ChapterPDF Available

First chronological, Palaeoenvironmental, and archaeological data from the Baden-Baden fossil spring complex in the western Free State, South Africa


Abstract and Figures

The Baden-Baden spring mound is one of the extremely scarce archaeological sites in the dry central and western interior of South Africa, where fossil fauna and also palaeobotanical material are preserved. This is the first and preliminary summary of ongoing palaeoenvironmental research at this spring mound complex, which is situated 70 km northwest of Bloemfontein near Dealesville. Topographic mapping, radiocarbon and OSL dating complement the archaeological, faunal and palynological records from Baden-Baden and compare it to other spring, pan and alluvial sites in the region like Florisbad, Deelpan and Erfkroon. OSL and radiocarbon dating places the available sequences within the last ~160 ka. Holocene archaeological and faunal remains were recovered from several excavations on the east side of the primary mound. These materials provide unique insights into prehistoric human adaptations in the grassveld. Pollen, extracted from a peat mound and buried organic layers beneath sand accumulations, suggests cooler, moist conditions during the late Pleistocene and drier conditions in the Holocene. These palaeoenvironmental proxy indicators offer the potential for better understanding of long-term climate and vegetation changes in the western Free State.
Content may be subject to copyright.
First chronological, palaeoenvironmental,
and archaeological data from the Baden-Baden
fossil spring complex in the western Free State,
Andri C. van Aardt
Department of Plant Sciences, University of the Free State,
Bloemfontein, South Africa
C. Britt Bousman
Anthropology Department, Texas State University, San Marcos, Texas
GAES, University of the Witwatersrand, Johannesburg, South Africa
James S. Brink
Florisbad Quaternary Research Department, National Museum,
Bloemfontein, South Africa
Centre for Environmental Management, University of the Free State,
Bloemfontein, South Africa
George A. Brook
Department of Geography, University of Georgia, Athens, USA
Zenobia Jacobs
Centre for Archaeological Science, School of Earth and Environmental
Sciences, University of Wollongong, Wollongong, Australia
Pieter J. du Preez
Department of Plant Sciences, University of the Free State,
Bloemfontein, South Africa
Lloyd Rossouw
Department of Plant Sciences, University of the Free State,
Bloemfontein, South Africa
Archaeology Department, Museum, Bloemfontein, South Africa
Louis Scott
Department of Plant Sciences, University of the Free State,
Bloemfontein, South Africa
Runge33-Book.indb 117Runge33-Book.indb 117 8/13/2015 7:42:31 AM8/13/2015 7:42:31 AM
118 Andri C. van Aardt et al.
ABSTRACT: The Baden-Baden spring mound is one of the extremely scarce archaeologi-
cal sites in the dry central and western interior of South Africa, where fossil fauna and also
palaeobotanical material are preserved. This is the first and preliminary summary of ongoing
palaeoenvironmental research at this spring mound complex, which is situated 70km north-
west of Bloemfontein near Dealesville. Topographic mapping, radiocarbon and OSL dating
complement the archaeological, faunal and palynological records from Baden-Baden and
compare it to other spring, pan and alluvial sites in the region like Florisbad, Deelpan and
Erfkroon. OSL and radiocarbon dating places the available sequences within the last 160 ka.
Holocene archaeological and faunal remains were recovered from several excavations on the
east side of the primary mound. These materials provide unique insights into prehistoric
human adaptations in the grassveld. Pollen, extracted from a peat mound and buried organic
layers beneath sand accumulations, suggests cooler, moist conditions during the late Pleisto-
cene and drier conditions in the Holocene. These palaeoenvironmental proxy indicators offer
the potential for better understanding of long-term climate and vegetation changes in the
western Free State.
Baden-Baden is a complex of spring mounds located 70km northwest of Bloemfon-
tein in the western Free State Province near a large pan known as Annaspan (Bousman
and Brink, 2008, 2012) (Figure1). The site is locally well known due to a pre-Boer War
bath house built over a large mineral spring on the east side of the largest mound. The
bath house is historically interesting because of the inscriptions written by bathers on
the outside walls. Historical records indicate that Dr. Brownlow from Europe believed
Figure 1. Locality map of Baden-Baden in its bioregion and three major surrounding biomes.
Insert: Annaspan and the positions of auger holes in the main Baden-Baden spring mound
and the dune south of the site.
Runge33-Book.indb 118Runge33-Book.indb 118 8/13/2015 7:42:31 AM8/13/2015 7:42:31 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 119
the spring water, which is rich in NaCl and MgSO
, had healing properties and estab-
lished a hospital near the site ca. 1905 (N. Nel, pers. comm., letter from Dr.Albert
Wessels). The palaeoenvironmental and archaeological potential of this site was first
recognized by Louis Scott in 1987.
Annaspan (Figure1) is one of the largest pans in South Africa measuring 7.5km
in its maximum dimension and covering approximately 1850 hectares. Pans are very
common in the western Free State and frequently occur in ancient abandoned stream
channels (Marshall, 1987, 1988). They form by wind erosion that removes unconsoli-
dated sediments down to bedrock (Butzer and Oswald, in press). This wind erosion
creates large aeolian dunes on the leeward sides of the pans. Frequently, when bedrock
allows, springs emerge on the slopes of pans and sandy mounds form around these
springs. Rarely these sand accumulations grow to form large spring mounds such as
found at Baden-Baden and Florisbad (Grobler and Loock, 1988; Visser and Joubert,
1991; Kuman et al., 1999).
The primary spring mound at Baden-Baden is a sandy hillock that stands eight
meters above the surrounding landscape (Figures2 and 3). Two short streams drain
the eastern and western side of the large mound and flow into Annaspan. On the
eastern drainage, a number of active springs and seeps emanate from a complex series
of small mounds. The western drainage starts at a different set of large and small
springs where a bathing pool, now abandoned, was built. Baden-Baden has in situ
faunal remains and artefacts dating to the Later Stone Age and sporadic surface Mid-
dle Stone Age occurrences. It is similar to the Florisbad spring mound about 30km to
the east and well-known for its fossil human cranium and faunal remains (Grün et al.,
1996; Kuman et al., 1999; Brink, 2005), but the primary mound contains fewer organic
layers than Florisbad (Bousman and Brink, 2008, 2012). The deposits at Baden-Baden
are important because of their archaeological, faunal and plant microfossil records
relating to human occupations and palaeoenvironments of the central interior of
South Africa.
Several research teams have worked at Baden-Baden. In 2003, Louis Scott and
his student, Bokang Theko, sampled for pollen analysis (Theko et al., 2003). Later
that same year and in 2006 a team from Texas State University and the National
Museum, headed by James Brink and Britt Bousman, undertook topographic map-
ping, archaeological excavations, and geological trenching. In 2005 and 2006 a team
from the University of Georgia under George Brook’s direction, cored the main
sandy spring mound and the sand dune to the south, and also collected samples
from Annaspan. The archaeological, faunal, palynological and dating results reveal a
complex set of Holocene and Pleistocene deposits with preserved Holocene human
occupations. Three blocks (North, Central and South) were excavated archaeologi-
cally along the eastern edge of the spring mound. Here we present a brief descrip-
tion of the site and report the combined initial results that can serve as a basis for
ongoing research and to assess the potential of the site for making contributions to
palaeoenvironmental studies in the central South African interior. There seems to be
renewed interest in this often neglected region with activities focusing at sites near
Baden-Baden like Erfkroon, Florisbad and Deelpan and further to the west in the
Southern Kalahari at Equus Cave, Wonderwerk Cave and Kathu Pan (Scott, 1987;
Beaumont, 2004; Gil Romera et al., 2014; Scott and Thackeray, 2014; Butzer and
Oswald, in press). Together with these efforts, future Baden-Baden studies promise to
expand the palaeoenvironmental information in the region where previous research,
such as at Alexandersfontein, indicated marked environmental changes, including the
existence of a palaeolake in that pan basin during the late Quaternary (Butzer etal.,
Runge33-Book.indb 119Runge33-Book.indb 119 8/13/2015 7:42:32 AM8/13/2015 7:42:32 AM
120 Andri C. van Aardt et al.
Figure2. A, The Phragmites australis reed community which dominates the moister parts of
the secondary spring mounds with some planted exotic trees. B, Profile cross-section showing
elevations for described backhoe trenches, excavation units, augers and the Pollen Pit. Backhoe
trench and excavation sediment textures and colours are correlated by sedimentary units (according
to Tables1–7). Augers and the Pollen Pit sections illustrated by elevation but not correlated into
sedimentary units. OSL and calibrated radiocarbon ages and sample locations plotted by elevations on
appropriate profiles.
Runge33-Book.indb 120Runge33-Book.indb 120 8/13/2015 7:42:32 AM8/13/2015 7:42:32 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 121
The study area falls within the Grassland Biome of South Africa (Mucina and Ruther-
ford, 2006) (Figure1). The altitude in the area is 1 300m amsl (above mean sea level),
sloping towards the west (Holmes et al., 2008). The interior of South Africa experiences
temperatures below freezing during the winter months, and frost at the site is a com-
mon phenomenon (Schulze, 1972; Bousman, personal observation). This is a summer-
rainfall area (MAP 380–530mm) with high summer temperatures (Geldenhuys, 1982;
Schulze, 1997). Evaporation increases towards the west of the Province—estimates are
that 91% of South Africa’s MAP is returned to the atmosphere (Schulze, 1997).
Baden-Baden is in the western portion of the Grassland Biome within the Dry
Highveld Grassland Bioregion (Mucina and Rutherford, 2006, pp. 32–51). The mound
is positioned at the boundary between the Vaal-Vet Sandy Grasslands and the Western
Free State Clay Grasslands (Mucina and Rutherford, 2006). The Western Free State
Clay Grasslands are a species-poor dry grassland community growing on thin clayey
Figure 3. Topographic map of the Baden-Baden site showing the main mound in the centre (shaded)
the secondary mounds (centre right) and the adjacent dune (shaded, far right).
Runge33-Book.indb 121Runge33-Book.indb 121 8/13/2015 7:42:34 AM8/13/2015 7:42:34 AM
122 Andri C. van Aardt et al.
soils overlying Ecca shales that form the lower-lying flat plains. The Vaal-Vet Sandy
Grasslands, dominated by Themeda triandra, occurs in areas of higher elevation on
aeolian and colluvial sandy hills and ridges overlying Ecca shales. At Baden-Baden
this type of grassland occurs on aeolian dune deposits eroded from Annaspan. Vegeta-
tion surrounding Annaspan is classified as Highveld Salt Pan Vegetation.
Several local vegetation communities occur at Baden-Baden on the spring mounds
and in the valley-bottom. The primary spring mound and the drier parts of the sec-
ondary spring mounds are covered by a Cynodon dactylonEragrostis lehmanniana
grass community which include the karroid shrub Chrysocoma ciliata and the grasses
Digitaria eriantha and Aristida congesta. The elevated mounds have deep well-drained
sandy soils that are leached of salts and minerals. Burrowing animals such as aard-
varks (Orycteropus afer) and suricates (Suricata suricatta) caused some disturbance on
the mounds. On parts of the primary and secondary spring mounds where water seeps
to the surface or where the water table is relatively close to the soil surface a Phrag-
mites australis community dominates (Figure2A) accompanied by aquatic forbs like
Ranunculus multifidus.
Depending on variations in salt concentrations, soil texture, surface water avail-
ability, leaching and the degree of disturbance a mosaic of three different plant com-
munities occurs at the base of the spring mounds in the valley-bottom wetland. For
example, in patches where soil pH and salt concentrations are high a Limonium drege-
anum community occurs. Areas with high salt concentration that result from seep water
evaporation are relatively hostile to most wetland plants except for the halophytic rush
Juncus rigidus and some Sporobolus virginicus. Large dry parts of the valley floor have
bare patches or are covered by reddish sands that support a community dominated by
Suaeda fruticosa dwarf shrubs and fewer Salsola aphyla. Around deep water along the
fringe of the impoundment a sedge community with Schoenoplectus triqueter (Scirpus
triqueter) dominates together with some stands of bulrush (Typha capensis).
Topographic surveying started during 2003 when elevation measurements were col-
lected in order to construct a detailed map of the site (Figure3). A local arbitrary
XYZ grid was established using a Sokkia SET 600 Total Station with a Sokkia data
collector. Over 1855 surface elevation points were measured including a trig marker
with a known elevation approximately 2km to the southeast of the primary mound.
The trig marker elevation was used to calibrate the arbitrary heights to elevations
above mean sea level. Elevations of all excavations and backhoe trenches were also
recorded and used to construct a topographic map in Surfer 8.0.
8.3.1 Archaeological excavations
A series of 1 × 1meter excavation units were dug at the base and on the edge of the
primary mound at Baden-Baden. Most were grouped in northern, central and south-
ern blocks, but a few isolated units were excavated also. All artefacts, bone, sediment
samples and other materials were plotted with a Sokkia Total Station and locations
recorded with a data collector. All sediment was wet sieved in 5mm screen and recorded
by arbitrary 10cm units and natural stratigraphy. Sediment, pollen and dating samples
were collected from each block. Seven backhoe trenches were excavated at various
locations to provide geological observations in areas not sampled by archaeological
Runge33-Book.indb 122Runge33-Book.indb 122 8/13/2015 7:42:35 AM8/13/2015 7:42:35 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 123
excavations. The artefacts and fauna from the archaeological excavations were cleaned,
analysed and curated at the Florisbad Quaternary Research Station.
The sediments in the archaeological excavations and backhoe trenches were
recorded using a combination of pedogenic and geologic methods. Sediment textures
were determined in the field using Olsens (1981, pp. 23–24) methods. Sediment colour
was determined on moist sediment using a Munsell Soil Color Chart. Soil structure
and other features classification follows the definitions in the Soil Survey Manual and
Keys to Soil Taxonomy (Soil Survey Staff, 1993, 2010), and sedimentary structures
follow those outlined in Depositional Sedimentary Environments (Reineck and Singh,
1975) and The North American Stratigraphic Code (North American Commission on
Stratigraphic Nomenclature, 2005).
8.3.2 Dating
To provide a provisional chronology for the site’s deposits, samples were collected
from different backhoe trenches, excavation units, cores and pollen profiles for radio-
carbon (
C) and Optically Stimulated Luminescence (OSL) dating at locations shown
in Figure4. Four
C ages have been obtained from the North and Central Blocks from
Quaternary Dating Research Unit (QUADRU) at the CSIR in Pretoria, South Africa,
in 2003. A further two samples of organic spring deposits from a pit (the Pollen Pit) in
one of the secondary spring mounds, were submitted and dated by Beta Analytic. The
pit was dug separately for exploratory pollen analysis by LS and not recorded by the
same criteria of sediment description as given above.
The first two sediment samples for OSL dating (OSL#4 and OSL#6) were processed
at QUADRU. These samples came from a backhoe trench adjacent to the archaeological
and palaeontological excavation from which samples for
C dating were collected. OSL#4
was collected from Zone 6, 143cm below the present ground surface, and OSL#6 from
Zone 10, 222cm below the present ground surface (Figure4). The samples were taken
by inserting a metal pipe into the section wall to prevent the sample from being exposed
to light. Both ends of the pipe were sealed and the sample tube was labelled.
Auger samples for further OSL dating were processed at the Luminescence Labo-
ratory of the University of Georgia (UGA). In July 2005 the spring mound at Baden-
Baden (28° 34.402 S; 25° 49.822 E) was augered using a sand bucket auger to 5.2m
depth. Samples of sediment and samples for OSL dating were collected at 0.97, 2, 3
and 4.5m depth (BB-1 to BB-4). Further sampling was undertaken in July 2006 on a
relict dune ridge at the edge of a corn field (28° 34.623 S; 25° 50.016 E). The dune was
augered to 5m depth and samples for OSL dating were collected in a bucket auger at
1, 2, 3, 4, and 4.55m depth (Figure4).
When samples were taken, the auger was twisted deep into the sand to force new
sand upwards through the bucket. At the surface a piece of PVC pipe or a tin can with
a pierced base was pushed into the top of the sand in the bucket after cleaning away
a few centimeters of sand. The auger and pipe were turned upside down; the pipe was
removed and capped at both ends, wrapped in thick, opaque black plastic, sealed with
duct tape and labeled. In the laboratory, the potentially light-exposed portions at both
ends of the sampling tubes were removed.
In both laboratories, all sample preparation and luminescence measurements were
carried out in subdued red light. Raw samples were treated with 10% HCl and 20%
to remove carbonate and organic matter. After drying, the samples were sieved
to select the grain size in the size ranges of 90–125μm, 125–180μm and 180–250μm.
Heavy liqui ds of densities 2.62 and 2.75g/cm
separated the grain fraction to obtain
Runge33-Book.indb 123Runge33-Book.indb 123 8/13/2015 7:42:35 AM8/13/2015 7:42:35 AM
124 Andri C. van Aardt et al.
quartz and feldspar. The quartz grains were treated with 40% HF for 60min to remove
the outer layer irradiated by alpha particles and remaining feldspars. The grains were
then treated with 1mol/L HCl for 10min to remove fluorides created during the HF
etching. Pure quartz fractions with grains in the size range 125–180 μm (UGA) or
180–212μm (QUADRU) were acquired finally.
The OSL measurements were carried out in both laboratories, using an auto-
mated Risø TL/OSL-DA-15 reader (Markey et al., 1997). Light stimulation of quartz
mineral extracts was undertaken with an excitation unit containing blue light-emitting
diodes (λ=470±30nm) (Bøtter-Jensen et al., 1999). Detection optics comprised two
Hoya 2.5mm thick U340 filters and a 3mm thick stimulation Schott GG420 filter
coupled to an EMI 9635 QA photomultiplier tube. Laboratory irradiation was carried
out using calibrated
Y sources mounted within the readers.
The equivalent dose (D
), the numerator in the OSL age equation, of quartz was
determined by the SAR protocol (Murray and Wintle, 2000). The purity of quartz
Figure4. Topographic map of the study area adjacent to the main mound at Baden-Baden showing
the positions of the auger hole for OSL dating and the excavation areas of the North,
Central and South Blocks.
Runge33-Book.indb 124Runge33-Book.indb 124 8/13/2015 7:42:35 AM8/13/2015 7:42:35 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 125
was checked by IRSL at 50°C (UGA) or the OSL-IR depletion ratio test (QUADRU)
and the results showed that none of the samples contained feldspar in the quartz frac-
tion. The OSL dating recuperation test and dose recovery test showed that the SAR
protocol was reliable for the samples examined in this study. Data were analyzed using
the ANALYST program of Duller (1999). It is possible that some sediments may be
insufficiently and/or unevenly bleached prior to burial. Whether the samples were fully
bleached prior to burial can be detected by measuring aliquots that contain a small
number of grains and examining the distribution among D
values obtained from many
individual aliquots from the same sample. Overdispersion (OD) values ranging from
5±2% to 55±10% were obtained and were spread symmetrically around a common
value; OD is the amount of scatter left after all sources of measurement uncertainty
are taken into account. The individual D
values were all statistically consistent with
each other and showed no clear evidence for partial bleaching and were combined, to
obtain a single value of D
for age determination, using the central age model of Gal-
braith et al. (1999). Overdispersion values much greater than 20% (at two sigma limits)
for auger samples UGA06OSL-331 and 332indicate mixing or grains of various ages
or partial bleaching of grains. Under conditions of partial bleaching a minimum age
model can be applied but in this case we believe the samples contain a mixture of grains
of different age, some of the mixing possible due to incorporation of younger grains
that fell into the auger hole during the augering process, and were incorporated into
older samples despite the precautions taken. Because of this, central, rather than mini-
mum ages were calculated for these samples. However, the very high over-dispersion
values for these two samples question their accuracy and here we use them only as
approximate estimates of age, despite the ages being in correct stratigraphic order.
The environmental dose rate, the denominator in the OSL age equation, is cre-
ated by the radioactive elements existing in grains of the sample and the surrounding
sediments, with a small contribution from cosmic rays. For all the samples measured in
both laboratories, thick source Daybreak alpha counting systems were used to estimate
U and Th for the dose rate calculation. K contents were measured by ICP90, using the
sodium peroxide fusion technique at the SGS Laboratory in Canada for the samples
from UGA, and by X-ray fluorescence for the samples from QUADRU. All measure-
ments were converted to alpha, beta and gamma dose rates according to the conversion
factors of Aitken (1985, 1998). The dose rate from cosmic rays was calculated based on
sample burial depth and the altitude of the section (Prescott and Hutton, 1994). Water
content of the sediment samples examined here must have changed drastically over
time after they were buried. This is because conditions varied between more arid to less
arid resulting in alternating periods of decreased and increased surface and ground-
water flow. Therefore, estimation of the water content was based on what is known
about the history of the deposits and on as-collected values of moisture content which
were all less than 5% of dry sample weight. Since it is not possible to accurately deter-
mine the mean water content during the sediment burial period, a slightly higher value
for water content (5±2.5%) than that measured in the samples was assumed in age cal-
culations for the sediments measured at UGA; the current measured water contents of
17±5 and 9±3% were used for QUADRU samples OSL#4 and OSL#6, respectively.
8.3.3 Palynology
Pollen was extracted from deposits using standard methods, which included digestion
in HCl (10%) for carbonates or KOH (5%) for peaty samples, and mineral separation
using ZnCl
solution (specific gravity 2) and mounted in glycerine jelly for microscope
Runge33-Book.indb 125Runge33-Book.indb 125 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
126 Andri C. van Aardt et al.
8.3.4 Phytoliths
Grass short-cell phytoliths were extracted from samples BB-1, BB-2, BB-3 and BB-4,
which were taken at the spring mound auger site. Standard laboratory steps included
deflocculation, the elimination of carbonates using HCl in low concentration (10%),
removal of clays by means of sedimentation and mineral separation with a heavy liq-
uid solution of sodium polytungstate (specific density=2.3). Fractions were mounted
on microscope slides in glycerin jelly and scanned under a Nikon 50i polarizing micro-
scope at x400magnification. Provisional observations focussed on the identification
of non-lobate grass short-cell phytoliths as indicators of C
and C
tal conditions (Rossouw, 2009).
The topographic analysis of the site shows that the crest of the primary mound forms a
bi-lobed oval and this is likely due to the presence of two large mounds that have merged
into a single mound (Figures3 and 4). Two streams drain the numerous springs. The
larger drainage is on the eastern side of the primary mound where a bathhouse was con-
structed at a large spring eye on the eastern flank of the primary mound. The eastern
stream flows north then turns northwest toward Annaspan. A dam was constructed at
that point. The eastern stream also extends upstream to the southwest where it is joined
by a minor tributary. A number of smaller secondary mounds are also evident on the
eastern stream. Some minor mounds flank the primary mound on its southeast side and
others occur on an eastern tributary drainage. These smaller mounds adjacent to the trib-
utary drainage appear to be coalescing into a new mound complex and they merge with a
large dune deposit to the east and south. The number and distribution of smaller mounds
illustrates the complex nature of mound genesis. A second major spring is southwest
of the primary mound. This is drained by a smaller stream that flows to the northwest
toward Annaspan. A bathing pool was also constructed on the south side of the primary
mound at the second major spring but it no longer holds water and is not in use.
8.4.1 Geology
The archaeological excavations, geological profiles, pollen profiles and geological cores
for dating are on the eastern and southern side of the mound (Figure4). Previously
the slopes on the eastern side of the primary mound were removed creating a steep cut
bank. This cut bank provided a vertical view of the mound’s lower slope sediments and
exposed archaeological materials eroding from the deposits. The surface archaeological
materials were used to determine the location of the archaeological excavations. In addi-
tion to the excavation units, a number of backhoe trenches were excavated to gain a bet-
ter understanding of the depositional history of the site. A description of the geological
profiles demonstrates the nature of sediment accumulation in the mound. The main
sediment units and relative elevation of various sequences are illustrated in Figure2B.
8.4.2 Auger holes
Major changes in sediment characteristics revealed by augering (Table1, Figure2B)
showed that the spring mound was almost pure fine sand to 4m with cemented sand
Runge33-Book.indb 126Runge33-Book.indb 126 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 127
‘nodules’ between 75 and 90cm depth and again between 300cm and 380cm with
the nodules larger and more abundant at greater depth. Between 400cm and 520cm
depth the sandy sediment was light brownish grey (2.5Y 6/2) with yellowish red mot-
tles (5YR 5/6). Dark grey organic-rich horizons were apparent at 410 and 470cm. At
505cm depth the sediments consisted of very fine sand with abundant silt and clay
all cemented by CaCO
. The sand ‘nodules’ were clearly produced by evaporation of
water from soil water containing CaCO
in solution. In the upper part of the profile
this is probably not related to a ground water table, rather to evaporation as rainfall
percolates downward through the sand. However, below 3m the ‘nodules’ and mot-
tling are clearly related to the capillary zone above a water table, perhaps 1–2m below,
and evaporation in this zone precipitated the CaCO
that formed the ‘nodules’. When
first brought to the surface, sediments deeper than 450cm were a green colour that
rapidly fades to light brownish grey after exposure to the atmosphere (see Visser and
Joubert, 1991, p. 126).
The sand dune south of the mound consists of brown, red and then yellowish red
fine sand with some silt and clay to 400cm depth. Below this depth, at 455cm, the fine
sand is strong brown in colour with higher silt content. At 500cm there is cemented,
light yellowish brown very fine sand with silt and clay (Table1). Augering was termi-
nated at 500cm. The sediments from 455 and 500cm were ‘green when first exposed
to the atmosphere but this colour quickly faded after exposure to the atmosphere.
At both the spring mound and the dune the green colour and finer texture of the
Table1. Sediment descriptions for the spring mound and dune auger sites.
Sample Depth (cm) Description
BB-1 97 Greyish brown (10YR 5/2) fine sand with rare
cemented sand ‘nodules’
BB-2 200 Light yellowish brown (2.5YR 6/4) fine sand
BB-3 300 Strong brown (7.5YR 5/8) fine sand with scarce
cemented sand ‘nodules’
BB-4 450 Light brownish grey (2.5Y 6/2) very fine sand,
silt and clay with yellowish red mottles (5 YR
5/6). The sediment is entirely cemented and was
‘green when first exposed to the air.
BBDR-1 100 Strong brown (7.5YR 4/6) fine sand cemented by
BBDR-2 200 Red (2.5YR 4/6) fine sand cemented by CaCO
BBDR-3 300 Red (2.5YR 5/6) fine sand
BBDR-4 400 Yellowish red (5YR 5/8) fine sand
BBDR-4.55 455 Strong brown (7.5YR 5/8) fine sand and silt
partially cemented by CaCO
Slight ‘green
colour when first exposed to the air
BBDR-5m 500 Light yellowish brown (10YR 6/4) very fine sand,
silt and clay. ‘Green colour when first exposed
to the air
Runge33-Book.indb 127Runge33-Book.indb 127 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
128 Andri C. van Aardt et al.
cemented sediments at depth suggests wetter conditions while the overlying, generally
loose aeolian sands record a much drier period.
8.4.3 Backhoe trenches
Trench 4 was excavated near the crest of the primary spring mound. Backhoe
trenches allowed the sediments to be observed in place and a range of other observa-
tions could be made that are not possible in auger samples (Figure2B). Twelve soil
horizons were identified in three sedimentary units (Table 2). An A-C pedon was
recorded in the upper 14cm of sandy aeolian sediment. This sat unconformably on
a buried A horizon that formed the top of the second sedimentary unit. Sedimen-
tary Unit 2 had a 2Ab-2B1-2B2-2B3-2B4 pedon spanning 161cm with thin CaCO
films on ped faces in yellowish-red to brownish yellow sands. Sedimentary Unit 3
is marked by a marked increase in CaCO
accumulations as large nodules grading
down to a rubified weathered horizon. A pale olive coarse sand horizon was found
below and it is tentatively grouped with Sedimentary Unit 3 but it could also be part
of an older sedimentary unit. All of these deposits are well sorted sandy loams to
sand and are differentiated by the degree of pedogenesis, soil carbonate accumula-
tion and weathering.
Trench 1 (Table 3) was excavated on the eastern edge of the spring mound.
Trench 2 was nearby and sampled similar deposits so it is not described. These
deposits provide the broader sedimentary context for the Central Block Excavations.
Trench 1 was excavated from the top of the cut bank lip formed by the removal of
sediment on the east side of the mound. The sediments from Trench 1 are divided
into 13 zones and two sedimentary units. At the top of the trench grasses grow on
sands. This (O-C horizon sequence) sand layer appears to be a very recent aeolian
deposit that caps the entire mound, and sits unconformably on a buried A horizon
inceptisol formed between 23–35cm below the surface and sitting on a C horizon. An
unconformity separates the first from the second sedimentary unit, which is charac-
terized by an A-B-Bg sequence with gleyed sediments in the lower meter of deposits.
Trench 3 (Table4) was excavated on the floor of the drainage adjacent to Excava-
tion Unit 2. The sediments are correlated to Trench 1 following the soil horizon and
sediment unit designations. The sediments begin at the top of Trench 3 with a trun-
cated gleyed B horizon followed by additional B horizons consisting of light yellowish
brown to greyish brown back to light yellowish brown sand shifting to a light olive
grey sandy loam at the bottom of this sedimentary unit. These B horizons sit uncon-
formably on a dark grey sandy loam A horizon recorded as Zone 5. This grades down
into a series of greenish grey sandy loam B horizons which form Sediment Unit 3.
Below this are two B horizons consisting of grey and greenish grey clay loam to sandy
loam soil horizons. The final soil horizon is a greenish grey sandy clay.
Trench 5 was excavated in one of the larger secondary mounds on the east side
of the site near the so-called Pollen Pit. It was placed in an attempt to record the sedi-
ments for the Pollen Pit in more detail but it is not entirely clear to what degree the
sediments vary from one spot to the other. In any case, this profile does provide an
indication of the sediment sequence and variability in these smaller mounds (Table5).
Two depositional units are visible. The upper depositional unit is capped by a yellow-
ish brown sandy loam that may represent fairly recent sediments. The lower sedimen-
tary unit is mostly gleyed with varying amounts of organic matter.
The Pollen Pit was 78cm deep and revealed mainly black “peat” with dark sandy
layers at 10–20cm and 40–45cm. Rootlets were observed down to the bottom.
Runge33-Book.indb 128Runge33-Book.indb 128 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 129
Table2. Profile description for Trench 4.
(cm) Description
1 (A) 0–2 Strong brown (7.5YR 4/6) friable sandy loam, common
rootlets, sod layer, very abrupt smooth lower boundary.
2 (C) 2–14 Brown (7.5YR 5/4) friable sand with fine cross bedding,
aeolian, few rootlets, abrupt smooth lower boundary.
3 (2Ab) 14–30 Dark brown (7.5YR 4/4) slightly firm sandy loam, weak
coarse subangular blocky structure, few roots and rootlets,
clear smooth lower boundary.
4 (2Bb1) 30–70 Yellowish red (7.5YR 4/6) slightly firm sandy loam, weak
medium subangular blocky structure, CaCo
films on ped
faces and root pores, insect burrows filled with grey and
green sand, gradual smooth lower boundary.
5 (2Bb2) 70–110 Strong brown (7.5YR 5/6) sandy loam, weak coarse
subangular blocky structure, few small roots, CaCo
on ped faces and root pores, in lower 10cm few very firm
yellowish red (5YR 5/8) iron oxidized clasts (5cm) that
seem to have formed in place, abrupt wavy lower boundary.
6 (2Bb3) 110–156 Yellow (10YR 7/6) firm sand with few yellowish red (5YR
5/6) mottles, coarser sand than above, medium moderate
subangular to angular blocky structure, very small Mg
flecks, few rootlets, clear smooth lower boundary.
7 (2Bb4) 156–175 Brownish yellow (10YR 6/6) friable sandy loam, few insect
burrows, very few roots, very abrupt highly irregular
sloping (toward south) boundary.
8 (3Bb1) 175–190 Very pale brown to yellow (10YR 7/4 to 7/6) slightly firm fine
sand with 20% slightly firm CaCO
pale brown (10YR 8/3)
nodules formed in globular to vertical patterns, small Mg
flecks, irregular clear to abrupt lower boundary.
9 (3Bb2) 190–228 Pale yellow (2.5YR 5/6) sand with brown (7.5YR 5/6) sand
filling insect burrows, Mg films, 5–7% very firm light
brownish grey (10YR 6/2- in the interior) CaCo
up to 15cm long and 4cm thick that form discontinuous
nodules, clear smooth lower boundary.
10 (3Bb3) 228–245 Slightly firm strong brown (7.5YR 5/6) sand with weak
medium subangular blocky structure, pale olive (5Y 6/3)
sandy loam filling insect burrows extending down from
above zone, very abrupt and irregular lower boundary.
11 (3Bb4) 245–256 Pinkish white to reddish yellow (7.5YR 8/2 to 6/8) very
firm sand CaCo
crust with Mg concretions, very abrupt
irregular lower boundary.
12 (3Bb)
Pale olive (5Y 6/3) coarse sand with strong brown (7.5YR
5/6) mottles and filled insect burrows, abundant Mg flecks,
20–30% insect burrows, 15% larger very firm CaCo
nodules with Mg flecks in nodules, lower boundary not
Runge33-Book.indb 129Runge33-Book.indb 129 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
130 Andri C. van Aardt et al.
8.4.4 Archaeological excavations
In the Central Block, six soil horizons were evident in the west wall of Unit 4 and
7 (Table6). During the excavation informal labels were used and those are presented
in parentheses at the end of the descriptions. These were deposited and altered by
spring activity and have been correlated to the sedimentary units in Trenches 1 and 3
described above. The uppermost 6cm of sediment is a light brown sandy loam with
recent historic artefacts. Below this is a series of organic-rich A horizons separated by
C horizons. The A horizons range from greyish brown to black sands and sandy loams
while the C horizons are pale yellow to light grey sands and sandy loams. Zone 5 is
a spring deposit that extruded through the A horizon that forms Zone 6 and covered
it with sand. There is an unconformity that has truncated the Zone 4A horizon but
the sediments above and below this unconformity are very similar and not different
enough to place into separate sedimentary units.
Sediments in the North Block are divided into two sedimentary units (Table7);
however both are correlated to Sedimentary Unit 1. The uppermost consists of a sandy
cap of recent deposits with historic artefacts sitting unconformably on a buried A
Table3. Trench 1 profile descriptions. Sediments from Trench 2 are similar.
(cm) Description
1 (O) 0–2 Grass and sod layer, very abrupt lower boundary.
2 (C) 2–23 Reddish yellow (7.5YR 6/6) very friable fine medium sand,
common rootlets decreasing slightly down profile, no
structure, few reddish yellow (7.5YR 5.5/6) sand infilled
burrows, southerly sloping very abrupt lower boundary
with occasional insect burrows on boundary.
3 (Ab1) 23–35 Brown (7.5YR 5/4) friable sand, few roots, abrupt smooth
lower boundary.
4 (Cb1) 35–54 Light brown (7.5YR 6/4) slightly firm to friable sand,
common rootlets, no structure, abrupt wavy lower
boundary turbated by insect burrows.
5 (Ab2) 54–72 Yellow brown (10YR 5/4) slightly firm sandy loam, common
rootlets, dark yellow brown (10YR 4/4) clay films and
bodies along root pores, clear smooth lower boundary.
6 (Bb2) 72–98 Brown (10YR 5/3) friable to slightly firm sandy loam with
slightly more silt than above, common rootlets, few
observable insect burrows, very weak coarse subangular
blocky structure, clear smooth lower boundary.
7 (C2) 98–130 Pale brown (10YR 6/3) friable sand, common rootlets, abrupt
wavy lower boundary.
8 (2Ab1) 130–142 Very firm yellowish brown (10YR 5/6) sandy loam with weak
coarse subangular blocky structure, few roots and rootlets,
stone artefacts in zone adjacent to trench, clear smooth
lower boundary.
Runge33-Book.indb 130Runge33-Book.indb 130 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 131
horizon. This Ab horizon is very dark grey clay loam and contained abundant mam-
mal bones (see Faunal and stone tool analysis) and a few prehistoric ceramic sherds and
fractured dolerite cobbles. No chipped stone artefacts were present which make the
assignment of cultural affiliation difficult. The Ab horizon is the second sedimentary
unit and it represents a marsh deposit.
Table4. Trench 3 profile descriptions. Soil horizon designations correlated to Trench 1.
Zone Depth
1 (2Bg4) 0–15 Light yellowish brown (2.5Y 6/4) firm structureless sand,
few medium faint reddish yellow (7.5YR 7/6) mottles,
few distinct fine dark greyish brown (10YR 4.2)
stained root pores, clear smooth lower boundary.
2 (2B5) 15–30 Greyish brown (10YR 5/2) firm sand, few distinct fine
dark greyish brown (10YR 4/2) vertical stained root
pores, clear smooth lower boundary.
3 (2B6) 30–90 Light yellowish brown (2.5Y 6/3) friable massive moist
sand with fine common distinct dary greyish brown
(10YR 4/2 vertical and horizontal stained root pores,
few faint fine light yellowish brown (2.5Y 6/4)
mottles, gradual smooth lower boundary.
4 (2B7) 90–115 light olive grey (5Y 6/2) moist friable sandy loam, few
fine faint yellow (2.5Y 7/6) mottles, few vertical very
dark greyish brown (10YR 3/2) mottles along root
pores, abrupt smooth lower boundary.
5 (3Ab) 115–125/134 Dark grey (N 4/0) sandy loam with few faint medium
greenish grey (5G 6/1) mottles, within this zone three
very dark greyish (N 3/0) 1–2cm thick bands, lower
boundary is clear smooth and sloping toward center
of drainage.
6 (3Bb1) 124/134–165 Greenish grey (5G 6/1) sandy loam with few fine distinct
bluish grey (5B 6/1) mottles with few fine distinct grey
(N 5/0) mottles around root pores, OSL 4sample at
143cm (25.6±1.4 ka), clear smooth lower boundary.
7 (3Bb2) 165–177 Greenish grey (5G 5/1) sandy clay with few medium
distinct grey (N 5/0) mottles, abrupt smooth sloping
lower boundary.
8 (3Btb3) 177–185 Grey (N 5/0) clay loam, with fine weak subangular
blocky structure, clear sloping lower boundary.
9 (3Bb4) 185–205 Greenish grey (5G 5/1) sandy loam, clear smooth
lower boundary.
10 (3Bt5)
Greenish grey (5BG 5/1) sandy clay, moderate medium
prismatic structure with dark brown stains around
root pores, OSL 6sampled at 222cm (76.7±6.2 ka),
lower boundary not observed.
Runge33-Book.indb 131Runge33-Book.indb 131 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
132 Andri C. van Aardt et al.
8.4.5 Chronological information
The age relationships of various sequences are summarized in Figure2B. Radiocarbon
samples were collected at different times by a number of researchers. Three radiocar-
bon laboratories processed the samples: Pretoria, Beta Analytic and Groningen. Sam-
ples analysed at the Pretoria lab used traditional Beta count methods while the other
two labs used AMS techniques. Radiocarbon ages in
C yr BP (
C BP) were calibrated
at the 2σ probability level to calendar years BP (cal BP) using CALIB 7.0 (Stuiver and
Reimer, 1993) and the Southern Hemisphere (SHcal13) atmospheric calibration curve
of Hogg et al. (2013) (Table8). All δ
C values for the dated organic material ranged
from 21.7 ‰ to 27.5 ‰ indicating that C
plant material was dated, most likely
aquatic plant material growing in the marshes associated with spring outflow in the
Table5. Trench 5 profile descriptions.
Zone Depth
1 (A/B) 0–45 Yellowish brown (10YR 5/4) sandy loam, abundant roots
and rootlets that decrease in frequency down profile,
clear smooth lower boundary.
2 (B) 45–60 Brown (10YR 5/3) loose sandy loam, common distinct coarse
yellowish red (5YR 4/6) mottles, small faint greenish grey
(5GY 5/1) mottles, clear wavy lower boundary.
3 (Ab1) 60–80 Dark grey (10YR 4/1) sandy loam with thin 2–3cm very
pale brown (10YR 7/4) sand bedsets of alternating
very pale brown (10YR 7/4) sand, common roots and
rootlets, abrupt smooth lower boundary.
4 (Ab2) 80–127 Black (10YR 2/1sandy loam alternating with dark grey
(10YR 5/1) sandy loam thin (2–10cm thick) beds,
few rootlets, abrupt irregular lower boundary,
5 (2Bg1) 127–138 Light olive grey (5Y 6/2) clay loam, few rootlets, dark grey
(10YR 4/1) stains around root pores, clay films in root
pores, clear smooth lower boundary.
6 (2Bg2) 138–205 Greenish grey (5G 5/1) clay loam, many vertical roots still
decaying, grey (N5/0) root pores, clear smooth lower
7 (2Bg3) 205–220 Greyish green (5G 5/2) sandy clay loam, common rootlets,
grey staining around root pores as above, clear smooth
lower boundary.
8 (2Bg4) 220–230 Gr-10eenish grey (5GY 5.5/1) sandy loam, common
rootlets, clear smooth lower boundary,
9 (2Bg5) 230–260 Light grey (10YR 7/2) sandy loam, common rootlets, water
flowing at lower boundary, clear smooth lower boundary.
10 (2Bg6) 260–295 Greenish grey (5GY 5/1) sandy loam, abundant rootlets,
gradual smooth lower boundary.
11 (2Bg7)
Very dark grey (N 3/0) clayey sand, lower boundary not
Runge33-Book.indb 132Runge33-Book.indb 132 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 133
past. Radiocarbon ages ranged from 490±40
C BP (545–340cal BP) to 7570±40
BP (8412–8205cal BP).
Organic material recovered from levels at 30–35cm and 73–78cm depth in the
Pollen Pit excavated in a secondary peat spring mound included plant fragments as
well as fine organic sediment, both of which were radiocarbon dated. In both cases,
the plant material was significantly younger than the organic sediment (Table8). In
the 30–35cm level the plant material dated to 656–557cal BP and the organic sedi-
ment to 861–1638 cal BP; in the 73–78 cm level the plant material gave an age of
3684–3475cal BP and the organic sediment 7672–7575cal BP. As rootlets have been
observed in the pit section, a likely explanation for this difference is that the plant
material dated consisted of these roots that had penetrated the older deposit, thus
providing a younger age. A second possibility is that the fine organic sediment is old
organic material transported to the site, thus providing an age that would be too old.
Particularly in dry climates organic material can be preserved in the environment for
thousands of years and if this material is introduced into younger deposits it can give
misleading ages for the deposit (e.g. Martin and Johnson, 1995). As 7 of the 9 ages in
Table8 are on organic sediments this would present a problem in ascribing an age to
the pollen recovered from these sediments. In fact, we believe the plant material was
most likely roots and that the organic sediment ages are reliable but further work is
needed to prove this.
OSL ages show that the spring mound sediment sequence dates from recent to
113±13 ka, and the dune sequence from recent to 164±15 ka (Table9). The old
Table6. Profile descriptions for Central Block Excavations in Unit 7 and 4.
Soil horizon designations correlated to Trench 1 and 3.
Zone Depth
1 (A1) 0–6 Light brown to brown mottled (10YR 6/4) loose sandy
loam, common historic artefacts, abrupt lower
boundary, slope wash.
2 (2Ab) 6–16 Friable very dark greyish brown to dark greyish brown
(10YR 3/2 to 4/2) sandy loam with very dark grey
(10YR 3/1) sandy lenses with brown (10YR 4/3) mottles,
abrupt lower boundary, “Upper Black Sand”
3 (2Cb) 16–32 Friable light grey (2.5Y 7/2) sandy loam, abrupt lower
boundary, “White Sand”.
4 (2Ab) 32–50 Slightly firm black (N2.5 /0) sand, abundant lithic
artefacts and bone, very abrupt wavy lower boundary,
unconformity, “Lower Black Sand”.
5 (2Cb) 50–85 Friable pale yellow to light brownish grey (2.5Y 7/3 to
2.5Y 6/2) sand, common dark reed fragments, spring
eye deposits extend down into zone 6in Unit 4 forming
a circular vertical spring vent deposit, very few lithic
artefacts or bones, “Green Sand”.
6 (2Ab) 85–110 Slightly firm greyish brown (2.5Y 5/2) sand,
lower boundary not observed.
Runge33-Book.indb 133Runge33-Book.indb 133 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
134 Andri C. van Aardt et al.
Table7. Profile descriptions for North Block Excavations in Unit 11, west wall profile.
Zone Depth
1 (A) 0–15/20 Greyish brown (10YR 5/2) loose sand lenses separating
thin lenses of finely chopped plant matter (1–2mm
thick), extremely common rootlets, fine moderate platty
structure, abrupt highly irregular lower boundary,
2 (A/B) 15/20–34 Yellowish brown (10YR 5/4) friable sandy loam, common
rootlets and insect burrows, much scattered decomposed
organic matter, abrupt wavy to irregular lower boundary.
3 (B) 34–42 Pale brown (10YR 6/3) sand with common very dark
greyish brown (10YR 3.2) irregular lenses of slightly
firm sandy infilled insect burrows, abrupt smooth to
irregular lower boundary.
4 (Ab) 42–61 Very dark grey (10YR 3/1) firm clay loam, medium
moderate subangular blocky structure, vertical
pedogenic cracks filled with yellowish brown (10YR 5/4)
sand, bone common in upper 6–7cm, lower boundary
not observed.
Table8. Radiocarbon ages dates.
Lab no.
Age cal
BP (2σ)
21.7 490 ± 40
545–340 North Block, Unit 11,
west wall, Black Clay Loam,
23.8 5600 ± 90
6601–6128 Central Block, Unit 8,
west wall, Lower Black Sand,
24.4 5630 ± 45
6465–6291 Central Block, Unit 7,
west wall, Upper Black Sand,
Gr-A 27637
24.46 7570 ± 40
8412–8205 Central Block, Unit 10, Upper
Dark Green Sand, 89–93cm
22.6 2420 ± 70
2718–2209 Central Block, near Unit 9,
Black sand (“peat”),
71–79cm (Theko et al. 2003)
27.5 680 ± 30
656–557 Pollen Pit, 30–35cm,
plant material
24.5 1870 ± 30
1861–1638 Pollen Pit, 30–35cm.
organic sediment
27.0 3380 ± 30
3684–3475 Pollen Pit, 73–78cm,
plant material
26.0 6810 ± 30
7672–7575 Pollen Pit, 73–78cm,
organic sediment
Runge33-Book.indb 134Runge33-Book.indb 134 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 135
Table 9. Equivalent dose (De) and dose rate information, together with OSL ages for the spring mound and dune sediments at Baden-Baden.
UGA06OSL-331 BB-1 97 150–180 13 4.2±0.46 36±7 1.40±0.22 1.94±0.79 0.74±0.1 5±2.5 0.22±0.02 1.36±0.13 3.09±0.45
UGA06OSL-332 BB-2 200 150–180 15 17.64±2.56 55±10 1.60±0.25 2.03±0.87 0.70±0.1 5±2.5 0.19±0.02 1.34±0.13 13.15±2.30
UGA06OSL-330 BB-3 300 150–180 16 51.86±2.48 17±3 1.47±0.11 2.78±0.41 0.71±0.1 5±2.5 0.17±0.02 1.36±0.11 38.28±3.61
UGA06OSL-333 BB-4 450 150–180 19 189.32±10.52 25±4 1.41±0.41 4.75±1.41 0.95±0.1 5±2.5 0.14±0.02 1.67±0.17 113.11±13.27
UGA09OSL-655 BBDR-2 200 180–250 19 51.02±2.18 16±3 1.49±0.22 4.02±0.78 0.72±0.1 5±2.5 0.19±0.02 1.46±0.13 34.92±3.38
UGA09OSL-656 BBDR-4.55 455 180–250 18 221.29±3.88 5±2 0.92±0.20 3.75±0.74 0.81±0.1 5±2.5 0.14±0.02 1.35±0.12 163.95±15.22
QUADRU-OSL#4 OSL#4 Z6 143 180–212 24 38.48 ± 0.59 6±1 1.52±0.23 5.09±0.74 0.84±0.1 17±5 0.20±0.02 1.49±0.07 25.82 ± 1.35
QUADRU-OSL#6 OSL#6 Z10 222 180–212 48 172.33±11.96 19±6 1.64±0.28 6.81±0.90 1.47±0.01 9±3 1.49±0.07 2.25±0.08 76.70 ± 6.18
* UGA06OSL 330 to 333from primary spring mound, UGA09OSL 655 to 656 samples from the dune, and QUADRU OSL #4 and #6 from Trench 3.
UGA samples were dated at the University of Georgia Luminescence Dating Laboratory in Athens, Georgia, U.S.A.; QUADRU samples were dated at the Quaternary
Dating Research Unit (QUADRU) at the CSIR in Pretoria, South Africa. All ages were determined on quartz sand.
Runge33-Book.indb 135Runge33-Book.indb 135 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
136 Andri C. van Aardt et al.
ages are not surprising given dates from nearby Florisbad spring of up to 279±47 ka
(Grün et al., 1996). Trench 3 provided two OSL ages of 26±1.4 ka at 143cm depth
and 77±6 ka at 222cm depth. All of the OSL ages appear reliable although the older
ages have large uncertainties that might suggest some mixing of sands of different ages
at some point during their burial history.
The age data for the spring mound and dune are not sufficiently detailed to
allow major inferences about past conditions; however, some tentative conclusions
are possible. The mound ages indicate that there was about 1m of sand accumulation
between 13.15±2.30 ka and 3.09±0.45 ka or about 100cm in 10 ka and a further
1m of accumulation between 3.09±0.45 ka and present (100cm in 3 ka). In con-
trast, there appears to have been only 1m of accumulation between 38.28±2.3 ka
and 13.15±2.30 ka, or 100cm in 25 ka. The much slower accumulation in this older
period might indicate much wetter conditions in the Baden-Baden area at this time.
In addition, the dune age of 34.92±3.38ka at 200cm corresponds closely with the
38.28±2.3 ka age from the spring mound and both could be evidence of an aeolian
period that preceded a long interval of wetter conditions. The age of 25.82±1.35 ka
for sample OSL#4 from 143cm in Zone 6in Trench 3 also suggests active spring flow
and possibly wetter conditions in the ca. 38–13ka interval (Table9).
In fact Burrough et al. (2009) report three high stands of Palaeolake Mak-
gadikgadi between ca. 39 ka and 17 ka (17.1±1.6 ka, 26.8±1.2 ka, and 38.7±1.8
ka), while Brook et al. (2013) obtained an organic age for Stromatolite MART1-14
from Etosha Pan, of 34.2–32.9cal ka, for a major wet phase in northern Namibia
when the lake filling Etosha Pan was at least 8m above the present floor. Isotope
and pollen evidence from a stalagmite in Wonderwerk Cave near Kuruman in the
Northern Cape suggested wetter conditions than today at ca. 33 ka, from 23 to 17 ka
(Brook et al., 2010). At Soutpan, Peats III and IV record two periods of increased
discharge at the Florisbad spring site at ca. 23.2 ka and 6.3 ka (Butzer, 1984a). At
Deelpan, Liebenbergspan and Alexandersfontein Pan in the western Free State and
Northern Cape, high lakes and spring activity are indicated from ca. 19.2–16.5 ka,
and possibly from 25.1–16.5 ka (Butzer, 1984a, b). Alexandersfontein is today only
an evaporation pan but according to Butzer et al. (1973) it contained a 44km
with an average depth of 8m around 19 ka if ages on carbonate are reliable. Using
water balance calculations and assuming a temperature depression of 6ºC during
the LGM, Butzer et al. (1973) estimated that rainfall must have been about twice
that of today to support such a large lake. OSL ages for sediments at Witpan in the
Northern Cape confirm a wet LGM climate in the South African summer rainfall
zone (Telfer and Thomas, 2007). Pan sediments deposited when water occupied the
pan indicate two major lake periods: one at ca. 32 ka, the other during the LGM at
20 ka (Telfer et al., 2009). There is thus strong evidence that the period 38–13 ka was
wetter than today in the Western Free State.
More rapid deposition of sand on the spring mound between 13.15±2.30 ka and
3.09±0.45 ka, and then between 3.09±0.45 ka and present, suggests drier conditions
overall at Baden-Baden during the Holocene. This agrees with evidence of lunette
development at several pans in the Western Free State at 12–10 ka, 5.5–3 ka, 2–1 ka,
and 0.3–0.07 ka (Holmes et al., 2008) and with evidence of drier conditions in the early
Holocene at Wonderwerk Cave, near Kuruman in the Northern Cape (Brook et al.,
2010). Burrough et al. (2009) report a major high-stand of Palaeolake Makgadikgadi
centred on 8.5±0.2 ka, which is in the interval between periods of pan lunette devel-
opment in the Western Free State at 12–10 ka and 5.5–3 ka reported by Holmes et al.
(2008). In fact, 4 of 7radiocarbon ages on organic sediment (not plant material) for
the Baden-Baden archaeological sites and Pollen Pit fall in the range ca. 8.4–6.1 ka
Runge33-Book.indb 136Runge33-Book.indb 136 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 137
(8412–8205, 7672–7575, 6465–6291 and 6601–6128cal BP) suggesting increased spring
flow in the Western Free State around the time when Palaeolake Makgadikgadi was at
a high level (8.5±0.2 ka). More confidence will only be gained with more dating of the
secondary spring mounds and when evidence from more sequences in the wider region
becomes available to test this hypothesis.
The age of 76.70±6.18 ka for sample OSL#6 at 222cm depth in Zone 10in
Trench 3 is evidence of active spring flow and possibly wetter conditions during MIS 4.
Burrough et al. (2009) report a high lake stand of Palaeolake Makgadikgadi in
Botswana at 64.2±2.0 ka possibly synchronous with wet conditions in the Free State.
However, at 205cm depth there is an unconformity recording either cessation of spring
flow because of reduced groundwater levels or a change in the location of the spring
outlet. As flow had resumed prior to 25.82±1.35 ka, former lower ground water levels
seem more likely. In fact, Riedel et al. (2009) report that fossil gastropods in the now
mainly dry Boteti River indicate that it was permanently flowing through the western
Makgadikgadi Pans at ca. 46 ka, indicating only slightly wetter conditions than today
as Lake Palaeo-Makgadikgadi must have been comparatively small at this time as the
Boteti River flowed for long distances across the exposed bed of this former lake. Rie-
del et al. (2009) indicate that the hard water effect could not be estimated for the shell
and suggest that the true age could be a few thousand years younger. Nevertheless, this
approximate time is when the hiatus between Zones 9 and 10in Trench 3 occurred,
supporting an interpretation of relatively dry conditions.
Perhaps the most significant results of the OSL dating are the apparent very old
ages for the cemented green, silt/clay-rich fine sands at the base of each auger hole. At
the spring mound, the age at 450cm is 113.11±13.27 ka; at the dune, the age at 455cm
is 164±15.2 ka (Table8). Because of the large uncertainties in both of these ages, it
is unclear in which Marine Isotope Stage (MIS) the respective sediments belong. The
spring mound age suggests that the wet period recorded by the clays is MIS 5. In fact,
Burrough et al. (2009) report high levels of Palaeolake Makgadikgadi at 104.6±3.1
ka and 131±11 ka, either of which might correlate with our spring mound age of
113.11±13.27 ka. In contrast, the dune age suggests wetter conditions during MIS 6
and given that MIS 2 and MIS 4may have also been wet it would not be surprising if
MIS 6 was wetalso. Given the height of the clays at the dune site, it is likely that there
was abundant spring flow and a high ground water table at this time and possibly also
a sizeable lake in Annaspan.
The accumulation of sediments on the east side of the primary mound is com-
plex. All of the deposits are covered with a recent sand layer containing historic arte-
facts. The marsh deposit in the Northern Block that contains abundant mammal bone
is the youngest dated deposit. It is also covered by a very recent sand deposit with
historic artefacts.
8.4.6 Faunal and stone tool analysis
The Northern Block, also known as the Bone Bed, was discovered during the last week
of the 2003season. A single 1 × 1m unit (Unit 11) was excavated then and eight addi-
tional units (not all a full 1 × 1m) were excavated in 2006. The only artefacts recovered
in the Northern Block were 20 stone cobbles and four fibre tempered plain sherds
(Table10). A single radiocarbon date (Pta-9195) produced an age of 490±40
(ca. 496cal BP). The most common species in the Bone Bed assemblage is black wilde-
beest, with some hartebeest, springbok, impala and warthog, but there is no evidence
for plains zebra (Table11). We know from historical records and from late Holocene
Runge33-Book.indb 137Runge33-Book.indb 137 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
138 Andri C. van Aardt et al.
Table10. Prehistoric artefacts recovered from excavation blocks at Baden-Baden.
Northern block Central block Southern block Total
Flakes 0 113 288 401
Adzes 0 6 1 7
Edge Modified Flakes 0 19 0 19
Ceramic sherds 4 0 0 4
Grindstones 0 2 0 2
Bored Stone 0 1 0 1
Hammerstones 0 0 1 1
Cores 0 11 1 12
Cobbles 20 23 13 56
Total 24 175 304 503
Table 11. List of identifiable vertebrate remains by excavation block, as recovered in the 2003
and 2006 field seasons. The counts are the number of identified specimens (NISP).
block Total
Tortoise 1 1
Indet. 1 1
Indet. 1 1
Equus quagga subsp. (plains zebra) 1 1
Phacochoerus aethiopicus/
africanus (warthog)
Alcelaphus buselaphus (hartebeest) 7 7
Damaliscus pygargus (blesbok) 2 2
Connochaetes gnou (black wildebeest) 68 6 74
Antidorcas marsupialis (springbok) 11 11
Aepycerus melampus (impala) 3 3
Ovis aries (sheep) 11 11
Bovidae indet.
Large-medium 88 11 99
Small-medium 10 1 11
Small 2 2
Total 212 23 235
Note: Bone preservation in the South Block was poor and only nine bone specimens were recorded,
but these were not identifiable and are not included in this table.
Runge33-Book.indb 138Runge33-Book.indb 138 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 139
sites, such as Deelpan (Scott and Brink, 1992) that plains zebra was a common fau-
nal element in the central interior. Given the abundance of black wildebeest in the
Bone Bed and the known ecological association between plains zebra and wildebeest,
the absence of plains zebra is more likely explained by human prey selection and is
probably not a reflection of palaeoenvironmental conditions. The size of the black
wildebeest suggests a late Holocene age, which accords with the radiocarbon date and
with the presence of domestic sheep in the assemblage. Most of the Northern Block
Bone Bed specimens had been intensively processed and were highly fragmented in the
process of extracting marrow. The artefacts suggest a very narrow range of activities
primarily focused on the fracturing of bone and cooking.
The faunal remains from the Northern Block demonstrate intensive processing
activities by LSA hunter-gatherers. The degree to which bones were broken, so that
even phalanges were split, suggests marrow-extracting or fat rendering activities. Since
chipped stone tools are lacking, there is little evidence for this being the location of the
initial stage of butchering and none to suggest this was the exact locale where the kills
were made, although both were probably very close. The narrow focus on black wil-
debeest suggests that this species was systematically hunted in an organized manner.
The preference for black wildebeest as prey is in contrast to the evidence for Middle
Stone Age hunting behaviour in a similar setting at Florisbad, where there is a focus
on medium-sized antelope, mainly blesbok and springbok and their extinct relatives
(Brink 1987; Brink and Henderson, 2001). The presence of domestic sheep in hunter-
gatherer subsistence is not unusual and has been recorded previously in the Riet River
occupations (Brink et al., 1992).
Of particular interest is the presence of impala in the Bone Bed (Figure 5)
assemblage, since there is uncertainty on its occurrence in historic times in the central
Figure5. A view of the dense clustering of bone specimens in Unit 11in Northern Block.
Runge33-Book.indb 139Runge33-Book.indb 139 8/13/2015 7:42:36 AM8/13/2015 7:42:36 AM
140 Andri C. van Aardt et al.
interior (Skinner and Smithers, 1990). Impala is usually associated with savanna
grassland, where there is some degree of tree cover. However, the modern vegetation
around Baden-Baden is a dry, open grassland (see above), and would not provide
suitable habitat for impala. It is noteworthy that the bone assemblages from Deelpan
do not contain impala (Scott and Brink, 1992; Brink, 2005; Butzer and Oswald, in
press), and its presence in the Bone Bed assemblage may suggest a slight shift in the
local vegetation during recent times with possibly localised patches of tree cover (see
pollen results). The predominance of black wildebeest in the Bone Bed assemblage
would not suggest tree cover on a wide scale, since the species requires open habitat
(Brink, 2005).
In the Central Block seven 1 × 1m units were excavated in 2003 and three sin-
gle isolated units (Units 2, 5 and 8) were placed nearby. LSA stone artefacts, stone
cobbles, and faunal remains were recovered. The Central Block and Unit 8 produced
material dated to the early-mid Holocene. These occupations are interesting because
they are not microlithic Wilton components. The most distinctive tools are adzes and
a single bored stone (Table10). The lack of microlithic tools is unusual as this mate-
rial overlaps chronologically with Wilton occupations, however, similar artefacts have
been found at Voigstpost and at Florisbad dating to the same approximate period.
It is possible that this presents a terminal Lockshoek or some other non-microlithic
tradition (Horowitz et al., 1978; Kuman et al., 1999). The predominance of black
wildebeest in the faunal remains from the central block confirms the focus on this spe-
cies as the preferred prey also earlier during the Holocene.
The bones in the Central Block were generally very poorly preserved compared
to the Northern Block bones, which is probably due to the proximity of the spring
drainage area. At Florisbad poorly preserved bones and stone artefacts were found
in a similar setting close to the drainage of the spring. Bone preservation was mark-
edly better toward the centre of the mound than on the eastern side with the spring
and drainage, suggesting that the quality of bone preservation will improve into
the mound and away from the spring drainage area. As in the Northern Block, the
species composition in the Central Block is dominated by black wildebeest, with
warthog also frequent, however, the evidence of intensive processing is absent in the
Central Block.
The Southern Block also had very poor bone preservation probably because the
occupation was in a fine friable sandy deposit. Most of the artefacts, made on horn-
fels, were flakes but a hammerstone, adze and core were recovered in the excavation.
Immediately below the cut bank where the Southern Block artefacts were exposed we
recovered a cryptocrystalline straight-backed bladelet that must have eroded out of
the deposit. No absolute dates have been obtained on this occupation but the bladelet
and the general degree of pedogenic development suggests a late Holocene age and a
Wilton occupation.
8.4.7 Palynological potential and preliminary results
Pollen samples were taken from the 70cm of organic material in the North Block bone
bed; no pollen was recovered. The lack of pollen could be due to exposure and dry-
ing out but these deposits have potential for phytolith analysis. The secondary spring
mounds, which showed the above-mentioned discrepancy between fractions of organic
material in one mound, have good potential for further palynological research. The
89cm sequence from the Pollen Pit, which consists of peat with some sandy horizons at
15 and 42cm, contained rich pollen assemblages mainly including Poaceae (Figure6).
Runge33-Book.indb 140Runge33-Book.indb 140 8/13/2015 7:42:38 AM8/13/2015 7:42:38 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 141
Figure6. Pollen diagram summarizing the main pollen types in organic layers of the Pollen Pit
in a secondary mound (top), Unit 7in the archaeological excavation in the Central Block (centre),
and from Trench 3 at Baden-Baden (bottom). Dots indicate presence where percentages were too low
to be illustrated effectively as bars.
Runge33-Book.indb 141Runge33-Book.indb 141 8/13/2015 7:42:38 AM8/13/2015 7:42:38 AM
142 Andri C. van Aardt et al.
Middle to late Holocene polleniferous deposits in the Central Block are dark
organic sands or peat that formed part of an energized spring flow. Similar to the
Pollen Pit, the results suggest essentially treeless vegetation with very low num-
bers of woodland pollen, possibly transported a long distance by wind from the
neighbouring Savanna Biome (Figure 1), but this should be supported by more
analyses at higher resolution. Two sections are available, viz., the upper section, P1,
of Theko et al. (2003) with an age of 2420±70
C BP (Pta-8840, ca. 2452cal BP,
Table7) corresponding to an area immediately above Unit 9 (Bousman et al., n.d.
interim report; details of pollen not shown here), and the deeper adjacent excava-
tion Unit 7 with a radiocarbon date of 5630±45
C BP (GrA-25206, ca. 6362cal
BP). Pollen composition in the former consists of a grass-rich assemblage that
overlies grey sand with lower organic contents which are richer in Chenopodiaceae
and Amaranthaceae pollen (Cheno/Ams). The pollen in the latter section (Unit 7)
yielded results with a strong presence of Asteraceae pollen; indicative of a grassy
karroid veld ca. 6362cal BP (Figure6). The high presence of Cheno/Ams in the
lower levels suggests drier conditions. Further work will be needed to establish the
nature of conditions ca. 8.5 ka, which as suggested by the geological results (above),
might have been relatively wet.
A sequence (Pollen 2, Figure3) some 15m south of the spring structure of the
Central Block with less organic material in the sands contained much lower pollen
concentrations (Theko et al., 2003). The Last Glacial period is represented in Trench 3
Figure7. A two-dimensional matrix of data generated by a Detrended Correspondence Analysis based
on the morphometrical dimensions of non-lobate morphotypes recorded in samples BB-1–BB-4,
compared to the morphometrical dimensions of non-lobate morphotypes produced by modern C
and C
grasses. 1=BB-1; 2=BB-2; 3=BB-3; 4=BB-4; Mc=C
mesophytic Microchloa; Ec=C
mesophytic or
xerophytic Eragrostis; Su=C4 xerophytic Stipagrostis and Fs=C
mesophytic Festuca.
Eigenvalues for Axis 1 and Axis 2 are 0.02792 and 0.002662, respectively.
Runge33-Book.indb 142Runge33-Book.indb 142 8/13/2015 7:42:41 AM8/13/2015 7:42:41 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 143
Table 12. Morphometrical analysis of non-lobate short-cell morphotypes from BB-1, BB-2, BB-3 and
BB-4 based on the assessment of three different measurements. The non-lobate category includes
diagnostic short-cell morphotypes commonly recognized as cubical, round, oblong, saddle-shaped or
elongated. NECKX = minimum width of the anterior aspect of the silica body; AX = maximum width
of the anterior aspect of the silica body; AY = maximum length of the anterior aspect of the silica body.
Values for NECKX, AX and AY are expressed in microns (µm).
Sample ID Slide # Ref # Morphotype NECKX AX AY Ratio x/y
BB-1 BB 378 249 Saddle 5.47 4.84 6.04 0.801
252 Saddle 6.36 5.89 8.41 0.700
253 Saddle 7.29 7.64 6.81 1.122
245 Saddle 5.96 6.12 8.2 0.746
BB-2 BB 379 255 Saddle 8.46 8.86 5.79 1.530
256 Saddle 4.95 6.58 5.16 1.275
257 Saddle 5.51 6.03 8.22 0.734
259 Saddle 5.36 6.18 4.91 1.259
BB-3 BB 380 267 Oblong 6.8 7.33 9.29 0.789
269 Oblong 5.02 4.97 10.15 0.490
269 Oblong 5.43 4.49 12.09 0.371
BB-4 BB 382 279 Saddle 3.24 4.11 4.91 0.837
280 Saddle 5.45 5.47 7.72 0.709
284 Saddle 5.91 6.13 5.11 1.200
286 Saddle 4.72 4.13 4.82 0.857
288 Saddle 4.26 4.65 5.82 0.799
289 Saddle 5.93 6.22 4.47 1.391
with organic lenses estimated to be ca. 25 ka old (OSL dating) that contain pollen
types such as Passerina and Stoebe, indicating that fynbos elements occurred, which
are probably associated with cooler grassy conditions. Data from this age are not avail-
able for Florisbad although this site contains limited pollen evidence from a previous
cooler period (possibly MIS 6 or 8) in levels that have yet to be dated effectively (van
Zinderen Bakker, 1989; Scott and Rossouw, 2005). They seem to have been deposited
under wetter conditions as indicated by Ericaceae pollen grains.
8.4.8 Grass short-cell phytolith potential and preliminary results
Based on the OSL ages provided for the spring mound auger site, as well as a multi-
variate analysis (DCA) of non-lobate morphotypes recorded in the samples (Figure7,
Table12 and Plates 1–4), the phytolith data indicate an overall prevalence of warm,
grassy conditions at 3.09±0.45 ka (sample BB-1), 13.2±2.3 ka (sample BB-2) and
again at 113.1±13.3 (sample BB-4). This inference is based on the presence of medially
indented, as well as medially convex saddle-shaped morphotypes (see measurements
in Table 12), which are exclusively produced by chloridoid and aristidoid grasses
adapted to warm and moderately wet to arid conditions (Rossouw, 2009). Conversely,
a predominance of C
-affiliated cubical and oblong—shaped silica bodies at the cost of
saddle-shaped morphotypes at 38.3±3.6 (sample BB-3) concur with the inference for
regionally cooler and wetter conditions around 39 ka and 34 ka as mentioned earlier
in the paper.
Runge33-Book.indb 143Runge33-Book.indb 143 8/13/2015 7:42:41 AM8/13/2015 7:42:41 AM
144 Andri C. van Aardt et al.
Plate 1. Spring Mound Auger Site sample BB-1. (A) cuneiform bulliform silica body; (B) short-necked
bilobate, anterior and side view; (C) medially indented saddle morphotype, anterior view; (D) medially
convex saddle morphotype, posterior view; (E) medially convex saddle morphotype, anterior view;
(F) medially indented saddle morphotype, anterior view.
Runge33-Book.indb 144Runge33-Book.indb 144 8/13/2015 7:42:41 AM8/13/2015 7:42:41 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 145
Plate 2. Spring Mound Auger Site sample BB-2. (A) cuneiform bulliform silica body;
(B) short-necked bilobate, anterior view; (C–F) medially indented saddle morphotype, anterior view.
Runge33-Book.indb 145Runge33-Book.indb 145 8/13/2015 7:42:43 AM8/13/2015 7:42:43 AM
146 Andri C. van Aardt et al.
Plate 3. Mound Auger Site sample BB-3. (A–D) trapezoid silica body, anterior and side view;
(E) oblong silica body, anterior view.
Runge33-Book.indb 146Runge33-Book.indb 146 8/13/2015 7:42:46 AM8/13/2015 7:42:46 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 147
Plate 4. Spring Mound Auger Site sample BB-4. (A) cuneiform bulliform silica body;
(B) globular silica body with rugose surface (C) short-necked bilobate, anterior view;
(D–H) medially indented saddle morphotype, anterior view.
Runge33-Book.indb 147Runge33-Book.indb 147 8/13/2015 7:42:49 AM8/13/2015 7:42:49 AM
148 Andri C. van Aardt et al.
The OSL ages for the spring mound, dune and Trench 3sediments are a clear indi-
cation that Baden-Baden could provide a very long palaeoenvironmental record for
central South Africa associated with human occupations. In both auger holes, auger-
ing was stopped when cemented, green, silt/clay-rich sand was encountered, and yet
deeper, and older, samples could be obtained from both sites. A future research pro-
gram linking spring activity at Baden-Baden to conditions in Annaspan could provide
one of the longest records of climate change for central South Africa, particularly
if it was supplemented by pollen and phytolith studies. The oldest OSL age we have
obtained so far is 164±15 ka and we have only examined the top 5m of sediments at
the site. There is a strong possibility that the sediments at Baden-Baden could provide
information beyond MIS 6 and even to MIS 8.
The middle to late Holocene deposits of the secondary spring mounds show good
potential for palaeoenvironmental reconstruction based on palynological research in
the secondary spring mounds provided that the discrepancy in dating between plant
material and fine organic fractions can be resolved. Microscopic charcoal data (not
shown here) also indicate potential for investigating past burning in the area. In the
Central Block palynological potential is marginal since some levels do not contain pol-
len. It will therefore be necessary in future also to rely on phytolith analysis in these
deposits as phytoliths have been shown to be useful in the semiarid Free State environ-
ments (Scott and Rossouw, 2005). Overall the pollen sequences from the Baden-Baden
Holocene deposits indicate changing plant communities that reflect alternating vari-
ations in moisture conditions—parallel to the results from Florisbad and Deelpan in
the western Free State (Scott, 1988; Scott and Brink, 1992; Scott and Nyakale, 2002;
Butzer and Oswald, in press) that can be integrated to improve the Holocene palaeo-
environmental reconstruction of the region. A significant finding at Baden-Baden is
that it provided pollen of last “glacial” age in Trench 3 at the base of the primary
mound. The levels containing fynbos and grass pollen of this age are limited to a few
horizons. Nevertheless it makes an important contribution about vegetation condi-
tions in the relatively dry central Free State because coeval levels from the other major
site in the region, Florisbad, are absent. Further palynological research and dating at
Baden-Baden, promises to fill some spatial and temporal gaps in the long-term palaeo-
environmental history of the interior region of South Africa by providing a more
complete picture of changing climates.
We would like to thank Mr Nellis Nel, owner of Baden-Baden, for access to the
site, logistical assistance, historical notes, and his generosity. We especially wish to
thank Sarel Greyling for his help in during the archaeological excavations. The town
of Dealesville provided the backhoe and both Hanlie Greyling and “Witkop” are
thanked for their generosity to the archaeological crew. The pollen work is based on
the research supported by the National Research Foundation. Any opinion, finding
and conclusion or recommendation expressed in this material is that of the authors
and the NRF does not accept any liability in this regard. Brook’s research was sup-
ported by NSF grant NSF-0725090. Bousman and Brink’s research was funded by
a grant from the Leakey Foundation, Texas State University, the National Research
Foundation and the National Museum. Bokang Theko, Petrus Chakane, Frank Neu-
mann, Abel Dichakane, Peter Mdala, Willem Nduma, Lebohang Nyenye, Adam
Runge33-Book.indb 148Runge33-Book.indb 148 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 149
Thibeletsa, Koos Mzondi, Zoe Henderson, Sharon Holt, Duma Mbolekwa, Bonny
Nduma, Peter Ntulini, Isaac Thapo, Gary Trower and helpers from the National
Museum, Bloemfontein, assisted with fieldwork. Thanks to Tania R. Marshall for
copies of articles.
Aitken, M.J., 1985, Thermoluminescence Dating, (Orlando: Academic Press).
Aitken, M.J., 1998, An Introduction to Optical Dating, (Oxford: Oxford University
Beaumont, P.B., 2004, Wonderwerk Cave. In The Northern Cape: Some Key Sites,
edited by Morris, D. and Beaumont, P.B. (Kimberley: Archaeology, McGregor
Museum), pp. 31–36.
Bøtter-Jensen, L., Duller, G.A.T., Murray, A.S. and Banerjee, D., 1999, Blue light
emitting diodes for optical stimulation of quartz in retrospective dosimetry and dat-
ing. Radiation Protection Dosimetry, 84, pp. 335–340.
Bousman, C.B, Brink, J.S. and Weinstein, A., n.d., L. S. B. Leakey Foundation Grant
Interim Report: A Search for Modern Human Behavior at Baden-Baden, South
Africa (Unpublished report).
Bousman, C.B. and Brink, J.S., 2008, Holocene Geoarchaeology and Archaeology at
Baden-Baden; Symposium: New views on ancient Africa. Papers in honour of C.
Garth Sampson. Presentation, Society for American Archaeology, Vancouver.
Bousman, C.B. and Brink, J.S., 2012, Baden-Baden—A Smithfield Antelope Process-
ing Site in the western Free State, South Africa. Presentation. Society of Africanist
Archaeologists, Toronto, June 2012.
Brink, J.S., 1987, The archaeozoology of Florisbad, Orange Free State. Memoirs of the
National Museum, Bloemfontein, 24, pp. 1–151.
Brink, J.S., 2005, The evolution of the black wildebeest (Connochaetes gnou) and
modern large mammal faunas of central southern Africa. DPhil dissertation, Uni-
versity of Stellenbosch.
Brink, J.S. and Henderson, Z.L., 2001, A high-resolution last interglacial MSA
horizon at Florisbad in the context of other open-air occurences in the central
interior of southern Africa: an interim statement. In Middle Palaeolithic and
Middle Stone Age settlement systems, edited by Conard, N. (Tuebingen: Kerns-
Verlag), pp. 1–20.
Brink, J.S., Dreyer, J.J.B. and Loubser, J., 1992, Rescue excavations at Pramberg,
Jacobsdal, south-western Orange Free State. Southern African Field Archaeology, 1,
pp. 54–60.
Brook, G.A., Scott, L., Railsback, L.B. and Goddard, E.A., 2010, A 35 ka pollen and
isotope record of environmental change along the southern margin of the Kalahari
from a stalagmite and animal dung deposits in Wonderwerk Cave, South Africa.
Journal of Arid Lands, 74, pp. 870–884.
Brook, G.A., Cherkinsky, A., Railsback, L.B., Marais, E. and Hipondoka, M.H.T.,
2013, Radiocarbon dating of organic residue and carbonate in stromatolites from
Etosha Pan, Namibia: the radiocarbon reservoir effect, correction of published car-
bonate ages, and evidence of a >8m deep lake during the Late Pleistocene. Radio-
carbon, 55, pp. 1156–1163.
Burrough, S.L., Thomas, D.S.G. and Bailey, R.M., 2009, Mega-Lake in the Kalahari:
A Late Pleistocene record of the Palaeolake Makgadikgadi system. Quaternary Sci-
ence Reviews, 28, pp. 1392–1411.
Runge33-Book.indb 149Runge33-Book.indb 149 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
150 Andri C. van Aardt et al.
Butzer, K.W., Fock, G.J., Stuckenrath, R. and Zilch, A., 1973, Palaeohydrology of Late
Pleistocene Lake, Alexanderfontein, Kimberley, South Africa. Nature, 243, pp. 328–330.
Butzer, K.W., 1984a, Late Quaternary environments in South Africa. In Late Caino-
zoic Palaeoclimates of the Southern Hemisphere, edited by Vogel, J.C. (Rotterdam:
Balkema), pp. 235–264.
Butzer, K.W., 1984b, Archaeogeology and Quaternary environment in the interior of
southern Africa. In Southern African Prehistory and Palaeoenvironments, edited by
Klein, R.G. (Rotterdam: Balkema), pp. 1–64.
Butzer, K.W. and Oswald, J.F., in press, Dry Lakes or Pans of the western Free State,
South Africa: Environmental History and possible Human Impact at Deelpan. Pal-
aeoecology of Africa, 33.
Duller, G.A.T., 1999, Luminescence Analyst computer programme V2.18. Department
of Geography and Environmental Sciences, University of Wales, Aberystwyth.
Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H. and Olley, J.M., 1999,
Optical dating of single and multiple grains of quartz from Jinmium rock shelter,
northern Australia, part 1, Experimental design and statistical models. Archaeom-
etry, 41, pp. 339–364.
Geldenhuys, J.N., 1982, Classification of the pans of the western Orange Free State
according to vegetation structure, with reference to avifaunal communities. South
African Journal of Wildlife Research, 12, pp. 55–62.
Gil-Romera, G., Neumann, F.H., Scott, L., Sevilla-Callejo, M. and Fernández-Jalvo,
Y., 2014, Pollen taphonomy from hyaena scats and coprolites: preservation and
quantitative differences. Journal of Archaeological Science, 46, pp. 89–95.
Grobler, N.J., and Loock, J.C., 1988, Morphological development of the Florisbad
deposit. Palaeoecology of Africa, 19, pp. 163–168.
Grün, R., Brink, J.S., Spoons, N.A., Taylor, L., Stringer, C.B., Franciscus, R.G. and
Murray, A.S., 1996, Direct dating of Florisbad hominid. Nature, 382, pp. 500–501.
Hogg, A.G., Hua, Q., Blackwell, P.G., Buck, C.E., Guildersno, T.P., Heaton, T.J., Niu,
M., Palmer, J.G., Reimer, P.J.; Reimer, R.W., Turney, C.S.M. and Zimmerman, S.,
2013, SHCal13 Southern Hemisphere Calibration, 0–50,000 years cal BP. Radiocar-
bon, 55(4), pp. 1889–1903.
Holmes, P.J., Bateman, M.B., Thomas, D.S.G., Telfer, M.W., Barker, C.H. and Law-
son, M.P., 2008, A Holocene-late Pleistocene aeolian record from lunette dunes of
the western Free State panfield, South Africa. The Holocene, 18(8), pp. 1193–1205.
Horowitz, A, Sampson, C.G., Scott, L. and Vogel, J. C., 1978, Analysis of the Voigtspost
Site, O. F. S., South Africa. South African Archaeological Bulletin, 33, pp. 152–159.
Kuman, K.M., Inbar, I. and Clark, R.J., 1999, Palaeoenvironments and cultural
sequence of the Florisbad Middle Stone Age hominid site, South Africa. Journal of
Archaeological Science, 26, pp. 1409–1425.
Markey, B.G., Bøtter-Jensen, L. and Duller, G.A.T., 1997, A new flexible system for
measuring thermally and optically stimulated luminescence. Radiation Measure-
ments, 27, pp. 83–90.
Marshall, T.R., 1987, Morphotectonic analysis of the Wesslesbron panveld. South
African Journal of Geology, 90(3), pp. 209–218.
Marshall, T.R., 1988, The origin of the pans of the western Orange Free State, a mor-
pholtectonic study of the palaeo-Kimberley River. Palaeoecology of Africa, 19, pp.
Martin, C.W. and Johnson, W.C., 1995, Variation in radiocarbon ages of soil organic
matter fractions from late Quaternary buried soils. Quaternary Research, 43, pp.
Runge33-Book.indb 150Runge33-Book.indb 150 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
Baden-Baden: A fossil spring site in the western Free State, South Africa 151
Mucina, L. and Rutherford, M.C., 2006, The vegetation of South Africa, Lesotho and
Swaziland (Pretoria: South African National Biodiversity Institute, Strelitzia 19).
Murray, A.S. and Wintle, A.G., 2000, Luminescence dating of quartz using improved
single-aliquot regenerative-dose protocol. Radiation Measurement, 32, pp. 57–73.
North American Commission on Stratigraphic Nomenclature, 2005, North American
Stratigraphic Code. The American Association of Petroleum Geologists Bulletin, 89,
11, pp 1547–1591.
Olson, G.W., 1981, Soils and environment, a guide to soils surveys and their applica-
tions (New York: Chapman and Hall).
Prescott, J.R. and Hutton, J.T., 1994, Cosmic ray contributions to dose rates for lumi-
nescence and ESR dating: large depths and long-term time variations. Radiation
Measurements, 23, pp. 497–500.
Reineck, H.-E. and Singh, I.B., 1975, Depositional sedimentary environments, with
reference to terrigenous clastics (Berlin: Springer-Verlag).
Riedel, F., von Rintelen, T., Erhardt, S. and Kossler, A. 2009, A fossil Potadoma (Gas-
tropoda: Pachychilidae) from Pleistocene central Kalahari fluvio-lacustrine sedi-
ments. Hydrobiologia, 636, pp. 493–498.
Rossouw, L., 2009, The application of fossil grass-phytolith analysis in the reconstruc-
tion of late Cainozoic environments in the South African interior. PhD Dissertation.
University of the Free State. Bloemfontein.
Schulze, B.R., 1972, South Africa. In Climates of Africa edited by Griffiths, J.F. in
World Survey of Climatology Volume 10, edited by Landsberg, H.E. (Amsterdam:
Elsevier Publishing Company), pp. 501–586.
Schulze, R.E., 1997, Climate. In Vegetation of Southern Africa, edited by Cowling,
R.M., Richardson, D.M. and Pierce S.M. (Cambridge University Press), pp. 21–42
Scott, L., 1987, Pollen Analysis of Hyena Coprolites and Sediments fron Equus Cave,
Taung, Southern Kalahari (South Africa). Quaternary Research, 28, pp. 144–156.
Scott, L., 1988, Holocene environmental change at western Orange Free State pans,
South Africa, inferred from pollen analysis. Palaeoecology of Africa and the sur-
rounding Islands, 19, pp. 109–118.
Scott, L. and Brink, J.S., 1992, Quaternary palaeoenvironments of pans in central
South Africa: Palynological and palaeontological evidence. South African Geogra-
pher 19 (1/2), pp. 22–34.
Scott, L. and Nyakale, M., 2002, Pollen indications of Holocene palaeoenvironments
at Florisbad spring in the central Free State, South Africa. The Holocene, 12(4), pp.
Scott, L. and Rossouw, L., 2005, Reassessment of Botanical Evidence for Palaeoenvi-
ronments at Florisbad, South Africa. South African Archaeological Bulletin, 60(182),
pp. 96–102.
Scott, L. and Thackeray, J.F., 2014, Palynology of Holocene deposits in Excavation 1
at Wonderwerk Cave, Northern Cape (South Africa). African Archaeological Review
(in press).
Skinner, J. and Smithers, R.H., 1990, The mammals of the southern African subregion.
(Pretoria: University of Pretoria, 2nd edition).
Stuiver, M. and Reimer, P.J., 1993, Extended
C database and revised CALIB radio-
carbon calibration program. Radiocarbon, 35, pp. 215–230.
Telfer, M.W. and Thomas, D.S.G., 2007, Late Quaternary linear dune accumulation
and chronostratigraphy of the southwestern Kalahari: implications for aeolian pal-
aeoclimatic reconstructions and predictions of future dynamics. Quaternary Science
Reviews, 26, pp. 2617–2630.
Runge33-Book.indb 151Runge33-Book.indb 151 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
152 Andri C. van Aardt et al.
Telfer, M.W., Thomas, D.S.G., Parker, A.G., Walkington, H. and Finch, A.A., 2009,
The palaeoenvironmental potential form a southeastern Kalahari pan floor. Palaeo-
geography, Palaeoclimatology, Palaeoecology, 273, pp. 50–60.
Theko, B., Scott, L. and Bousman, C.B., 2003, Preliminary Report on the Palynology of
Baden-Baden Spring Deposits in the Free State, South Africa. SASQUA Conference,
University of the Witwatersrand, Johannesburg.
Van Zinderen Bakker, E.M., 1989, Middle Stone Age palaeoenvironments at Florisbad
(South Africa). Palaeoecology of Africa and the surrounding Islands, 20, pp. 133–154.
Visser, J.N.J. and Joubert, A., 1991, Cyclicity on the Late Pleistocene to Holocene
spring and lacustrine deposits at Florisbad, Orange Free State. South African Journal
of Geology, 94, pp. 123–131.
Runge33-Book.indb 152Runge33-Book.indb 152 8/13/2015 7:42:57 AM8/13/2015 7:42:57 AM
Evidence for climate variability in the southern African monsoon region (SAMR) is limited by a spatially and temporally discontinuous palaeoclimatic dataset. We describe a 6680 year long, largely sub-decadal resolution δ¹⁵N record from a rock hyrax midden from southeastern Africa. The results provide a detailed reconstruction of regional hydroclimates since the beginning of the mid-Holocene. A long-term – albeit subtle – increase in humidity consistent with precessional forcing is observed, but the record is dominated by a strong ∼1750-yr cycle, a signal that is shared with other SAMR records. Considered in their regional context, these data suggest that changes coincident with the termination of the African Humid Period at ∼5500 cal BP do not express the abrupt transition observed in some records from the northern African tropics. Rather they indicate gradual changes, as observed at peri-equatorial sites. Notably, however, eastern and western subregions of the SAMR experience a rapid phase shift beginning ∼5500 cal BP, with initially in-phase hydroclimate anomalies transitioning to the establishment of a strong east-west dipole. This likely reflects a coeval strengthening of the Southeast Atlantic trade winds and decreased atmospheric pressure in southeast Africa, factors associated with increasing (decreasing) austral (boreal) summer insolation. The results highlight the distinct nature of southern African responses across this key period of African climate history.
Full-text available
Marine Isotope Stage (MIS) 5 and 4 are periods of major cultural innovations in the Middle Stone Age (MSA) of southern Africa. While extensive data is available for the coast, far less is known about the interior, in particular its central plateau. This is likely due to the large geographic extent of this area and a general paucity of caves and rock shelters that can provide long stratigraphic sequences and environmental records. The lack of information and systematic research has hindered our understanding of regional variation and patterns of human dispersal within the subcontinent. Our research at the open-air MSA site of Lovedale situated on the Modder River addresses this issue. Using sediment micromorphology, infrared spectroscopy of bones and sediments, phytolith and faunal analyses, as well as luminescence dating, we have reconstructed the evolution of paleoenvironments in this region at specific points over the last ∼80,000 years. Our results help contextualize human occupation and hunting strategies associated with a pre-Howiesons Poort technology that occurred in a wetland environment during a short-lived warm, dry period dated to ∼70 ka. These results show that humans settled the grasslands of the central interior at the onset of MIS 4 and confirm the importance of wetlands in human subsistence strategies, especially in times of climatic stress.
When considering the pan-African process of human evolution in the Pleistocene it is important to review our understanding of climate on a regional scale, including heterogeneity due to regional variability in the relative strengths of forcing mechanisms. Research in the Kuruman Hills at the southern edge of the Kalahari in South Africa has recently provided new insights into the complexity and distinctiveness of southern African palaeoclimate. Here we compare these records, obtained from Kathu Pan, Wonderwerk Cave, and Mamatwan Mine, focusing on the presence of water bodies from c. 2 million years before present to the end of the Pleistocene. Through the synthesis of multiple proxies we create a picture of localized water availability as it would have affected hominin adaptation. These records are then framed within a larger discussion of regional climate and environmental change. The emerging record from the southern Kalahari suggests that the distribution of zones of rainfall seasonality varied significantly throughout the Pleistocene and a hominin presence is particularly found in association with wetter phases which supported lakes, springs and pans.
We have reassessed the palynological record of Equus Cave in the Savanna Biome of the southern Kalahari, one of the longest Late Quaternary pollen records for the semi-arid central interior of South Africa. We combined published pollen results from the cave, derived from hyena coprolites and the rubified deposits in which they occur, into a single sequence. By re-considering the chronology of this sequence, we critically evaluated the palaeoenvironmental record for the site. We compared the pollen evidence from Equus Cave to that from the longer Wonderwerk Cave records (stalagmite, sediments and dung), also located in the Savanna Biome. Then, we contrasted Equus and Wonderwerk records with other previously published pollen sequences derived from a range of sources from several sites in central South Africa. These sites follow a broad northwest to southeast transect of c. 500 km through the Grassland and Nama Karoo Biomes of the Free State and Eastern Cape. Applying Principal Components Analysis to the pollen data, we derived climatic signals at a regional scale to refine reconstructions of Late Quaternary changes for central South Africa.
Full-text available
The interior regions of South Africa have had less attention devoted to archaeological research than coastal regions, and palaeoenvironmental studies are also more limited. As such, little is known about the interaction between human behaviours and past environments in these semi-arid regions. Here, we present an archaeobotanical and mineralogical study from the Middle Stone Age site of Mwulu’s Cave, Limpopo Province. Our study shows the importance of using taphonomical approaches prior to interpreting archaeobotanical assemblages, while provides with novel information on the plants used by ancient inhabitants of Mwulu’s. The grass phytolith composition is of environmental significance, where a shift from C4 Panicoideae to C3 grasses is observed in the last occupation event. This tentatively suggests a shift in rainfall regime, from summer rainfall conditions to an increase in winter rain, during Marine Isotope Stage 5b in the Polokwane region, or a decrease in rainfall seasonality. Although we are unable to chronostratigraphically associate this change in the plant composition, our study adds evidence in support of previous propositions for an expansion of the winter rainfall zone into the interior regions of South Africa.
Thermoluminescence dating: By M. J. Aitken.
The first edition of this book was published in 1973 and was abstracted as 74E/1386. Emphasis is placed on contemporary environments and sediments found in them and few examples from ancient sediments are introduced with care since these are so infrequently reinterpreted. The book is divided into 2 principal sections: framework structures and textures; modern environments.-K.Clayton
Although the Florisbad spring site in the central Free State is famous for its unique evidence of Pleistocene humans and fauna, it also provides microfossil evidence, especially pollen, of long-term vegetation and climate change. This brief account of the palaeobotanical evidence provides a summary of the pioneering work of the late Eduard M. van Zinderen Bakkerat Florisbad. The aim is to briefly review some past environmental interpretations, identify inherent site problems, and to point out the direction of future palynological research. It is suggested that these studies will greatly benefit from fossil plant phytolith investigations and a more in-depth pollen analysis (especially of the Asteraceae pollen component) that will link pollen data more effectively to palaeontological and archaeological evidence.