Charcoal from pre-Holocene Stratum 5,
Wonderwerk Cave, South Africa
Marion K. Bamford
Evolutionary Studies Institute and School of Geosciences,
University of the Witwatersrand, Johannesburg, South Africa
ABSTRACT: Charcoal from the archaeological layer, Stratum 5 of Wonderwerk Cave in the
Northern Cape Province of South Africa is described and identified. The stratum has been
dated by several methods and is pre-Holocene in age, up to about 15 ka. About three quarters
of the 134 pieces are identifiable and belong to eight types (six species and two to generic level)
of woody plants: Ozoroa paniculosa and Searsia (Rhus) lancea (Anacardiaceae), Ehretia sp.
(Boraginaceae), Commiphora sp. (Burseraceae), Dombeya rotundifolia (Pentapetalaceae), Olinia
ventosa (Oliniaceae), Berchemia discolor (Rhamnaceae) and Halleria lucida (Scrophulariaceae).
Of these species only Searsia lancea occurs in the area today. The majority of the woods found
can tolerate dry conditions and a wide range of temperatures but the presence of Berchemia
discolor and Halleria lucida indicates that conditions between c. 14,985 and 13,952 years cal. BP
may have been slightly wetter than today.
Wonderwerk Cave in the Northern Cape Province of South Africa contains a record of
approximately the last 2million years of occupation by hominins and humans although
to date no skeletal material has been found. The site was first studied by Malan and
Cooke (1941) and later extensively excavated by Peter Beaumont and colleagues from
the 1970s to early 1990s (Beaumont, 1990, 2004; Camp, 1948; Thackeray, 1983; Avery,
1981, 2007; van Zinderen-Bakker, 1982) producing large amounts of artifacts, bone
and botanical remains from seven excavations within the cave. In the mid 2000s a new
team, led by Michael Chazan and Liora Kolska Horwitz, has used high resolution
mapping and dating techniques in the cave. Beginning with Excavation 1 which is 30m
from the cave entrance, their multinational and multidisciplinary team is in the process
of re-analysing old material and collecting new material (Figure1), (Chazan et al., 2008,
2012; Rüther et al., 2009; Matmon et al., 2011). Some of the results produced by this
team are being collected in a special issue of African Archaeological Review (AAR for
2015). A description of the relatively meagre macrobotanical remains from Excavation 1
Strata 12 to 5includes only the older remains (Bamford, AAR in 2015) where the
rare charcoal pieces are unidentifiable. These older remains comprise grass and sedge
culms and seeds of a palm and a legume. Charcoal, however, is abundant and well
preserved in the younger strata, 5 to 2, from pre-Holocene to Present. This paper is
the first in a series of detailed descriptions and palaeoecological interpretations of the
charcoal remains from the various Holocene strata. In due course the charcoals from
the other strata will be compared and both the anthropogenic selection of woods and
palaeoecology will be assessed.
Runge33-Book.indb 153Runge33-Book.indb 153 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
Palaeoecology of Africa, Volume 33 (2015), pp. 153-174,
154 Marion K. Bamford
9.1.1 Research history
Wonderwerk Cave is situated in the Northern Cape Province of South Africa at the
edge of the Ghaap plateau in the Kuruman Hills (27°50′46″ S, 23°33′19″ E; about
1680meters above sea level), (Figure1). With an average annual rainfall of 420 mm
(80% occurring during the summer) and relatively flat topography the climate and veg-
etation are dry to semi-arid (Schulze, 1984). The vegetation biome is savanna but the
cave is very close to the grassland biome. According to Mucina and Rutherford (2006)
the vegetation type is Kuruman Mountain Bushveld (SVK10) with a wide variety of
grasses and some small trees and shrubs on shallow, sandy soils. Trees include Searsia
lancea, Searsia pyroides, Searsia tridactyla, Searsia ciliata, Diospyros austro-africana,
Euclea crispa, Olea europaea, Tarchonanthus camphoratus, and shrubs Anthospermum
rigidum, Helichrysum zeyheri, Wahlenbergia nodosa. Near the cave entrance are Grewia
flava and Boscia albitrunca, and the common grasses are Themeda triandra, Cymbopo-
gon plurinoides and Aristida spp.
The cave is long, narrow and almost horizontal with one entrance that faces
north-north-west, overlooking the Ghaap Plateau. It is a solution cavity in stratified
dolomitic limestones but is overlain by early Proterozoic to late Archaean banded
ironstones (2–3Ga; Kent, 1980). The cave is about 140m long and 11–17m wide with
Figure 1. Map of the locality and other sites mentioned in the text. WW=Wonderwerk Cave,
WK=Wonderkrater, RC=Rose Cottage, S=Sifiso rock shelter, EB=Elands Bay Cave.
Runge33-Book.indb 154Runge33-Book.indb 154 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 155
a domed roof. Over time the entrance has receded but light penetrating the cave is par-
tially blocked by a large stalagmite that is about 20m in from the entrance. Excavation 1
is immediately behind the stalagmite.
One piece of charcoal was recovered from the lowermost layer, Stratum 12,
approximately 2 Ma but is too poorly preserved for identification (Bamford, AAR
in 2015). The earliest evidence for the controlled use of fire based on four lines of
evidence, including burned grasses and sedges, comes from Wonderwerk Cave at
1 Ma (Berna et al., 2012). The subsequent layers have some calcified plant material
preserved but none of it has been burned (Bamford, AAR in 2015). The inconsistent
macroplant record may be due to gaps in occupation of the site, different uses of the
site and/or taphonomic biases. There is a large time gap prior to Stratum 5 (Chazan
etal., 2008) and then a rich record of charcoal remains through the Holocene (Table1).
The Holocene strata 1–5in Excavation 1 were isotopically dated by John Vogel using
charcoal remains (Thackeray, 1983; Vogel et al., 1986). Later equid teeth were dated
from the Holocene layers and produced slightly different ages (Thackeray and Lee
Thorp, 1992) and these ages are given in Table1 together with more recent results
based on ostrich egg shell (Lee Thorp and Ecker, AAR in 2015). Ecker and Lee Thorp
are in the process of dating more charcoal samples from the Holocene layers but the
results have not yet been published.
Table1. Charcoal-rich strata from Wonderwerk Cave with ages
and estimated abundances of charcoal pieces.
age years BP)
of pieces of
1 Last 100 years Modern
Last 100 years
five boxes of
12,500 14,989–13,952 Late Stone
Runge33-Book.indb 155Runge33-Book.indb 155 8/13/2015 7:42:58 AM8/13/2015 7:42:58 AM
156 Marion K. Bamford
The Holocene strata (Strata 1–5) of Excavation 1 were excavated by Peter Beaumont
in 1978 and Anne and Francis Thackeray in 1979 (Thackeray, 1983). Cave sedi-
ments were dry sieved and charcoal fragments were separated from other finds in the
McGregor Museum (Thackeray, 1983:40) but there are no further details recorded
on the collection methods. Since the charcoal samples have been packaged according
to archaeological stratum, square number and depth within the square, and noted in
centimetres or as spits, that notation will be used in this work.
Since the sample size was relatively small all charcoal fragments were studied, ini-
tially under a binocular microscope (Olympus SZX16, magnification 7–112x) and ori-
ented for hand fracturing to reveal fresh surfaces of the three planes required for study
(transverse, radial longitudinal and tangential longitudinal). Under higher magnification
(200–500x; Zeiss Axiolab-A1 bright phase/dark phase reflected light microscope) the
wood anatomy was studied further and photographed with a digital camera (Olympus
DP72 with Stream Essentials® software). Measurements were made of the cell dimen-
sions but precise and averaged measurements of the cells are less useful for identification
and comparison with modern taxa because woods shrink between 8–20% depending on
the wood type, moisture content and temperature of the fire. None theless when charcoal
is compared with charcoal these values are useful. Since diagnostic features such as ray
cell type, inter-vessel and vessel-ray pits were visible under the light microscope it was not
necessary to use scanning electron microscopy. Charcoal was identified using the mod-
ern comparative charcoal collections housed in the Archaeology Department and the
Bernard Price Institute (now Evolutionary Studies Institute), both at the University of
the Witwatersrand. The computer database for modern woods, InsideWood, was also
used. Terminology follows that of the IAWA Committee (1989). The fractured pieces
were placed in small individual ziplock bags so they can be retrieved and re-examined.
Each fragment has been given a number: locality+(depth)+letter, eg. WW-5 M29 (0–5cm)
I. On completion of the project the material will be returned to the McGregor Museum,
Kimberley (Wonderwerk Cave 6508).
Stratum 5 has over 134 charcoal fragments from many of the grid squares. Most pieces
are less than 7×7×7mm but a few are up to roughly 20mm cubes but often the wood
is distorted. Approximately one quarter of the pieces are twisted, weathered or too
small to fracture and reveal the three planes needed for identification. These specimens
have been recorded as “not identifiable.” The abundances and identifications of the
charcoal are given in Table2 and the descriptions of the charcoal types are below.
Descriptions of charcoal types from Stratum 5
Described sample: WW-5 M29 (0–5cm) I
Illustration: Plate 1, Figures1–4
Total number of pieces: 40
DESCRIPTION: The wood is diffuse-porous and no growth rings are visible (Plate 1,
Figures1, 2). Vessels are arranged in short radial multiples of 1–3, sometimes 4 vessels,
Runge33-Book.indb 156Runge33-Book.indb 156 8/13/2015 7:42:58 AM8/13/2015 7:42:58 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 157
and in clusters. Vessels have simple, horizontal to oblique perforation plates and the
average vessel diameter is about 50μm, vessel elements are 200–500μm long, and there
are more than 40 vessels per square mm. Paratracheal parenchyma is scanty to vasi-
centric (Plate 1, Figures2, 3). Rays are 1–3seriate, heterocellular with mixed square
cells and upright cells. Amorphous dark contents are present in many of the square
cells, one per cell and the cells are not enlarged; they do not appear to be crystalline.
Fibres are septate. Inter-vessel pits are alternate and minute (2–4μm); vessel-ray pits
are elongated to scalariform (Plate 1, Fig 4). Radial canals were not seen.
IDENTIFICATION AND COMPARISON: Woods of the Anacardiaceae in south-
ern Africa typically have vessels arranged in short radial multiples and not solitary;
rays are commonly 1–2seriate with square and/or upright cells, seldom procumbent;
crystals are common in the ray cells; septate fibres; parenchyma is not abundant and
can be banded and/or paratracheal. The piece of charcoal described here has these fea-
tures and so was compared with the modern reference collection and found to be very
similar to Ozoroa paniculosa (University of the Witwatersrand reference material: Lucy
Allott collection sample number A4, and Zimbabwe collection sample number Z97).
Table2. Complete list of identified charcoal pieces from Stratum 5, Excavation 1, Wonderwerk Cave
and the number of pieces from each grid square. Note that every piece was studied and is listed here.
Identity of charcoal pieces
Searsia (Rhus) lancea
Total number of pieces
5 K 27 0–5 3 3 2 8
M 29 0–5 12 5 8 1 2 28
M 29 5–15 5 5
N 22 0–5 1 1
N 27 0–5 2 1 3
O 26 0–5 1 2 2 5
O 28 5–10 3 4 1 8
Q 29 0–5 8 4 12
Q 29 Veg 9 1 11 21
Q 29 5–10 1 2 3 6
Q 29 10–15 1 1 2
Q 31 0–5 3 9 4 16
Q 33 0–5 1 1
R 26 0–5 14 14
R 32 0–5 2 1 1 4
5A O 24 1 1 2
Totals 40 8 2 3 7 10 29 2 35 134
Runge33-Book.indb 157Runge33-Book.indb 157 8/13/2015 7:42:58 AM8/13/2015 7:42:58 AM
158 Marion K. Bamford
Plate 1. Photomicrographs of charcoal from Stratum 5, Excavation 1, Wonderwerk Cave,
taken under bright field at 200x and 500x magnification. Cells are labeled on the photographs
of Plates 1–8 for guidance: V=vessel; R=ray, P=parenchyma, X=prismatic crystal in ray cell,
F=fibre. TS=transverse section; RLS=radial longitudinal section; TLS=tangential longitudinal
section. 1–4 – Ozoroa paniculosa (Anacardiaceae). 1, 2 – TS, vessels in short radial multiples of 1–3;
longitudinal rays with crystals visible and vasicentric parenchyma (few cells surrounding the vessels;
ground tissue comprises fibres). Scale bar 1=100 μm; 2=40 μm. 3 – RLS, square to upright ray cells
containing crystals (angular dark contents). Scale bar=40 μm. 4 – TLS, 1–2seriate rays and vessels
with simple and oblique perforation plates and small, alternate inter-vessel pits at the bottom
of the photograph and more elongated vessel-ray pits above the letter V. Scale bar=40 μm.
Runge33-Book.indb 158Runge33-Book.indb 158 8/13/2015 7:42:58 AM8/13/2015 7:42:58 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 159
There are fourteen species of Ozoroa in southern Africa (Coates Palgrave, 2002)
and most of them can be excluded on a geographical basis as they each have a very
restricted distribution in either Namibia or the far northwestern part of South Africa
or in the eastern and northern parts. Ozoroa insignis and Ozoroa paniculosa have a
much wider distribution and are therefore considered further. The wood or charcoal
of these two species can be differentiated by the presence of crystals in the ray cells of
O. insignis and tyloses in the vessels. The archaeological material has neither of these
features so is identified as O. paniculosa.
Searsia (Rhus) lancea
Described sample: WW-5 R32A
Illustration: Plate 2, Figures1–4
Total number of pieces: 8
DESCRIPTION: Growth rings are not visible, wood is diffuse-porous and the simple
perforation plates are horizontal or oblique (Plate 2, Figures1, 2). Vessels are arranged
in short radial multiples of 1–3cells with occasional clusters. The average vessel diam-
eter is about 80μm, average length is around 100μm with about 40 vessels per square
mm. Parenchyma is scanty paratracheal. Rays are 1–2seriate, 200–400μm tall and
heterocellular with mixed bands of square, upright and procumbent cells (Plate 2,
Figure3). Radial canals are rare (Plate 2, Figure4). Prismatic crystals occur in the
square ray cells. Fibres have medium-thick walls and are septate. Inter-vessel pits are
alternate, minute to very small (4μm diameter; Plate 2, Figure4) and vessel-ray pits
are elongated to scalariform.
IDENTIFICATION AND COMPARISON: This charcoal also has the typical fea-
tures of woods of the Anacardiaceae. This wood is very similar to that of Ozoroa
paniculosa in transverse section but the rays are more heterocellular with procumbent
cells included and prismatic crystals in the ray cells. Radial canals are present and there
is no diffuse parenchyma. This wood is identical to Rhus lancea charcoal reference
There are 47species of Seasia (Rhus) in southern Africa (Coates Palgrave, 2002;
van Wyk and van Wyk, 2013), most of them with restricted distributions that are far
distant from Wonderwerk. The 11species that today either occur in the area or are
within a few hundred kilometre radius have been compared with these samples. Only
two of these have crystals in the ray cells, Searsia gueinzii and Searsia lancea. The
former species can be excluded as it has diffuse parenchyma whereas S. lancea has
vasicentric parenchyma, like the archaeological specimens. Additionally, S. gueinzii
has smaller vessels than S. lancea.
Described sample: WW-5 Q29 (5–10cm) B
Illustration: Plate 3, Figures1–4
Total number of pieces: 3
DESCRIPTION: Wood is diffuse-porous and perforation plates are simple and hori-
zontal. Vessels are arranged in short radial multiples of 1–3 cells and some clusters.
Average vessel diameter is about 80μm, length about 200μm and there are about
40 vessels per square mm (Plate 3, Figures1, 2). Parenchyma is aliform to confluent
Runge33-Book.indb 159Runge33-Book.indb 159 8/13/2015 7:43:01 AM8/13/2015 7:43:01 AM
160 Marion K. Bamford
Plate 2. Searsia (Rhus) lancea (Anacardiaceae). 1 – TS, vessels in short radial multiples.
Scale bar=100μm. 2 – TS, vessels in short radial multiples. Scale bar=40 μm. 3 – RLS,
mostly procumbent ray cells shown with a few crystals present. Scale bar=40μm. 4 – TLS,
1–2seriate rays and central ray has a radial canal (large dark rimmed circle).
Inter-vessel pits are small and alternate (below letter V). Scale bar=40μm.
Runge33-Book.indb 160Runge33-Book.indb 160 8/13/2015 7:43:01 AM8/13/2015 7:43:01 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 161
Plate 3. Ehretia sp. (Boraginaceae). 1, 2 – TS, vessels in short radial multiples and clusters.
Rays are wide and seen as longitudinal lines of brick-shaped cells. Note bands of thick-walled fibres
alternating with thin-walled parenchyma cells. Scale bar 1=100μm; 2=40μm. 3 – RLS,
mostly procumbent ray cells with 1–2 rows of marginal upright cells (bottom of photograph).
Scale bar=100μm. 4 – TLS, rays short and 5–6cells wide and surrounding thick walled fibres.
Runge33-Book.indb 161Runge33-Book.indb 161 8/13/2015 7:43:05 AM8/13/2015 7:43:05 AM
162 Marion K. Bamford
to almost banded (Plate 3, Figure2). There are alternating bands of axial parenchyma
and thick-walled fibres. Parenchyma strands have 4–8cells. Rays are clearly visible
in transverse section and are 4–8seriate, short, wide and abundant. They are weakly
homocellular comprising procumbent body cells and one, rarely two, rows of upright
to square marginal cells (Plate 3, Figures3, 4). Inter-vessel pits are alternate and small;
vessel-ray pits are the same as the inter-vessel pits (Plate 3, Figure3).
IDENTIFICATION AND COMPARISON: Woods of the Boraginaceae that occur
in Africa typically have a few small vessel members, weakly heterocellular rays around
6cells wide and banded parenchyma. This piece of charcoal is very similar to Ehretia
spp (InsideWood database) but the four species occurring in South Africa have not
been studied. Without more comparative material it is not possible to distinguish the
species, Ehretia alba, Ehretia amoena, Ehretia rigida and Ehretia oppositifolia so the
charcoal is identified only as Ehretia sp. Geographically, the charcoal is likely to be
Described sample: WW-5a O24A
Illustration: Plate 4, Figures1–4
Total number of pieces: 2
DESCRIPTION: Wood is diffuse-porous and the perforation plates are simple and
oblique. Vessels are arranged in pairs or are solitary, about 50μm in diameter and
about 20 per square mm (Plate 4, Figures1, 2). Parenchyma is rare to absent. Rays
are 1–3cells wide and 10–15cells high, heterocellular with mixed procumbent, square
and upright cells. There are prismatic crystals in the ray cells and sometimes 2–4 crys-
tals per upright cell (Plate 4, Figure3). Radial canals are rare. Fibres are medium to
thick-walled and septa are visible. Inter-vessel pitting is alternate, crowded and small
(4–8μm). Vessel-ray pits are elongated to scalariform (Plate 4, Figure4).
IDENTIFICATION AND COMPARISON: There are about 35species of Commi-
phora in southern Africa (van Wyk and van Wyk, 2013) but we do not have charcoal
reference material of all of them. The charcoal from Stratum 5 has the typical features
of all the species with relatively small and few vessel members, rays 1–4seriate and
heterocellular; and parenchyma rare to absent. The important identifying features are
the presence of radial canals and more than one crystal per upright cell. It is tentatively
identified as Commiphora sp.
Described sample: WW-5 O28 (5–10cm) E
Illustration: Plate 5, Figures1–4
Total number of pieces: 7
DESCRIPTION: Growth rings are not visible, wood is diffuse-porous and the perfora-
tion plates are simple and horizontal. Vessels are arranged in short radial multiples of
1–3cells, average diameter 80–100μm with about 30 vessels per square mm (Plate 5,
Figure 1). Vessel elements are comparatively short, about 100–150μm. Parenchyma
is scanty. Rays are 1–3seriate, rarely 4-seriate, 10–20cells or 200–500 μm high with
mixed upright and square cells (Plate 5, Figures2–4). Narrow rays seem to be storied.
Runge33-Book.indb 162Runge33-Book.indb 162 8/13/2015 7:43:08 AM8/13/2015 7:43:08 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 163
Plate 4. Commiphora sp. (Burseraceae). 1, 2 – TS, vessels in 1–2s and no parenchyma seen.
Scale bar 1=100 μm; 2=40 μm. 3 – RLS, ray cells with crystals. Note 3–4 crystals per upright cell
(bottom centre). Ray-vessel pitting (lower right, above the scale bar) elongated to scalariform.
Scale bar=40μm. 3 – TLS Rays 1–2seriate. Inte-r-vessel pitting is small
and alternate (centre of vessel). Scale bar=40μm.
Runge33-Book.indb 163Runge33-Book.indb 163 8/13/2015 7:43:08 AM8/13/2015 7:43:08 AM
164 Marion K. Bamford
Plate 5. Dombeya rotundifolia (Pentapetaceae). 1 – TS, vessels in short radial multiples
and fibres are thick-walled. Scale bar=100μm. 2 – RLS, upright and square ray cells.
Scale bar=40μm. 3 – TLS, rays of two sizes, 1 and 3–4seriate. Fibres are thick-walled.
Scale bar=40μm. 4 – RLS, upright and square ray cells, inter-vessel pits alternate,
and bordered pits in fibre walls. Scale bar=40μm.
Runge33-Book.indb 164Runge33-Book.indb 164 8/13/2015 7:43:11 AM8/13/2015 7:43:11 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 165
Fibres are thick-walled and have bordered pits. Inter-vessel pits are round, alternate and
4–6μm in diameter. Vessel-ray pits are very similar to inter-vessel pits (Plate 5, Figure3).
IDENTIFICATION AND COMPARISON: This wood, with vessels arranged in
short radial multiples, with scanty paratracheal parenchyma and heterocellular rays
1–3seriate, is typical of many families. Having rays of two sizes and with the unseriate
rays being more or less storied is a useful diagnostic feature for the genus. The charcoal
is the same as Dombeya rotundifolia in the modern reference collection (A162). It is
very similar to Dombeya tiliacea (A163) but this species has diffuse parenchyma, not
scanty parenchyma, and the vessels are fewer in number per square mm. According
to the InsideWood database for Dombeya spp. parenchyma ranges from diffuse, to
diffuse-in-aggregate to scanty paratracheal. The rays are composed of mixed square
and upright cells in the reference material, and not with procumbent body cells and
1–4 rows of marginal upright or square cells as described in the more general descrip-
tion in the InsideWood database.
There are nine species of Dombeya in southern Africa, mostly with very restricted
distributions (Coates Palgrave, 2002; van Wyk and van Wyk, 2013) and very far from
Wonderwerk. Only Dombeya rotundifolia var. rotundifolia is widespread. Based on the
limited comparative material and the geographical distribution the charcoal is identi-
fied as D. rotundifolia.
Described sample: WW-5 M29 (0–5cm) F
Illustration: Plate 6, Figures1–4
Total number of pieces: 10
DESCRIPTION: Wood is diffuse-porous and vessels have horizontal to oblique, sim-
ple perforation plates. Vessels are 40–50μm wide, about 200μm long with more than
40 vessels per square mm. They are arranged in short radial multiples up to 4 vessels
and clusters are common (Plate 6, Figure1). Inter-vessel pitting is small and alternate.
Parenchyma is scanty paratracheal to vasicentric and with irregular, narrow bands.
Rays are numerous, 1–3seriate, about 10–20 cells high, heterocellular with mostly
upright cells but also square and procumbent cells (Plate 6, Figures2, 3). The multi-
seriate sections are as wide as the uniseriate sections and the large cells, as seen in
tangential longitudinal section, are perforated (Plate 6, Figure4). There are septate
IDENTIFICATION AND COMPARISON: The suite of features, vessels arranged in
short radial multiples that have a tendency towards an oblique pattern, heterocellular
rays that are 1–2seriate, with the uniseriate portions as wide as the multiseriate por-
tions, and parenchyma that is scanty paratracheal with narrow bands are typical of
the Oliniaceae. This charcoal is the same as Olinia ventosa (syn. O. cymosa) in the
modern comparative collection (A111). Photographs and descriptions in Kromhout
(1975) confirm this identification. Olinia usambarensis (InsideWood database) is also
very similar but in this species the rays have procumbent body cells with 1–4 rows of
marginal upright cells.
In southern Africa there are seven species of Olinia (Coates Palgrave, 2002; van
Wyk and van Wyk, 2013) occurring in low altitude forest or montane forest. It is unex-
pected to find this charcoal at Wonderwerk so it is treated with caution and will be
compared with charcoal from the upper layers.
Runge33-Book.indb 165Runge33-Book.indb 165 8/13/2015 7:43:14 AM8/13/2015 7:43:14 AM
166 Marion K. Bamford
Plate 6. Olinia ventosa (Oliniaceae). 1 – TS, vessels in radial lines of up to four cells
and in clusters, patches of thick-walled fibres. Scale=100 μm. 2 – RLS, mostly upright ray cells
with bands of square or procumbent cells. Scale bar=40 μm. 3, 4 – TLS, rays with uniseriate
portions as wide as the 2–3seriate portions. Marginal ray cells are larger and perforated
(spotty appearance). Scale bar 3=100 μm; 4=40 μm.
Runge33-Book.indb 166Runge33-Book.indb 166 8/13/2015 7:43:14 AM8/13/2015 7:43:14 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 167
Described sample: WW-5 M29 (0–5cm) G
Illustration: Plate 7, Figures1–4
Total number of pieces: 29
DESCRIPTION: Wood is semi-ring to diffuse-porous with possible ring boundaries
made of terminal/initial parenchyma bands. Perforation plates are simple and hori-
zontal. Vessels are arranged in short radial multiples of 1–4cells, and clusters (Plate 7,
Figures1, 2). Average vessel diameter is 20–40 μm and there are more than 80 vessels
per square mm. Vessel clusters comprise small and larger vessels. Parenchyma is vasi-
centric to aliform and prismatic crystals are visible in transverse section in both the rays
and parenchyma cells (Plate 7, Figure3). Rays are 1–2(−3) seriate and up to 200μm
high. Ray cells are predominantly procumbent but there are 1–2(−4) rows of marginal
upright cells that frequently contain crystals. Crystals also occur in the procumbent
cells. Inter-vessel and vessel-ray pits are small and alternate (Plate 7, Figure4).
IDENTIFICATION AND COMPARISON: Typical features of southern African
woods of the Rhamnaceae are short to long radial multiples, sometimes with very
numerous vessel members, rays 1–3 seriate and paratracheal parenchyma. Local
Ziziphus species have uniseriate rays so can be excluded. Rhamnus species also have
narrow rays, 1–2seriate. The charcoal is identified as Berchemia discolor based on
comparison with the modern reference material of B. discolor (A119) and Berchemia
zeyheri (Kromhout, 1975) as well as InsideWood descriptions. The charcoal is also
very similar to B. zeyheri but the latter has only scanty parenchyma whereas B. discolor
also has narrow bands of parenchyma.
There are only two species of Berchemia in southern Africa and they occur in
bushveld (Coates Palgrave, 2002; van Wyk and van Wyk, 2013).
Described sample: WW-5 Q29 (5–10cm) D
Illustration: Plate 8, Figures1–4
Total number of pieces: 2
DESCRIPTION: Perforation plates are simple and horizontal and the wood is diffuse-
porous. Vessel arrangement is short radial multiples of 1–3 pores but predominantly
solitary, with an average diameter of about 80 μm and less than 30 per square mm
(Plate 8, Figure1). Parenchyma is rare to absent. Rays are 1–5seriate, heterocellular
with mostly square cells but procumbent and upright cells are present. Rays are
400–800 μm high (Plate 8, Figures2–3). Inter-vessel pitting is alternate and minute
(Plate 8, Figure4); vessel-ray pitting was not seen. Fibres are thick-walled.
IDENTIFICATION AND COMPARISON: Woods of the Scrophulariaceae are
quite variable but tend to have fairly wide and tall rays with vessels arranged in a
tangential or oblique pattern. The charcoal is the same as Halleria lucida (A158) in
the modern reference collection and this was confirmed in the InsideWood key where
Halleria abyssinica is a synonym of H. lucida.
Runge33-Book.indb 167Runge33-Book.indb 167 8/13/2015 7:43:17 AM8/13/2015 7:43:17 AM
168 Marion K. Bamford
Plate 7. Berchemia discolor (Rhamnaceae). 1, 2 – TS, vessels in short radial multiples and clusters,
wood semi-ring porous (horizontal bands). Scale bar 1=200 μm; 2=40 μm. 3 – RLS,
1–2 rows of marginal upright cells and body procumbent ray cells some of which have
3–4 crystals per cell CHECK. Scale bar=100 μm. 4 – TLS, rays 1–2seriate.
Vessels long with small alternate inter-vessel pits. Scale bar=40 μm.
Runge33-Book.indb 168Runge33-Book.indb 168 8/13/2015 7:43:17 AM8/13/2015 7:43:17 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 169
Plate 8. Halleria lucida (Scrophulariaceae). 1 – TS, vessels in short radial multiples,
fibres thick walled (section cracked). Scale bar=100 μm. 2 – RLS, square ray cells.
Scale bar=100 μm. 3 – TLS, tall rays 3–5seriate. Inter-vessel pitting is alternate and minute
but appears scalariform at lower magnification. Scale bar=100 μm. 4 – TLS, small,
alternate inter-vessel pits. Scale bar=40 μm.
Runge33-Book.indb 169Runge33-Book.indb 169 8/13/2015 7:43:20 AM8/13/2015 7:43:20 AM
170 Marion K. Bamford
9.4.1 Distribution of charcoal within Stratum 5
Numbers of pieces of charcoal cannot be taken to reflect the abundance or prefer-
ential use of that wood because charcoal is fragile and can easily be fractured during
excavation thus inflating the number of pieces. Furthermore, different woods preserve
differently, some forming robust charcoal and others disintegrating or leaving no rec-
ognizable trace. Nonetheless presence and absence of taxa, and changing taxa over
time can be interpreted to reflect changes in selection or wood use, and/or changes in
wood availability or vegetation.
The most common charcoal in Stratum 5 is Ozoroa paniculosa with 40 pieces
occurring in 6 of the 19 grid squares excavated (Table 2). Most pieces occurred in
square M 29 (0–5cm depth). This square also had the most pieces of charcoal and
highest diversity (total 28; diversity 4species). The second most common charcoal was
Berchemia discolor with a total of 29 pieces in 6grid squares. The least abundant or
rare taxa are Commiphora sp. and Halleria lucida with only two pieces each. One piece
of Commiphora sp. occurs in sub-Stratum 5a but this taxon is abundant in Stratum 4.
With more material to work with from this and possibly other strata the species should
eventually be identifiable. Since the total number of charcoal pieces in Stratum 5 is low
compared with those of the upper strata no spatial analysis will be done at this stage.
9.4.2 Geographical range, ecology, wood quality and uses of charcoal
taxa from Stratum 5
There are two genera belonging to the Anacardiaceae, Ozoroa and Searsia (Rhus).
Ozoroa paniculosa is a small tree or shrub and grows in bushed grassland or on rocky
hillsides in dry, summer rainfall areas (Coates Palgrave, 2002; van Wyk and van Wyk,
2013). Searsia lancea is also a small tree to shrub, widespread in southern Africa and
grows in a variety of habitats. It occurs today in the Wonderwerk environs (ibid.;
Mucina and Rutherford, 2006; pers. obs.).
The three species of Ehretia (Boraginaceae) are widespread in the drier regions
of southern Africa and grow in a wide range of habitats. Ehretia rigida or the puzzle
bush, for example, occurs in wooded grassland, karroid vegetation and bushed grass-
land (Coates Palgrave, 2002; van Wyk and van Wyk, 2013). It is a deciduous shrub or
small tree and occurs today in the Wonderwerk Cave area.
Commiphora species are common in dry to very dry areas and are often leafless,
but some species occur in bushy grassland and bushland. Many species of the Burser-
aceae produce a scented sap or resin. The wood is light and generally not used for fuel-
wood but twigs from Commiphora schimperi are used as fire sticks (Coates Palgrave,
2002). This species and five others occur today not too far from Wonderwerk Cave and
may have had a wider distribution in the past.
Dombeya rotundifolia, the common wild pear, is a small tree or shrub, widely
spread in bushland, and on rocky outcrops, from Namibia, northern Botswana,
Zimbabwe, Limpopo Province of South Africa and Mozambique. It does not occur in
the Wonderwerk Cave area today.
There are seven species of Olinia in southern Africa (Coates Palgrave, 2002; van
Wyk and van Wyk, 2013) and they each have restricted distributions but occur in mon-
tane or evergreen forest. Olinia ventosa occurs along the southern Cape coast and is an
evergreen tree with strong, hard wood. Its current habitat and distribution place the
Runge33-Book.indb 170Runge33-Book.indb 170 8/13/2015 7:43:23 AM8/13/2015 7:43:23 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 171
identification of the charcoal in doubt so this requires further research and sourcing
of more comparative material. Two other species also occur in forest or forest margins
but much closer to Wonderwerk. More reference material is needed to confirm the
identification of the species.
There are about 20indigenous species of the Rhamnaceae in southern Africa
(Coates Palgrave, 2002) including Berchemia discolor and Berchemia zeyheri. Today
neither species occurs in the Wonderwerk Cave area but B. discolor occurs in low
altitude bushveld across northern Namibia, Botswana and Zimbabwe and south to
Mozambique and eastern Limpopo. B. zeyheri also occurs in bushveld in the Northern
Province, Limpopo and northern KwaZulu-Natal. The woods of both species are hard
and used for making furniture.
Halleria lucida is one of the few woody members of the Scrophulariaceae. It
is a shrub to small tree occurring in forest margins, grasslands and along streams in
rocky places. It ranges from Northern Province, Limpopo, KwaZulu Natal to the
Cape Coast but does not occur in the Wonderwerk area today (Coates Palgrave, 2002;
van Wyk and van Wyk, 2013). The wood has many uses including fire sticks (Coates
9.4.3 Comparison with other records
Older charcoal records from South Africa have been discussed in Bamford (AAR in
2015). Sites that contain preserved charcoal that are contemporaneous with Stratum 5
(i.e. slightly older than 12 ka) are few in number. Elands Bay Cave on the southwestern
Cape coast has charcoals of proteoid fynbos in the levels dated to 13.6–12.45 ka BP
(Cowling et al., 1999) which is different from the Wonderwerk flora. From Rose Cottage
near the Caledon River the charcoals from 12–10 ka indicate cool conditions (Ester-
huysen et al., 1999). Charcoal from Wonderkrater in the savanna biome of northern
South Africa at ∼12–13 ka has predominantly Phragmites sp and grass charcoal plus
some Diospyros austro-africana wood probably indicating fluctuating wetter and drier
periods (Backwell et al., 2014). Prior and Price-Williams (1985) described charcoal
from early Holocene layers in Siphiso rock shelter in northeast Swaziland: from Stra-
tum 6 (12–9 ka) and indicate moist conditions from Leguminosae, Combretum apicu-
latum, Androstachys johnstoni, Maytenus and Diospyros species.
The Rose Cottage charcoal taxa from Level DB dated at 12,690 years are: Protea
spp., Leucosidea sericea, Maytenus spp., Scolopia mundii, Heteromorpha trifoliata and
Passerina montana (Wadley et al. 1992; Esterhuysen et al. 1999). The site is on the
northwestern side of the Drakensberg Mountains, eastern Free State, at a higher alti-
tude than Wonderwerk, and represents a region of higher rainfall.
Pollen records for the pre-Holocene have been summarised by Scott et al. (2012)
for the various biomes. Wonderwerk falls within the savanna/dry woodland biome
and pollen from here and Equus Cave indicate lower temperatures, relative to todays,
before 17 ka with general warming thereafter. They also show strongly oscillating
moisture between about 25 and 11 ka (with peaks at 17 ka, 14,6 ka and 12.6 ka). After
12.6 ka the drying trend reached a peak in dryness between 12–11 ka (the Younger
Dryas) with evidence of a sub-humid woodland sequence (Scott et al., 2012). Recent
pollen studies at Wonderwerk Cave show that from about 12,500 to 6500cal years BP
(Strata 5–4c), there were some fluctuations in moisture and temperature, based on the
changing proportion of grass, fynbos and Asteraceous elements (Scott and Thackeray,
AAR in 2015). They add that the vegetation would have resembled the “renosterveld”
that forms the current transition between the Nama Karroo and Fynbos Biomes and
Runge33-Book.indb 171Runge33-Book.indb 171 8/13/2015 7:43:23 AM8/13/2015 7:43:23 AM
172 Marion K. Bamford
that the vegetation seems to have been growing under moderately humid conditions,
although not as humid as in areas in the eastern parts of the country during this
9.4.4 Pre-Holocene Wonderwerk vegetation
Charcoals from Wonderwerk would have been the product of human selection and
activity. The woods represented are, therefore, a subset of the plants growing in the
vicinity and represent an unknown proportion of the flora. Nonetheless they can be
used to deduce the past climate. There are about 134 fragments and they represent
eight identifiable woody species and 34 unidentified pieces. The diversity is relatively
low but seven families are represented. The most common woods are Ozoroa panicu-
losa and Berchemia discolor. Of the eight identified charcoal taxa, only two are listed
in the modern vegetation (Searsia (Rhus) lancea and Ehretia rigida) and they toler-
ate low rainfall in summer and frost during winter. One species occurs in dry to arid
regions (Commiphora sp.). Two species occur in fairly dry regions (Ozoroa paniculosa,
Dombeya rotundifolia). Two species occur in more mesic areas (Berchemia discolor and
Halleria lucida) and one species occurs in moister and more forested conditions, Olinia
ventosa. This may correspond with the peak in moisture reflected by the pollen spec-
trum at Wonderwerk at 12.6 ka (Scott et al., 2012). There is an anomaly in the dates
given by Lee Thorpe and Ecker (AAR in 2015) for Stratum 5 which they cannot fully
explain but it could also be a problem with mixing of samples.
Considering the range of woods represented by the charcoal it is possible that
their past range was wider than it is today as it is unlikely that the cave inhabitants
would have walked great distances to collect firewood. The majority of the woods
found can tolerate dry conditions and a wide range of temperatures but the pres-
ence of Berchemia discolor indicates that conditions between c. 14,985 and 13,952cal.
BP may have been slightly wetter than today. Furthermore, there seem to have been
important climatic shifts during the pre-Holocene but higher resolution records are
needed to determine the degree and rate of climate change.
I would like to thank Liora Kolska Horwitz, James Brink and Lloyd Rossouw for
organizing the Tribute to Louis Scott in Bloemfontein in July 2014 and for inviting me
to participate. I also thank Michael Chazan and Liora for including me in the Wonder-
werk project, and PAST for funding for the modern plant reference collection at the
Evolutionary Studies Institute.
Avery, D.M., 1981, Holocene micromammalian faunas from the northern Cape Prov-
ince. South Africa. South African Journal of Science, 77, pp. 265–273.
Avery, D.M., 2007, Pleistocene micromammals from Wonderwerk Cave, South Africa:
practical issues. Journal of Archaeological Science, 34, pp. 613–625.
Backwell, L.R., McCarthy, T.S., Wadley, L., Henderson, Z., Steininger, C.M., deKlerk,
B., Barré, M., Lamothe, M., Chase, B.M., Woodborne, S., Susino, G.J., Bamford,
M.K., Sievers, C., Brink, J.S., Rossouw, L., Pollarolo, L., Trower, G., Scott, L. and
Runge33-Book.indb 172Runge33-Book.indb 172 8/13/2015 7:43:23 AM8/13/2015 7:43:23 AM
Charcoal from pre-Holocene Stratum 5, Wonderwerk Cave, South Africa 173
d’Errico, F., 2014, Multiproxy record of late Quaternary climate change and Middle
Stone Age human occupation at Wonderkrater, South Africa. Quaternary Science
Reviews, 99, pp. 42–59.
Bamford, M.K., (2015), Macrobotanical remains from Wonderwerk Cave (Excavation
1), Oldowan to Late Pleistocene (2Ma to 14ka BP), South Africa. African Archaeo-
Beaumont, P., 2004, Wonderwerk Cave. In Archaeology in the Northern Cape: Some
Key Sites, edited by Beaumont, P. and Morris, D. (Kimberley: McGregor Museum),
Beaumont, P.B., 1990, Wonderwerk Cave. In Guide to archaeological sites in the north-
ern Cape, edited by Beaumont, P. and Morris, D. (Kimberley: McGregor Museum),
Berna, F., Goldberg, P., Horwitz, L.K., Brink, J., Holt, S, Bamford, M. and Chazan,
M., 2012, Microstratigraphic evidence for in situ fire in the Acheulean strata of Won-
derwerk Cave, Northern Cape Province, South Africa. PNAS Plus, 9(20), E1215–
Camp, C.L., 1948, University of California African Expedition – southern section.
Science, 108, pp. 550–552.
Chazan, M., Avery, D.M., Bamford, M.K., Berna, F., Brink, J., Holt, S., Fernandez-
Jalvo, Y., Goldberg, P., Matmon, A., Porat, N., Ron, H., Rossouw, L., Scott, L. and
Horwitz, L.K., 2012, The Oldowan horizon in Wonderwerk Cave (South Africa):
Archaeological, geological, paleontological and paleoclimatic evidence. Journal of
Human Evolution, 63, pp. 859–866.
Chazan, M., Ron, H., Matmon, A., Porat, N., Goldberg, P., Yates, R., Avery, M.,
Sumner, A. and Horwitz, L.K., 2008, Radiometric dating of the Earlier Stone Age
sequence in excavation 1 at Wonderwerk Cave, South Africa: preliminary results.
Journal of Human Evolution, 55, pp. 1–11.
Coates Palgrave, M., 2002, Keith Coates Palgrave Trees of Southern Africa, 3rd Edi-
tion, (Cape Town: Struik Publishers).
Cowling, R.M., Cartright, C.R., Parkington, J.E. and Allsopp, J.C., 1999, Fossil wood
charcoal assemblages from Elands Bay Cave, South Africa: implications for Late
Quaternary vegetation and climates in the winter rainfall Fynbos biome. Journal of
Biogeography, 26, pp. 367–378.
Esterhuysen, A.B., Mitchell, P.J. and Thackeray, J.F., 1999, Climatic change across
the Pleistocene/Holocene boundary in the Caledon River, southern Africa: results
of a factor analysis of charcoal assemblages. Southern African Field Archaeology, 8,
IAWA Committee, 1989, IAWA list of microscopic features for hardwood identifica-
tion. IAWA Bulletin, n.s. 10, pp. 219–332.
InsideWood (accessed 2011–May 2014): http://insidewood.lib.ncsu.edu/
Kent, L.E., ed, 1980, Stratigraphy of South Africa. 8. Part I. Geological survey of
South Africa Handbook (The South African Committee for Stratigraphy).
Kromhout, C.P., 1975, ‘n Sleutel vir die mikroskopiese uitkenning van die vernaamste
inheemse houtsoorte van Suid-Afrika. South African Department of Forestry, Preto-
ria, Bulletin, 50, 124 pp.
Lee-Thorp, J. and Ecker, M., (2015), Holocene environmental change at Wonderwerk
Cave, South Africa: Insights from stable light isotopes in ostrich egg shell. African
Malan, B.D. and Cooke, H.B.S., 1941, A preliminary account of the Wonderwerk
Cave, Kuruman District, South Africa. South African Journal of Science, 37,
Runge33-Book.indb 173Runge33-Book.indb 173 8/13/2015 7:43:23 AM8/13/2015 7:43:23 AM
174 Marion K. Bamford
Matmon, A., Ron, H., Chazan, M., Porat, N. and Horwitz, L.K., 2011, Reconstruct-
ing the history of sediment deposition in caves: A case study from wonderwerk Cave,
South Africa. Geological society of America Bulletin, 124, pp. 611–625.
Mucina, L. and Rutherford, M.C., eds, 2006, The vegetation of South Africa, Lesotho
and Swaziland, Strlitzia 19 (Pretoria: South African National biodiversity Institute),
Prior, J. and Price Williams, D., 1985, An investigation of Climatic change in the Hol-
ocene epoch using archaeological charcoal from Swaziland, South Africa. Journal of
Archaeological Science, 12, pp. 457–475.
Rüther, H., Chazan, M., Schroeder, R., Neeser, R., Held, C., Walker, S.J., Matmon, A.
and Horwitz, L.K., 2009, Laser scanning for conservation and research of African
cultural heritage sites: the case study of Wonderwerk Cave, South Africa. Journal of
Archaeological Science, 36, pp. 1847–1856.
Schulze, B.R., 1984, Climate of South Africa. Part 8: General survey. WB 28, fifth ed.
(Pretoria: South Africa Weather Bureau).
Scott, L., Neumann, F.H., Brook, G.A., Bousman, C.B., Norström, E. and Metwally,
A., 2012, Terrestrial fossil-pollen evidence of climate change during the last 26 thou-
sand years in Southern Africa. Quaternary Science Reviews, 32, pp. 100–118.
Scott, L. and Thackeray, J.F., (2015), Palynology of Holocene deposits in Excava-
tion 1 at Wonderwerk Cave, Northern Cape (South Africa). African Archaeological
Thackeray, A.I., 1983, Archaeological sites in the Kuruman Hills area. In Ghaap and
Gariep. Later Stone Age Studies in the Northern Cape, edited by Humphries, A.J.B.
and Thackeray, A.I. (Cape Town: The South African Archaeological Society),
Thackeray, J.F. and Lee-Thorp, J.A., 1992, Isotopic analysis of equid teeth from Won-
derwerk Cave, northern Cape Province, South Africa. Palaeogeography, Palaeocli-
matology, Palaeoecology, 99, pp. 141–150.
Van Wyk, B. and van Wyk, P., 2013, Field Guide to Trees of Southern Africa. Second
edition, (Cape Town: Struik Publishers).
Van Zinderen Bakker, E.M., 1982, Pollen analytical studies of the Wonderwerk Cave,
South Africa. Pollen et Spores, 26, pp. 235–250.
Vogel, J.C., Fuls, A. and Visser, E., 1986, Pretoria radiocarbon dates III. Radiocarbon,
Wadley, L., Esterhuysen, A. and Jeanneret, C., 1992, Vegetation changes in the east-
ern Orange Free State: the Holocene and later Pleistocene evidence from charcoal
studies. South African Journal of Science, 88, pp. 558–563.
Runge33-Book.indb 174Runge33-Book.indb 174 8/13/2015 7:43:23 AM8/13/2015 7:43:23 AM