Content uploaded by Genevieve Dewar
Author content
All content in this area was uploaded by Genevieve Dewar on Oct 11, 2017
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=uica20
Download by: [Brian Stewart] Date: 01 September 2016, At: 05:08
The Journal of Island and Coastal Archaeology
ISSN: 1556-4894 (Print) 1556-1828 (Online) Journal homepage: http://www.tandfonline.com/loi/uica20
Early Maritime Desert Dwellers in Namaqualand,
South Africa: A Holocene Perspective on
Pleistocene Peopling
Genevieve Dewar & Brian A. Stewart
To cite this article: Genevieve Dewar & Brian A. Stewart (2016): Early Maritime Desert Dwellers
in Namaqualand, South Africa: A Holocene Perspective on Pleistocene Peopling, The Journal of
Island and Coastal Archaeology, DOI: 10.1080/15564894.2016.1216476
To link to this article: http://dx.doi.org/10.1080/15564894.2016.1216476
Published online: 31 Aug 2016.
Submit your article to this journal
View related articles
View Crossmark data
The Journal of Island & Coastal Archaeology, 00:1–21, 2016
Copyright © 2016 Taylor & Francis Group, LLC
ISSN: 1556-4894 print / 1556-1828 online
DOI: 10.1080/15564894.2016.1216476
Early Maritime Desert Dwellers in
Namaqualand, South Africa: A Holocene
Perspective on Pleistocene Peopling
Genevieve Dewar1,2 and Brian A. Stewart2,3
1Department of Anthropology, University of Toronto Scarborough, Toronto,
Ontario, Canada
2Rock Art Research Institute, University of the Witwatersrand, Johannesburg,
South Africa
3Museum of Anthropological Archaeology, University of Michigan, Ann Arbor,
Michigan, USA
ABSTRACT
South Africa’s northern Namaqualand coastal desert is the southern
extension of the Namib. Today, this region is semi-desert with patchy
subsistence resources and scarce, unpredictable rainfall. Yet this an-
cient desert landscape possesses residues of human activity stretching
back into the Middle Pleistocene, evidenced by heavily weathered sur-
face finds, including handaxes and Victoria West cores. Such old finds
in so harsh an environment raise important questions: how do human
movements into this area relate to local palaeoenvironmental changes,
and how has this relationship changed through time? While no dated
Middle Pleistocene sites presently exist to reconstruct the earliest ho-
minin dispersals, several late Pleistocene sites now have chronostrati-
graphic sequences that can be brought to bear on these questions. This
article presents chronological and subsistence-settlement data for one
such site, Spitzkloof A Rockshelter in northern Namaqualand’s rugged
Richtersveld. Humans are shown to have visited the site very sporad-
ically between ∼50,000 and 17,000 cal BP. Unlike most of the sub-
continent, the most intensive occupations occur during early Marine
Isotope Stage 2, when multiple proxies suggest enhanced humidity as-
sociated with intensified winter rainfall. We examine these data using
the region’s better-developed Holocene archaeological record to create
predictions about the earliest coastal desert dwellers.
Received 3 November 2015; accepted 14 July 2016.
Address correspondence to Genevive Dewar, Department of Anthropology, University of Toronto Scar-
borough, 1265 Military Trail, Toronto, Ontario M1C 1A4, Canada. E-mail: genevieve.dewar@utoronto.ca
Color versions of one or more of the figures in the article can be found online at
http://www.tandfonline.com/uica.
1
Genevieve Dewar and Brian A. Stewart
Keywords coastal desert, Later Stone Age, Middle Stone Age, MIS 2, MIS 3, Namaqualand,
South Africa
INTRODUCTION
Relative to the Mediterranean-like coastal
ecosystems of the southern and western
Cape, southern Africa’s drier coastal zones
remain archaeologically under-investigated.
Moving north from the Cape along south-
ern Africa’s west coast, environments be-
come increasingly water-stressed, culminat-
ing in the hyper-arid Namib Desert, the sand
seas of which drape the Skeleton Coast. Be-
tween the Cape and the Namib lies the Na-
maqualand coastal desert of northwestern
South Africa, a semi-arid landscape of dy-
namic coastal sand dune systems flanking
ancient, heavily degraded mountain chains.
Like other coastal deserts worldwide, Na-
maqualand would have presented prehis-
toric foragers (and, in the last 2,000 years,
herders or herder-foragers) with a mix of ter-
restrial resources that were unpredictable in
space and time and much more stable littoral
and marine foods along the adjacent shore-
line. Exploitation of the latter has generated
what are by far Namaqualand’s most visible
archaeological signatures—hundreds (and
likely thousands) of coastal shell middens in
Holocene dunefields, dozens of which have
been documented in the course of cultural
resource management mitigations of the re-
gion’s voracious open-cast diamond mining
operations.
By contrast, virtually nothing is known
about Namaqualand’s inland archaeologi-
cal record, which includes diagnostic arti-
facts suggesting human occupation of this
parched region extended deep into the Mid-
dle or perhaps Lower Pleistocene (Dewar
and Stewart 2016). Considering Africa’s key
role in charting human behavioral evolution,
an understanding of the timing, motives, and
modes of human dispersals into challenging
environments like Namaqualand can provide
us with fundamental insights into biogeo-
graphical processes that complement other
measures of adaptive change, such as techno-
logical and subsistence innovations and the
proliferation of non-utilitarian technologies
(e.g., Clark and Kandel 2013; d’Errico et al.
2005; Henshilwood et al. 2002, 2009; Texier
et al. 2010; Wadley 2010, 2013; Wadley et al.
2009). Our research goal is to identify the
tempo, nature, and causes of early human
occupation in the region. Related to this,
what was the environment and thus the avail-
able resource like, and which biocultural in-
novations allowed people to adapt to that
niche? In this article, we use the region’s
better developed Holocene archaeological
record to create predictions about these
questions as they apply to early coastal desert
dwellers. We evaluate these predictions us-
ing palaeoenvironmental and subsistence-
settlement data derived from faunal assem-
blages at Spitzkloof A, a late Pleistocene rock-
shelter sequence in inland Namaqualand on
which our work has focused (Dewar and
Stewart 2012, 2016).
NAMAQUALAND: LANDSCAPE AND
ENVIRONMENT
The Namaqualand coastal desert is known
for its spectacularly strange and desolate
landscapes, its extraordinarily diverse plant
and animal life, and, though today the lo-
cal inhabitants are generally impoverished,
its immense mineral wealth. The majority
of the region’s inhabitants live in remote
towns built by the open-cast diamond min-
ing industry or small villages, sedentary cen-
ters of the region’s previously transhumant
Nama-speaking pastoralist population. Na-
maqualand’s rocky hinterland is a biodiver-
sity hotspot for low-lying succulents and
small reptiles (Desmet 2007), while open-
air shell middens in mostly Holocene-age
dune systems cover the coastline between
the scars of diamond pits scoured into pale-
oriver channels.
Today, minimum temperatures in win-
ter fall below 0◦C and dry summer heat
reaches above 40◦C when evapotranspira-
tion exceeds precipitation and the only
source of water are coastal fogs with
2 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
Figure 1. Map of South Africa showing the geographic location of the rockshelters Spitzkloof A,
Apollo 11, and V003. The boundaries of the Winter Rainfall Zone, Summer Rainfall Zone,
and the extent of the continental shelf during the last glacial maximum (LGM) are shown
as lines.
relative humidity levels of between 70% and
100%. Rain is infrequent, but seasonally pre-
dictable in that more than 66% of the an-
nual precipitation falls in the austral win-
ter months. However, precisely where rain
falls on the landscape is extremely erratic.
Mean annual rainfall is low, ranging from
150 mm at the Olifants River, which de-
lineates the region’s southern boundary, to
50 mm at the Orange River in the north
(Figure 1). This range of precipitation val-
ues defines the region today as semi-arid. Na-
maqualand can thus be seen as a transitional
zone between the hyper-arid Namib Desert
to the north and South Africa’s relatively
well-watered Western Cape Province to the
south.
Namaqualand’s aridity is exacerbated by
the meeting of frigid upwelled ocean waters
with the hot, dry air of the subsiding sub-
tropical cell. Cold upwelling waters of the
Atlantic Benguela Current produce sea sur-
face temperatures ranging from 11◦Cto17
◦C
(Eitel 2005). This cools a thin layer of sur-
face air, impeding convection and prevent-
ing precipitation. The result is a rain shadow
effect (Mucina and Rutherford 2006) extend-
ing some 30 km inland, where fine-grained
sand deposits and quartz gravel plains even-
tually meet the foothills of the Stinkfontein
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 3
Genevieve Dewar and Brian A. Stewart
Mountain group and precipitation can reach
250 mm per annum. As the Benguela Cur-
rent was established in the late Miocene
7 to 10 mya (Deacon and Lancaster 1988;
Eitel 2005), this coastal desert has likely var-
ied from hyper-arid to semi-arid over time,
but has overall maintained its desert-like
characteristics (Eitel 2005). The northern
and southern boundaries of Namaqualand
are marked by the only two rivers with year-
round flow. The Orange River with its head-
waters in the Maloti-Drakensberg Mountains
of southeastern southern Africa (Stewart et
al. 2016), over 1500 km east of Namaqualand,
receives input from predominantly summer
rains, while the Olifants River to the south
receives winter rainfall flow from the near-
coastalCederbergMountains.Fiveadditional
riverbeds bisect the coastal plateau, yet only
flow in years with well above average winter
rainfall, although the estuaries of two—the
Buffels and Spoeg—do contain permanent
standing water.
Given the scarcity of available water in
the region, the flora and fauna are neces-
sarily arid-adapted specialists with many of
them endemic to Namaqualand. The distri-
bution of plants and animals is also unpre-
dictable and patchy due to shifting rainfall
patterns (Desmet 2007). This area is a Succu-
lent Karoo biome dominated by dwarf succu-
lent shrubs; Aizoaceae are prominent, as are
Euphorbiaceae, Crassulaceae, and succulent
members of Asteraceae, Iridaceae, and Hy-
acinthaceae (Mucina and Rutherford 2006).
Extravagant mass flowering of Asteraceae
daisies occurs in spring, transforming the
barren landscape into a mosaic of rich col-
ors. Sweet, palatable grasses are rare, while
spiny grasses are common on aeolian dunes;
trees such as Acacia karroo are present only
along riverbanks. The larger non-domestic
fauna includes remnant herds of springbok
(Antidorcas marsupialis)andgemsbok
(Oryx gazella), while smaller solitary
species include steenbok (Raphicerus
campestris), duiker (Sylvicapra grimmia),
klipspringer (Oreotragus oreotragus), Cape
hare (Lepus capensis), rock hyrax (Pro-
cavia capensis), Namaqualand speckled
padloper (Homopus signatus signatus), Na-
maqua tent tortoise (Psammobates ten-
torius trimeni), and angulate tortoise
(Chersina angulata). Historical sources in-
dicate that much larger species were once
supported on the landscape, including ele-
phant (Loxodonta africana), hippopota-
mus (Hippopotamus amphibious), black
rhinoceros (Diceros bicornis), eland (Tau-
rotragus oryx), zebra (Equus quagga), and
Cape buffalo (Syncerus caffer), as recorded
by early colonial traveler Robert Gordon
commenting on his sojourn to the Orange
River from Cape Town in September 1779
(Cullinan 2006; Skead 1980). The paucity of
large herbivores on the landscape today is
partially due to the degradation and loss of
vegetation and soil quality through overgraz-
ing by domesticated animals as seen in other
regions of South Africa (Hoffman et al. 1999).
Ironically, however, the mining industry ac-
tually preserved the natural landscape as only
10% of the region was available to grazing an-
imals (Desmet 2007). As such, researchers
must also look for evidence of shifting en-
vironments and overhunting to explain the
drastic reduction in large herbivores on the
landscape today.
In his travels, Gordon also met many lo-
calindigenouspeopleonhisjourneythrough
Namaqualand, including “kleine Nama” pas-
toralists and “Bushman” foragers (Cullinan
2006). Thousands of shell middens along the
coastal plateau confirm Gordon’s observa-
tions that the region was more heavily oc-
cupied in the past, at least at certain times
(Dewar 2008).
EARLY OCCUPATION OF
NAMAQUALAND: INSIGHTS FROM THE
HOLOCENE
Gordon visited Namaqualand during a partic-
ularlywetperiodattheendoftheLittleIce
Age, yet he continuously lamented the lack
of available freshwater and large trees for fuel
tostaveoffthecoldertemperatures(Cullinan
2006). An image painted by one of Gordon’s
companions is very telling of how the local
“Bushmen” managed to survive when fresh-
water was rare: women carry net bags full
of ostrich eggshell flasks, watertight contain-
ers that can be plugged and buried for future
4 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
use (Cullinan 2006:gallery page 15). So the
ability to source, cache, and carry water was
clearly important historically.
In terms of archaeological research, the
general picture is still in its infancy. To date,
justover100sitesandburialshavebeenexca-
vated, with a total of 108 radiocarbon dates.
The vast majority (91%) represent Holocene
occupations (Dewar and Orton 2013). How-
ever, these dated sites represent less than 5%
of known sites, with the majority represent-
ing shell middens situated within a kilometer
of the current coastline (Dewar 2008; De-
war and Orton 2013). Moreover, with a dy-
namic marine environment including shift-
ing sea levels and marine dunes there are
likely to be many more sites both underwa-
ter and deeply buried beneath terrestrial sed-
iments. Archaeological survey has recorded
diagnostic stone tools that identify a human
presence stretching back into the Early Stone
Age(1.8mya–250kya):numbersofhandaxes
and Victoria West cores have been identified
in the region, typically at silcrete quarries
and/or as isolated occurrences along river
terraces. It is extremely difficult to interpret
such finds beyond inferring that they likely
represent relatively short-lived incursions by
very small hominin groups. Indeed, their ex-
tremely low densities in unstratified contexts
render them unsuitable for addressing ques-
tions of when humans first dispersed into Na-
maqualand, what palaeoenvironmental con-
ditions prevailed at the time, or whether arid-
environment adaptations were necessary for
its exploitation.
These questions are the focus of two re-
cently initiated, ongoing research projects
investigating rockshelter sites that provide
deep chronologies and palaeoenvironmen-
tal sequences: the Spitzkloof shelters in
northern Namaqualand (Dewar and Stew-
art 2011, 2012, 2016) and Varsch Rivier 3
(VR003) in southern Namaqualand (Steele
et al. 2012). Unlike open-air midden sites,
virtually no work has previously been un-
dertaken in rockshelters in Namaqualand,
and none whatsoever on shelters with Pleis-
tocene deposits. Our understanding of the
timing, motives, strategies, and processes by
which people first began living in this land-
scape is therefore nascent. We can, however,
use the more substantial and better-resolved
datasets from the Holocene occupation of
the region to model subsistence and settle-
ment strategies and to predict which envi-
ronmental conditions would have been suit-
able for supporting continuous occupation
of the region.
The Holocene occupation of Namaqua-
land was pulsed, with palaeoenvironmental
proxies indicating that people primarily
lived in the region when it was colder
and wetter than present day (Dewar 2008;
Dewar and Orton 2013; Dewar and Stewart
2012, 2016). The main periods of occupation
during the Holocene occurred during the
Neoglacial period from 4200 to 1400 BP and
the Little Ice Age from 650 to 150 BP (Dewar
and Orton 2013). While there is some evi-
dence of light occupation during the Middle
Holocene Altithermal (7000 to 4200 BP)
and the Medieval Warm Epoch (1390 to 650
BP), these warm and dry periods are mainly
represented by isolated human burials
(Dewar 2008; Dewar and Orton 2013). Ac-
cess to freshwater is the most likely limiting
factor, with ostrich eggshell flasks present
in archaeological assemblages through the
Holocene, although sites were unexpectedly
untethered to rivers (Dewar 2008). Cold and
wet periods would have been beneficial to
the food chain, increasing grass coverage and
attracting larger game to riverine and inland
locations.
During the Neoglacial subsistence and
settlement patterns indicate an opportunis-
tic residential foraging strategy (see Binford
1980; Rowley-Conwy 1999, 2001) focusing
on the immediate vicinity of sites. Archaeo-
logical sites have produced a relatively high
diversity of faunal species, reflecting a gen-
eralized diet focusing on small meat pack-
ages (Dewar 2008). Low-ranking species that
dominate these assemblages include easily
snared or trapped small bovids, tortoises, and
shellfish. Sites are small, single occupation
processing areas or kill sites where primary
butchery was practiced and the best cuts of
meat taken elsewhere. Group sizes were ei-
ther very small and people highly mobile, or
else base camps have yet to be identified (De-
war 2008). Maintaining clothing was clearly
important as most deposits include bone
needles or awls, as well as the remains of
pelt-bearing species including black-backed
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 5
Genevieve Dewar and Brian A. Stewart
jackal (Canis mesomelas) and bat-eared fox
(Otocyon megalotis).
There seems to have been a population
crash in Namaqualand from 1400 to 650 BP
corresponding to the Medieval Warm Epoch,
but with the onset of the Little Ice Age the re-
gion rebounded and people again become
archaeologically visible. When humans re-
turned to Namaqualand they continued to
practice an opportunistic residential forag-
ing strategy focused on the immediate sur-
roundings of the sites they occupied. A rela-
tively high diversity of species reflects a gen-
eralized diet, yet based on the available calo-
ries an unusual pattern is evident in which
both low-ranking and high-ranking species
dominate the assemblages. Additionally, we
see the introduction of mass harvesting of
lower ranked species, suggesting an overall
subsistence strategy with significant process-
ing (mass harvesting small bovids, penguins,
rock lobster known locally as crayfish) (De-
war2008;Dewaretal.2006) and search costs
(hunting large bovids). Base camps appear at
the coast and include examples of site “furni-
ture”(largeelaborategrindstones,seeDewar
and Orton 2013) and evidence for a social di-
vision of labor, including high frequencies
of exchange items such as ostrich eggshell
beads and pendants and marine shell pen-
dants. Either populations were much larger
or people were staying for extended peri-
ods of time since the mean size of shellfish
takendrops,implyingadegreeofpressureon
available resources (Dewar 2008). Maintain-
ing clothes continued to be important, with
sewing tools and now also high frequencies
of small and large antelopes and penguins
available for both meat and pelts.
The above synthesis of the Holocene ev-
idence from Namaqualand makes it tempt-
ing to categorize the region as a more pre-
dictable and plentiful environment during
coldwetphases.Yetthebehaviorsregistered
in the archaeological record are character-
istic of human adaptations to relatively un-
predictable desert environments (see Smith
2005; Veth 2005; Yellen 1977). For exam-
ple, people appear to have had: 1) high de-
grees of mobility; 2) generalist broad (oppor-
tunistic) subsistence strategies; 3) flexible so-
cial organizations; and 4) open inter-group
boundaries with little territoriality (see De-
war 2008). Taken together these traits sug-
gest that the environment was still patchy
and unpredictable in terms of its distribution
of water and other resources.
PREDICTED USE OF THE PLEISTOCENE
LANDSCAPE
The subsistence and settlement data pre-
sented above for the Holocene provide some
patterns that can be used to model or predict
Pleistocene dispersals and landscape use.
First, it is clear that increased human occu-
pation of Namaqualand during the Holocene
correlates with phases when the region was
colder and wetter than present day. Based
on palaeoenvironmental studies of the win-
ter rainfall zone (Chase and Meadows 2007;
Dewar and Orton 2013; Dewar and Stewart
2016), glacial periods should produce evi-
dence for more intense occupation, as this
is when this desert region was also the most
humid. Unfortunately, regressing shorelines
during glacials means that coastal sites from
those periods are currently submerged, but
inland sites could still be preserved. Colder
periods with greater moisture availability
would have been beneficial for the growth
of grasses providing fodder for large prey
in inland regions. Recall that it was at the
end of the Little Ice Age that Gordon visited
Namaqualand, and his observations make
clear that the landscape at that time could
support large grazers. Second, during these
cool Holocene occupational phases, people
seemed to have maintained fairly high levels
of mobility following a residential foraging
pattern. They focused on species in the im-
mediate vicinity of their encampments, espe-
cially during mass harvesting episodes in the
Little Ice Age, but overall maintained a broad
generalist diet. Using Gordon’s historical ob-
servation, we would also expect to see the
useofostricheggshellflasksaswatercontain-
ers since they were vital to hunter-gatherers
even during the Little Ice Age, a considerably
wetter period than today.
We now turn to northern Namaqualand
and the recent excavation of Spitzkloof A
Rockshelter (28◦51.790S17
◦04.65270’E)
6 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
Figure 2. Photograph of Spitzkloof Rockshel-
ters showing the three hollow domes
and the dry tributary. The shel-
ters are identified upwards as A
through C.
located in the foothills of the front range of
the Stinkfontein Mountains. Today, Spitzk-
loof A is an inland site 30 km east of the At-
lantic Ocean and 30 km southwest of the
Orange River (Figure 1). It is the largest of
three sequential domed shelters on the west
side of a deeply incised dry riverbed of a
tributary of the Holgat River (Figure 2). The
excavated deposit is currently 2.0 m deep,
though it has not yet reached bedrock, and
has produced well-preserved organic mate-
rial, including charcoal, bone, and ostrich
eggshell.
RESULTS AND DISCUSSION
Chronology
The calibrated radiocarbon ages cur-
rently available from Spitzkloof A range from
17kacalBPatthesurfaceto52kaBPata
depth of 1.45 m below it (Table 1, Dewar
and Stewart 2016). Unfortunately, the bot-
tom 0.55 m remain undated as these levels
are beyond the radiocarbon limit, and the
surrounding bedrock is so ancient (>4bil-
lion years ago) and radioactive that the quartz
electron traps here are saturated, preventing
the use of optically stimulated luminescence
(OSL). However, a study is currently under-
way to address this issue and hopefully basal
dates can be obtained in the near future.
Based on a suite of 10 radiocarbon dates
on ostrich eggshell, the nature of occupation
of Spitzkloof A was clearly pulsed, with lay-
ers dated at 52 to 51 ka BP, 23 ka cal BP, 19
ka cal BP, 18 ka cal BP, and 17 ka cal BP
Table 1. Radiocarbon ages from Spitzkloof A, Namaqualand, South Africa (Dewar and
Stewart 2016). Dated samples consisted of inorganic calcium carbonate from
ostrich eggshell.∗
Lab no. Context Date in 14C BP Calibrated dates BP
UBA-17609 Layer Nick 14,350 ±10 17,270–17,090
UBA-17610 Layer Nick 14,400 ±70 17,390–17,130
UBA-17611 Layer Nadja 15,200 ±50 18,300–18,110
UBA-17612 Layer Jaird 16,250 ±60 19,460–19,240
UBA-17613 Layer Dave 19,550 ±60 23,420–23,130
UBA-17614 Layer Mark 19,750 ±80 23,670–23,390
UBA-17615 Layer Julie 19,550 ±60 23,420–23,130
UBA-17617 Layer Brian 52,150 ±800 N/A
UBA-17618 Layer Brian 51,150 ±850 N/A
UBA-17616 Layer Brian >59,250 N/A
∗Experiments have shown that fossil ostrich eggshell is typically 180 ±120 years too old (Vogel
et al. 2001) and so 180 years were subtracted before calibration. Dates were calibrated using the
software Calib 7.0 and the calibration curve Shcal13.14c for the Southern Hemisphere (Hogg et al.
2013). Note that the geological layers Dave, Mark, and Julie represent a single chronological layer.
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 7
Genevieve Dewar and Brian A. Stewart
Figure 3. Profile of the east section of unit G3 from Spitzkloof A Rockshelter with the associated
calibrated radiocarbon dates. The various colors identify the gross layer boundaries of the
deposit.
(Table 1, Figure 3). The thickest deposits
with potential for continuous occupation
are the bottommost radiocarbon-dated lay-
ers (52–51 ka BP and 23 ka cal BP), indicating
that people had indeed colonized the region
by the Middle Stone Age (MSA). The upper-
most layers, ranging from 19 to 17 ka cal BP,
may indicate a particularly sustained occupa-
tion of the site (Table 1, Figure 3).
The radiocarbon chronology for Spitzk-
loof A confirms that people occupied north-
ern Namaqualand during periods that were
wetter and colder than today based on a
range of palaeoenvironmental proxy data
8 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
(Table 2, summarized in Dewar and Stewart
2016). More specifically, these dates reflect
occupation during Marine Isotope Stages
(MIS) 3 (57 to 29 ka) and 2 (29–14 ka) re-
spectively, periods that appear to have wit-
nessed considerable climatic and population
instability across much of the southern Africa
(Beaumont 1986; Mitchell 1990, 2002). The
nearest comparative site is Apollo 11 at
roughly 120 km due north, located within
the desert biome. Recent amino acid racem-
ization dates identify two very old occupa-
tion pulses with minimal ages of 180,000
±10,000 years and 236,000 ±18,000 years
and a MIS 5a date of 80,000 ±12,000 years
(Murray-Wallace et al. 2015). These data sup-
port the previously published 14CandOSL
dates that suggest Apollo 11 was also oc-
cupied sporadically (Vogelsang et al. 2010).
More importantly, Apollo 11 has also pro-
duced deposits dating to MIS 3 and 2. How-
ever, during MIS 3 the occupation of Apollo
11 is not in sync with Spitzkloof, with OSL
dates at 58 ±3 ka (AP9), 57 ±3 ka (AP3),
43 ±3 ka (AP2), and 30 ±1.4 ka (AP11)
(Vogelsang et al. 2010), suggesting region-
ally complex patterning (Mackay et al. 2014).
MIS 2 dates are more similar between the
two sites with recurrent pulses from 22 to
17 ka (Dewar and Stewart 2016; Vogelsang
et al. 2010). Of note is the temporary aban-
donment of both shelters during the height
of the Last Glacial Maximum (LGM), perhaps
indicating that the area was simply too cold
even if moisture availability was then high.
As our most robust evidence for the occupa-
tion of Namaqualand is at 52–51 ka BP and
23 ka cal BP, we now evaluate the evidence
for subsistence and settlement from those
deposits.
Subsistence and Settlement
During the 52–51 ka BP occupation, the
people inhabiting Spitzkloof A utilized a rel-
atively broad range of species (n=16) dom-
inated by small meat packages (Table 3).
Although compared to the coast and bet-
ter watered regions further south, the di-
versity of species is quite low (see Dewar
2008). Based on the number of identified
specimens (NISP), small fur-bearing mam-
mals, tortoises, and large ungulates dominate
the assemblage in that order (Dewar and
Stewart 2012). While all identified species
would have been available in the immedi-
ate vicinity of the site, only the rock hyrax
prefers rocky terrain, with all other identified
species preferring plains or sand dunes. To-
gether, these data reflect a residential mobil-
ity foraging strategy focusing on the immedi-
ate surroundings (see Binford 1980; Rowley-
Conwy, 1999, 2001). Based on NISP, the ma-
jorityofanimalsbroughtbacktothesitewere
small, easily snared, trapped, or picked up,
characteristics that make all of them low-
ranking species in terms of available meat
and low search costs. The presence of a Size
4 (296 to 900 kg; see Brain 1981) ungu-
late indicates the active hunting of at least
one large high-ranking species. Evaluating
the proportion of large to small animals,
the large ungulate index (NISP medium +
large ungulates/NISP small +medium +
large ungulates ∗100%) shows a low repre-
sentation of the highest ranked species at
16% of the ungulates. This confirms that peo-
ple practiced a broad diet with a high con-
tribution of small mammals and tortoises.
However, applying optimal foraging theory
(Bettinger and Baumhoff 1982; Charnov
1976; Kelly 1995) and calculating the avail-
able kilojoules based on the minimum num-
ber of individuals (MNI) present in the as-
semblage (Table 3), the two large ungu-
lates (a gemsbok and a zebra-sized ungulate)
provided 65% of the total energy available,
while the small mammals/bovids together
contributed 34% and the tortoises <1%.
This assemblage suggests a subsistence strat-
egy that included both hunting the highest
ranked species and following a broad gen-
eralized diet. The subsistence strategy for
the albeit younger MIS 3 layers at Apollo
11 is similar with evidence for a broad
diet based on local species and a high fre-
quency of large ungulates (Dewar and Stew-
art 2012). Differences in available prey be-
tween the sites are based on local geology
and ecology with more rocky loving species,
fewer tortoises, and a higher proportion
of zebra at Apollo 11 (Dewar and Stewart
2012, 2016; Thackeray 1979; Vogelsang et al.
2010).
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 9
Table 2. The proxy palaeoenvironmental data for Marine Isotope Stages 3 and 2 as presented in Dewar and Stewart (2016), the
associated dates, implications, and sources of the data. The dates are presented as they were in their published form while the
calibrated dates column reflects calculations for Dewar and Stewart (2016).
Marine isotope
stage Data Signal Date
Calibrated dates at
1σcal BP (for
this study)2Implication Reference
Stage 3: Stadial 57–29 ka
OSL and charcoal Apollo 11: Single species of Xeric
taxa Chenopodiaceae during
occupation hiatus
Between 57.9 ±2.6
ka (AP3) and 42.9
±2.7 ka (AP2)
Arid Vogelsang et al. 2010
Pollen1Decreasing but fluctuating levels
of Restionaceae and
desert/semi-desert species
∼57–50 ka Drying? Shi et al. 2001
Southern Cape sea
level
Paleoscape model:
Bathymetry and
GIS1
Coastline shifts to ∼18 km from
the modern shore
∼57–40 ka Slight regression Fisher et al. 2010
Southern Cape sea
level
Strontium isotopes Slight decrease in 87Sr/86Sr ratios ∼55 ka Slight transgression Fisher et al. 2010
Pollen1Increasing proportion of
Restionaceae and
desert/semi-desert taxa with a
peak at 32 ka
50–29 ka Increasing humidity Shi et al. 2001
Southern Cape sea
level
Steady increase in 87Sr/86Sr ∼55–27 ka Regression Fisher et al. 2010
Terrigenous
sediments1
Rapidly fluctuating aeolian dust
input and trade winds
Instability? Stuut et al. 2002
14C on ostrich
eggshell, fauna
and gypsum
SptzA arid adapted species,
primarily browsers but
Gemsbok suggests some grass;
Gypsum crystals present
∼52–51 ka People on the
landscape in a
(semi-arid)
modern-like
environment but
with increased
humidity
Dewar and Stewart
2012, 2016
10
Eastern Cape
sea level
14C dated wetland peats −52 m a.m.s.l. lowstand 45,200 ±2,000 14C
BP (Pta-4140)
49,970–47,070 Regression Ramsey and Cooper
2002
14C on ostrich eggshell
and fauna
Boegoeberg1: Large hyenas
and water-dependent
grazing species
37,220 ±5,010 14C
BP (GX-22191)
45,430–36,160 Cool/humid with grass Klein et al. 1999
Eastern Cape
sea level
14C dated wetland peats −46 m a.m.s.l. lowstand 39,100 ±1,530 14C
BP (Pta-4142)
44,290–41,880 Regression Ramsey and Cooper
2002
Amino acid
racemization dates
on eggshell
Erb tanks 45 ka Presence of people on
the landscape
McCall et al. 2011
OSL Apollo 11 43 ±3 ka (AP2) Presence of people on
the landscape
Vogelsang et al. 2010
14C on ostrich eggshell
and fauna
Boegoeberg1: Large hyenas
and water-dependent
grazing species
34,990 ±3,110 14C
BP (GX-21190)
42,060–36,100 Cool/humid with grass Klein et al. 1999
14C on ostrich eggshell
and fauna
Boegoeberg1: Large hyenas
and water-dependent
grazing species
33,230 ±2,630 14C
BP (GX-21189)
40,120–34,760 Cool/humid with grass Klein et al. 1999
Southern
Cape sea
level
Paleoscape model:
Bathymetry and GIS1
Slight shift to ∼10 km from
modern shoreline
∼40 ka Shallow transgression Fisher et al. 2010
Calibrated 14C dates Apollo 11 occupational
pulses
∼37 cal BP Presence of people on
the landscape
Vogelsang et al. 2010
Southern
Cape sea
level
Paleoscape model:
Bathymetry and GIS1
Coastline moves to ∼25 km
from modern shore
∼32 ka Slight regression Fisher et al. 2010
Calibrated 14C dates Apollo 11 occupational
pulses
∼32–29 cal BP Presence of people on
the landscape
Vogelsang et al. 2010
14C dated peat bed Peat bed at Kannikwa near
Port Nolloth
27,900 ±310 14C
BP
32,000–31,270 High humidity Beaumont 1986
(Continued on next page)
11
Table 2. The proxy palaeoenvironmental data for Marine Isotope Stages 3 and 2 as presented in Dewar and Stewart (2016), the
associated dates, implications, and sources of the data. The dates are presented as they were in their published form while the
calibrated dates column reflects calculations for Dewar and Stewart (2016). (Continued)
Marine isotope
stage Data Signal Date
Calibrated dates at
1σcal BP (for
this study)2Implication Reference
Orange River
Mouth sea level
14C marine shell −78.4 m a.m.s.l. lowstand 27,800 ±440 14C
BP (Pta-1104)
30,880–31,5104Regression Vogel and Visser 1981
OSL and fauna Apollo 11: Arid adapted species
+equids and warthog
(grazers)
30 ±1.4 ka (AP11) Landscape is slightly
more humid than
today. Some grass
available?
Vogelsang et al. 2010
Southern Cape sea
level
Paleoscape model:
Bathymetry and
GIS1
Shore returns a few km to
∼22 km from modern coast
∼30 ka Shallow
transgression
Fisher et al. 2010
Stage 2: Last Glacial Maximum 29–14 ka
Eastern Cape sea
level
Durban Bay: 14C
dated wetland
peats
−22 m a.m.s.l. stand 24,950 ±950 14C
BP (GaK-1390)
29,950–28,000 Transgression Ramsey and Cooper
2002
Pollen and
charcoal
Elands Bay Cave: Woodland taxa
peak and xeric taxa minimum
and drought intolerant species
20.5 to 17.8 14CkBP ∼25,100–24,310 to
21,910–21,1103
High humidity Meadows and Baxter
1999
Pollen1Peak percentage of Restionaceae ∼24 ka High humidity Shi et al. 2001
(Continued on next page)
12
Terrigenous sediments
and trade wind
proxies1
Peak fluvial activity and
trade winds
∼24 ka High humidity Stuut et al. 2002
14C Hyrax dung Pollen Olea, Stoebe type,
Artemesia, and fern
pollen co-occurring with
dwarf shrubs
17,000 ±190 14C
BP (Pta-8902)
20,740–20,250 Cool and moist or
increased
evapotranspiration
Scott et al. 2004
Southern
Cape sea
level
14CPecten sp.shell −130 m a.m.s.l maximum 16,990 ±160 14C
yrs BP (Pta-182)
20,050–19,6104Last Glacial Maximum
peak
Vogel and Marais
1971
Pollen1Restionaceae percentages
declining
∼19–14 ka Declining humidity Shi et al. 2001
Terrigenous Sediments
and trade wind
proxies1
Fluvial activity and trade
wind curves declining
∼19–14 ka High humidity
declining through
time
Stuut et al. 2002
Orange River
Mouth sea
level
14C marine shell −87.2 m a.m.s.l. lowstand 16,100 ±160 14C
BP (Pta-1105)
18,980–18,6104Regression Vogel and Visser
1981
Pollen Cederberg Mountains:
increasing fynbos, thicket
and succulent vegetation
13000 ±130 14CBP
(Pta-5896) to
11390 ±100 14C
BP (Pta-6041)
15,700–15,290 to
13,280–13,100
Increasing
temperatures and
reduced
precipitation
Scott and
Woodborne
2007a, 2007b
Pollen Eksteenfontein spring:
Stoebe/Elytopappus
indicate cool
temperatures and
increase of Karoo-like
environment
15.2 to 13.6 cal BP
(extrapolated
dates)
Cool and humid
replaced by aridity
Scott et al. 1995
13
Genevieve Dewar and Brian A. Stewart
Table 3. Species list of identified fauna from the MIS 3 (52–51 kyr BP) and MIS 2 (23 cal kyr
BP) layers at Spitzkloof A. NISP is number of identified specimens and MNI is
minimum number of individuals.
52–51 kyr BP 23 cal kyr BP
Species NISP MNI NISP MNI
Black-backed jackal Canis mesomelas 2121
Carnivore size 1 1 / 1 /
Cape Hare Lepus capensis 1121
Namaqua molerat Bathyergus janetta 21
Rock hyrax Procavia capensis 1111
Steenbok Raphicerus
campestris
21
Klipspringer Oreotragus
oreotragus
11
Bovid size class 1 11 1 5 1
Duiker Sylvicapra grimmia 21
Springbok Antidorcas
marsupialis
Bovid size class 2 21 1 56 3
Gemsbok Oryx gazella 2 1 14 3
Red hartebeest Alcelaphus
buselaphus
11
Bovid size class 3 2 / 61 1
Equid sp. zebra 3 1
Mammal size class 1 287 / 349 2
Mammal size class 2 175 1 199 /
Mammal size class 3 10 / 121 /
Mammal size class 4 1 1 37 1
Brant’s whistling rat Otomys brantsii 31
Mus sp. 11
Micromammal 51 2 179 1
Mammal unidentified
size class
3,123 / 4,555 /
Namaqualand speckled
padloper
Homopus signatus
signatus
31 3
Namaqua tent tortoise Psammobates
tentorius trimeni
24 3 70 4
Angulate tortoise Chersina angulata 95 3 873 14
Tortoise 792 3 5,912 6
Unidentifiable fragments
mammal or tortoise
3,029 / 1,346 /
Aves size 1 3 1
Avessize2 1 1202
(Continued on next page)
14 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
Table 3. Species list of identified fauna from the MIS 3 (52–51 kyr BP) and MIS 2 (23 cal kyr
BP) layers at Spitzkloof A. NISP is number of identified specimens and MNI is
minimum number of individuals. (Continued)
52–51 kyr BP 23 cal kyr BP
Species NISP MNI NISP MNI
Largesnake 1111
Medium snake 4 1
Small snake 2 1 4 1
Snake 4 / 2 /
Small fish 1 1
Land snails 4 2
Black mussel Choromytilus
meridionalis
21
Total 7,649 28 13,857 54
Diversity of species 16 19
Large ungulate index 16% 48%
Small mammal MNI ×KJ 800,000 35% 900,000 14%
Large mammal MNI ×KJ 1,451,400 64% 5,250,000 85%
Tortoise MNI ×KJ 18,000 1% 54,000 1%
Size classes of mammals are based on Brain 1981 such that size 1 <23 kg, size 2 23–84 kg, size 3
85–295 kg, size 4 296–900 kg, size 5 >900 kg. Large ungulate index is calculated as (NISP medium +
large ungulates/ NISP small +medium +large ungulates) ∗100%
While no clear ostrich eggshell flask
mouths were recovered at Spitzkloof, we
did retrieve 283.1 g of ostrich eggshell frag-
ments likely representing a minimum of two
eggs (average weight of an empty eggshell is
259 g; Dewar, 2008). There is no evidence
that people were accessing the resources
offered by riverine environments and it is
therefore likely that the tributary outside the
shelter was not in flow. While palaeoenvi-
ronmental proxies indicate that MIS 3 was
overall cooler and wetter than today (De-
war and Stewart 2016), this period was also
one of great climatic flux (Table 2; Stewart
et al. 2016). While the presence of a zebra-
sized ungulate suggests that fodder was avail-
able and potentially even grass, the identi-
fied species are all arid-adapted. There is thus
little support for a particularly wet environ-
ment, and conditions may have been not dis-
similar to today, if slightly more humid and
unstable. There are no sewing tools in the as-
semblage, but there are fur-bearing species,
a bone bead, and two ostrich eggshell bead
preforms that could have been used for mak-
ing clothing or personal ornamentation (De-
war and Stewart 2012). There is no evidence
for personal ornamentation or bone tools at
Apollo 11 from the MIS 3 layers, but there
arefur-bearingmammalsandostricheggshell
(Murray-Wallace et al. 2015).
Thus far the environmental context is in-
conclusive, but the large ungulate at Spitzk-
loof and zebra at Apollo 11 may suggest a
slightly cooler and wetter period than today.
While people seemingly reused the site from
52 to 51 ka BP (Dewar and Stewart 2012), it
was likely visited by small groups of highly
mobile people pursuing an opportunistic,
broad subsistence strategy. Terminal MIS 3
was a period of particularly pronounced in-
stability (Table 2), which may explain the
apparent occupational hiatus between 51 ka
and 23 ka cal BP.
Evaluating the 23 ka cal BP occupa-
tion at Spitzkloof A, important differences
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 15
Genevieve Dewar and Brian A. Stewart
are apparent in the intensity of site use
and in the subsistence and settlement strate-
gies that people pursued there. The diver-
sity of species increases (n=19) and the
total NISP is nearly double that of MIS 3
(Table 3). We now see more large ungulates,
including zebra, but also red hartebeest (Al-
celaphus buselaphus) and gemsbok, both
large grazers. We also see evidence for con-
tact with the coast, which would have been
well over 60 km distant (Dewar and Stewart
2016), in the presence of two black mussel
(Choromytilus meridionalis) shells. A sin-
gle fish vertebra indicates access to river-
ine environments, although the lack of other
freshwater species and the absence of other
remains suggest that the fish did not derive
from the tributary immediately in front of
the site. Based on NISP it would seem that
tortoises were the most important contribu-
tions to the diet at 78% of the faunal assem-
blage, followed by small mammals at 17% and
large ungulates at 5%. Evaluating the propor-
tion of large to small animals, the large un-
gulate index reveals a higher contribution to
the diet from the larger animals at 48% of all
ungulates. This indicates that people prac-
ticed a broad diet with a fairly even contribu-
tion of small and large ungulates. However,
calculating the available kilojoules based on
MNI (Table 3), the large ungulates provided
85% of the total energy available, while the
small mammals and bovids contributed 14%
and the tortoises <1%. Clearly, the large un-
gulates contributed much more to the diet
than the tortoises. There are also more rocky
loving species in this assemblage with the
addition of klipspringer (Oreotragus oreo-
tragus), and the Namaqualand speckled pad-
loper (Homopussignatussignatus,ator-
toise) in addition to the rock hyrax, a pattern
more similar to Apollo 11. While the remain-
ing prey are plains species and are available
in the immediate vicinity of the site, espe-
cially if grasses were abundant, these rocky
loving species suggest that people were also
ranging further afield while living at Spitzk-
loof A, perhaps into the Kamiesberg Moun-
tains to the northeast, beyond which flows
the Orange River. There is also evidence of
a shift in hunting strategy, with a drop in
the number of animals that are more easily
collected, and an increase in those large un-
gulates that typically require active hunting
and tracking. The faunal material from the
MIS 2 layers (early LSA) at Apollo 11 suggest
a similar strategy. While based on NISP the
diet is less reliant on tortoises (3%) and more
focused on small mammals (80%), the major-
ity of calories overall are still coming from
the large ungulate with zebra, representing
16% of the total assemblage (Vogelsang et al.
2010).
At Spitzkloof, the 23 ka cal BP strata
yielded one identified ostrich eggshell flask
mouth and 1179 g of eggshell fragments, rep-
resenting five ostrich eggshells. There is also
more evidence for social interactions, with
ochre appearing for the first time, along with
a bone bead, two large ostrich eggshell bead
preforms, and two complete beads. Two en-
graved ostrich eggshell flask fragments (sin-
gle wavy lines) reinforce the evidence for
enhanced social signaling. The substantial
volume of sediment deposited during this
time period suggests either a larger group
of people used the site more frequently or
for much longer periods of time. At Apollo
11 we also see an increase in social artifacts as
Wendt (1976:7) refers to the presence of os-
trich eggshell and bone beads as well as frag-
ments of ostrich eggshell containers, some
bone tools, fragments of seashells, pigments,
and minerals in the early LSA layer D.
The 23 ka cal BP assemblage repre-
sents a subsistence strategy that suggests an
opportunistic and broad diet (tortoise, small
mammals, birds, fish, and shellfish). Yet peo-
ple were often also hunting the highest
ranked or largest species, including zebra,
red hartebeest, and gemsbok. This strategy
suggests a highly mobile group taking advan-
tage of whatever they come across as they
focused on hunting large grazing animals, a
pattern mirrored at Apollo 11. The palaeoen-
vironmental proxy data for this period (Ta-
ble 2; Chase and Meadows 2007) indicate
that it was one of the coldest in human prehis-
tory. Decreased temperatures would have in-
creased the effectiveness of rainfall by reduc-
ing overall evapotranspiration (Dewar and
Stewart 2016). This seems to have produced
favorable environmental conditions for sup-
porting a range of large grazers. People
16 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
were extracting resources from riverine, in-
land, and coastal landscapes during what
were possibly seasonal rounds. However, ev-
idence for high levels of mobility notwith-
standing, some aspects of the behavioral pat-
tern for this time period are slightly sugges-
tive of an adaptation to a more predictable en-
vironment. People were highly mobile and
overall practiced a generalist diet, but they
were also narrowing their diet by focusing on
highly ranked (large) ungulates that require
an increase in travel over processing time.
Though speculative, it is possible that they
were also social signaling through their use
of pigments and decorated ostrich eggshells
inordertodelineateterritory,somethingper-
haps reflected in their more recurrent use of
the Spitzkloof A site.
CONCLUDING THOUGHTS
At the outset of this article we asked: 1) what
was the tempo, nature, and causes of the Na-
maqualand coastal desert; 2) what was the
environment like; and 3) what pre-existing
socio-technological strategies or new inno-
vations did people use to adapt to this desert
environment? The transition from pioneer
to colonizer in this unpredictable environ-
ment occurred multiple times as the environ-
ment ameliorated from overwhelming dry
heat or intense cold. We have seen that the
region was colonized and then abandoned
often, and sometimes for many thousands of
years, as shown at Spitzkloof A, Apollo 11
and again—from the evidence of multiple
sites—during the Holocene (Dewar 2008).
Pleistocene peoples would therefore have
had to relearn the landscape and its resources
multiple times before becoming relatively
“adept” Namaqualand foragers, and once the
region’s archaeological record is better re-
solved these recurrent signatures should be
detectable.
Humans were certainly, if very sporadi-
cally, using Namaqualand prior to the MSA,
as shown by rare handaxes and Victoria West
cores. But the earliest colonizations in the
region—occupations that resulted in pro-
longed presence (centuries or millennia?)—
probably occurred during the MSA. We have
evidence for continuous re-occupation of
Spitzkloof A from early/mid MIS 3 (52 to 51
ka BP). In southern Namaqualand, two re-
cently described undated sites—Soutfontein
1 and VR003—contain industries with tool
forms diagnostic of the Still Bay (∼70 ka)
and Howiesons Poort (∼65–60 ka), respec-
tively (Mackay et al. 2010; Steele et al. 2012).
These assemblages at both sites are large,
and VR003 possesses a deep sequence that
likely predates the Still Bay (Steele et al.
2012). Thus, it seems that Namaqualand’s
southernmost extreme was colonized by MIS
4 and probably earlier, which is consistent
with its close proximity to the better wa-
tered Cape regions where population lev-
els were presumably more stable. In north-
ern Namaqualand, Still Bay and Howiesons
Poort tools have only been recovered as iso-
lated finds. While this might be marshaled
to support a pioneer interpretation for these
artifacts, it is important to note that Apollo
11 contains deposits with both Still Bay–like
and Howiesons Poort artifacts in high den-
sities, as well as two deeper “Early MSA”
industries dated to 180,000 ±10,000 years
and 236,000 ±18,000 years (Murray-Wallace
et al. 2015). It is far more likely, therefore,
that northern Namaqualand, situated be-
tween the Cape and Namibia, was colonized
much earlier in the MSA, underscoring the
importance of dating Spitzkloof A’s deepest
layers.
While awaiting better palaeoenviron-
mental proxy data, we have turned to Na-
maqualand’s better known Holocene ar-
chaeological record to make predictions
about whether and which particular envi-
ronmental conditions corresponded to Pleis-
tocene human dispersals into the region.
The most intensive Holocene occupations
took place during the Neoglacial and LIA—
both stadial phases with reduced temper-
atures and greater humidity—with little
to no demonstrable human occupation at
other times. The occupations at Spitzkloof
A largely meet our expectations of the site
having been used when the region was wet-
ter and colder than today. The presence of
large ungulates during both Pleistocene oc-
cupation pulses examined here (52–51 ka
BP and 23 ka cal BP) indicates higher levels
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 17
Genevieve Dewar and Brian A. Stewart
of humidity than prevail at present. This pat-
ternofhumanincursionsduringglacialorsta-
dial phases is also consistent with the above-
mentioned human presence in the broader
region during MIS 4, which seems to have
been a period of widespread humidity across
southern Africa (Chase 2010). Looking fur-
ther back in time, it might also hold implica-
tions for understanding the timing and routes
of the earliest colonizations, as perhaps cap-
tured in the deepest, as yet undated levels of
VR003 and the MIS 6 and early MIS 7 dates
from Apollo 11. Humans dispersing into new
landscapes tend to follow conspicuous land-
scape features, including coastlines, moun-
tains, and rivers (Kelly 2003; Meltzer 2009;
Veth 1989). Evidence for coastal dispersals
from the Cape northwards or Namibia south-
wards during glacial/stadial phases of low-
ered sea levels will have been eradicated by
subsequent regressions. If humans opted to
travel along the Cape Fold Belt and then—
from Namaqualand onwards—the Great Es-
carpment, this evidence should be preserved
at sites similar to Spitzkloof in the mountains’
front ranges. A third potential route is the
Orange River, which is oriented east-west
rather than south-north, draining the sub-
continent’s semi-arid interior after rising in
the high rainfall Lesotho Highlands. We sus-
pect that this may have been an important
dispersal corridor into the region, not least
because it crosses both southern Africa’s
summer and winter rainfall zones, precip-
itation in which appears to have been
out of phase during the Pleistocene not
only seasonally but also over long-term
(glacial/interglacial) climatic cycles (Chase
and Meadows 2007). We stress, however,
that these routes need not be seen as mu-
tually exclusive.
Finally, testing our predictions of sub-
sistence and settlement revealed many sim-
ilarities between Holocene and Pleistocene
occupations. But more interestingly, it also
highlighted important differences that can
be investigated further. For example, while
relatively broad-spectrum diets appear to
have characterized both Holocene and Pleis-
tocene groups, there is no evidence the
latter engaged in mass harvesting of situa-
tional or seasonal bonanzas. If such episodes
can be related to broader pan-southern
African trends of Holocene subsistence in-
tensification (Mitchell 2002), this may sug-
gest that Pleistocene population densities
were considerably lower. This is interest-
ing with regards to the subsistence signa-
tures we obtained from the 52–51 ka BP
and 23 ka cal BP occupations, which sug-
gest broad, arid-adapted diets, but with regu-
lar access to high-ranked foods like large un-
gulates. It would seem that access to these
calorific resources during Pleistocene occu-
pational pulses, together with the evidence
for greater overall environmental productiv-
ity (particularly in MIS 2), would have pro-
moted larger absolute population sizes than
during the Neoglacial or LIA. One possibil-
ity is that limitations to the growth of Pleis-
tocene populations, whether resident in Na-
maqualand or situated nearby, were under-
written by rapid climatic changes even dur-
ing phases that promoted greater access to
challenging environments. While this could
have theoretically inhibited the rate of in
situ socio-technological innovations or their
successful transmission (Powell et al. 2009),
local foragers may have maintained large
enough effective population sizes for inno-
vative technologies and strategies to develop
by maintaining social connections to groups
outside the region (Mackay et al. 2014).
The most significant such innovation may
have been the unassuming ostrich eggshell
flask (Texier et al. 2010). Testing this and
other hypotheses raised in this paper will
require many more data points across this
vast region. In the meantime, we hope to
have shown that in marginal environments
with sporadic histories of occupation like
Namaqualand, the archaeology of recent for-
agers can help us forage for food for thought
about the region’s deep past.
ACKNOWLEDGEMENTS
We thank Peter Mitchell and Gustavo
Mart´
ınez for inviting us to contribute to
this special issue on the archaeology of
coastal deserts. A version of this article was
presented at the 4th Southern Deserts Con-
ference in Mendoza, Argentina, an excel-
18 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
lent meeting for which we thank Ramiro
Barberena and the other members of the
organizing committee. We are extremely
grateful to all of the students and col-
leagues who have assisted us with our ex-
cavations at Spitzkloof through the years,
and we are especially indebted to Ben-
jamin Collins, Mark McGranaghan, and
Hugo Pinto. We also thank Ms. Ethel Cloete,
Head of Communications for the Northern
Cape Province’s Richtersveld Municipality,
South Africa, for her assistance with obtain-
ing key supplies in field.
FUNDING
Our work at Spitzkloof Rockshelter A is
funded by Social Sciences and Humanities
Research Council of Canada (SSHRC), grant
numbers 410-2009-1148 and 435-2015-0975.
REFERENCES
Beaumont, P. B. 1986. Where did all the young
men go during 18Ostage2?Palaeoecology of
Africa 17:79–86.
Bettinger, R. and M. Baumhoff. 1982. The Numic
spread: Great Basin cultures in competition.
American Antiquity 47:483–503.
Binford, L. R. 1980. Willow smoke and dog’s
tails: Hunter-gatherer settlement systems and
archaeological site formation. American Antiq-
uity 45:4–20.
Brain, C. K. 1981. The Hunters or the Hunted? An
Introduction to Cave Taphonomy. Chicago:
University of Chicago Press.
Charnov, E. L. 1976. Optimal foraging, the
marginal value theorem. Theoretical Popula-
tion Biology 9:29–136.
Chase, B. M. 2010. South African palaeoenviron-
ments during Marine Oxygen Isotope Stage 4:
A context for the Howiesons Poort and Still Bay
industries. Journal of Archaeological Science
37:1359–1366.
Chase, B. M. and M. E. Meadows. 2007. Late Qua-
ternary dynamics of southern Africa’s winter
rainfall zone. Earth-Science Reviews 84:103–
138.
Clark, J. L. and A. W. Kandel. 2013. The
evolutionary implications of variation in hu-
man hunting strategies and diet breadth
during the Middle Stone Age of south-
ern Africa. Current Anthropology 54:S269–
S287.
Cullinan, P. 2006. Robert Jacob Gordon 1743-
1795: The Man and his Travels at the
Cape. http://web.uct.ac.za/depts/age/people/
Gordon.
Deacon, J. and N. Lancaster. 1988. Late Quater-
nary Palaeoenvironments of Southern Africa.
Oxford: Oxford University Press.
d’Errico, F., C. Henshilwood, M. Vanhaeren, and
K. van Niekerk. 2005. Nassarius kraussianus
shell beads from Blombos Cave: Evidence for
symbolic behaviour in the Middle Stone Age.
Journal of Human Evolution 48:3–24.
Desmet, P. 2007. Namaqualand—A brief
overview of the physical and floristic en-
vironment. Journal of Arid Environments
70:570–587.
Dewar, G. I. 2008. The Archaeology of the Coastal
Desert of Namaqualand, South Africa: A Re-
gional Synthesis. Oxford: Archaeopress.
Dewar, G. I., D. Halkett, T. Hart, J. Orton, and
J. C. Sealy. 2006. Implications of a mass kill
site of springbok (Antidorcas marsupialis)
in South Africa: Hunting practices, gender re-
lations and sharing in the Later Stone Age.
Journal of Archaeological Science 33:1266–
1275.
Dewar, G. I. and J. Orton. 2013. Subsistence, set-
tlement, and material culture on the central Na-
maqualand coastline In The Archaeology of the
West Coast of South Africa (A. Jerardino, D.
Braun, and A. Malan, eds.):109–123. Oxford:
Archaeopress.
Dewar, G. I. and B. A. Stewart. 2012. Preliminary
resultsofexcavationsatSpitzkloof Rockshelter,
Richtersveld, South Africa. Quaternary Inter-
national 270:30–39.
Dewar, G. I. and B. A. Stewart. 2016. Paleoen-
vironments, sea levels and land use in Na-
maqualand, South Africa during MIS 6-2. In
Africa From MIS 6–2: Population Dynam-
ics and Paleoenvironments (S. Jones and B.
A. Stewart eds.):195–212. Dordrecht: Springer
Press.
Eitel, B. 2005. Environmental history of the Namib
Desert. In 23◦S: Archaeology and Environ-
mental History of the Southern Deserts (M.
Smith and P. Hesse, eds.):45–55. Canberra: Na-
tional Museum of Australia.
Fisher, E., M. Bar-Matthews, A. Jerardino, and C.
W. Marean. 2010. Middle and Late Pleistocene
paleoscape modeling along the southern coast
of South Africa. Quaternary Science Reviews
29:1382–1398.
Henshilwood, C. S., F. d’Errico, and I. Watts. 2009.
Engraved ochres from the Middle Stone Age
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 19
Genevieve Dewar and Brian A. Stewart
levels at Blombos Cave, South Africa. Journal
of Human Evolution 57:27–47.
Henshilwood, C. S., F. d’Errico, R. Yates, Z. Jacobs,
C. Tribolo, G. A. T. Duller, N. Mercier, et al.
2002. Emergence of modern human behavior:
Middle Stone Age engravings from South Africa.
Science 295:1278–1280.
Hoffman, M. T., B. Cousins, T. Meyer, A. Peterson,
and H. Hendricks. 1999. Historical and con-
temporary land use and the desertification of
the Karroo. In The Karoo: Ecological Patterns
and Processes (W. R. J. Dean and S. J. Milton,
eds.):257–273. Cambridge: Cambridge Univer-
sity Press.
Kelly, R. L. 1995. The Foraging Spectrum: Diver-
sity in Hunter-Gatherer Lifeways. Washing-
ton, DC: Smithsonian Institution Press.
Kelly, R. L. 2003. Colonization of new land
by hunter-gatherers: Expectations and implica-
tions based on ethnographic data. In Coloniza-
tion of Unfamiliar Landscapes: The Archaeol-
ogy of Adaptation (M. Rockman and J. Steele,
eds.):44–58. London: Routledge.
Klein, R., K. Cruz-Uribe, D. Halkett, T. Hart, and J.
Parkington. 1999. Paleoenvironmental and hu-
man behavioral implications of the Boegoeberg
1 late Pleistocene hyena den, Northern Cape
Province, South Africa. Quaternary Research
52:393–403.
Mackay, A., J. Orton, S. Schwortz, and T. Steele.
2010. Soutfontein (SFT)-001: Preliminary re-
port on an open-air site rich in bifacial points,
southern Namaqualand, South Africa. South
African Archaeological Bulletin 65:84–95.
Mackay, A., B. A. Stewart, and B. M. Chase. 2014.
Coalescence and fragmentation in the late Pleis-
tocene archaeology of southernmost Africa.
Journal of Human Evolution 72:26–51.
McCall, G. S., T. Marks, J. Thomas, M. Eller, S.
W. Horn III, R. A. Horowitz, K. Kettler, and
R. Taylor-Perryman. 2011. Erb tanks: A Middle
and Later Stone Age rockshelter in the central
Namib Desert, western Namibia. PaleoAnthro-
pology 2011:398–421.
Meadows,M.andA. Baxter. 1999. Late Quaternary
palaeoenvironmentsofthesouthwestern Cape,
South Africa: A regional synthesis. Quaternary
International 57/58:193–206.
Meltzer, D. J. 2009. First Peoples in a New World:
Colonizing Ice Age America. Berkeley: Univer-
sity of California Press.
Mitchell, P. J. 1990. A palaeoecological model
for archaeological site distribution in southern
Africa during the Upper Pleniglacial and Late
Glacial. In The World at 18 000 BP, Volume
II, Low Latitudes (C. S. Gamble and O. Soffer,
eds.):189–205. London: Unwin Hyman.
Mitchell, P. J. 2002. The Archaeology of South-
ern Africa. Cambridge: Cambridge University
Press.
Mucina, L. and M. C. Rutherford. 2006. The Veg-
etation of South Africa, Lesotho and Swazi-
land. Pretoria: South African National Biodiver-
sity Institute.
Murray-Wallace, C. V., J. Richter, and R. Vogelsang
2015. Aminostratigraphy and taphonomy of os-
trich eggshell in the sedimentary infill of Apollo
11 Rockshelter, Namibia. Journal of Archaeo-
logical Science Reports 4:143–151.
Powell, A., S. Shennan, and M. G. Thomas.
2009. Late Pleistocene demography and the ap-
pearance of modern human behavior. Science
324:1298–1301.
Ramsay, P. and J. Cooper. 2002. Late Quaternary
sea-level change in South Africa. Quaternary
Research 57:82–90.
Rowley-Conwy, P. 1999. Economic prehistory
in southern Scandinavia. In World Prehistory:
Studies in Memory of Graham Clark (J. M.
Coles, R. M. Bewley and P. A. Mellars, eds.):125–
159. Oxford: The British Academy.
Rowley-Conwy, P. 2001. Time, change, and the
archaeology of hunter-gatherers: How original
is the “Original Affluent Society?” In Hunter-
Gatherers: An Interdisciplinary Perspective
(C. Panter-Brick, R. Layton and P. Rowley-
Conwy, eds.):39–72. Cambridge: Cambridge
University Press.
Scott, L., E. Marais, and G. A. Brook. 2004. Fos-
sil hyrax dung and evidence of Late Pleis-
tocene and Holocene vegetation types in the
Namib Desert. Journal of Quaternary Science
19:829–832.
Scott, L., M. Steenkamp, and P. Beaumont.
1995. Palaeoenvironmental conditions in South
Africa at the Pleistocene–Holocene transition.
Quaternary Science Reviews 14:937–947.
Scott, L. and S. Woodborne. 2007a. Pollen analy-
sis and dating of late Quaternary faecal deposits
(hyraceum) in the Cederberg, Western Cape,
South Africa. Review of Palaeobotany and Pa-
lynology 144(3-4):123–134.
Scott,L.andS. Woodborne. 2007b. Vegetation his-
tory inferred from pollen in late Quaternary fae-
cal deposits (hyraceum) in the Cape winter-rain
region and its bearing on past climates in South
Africa. Quaternary Science Reviews 26:941–
953.
Shi, N., R. Schneider, H.-J. Beug, and L. Dupont.
2001. Southeast trade wind variations dur-
ing the last 135 kyr: Evidence from pollen
spectra in eastern South Atlantic sediments.
Earth and Planetary Science Letters 187(3-4):
311–321.
20 VOLUME 00 •ISSUE 00 •2016
Early Maritime Desert Dwellers in Namaqualand
Skead, C. J. 1980. Historical Mammal Incidence
in the Cape Province Volume 1: The West-
ern and Northern Cape. Cape Town: The De-
partment of Nature and Environmental Conser-
vation of the Provincial Administration of the
Cape of Good Hope.
Smith, M. 2005. Moving into the southern deserts:
An archaeology of dispersal and colonisation.
In 23◦S: Archaeology and Environmental His-
tory of the Southern Deserts (M. Smith and
P. Hesse, eds.):92–107. Canberra: National Mu-
seum of Australia.
Steele, T., A. Mackay, J. Orton, and S. Schwortz,
S. 2012. Varsche Rivier 003, a new Middle
Stone Age site in southern Namaqualand, South
Africa. South African Archaeological Bulletin
67:108–119.
Stewart, B. A., A. Parker, G. Dewar, M. W. Morley,
and L. F. Allott. 2016. Follow the Senqu: Maloti-
Drakensberg paleoenvironments and implica-
tions for early human dispersals into mountain
systems. In Africa From MIS 6-2: Population
Dynamics and Paleoenvironments (S. Jones
and B. A. Stewart eds.):247–271. Dordrecht:
Springer Press.
Stuut, J.-B., A. Maarten, R. Schneider, G. Weltje,
J. H. Jansen, and G. Postma. 2002. A 300
kyr record of aridity and wind strength in
southwestern Africa: Inferences from grain-
size distributions of sediments on Walvis
Ridge, SE Atlantic. Marine Geology 180:221–
233.
Texier, P.-J., G. Porraz, J. Parkington, J. P. Rigaud,
C. Poggenpoel, C. Miller, C. Tribolo, et al. 2010.
A Howiesons Poort tradition of engraving os-
trich eggshell containers dated to 60,000 years
ago at Diepkloof Rock Shelter, South Africa.
Proceedings of the National Academy of Sci-
ences 107:6180–6185.
Thackeray,J.F.1979.An analysis of faunal remains
from archaeological sites in southern South
West Africa (Namibia). The South African Ar-
chaeological Bulletin 34(129):18–33.
Veth, P. 1989. Islands in the interior: A model
for the colonization of Australia’s arid zone. Ar-
chaeology in Oceania 24:81–92.
Veth, P. 2005. Cycles of aridity and human mo-
bility risk minimization among late Pleistocene
foragers of the western desert, Australia. In
Desert Peoples: Archaeological Perspectives
(P. Veth, M. Smith and P. Hiscock, eds.):100–
115. Oxford: Blackwell Publishing.
Vogel, J. and M. Marais. 1971. Pretoria radiocar-
bon dates. Radiocarbon 13:378–394.
Vogel, J. and E. Visser. 1981. Pretoria radiocarbon
dates II. Radiocarbon 23:43–80.
Vogel, J. C., E. Visser, and A. Fuls. 2001. Suitabil-
ity of ostrich eggshell for radiocarbon dating.
Radiocarbon 43(1):133–137.
Vogelsang, R., J. Richter, Z. Jacobs, B. Eichhorn,
V. Linseele, and R. Roberts. 2010. New excava-
tions of Middle Stone Age deposits at Apollo 11
Rockshelter, Namibia: Stratigraphy, archaeol-
ogy, chronology and past environments. Jour-
nal of African Archaeology 8:185–218.
Wadley, L. 2010. Were snares and traps used in
the Middle Stone Age and does it matter? A Re-
view and a case study from Sibudu, South Africa.
Journal of Human Evolution 58:179–192.
Wadley, L. 2013. Recognizing complex cognition
through innovative technology in Stone Age
and Palaeolithic sites. Cambridge Archaeolog-
ical Journal 23:163–183.
Wadley, L., T. Hodgskiss, and M. Grant. 2009. Im-
plications for complex cognition from the haft-
ing of tools with compound adhesives in the
Middle Stone Age, South Africa. Proceedings of
the National Academy of Sciences 106:9590–
9594.
Wendt, W. E. 1976. ‘Art Mobilier’ from Apollo 11
Cave, South West Africa: Africa’s oldest dated
works of art. South African Archaeological
Bulletin 31(121/122):5–11.
Yellen, J. 1977. Long term hunter-gatherer adapta-
tion to desert environments: a biogeographical
perspective. World Archaeology 8:262–274.
JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 21