhood of the Stormberg, while type vii occurs only at Colesberg,
Aliwal North, and Burghersdorp. Type v is extremely persistent all
through the South-Hastern Region and eventually seems to become a
lowland type, though not associable with true coastal middens.
Starting from the Free State, where it appears to be a fairly general
form, it passes through the present area to Knysna, apparently by
way of Graaff Reinet, where a single analogy is known. After Knysna
it passes along the coast into the Belvedere-Heidelberg area, but gets
no farther than Still Bay.

It is important to note that the Upper Thornveld feeds the Southern
Transkei directly, without any types common to both passing through
the Lower Thornveld.

Only two examples have a thickness between 7:0 cm. and 10-4 cm.
Diameters are fairly evenly distributed from 5-0 cm. to 14:0 cm.

CoLESBERG.
MMK 1304 . 5:0 3:8 76:0 Ooms? yack Ss. Hollow centre.

NAAUWPOORT.
PEM 615Q . IL-1 6-1 55-0 Zine Gr3) 9 1 3%8:

ALBERT DISTRICT.

AMG 132 . 10:
SAM Alb. )

(e2)
Or
oO
—
S
He
j=)
Ow
iS
ke! DO
QO ww

74 Annals of the South African Museum.

BURGHERSDORP.
SAM 84. « 12°8: S1E3 2) 88°3> F850" 2-6 =e
140 84 600 1645 50 2:8
TMP 1849 . 7-0 4:3 61-4 Paki li scl le Sle
SAM Kann.. 60 42 70:0 151 26 1:8
84. Rape 3 3:5 66:0 99° 2:3) 15
Attiwat Norra. .
ELM 50 . 120 11:5 95:8 1655 5:2 3-2 °#£Dolerite.
Drennan Coll. 11-0 70 63-6 845 45 34 LimySs. Broken.
SAM 1316. 9-0 5-6 62:6 455 4:1 2-9
8:0 6:8 85:0 “sey eT “7
7-5 5-4 72-0 305 3:0 2-2
6-3 6-3 100-0 250.0) gS eae Pyriform mace head.
5:0 3:9 78:0 OT 2 bee
Rossouw.

Univ. Stell. . 13-4 I1-1 85:4 1995 44 2:8

QUEENSTOWN-DORDRECHT RD.

ASJ 2/36 a 1387 ol0:45 375:9) 1950) 5-0,
DORDRECHT.
NMP 96 fs 8-5 6-4 75:3 460 46 2-2

CLARKS SIDING, DORDRECHT.

SAM1756 . 7:7 (5-0) 649 295 4:8
70 (5:0) 71:4 245 33
3-8

3-1
1:8 p. 138.)
73 3-8 | 52:0 205 1-6

LTV. Lowrer THORNVELD.

This area covers the belt of high country between the south-eastern
limits of the Upper Thornveld and the Winterberg range. It includes
most of the Upper Fish River and Upper Kei River basins. The climate
is far more agreeable here, the rainfall is between 20 and 30 inches, and
this has once been excellent game country. The rivers flow regularly
except in the driest months of the year, when pools are available.
Types are fairly regular.

1. (LS 9-0 78-2 1190) UTYV. ii, BMR. u.
i. 10-6 8-1 76-4 910 KB. iv.
ii. 10-6 5-7 53°8 640 NV. iv.
UVic Ss 5:6 65-9 405 WV. vii, KA. ii, KTZ. xv, EL. iii.
View aO) 4-7 67-1 220 UTV. vi, TV. iii, KB. vii.

Probably types i and ii are the same. This area appears to show
greater affinity with the western part of the Free State than with the
eastern. This suggests that Norval’s Pont and Bethulie were the
paths by which these types reached their present area from an origin
near Koffiefontein. Only two types show any real affinity with the
Upper Thornveld.

Notice the very marked gap between specimens 6-2 cm. thick and

(See Burkitt, op. cit.,

ae “ae a > i ae

The Bored Stones of South Africa. 75

8-0 cm. thick, and also the lack of any examples over 9-0 cm. in height.
Diameters vary from 6-6 cm. to 11-8 cm.

GLEN GREY.

SAM 3008 . 6:8 5:0 73:5 231 3:2 2-0 Smithfield B imple-
ments.
ADAMSON’S BANK, QUEENSTOWN.
SAM 4564 . 7:3 4:6 63:0 255). ar, | Le?
THORNEYCROFT, QUEENSTOWN.
ELM 78 6 UTED) 5:0 45:4 605 50 2:0 0-92 kg.
QUEENSTOWN.
SAM 4565 . 11:8 8:9 75:4 1235 48 2-7
WUA 1159 . 10:0 6:2 62:0 620 56 2:6 Gritstone.
SAM 4565 8-6 5:4 62:8 400 34 1:8
TARKASTAD.
SAM 667 . 11-2 90 803 1130 3-2 41:5 Ss.
STEYNSBURG. .
ASJ 61/36 11-2 8°29 73:2 71030 4:7 _ 2-6
3 10:7 8:2 76:6 940 3-7 :
ELM 45 . 10-0 8:0 80-0 800 40 1-7 0:91 kg.
ASJ 61/36 . 88 5:9 67:0 455 3:7 2:6
ay Sl 5-6 69-2 365 39 1:8
55 7:9 5:5 69-8 34599) 3-5) | 2-1
F. Malan 6-6 4] 62-1 179 3:8 # 41:9
THEBUS.
ELM 84 - 10-5 6-0 57-1 660 40 3:2 1-02 kg.
HOFMEYR.
Unive stell. - 10-7 56 52:3 640 4:0 0-9 Grinder.
9-1 5:0 55-0 ALS *3-8) 227 9 Ss.
8-3 5-7 «68-7 B90 Boro 2k

MIDDELBURG.
Drennan Coll. 7:3 50 685 2Gornat | ole7 Shale.

NTK. anp STK. TRANSKEI.

Only two sites belong to the Northern Transkei, and neither of the
specimens show any relationship to either the Southern Transkei or
to Natal. These are listed below under their sites, but are not included
in the present type averages. The Southern Transkei is related to the
lower reaches of the Great Kei River, and the averages given are
deduced from this area (STK.) only.

i727) 11-0 863 1775) UTV.

i
Hee LS-5 7:9 59-0 1440 UTV.

lili. 12-4 8:3 68-0 1265 KA.

iv .
Vv

ete FS ee
e e e

7:8 84-0 675 CES.) i:
(5-4 4-0 74:3 125) UTV. viii, KA. v, SEES. viii.

Almost all these Southern Transkei examples lie between 7-5 and
8°5 cm. in thickness. Only three specimens lie outside this range

76 Annals of the South African Museum.

(typesiandv). The relationship between this area, the Central Free
State, and Koffiefontein A is marked, but almost all types show a
relationship with the Upper Thornveld. The complete absence of
any affinity with the Lower Thornveld is important. Obviously the
routes passing through the Lower Thornveld missed the Southern
Transkei completely.

The Northern Transkei provides a different set of problems. Pre-
sumably the routes that fed this little-known area came up the Orange
River to Qacha’s Nek. This question, which will be enlarged upon
under Migrations, can only be cleared up in Basutoland itself, where
archaeological research is at present difficult. Careful collection of
material from both the Southern and the Northern Transkei should
clarify the relationship between these two areas.

Nortu (NTK.).

THABANKULU.
TMP 7375 . 14:0 70 50:0 1370 5:4 2:0

XUKA DRIFT.
SI30 . . 10-7 74 69-1 850 40 2:8

SoutH (STK.).

Fort MALAN.
AMG — yy Bhd) 3:4 59-4 82 20 1:8 ‘*Choke bored.”

CATHCART Downs.

ELM 98 - 127 PO 86°39 1775 o- 25) 837 eeee
100 4 diesels) 80 50:3 1460 42 30 2:26 kg.
97 - els 85 68:0 1330 45 25 18kg. Redss.
96 ce (Sl) SO OOL8 605 45 3-2
Happy VALLEY, CATHCART.
ELM 94 > 14-2 8-0 56:3 1615 40 2:7 2-27 kg.
93 py ag B24 75 79:0 675 3-2 2-7 GO: Oikas
KINGWILLIAMSTOWN.
Drennan Coll. . 12:7 8:4 66:1 1335 3:0 38 Dolerite.
12-9 7-6) 1 58°9) 1265) 3:9" 2 ess:
11-9 79 664 $1120 46 25 #£zPecked.
East Lonpon B.
SAM 520 a) O28 8:0 81-6 770 44 2-4 Midden site. Shale.
Drse NEK.
ELM 44 a? On; 7:7 83:8 650 3:5 2:2 0-91 kg.
BALLINAFAD, COMMITTEER’S DRIFT. -
AMG — - 60 4-7 79-1 7a 28-1

We now turn along the coast to study lowland types, and the
intrusive elements brought in at various points by rivers whose basins

The Bored Stones of South Africa. 17

lie partly behind the parallel rows of mountain ranges, but which
debouch at points along the coast.

VIII. SOUTH-EAST LOWLANDS REGION.

This consists of the lowlands between Port Elizabeth and Hast
London. The region is difficult to understand, but certain fairly
localised areas occur, and should provide a skeleton for future research
by the four museums which happen to be included in this area. One
point that should certainly be watched is the distribution of types by
index, those with an index below 40 per cent. being apparently coastal.
Diameter too seems important. I suggested in the Preliminary Survey
that there were three parts to this region: Hast London to the Great
Fish River, where diameters between 8°5 and 12:8 cm. are met;
Fish River to Sundays River, where diameters from 4:9 to 8-3 cm.
and from 10-8 to 14-9 cm. are encountered; and finally in the Port
Elizabeth area, where diameters lie between 8:9 and 15-2cm. A little
new evidence has changed that picture slightly, and I have been able
to divide the Region into five areas, some of only slight importance
and size, but all highly suggestive.

We are reaching a part of the country where we have a difficult
double flow of migration. There are the feeding lines from the
inland areas down the great rivers to the coast, and across these are
obvious lateral movements along the coast. In this region the rivers
are still more important than the coastal highways, but as we go
south and west this changes, and coastal or lateral movements become
more dominant than the river routes. The reason for this obviously
lies in the parallel rows of mountains, enclosing narrow valleys, typical
of the southernmost littoral, forming a more effective barrier between
the inland plateau and the coast.

EL. East Lonpon AREA.

Types here show bad grouping, but all are limited to 8-3 to 12-7 cm.
in diameter, and to 3-0 to 7:5 cm. in height.

Pol) 74 163-2) 1015) | BMR iv,

ii. 120 5:4 45-0 780 NL. ii.

ii, (88 5-7 64:8 440) LTV. iv, UTV. v.
iv. (103 35 34-0 375)

v. (85 40 47-0 290) ZMM.i.

The comparison with the single specimen from the Natal Lowlands
is given for what it is worth. There is no relationship whatsoever

(ito aay Annals of the South African Museum.

discernible with the neighbouring Transkeian material, which lies
between Natal and Kast London.

KWwELEGHA R.

ELM 83 - 12:0 75 62:5 1080 45 2:2 1:9kg. Quartz.
NaHOoon.
SAM 1495 . (12-5) 5-2 41-6 820 (4:4) (2-2) Midden. Fragment.
ELM 71 Ae) 5:2 61:8 375 40 2:2 —0-57 kg.
Hast Lonpon A.
SAM — . 109 51 46:9 605 50 5-5 Rocklands. Sub-
triangular.
ELM 34 - 10-0 4:0 40-0 400 30 1:2 0-57 kg. Park Gates.

CovE Rock.

ELM 89 Ae Iie 75 63°38 1025 45 3-0 1:36 kg. Dolerite.
75 aah LA) 5:0 39-2 805 30 2-0
— Sai 4:0 45-7 305 2-7 1:2 0-45 kg. Dolerite.

GuLu RIVER.
ELM 107 Br likey 60 51-1 820 45 2.
3:0 . “yen all

5 1:14kg. Grinder.
‘5 0-57 kg.

103 5 les 28:5 3303 Dolerite.
Fort BEAUFORT.
AMG — . 10:9 70 64:2 830 66 2:7
11:8 4:8 40-7 670 4:0 2-0
UNIONDALE, ALBANY.

AMG — a0) So3 4:0 48-2 275 49 2-0 Shelter. Wilton layer
underlying Smith-
field.

PAARDEKRAAL,
AMG 2887 . 9:2 6:3 68:5 5385 1-8 2:5 Choke bored.

BMR. BusHmMans RIvEeR AREA.

Types here are again limited in size. Diameters lie between 10-2
and 14-8 cm. and heights between 6-0 and 9-8 cm., apart from a single
thick specimen which is either anomalous or an erratic from the
Knysna-Tzitzikamma area.

i. (14-9 UX 48-3 1600) Burghersdorp. BH. i.
li. 12-0 9-4 78:3 1355
iii. (10-6 11:9 112:3 1340) KYTZ. iii.
thy, | Liter 6-0 59-0 945 Aliwal North. KTZ. vi, EL. i.
v. 10-4 Pal 74-0 835 NO We thx

Further comment on this area will follow the Port Elizabeth material.
There are few affinities with surrounding areas. Types ii and v here
suggest a general similarity to LTV. 1 and ui respectively, but not
close. The main affinities are with Knysna-Tzitzikamma, an area
some little distance away (p. 85).

The Bored Stones of South Africa. 79

ADELAIDE.
ELM 87 pene) 70 56:0 1095 40 #£2-2 1-47 ke.

Borua’s Post.

AMG 634 . 106 1

19 112-3 1340
699 ies 10:3 7:7

5-5 ‘0 Erratic. KTZ. iu.
74:8 815 4-6 9
Fort Brown.
AMG — 51 MELA: 69 56-1 1045 50 2:5
656 . 10:5 Weak = 7ie33R3 850 5:7 3:0

Fiso RIveEr.

AMG. 703 5) a llZs®) leZenAs8:3 NG00) 5-2). 3-5) 12-0 ket
636 = liber 8-9e 6-1 L220) (95:5) 3
SIDBURY.

AMG 1568 . 11:8 PAE DO IBD ne G8) 9 Biol

ALICEDALE.
AMG 1643 . 11-5 80 69-1 1060 54 2-0
et . 11:0 68 61-9 825 56 2-0
RBM 6I4 Y . 11-7 6-1 52-3 835 5:8 3:2

SANDFLATS.
AMG 1561 . 11-7 Oe 82-9 s3s0e 4-50 223

BUSHMAN’S RIVER.
PEM 614B . 13-1 9-2 0-1 80 5-2 oor 20) koe

PE. Port ELizABETH AREA.

A scattered and wide distribution that should react to future
careful collection and association with coastal midden materials,
inland caves, river sites, and soon. It is obvious that the local rivers
make the picture difficult to see.

re (o°9 10°8 77:8 2085) UTV.i.
ii. 13-5 eu 57:0 1405 Kai WV er.

iii, (15-2 530) 34-7 |) 1995)

iv. (10-4 68 65-4 735) ZWB. iv.
v. 11-4 53 46-5 690 EL. ii.
vi. (10:8 3:9 36-2 455) BL. iv.
vii. 9-3 46 49-5 400 KZ. xiii.
viii. 7-7 5-1 66-2 305 UTV. vii.
ix. (3:3 3-5 108-1 35)

There are general similarities to the Upper Thornveld and to Hast
London. The link with the former is obviously by way of Cradock
and the Fish River, from a common origin.

CRADOCK.
AMG — sueleo MeO EDG:S aoe Orin! 238
ASJ 4/35/2 . 9-0 4:5 50:0 365 (4:0) 2-1
4/35/— . 8-1 5-5 67-9 360 3=64:0 an
4/35/28 . 7:5 46 61:3 PAY BB} 18)

80 Annals of the South African Museum.

SOMERSET FEAST.

AMG 1618 . 13-9 8-5 683 1640 5:5 3:4 One pole flattened.
11:0 (5-5) 50-0 665 5:0 2:5 Both poles flattened.
9-7 4:8 49-5 455 4:4 1-9 Battered!
7:2 4-8 66:7 250 4:5 2-0 Pecked. Grinder.
GRAHAMSTOWN.
AMG E580). 2 114235 10-5) 2 73:4 2150) 822) 2-0
NMP 100 . 10-4 6-7 64-4 720 3:4. 1-9 Idee
TOOTABI.

PEM 614Q . 10:8 3:9 36-2 455 45 25 5-8 kg.

Port ELIzABETH.
AMG 630 2 1325) 1 82-2 2025 etd ee

HumeEwoop, P.E.

PEM 615T . 13-4 73 542 13810 4:7 13 #£4241:8kg.
59 5 thes} 50 44-0 640 45 2-5 Oval hole.
i 8-9 4-6 . 51-7 365 3:2 19 0-45 kg.
a 8-2 50 61-0 339 25 .. Obekee
South Enp, P.E.
PEM 68/ 12-0 54 45:0 780 50 2-8 0-95kg. 14feet deep.
5 10-4 6-8 65-2 735 5:0 2:7 “Oke -
ss 3-2 3:3 103-1 34 2:38 16 0-09 kg.
CaPE RECIFE.
PEM 614A . 13-4 75 56-1 13850 46 2-2
15-2 53 34-7 1225 53 2-6 Anomalous. Angu-

lar.

The Port Elizabeth area appears to be a migrationary intrusive
route from upcountry to the coast. It shows only one type similar
to those from Knysna-Tzitzikamma, and, like the Sundays River
route and the localised Zuney-Maitland Mines material, this area cuts
off the Bushman’s River types (where three averages conform to the
Tzitzikamma range) from farther west. This would suggest that the
Bushman’s River material was later, and spread across existing
populated areas to flourish again on the far side of them—presumably
the spread was from BR. to KTZ., though this is by no means certain,
as the opposite migration of a limited number of types would be
perfectly plausible. My presumption rests upon the supposition that
areas of attraction will tend to collect surrounding types early, digest
and assimilate them into a secondary cultural centre, then launch
them forth once again as new migrations in new patterns.

BIBLIOGRAPHY AND COMMENTS.
J. C. Rickard,* one of the most efficient archaeologists who has
worked in this country, drew material from here and from Kast London.

* J. C. Rickard, ‘“‘ Notes on some Neolithic Implements from South Africa,”
Cambs. Antiq. Soc., Comm. V, 7, 1881, pp. 67-74.

?
+
oe
te
2

ie” ma ae we

The Bored Stones of South Africa. 81

He associated bored stones with his Earlier Midden (pottery scarce
or absent) period, at a site at North End, Port Elizabeth. At Hast
London he likewise associated a small bored stone with his Later
Midden (no cutting implements, but ornamented pottery, etc.) period.
This and the associated paper are valuable, and have never received
the attention they have merited in this country.

ZMM. ZuNEY-MAITLAND MINES.

These few examples certainly form a series of extremely small types.
Diameters lie between 5:0 and 8-0 cm. only, and the heights do not
exceed 4-5 cm.

i. (7:
i. 5:

Possibly this area should be included in the South Coast Region,
but until further material is available it may well remain in association
with Port Elizabeth. Only the Knysna-Tzitzikamma and the
Belvedere-Heidelberg parallels are important, Hast London and the
Transkei less so, while the suggested parallel with the Trekveld has
no real value. There appears to be no relationship with Sundays
River, Tzitzikamma caves, or with Bushman’s River.

Much that has been said immediately above about the Bushman’s
River area in relation to Port Elizabeth would probably apply here.

© 00

9 52-7 Zo) ICD. xvi, BE xa EL. vy EV. ii.
if

3:
3° 56-4 94 KEE Zs avilt) Dy xiv ee yi;

ZUNEY, ALEXANDRIA.

Univ. Stell. . 7-5 4-4 58-9 24s Oe At
5:6 2-6 46-0 81 2-8 0-9 Slightly oval stone.
ZO eL7

Mairtanp Mines Farm.

PEM 615/A. 79 3:
5S) 3:

1
3 0-17 kg.

CO Or
for)
@¢
i)
—
(5%)
jm

Obviously all knob-kerrie heads—no digging stones.

IX. SOUTH COAST REGION.

Rivers are important in a somewhat different way in this region.
Before they had been highways. Now they appear to be backwaters
of distribution. Along the South Coast the coast provides a 300-foot
shelf, rising gently from the narrow strip of sand or soil that con-
stitutes the Middle Stone Age 20-foot raised beach. This open belt
is wide enough to keep wandering peoples near the coast, and varies
from 5 to 15 miles in depth. Behind it stand serried rows of moun-
‘tains, two, three, or four deep, rising to heights of 5000 feet, and

VOL. XX XVII, PART 1. 6

82 Annals of the South African Museum.

dropping steeply to enclose valleys 1000 or 2000 feet above sea-level.
Finally, behind all this stands the lip of the plateau starting at 3000
feet and rising slowly until 4000 feet is reached, with higher mountains
standing out of the plains. The mountains thus form highly effective
barriers between the inland plateau and the coast. The result of this
has been that the enclosed parallel valleys are fed by T-shaped or
cross-shaped streams, running from east and west, then meeting to
pass southwards to the coast through some lovely entrenched valleys
in the mountain ranges. Peoples seem here and there to have been
caught up in the valleys of these T-streams, and types have a local
distribution, only partly reflected at the coastal peneplain where the
river debouches into the sea and immediately on either side (e.g.
Grootrivier).

Added to this, the whole stretch of country from Riversdale to the
Gamtoos River mouth was thickly forested before 1875 when a
devastating fire denuded all but the deepest valleys from Riversdale
to the Knysna River, while even across that river game was roasted

where it stood. There are thus three types of environment present

in a confined area—upland plain, forest, and coastal midden country.
This will be referred to once again when we deal with the various areas.

SR. Sunpays River AREA.

This river drains the plateau country and has its sources in the
Nieuwveld and Goup. In the Preliminary Survey I included the
whole basin as a single area, but the Nieuwveld appears to have only
one point of contact, Aberdeen district. As I explained above, the
PEM 614 F series has been divided into two, and I have found it
necessary to do the same with the ELM 70 specimens. The museum
divides this group into 70, 70 A,and 70 B. This last has been included
under the Knysna series. The series is mainly coastal midden.

i (98° 73 745 700) KTZ, ix, ZWBoive Wives
Way SPC 7-0 80-4 530
Le) S26 5:0 56:8 390 KTZ. xiii.
Vier (do ol 70:8 265) UTV. vi.
Specimens are confined to heights from 4-5 to 7-5 cm., and diameters
between 7-0 cm. and 10°0 cm.

ABERDEEN B.

PEM 614F . 8-9 75 84:3 595 40 2-2
9-3 5-6 60:7 485 40 ae
8:5 64 75:3 460 40 2-1
9-0 5:0 955-6 405 32 2-0
8:5 4:7 55:3 340 38 2-2
7:0 5:2 74:3 255 34 2:2

The Bored Stones of South Africa. 83

Kirin Rivier A, HANKEY.

ELM 70A . 8-7 70 80-0 530) 3°2) = Ey,

70 Be (is) 5:0 66-7 280 3:2 1:5
GamToos River, HANKEY.

PEM 615C . 9-7 75 77-4 705 44 1-6

Po 3979 Ge aed 695 44 2-4

Cre trs'9 45 50:5 350) 2 1-7

TZITZIKAMMA AND KNnySNA AREAS.

The next two areas need a little explanation. It was pointed out
earlier that before 1875 the Riversdale-Gamtoos belt was a single
forest area. To-day there is a distinct change in environment at
about the George-Mossel Bay line, which was absent or only slightly
important to the makers of the bored stones. Types thus cut directly
across present boundaries, but the wide stretch of the Knysna River
seems to have provided some sort of barrier, and may be taken as a
boundary between areas.

The vast numbers of specimens that have been marked “ Knysna”
have made it difficult to discover anything very definite about that
locality. We seem to be presented with so many types that they
suggest that our system has broken down, and that we are dealing
with a completely haphazard distribution of types. Further analysis
shows that this is not the case.

Only one mode of analytical approach can be fairly employed at
present, the subtractive method, based upon our knowledge of indi-
vidual sites. I have therefore subtracted the following sites from the
whole, and present them as a single area, overlapping the Knysna-
Tzitzikamma area :—

ASSEGAIBOS. ROBBERG CAVES.
TZITZIKAMMA Mountains. HAsTERN HEADS CAVE.
TZITZIKAMMA CAVES. Heaps Roap.

Forrest Hau SHELTER. WOODBOURNE.

To these I have added one example from the Tzitzikamma Middens,
one from Hankey, and three from Knysna, all of which agree with the
types developed from the above series.

As to the remainder, I have perforce had to group all of them
together. If we had more information we should probably find that
the list given above covers the cave or shelter types, while others
that are represented in the Belvedere-Heidelberg area may eventually
yield a Coastal Midden series and a Coastal Plain or Forest series.

84 Annals of the South African Museum.

This possibility should be looked for as further specimens are collected;
it is discussed later under the appropriate areas.

Here we have a supreme example of the need for an exact provenance
for each specimen. This area is split by gorges and divided by almost
impenetrable forests, traversed only by elephant paths. It yields
three distinct environmental belts, each roughly related to altitude,
the 20-foot coastal belt, typified by open middens and cave middens;
the forested gorges and country rising to the 300-foot level; and finally
the grassy foothills immediately below the mountain ranges. This
differentiation should be reflected with varying intensity from the
Gamtoos to the Gouritz Rivers and perhaps farther along the coast
in both directions.

GR. GROOTRIVIER.

The Grootrivier is the western affluent of the Gamtoos River. It
seems to provide a backwater where only one type occurs:

13-0 11-8 90-8 1995 KTZ. i.

This obviously links with type i from Knysna-Tzitzikamma and

is probably related to examples at Port Elizabeth, Ladismith,
Victoria West, and elsewhere. It suggests a movement along the
valleys at the Cape, which only now and then found its way to the
coast.
WILLOWMORE.

PEM 1429/87) Us: We 85-4 29 aol 2A

FULLARTON.
PEM 614X. 12:9 116 90-1 20830 48 23

MILLER.
PEM 614Z . 12-4 12:2 984 1875 43 £43:8

TZC. TzirzikAMMA CAVES.

This area partly overlaps the Knysna-Tzitzikamma area, and
enough has been said above to explain the position.

i, 11-8 4:9 “41-5 | 680 PE. v, BH. iv.

ii, 140 3:4 245 (665)

ih, OH GE SR a Cl ae

iv. 10-6 2-2 | 20:7) (045) BEL. xii.

v. (75 45 59:0 225) ZMM.i, SR. iv, DBV. vi.

Vin Or led Tee 65)

It will be observed that most of the indices lie below 41-0 per cent.,
while all specimens are Jess than 5°5 cm. in thickness.. Only one type

ih,
Pe

Sa ee.

The Bored Stones of South Africa. 85

runs through the Belvedere-Heidelberg area, where it is represented
at the two extremes and at Mossel Bay (type TZC.i). This latter I
have compared with BH. iv (11-7 x 5-4 cm.) and type TZC. iv with
BH. xii (8-7 x 2:6 cm.) as I believe the BH. series to be derived from
these; the larger type persists farther west as BOT.1i, which suggests
a coastal type.

Kern River B, HANKEY.
ELM 70B . 13-7 3°5 25:5 (660) 3:7 3-2 0-91keg. Limestone.

ASSEGAAIBOS. i

PEM 615/F . 12-2 S1 41-8 UGOP e183) 1-6) 70777) ko. One face
concave.

TZITZIKAMMA MOUNTAINS.
PEM 615/0 . 7-4 4-4 59-4 240 36 41:9 0-3kg. With female

skeleton.
9:7 3-1 20:5 (290) 30 2:6
TZITZIKAMMA CAVES.
PEM 1410 . 14:6 4-2 28:8 (895) 41 2-9
Pee! Le 30 27-1 (exe “aires ilar)
TZITZIKAMMA MIDDENS.
PEM 94 ey ess 4:4 38-5 580 42 2-2
Forest Hatt SHELTER.
Univ. Stell. . 10-0 4:0 40:0 400 5-1
RoBBERG CAVES.
ASJ 11/42 . 11:5 5:2 45-2 G9IOs Wade 2-2
‘ 13-6 2 19-1 (480) 5-0 : Trreg.
SAM 2981 . 11-0 2-37 120-9 (278) 2:00 eit
11-0 2-1 19-1 (254) 2-4 2-2
ASJ 11/42 . 10-2 2-2 21-6 (229) 5:0 3-2 Irreg.
EASTERN HEADS CAVE.
SAM Bain . 6-2 Ui) 271-4 65 23 #£421:8

Heaps Roan.
PEM 615/P . 10:3 3:1 30-1 330 3:9 2:6 With skeletons.

WooDBOURNE.

A.W.Robinson 7:6 4-6 60-5 265 3:5 2-4 Chalky limestone.
Odd grooves, ap-
parently not re-
lated to artefact.

KNYSNA.
SAM 1832 . 10-5 2-0. 19°0 — (220) 2:0- 1-0
Ps). 12-0 54 45-0 (80 onl 5 2-7
A.W. Robinson 12-2 4:8 39:5 WS) 9-7. 271) Ss. pebble:

KTZ. Knysna-TZITZIKAMMA AREA.

The multiplicity of types here is somewhat terrifying, but most of
them are reproduced in surrounding areas, either as directly analogous

86 Annals of the South African Museum.

groups, or as the results of divergent copying. This area is therefore
to be regarded as focal.

i, (13-3 12-0 89-8 2120) \ GRS, ULV

Toh, | 1a 9-0 68-2 1570 SWB. i, BH. ii, WV. iii.
iii. (10-5 12-2)" 116-2 1345) BMR. iii.
iv. 12-1 8-7 71-9 1275 SWB. ii, BH. ii.
v. 12-8 7-4 57-8 1210 IBID Ti
vi. 11-4 Tee 63-1 935 BMR. iv, EL. i.
vii. 10-2 8:7 85:3 805 BH. iii.
viii. 11-0 6:2 56-4 750 UTV. iv, NV. iv, BDV. iii.
ix. 10-0 7-4 74-0 740 SWB. iv, BMR. v.
xe tO 7:9 86-8 655 STK. iv.
Kis O22 6-6 71-7 560 BH... vi, GRKoiv,, BOR sir
xii. 10-4 4-4 42-3 480 PE. vi, BDV. v.
Kili, 8-5 6-2 72-9 450 UTYV. v, BH. vii, WV. vii.
xiv. 9:3 5-0 53:8 430 PE. vii, SR. iii, BH. viii.
xv. 80 5:3 66:7 330 PE. viii.
Xvi. 9-2 3:0 32:6 255 BH. xii.
xvii. 7:8 4-0 64-1 245 BH. xi, ZM. iv.
Qiang) (7/2! 2-8 40-0 140) CF. vi.
xix. (4:5 3:0 66-7 61) BH. xiv.
xx. (5-0 1-4 28-0 35)

Types xi and xii are divided into two groups partly on analogy |
with Belvedere-Heidelberg, but also because there appears to be a
double distribution presented here: one linking with the Upper
Thornveld, where the distribution of type xiii was shortly discussed.

Types vil, xX, Xlv, xvi, and xviii all include material from the
Tzitzikamma middens, and are also related to the Sundays River
middens or to those of the Belvedere-Heidelberg area. I would
therefore suggest that they represent a mdden series, separable from
the remainder.

Types ii, iv, and ix are related to the Zwartberg area, and so repre-
sent an inland series. Type 1 is too closely related to the Grootrivier
to be mistaken. it is also an extreme type and therefore more
certainly analogous. The comparison with the Upper Thornveld is
not so exact, but is suggestive. Type viii is related to Nieuwveld
material and is also an inland type.

Types i, lll, V, Vi, ix, x are absent in area BH.; see the note given
at the conclusion of the figures for the present area.

TZITZIKAMMA MIDDENS.

PEM 94 . 98 89 90:4 885 3:8 2-1 1-2 kg.
88 75 864 580 3:3 Lo 0-8 kg.
93 49 52:2 425 3:9 2-4 0-53 kg.
9:27 53-3) oD oO 4-3) oes
TIA DAES esta lS SI hS7/ 4 0-18 kg.
Dr Poort.
ASJ 8/36 . AN rete ahete WAlda SSO). veksy sos
Ti7 8-9") 76:1 -1220) 3:9) -2°5

The Bored Stones of South Africa. 87

Der Poort—continued.

ASJ 8/36. 11-2 7:8 69:6 980 4:4. 2-9
115 7:3 63:55 965 50 2:5
9-9 8-4. 84:8 825 3:9 2-1
11:0 63 57:3 760 43 £2-4
9-] 5:3 58:2 440 3:3 2:3
7:9 6:5 82:3 405 3:5 2:9
92 46 50:0 '390 38 2:6
84 5-1 60:7 360 3:7 2:0
82 41 500 275 30 19
Knysna AREA.
AMG 1578 5 MIB) pace! 91-2 2295 4:5 2-9 Pecked.
SAM 1114 - 13:0 11-5 88:5 1945 4-2 2-2 Ss. Pecked.
12:9 9-5 73:6 1580 4:5 2°8 Ss. Pecked.
Drennan Coll. . 13:6 85 62:5 1570 41 2:7 Ss. Pecked. Slight
cuts at narrower
mouth.
A.W. Robinson 13:0 93 £715 1570 3-8 $1:8 Coarse Ss. Pecked.
10-5 12-2 116-2 1345 48 2-2 Coarse Ss. Pecked.
13 2 OS2d UL o40) seo: an 2k Coarse Ss.
12°66 8:4 66:7 335 4:9) 2-2 Coarse Ss.
SAM Bain 5 B20) 7:4 56:9 1250 4:0 2-6
A.W.Robinson 11:6 86 74:1 1155 59 2-2 Ss. Pecked.
SAM 1114 - 10:5 9:0 85:7 995 4:8 2:5
Sharples 17 a 0-485) 802275920) 4:3 hs Pyriform mace.
Hole. 5 cm. deep.
SAM 1832 see OD-Sa1 6 G05) taco 19,
A.W.Robinson 11:5 68 £583 900 49 2-9 Coarse Ss.
UCT 23/18 ee 1-3 16:9) 61-0) 850) 3-9) 2:1
TMP 1868 . 10-2 7:6 74:5 790 4:7 1:9
A.W. Robinson 10:2 7:6 74:55 790 40 1:8 Coarse Ss.
NO-Gee G23 spoueSie soO0s. 420) | 225 Coarse Ss.
SAM Bain . 10:0 7:5 75:0 750 4:5 2:2
Ep Seis! tsi 6-0 54:5 730 5:0 2:5 Coarse Ss.
SAM 1832 fGen 8:08 83:3 foo) 2:6 2-2
UCT 23/18 Pel Oo en G:O) )vo4-5) 25 3-6) 2-1
SAM 1114 9-47 78:0 “85-1 705 4-1 . 2-4
SAM Bain Oi ein) 14:2) * 680) = 43° >. 2-1 Coarse Ss. Pecked.
A.W.Robinson 9:7 7:2 74:2 680 4:0 2:6 Very coarse Ss.
92 79 85-8 670 38 2:0 Ss.
SAM 1114 90 80 889 650 40 3:5 Coarse Ss.
A. W. Robinson 86 80 93:0 590 3:5 2:2 Coarse Ss.
Sharples 19 9-4. 6:7 70:2 590 4-] 1:8
A. W. Robinson 94 66 70:1 585 34 21 Coarse Ss.
SAM 1832 89 73 82:0 580 4-3 ms Coarse Ss. Pecked.
9-9 5:4 54-5 530 4:3 5 ” 2”
SAM 1114 9-055 6:6) 73:09 -535)-4-0 Se Ss. Pecked.
A. W. Robinson O22 Gea 68-5) 2 535n 4k 1:8 Coarse Ss.
103 48 466 510 62 3:2 Ss.z
10:8 43 40:0 500 4-2 2:0 Ss.
Sharples 5 9-7 53 544 500 36 1:8
SAM 1114 10-4 44 42-3 475 5:4 1-8
A. W. Robinson 93 54 58-1 465 3:9 1-8 _ Ss.
SAM 1114 88 58 659 450 40 2-6
85 6:2 72:9 450 4:5 2:1
UCT 23 92 10:0 44 440 440 40 2-1 =TM.Ss.
93 51 548 440 29 1:9 TM.Ss.
A. W. Robinson 86 59 683 435 3-6 2:0 Coarse Ss.
10:37 4:0) 38:8") 425° 93-7 2-2 vs
8:9) 95-2.) 158-4, 7410) 3:9 2:3 Coarse Ss. Pecked.

88 Annals of the South African Museum.

Knysna AREA—continued.

UCT— . . 95 4:5 47-4 405° 4:0 2:2) TM. Ss:
SAM 1635 - 90 50 555 405 44 2:4
A.W.Robinson 9:0 4:9 54:4 400 (4:6 2-8) Coarse Ss. Conical
bore.
(8:8) 53 60:2 400 3:5 2-1 ‘Fragment. Coarse
Ss.
84 55 65:4 390 3-2 1:9 Coarse Ss.
91 44 483 365 3-4. 41:6 Ss.
SAM 1832 80 54 67:5 355 3:9 2:0
A.W.Robinson 9:6 3-7 37:5 340 2-5 2:2 Ss. pebble.
8:0 50 641 305 24 1:8
95 3:0 31: 270 3-0 2-4 Slatey shale.
8-0 41 51-2 260 3:5 41:7 Coarse Ss.
7-6) 23°87) 50°05 9220) 2-Sae len Sandy shale.
SAM 1114 eon dO 9 Ol On 2202-8
O22 2-4 26205) 2a ee
Sharples 16 8:6 2:8: 33:0) 200) 3-0 >) 2°3))))Brokenr
A.W.Robinson 7:0 3:3 47:1 160 3:2 41-9 Ss. Grinder.
SAM 1474 oi 457 3-0 an G6A7 61 1:7 0-8 Iron bored. Ss.
Anomalous.
1114 50 1:6 32:0 40 25 1:3
1832 50 1:2 24-0 30 2:0 1:2
WILDERNESS.
MMK 508 yp 23 Tl bier Ota) 4A See

The South African Museum series, numbered 1833, presented by
J. 8. Henkel in 1915, was said by Dr. Péringuey to have come from
sites “between Knysna and Plettenberg Bay.’ He asked Henkel
for further information, but no reply was received on this
question. The scatter of the various types strongly suggests a
midden series.

In comparing graphs of the Knysna-Tzitzikamma and the Belvedere-
Heidelberg series the following points of difference are important :—

i. None of the BH. specimens is thicker than 9-5 cm.

u. The KTZ. shows a marked gap between 9-5 cm. and 11-5 cm.
Types KTZ. i and i are therefore absent in BH.

il. In the BH. specimens an analogous marked gap is seen
between 7:0 and 8-0 cm. in thickness.

iv. The KTZ. series shows types v, vi, 1x, and x between these
two limits. These are again absent in BH. They seem to have
their origins to the north-east.

These points suggest strongly that the sources for K TZ. 1, iii, v, vi, ix,
and x were not in touch with BH. Unluckily most of these sources
are not known to us with any certainty.

My thanks are due to Mr. A. W. Robinson of Plumstead for his
courtesy and kindness in allowing me access to his collection and
notes.

The Bored Stones of South Africa. 89

BIBLIOGRAPHY.

There is little comment by early writers on this part of the country.
Sparrman * is the only writer to mention the bored stone from any-
where in this region. He describes a chance meeting with some
Bushmen in the Lange Kloof leading from George to Uniondale.
““Most of these refugees carried a thick, stout staff, generally headed
with a heavy gritstone of 2 lb. weight or more, rounded off, and with
a hole bored through the middle of it, to increase the force of the stick
for digging up roots and bulbs from the ground.”

SWB. SwaRTBERG AREA.

This inland area, related to the valley between the Kammanassi
and Swartberg ranges, seems to be another backwater.
i, (13-8 9-5 68-5 1795), TAS i.
ue 125 8-2 65-6 1280 KTZ. iv, BH. ii.
ii. 11-4 8-4 73°7 1090 BH. iii, NW. iii.
iv. (10-2 7-1 70-0 730) K'TZ. ix, WV. v.

The relationship between this area and Knysna-Tzitzikamma is
very evident, and helps to explain the multiplicity of types in that
area. These are, of course, all inland types, and therefore provide
a clue to the composition of KTZ. No examples show a thickness
less than 7-0 cm., or greater than 10:0 cm. This again precludes any
relationship between this series and KTZ. i and iii.

SwARTBERG.
SAM 516. . 123 85 69-1 1285 46 3-0
MP. . 11-6 8:3 71:5 1120 4-2 2-1
ae LOM ie Orn GOs wlio) 403 eo
VOORUITZICHT.
Oudts. Mus. . 13:0 80 61:5 1350 4:7 2-7
OUDTSHOORN.
PEM 614M . 136 99 72:8 1830 50 2-3. Kammatie.
A. B. Pocock . 14:0 9-0 64:3 1765 46 # 3:5
MMK 588 Pe Ge Oe 8-4 225.9 5-1 2223 Koetzee’s Poort.
Oudts. Mus. . 11:0 80 72:7 970 4-0 :
PEM 614M . 102 7:2 706 740 = 6:3
‘HAZENJACHT.
Oudts. Mus. . 12:0 79 65:8 1140 4:0 2-6

BH. BELVEDERE-HEIDELBERG.

This area shows fundamental similarities to the Knysna-Tzi-
tzikamma area, but other factors affect it, notably the presence of the

* Sparrman: A Voyage to the Cape of Good Hope. 1786. First edition, p. 328;
second edition, p. 307.

90 Annals of the South African Museum.

Gouritz River. I have already assessed the environmental factors
here, and need hardly recapitulate. The Grootrivier element is absent
in this area.

i. (15-9 6-9 43-4 1745) BMB.¥i, PE. iii.
ii. (12-6 8:7 69-9 1385) KTZ. ii and iv, ZW. ii, WV. iii.
ii. 10-7 8:2 75:0 1000 KTZ. vii, NV. iii.
iv. 11-7 5-4 46:1 740 BOT. i (TZC. i).
Won keke) 93 110°8 670 WV. vi.
vi. (9:6 6:4 66:7 590) BOT. ii, GK. iv, KTZ. xi.
vii. 8-7 6:3 72-4 475 KTZ. xiii, UTV. v.
viii. 9-1 4:7 51:6 390 KTZ. xiv, BOT. iv.
vba (78) 6:1 80:2 350)
x eG 5:2 68:4 300
xe) Se) 4-3 53:1 280 WV.* viii.
Xi, | O-7 2:6 30:0 200 KTZ. xv (TZC. iv).
xii. (7:0 3°8 54:3 186) BOT. v.
XLV 3°3 62:3 93 KTZ. xviii, UTV. viii.

There appear to be two main groups here, those with heights less
than 7-0 cm., and those with heights greater than 8-0 cm. (types i, u,
ill, and v). These latter appear to be znland types. While I have
added comparisons for type 1, these are unsatisfactory. Affinities with
the Swartberg and with the Winterveld and the Nieuwveld areas
suggest that ii, 111, v, and xi are inland types. Type vill is a persistent
coastal midden type, and appears in the Port Elizabeth, Sundays River,
and Knysna areas. Type vii is also very persistent and can be traced
from the Free State, through the Upper Thornveld (where the dis-
tribution is more fully given) to here. It is almost certainly a lowland
type here, though not necessarily associable with middens. Still
Bay is the last site at which it occurs, and it may be regarded as almost
certainly Wilton. Type vii may be related to type vi, which occurs
in the Gouritz-Karoo area and Bot River.

The gap between 7-0 cm. and 8-0 cm. in thickness is not evident in
the KTZ. series. The thickest example from here is 9-6 cm. Contrast
this with KTZ.

BurFFALO POINT.

A.W.Robinson 5:1 3:8 74:5 oe) Bee a Sandy shale.
WESTFORD, BELVEDERE.
SAM 4895 - (G1-O))) 8:4 76:47 OTD 4-555 2:5 ess:
10:2. 8-1 79:4 840 4:3. .2°2) “Sst
12:1 (5:0) 41:3 730 42 2:7 Bad condition.
88 41 466 320 3:3 19 Circumference
abraded TM. Ss.
90 40 444 325 44 23 ~& Ssz
OAKHURST CAVE.
UCT — 91 96 105-4 795 3:5 2:6 Grave VI A.
76 60 78:9 345 3:7 2:2 Grave X.

— a

The Bored Stones of South

BERGPLAATS CAVE.

SAM 1635 ? 90 5:7 63:3 460 3:7
KAAIMANS RIVER.
SAM 1635 : 94 51 542 450 43
LIMBERLOST, WILDERNESS.
AMG 3001 ; 91 29 319 240 5-1
GEORGE COAST.
TMP 9174 15-9 69 43:4 1745 46:3
L-l 8:2) 73-8 1010 °5-4
8-9 85-1 eboi-2 405 4-2
GREAT BRAK.
SAM 4838 : 86 64 74:0 475 3:8
HARTENBOS.
SAM 1635 124054) 775-0 = 2807 | 23-6
Mosset Bay.
MMK 865 12-8 92 71-9 1510 5-0
SAM 1777 p24 8-2) 66-1 | 1260 4-6
1635 10:8 81 75:0 945 4:0
M.B. Library PP 85-3) 47-8) 1655. = 457k
97 64 660 610 4:3
MMK 705 83 6:7 80-1 460 3-9
SAM 1635 8-7 44 506 305 3:5
76 50 65-8 290 3-2
$-O8) 25-0) 962-5 9325, 3-5
8-2 4-5 54-8 305° | 3-5
1586 8-0 4-5 56-2) 290 4-5
TMP 1853 Sou ee wrol-s. FO5= (3-9
Gouritz River Mout.
SAM 506. ie 29-G) 16:4 66:7" 9590: | 3:7
9-4 5-2 55:3 460 —3-4
8:4 46-5 77:33 460 4:0
92 48 55-8 405 4-5
OF 4:51 44:5. 350 5-1
mOn 7-3-8) 54-3" 8b 2-8
| 50 2-4 48-0 60 2-0
StrL Bay.
HHC 1057 ; 87 60 69:0 455 3:0
BLOEMBOS.
Heese Coll. 7. 76 60 $78:9 345 £3:8
18. S-1- 4-1 50-6" 270° 4-6
14. Sule 24cm 150) 62-8
10: 5-9 (3:7) 162-7. 130° 3:1
Se 75-0) 3-6) 72:0 90 2-0
RIVERSDALE.
PEM 615 H 74 94 1272 515 4-5
Heese Coll. 6. 89 89 100:0 705 3-4

HEIDELBERG.
SAM — . ek -Our 5-60 s46-08 | S05. 5-6

Africa.

2-0

2:3

1-8
1-9
1-2

ee ND
ke GO Hy H= 00

GO 1

91

TMS. pebble.

Oval hole.

Coastal midden.

Ss.

Pecked.
Ss. Squarish.

Slightly triangular.
Found at depth.

92 Annals of the South African Museum.

BIBLIOGRAPHY.

There is little local literature of value on the bored stone, though
papers by Dr. C. H. T. D. Heese are important in throwing light on
associated Wilton material from Bloembos and Riversdale. The two
specimens from Oakhurst are associated in Goodwin * with Smithfield
B material, Grave VI A with certainty, Grave X with slightly less
certainty. They are illustrated in Plate V, C (centre of picture) and
Plate VI, A (left of picture).

GK. Gouritz Karoo AREA.

The Gouritz River is a cross-shaped stream, draining the Little
Karoo by the Touws River affluent; the Great Karoo by the Gamka
River; and the Armoed valley by the Olifants tributary. It is a
wide area, and therefore grouping is bad, but there is a certain unity,
and types link with sites near the Gouritz mouth in the Belvedere-
Heidelberg area. This is most noticeable at Bloembos. The most
provoking specimens are from Bellair, they fail to link with any local

types.

i. (125 7:6 60:8 1285) SWB. ii (Bellair).
ii, (132 5:5 41-6 960) (Bellair).

iii. 10-4 58 655-7 625. NV. iv.

iv. (92 64 70-0 540) BH. vi, KTZ. xi.
vy. (102 35 348 380)

vi) (con) G20 S27 355) BH. ix.

The larger specimen from Bellair shows some resemblance to a
Swartberg type, while type ili suggests an affinity with an Aberdeen
specimen from the Nieuwveld area. Only one specimen (from
Bellair) is thicker than 6-5 cm. The peculiar coincidence that both
Bellair specimens are anomalous, though submitted at different times,
probably means that they are single representatives of local and
normal types not yet adequately represented from other sites.

NELSPOORT.

Sharples E
Sharples 18 . 10-2 6-1 60-0 635 5-1 25

LAINGSBURG.
Heese Coll. 19 9-4 6-1 64-9 540 49 2:5

Hex RIver.

SAM 765 : 9:0 65 72-2 525. are, 2:0
ANYSBERG.
SAM 175 Lh) 6-4 58-2 Tio. 4:4 | 2-4

* Goodwin, “The Oakhurst Shelter,” Trans. Roy. Soc. S. Afr., vol. xxv, 1938,
pp. 251 and 254.

1

The Bored Stones of South Africa. 93

BELLAIR.
SAM 367 = ais 76 60-8 1285 5:5 2-5 Anomalous.
178 a, la 5-5 41:6 960 5:0 2-4 Anomalous.
SANDFONTEIN.
SAM 176 ‘ 7:5 6-4 85:3 360 46 2:8
BUSHMANSFONTEIN.
SAM 180 . 10:0 5-5 55:0 550 4:0 2-2
PRINCE ALBERT.
SAM 694 Bn LO=2 3:00 2 oDre 375 44 2-8
7:6 6-1 80:3 350 46 3:0

ARMOED, GOURITZ BASIN.
MMK 504 . 10-6 5:8 50°3 650) 95:85) 223

VREYERSBERG CAVE, GouRITZ R.
EH EC1050 = 29°5 5:5 =: BT -9) A957 93:9) 7 1-95) 0-72) ke.

MELKHOUTFONTEIN, GOURITZ R.
HHC 1055 . 10:5 30M oor 385 3:9 1-9

X. AGULHAS REGION.

This region lies east of the Hottentots Holland range and includes
several wide valleys between mountains. I have divided it into
three areas—West Langeberg, the Breede River, and the Bot River
Valley. To the north the boundary appears to be the Keeromberg
and the Langeberg as far as Heidelberg (Cape), where the coast is
reached.

WL. West LANGEBERG.

This consists of sites surrounding the western end of the Langeberg
range. It is surprising to find a mountain range forming the centre
of an area, but considerable new material submitted by Mr. G. Keet,
from Montagu, SAM 5018, has changed our knowledge of this area
very considerably. I would like to take this opportunity to thank
him for the trouble he has taken in this matter.

is ISOM 10:8" 83-1 S25 CHn ii

ii. (12-0 SO | Wee) Ma) IDA ie a
iii. (10-0 Wl | a 770) SV. iii.

iv. 11-4 4-3 42-1 625. BH. iv.

v. 12:5 3:2 25-8 510 CH. ix.

vi. (7-5 49 65-4 275) BDV. v.
vii. (7-6 21) 7 ZIGPI| 185)

viii. (5-9 5-0 848 175)

Type i is reminiscent of the Grootrivier material, and perhaps
Ladismith supplies the link here. The gap in thickness, a notable

94 Annals of the South African Museum.

feature, occurs between 5:0 and 7-5 cm. The greatest thickness is
Lt-0 sem:

WORCESTER.
SAM Bain . 14:0 105 75:0 2060 52 2:3
10-0 ie lew 770 =4-4 *
INMIP 2631) ioe eles 5:1 45:1 640 3-8 2-1
SAM Bain . 12-4 3:4 27-6 525 37 2:0
12-0 3:5 29-2 505 = 33-0 a
246 7:2 4:8 66-7 250i os ae
164 a Ook 50 81-9 185 26 1-8
Heese Coll. 15 7-6 3-2 42-1 185° 2°74 O28 alee
ROBERTSON B.
SAM Bain. 13:2 2:7 20-5 500 50 3:5
MontTacu.
SAM 5018 12:7 I1-1 87-4 1790 46 2-6 Fine shale.
12-6 104 825 1650 49 2-6 Shale. Shaped.
12-0 8:8 73:3 1270 44 1-9 Reddish Ss. Erratic.
39 8-0 80:8 785 46 2-4 Ss. Shaped.
11-5 5-1 44-4 675 774-500 r TMSs. pebble.
11-2 4:5 40-2 565 40 1:4 Shale.
7:8 50 64-1 300-7 2:4) Eee: Shaped. Shale.
Signs of a groove
on one face, as
though broken and
rebored.
BARRYDALE.
Univ. Stell. . 5:7 5:0 = 87:7 162 3-1 (2:2
LADISMITH.
MIMKG 503" 27 13:0) lt 85-4 Sie to:eee-O,

BDV. BREEDE VALLEY.

As is to be expected, the long wide valley of the Breede, running
from Tulbagh and Ceres almost to Heidelberg, shows similarities with
the West Langeberg, which forms the northern section of its catchment
area for part of its length.

Pas) 9-0 70-3 1375 BH. 1, Wikia SWoae
tly, Pts} 8-1 63:3 1330 BE au:
ils Lake) 6-6 60-0 800 CH. v, SV. ii.
tv (9:0 7:8 86-7 630) BOT. ii, CH. viii.
v. 10-0 4-3 43-0 430 BOD ive S Vena.
val | 14 4-2 48-3 220 WL. vi, BH. xi, BOT. vy, Ciena

Types i and 11 are obviously the result of divergent copying. An
analogous example in the same shapes and sizes is to be seen in KTZ.
iiandiv. There is a gap in thickness from 5-0 to 6:0 cm.

CERES.
Univ. Stell. 12:8 8 204 sald oa 2
NMP 2585 . 11-0 65 60-0 785 41 aw
9-2 79 85-9 670 46 2-8
F. Malan Coll. 6-4 3h) ons 1527 veel Ratelfontein.

The Bored Stones of South Africa. 95

GouDINI.
SAM Botha . 13-0 8-2 63:0 1385 4:8 3:2
13-0 80 61:5 1350 5:8 2-9
12-3 (8-0) 65-0 1200 4:9 ay
74 (45) 60:8 245 2:5 1:8 Shale. Broken.
VILLIERSDORP.
UCT 43/14 12:3 90 73-2 1500 40 2-6 Tertiary Ss.
Univ. Stell. 11-5 60 52-2 795 4:0 ae
UCT 43/14 109 4:7 43:1 560 46 2-7 Surf quartzite.
96) 94:2) 4377, 390. 32) 1b Se
75 42 56:0 235 35 1:4 TMSs.
RoBeErtson A.
SAM 395 . 13:5 88 65-2 1605 48 3-2 Radial cuts at one
mouth.
LECHASSEUR, ROBERTSON.
WCi43/s .) 7-0 4-1 58:6 200 3:4 41:6 10 feet deep in lime-
stone.
SWELLENDAM Dist.
SAM 179 . 13-0 90 69-2 1520 48 2-5
is) 13-0 880 61-5) 1350 4.2 21
Pie olIc3evn7-O) 161-9) 18954 4:45 2:2
181 . 105 70 667 770 40 22
etal . 88 76 86:3 590 36 2-0
1799 . 100 40 400 400 34 21
369 . 69 4:3. 62:3 20574) 221 7) 1-4

Kruis River, HEIDELBERG.
HHC 1051 . 10-0 4:5 45:0 450 40 2-2 Grinder.

Cape INFANTA.

F.MalanColl. 4-9 4-9 100-0 fe) GRD on Anomalous. May be
a hammerstone.
TMSs.

BOT. Bot River VALLEY.

This little enclosed area appears to have types of its own.

a (12-2 58 47-5 865) BH. iv, CH. vi, CF. ii, CM. vi.
li. 10-4 6-5 62-5 705 BH. vi.
mal, 2 (Pell 7-1 78:3 590) BDV. iv, CH. viii.
iv.) 9:6 4-2 43-5 385 BDV. v, BH. viii.
Vere ec 38 51:3 210) WL. vii, BDV. vi, BH. xiii, CF. vii, CM. xv.

There is a gap in thickness between 4:1 cm. and 5-7 cm. The
thickest specimens measure 7-1 cm. This shows a considerable
relationship with most surrounding areas, and probably represents
the most important components of those nearby areas showing
heights below 7:1 cm.

The midden-filled cave at Kip Kop has been described somewhat
briefly by the writer.*

* Goodwin, Annals S. Afr. Museum, xxiv, 5, 1938.

96 Annals of the South African Museum.

CALEDON.
SAM 786. 2) LOM 6-9) SiG2 Hie) Soount4- 5 a le
100 68 68:0 680 3:8
Kuiie Kop Cave, HERMANUS.
SAM 4776 . 12°29 5:8 47-5" 865: ~6-0 Be Midden associations.
Patterson . 87 j%T4 86:4 550 59 2-2 #£White Ss.
10-4 3:9 375 420 3:6 2-3 Surface quartzite.
SAM 4776 a ee ee 5 Bille} SYN el) Te(s}

Diz Monp, HERMANUS.

L.B.Goldschmidt 7:4 3:8 51:3 210 3-8 1:8 TMS. pebble, W.
bank of lagoon.
Midden. Erratic.

HawstTon.
MMK 1253 a ALOSGs= 95-8) 547 GSO s ao.S Wms Ss.
KLEINMOND.
ASJ 4/35/3 J OA il) 17525) O80) 2) oso aad
88 71 806 550 45 2-6
Betty’s Bay.
Miss Pirie : 95 4:9 52:6 445 3-8 2:0 £Grinder.
HANGKLIP.
SAM 786. A OZ 6:4 we G27 Guo? mime mS
ODA AA Ole oO Mn Somme
COMMENTARY.

Once again there is an absence of literature on the bored stone in
this region. Some of the South African Museum specimens marked
as from Swellendam may have come from north of the Langeberg,
but those marked with farm names in addition have been correctly
allocated here. |

A more important difficulty arises with the medium oblate (12-5 x 3-2
cm.) group from West Langeberg area. This group depends upon
material submitted by one man, J. M. Bain, who often presented
accumulations of unmarked specimens to the South African Museum.
These were marked “Bain, ’05” or “‘ Bain ’08,”’ and are discarded as
valueless. Other examples have a district name written on them by
museum authorities. Where there is positive evidence of error these
have been discarded. There are, for instance, a specimen marked
“Montagu,” but made from Knysna district material, and a specimen
marked “‘Cape Flats”’ over Bain’s original allocation to Storms River,
Tzitzikamma. These are both rejected as unsuitable, though the
Storms River example will be made use of in a paper discussing
dagga-pipes. Those typological groups that bear Bain’s name are
thus open to suspicion, though where the evidence is supported by
other examples, they may be regarded as completely acceptable.
This difficulty arises once again in the Cape Region, and should be
noted.

The Bored Stones of South Africa. 97

I have drawn attention to the gap which occurs regularly in the
graphical distribution in different areas. I do not know what it
means, but I presume that it marks a development in the technique
of boring stone. It may also help us in our knowledge of migrations.
This phenomenon continues in the Cape Region.

XI. CAPE REGION,

This region is enclosed by the Hottentots Holland range and by
the Drakenstein to the east, by the Olifants River range to the north-
east, and probably by the Piquetberg range to the north. It thus
approaches the Bokkeveld area to the north, and the Winterveld
area to the north-east, but is cut off from both by the archaeologically
unknown valley of the Doorn River. On the east it joins the Agulhas
Region, or is cut off from it by narrow mountain ranges. The Cape
Peninsula is treated as a sub-region.

We are now well within the area covered by Péringuey in his “Stone
Ages of South Africa,” * and the reader may turn to that volume for
the chapter on the bored stone.

The materials used in this whole region consist of shales from the
Malmesbury Beds, and sandstone from the Table Mountain series,
and in a few instances of granites from intercalated lavas in the
Malmesbury Beds. These materials seem mainly to have been used
in their beach-pebble form, and this must have been a factor in the
choice of size and shape. The slates and shales naturally tend to
achieve flatter pebble forms than the harder and less schistose sand-
stones and granites. The latter rock is rare or very coarsely weathered
in beach or river gravels.

In all these areas there is a marked gap either between 7-0 and 8-0
em. in thickness, or between 9-0 cm. and 10-0 cm. This should be
watched. At times this is replaced by a complete absence of types
thicker than 7-0 or 9-0 cm., again suggesting that there is some
importance in this gap.

CH. Carr HINTERLAND.

This rather confined area consists of the western slopes of the
Hottentots Holland range and the areas immediately adjacent. One
small sub-area, typified by a single type of small bored stone, is included
here as type xi. To the west it merges with the Cape Peninsula,
and is only separated from it by the low-lying sandy expanse, broken

* Annals 8. Afr. Mus., viii, 1912.
VOL. XXXVII, PART I. 7

98 Annals of the South African Museum.

here and there by granitic, Table Mountain Sandstone, or surface
quartzite outcrops, that is known to-day generally as the Cape Flats.
This stretch of country irked the early settlers and proved difficult
to cross by ox-waggon or on foot. The 30-mile journey by perfect
modern roads to-day does not in any way represent the two days’
heavy walking through dune sand that originally provided a marked
barrier. It is obvious that the firmer seasand on either side of the
peninsula would have provided the only easy ways of access to the
peninsula proper.

1 (16-7 2:9 59-3 2760) Cf. Hout Bay. ? MaNtatisi.
ii. 12-8 10-3 80-4 1690 WL. i.
iii. 12-2 7-4 60-6 1010 GK. i, BV. ii, CM. iii.
yg) LI) Ty) 71:8 965 BH. iii, CF. i.
ven EO 6-8 61-8 825 BDV. i, CM. vii, BOT. ii.
va (dd-9 5-7 47-9 810) BH. iv, BOT. i, CF. ii, CM. vi.
vil. (12-7 4-3 33°8 695) Anomalous.
vain, 8)33 7:5 80-6 650 BDV. iv, BOT. iii.
xy | O° 5:1 52-6 480
xe Oat 5:8 63-7 480 GK. iv, CM. xi.
Rin Ss 6-0 73-2 405 GK. v, CM. xii.
XAT Sac 4-3 51:8 285 BE aa:
xiii. (4:0 2-4 60-0 38)

Type xii is local, and occurs in Ida’s Valley (Stellenbosch), at
Vlottenberg and Somerset Strand only. It is absent at Kogel Bay
and from the far more representative series from Gordon’s Bay.
Types iv and v are a pair, and type vill may eventually yield to
subdivision. There is a marked gap between types ii and iil.

Major Jardine’s collection, marked as from Gordon’s Bay, is partly
from his farm, “‘Applegarth,” situated immediately north of Sir
Lowry’s Pass on the western slope of the Hottentots Holland. It is
thus within four or five miles of Gordon’s Bay, but does not necessarily
consist of midden types. All other materials from Gordon’s Bay are
from midden sites which show no formal implements. Pottery
occurs on the recent surfaces only and is absent from most middens.
My thanks are due to Major Jardine for his courtesy in permitting
me to make use of his material.

There is a complete absence of types between 8-6 and 9-6 cm. in
thickness.

HELDERBERG.
UCT 44/1 . 4:0 2-4 60-0 38 1:9 1-1 Extremely small.
STELLENBOSCH.
AMG 1549 . 16-7 9:9 593 2760 5-3 2-8 Compare Hout Bay.
UCT 27/10: . 12:0) 10-0)" 91-7) 1585 3:3 1:6.) Brobably jmmieraredt

bored.
SAM 3428 . 11:0 -7:6 69-0 920 4:0 te TMSs. pebble.

~The Bored Stones of South Africa. 99

STELLENBOSCH—continued.

SAM 4959 . 10-7 6:7 62-6 770 41 #=$1:9 Gustrouw.
ats) 9-2 5-6 60-8 A> 3-1 2:8
Wem zr ls. . 8-2 6:28) 75:6 Alay a:9) 21 TMSs. pebble.
43/20 9-5 5-1 653-4 460 3:8 2:0 TMS.
Ipa’s VALLEY.
SAM — : 8-1 AE OO an etOL i arO. 1420
VLOTTENBERG.
Wernatis . 12:3 72 585 1090 4:2 2-1 Used as hammer.
Univ. Stell. . 8-4. 4-5 53-6 S20 osc
SoMERSET STRAND A.
SAM — : 8-2 4:0 48:8 QiOw mute wlll,
Gorpon’s Bay.
Jardine Coll.. 13-4 OG eile (725. 73:0 ik
12-4 78 62:9 1200 4-1 M72
12-3 G2.) 58-5" (1090) > 3°8) 3-2
SAM ws '. 11:8 1-4) 62-7" ~ 1020 5-4 ;
Jardine Coll.. 11:1 TSS) pil c OTD: 3S 22
10-8 8-2 75-9 960 2-9 ae
12-7 5:8 45-0 925 (4:5 x 2:5) Oval, tubular hole.
10:8 76 70-4 885 41 2-1
11-1 70 63-1 860 41 2-0
Univ. Stell. . 11-0 fal 64-5 860 4:3 3:0
Jardine Coll.. 11-4 6:6 57-9 Shon to Sh eee
Cmrasi21 <li 6:7 60-0 825) ) aay 2%. Grinder.) “IMS:
Jardine Coll... 11-9 5:1) 47-9 810 5:1 2:5
10°5 65 61:9 alin Wiese D4)
Univ. Stell. . (9:1) 8-6 (94-5) 710 4:0 2-0 Coarse Ss.
AMG 1646 . 12:7 4:3 33:8 695 50 2:3
Jardine Coll. . 9-8 70 71:4 G70) 9-2-9) 1-9
9-2 76 82-6 645 70 1-1
9-2 7-5 ©6815 635 3°5 itt
UCT — 4: 70 74-4 620 4:2 oe
Jardine Coll.. 10-0 A PAL) 520) 3-8). 2-5
Univ. Stell. . (9-3) 6:0 64:5 520° 3:8 3-5
8:3 6:6 80-0 Aa 1325 Be
Jardine Coll. . 8:9 5-5 61-8 ASD en ooo
SAM 1775 8-4 6:0 71-4 425 AY fe Both holes started.
8-4 5:4 64:3 380 39 3-6
Kocet Bay, MIppDENS.
WiWAS2545 2 °13-0 10:3 78:5 1740 44 (1-3 2:-3ke. Surf quartz.
2500 7.2 9-0 5:9 65-6 480 40 2-3 Calcareous Ss.

CAPE PENINSULA SUB-REGION.

Just as it was found possible to suggest the presence of various
cultural groups in the Knysna-Tzitzikamma and Belvedere-Heidelberg
areas, so we find that the Cape Peninsula provided a focus of attraction
for various peoples. Here too the country was at one time heavily
wooded, as such names as Hout Bay, Rondedoornbosjen, etc., would
imply, and we do not get a true picture of the prehistoric scene among
to-days’ oak and pine forests. From Constantia Nek to Hout Bay was

100 Annals of the South African Museum.

a single forest of mixed type. The same was true of Tamboer’s Kloof
above Cape Town. To reach Simonstown, van der Stel was forced
to hack his way through thick natural scrub and bush, forming a
thick talus growth typical of both sides of the Peninsula. The Fish
Hoek valley, cutting through the mountains to Noordhoek, was a
mass of low trees, 6 feet in height, forming a natural warren of
sheltered paths or run-ways, 4 feet high.

In contrast to this picture of wooded valleys and mountain slopes,
the Cape Flats appear to have been more bare than they are to-day.
Peaty deposits show that this area was dotted with undrained fresh-
water vleis or springs, often almost at sea-level. These were bordered.
by water-loving reeds and grass, giving way to the ubiquitous white
sand.

The Peninsula therefore provided a wide ecological range for man,
and must have attracted a variety of cultural groups. Some of these
appear to have clung to the eastern slopes of the mountains and to
have made use of the vleis and springs of the Cape Flats proper.
Others passed over the first range of mountains to the Fish Hoek-
Noordhoek valley, while only a few crossed to Simonstown and the
treeless area about Cape Point.

The Peninsula is here dealt with in two sections: the Cape Flats.
and the Cape Middens. These two would seem to be partly cultural
and partly distributional groups. If the two sections are plotted on a
graph paper as one it is obvious that nothing can be made of local
types. If, however, we group together known midden material and
specimens from the mountain slopes immediately overlooking the
coast, and contrast this with the distribution of Cape Flats types, a.
very clear difference appears, and only slight adjustment of specimens.
from uncertain sites becomes necessary.

The attractions of these two environments with their different.
modes of life are obvious. In the midden cultures we have an environ-
ment with abundant supplies of lime, fish foods, iodine, salt and kelps,
all with a high nutritive content, sufficient in itself to account for the
regularly taller stature of the coastal midden peoples leading a
strandloper mode of life. On the Flats, in contrast to this, we have
an environment with a natural fauna of hippo and ungulate types, and
abundant wild water-uintjes,* which must have supplied a source of
essential vegetable foods at that time, just as 1t does to a section of the
Cape coloured community to-day.

In the Preliminary Survey the Cape Middens material was referred

* Avponogeton distachyon.

The Bored Stones of South Africa. 101

to as the Cape Mountains area. I have changed this as the original
term was misleading; there are few examples known from heights
greater than 500 feet above sea-level, while those that occur on lower
mountain slopes are related in almost all instances to midden deposits.
The sites not relatable to midden deposits are added at the end of the
main list.

CF. Care Farts.

The term Cape Flats was fairly generally applied fifty years ago to
the low-lying sanddune country immediately to the east of Claremont,
Wynberg, and Plumstead. To-day the term has a wider meaning, but
we can generally refer all the material so marked to that older and
more confined stretch. One specimen, presented by J. M. Bain, is
marked “Cape Flats’’, but is almost certainly an erratic related to
types from Kloof, etc., in the Middens series. It is therefore listed
under Cape Middens. An exactly analogous case is a specimen
marked SAM 402, also presented to the South African Museum as
being from the Cape Flats.

The normal routes from the mainland of South Africa to the
Peninsula were along the Table Bay or the False Bay beaches of hard,
wet sand rather than across the Flats. This would account for the
present area being isolated from the Cape Middens area, which is
connected with sites on the Table Bay side, and with the Cape
Hinterland.

The natural False Bay coastal road leaves the heavy sand at
Somerset Strand, thus accounting for the Somerset Strand B material
that resembles Cape Flats rather than MHinterland types. The
following types should therefore prove to be typical of both the Cape
Flats and the False Bay coast, and seem to represent migrations

between BH., WL., BOT. on the one hand, and CF. on the other.

ie 1S) 74 67:8 865 CH iv, Vv.
ries ie Ta 5-7 50-0 740 CM. vii.
ii. (11-6 31 26-7 415) WL. Vv.
hye ey 5:5 71-4 325 BH. x.
v. 68 4-4 64:8 205 BDV. vi.
Vise vic 3:2 45-9 175 BOT. v, WL. vii.

There appear to be no specimens thicker than 7:5 cm., while diameter
varies from 6-5 to 12-0 cm.

SOMERSET STRAND B.

UCT 43/22 . 7-9 Dehepey lacm 355
AAI 6°7. 4-6 68-7 207

Diep RIVER.
SAM — - 11:0 74 67:2 895 4:2” 35

2-9

oa
SS)

102 Annals of the South African Museum.

CAPE FLATS.

SAM Bain . 12:0 6:5 54-2 93D Aad aso
692 be dled 70 60-4 925 46 4-0
10-8 75 69-4 875 40 2:8
Bain . 11:8 60 50°8 83D) oo co

at « 11-0 5:6 50-9 680 4:5 3:2 Non-radial cuts at

mouth.
a Said Wf Lots 2-8 23-8 (390) 3:8 3:0
i Ae rfc) 4:6 65:7 225 oO 2.2
ue 8 7:0 2:6 37-1 Ath a2 8 acl)
PHILIPPI.

Drennan Coll. 6-7 4-1 61:0 184 3:8 #41:7

VicToRIA HOSPITAL.

ASJ —. LO (7:5) (68:3)in 90 Den ka:or mers
orl 5-5 71-4 330 =63°9' 2-0
75 3:5 46-7 HOE Billy. iey/
UCT 30/21 7:3 3-5 47-9 187 3:2 2:0 One face ground.
PINELANDS.
Jager Coll. . 10-0 74 74-0 140 392 eT iss

EERSTE RIVIER.
F. Malan Coll. 11:3 5:8 51-3 740 4-7 3-1

CONSTANTIA VALLEY.
TMP 2334 . 11-4 3:3 28-9 (429) 43 2:3

CM. Carre MIDDENS.

It is obvious that the Peninsula as a whole was a pocket, an
ethnological marginal area, into which any passing culture might
throw off groups which would remain isolated in their new home. It
is to be expected therefore that we will find a multiplicity of types.
Some of these have gone farther afield than others, only one type of
stone reached Cape Point, while only three types occur at Simonstown.
South of the Fish Hoek-Noordhoek valley the marginal effect is to
localise and isolate types, while north of that line its effect is to catch
and hold a multiplicity of passing types from a variety of sources.

But that alone would not account for all the types. Some may
prove to be pairs, due to divergent copying. For instance, types 1
and iv seem to be related.

It is to be noted that the first of the Groote Schuur specimens does
not fit comfortably into this distribution, nor into the Cape Flats
distributions. It is, however, added here.

i. (16-0 10-5 65:7 2690) Hout Bay only. CH.i.
li. (12-9 8:7 67-4 1445)
iii, 12-4 7:4 60-0 1140 BDV. ii, CH. iii.
iv. 11-4 8:5 74:5 1105 ZW. iii, BH. iii.
v. (13:7 5-1 37°3 955)
vi. 12-2 5:7 46-7 850 BOT. i, CH. vi.

The Bored Stones of South Africa.

vii. 10:5 7:0 66-7
viii. 9-1 8:4 92-3
ix. 10-4 5:8 55°8
xen Le? 4:8 42-8
xi. 9:0 5-4 60:0
xii. (8:3 6-2 75:0 °
xiii. (7:0 71 101-4
xiv. 9:0 3°5 39-1
Sa. Wes) 3:6 46:2
xvi. (10-2 2-0 19-6

Shows a close relationship to the Cape Hinterland area.

770
695
625
600
435
430)
350)
285
220
208)

BDV. iii, CH. v.
CH. viii.

WL. iv.
CH. ix.
CH. xi.

BOT. iv, SV. vi.
BOT. v, WL. vii, CH. xii.

103

There is

a distinct gap in thicknesses, from 9-0 cm. to 10:5 cm., and only one
specimen occurs in this latter thickness, the Hout Bay example. The
connection of this area, through Darling, Tygerberg, and Blaauwberg,
with the natural route along the hard sands of Table Bay is obvious.
A similar hard route along the beach links the Hinterland and the

Middens area.

DARLING.
MMK 282/a . 10:8 8-7
3-7 © 5:3
115 3655-0
TYGERBERG.
SAM 5001 - 93 3:3
BLAAUWBERG.
SAM ’09 . 2 t2-0.8 18-2

GREEN POINT.
SAM Péringuey 11:6 8-4

10:2 2-0
STRAND Sr.
UCT 43/18 LON 5:5
PARLIAMENT FOUNDATIONS.
SAM 0/11 J Belly = taye7/
KLOorF.
SAM 5002 4 78 3:8
Hout Bay.
UCT 27/18 . 16:0 10:5
119 7-4
SAM 4750 2 el 225 bet
4777 : 92 81
WUA 1182 . 106 7-2
SAM 4777 p 9-1 5:6
UCT 27/18 ; 83 6:3
SAM 4777 83 62
70 71
UCT 27/18 ‘ 9-3 (4:0

80-5
38-0
43-5

35°5

68-3

72-4
19-6

54:5

47-1

1015
995
660

285

1180

1130
(208)

560

835

231

2690

1045
895
685
585
465
435
430
350
345

4-2

4-2

4-]

4-9

>
=

—

CO tim the CO Tne ag Oo) CANE ee
WEE ORWAHS ©

2-2
2-4
27

1-0
3-1
2-1
2-3
2-9
3-1
18

Coarse Ss.

Ss.

Irregular pebble.

3 feet 4 inches below

surface.
Slate.

30 feet deep in

foundations.

Malmesbury
Metal bored.

shale.

104

NooRDHOEK.

UCT 30/13
SAM 1287
UCT 30/13

KOMMETJE.
SAM 3831

Univ. Stell.

CAPE PoINnt.
SAM Martin

SIMONSTOWN.
SAM Martin

Fish Hoek.
SAM 1287

Jager Coll.
SAM —

Drennan Coll. .
J. Gordon Coll..

SAM (Peers)
SAM— .

PEERS’ CAVE.
SAM (Peers)

CLOVELLY.

Drennan Coll. .

St. JAMES.
UCT 30/28

MUIZzENBERG.
UCT 43/23

GROOTE SCHUUR.

SAM 4362
TMP 2354

(M. C. Burkitt .

ee
HOH SWyly
aontowocooeo

—
i
bo

10-9

13-0

9-0

12-9
10-8
9-3

8-9
6-0
4-4

IU
ko bo 00

eo o9 69
= QO 00

PAD AIH
oH OR

OO tres is Ms Se SY
DRAM OOD

ot
a

8-4

7-1

5:0

Non-M1IppEN SITEs.

Si
o-4
5:3

104-7
65-2
41-9

61-7
47-3
40-8
56-3
40-9
51-7
40-5
48-0

45-5

77-1

54:6

55-6

67-4
90-0
57:0

645
510
485

825
685
365

310
250
214

760
730
645
480
600

1065
805
705
610
545
340
330
205

640

1000

1200

405

1445
640
460

G2 Oo OD Oo G2 TT
eae Kens io) mean an)

H
wo

38

4-8

3°8

5:3
4-5

5-2

Ou: Oe |

w. boron tony,
Ore em oO. noe:

bo RD DOWD LD NLT S
SONAownrFr~

—_
CO

2:2

2:9

1-9

2:3
4-9
2:5)

Annals of the South African Museum.

Malmesbury shale.

Shale.

Two faces ground.
0:96 kg.
0:77 ke.

Broken, rebored at
right angles, then
rebored again with
a larger rimer,
leaving part of the
narrower bore still
visible down the
side of the new.

Irregular. Grinder.

Shale.
TMS.
Slate.

Ss. Midden deposit.

Symmetrical bore.

TMSs.

4 feet below surface.
UCT.

3 feet deep. Lion’s
den. Zoo.

Estimated from
illustration. UCT
building site.

The Bored Stones of South Africa. 105

ConsTANTIA NEK.
Jager Coll. 2b) 7-6) «60:8 90) bb) 3-7

WeETTON FLATS.
G. A. L. Green 13-7 48 35:0 900 4:3 £41-7 Ss.

Coll.
5 a 8-7 34 25:6 255 42 .. ##Malmesbury. Ss.
Care Fats.
SAM 402. a IEO. 4-8) 9 43-6 580 4:9 2:8
Bain - 95 3:6 37:3 325 40 3:5 Almost cylindrical
bore.
CoMMENTARY.

Of these types only one (the slate example from Green Point),
type xvi, does not appear to have crossed the mountains. It may or
may not be related to type i from the Cape Flats, but it is certainly
a product of the material used. In contrast types vii, 1x, and xi are
the only ones that reached as far as Simonstown, and only type xiv
managed to traverse the length of the Peninsula to the Cape of Good
Hope. This is the type of distribution that would be expected on a
peninsula. Types i, x, xii, and xiii seem to be confined to the Fish
Hoek-Noordhoek valley and to Hout Bay. This at first seems
surprising, as Chapman’s Peak stands across the present coastal road
between Hout Bay and Noordhoek as an almost sheer cliff, 1500 feet
high. In practice the footway passing behind the peak through the
Silvermine Valley is an easy walk and also provides access to the
Tokai portion of the Constantia valley, from where I have no specimens.
This footpath was in fact the only wagon-road to the Cape of Good
Hope two centuries ago. Type vii is confined to this same area and
to Simonstown, so far as our evidence goes.

This mountainous area shows the possibilities which may be brought
out by careful use of the distributional method. In this last area
alone, five possible cultural waves may be recognized, the least
adventurous reaching only to the Constantia Valley and the northern
slopes of the mountains. The Hout Bay-Fish Hoek group provides a
second wave; a third seems to have been more general in its distribu-
tion; the fourth reached the enclosed bay at Simonstown; while the
most adventurous passed even farther south to Cape Point. We have
not an iota of evidence to give us any clue as to the order in which
these presumed spreads occurred, but the evidence of distribution
does supply us with an important new source of material for future
research workers. With increasing evidence from various areas,
museum and field workers now have a new methodological tool.

We now return to the Cape Region proper.

106 Annals of the South African Museum.

SV. SANDVELD.

This covers the inland stretch from Wellington to Clanwilliam.
Some of the groups are wide and therefore unsatisfactory, but more
material should yield a better approximation. It is a fertile, low-
lying downs country, and seems never to have consisted of anything
except open grassland, with a scrub flora in the valleys. In late
Quaternary times the fauna seems to have included such plains-
loving forms as Bubalus bain and Equus capensis, pointing to an
environment much the same as that encountered to-day.

i. 12-0 9-2 76-7 1325 WL. ii, BDV. i.

ly | Ale 6-6 56-4 905 CF. ii, BDV. iii.

ii. (10-4 8-0 77-0 865) WL. iii, BH. iii.

iv. 11-4 3:9 34-2 510 CF. iii.

We) ey) 57 64-0 450 CH. ix, CM. xi.

Wile. BED 3°8 42-2 310 BOT. iv, CM. xiv, BDV. v.
vil. (8-4 2-3 27-4 160) BH. xii.
vil. 5:8 3-1 51-7 105 BH. xiv.

Shows considerable affinity with Breede River and West Langeberg.
No specimens thicker than 9-5 cm. thick. i

CLANWILLIAM A.

SAM Bain Aees220 NS -40 Go: OuloZ oul k4 hen acG
12°09 -28) 16349 325) 4-4 al
106 95 89-6 1070 42 2-6
11:9 70 58:8 990 4-1 2:1
CLANWILLIAM B.
SAM 1120 - 100 83 £830 8380 45 3:8
12-0 4:5 37-5 650 4-2 21
8:5) 15-4) 6325) S90 19 y3:08 2-2

OLIFANTS RIVER.

SAM 420. . 10: 5 55:0 550 £4:°5 es
9 0 66:7 485 35 2-0
SANDKRAAL.
F. Malan Coll. . 10:7 7:6 71:0 870 4-6 3:8 Reddish sandy shale.
Ochre on one side.

Signs of sharpen-

ing.

ExLanps Kioor, TuLBAGH.
NMP 2578 » Il? 64 (57-1 805 4:0 2:2
2584 5 10:34 13-00 29-1 5 320) 2-8) 21s

MooRREESBURG.

SAM Bain 12:0 64 53:3 920 40 2-2
710. 11-2 38 33:9 475 35 3:0
Uke), eS) kos} ais) | pil Ne
8:2 3:3 40:2 222 2: 1:9

Batn’s Kioor.
SAM — . - 84 (2:2) (26-3) 155 (2:7) 1:9 Fragment.

GoEDHOOP, WELLINGTON.
F. Malan Coll... 93 59 63:4 510 43 £26

The Bored Stones of South Africa. 107

WELLINGTON.
SAM 3425 - ING 65 560 875 58 cs
90 3:8 422 310 30 1-9
64° 3:0 46:8 9123) “32h 2-1
OLYVENBOSCH, WELLINGTON.
F. Malan . . 12:2 67 54:1 1000 5-2 2:5 Malmesbury shale.
NABYGELEGEN, WELLINGTON.
HevalaniColl .. 5:9, 3:5 59:3 122) 2-7, 1-5
50 2:3 46:0 59 30 1-4
GROENVLEI, WELLINGTON.
HeMalamColls . 12:2 43 35:2 640 3:2 2:1 =%‘TMSs.
99 44 444 480 47 3-4 TMSs.

SWL. SwARTLAND.

Apart from a specimen with Wilton associations, a single type
appears to cover the coastward stretch between Saldanha Bay and
Parow. Probably the distribution will be found to approximate to a
coastal midden people. Its importance (like that of the Grootrivier
type) lies in its localisation and in the uniformity of size.

i. 14:3 7-4 51-0 1520 No parallels known.
li. (10-3 6:8 66-0 7120) (CM vu, BOUT i;:CH. v.

Type 1 shows no relationship with any surrounding types, but
type li suggests relationship with the Cape Middens material. The
nearest affinity to type i is in the Southern Transkei, with no inter-
vening links.

GEELBEK, SALDANHA.
L. B. Goldschmidt 13:8 7:2 52:2 1370 3:6 2:4 #=‘TMS.

KLIPVLEI, SALDANHA.

P. Bateman . 10:3 68 66:0 720 3-8 2:5 ‘TMS. Shalecontact.
Wilton site.

MALMESBURY.

F. Malan Coll. . 14:7 7:0 47:6 1515 3-9 2:5 Malmesbury shale.

YZERFONTEIN POINT.

SAM 4997 yA Se e7-O OS: 7) ep looO! retin) 222
KuipKop, PAarow.
SAM 5006 - 143 47:6 53:1 1555 4-2 3-8 Shale. Cylindrical

bore.

While nothing is known of the Wupperthal area, Sir Langham Dale,*
writing in 1870, says: “From Wupperthal I hear that oval perforated
stones were used by the old Hottentot warriors as weapons of war,
a stick of hard wood was thrust into the hole. For digging up roots
the stone was grasped in the hand, the end of the stick being sharpened
for picking up the ground.”’

* A (Delta) “Stone Implements in South Africa,” Cape Monthly Mag., New
Series, i, 1870, pp. 236-239, illustrated.

108 Annals of the South African Museum.

CONCLUSIONS.
These two parts have been designed to bring out the following
premises :—
I. Identity of intention. It was intended to produce a limited
number of types of bored stone.

II. There is no reason to believe that any one cultural group used
more than two or three of these types in any one area and
at one period.

III. Dissimilarity is due to inexact methods of achieving identity.
The approximation is generally sufficient to show a clear
graphical scatter (except in the focal areas), but seldom
sufficient to yield exact replication.

IV. This led to an eventual multiplication of types (most marked
in the focal areas), in some instances approximating to
previous types.

V. Only certain of these new developed types occur in new migra-
tions, when the whole process above is repeated.

VI. The numbers of anomalous specimens are small enough in
most areas to be negligible, and to permit us to presume that
they were rare individualist products.

We can conclude from these six premises that it should be possible
to trace migrational routes from one focal area to another by means of
a study of types. It is not possible, however (by reasons given as
premises iil, iv, and v), to study the distribution of any one type
throughout the entire area considered.

Not one of these premises would seem to apply to the Transvaal
Bantu bored stones.

KEY AND INDEX TO AREAS (CIRCULAR SPECIMENS).

PAGE PAGE
BDV. Breede Valley. 94 NG. Nieuveld-Goup. 70
BH. Belvedere-Heidelberg. 89 NH. Natal Highlands. 31
BMR. Bushmans River. 78 NL. Natal Lowlands. 33
BOT. Bot River. 95 NQL. Namaqualand. 66
BS. Bloemfontein-Sepani. 39 NTK. Northern Transkei. 75
BV. Bokkeveld. 64 PE. Port Elizabeth. 79
CF. Cape Flats. 101 SCV. Schoonspruit- Vaal. 36
CFS. Central Free State. 52 SEFS. South-Eastern Free State. 54
CH. Cape Hinterland. He SR. Sundays River. 82
CM. Cape Middens. 102 ST. Southern Transvaal. 29
CT. Central Transvaal. 26 STK. Southern Transkei. 75
EL. East London. 77 SV. Sandveld. 106
ET. Eastern Transvaal. 25 SWB. Swartberg. 89
GK. Gouritz-Karoo. 92 SWL. Swartland. 107
GR. Grootrivier. 84 TV. Trekveld. 69
KA. KB. Koffiefontein. 48 TZC. Tzitzikamma Caves. 84
KTZ. Knysna-Tzitzikamma. 85 UH. Upington-Hartbees. 62
KV. Kaaienveld. 65 UTY. Upper Thornveld. 73
LTV. Lower Thornveld. 74 VH. Vaal-Harts. 4]
LV. Lower Vaal. 44 VMR. Vet-Marico Rivers. 37
MF. Modder-Fauresmith. 53 WL. West Langeberg. 93
MG. Molopo-Groenwater. 61 WV. Winterveld. 67
MV. Middle Vaal. 42 ZMM. Zuney-Maitland Mines. 81

The Bored Stones of South Africa.

"T avi

Pursue \ Ny Pure peysdooH of e uyauoy | 86
ho ( ®Is7 7 ° RULNSHYD _eudooy
SUCINYSURIACT iciope doy
—~ prjsuoory NS S. eyo uiaold j
adj ® poomssuly
9
-S,
7; aaa
> to Sony as * ayauay aziamyag
U0}.19pueqS 1% | PS
Lo- y ‘ -LG
diopsylo[ 1 © ZinqAy
=f Y [eo shitd 2 ¥2 A
=, @
: Sioqepiayy & ~S SSuIsUIeI0A W001} SJaYII0g
(UdM|NZA © se OSUIAP SIU Scephug AN
2 S“oissisyd O41. leyog: S pores
~ Oe, <
= sSuridg “'
ISnsHEL Uojiwuien 0 #Nqsouueyor
s ~, K : = ye qunqiay ho ‘ © Zunquayyory ie
ve e eek SE 2d ol, qnsds}ssoyHUuoIg ar o Ua} GO; Ng © 2uIHaeW :
uo}soqieg 2, 2 uy fd Ko usHatqvyaysson® Aa eE ES ea “'sinquojsny if SaneaT maar odo lop
e SURIAIPINS: < SG lk Ses wooqsapuoyy® ee?
U a[!P 0015 Sen om Pow Es
ae cn Sunes uoiqayy ?
Se Rye Wg sila
Fees dowued WLOUOTUA » o WOepeqIoM oe
i © SUJEqUiTe AA
NOISaa “~WVASNVUL® us
Arran
ISNISIO9ITIOg 7G
Fo-
eyedy MON 7] “ayy ag Aint
Y egeiay ee :
= Uo
he poke WZ =
wee es = %, :
swalilibip] oe Sahiyy 84 - 2 %y a"
£3 LS Dugg] SNVALNOS St is ~€
° | | ; | ! | ! I
8G LG 9G GS

Ie of 6a
— 2 SS 5 - = so ere oe as Sea

Annals of the South African Museum.

110

Se =e % AV
(ce :
/NOLOSaY LSV
wl 3
‘y nqnaunzmy) aoe
= xb A HaOD wy Z

3 é ww: SWaOLS: *h % mie Le

: WG ean Melson OBL os aoe" PouaE °
ies on eS estes ee ss *diopsigyaing N -1¢

qo yeajeW 5 ae i, Zinqsajog ®

eIION TEMITY

rd
x

oO
A}
e ongieg 7

ONW7LSYIHL

- auapliey -0€
0 ueqing ZOYHON'To .
i) yemed °
DS ee e Uslusomeg
3 °: e aquery
> pee ES juoWpog
TUeMOs|e_ © ais yaAnayMnerlg °
© * HIOK ° BM l cere Tra} UOJatosy *
7b “., ayepsiouug ion é
6z- <b waa AA &: Fe ees \ wisjubyuroora® snewwg | 6%
age ree Ny a 41n09}Sq a Ss Zingshoold °
sue I oe i= :
a ANT ho be snc omNAgiog «% a 70P Poy Ula} uOJssapuexely
ra ais : IRS a _e euingeying Sots2dan f) joludg
“5 : EN 3 : aoe sinquime Y taj 189 An}
= 4, a has eke Joysog o | e
= : oypuesg ° ANS uoyospur yey UOMOATY
: ? beeen A spur[ Man
tee Le Q. i N “86
4 Up ° U9} UOJ
By wh 1S, pejsdooy 2
i by ° suesysuelneg Te Srernoon
: es o Na
: Be g pejsmooly Ofte
E ’ Oo pooms3ury
S “ ee ®ayouoy JazIoMyIS
= i. Se uojsepurys eel :
Le-| ae ne Za : LG
| I | I ' | 1
I& O& 66 86 LG 96 GS

111

The Bored Stones of South Africa.

"€ 4VN

[Ss J © sonouey / UOAITUIED © ayserjseng ° ay (ieee a 'G
7 NOIS
= hover Vi Ty
Bingson \_ [+ PUSHEAL wea q N
© ailjioog Ee 1o[APUCIG © b
JOJASTeLIPUIO t wep2epe arfjoodyoqSutadg \ See > dypyapuoy
e UMORSdIIG <u { WIOOpjooly Wnquaree y nS PAS ¢ diopsomog °
O&- Dp SS a Se SS . ={)}S
Z yoquesyeig os ¢
Ps
umojadoyH “es vySalid e
& Oo
wupduy © Lae % daryog ¢
uedsoodopn Mm e - CIpl0dudg °
uapuonagjang °dooH S,310H0IN\Y % pueyueutsng \<
St|IN0g «, Os svZ00y 0} sep aren jdoyutas °
66- * uyoju0}S|eZe Hl, * Seynen’ 10019 -63
Syupsumyss = deze Rss “oAty ebueiO
O * 32°, 4, os
JIOOdIo0y : ai e © fy
£3 5 {7
Cdoyseeqiom~ ae Zrocoauian
df SPUL[MAN Wee TIOM ae rN
a Finqseunysod “ae h
craps [/ 3a 83
: 68: VIA
oy & ?:
ueunin : ; cs LSAM HLNOS
x to s
: o
‘awe ©
Le- oy “LE
© BInqhiA
| I !
61 8T LI

————O

Annals of the South African Museum.

112

‘b aVIW

he
c) te
x % S, >
7, 49, “Op xy
ey % bas Mn, KA Gs
4s. y We, 0, ey ‘0
NOIOsY LSVA-HLNOS Peat a ey Y.
ayloay adeg “s y = =
ve- ypeqeziig 110d \e PB ie! \ood 2a +
: 2 Pung 4suNyyeO
Y SAepygn S : Aaquey (ae (ue cite WY
y e Y mie Sean, SS
Y fevryapieed Aaunz, oe he Dinca eo cooeD
%p e poe , ” ‘ey. ~ asiuisasnens pei
y v oS}e[Jpurs 8 Mba ate ay
“opto e1qe}00L Ohi bp Pe sinus
“one g Manis soepeoity Cia ee Reps UG cioenoiiny
dy, 70 © UMOJSUIe yer) © monioyng ~ ILA
UMOIQ, 110, eM Lennar e tet Garten (ae
: . : Hive
tag= ¢- YON 299d © 180d S,ey0g ajiauasuer ° JON
7%, gUMOL SWRI Suryy © Wosnvag (OT :
esi wog ; z
ASpoaypIUUUOD ge tCi ig e omeces woapsa
Kayren AddeH Sine & set a vey prqy
sumMOog jLeoyyeQ "> iaallces
f fi spur peey
. Ure 440 PEIseAIeL, : :
ce 0, Ms » UMO}sUIEND “as ° dingskeLiny
%, 4 NG
oe Aaiy uayD AKoujoH We
wyiigs eynX ae J 3
x oh, 4 MOL: anqlopPlial rece
; es o SNQOUL ost
L See Sete epee Binqsukas-py ad ranns: R
nyuyurqeas, i E Beg ag et ee MoodMneeN ( soaoucy
L&= A) = - é

{ ] ] ] | |
0€ 6G 86 LG 9Z GG PG &@

—9§
=rS
—€
—GE
-T€
|
GG

The Bored Stones of

Olifants R.

Moorreesburg e
zs
ra ek
Ma =

c2)

Swellendam
% Caledon

on te

VOL. XXXVIT, PARTOL.

Sef
Pa aa

EeEaenaite Barrydale ! ene yee °

South Africa. 113

Brandvlei Poortjie °

Eliasiaagie oe Carnarvon

a -31

LO <n GaNS
3 /
a w, esata ae oe | —-32
cleowilliam S 2, ° i » Beaufort West a as
4,0, eta : ee
2 "4, MTS, = seest ~ i

‘e, Prince ee

erat

Sw there 2 5 a Oe |
Voortttsis eg

© Oudtshoorn

3 Anysberg( 2...
p(s Ye OUTEN
Ladismith

“Riversdale e PGCE paemceeo™ oy —34
Heidelberg. albertinian. goa F eS
Tay a ] 0 : 3 me e xe
Sager ee IS
Ieee 2 =
j oy ee ee.
y %, © e =
Shee NM te
3 oe
3 D He)
O,
‘?

Annals of the South African Museum.

114

Wellington

ate
2
Blaauwberg
ee)
8 =i
‘ yle,
e ues
Tygerberg 3
Green Point “ye
4 Parow %
) i; Stellenbosch :
Pinelands S @ 3 ap, +
@ Qy “mn "Ne
o "ny 8
o Zp A) uty
S/n, © Groote Schuur “Me,
S34, yo e a
es Eerste River

Wetton

e Diep River, © Philippi

Constantia Nek
: =)

ease =

Somerset Strand . .
ae

e

hs
Pe
Av *

Cape Point

, Miles Hangklip

CAPE PENINSULA SUB-REGION

Map 6,

:
>
.
LLU
aye!
vet
J
2
=
-
id
“1,
o
“"
Nn
ty
oe ats
= ae
= uh
= ,
- 4
as!
“=
Ate
LP
a
B
(Or
% ‘
iz
A
e
’ :
‘tees, ome
4, ie
—' .
& = 3
- = &
= 2S a
= ah
z 2:
= =
<= : =
= : a
© : .
= . ee
+ as A)
= .
= Oy eret
=
= to, Wank
Zerent 2B =

Bain’'s Kloof ~~

Hermanus |

The Bored Stones of South Africa. 115

PART III.—MIGRATIONS.

In Part I of this work, and also on page 108, we referred to
the possibility of making certain deductions regarding migrational
routes. This third part will consist of an analysis of a selection of the
material listed above, with a view to reconstructing the routes by
which Later Stone Age man spread inside the Union. It is to be
hoped that further and more detailed work may be done on the
remaining material, as more exact and fuller knowledge is at our
disposal.

It has been shown conclusively that types exist, and areas and
regions have been selected in such a way as to isolate habitually
associated types in relation to geography. This arrangement is not
of such great significance as I had at first thought. My prejudices in
1926 when I laid down my task, and in 1939 when I took it up once
again, led me to presume that I should be analysing tribal or hunting
areas, each of which would be neatly typified by certain forms of
bored stones, and be limited to recognisable boundaries. That may
be the sort of thing these areas do represent, but it is obvious that,
if so, they were neither static nor stable, but that they overlapped and
cut into one another from time to time, probably in relation to the
waxing and waning of tribal strength.

It might be argued by critics that the types result from the group-
ing into areas, and that they do not exist except as an over-careful
geographical patterning on the part of the author. There are several
arguments that should make it abundantly clear that this is not so,
or that such a probability is so remote that it can be ignored with
safety.

1. The arrangement acts in all parts of the Union for the Prehistoric
Series only. In contrast, no manner of adjustment will permit us
to fit the culturally different Bantu Series into any type-area arrange-
ment. If unconscious prejudice had permitted the evidence for the
more widespread series to be so consistently twisted, it should
have been easier to do the same for a smaller and more confined
series.

2. The grouping developed is directly related in every instance to
geography; not as the scientific cartographer sees it, but as an
immigrant pedestrian people would find it on approaching consistently
from the north, progressing by timid trial and fatal error. The
present arrangement into areas in many instances cuts across my own

116 Annals of the South African Museum.

prejudices that expected a neat arrangement of static areas, based
upon my “ecological”? concepts of where those areas should be.

3. It therefore becomes obvious that with a migrating people, the
route and the point of entry are more important than the static
area. In this instance, except in the heavily colonised focal areas,
routes are the more important. The complex of migrational routes
revealed was logical and consistent. Each path was directly related
to, and governed by, geographical factors. Each in turn emanated
from a previous stage, often marked by a period of sojourn in a focal
area. It is the logic of these routes that most decidedly proves the
existence of types per se.

THE SELECTION OF MIGRATIONAL TYPES.

The problem ts not simply to follow migrations, but is complicated
by the multiplicity of types. Two, three, or even more cultural
migrations into a single area have occurred in an unknown order.
At Koffiefontein A and B there is every suggestion that over a long
period there have been at least two cultural waves, and that types
have persisted there long enough to diverge from series A to series B
or in the opposite sense.

It was shown earlier that consecutive, serial, or divergent copying
might either yield divergence or permit errors to cancel out partially
so as to maintain the original grouping. There is thus some risk
involved in selecting closely related types for the purposes of following
migration. It has been shown that within a single area types may
show a scatter of 2-5 cm. (about an inch). The scatter is often less
then 1:3 cm. In working from one area to another, only examples
within that smaller range will be considered. The main deviation
from this rule lies in those exceptional shapes in which the possibility
of assimilation into an adjacent graphical type (that is, the chances
of sliding off one group and on to another) can be eliminated. For
these reasons several members of local groups, as given in Part II
above, will be ignored. Those conforming most closely to the average
will be considered. The general average given with each migration
will usually diverge less than 10 mm. from any specimen within the
group, unless a change of average is given.

ENVIRONMENTAL FACTORS.

In general we can group those environmental factors controlling
man’s sojourn or his migrations under two general heads: barriers

The Bored Stones of South Africa. 117

or limiting factors, either complete or partial in their actions, and
attractors, factors that draw migration to habitable areas or to focal
points. We are dealing in this Prehistoric Series with hunting and
collecting peoples, and in the later Bantu Series with cattle-keepers
and agriculturalists, so that the applicability of these two groups of
factors may not be identical, but variation should not be great as an
area attractive to a hunter is generally suited to a herder.

Barriers.—It is obvious that the major barrier to movement in
South Africa has always been the Basutoland massif. It accounts
to-day for the relatively sharp demarcation between Nguni and
Sotho tribes, and appears to have a general relationship to the dis-
tribution of cave paintings and petroglyphs in the Later Stone Age.
It may have proved a factor dividing the Smithfield and Wilton
people from one another, though the present paper yields little
evidence on this point. This very effective barrier is further extended
northward and south-westward in the series of mountains forming
the main watershed of South Africa. This is markedly seen in the
elongated forms of the bored stone, where upcountry and coastal
forms appear to present two partially separate distributions. The
unexpected importance of the Sneeuwberg as a barrier will be touched
upon later.

In certain areas rivers appear to form barriers; thus the Orange
River at first appears to be an important divide in the distribution
of circular types, and seems to have held back movement southward
for a long time. A river may be a barrier for various reasons, by
reason of its width, depth, volume or strength of flow. But in the
case of the Orange it would seem that the area of further expansion
to the south was too unattractive to draw these early hunters across
the Thirstland to the unknown coast. It needed what Horace calls
the aes triplex, the thrice brazen courage of the navigator, to press
forward across these dry expanses. A final but very important
element is agreement. If two ethnic groups agree to confine them-
selves to their respective sides of a river, then, however small the
stream may be, it will provide a political boundary as effective as
man is willing to make it. The tendency among the early Bushmen
(and indeed among all known hunting peoples) to confine their
activities to delimited hunting and collecting grounds, and to certain
water-rights, is too well known to be ignored.

Rivers have been shown on the maps used, and certain mountains.
The selection of the latter is not related to height above sea-level,
but to “effective height,” which includes the steepness, ruggedness,

118 Annals of the South African Museum.

height above the immediate surroundings, etc., of a range. Ethno-
logically the importance of mountain ranges consists in the drainage,
gaps and neks through which migration may take place, and their
effective height; 7.e. their ability to prevent, inhibit, divert, or permit
man’s migrations in a particular direction.

Attractors.—In general, drainage areas provide the main areas of
attraction in South Africa. In a country where water is often scarce,
animals and men arrange their lives in relation to the distribution of
water. Abundant water will attract, while arid areas will repel or
may provide migrational routes, skirting their edges, or along the
valleys of intermittent rivers, such as the Seekoei River, the Groen-
water, the Hartbees and the Ongers Rivers. Rivers are indeed more
often foci of attraction than barriers to man’s movements. To a
hunter a river that can be quickly and easily crossed by a hunted
animal can hardly be permitted to act as a barrier. Where a gorge
or swift stream intervenes the position may be different.

The Free State has provided so great an attraction to Later Stone
Age man that I have employed the terms “‘focus”’ or “focal area” to
describe it. Here we have a dozen rivers converging on the Vaal and
Orange Rivers, few of them large enough to provide a barrier for most
of their length, and all attractive for the greater part of the year.
The watersheds are never high, consisting mainly of wide downlands,
where cross currents from one river to another have been possible.
We have the Harts and the Vaal, the Vet, Modder, Riet, Caledon and
the Orange. In addition there are northward-flowing streams of
importance, the Valsch, the Wilge, and so on. Of the material from
the area fed by these last I have been unable to obtain much informa-
tion. Most of it is housed in the National Museum, Bloemfontein,
where it is presumably being ably dealt with by Dr H. C. N. van
Hoepen, but has not been made available to me.

The possibility that mountains may, under certain conditions,
prove attractive to hunters and collectors must not be overlooked.
In most areas the deep-cut mountain-sides and the high precipitation
of rain or dew provide the main concentrations of wild fruits, roots,
and edible plants, and overhanging kranses provide shelter from
prevailing winds. The Western Langeberg area gives us an example.
Here the mountain-side on the north of the Breede River valley shows
distinctive types, differing from those in the open valley, west and
south of the river. Basutoland will eventually show a similar
localisation of types.

An analogous state of affairs seems to exist in the kloofs cutting the

The Bored Stones of South Africa. 48)

peneplain towards the southern coast, where deep and often wooded
valleys similarly yield protection, water, vegetable foods, and a
certain degree of isolation. These valleys may link back to the
_ mountain ranges immediately behind, or may run back through the
first range to further valleys running parallel with the coast. The
Gouritz River provides an example of a river cutting across parallel
valleys, and the Sundays River, Bushmans River, etc., are all very
similar.

Remarks on these two opposing functions of geographical forces—
attraction and repulsion—have been touched upon in considering
individual areas above, and will be dealt with further in discussing
specific migrational routes.

MIGRATIONAL ROUTES.

These may be merely points of access between areas of attraction,
fords over rivers or passes through mountains They may, however,
in a dry and barren land, be river-beds only usable in the rainy season
or in years in which there happens to be an episodic rainy season.
The Seekoei River, and probably the Kuruman and Molopo Rivers (of
which we know little in the prehistoric period) were of sucha type. In
many cases it 1s obvious that the headwaters of one river will lead to
within a mile of the headwaters of another, across the low divide of
a watershed. In this way the Limpopo served the upwaters of the
Molopo and Kuruman Rivers, which in turn led across the Orange
River, up the Hartbees River to its headwaters. This route was
chosen by the MaNtatisi movement, and previously provided the
Western Bypass. An analogous action seems to have taken place in
prehistoric times, leading migrations down the Nylstroom from the
Limpopo River basin to the Upper Vaal. The Klip River carried the
early colonisers of Natal from the Vaal to the upwaters of the Buffalo,
at Laing’s Nek. Cases of the same sort will be dealt with as we
proceed.

THE DIRECTIONS OF MOVEMENT.

Except where the general direction of a movement is obvious (e.g.
from the Transvaal to Natal, and from the Transvaal to the Free
State, and those cases in which migration obviously started from the
Free State and spread radially from certain points on the Orange), it
is difficult to discern the direction in which migration occurred. There
is a fundamental presumption in the following pages, that all move-

120 Annals of the South African Museum.

ment tended to proceed from the north southward. It is presumed
that the Transvaal peopled the Free State, which in turn provided the
main source for the peoples of the Cape Province, and the possibility
of any return movement is ignored. Once the southern coast is
reached no real presumption can be made, though there are certain
indications that can be made to supply a clue. When discussing
that stretch of coast I pointed out that lateral movements occurred
between the serried ranks of mountains, in relation either to the
streams running east and west or to the coast-line itself. It is easily
possible in that area for a migration of some magnitude to pass across
the southernmost Cape from east to west, or from west to east,
without affecting a neighbouring movement between adjacent ranges,
or along the coastal belt. This would be even easier along the natural
forest belts near the coast, where a sharply contrasting environment
would possibly deter the movements of an inland people. This area
therefore is to be regarded asa belt along which parallel or even contra-
parallel migrations might take place without mutual interference.

I regard the Cape of Good Hope as generally late—the end of a
series of movements through the Hinterland area along the south
coast. But what of the south coast itself? Was this fed by way of
the Sundays River; was it fed through the Breede River valley;
was it perhaps fed through Oudtshoorn to George and Mossel Bay?
The answer involves all three routes, and we can take it that the
general tendency of coastal spread was from east to west, though
return movements are to be expected. But the almost cross-shaped
catchment areas of the river beds also permit of transverse movements,
either from the coast inland or in the opposite sense. A local and
intensive study of other analysable factors, such as pottery, styles of
painting, implements, etc., may eventually yield a clue to the question.

Migration will be given from east to west as the important sources
of movement seem to lie to the east, but there are subsidiary routes
constantly cutting in on this east to west migration. If the main
southward movement enters the coastal strip to the east, or to the west,
the direction of flow is obvious. If it enters towards the centre
of the coastal distribution, then we may presume a lateral spread in
two directions. Even in the spread of a single type of bored stone all
three possibilities have to be taken into account, as a glance at the
charts provided later will show (pp. 143-145).

Where definite barriers have limited spread, the question is some-
what simplified. Only two such barriers, apart from the sea, exist.
The first is the Basutoland mountain massif. This has provided us

The Bored Stones of South Africa. 121

with a certain key to two areas, Natal and the lower Thornveld, and
a strong basis for presumption regarding the Northern Transkei.
It so happens that the two migrational routes separated by the
Basutoland divide lead from the Transvaal to the heart of the Free
State Smithfield, and to the heart of the Natal Smithfield, obviously
parallel developments from a single northern source. We thus have
a valuable key to the origins of this complex cultural development.

The second important divide, probably the product of chance as
much as anything, is the Sneeuwberg range, which divided the
migrations over the Western crossing of the Orange from those over
the Central crossing of the Orange. The presence of this divide must
in future be regarded as of the utmost ethnological importance in the
Later Stone Age.

In the extreme south it has been suggested that movements to and
from the coast by way of rivers cutting directly across the line of the
valleys running parallel to the coast played an interesting part.
These movements may have been purely migrational, but there is a
possibility that they may have been annual movements from the
coast inland during certain seasons of the year. We must therefore
expect to find seashell ornaments, etc., carried back from the coast
along these rivers to the inland areas. The valleys running parallel
to the coast account for the distribution of a type at Ladismith (Cape)
and in the West Langeberg area, and of a prolate type at Oakhurst
and at Riversdale, with no known intervening examples.

TyPrEs oF ROUTE.

The question of routes dissolves itself into a problem exactly
analogous with that of the modern communications of a country.
There are three main types of route: highways along which migration
normally flows; bypasses that skirt the areas fed by the highways
and so provide a simpler picture; and lateral or feeder routes, that
transgress the more important highways and hypasses constituting
the normal orientation of a country’s flow, and intrude odd elements
from laterally situated sources.

This analogy of the modern highway may be carried further. A
highway starts at a primary distributive centre, from there goods are
carried for distribution in the immediate neighbourhood and yield a
typical strip development, then continue along highways and by-
passes to secondary centres of distribution. This analogy fits the
prehistoric and primitive highway.

_ This secondary distributive centre is an area of focus where migra-

122 Annals of the South African Museum.

tion consolidates, and where the new population may increase to the
limits of local saturation. Then there follows a series of divergent
distributive routes, radial, and often cutting directly across other
highways or joining with them. The cross-road is difficult to dis-
entangle. It depends for its understanding upon the careful analysis
of surrounding areas and the use of the subtractive approach. By this
I mean that once a group is understood in terms of an adjacent area,
it may be subtracted, and the residue dealt with by the same approach.
This method is employed for the crossing of the Namaqualand high-
way and the Western Bypass at Prieska. I would hardly have dared
to untie this Gordian knot had it not been for the considerable help
supplied to me by Mr C. Wilmot of Prieska. The tangle now resolves
itself into a neat pattern of intersecting strands of migration.

In relation to environment various types of route can be dis-
tinguished. Those passing from the headwater of one stream to the
source of another flowing in the opposite direction. Those following
river-beds through dry regions, or along such great streams as the
Vaal and Orange. Routes skirting mountains at a fairly high level—
the southern extension of the Natal Highlands area belongs to this
class. Routes keeping to a low level, and generally related to a wide
coastal belt, such as the Natal Lowlands. Routes confined to narrow
valleys, such as the Groot River, the Breede and the Berg, often cut
off from outside influence.

ORDER OF TREATMENT.

While I have spoken of the Transvaal as the prototype region, a
primary distributive centre, this is only true for South Africa. The
bored stone is too well known in the body of Africa for us to ignore
the fact that itisimmigrant. We will discuss possible points of entry
first. From the Transvaal there are two great routes, the Western
Bypass to the Bokkeveld, and the road to the cul-de-sac of the Natal
Highlands. These will be discussed next, with whatever lateral
byways may have fed them. There is also a central highway down
the Vaal River, from which lateral or feeder roads spread over the
Free State along its southern river system. In this Focal Region
movement seems to have been held back during a considerable period
of populational stability and expansion, if the quantities of bored
stones, implements, rock engravings and paintings provide any criteria.
This may therefore be called the Deuterotype Region, and the new
migrations that passed southward by three or four main crossings
of the Vaal and Orange Rivers may be regarded as having their

The Bored Stones of South Africa. 123

immediate origins here. These crossings of the Thirstland towards
the south, and the mountains towards the east, will be dealt with in
turn as new stages in migration.

South of the Orange the Sneeuwberg seems to provide an important
divide. From here migrations will be followed to the coast and along
it wherever possible. Finally a few movements that do not link
very securely with inland types will be discussed. Treatment will
not be exhaustive, but suggestive examples of major migrations will
be given.

ENTRY INTO THE UNION.

(A) The junction of the Crocodile and Marico Rivers at about
27° H. by 24° S., just north of the Dwarsberg, at once suggests itself
as an important point of entry into the Union, for a variety of reasons:

1. It is nearly related to the Victoria Falls, above which most
major movements have crossed the Zambesi into southern Africa.

2. It is an area in which four groups of bored stone seem to have
diverged from two original types (probably 11:9x8:3 cm. and
10-5 by 8-0 cm.), and spread southwards over the entire Vet-Marico
area, to cut deeply into the whole Transvaal.

3. Lateral movement into the Transvaal in relation to the east-
west valleys and mountain folds of the central Transvaal would be
natural and attractive. In recent years this was the natural line of
movement used by various Sotho tribes.

One cogent argument can be adduced against our selection of this
point as the main doorway to the Union. It is probably important,
and certainly obscure. Of perhaps a dozen specimens from this area
four have a conical bore. If this is a result of the use of metal, as I
suspect, then this might be regarded as an argument in favour of
fairly recent immigration by this route, by a people who were them-
selves users of metals, or were in contact with Hottentot or Bantu
metal users. It is just possible that these stones were bored from one
face only with a stone borer, and that this difficult technique was later
dropped in favour of the easier biconical bore. These examples cut
across other groupings, and can be differentiated into two groups,
12-0 x 8-2 cm. (one example only) and 11-0 x 7-2 cm., close enough to
each other to be regarded as a single group. If these four examples
are removed from the Vet-Marico series as Bantu, the averages are
not very greatly affected.

(B) Even if we regard (A) as a possible point of entry it would not
account for all the Transvaal types. One Central Transvaal type
(14-1 x 10-5 cm.) that certainly developed into the Southern Transvaal

124 Annals of the South African Museum.

type (12-9 x 10-2 cm.) may have entered by way of the Mogalakwena
River or along the Sand River (29° 8. x 30° E.) to the Pietersburg
Highlands.

The 8-0 x 5-5 cm. group, present in the Central Transvaal, and in a
derived form (8-8 x 5-2 cm.) in the Southern Transvaal, is absent from
the Eastern Transvaal and the Vet-Marico River area. I would
therefore suggest a similar origin for this. The same may be said of
the small type (about 6-0 x 4:0 cm.) that occurs at Button’s Kop, and
has spread throughout the Southern Transvaal, presumably as a
knobkerrie head.

(C) The 10-0 x 5-5 cm. and the 10-7 x 4-4 cm. groups, common in the
Central Transvaal, miss the Southern Transvaal area, but reappear in
the little-known Eastern Transvaal. Possibly these two types are
related, but they do not seem to have been associated with the same
migrations that brought the three types mentioned above under (B).
They may be associable with (A), or they may have entered across the
Limpopo, east of the Magato mountains, perhaps by way of the Pafuri
River (31° EH. long.) accessible to both the Central and the Hastern
Transvaal. This route is highly speculative.

These seven prototypes I regard as sufficient to account for most
of the Transvaal averages, and for those of the Vet-Marico River area.
It is perfectly possible that they may be held to account for all the
South African types, except those of the Natal Lowlands and a few
(very few) of the types on the south coast. I refuse to make any
attempt to enter this somewhat speculative field.

Several important points are worthy of consideration. The
Southern and the Eastern Transvaal are mutually exclusive, while the
types in the Vet-Marico area agree with those of the Eastern Transvaal,
and therefore exclude Southern Transvaal types. Expressed differ-
ently, the Central Transvaal includes all the prototypes; the Southern
Transvaal includes three only; the Vet-Marico area shows three, and
the Eastern Transvaal four of the remainder. This suggests the
crossing of two routes in the Central Transvaal. The one from VMR.
to ET., the other through CT. to ST.

Prototype. VMR. Ciz Si KT. Route.
1. 138-5 x 10-5 141x105 12-9x10-2 B
2.12:0x 82 116x8-4 11:7 x 7:8 12-3 x 8-6 A
3. 10:5x 81 10-7x7-6 10:2 x 8-2 10-5 x 8-0 A
4. 10-0x 5:5 10:0 x 5-4 10-2 x 5-9 C
5. 10:7x 4:4 ( 9:4x3-7) 10:8 x 4:2 10:7x4:6 AorC
6. 80x 5:5 io x 1o6 8:8 x 5:2 B
7. 60x 4:0 5-7 x 3:6 6:5 x 4:2 B

The Bored Stones of South Africa. 125

In the Central Transvaal two specimens averaging 12-2 x 9-2 cm.
are obviously related to divergents from prototype 1 as it occurs in
the Southern Transvaal; this suggests a return movement. Two
other specimens from OT., averaging 11-5 x5-4 cm., are seemingly
divergents from prototype 4. One of the VMR. examples approxi-
mates to this type (Bloemhof, 11-7 x 5-8 cm.) and is probably related
to a nearby specimen from CT. at Potchefstroom. The VMR. stone
listed above as a development from prototype 5 is a single specimen,
and the fairly wide divergence from the average suggests that it may
be a secondary development of atypical dimensions.

THe Natat Rovute.*

In general the migrations from the Transvaal southwards along the
Natal route, the Western Bypass, and down the Vaal are all relatively
easy to follow. Almost all the prototypes can be followed along these
three great routes. The opposite is not true, for we can hardly
speculate on the origins of the Natal Lowlands types, and the affinity
these show to examples from the Schoonspruit-Vaal area cannot be
regarded as more than superficial, as the latter disc types are products
of the local schistose rocks.

Some interesting and important points emerge in our study of the
Natal Highlands, each of which needs careful consideration in any
future work in that area.

1. The Natal Highlands types show no relationship to forms south
of the Tsitsa-Tina and Umzimvubu basin, from where only two
examples are known. Spread southward seems therefore to have
ceased at Mount Currie and Mount Ayliff.

2. The Natal Highlands seldom transgresses the Natal Lowlands
area, and there appears to be a marked divide at about the 2000-foot
level (650 m.). The Lowlands area was certainly fed from a different
source, probably from the Portuguese East African, St. Lucia, and
Zululand Lowlands.

3. The Highlands are situated along the eastern slopes of the
Drakensberg, and are cut off from the Free State by the Basutoland
massifs.

4, The area is further cut off from the Free State focal area by an
important field of distribution, the southern portion of the Vet-

Marico area, that contains no types similar to those from the Natal
Highlands.

* Maps of routes are added at the end of this part, following page 142.

126 Annals of the South African Museum.

5. It is highly probable that this Vet-Marico area was not a high-
way, but an area of spread or colonisation. It1is obviously a marginal
area with few types, lying between more attractive migrational routes,
the Vaal running west and south, and the Klip River running to the
Majuba Hill, where it almost meets the Buffalo-Tugela complex some
8 or 9 miles away.

6. Excluding the Durban anomaly, five of the remaining six Natal
Highlands averages are represented in the Transvaal prototype areas,
and show affinities with prototypes 1, 2, 4, and 7. Two derivatives
of this last are represented, one related to the Southern and one to the
Central Transvaal. Prototypes 1, 4, and 7 are absent in the VMR..,
while the Natal Highlands derivative of type 2 is allied to the Hastern
Transvaal form, not to the VMR. average.

7. Certain Natal Highlands types are associable with Smithfield N
material, a variation showing marked differences from the Smithfield
series of the Free State, and one which I have shown has characteristic
affinities with Southern Rhodesia that are absent in the Free
State.*

8. The closest affinity (apart from the prototype area) with the NH.
area is at Bloemfontein and Sepani, where one type only shows
analogies. The BS. area is cut off from NH. by the Basutoland
massifs and by the VMR. area.

We thus have at our disposal the material for tracing the migrations
of the Later Stone Age people from the Transvaal to the Natal
Highlands. This suggests a possibility of an early pre-Smithfield
culture in the Transvaal, parent to both the Free State and the Natal -
variants.

1.¢ (13-5 x 10-5 cm., Prototype 1.)

Potgietersrust.
| Brits.
Brits-Rustenburg Road. Pretoria.
Rustenburg. Bronkhorstspruit.
- Standerton.
Estcourt.
Maclear.

Prototype 1 may have originated either in the Western or the
Northern Transvaal, but its journey from Pretoria is clearly through
Newcastle or Botha’s Pass, or another nearby gap in the mountains
about Majuba (29° 50’ E., 27° 30’ 8.).

* Goodwin, Proc. Rhod. Sci. Ass., xxxiv, 1934.

{+ Examples of migrational routes are numbered consecutively for ease of
reference. The same initial number is used on the maps for each route.

The Bored Stones of South Africa. 127

2. (11-9 x 8:3 cm., Prototype 2.)

Lichtenburg.
Ventersdorp.
Eerstefabrieken.

Kransfontein. Ermelo.
Acton.— ————_ —————

ore

Utrecht.
Stanger.
Durban.

Prototype 2 had a different story; it seems that one movement
went along the valleys of the Transvaal from west to east as far as
Ermelo, then through Pietretief to Utrecht. An alternative route
from EHerstefabrieken passes Harrismith and the van Reenen’s Pass.

The remaining three examples given below all suggest the Majuba
route. It is to be observed that Prototype 4 is the most clearly
associable with the Smithfield N material at Balgowan.

3. (10-0 x 5-5 cm., Prototype 4.)

——WMolsgat.
W onderboom. Belfast.
ee neeneton Bethal. Jakslust.
: Balgowan. Barberton.
Durban.
Matatiele.
Kokstad.

4, (6:0 x 4:0 cm., Prototype 7.)

eS oe. Oa ee

Magaliesburg. Germiston.
Johannesburg. Bergville.
Pietermaritzburg.
Underberg.
Cedarville.

Analysis of a further, final series is unconvincing. The type
(7-5x2-5 cm.) is evident only in the Suikerboschrand valley at
Vereeniging and at Barnard’s Kop, Heidelberg (Transvaal), and
appears to link with the Harding specimen from Southern Natal, a gap
of 300 miles.

Natat LOWLANDS.

Of the origins of the Natal Lowlands types with indices below 40 we
know nothing. Far more will have to be known of the Northern
Transkei, especially the coastal belt, before we can presume a migration
from the south. It is more likely that this Natal series inks through

128 Annals of the South African Museum.

Zululand lowlands with Portuguese Mocambique, areas as little known
as those to the immediate south.

WESTERN ByPaAss.

Five migrations, starting at various points and continuing for
varying distances along this Bypass, can be followed. The Bypass
consists of three areas skirting the eastern edge of the Kalahari and
the western watershed of the Harts and Vaal Rivers. It seems to
run from the sources of the Molopo River southwards to within 150 miles
of Cape Town, and there ceases. The three areas are the Molopo-
Groenwater, the Upington-Hartbees, and the Bokkeveld areas.

5. (13-7 x 11-2 cm., related to Prototype 1.) Obviously originated
in the Central Transvaal, from whence the Vaal-Harts area is reached,
and examples occur at Bloemhof and Newlands. From there the
route follows the Vaal to Rietpan, Barkly West, and to Fauresmith
and the Modder River. A parallel road, continuing farther to the
south, and presumably starting at Bloemhof, runs as follows:—

Vryburg A.
Postmasburg.
Niekerk’s Hope.
Uitdraai.
———Springbokpoortje.
Groetvloer. Omdraaivlei.
Vanwyksvlei. Vosburg.
Blaaukrans (Carnarvon). Victoria West.

This type does not reach the Bokkeveld area.

6. (12:3 x7-2 cm., Prototype 2.) Perhaps an offshoot from the
Central Transvaal type ui, where examples occur at Ventersdorp and
at Lichtenburg, from where the route runs:

Vryburg A.
Kuruman.
—_— Bultfontein. |
Upington. Redlands (Prieska).
Kenhardt.————_ Keikamspoort (Prieska).
Hondeklip Bay. Calvinia. eet
Klaver. Ie

7. (10:9x 7-2 cm., related to VMR. iv.) Sites at Klerksdorp,
Kroonstad, and Brandfort provide a fan-shaped distribution, con-
verging on Bloemhof from where the distribution runs:

Postmasburg.
Groot Nauta.

Upington. Middelwater.
Calvinia.

The Bored Stones of South Africa. 129

Kimberley and Douglas show analogies, and it is possible that there
was a lateral movement along the Orange. This route obviously
resembles example 5.

8. (8:0 x 6-0 cm., Prototype 6.) A lateral drift that does not seem
to have followed the Bypass in either direction, it may represent part
of a roadway to the Southern Kalahari.

Koffiefontein.
Kimberley.
Gonggong (Barkly West).
Winter’s Rush.
Douglas.
Kuruman.
Wolhaarkop.
Koegas.

9. (6°38 x 4-8 cm., Prototype 7.) This lateral movement follows
much the same route, but turns southward across the Orange to the
Trekveld and Winterveld areas, presumably from Barkly West.

| ——wWindsorton. |
Kimberley. Botterhoek.
Kanteen Kop.
Kuruman. Barkly West.
Goup.
Wildebeespan.
MOSPUCS coeaaateaa |
Vanwyksvlei. Britstown.

Victoria West.

This may provide the origin of the Trekveld iii type that shows a
local distribution at Hliaslaagte, etc.

NAMAQUALAND RovtTE.

Although this route appears to have some relationship with the
Vaal-Orange confluence, I shall deal with it here, partly to clear up
the western Thirstlands, and also because I suspect the possibility
of a direct route along the Molopo River to Namaqualand. The
evidence is slight, and consists at present of a marked gap in example
12 below. The Orange River route would have been far easier to
follow, once the confluence of this river and the Vaal had been
reached. A long stretch along the Orange is quite unknown to us at
present.

10. (12:0x9-5 cm.) This starts from Newlands, and, for all we

NiOlia XOX Dy PART. I; 9

130 Annals of the South African Museum.

know, may also have passed directly along the Molopo from there.
The southward route is quite clear.

Newlands.
Riverton.
Barkly West.
Pniel.
Schmidt’s Drift.
Prieska.
Witfontein.
Steinkopf.
Concordia.
Kamieskroon.
“Namaqualand.”

11. (10-5 x 8-5 cm.) Almost certainly related in ultimate origins

to 10.

Windsorton.
Witboom, Hay.
Katlani.
Grootdoorn.

_ Prieska.
Marydale.
Steinkopf.
Garies.

12. (90x75 cm.) A shorter form of the above route. Four
examples occur at Prieska. There are three possibilities here. Hither
this is a movement directly across to Namaqualand from Fourteen
Streams, with a return movement along the Orange; or it is a move-
ment down the Vaal similar to 10 and 11; or it links with the Free
State, a complex that can be seen under example 23 below.

Hoopstad.
Sheppard Island.
Fourteen Streams.

(gap)
St. Clair.
Prieska.
Kaaienbult.
Concordia.
Steinkopf.

Finally, the small Steinkopf specimen (4-2 x 3-3 cm.) has analogies
at Smithfield, Bethelpella, Koffiefontein B, Belmont, and in the
Winterveld. The similarity is suggestive, but carries no weight at
present.

SoUTHERN HIGHWAYS.
Enough has been said elsewhere to show that most of the pro-

totypes of the Transvaal have passed southward along the Vaal
River by way of the Vaal Harts, the Middle and the Lower Vaal areas,

The Bored Stones of South Africa. 131

or across a wide front on the Free State section of the Vet-Marico area.
The next series of movements to be discussed were from the Vaal,
generally eastward along the complex pattern of rivers in the Free
State, and transversely across the direction of flow, over low hills or
- denuded plains set with isolated doleritic kopjes. It would be useless
to trace these movements in detail. The references given in the
column of comparisons, after the type averages in Part II, should
speak amply for the cross-currents in the confined focal areas.

This Free State Region is important. It supplied a focal area for
consolidation, internal expansion and finally for movements to less
thickly populated areas. The Orange River and the Basutoland
mountains with their southern extensions provided a barrier to the
south-east, while the increasing aridity encountered by any movement
towards the west from the Vaal basin across the Kaap Plateau, and
the general dryness of the land to the south of the Orange, precluded
easy migration in these directions. We must therefore presume that
there was little incentive for the Free State people to break through
these barriers until pressure forced such a drastic step. Pressure may
have been the product of over-population, but is more likely to have
been due to the pressure of new peoples from the north, be they
Hottentot, Wilton, or Bantu. The very considerable lag due to
consolidation in the Free State may account for the obvious lateness
of the Later Stone Age along the coastal belt, especially towards the
Cape.

The breaking through of the Orange River ethnic barrier (or better,
the initiation of real migrations across the Thirstland, in contrast
to confined hunting movements) provided an effective filter. Only
certain types transgressed towards the south and south-east, and the
routes they took seem to have been limited.

We can recognise three points of origin for the various movements
that left the Free State and Vaal River Regions. This small number
of breaks-through on a front seven hundred miles long is surprising,
until it is realised that the Orange did not in itself provide the major
barrier, but the repellent country farther south and east. The Orange
may have been crossed at any point, but migrations across the Thirst-
land tended to follow the same routes in their new migrational
pattern.

The first point seems to have been near the present main railway line
from the Cape to Kimberley, between Belmont and Hopetown. Next
were points between Bethulie Bridge and Aliwal North, forming the
Central Crossing. Farther east there is a more speculative break-

132 Annals of the South African Museum.

through from the eastern sources of the Orange, high in the Basuto-
land Massifs, through Qacha’s Nek and neighbouring gaps to the
Matatiele district, the Hastern Exit. These three crossings will be
dealt with in order from West to Hast.

THE WESTERN CROSSING.

As we have already seen, the main routes to Namaqualand and the
feeders to the Western Bypass seem to have crossed the Vaal at
Douglas. This crossing may be said to be the gateway to south-
western Africa. In contrast the southward routes appear to have
crossed at a point coinciding with the origins of the Winterveld area,
between Belmont and Hope Town. If Douglas fed Namaqualand, the
Orange River basin and the Trekveld, then the general area of Belmont
was the gateway to the Winterveld and certain points on the south
coast of the continent.

We may follow the more important Winterveld types. They appear
to yield somewhat sketchy routes, but they are sufficiently consistent
to prove that this route across the dry semi-desert country was used
by man. It has good rains in certain years, and was once the haunt
of great herds of springbok and other mobile types of game before
man interfered with essential migrations by erecting fences. Oddly
enough, the Vosburg specimens were found in association with a large
unidentified pachyderm (possibly rhinoceros), suggesting that even
these animals migrated across the dry Winterveld at times. Three
environmental factors determined this route, the Ongers River, the
Seekoei River, and the Sneeuwberg mountains.

13. (14:6 x 13-0 cm.) This distribution does not suggest a true
migration, but rather a local spread. The type occurs at Rietpan,
Botterhoek, Belment, then crosses the Orange to Philipstown,
Omdraaisvlei, and Britstown. The size suggests that it may quite
possibly be Bantu, the product of the final scattering of the MaNtatisi
refugees.

14. (14:7 x 10-0 cm.) This has a scattered distribution. Unlike
example 13, it is present in the Free State proper at Koffiefontein B.
The further distribution includes two sites near Prieska and one near
‘Britstown, south of this it occurs at Victoria West, Three Sisters,
Beaufort West, and Aberdeen.

15. (12:9» 9-0 cm.) The graphical scatter is remarkably compact
when worked out on a graph paper. It is represented by the first
specimen listed at Koffiefontein A, then follows this route southward:

The Bored Stones of South Africa. 133

Koffiefontein A.
Welgevonden. Britstown.
Poortje. Graaff Reinet.
Jansenville.
Swartberg.
Mossel Bay. Knysna.

Breede River.
Cape Middens.
Sandveld.

The southern portion of this distribution seems to run behind the
mountains, parallel to the southern coast, and to avoid the coast for
most of the route.

16. (10:0x7-5cm.) This probably belongs to the same route as 15,
but there are gaps.

Smithfield.
Koffiefontein B.
Britstown.
Hanover.

From here it occurs in the Swartberg area, and from there links
with Knysna (KTZ. ix), SR. i, BMR. v, and WL. 1. The journey
from the Swartberg to the West Langeberg may be associable with a
mountain people. (Compare example 26.)

17. (3-7 x 2-5 cm.) Beadlike or ornamental forms that show a
wide variation in size occur at Smithfield, Koffiefontein B, Belmont,
Hope Town, Middelwater, Britstown, and Hanover. South of this
the smaller examples disappear except for a single example from the
Helderberg, and any deduction would be speculative. It seems likely
that these too may be MaNtatisi in origin, and they have much the
same dispersal as example 13, but compare the note below example 12
(above).

It will have been observed that the Koffiefontein B series on the
northern side of the Orange plays an important part in the study of
those groups that are sufficiently numerous to follow. This does not
mean that Koffiefontein was the original home of these Winterveld
people, but it does provide an important clue to the point at which
the river was crossed by the migrations passing west of the Sneeuwberg
range, and also it is highly suggestive of the type of culture that was
carried along this route. We should expect a new point of dispersal
in the Swartberg area, feeding the coast at Mossel Bay and Knysna,
and feeding westward to the Western Langeberg along hidden valleys,
then across to the Sandveld and Swartland areas by way of the Breede
River.

The importance of Britstown in all these distributions is quite

134 Annals of the South African Museum.

certainly due to its association with the Ongers River and to its close
proximity to the Seekoei River, north of the Sneeuwberg.

CENTRAL CROSSING.

In much the same way that Koffiefontein provides a point of origin
for migrations across the Orange to Britstown, so Smithfield provides
a bridgehead for the Central Crossing that leads to the Upper Thorn-
veld area, From here two alternative routes seem to have been
presented, either directly down the Tsomo and Great Kei Rivers, or
else down the Fish and Sundays Rivers. At times both may have been
taken. Enough examples will be considered to prove the presence
of these routes and to show the general line of their direction, con-
sistently east of the Sneeuwberg mountains.

18. (13-0x 11-0 cm.) This rather complex series probably links
with example 5 above, either in its origins or in its middle phases,
but intervening sites are absent.

Middle Vaal.
Modder River.
Fauresmith.
We os a) Gee
Burghersdorp Rossouw.
2 Cathcart Downs.

(Murraysburg). |
Grahamstown (14-2 x 10-5 cm.).
Port Elizabeth.

This may or may not prove to be linked with a more local move-
ment with some changes in size. The evidence is added here in case
later evidence clarifies the position. The direction taken is speculative,
and while I suggest that Murraysburg provides the link, it may possibly
occur between Port Elizabeth and Knysna.

a

Murraysburg (12-1 x 11-0 cm.).

Beaufort West. Grootrivier (13-0 x 12-0 cm.).
Ladismith. Knysna.
Montagu. |
Worcester. Kogel Bay.

Gordon’s Bay.

19. (8-5 x6-5 cm.) Apart from one digression to Paardekraal,
there seems to be no evidence of this type in the South-Hastern
Lowlands area. It occurs at several sites in the Central Free
State.

The Bored Stones of South Africa. 135

Winter’s Rush.

Barkly West.
Kimberley.
Koffiefontein A.
Smithfield.
pe North.————
Steynsburg. Dordrecht.
Graaff Reinet. Paardekraal.
Aberdeen A.
Knysna.

Great Brak.

Mossel Bay.
Gouritz River.
Still Bay.

20. (9:1 x4:7 cm.) This route first strikes across to Port Elizabeth,
then along the coast as far as the Cape.

Smithfield.
Hofmeyr.

| |
*Aberdeen B. Cradock.
Somerset East.

Humewood.
Hankey.
Tzitzikamma Middens.
De Poort.
Knysna.
Westford.
George Coast.
Mossel Bay.
Gouritz River mouth.

At Gouritz the type changes to 10-0 x 4:3 cm. and continues:

Gouritz R. mouth.
Heidelberg.
Swellendam.
Villiersdorp.

Cape Hangklip.

There is a second change here to 9-7 x 5-1 cm., and this form occurs
at Gordon’s Bay and Stellenbosch, possibly ending its migration at
Muizenberg.

It is likely that the 10-4x4-4 cm. type at Knysna represents a
return movement of the changed type from Heidelberg. It does not

occur farther east.
21. (80x 4:5 cm.) This type shows a variation of less than two-

fifths of an inch in its wanderings along much the same route as that
outlined in 20 (and similar to 24). The type is slightly smaller than
in 20. The central Free State distribution is not discussed.

* Possibly links with Knysna, also Hanover and Victoria West.

136 Annals of the South African Museum.

Central Free State.
Burghersdorp.

|
Cradock.

Aberdeen. Somerset EK.
Zuney.
Hankey.
Tzitzikamma Mountains.
Woodbourne.
De Poort.
Knysna.
Mossel Bay.
Bloembos.
Montagu.
Worcester.
Stellenbosch.
Ida’s Valley.
Vlottenberg.

|
Clarke’s Siding.

Notice the consistent movement inland (but still within the coastal
belt), represented here by the step from Blombos to Montagu and
Worcester. An analogous series is represented at Hanover, and
occurs through the Trekveld area—what the relationship may have
been I do not know. Probably both series originated in the Free
State.

SOUTHERN TRANSKEI.

A few additional examples outlining the routes to the Transkei may
prove helpful, though the lines are short and links are sometimes

unsatisfactory.
22. (12:4 x 8-3 cm.)

Koffiefontein A.
Smithfield.
Cathcart Downs.
Kingwilliamstown.

23. (9:2 x7-8 cm.)

Avalon.
Fauresmith.
Cathcart Downs.
Happy Valley.
East London.
Debe Nek.

This is similar to the Kaaienveld type outlined as example 11, and the
types may be linked in their origin in the Free State.

The Bored Stones of South Africa. 137

24, (11-4 x 5-3 cm.)
Hofmeyr.
Somerset East.

Humewood. Fort Meroe
South End. East London.
Tzitzikamma Caves.
Belvedere.
_ Mossel Bay.
Heidelberg.
West Langeberg.
Gordon’s Bay.
Cape Flats.
Cape Middens.

These should be sufficient to demonstrate a route through Cathcart
towards Kingwilliamstown, thus helping to fill the gap that the
distributions given above may show. In the final example above,
the sequence of the sites from Hofmeyr and on to Mossel Bay is much
the same as that given in example 20 above, but the eastward route
to East London is new.

EASTERN EXIT.

Our lack of knowledge of the prehistory of Basutoland makes it
impossible to prove the existence of this route, but enough knowledge
exists to permit the suggestion that the two Northern Transkei
examples came by this road, and that the migrations did not proceed
northward to feed Natal from there.

25. (14:0x 7-0 cm.) This is an unusual form, and in such a case
we can expect considerable variation. Examples of analogous shapes
exist at Hoopstad and Douglas, but the Koffiefontein B 1 (12-8 x 7-3
cm.) is the nearest to this Thabankulu stone. The difference in
diameters is less than half an inch.

26. (10-7 x 7-4 cm.)

Central Free State (11-0 x 7-0 cm..).
Smithfield.

Aliwal North.
Xuka Drift.

The variation is 4 mm. or about one-sixth of an inch, and all specimens
are related to the Orange River basin. This example is more valuable
than the first, as distances are nowhere very great.

The suggestion is that the route lay along the upwaters of the
Orange, and then cut across at Qacha’s Nek, Ongeluk Nek, or over
the Pack Ox Pass to the headwaters of the Kenegha and Tina affluents
of the Umzimvubu. Such a hypothesis would explain the fact that

138 Annals of the South African Museum.

the Pondoland Smithfield types are closer to the Free State forms than
are the Smithfield N series. No bored stones are associable, so far
as I know, with the Pondoland material.

THE CoastaL HiGHway.

A few of the lateral movements along the southernmost coast of the
continent may be considered briefly. Additional data may be sought
under the examples given above, but we must here consider those
movements that do not clearly emanate from inland sources. The
example given under 20 above, suggesting the return of the 10-4 x 4-4
cm. type from Heidelberg (Cape) is instructive, as it demonstrates
how an inland type may develop until it becomes a purely “coastal”
form.

The south coast seems to have had a difficult history. There appear
to have been two important focal areas of attraction, Knysna and the
Cape Peninsula. The route between these will prove easier to dis-
entangle than the Free State, as the valleys running parallel with the
coastal belt are more definite than the low watersheds of the inland
plateau.

Movement seems to have been partly lateral along the coast, partly
lateral behind the mountain chains that provided independent path-
ways for migration, and partly transverse, by way of such rivers as
reached the coast through the lattice of mountain ranges, the Sundays,
Gamtoos, Groot, Gouritz, and others. The Breede River is an impor-
tant highway, partly feeding the Cape of Good Hope through the
Hinterland area, but at the same time providing a bypass directly
across to the Sandveld and Swartland areas. In contrast the little
Bot River valley (including the Palmiet and Swart Rivers), hemmed
in by high mountains, is an isolated area with little real relationship
to the outside world.

27. (10:0x 7-5 cm.) This type has been discussed very shortly
under example 16 above, and therefore must be regarded as having a
possible inland origin.

Botha’s Post.
Fort Brown.
Gamtoos River.
Knysna.
Swartberg.
Oudtshoorn.
Worcester.
Pinelands.

The journey between Oudtshoorn and Worcester quite certainly

The Bored Stones of South Africa. 139

passed behind the coastal ranges, and specimens of this type are
absent at coastal sites west of Knysna.
28. (7-8x3-9 cm.) This is apparently a midden form, though

reminiscent of an inland type, Winterveld viii at Hanover.

Cove Rock.

Uniondale.

Zuney.

Maitland Mines.

Tzitzikamma Mountains.

Woodbourne.

Mossel Bay.
Bloembos.

CoNCLUSIONS.

Enough has been said in the previous pages to permit us to make
certain generalisations regarding the peopling of South Africa in the
Later Stone Age. So far as I know, the peopling of no country in the
prehistoric world has ever been so exactly definable; and nothing but
the bored stone could have yielded a series of routes, each capable
of proof again and again, and each showing a remarkably logical
course. Few of these routes could have been predicted from previous
knowledge, yet each is demonstrably the obvious road for people
with a localised knowledge, extending perhaps a hundred miles in any
one direction, and entirely dependent upon hope and travellers’ tales -
for knowledge of water-supply, vegetable foods, etc., in any direction
of new migration. We may therefore epitomise some of our main
discoveries, with a view to laying the foundations of a new branch
of science, the study of migration through the medium of the bored
stone.

1. There would seem to have been seven prototypes present in the
northern parts of our known distribution. These with a few deriva-
tives will adequately account for types in the Transvaal Region, the
Natal Highlands and the Vet-Marico area.

2. The distribution of these seven prototypes in the Transvaal can
be explained by two main migrations, one from the Vet-Marico to the
Eastern Transvaal area, and one from the Central Transvaal to the
Southern Transvaal area. The Central Transvaal reflects all types,
and is therefore the cross-road.

3. It is possible that all types (excluding those from the Natal
Lowlands) may have originated from these seven, though no attempt
is made to demonstrate this in the present paper.

4, The peopling of the Natal Highlands took place almost exclu-

140 Annals of the South African Museum.

sively along the Wilge and Klip Rivers, to the upwaters of the Tugela
and Buffalo Rivers from the Transvaal Region. These types did not
originate in the Free State, and they do not exist south of the great
Umzimvubu basin.

5. Nothing can yet be reed regarding the Natal Lowlands,
which may have been peopled from Mocgambique, by way of the
Zululand Lowlands.

6. While the Western Bypass provided a highway from Vryburg
to Klaver and Calvinia, this route was fed by lateral movements over
crossings of the Vaal between Barkly West and Douglas, and the
neighbourhood of the Vaal-Harts junction, related to the Klein Riet
River, and along the Kaap Plateau escarpment.

7. These last, and similar crossings of the Vaal, provided the types
that reached Namaqualand. Their route follows the Orange River,
though no demonstrable proof can be shown for continuity between
Upington and Steinkopf. Identity of types at each end of this hiatus
permits us to accept this route as a working hypothesis.

8. An alternative route from the Vaal-Harts, along the Kuruman
and Molopo Rivers to Namaqualand, is suggested, but only negative
evidence can here be adduced.

9. There is a great and important highway from the Transvaal,
by way of the Vaal River, which has not been specifically discussed
as the evidence is too obvious to be further laboured.

10. From this Vaal highway movements back and forth along the
rivers of the southern Free State are common, and provide the
material in the focal area of the Free State, which is cut off from
the Transvaal by the Vet-Marico area.

11. Crossings of the Orange River, in the stretch oppesite Hope
Town (related to the Ongers and the Seekoei Rivers), fed only those
routes that pass west of the Sneeuwberg (24° 30’ E. by 31° 45’8.). From
there access was gained to the Sundays River near Aberdeen and
Jansenville, and to the Doorn River near Beaufort West. The types
represented on these routes spread both east and west between the
parallel mountain folds of the south, and seem only to have reached
the coast at isolated points, such as the Sundays River mouth, Knysna,
and Mossel Bay.

12. Crossings in the neighbourhood of Smithfield and Aliwal North
provided routes that ran east of the Sneeuwberg, to feed the South-
Kastern Lowlands by two routes, one related to the Great Kei, the
other to the Great Fish River and the Bushmans River. From these
routes the greater part of the southern coastal belt was fed, and these

The Bored Stones of South Africa. 14]

migrations eventually give rise to most of the so-called “coastal
midden types.”’

13. There is little real reason for divorcing these southern midden
types from inland forms on a basis of either size or index. The differ-
ences appear to lie rather in the numerical proportions of the various
types that are presented in each environment.

14, The Central crossing also provided routes that cut directly
behind the south-eastern mountain ranges and south of the Sneeuwberg
to build paths that often transgressed the routes feeding the south
coast, running towards the Langeberg and the Breede River valley.

15. It is suggested that movements from the Free State, by way of
the sources of the Orange River, passed over Qacha’s Nek to the
Matatiele district, to provide the Northern Transkei specimens,
which do not agree with any Natal types immediately to the north.
These movements missed the Southern Transkei, and did not penetrate
Natal from the south. The Northern Transkei distribution will
therefore probably prove itself to be an intrusive divide between the
Southern Transkei and Natal, and is therefore important.

16. Types from the Prieska area and from the Western Bypass
reached the Cape of Good Hope. The Prieska material does not
transgress the 32° 30’ S. parallel.

17. Routes laterally from east to west (or sometimes conversely)
can be traced along the southern coastal belt in two or three parallel
paths, each confined along part of its course by either the coastal belt
or by valleys between the first, second and third mountain folds.

18. These routes generally fed either the Breede or the Bot River
valleys, both important as sources for types occurring in the Cape
Hinterland, and at the Cape of Good Hope, either as Flats or as
Midden types.

19. The same source supplied the Sandveld area; but it is note-
worthy that the Swartland area shows no affinities with any known
types in the rest of the Union.

20. The Cape of Good Hope seems to have been fed by a variety
of streams from the east, coming either by way of the Heidelberg
district, through the Bot River Valley, or from various similar sources
through the Breede River valley and the Cape Hinterland. No
material seems to have reached the Cape by any direct route from the
north.

21. These migrations covered a long period of time. A first halt
appears to have occurred at the Vaal River, after which spreads took
place by way of the Vaal, and to Natal. A second and more important

142 Annals of the South African Museum.

halt on the main highway occurred at the Orange River, after which
spread continued along the routes outlined in 11 and 12 above.

22. The period involved by the second halt is probably sufficient
to account for the apparent late survival and advancement of Middle
Stone Age types south of the southern mountain belt, such as the
Still Bay, Howieson’s Poort, and perhaps the Mossel Bay culture.

23. The Berg River valley, and the Olifants and Doorn Rivers,
provide a wide and unknown belt between the Bokkeveld and areas
farther south. No deductions can be made regarding this whele
stretch, but it may eventually explain the anomalous Swartland types.

24. Much local work is needed to discover the habitual associations,
not only of cultures, but also of each type of bored stone. I have
suggested that two or even three different types may normally repre-
sent the stones made by a single culture. While the question of the
differing sex of the users of digging-stones and knobkerries clouds the
issue, this should eventually be cleared up in the field.

PART IV.—ELONGATED PREHISTORIC TYPES AND
MIGRATIONS.

We have dealt with the circular and subcircular examples of the
Prehistoric Series of bored stones, and have followed them in their
migrations through southern Africa. We now discuss the elongated
specimens. These are defined as those in which the length is greater
than the width when casually inspected in plan. It has been found
that 8 per cent. represents the difference immediately observable, and
all the examples with a difference greater than this are listed here.
In a few cases it is difficult to be certain whether a stone was intended
to be circular or elongated, as subsequent breakage or use as a grinder
may have reduced the width. In such cases a careful estimate of
the original shape is made, and the missing figure given in brackets.

Certain questions arise in discovering the real relationship between
the circular series and the elongated. How far can these be divorced
from the circular series? Do they represent a variant that may
appear at any period, or in any culture or environment? Do they
represent a distinct cultural group or migration with a more limited
distribution, and particular distributional routes? Were they made
for the same uses as the circular examples? Is the elongated form
the result of a scarcity of round pebbles or is it the result of deliberate

The Bored Stones of South Africa. 143

[ft a

| a ¥y
¢ y

WESTERN BYPASS. | 7".

e
eo?
en
OG-0lo oid 0D OY

3 : "24710 2 2\2 2\35 2

Mars 7 AnD 8.—Migrational Routes to Natal and to the Western Bypass.

2\l

WESTERN
CROSSING

31

CENTRAL CROSSING re Be

++ BARRIERS

be
rim; FIELD
iH t “KL
pry: es
4), e eh d pan
3a gy i an *
33 18 Z\, ae
e arr a We

) : EGS = aa
721 Co :
poke ASTAL. B Ett

Maps 9, 10 anv 11. Fe Routes to N eS 5 and FF way
of the Western and Central crossings of the Orange River.

145

The Bored Stones of South Africa.

‘suowtoods Ie[NoILO
uMmouy Jo UI] syuesorder uly poyjop AavoyT “possnostp wore pue

siolieq ‘soynor urew Jo ArvwmuUNg—'ZT IVI

10

PART l,

5)

VOL. XXXVII

146 Annals of the South African Museum.

choice? Are we perhaps dealing with a coastal type, used by people
who have penetrated inland at various points?

I shall attempt to answer some of these points, or to suggest factors
that may determine the eventual answer. For instance, it can be
said at once that these are not variants that appear universally.
The elongated series is the more localised, and in certain areas it is
completely absent. We thus have some 44 examples from Smithfield
and the Albert district presented by Dr. Kannemeyer, all of these
are circular, none are elongated. In the same way, elongated forms
only exist along one-sixth of the Western Bypass.

The suggestion that the elongated forms may be symptomatic of a
separate migration is not excluded by the finding of round and
elongated types in apparent association on the surface of sites or
kitchen middens. The very few cases in which association is suggested
by identity of museum accession number may be variously accounted
for. Quite apart from culture contacts, merging, marriage or
imitation, there are the difficult questions of chance association, and
the habit of presenting examples to a museum in batches whose true
provenance has been forgotten or muddled. This question of associa-
tion must be watched by collectors, and documented by our museums.
The limited distribution does suggest that we are dealing with a
distinct cultural spread.

The elongated series does not present a coastal type. As in the
case of the discs and flattened forms, this is a question of the pro-
portions of types. Inland the elongated types are rare; at the coast
they are somewhat more common. This might suggest that we are
dealing with a coastal people who may have spread inland. Evidence
from other parts of the continent would have to be considered, and
F. Cabu and M. Van den Brande * show that the proportions of
elongated to circular examples in the Congo is high. Material
collected in the Southern Rhodesia Museum, and kindly submitted
by Neville Jones, gives a very similar impression. These facts
suggest that both sorts of bored stone originated inland and to the
north of the Union.

Two facts point to the elongated forms being the result of a deliberate
selection of elongated stones. One is that the circular examples
were themselves selected for size; diameter and thickness were both
important, and seemingly part of a tradition. The same can be shown
for the elongated series, in which length and width were both im-

* “Contribution 4 l’étude de la répartition des Kwés au Katanga,’ Ann. Mus.
Congo Belge, D, Série I, tom I, fasc. 4, 1938.

The Bored Stones of South Africa. 147

portant, though thickness was almost completely haphazard. It is
obvious then that this form was a chosen type. In addition the
distribution of the circular series includes the area of spread of the
elongated. The abundance of circular specimens shows that circular
pebbles were everywhere obtainable, if therefore there is any question
of choice then circular types would have been the compromise forms,

METHOD.

We have not yet enough material to make use of the region and
area system developed for the circular types, but specimens are
grouped under a new set of regions that I have left somewhat indefinite,
as future material is likely to increase these very considerably in
almost any direction not limited by an adjacent region. No attempt
has been made to group examples into types. Type averages are
given, and grouping can generally be deduced by eye, or recourse to
millimetre graph paper will supply the exact grouping.

These examples have three variables—length, width, and height
or thickness. They should therefore be plotted three-dimensionally.
This was originally done by grouping the specimens by one variable
(the height), and plotting each group so obtained by the other two
variables (length and width). Two very curious facts emerged from
this study. It was first observed that height, or the thickness of the
original stone, has little significance, and need only be taken into
account when the stone is extremely thin, 1 or 2 cm., and the type
is an elongated disc. The second discovery was that the proportion
of length to width is slightly more important than the exact measure-
ment of either. In plotting graphical distribution this means that
the scatter tends to be leaf-shaped rather than circular, with its axis
parallel to a diagonal representing certain proportions (generally 6 : 5).

Types have therefore been plotted by the first two figures only,
height being ignored. It would possibly be helpful to develop an
index, on the formula

Width x 100
Length

and to plot this against either length or width. The cube is useless
in assessing the type, as it is a product of height, which seems to have
had as little significance to the makers as the height of a plug has to
ourselves.

Regions have been altered to suit the distribution of elongated

148 Annals of the South African Museum.

types. The Thirstland here includes Namaqualand, while part of the
previous Thirstland now becomes the Midlands area. I do not
differentiate between the terms “region” and “‘area”’ as was done in
the previous section, as I do not regard the difference as important
here. The areas are generally wide enough to be regarded as regions,
and any further subdivision can wait until more material accumulates.

THE MATERIAL.

The columns given below represent length, width, height, cube,
maximum diameter of mouth and maximum diameter of the centre
of bore. To these may be added the weight and any comments.

Before proceeding to the bulk of the material at our disposal, we
may list a few examples from isolated areas.

MBABANE, SWAZILAND.

MMK 911 . 9-3 8-1 6-1 460 3-8 2-7 Schistose rock.
8-5 6-9 58 340 30 2-6 Rolled. Igneous
rock.
KURUMAN-GROENWATER RIVERS. KURUMAN.
MMK 961 . 142 12:9 63 1155 4:5 2-5 Fine white Ss.
GRIQUATOWN.
MMK 165. 142 11:3 7:0 +9125) 54-04 1322

The larger Mbabane example agrees well with the second Transvaal
type. The Kuruman and Griquatown specimens form members of a
single group that may be related to Natal and Cape Peninsula types,
but might just as easily belong to the MaNtatisi migration, to be
discussed later.

TRANSVAAL.

Material from the Transvaal proper and from the Upper Vaal has
been grouped together. I have here included seven examples that I
originally listed as Bantu, as they fit more clearly into the present
series. Only one of these was actually found with a Bantu specimen,
from Ohrigstad.

MEssina.*
HAC 1248 . 16:5 13:8 5:0 1140 45 3-6 #£=1:5 kg. Weathered
schist.
W247 el 5-3) 2:0: 3°5 645 34 14 .1]-2kg. Basicschist.
(1249 . 13-0 9-4 4:5 550 3-5 1:8 0O-8kg. Quartzite.)
(1250 . 12:2 9-7 4-0 475 3:8 20 0-57 kg. Yellow

schist.)

* These four specimens formed part of the original Geological Museum Collection.
The first pair are known to be from Messina district, twelve miles from the Limpopo

The Bored Stones of South Africa.

OHRIGSTAD.
PEM 615/J 11-8 10-2
ROOSSENEKAL.
WUA 1155 15:6 12:0
Univ. Stell. 16-2 12-2
LYDENBURG.
WUA 313 10-4 7:9
A/84 . 6:5 5-5
SANNIESHOF, LICHTENBURG.
ASJ 29/42 11:7 10-3
GREENLANDS, VREDEFORT.
ASJ 8/41 riesaie7(6l| 6:3
HEILBRON.
ELM 36c 11:7 10-0
TAUNGS.
AMG 1890 11:3. 10-0
SHEPPARD ISLAND.
ASJ 7/35/48 . 9-7 8-0
Averages: i. 16:5 13-8
Mi. lode 2-1
ie 12-0 9-9
iv. 10-0 8:0
v. 6:8 5:9

4-2

ce
bo >

3:9

6-0

7:0

9-3

5:0
6-4
5-4
au
4-]

505

1600
1425

500
145

820

175

700

790

720

1140
1225
640
610
160

3:2

oo bo
QW D>

2-4

3°5

35)

3-2

1-9

2-0
2-2
1-2
2-0
3:3

2:9

149

Possibly Bantu, sub-
mitted to P.E.
Mus. with a Bantu
example.

1:15kg. Ironstone.
0-23 kg.

Wonderstone.
1 foot deep. Un-

associated.

1-03 kg.

Under 3 feet of
gravel.

It seems clear that this small series is by no means representative
of a prehistoric population of normal density. Either these people
merely passed swiftly through this area, or a vast work still remains

to be done here.

SOUTHERN ORANGE FREE STATE.

WINDSORTON.

SAM 1829 13-0

RIETPAN.
MMK 684 . = 6:3

DOUGLAS.
SAM 06 5 Ors}

PHILIPPOLIS.
JNSS) = 5 . £89

River.

11-4

5-4

5:0

7-0

9-3

3°6

4-5

4-3

1280

122

142

270

6-4

2°8

3°6

3°6

3°8

0:5

2:5

2-4.

Ss.

From the entry Dr. Harger presumes that 1249 and 1250 are from the

same part. The provenance is therefore uncertain, but we may include them here
as products of the same geological excursion to the Northern Transvaal.

150 Annals of the South African Museum.

KOFFIEFONTEIN.
MMK 1412 . 13:8 11-5 305 530 3:2 1:8
TpsHOM ee) Neri) 6-9 6:0 320) ssD) 92-0 Sst
WUA 1100/C 8-5 7-6 4:9 320. 38 1-7 O-ATike:
MMK 1657 6-1 5-1 3°8 118 35 1:8 Ss. Flared bore.
1432 52 (46) 42 100 28 12 £Ss. Damaged.
STRYDPOORT, JAKOBSDAL.
WUA 975 . 80 6:8 6-0 430 39 2:2 0:42 kg.

AVALON.
ASJ 9/38/10 . 9-0 6-1 7:3 400 46 2:8 Two opposed faces
flattened. Eccen-
tric hole. Smith-

field Bassociations.

MAFETENG.
SAM 3183 . 12:9 10:6 4 LOLOD a4 2 ute Started.
BuTHABUTHA.
SAM 1758 7-7 6:8 4-] O15, , 4:05 21 Shale.
Averages: i. 13:2 11-2 6-7 950

i. 369 6:9 5-5 350

Wi. aes 6:8 5:4 290

iv. 6-0 5:0 4:5 isp

There is little agreement between these averages and those above,
but the second Swaziland example fits type ui, though no deduction
can be made, on account of distance. Types 1 and i seem to form
a pair.

THIRSTLAND.
PRIESKA.

Sharples5 . 81 6-6 (1-5) as 14 0:6 Micaceous shale.
Deep radial scor-
ing by rodent.
Anomalous.

GROOTENAUTE, PRIESKA.

SAM 5028 . 12:0 10:5 6:7 850 5:3 2:6

BUSHMANLAND.

SAM Bain . 11:7 9-0 4:7 495 4:0 3:2 Hammer.

STEINKOPF, NAMAQUALAND.

NMP 2342/1. 11:5 10-0 7-6 875 47 29 1-5kg.

2349) 2 9-1 7-6 7-2 500 50 2:3 40:76 kg.
2586 . 10-2 6:5 7:0 465 45 2-6 0-65 kg.
2580 . 8-4 6:7 7-5 420 3:8 2:2 0-69 kg. Oval hole,
3-1 x 2:2 cm.
2342/A. 7:8 6-2 5:7 275 34 18 £0-47 kg.
BRITSTOWN.
Heese Coll. 20 11-3 9-7 6-9 755. 35 2-1 | 1:16) )kes@mnder

Hammer.
16 11:3 #£210-4 6:2 HERO RD) Pep 1:05 kg. Quartzite.
24 10:0 7:3 4] 300 36 2-1 0-45 kg. Pebble.

The Bored Stones of South Africa. 151

CARNARVON.
AMG 689 a) hse 9-2 8:3 920 62 42 Two sides flattened
by hammering.
Elliptical hole.
Striated from top
to bottom.
SUTHERLAND.
Drennan Coll. 8-7 7:2 5-0 315 3:3 .. Ss. pebble.
BEAUFORT WEST.
MMK 1103. . 4:7 3°9 2-0 edt es dal Anomalous.

The Beaufort West example is anomalous, and is most probably
a circular type (the bead type so typical of the area) made from an
elongated stone. The anomalous types are not employed in taking
averages.

Averages: i. 11:5 10-2 6-9 810
un. 11-8 9-1 6:3 708
mu, 10-1 6-9 55 390
iv. 8-5 6-9 6-3 380

These averages immediately suggest the route by which Namaqualand
was reached, through the dry country to the west, or along the Orange
to Steinkopf.

NATAL.

Umautotr Bracu, DURBAN.
A. G. Wills 4. 10-2 8-1 4-8 400 3-1 .. Gritty Ss.
WEENEN.
Wore. 2824 «12-3 9:3 8-2 925 5:7 2-4

908 11-9 9-6 5:6 645 5-1 2-1
KRANSKOP-GREYTOWN ROAD.
SET: . 113 9-5 4:3 455 3-1 08
RICHMOND.
SEI2 . ¢ bb, LZ 66 1145 46 2-1
*“NATAL.”’
NMP — . 140 12-4 76 1320 55 1:5
Averages: i. 14:7 11:8 74 1230

ii. 11-8 9-5 6:0 675

The first group is unsatisfactory, but straddles the scatter of the
second group in the south-eastern area almost exactly. This may
be important in discussing migration.

MIDLANDS.

This area is essentially an inland area of migration, and has been
chosen to represent the variation within a single type. The type

152 Annals of the South African Museum.

(Nucleus 4) is reflected in slightly differing averages in surrounding
areas to the east and south, and this Midlands area overlaps the
South-Eastern and the Southern Forests Regions to some extent.

VENTERSTAD.

Univ. Stell. . 11:55 10:0 (43) 500 38 2:8 ##Both faces worn.
Soapstone.

STEYNSBURG.

ASJ 61/386 . 10-9 9-8 5:9 630 4:2 2-7. Shale.

Lapy FRERE.

ELM 42 . SLI-b 9-7 5:0 560 3:5 2-2 1:02kg. Grinder.

Huvux.ty, STUTTERHEIM.

ELM 76 ~ 105 9-0 4:5 435 40 22 £0°68 kg.

GRAAFF REINET A.
AMG 2448 . 10-9 9-5 71 735 3:6 1:8 Hammer. One side
flattened.
ABERDEEN.

ELM 72 « Abb 9-8 4-5 505 3: 2-2

oO

Kirkwoop, SuNDAYS RIVER.
Mossop Coll... 11:9 10:6 6:6 835 45 3-0

ScHOENMAKERS, Port ELIZABETH.
PEM 615/L . 11:3 10:3 7:3 840 4:7 2:7 1:02 kg.

MatmtLtanp Mines Farm.
PEM 615/A . 10:8 9-4 5:3 540 48 21 0-67 kg.

FERNDALE, HUMANSDORP.
ELM 80 - 11:0 #£10-0 5-0 550 35 22 08 kg.
Average: 11:2 10-0 5'8 640

The limits of variation from the average are seen to be: Length,
0:7 cm.; Width, 1:0 cm.; Height, 1-5 cm.

Note the slight change in width, and the wide range in height, 4-3
to 7-3cm. Analogous types can be followed at Alicedale, Cambridge,
etc., with a wider variation in some dimensions. They have been
included under other regions, but prove the presence of a single
widespread culturaltype. This suggests that any attempt to associate
and isolate this type definitively must be undertaken in the Hastern
Province of the Cape of Good Hope. A careful listing of associable
material, including paintings, etc., should be of very great help to us
in the future.

SouTH-EASTERN REGION.
ROSMEAD.
PEM 614/F . 14:2 12-3 6-9 1205) 94-2 1-7 1-65 kg.

STRATTON, QUEENSTOWN.
ELM 91 5. 16:0 9-0 70 1010 45 2:2 2-04 kg. Anomalous.

The Bored Stones of South Africa. 153

THORNEYCROFT, QUEENSTOWN.

ELM 92 - 150 13:2 6-5
74 » 13856 115 6-0
Harry VaLuey, CATHCART.
ELM 95 . 162 143 75
BREAKFASTVLEI, KINGWILLIAMSTOWN.
ELM — . 145 12-0 5-5
FERNDALE, KomGHa.
ELM 105 » 160 137 6-5
39 = 7-0 5:5 4-7

BasHEE RIVER.
TMP 7538 . 95 5-1 6-9

CAMBRIDGE, East Lonpon.
ELM 35 A. . 18:0 11-8 4-0

35B. . 11:5 9-5 4-0

East Lonpon.

MMK 2805 . 15:0 11:3 4-0
ASJ — » 120 9-3 5-0
ELM 46 2. &7 6:8 4-5

— . 60 5-0 35

Fort Bravurort.
ELM 48 .. FO 6-0 3-5

BEDFORD.
PEM 614/E . 7:5 6-7 4:3
GRAHAMSTOWN.
AMG 2980... 15: 13-6 6-5
. 13-7 12:0 8-0

UNIONDALE, ALBANY.

AMG — cm fe eee 7:2 4:0
ALICEDALE.
PEM 614/Y . 149 12:3 12-4
mee 8 3) 9-7 6-9
AMG 1643 . 11:5 8:8 6:8
Mimosa.
SAM — o32T — 2-8 7:5

Kierpoort-CorGa Roap A.
PEM 614/c . 12:7 11-4

Averages: i. 16:0 13-9
ms 14-7 12-5

1290
930

a
oo

5:0

n=
tos

2-22 kg.
0-45 kg.

Anomalous.

1-3 kg. Horseshoe.
Anomalous.
0-93 kg. Quarry.

0-34 kg. Fort Gla-
morgan.

0-23 kg. Buffalo
River, West.

0-23 kg. Ss.

0-29 kg.

Unshapely.

Irregular.

2-2 kg. Very thick.
1-06 kg.

Zuurberg lava.

1-01 kg.

154 Annals of the South African Museum.

Note the extraordinary range of thickness in type ii (5:5 to 12-4 cm.).
The Kleinpoort-Coega Road site has been split, as this is obviously
at about the point of junction of the two areas, South-East and
Southern Forests. The Rosmead example comes from some distance
outside this area, but it fits well with the Queenstown, Breakfastvlei,
Komgha, and East London specimens in this group. It shows no
relationship to the Midlands type. )

SOUTHERN FORESTS.

KLEINPOORT-CoEGA Roap B.

PEM 614/c. . 9-0 78 4-9 505 3:5 2:2 0-8 kg.

TZITZIKAMMA MIDDENS.

PEM 94 183055 9-3 4:5 565 (83 2-3) 0-8kg. Conical bore. $
Anomalous.

Si aac O 5:3 325 46 23 40-43 kg.

PLETTENBERG Bay.

SAM 1660 . 12-4 9-0 1:6 180 3:0 2-0 Shaledisc. Anoma-
lous.
ROBBERG CAVES.
SAM 2849 . 9-4 7-4. 5°5 385 3:9 1:9 Ss. Midden.
Univ. Stell. . 12-4 9-0 2-6 290 2:8 1:5 “Propellor shaped.”
ASJ 11/41. 8-9 7:4 2°5 165 44 2-1 Elliptical hole.
KNYSNA.
Robinson Coll. 15:3 12:3 fl) MIG) aka sai Coarse Ss. Erratic.
SAM Bain . (12-5) 10-2 8-1 1035 4:2 2-4 Coarse Ss. Broken.
One face flattened
before boring.
UCT — ot ee Oat: 6-0 710 4:0 2-6 Coarse Ss.
Robinson Coll. 12-5 9-4 5:8 680 3-7 2:0 Sandy shale.
11-1 9-6 4-2 450 3:9 2-8 Sandy shale.
SAM 4895 11-0 8:3 4-9 AAD 4026 Sse
Robinson Coll. 9:8 7:0 4:3 295 40 2-1] Ss.
9-1] 7-9 3-2 23008 2-9 7-8) ass
Sharples 3/ 9-9 5:8 4-] 230 2:5 1:9 Anomalous.
DRIPPING WELL, KNYSNA.
- Robinson Coll. 11:3 10:0 7-2 815 4:6 Surface quartzite.
WILDERNESS.
UCT —/43 . 12:0 10-1 (6:0) 730 4:7 2-9 Split. Witteberg
Ss. Midden.
OUDTSHOORN.

Pocock Coll. . 11-7 7:7 6:3 570 «63:6 °° 2-1

GRAAFF REINET B.
F. Malan Coll. 12:1 9-0 5:4 600 42 2:0 Grinder. Ss.

MossreL Bay.

HHC (Black). 11:

15 430
SAM — 5) bis,

280

Se

©
>
Toon)

2-4 TMS pebble.

Corer
bo =

The Bored Stones of South Africa. 155

Doorn River, BARRYDALE.
SAM 173 . 12-4 9-6 6-5 775 55 2-4 Compare South Coast,

average i.
Averages: i. 15:3 12:3 75 1410
i. Lis 9-9 5-7 670
ii. 12:3 9-2 5-1 585
iv. 11:3 8-0 5:6 510
Wel SHO 7-4 4-4 312

Types 1 and ii are obviously the commencement of ..a division in
the Midlands type (11-2 x 10-0 x 5-8 cm.). We will watch the larger
of the two develop further in the South Coast area. Note the inclusion
of Barrydale in the present area. Type v appears to be a consistently
midden variety.

SoutH Coast AREA.

BRAaNDWacuHT, Mosset Bay.
SAM . 12:4 10-3 5:5 700 = 55

MELKHOUTBOS, ALBERTINIA.
HHC 1056 . 10:5 9-0 7:8 345 4:0 2:2 0-52kg. TMS.

BLOEMBOs.
Heese Coll. 23 9-0 7:8 71 500 3:8 2:5 Quartzite pebble.

RIVERSDALE.
Heese Coll. 22 13:0 11-0 6-8 905 43 2-2 Quartz pebble.

25 85 7-4 3:8 140 3:2 2-5. 0-28kg. Ss. pebble.

13. 7-0 6:3 2-4. 106 2-1 1:3 0-17-kg. Quartzite.

CapPE INFANTA.
F. Malan Coll. 11-1 8-4 5-5 515 84-2

CaprE AGULHAS.

ASJ 29/46/2. 12-7 10:8 4-4 550 46 2:6
12:4 10:8 3:9 520 4:2 2:5
Het BosMANSPAD, SWELLENDAM.
SAM 183 oes OS 79 1020 46 2-6
SWELLENDAM.
SAM 181 . 12-4 10-6 4-6 605 3-8 1-9 Compare Barrydale.
10-2 8:6 5:4 475 42 2-0
9-0 7:7 3:5 245 28. 2-0
6-0 4-1 0-5 (12) 2:0 0:5 Anomalous disc.
Montacu.
SAM 5018 . 12:0 10:5 5:6 705 45 1:6 Ss. Grinder.
9-] 7:6 3°6 PANO) Ad EI Sep
CERES.
SAM 50138. 12:5 #8 10-4 3°6 470 45 25 TMS.
HERMANUS.

SNe en Qe) BONN 9.7 090) a9

156 Annals of the South African Museum.

Gorpon’s Bay A.

Univ. Stell. . 12:2 10-8 8-7 1145 3-6
F. Malan Coll. 12:3 10-7 Ter? 9507 75-9 aus
(6:0) 5-1 1-1 af 2:3. 1:2 Oblong slate disc.
Battered.

Jardine Coll.. 7:3 5-1 2:9 108 25 =—18
Averages: i. 12:4 10-6 5:0 645

li. 10:7 8-5 5:4 495

iis 950 7:6 3°6 245

iv. 6-4 4:8 1-5 50

I have divided Gordon’s Bay into two series. The last two examples
fit in well with the third specimen given from Swellendam. It might
have been as appropriate to divide Swellendam, and bring this single
example into the Steenbras-Palmiet Area.

STEENBRAS-PALMIET RIVERS.

VILLIERSDORP.
UCT —/48 . 11:2 10:0 8:3 930 55 2-5 ‘Tertiary Ss.
115 10:0 6:2 715 44 25 ‘TMS.
A. M. de
Villiers . 11:0 9-5 6-1 635 4:2 2-2 ‘TMS pebble.
UCT —/43 . 10°5 9-3 4:6 495 3-7 19 Ss. pebble. Two

faces ground to

an edge.
A. M. de

Villiers . 81 6-8 4:7 315 3:5 2:0 ‘TMS pebble.

KoGeEt Bay MIDDENS.

WUA 256 . 9-9 7:3 5:0 365 44 1:5 £0-51 kg.
HANGELIP.
Univ. Stell. . 13:5 10-2 3°8 525) 4-0 ed Erratic.
HAwWSTON.
MMK 1253. 17-8 11-2 68 1995 4:8 2:4 Ss. Anomalous.
. 10-4 7:3 6-2 AO 422 el Ss. Hammer at

ends.
CALEDON COAST.
Drennan Coll. 13:6 12:0 5:5 900 46 2:4 Ss. Midden, found
with small conical

based pot.
GorpDon’s Bay B.

MMK 420. 15:4 11:8 79 1815 4:7 2:3 Erratic, cf. Knysna.
F. Malan Coll. 11:3 9-4 6-1 648 4-1 sf
Univ Stell... 12:7, 9-2 5:3 615 4:0 1-1 Anomalous.
F. Malan Coll. 9-2 7:3 6-8 455 4-6 .. Uneven grey Ss.
11-0 8-9 4-5 440 2:8 18 ‘TMS.
Univ. Stell. 9-4 7-2 5:8 395) 73:8 6
9-0 7:3 5:3 350 46 2-0
Jardine Coll. 11-4 9-2 3:2 335 4:1 2-2
Univ. Stell. 10°5 9-7 2:7 215) 3-9 el
7:8 6-5 5:3 275 = 2-7 5 Started.
10-1 8-0 3:3 265) 2 leer Merely started.
F. Malan Coll. 8-8 6-8 4-4 265 3-8 2:0 Grey quartz-veined

shale.

The Bored Stones of South Africa. 157

STELLENBOSCH.

PEMIGIZ/K. 13-2 11:6 8:0 labo Sell ad Ske,
Drennan Coll. 13:9 12-4 58 1020 46 2:4 ‘TMS. pebble.
Her 27/15. 10:6 75 6-5 515 3-6 TMS.
Averages: i. 13:5 12-0 6-8 1090

li. 13-1 oT 4-6 570

ite 1-0 9-5 5-2 560

iv. 10-4 7:6 5:3 415

Ve o-2 7:3 6-0 400

Vien Ore 6-7 4-8 285

The anomalous and erratic examples are not included.

SwWARTLAND-SANDVELD.

CLANWILLIAM.

SAM Bain . 11:5 10-0 5:5 635 5°5 , 23

8:0 5:8 4:8 225) 3-4.) 2-8

9-0 8-0 2°8 200 2-2 1-8 Subrectangular.
SALDANHA.

UCT — . 13:6 10:6 5:7 820 3-1 2-6

MMK 1499 71 5-4 5:8 220) 3-459 2:1 Ss. Grinder.
MELKBOSSTRAND.

SAM 5016. 9-4 8-0 3:5 265 4:7 2:5 Malmesbury shale.
WELLINGTON.

SAM — peedd-5 10-0 5-1 585 5:0 2-4 TMS. pebble.

Grinder.

GROENBERG, WELLINGTON.
F. Malan Coll. 13-7 11-0 85 1280 48 1-7 TMS.

NABYGELEGEN, WELLINGTON.
F. Malan Coll. 5-1 4:2 1-0 (21) 1:7 0-8 Disc. Anomalous.

GoOEDEHOOP, WELLINGTON.

F. Malan Coll. 10:5 8-9 6-4 590 48 2:0 Erratic. Steenbras-
Palmiet.
7-0 5:7 3°8 T52e roca leS
Averages: i. 13-7. 10:8 (les {L050
me 14-5 10-0 5:3 610
lii. 9-2 8-0 3:2 230
riya iieel 5:6 4-7 200

There is a general resemblance to the Steenbras-Palmiet area (to
which type ii shows a particular resemblance) and also with the
Thirstland and Midlands. Type iresembles a Cape Peninsula type.

CapE PENINSULA.
LAMBERT’S Bay.

SAM — 14-6 12-5 3°5 640 3:8
BLOEMBOS, DARLING.
SAM 784 ey 12-0 10:0 5-0 600 6:0 2-1 ‘TMS. pebble.

Wilton.

158

MALMESBURY.
NMP 2599 14:5 11:0
Batn’s Kioor.
NMP 2581 13:0 11:0
WOLSELEY.
PEM 615/E . 12-2 9-]
VLOTTENBERG.
UCT 44/4 14-2 11-0
KuipKop, Parow.
SAM 5011 13:7. 10-2
5012 12-8 11-1
5011

11-8 9-3

ORANJEZICHT, CAPE Town.

SAM 4999 . 8-7 6:4
Hout Bay MIDDENS.
SAM 4777 1255 10-2
17-2 8612-5
4750 13:9 10:5
4777 13:0 11:3
UCT 27/18 16:0 10°5
SAM 4777... 9-0 75
NooRDHOEK MIDDENS.
Mossop Coll.. 10-4 (8-6)
KoMMETJE MIDDENS.
SAM 3831 10:6 9:0
Cape PoInt.
SAM Martin. 14:0 12-4
SIMONSTOWN.
SAM Martin. 146 12:0
Fiso Hogrk.
J.B.Gordon. 10:5 8-8
SAM Peers 10-2 7:3
8:3 7:0

Smit’s Farm, Fish Honk.

Mossop Coll. . 10-0 7:5
Sr. JAMES.
UCT 44/5 . 8&4 7:6

5:6

9-2

4-4

6-7

4-7
4-5
4-5

3°6

Se be iS ©
mo OwWId-E

tS

7:8

6-4

4-8

5:3

950

1315

490

1045

655
640
495

200

1025
975
880
655
300

196

700

610

835

1155

425
255
198

330

340

3°6

7:8

3°8

4-2

4:3
4-4
4-5

3:2

bt He OU OU Or
bo CO HW OL

ee

4-0

4-3

3'8

3°8

les)

1-9

3

1-6
2-3

2-2

2-4

2-3

Annals of the South African Museum.

1-7 kg.

2°13 kg. :

0-69 kg.

Ss.

TMS. pebble.
Burial.

TMS. pebble.
Burial.

TMS. pebble.
Burial.

Ss. pebble. Grinder.

TMS. pebble.

Shale pebble.

TMS. pebble.

TMS. pebble.

Slate disc. Anoma-
lous.

TMS pebble.

Purple Ss.

TMS. pebble.

Greenish shale.

TMS.
Shale.
Shale.

Grinder.
Waterworn.

5 feet deep. Grinder.

Ss. Used as hammer,
Station House
Garden.

The Bored Stones of South Africa. 159

CaPE FLATS.

SAM 605 -. 165 145 10:0 2395 3-5 2:5 Outsized. Anoma-
lous.
Lister 15:8 12:0 3°8 720 48 1-9
12438 . 12:0 10:0 6:0 720 36 1:5
AMG 2419 . 12:7 9-0 5:7 650 48 2:0 Pointedstone. Oval
hole.
SAM Bain. 146 (11-4) 3:5 585 3:5 3-1 Laterally flattened.
402 ae O-o 9-2 5:7 a0) aie). Dopo
Bain . 10:5 9-1 5-1 490 35 2-1
9-5 7-5 3:7 265 4:0 2-0
10-0 8:8 2-0 IEray 9 PAIL le 7/
PHILIPPI.
Arends Coll. . 13:0 11-2 6-8 990 4:2 2-5 12 feet deep in sand.
Oval hole.
PRINCESS VLEI.
Mossop Coll.. 14:5 10:3 4-6 685 4:0 2-2 Skew bored.
Victoria HospitaL, PINELANDS.
ASJ — . 10-0 8:8 2-2 195% 3:5 2-5

NEWLANDS.
SAM Camp . 13:3 10-4 4-0 550 4:0 1:6 Shale pebble.

CONSTANTIA.
Jager Coll. . 12-0 95 (7:6) 865 50 3-0
ToKal.
SAM — . 140 10-0 3:0 420 45 2-8 Irregular.
Averages: i. 16:5 12-2 4-2 850

iis. Lieto) TTY) 4-8 535

aie ZO) 2, 6-1 675

Ves Ou eLO:3 4-4 638

Mf, dap 9-6 5-9 565

wig 1iOr3} 8-9 4:8 450

wills ) Oey 74 3°6 260

Vili. 8-5 7-0 4-] 245

Only type 1 seems to have a peculiar distribution, z.e. Lambert’s
Bay, Malmesbury, Cape Point, Simonstown, and a single example
from the Cape Flats. Type vill may be a divergent from type vil.
The St. James example, presented by A. R. H. Walker, is nearly
circular, and may be a slightly elongated variant of the circular type,
CH. xi.

Comparative Distribution of Elongated and Circular Forms.

Material is listed opposite the areas used above for the elongated
forms. These are not directly comparable with the regions and areas
used for circular forms, but the areas given here for the circular types
are those that most nearly coincide. The first column of figures gives
the total of elongated specimens; the central column shows the

160 Annals of the South African Museum.

distribution of circular forms in component and comparable areas,
while the final figure gives the total. A few additional circular forms

have been submitted too late for inclusion in the list. No new
elongated forms are known.

Area. Elongated. Circular by Areas. Circular.
Kuruman-Groenwater 2 Uzi 17
Transvaal-Swaziland 15 9+29+18+17+26+11+420 130
Southern Free State . 13 59+18+23+33+17+5+4+18 1B
South-East Free State . 0 45 45
Upington-Hartbees 0 14 14
Bokkeveld ? 0) 6 6
Thirstland 14 24+154+25+8 12
Natal . 6 23+9 a2
Midlands 10 7+9+22 38
North Transkei 0 2 2
South-East region 25 19+13+14+4+13+421 80
Southern Forests . 23 5+114+3+4+19+82+10 130
South Coast . 22 44+418 62
Gouritz-Karoo ) 13 13
Steenbras-Palmiet 25 24+10+41 75
Swartland-Sandveld ll 27+5 32
Cape Peninsula 39 17 +53 70

TOTALS . 205 991
MIGRATIONS.

It will be remembered that, in discussing the Transvaal circular
types, 1t was shown that there are seven prototypes to be discerned
in the Transvaal and the Vet-Marico area. These persist and can
still be recognised in Natal and the Vaal River areas, but in the Free
State the picture becomes blurred. There seems to have been a
stoppage, a period of lateral movements that gave rise to new forms,
with little or no attempt to cross the Thirstland to the south of the
Orange. It was therefore found convenient for us to regard those
types that eventually crossed the Orange southward as deuterotypes,
though these were not clearly isolated.

In our study of the elongated series the picture is very different,
and to account for those differences we shall have to assume that the
cultural migrations that first brought the elongated types into the
Union were different in origin from those that brought the circular
types, whatever may have been the eventual results of mixture.

In the elongated series there are (oddly enough) seven prototypes.
These are arrived at by a method somewhat different from that

The Bored Stones of South Africa. 161

employed to develop the Transvaal prototypes for the circular series,
and I have therefore called them “nuclei.” These may not all have
entered the Union at the same point or within the same area, but our
knowledge is limited by present lack of evidence.

The first point of interest is that these nuclei persist over a far
wider area than the seven prototypes. As will be seen when we
discuss the ““Age and Area” hypothesis, there can only be one
explanation for this persistence; that spread was faster and was not
hampered by any period of delay in the Free State, nor by deferment
of the crossing of the Thirstland southward. This itself needs ex-
planation. The earlier spread was hampered by the dangers of the
Thirstland and the unknown peril of pressing farther and farther into
what might prove to be an increasingly arid area. In addition, it
would be faced by hostile remnants of the Middle Stone Age peoples
to the south. Some period of consolidation and of reconnaissance
was essential. In contrast, any later spread by a friendly incoming
culture could have followed freely without these inhibitions, probably
by what had become a well-marked and well-known route. This
gives us reason to presume that the circular types are the earlier,
and the elongated are later. Slhght contributory evidence les in
the fact that a greater proportion of circular specimens he buried
at depth in different parts of the Union. To the north, in the Congo,
both types of bored stones seem to be found persistently at depth,
and to precede the later Neolithic period, according to Dr. F. Cabu,
and no differentiation has yet been attempted. M. Janmart finds
the bored stones of N.E. Angola under the youngest gravels on
bedrock.

To take the matter a highly speculative step farther, it may eventu-
ally be possible to suggest that the circular types were the tools of
the vague physico-cultural-linguistic ““Bushman” group, while the
elongated types may similarly be attributable to the equally vague
cultural-physical-linguistic “‘Hottentots.” Whether or not they
will ever prove to be associable with Smithfield-bearing and with
Wilton-bearing cultural groups remains to be seen. This is here put
forward for what it may eventually prove to be worth. Much still
remains to be said for the logical steps that have led me towards
making the more legitimate suggestions set out above.

THE SEVEN NUCLEI.

We exclude the South Coast, Steenbras-Palmiet, Swartland, and
Cape Peninsula areas for the present, as these are the farthest from
VOL. XXXVII, PART I. 11

169 Annals of the South African Museum.

any point of entry into the Union, and are therefore to be regarded
as later and derived forms. Our discussion will for the moment be
confined to the remaining areas, east of Mossel Bay, and north of
about 33° S.

If we take all the averages that we have already developed, exclud-
ing anomalous types, for this part of the Union we find that there are

seven nuclei of averages on our graph. These nuclei are made up
as follows :—

q, 16:5 13-8 5-0 1140 #£xTransvaal.
16:0 13-9 70 1560 South-East.
Nucleus]. 16-2 13-8 6-0 1350
2: 14-2 12:1 6-6 1140 Molopo-Groenwater.
15:7 Le 6-4 1225 ‘Transvaal.
14:7 11-8 7:4 1230 8=6Natal.
15-3 12:5 75 1395 South-East.
15-3 12:3 7-5 1410 Southern Forests.

Nucleus\2). 15:0 © 12:1 71 1280

3. 13222) le? 6-7 950 Orange Free State.
1S le 6-4 910 South-East.

Nucleus 3. 3-6, LS 6:5 930

4. 12:0 9-9 5-4 640 Transvaal.
11:8 9:5 6:0 675 Natal.
Lay 1022 6-9 810 Thirstland.
11:5 9-] 6:3 710 #£Thirstland.
11-2 10:0 5:8 640 Midlands.
11-6 9-3 5:7 615 South-East.
11-8 9-9 5:7 670 Southern Forests.
12°3 9-2 5-1 585 Southern Forests.
Nucleus 4) ai” toe Vs'8 670
5 10:0 8:0 7:7 610 Transvaal.
10-1 6-9 5°5 390 ~=6©Thirstland.
11:3 8-0 6-4 510 Southern Forests.
Nucleus 5 . 10:5 7:6 6:3 505
6. 8-5 6:9 58 340 Swaziland.
8-9 6-9 5:5 350 Orange Free State.
7:8 6-8 5-4 290 Orange Free State.
8-5 6:9 6:3 380 Thirstland.
8:2 6-9 4:3 240 South-East.
9-0 7-4 4-4. 310 Southern Forests.
Nucleus 6 . 8:5 7:0 5:2 315
7. 6:8 5:9 4-] 160 Transvaal.
6-0 5:0 4-5 135 Orange Free State.
6-7 5:5 3°9 145 South-East.
Nucleus 7 . 6:5 5:5 4-2 150

The Bored Stones of South Africa. 163

These nuclei are not true averages, as I have not loaded the figures
correctly, but the error is not more than a few millimetres in each
case. This table gives the distribution by areas, the nucleus figures,
and the variation from the averages.

Unhappily it is impossible for us to say how many different objects
it was intended to make. We might suggest that the makers were
aiming at digging-stick stones, knobkerrie heads, and dagga-pipes;
but they might just as easily have intended a variety of objects outside
our ken, such as traps for small birds, stands for the centre-poles of
huts, etc. Plotting the distribution of all individual types from the
areas cited above, it will be found that there are three marked aggrega-
tions, coinciding with nuclei 2, 4, and 6. This suggests that three
objects were intended, and leaves us with two possibilities: either
nuclei 3, 5, and 7 are divergents, or we are considering two separate
cultural spreads, each with three elements. Nuclei 1 and 2 are
certainly related to each other in the Transvaal.

This coincidence of types over the greater part of the Union supplies
us with proof that our groups are a valid and reasonable source for
deductions as to migration and spread. Further proof can be obtained
by the application of these types to geographical distribution with
the intention of discovering localised types and migrational routes,
reasonably related to terrain and environment. I have therefore
plotted the distribution of those examples most closely approximating
to the average in each area, with curious and unexpected results.

For the geographical plotting I employ a grid of one inch to the
degree, using graph paper. This simplifies plotting, as one-tenth of
an inch represents 6’ of degree, simplifying the making of distribution
maps. The distortion resulting from the use of a square grid is
negligible in the Union, and occurs at points where it 1s most needed,
at the southern end of the Union where routes are congested.

In making deductions from distribution maps various points have
to be constantly remembered, and we may list those that affect the
present material.

1. Our knowledge depends to some extent on the original distribu-
tion of the prehistoric makers, and to some extent on the actual
distribution of people sufficiently intelligent to recognise, collect,
and submit specimens. Happily the two distributions roughly
coincide, except in the sparsely populated Northern Transvaal and
in the native territories.

2. Too little is known of the elongated series from the Transvaal,
Natal, Bechuanaland, the Transkei and Basutoland to yield evidence

164 Annals of the South African Museum.

of value. Our deductions will be of the greatest value in the southern
Free State, Griqualand West, and the Cape Province.

3. We can make three fairly safe deductions concerning the direc-
tion of migrations. South of 28°S8. migration seems to have been
radial, west, south-west or south-east from a point west of the Basuto-
land massifs, directed towards the coast-line. It is just possible
that the elongated types (in contrast to the circular) may in some
instances have passed through Natal to the South-Eastern area.
Finally, once the coast was reached at any point migration might
have been in either or in both possible directions.

4. Owing to the factor of “serial copying” we must expect to find
a change in type average, or a doubling of types, as we go farther from
the sourch of each type. This effect becomes marked when the
Southern Forests are passed.

5. Areas immediately west of the Southern Forests seem to have
been fed through that region, and the type averages at the Cape and
the areas round about are somewhat different from those elsewhere.

6. The Southern Forests area thus seems to have been a point of
sojourn and consolidation, a focus where types had time to multiply
before passing farther west.

7. In studying migrational routes it may be possible from our
knowledge of terrain to suggest the route followed along a particular
stream, or through a gap in the hills.

8. Anomalous types (so marked on the list) have been tried out,
but their distribution is of no value whatsoever. We must regard
these as the products of individuality.

THE RovutesEs.*

Nucleus 1. 16-2 x 13-8 x 6:0 cm.

Confined to Messina in the northern Transvaal; then recurring at
Happy Valley, Cathcart, and at Queenstown and Grahamstown. Two
or three examples from the Cape show a similar type. The gaps of a
thousand and six hundred miles respectively preclude any deductions.

Nucleus 2. 15:0 x 12:1 x 7-1 cm.

Examples from Messina, Roossenekal, Richmond, and ‘“*Natal’’
show a distribution towards Natal. A clearer route can be plotted as
follows: Kuruman, Griquatown, Koffiefontein, Rosmead, Queens-
town, Komgha, Kingwilliamstown, Hast London, Grahamstown,

Alicedale.

* Maps of routes are given on pages 171 and 172.

The Bored Stones of South Africa. 165

Between Alicedale and Knysna there is a gap, and a slightly smaller
type (13-9 x 11-2 x 6-5) runs to Cape Hangklip, Gordon’s Bay, Simons-
town, Cape Point, and to Malmesbury and Clanwilliam. It is highly
probable that this southern variant is a derivative of nucleus 3.

Nucleus 3. 13°3x11-5x6-5 cm.

No examples are known north of 28° 20’ S. or east of 27° 30’ E.
The distribution runs as follows: Windsorton, Koffiefontein,
Mafeteng, Queenstown, Fort Beaufort, Grahamstown, Mimosa,
Coega, Assegaibos, Tzitzikamma, Mossel Bay, Riversdale, Barrydale,
Swellendam, Agulhas, Hermanus, Hawston, Caledon Coast, Hangklip,
Gordon’s Bay, Vlottenberg, Stellenbosch, and the Cape Peninsula.
From about Stellenbosch there is a branch that continues to
Wellington, Bain’s Kloof, Goudini, Saldanha Bay. Towards the Cape
the type changes to a slightly smaller average (12-6 x 10-9 x 5-8 cm.).

Nucleus 4. 11-7x9:6x5-8 em.

‘ This seems to be the most widespread and persistent type. We

can really learn something from this distribution. There are some
marked variants, but only those obviously related to local averages
are employed here. Isolated examples from Natal suggest a dis-
tribution there, but the main spread is complicated and interesting.

| Ohrigstad. |
Lichtenburg. Heilbron.
Taungs. Weenen.
Britstown. Venterstad. LSS
Carnarvon. | Steynsburg. |
me fe ee Graaff Reinet. Lady Frere.
ees Alicedale. Stutterheim.
bee P. Elizabeth. East London.
Mossel Bay. ae eS
Albertinia. rae y Apes
Bloembos. a
Swellendam.
C. Infanta.
Hermanus.
Gordon’s Bay.
Wellington.
Clanwilliam.

In addition there is a local distribution in the Steenbras-Palmiet
area and the Cape Peninsula, where the average changes to 11-4
x 9-5 x 5-0 cm., related to Hermanus and Gordon’s Bay.

Nucleus 5.

10-5 x 7-6 x 6:3 em.

This is not of great value, but shows an additional route to
Namaqualand. It runs from Lydenburg, Sheppard Island, Britstown

166 Annals of the South African Museum.

to Steinkopf. A separate route (possibly linking with Britstown by
way of Graaff-Reinet and Aberdeen) runs from Knysna to Kogel Bay,
Gordon’s Bay to Fish Hoek. Compare the Umhloti Beach example
from Natal.

Nucleus 6. 8-5x7-0x5-2 cm.

The type occurs at Mbabane, then at Jakobsdal, Koffefontein,
Philippolis, Prieska, and Steinkopf, suggesting another route to
Namaqualand.

The coastal route is so far separate from this, and runs from Hast
London, through Coega, Jeffrey’s Bay, Tzitzikamma, Robberg,
Mossel Bay, Riversdale, Bloembos, Swellendam, Cape Infanta, Kogel
Bay, Gordon’s Bay, Hout Bay, Oranjezicht (Cape Town), and Fish
Hoek, with a branch to Clanwilliam. The presence of an example at
Sutherland suggests that in this instance there may have been a route
crossing from Prieska to the Cape through the Thirstland, thus
reversing all or part of the coastal route given here. For a time the
average approximates to 9-2 x 7-1 x 5-0 cm., with little variation.

Nucleus 7. 6:°5x5:5 x 4-2 cm.

Excluding the discs from Gordon’s Bay, Swellendam, and
Wellington, we can deal with these as two slightly different groups.
The larger series, conforming to the figures given above, occurs at
Lydenburg, Rietpan, Koffiefontein, Douglas, then at Komgha and
Kast London, and finally at Wellington and Saldanha Bay. Slightly
smaller examples (6:1 x5-0x3-0 cm.) occur at Koffiefontein and
Victoria West, once again suggesting the possibility of a Thirstland
route. Further knowledge of the western province of the Cape would
help our knowledge very considerably.

DEDUCTIONS.

In making deductions from our figures I intend to use the “Age
and Area” hypothesis. This approach was first developed as a
botanical tool, but has been applied to ethnology by Wissler in
America, and by others. The fundamental premises, which I have
augmented, may be expressed thus:

i. The area showing the greatest number of basic varieties of
comparable objects was most probably the original centre
of dispersal.

il. The wider the spread of a recognisable type, the older the move-

The Bored Stones of South Africa. 167

ment is likely to have been. Conversely, the more confined
the distribution, the younger the spread.

i. The farther we go from the original centre of dispersal, the

greater the divergence will be.

iv. The slower a spread, the greater will be the divergence and

creation of new types; the smaller the variation in types over
a given distance, the swifter will the spread have been.

v. A limited number of types will pass through a filter or barrier,
to continue as derived variants on the far side. New and
later varieties past the barrier will be unrelated to variants on
the proximal side.

The first and third would appear contradictory, but the third is

generally governed by the fifth hypothesis.

Though our knowledge of the bulk of Africa is scanty, we may
accept the assumption that the original source of both circular and
elongated types lay to the north of the Limpopo River, outside our
field. There the incidence of these two general forms is much the
same, circular and elongated forms appearing in approximately equal
numbers. This area falls outside the scope of the present paper, and
little need be hazarded on the relative age there of the two types
discussed.

Once the Union was reached, seven “‘circular” prototypes survived
the Limpopo barrier or filter, while only five “elongated” nuclei
appear in the Transvaal, eventually increasing to seven. The number
of circular examples in the Transvaal is vastly greater than the
number of elongated. We therefore deduce that the makers of circular
types persisted longer in the Transvaal than did the users of elongated
forms.

The same must be said in the southern Free State and the Vaal
basin. This is the area of greatest divergence of circular types. We
also have the accumulated evidence of Prof. C. van Riet Lowe to prove
that it is extraordinarily rich in material, especially Smithfield B
sites associable with the circular series. We must therefore accept
this as an area where a period of long sojourn provided variations
and new types, a secondary centre of distribution for the circular
forms. The same cannot be said of elongated types, which passed
through the Free State relatively swiftly, leaving behind few examples
and only four types.

Excluding the Western Bypass and Natal, both dependent upon
the Transvaal or the Vaal for their origins, only a few circular types
passed the Thirstland filter or barrier by routes we have already dis-

168 Annals of the South African Museum.

cussed. The Orange River provided a route to the west, the Ongers
and Seekoei Rivers gave a roadway towards the fertile south, while
other routes led through the Thornveld to the south-east coast.
Movement due east was always debarred by the Basutoland massifs,
though elements seem to have filtered through to the northern
Transkei. South of the Orange River the circular types multiplied —
once again, but it is of great importance to note that there was no
such increase in elongated forms until the Southern Forests were
reached.

All this leads us to presume that the elongated types were later
than the circular, and that they spread more swiftly until the Knysna
area was reached. Low density of distribution and limited variation,
up to this latter point, all lead to these deductions.

We may attempt to apply these principles to the spread of individual
elongated nuclei, though we have little reason to suppose that there
were more than two cultural groups involved. However, for what
such deductions may eventually prove to be worth in the light of
future research, here are the probabilities.

Nucleus 4. Most widespread, probably the oldest.

6. Slightly less widespread.

2. Merely spread to the coast. The Cape Region variant
may be an independent and convergent development.

5 and 7. Hach is possibly the product of two or even three
independent developments, each relatively late.

3. Less widespread. |

1. Spread slight, probably a restricted variant of type 2.

Such evidence is deductive and circumstantial, and before we may
hope to presume on such a suggested order of movement very con-
siderable work will have to be done on the stratigraphy of the bored
stone, and its association with other cultural material, such as paint-
ings, petroglyphs, pottery and the like. It is, however, interesting to
note that the probabilities outlined above once again group the
second, fourth, and sixth nuclei together, and the third, fifth, and
seventh.

It will have been observed that there is a general similarity between
the routes taken by the elongated types and some of the routes taken
by the circular. There are two possible explanations; that these two
main forms are coupled in certain instances, and therefore took the
same paths; but the alternative seems preferable, that the elongated
series was the later, and followed tracks already beaten by the earlier
folk. This alone would explain the lack of concentration of elongated

The Bored Stones of South Africa. 169

forms in the Transvaal and Free State. But the possibility of linkage
cannot be overlooked.

So far I have been unable to find any instance in which elongated
and circular types occur together in undisturbed strata, showing either
absolute association or dissociation. At Oakhurst where two types
occurred in different early Wilton graves, both were circular.* At
Keurfontein, Vosburg (where association only occurred in a spring),
two vastly different circular examples occur together with Smithfield
B implements, and so on.

Along the southern coast it has been my general conclusion from
observation that elongated types occur on midden sites with pottery—
that is, with late midden material; while circular examples belong to
older non-pottery-bearing middens. This needs directed research,
our main difficulty being that round forms tend to roll down middens
and sandhills, while shifting dunes re-associate materials in incongruous
patterns.

Careful excavation in untouched cave-midden sites is our only
real hope. Hzact distribution (in depth) of pottery and other elements,
and the exact relationship of whole or broken examples of both general
types of bored stone, should be watched far more carefully than
heretofore.t J. F. Schofield’s attribution of coastal pottery to the
““Hottentots” may eventually yield proof that the elongated types
belong to that complex, whatever we may call it. At Oakhurst the
pottery belonged to a final stage of developed Wilton, long post-
dating the circular bored stones there. No elongated bored stones
occurred.

SuMMARY OF MaIn CONCLUSIONS.

1. As a whole the elongated forms show signs of being later than
the circular types.

2. There is reason to assume that the elongated forms entered the
Union somewhere in the region of Messina, across the Limpopo
River at 30° Hast. long., either together or as two groups (2, 4 and 6,
and 3, 5 and 7—type 1 being a probable derivative of 2).

3. The bearers of the circular types had a Smithfield culture, while
they were held back in the southern Free State. The users of the
elongated forms passed directly to the coast, and reached it at fewer
points than the earlier people.

* Goodwin, Drennan and Schofield, ‘“‘Oakhurst Shelter,’ Trans. Roy. Soc. S. Afr.,
vol. xxv, pp. 251 and 254 (Plates Vc and Vla show graves Vila and IX and IXa

respectively, with bored stones in sitw), 1937.
t Idem, p. 300.

170 Annals of the South African Museum.

4. Little variety occurs in elongated forms until the Southern
Forests area is reached. Movement from the point of entry into
the Union must therefore have been swift as far as Knysna, where
the first period of consolidation took place.

5. Present deductions concerning the order of spread of elongated
types are highly speculative.

6. Future work must differentiate clearly between circular and
elongated types, and the association between each of these and other
cultural elements will accumulate to provide the clue to the true
position of each. Associations with petroglyphs and with paintings
based upon large numbers of accumulated instances are no less
important.

7. The possibility of certain elongated types being linked with
certain circular forms must not be overlooked.

8. There is as yet no reason to suppose that the elongated series
were carried by routes different from those used by the bearers of
circular forms. The main differences lie in the paucity of routes
used by those former, their more restricted spread, and the greater
speed of distribution with the concomitant poverty of variants.

9. There has so far been no case of absolute association of an
elongated type with specific implement types or with pottery, but
both pottery and elongated forms are consistently found together on
the surface of recent deposits along the south coast. This suggests a
‘“ Hottentot” culture.

PART V.—THE BANTU SERIES.

Having discussed the prehistoric material, we may now turn our
attention to the group of bored stones, confined to the Transvaal
and immediately surrounding areas, that can with certainty be
attributed to the Bantu of the early historical period. These may
be dated as belonging to the first years of the last century and earlier.

In the main these people seem to have belonged to the Sotho
group of tribes, to-day covering Bechuanaland, Basutoland, and
parts of the Transvaal and Orange Free State. Their use of these
stones seems originally to have been confined to the eastern Transvaal,
and to have spread from there to the sources of the Molopo and
Kuruman Rivers, and down the Wilge towards the Natal border.

These Bantu were simple farmers, the women practising horti-
culture, the men tending cattle or devoting themselves to the arts of

The Bored Stones of South Africa.

Maps 13 anp 14.—Migrational Routes of elongated types.

71

172 Annals of the South African Museum.

The Bored Stones of South Africa. 173

war. They seem to show no real alliance with the prehistoric users
of bored stones, but one puzzling fact associates the Bantu and
“Hottentots” of the Transvaal. This is the attribution at Mapun-
gubwe * of a series of skulls to ““a Bush-Boskop people, showing
sporadically a few Negro features,” associated with a cultural complex
that is essentially Bantu and includes the use of terracing. Un-
happily there is as yet no link between that offshoot of Zimbabwe
culture and the bored stone.

The Sotho and allied tribes are sedentary, and live in villages near
their fields for the agricultural part of the year. In the Transvaal
these villages do not appear to have been large, and it is only when the
eastern side of the Kalahari is reached that a true town economy,
related to the immediate abundance of water, is practised. During
the remainder of the year these tribes are generally transhumant,
spreading far afield to pasture lands where they live in smaller units.
So far as other customs are concerned, they form exogamous groups
practising polygamy with patrilocal marriage. Their way of life is
easily adapted to accommodate simple domestic slavery with con-
cubinage, and indeed such habits are only now disappearing in the
farther parts of Bechuanaland. The condition of these people 150
years ago was not very different, save that at that time they con-
sistently practised terraced agriculture in the eastern Transvaal,
and in that mountainous area the bored stone is associable with this
peculiarly Mediterranean form of cultivation. Whether the associa-
tion between bored stones and terraced fields is to be found in Southern
Rhodesia at all consistently I do not know.

Filtering slowly down the subcontinent from the general region of
the Great Lakes and the Rift Valley, the southern advance guard of
the eastern branch of the Bantu-speaking peoples crossed the great
divide of the Zambesi valley. Their cattle differed from those used
by the earlier Hottentot migrations, but were quite as vulnerable to
nagana. Hemmed in by the dangerous tsetse-fly country, to-day
Portuguese East Africa, they were forced inland along the northern
bank of the Zambesi, and must have crossed at a point free of fly,
and west of the Victoria Falls.

Judging from cultural evidence, two great groups entered the
Union by the same general highway, the Nguni and the Sotho. They
show marked differences in pottery, hut-form, social and territorial
organisation, and in language, but the fundamental similarities

* Fouché, Mapungubwe (Cambridge, 1937).

174 Annals of the South African Museum.

between the two groups are great.* Both had the same mode of life,
similar concepts of chieftainship, grew the same crops ¢ and herded
the same stock.

The Nguni spread eastward, turning south when they reached
the escarpment leading down to the tsetse stricken lowlands of
the subequatorial east coast. They passed southward, east of the
Basutoland massifs, to the healthy south-east coast, colonising
Zululand, Natal, and the Transkei, and coming into marked contact
with the Hottentot tribes along the Kei River and farther south.
The Sotho people spread down the central plateau to populate the
bulk of the country west of the Basutoland mountains, and continued
as far south as the Orange River. Their main contact with the
Hottentots seems to have occurred here, in the Douglas-Kuruman
area, and not before.

The Bantu seem in all cases to have been the newcomers, the north-
ward element, pressing down upon the earlier inhabitants. In the
course of their movements southward they encountered the indigenous
“Bushman” and “ Hottentot”’ peoples, and these contacts showed
themselves in various ways. The Bushmen were then a thriving
people, simple hunters and collectors, living in rude shelters or under
the overhang of rocky outcrops, and painting on their cave walls.
They were by no means confined to the arid parts of the country.
They depicted the Sotho tribes and their cattle, characterising them
by their darker skins, their national shields and skin-cloaks, often
wearing skin head-dresses with the tail still attached. The Ambo
and Herero tribes of South West Africa (unrelated to either Nguni
or Sotho, but representing a separate Bantu stock) learnt the art of
making ostrich eggshell beads, and the trick of weaving or netting
these into bracelets, necklets and headbands. The Neguni similarly
took over the technique, but seem only to have applied it to the small
trade beads of the east coast. They took over three basic clicks, with
their nasalised, aspirated, voiced and unvoiced forms from their
contacts with the Hottentots, presumably along the Great Kei River,
where the two distinct ethnic groups lived beside each other for some
time.

* See I. Schapera and A. J. H. Goodwin, “‘Work and Wealth,” both in Schapera’s
Bantu-speaking Tribes of South Africa. Routledge, 1937. Also I. Schapera in
his Western Civilization and the Natives of South Africa, ‘“‘The Old Bantu
Culture.” Routledge, 1934.

+ A. J. H. Goodwin, “The origins of certain African food plants,” S.A.J.S.,
xxxvii, 1939.

The Bored Stones of South Africa. 175

In contrast the Sotho tribes seem to have been in closer touch
with the Bushmen. They took over only one click, but adapted the
Bushman mural parietal arts, and translated their own customary
pottery and basketry designs into wall-paintings on their cylindrical
hut walls, using the traditional Bushman earth colours. They appear
to have been less antagonistic to the Bushman than other South
African peoples, for it is important to note that the Bushmen survived
longer in the Sotho territories than they did elsewhere. At Lake
Chrissie, in the Tati area, in the Kalahari and in Basutoland the
Bushmen were permitted to live and survive into the historic period,
though generally as serfs.

These borrowed elements are not found among the Bantu north
of the Limpopo until the Ndebele, Ngoni and Kololo movements,
which shows that they originated in the Union before the end of the
eighteenth century. There is therefore little reason for surprise that
the bored stone and the techniques accompanying it were similarly
adopted in certain areas of the country. The Sotho tribes have
habitually incorporated Bushman or Hottentot girls into their economy,
and these girls would readily have reverted to their own types of
digging-stick, at periods when iron was scarce or unobtainable owing
to the need for the munitions of war.

Where the introduction of the bored stone first took place, and how
widespread the casual use of the stone-boring technique among the
Bantu may have been we shall never know, but the region in which
the Bantu definitely adopted the bored stone as an agricultural
implement is definable. Broadly, the area seems to have been
confined to the highlands north of the wide upper Limpopo Valley,
to the highlands between the upper Limpopo and the upper Vaal
valleys and to the Wilge tributary. The distribution commences at
31° 30’ H., and specimens are most numerous near the eastern limit,
where they are consistently associable with terraced agriculture.
The distribution thins out considerably towards the west, until only
a few scattered examples appear in the central Transvaal, and fewer
still in the western part of Bechuanaland. Applying the “Age and
Area”’ hypothesis, we can deduce that the eastern Transvaal (judging
from the variety of original forms) was the home of this instance of
Bantu borrowing, and that subsequent spread within the Union was
west or south from there.

The Transvaal and surrounding areas are thus the common field
of two distinct cultural distributions. The general prehistoric types
include the Transvaal, which also provides the distributional centre

176 Annals of the South African Museum.

for the Bantu types. Our first difficulty is therefore to differentiate
between Bantu and prehistoric types. Complete differentiation is
impossible, but we can allocate forms with a low percentage error.
I use the word “forms” advisedly; as has already been shown (p. 10),
certain examples in the Vet-Marico Rivers area were almost certainly
conically bored with a metal borer. It will be shown later that there
is reason to believe that many of the Transvaal prehistoric types were
so bored, and that Hottentots, Bantu, and Bushmen were all using
the digging-stick contemporaneously.

This could only mean that the users of the prehistoric types carried
on their traditions of size and form, in some sort of association with
the Bantu metal-users. The users of our prehistoric types were
“Bushman” women, and we can only suppose that these girls were
incorporated into the agricultural system of the Sotho peoples, and
had access to metals for boring their stones. This point will be brought
out more clearly when we discuss the histograms of the bore.

DIFFERENTIATION OF TYPES.

We have various criteria at our disposal for differentiating between
Bantu and prehistoric types. Perhaps the most striking is the
development of histograms and graphs, representing the various
dimensions of specimens found within the Union. The second is the
attribution of stones and groups of specimens to Bantu users, either
by the Bantu themselves, or from their association with known
Bantu settlements and terraced fields. In this latter group are
sometimes found specimens that might fall within the Prehistoric
Series. This is less important in the light of what has been said
above about Sotho-Bushman concubinage. In most cases therefore,
where specimens fit clearly into the prehistoric types they have been
included there, even if found in apparent association with Bantu
sites.

A further check is provided by geographical distribution. By
comparing the Transvaal series with material from the remainder of
the Union it becomes clear that a new cultural element is involved
and can be isolated by statistical methods. In addition, as would
be expected where two populations have inhabited the same area
consecutively with a period of overlap, the implements of the earlier
people show in several instances that they belong to older deposits,
while Bantu types are more exclusively (though by no means com-
pletely) confined to surface or near-surface deposits, and are associated

The Bored Stones of South Africa. Lay

with still recognisable terraced fields and settlements. This latter
source. of evidence is only worthy of consideration in dealing with
fair numbers, as the dry soiland heavy summer rainfall of the Transvaal
area lead to erosion, that in turn produces peculiar reversals of stratifica-
tion, especially where cylindrical or subspherical objects are concerned.

A graphical plotting of the two circular groups, Bantu and pre-
historic, reveals statistical differences. It will be found that the
prehistoric types of most of the Transvaal conform to the seven proto-
types given in Part III above (p. 162), while the Bantu series shows
an almost completely haphazard distribution. This shows that once
the bored stone had been completely assimilated into Bantu culture,
the concept of “‘types’”’ was dropped. There will therefore be no
attempt to divide those examples into groups by form or size, but
similarities between specimens will be noted, for what they may be
worth.

I attribute this breakdown of the original types, so carefully main-
tained by the prehistoric folk, to the distributive effects of two
Transvaal Bantu customs, exogamy and transhumance, both of which
meant movement on the part of the women, users of the bored stones,
away from their homes where they had learnt their agricultural
methods. These two customs may even have acted together, so
that women who married from cattle-posts might have moved even
farther afield than would normally be expected. The refugee move-
ments engendered by the Zulu explosion in the early years of last
century played their part, and this will be further dealt with in dis-
cussing the MaNtatisi (Tlokwa) movement.

SOURCES.

Initial numbers refer to sites on the key map. As circular specimens
are listed first, gaps will be found. These are left to include elongated
examples, given later. Too much attention must not be paid to
identical measurements in this Transvaal series. Hxamples from the
Transvaal Museum, Pretoria, appear to have been measured to the
nearest eighth of an inch (3 mm.), and the figures transcribed to milli-
metres. The apparent identity thus represents an error of 1-5 mm.
in that series.

2. “Zimbabwe specimen” (South African Museum collection).

This unusual specimen was certainly made for ornament and not
for use. It has been illustrated both by Bent and by Hall.* The

* Bent, Ruined Cities of Mashonaland (p. 170 in first edition, p. 202 in 1893
edition); and Hall, Prehistoric Rhodesia (1909, p. 350).

VOL. XXXVII, PART I, 12

178 Annals of the South African Museum.

perimeter and one face are decorated with regularly disposed raised
circular bosses, superficially resembling beaten metal work. Each
boss is about 1-5 cm. in diameter and 3 to 4 mm. high, each carefully
shaped. In contrast the underface is crude and irregular. The
central bore has been chiselled to shape and shows no signs of usage.
The hole in the upper face is well made (diam. 5-3 cm.), but the
opposing mouth is larger and cruder (diam. 6-9 cm.). —

Including bosses, the diameter is 21-6 cm., excluding bosses 21-0
cm. Including bosses, the height is 9-8 cm., excluding bosses 9-4 cm.
The last measurement was taken on the central rim surrounding the
hole.

The original gives the odd impression that it might have been a
template or forme for the making of beaten gold objects. Had it
been so used, it would have been impossible to remove it from the
finished object without first breaking it. Two radial cuts, one deep
and wide, the other shallower and narrower, set roughly at right
angles to each other, mar the decorated face. They were made after
completion and were present in Bent’s illustration. No reason for
their presence has been deduced. The material is a soft soapstone,
easily carved or scratched, but too soft for hard use.

It is important to realise that this specimen was found by A. and W.
Posselt, in 1889, in the “Zimbabwe Region,” and was not found in the
Zimbabwe “‘Temple” ruin. Hall stated that it was “found in close
association with the great conical tower and platform, and with carved
birds and scores of phalli.” It was not. At the same time it must
not be forgotten that the carving of the soapstone fits in well enough
with the soapstone birds, etc., to be regarded as a part of the same
general complex, as does the material. Additional remarks may be
found in Miss Caton-Thompson.*

NORTHERN TRANSVAAL.
7. MaRa.
ASJ 43/49 a VL Os 12:0 609 4655 56 2-4

8. SanD RivER, ZOUTPANSBERG.
TMP 7392 20 18:4). 213:0. 370-6)" 4400 753) peo

9. SCHOEMANSDAL, LOUISTRICHARDT.
ASJ 51/38 2) USD 1IE2 6574 4260". ST 20

* Zimbabwe Culture (pp. 159 et seq., footnote). Mr. J. Schofield of Durban tells
me that the specimen was found on the surface of the Acropolis ruin, in plain
association with the stone birds, by W. Posselt.

The Bored Stones of South Africa. 179

10. SPELONKEN.
SAM 3189 17:0 10:0 588 2890 50 2-4

11. DUIWELSKLOOF, TZANEEN.
TMP. 9122 - 19:0 11:0 57:99 3970 44 2:5

NortH CENTRAL TRANSVAAL.

13. PIETPOTGIETERSRUST.

TMP 7436 . 17-8 10:3 57:9 3265 3:3 2-2 Groenfontein 195.
7561 « 15:2 9-4 es Re 4-6 3-8 Bored, split, re-
bored at right
angles.
15. KALKFONTEIN 146.
SAM 3819 . 194 108 556 4065 68 1:7 6 feet deep in
alluvial soil.
3817 . 195 106 54:3 4030 70 2:3
3818 » 18-057 11-0 “GIL 3565 4:0. 2:3
3817 . 20-4 TA G-d) pT OUSO a 42aNm ore
17-8 95 534 3010 42 26
15:8 (8) 249-35) 1945) 9 5:2) 2:3
16. NEw SMITSDORP.
TMP 1890 Peel -G 1O-2h 47-20 4760) 3-6 9 1-3
1887 a4: (-Onet2i-Or) 4515) 3-9) 2-7
1885 se li:77 8-2 416 3180 42 3-0 Badly fashioned.
1888 . 165 108 65:4 2940 35 1-7
1889 . 159 8:9 55:9 2250 2:9 1:3
1886 5 day 73 47:3 2055 3-8 2-5
1884. . 16:8 6-4 381 1805 1:9 1-4 Uneven, badly
fashioned.
5895 lb? 73 47-9 1690 3:6 1-7 Irregular.
PBB 2) 5-2 70 460 1615 35 1:9
1883 . 19:0 43 226 1550 3:4 2-5
5894 5 ASR 4-3 32:6 760 3:8 2-7
17. EERSTELING.
TMP 9176 5 | ee 75 49-2 1735 3:3 2-4 Surface find.

EASTERN TRANSVAAL.
18. OHRIGSTAD.

MMK 1255 2 19-1. 10:3: 53°99 3755 44 # 2:0

PEM 615 J . 172 10-7 62:0 3165 43 41:8 49 kg.

19. RUSTPLAATS.

TMP 7880 . 16:5 G25 36a 16908" 8:75 °7-7, |, Duge from “deep
underground.”

20. KRUGERSPOST.

TMP 9123 - 19:0 14:7 774 53805 5:9 2:5 Huge. Surface
find.

8039 - 16:2 12-4 76:4 3255 6:0 2:9

21. WATERFALL VALLEY.

TMP 2339 . 19:0 114 600 4115 50 3:0 “Makatese.” Fine
workmanship.

180 Annals of the South African Museum.

22. BaARKLY VALE. ALKMAAR.

HAC 1041 - 15:8 85 53-2 2120 50 2-8 #£xDiabase. Surface
of old river ter-
race. Weathered.
3:6 kg.

23. ScHaGEN, NELSPRUIT.

WUA 1086 . 19:9 146 73:4 5780 3:6 As 9:7 kg. 2?Gabbro.
1085 F . 18:0 12:3 683 3985 4:5 3:8 6-2kg. Taleschist.
1085 E . 18:2 11:7 64:3 3875 63 2:9 63 ke: Amphi-
bolite.
1085 B . 16:8 13:1 784 3695 55 2:4 55 kg. ?Gabbro.
1085 A. 17-7 95 53:7 2975 3:8 1:8 4:1 ke. ?Gabbro.
1085C . 13:9 84 604 1625 43 2:8 2:2 kg. Tale ser-
pentine.
1085 H . 14:8 3:4 23-0 145 4] 2:4 Brokensiawe
schist.
24, BARBERTON.
ASJ 49/36/8 . 186 140 75:2 4845 40 2-6
TMP 1914 5 PANES} 98 48:2 4040 3:3 1:7 Lot 133. Surface.
1911 . 165 106 64:2 2885 49 2-7 Lot 133. Native
smelting work.
1912 . 16-5 5:9 57:6 2585 52 2-4 Kaap River wash,
near Honeybird
Creek, with a
prehistoric
specimen.

SAM 3017 . 150 11:0 73:3 2475 70 2:9 Surface. Hole has
distinct shoulder
near mouth.
Fine workman-
ship.

26. WELGEVONDEN, Komati VALLEY.
TMP 8031 5 élelL 85 35:3 4935 81 3-4
8027 . 19:0 10:6 55:8 3825 56 3:0
8024 . 20-0 95 47:5 3800 53 2:5
8026 - 18-7 108 57-9) 3775 5-6 2:0 “Trreculam
8022 » 17-8 14) 964-0) 3610) 16:0) 53-5
8021 2 183% 10:3) Poo S600 a1 ase
8032 . 20:3 81 40:0 3340 64 3-0
8030 parc 79 38:9 3255 44 2-9
8019 . 17-4 $102 586 3090 49 28
These last three
8020 Niel G7. (S22) 6:3 2k specimens seem
8025 ~ 20:0 (7-5) 5:3 3:4 to have been
8023 9 Oh (on) 5:2 2-4 cut (?) horizon-
tally.
STEENKAMPSBERG.
27. ROOSSENEKAL.
MMK 1155 = G5 97 59-1 2460 42 2-9
28. BADFONTEIN.
TMP 2344 ZO ler 1 pG2d) ) ho250 1 1i3:Oli2ss
2338 . 184 11:7 63:5 3960 48 3:0 #£‘*Makatese.” Fine
workmanship.
O34, ae Sa a10:60.59-9) | oc6O) yao d
2345 : 278 100 662 S170 5:6 7
2346 a es 94 52:8 2980 3:5 2:5

The Bored Stones

29. SCHOEMANSKLOOF.

TMP 2341 17:8
2340 16:5

30. MACHADADORP.
ASJ 15/35 /2 18-4

31. SEVENFONTEIN 39.

ASJ 12/36 20:3
32. WATERVAL 97.

ASJ 12/41 19:1

19:5

19-5

17-7

17-2

16-6

19-2

33. BLOEMFONTEIN 101.

ASJ 12/36/B 18:7

13-0

12-4

=" —
7
~]

a
POomsosy
Crim bo Oo OS Aw

12-0

73-0

75:7

64-2

of South Africa.

4120 49 2.2
3375 49 2-9
4065 63 24
4410 5:0 23
4485 4:7 1:6
4070 75 35
3650 5:0 2-7
3420 5:3 2:2
3315 44 15
2510 5:0 2-7
2395 52 3-5
4195 59 23

181

*“Makatese.” Sur-
face. Possibly
unfinished.

‘*Makatese.”’ Sur-
face find.

Amphibolite.

Ss.

Diabase.
Serpentine.

Hole bored skew.

(Both Waterval 97 and Bloemfontein 101 material is associated
with Bantu terraced agriculture and ruins in the Komati valley.
They are adjacent farms.)

34. CAROLINA.

SAM 3020 170 10:8 63:5
(7-2 6:3 87-5
(7-2 6:3 87-5
SOURCES OF
35. BREYTEN.
SI 20 20-2 11:2 55-4
36. ERMELO.
WUA 1036 18-6 9-1 48-9
NMP. 2066 17:8 9-8 55:0
37. WITKRAANS 58.
TMP 2348 1:7 10°3 55-1
WUA 1036 18-6 9-1 48-9
38. Mount Prospect, AMAJUBA.
JE Briscoe E7-Se ls25.0 15:8
39. HEILBRON.
ELM 36 . 205 115 £561
36/B 18°55 11:0 59-5

3120 40 2:8
325 3-2 2-7)
325 2-8 1-5)

THE VAAL.

4570 4-2 2-6

3150 5:0 3-5

3105 5:2 2:7

3600 5-1 2-4.

3150 50 3:5

4275 4-4 2-3

4835 5:0 2:5

31G5 0:0) 3-0

Unfinished dagga-
pipe ?

Unfinished dagga-
pipe? Sides
flattened.

Glen Avon Farm.

5-2 ke.
3-9 kg.

?Gabbro.

Vaal River, south
of Ermelo.
5-2 kg.

Surface deposit.

Rhenoster River
affluent.
Cf. first Witkraans

spec.

182 Annals of the South African Museum.

East CENTRAL TRANSVAAL.

40. MAPOCHSGRONDEN.

TMP 2342 - 19-7 11:6 589 4500 59 2:7
19:0 89 416 3215 48 2-7 #£2Broken.

41. WITPOORT.

TMP 7543 « 1957 eT 59-4 4540) 57 2-7
42. LAGERSDRIFT.

TMP 2775 § lve! 9-5 955:6> 2775 5:8) 28
43. KLIPBANKSPRUIT.

TMP 5893 . 178 98 55:0 3105 S51 3:6

CENTRAL TRANSVAAL.

44, NYLSTROOM.

SAM 3843 15-0 9-5) 163-3 21409) sale eZ

45 TURFFONTEIN.

TMP 1892 - 140 13:3 95:0 2605 29 I:-1 Found in Bantu
kraal.

46. WARMBATHS.
MMK 1153 . 153 #118 §$.76 2760 48 1:5 Ss.

47. WONDERBOOM.
TMP 1894 - 190 10:2 53:7 3680 4:3 19 Badly trimmed.

SOUTHERN TRANSVAAL.

48. SELONS VALLEY.
TMP. 5908 2 20:3 AO) waren oLSO) Ou) eo

49. MIDDELBURG.
ADJ 49/36/9 . 18-7 93 496 3250 60 2:5

50. DEvon.
TMP 1417 - 162 11:7 =72:9 3070 5:3 2:0 Suikerbosrand
affluent.
1418 . 19-7 (4:5) ace aie 4:5 3:0 ?Cut horizontally.

51. GREENLANDS.

ASJ 8/41 ~ 152 111 72:2 2565 45 2:6 18 inches deep.
No associations.
Cf. first Devon
specimen.

ELONGATED SPECIMENS.

1. GwAAI RESERVE, S. RHODESIA.
SR/WI 3/653 17-5 15-4 8:9 2400 45 2:3

3. MELSETTER.
SR/WI 3/654

4, MESSINA.

HA Coll. 1248 .
OAS.

5. MALABoc#.
ASJ 15/35/1

6. BLAAUWBERG.
ASJ

23. ScHacEen, NELSPRUIT.

WUA 1085 D
1087

25. STEYNSDORP.
TMP 1449

26. WELGEVONDEN.

TMP 8029
39. HEILBRON.
ELM 36 A

49. MIDDELBURG.
ASJ

52. RIcKETSDAM, MaRIco.

WOUA 397

14, PIETERSBURG.

SAM 3821

The Bored Stones of South Africa.

18-4

16:5
15-3

20-9

20-7

18-3
18-5

27-9

24-0

20-0

7-5

13-7

15-4

16:5

16-3

6-7

12-0

(14-0)

(6-0) 1700
5-0 905
3°5 645
9-8 3050
9-2 2895
9:0 2370

(8:7) 2140
3-5 1740
8:7 3090
5:3 ©1025

11-5 4555
7-0 2280

VARIOUS.
5:7 285
3-4
4-8

3°5

4:5
3:5

6-1

5-4

gro
ke o>

Ll
&

4-8

3-2

3:7
5-0

2-4

3:6
1-4

roy
oO

bo
> 00

3:0

2°3

2°5

1-9
ol.

183

Soft red Ss.
Riverhill farm.
One face worn
recently.

1-5 kg. Weathered
schist.

12 kg. Basic
schist.
Amphibolite.

Eccentric bore.

4-6 kg. ?Gabbro-

?Gabbro.

Irregular shape.
Irregular shape.

6-2 kg.

0-62 kg. Iron-
stone. Found
with ash-heaps,
metal industry
and pottery.
(Elongated.)

Soapstone. Elon-
gated.

Soapstone. Sub-
circular.

Mr. T. J. Trevor, when submitting these last to Dr. Péringuey in
1917, wrote: “In other parts of the country where such stones are
plentiful, I have been informed by the natives that they used to carry
these on a stick, in order to have them handy to throw at a buck—

184 Annals of the South African Museum.

the idea being to throw them on edge like a skipping-stone and
attempt to break the buck’s leg.” Both example8 are formless slabs
of soapstone, quite untooled. Their only right to inclusion lies in the
central bore. Butcomparethe example described by P. J. Kloppers *
from Natal, referred to in a previous part of this present work (p. 33).
Two examples away from the Transvaal are almost certainly
Bantu in origin, and probably quite modern. Both appear to have
been bored with a fine metal drill, probably fencing wire, possibly
Bantu-made wire, or the tang of a light spear. Only the actual
dimensions are given. They are probably beads of some sort.

61. NewLanps, BarkLy WEST.
MMK 1536 Ben!) Sl a tes 0:6 0-4 #£Broken.

62. RAMALITSE, THABANCHU.
SAM 4873 «1 205 1-7 =a Se 0-7 0-5

We now come to a series of figures that appears to be related to the
path taken by the BaTlokwa and other Bantu peoples, under the
leadership of their queen, MaNtatisi. This extraordinary movement
has got the name of ‘“‘ Makatisi,”’ etc., in the Transvaal, where these
people lived before the hopeless refugee movement that ended in their
destruction and scattering in the neighbourhood of Litaku and Kuru-
man. I have started with the Ventersdorp and Wolmaransstad sites,
as this material accords better here than with the remainder of the
Transvaal.

54. VENTERSDORP.

TMP 3721 . 14-9 56 376 1245 3:9 25 Schoonspruit
affluent.
ASJ 17/48 5) hs) 8:3 53°55 1955 6:3 (1:8) Wonderstone.

Conical bore.
Stompjie’s Dig-

gings.
55. WoOLMARANSSTAD.

TMP 8047 BL bo7| 2-4 20-0 405 3:0 1:9 Makwassiespruit

affluent.
56. MAFEKING.

ASJ 38/38 2 2 3:9) 124-1 1745) 4-9) 2 Probably metal

bored.
57. VRYBURG.

SAM 763 . . 140 142 101-55 2788 3:7 24 $3feet deep. Hass-
forth, Metal
bored.

Mossop Coll. . 14-4 8-0. 556 1659 5:2 2 Pecked. Farm
England.

* §. Afr. J. Sci., xxxi, 1935, pp. 474-475, illust.

The Bored Stones of South Africa. 185

58. RIVERVIEW EstTaTE, VAAL RIVER.

ASJ 5/35/55 . 17:0 10:5 61-8 3035 6-0 .. Site VI. Sohnge,
Visser and Lowe,
op. cit.

59. PosTMASBURG.
MMK 252 lar ZO 83-9 (24550 Oo oe) Ss
Sa - 14:8 G:omw aol 42 ona oD Oe SS:
14:3 4:8 33-6 980 2-4 1-8 Ss. Metal bored.
583 5 118383 4:9 28-2 935 2-5 1-7 Ss. Metal bored.
7, 191 » 14:3 be i ae S65 2:0 Very worn soap-
stone.
60. KAMEELFONTEIN.
MMK 1283 . 148 15:4 104:1 3375 (4:2) 2-5 Metal bored.

It will be noted that, in general, the average diameter (14:5 cm.) is
a little below the minimum for the Transvaal Bantu types, and only
one example, from Riverview, exceeds that minimum. Height is
erratic, showing extremely thick examples (in three instances with an
index in excess of 100-8 per cent.) and very flat types. None of these
fit comfortably into the Prehistoric Series. The Wolmaransstad
specimen recalls the Natal Lowlands.

The first Vryburg specimen, from Hassforth, is somewhat amusing,
as it has obviously been made to resemble an ostrich egg in shape,
texture, and size. The material is marble.

The Tlokwa migration and eventual dispersal will be discussed more
specifically later (p. 202).

GENERAL DIMENSIONS.

If the diameters of all Transvaal specimens are considered, we find
that an absolute differentiation is possible (Histograms I and II).
All those specimens from the neighbourhood of the Transvaal with a
diameter greater than 15-1 cm. are Bantu. All those examples from
the same neighbourhood with a diameter less than 13-4 cm. are pre-
historic. From perhaps 14:0 cm. to 15-0 cm. there are overlapping
types, some Bantu, some prehistoric. These last are most consis-
tently represented in the Vryburg-Postmasburg area, where we find
several examples that fail to fit the prehistoric types, while some show
distinct use of a metal borer. These I attribute to the Tlokwa
migration.

For purposes of direct comparison I have used the figures of all the
Bantu specimens, and prehistoric material from the same area. It
will be seen that the Bantu Series yields a simple but asymmetrical
curve, with a peak between 17-5 and 19-5 cm. (mode 17:8 cm.). The
range is from 12-3 to 21-5 cm., though there are three examples with a
diameter greater than 20-5 cm. In contrast to this the Prehistoric

186 Annals of the South African Museum.

I. Bantu circular examples by diameter.

5 v 1 z By 7G cms
RCRD ERED ERS SHOR ERABAP
eee Py
Ret ee fae Tha

a
20

III. Bantu circular and elongated examples by height.

IV. Prehistoric Transvaal circular examples by height.

Fia. 7.

The Bored Stones of South Africa. 187

Series produces a trimodal curve, with modal points at 8-0 em.,
10-4 cm., and 12-8 cm. respectively. These probably represent three
uses of the bored stone, but mathematically they are made up of
the seven prototypes and their variants already analysed. These
examples cover an effective range of 4:5 to 14-5 cm.

The study of height alone (Histograms III and IV) provides little
new evidence, but corroborates the study of diameters. The pre-
historic Transvaal material yields an irregular curve, again showing a
trimodal distribution, with modes at 4:3 cm., 7-7 cm., and 9:5 cm.
This is obviously a product of the same causes outlined above. In
contrast the Bantu Series shows a simple asymmetrical curve, with
the mode at 10-7 cm. This asymmetry can be attributed to the
desire to make an implement as close to the limits of thickness as
could be efficiently bored. The effective ranges in the two histograms
are about 2-0 to 12-0 cm. for the prehistoric, and 3-5 to 13-5 cm. for
the Bantu, merely showing that, while the Bantu and prehistoric
people were capable of boring through similar thickness of stone,
the Bantu could extend their range slightly towards thicker
stones.

I have not given histograms of index or cube, but these can be
very easily worked out on graph paper from the data. The study
of index shows that Bantu types have an effective range of 20 to
78 per cent., with a mode at 56 per cent. They seldom exceed 66 per
cent., and only one or two examples reach 100 per cent. In contrast,
the Prehistoric Series from the whole Union quite frequently exceeds
100 per cent. and even 110 per cent., to form prolate anomalies, with
no distinctive distribution.

The study of cube is comparable to that given above for the
diameters. The prehistoric range from 50 cu. cm. to 2000 cu. cm.,
and the Bantu from 2400 cu. cm. to 5300 cu. cm. The inter-
vening depression covers a few isolated examples of ambiguous
origin, but not too much faith should be placed in the “‘cube” as
evidence.

This gives us sufficient data to define the great proportion of our
Bantuexamples. They consist of those from the neighbourhood of the
Transvaal with a diameter greater than 15-1 cm., a cube greater than
2400 cu. cm., a weight greater than 3-7 kg. (8 lb. 3 oz.), and a height
that does not usually exceed 78 per cent. of the diameter. While this
does not cover all known Bantu examples, it excludes prehistoric
types. To complete the Bantu Series we should include smaller
forms from this general locality that do not conform to the known

188 Annals of the South African Museum.

prehistoric types. Their allocation becomes more certain when
they fit comfortably as extensions of graphs or histograms based
upon known Bantu material. This last applies especially to the
Tlokwa migration.

STUDY OF THE Bork.

I have attempted to discover whether there is any means by which
we can recognise examples that have been bored by metal tools.
Almost cylindrical bores are rare, and are certainly products of usage,
and not of technique. The Zimbabwe area specimen is an exception
to this rule, just as it is to many others. Here tool marks visible in
the hole through the soft soapstone show that this was partly chiselled
to shape, and there are no signs of usage.

In such a study the centre of the hole can yield no result, as the
bore may be unfinished or may have been smoothed and enlarged by
use. In the biconical holes there should be some relationship between
the base of the cone (represented by the mouth of the hole) and the
thickness of the stone ; or, more accurately, half the thickness of the
stone. I have therefore developed histograms of the Index of Bore:

mouth of hole x 100
% height

At best the results obtained from such an index can only be very
general, as we are dealing with a variety of irregular cones cut into
materials of varying hardness. In some cases the lip of the conical
bore is difficult to recognise, in others the mouth of the bore varies
on the two faces of the specimens. Measurements are taken from the
wider mouth of the bore, so that this is to some extent accounted for.
I have refrained from giving the IB of specimens in the lists of sources
of data, as individual cases are merely misleading. However, pro-
vided that we do not argue from the general to the particular, we can
draw some conclusions from the accumulated figures.

We must exclude examples with a conical bore. Discs or specimens
less than 3-5 cm. thick have to be ignored as well. It was found that
18 such specimens from the Cape show an average IB of 116-5 per
cent., while 22 from the Knysna area yield an average IB of 114-1
per cent. This suggests that boring in the case of smaller stones was
done by pecking two wide cup-shaped depressions in the opposing
faces of the stone, then merely using the rimer to smooth and enlarge
the hole. This is a very slight technical variation, and may or may

The Bored Stones of South Africa. 189

not be cultural. Perhaps a more accurate gauge might be developed
by plotting IB against height or thickness, on squared paper, but from
the few samples I have so far tested there seems to be little new to be
learnt from this.

Histograms of all Bantu and all Transvaal prehistoric examples
are given (Histograms V and VI), and as a check I have de-
veloped the histogram (No. VII) of 106 examples from the Free
State, chosen mainly from the Koffiefontein sites, but otherwise at
random. )

The 88 Transvaal Bantu specimens show an effective range of
20 to 85 per cent., an almost symmetrical curve with the mode
at 45 per cent. The 86 prehistoric examples from the same general
area yleld an asymmetrical curve over a range from 25 to 100
per cent., with the mode at 43 per cent.—even lower than the
mode of the Bantu Series. It becomes obvious that the two curves
are directly comparable, and that specimens were bored by the same
techniques. I had hoped to find a bimodal curve, representing those
bored with metal (low index) and those bored with stone (high index),
but instead we find that the bulk of specimens in each case show an
IB greater than the mode (7.e. the average is higher than the mode,
yielding an asymmetrical curve), suggesting that many specimens
in both series were bored with stone, but that a high proportion of
those below the mode were bored with metal. This fits perfectly with
the deductions made earlier concerning the conical-bored specimens
from the VMR. area in the western Transvaal.

This deduction must be viewed in the light of comparative material
from elsewhere, and, knowing that the IB in the Free State shows a
different state of affairs, I give the histogram of 106 examples from
there. The range is from 30 to 105 per cent. The main mode
is at 60 per cent., and there is a subsidiary mode at 77 per cent.,
due to the absence of figures from 73 to 75 per cent., possibly a chance
gap, but possibly due to differences of material. The stones used in
the Free State range from hard igneous rock or indurated shale to
somewhat softer sandstones or shales.

In an attempt to control these figures by means of a third series, a
sample was taken from the Tzitzikamma-Knysna-Mossel Bay area,
where coarse-grained tertiary sandstones and fine-grained quartzitic
sandstones are used. The sample (Histogram VIII) shows a range
from 33 to 129 per cent., a fairly symmetrical curve with a mode at
63 per cent. A final sample from the Cape of Good Hope shows an
exactly comparable symmetrical curve, from 41 to 120 per cent., with

190 Annals of the South African Museum.

p-4

a=
Hl
8 o|6 | ool! fe)
Kn 3
i
zt aed eS BEES (FET Sea

ae
G
Ss = Oo!
| GI
| JIU |ad¥ TSI oy
x i St
@ W OlS g 6 ool/ oli

Fia. 8.—Index of Bore.

The Bored Stones of South Africa. 191

the mode at 66 per cent. The materials used here are quartzitic
Table Mountain Sandstones, Malmesbury shales, etc.*

From the comparative evidence yielded by these histograms we
can state with certainty that some distorting factor is present below
the 50 per cent. line in both the Transvaal histograms that is absent
elsewhere. It is reasonable to assume, in the light of the conical-
bored specimens in the Prehistoric Series, that this factor was the use
of metal. Developing this assumption, we may further assume that
all examples with an IB below 30 per cent. (the lower limit of almost
all IB from outside the Transvaal) have certainly been bored with iron.

At the same time it is essential to remember that many of the
materials used in the Transvaal were of a softness and a coherent
texture that would permit of a narrower bore than harder substances.
The prevalence here of soapstones, schists, marble, wonderstone, soft
shale, etc., cannot reasonably be ignored. Unluckily the exact
knowledge we need is not available, and we can draw no conclusions
with certainty. What is wanted is not an exact analysis of the
stone, but rather some gauge, akin to the rule of thumb used by
geologists to denote hardness and softness, to provide us with know-
ledge of reaction to mechanical corrasion. A gauge of this sort has
recently been developed by M. J. Janmart in Angola in an attempt to
assess relative wear in mixed gravels. It is to be hoped that some of
his methods and results will be published.

In general we can take it that, in those areas and periods that show
symmetrical curves in the histograms of IB, these represent the desire
to bore a hole through the stone, without any interest in the size of
the mouth of the hole. In cases where there is a marked asymmetry
towards the lower indices, we must suppose that some real effort has
been made to obtain as narrow a mouth as possible. The distortion
in the curve represents the degree of effectiveness of this additional
effort, by whatever means the narrower bore was accomplished.

AFFINITIES.

In plotting these Bantu figures graphically I have found that the
most illuminating presentation is obtained by using height as the

* Not included in this last sample are two striking and perhaps important
specimens. One is from Hout Bay, with an IB of 18 per cent., the other from
Stellenbosch with an IB of 30 per cent. Both are in the University of Cape Town
Collection. The Hout Bay example is certainly metal-bored, and the Stellenbosch
specimen lies on the 30 per cent. line, which we can safely regard as the line of
absolute demarcation between the two methods of boring.

192 Annals of the South African Museum.

horizontal component, and index as the vertical. There is little that
can be said with any certainty, and there seems little relationship
between type and geographical distribution. Larger numbers of
specimens and a greater and more exact knowledge of association,
use, and provenance are essential before this question can be finally
dissolved.

1. Specimens from Greenlands and Devon, both on the Suikerbos-
rand stream, show distinct affinities. Both belong to a small group
with indices above 70 per cent. and heights between 11-0 and 12-0 cm.,
thus standing somewhat separate from the usual types.

2. Heilbron shows affinities with Breyten, Ermelo, and Witkraans,
and perhaps with some of the Welgevonden and Waterval Series.

3. In the northern Transvaal the following sites show strong
affinity, Mara, Louis Trichardt, Sand River, Spelonken, and Tzaneen
(sites 7 to 11). All have diameters between 17-0 and 19-7 cm. and
indices ranging from 57-9 to 70-6 per cent.

4. The farm England in the Vryburg district shows affinity with
four specimens from New Smitsdorp (16), one from Kalkfontein (15),
and the Eersteling specimen (17), suggesting that the Vryburg
example is an offshoot from the northern Transvaal.

5. The Mount Prospect specimen (38) is very similar to the first
example from Schoemanskloof (29). Mount Prospect lies in the
original area of the BaTlokwa and BaNSia tribes.

6. In the elongated series the two Rhodesian sites show marked
affinities, and Malaboch (5) and Blaauwberg (6) show much the same
type.

In general we can gather that no type of implement is localised
to any one area, the makers seem at times to have aimed at divergent
shapesand sizes. Thiscan be seen by comparing the eastern Transvaal
figures and those from the sources of the Vaal with material farther
west.

DISTRIBUTION.

The distribution of known specimens is of extreme interest. Mr.
T. J. Trevor, writing to Péringuey on 28th August 1917, notes that
the Bantu specimens “‘are plentiful in all the districts of the Transvaal,
with the exception of the Low Country below the Drakensberg on the
east, where I only found one.” Apart from that short reference,
nobody seems to have realised the relationship existing between
distribution and altitude. The distribution of all specimens known
corresponds in a most remarkable manner with the highlands of

The Bored Stones of South Africa. 193

Southern Rhodesia and the Transvaal, and careful collation with the
physiographic map shows that the makers of these stones seldom
transgressed below the 2500- to 3000-foot contours (800 to 900 m.).
The Letaba, Pilgrim’s Rest, Nelspruit, Barberton, and Swaziland
lowlands are without specimens, while the relatively low-lying Gwanda
and Chibi districts in Southern Rhodesia are similarly barren of Bantu
bored stones.

In contrast to this, specimens are scattered all along the adjacent
highlands. In Rhodesia the three Bantu examples are from the Gwaai
reserve, the Zimbabwe area, and Melsetter district. In the northern
Transvaal one questionable group alone comes from Messina in the
Limpopo valley. The remainder are concentrated along the Blaauw-
berg-Zoutpansberg ridges. South of this there is a gap coinciding
with the Great Letaba valley. South and south-west of that river
the distribution increases to its maximum concentration, along the
Strydpoortberg and down the Drakensberg. Distribution continues
down either side of the Steenkamsberg, and south to Witkraans,
Heilbron, and Amajuba, a solitary outpost on the Natal border.
Another extension runs westward through Nylstroom, Warmbaths,
Middelburg, and Wonderboom, and a southward drive passes along
the Suikerbosrand to Devon and Greenlands. The further extension
west and south, related to the Vaal River, consists of sites at Venters-
dorp, Wolmaransstad, Vryburg, Mafeking, and others marking the
MaNtatisi spread.

This curious relationship between altitude and the Bantu bored
stones can imply only one thing, namely, that we are dealing with a
mixed pastoral and agricultural people, limited strictly by the original
distribution of the tsetse-fly and cattle nagana. All the extreme
eastern Transvaal and the lower and middle reaches of the great
eastward-flowing rivers were fly-infested before the “Great Rinder-
pest” of 1892-1894 so reduced the cattle and big game that the fly
was thinned out through lack of the essential hosts. The tsetse-
borne nagana, while consistently fatal to cattle, appears to infect big
game less fatally. Both fly and nagana appear to be maintained by
game, so that these areas provide an almost complete cattle-barrier.
This does not necessarily mean that this barrier was absolute so far
as Bantu agriculturalists were concerned: a native can exist as a pre-
carious tiller of the soil without cattle, and indeed the impoverishment
of the native tribes of Portuguese Hast Africa is fundamentally
due to this one-sided mode of life in a cruel environment. To enter
the tsetse belt means poverty and struggle, so that the distribution,

Viol, XO, PART I: 13

194 Annals of the South African Museum.

DuRBAN © 5D}

Map 16.—Distribution of Bantu Bored Stones. Areas with general Tsetse history shaded.
3000 ft. contour marked. Sites referred to by number.

The Bored Stones of South Africa. 195

with the 150-mile gap separating Southern Rhodesia from the specimens
of the northern Transvaal, is extremely suggestive. It is curious
to note that the Mapungubwe culture in the Limpopo valley lay at
altitudes from 1700-2000 feet.

We have here a clue, not only to those areas at one time infested
by fly, but also to those areas that might once again become infested
should vigilance be relaxed. The former extent of the tsetse country
has been described by E. E. Austin * and by R. Newstead,t while
the more detailed distribution in Southern Rhodesia is given by

R. W. Jack.t

UsrEs AND USERS.

While the Bantu have imitated the Hottentot or Bushman bored
stone, and seem to have continued the use of this implement after
losing contact with the Bushmen, yet our knowledge of the uses to
which the tool was put by the Bantu is less well authenticated than
for the Bushmen. This I attribute almost entirely to the fact that
this was a woman’s digging tool, and that no attempt whatsoever
has been made to obtain information from old native women in the
Transvaal. All the information seems to be from the men.

Apart from a newspaper article by a Miss E. Bezuidenhout,§ Mrs.
Hoernlé is the only person who seems to have placed on record any
detailed evidence on the subject. Unhappily neither account can
be made to conform with the evidence of the stones themselves.

Miss Bezuidenhout describes an account from an old MuVenda,
who told her that these vast stones were threaded on to the horns of a
selected beast at the time of the full moon. A fanciful illustration
accompanying the article shows no less than fourteen such stones so
threaded on the horns of a proud beast. Any additional weight on a
cow’s horns will depress the head and is irksome. The addition of
a weight of at least seventy pounds would be an impossible burden.
More important is the evidence of the bore. No hole centre is greater
than 4-0 cm., while the maximum distributional frequency is less than
an inch (2:4 cm.) and would not permit the passage of even an inch
of horn. The tale must be dismissed as a complete fabrication on the
part of the genial and possibly mercenary Venda.

* HK. E. Austin, A Handbook of the Tsetse-flies, Brit. Mus., 1911.

+ R. Newstead, Guide to the study of the Tsetse flies, Liverpool Univ., 1924.

t R. W. Jack, Tsetse-fly in Southern Rhodesia, Dept. Agric. Salisbury, S.R.,

1918.
§ E. Bezuidenhout, article in Natal Daily News, 26th April 1941.

VOL. XXXVil, PART 1¢ isis

196 Annals of the South African Museum.

Mrs. Hoernlé’s account * has a closer relationship to fact, but we are
again dealing with hearsay evidence. The informant in this case was
Mr. W. G. Barnard, and his source of information is an account given
by members of the BaRoko and BaKoni tribes, men well over seventy
years old. In spite of the fact that the BaRoko were themselves iron-
workers and became smiths to their Pedi masters, they say that iron
was a rare metal in the Transvaal and only the Pedi overlords could
afford to use the metal for such effeminate uses as tilling the soil.
“The ordinary tribes-people used very large wooden hoes weighted
with bored stones which were fixed to the haft. Both BaKoni and
BaRoko men well over seventy years of age have stated that they
saw these hoes in use up to sixty years ago. It was only after the
opening of the diamond mines. . . (after 1870) . . . that iron hoes
became universally used throughout Sekukuniland.”’ A wooden hoe
was made to the order of Mr. Barnard, and the illustration shows a
posed native using the hoe, upon the tang (not the haft) of which is
a bored stone. The specimen is at present in the collection of the
Archaeological Survey, Johannesburg, and has also been sketched by
Professor van Riet Lowe,t but it is nowhere stated that the specimen
was made to order, and exactly copies a modern native iron hoe.

Directly the specimen is handled it becomes obvious that it could
not possibly be used in a Transvaal field. The tang on such an
implement is limited to 2-4 cm. by the average size of the aperture in
the stone. The tang passes through a node of wood in the haft, which
is about 7°5 cm. thick at this point. This fragile wooden structure
has to bear, not the static weight of from ten to nineteen pounds, but
the effects of a violent blow on hard soil, backed by that weight. The
stone itself lies above the centre of gravity of the hoe, and the whole
would be almost impossible to manipulate, as the handle tends to
twist in the grip.

If it is presumed that the weight was on the haft, not on the tang,
the structure becomes even more flimsy, as the haft itself is reduced
to less than an inch in diameter in order to pass through the stone,
and would be further weakened by the hole necessary to take the hoe-
head.

If we go three hundred miles south, we have three descriptions of
digging-stick used by the Bantu Mpondo women of a century ago.

* A.W. Hoernlé, ‘A Note on Bored Stones among the Bantu,” Bantu Studies,
vol. v (1931), pp. 253-255, illus.

+ C. van Riet Lowe, ‘“‘Bored Stones in Nyasaland,” S.A.J.S., vol. xli (1941),
pp. 320-326.

The Bored Stones of South Africa. 197

Here the shortage of iron was much the same. The AmaMpondo
depended upon itinerant smiths for their iron, and iron was not worked
until the Insimbini mine was started for native use by a European
trader. Our first account comes from the pen of Rev. Stephen Kay.*
“The latter (digging tools) consist of a kind of wooden spades, which
are usually made by the men, and so formed as to render both ends
useful.’ The other two descriptions were collected by myself in 1924.
The first was taken from the wife of Pangela, a seventy-year-old
member of the AmaMolo clan, living at Hluleka. She described the
digging-stick used when she was a child as being made of sneezewood
(Ptaeroxylon utile) and shaped like a double-ended paddle about 3 feet
in length. The blades were oval in plan and flat in section, the haft
round. It was used kneeling, a position also described by Kay.
My second account came from Chief Victor Poto Bokleni of Western
Pondoland, who gave a similar description, adding that a specimen
had been sent to the British Empire Exhibition at Wembley. An
actual Mpondo digging-stick of exactly this type is in the South
African Museum Collection, Cape Town.

It is thus highly probable that these digging-stick weights were
employed on the only wooden implement which could reasonably be
expected to bear the jar of their weight, namely, a straight digging-
stick, or single-ended spade, used as a crowbar for piercing and break-
ing the soil for a depth of a few inches.

The narrow bore found in many of these specimens suggests one of
two things: either that these are unfinished specimens, or that they
were at times employed in the manner depicted in Mrs. Hoernlé’s
article, but on a metal-tanged hoe, and not on a wooden tang. For
example, a stone weighing some sixteen pounds (TMP. 4760, New
Smitsdorp) would have been of little practical use mounted on a
digging-stick 1-3 cm. (half an inch) in diameter. On the other hand,
it might well have been designed for use on an iron-tanged hoe. Such
use would quickly show conical wear in the hole, but as there are no
signs of such wear we are thrown back on the suggestion that these
are unfinished specimens.

About half a dozen suggestive specimens (p. 11), all of the smaller
Bushman size and therefore not included above, are scattered through-
out southern Africa. These show conical holes with one wide and
one narrow mouth, instead of the typical hour-glass hole. It is
possible that the wider-mouthed specimens may have been bored
with stone from one face only, but one or two specimens strongly

Sole Kay, Travels and Researches in Caffraria. London, 1833.

198 Annals of the South African Museum.

suggest that an iron borer was used either to make the hole or to
ream it out afterwards, or perhaps may have helped to shape the bore
in later usage. The tang of a metal hoe would provide an excellent
instrument for the purpose. Unhappily, as the dimensions conform
with those of the VMR. series we have dealt with them as of Bushman
rather than of Bantu origin. It is perfectly possible that we are
dealing with a local type which will eventually be associable with
tribalised Bushman serfs working under the Bantu. The constant
and seemingly happy association of Sotho and Bushman at the
beginning of last century must not be overlooked.

Mr. Barnard has stated that the BaKoni smiths used the bored
stone for a variety of other purposes, such as bellows’ nozzles or
tuyeres. Kew specimens show evidence of fire, and those that do
might more probably have been subjected to the annual Transvaal
veld-burnings. <A pottery tuyére of the normal Bantu and African
type, such as occurs from the Ubanghi to the Orange, would seem to
be a far more efficient implement.

The BaLemba, according to his article, ““used a disc-shaped stone
with a small hole in it for burnishing copper and other wires... .
Further, this wire was flattened on one side for bangle-making and on
both sides for bead-making, by being pulled backwards and forwards
through a bored stone firmly fixed to a pole or tree.”

Here again the evidence of the stones themselves is against the
suggestion. In a few instances (e.g. the “Zimbabwe” specimen)
there are radial cuts or grooves from one mouth of the stone, but in
no instance could these have been made by a wire passing through
the central hole, as there is always a sharp lip at the junction of hole
and groove, showing that the groove was made by a sawing motion
across the mouth. Mr. Neville Jones describes one such Bushman
specimen (S.R.W.I., 3/356) from Tyger Kloof, and suggests that the
grooves are “the result of sharpening tools, probably of metal.” The
specimens I have seen suggest ownership marks. No disc-shaped
stones with a small hole are known from the Transvaal except for
the two Pietersburg examples.

Native wire does not show signs of having been flattened by this
means. In all the specimens I have examined the longitudinal
striations are raised, and must therefore have been a product of the
drawplate. The semicircular sectioned specimens are also a product
of the drawplate, and the flattened wire is palpably hammered flat.
While we have no evidence either from the stones themselves or from
the wire used in native ornaments, there is reason to believe that

The Bored Stones of South Africa. 199

stones of this sort were sporadically used for a great variety of
purposes.

Mr. Barnard’s evidence, while given in good faith, appears to show
that the male native informants were not clear as to the differences
between the hoe and the digging-stick, which is not surprising when
we realise that these were both essentially associated with woman’s
work, and as such would be scorned by the men. Similarly, these
natives appear to have gone out of their way to describe possible
uses for the bored stones rather than to trust to their memories. I
would suggest that the carefully collected testimony of a few old
women would yield some interesting facts concerning the use of the
Bantu bored stone.

From Abyssinia J. Desmond Clarke sent me a manuscript and
photograph,* describing the actual use of the bored stone there. In
addition the Abbé H. Breuil has described to me how three workers
will dig their straight digging-sticks (not hoes), weighted with bored
stones, into the hard earth, each at the angle of a rough equilateral
triangle. The whole mass of earth so enclosed is then levered out
with one effort, and the solid ground broken up by hammering with
the mounted stone. This produces characteristic wear in the stone
itself, though whether this eventually produces the peculiar smoothing
so typical of many Transvaal Bantu specimens, or whether it leaves a
finely pecked surface, I do not know.

M. J. Janmart recently described specimens to me that he has found
in north-eastern Angola and the south-west Congo. In some cases
these suggest usage about the lips of the bore, while the bore itself
is smoothly polished as though by use. In certain cases they suggest
that they were used as weights or pulleys, and in one instance a marked
groove has been worn into one side of the bore. As the stone hangs
naturally from this groove when suspended freely, he regards the
groove as a result of long usage on a rope.

Confirmation of the use of bored stones by metal-working peoples
near Bethal comes from an independent source. Professor Dart f
says: “The fact that both stone cones and perforated stones occur
together, as at Heilbron, is suggestive.” The stone cones referred to
here are highly symmetrical conical objects, about 6 cm. in diameter
at the base and tapering from there to a point, forming an elongated
ogive from 15 to 25 cm. in length. The most acceptable explanations

* Since published: Man, 1944, Article 25.
7 RB. A. Dart, “‘Phallic objects in South Africa,” 8. Afr. J. Sci., vol. xxvi (1929),
pp. 553-562, illus.

200 Annals of the South African Museum.

for their use are either that they are marline spikes for the making of
twisted wire bracelets, or Dart’s far more plausible suggestion that
they were forms or matrices about which clay was wrapped in order
to make the typical tuyéres or bellows-nozzles used by the Bantu smith.
Specimens of twyéres | have examined show in all cases that the clay
was moulded about some such base, dried, and removed for firing.
A wooden matrix would swell and split the wet clay. Dart shows
that the stone cone and the bored stone are associated together at
Bethal and at Heilbron, and that both are there related to a metal-
using people.

There seems to be no authenticated record of the Bantu employing
stone-headed clubs. Their use of metals probably allowed them to
make the technically simpler wooden knobkiri. It is worthy of note,
too, that the root “kiri” appears to be absent in the Bantu languages,
- and both Afrikaans and English appear to have taken the root over
directly from Bushman or Hottentot sources.

As a result of the Preliminary Survey, Mnr. J. B. de Vaal got into
touch with Mr. B. D. Malan of the Archaelogical Survey, and the
latter has very kindly permitted me the use of his note, from which
I translate.

‘““At the request of Mr. B. D. Malan of the Archaeological Bureau
I have tried to discover from various old Venda for what the bored
stones, so commonly found on different farms in the Soutpansberg,
were used. I showed them three stones, from my collection, that I
had found on the farms Mooiplaas, Maanskyn, and Perth. Three
Venda told me individually that they used them to tame cattle. One
old fellow, Johannes Masindi, about 60 years old, born and bred on the
farm Happy Rest, told me the following: Bored stones are used to tie
up obstreperous cows when they are milked. A twisted or plaited
rope is passed through a hole pierced in the cow’s nose, and tied back
on the head behind the horns. When the cow has to be milked the
rope is fixed with a peg. Where she is to stand a thong is tied to the
rope, the bored stone is kept near by, the end of the thong is threaded
through the hole, and the cow tied up as tightly as possible. If the
cow pulls, she hurts her nose, and has to keep still.

“In addition my informants told me that whenever a young ox
has to be broken to the yoke, one of the bored stones is hung about
his neck. He is allowed to run the whole day like this, and thus learns
to carry a weight, for which a yoke is substituted later... .Up
to the present I have not been able, following Mr. Malan’s request, to
discover anything from any old Venda woman on the subject.”

This is obviously a secondary use, more recent than the Bantu
use of the yoke and the use of the ox as a beast of traction, both

The Bored Stones of South Africa. 201

of which came in as European or Indian innovations within the last
century.*

THE USERS.

We have noted above that Mr. Barnard attributes at least some of
the bored stones of the Transvaal to the BaRoko and BaKoni tribes,
and perhaps to the MaTlala and BaTau, all of whom “were in the
country long prior to the arrival of the BaPedi,”’ and some of whom
were subject to this later folk. Mr. Barnard at the same time states
that the Pedi agriculturalists used iron hoes made for them by the
BaRoko. We may presume therefore that the digging-stick stone
is not Pedi.

The very definite ancestral and sexual symbolism which the Ba Venda
attach to old native iron hoes or to rings made from ancient hoe-
heads + proves that the Venda regard the iron hoe as sacred to the
ancestors. In addition, while certain stones are similarly sacred to
the ancestors, these are not bored, nor are they in any way linked with
the hoe. We may thus exclude the Venda peoples as users of the
bored stone.

Professor van Riet Lowe has found these implements (sites 32 and
33) in definite association with Bantu terrace agriculture. We may
presume that the Welgevonden (26) material, from a nearby farm,
south of the Komati valley, can be similarly associated. In Rhodesia
the association between bored stones and terrace agriculture is less
exact, and Hall has suggested that the bored stone from the Zimbabwe
region belonged to that cult, and that we are dealing with a single but
significant element of the Zimbabwe culture. No specimens appear
to have been reported in true association with any sites of this culture-
complex, either in Rhodesia or the Transvaal, though carved soap-
stone dagga-pipes have been reported.{

While specimens are sometimes attributed to the “Secocoeni,’
this term is generally applied to include the natives of the vague
area about Lydenburg, once known as Sekukuniland, and agrees quite
well with Mr. Barnard’s attribution and with the known distribution
of specimens. In the same way, many specimens have been called
““Makatees” (MaNtatisi); here again we are dealing with a mixed
group of broken-down tribes, with a nucleus of BaTlokwa, who
followed the chieftainess MaNtatisi in her amazing sweep half across

b

* Goodwin, Communication has been Established, Methuen, 1937, p. 76.
+ H. A. Stayt, The Bavenda, Oxford, 1931, pp. 242-248.
{ #.g. Randall-Maclver, Medieval Rhodesia, London, 1906, p. 58.

202 Annals of the South African Museum.

the southern continent. In the turmoil of refugees, conquerors, and
renegades so typical of the beginning of last century, remnants of any
Transvaal tribe might be found among her followers.

THe MaNtTatist on TLokwa MIGRATION.

The story of this extraordinary tragedy may be followed in works
by Ellenberger and MacGregor,* by Broadbent,+ Robert Moffat,
and Thompson.§ A final footnote to the story is to be found in the
letters of Lady Duff Gordon.

The BaTlokwa were a tribe originally allied to the BaPedi, in the
Magaliesberg area of the central Transvaal. The tribe split into
three groups, eventually known as the BaMoKhalong, BaMoKotleng,
and the BaMaLakeng, and migrated to the Wakkerstroom district
in the south-eastern Transvaal. During the Sekoboto year (the Year
of Famine, 1803) these groups were constantly at war, and ended by
having no grain or foodstuffs. In the course of these troubles Mokotjo
had made himself chief of the Kotleng branch, and took as his queen
the daughter of the chief of the BaSia, immediately to the south.
She was his first cousin, MoNyalue. When her first child was born
it was a daughter, Ntatisi, so according to custom the mother took
the name MaNtatisi. That name was eventually to terrorise the
central plateau of the northern Free State. To her followers she was
known as “‘Mosayane,” the little woman. She bore her husband a
son, and in 1813 Mokotjo died, leaving the son an infant; MaNtatisi
assumed the regency though still a young woman. She ended by
becoming the absolute ruler in the face of strong opposition from her
people. The Tlokwa were then living on the immediate west of the
Drakensberg, in the Harrismith district near the Elands River, an
affluent of the Wilge River, and the population has been roughly
estimated as about thirty thousand.

In 1822 began the Lifaqane wars, during which the central plateau
of the Free State was invaded in turn by the AmaHlubi, AmaNegwane,
BaTlokwa, AmaNdebele, and AmaZulu. At the same time the
Hottentot Griquas and Koranas grew restive, and they too invaded
part of the same plateau.

South of the Tlokwa lived the BaSia, south-west lived the BaFokeng

* History of the Basuto. London, 1912.

+ Rev. S. Broadbent, Narrative of introduction of Christianity among the
Barolong. London, 1865.

t Rev. R. Moffat, Missionary labours and scenes. London, 1842.

§ G. Thompson, Travels and adventures in Southern Africa. London, 1827.

The Bored Stones of South Africa. 203

and other tribes of Sotho extraction, who covered what is now the
Free State and Basutoland. Most of these tribes are known to have
been in contact and alliance with local Bushman tribes. When the
AmaHlubi chief Pakalita crossed the Drakensberg from the turbulent
east, the surprised Tlokwa fell back on the BaSia, and in a fit of
quixotism refused to form an alliance as there was not enough food
for both tribes. They fell back farther west, and overwhelmed the
Fokeng, intending to settle there; but Pakalita was fast on their
track, and in his turn he was being followed by the victorious Ma-
tuwane. The Tlokwa turned and fought, and managed to divert the
Hlubi southward across the Caledon River. The Tlokwa followed
them and attacked Buthabutha, the stronghold of the young chief
Moshweshwe, who was eventually to draw the Basuto people together
into a solid nation. Moshweshwe fought a good rearguard action,
and retreated into the Basutoland mountains, where he lived with a
friendly Bushman tribe.

The Tlokwa now moved down the one bank of the upper Caledon
River, while Pakalita and his Hlubi took the opposite bank, each
waiting an opportunity to attack the other. After inflicting a defeat
on the Hlubi and incorporating members of that tribe, the Tlokwa
turned northward once again, defeated a few more tribes, and returned
once again to the upper Caledon. Backwards and forwards this extra-
ordinary migration of human termites moved: eating everything that
came their way, increasing their numbers with conscripts and prisoners,
and depleting their numbers again as the old and the weak fell by the
wayside.

The Tlokwa eventually crossed the middle Vaal in January 1823,
and engaged the BaRolong on the banks of the Bamboesspruit on the
eighteenth of that month. The Tlokwa were checked and weakened
at the battle of Tsabalira, and turned south and west as a hopeless
rabble looking for some weaker tribe from whom they could recuperate
their supplies and warriors. On 26th June they met their Moscow
at Takun, between Kuruman and Litaku, where they were routed by
60 mounted and armed Griqua under their chief Waterboer, and
pillaged by 2000 cowardly BaTlaping. At that battle the Tlokwa
numbered between 40,000 and 50,000 men, women, and children.

The whole story of this last march can be readily constructed from
the Journals of Broadbent, Thompson, Moffat, Andrew Smith, and
finally Lady Duff Gordon’s letters from the Cape. Lady Gordon,
writing in 1862, says that “the Mantatees . . . are scattered all over
the Colony,” and she specifically mentions their presence at Caledon

204 Annals of the South African Museum.

and at Genadendal, both less than a hundred miles from the Cape of
Good Hope. Scattered descendants of these people are still to be
found in the south-western districts of the Cape, though many of them
have forgotten their immediate ancestry, and have been merged into
the heterogeneous “‘Cape Coloured.’”’ I have even heard it said that
some call themselves ‘“‘Griquas,”’ having taken the name of their last
conquerors. In the neighbourhood of Malmesbury, Stellenbosch,
and the Cape Hinterland generally, the name “Makatees” was
regarded as a guarantee of a good and willing farm-labourer up to a
few years ago.

The route from the Bamboesspruit near Wolmaransstad to Postmas-
burg, south. of Kuruman, is adequately covered by the list of sites
given earlier (p. 194). Whether the bored stones belong to the
Tlokwa nucleus or to tribal refugees incorporated along the road does
not really matter. The stones do not fit any prehistoric series, but
fit excellently as extensions of Bantu graphs and histograms, and they
agree with the final route of the defeated army. The average diameter
is 14-4 cm., but the height is very variable. Several examples have
been bored withiron. The proofs of the association of these examples
with the Tlokwa would be threefold: the discovery of numbers of
bored stones of these types on the Bamboesspruit, a task that might
well be undertaken from Potchefstroom; the discovery of similar
stones in the Takun area, research best left to the McGregor Museum,
Kimberley; and finally a study of the south-eastern Transvaal with a
view to discovering what types of bored stone can be associated with
the Tlokwa of Wakkerstrom district, a task best left to Pretoria or
Johannesburg.

I have purposely refrained from raising the question of the possible
affinities of the largest elements in the Western Bypass (Prehistoric
Series) with this movement, although the IB along that route is
generally high; the possibilities should not be completely overlooked.
The routes are partially the same, and it is highly possible that my
analysis here is neither final nor correct in all its details. It depends
upon the evidence available to me. Association must help us here.

The consequent logical question, as to whether our Prototype 1
in the Prehistoric Transvaal Series is truly Bushman or is in fact
Bantu (or even both), would follow; and finally the possibility that all
large specimens, wherever they may be found, are suspect. These
questions areignored. To presume to attempt any answer here would
indeed be pyramiding hypotheses, and these questions can only be
solved in the light of association and clear thought.

The Bored Stones of South Africa. 205

The northern Free State, where the pendulum-swing up and down
the Caledon River took place, may well be adequately represented in
the National Museum, Bloemfontein, but as Dr. van Hoepen himself
proposes to work on his Free State material, I have been unable to
gain access to those sources.

We now come to more contentious ground: the further spread of the
scattered MaNtatisi movement from Takun. Along the route south-
ward several large bored stones occur, and were included in the Winter-
veld Prehistoric Series. To cite a few, the two largest Britstown
specimens have a low IB (34-1 and 41-3 per cent. respectively) while
two beads in the immediate neighbourhood, from Middelwater, have
certainly been bored with metal. At Victoria West the largest
example has an IB of 36-7 per cent., while at Three Sisters the large
specimen has an IB below 40 per cent., and so it goes on. When the
Cape is reached we find two outstanding examples, one from Stellen-
bosch (IB 18 per cent.), the other from Hout Bay (IB 30 per cent.).
Both have quite certainly been bored with metal. The Hout Bay
specimen (16-0 x 11-0 cm.) agrees with the Transvaal series, but the
Stellenbosch example only measures 12:0 cm.x11-0 cm. In addition
the only stone comparable with the Hout Bay example is also
from Stellenbosch, though it was almost certainly bored with stone
(IB 53-5 per cent.).

There is thus some reason to believe that a trail was laid by the
defeated armies of MaNtatisi from the south-western Transvaal to
the Cape and perhaps elsewhere. Much remains to be done, but I
sincerely hope that this will prove a fertile field for future ethnological
research, though clear association between Bantu and bored stone
outside the Transvaal is needed. Careful and detailed research must
eventually lay bare a story at which I can only hint.

SUMMARY OF CONCLUSIONS.

1. We may accept as Bantu all circular bored stones, from the
general neighbourhood of the Transvaal, with a diameter greater than
15-1 cm. In addition we must include certain stones associable with
these; stones known to have been used by the Bantu, or those
associated with Bantu sites and ruins; and finally, speculative
examples that fall well outside the normal graphical scatter and type-
groups of the local Prehistoric Series.

2. The large bored stones are too heavy to have been used on a
wooden hoe, though their use on an iron-headed hoe is not precluded.

206 Annals of the South African Museum.

Mpondo evidence shows that a spatulate digging-stick, used kneeling,
was their substitute for the hoe. It is suggested that the Transvaal
substitute was similar. Similar evidence is known from Abyssinia.

3. There is evidence that these peoples were essentially cattle-
keepers, as their distribution corresponds inversely with that of the
tsetse-fly.

4, These stones can be attributed to various Bantu-speaking peoples,
among them we can include such tribes as the BaRoko, the BaKoni,
MaTlala, BaTau, and BaTlokwa, and possibly the BaSia. We can
exclude the BaPedi and BaVenda. Dart further associates examples
at Heilbron and Bethel with metal-using peoples, while Lowe associates
them at sites 32 and 33 with terrace agriculture.

5. In Southern Rhodesia we may assume that they have been used
by an analogous Bantu-speaking people, possibly with terrace agri-
culture. There is no reason to associate normal bored stones with the
Zimbabwe complex. Even if the famous “‘ Zimbabwe area” stone
were so associable it is abnormal and had some very different function.

6. While isolated bored stones may well have been used variously
as tuyeres, wire-stroppers, etc. and for controlling restive cows, the
stones themselves show no such usage, and there is reason to believe
that these uses were incidental, not reflecting the original function.

7. The use of bored stones to decorate cattle horns is precluded by
weight and bore. Their use for training young oxen to the yoke must
post-date the introduction of the yoke itself from Europe or India.

8. There is as yet no reason to believe that the Bantu made stone-
headed knobkiries. Stone beads are attributable to a metal-using
people.

9. The use of bored stones as dagga-pipes is mentioned in Appendix
I, and the possibility that unfinished pipes may sometimes be included
as bored stones is noted.

10. The number of elongated Bantu examples so far available is
too small to permit of any deductions as to their relationship to one
another, or to the circular specimens.

11. An Index of Bore is developed, and reasons given to assume that
stones (over 3-5 cm. in diameter) in which the mouth of the hole is
less than 30 per cent. of the total thickness pierced have been bored
with metal, probably the tang of an axe or hoe-head.

12. The thickness of the stones pierced and the methods of boring
are much the same for the Transvaal prehistoric and Bantu specimens.
Cylindrical bores are attributed to usage.

13. The known association of Sotho with Bushmen at the beginning

The Bored Stones of South Africa. 207

of the last century cannot be ignored. This appears to show itself
in parts of the Transvaal through the conical boring of prehistoric
types with metal tools, and the similarity between Bantu and pre-
historic Transvaal histograms of IB. The partial contemporaneity
and contact of Bushmen and Bantu necessarily follows.

14. The history of the MaNtatisi movement in the first quarter of
the last century is outlined; and the incidence of stones, often bored
with metal and analogous to Bantu types, is noted along the final
stages of their migration.

15. Speculations are raised on the possible relationship between the
final scattering of the MaNtatisi and certain metal-bored examples in
the Province of the Cape of Good Hope. The relationship of these
movements with that outlined in the prehistoric “Western Bypass”’
is not raised, nor the deductions that might logically follow.

208 Annals of the South African Museum.

STORMS RIVER
Eo ONE DAGGA PIPES

SMITHFIELD

DISTRIBUTION OF
EDGED DISCS

BELOW 3-0 CMS. BORE.®
OVER 4°4CMS. BORE SO

Fia. 9.—Three stone pipe bowls for smoking hemp (above). Distribution of edged discs (below).

The Bored Stones of South Africa. 209

APPENDIX I.
Tut BorepD STONE AS A DaAGGa-PIPE.

Mr. Barnard’s statement in the paper quoted earlier, that the bored
stone was used by the BaRoko “as a bowl for their dagga-pipes,”’
is a misleadingly bare presentation of the facts. Bored stones of a sort
were certainly used as dagga (hemp, chanvre, Cannabis sp.) pipes by
both Bantu and Hottentot peoples, and are still so used, but the pipe
bowl is somewhat differently made. While authenticated examples
occur as far afield as Barotseland and the Cape, and are in use in
Pondoland, Basutoland, Zululand, and elsewhere, nothing has so far
been done on their typology or distribution. Material might well
be based upon examples confiscated in magistrates’ courts. Their
manufacture in Southern Rhodesia may go back a few centuries, and
Randall-Maclver mentions an example found in a midden of Zimbabwe
culture. But even if the hemp was originally introduced by the
Arabs, there is no evidence that this occurred early, or that the method
of making pipes was ever anything but an adaptation of the usual
technique used in making the bored stone.

A study of both complete and incomplete specimens is needed. I
here illustrate three, all markedly different. These show that the
section of the bore does not resemble that of the normal bored’ stone.
There are here two portions: the deeper hole provides the bowl of the
pipe, and the shallower is only wide enough to permit the insertion
of a reed stem. These two are joined by a narrow aperture, about
0-5 cm. in diameter. The holes are more tubular, and the examples
given all suggest that the hole was enlarged and shaped, after the two
pecked depressions met in the heart of the stone, by tooling. There
is thus a marked step in the sections, towards the aperture.

The two larger specimens have had a pebble origin, and were
originally far larger than they are now, but have been rubbed and
ground to shape. The Smithfield example is an almost perfect barrel
shape, thinning slightly to one end. The Mpondo specimen is less
regularly ground, and was probably rubbed down on a flat stone,
so that the remnants of faceting are visible down the outer face of the
cylinder.

The pebble was therefore bored, then ground laboriously to shape,
and finally the interior of the bowl was dug out with some metal
tool until it had a reasonable capacity. To have shaped the stone
before boring would have been fatal.* The Smithfield specimen (sent

* Since writing this note, a dagga-pipe, in process of manufacture, has come
to my notice. It is from Weenen, Natal, and was shaped to a “‘Christmas cracker”
shape (a cylinder with two waists) and had broken directly boring had been
commenced.

210 Annals of the South African Museum.

by Dr. Kannemeyer to the South African Museum) is the most
perfect, though the example from Hluleka on the western Pondoland
coast is good. Both these are heavily impregnated with black hemp-
juice from use. Known dagga-pipes are ignored in the body of this
paper. A few thistle-shaped examples are known to exist among the
Zululand tribes, probably more closely related to the Barotseland
forms. These occur (I believe in both areas) in stone or in pottery.

These various forms are not exclusively Bantu, and are known to
have been in use among the Hottentots, though there is not the
shghtest evidence that their use was earlier than the historic period
in South Africa. There are two other forms of pipe in use among the
Bushmen and Hottentot peoples, the “trap-drain”’ form, used exclu-
sively for dagga, and the “cigar-holder”’ type which seems to have
come in with tobacco, and is not suited to hemp at all.

APPENDIX II.
EpcEep Discs anp ARM-RINGS.

The figures and available data on these examples of the bored stone
have been published elsewhere,* and the reader is invited to turn to
that reference for a complete survey of the question, as I have no
intention of repeating the body of that paper here.

The edged disc is a discus-like stone, bored in the centre, and either
planoconvex or biconvex, so that a polished edge forms the outer
boundary of this circular stone. This has in certain instances been
reamed out in such a way as to increase the central hole to dimensions
that suggest its use as an arm-ring. In only one instance could the
stone have been used as an arm-ring for an adult. It was found that
two types could be differentiated; those with an aperture under 3 cm.,
and those with an aperture greater than 4-4 cm., and the list of sites
was given in that paper. Figures for two further specimens have
come to hand; one from Boetsap falls into the small-bored series, and
another from a midden site at the Bashee River mouth falls into the
large-bored series. The accompanying map, not hitherto published,
includes these two. The small-bored examples are marked as dots,
the large-bored as circles.

* A. J. H. Goodwin, South African Journal of Science, vol. xl, pp. 296-302,
1943.

CORRIGENDA.

Map 2, p. 110. For Colesburg read Colesberg.
Map 4, p. 112. for Fullerton read Fullarton.

PARTS OF THE ANNALS PREVIOUSLY ISSUED—

Vols. I, 11 and VIIL.—Out of print. :
Vol. 11I1.—Part 1, 2/6; Part 2, 1/6; Part 3, 5/6; Part 4, 3/-; Part 5,
5/6; Part 6, 7/-; Part 7, 1/6; Part 8, 3/-; Part 9, 1/6; Index,
Title, etc., 1/6.
Vol. IV (containing Palaeontological papers published in conjunction with the
Geological Survey).—
Part 1, 11/-; Part 2, 7/-; Part 3, 4/6; Part 4, 4/6; Part 5, 2/6;
Part 6, 4/6; Part 7, 14/-; Part 8 (with Index, Title, etc.), 8/—.
Vol. V.—Part 1, out of print; Part 2, 8/6; Part 3, 2/6; Part 4, 1/6; Part 5,
2/-; Part 6, 5/-; Part 7, 3/-; Part 8, 4/6; Part 9, out of print;
Index, Title, etc., 1/6. ;
Vol. VI.—Part 1, 13/6; Part 2, 4/6; Part 3, 3/6; Part 4, 30/—; Index, Title,
Cte. Oo:
Vol. VII (containing Palaeontological papers published in conjunction with the
Geological Survey).—
Part 1, 3/-; Part 2, out of print; Part 3, 5/-; Part 4, 8/-; Part 5,
5/6; Part 6, 1/6; Index, Title, etc., 1/6.
Vol. [X.—Part 1, out of print; Part 2, 5/6; Part 3, 10/-; Part 4, 6/6; Part
5, 3/6; Part 6, 11/-; Part 7, 9/-; Index, Title, etc., 1/6.
Vol. X.—Part 1, 3/-; Part 2, 2/6; Part 3, 2/-; Part 4, 3/-; Part 5, 20/-;
Part a 3/-; Part 7, 10/-; Part 8, 2/6; Part 9, 5/-; Part 10, 2/6;
Part 11, 20/-; Part 12, 7/-. Complete. Title, etc., contained in
Part 11.
Vol. XI.—Part 1, 3/6; Part 2, 2/-; Part 3, 13/6; Part 4, 1/6; Part 5, 17/-;
Part 6, 1l/-; Part 7, including Index, Title, etc., Appendix and
Plate III, 3/-.
Vol. XII (containing Palaeontological papers published in conjunction with the
Geological Survey).—
Part 1, 15/6; Part 2, 3/6; Part 3, 4/6; Part 4, 3/-; Part 5, 7/-;
Part 6, 6/—; Part 7, 20/-3 Part 8, 20/—; Index, Title, etc., 1/6.
Vol. XII.—Part i 6/-; Part 2, 2/6; Part os 3/-; Part 4, 8/6; Part 5, 1/6;
Part 6, 5/—; Part 7, 30/-; Part 8, 1: Index, Title, etc., 1/6.
Vol. XIV.—Part 1, 8/6; Part 2, 8/-; Part 3, 6/-; Part 4, 17/6; Part 5, 5/-;
Part 6, 9/-; Index, Title, etc., 1/6.
Vol. XV.—Part 1, 17/-; Part 2, 17/-; Part 3, 14/-; Part 4, 12/6; Part 5, 5/6;
Part 6, 3/6; Index, Title, etc., 1/6.
Vol. XVI.—Part 1, 30/6; Part 2, 4/-; Part 3 (with Title, Index, etc.), 25/-.:
Vol. XVII.—Part-1, 12/-; Part 2, 9/6; Part 3, 3/-; Part 4, 15/—; Part 5, 15/—
Part 6, 2/6; Index, Title, etc., 1/6.
Vol. XVIII.—Part 1, 20/-; Part 2, 7/6; Part 3, 30/-; Part 4, 12/6; Index, Title,
etc., 1/6.
Vol. XIX.—Part 1, 22/-; Part 2, 17/6; Part 3, 11/-; Part 4 (with Index, Title,
etc.), 5/6. | :
Vol. XX.—Part 1, 8/6; Part 2, 12/6; Part 3, 4/-;. Part 4, 10/-; Part 5, 4/-;
Part 6 (with Title, etc.), 4/6.
Vol. XXI.—Part 1, 25/-; Part 2 (with Title, etc.), 30/-.
Vol. XXII (containing Palaeontological papers eabished in ooaleee with tthe
Geological Survey).—
Part 1, 20/-; Part 2, 10/-; Part 3 (with Title, etc.), 3/6.
Vol. XXITI.—Part 1, 12/6; Part 2, 8/-; Part 3 (with Index, Titie, ete.), 9/-.
Vol. XXIV.—Part 1, 10/-; Part 2, 2/6; Part 3, 4/6; Part 4, 11/-; Part 5 (with
Index, Title, etc.), 16/-.
Vol. XXV.—Part 1, 12/6; Part 2, 10/-; Part 3 (with Index, Title, etc.), 10/-.
Vol. XX VI.—Complete, 25/-. .
Vol. XX VII.—Complete, 25/-.
Vol. XXVIIT (containing Palaeontological papers published in conjunction with the
Geological Survey).—
Part 1, 10/-; Part 2, 17/6; Part 3, 6/-; Part 4 (with Index, Title,
etc.), iO
Vol. XXIX.—Part 1, 20/-; Part 2 (with Index, Title, etc.), 20/-.
Vol. XXX.—Part 1, 15/6; Part 2, 20/-; Part 3, 8/-; Part 4, 9/-; Part 5 (with
Index, Title, etc.), 9/-.
Vol. XX XI (containing Palaeontological papers published in conjunction with the
Geological Survey).—
Part 1, 20/-; Part 2, 13/6; Part 3, 17/6.
Vol. XX XII.—Part 1, 2/6; Part 2, 7/6; Part 3, 17/-; Part 4, 1/-; Part 5, 20/-;
Part 6 (with Index, Title, etc.), 10/-.
Vol. XX XITI.—Part 1, 35/-.
Vol. XX XIV.—Complete, 40/-.
Vol. XXX VI.—Part 1, 9/6; Part 2, 10/-; Part 3, 5/-; Part 4, 3/-.
Vol. XXX VII.—Part 1, 26/-.

INDEX, Subjects and Authors, to Vols. I-X XX, 1/-.

The Annals of the South African Museum will be issued at irregular intervals,
as matter for publication is available.

Copies may be obtained from—
Messrs. WHELDON & WESLEY, Lrp.,
721 Norra CrracutarR Roap, WILLESDEN, loner N.25 of,
The LIBRARIAN, Soura Arrican Muszum, Carr Town.

Bo, 68

ANNALS

OF THE

SOUTH AFRICAN MUSEUM

VOLUME XXXVII.

PART II, containing :-—
2. Cramometric Survey of the South African Museum Collection
of Bushman, Hottentot and Bush-Hottentot Hybrid Skulls.—
By J. A. Keen, Department of Anatomy, University of
Cape Town. (With Plates I-IV.)

ISSUED AUGUST 1952. PRICE 10s.

PRINTED FOR THE
TRUSTEES OF THE SOUTH AFRICAN MUSEUM, CAPE TOWN
. BY NEILL AND CO. LTD.,
212 CAUSEWAYSIDE, EDINBURGH.

Li

Genie)

2. Craniometric Survey of the South African Museum Collection of
Bushman, Hottentot and Bush-Hottentot Hybrid Skulls.

By J. A. Keen,
Department of Anatomy, University of Cape Town.

(With Plates I-IV.)

THE collection of crania in the South African Museum representing
the aboriginal peoples of the Union was started at a time when Mr.
W. L. Slater was the Director (1896-1906), and was continued by
Dr. L. Peringuey (1906-24), under whose directorship the collection
grew to considerable dimensions. In 1923 Dr. Peringuey decided to
submit the collection, about a hundred complete crania and other
skeletal material, to Professor H. Pittard, the eminent anthropologist,
of the University of Geneva. Professor Pittard undertook to make
a detailed study, and for many years this valuable Bushman and
Hottentot material remained in Switzerland.

During the directorship of Dr. E. L. Gill (1925-42) the Bushman
and Hottentot anthropological collection continued to increase
through gifts of accidental finds and also through the exhumation
of graves, as occasion arose. In 1939 a portion of the skeletal material,
originally sent away by Dr. Peringuey, returned to Cape Town, the
remainder not until 1948, owing to the outbreak of the Second World
War. The whole collection is now again housed in the South African
Museum and consists of 205 adult crania, 121 of these with mandibles.
I am indebted to Dr. K. H. Barnard, the Director of the South African
Museum, for permission to study these crania and to present a general
survey. Hach skull was measured, following a standardised method
which had previously been applied to groups of European, Bantu
and Cape Coloured crania, and also to a very similar collection of
Bush-Hottentot material in the Nasionale Museum, Bloemfontein
(Keen, 1947).

Previous Studies—A subgroup of the collection, viz., 53 crania
exhumed from graves near Colesberg in the Cape Province, was studied
by Slome (1929) and the results published under the title of the
VOL. XXXVII, PART 2. 14

212 Annals of the South African Museum.

““Osteology of a Bushman tribe”. During the time that the Peringuey
collection of skulls was in Switzerland, Professor Pittard and his
co-workers published many articles bearing on special points in
craniology which interested them. However, a general craniometric
survey of the collection, as a whole, had not been made previously.

%

METHODS.

The skull measurements.—The list of skull measurements which was
selected is seen in Table I. It includes all the usual cranium and face
dimensions and indices which are employed in physical anthropology.
The first seven measurements and four indices concern the vault of the
skull. The remainder concern the facial part of the skull, and are
designed to study the configuration of the orbit, nose and mandible.
Most of the measurements were made according to the instructions
given in R. Martin’s text-book (1928), but a few require special
mention.

The length of mastoid process is obtained by drawing a line across
the root of the mastoid process parallel to the Frankfort plane, at the
level of the upper margin of the suprameatal triangle. On that line
take a point which is vertically above the tip of the mastoid process,
and then measure the distance between that point and the tip
of the mastoid process. If the suprameatal triangle is ill-defined,
the level of the upper margin of the external auditory meatus is
used.

The wnterorbital interval (“Zwischenaugenbreite”) is defined by
Martin (1928) as the straight distance between the two dacrya, the
dacryon being the point on the inner margin of the orbit where the
frontal bone, the frontal process of the maxilla and the lacrimal bone
meet. The sutures tend to disappear with advancing age, especially
that between the maxilla and the lacrimal bone, and in many skulls
it is impossible to find the dacryon. On account of this difficulty
I have consistently interpreted the interorbital distance as being the
maximum distance obtainable across the bridge of the nose when the
caliper points are placed on the upper part of the lacrimal bone, or on
the fronto-lacrimal suture, if this is visible.

The minimum ramus breadth is the smallest distance obtainable
between the front edge and the posterior border of the ramus of
the mandible. The monimum ramus height is the distance between the
lowest point of the mandibular notch and the lower border of the
mandible, the measurement being made along a line at a right angle
to the line of the minimum ramus breadth measurement. These
two mandible dimensions and an index obtained from them have
been shown to have a special significance in the differentiation
between Bushman, Hottentot and Bantu crania. The angle at the
nasion is the forward angle formed at the nasion between straight
lines joining nasion to glabella and nasion to rhinion; a simple method
of measuring this profile angle has been described (Keen, 1949).

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 218

Methods of calculation.—The frequency distribution of each measure-
ment was tabulated, after choosing a suitable class interval, and the
means (M) and standard deviations (a) were determined in the usual
manner. In order to compare the data obtained in a related popula-
tion group, calculations were made to determine the critical ratios
between the means of each dimension or index. This calculation
involves first of all the determination of the standard error of the mean

(c), according to the formula e=o/\/n (xn =number of observations).
The standard error of the mean for one group may be written as e¢,,
and for another group as «,. The next step was to find the standard
error of the difference (caig), using the formula ogig=Ve2 +€,2. The
critical ratio is given by the formula:

actual difference between the two means pe M,-M,
Odift VJ ey? + €52

According to the laws of probability, the significance of the numerical
values obtained for the critical ratio is as follows. If the critical
ratio is 2-5, a difference between the two means equal to or greater
than the observed difference is not likely to be found in more than
6 out of 1000 comparisons of groups drawn from similar population
sroups. Hence we are justified in concluding that the observed
difference is almost certainly due to systematic and not to chance
factors; in other words, that the observed difference is significant.
Such a conclusion is all the more justified if the critical ratio exceeds
2-5, the significance becoming more and more important as the
critical ratio increases.

Some observers are prepared to accept critical ratios of 2 or even
slightly less as indicating differences which are statistically significant.
However, in Table III the differences between the means have been
noted as significant only when the critical ratio was 2-5 or more.

critical ratio =

RESULTS.

The results of the analysis of the dimensions and indices of the
205 crania in the collection are presented in Table I. Included in
the series are 12 crania which are labelled Griqua in the Museum
records. A study of the osteology of the Griquas was made by Brink
(1923), who described this population group as a Hottentot clan in
which some European admixture had occurred. On examination,
these Griqua skulls in the collection showed no evidence of such
admixture. Nine out of the 12 ‘ Griqua’ skulls could be classified
unhesitatingly as Bush-Hottentot hybrid skulls. Of the remaining
three, two were very large skulls showing the characteristics of the
pure Hottentot. The third one, although resembling the Hottentot
type in most of its dimensions, had a gnathic index of 104, which
raised the question of a possible Negro admixture.

214 Annals of the South African Museum.

TaBLE ].—SHOWING THE RANGES, MEANS AND STANDARD DEvia-
TIONS OF THE CRANIAL AND FactaAL MEASUREMENTS AND INDICES
OF THE SouTtH AFRICAN MuskumM CoLLECTION OF BUSHMAN,
Hottentot AND BusH-HotTtrentot Hysrip CRANIA.

(205 skulls; measurements in millimetres unless stated otherwise.)

Range Mean +€ o
1. Maximum length 157-200 179-3+ -d (PL:
2. Maximum breadth 119-149 132-34 -4 5-6
3. Minimum frontal
breadth 82-111 94-5+ -3 4-6
4. Basion-bregma height . 112-146 126-64 -4 6:3
Breadth/length
index (2/1) 64-4-86-6 73°94 -3 3:8
Height/length
index (4/1) 60-7-78:-7 (pt se 2 35
Height/breadth
index (4/2) 81-7-113-7 95-94 4 6-0
Transverse fronto-
parietal index (3/2) 62-6-83-4 M154 °3 4-0
5. Horizontal circum-
ference 456-560 502-5 + 1-2 16:7
6. Cranial capacity (c.c.°) . 989-1615 1238-8+47-7 | 109-7
7. Length of mastoid
process 13-36 24-44 +3 4-1
8. Length of base 85-110 96-84 -3 4-9
9. Length of face. 82-115 95-6+ -4 6-0
Gnathic index (9/8) 89-6-108-0 Sleeiee 33) oT
10. Maximum bizygomatic
diameter . 108-142 123-5+ -5 6:8
11. Total face height 82-123 1045+ -7 75
12. Upper face height 50-77 62°32 5-2
Total facial index
(AO) 70-1-99-2 80-4 == 35 5:5
Upper facial index
(12/10) 41-5-60:5 51-04 +3 3°6
18. Breadth of orbit . 30-47 40-9+ -1 2-2
14. Height of orbit 25-37 SIIPIESE ol 2-2
Orbital index (14/13) 63:0-89-7 T6-34 -4 5-1
15. Interorbital interval 18-34 26-2+ :2 3:0
16. Nasal width 20-32 20°00 ee 2°3
17. Nasal height : 36-52 45-44. +2 od
Nasal index (16/17) 46-7-69-4 579+ °3 4-6
18. Minimum ramus breadth 27-45 349+ +3 34
19. Minimum ramus height 32-53 41-24 -4 4-8
Index of ramus (18/19) 62:5-114:3 Oma oY 9-7
20. Angle of nasion 123-169° 151-1+ -6° 9-0

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 215

Further, one must admit that an occasional skull of a Cape Coloured
may have been added to the collection, the cranium being labelled
Hottentot or Bush-Hottentot hybrid. With these reservations in
mind, we may affirm that the ranges, means and standard deviations
of the 20 skull dimensions and 10 indices, seen in Table I, are repre-
sentative of the Bushman, Hottentot and Bush-Hottentot hybrid
peoples, when they are placed together as one entity.

Provenance and Classification of the Crania.

Nearly every skull in the collection had an entry in the Museum
records to indicate when it reached the Museum, who presented it,
and from which locality it came. The skulls originated from an
extensive area which included the whole of the Cape Province, as well
as the territory along the west coast of South Africa which formerly
was the German colony of South West Africa—that is, Ovamboland,
Namaqualand north of the Orange River, and the adjoining Kalahari
Desert. An obvious method of subdividing the crania was on a
geographical basis, as follows :—

(a) Crania which came from the interior of the Cape Province: 71.

(6) Crania which came from Ovamboland, Namaqualand, and the
Kalahari Desert: 63.

(c) Crania which came from localities within a fifty-mile radius of
Cape Town: 25.

(d) Crania from localities situated in the southern coastal strip of
the Cape Province: 19.

With the remaining 27 crania there was no information as to the
place of origin. When Dr. Peringuey submitted the material to
Professor Pittard in 1923, the collection of crania was about half the
size of the present one. Dr. Peringuey divided the skeletal material
partly on a cultural and partly on a geographic basis, and suggested
the following subdivisions: (1) Crania obtained from coastal rock
shelters. (2) Crania obtained from sand-dunes and coastal middens.
Both these groups, according to Peringuey, would represent the
Strandloper race. (3) Crania from the interior of the Cape Province,
representing mainly the inland Bushman. (4) Crania from Ovam-
boland, Namaqualand, and the Kalahari Desert, which were also
said to be mainly Bushman material.

From the letters exchanged with Professor Pittard, which are
preserved in the files of the South African Museum, it is clear that

216 Annals of the South African Museum.

Dr. Peringuey considered that the Hottentots were a subdivision of
the Bantu, the latter being defined as central African Negros who had
undergone a “‘semitic infiltration”. The Hottentots are said to have
invaded South Africa, where they encountered an aboriginal race,
the Bushmen, whom they conquered. The Hottentots mixed with
the Bushmen and as a result their physique became considerably
changed. This was the original view put forward by Stow (1905).
The resultant mixed group settled along the sea-coast, while the
original Bushmen were displaced inland. For the mixed Bush-
Hottentot group, which after a time became relatively homogeneous,
Dr. Peringuey preferred the name Strandloper, no doubt having been
influenced by Shrubsall (1907). This author insisted on the very
distinct racial characteristics of the Strandlopers, who were said to
be a purer group than the Bushmen, and to be distinct from the
Hottentots.

In the numerous anthropological studies by Professor Pittard
and his co-workers, the grouping suggested by Dr. Peringuey was
rigidly adhered to, on the assumption that there was some inherent
fundamental difference. After discarding damaged crania and
juvenile ones, and further subdividing them rather arbitrarily into
male and female, the resultant subgroups became very small (often
less than 12), so that the means of the various dimensions and indices
given in Professor Pittard’s writings were not particularly helpful

A collection of crania of the Bushman and Hottentot population
groups has also been made by the Nasionale Museum, Bloemfontein.
Here the skulls are subdivided into “typical” Bushman, “typical”
Hottentot, and the intermediate Bush-Hottentot hybrid. The
Hottentot material had been obtained in the course of several expedi-
tions to collect old Hottentot skulls from graves along the Orange River
between Kakamas and the Falls, the Hottentot graves being marked
by conical mounds of stones (Dreyer and Meiring, 1937). This Orange
River region was occupied by relatively pure Hottentot tribes within
historical times (Maingard, 1932).

The “typical”? Hottentot cranium is large (mean maximum length:
192 mm.), narrow in the forehead region and markedly dolichocephalic
(mean breadth/length index: 70-5); while the “typical” Bushman
cranium is small (mean maximum length: 174 mm.), relatively wide
in the forehead region and with a tendency to brachycephaly (mean
breadth/length index: 75-8).

An analysis of the skull dimensions and indices of the Bloemfontein
crania (Keen, 1947), and a comparison with the data obtained from

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 217

the present collection, reveals differences in certain directions. But
these differences are easily explained by a preponderance of large
Hottentot crania in the Nasionale Museum collection. In the Bloem-
fontein series the large “typical”? Hottentot crania formed 35 per
cent. of the total number (35 out of 99 skulls), while in the present
collection the “typical”? Hottentot crania are only 19 per cent. of the
total (38 out of 205).

The experience gained with the material in the Nasionale Museum,
Bloemtontein, enabled me to subdivide the present series of crania
into the three groups of Bushman, Hottentot, and Bush-Hottentot
hybrids, and the results were as follows. Among the 205 crania,
46 were small mesocephalic skulls which were called “typical”
Bushman, 7.e. 22 per cent. On the other side were 38 large doli-
chocephalic crania which were called “typical”? Hottentot, 7.e. 19
per cent. The remaining 121 crania (59 per cent.) were classified as
Bush-Hottentot hybrid. The skulls were also divided into male and
female; but little importance is attached to this, because the sexing
of crania, in the absence of other skeletal data, is recognised to be
very uncertain. :

Where the dividing lines are drawn between the characteristic
Hottentot, Bush-Hottentot hybrid, and Bushman is a matter of
choice which rests with the observer. But the recognition of the
tendency in one or other direction, and the definition of the extreme
types at either extremity, are tasks which present no particular
difficulty. 7

If an analysis is made of the two extreme groups, based solely on
the two factors of maximum length and breadth/length index, we
obtain the following results. The 46 Bushman crania in this series
gave a mean for maximum length 172 mm. (range 157-177 mm.),
with a mean breadth/length index of 7S—that is, in the mesocephalic
class (76-0-80-9). The 38 Hottentot crania gave a mean for maximum
length of 189 mm. (range 183-200 mm.), with a mean breadth/length
index of 70—that is, well into the dolichocephalic class (x — 75-9). In
between lay the large group of Bush-Hottentot hybrids. These data
show a close approximation to the criteria established when defining
“typical” Hottentot, “typical” Bushman, and Bush-Hottentot
hybrid crania in the previous study of the Bloemfontein series.

If we accept the method of classifying the present series into the
three types, it becomes of special interest to study their relative
distribution in the various subdivisions based on the geographical
origin of the crania (see Table II). Among the crania originating

218 Annals of the South African Museum.

from the interior of the Cape Province the small mesocephalic Bushman
skull was found more often than the large dolichocephalic Hottentot
cranium. In the material from South West Africa and the Kalahari
Desert the proportions are reversed (Table II). Among the crania
from the southern coastal strip of the Cape Province and near Cape
Town the characteristic Bushman type is often seen, the Hottentot
type much less often. In all three geographical groups the Bush-
Hottentot hybrid skulls are the most numerous, as applies to the
collection as a whole.

TABLE IL.—SHowING THE PERCENTAGE DISTRIBUTION OF THE THREE
TYPES WHEN THE CRANIA ARE ARRANGED ACCORDING TO THEIR
GEOGRAPHICAL ORIGIN.

Bush-Hottentot

Bushman. Hottentot. Heads

From the interior of Lf 13 4]
the Cape Province: 24% 13%, 58%
71 skulls.

From Ovamboland, 9 14 40
Namaqualand, and 14 R 22% 64%,
Kalahari Desert:

63 skulls.

From southern coastal 17 4 2

area and near Cape 39% I 52%

Town: 44 skulls.

Dr. Peringuey had informed Professor Pittard that the crania from
the interior of the Cape Province and those from South West Africa
and the Kalahari Desert were representative of the Bushmen, placing
the hybrid crania, which he preferred to call Strandlopers, along the
coastal strip. Actually the South West Africa and Kalahari Desert
group contains the highest proportion of “pure” Hottentot crania.
Thus we can explain such disturbing statements as that made by
Professor Pittard (1927) to the effect that the Bushmen of the Kalahari
are dolichocephalic. The analysis was based on the examination of
27 crania. The material, as becomes obvious, was largely Hottentot
and Bush-Hottentot hybrid, and not Bushman. The characteristic
Bushman type of cranium appears to be best represented among

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 219

those obtained from middens and rock shelters along the coast and
near Cape Town.

Australoid Type of Hottentot.—A smooth forehead region, with
absent or only slightly marked eyebrow ridges, is characteristic of
Bushman and Hottentot crania. Occasionally the glabella is
prominent and continuous at the sides, with definite supra-orbital
ridges. Among the 205 crania there were 7 (3 per cent.) where this
feature was noted. Of the 7 skulls with marked supra-orbital ridges,
3 were large crania which had been called Hottentot, and the other
4 had been grouped as Bush-Hottentot hybrid.

Well-marked supra-orbital ridges were described by Broom (1941)
as occurring among the Korana. In order to explain the existence of
such australoid crania among them, it was postulated that an unknown
australoid type, now quite extinct, mixed with the Korana race. The
Korana are considered by Broom to be a purer type than the Hottentot.
The latter, according to this author, were a hybrid race resulting from
the crossing of Bushmen and Koranas. It has been argued elsewhere
(Keen, 1943) that the Koranas should be considered to be a tribe of
the Hottentots, on historical grounds and also because their cranial
dimensions conform closely to the criteria established for the Hottentot

group.
Comparison between Bushman and Hottentot Crania.

Differences—Many of the differences between the “typical”
Bushman and the “typical” Hottentot are accounted for by the fact
that the Hottentot skull is considerably larger than the Bushman
skull, and this is mainly shown in the dimension of maximum length.
As regards maximum breadth, the Bushman cranium is relatively
wider than the Hottentot one. In consequence, the breadth/length
index is a larger figure in the Bushman, bringing these skulls into the
mesocephalic class (76-:0-80:9), while the Hottentot skulls are decidedly
dolichocephalic (Plate ITT).

The facial part of the skull, even more than the vault, reflects the
difference in general size. Of the two facial dimensions, total face
height and upper face height, the latter is relatively more affected.
Accompanying the smaller upper face is a reduced nasal height in the
Bushman. The Bushman face, as compared with the Hottentot, is
compressed from above downwards, and this reduction in height
affects more particularly the part of the face between the nasion
and the prosthion. Both facial indices are smaller in the Bushman,
giving the face a shape which tends to be somewhat oblong from

220 Annals of the South African Museum.

side to side, while the Hottentot face is more square-shaped. The
nasal index, on the average, shows a higher figure in the Bushman, this
being due to a relatively greater reduction in nasal height than in
nasal width, as compared with the Hottentot. The Bushman is more
platyrhine than the Hottentot (Plate I).

Resemblances.—Both Bushman and Hottentot crania have a low
vault of the skull, the mean for basion-bregma height being 126 mm.,
as compared with a mean of 133 mm. in a group of Bantu crania.
Bush-Hottentot crania, as a rule, have a short mastoid process, and
this is very evident if they are contrasted with HKuropean or Negro
crania. Bushman and Hottentot crania are orthognathous with a
mean for the gnathic index of 98-7, as compared with a mean of 101-4
for the Bantu (Plate II).

The orbital index is small, on an average, in Bush-Hottentot skulls,
indicating that the orbital aperture tends to be narrower from above
downwards. The zygomatic bones are prominent and push the
inferior orbital margin upwards to some extent. The prominent
cheek-bones in the faces of the South African aboriginal peoples was a
feature which was noted by the early Dutch settlers. The narrower
orbital aperture in a vertical direction contrasts with the more rounded
orbital aperture in European and Negro crania.

An important characteristic of the Bushman and Hottentot crania
is seen in the shape of the ramus of the mandible, which tends towards
asquareshape (PlateIV). Thisgives a high index of ramus, in contrast
with the more elongated ramus of the Negro and Huropean mandible
with a much lower index. But perhaps the most striking and im-
portant character seen in the facial part of the skull is the “flatness”
of the nasion region, so very obvious in all Bushman and Hottentot
crania (Plate I). A flat bridge of the nose goes pari passu with a large
profile angle at the nasion; the mean for the whole group of crania
was 151° (Table I). The mean for the angle at nasion in 220 Bantu
crania was 138°, and in 186 European crania it was 127°. The flatness
of the nasion region is seen typically in Mongolian skulls, and the
Bushmen and Hottentots further share skin colour, the Mongolian
fold at the inner margin of the eye, and the prominent cheek-bones
with certain Asiatic population groups.

The factor of “flatness” of the nose appears to be linked with a
wide wnterorbital interval, which was stated by Shrubsall (1907) to be
a characteristic of the Bushman cranium. A priori, one expected that
the factor of “flatness” of the nose, as expressed numerically by the
size of the profile angle at the nasion, would be positively correlated

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 221

with a wide interorbital interval. But when the coefficient of correla-
tion was calculated on the data provided by these 205 crania, no
correlation was found to exist between the two factors.

The difficulties of measuring the interorbital interval have been
referred to previously. When the interorbital interval was measured
as described earlier, the various racial groups showed very slight
differences in the means: 26 mm. in the present series of Bush-
Hottentot skulls, 27 mm. in the Bantu crania, and 25 mm. in the
Kuropean series, in spite of the very large differences between them
in the means for the angle at nasion. The nasal bones and the frontal
processes of the maxillae are flattened out, and the bridge of the nose
itself is wider in the Bush-Hottentot crania. But there is no corre-
sponding increase in the distance between the front parts of the inner
walls of the right and left orbits.

The various points of difference between Bushman and Hottentot
crania, and the characteristics common to both groups, can be seen
to a greater or less extent in the photographs of three representative
crania (Plates I-III). Only one of these had a mandible available,
and therefore they are shown without the lower jaw.

Comparison with the Cape Coloured Crania.

Of the present-day population groups in the Union of South Africa
the Cape Coloured people are the most closely allied to the aboriginal
Bushmen and Hottentots; the two latter population groups are
becoming extinct. It seemed of interest to make a comparison
between the crania of the aboriginal peoples and those of the present-
day Cape Coloureds. The data for the Cape Coloureds came from a
recent craniometric study (Keen, 1951) in which male and female
crania were analysed separately. For purposes of this comparison
the midway mean between the male and female means was used, on
the assumption that in the Bush-Hottentot crania of the S.A. Museum
collection we were also dealing with male and female in equal
proportions.

The means for the various dimensions and indices of the two series
of crania are shown in Table III, together with the results of calcula-
tions of the critical ratios of the differences in the means. A glance
at the column of the critical ratios shows at once that the significant
differences affect the dimensions and indices of the facial part of the
skull much more than those of the vault part. The importance of the
facial part of the skull in racial differentiation has been stressed by

222 Annals of the South African Museum.

TABLE IIJ.—SHOWING THE MEANS OF THE VARIOUS DIMENSIONS AND
INDICES IN THE BusH-HoTTENTOT CRANIA OF THE S.A. MusEuM
COLLECTION SIDE BY SIDE WITH THE SAME MEASUREMENTS IN
CaPpE COLOURED CRANIA, TOGETHER WITH THE CRITICAL RATIOS
OF THE DIFFERENCES, PROVIDED THESE WERE 2:5 OR MORE.

(Measurements in millimetres unless stated otherwise.)

S.A. Museum | Cape Coloured Critical
Collection Group Ratio of
(205 skulls). (201 skulls). Difference.
Maximum length : 179-3: -5 S023 =e 2°5
Maximum breadth 1382-34 -4 134-14 -;
Minimum frontal
breadth . ; 94-54 33 94-6+ °5
Basion-bregma height . 126-64 -4 129-24 -5 4-0
Breadth/length index . (gepse 3) T4-24 -4
Height/length index. Wise 2 (laa -&
Height/breadth index . pire) de AE 96-0 == <5
Transverse fronto-
parietal index . T15+4 -3 10:32 | 4
Horizontal circumfer-
ence : : 502-5 + 1-2 504:9+ 1-6

Cranial capacity (c.c.3). | 1238:847-7 | 1264-6413-8
Length of mastoid |

process. : : 24-44 33 Mf i 6-6
Length of base . 96-84 °3 97-64 -D
Length of face. 99:64: -4 96-44 -7
Gnathic index . : 98-:7+ -3 98:9+ -5
Maximum bizygomatic

diameter 123°5+ -5 123-7+ -6
Total face height 104-54 -7 112-24 1-0 5-2
Upper face height 62-34 -4 665+ -6 D1
Total facial index 85-4+ +5 906+ -7 6:5
Upper facial index 510+ 3 53:64 -4 5-2
Breadth of orbit : 40:94 -1 422+ 2 5:9
Height of orbit . Sl-1l+ +1 32-384 +3 2°8
Orbital index : : 76-34 -4 77:84 -4 2°6
Nasal width , : 25:8+ -2 25:-T7T+ °3
Nasal height 45-44 -2 AT-94+ -4 5:5
Nasal index ; : SO) a ok 54-44 +6 5:2
Minimum ramus

breadth . ; : 34°94 -3 3154 -4 6:8
Minimum ramus

height. : ; Al-24 -4 450+ -6 53
Index of ramus . 85:24 -9 70-6+ I: 9-7

Angle at nasion . 1D51-1+ -6°} 141-14 1-2° Con

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 223

many anthropologists, and we may quote Sir Arthur Keith (1929) who
said ““ When we have to distinguish one race of mankind from another,
we obtain more assistance from the facial than from the cranial parts
of the skull”’.

Where the average dimensions of the Cape Coloured crania differed
from the Bushman and Hottentot dimensions, these Cape Coloured
means had moved towards the Kuropean and Negro means. When
the genetic influence of the two last groups tended in opposite direc-
tions, as e.g. in nasal width, the Cape Coloured means remained
practically identical with those of the Bushman and Hottentot
crania (Table III).

Although it is seen that many skull dimensions and indices show
significant differences when a comparison is made between the Cape
Coloured on the one hand, and Bushman and Hottentot crania on
the other hand, these differences, as a rule, are less than those which
appear in comparisons between Cape Coloured and Bantu, or Cape
Coloured and European crania. Thus we see that the Bush-Hottentot
inheritance exerts a powerful genetic influence on the Cape Coloureds,
- é.g., the flat nasion region with a wide bridge of the nose, the prominent
zygomatic bones, the absence of prognathism and the square-shaped
ramus of the mandible (Plate IV).

DISCUSSION.

In the seventeenth-century records of the Dutch settlers different
names were used when the aboriginal tribes are mentioned. Apart
from the name Hottentots, they were sometimes called the Watermans,
sometimes the Strandlopers (beachrangers). The predatory habits of
the inland Bushmen were well known, but it was supposed that they
were merely another Hottentot clan (Theal, 1888). From this original
error much confusion has arisen, and even to-day anthropologists are
not agreed. One school believes that the Hottentots and the Bush-
men are so much alike in physical characters, and so distinct in
appearance from the other inhabitants of southern Africa, that one
cannot separate them into two races. They should be considered as
subgroups of one race, viz., the Khoisan race (Shapera, 1930; Galloway,
1933); the term Bush race is often used with the same meaning, 7.e.
to include both Hottentots and Bushmen. The other school of
anthropologists (Broom, 1923; Dreyer, 1937; Drennan, 1938) con-
siders that the Hottentots and the Bushmen are sharply differentiated
racial groups, although extensive hybridisation of the two peoples

224 Annals of the South African Museum.

had produced the intermediate Bush-Hottentot type, and clear-cut
separation is absent.

In a previous study of Bushman and Hottentot crania, I followed
the view that the Bushmen and Hottentots were distinct population
groups. The characteristic or “typical” Bushman crania are easily
separated from characteristic Hottentot skulls, as they show very
opposing tendencies. Crania which in the past have been called
Strandloper, on account of their being found in association with coastal
middens and rock shelters, are proving to be Hottentots or Bush-
Hottentot hybrid types (Drennan, 1938; Keen, 1941). For this
reason the term Strandloper should no longer be used in a racial sense,
as denoting a distinct population group. The Strandlopers are a
subdivision of the Hottentot group with a propensity for living along
the seashore. The same criticism applies to the term Korana; they
should be considered as a clan of the Hottentots. ‘Two Korana skulls
obtained from graves in the Orange River area near Upington proved
on analysis to conform very closely to the criteria established for the
Hottentot race. Further, it was from graves along the Orange River
between Kakamas and the Falls that the Bloemfontein Nasionale
Museum obtained its collection of “typical”? Hottentot crania.

The difficulties which face anthropologists in their study of the
aboriginal peoples of the Union of South Africa will be lessened
considerably if a clear differentiation is made between the two racial
groups, the Bushmen and the Hottentots, recognising at the same
time that there was extensive hybridisation between the two groups,
and that many crania represent the intermediate type, viz., the
Bush-Hottentot hybrid.

SUMMARY.

The author made a craniometric study of a collection of 205 skulls
in the South African Museum, Cape Town, which appertain to the
aboriginal peoples of the Union of South Africa, the Bushmen and the
Hottentots. The crania were submitted to a standard set of 20
measurements from which 10 indices were calculated. The means
of the dimensions and indices are given in tabular form, together with
the ranges and standard deviations.

The crania were divided into three types, Bushman, Hottentot and
Bush-Hottentot hybrid, and an analysis was made of the distribution
of the three types among the crania originating from different areas.
It was found that the Hottentot type occurred more frequently among
the crania originating from South West Africa and the Kalahari
Desert, the Bushman type more often among the crania originating

Bushman, Hottentot and Bush-Hottentot Hybrid Skulls. 225

from the interior of the Cape Province and along the southern coastal
area.

The differences and resemblances between characteristic Bushman
and Hottentot crania are described. A comparison was made between
the Bush-Hottentot crania and the allied Cape Coloured type. The
significant differences in the means could be explained by the genetic
influence of the other racial groups on the Cape Coloured. The
resemblances demonstrated the strong influence of the Bush-Hottentot
inheritance.

The author supports the view that the Bushmen and the Hottentots
are distinct racial groups, in spite of extensive hybridisation which has
taken place and which complicates the problem. Lastly, it is argued
that the terms Strandloper and Korana should no longer be used in a
racial sense.

I gratefully acknowledge the help received from a Research Grant
of the University of Cape Town, and wish to thank Mr G. McManus, of
the Department of Surgery, for the photographs of the skulls.

REFERENCES.

Brink, V. H. (1923). Trans. Roy. Soc. 8S. Afr., xi, 145.

Broom, R. (1923). Journ. Roy. Anthrop. Inst., li, 132.

Broom, R. (1941). Annals Transvaal Museum, xx, 217.

Drennan, M. R. (1938). Trans. Roy. Soc. 8. Afr., xxv, 259.

DreveER, T. F., and Metrine, A. J. D. (1937). Sodl. Nav. Nasionale Museum,
Bloemfontein, I, 81. :

GatLoway, A. (1933). S. Afr. Journ. Sci., xxx, 585.

Keen, E. N. (1949). S. Afr. Journ. Med. Sci., xiv, 7.

Keen, J. A. (1942). S. Afr. Journ. Sci., xxxviii, 301.

Keen, J. A. (1943). S. Afr. Journ. Sci., xxxix, 247.

Keen, J. A. (1947). Sodl. Nav. Nasionale Museum, Bloemfontein, I, 16, 191.

Keen, J. A. (1951). Trans. Roy. Soc. 8. Afr., xxxiii, 29.

Kern, Sir A. (1929). The Antiquity of Man,i,11. London: Williams & Norgate
Ltd.

Matncarp, L. F., and Maincarp, J. F. (1932). Bantu Studies, vi, 103, 163.

Martin, R. (1928). Lehrbuch der Anthropologie. Ed. Gustav Fischer (Jena),
p- 658.

PitTaRD, E. (1927). Le Globe, Ixvi, Mém. p. 2.

Scuaprera, I. (1930). The Khoisan Peoples of South Africa. London: G.
Routledge & Sons.

SHRUBSALL, F. C. (1907). Ann 8.A. Museum, v, 227.

StomeE, D. (1929). Ann. S.A. Museum, xxiv, 33.

Stow, G. W. (1905). The Native Races of South Africa. London: Swan,
Sonnenschein & Co.

TuHEAL, G. McC. (1888). History of South Africa (1487 to 1691). London:
Swan, Sonnenschein & Co.

226 Annals of the South African Museum.

EXPLANATION OF PLATES.

Plate I.—Photographs of three representative crania, seen from the front.
H, typical Hottentot. B-H, Bush-Hottentot hybrid. B, typical Bushman.
Plate II.—Photographs of three representative crania, seen from the side.
H, typical Hottentot. B-H, Bush-Hottentot hybrid. B, typical Bushman.

Plate I1I.—Photographs of three representative crania, seen from above.

H, typical Hottentot. B-H, Bush-Hottentot hybrid. B, typical Bushman.

Plate IV.—Side view of three mandibles belonging to different population
groups, to illustrate the shape of the ramus of the mandible.
A, typical Hottentot: index of ramus, 91-7.
B, Cape Coloured: index of ramus, 82:2.
C, European: index of ramus, 58-2.

Ann. S. Afr. Mus., Vol. XX XVII. Plate I.

J, A. Keen Neill & Co. Ltd

Plate II.

Ann. S. Afr. Mus., Vol. XX XVII.

Co. Ltd

oS}
~
SS

Y

J. A. Keen.

Ann. S. Afr. Mus., Vol. XX XVII.

If, Alo UGa0«

Plate III.

Neill & Co.

Ltd,

Ann. S. Afr. Mus., Vol. XX XVII. Plate IV.

J, A. Keen Neill & Co. Ltd.

The ANNALS OF THE SOUTH AFRICAN MUSEUM are issued in parts at

irregular intervals as material becomes available.
devoted exclusively to a particular subject (Zoology, Botany, etc.).
volumes may be in course of publication concurrently.

As far as possible each volume is
Two or more

Most of the Geological and Palaeontological papers are issued in conjunction with

the Geological Survey of the Union of South Africa.

Some volumes and parts are out of print, and others are only sold as parts ‘of a set,
or volume, respectively. The prices of parts published prior to 1940 have been

inereased.

Out of print: Vols. i. II, V (Part 9), VII, VIII, 1X (Part 1), XII (Part 7), XXII,

XXIV (Part 2), XXXI (Parts 1, 2, 3).

Vol. £ 8. a,
III. 1903-1905 Zoology E19. 30
IV. 1903-1908 Palaeontology 3 oy de®
VY. 1906-1910. Geology, eae Zooloey, Anthropology Lo 6
VI. 1908-1910 Zoology 3.3 6
IX. 1911-1918 Botany lexoll Part 1) 216 0
X. 1911-1914 Zoology wae “4060
XI. 1911-1918 Zoology m8 wes wed ie EG)
XII. 1913-1924 Palncontalogs whi Gealoes .. (excl. Part 7) 313 0O
XIII. 1913-1923 Archaeology and a aes ie oo LORE
XIV. 1915-1924 Zoology 3 6 6
XV. 1914-1916 Zoology Ao BU)
XVI. 1917-1933 Botany Biers ell es 0)
XVII. 1917-1920 Zoology & 3 OO
XVIII 1921 Zoology . 4 5 6
XIX. 1924-1925 Zoology =%, one
XX. 1924-1926 Zoology ooo
XXI. 1925-1927 Zoology ; po Owe)
XXIII. 1925-1926 Zoology AS ar an LE
XXIV. 1929-1938 Rethtecolny ana Tiheeleey .. (excl... Part 2) 2 oae
XXV. 1927-1928 Zoology et ct =o, OL ROG)
XXVI. 1928 Zoology .» 110730
XXVII. 1929 Anthropology 2 dlOea0
XXVIII. 1929-1932 Palaeontology ere)
XXIX. 1929-1931 Zoology 2 Se i0
XXX. 1931-1935 Zoology 3°13 6
INDEX of papers, authors, and subjects, published: in ved I-XXX 0 1 6
XXXII. 1934-1950 Palaeontology ..(Part 4 only) 014 0
XXXII. 1935-1940 Zoology wis ote). Ole ee
XXXITf. 1939 Zoology De
XXXIV. 1938 Zoology 2 8 0
XXXV. Reserved for conclusion of iionomtanly a in a vor XXXIV.
XXXVI. 1942-1948 Zoology A 2 yk
XXXVII. 1947-1952 Archaeology and Anthropology 1 16
XXXVIII. 1950 Zoology -s Be Bree 3)
XXXIX. 1952. Zoology 2 14

Se oy

ee ee a

Copies may be obtained from—
Messrs. WHELDON & WESLEY, Lop.,
2, 3, and 4 ArTHuR STREET, NEw Oxrorp Street, Lonpon, W.C. 2; or,

The LIBRARIAN, Soutn Arrican Musrum, Care Town.

Except the Geological and Palaeontological parts which are obtainable from the
GOVERNMENT PRINTER, PRETORIA.

A200

Pe

=e —