ArticlePDF Available

Paleoanthropological investigations of Middle Stone Age sites at Pinnacle Point, Mossel Bay (South Africa): archaeology and hominid remains from the 2000 field season

Authors:

Abstract and Figures

The Middle Stone Age (MSA) in South Africa has gained increasing attention due to the discovery of bone tools at Blombos Cave, the abundance of ochre suggesting artistic expression, the presence of a variety of lithic assemblages with advanced technological characteristics, and debates over the interpretation of the fauna. Linked to these findings are debates over the antiquity of modern human behavior, with some researchers arguing that the South African evidence suggests an early origin of modern behavior, while others suggesting a late origin. Resolution of these debates relies on two advances: improvements in our theoretical approach, and an improvement of the empirical record in Africa. We initiated fieldwork at Mossel Bay on the southern coast of the Cape to address the latter deficiency. Our survey to date has covered a 2 km section of 8km of coastal cliffs, penetrated 1 km inland, and resulted in the discovery of 28 archaeological sites, 21 of which are MSA, and 15 of those are caves or rock shelters. Test excavations were carried out at 3 of these caves, all at Pinnacle Point. Two (13A and 13B) yielded rich MSA horizons with outstanding preservation of fossil bone,and lithic assemblages. Two hominin fossils were found, and the incisor is metrically intermediate between Middle Pleistocene hominins and modern humans. The lithic assemblages are dominated by the early stages of artifact manufacture, and resemble typologically and technologically the regional variant of the MSA (the Mossel Bay Industry), but blades and points are significantly smaller than reported elsewhere. The fauna from both sites display a surprising rarity of small mammal, micro-mammals, and tortoise, unlike at most other MSA sites from the Cape. The shellfish remains indicate collection from nearby rocky outcrops and very occasionally also from sandy beaches. Collecting rounds targeted the mid- and lower reaches of the inter-tidal zone.
Content may be subject to copyright.
Paleoanthropological investigations of
Middle Stone Age sites at Pinnacle Point,
Mossel Bay (South Africa): Archaeology and
hominid remains from the 2000 Field Season
CURTIS W. MAREAN
Institute of Human Origins, Department of Anthropology, P.O. Box
872402, Arizona State University, Tempe, AZ 85287-2402 USA
PETER J. NILSSEN
Archaeology Division, Iziko - South African Museum, P.O. Box 61,
Cape Town 8000, South Africa
KYLE BROWN
William Self Associates, Inc., P.O. Box 2192 61d, Avenida de
Orinda, Orinda, CA 94563 USA
ANTONIETA JERARDINO
Department of Archaeology, University of Cape Town, Rondebosch
7700, South Africa
DEANO STYNDER
Department of Archaeology, University of Cape Town, Rondebosch
7700, South Africa
ABSTRACT
The Middle Stone Age (MSA) in South Africa has gained
increasing attention due to the discovery of bone tools at
Blombos Cave, the abundance of ochre suggesting artis-
tic expression, the presence of a variety of lithic assem-
blages with advanced technological characteristics, and
debates over the interpretation of the fauna. Linked to
these findings are debates over the antiquity of modern
human behavior, with some researchers arguing that the
South African evidence suggests an early origin of mod-
ern behavior, while others suggesting a late origin. Res-
olution of these debates relies on two advances:
improvements in our theoretical approach, and an
improvement of the empirical record in Africa. We ini-
tiated fieldwork at Mossel Bay on the southern coast of
the Cape to address the latter deficiency.
Our survey to date has covered a 2 km section of 8
km of coastal cliffs, penetrated 1 km inland, and resulted
in the discovery of 28 archaeological sites, 21 of which
are MSA, and 15 of those are caves or rock shelters. Test
excavations were carried out at 3 of these caves, all at
Pinnacle Point. Two (13A and 13B) yielded rich MSA
horizons with outstanding preservation of fossil bone,
and lithic assemblages. Two hominin fossils were found,
PaleoAnthropology 2004.05.02.14–83
Copyright © 2004 PaleoAnthropology Society. All rights reserved.
MIDDLE STONE AGE SITES AT PINNACLE POINT 15
INTRODUCTION
The Mossel Bay Archaeology Project
(MAP) is a long-term field study of the
Middle Stone Age (MSA) in the Mossel
Bay region. Our field research to date
demonstrates that the Mossel Bay region
has an unusually rich MSA record, and
here we report on the first several years of
survey and test excavation. Our primary
research goals are to test several compet-
ing models concerning the behavioral
modernity of MSA people in Africa and
thus contribute to our knowledge of the
origins of modern humans. In particular,
we plan to focus on resolving several
chronologic and chronometric questions
about the South African MSA, raw mate-
rial exploitation strategies, and faunal
exploitation strategies. To that end we
plan a longitudinal study that will involve
two missions. The first ongoing effort
involves survey (both archaeological and
geological for raw material sources) and
test excavations of discovered sites. The
second will involve intensive excavations
at those previously tested sites identified
as having high potential for helping us
resolve our research problems, which has
just begun. As reported here, the lithic
and faunal assemblages, though small,
already expand the range of variation doc-
umented for the MSA in South Africa.
HISTORY OF RESEARCH
Archaeological research in the Mossel Bay
region has not been intense, despite the
early initiation of work by George Leith in
1888 (Leith, 1898) at Cape St. Blaize Cave
(CSB), located in the town of Mossel Bay
(Figure 1). Last excavated by Goodwin in
1932 (Goodwin & Malan, 1935), this site
yielded a series of selected lithic collec-
tions central to the definition of the Mos-
sel Bay Industry (Goodwin, 1930;
Sampson, 1971). Goodwin argued for the
presence of an inter-stratified Howiesons
Poort occupation at CSB, but this was
based on the presence of point-types
thought at that time to be characteristic of
the Howiesons Poort, while more recent
definitions tend to rely on the presence of
backed pieces (Volman, 1981; Thackeray,
1989). MSA research in the Mossel Bay
region effectively stopped after these
investigations.
and the incisor is metrically intermediate between Middle Pleistocene
hominins and modern humans. The lithic assemblages are dominated by the
early stages of artifact manufacture, and resemble typologically and technolog-
ically the regional variant of the MSA (the Mossel Bay Industry), but blades and
points are significantly smaller than reported elsewhere. The fauna from both
sites display a surprising rarity of small mammal, micromammals, and tor-
toise, unlike at most other MSA sites from the Cape. The shellfish remains indi-
cate collection from nearby rocky outcrops and very occasionally also from
sandy beaches. Collecting rounds targeted the mid- and lower reaches of the
intertidal zone.
16 •PaleoAnthropology 2004
In 1997 Kaplan and Nilssen conducted
an environmental impact surface survey
of the Pinnacle Point area (Figures 1 and
2), a section of coastal cliffs west of the
town of Mossel Bay (Kaplan, 1997). Their
work was in response to a
hotel/casino/golf course development
proposal for the area above the cliffs,
which is now in the construction stage.
They covered an area of approximately 2
km of the coast at Pinnacle Point and
about 1 km inland and discovered 28
archaeological sites (21 MSA), 15 of which
are caves/rockshelters. They named sites
in numeric sequence of discovery, and
sites deemed to be related in formation
were given the same number with letters
used to designate distinct caves or rock-
shelters. In March of 1999 Nilssen and
Marean revisited Pinnacle Point and Mos-
sel Bay to survey the area and investigate
the potential of the sites. Four caves were
selected for test excavation (Cave 9, 13A,
13B, and 13C). Since most of the
caves/rockshelters in this area suggest
high potential due to abundant scatters of
MSA lithics, we chose these four based on
210
200
190
5 Kilometers
Sites 13A,B, CSite 9
Mossel Bay
Pinnacle Point
Town of
Mossel Bay
34
0
33
0
34
0
33
0
21
0
22
0
23
0
24
0
26
0
28
0
27
0
25
0
20
0
19
0
18
0
21
0
22
0
23
0
24
0
26
0
28
0
27
0
25
0
20
0
19
0
18
0
Cape Town
Hermanus Stillbay Mos sel Bay Plettenberg
Bay
Port Elizabeth
East London
100 km
Die Kelders Blombos
Klasies
River
Cape St. Blaize
Cave
Figure 1. Map showing the Cape of South Africa (above), and a blow-up of the Mossel Bay
peninsula, Pinnacle Point, and the surveyed area (grey shading).
MIDDLE STONE AGE SITES AT PINNACLE POINT 17
0 1 2 km
25
1
2
24
22
23
26
7
6
5
16
15
14
17
8
9
10-12
13
21
20
19
18 3
4
Figure 2. Aerial photograph
of the survey area (enclosed
yellow line), and the sites
found to date: Red =
cave/rockshelter, Black =
open-air lithic scatter, and
yellow = shell midden.
18 •PaleoAnthropology 2004
their proximity to one-another and the exceptionally high potential
of 13B. Marean and Nilssen conducted another survey in March of
2000, established a coordinate system (grid), laid numerous per-
manent survey control points, and mapped Cave 13B.
In July of 2000 we conducted test excavations, preceded by
several weeks of logistical work. The caves sit near the base of a
nearly vertical cliff face roughly 60 m from base to top (Figures 3
and 4). Scaling the cliff is a hazardous climb, and probably
explains the lack of discovery of these sites in the past. Bringing a
team up and down the cliff daily would be impossible, so we built
a 176 step wooden staircase from the top to bottom of the cliff. We
installed piping from a municipality water line located above the
cliff down to the excavations so that all materials could be wet-
sieved with fresh water.
During the July fieldwork we decided to test only 3 caves, pri-
marily because 13B required several test-pits to properly evalu-
ate. We conducted 3 weeks of excavation with a field team of 15
and a lab team of 3, and followed by 2 months of curation and
study at the South African Museum. In March of 2001 we revis-
ited the sites with two teams for the purposes of luminescence
dating. One team (Ann Wintle, Geof Duller, Helen Roberts, and
Zenobia Jacobs) sampled sandy sediments from Caves 9 and 13A
for the purpose of OSL dating, took background gamma meas-
urements, and implanted dosimeters. A second team (Helene
Valladas, Norbert Mercier, Chantal Tribolo) placed dosimeters
into various locations of Cave 13B for TL dating of burnt lithics
and took gamma measurements. In March of 2002 we continued
A)
B)
Site 13 Cave Complex Calcretes
Figure 3. Pinnacle Point and the cliffs, showing A) Pinnacle Point
from the north-east with the Site 13 Cave Complex indicated by the
arrow, and B) the cliffs to the east of the site 13 complex and the
stairs down the cliffs.
MIDDLE STONE AGE SITES AT PINNACLE POINT 19
with survey and mapping of the surround-
ing area. Since then, construction began
on the development, and MAP maintains
a team at Mossel Bay that monitors con-
struction activity for archaeological
remains. Significant quantities of
Acheulean material have been found, and
these are under study by Erin Lassiter,
and will be reported elsewhere.
GEOLOGICAL BACKGROUND
The coastal cliffs in this area are expo-
sures of the Skurweberg Formation of the
Table Mountain Group. This is a coarse-
grained, light-gray quartzitic sandstone,
often covered with lichens, with beds of
varying thickness and consolidation (see
Figure 3). Small patchy exposures of the
Robberg Formation of the Uitenhage
Group outcrop above the Skurweberg For-
mation near the Mossel Bay point. The
regional dip varies strongly from west to
east between 10 to 75 degrees (South
African Geological Series 3422AA 1993).
The vast majority of the caves we have
examined occur in the nearly vertical
coastal cliffs, and this is well illustrated in
Figure 4. The location of several of the Pinnacle Point sites, plotted on an orthophoto that has
been rectified to the MAP grid. The sites include site 9 (in red), 13A (in light blue), 13B (in
green), and 13C (in dark blue) plotted on it. Orientation is magnetic north.
20 •PaleoAnthropology 2004
Figure 4, which is an orthophoto rectified
to the MAP grid (see below), with maps of
site 9 (in red), 13A (in light blue), 13B (in
green), and 13C (in dark blue) plotted on
it. We noticed a trend for caves and shel-
ters to be present when the bedding plane
was more horizontal (10-40 degrees) and
less resistant layers of quartzite eroded
out, likely by relatively higher sea levels,
leaving more resistant layers intact above
and below forming a cave or rock shelter
(Figure 3). Many also occur in faults.
Thus, caves and rock shelters occur in
clusters that are predictable from the geo-
logically mapped dip and faulting.
Calcretes and dunes cap the quartzite
throughout the area (Figures 3A and 4).
The calcretes are highly variable in thick-
ness and form. Some are up to a meter
thick or more, while others are thin flows
of very hard calcite adhering directly to
the quartzite. In several road cuts we
observed calcretes and caliche horizons in
the sands, and in several locations we
observed MSA artifactual horizons in
these cuts. Various calcite formations are
present in the caves and rockshelters, par-
ticularly at joints and bedding planes.
Small stalactites hang from the roofs of
many of the caves, and flowstone forma-
tions are present in many and in some
cases, cover archaeological deposits in the
caves (see discussion of site 13B). In
numerous locations there are fossilized
and cemented dunes adhering to the
quartzite, both inside and outside caves,
and are testament to ancient dune activity
that once abutted the cliffs. While the
caves are mostly dry today, it is clear that
water seeps through the joints and bed-
ding planes of the quartzite. While
quartzite is acidic and typically creates a
sedimentary environment hostile to bone
preservation, the water entering these
caves has been buffered by the calcretes
capping the quartzite, raising the ph of the
sediments and resulting in outstanding
bone preservation, as at Klasies River
(Singer & Wymer, 1982: 2).
EXCAVATION PROCEDURES
We employ a three dimensional coordi-
nate system (grid) that encompasses all
our excavated sites, and will eventually
envelope our entire research region (Fig-
ures 4 and 5). Our grid began in site 13B,
and there is a MAP control point
cemented into the cave that acted as our
starting elevational datum. The x-y-z
coordinates for that control point were
originally and arbitrarily set at 100, 100,
0. Since then we have tied our grid into
the South African national coordinate ref-
erence system. While South Africa’s coor-
dinate system was once based on a local
framework (known as the Cape Datum),
as of January 1, 1999, South Africa offi-
cially shifted to the World Geodetic Sys-
tem 1984 (WGS84) as their base, with the
new system being called the Harte-
beesthoek94 Datum. Our grid has been
tied to this updated system through South
Africa’s network of passive trigonometric
beacons (via total station shots to 16 MAP
control points), and all our coordinates
can be transformed to it using a two-
dimensional conformal coordinate trans-
formation (Wolf & Ghilani, 2002), and
from there to latitude and longitude fol-
lowing the Gauss Conform Projection
used by South Africa. Since the numbers
MIDDLE STONE AGE SITES AT PINNACLE POINT 21
used for planar coordinates in the South
African system are large and unwieldy for
day to day excavation tasks, we have
elected to maintain our coordinate system
(henceforth the MAP coordinate system)
for use and publication of all planar coor-
dinates, but have shifted our elevational
measurements to conform to the South
African system of true orthometric height
above mean sea level, which is determined
in Cape Town and verified by a variety of
tide gauges distributed across the country.
Recording and providing elevational data
in sea level has obvious advantages, and
the height of our starting elevational
datum (originally set at 0, see above) in
13B is 20.988 m above mean sea level.
About 40 permanent control points
have been shot-in throughout this area,
typically as bolts in concrete in protected
rock crevasses assessable to a prism, or
holes drilled into the quartzite, both
inside sites and on the landscape. These
are numbered sequentially, digitally pho-
tographed, and available at the South
African Museum and SAHRA. Any future
researcher can access these, as well as the
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
80
90
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
55
65
75
85
95
105
115
125
135
145
155
165
175
185
195
205
215
55
65
75
85
95
105
115
125
135
145
155
165
175
185
195
205
215
North
meters East
meters North
Indian Ocean
Coastal Platform
9
10
12
13C
13B
13D
13A
Figure 5. The position of the sites referred to in the text relative to the MAP grid. The site out-
lines are indicated by the limits of the drawn contour lines, and the scale is indicated by the
axes labeled in meters. The approximate location of the cliffs and ocean are indicated. Orien-
tation is magnetic north.
22 •PaleoAnthropology 2004
corresponding South African grid coordi-
nates of 16 of these, and establish their
precise position within our coordinate
system (and the national coordinate sys-
tem) using a standard surveyor’s three-
point resection (Wolf & Ghilani, 2002).
The MAP coordinate system is oriented
to magnetic north and grid coordinates
advance positively to the north (x axis)
and east (y axis). A 1 m square is named by
the planar coordinates of its south-west
corner. We excavated within 50 cm quad-
rants within squares, and these are named
by their bearing: NE, NW, SE, and SW.
Excavations were conducted within these
quadrants following natural stratigraphic
units (layers, features, etc.), and thus
square-quadrant-stratigraphic unit prove-
nience designation is the minimum
assigned to any find, while most finds also
have a precise 3D coordinate as well. Sed-
iment volumes were measured during
excavation, and bulk samples of sediment
were taken from every unique strati-
graphic unit. All finds that were identifi-
able or had a maximum linear dimension
of 1 cm or more were plotted (henceforth
called plotted finds) in x-y-z coordinates
by total station directly to a hand-held
computer using Survey Pro software. We
used a specimen number-based system:
all excavators were given pre-printed
adhesive specimen numbers. Finds were
dropped individually into plastic bags, the
specimen number adhered to the inside,
and the specimen number was typed into
the hand-held computer, thus linking it to
the coordinates and stratigraphic unit. All
spatial data is linked within ArcView GIS,
which is also linked to external relational
databases with detailed descriptions of all
finds.
All recording was done to forms (sup-
plemented by notebooks), and those were
typed into form-based database systems.
All features and stratigraphic units were
drawn to graph paper, and their shape and
topography were shot in directly by total
station to hand-held computer. We used
both digital and film photography. Draw-
ings were digitized to computer, and
ArcView GIS with the Image Analysis
extension is used to integrate the coordi-
nate and image data.
All non-plotted materials were gently
wet-sieved with fresh water through a
nested 10-3-1.5 mm sieve. These were
dried, packed in plastic bags, and trans-
ported to the field laboratory where they
were sorted into major analytical cate-
gories, and some preliminary analyses
performed. All plotted finds were labeled
with their specimen number in black
India ink. All materials were then trans-
ported to the South African Museum for
study and storage.
We have completed the analysis of all
the plotted finds, and that sample of arti-
factual and ecofactual material is reported
here. The following is a general descrip-
tion of the excavated sites, and a descrip-
tion and preliminary interpretation of the
major depositional events. Each square is
described as a section of sediments, and
the sediments are classified into a series of
facies that are defined by their having
similar characteristics that suggest they
were formed under similar conditions.
Broadly speaking, both natural and
behavioral processes have contributed to
the facies, typically with one dominant
MIDDLE STONE AGE SITES AT PINNACLE POINT 23
over the other. When anthropogenic sedi-
mentary processes are significant, the
name includes “MSA” to signal that char-
acter. No LSA (Later Stone Age) material
was found in these sites. All descriptions
proceed from bottom to top.
SITE 9: SUMMARY OF THE
EXCAVATIONS
Site 9 is a large cave (Figures 5 and 6) in a
heavily metamorphosed layer of quartzite
that appears erosion resistant and shows
few signs of spalling or other activity.
There is a large debris field of cliff collapse
outside (Figure 7a) and inside (Figure 7b)
the mouth of the cave that nearly seals the
entrance. One enters the cave by partially
crossing this boulder field, ascending a
steep hill of cliff collapse, and then
descending down the other side into the
cave. This collapse must be geologically
recent, because the boulders are sharp
and unweathered (unlike the typical boul-
ders in this area), and lichens have yet to
invade the zone of cliff detachment. Sur-
prisingly, this cliff collapse did not seem
to cause any roof collapse in the cave, and
Site 9
meters East
meters North
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
8
9
10
12
Figure 6. The outline and
topography of Site 9, and the
position of the test pit (the
grey box). The contours are in
meters asl, at 10 cm incre-
ments. The far northern part
of the cave has a crawl space
too low to the sediment to
properly map, and thus the
extent is estimated and indi-
cated by a dotted line. There
are several small caves within
the walls and roof of Site 9
that are not mapped here. Ori-
entation is magnetic north.
24 •PaleoAnthropology 2004
the debris from the collapse failed to roll into the cave beyond the
immediate area of the mouth. The cave is currently dry, and the
few signs of flowstone are concentrated near the mouth. This cave
appears to be very ancient, as suggested by the resistant rock into
which it is cut.
Smaller caves are present to the west side of the mouth when
looking in, and above the mouth as well. They have sandy sedi-
ments eroding out of their mouths and are difficult to access. We
have recently explored one of these and found MSA sediments
just several centimeters below dune sand.
The cave floor is covered by an undisturbed sand dune deposit
(Figure 8). Despite the rarity of lithic material in and around this
cave, we expected it to have a rich deposit of archaeological mate-
rial, for several reasons. First, the cave presents a very agreeable
living space: it is warm and protected from the elements. Second,
the cave is likely to be very old, and thus could harbor even ESA
material. We placed a 1 x 1 m excavation square near the back of
the deposit (Figure 6). This appeared to be a likely location where
we would miss potential cliff-collapse, but the obvious negative
aspect of this placement was the potential for thick dune sands.
Square N204E225
Our excavations in this cave penetrated nearly 2 meters into the
deposit (Figure 9) and failed to encounter any dense archaeolog-
ical deposit, while a thin deposit of 7 flakes were encountered
roughly 1 m below the surface. The profile is massive dune sand
(Figure 9). The top 50 cm display included that we interpreted as
A)
B)
Site 9
Figure 7. Site 9 from the outside and to the south-west (A) showing the large fresh blocks
of stone from the recent cliff collapse, and (B) site 9 from the inside looking out showing
the extent of the cliff collapse that fell into the cave.
MIDDLE STONE AGE SITES AT PINNACLE POINT 25
rock hyrax, as well as moderate stratifica-
tion caused by these organic inclusions
and perhaps some mild horizonation.
Below this no stratification was obvious
during excavation, but in section very sub-
tle changes in the sands that may repre-
sent either temporary surfaces or could be
postdeposional features caused by water
action.
At 2 m, we deemed any further excava-
tion to be hazardous, and thus terminated
the excavation. However, we did probe the
base of the square with a soil auger with a
10 centimeter diameter, saving all the
materials from the auger for study. The
auger did not encounter any obvious
archaeological materials, but did strike
rock at roughly 1 m below the base of our
excavated square. We believe that this rock
is likely to be roof fall, as the auger did not
go through any zones of debris as one
would expect if we were encountering a
cave floor that had been exposed to erosion.
SITE 13A: SUMMARY OF THE
EXCAVATIONS
This site presents a loose sandy dune
against a 2-sided corner in the cliff wall
(Figure 10), protected by an overhang in
the cliff. The entire dune is within the drip
zone of the overhang. The back of the
dune against the cliff wall reveals a small
Figure 8. Site 9 from top of cliff collapse, showing the dune that comprises the surface of the
cave, and our excavations near the back.
26 •PaleoAnthropology 2004
A)
B)
20 cm
Figure 9. Site 9, square N204E225, south section. Figure 10. Site 13A (A) from the south showing two loca-
tions of cemented dune with MSA lithics (red arrows) and
(B) a close-up of the lower cemented dune with lithics.
MIDDLE STONE AGE SITES AT PINNACLE POINT 27
cave or tunnel that currently is used as a
latrine by fishermen and/or other visitors.
The cliff wall and overlying ledge are
quartzite, some of which is poorly meta-
morphosed and sometimes even friable.
Adhering to the wall in two places is
cemented dune sand that includes MSA
lithics. There is a sample of this material
adhering to the wall just near the top of
the dune sands (Figure 10B), and there is
another bit of cemented dune adhering to
the wall several meters above the sedi-
ment surface (Figure 10A).
Just below the surface of the deposit is
a narrow ledge of harder sandy matrix,
partially stabilized with vegetation. On the
ledge are numerous quartzite MSA lithics
including points, blades, and cores. There
is some patchy dark sediment eroding
from the base of the dune near these arti-
facts that may represent eroding humic
archaeological sediment. Below the ledge
is a steep slope down to a beach boulder
field, and the slope is covered in vegeta-
tion. The site overlooks the ocean and
receives sun in the morning, but for most
of the day the deposit was shaded, and
reasonably well protected from wind.
We placed a 1 x 1 m excavation square
near the center of the deposit (Figure 11).
Square N69E100
The entire section sampled by our excava-
tions was composed of loosely-consoli-
dated dune sands with human occupation
being represented by thin lenses of finds,
often with organic material interstratified.
We recognized three facies in our excava-
tions, each composed of several thin and
horizontally discrete stratigraphic units,
including hearths and other organic-rich
layers (Figure 12).
At the base of our excavations was a
dune sand of coarse undecalcified, light
brown to white, poorly sorted sands
(White Sands MSA Facies). Its thickness
is unclear as it continued deeper, yet we
terminated excavation because the loose
sandy sections became unstable. Artifacts
and fossil bone are abundant in lenses of
finds that show little or no lithologic dis-
tinctions. A rotting roof block lies in this
layer in the south-east part of the square,
but this could be easily removed with fur-
95
96
97
98
99
100
101
102
103
104
105
106
107
108
95
96
97
98
99
100
101
102
103
104
105
106
107
108
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
13
12
11
Figure 11. The outline and topography of Site
13A, and the position of the test pit (the grey
box). The contours are in meters asl, at 10
cm increments. Orientation is magnetic
north.
ther excavation. These sands were dug as
2 separate stratigraphic units (J and K),
recognizing a slight change in K to a sand
with grayish nodules, very soft, within the
sands and increased artifactual content.
The lower white sands grade into a
series of inter-bedded brown-humic sand
lenses and white-yellow non-humic sand
lenses named the Brown Humic Sands
MSA Facies. This facies ranges in thick-
ness from approximately 10 to 15 cm.
Excavators struggled to separate these
during excavation and were reasonably
successful. These seem to represent
events of clean sand deposition (the
white-yellow) along with the deposition of
humic materials with sand (the brown
sand). We also recognized very subtle
lenses of burning, most of which are not
visible in section, except for the reddish
lense in Figure 12B.
The section is capped by a light brown
surface sand of very loose and disturbed
material called the Light-brown Surface
Sands Facies that ranges in thickness
from 3 to 15 cm. This unit is mostly sterile,
and displayed clear signs of disturbance
by foot traffic.
SITE 13B: SUMMARY OF THE
EXCAVATIONS
Site 13B is a cave (Figure 13) with the cir-
cular mouth of the cave facing east and
overlooking the ocean. From the back of
the cave the mouth resembles a ship’s
porthole looking out to sea (Figure 13B).
The roof of the cave is about 7 m high at
the front and narrows as the sediments
slope up toward the back. The entire cave
has sufficient room to allow a person to
walk or crawl except for the few meters at
N70E100 N70E10 1N70E100.5
13.14 m asl
12.64 m asl
Not Excavated Not Excavated
N69E100 North Section
White
Sands
MSA
Facies
Reddish Lense
Brown
Humic
Sands
MSA
Facies
Light
Brown Surface
Sands
Facies
Pebble Inclusions
A)
Figure 12. Site 13A N69E100 north section photo (A) and drawing (B).
28 •PaleoAnthropology 2004
MIDDLE STONE AGE SITES AT PINNACLE POINT 29
the back. The dimensions of the cave are
roughly 30 m long by 8 m wide, and it is
oriented with its long axis in a roughly
East/West direction (Figure 14). The floor
of the cave at the mouth is 13 m above the
high spring tide mark, and 15 m above
mean seal level.
Several layers of quartzite of varying
erosional resistance form the cave walls
and ceiling. Much of the ceiling quartzite
is friable, and seems to be regularly pro-
ducing roof-spalling. The walls are more
heavily consolidated and resistant to ero-
sion. The friable nature of the roof makes
this a very active cave, perhaps even rea-
sonably young in age. Some of the cave
roof and wall is covered by flow-stone, and
small stalactites hang from several loca-
tions. On either side of the cave against
the cave wall is a lightly consolidated MSA
deposit (Figure 14, indicated as LC-MSA)
that is capped by a flow-stone that
emanates from several joints in the
quartzite (Figure 15). This flowstone
capped deposit was clearly cut by an ero-
sional event, now leaving an exposed sec-
tion. We initially recognized four distinct
strata in the exposed section on the south-
ern wall, but these are not as clear in the
northern sections. Well-preserved and
abundant fossil bone and MSA lithics are
visible in this section.
The area within 13B is well protected
from the elements. We noticed that the
cave tended to be warmer inside than out-
side, and it is currently dry, though there
is evidence that it was wetter at times dur-
ing the past (see below). Winds rarely
blew into the mouth, and during even the
most intense of storms the inside of the
cave was warm, dry, and sheltered from
B)
Figure 12. Site 13A N69E100 north section photo (A) and drawing (B).
30 •PaleoAnthropology 2004
13B
13C
A) B)
Figure 13. Sites 13B and 13C, with the stylized yellow figure pro-
viding scale (A), and the view out the mouth of 13B (B).
the wind. During the mornings the sun
shines on the front 25% of the floor for
several hours, and then disappears
behind the cliffs. Overall, the cave pro-
vides a well-sheltered environment.
Our strategy for sampling this cave
included testing three areas: front, back,
and the side where the exposed section
was located (Figure 14). Our excavations
in the back at N97E97-N97E96 were
placed to sample the area where we
believed a thick midden was present,
based on the exposure of MSA finds on
the surface resulting from a disturbance,
likely caused by fishermen using the cave
as a camping spot. The initial square
(N97E97) intersected this disturbance so
as to evaluate the damage, and we
extended it 50 cm to the west (N97E96,
eastern quadrants) to get a better sample
of the rich MSA layer. Our excavations in
the front of the cave (N91E108) were
placed in a fairly level area that looked
suitable for domestic activities; currently
fishermen that use the cave build their
fires in this area. We decided to sample the
LC-MSA by placing an approximately 50 x
50 cm square into the exposed section.
Square N97E97 - N97E96
Excavations in this area provided the
longest section, but dense MSA remains
were found only in the top layers (Figures
16 and 17). Water-worn boulders line the
base of the excavated area, and we have
labeled it the Boulder Facies. Spaces
between the boulders are filled with very
loose silty sediment, and in some cases
there are empty holes between the boul-
ders. There are also some empty spaces
between the boulders and the base of the
MIDDLE STONE AGE SITES AT PINNACLE POINT 31
Site 13B
LC-MSA
LC-MSA
LC-MSA
Disturbed
Area
21
20
19
18
16
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
Meters East
Meters North
North
Figure 14. The outline and topography of Site 13B, and the position of the test pits (the grey
boxs). The contours are in meters asl, at 10 cm increments. LC-MSA is the lightly consolidated
MSA deposits. Orientation is magnetic north.
32 •PaleoAnthropology 2004
sections, perhaps indicating that the tops
of the boulders are still an active point of
erosion. The loose sediment between the
boulders includes some roof-fall up to 20
x 20 cm that is not water-worn but
exhibits a thin coating of clay, sometimes
patchily adhering to the roof-fall. We
found rolled fossil bone in this loose sedi-
ment, and some of it was bird. In antiq-
uity, it is likely that there was a very active,
high-energy stream running through the
cave, as indicated by the large water-worn
boulders, or else the worn boulders repre-
sent a very high sea stand.
The Boulder Facies is capped by sedi-
ments that continue approximately 1 m up
the section and are named the Laminated
Facies. This is a predominantly sterile
zone of grayish-brown silty sediment with
thin multiple laminae of sand (Figure 16
and 17). The few finds from these units
were small, polished, pieces of fossil
mammal bone, and there was very little
roof fall in these units. The laminated sed-
iment is clearly the result of water activity,
and we currently are analyzing it to estab-
lish the precise process of deposition.
The Laminated Facies is overlain by
yellowish sediments with some stringers
of brown organic material, the lower half
of which displays randomly oriented roof
spall. When first encountered, we inter-
preted this to be an erosion gulley cutting
into the Laminated Facies, and named it
the Erosion Gulley Facies. However, the
Flowstone
Exposed MSA Section
Figure 15. The flowstone on the north-east wall of site 13B capping the MSA deposit (LC-
MSA), the MSA section eroded and exposed by natural erosional processes, and the location of
our test excavation into that section.
MIDDLE STONE AGE SITES AT PINNACLE POINT 33
Laminated Facies displays a clear fault in
the north section (Figure 16 and 17). This
fault resulted in substantial (> 30 cm)
downward vertical displacement of the
western Laminated Facies sediments, as
well as some sediments that overlie the
Laminated Facies. Thus, the intrusive-
appearance of the Erosion Gulley Facies
may in fact be entirely due to the fault,
with the layer above infilling the post-fault
depression caused by the subsiding sedi-
ments. For now, we will retain the Erosion
Gulley Facies since it has some features
characteristic to water action, and in fact
erosion may have followed the slope dip
caused by the subsidence.
A yellowish layer of decomposing roof
spall (the Roof Spall Facies) overlies the
Erosion Gulley Facies, and postdates the
fault as it is not cut by it. It averages about
10 cm thick. This layer is sandy quartzitic
debris that resembles the surface sedi-
ment covering the back of the cave, but is
finer grained. Only one burnt lens (WF)
occurs within this facies.
Overlying this predominantly sedimen-
tary set of facies is a more anthropogenic
N98E96.5 N97E98N98E97
20.45 m asl
19.95 m asl
Roof Spall
Facies
Brown
Sands
MSA
Facies
Surface Disturbed
Facies
Not
Excavated
Not
Excavated
Erosion Gulley
Facies
Laminated
Facies
Boulder
Facies
18.95 m asl
19.45 m asl
Hollow
Space
Fault
Large
Boulders
N97E96.5-97 North Section
Sediment Removed
by Disturbance
Figure 16. Section drawing for the northern section of N97E96.5 and N97E97, site 13B.
34 •PaleoAnthropology 2004
Sediment Removed by
Disturbance
Figure 17. Section photos for the northern sec-
tion of N97E96.5 (left) and N97E97 (right),
site 13B, with the photos scaled and overlain.
The small increments on the scale are 1 cm,
while the large increment is 10 cm.
MIDDLE STONE AGE SITES AT PINNACLE POINT 35
facies named the Brown Sand MSA Facies
that ranges in thickness from 3 to 20 cm.
This is an organic-rich, black to dark
brown, sandy sediment rich in lithics and
fauna. In N97E97 we encountered a thin
and horizontally restricted layer of this
facies (named WB) that projected just 30
cm into the square. The rest had been
removed by the recent disturbance, which
cut all the way into the top of the Lami-
nated Facies (Figure 16 and 17). In the
west section of N97E97 WB appeared thin
and homogeneous. To sample WB better,
we extended our excavation with two 50
cm quads to the west into N97E96. As we
excavated west, WB thickened and
increased in complexity (Figure 16 and
17). Several discrete lenses became visible,
some of which appeared in the west sec-
tion of the 2 newly excavated quads. While
dark, they do not appear to be hearths, but
may be organic-rich dumps of material.
This facies appears to extend far into the
back of the cave, as it is visible under a
small block of roof fall to the west of our
excavation.
Capping this rich MSA facies is a final
layer of roof spall called the Surface Dis-
turbed Facies. This material is coarse with
sandy yellow roof detritus, and ranges
from 2 to 4 cm thick.
Square N91E108
The base of this square is a clast-sup-
ported matrix of small (1 cm and smaller)
roof spall with fresh edges, variably
cemented and uncemented, called the
MSA Roof Spall Facies. We do not know
the thickness since we did not excavate to
its full depth. Uncemented zones were
excavated and appear as valleys in the
floor of the square between the cemented
zones. Stratified within this matrix are
small, thin, well-preserved hearths with
lithics and fauna laying in and beside the
hearths (Figure 18 and 19). The hearths,
as shown in close-up digital photography
(Figure 19B), have discrete bands of ash,
charcoal, and baked (reddish-brown) sed-
iment. We expect these hearths to con-
tinue in all directions, as they are clearly
visible in section.
N91E108 N92E108
N91.5E108
16.56 m asl
Not Excavated
Not Excavated
N91E108 West Section
MSA
Roof
Spall
Facies
Surface Disturbed Facies
16.26 m asl
Hearths
= rock
Figure 18. Section drawing for the western section of N91E108, site 13B.
36 •PaleoAnthropology 2004
A very thin surface material of wind-blown dusty sediment and
roof spall overlies the MSA Roof Spall Facies, includes signs of
recent burning, and is called the Surface Disturbed Facies. It
ranges from 2 to 5 cm thick.
Square N94E109
The excavation of N94E109 was complicated by the cement-like
flowstone capping the archaeological horizons, and the semi-
cemented nature of the deposit. We used an angle-grinder to cut
through the flowstone, and lifted it off, and it separated reason-
ably cleanly from the underlying archaeological sediment. Slow
careful excavation with metal tools allowed us to excavate the
semi-cemented archaeological horizons sufficiently cleanly to
allow stratigraphic excavation and 3-D plotting of the finds.
The main part of the section is grouped as a single facies,
despite the fact that it has multiple lenses of material that were
excavated as separate stratigraphic units (Figures 20 and 21). The
Lightly Consolidated MSA Facies has multiple layers of burning
and organic deposition clearly visible in photography, and ranges
in thickness from 20 to 35 cm. Lenses of burnt material may rep-
resent hearths, and they are often associated with burnt fauna,
lithics, and ochre (for example, NEJ at the base of the section,
Figure 21). The entire section is partially cemented with calcium
carbonate, and gypsum occurs as lenses, nodules, and pipes run-
ning into the section, the latter likely to be gypsum infillings of
holes left by tree roots that once snaked through the section.
Capping this anthropogenic facies is a mostly sterile facies of
A)
B)
Figure 19. Site 13B N91E108 west section photo (A) and close-up
photo of hearths (B). The small increments on the scale are 1 cm,
while the large increment is 5 cm.
MIDDLE STONE AGE SITES AT PINNACLE POINT 37
very hard flowstone (5 to 10 cm) that has
seeped down from fissures in the quartzite
above named the Flowstone Facies. It
does include a few lithics and fossil bone,
but these are likely in secondary context.
The flowstone appears to have multiple
events of flowstone deposition. As noted
above, there is also a flowstone on the
south-west wall of the cave that caps an
archaeological horizon. It is possible that
these two flowstones once connected, cap-
ping an LC-MSA deposit that covered the
entire cave floor, and was subsequently
eroded out.
PRELIMINARY SEDIMENTARY
HISTORY OF THE THREE SITES
There is a fairly clear sequence of sedi-
mentary, erosional, and anthropogenic
processes that are evident in all three
caves that we believe can be cross-corre-
lated to some degree. The lack of LSA
remains suggests all anthropogenic activ-
ity is MSA. As will be clear below, we
believe that the series of caves and shel-
ters between 13A and 13C are all part of a
complex that have shared depositional
and formational histories, and henceforth
we use the term “site 13 cave complex” to
refer to them collectively. Below we sum-
marize this sequence as we now under-
stand it, and suggest some tentative
chronometrically meaningful correlations
to the OIS system. Many of our prelimi-
nary interpretations below are based on
our current observations wedded to other
recent stratigraphic and chronometric
results from other coastal caves such as
Die Kelders Cave 1 (Marean et al., 2000)
and Blombos (Henshilwood et al., 2001;
Henshilwood et al., 2002). Forthcoming
radiometric dates for the sites will allow
us to check our correlations and tie these
sequences to the OIS and sea level
sequence.
The site 13 cave complex, along with
site 9, shows substantial evidence for
dune formation, dune cementation, non-
Flow Stone
Lightly Consolidated
MSA
Rock Rock
N94.5E109 N94E109.5
N94.5E109.5
17.99 m asl
N94E109 East Section
N94E109 North Section
17.49 m asl
Figure 20. Site 13B N94E109 north (on left) and east (on right) section drawing.
38 •PaleoAnthropology 2004
North Section
East Section
Figure 21. Site 13B N94E109 north (on left) and east (on right) section photos. The small increments on the scale are 1 cm, while the large
increment is 5 cm.
MIDDLE STONE AGE SITES AT PINNACLE POINT 39
dune sediment deposition (often domi-
nated by roof exfoliation mixed with
anthropogenic formation), flowstone for-
mation (often capping dune and non-
dune surfaces), cementation of MSA
material to the cave walls and cliff face,
and subsequent erosional episodes. Just
east of 13A we have mapped a set of
cemented dune deposits (with MSA
lithics) clinging to the cliff face that also
extend into a very shallow and high cave
above 13A that lacks any deposit on its
floor. Our analyses of the elevational data
show that the flowstone and dune surfaces
of the cemented material in 13B stands
roughly at the same elevation as the high-
est cemented material in 13A and the cave
above. It seems likely that these calcite
formations represent a related event of
discharge of carbonate charged water
from the surrounding quartzite that in
some cases cemented MSA deposit to
abutting cave and cliff wall (13A and 13B)
and in other areas formed a flowstone
capping that same MSA deposit (13B at
N94E109).
Thus, we think it likely that the site 13
cave complex once held a connected set of
MSA deposits that wrapped around the
cliff walls, penetrated into the various
caves and shelters, and likely spilled far
down the current slope toward the ocean.
Remnants of this set of deposits is pre-
served as the LC-MSA Facies in 13B
(N94E109), as well as in the small and
generally isolated sets of MSA cemented
deposits adhering to the cave and cliff
walls in and around 13A and in 13B. We
think it likely that this set of deposits
dates prior to OIS5e. One explanation for
their current eroded state is that rising sea
levels of OIS5e cut and undermined this
massive set of deposits. Wave action
eroded away the exposed deposits on the
slope, ripping out much of the sediments
from the cave mouths, assisted possibly by
stream action (the large cobble-layer
named the Boulder Facies at the base of
N97E97), arising from the generally
greater rainfall (Deacon & Lancaster,
1988) during OIS5e times. The remnant
MSA deposit owes its survival to being
cemented to adjacent walls by calcite dep-
osition near the walls.
Some of the lithics in the lag deposit of
13C may be the remnants of this eroded
LC-MSA. In favor of this hypothesis is the
fact that MSA artifacts are currently erod-
ing out of the 13B cave mouth above the
seaward end of the lag deposit in 13C.
Against this hypothesis are several facts.
First, some of the lag deposit occurs under
13B and to the north, and if these were
from 13B then they must have been
washed back (west) into 13C and north.
However, many of these lithics are fresh
and show no signs of water polishing. Sec-
ond, there is a substantial deposit of mate-
rial in the back of 13C, and MSA lithics are
eroding out of this deposit, so this may
also be a source of lithics for the lag
deposit. And third, the erosional event
must be very ancient, and this means that
the lithics would have been on the surface
of the beach for a very long time. While
some lithics show water polishing, the
majority have sharp, unrounded edges.
Following this erosional event (OIS5)
was a period of deposition resulting from
a combination of aeolian, possibly fluvial
(ground water seepage) transport of silty
sediments as evidenced by the laminated
sediments in N97E97 in 13B, and anthro-
pogenic processes. The laminated sedi-
40 •PaleoAnthropology 2004
ments in 13B may have been cut by ero-
sional events of stream activity emanating
from the back of the caves, though these
sediments were clearly faulted. The lack of
dune deposits in 13B suggests that dunes
were vegetated and inactive during this
time, and/or that 13B was sealed during
the time that dunes were active on the
coastal shelf. A period of fine-grained roof
spalling terminated the formation of lam-
inated facies, after which MSA people re-
occupied the cave and anthropogenic
sedimentation once again became a major
contributor. The occupation in the back of
13B appears to include only one event of
unknown duration, as indicated by the
vertical distribution of lithics. Figure 22
plots the lithics from the most northerly
25 cms of the excavated area: this is done
because the deposits slope strongly to the
south-east, and plotting all would falsely
represent their vertical distribution. As is
clear, there appears to be a single horizon
of lithics that begins at the top of the
Brown Sands MSA Facies, then scatters
downward, with substantial vertical dis-
placement in the fault zone.
N98E96.5 N97E98N98E97
20.45 m asl
19.95 m asl
Roof Spall
Facies
Brown
Sands
MSA
Facies
Surface Disturbed
Facies
Not
Excavated
Not
Excavated
Erosion Gulley
Facies
Laminated
Facies
Boulder
Facies
18.95 m asl
19.45 m asl
Hollow
Space
Fault
Large
Bouldars
N97E96.5-97 North Section
Sediment Removed
by Disturbance
Figure 22. Site 13B northern section of N97E96.5 and N97E97, with lithics from the most
northerly 25 cms in purple.
MIDDLE STONE AGE SITES AT PINNACLE POINT 41
We believe the front of 13B was utilized
for domestic tasks, as indicated by the
hearths in N91E108. Artifact production
and mammal-bone processing occurred in
this area. The back of the cave seems to
have been used as a midden, as repre-
sented by the Brown Sand MSA Facies. It
is possible that this stage of sediment dep-
osition is represented by the darker sedi-
ments eroding out below the dune in 13A,
but not reached by our excavations, and
we think it likely that the dune hides a
much deeper, buried cave that extends
farther into the cliff wall, and may be par-
tially filled with sediment.
The dunes in 9 and 13A document a
period of active dune formation. Both
sites were occupied, 9 only sporadically
and 13A more regularly. 13B was not occu-
pied, nor did dunes blow into the site,
suggesting that the cave may have been
sealed. We believe it likely that this period
of dune formation dates to OIS4 or 3, as it
does at Die Kelders Cave 1 (Mareant al.,
2000) and Blombos (Henshilwood et al.,
2002). A dune blew into site 9, fairly rap-
idly, as suggested by the massive and
homogenous nature of the profile. The
protected nature of the cave retarded any
wind erosion of the sands, which subse-
quently stabilized in the cave. The resist-
ant and dry nature of site 9 minimized any
further sedimentary processes, and there
was little diagenetic modification of the
profile. A shallow dune also blew up
against 13A, which was more exposed and
subject to erosion. Sedimentation and
erosion was interspersed with occupation
as indicated by the White Sands MSA
Facies and the Brown Humic Sands MSA
Facies. Both sites 9 and 13A were then
abandoned for a substantial period of
time, with further dune sand formation
capping the deposit. There was no LSA
occupation at any of the sites.
DESCRIPTION OF THE LITHIC
ASSEMBLAGES
The following section provides a descrip-
tion of the MSA lithics found at 13A and
13B during the 2000 excavation season, as
well as a sample taken from the lag
deposit in 13C. Figures 23 though 26 illus-
trate complete specimens that were cho-
sen for illustration for being representa-
tive of the major categories discussed
below, and Appendix 1 lists the details of
those specimens. The sample from 13C
includes 2 randomly placed 50 cm x 50 cm
squares, with 100% collection in those
squares. An attempt is made here to pro-
vide both a typological and technological
review of the samples from each strati-
graphic aggregate, as described above, as
well as to examine any broader differences
that may exist between 13A and 13B. We
also discuss the nature of the lithic assem-
blages relative to the definition of the
“Mossel Bay Industry” (sensu Goodwin
1930). The small sample of 7 lithics from
site 9 is not described.
Raw Materials
At 13A and 13B the predominant raw
material is quartzite, generally fine-
grained, that varies in color from light
grey (Figure 23A-F), dark grey, light
brown (Figure 23G,H), to red. Quartz,
crystalline quartz (grouped with quartz for
this analysis), and the local cave quartzite
(a coarser material than the preferred
42 •PaleoAnthropology 2004
fine-grained quartzite) are used to a lesser degree, while silcrete,
crypto-crystalline silicates (includes chert and chalcedony), and
hornfels are less commonly used (Figure 27). At many MSA sites
including Klasies River (Singer & Wymer, 1982), Nelson Bay Cave
(Volman, 1981), DKI (Thackeray, 2000), and Montagu Cave
(Keller, 1973) there is a visible change in the frequency of fine-
grained versus course grained raw material represented in the
stratigraphic column. Often a shift towards fine-grained raw
material use, and especially silcrete, has been equated with the
Howiesons Poort Industry (Thackeray, 1992). There is no obvious
change in raw material frequency percentages between strati-
graphic aggregates at 13A and 13B. However, the Brown Humic
Sands MSA Facies of Cave 13A does appear to have a slightly
greater percentage of silcrete when compared with other aggre-
gates, as well as a high diversity of raw materials (Table 1, Figure
27). An increased sample might produce a more striking pattern.
We initiated a lithic raw material survey in 2000 targeted
toward developing an understanding of both primary and sec-
ondary raw material sources. Raw material and its transport has
figured prominently in discussions of the MSA of South Africa
(Ambrose & Lorenz 1990; McBrearty & Brooks, 2000), yet there
are no systematic studies of raw material distributions. Despite
this, silcrete is often classified as “exotic” (e.g. Ambrose & Lorenz
1990). We have initiated a survey to locate and sample possible
source locations for major categories of utilized raw materials
and we are testing several noninvasive methods to chemically fin-
gerprint the sources. We have been seeking to find and charac-
Figure 23. Photographs of lithics from sites 13A and 13B. All are dor-
sal view unless otherwise indicated: A) convergent blade, B) quadri-
lateral blade with view of butt, C) flake, D) flake, E) proximal
fragment with view of butt, F) point, G) point, and H) quadrilateral
blade. The number next to the lithic is its MAP specimen number.
Additional information may be found in Appendix 1.
MIDDLE STONE AGE SITES AT PINNACLE POINT 43
terize sources that are both primary
(sources in their original formation loca-
tion) and secondary (transported by natu-
ral processes).
Initial efforts to find secondary sources
focused on the silcrete rich Grahamstown
formation of the Bredasdorp geological
group (South African Geological Series
3420 1993; 3422A 1993) at the confluence
of the Gourits and Drour Rivers some 12-
15 km NE of Pinnacle Point. This location
was chosen as being an area that would
naturally concentrate silcrete and
quartzite cobbles of the Bredasdorp and
Table Mountain Groups respectively. A
survey of the Gourits River resulted in the
collection of three silcrete cobbles of vari-
able quality within a period of about one-
TABLE 1: THE COUNTS OF RAW MATERIAL TYPES IN THE SITE 13 CAVE COMPLEX LITHIC SAMPLES.
SITE NAME STRATIGRAPHIC AGGREGATE QUARTZITE QUARTZ TMS SILCRETE CCS HORNFELS OTHER GRAND
TOTAL
13A Brown Humic Sands MSA Facies 93 2 2 10 1 4 2 114
White Sands MSA Facies 28 4 1 33
13A Total 121 2 6 11 1 4 2 147
13B Surface Disturbed Facies 5 5
Brown Sand MSA Facies 54 14 8 2 1 2 81
Roof Spall Facies 1 1
Erosion Gulley Facies 3 3
Laminated Facies 1 2 3
MSA Roof Spall Facies 110 6 13 6 1 136
Lightly Consolidated MSA Facies 134 23 5 4 2 1 1 170
13B Disturbance Sample 76 2 9 4 1 92
13B Total 384 45 37 14 5 2 4 491
13C Total Lag Deposit Sample 98 5 5 3 1 4 4 120
Grand Total 603 52 48 28 7 10 10 758
44 •PaleoAnthropology 2004
hour. The beach also contained quartzite
and quartz cobbles as well as shale and
mudstone. A survey of the Drour River (an
ephemeral streambed that was dry at the
time of our visit) discovered a much
higher concentration of silcrete and it was
possible to locate relatively high quality
cobbles and pebbles within a few minutes
of searching. The survey encompassed
approximately 3-4 linear kilometers and
turned up not only silcrete and quartzite
cobbles and pebbles, but also some small
golf ball-size chert and hornfels pebbles.
In general, the higher quality fine-grained
materials occurred in smaller nodules.
Primary context silcrete and fine-
grained quartzite deposits are located
within 10km of Pinnacle Point. A silcrete
outcrop of the Grahamstown Formation
located on a ridgetop above the R327 road
near Hartebeeskuildam, some 9-10 km
from Pinnacle Point was visited. Silcrete
here is of moderate to good quality, and
several samples were collected. One such
sample is virtually identical in terms of
matrix and cortex appearance, color, and
texture to a small piece of silcrete debitage
found in 13B. High quality quartzite simi-
lar to the buttery and dark grey quartzites
found in the Pinnacle Point samples are
located 7 km to the east at Mossel Bay,
outcropping at the Mossel Bay point near
Cape St. Blaize Cave, and extending
beneath the surface of the Indian Ocean.
In several locations below the cliffs there
are a variety of secondary quartzite
sources in the form of beach cobbles.
Along the cliff tops there are raised
beaches with quartzite cobbles along with
several scatters of MSA lithic debris that
may suggest these were a source for raw
material. While the caves of the site 13
cave complex are in quartzite, this coarser
material fractures erratically and does not
seem to have been a preferred raw mate-
rial. We do not know of any other poten-
tial sources for the other raw materials.
An examination of cortical debitage can
give some indication of the primary or
secondary depositional nature of the raw
material source location. Three types of
cortex were noted and recorded: cobble
cortex, outcrop cortex, and rind cortex.
Cobble cortex is smooth, very round,
sometimes dimpled, and is often of uni-
form shade and color. Outcrop cortex is
generally more rough, irregular, and var-
ied in shade and color. Rind cortex is the
chalky matrix that often surrounds sil-
crete nodules, and is found on silcrete
from both primary and secondary con-
texts. Most of the complete quartzite cor-
tical flakes at Pinnacle Point have cobble
cortex (Table 2, Figure 24A, 28). Silcrete
cortical flakes are of both cobble and rind
type. The majority of the CCS, hornfels,
and other raw materials have cobble type
cortex. This suggests that while some of
the raw materials are coming from a pri-
mary context, most, including many of the
more fine-grained raw materials, are com-
ing from a secondary high-energy context.
Source locations for such cobbles would
include a storm beach at a river mouth, an
active riverbed or streambed, a raised
beach, or a geologically uplifted riverbed
deposit.
This limited survey and analysis
demonstrates that the major categories of
raw material used during the MSA for
flake production can still be found within
MIDDLE STONE AGE SITES AT PINNACLE POINT 45
10-15 km of the Pinnacle Point, in both
primary and secondary contexts, but that
secondary contexts were preferred. There
could be submerged raw material sources
in the vicinity of the caves that would have
been available to MSA hominins. Future
surveys will focus on the nearby Gourits,
Grootbrak, and Kleinbrak river mouths,
and surrounding river terrace deposits as
well as outcrops of the Bredasdorp and
Table Mountain Groups.
Technology and Typology
The total sample of plotted lithics from
the 2000 Pinnacle Point excavations con-
sists of slightly more than 600 pieces.
While there is much more material in the
sieved and unplotted sample, it is small
(generally < 1 cm) and fragmented. The
largest excavated samples from 13A come
from the Brown Humic Sands Facies, and
at 13B from the Brown Sand MSA Facies,
MSA Roof Spall Facies, and Lightly Con-
solidated MSA Facies (Table 1). The vast
majority (95%) of the lithic artifacts in all
aggregates are unretouched pieces: flakes,
blades, points, flake and blade fragments,
TABLE 2: THE FREQUENCY OF CORTEX TYPES FOR COMPLETE PIECES IN SITE 13 CAVE COMPLEX LITHIC SAMPLES.
SITE NAME STRATIGRAPHIC AGGREGATE COBBLE OUTCROP RIND NO CORTEX GRAND TOTAL
13A Brown Humic Sands MSA Facies 18 3 0 32 53
White Sands MSA Facies 2 1 0 10 13
13A TOTAL 20 4 0 42 66
13B Surface Disturbed Facies 10 1 0 29 40
Brown Sand MSA Facies 20 1 0 41 62
Roof Spall Facies 0 0 0 2 2
MSA Roof Spall Facies 0 0 1 0 1
Lightly Consolidated MSA Facies 18 2 0 25 45
13B TOTAL 48 4 1 97 150
13C Lag Deposit Sample 13 2 1 19 35
ALL GRAND TOTAL 81 10 2 158 251
46 •PaleoAnthropology 2004
and chunks/block shatter. Retouched
pieces (.76%), cores (2.5%) and hammer-
stones (1.4%) are far less frequent. Com-
plete pieces (flakes, points, blades, and
indeterminate pieces with only minor
damage) account for 33% of the unre-
touched sample. Our analysis will focus on
the unretouched complete pieces due to
the small size of the retouched sample.
All stratigraphic aggregates show a sim-
ilar pattern where flakes dominate, fol-
lowed by blades and then points (Table 3,
Figure 29). Cortex is common (37%) on
the complete piece sample. Primary pieces
(67-100% cortex coverage) are present in
all complete piece samples greater than 10
pieces, suggesting that primary reduction
occurred to some extent in all of the major
stratigraphic aggregates. All aggregates
with reasonable samples show a similar
pattern: the representation of non-cortical
pieces dominates, with a steady decrease
in frequency as the cortex covers more
area (Table 4, Figure 30). We also coded
all lithics with a slightly modified version
of Geneste’s (1988) techno-typology for
Middle Paleolithic stone tool assemblages.
The vast majority of both whole and frag-
mented pieces fall into Stages 1 (21%) and
2 (54%) of the techno-typology system,
suggesting a dominance of the early stages
of manufacture, a result consistent with
the cortical flake frequencies noted above.
We coded platform type following
Thackeray and Kelly (1988:24) with the
addition of the simple facetted category to
describe flake platforms with one to two
faceting scars. Plain platform (Figure 23B,
25E) pieces dominate the sample as might
be expected, with a smaller number of
flakes and blades with facetted (Figure
23E, 23H, 25A, 25C, 26A), dihedral, sim-
ple facetted, crushed, and cortical plat-
forms (Table 5). The dorsal scar pattern
on complete flakes and blades demon-
strates that most of the complete debitage
is produced with parallel (36.4%) and
convergent (24.4%) flaking with a small
percentage produced using radial flaking
(4.2%) and other methods.
Cores are rare but are more numerous
than retouched pieces. There are a total of
20 cores, of which 16 are complete. The
majority of all cores are produced on
Figure 24. Photographs of lithics from sites
13A and 13B. All are dorsal view unless oth-
erwise indicated: A) cortical flake, B) multi-
ple platform core on flake, and C) disc core
with side view. The number next to the lithic
is its MAP specimen number. Additional
information may be found in Appendix 1.
MIDDLE STONE AGE SITES AT PINNACLE POINT 47
quartzite and quartz. Core classification
follows Volman (1981). Disc cores (Figure
24C, 26D) make up nearly half of the core
sample (9), followed by single platform (4,
Figure 26E), and multiple platform cores
(2, Figure 24B is one on a flake). Other
core types represented include conical,
convergent, core-on-flake, and minimal
cores. Over half of the cores have some
percentage of cortex on their total surface.
The direction and orientation of negative
flake scars on the core surface give some
idea of flake production technique. Radial
flaking is the predominant core prepara-
tion method followed by unidirectional
parallel flaking and multiple direction
flaking. This is interesting in light of the
fact that most of the complete flakes show
parallel and convergent flake preparation.
It appears that in their final stage of
reduction, a large number of cores are
flaked radially before they are discarded.
Only 5 retouched pieces were found in
the excavations, all at site 13B. Four of the
five are notched pieces, and the fifth is a
denticulate. One fragment of a Still Bay
point was found on the surface of 13C.
There are nine hammerstones, all of
which are found in Cave 13B. Battered
pebbles are included in this category. The
TABLE 3: THE COUNTS OF MAJOR SHAPE TYPES FROM THE COMPLETE PIECES IN THE SITE 13 CAVE COMPLEX LITHIC SAMPLES.
STRATIGRAPHIC AGGREGATE FLAKE POINT BLADE INDETERMINATE GRAND TOTAL
13A Brown Humic Sands MSA Facies 33 5 14 1 53
White Sands MSA Facies 8 1 3 1 13
13A TOTAL 41 6 17 2 66
13B Surface Disturbed Facies 2 0 0 0 2
Brown Sand MSA Facies 27 5 7 1 40
Roof Spall Facies 0 0 0 1 1
MSA Roof Spall Facies 27 1 13 4 45
Lightly Consolidated MSA Facies 51 3 5 3 62
13B 13B TOTAL 107 9 25 9 150
13C Lag Deposit Sample 15 1 11 8 35
ALL GRAND TOTAL 163 16 53 19 251
48 •PaleoAnthropology 2004
hammerstones are all either quartzite cobbles or pebbles.
During excavation we thought we noticed a difference in the
size of artifacts between 13A and 13B. While this is true (Table 6),
there are no statistical differences in the size of flakes, blades, and
points between the two sites (Figure 31). However, there is a dif-
ference in the proportions between length and width of points
between the two sites. At 13A, length and width increase in con-
cert, while at 13B as points get longer they also become less wide.
The samples, however, are small and this pattern remains uncon-
firmed.
Discussion of Lithics
The lithic assemblages from the sites 13A and 13B are composed
predominantly of fragmented pieces, followed by whole pieces
(flakes, blades, and points), and a small sample of cores and
retouched pieces. The abundance of cortical pieces represented in
all technological categories, as well as an almost complete lack of
finished retouched products suggest that the lithics are represen-
tative of early and middle phases of lithic tool production. Raw
materials are diverse but are dominated by high quality
quartzites.
While research near Mossel Bay has not been intense, early
research in this area was fundamental to the definition of the
MSA in South Africa and the development of nomenclature still
in use today. Leith’s excavations at Cape St. Blaize in 1888
resulted in a described sample of quartzite blades and points with
few stratigraphic details (Leith, 1898). This same sample was
used to define the Mossel Bay Culture (van Hoepen, 1926). The
1524
1643
2191 3275
1 cm
1 cm
1733
A)
B)
C)
D)
E)
Figure 25. Drawings of lithics from sites 13A and 13B. All are dorsal
view unless otherwise indicated: A) blade with facetted butt, B) blade
with cortical butt C) flake with facetted butt, D) flake with facetted
butt, and E) blade. The number next to the lithic is its MAP specimen
number. Additional information may be found in Appendix 1.
MIDDLE STONE AGE SITES AT PINNACLE POINT 49
Cape St. Blaize sample was then supplemented with visits to the
cave and surface collections at Gouritz River and Knysna Heads
that resulted in the recognition of the Mossel Bay Variation
(Goodwin & Van Riet Lowe, 1929) later termed an Industry
(Goodwin, 1930).
Cape St. Blaize was re-excavated by Goodwin and the lithics
and stratigraphy received its first comprehensive description
(Goodwin & Malan, 1935), but the analysis focused on a subjec-
tively defined sub-sample and the illustrations were of special
pieces that are hardly characteristic of the total sample (Marean,
pers. obs.). Keller’s analysis of Goodwin’s sample provided a
quantitative analysis (Keller, 1968), but Goodwin’s layers were
collapsed into one aggregate and the analysis focused only on
retouched pieces, which comprise a tiny percentage of the assem-
blage. Sampson (1974) maintained the use of the Mossel Bay
Industry in concert with accepted terminological protocols, but
noted its poor documentation. The Mossel Bay Industry was later
subsumed into the MSA2 (Volman, 1984) though recently the old
formal name has been used to describe materials at Klasies River
(Wurz 2002). Wurz (Wurz, 2002; Wurz et al., 2003) suggests
continuing the use of the term Mossel Bay, though she prefers
“sub-stage” over “Industry”. Wurz considers the Klasies River
lithics from the SAS member to belong to the Mossel Bay sub-
stage (= MSAII of Singer and Wymer and MSA2b of Volman),
largely on the basis of metric criteria as a reflection of technology.
Through this long history the description and definition of the
industry has fluctuated as the nature of the samples, and the
3295
3192
326
1772 2251
1 cm
1 cm
A)
B) C)
D) E)
Figure 26. Drawings of lithics from sites 13A and 13B. All are dorsal
view unless otherwise indicated: A) point with facetted butt, B) flake
C) flake fragment with retouch, D) disc core, and E) single platform
core. The number next to the lithic is its MAP specimen number. Addi-
tional information may be found in Appendix 1.
50 •PaleoAnthropology 2004
preferences of the researchers to sample,
have changed. We have examined the
Cape St. Blaize collection from Goodwin’s
excavations and note that the lithics in the
collection are large, probably a reflection
of the rather coarse screening methods
Goodwin used. It seems questionable to
define an industry on a skewed sample
that has yet to be completely analyzed and
published. The use of the term “Mossel
Bay Industry” assumes the presence of a
regional pattern, and it is by no means
clear that this exists, and so it is critical to
more precisely define the characteristics
and variability of the MSA lithics from the
Mossel Bay area. While the site 13 cave
complex collections are small, they do
provide one of the few unbiased collec-
tions of MSA lithics from the Mossel Bay
area.
The original definitions of the Mossel
Bay were grounded in typology: “All the
implements are of one general shape; they
may be described thus: Longitudinally
trimmed flakes, trimmed by the removal
of two or at most three convergent flakes”
(Goodwin & Van Riet Lowe 1929:136). He
continues, “The flakes generally form
points with a general shape much [like]
that of an acute-angled isosceles triangle.
TABLE 4: THE FREQUENCY OF CORTEX COVERAGE TYPES FOR COMPLETE PIECES IN SITE 13 CAVE COMPLEX LITHIC SAMPLES.
SITE NAME STRATIGRAPHIC AGGREGATE 0 1-33% 34-66% 67-99% 100% INDETERMINATE GRAND TOTAL
13A Brown Humic Sands MSA Facies 30 12 7 2 0 2 53
White Sands MSA Facies 9 2 1 1 0 0 13
13A Total 39 14 8 3 0 2 66
13B Surface Disturbed Facies 2 0 0 0 0 0 2
Brown Sand MSA Facies 30 6 4 0 0 0 40
Roof Spall Facies 0 1 0 0 0 0 1
MSA Roof Spall Facies 25 10 7 2 1 0 45
Lightly Consolidated MSA Facies 40 13 6 1 2 0 62
13B Total 97 30 17 3 3 0 150
13C Lag Deposit Sample 22 8 5 0 0 0 35
ALL GRAND TOTAL 158 52 30 6 3 2 251
MIDDLE STONE AGE SITES AT PINNACLE POINT 51
A number of parallel-sided rectangular
flakes…point to this form being also a
desired type. Secondary trimming seems
uncommon, but is present in some speci-
mens. Some few oak-leaf types appear. In
all instances the butt of the implement or
flake is facetted [sic] and gently rounded”
(Goodwin 1929:136). The latter comment
about butt-faceting is inconsistent with
TABLE 5: THE FREQUENCY OF PLATFORM TYPES FOR COMPLETE PIECES IN 13 CAVE COMPLEX LITHIC SAMPLES.
SITE STRATIGRAPHIC PLAIN DIHEDRAL FACETTED SHATTERED SIMPLE NO CORTICAL CRUSHED INDETERMINATE
NAME AGGREGATE FACETTED PLATFORM
13A Brown Humic Sands 22 3 12 0 3 0 4 4 5
MSA Facies
White Sands MSA 2 3 1 1 1 0 1 2 2
Facies
13A TOTAL 24 6 13 1 4 0 5 6 7
13B Surface Disturbed 0 0 0 0 0 0 0 0 2
Facies
Brown Sand MSA 14 3 7 1 5 0 1 4 5
Facies
MSA Roof Spall 7 0 3 0 2 0 2 0 31
Facies
Roof Spall Facies 0 0 0 0 0 0 0 0 1
Lightly Consolidated 26 9 6 0 9 3 2 7 0
MSA Facies
13B TOTAL 47 12 16 1 16 3 5 11 39
13C Lag Deposit Sample 16 3 4 1 2 1 3 3 0
ALL GRAND TOTAL 87 21 33 3 22 4 13 20 48
52 •PaleoAnthropology 2004
our results at Pinnacle Point (Figure 23B,
25E, Table 5), and during our examination
of the Cape St. Blaize assemblage from
Goodwin’s excavations we discovered
plain platforms (as did Keller, 1968).
From a typological perspective, our sam-
ples are similar to those from Cape St.
Blaize as described by both Goodwin
(Goodwin & Malan, 1935) and Keller
(Keller, 1968) if one keeps in mind that
those original samples are likely some-
what biased.
The metrics tell a different story. Good-
win (Goodwin & Malan, 1935:131) pro-
vided metric data for points only, while
Keller provided metric data on a variety of
classes but only retouched pieces, and he
did not sub-divide the sequence by layer.
Goodwin’s point measurements range in
mean length from 83 mm in the lower
(C4) layer to 65 mm in the upper (C1)
layer: there is a steady decrease in point
length over time (Figure 31). The Klasies
River points (Wurz et al., 2003, Table 3:
Figure 27. The percentage of raw
materials of all lithics by site
and stratigraphic aggregate
from the site 13 cave complex.
CCS = crypto-crystalline sili-
cates, while TMS (Table Moun-
tain Sandstone) refers to the
local coarse-grained quartzite.
MIDDLE STONE AGE SITES AT PINNACLE POINT 53
1118) fall comfortably within the range of the upper layers at Cape
St. Blaize (Figure 31), while those from the site 13 cave complex are
significantly smaller than both. Similarly, the blades from the site
13 cave complex are significantly smaller than those at Klasies
River (compare our Table 6 with Wurz et al. 2003: Table 1: 1110).
From these metric data, and tentatively, there are several possible
alternative conclusions: 1) the site 13 cave complex material is not
the same as the classic Mossel Bay Industry, 2) the Mossel Bay
Industry shows a time vectored diminution, as first recognized by
Goodwin at Cape St. Blaize and reaffirmed by Wurz’s recent results
from Klasies River, and the site 13 cave complex material is at the
end of this diminution, or 3) the Cape St. Blaize material is biased
toward larger pieces and does not adequately represent the range
of variation in the Mossel Bay region. An unbiased sample from a
dated new excavation at Cape St. Blaize would provide a starting
point to answer this question.
ANALYSIS OF THE MOLLUSK REMAINS
The following section provides a description of the shellfish
remains from sites 9, 13A, and 13B. Observations include species
identification, MNI counts and percentage frequency of MNI and
weight (grams). For the purpose of this analysis, observations were
obtained from quantified marine shells plotted during excavations
and recovered from the 10 mm sieve. Marine shell was also
retained in the 3 mm and 1.5 mm sieves. The analysis of this finer
material might provide a more complete picture of the shellfish
Figure 28. The percentage of cortex coverage categories for com-
plete pieces by site and stratigraphic aggregate from the site 13
cave complex.
54 •PaleoAnthropology 2004
Figure 29. The percentage of
major shape categories of
complete pieces by site and
stratigraphic aggregate from
the site 13 cave complex.
Figure 30. The percentage of
cortex coverage classes for
complete pieces by site and
stratigraphic aggregate from
the site 13 cave complex.
MIDDLE STONE AGE SITES AT PINNACLE POINT 55
assemblages, particularly once the sample
size of marine shells is increased with fur-
ther excavations. So far, the amount of
marine shell recovered from all sites com-
bined does not exceed 0.6 kg, which is a
very small sample. The few observations
described and discussed in this report are
strictly preliminary.
Methodological and taphonomic
considerations
Analysis was undertaken on all shell mate-
rial, including that of terrestrial origin and
macro sediments, such as water-worn
shells. Quantification (counting and
weighing) was only undertaken on shell
remains that resulted from collections
made by people in the past. Although
quantified, fresh-looking shells (e.g. the
presence of periostrum on marine shells)
were excluded from the final tables as
these were likely deposited during visits to
the caves by local 20th Century fishermen.
Such remains are present on the cave floor
today, and occur in the surface layers.
Although present in fair quantities and
0
10
20
30
40
50
60
70
80
90
100
C1 C2
C3
C4
MSAI
MSAII L
MSAII U C1C2
C3
C4
MSAI MSAIIL
MSAIIU
CSB Mean Length
KR Mean Length
13A Length
13B Length
13A Width
13B Width
KR Mean Width
CSB Mean Width
Figure 31. The median (solid bar), 25th/75th percentile (box), 10th/90th percentile (whisker),
and maximum outlier (dot) of point lengths and widths from 13A and 13B. Also included are
the mean point lengths and widths from Cape St. Blaize cave (CSB) and Klasies River (KR)
from Goodwin and Mason (1935) and Wurz et al. (2003), respectively. These are plotted for
each layer (CSB) or analytical aggregate (KR), with CC1-4 from CSB and MSAI-IIU for KR.
56 •PaleoAnthropology 2004
recorded in spreadsheet files, water-worn
shells were also excluded from the final
tables as they were deemed to not be the
direct result of shellfish consumption by
people in the past. Based on earlier work
with LSA shell-bearing deposits along the
South African West Coast (Jerardino,
1997), it is clear that the presence of
water-worn shells in archaeological sites
do not reflect the deliberate transport of
these to camp sites. Instead, their pres-
ence in archaeological sites is likely the
result of water-worn shells clinging to the
byssus (threads of attachment) of rocky
shore mussels (e.g. black, brown and
ribbed mussels) which were collected by
people and subsequently taken back to
sites for processing and consumption.
During the washing and labeling of the
shell material it became apparent that
land snail shells were less frequently bro-
ken than those of marine origin. If people
collected land snails while also collecting
marine shellfish during the same visits,
then both land snail and marine shells
would have been discarded and affected in
the same way by post-depositional factors.
For this reason, it was expected that land
snail and marine shells should show a
similar degree of breakage. Upon initial
visual inspection, this expectation was not
confirmed. After quantifying the number
of broken land snail and marine shells for
each site, it became clear that marine
shells are markedly more fragmented than
land snails (Table 7). Although the overall
Minimum Number of Individuals (MNI)
for each site is very small, the percentage
of unbroken land snail shells is markedly
higher (between 77% and 50%) than that
for marine shells (between 20% and 0%).
Unless some significant differences in
shell density and toughness exist between
land snail and marine shells, this differ-
ence (if representative of the larger por-
tion of shells at these sites) could reflect
different depositional histories.
The sequence of taphonomic events
regarding the presence and preservation
of land snail and marine shells was likely
as follows. After shellfish were brought
back to site and their flesh consumed,
shells were discarded and subsequently
fragmented as a result of food processing
(heating, opening of bivalves and scoop-
ing of limpet contents) and trampling.
After the site was abandoned, land snails
moved into the site as part of their usual
foraging rounds and some subsequently
died there. Other agents responsible for
the introduction of land snails to the site
seems unlikely, as one of their usual pred-
ators (field mice) are always capable of
breaking up their shells (Jerardino, per-
sonal observation, Namaqualand 1989-
1996). Moreover, during field surveys,
whole and empty snail shells are not rare
among bushes (on bare surfaces and on
those of open sites) and away from mice
nests. This clearly shows that land snails
can die with their shells undamaged after
moving across the landscape during their
short life spans. Consequently, land snails
do not appear to have been collected by
people in the past.
For the above reasons, land snails are
not included in the final results of the
shellfish analysis at this stage. Neverthe-
less, as additional shell samples become
available with further excavations, this
conclusion should be revised before pro-
MIDDLE STONE AGE SITES AT PINNACLE POINT 57
TABLE 6: SUMMARY STATISTICS FOR THE COMPLETE FLAKES, BLADES, AND POINTS IN SITE 13 CAVE COMPLEX LITHIC SAMPLES.
13A 13B
FLAKES
Length Thickness Width Length Thickness Width
Mean 39.6 7.9 28.1 40.1 7.4 25.5
Std 17.4 4.0 9.5 18.0 3.5 11.6
95% Confidence Limit 6.0 1.4 3.3 3.5 0.7 2.2
Sample Size 41 41 41 107 107 107
Minimum 13 2 7 11 2 8
Maximum 89 18 53 91 20 75
BLADES
Length Thickness Width Length Thickness Width
Mean 35.3 6.9 24.7 39.1 9.2 29.2
Std 21.2 4.9 9.0 16.6 6.7 10.7
95% Confidence Limit 11.3 2.6 4.8 6.8 2.8 4.4
Sample Size 17 17 17 25 25 25
Minimum 9 3 12.96 15 3 16
Maximum 84 17 41 86 38 59
POINTS
Length Thickness Width Length Thickness Width
Mean 36.8 7.5 23.7 41.2 8.6 28.4
Std 13.7 3.8 6.1 13.1 3.5 6.2
95% Confidence Limit 14.4 4.0 6.4 10.1 2.7 4.7
Sample Size 6 6 6 9 9 9
Minimum 17 3 16 25 6 17
Maximum 59 14 32 60 15 37
58 •PaleoAnthropology 2004
ceeding with the analysis of shell remains.
Also, the preliminary interpretation of the
presence of land snails should also be
tested with experiments designed to com-
pare possible differences in the strength
and toughness of both land snail and
marine shells.
Site 9
An extremely small amount of relatively
well preserved shells (total MNI = 12; total
weight = 27.4 g) were recovered from site
9. This small amount of shell is further
dwarfed by the considerable amount of
excavated deposit (2 cubic meters, Table
8). The low density of marine shells along
with the presence of few cultural remains
appears to indicate very few and brief vis-
its to the cave by people. Only three shell-
fish species are represented in the Site 9
sample (Table 8), namely one species of
limpet (Patella oculus), the brown mussel
(Perna perna), and the black mussel
(Choromytilus meridionalis). At least two
brown mussels and all four black mussels
recovered from PP9 are sub-adults. It is
highly unlikely that people would have
collected sub-adult individuals while
ignoring larger adult-sized mollusks. In
shell middens of clearly human origin
along the South African west coast (Jer-
ardino, 1997), sub-adults are always out-
numbered by adult individuals by at least
one, and frequently two orders of magni-
tude. Consequently, it is likely that agents
other than humans incorporated much of
the marine shell into the site 9 deposits.
Only the presence of one limpet can be
attributed directly as a result of foraging
rounds by people along the nearby shore.
The remaining brown mussels are too
fragmented for their identification as
either adults or sub-adults. In sum, shell-
fish collection by humans using site 9 was
almost non-existent.
Site 13A
Compared to site 9, a much larger number
of marine shells were recovered from 13A
(Tables 9 and 10). A total of twelve species
are present in this assemblage, although
not all of these species are present
throughout the stratigraphic sequence
TABLE 7. PERCENTAGE FREQUENCY OF WHOLE (UNBROKEN) LAND SNAIL AND
MARINE SHELLS FOR EACH SITE ASSEMBLAGE. TOTAL N IS TOTAL MNI AND NOT
TOTAL NUMBER OF COUNTABLE SHELLS (BIVALVES AND LIMPETS DIFFER IN THE
EQUIVALENCE OF THESE TWO NUMBERS).
SITE TYPE OF SHELL FREQ. (%) TOTAL N
9 land snail 62.3 61
marine shell 20.0 10
13A land snail 76.9 13
marine shell 8.9 55
13B land snail 50.0 2
marine shell 0.0 25
MIDDLE STONE AGE SITES AT PINNACLE POINT 59
and at this stage of fieldwork. Sub-adult
brown and black mussels are present,
although in very small numbers, through-
out the stratigraphic sequence. Limpet
species collected from 13A include P. bar-
bara, P. oculus, P. granularis and possi-
bly P. tabularis. In a few instances, weight
of shell was estimated, as varying quanti-
ties of sand was found cemented to a num-
ber of shell fragments. A total of 1.5 grams
of shell was not possible to identify, and
thus named “unidentified”. Marine shells
from 13A appear reasonably well pre-
served, although somewhat brittle and/or
chalky at times when handled during
analysis (particularly P. perna).
P. perna and Turbo sarmaticus domi-
nate the 13A shell assemblage at this stage
of field work. According to the list of
species (Tables 9 and 10), the large major-
ity of shellfish was collected from the mid-
and lower intertidal rocky shores. Among
these species, T. sarmaticus, Haliotis spp,
chitons and Cymatium cutaceum are
found in the lower reaches of the intertidal
and below. Large adult brown and black
mussels (P. perna and C. meridionalis)
are also found in the low intertidal and
subtidal shores (Kilburn & Ripey 1982,
Jerardino personal observation). How-
ever, size observations for black and
brown mussel shells are not available due
to extensive breakage of shell remains.
The sample size of marine shells so far
recovered from 13A is very small, and thus
little can be said at this stage. For
instance, the Brown Humic Sands facies
appears to have a relatively larger number
of shells and higher richness of species
when compared to the older and younger
stratigraphic components. Nevertheless,
this higher species richness in Brown
Humic Sands could just be the result of
larger quantities of deposit recovered dur-
ing excavations, and thus an artifact of a
larger sample size.
Site 13B
Despite excavation efforts in three areas
of this cave, a very small amount of
marine shell was recovered (nearly 97
grams overall) (Table 11). This mass is
three times less than that excavated from
only one square in 13A. An additional
amount of marine shell (about 180
TABLE 8. MNI COUNT AND WEIGHT (G) OF MARINE SHELL REMAINS AT SITE 9
THROUGHOUT ITS STRATIGRAPHIC SEQUENCE ORDERED TOP (LEFT) TO BOTTOM
(RIGHT).
STRATUNITS ND-NE STRATUNITS NH-NH1 STRATUNITS NI-NI7
SPECIES MNI w MNI w MNI w
Patella oculus 1 5.3 0 0 0 0
Perna perna 0 0 3 8.5 4 7.4
Choromytilus 0 0 0 0 4 6.2
meridionalis
60 •PaleoAnthropology 2004
grams), however, was obtained from the
disturbed material swept from the surface
of the cave and believed to result from the
disturbance by fishermen (Table 12). Shell
preservation appears to vary from one
area of excavation to the next. In
N91E108, a small amount of burnt shell
was found and only in a few instances the
shell weight was estimated because of
clumps of sand cemented to other shell
fragments. In N94E109, burnt and chalky
shell is noticeable, and clumps of sand
were removed from shell fragments when-
ever possible before weighing, although
this was not always possible. In N97E97
and N97E96, shell was relatively well pre-
served and in only one instance the shell
weight was estimated because of sand
cemented to shell fragments.
Overall, P. perna and T. sarmaticus
appear in 13B with highest frequency,
with Patella spp. (P. argenvillei and P.
longicosta) and the sandy-bottom white
mussel Donax serra in moderate quanti-
ties (Table 11). The analysis of the marine
shell from the disturbance shows the same
species comprising the large majority of
the assemblage, with Patella spp. domi-
nating above the other species (Table 12).
Very little can be inferred, at this stage,
from the available observations due to the
small amounts of shell material. According
to the list of species (Tables 11 and 12), the
large majority of shellfish taken back to
13B was collected from the mid- and lower
intertidal rocky shores, with some visits to
sandy beaches where D. serra was col-
lected.
Haliotis spp, chitons and Crepidula spp
are not present in 13B (but present in 13A),
and their absence could just be the result
of the very small quantities of shell recov-
ered from this site at this stage. D. serra
TABLE 9. MNI COUNT AND PERCENTAGE FREQUENCY (%MNI) OF MARINE SHELL
REMAINS AT SITE 13A THROUGHOUT ITS STRATIGRAPHIC SEQUENCE. LBSS:
LIGHT-BROWN SURFACE SANDS FACIES; BHS: BROWN HUMIC SANDS MSA
FACIES; WS: WHITE SANDS MSA FACIES.
LBSS BHS WS
SPECIES MNI % MNI % MNI %
Perna perna 0 0 31 81.6 13 68.4
Choromytilus meridionalis 0 0 1 2.6 0 0
Turbo sarmaticus 1 33.3 3 7.9 0 0
Haliotis spp 0 0 1 2.6 0 0
Patella spp 1 33.3 1 2.6 4 21.1
Crepidula sp. 0 0 1 2.6 1 5.3
Chiton 1 33.3 0 0.0 0 0
Turritellidae 0 0 0 0 1 5.3
TOTAL MNI 3 100 38 100 19 100
MIDDLE STONE AGE SITES AT PINNACLE POINT 61
and barnacle, however, are present in 13B,
and are absent from 13A. Either this
observation reflects contrasting behavioral
and environmental data between 13A and
13B, or it might be an anomaly resulting
from the small sample size.
Discussion of Shellfish Remains
MSA inhabitants of sites 13A and 13B col-
lected marine shellfish from nearby rocky
reefs and very occasionally also from sandy
beaches. Their collecting rounds targeted
the mid- and lower reaches of the inter-
tidal from where brown mussels (P.
perna), arikreukel snails (T. sarmaticus),
limpets (Patella spp), chitons (Polypla-
cophora) and black mussels (C. meridion-
alis) were collected during low spring
tides. Other species were also gathered
(e.g., ear shells or Haliotis spp, white mus-
sel [D. serra] and corrugated white mussel
[Venerupis corrugate]), but in much
smaller quantities.
Although present in the caves, terres-
trial mollusks (land snails) were appar-
ently not collected by MSA people. These
mollusks are likely to have entered the
caves (perhaps attracted by the garbage
left behind) after the abandonment of
these sites on several occasions during
sediment deposition. The relative fre-
quency of land snail shells differ from site
to site, as does the density of marine shells.
TABLE 10. SHELL WEIGHT (G) AND PERCENTAGE WEIGHT OF MARINE SHELL
REMAINS AT SITE 13A THROUGHOUT ITS STRATIGRAPHIC SEQUENCE. LBSS:
LIGHT-BROWN SURFACE SANDS FACIES; BHS: BROWN HUMIC SANDS MSA
FACIES; WS: WHITE SANDS MSA FACIES.
LBSS BHS WS
SPECIES w % w % w %
Perna perna 3.5 19.6 164.7 73.6 48.4 53.7
Choromytilus meridionalis 0 0 3.7 1.6 1.6 1.7
Venerupis corrugata 0 0 0.1 <0.1 0 0
Turbo sarmaticus 10.2 57.0 37.8 16.9 0 0
Haliotis spp 0 0 0.7 0.3 0 0
Burnupena sp. 0 0 1.2 0.5 0 0
Cymatium cutaceum 0 0 0.4 0.2 0 0
Patella spp 0.4 2.2 3.5 1.6 39.1 43.4
Helcion sp. 0 0 0.2 0.1 0 0
Crepidula sp. 0 0 0.4 0.2 0.3 0.3
Chiton 3.8 21.2 9.5 4.2 0 0
Turritellidae 0 0 0 0 0.8 0.9
Unidentified 0 0 1.5 0.7 0 0
TOTAL WEIGHT (G) 17.9 100 223.7 100 90.2 100
62 •PaleoAnthropology 2004
ANALYSIS OF REPTILE, BIRD,
AND MAMMAL REMAINS
As stated above, one of our long-term
research goals is to test several competing
models of human behavioral modernity in
the MSA. Faunal exploitation has become
a critical point of debate in this issue
(Marean & Assefa, 1999). South African
faunal assemblages have been used to
develop “the less effective hunter” model
championed by Klein and colleagues
(e.g.Klein, 1995; 2000), though the MSA
component of the empirical record under-
lying that model is somewhat thin. For
that reason, a goal of this project is to
improve our understanding of MSA faunal
exploitation through the production and
study of well-excavated and well-pre-
TABLE 11. MNI COUNT AND WEIGHT (G) OF MARINE SHELL REMAINS AT SITE 13B IN SQUARES N91E108, N94E109, AND
N97E97-N97E96. SDF: SURFACE DISTURBED FACIES; MRS: MSA ROOF SPALL FACIES; LC-MSA: LIGHTLY CONSOLIDATED MSA
FACIES; AND WA AND WB ARE THE MAJOR STRATIGRAPHIC UNITS COMPRISING THE BROWN SANDS MSA FACIES.
N91E108 N94E109 N97E97-N97E96
SDF MRS LC-MSA WA WB
Species MNI w MNI w MNI w MNI w MNI w
Perna perna 1 1.0 1 10.2 2 7.6 0 0 1 2.8
Choromytilus meridionalis 0 0 0 0 1 2.9 0 0 0 0
Donax serra 2 9.3 2 21.8 0 0 0 0 0 0.3
Veneridae 0 0 0 0 0 0 0 0 0 0.3
Bivalve 0 0 0 0 0 0.3 0 0 0 0
Turbo sarmaticus 0 4.3 0 1.5 1 1.3 0 0 1 6.7
Patella spp 0 0 0 2.3 0 0 1 19.0 0 0
Nucella squamosa 0 0 0 0 1 4.3 0 0 0 0
Barnacle 0 0 0 0 0 0 0 0 0 0.7
Unidentified 0 1.8 0 0 0 0 0 0 0 0
TOTAL (MNI/WEIGHT) 3 16.4 3 35.8 5 16.4 1 19.0 2 10.8
MIDDLE STONE AGE SITES AT PINNACLE POINT 63
served faunal samples. At this stage, the
small size of the faunal assemblages from
our excavations restricts us from making
any confident conclusions, but our
descriptions here show that enlarged sam-
ples hold great promise, and the current
small samples already extend the range of
variation seen in MSA faunal assem-
blages.
Bone preservation is excellent in all 3
caves that were tested due to calcium car-
bonate enriched water seeping through
the walls and roof of the caves. Only a few
South African sites dating to the MSA pre-
serve animal bones and the sites of Klasies
River and Die Kelders Cave 1 have yielded
the largest samples thus far reported. Fau-
nal remains are dense relative to exca-
vated sediment at 13A and 13B. Thus far
we have examined and analyzed all plot-
ted and identifiable bone from the three
tested caves and the following account
pertains to those finds.
Methodology
Excavated materials were washed on site
with fresh water and where necessary,
specimens were washed again in the labo-
ratory at the South African Museum in
Cape Town. Bones coated in adhering
matrix were carefully cleaned by soaking
for several days in luke-warm water and
brushing off adhering matrix with soft
nylon-bristled brushes. In cases where
adhering matrix was very tough and diffi-
cult to remove, bone surfaces were
cleaned piecemeal with a mechanical tool
commonly used by palaeontologists to
prepare fossils. Special care was taken not
to damage cortical surfaces. Specimens
were identified to taxon and/or species
and skeletal element where possible. Each
specimen was examined macro- and
microscopically (using a 10-40X zoom
binocular microscope) for fracture fea-
tures, weathering, burning and bone sur-
face modification such as marks produced
by animal gnawing as well as percussion
and cut marks produced by humans.
Information from detailed examination
was computer coded using a system devel-
oped by Marean (Marean et al., 2001).
Here we present a general overview of the
TABLE 12. MNI COUNT AND WEIGHT (G) OF MARINE SHELL REMAINS AT SITE 13B
FROM THE DISTURBED DEPOSITS.
SPECIES MNI %MNI W %W
Perna perna 2 14.3 8.9 4.9
Donax serra 1 7.1 7.4 4.2
Turbo sarmaticus 3 21.4 10.2 5.7
Patella spp 7 50.0 142.4 79.4
Burnupena sp. 1 7.1 10.1 5.6
Helcion sp. 0 0 0.1 0.1
Unidentified 0 0 <0.1 <0.1
TOTAL (MNI/WEIGHT) 14 179
64 •PaleoAnthropology 2004
nature of faunal remains in the tested
sites. Minimum Numbers of Individuals
(MNIs) are not given as the samples are
very small and thus such values are of no
real significance.
Site 9
A mere 9 bone specimens were plotted
and they derive from; unidentified micro-
mammal, Cape gray mongoose, chacma
baboon, unidentified bovid and unidenti-
fied carnivore (see Table 13). It is interest-
ing that no tortoise remains were
recovered since tortoise remains are
found in high frequencies in most South
African archaeological deposits. None of
these fragments retain any characteristics
indicative of humans having been the pri-
mary accumulating agents. Interestingly,
the majority of identified fragments are
from carnivores.
Site 13A
The archaeological deposit in this site
consists of two depositional horizons as
described above. Unlike the deposit in
Cave 9, that in Site 13A consists of mate-
rial remains deposited predominantly by
humans. For Site 13A as a whole, 350 plot-
ted, identifiable bone specimens are pre-
sented in Table 13. Fragments from only
two species were identifiable, namely
angulate tortoise and grysbok/steenbok.
The assemblage is dominated by unidenti-
fied mammal, followed by unidentified
bovid and tortoise bone. Carnivore and
unidentified micromammal bone is rare.
Damage caused by animal gnawing is rare
while humanly induced damage occurs
more commonly, indicating that the fau-
nal remains in this site were mostly col-
lected and deposited by humans.
Interestingly, fish, snake and bird bone
are entirely absent.
Site 13B
Four of the seven identified Facies in Site
13B consist of deposits with high densities
of archaeological material and those are
the focus of our presentation here. Overall,
faunal preservation in 13B is excellent and
bone fragments are present in high fre-
quencies. Some fragments have a thin
matrix adhering to the surface (Figure
32C). This is easily removed mechanically.
A total of 1479 plotted, identified fragments
are presented in Table 13 for Site 13B.
A total of 262 plotted, identified bone
fragments were unearthed from the
Lightly Consolidated Facies representing
4 identifiable species: angulate tortoise,
grysbok/steenbok, mountain reedbuck
and eland. Of the bovids, eland are repre-
sented by the most specimens. No bird,
fish, snake or carnivore remains were
found in this Facies. The vast majority of
bones are from larger mammals with
micromammals and small mammals only
represented by 10 specimens.
The sample from the Brown Sand MSA
Facies includes 417 plotted and identified
specimens that include angulate tortoise,
jackass penguin, Cape gray mongoose,
hyrax, springbok, Cape fur seal, black
wildebeest, vaal rhebuck and eland.
Again, of the bovids, eland are repre-
sented by the largest number of frag-
ments. Micromammals and small
mammals are only represented by 9 spec-
imens and carnivore, fish and snake bones
are entirely absent.
MIDDLE STONE AGE SITES AT PINNACLE POINT 65
1 mm
A)
B)
1 mm
C)
1 mm
Figure 32. Photographs of A) a stone tool cutmark, B) a hammer-
stone percussion mark, and C) matrix adhering to the surface of a
fragment from site 13B. All photographs taken with a 10-40X zoom
binocular microscope, and the scale is 2 mm long.
66 •PaleoAnthropology 2004
TABLE 13. THE NUMBER OF IDENTIFIABLE SPECIMENS (NISP) PER STRATIGRAPHIC AGGREGATE FOR THE THREE EXCAVATED SITES.
REPTILE
Chersina angulata (angulate tortoise) 5 27 32 11 2 0 0 34 32 5 84 0
BIRD
Spheniscus demersus (jackass penguin) 0 0 0 0 0 0 0 1 0 0 1 0
Unidentified Bird 0 0 0 0 0 0 0 1 1 0 2 0
MICROMAMMAL
Chrysochloridae (mole) 0 0 0 0 0 0 0 0 1 0 1 0
Unidentified Micromammal 1 1 2 8 2 1 3 7 7 2 30 1
SMALL MAMMAL
Lepus sp. indet. 0000000010 10
Herpestes pulverulentus (cape grey mongoose) 0 0 0 0 0 0 0 1 0 0 1 1
Procavia capensis (hyrax) 0 0 0 0 0 0 0 1 1 0 2 0
Unidentified Small Mammal 0 0 0 2 0 0 0 0 0 0 2 0
WHITE SANDS MSA FACIES
BROWN HUMIC SANDS MSA FACIES
SITE 13 A TOTAL
LIGHTLY CONSOLIDATED MSA FACIES
PONDING FACIES
EROSION GULLEY FACIES
ROOF SPALL FACIES
BROWN SAND MSA FACIES
MSA ROOF SPALL FACIES
SURFACE DISTURBED FACIES
SITE 13B TOTAL
SITE 9
MIDDLE STONE AGE SITES AT PINNACLE POINT 67
TABLE 13. THE NUMBER OF IDENTIFIABLE SPECIMENS (NISP) PER STRATIGRAPHIC AGGREGATE FOR THE THREE EXCAVATED
SITES. (CONTINUED)
LARGE MAMMAL
Antidorcas marsupialis (springbok) 0 0 0 0 0 0 0 1 1 0 2 0
Artocephalus pusillus (Cape fur seal) 0 0 0 0 0 0 0 1 0 0 1 0
Connochaetes gnou (black wildebeest) 0 0 0 0 0 0 0 1 0 1 2 0
Papio ursinus (chacma baboon) 0 0 0 0 0 0 0 0 0 0 0 1
Pelorovis antiquus (giant buffalo) 0 0 0 0 0 0 0 0 1 0 1 0
Pelea capreolus (vaal rhebuck) 0 0 0 0 0 0 0 2 1 0 3 0
Raphicerus campestris (steenbok) 0 0 0 0 0 0 0 0 2 0 2 0
Raphicerus sp. indet. (steenbok/grysbok) 0 1 1 2 0 0 0 0 6 0 8 0
Redunca fulvorufula (mountain reedbuck) 0 0 0 4 0 0 0 0 1 0 5 0
Taurotragus oryx (eland) 0 0 0 5 1 0 0 6 2 5 19 0
Unidentified Bovid 11 28 39 30 2 1 0 65 111 9 218 1
Unidentified Carnivore 1 1 2 0 0 0 0 0 1 1 2 5
Unidentified Primate 0 0 0 0 0 0 1 0 0 0 1 0
Unidentified Mammal 101 173 274 200 15 7 32 296 502 39 1091 0
WHITE SANDS MSA FACIES
BROWN HUMIC SANDS MSA FACIES
SITE 13 A TOTAL
LIGHTLY CONSOLIDATED MSA FACIES
PONDING FACIES
EROSION GULLEY FACIES
ROOF SPALL FACIES
BROWN SAND MSA FACIES
MSA ROOF SPALL FACIES
SURFACE DISTURBED FACIES
SITE 13B TOTAL
SITE 9
68 •PaleoAnthropology 2004
The sample from the MSA Roof Spall
Facies includes 671 plotted, identified
specimens representing the following
species; angulate tortoise, unidentified
mole, unidentified hare, hyrax, springbok,
extinct giant buffalo, vaal rhebuck, steen-
bok, grysbok/steenbok, mountain reed-
buck, eland and unidentified carnivore
(Table 13). Raphicerus sp. is represented
by the most identifiable specimens among
the bovids. This Facies includes the widest
variety of bovid species when compared to
all facies in this site and the combined
specimens from the other tested sites.
Only one bird bone is present in this
Facies while fish and snake are entirely
absent.
The Surface Disturbed Facies has a
total of 62 plotted, identified bone frag-
ments representing the following identi-
fied species; angulate tortoise, black
wildebeest, eland and unidentified carni-
vore. Eland are represented by the largest
number of fragments. Bird, fish, snake
and small mammals are not represented
by plotted or identified specimens.
DISCUSSION OF REPTILE,
BIRD, AND MAMMAL REMAINS
While the faunal assemblages from sites
13A and 13B are small, they do provide
high expectations for enlarged samples,
and show several interesting patterns.
Most of the fauna from these sites appear
to have been collected by hominids as evi-
denced by the frequent occurrence of cut-
marks (Figure 32A), hammerstone
percussion marks (Figure 32B), and burn-
ing. Hammerstone percussion marked
fragments are far more numerous than
carnivore tooth marked specimens (Fig-
ure 33), and the latter frequencies are far
below what are typical of carnivore frag-
mented assemblages (see Marean et al.
2000, Table 3). The very low frequency of
carnivore tooth marks suggests a minimal
involvement of carnivores as either pri-
mary accumulators of faunal material or
secondary ravagers of hominid discarded
bone.
Though the samples are small and thus
the pattern is not conclusive, the repre-
sentation of fauna from both 13A and 13B
is rather unique relative to other MSA
assemblages in South Africa. Unlike most
other MSA cave deposits, remains from
small mammals are not well represented.
Figure 34 groups the remains into tor-
toise, micromammal (mouse-sized
species), small mammal (hare size
species), and large mammal (from small
bovid and larger). Most South African
MSA assemblages have substantial quan-
tities of tortoise. At Die Kelders Cave 1 tor-
toise is either the first or second most
abundant taxon in all MSA layers, being
several orders of magnitude more abun-
dant than any large mammal taxon (Klein
& Cruz-Uribe, 2000). Tortoise appears to
be the most abundant taxon in the Blom-
bos Cave MSA layers (Henshilwood et al.,
2001). While tortoise is abundant at 13A
and 13B, its numbers do not approach
those present at most MSA sites in South
Africa.
Similarly, small mammals are either
the first or second most abundant taxon at
both Die Kelders Cave 1 and Blombos, but
are rare to absent at 13A and 13B. Contin-
MIDDLE STONE AGE SITES AT PINNACLE POINT 69
uing this pattern is a rarity of micromam-
mals, which tend to be very abundant in
cave assemblages throughout South
Africa. Large mammals are, almost
uniquely, the dominant faunal component
at 13A and 13B.
Raptors, owls, and small carnivores are
typical accumulators of micromammal,
small mammal, and tortoise remains,
while hyenas and other large carnivores
regularly accumulate the remains of large
mammals in caves. The rarity of micro-
mammals, small mammals, and tortoise
suggests that there was little owl or raptor
involvement in the accumulation of the
fauna at 13A and 13B. The lack of carni-
vore tooth marks, and the abundance of
percussion and cut marks, suggests that
people were the primary accumulators of
the large mammals. If supported by larger
samples, the assemblages from 13A and
13B may provide a faunal collection with a
very “clean” taphonomic signature domi-
nated by human accumulation, unlike
most other South African MSA sites.
The sample size is too small to say any-
thing definitive about the species repre-
sentation, but the presence of several
species suggests an environment different
from today. The combined presence of
wildebeest, springbok, and eland is a clear
indication of open grassy conditions.
Pelorovis antiquus is typically associated
with dry open conditions, and its presence
may suggest an arid grassland ecosystem.
DESCRIPTION OF HOMININ
SPECIMENS
Two hominin fossil fragments were found
during the 2000 season: a cranial frag-
ment and an incisor. Both were found in
deposits derived from the disturbance
(see discussion above). The cranial frag-
ment was found in the middle of the fish-
ermen disturbance, likely very close to its
original in situ position. The incisor was
found down-slope of the disturbed area in
the sediment that had been dug away from
the disturbed area. Their scientific value is
slightly reduced by this secondary context.
However, there are several facts that pro-
vide strong confidence that these remains
are MSA, and further that they derive
from the Brown Sand MSA Facies. First,
there is no LSA deposit in the cave. Sec-
ond, their condition is mineralized and
strongly resembles in color the fossilized
material from the Brown Sand MSA
Facies. Third, in the disturbed area there
is only one fossiliferous horizon, and thus
only one potential source – the Brown
Sand MSA Facies.
Specimen 4500 – Parietal Frag-
ment
This is a well-mineralized left parietal
fragment measuring ca. 65 mm by ca. 55
mm. Approximately 45 mm of the speci-
men’s sagittal suture (sutura sagittalis) is
preserved (Figure 35). The maximum
thickness along the sagittal border is 6.5
mm and the minimum thickness is 5.6
mm. The serrations are open and there is
no evidence of fusion with the opposite
side. Although weathered, the surface of
the inner table preserves substantial mor-
phological detail. The groove for the sagit-
tal sulcus is weakly developed in this
specimen. A pronounced depression of
the arachnoid granulations (pacchionian
bodies) occurs parallel to the sagittal bor-
70 •PaleoAnthropology 2004
der. Faint arterial impressions (impres-
siones arteriosae), representing impres-
sions left by branches of the anterior
ramus of the middle meningeal artery,
radiate towards the sagittal border.
Foramina for emissary veins are visible
across both the internal and external sur-
faces. The surface of the external table has
a mottled roughness to it, a sign of exten-
sive weathering, and there are calcite crys-
tals on this surface. Staining, possibly
manganese, is present on both the inter-
nal and external surfaces.
The Klasies River human skeletal col-
lection indicates that modern parietal
bone morphology was present in South
African MSA populations (Singer &
Wymer, 1982). Thus as a guide, a sample
of parietal bones from a South African
archaeological population was used to
facilitate accurate positioning of the frag-
ment in the skull. The presence of the
sagittal suture made positioning relatively
easy. The parietal bones in the compara-
tive archaeological sample display distinc-
tive changes in bone thickness, curvature
and internal morphology along the sagit-
tal border. Typically, relatively thin bone
with minimal curvature, combine with
depressions of the arachnoid granulations
near the frontal angle, while thicker, rela-
tively curved bone, combine with a pro-
nounced groove for the sagittal sulcus
approximately half way between the
frontal and occipital angles. Specimen
4500 is a relatively thin fragment with
minimal curvature. This, in addition to
% of Total NISP
0 5 10 15 20 25 30
Mammal Size Category (Brain 1981)
1
2
3
4
5
% Tooth Marked
% Percussion Marked
Figure 33. The percentage of total NISP by Mammal Body Size Category (Brain 1981) of tooth
marked and percussion marked fragments at sites 13A and 13B.
MIDDLE STONE AGE SITES AT PINNACLE POINT 71
the weak development of the groove for
the sagittal sulcus and the presence of a
depression of the arachnoid granulations
on the inner surface, suggest that this
fragment probably originated close to the
frontal angle on the parietal bone from
which it derives. The internal morphology
of the fragment, i.e. the depression of the
arachnoid granulations and the groove for
the sagittal sulcus, makes anatomical
sense only when positioned on the left
parietal. On this side, the arachnoid gran-
ulations are closer to the anterior angle
and the sagittal sulcus closer the the
occipital angle. We therefore suggest that
this is a fragment of a left parietal (see
Figure 35).
The lack of fusion along the sagittal
border suggests derivation from either a
juvenile or young adult. The thickness of
this specimen, however, favors derivation
from a young adult rather than a juvenile.
The pronounced depression of the arach-
noid granulations adds weight to this
view, as these depressions are rarely visi-
ble in juveniles, are more pronounced in
adults and are most pronounced in old
individuals (Gray et al. 1977).
Specimen 4501 - Central Incisor
This is a well-preserved, permanent right
central mandibular incisor (Figure 36).
The specimen is complete, except for a
slightly damaged root apex. Estimated
mesio-distal (MD) crown diameter is 5.9
mm and the bucco-lingual (BL) diameter
is 6.2 mm. Estimated crown height is 7.5
mm and root length is ca.12.7 mm. Well-
defined distal and mesial interproximal
% Total NISP
0 20406080
Tortoise
Micromammal
Small Mammal
Large Mammal
13A
13B
Figure 34. The percentage of total NISP of tortoise and mammal taxa grouped as tortoise,
micromammal (mouse-sized species), small mammal (hare-size species), and large mammal
(from small bovid and larger) at sites 13A and 13B.
72 •PaleoAnthropology 2004
wear facets (IWF) are placed slightly lingual to the midline of the
tooth. Wear is comparable to wear stage 5 in the Murphy System
[as modified by Smith (1984)], and is characterised by a broad strip
of dentine exposed along the incisal edge. The crown outline, as
seen in the lingual view, is marginally asymmetrical, with wear
tapering slightly towards the disto-incisal edge. Consistent with the
identification of this tooth as a central incisor, the distal crown
edges do not flare distally to the degree that they would in lateral
mandibular incisors (White and Folkens 2000). Discernible mesial
and distal marginal ridges merge at a slight cervical enamel promi-
nence, creating a significant degree of shoveling.
Microscopic scratches cover the entire surface of the crown. Ver-
tically orientated scratches are especially evident close to the
incisal edge, on the labial face. Enamel polishing is present over the
entire surface of the crown. Remnants of dental calculus are evi-
dent on the cervical third of all four faces, but are especially pro-
nounced on the mesial and distal faces. The tooth possesses a fully
developed root, which tapers to a closed root apex. It is relatively
long and MD compressed.
Specimen 4501 is a relatively large tooth when compared in
absolute terms with modern African populations (Table 14, Figure
37). However, it falls at the lower end of the range of size variation
displayed by a comparative sample of fossil specimens from the
African Early, Middle and Late Pleistocene. Although worn, an
estimate of Specimen 4501’s crown height, suggests that it may
have had the shortest crown height amongst the fossil specimens in
the comparative sample. It compares more favorably in this
Figure 35. The hominid parietal fragment (4500) showing external
(A) and internal (B) surfaces, and its approximate location on a
human skull (C). The scale is 2 cm long, and C) is not to scale.
MIDDLE STONE AGE SITES AT PINNACLE POINT 73
dimension to modern African specimens.
The estimated MD diameter of Specimen
4501 falls at the upper limits of the range
of a comparative archaeological and mod-
ern African sample. Compared to the MD
diameters of the sample of fossil speci-
mens, it possesses the fourth most MD
compressed crown after the Middle Pleis-
tocene specimens of Atapuerca-SH (popu-
lation mean of 5.5 mm), Mauer (5.5 mm)
and Ternifine (Tighennif) (5.7 mm).
The BL diameter also falls at the upper
limits of the range of a modern African
sample, but is smaller than the mean val-
ues of an African archaeological sample.
When compared to the BL diameters of
the fossil human sample, it possesses the
third most BL compressed crown after
Dmanisi (5.9 mm) and Sangiran Bs 9706
(5.8 mm). In terms of overall crown shape
(as expressed by a CSI of 105.1, Table 14),
Specimen 4501 displays a level of relative
BL expansion that falls between levels dis-
played by Plio-Pleistocene/Lower Pleis-
tocene (e. g. Dmanisi, Sangiran 11and
Sangiran Bs 9706) and Middle Pleis-
tocene specimens (e.g. HDP1-3, Ternifine,
Rabat, Mauer and Atapuerca-SH), and
most closely resembles levels displayed by
contemporary southern African popula-
tions. Bermúdez de Castro et al. (1999)
identify relative BL expansion of
mandibular incisors as one of the evolu-
tionary trends of hominid evolution from
the Lower through the Middle Pleis-
tocene. The high levels of relative BL
expansion in mandibular incisors that
were attained by Middle Pleistocene pop-
ulations were not retained by contempo-
rary southern African populations such as
the San, who generally display quite
reduced levels. The dimensions of Speci-
men 4501 suggests that reductions in rel-
ative BL diameter may have occurred
relatively early on in this region. In root
length, this incisor also compares more
favorably with modern African homo-
logues than with fossil specimens.
Discussion of Hominid Remains
Specimen 4500 is a left parietal fragment
of probably a young adult. There is insuf-
ficient morphology preserved to confi-
dently assess the “modernness” of this
specimen. Specimen 4501 is one of only a
handful of measurable permanent central
incisors from the South African MSA.
Although mandibular central incisors
were recovered in similar contexts at
Klasies River and Die Kelders Cave 1, the
teeth from these sites are either
unerupted, or are too worn to provide use-
ful metrical data on crown dimensions.
Specimen 4501 is thus a valuable addition
to a very small sample.
DISCUSSION AND CONCLU-
SIONS
Pinnacle Point near Mossel Bay has not
been archaeologically investigated previ-
ous to our study, and our investigations to
date have demonstrated that this stretch
of coastline may have one of the richest
records of MSA archaeological remains in
South Africa. Caves, rockshelters, and
open air MSA sites are abundant and
archaeological sediments are well pre-
served. Calcretes and caliches are abun-
dant in and around the surface sites,
flowstones and dripstones are abundant
in the caves and some seal MSA deposits,
74 •PaleoAnthropology 2004
leading to the possibility of uranium-
series dating. Some of the caves are well
above the OIS5e high sea stand (Hendey
& Volman, 1986; Van Andel, 1989;
Richards et al., 1994; Van Andel &
Tzedakis, 1996)) and thus may preserve
OIS6 MSA material, a time poorly sam-
pled in Africa but clearly critical to the
MSA and the debate over the origins of
modern people. Our excavations were
productive at sites 13A and 13B, but less
so at site 9 because a large dune sealed the
deposit and we were unable to reach
archaeological deposits.
The MSA lithic assemblages do not vary
significantly between the sites and facies,
but the samples are still fairly small when
the assemblages are split into sub-assem-
blages. They display a diversity of raw
materials that include local quartzites,
non-local quartzites, silcretes of various
types, and even non-local hornfels. While
we have not yet developed a full under-
standing of the range and placement of
sources, we have found that most of these
raw materials are procurable fairly locally,
and that silcrete may not be as “exotic” as
is regularly claimed. The lithic assem-
blages include a high frequency of cortical
pieces, suggesting primary core reduction
on site. Uniformly, the samples represent
the early stages of artifact manufacture,
and “finished tools” as represented by
retouched pieces are rare, as is typical for
the MSA in southern Africa. The Mossel
Bay region is famous in Stone Age studies
because it lent its name to one of the first
formally recognized stone tool industries
in South Africa – The Mossel Bay Indus-
try. The key characteristics of the Mossel
Bay Industry are artifacts made on
quartzites, and long pointed and parallel-
sided flakes. Our excavations revealed
Figure 36. The hominid incisor (4501) showing A) buccal, B) distal, C) lingual, and D) mesial
views. The scale is 2 cm long.
MIDDLE STONE AGE SITES AT PINNACLE POINT 75
TABLE 14. MESIO-DISTAL (ESTIMATED) AND BUCCO-LINGUAL DIAMETERS (IN MM), CROWN SHAPE INDEX (CSI) AND ROOT
LENGTH OF SPECIMEN 4501 AND CENTRAL MANDIBULAR INCISORS FROM SELECTED MODERN, ARCHAEOLOGICAL AND FOSSIL
HUMANS FROM AFRICA.
MESIO-DISTAL BUCCO-LINGUAL
SPECIMEN (S) N MM S.D. N MM S.D. CROWN ROOT CSI REFERENCES
HEIGHT LENGTH
Mossel Bay 4501 5.9 6.2 ca. 7.5 ca. 12.7 105.1 This report
Atapuerca-SH 5 5.5 0.08 5 6.5 0.43 118.2 Bermudez de Castro (1986,
1988, 1993)
Border Cave 5 6.6 Stynder et al.(2001)
Dmanisi 6.2 5.9 95.2 Gabunia & Vekua (1995)
HDP1-3 6.4 7.2 ca. 8.1 15.1 112.5 Stynder et al.(2001)
KNM-ER-820 6.1 6.3 103.3 Wood (1991)
KNM-WT-15000 6.6 6.8 10.1 19.9 103 Brown & Walker (1993)
Mauer 5.5 7.1 129.1 Bermudez de Castro (1986)
Nubia, "Mesolithic", F 12 5.48 0.34 14 6.31 0.39 115.1 Calcagno (1989)
Nubia, "Mesolithic", M 8 5.8 0.35 11 6.43 0.39 110.9 Calcagno (1989)
OH16 6.5 7 10.6 107.7 Tobias (1991)
OH7 6.8 6.6 9.2 97.1 Tobias (1991)
Rabat 6 7 116.7 Bermúdez de Castro (1986)
S.A. Blacks 5.9 6 7.8 12.4 101.7 Shaw (1931)
S.A. Blacks, Female 61 5.3 0.35 5.71 0.42 107.7 Jacobson (1982)
S.A. Blacks, Male 181 5.33 0.42 5.83 0.39 109.4 Jacobson (1982)
S.A.Blacks, Male 56 5.43 0.45 6.2 0.32 114.2 Kieser et al. (1987)
San 5 5.2 6.8 12 104 Drennan (1929)
San, Female 5 5.4 0.84 5.6 0.61 103.7 van Reenen (1982)
San, Male 6 5.4 0.47 5.3 0.27 98.1 van Reenen (1982)
Sangiran 11 7.4 6.5 87.8 Wood (1991)
Sangiran Bs 9706 6 5.8 9.9 96.7 Baba et al. (2000)
Ternifine (Tighennif) 5.7 6.6 9 115.8 Bermúdez de Castro (1986)
76 •PaleoAnthropology 2004
material that is technologically and typo-
logically similar, but also has some inter-
esting differences in that the complete
blades and points are significantly smaller
than those from Cape St. Blaize (the Mos-
sel Bay type sample) and Klasies River
(the best described sample).
MSA inhabitants of caves at Pinnacle
Point collected marine shellfish primarily
from nearby rocky shores and very occa-
sionally also from sandy beaches. Their
collecting rounds targeted the mid- and
lower reaches of the intertidal zone from
where brown mussels, arikreukel snails,
limpets, chitons and black mussels were
collected most probably during low spring
tides. Other species were also gathered
(e.g., ear shells, white mussel and corru-
gated white mussel, but in much smaller
quantities). Comparison of shell densities
and species frequencies from MSA sites
with those from nearby Holocene LSA
sites might also help us estimate the dis-
tance between MSA shell bearing sites
and their contemporary shoreline.
Fossil bone is well preserved in the
deposits, and this is expected as the geol-
ogy of the area features a laterally exten-
sive, nearly continuous, calcrete (derived
from ancient calcareous and now decalci-
fied dunes) that caps the acidic quartzite.
High frequencies of cut marks and ham-
merstone percussion marks on long bone
fragments and the near absence of carni-
Mossel Bay 4501
HDP1-3
Rabat
Ternifine (Tighennif)
KNM-WT-15000
KNM-ER-820
OH7
OH16
Nubia Male
Nubia Female
San Male
San Female
San
SA Blacks Male
SA Blacks Female
SA Blacks M2
SA Blacks
Dmanisi
Sangiran Bs 9706
Mauer
Atapuerca-SH
4.9 5.4 5.9 6.4 6.9
Mesio-Distal mm
5
5.5
6
6.5
7
7.5
Bucco-Lingual mm
Figure 37. Scatterplot of the mesio-distal and bucco-lingual diameters in mm of the Mossel
Bay 4501 hominid central incisor relative to central mandibular incisors from selected mod-
ern, archaeological and fossil humans from Africa.
MIDDLE STONE AGE SITES AT PINNACLE POINT 77
vore tooth marks on bone surfaces,
strongly indicates that humans were
responsible for accumulating and deposit-
ing the bulk of the faunal remains at 13A
and 13B. Interestingly, there are no snake
bones in either 13A or 13B. This is unusual
and may have significant implications for
interpreting the presence of snakes and
other fauna in other MSA sites. Because
both sites present a strong human tapho-
nomic signature, the faunal assemblage
may be a good indicator of the species
range that hominids were and were not
exploiting. Tortoises, small mammals,
and microfauna are rare at both sites rela-
tive to other MSA sites in South Africa.
The hominin remains include two frag-
ments: a parietal and central incisor.
While the parietal is too fragmentary to
interpret, the incisor falls comfortably
between Middle Pleistocene hominins
and modern humans in metric attributes.
Given the rarity with which human fossil
material is found in MSA sites in South
Africa, this sample from such a small
excavation perhaps provides optimism for
further discoveries.
78 •PaleoAnthropology 2004
ACKNOWLEDGEMENTS
We thank the SAHRA for providing a permit (No.
80/99/04/01/51) to conduct test excavations at the selected sites.
We extend sincere thanks to the Mossel Bay community for assist-
ing during our trial excavations in July 2000 and the following
fieldwork and analyses. In particular we thank the staff of the Diaz
Museum Complex (especially Mr. Klaas Eland, Mrs. Linda
Labuscagne and Mr. John Thackray), Mossel Bay Municipality
(Mr. Dries Celliers and Mr. Dawie Zwiegelaar), Cape Nature Con-
servation (Dr. Annalize Schuttevlok, Mrs. Justine Sharples and Mr.
Johan Oelofse), Mr. Francois van der Walt for surveying assis-
tance, Mr. Ricky van Rensberg for building our staircase, as well as
the business community. We are also grateful to the local media for
popularizing our research. We thank the National Science Founda-
tion (USA) (grant # BCS-9912465 and BCS-0130713 to Marean)
and the Hyde Family Trust for funding the excavations, analysis,
and write-up. The financial assistance of the National Research
Foundation (NRF): Division for Social Sciences and Humanities
(DSSH) (South Africa) towards this research is hereby acknowl-
edged (grant # 15/1/3/17/0053 to Nilssen). Opinions expressed in
this document and conclusions arrived at, are those of the authors
and are not necessarily to be attributed to the NRF: DSSH. We
extend a very special thanks to our field and laboratory crew for
their outstanding dedication and hard work as well as specialists
for their analyses presented in this report. Tom Minichillo pro-
vided helpful comments on the lithic analysis, and Erin Lassiter
assisted with the editing and production.
MIDDLE STONE AGE SITES AT PINNACLE POINT 79
REFERENCES
Ambrose, S. H. & Lorenz, K. G. 1990. Social and
ecological models for the Middle Stone Age in
southern Africa, in Mellars,P. (Ed.), The
emergence of modern humans: an archaeo-
logical perspective, Edinburgh University
Press, Edinburgh, pp. 3–33.
Baba, H., Aziz, F., Narasaki, S., Sudijono, Kaifu,
Y., Suprijo, A., Hyodo, M., Susanto, E. E. &
Jacob, T. (2000). A new hominid incisor from
Sangiran, Central Java. Journal of Human
Evolution 38, 855–862.
Bermúdez de Castro, J. M. (1986). Dental
remains from Atapuerca (Spain). I. Metrics.
Journal of Human Evolution. 15, 265–287.
Bermúdez de Castro, J. M. (1988). Dental
remains from Atapuerca/Ibeas (Spain) II.
Morphology. Journal of Human Evolution 17,
279–304.
Bermúdez de Castro, J. M. (1993). The Ata-
puerca dental remains. New evidence (1987-
1991 excavations) and interpretations.
Journal of Human Evolution 24, 339–371.
Brown, B. & Walker, A. (1993). The dentition. In
(A. Walker & R. E. Leakey, Eds) The Narioko-
tome Homo erectus skeleton, pp. 161–192.
Berlin: Springer Verlag.
Calcagno, J. M. (1989). Mechanisms of human
dental reduction. A case study from Post-
Pleistocene Nubia. University of Kansas Pub-
lications in Anthropology 18. Lawrence: Uni-
versity of Kansas.
Deacon, J. & Lancaster, N. 1988. Late Quater-
nary Paleoenvironments of Southern Africa.
Clarendon Press, Oxford.
Drennan, M. R. (1929). The dentition of a Bush-
man tribe. Annals of the South African
Museum 24, 61–87.
Gabunia, L. & Vekua, A. (1995). A Plio-Pleis-
tocene hominid from Dmanisi, East Georgia,
Caucasus. Nature 373, 509–512.
Goodwin, A. J. H. 1930. The chronology of the
Mossel Bay Industry. South African Journal
of Science 27, 562–572.
Goodwin, A. J. H. & Malan, B. D. 1935. Archae-
ology of the Cape St. Blaize Cave and raised
beach, Mossel Bay. Annals of the South
African Museum 24, 111–140.
Goodwin, A. J. H. & Van Riet Lowe, C. 1929. The
Stone Age cultures of South Africa. Annals of
the South African Museum 27, 1–289.
Gray, H., Pickering, T. (Ed) and Howden, R.
(Ed). (1977). The classic collector’s edition:
Gray’s Anatomy. New York: Bounty Books.
Hendey, Q. B. & Volman, T. P. 1986. Last Inter-
glacial Sea Levels and Coastal Caves in the
Cape Province, South Africa. Quaternary
Research 2, 189–198.
Henshilwood, C. S., D'Errico, F., Yates, R.,
Jacobs, Z., Tribolo, C., Duller, G. A. T.,
Mercier, N., Sealy, J. C., Valladas, H., Watts,
I., & Wintle, A. G. 2002. Emergence of mod-
ern human behavior: Middle Stone Age
engravings from South Africa. Science 295,
1278–1280.
Henshilwood, C. S., Sealy, J. C., Yates, R. J.,
Cruz-Uribe, K., Goldberg, P., Grine, F. E.,
Klein, R. G., Poggenpoel, C., van Niekerk, K.,
& Watts, I. 2001. Blombos Cave, southern
Cape, South Africa: Preliminary report on the
1992–1999 excavations of the Middle Stone
Age levels. Journal of Archaeological Science
28, 421–448.
Jacobson, A. (1982). The dentition of the South
African Negro. Anniston: Higginbotham.
Jerardino, A. 1997. Changes in shellfish species
composition and mean shell size from a late-
Holocene record of the west coast of southern
Africa. Journal of Archaeological Science 24,
1031–1044.
Jerardino, A. & Yates, R. 1996. Preliminary
results from excavations at Steenbokfontein
Cave: Implications for past and future
research. South African Archaeological Bul-
letin 51, 7–16.
Kaplan, J. Archaeological Study: Proposed Pin-
nacle Point Development. 1997. Riebeek
80 •PaleoAnthropology 2004
West, South Africa, Agency for Cultural
Resource Management, unpublished report.
Keller, C. M. 1968. Mossel Bay: A redescription.
South African Archaeological Bulletin 33,
131–140.
Keller, C. M. 1973. Montague Cave in Prehistory.
University of California Anthropological
Records 28, 1–150.
Kieser, J. A., Groeneveld, H. T. & Cameron, N.
(1987). Evidence for a secular trend in the
Negro dentition. Annals of Human Biology
14, 517–532.
Klein, R. G. 1995. Anatomy, behavior, and mod-
ern human origins. Journal of World Prehis-
tory 9, 167–198.
Klein, R. G. 2000. Archeology and the evolution
of human behavior. Evolutionary Anthropol-
ogy 9, 17–36.
Klein, R. G. & Cruz-Uribe, K. 2000. Middle and
Later Stone Age large mammal and tortoise
remains from Die Kelders Cave 1, Western
Cape Province, South Africa. Journal of
Human Evolution 38, 169–196.
Leith, G. 1898. On the caves, shell-mounds and
stone implements of South Africa. Journal of
the Royal Anthropological Institute 28,
258–274.
Marean, C. W., Abe, Y., Nilssen, P. J., & Stone, E.
C. 2001. Estimating the minimum number of
skeletal elements (MNE) in zooarchaeology: a
review and a new image-analysis GIS
approach. American Antiquity 66, 333–348.
Marean, C. W. & Assefa, Z. 1999. Zooarchaeolog-
ical evidence for the faunal exploitation
behavior of Neandertals and Early Modern
Humans. Evolutionary Anthropology 8,
22–37.
Marean, C. W., Goldberg, P., Avery, G., Grine, F.
E., & Klein, R. G. 2000. Middle Stone Age
stratigraphy and excavations at Die Kelders
Cave 1 (Western Cape Province, South Africa):
the 1992, 1993, and 1995 field seasons. Jour-
nal of Human Evolution 38, 7–42.
McBrearty, S. & Brooks, A. S. 2000. The revolu-
tion that wasn't: a new interpretation of the
origin of modern human behavior. Journal of
Human Evolution 39, 453–563.
Richards, D. A., Smart, P. L., & Edwards, R. L.
1994. Maximum sea levels for the last glacial
period from U-series ages of submerged
speleothems. Nature 367, 357–360.
Sampson, C. G. 1971. The Stone Age Archaeology
of Southern Africa. Academic Press, New
York.
Shaw, J. C. M. (1931). The Teeth, the Bony Palate
and the Mandible in Bantu races of South
Africa. London: John Bull, Sons & Daniels-
son.
Singer, R. & Wymer, J. 1982. The Middle Stone
Age at Klasies River Mouth. The University of
Chicago Press, Chicago.
Stynder, D. D., Moggi-Cecchi, J., Berger, L. R.
and Parkington, J. E. (2001). Human
mandibular incisors from the late Middle
Pleistocene locality of Hoedjiespunt 1, South
Africa. Journal of Human Evolution 41,
369–383.
Thackeray, A. I. 1989. Changing fashions in the
Middle Stone Age: the stone artefact sequence
from Klasies River main site, South Africa.
The African Archaeological Review 7, 33–57.
Thackeray, A. I. 1992. The Middle Stone Age
south of the Limpopo river. Journal of World
Prehistory 6, 385–440.
Thackeray, A. I. 2000. Middle Stone Age arti-
facts from the 1993 and 1995 excavations of
Die Kelders Cave 1, South Africa. Journal of
Human Evolution 38, 147–168.
Tobias, P. V. (1991). Olduvai Gorge, Volume IV.
The Skulls, Endocasts and Teeth of Homo
habilis. Cambridge: Cambridge University
Press.
van Andel, T. H. 1989. Late Pleistocene sea levels
and human exploitation of the shore and shelf
of southern South Africa. Journal of Field
Archaeology 16, 133–155.
van Andel, T. H. & Tzedakis, P. C. 1996. Palae-
MIDDLE STONE AGE SITES AT PINNACLE POINT 81
olithic landscapes of Europe and environs,
150,000-25,000 years ago: an overview. Qua-
ternary Science Reviews 15, 481–500.
van Hoepen, E. C. N. 1926. Ouderdom van die
Suid-Afrikaanse Klipwerktuie. South African
Journal of Science 23, 797.
van Reenen, J. F. (1982). The effects of attrition
on root dimensions of San (Bushmen). In (B.
Kurtén, Ed.) Teeth: Form, Function and Evo-
lution, pp. 182–203. New York: Columbia
University Press.
Volman,T.P. (1981). The Middle Stone Age in the
Southern Cape. (Ph.D. University of Chicago).
Wolf, P. R. & Ghilani, C. D. 2002. Elementary
Surveying: An Introduction to Geomatics.
Prentice-Hall, Upper Saddle River, N.J.
Wood, B. A. (1991). Koobi Fora Research Project
IV: Hominid Cranial Remains from Koobi
Fora. Oxford: Clarendon Press.
Wurz, S. 2002. Variability in the Middle Stone
Age lithic sequence, 115,000–60,000 years
ago at Klasies River, South Africa. Journal of
Archaeological Science 29, 1001–1016.
Wurz, S., le Roux, N. J., Gardner, S., & Deacon,
H. J. 2003. Discriminating between the end
products of the earlier Middle Stone Age sub-
stages at Klasies River using biplot methodol-
ogy. Journal of Archaeological Science 30,
1107–1126.
82 •PaleoAnthropology 2004
APPENDIX 1: PROVENIENCE AND DESCRIPTIVE DATA FOR THE ILLUSTRATED LITHICS.
MAP FIGURE SITE SQUARE QUAD STRATUNIT FACIES DESCRIPTION PORTION RAW MATERIAL
124 23 13B N94E109 SE NEE Lightly Consolidated Convergent Blade Whole Flake Quartzite
MSA Facies
145 23 13B N94E109 SE NEE Lightly Consolidated Quadrilateral Blade Whole Flake Quartzite
MSA Facies
148 23 13B N94E109 SE NEE Lightly Consolidated Flake Whole Flake Quartzite
MSA Facies
149 23 13B N94E109 SE NEE Lightly Consolidated Flake Whole Flake Quartzite
MSA Facies
234 24 13B N94E109 SE NEG Lightly Consolidated Radial Disc Core Complete Quartzite
MSA Facies
326 26 13B N94E109 SE NEN Lightly Consolidated Flake Fragment Distal Quartz
MSA Facies with Retouch Fragment
1294 23 13B N91E108 SW EC MSA Roof Spall Indeterminate Proximal Quartzite
Facies Fragment
1303 24 13B N91E108 SW EC MSA Roof Spall Multiple Platform Whole Flake Quartzite
Facies Core on Flake
1524 25 13B N91E108 SW EH MSA Roof Spall Simple Blade Whole Flake Quartzite
Facies
1643 25 13B N91E108 NE EH3 MSA Roof Spall Facetted Blade Whole Flake Quartzite
Facies
1733 25 13B None None Surface 13B Disturbance Facetted Blade Whole Flake Quartzite
1772 26 13B None None Surface 13B Disturbance Disc Core Complete Silcrete
MIDDLE STONE AGE SITES AT PINNACLE POINT 83
APPENDIX 1: PROVENIENCE AND DESCRIPTIVE DATA FOR THE ILLUSTRATED LITHICS. (CONTINUED)
MAP FIGURE SITE SQUARE QUAD STRATUNIT FACIES DESCRIPTION PORTION RAW MATERIAL
2142 24 13B N97E96 NE WB Brown Sand MSA Flake Whole Flake Cave TMS
Facies
2191 25 13B N97E96 SE WB Brown Sand MSA Facetted Flake Whole Flake Quartzite
Facies
2251 26 13B N97E96 SE WB Brown Sand MSA Single Platform Complete Quartzite
Facies Core
2501 23 13B N97E96 NE WB Brown Sand MSA Point Whole Flake Quartzite
Facies
2529 23 13B N97E96 NE WB Brown Sand MSA Quadrilateral Blade Whole Flake Quartzite
Facies
3192 26 13A N69E100 NW F Brown Humic Sands Whole Flake Whole Flake Silcrete
MSA Facies
3275 25 13A N69E100 NE B Brown Humic Sands Simple Blade Whole Flake Quartzite
MSA Facies
3295 26 13A N69E100 NE B Brown Humic Sands Facetted Point Whole Flake Quartzite
MSA Facies
3396 23 13A N69E100 NE H Brown Humic Sands Point Whole Flake Quartzite
MSA Facies
... The Pinnacle Point complex, beginning with Pinnacle Point 13B (PP13B), has been under investigation by C.W. Marean and colleagues for the last 20 years (e.g. Marean et al., 2004). As Marean (2010) has pointed out, extracting information from these sites has benefitted from great advances in the development of field and laboratory techniques. ...
Article
Improvements in excavation methods, dating, analytical techniques and statistical applications have all led to a substantial increase in recoverable environmental evidence from micromammals. Because these animals are so small the information they provide is on a smaller geographical scale than that afforded by most other lines of evidence. However, with increasing amounts of data and greater interpretational precision in all spheres, the chances are improving of being able to mesh information from different scales. Blombos, Pinnacle Point and Klasies River on the southern coast of South Africa have clearly demonstrated that micromammalian data can contribute to multidisciplinary interpretations of past conditions, in this case during MIS 5 and 6. Little attention has been paid to the generally small samples from Iron Age sites but the presence of the House rat Rattus rattus may provide important information about human movements and may also contribute to our understanding of the Anthropocene once this has been formally defined. Micromammals have not yet been used as chronostratigraphic indicators in southern Africa but it may be possible to develop biochronologies using them and to incorporate this material into African Land Mammal Ages.
... We used an excavation and plotting procedure modified from that developed at the Pinnacle Point complex of archaeological sites 56 . Permanent control points were established on the bedrock and large roof-fall, and linked to the WGS84 UTM zone 34S grid. ...
Conference Paper
Homo sapiens exhibit extreme behavioural plasticity, mediated by culture and technology, that permits us to adapt rapidly to new environments and situations. Understanding the role that past climate change played in selecting for Homo sapiens’ adaptability is a key question in human evolution research. The arid and semi-arid Kalahari Basin in southern Africa is an ideal region for addressing this question because fossil, genetic, and archaeological evidence supports an early origin for Homo sapiens in southern Africa. The growing archaeological record of the Kalahari Basin reveals that significant behavioural innovations accumulated in the region over the course of the Middle and Late Pleistocene, including ochre use, hafted hunting weapons, fishing, and figurative paintings. Here, we report the results of interdisciplinary investigations at two locales in the southern Kalahari; Ga-Mohana Hill and Witberg 1. The archaeological and palaeoenvironmental record (based on U-Th dating of tufas) at Ga-Mohana Hill reveals that site occupation correlated with a previous period of increased effective precipitation ~110-100 ka, and preliminary results suggest a more complicated relationship between occupation and precipitation after that time. At Witberg 1, Middle Stone Age archaeology is associated with the shoreline of a previously unidentified palaeolake. Current investigations are focused on dating the Witberg deposits, analysing the lithic technology, and generating palaeoenvironmental archives using phytoliths and diatoms. Collectively, this research provides a rare opportunity to evaluate Middle Stone Age occupation across a changing landscape from both stratified rockshelters and sealed open-air sites, to explore the complex interactions between past climate change and early human behaviors, and to better understand the origins of Homo sapiens extreme adaptability.
... We used an excavation and plotting procedure modified from that developed at the Pinnacle Point complex of archaeological sites 56 . Permanent control points were established on the bedrock and large roof-fall, and linked to the WGS84 UTM zone 34S grid. ...