Early LGM Environments Accelerated the MSA/LSA Transition in Southern African Highlands: the Robberg’s Emergence at Melikane (Lesotho)

Abstract

Melikane, a large sandstone rockshelter in the Maloti-Drakensberg Mountains of highland Lesotho, preserves an 80,000-year-old archaeological sequence including an occupation pulse dated to the onset of the Last Glacial Maximum (LGM), ~27–23 kcal BP. Paleoenvironmental proxies indicate that temperature depressions of ~6 °C below present values provoked changes in vegetation distribution around the site. The onset of the LGM also coincides with a global shift towards microlithization, expressed in southern Africa as the Later Stone Age Robberg bladelet industry. Bousman and Brink’s (Quaternary International 495:116–135, 2018) rapid replacement hypothesis asserts that this technocomplex was adopted nearly simultaneously across the subcontinent ~24 ka cal BP, replacing the Early Later Stone Age technologies that preceded it. An alternative model, which we term the LGM acceleration hypothesis, suggests that the Robberg developed slowly as existing technologies were modified and expanded to function flexibly in a variety of LGM environments. In this paper, we test these hypotheses at Melikane through attribute and morphometric analyses of > 17,000 lithic artifacts. Intrasite continuities and gradual, asynchronous changes in flaking systems are inconsistent with rapid replacement. Instead, the subtle refinement of bladelet reduction strategies alongside climate shifts and a reorganization of mobility and settlement systems supports our LGM acceleration hypothesis. However, Melikane’s combination of highland-specific idiosyncrasies and shared flaking systems with sites in less marginal environments suggests a complex role for cultural transmission. We suggest that periodic isolation throughout the LGM encouraged the development of new flaking systems, the most flexible of which were adopted in a variety of environments when biogeographic barriers to transmission were lifted.

Full text

Vol.:(0123456789)
Journal of Paleolithic Archaeology (2023) 6:24
https://doi.org/10.1007/s41982-023-00150-2
1 3
RESEARCH
Early LGM Environments Accelerated theMSA/LSA
Transition inSouthern African Highlands: theRobberg’s
Emergence atMelikane (Lesotho)
KyraPazan1 · BrianA.Stewart2,3 · GenevieveDewar3,4
Accepted: 5 July 2023
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023
Abstract
Melikane, a large sandstone rockshelter in the Maloti-Drakensberg Mountains of
highland Lesotho, preserves an 80,000-year-old archaeological sequence including
an occupation pulse dated to the onset of the Last Glacial Maximum (LGM), ~27–
23 kcal BP. Paleoenvironmental proxies indicate that temperature depressions of
~6 °C below present values provoked changes in vegetation distribution around the
site. The onset of the LGM also coincides with a global shift towards microlithiza-
tion, expressed in southern Africa as the Later Stone Age Robberg bladelet industry.
Bousman and Brink’s (Quaternary International 495:116–135, 2018) rapid replace-
ment hypothesis asserts that this technocomplex was adopted nearly simultaneously
across the subcontinent ~24 ka cal BP, replacing the Early Later Stone Age tech-
nologies that preceded it. An alternative model, which we term the LGM accelera-
tion hypothesis, suggests that the Robberg developed slowly as existing technologies
were modified and expanded to function flexibly in a variety of LGM environments.
In this paper, we test these hypotheses at Melikane through attribute and mor-
phometric analyses of > 17,000 lithic artifacts. Intrasite continuities and gradual,
asynchronous changes in flaking systems are inconsistent with rapid replacement.
Instead, the subtle refinement of bladelet reduction strategies alongside climate
shifts and a reorganization of mobility and settlement systems supports our LGM
acceleration hypothesis. However, Melikane’s combination of highland-specific idi-
osyncrasies and shared flaking systems with sites in less marginal environments sug-
gests a complex role for cultural transmission. We suggest that periodic isolation
throughout the LGM encouraged the development of new flaking systems, the most
flexible of which were adopted in a variety of environments when biogeographic
barriers to transmission were lifted.
Keywords Last Glacial Maximum· Later Stone Age· Lithic technology· Hunter-
gatherers· Lesotho· Southern Africa
Extended author information available on the last page of the article

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 2 of 35
Introduction
Worldwide, the millennia preceding and encompassing the Last Glacial Maxi-
mum (LGM) — the phase of maximum global ice volume in early marine isotope
stage (MIS) 2 (~24–18 kcal BP; P. U. Clark etal., 2009) — coincide with signifi-
cant shifts in early human adaptive strategies and demography (Chan etal., 2015;
Clarkson etal., 2018; Kuhn & Elston, 2002; Soffer & Gamble, 1990a, 1990b).
Fluctuating atmospheric and oceanic circulations, a global temperature depres-
sion of 6 °C, and extensive glaciation transformed environments and resource
distribution across the globe (P. U. Clark etal., 2009; P. U. Clark & Mix, 2002;
Seltzer etal., 2021). A nearly synchronous global shift to microlithic stone tool
technology, beginning amidst the climatic oscillations of MIS 3, also reached its
full potential during the LGM (Kuhn & Elston, 2002).
Microlithic assemblages were diverse, like the environments they developed
within. The Solutrean technocomplex of southwest Europe (~25 ka) blended
macrolithic, bifacially flaked points with microliths and bladelets (Straus, 2015,
2016). Late Gravettian foragers (~25–23.5 ka) used backed bladelets alongside
regionally diverse points and knives (Polanská etal., 2021; Tomasso etal., 2018).
Siberian and Japanese microblade assemblages (~26–13 ka) also incorporated a
range of larger burins, endscrapers, and blades (Doelman, 2008; Goebel, 2002;
Iwase, 2016; Otsuka, 2017). In southern Africa, microlithization culminated at
the end of the Middle to Later Stone Age (MSA/LSA) transition with the Robberg
technocomplex, which appeared alongside the LGM ~24 ka. Systematic bladelet
production is the hallmark of Robberg assemblages, which often contain scaled
pieces, few formal tools, and significant numbers of unretouched bladelets and
bladelet cores (Table1) (Lombard etal., 2012).
Initially, archaeologists doubted that microlithization was a development internal
to southern Africa. Goodwin and Van Riet Lowe (1929, p. 149) tentatively suggested
that microlithic assemblages arrived via population incursions from further north.
Sampson (1974) modified this idea, asserting that microlithic industries had ori-
gins in southern Zambia and Zimbabwe. J. D. Clark (1959) and Humphreys (1972)
were among others advocating for migration theories. Models incorporating climate
became more popular in the second half of the twentieth century: J. D. Clark (1974)
and Phillipson (1976, 1977) suggested that microlithic technology evolved indepen-
dently in many areas, taking forms that were advantageous in given environments.
Similarly, H. J. Deacon (1976) proposed that microlithic techniques arrived though
diffusion and were then modified to suit environmental conditions. Views began to
change with Mitchell’s (1988a, 1988b) hypothesis of LSA origins. Specific to the
Maloti-Drakensberg mountains of highland Lesotho, it implicated the LGM as an
impetus for microlithization, positing that microlithic technologies were adaptive in
“time-stressed” environments (Torrence, 1983). By the end of the twentieth century,
gradualist theories rejected a sudden population replacement or migration entirely.
Built on sites with long occupation histories like Sehonghong, Umhlatuzana, and
Rose Cottage Cave (Fig.1), these hypotheses recognized intrasite continuities and
subtle temporal trends (Clark, 1997; Kaplan, 1990; Mitchell, 1994; Mitchell, 1996).

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 3 of 35 24
Table 1 Characteristics of final MSA, ELSA, and Robberg assemblages (Beaumont, 1978; Binneman
& Mitchell, 1997; Diez-Martín etal., 2011; Lombard et al., 2012, 2022; Mitchell, 1988a, 1988b, 1994,
1995, 1996; Muller & Clarkson, 2016; Pargeter, 2017; Pargeter & Eren, 2017; Pazan etal., 2022; Porraz
etal., 2016; van der Drift, 2012; Villa etal., 2012; Wadley, 1993, 1996)
Flaking systems Flake attributes Tools
Final
MSA
Prepared core technology • Generally macrolithic
• Blades, triangular flakes
• Prepared platforms (regionally), but plain
platforms and diffuse bulbs more common in
the Melikane MSA
Knives, points,
sidescrapers,
other formally
retouched
implements
ELSA Bipolar reduction, irregu-
lar cores
• Microflakes, occasional macrolithic elements
• Rare bladelets
• Evidence for bipolar: crushed/sheared
platforms, shatter, axial terminations, “blown
off”/crushed bulbs
Occasional
MSA-type
tools
Robberg On quartz: bipolar
bladelet reduction on
flat cores
On CCS: conical single-
platform bladelet cores,
high-backed bladelet
cores
• Emphasis on fine-grained raw materials
• Microlithic
• Frequent unmodified bladelets
• On quartz: evidence for bipolar
• On CCS: smaller platforms, fewer axial
terminations, point-type bulbs
Generally fewer
retouched
tools, no
formal MSA
tools
Fig. 1 Map of southern Africa indicating boundaries between the summer (SRZ), winter (WRZ), and
year-round (YRZ) rainfall zonesand locations of Melikane (MEL) and other sites mentioned in the text:
Rose Cottage Cave(RCC),Sehonghong (SEH), Border Cave (BC), Umhlatuzana (UMH), and Boomp-
laas Cave (BPA)

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 4 of 35
However, variations of the original population replacement model live on. Most
recently, Bousman and Brink (2018) proposed that the MSA/LSA transition actu-
ally consisted of two transitions centered around demographic events at ~43 and ~25
ka. The earlier — from the Middle Stone Age (MSA) to the Early Later Stone Age
(ELSA) — was first expressed at Border Cave in KwaZulu-Natal (d’Errico et al.,
2012; Villa etal., 2012). In this context, “ELSA” encompasses eastern and south-
ern African lithic assemblages dating to between ~45 and 20 ka. These often feature
bipolar cores, scaled pieces, and incipient microlithization and bladelet production,
but sometimes macrolithic components or MSA-type formal tools (Table1) (Bous-
man & Brink, 2018; Lombard etal., 2022). ELSA groups then slowly brought this
technology westwards over a period of 20,000 years. Bader and colleagues (2022)
recently challenged this model, proposing instead that the Border Cave “ELSA”
assemblages were a short-lived local adaptation unrelated to true LSA assemblages
appearing more than 15,000 years later. A subsequent transition posited by Bousman
and Brink (2018) — from the ELSA to the Robberg, spurred by a population bifurca-
tion event ~25 ka and centered on the Southern Cape and uKhahlamba-Drakensberg
Escarpment (Behar et al., 2012; Bousman & Brink, 2018). Bousman and Brink’s
(2018) model inherently challenges more recent gradualist interpretations and hark-
ens back to the migration and replacement models of the early twentieth century.
In this paper, we test current models of this second (ELSA/Robberg) transition
in southern Africa against two lithic assemblages dated to the onset of the LGM at
Melikane Rockshelter, Lesotho (Table2). If a rapid replacement occurred during the
~27–23 ka sequence, we would expect to see an abrupt change in flaking systems
and tool technologies without any foreshadowing or “ramping up” in the underly-
ing layers. If under this same scenario the replacement event had yet to occur, we
Table 2 Expectations of the rapid replacement and LGM acceleration models
Hypothesis Variant Expectations
Rapid
replace-
ment
Rapid replacement of the ELSA at Melikane
occurs later than ~23 ka
• ELSA industry
• No continuity with MSA assemblages
• Stable flaking systems, artifact types,
and frequencies over time
Robberg rapidly replaces the ELSA within the
sequence
• Simultaneous, sudden change from
an ELSA to a completely developed
Robberg flaking system
• No continuity with MSA assemblages
• Intersite similarities greater than intra-
site continuities
LGM
accelera-
tion
Robberg develops gradually in response to dete-
riorating environments
• No sudden turnover in flaking systems
or tool types
• Gradual increase in Robberg traits
• Changes most pronounced closer to
the LGM
• Intrasite continuities greater than
intersite similarities
Mobility and site use patterns change in response
to deteriorating environments
• Changes in artifact density
• Changes in retouch frequency
• Changes in tool type representation

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 5 of 35 24
would expect the assemblages to lack Robberg technologies altogether and resemble
those assigned to the ELSA by Bousman and Brink. The alternative, which we term
the LGM acceleration model, sees the Robberg emerge within southern Africa itself
from MSA technological antecedents. In this scenario, we would except to see intra-
site continuity in flaking systems and a more gradual accumulation of Robberg traits
during the lead-up to the LGM.
The LGM acceleration model builds on Mitchell’s (1988a) hypothesis. Time-
stressed foragers wanting to maximize resource collection during predictable but
limited windows of subsistence opportunity should prioritize curated, premade,
and reliable technologies (Bleed, 1986; Bousman, 1993; Torrence, 1983; Vaquero
& Romagnoli, 2018). However, these same technologies leave foragers unprepared
to take advantage of last-minute subsistence opportunities or cope with changes in
resource scheduling (Nelson, 1991; Vaquero & Romagnoli, 2018). Composite blade-
let technology, while reliable and effective, is also portable and flexible: identical
bladelets can be used with a variety of implements, allowing foragers to adapt to
unpredictable circumstances with a relatively limited toolkit (Bjørnevad etal., 2019;
Lombard & Parsons, 2008; Manninen et al., 2018; Yaroshevich et al., 2010). As
such, we hypothesize that Robberg technologies were adopted at Melikane as a solu-
tion to increasing subsistence risk during the early LGM. If so, we would expect that
instead of arriving as a “package,” different pieces of the Robberg would fall into
place at different times, accumulating gradually with heightened subsistence risk.
Furthermore, we would expect that environmental stress strong enough to instigate
lasting technological change would also provoke changes in mobility and site use
patterns, evidenced through variations in artifact density, retouch frequency, and
reduction intensity.
Background
Site Description andExcavation History
Melikane Rockshelter is located in the Qacha’s Nek District of eastern Lesotho on
the Melikane River, a tributary of the larger Senqu (Orange). It faces northeast at an
altitude of ~1860 m and measures 44 m long and 21 m deep, with an average roof
height of 7.7 m (Carter, 1978) (Fig.2). Melikane is part of the Clarens Formation,
or “Cave Sandstones,” resting just under their interface with the flood basalts of the
Drakensberg Group (Bordy & Head, 2018; Carter, 1978; Visser, 1984). Thermo-
genic processes at the contact zone produced a variety of raw materials used by for-
agers at the shelter. Cryptocrystalline silicates (CCS) and quartz derive from amyg-
dales at the base of the basalts (Gliozzo etal., 2019; Visser, 1984). Various sources
(Humphreys, 1972; Mitchell, 1995; Sampson & Sampson, 1967) claim that these are
easily accessible in small nodules (<4–5 cm) along river banks, but river-rolled cor-
tex is uncommon in highland lithic assemblages. Other available materials exist on
a continuum formed by contact metamorphism between dolerite (diabase) intrusions
and the surrounding sandstones: coarse-grained dolerite from the dykes and sills
themselves, fine-grained hornfels (lydianite) closer to the edge of the contact zone,

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 6 of 35
and partially metamorphosed sandstones and rare quartzites bordering the dyke mar-
gins (Fritzen, 1959; Sampson, 1968). Patrick Carter (1978, p. 145 plate 27A) pho-
tographed a lydianite outcrop “adjacent” to Melikane itself from which chunks of
material >10 cm are eroding. High-quality materials like this would have been read-
ily available at the foragers’ doorstep, although it is important to consider variability
in source exposure through time in such an erosive landscape.
Melikane was first excavated by Carter in 1974. He distinguished seven layers by
perceived depositional differences, but nonetheless excavated in 10 cm spits that cross-
cut the site’s visible stratigraphy (Carter, 1978). In 2007, Carter’s trench was reopened
to obtain samples for optically stimulated luminescence (OSL) dating (Jacobs etal.,
2008). The following year, the Adaptations to Marginal Environments in the Middle
Stone Age (AMEMSA) project led by B. A. Stewart and G. Dewar opened a new 2
× 3 m trench 1 m east of Carter’s, ultimately identifying 30 layers loosely correlating
with individual contexts (Fig.3). Excavations concluded in 2009 upon reaching bed-
rock (Stewart etal., 2012). AMEMSA used a single-context recording system, dividing
thicker contexts into 5 cm spits when necessary. Artifacts were sieved using 1.5 mm
mesh for the upper contexts (including those described in this paper) and 3 mm mesh
for lower ones, where sediment moisture was high (Stewart etal., 2012).
AMEMSA selected archaeological charcoals from secure contexts for accelerator
mass spectrometry (AMS) radiocarbon dating at the Oxford Radiocarbon Accelera-
tor Unit. Samples expected to date >25 ka were processed using rigorous acid-base-
wet oxidation stepped-combustion (ABOx-SC) pretreatment when possible, while
younger samples were processed using the standard acid-base-acid (ABA) protocol
(Brock etal., 2010; Stewart etal., 2012). Dates were calibrated using the IntCal09
data set and OxCal 4.1 (Ramsey, 2009; Reimer etal., 2009). Nine single-grain OSL
dates were also obtained from Carter’s trench, consistent with recalibrated radio-
carbon dates from his original excavations. Occupation pulses at Melikane cluster
around ~80, 60, 50, 46–41, 27–23, 9, and 3 ka (Stewart etal., 2012, p. 51). This
paper is concerned with the occupation pulse dating to 27–23 kcal BP. This corre-
sponds to AMEMSA contexts 6–8 (Layer 5) and context 5 (Layer 4) (Stewart etal.,
2012 p. 44, Fig.4).
Fig. 2 View of Melikane River Valley from the west (left) and interior of Melikane Rockshelter (right).
Adapted from Pazan etal.(2022, p. 117 fig.2)

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 7 of 35 24
AMEMSA contexts 6–8 comprise a single archaeological unit divided into three
arbitrary spits (Fig.4). A sediment sample (MLK 9) from Carter’s east section wall
corresponding to the interface of spits 3 (6–8:3) and 2 (6–8:2) produced an OSL
date of 27.1 ± 1.8 ka for the majority (61.8%) of its quartz grains. Spit 1 (6–8:1) was
directly dated by Stewart etal. (2012, p. 49 tbl. 1) to 24.2–23.6 kcal BP. AMEMSA
context 5 directly overlies contexts 6–8 and consists of five spits. It did not contain
sufficient organic material for a radiocarbon date and attempts to retrieve an OSL
date from this layer failed. However, three additional radiocarbon dates from Cart-
er’s excavation correspond to its interface with contexts 6–8 and provide a maximum
age of 23.8–23.3 kcal BP (Vogel etal., 1986, p. 1144). Multiple sedimentological
indicators also support a LGM age for context 5. This level yielded the sequence’s
lowest organic carbon and magnetic susceptibility values, a peak in mean particle
size, and large volumes of sand and imbricated gravels. All point to colluvial inwash
from a sparsely vegetated, increasingly periglacial landscape (Stewart etal., 2012, p.
51). Moreover, some parts of contexts 6–8 are separated from context 5 by tabular
sandstone slabs (Carter, 1978; Stewart etal., 2012). Similar to those in Layer OS at
nearby Sehonghong Rockshelter and dating 24.4–23.9 kcal BP (Mitchell, 1994; Par-
geter etal., 2017), these may derive from frost-shattering at the LGM’s onset.
Multiple lines of evidence suggest that both layers were strongly affected by
recurrent water ingress through large fissures in Melikane’s rear shelter wall. Con-
text 5 yielded the sequence’s highest mean particle size and lowest organic carbon
values, and its sedimentary matrix is rich in imbricated and rolled gravels. Together
with observations of sub-rounded burnt bone and charcoal observed in thin-section,
abraded artifact surfaces, and water channeling noted during excavation, this sug-
gests a colluvial origin for this layer, perhaps as violent summer storms swept a
Fig. 3 Melikane site plan, showing locations of the drip line (dotted line), Carter trench (blue squares),
AMEMSA trench (maroon squares), and squares chosen for attribute analysis (purple stars). Adapted
from Stewart etal. (2012, p. 45 fig.5)

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 8 of 35
scarcely vegetated landscape (Stewart etal., 2012, p. 51–52). Underlying contexts
6–8 also exhibit micromorphological features consistent with repeated sheetwash
and percolation of water, including abundant clay-coated channels as well as amor-
phous iron staining and sparitic calcitic coatings of lithics. Enhanced organic car-
bon levels towards the base of this layer indicate a buried occupation horizon that
was scoured and reworked by the erosive action of the overlying colluvium (Stewart
etal. 2012, p. 51–52). While these taphonomic forces may have acted to winnow out
smaller artifacts from these levels, vertical movement of archaeological materials is
unlikely give their thickness and quasi-cemented compaction.
Environmental Context
Melikane is located in southern Africa’s summer rainfall zone (SRZ), in which
>66% of precipitation falls during the summer months (Roffe et al., 2019). Gen-
erally, summers are warm with violent thunderstorms, and winters are cold and
dry. Snow can fall at any time of year (Grab, 1997). The mean annual temperature
around Melikane today is ~13 °C (Grab, 1997), ~6 °C greater than values during
the LGM (Holmgren et al., 2003; Kulongoski et al., 2004; Partridge et al., 1999;
Seltzer etal., 2021; Stute & Talma, 1998). By 24 kcal BP, depressed temperatures
combined with a shift in precipitation seasonality resulted in glacier and permafrost
formation on the cold, south-facing slopes of the Drakensberg Escarpment and high
plateau (Bregman & Knight, 2022; Grab etal., 2021; Grab, 1996; Mills & Grab,
2005; Wang & French, 1995).
Fig. 4 Diagram of the Melikane sequence showing locations of Carter’s recalibrated 14C dates (purple),
AMEMSA’s 14C and OSL dates (pink), Bousman Index values (% C3, green), tree cover density ratio
values (D/P, blue), and 13C isotope values (red). Spits within contexts are designated by tick marks on the
left (data fromStewart etal., 2016)

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 9 of 35 24
Vegetation composition in the Lesotho highlands is largely determined by temper-
ature, and thus correlates with altitude and aspect (Ehleringer etal., 1997; Patalano
etal., 2023; Vogel, 1983). Higher altitudes host primarily C3 plants, whereas lower
altitudes favor C4-dominant communities (Vogel, 1983). Melikane is surrounded
primarily by C4 vegetation today, but soil organic matter (SOM) δ13C values from
the site suggest that C3 grasses dominated the local landscape throughout the site’s
late Pleistocene sequence (Stewart et al., 2016). Four SOM and phytolith samples
were taken from Melikane’s ~27–23 ka levels at depth increments of 10 cm, corre-
sponding to contexts 6–8:3, 6–8:1, 5:3, and 5:1 (Fig.4; Table3). The ratio of woody
to grass phytolith types (D/P ratio) falls to its lowest point in the entire sequence
in samples MLK4 and MLK5 (contexts 6–8), reflecting a treeless, open grassland
landscape (Stewart etal., 2016). Slightly higher D/P ratios in MLK 2 and MLK3 are
still lower than values frommost other occupation pulses (Stewart etal., 2016). Cli-
matic indices (Ic%) based on phytolith counts indicate that C3 pooid(“sour”) grasses
(e.g., Festuca and Merxmuellera) became more common relative to C4 panicoid and
chloridoid types in samples MLK 4 and MLK3 (Mucina & Rutherford, 2006; Parker
etal., 2011; Stewart et al., 2016; Twiss, 1992). This is consistent with lower tem-
peratures at the onset of the LGM ~24 ka.
The expansion of C3 grasses into the Melikane River Valley would have changed
both the nutritive quality and seasonality of the landscape. Lesotho’s Highland
Basalt Grasslands currently support Themeda-Festuca communities between 1900
and 2900 masl. Lower elevations with greater proportions of C4 Themeda grasses
are heavily grazed year-round, but higher altitude, Festuca-dominant grasslands are
only preferred by grazers in the summer (Mucina & Rutherford, 2006; Otunga etal.,
2016). If the landscape around Melikane ~24 ka began to resemble the pooid-rich,
nutrient-poor higher altitudes of the Basalt Grasslands, its suitability for grazing
would have declined. In other words, while foragers and grazers at Melikane ~27
ka may have had year-round access to the lusher faunal resources of ahigh-nutrient
grassland, those resources may have only been available on a seasonal basis by ~24
ka.
Interestingly, the top sample in context 5, MLK2, has a the lowest δ13C values
in the sequence (indicating >95% C3) but a lower Ic% and significant contributions
Table 3 Carbon isotope ratios, Bousman indices, climatic indices (Ic%), tree cover density ratios (D/P),
elemental carbon (C%), panicoid %, and chloridoid % from the ~27–23-ka contexts at Melikane rock-
shelter (dates from Stewart et al., 2012, SOM, and phytolith data from Stewart et al. (2016, p. 258
fig.14.6)
Sam-
ple
Context/
spit
equiva-
lent
Date Mean
δ13C
Bousman
index (%
C3)
Ic% D/P C% Pani-
coids
Chlori-
doids
MLK2 5:1 <23.8 ka cal BP −24.0 >95 ~50 .1 0.71 ~15% 10%
MLK3 5:3 <23.8 ka cal BP −23.2 ~90 70–75 ~.05 0.27 ~15% ~1%
MLK4 6-8:1 24.2–23.6 ka cal BP −22.5 ~90 70–75 <.05 2.84 ~15% 0%
MLK5 6-8:3 27.1 ± 1.8 ka
(OSL[MLK9])
−23.1 85 ~50 <.05 5.78 ~15% ~2%

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 10 of 35
from chloridoids in its phytolith sample. These highly nutritious grasses typically
thrive in low moisture conditions and do retain nutritional value in the winter. As
context 5 remains undated, it is unclear whether increased chloridoids in MLK2
indicate warmer temperatures at the end of the LGM, increased aridity (intolera-
ble for pooids despite lower temperatures), or a response to lower atmospheric CO2
values, which favor C4 plants (Ehleringer etal., 1997, 2002; Mucina & Rutherford,
2006). Theoretically, their presence at the site could have attracted grazers during all
seasons and alleviated some of the stresses experienced by early foragers, but cold
temperatures and/or lowered snowlines may still have placed geographic barriers on
human movement. This is supported by sedimentological indicators projecting an
image of a dry, periglacial landscape (Stewart etal., 2016). Thus, the relationship
between LGM climates and vegetation may not have been straightforward in the
Lesotho highlands, and resource availability would likely have been unpredictable
even on short-term time scales.
Charcoal samples from contexts 6–8 show a narrowing of woody species rela-
tive to earlier contexts, biased towards hardy and frost-resistant taxa (Mucina &
Rutherford, 2006; Orlandi etal., 2005). Three species — Erica drakensbergensis,
Leucosidea sericea, and Olea europea — are well represented in the early MIS 2
assemblage and are good sources of fuel (Stewart etal., 2016). Notably absent are
Buddleja salviifolia and Rhamnus sp., both of which occur throughout the earlier
deposits. While Rhamnus sp. is resistant to frost, Buddleja salviifolia, a parallel
scrub community to Leucosidea sericea, exists at lower altitudes than the latter and
is less cold-hardy (Killick, 1978; Stewart etal., 2016). However, charcoal samples
are inherently biased towards fuel sources, and may not accurately represent the taxa
local to the site. Charcoal samples were not obtained for context 5.
Taken together, these data indicate that Melikane was surrounded by a nutri-
tious grassland just prior to the LGM that became only seasonally productive ~24
ka. Carter (1976) hypothesized that during particularly harsh periods, the highlands
may have been abandoned altogether. However, the landscape was clearly still pro-
ductive enough to support populations at the onset of glacial conditions, at least dur-
ing parts of the year. Productivity may have increased at times during the LGM, but
it is unclear how this relates to temperature and transhumance.
Materials andMethods
In total, 20,194 lithic artifacts were analyzed from contexts 6–8 and context 5 at
Melikane (Table 4). All lithics >10 mm in maximum dimension were sorted by
typology and raw material. Lithics <10 mm with retouch, use, or flake attributes
were also sorted. Undiagnostic flake fragments and shatter <10mm were counted
as chips, but raw material was not recorded. They are included in calculations of
artifact density and relative artifact type frequency but are excluded from further
analyses. The lithics from squares S5 and S6, representing a subsample of 5916 arti-
facts, were also selected for attribute analysis and measurement. Square S6 was cho-
sen because it is the only square excavated at the base of the Melikane sequence.
Square S5 nearly reaches the base of the site, and its position immediately next to S6

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 11 of 35 24
ensures maximal continuity of the contexts and spits. Artifacts were measured using
digital calipers and a digital scale. An engineering protractor was used to determine
platform and edge angles. Cores were identified using the core classification scheme
described in Conard etal. (2004). Cores with bladelet removals were further subdi-
vided into typological categories — irregular, high-backed, prismatic, flat, and small
— used by Mitchell (1995) and J. Deacon (1984b) in their analyses of Robberg
deposits at Sehonghong, Boomplaas, and other sites. Retouched artifacts were clas-
sified based on the typology outlined in Carter (1978) with modifications to allow
for comparison to more recently excavated assemblages. Types are consistent with
those used in analyses of Melikane’s older deposits (Pazan etal., 2022).
RStudio (version 4.0.5; R Core Team, 2021) was used for both statistical analyses
and data visualization. Permutation Welch’s t-tests were used for comparisons of 2
independent groups. Row-wise, pairwise, and 2-sample tests for proportions (z-tests)
were used for the analysis of contingency tables, using the Bonferroni correction for
multiple comparisons as necessary. P-values <0.05 were considered potentially sig-
nificant, pending effect sizes expressed through 95% confidence intervals. Justifica-
tions for statistical methods can be found in Online Resource 1 (OR1), and detailed
results of all operations can be found in Online Resource 2 (OR2).
Results
Artifact Density andRaw Materials
We anticipated that environmental change strong enough to encourage the adop-
tion of a new technocomplex would also prompt new site use and mobility patterns.
Greater density in contexts 6–8 (21,400.4 artifacts/m3) than in context 5 (12,199.8
artifacts/m3) suggests that the site was occupied for shorter periods of time dur-
ing the LGM proper. Tools (retouched artifacts) and shatter are also significantly
Table 4 Artifact type by context
*Significant in row-wise proportions test. Details in OR2
+ Raw material not recorded. Not included in subsequent tables or
statistical tests
Type Context 5 Contexts 6–8 Tot al
Frequency %Frequency %
Flakes* 4246* 43.5 3681* 35.3 7927
Cores* 983* 10.1 925* 8.9 1908
Tools* 2430* 24.9 3436* 33 5866
Flaked pieces 96 1120 1.2 216
Shatter* 598* 6.1 1006* 9.7 1604
Heat spall 4 <0.01 30 0.3 34
Chips*+ 1415*+14.5 1224*+11.7 2639
Total 9772 10,422 20,194

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 12 of 35
more frequent in contexts 6–8 (Table4, OR2). High retouch frequencies and artifact
densities are consistent with assemblages formed by foragers with low residential
and low logistical mobility. Less retouch and lower artifact densities in context 5
suggest, in contrast, that foragers moved more frequently and occupied the site for
shorter periods (Schoville etal., 2021).
It is notable that chip frequencies are relatively low in both contexts, and espe-
cially in contexts 6–8, which, as mentioned, underwent particularly significant
taphonomic changes. Chip frequencies reach ~59% during the MSA/LSA tran-
sition and 51% in the Robberg assemblages of Rose Cottage Cave, for example
(Clark, 1997; Wadley, 1996). On the other hand, Deacon (1984a) linked increas-
ing quartz use with rising chip frequencies throughout the LSA at Nelson Bay
Cave, Kangkara, and Boomplaas. At the latter site, chips comprise 52.7% of
“waste” in LGM Robberg member GWA. Their relative paucityat Melikane —
and subsequently inflated retouch frequencies and depressed flake counts — is
therefore likely a symptom both of taphonomic differences and the physical prop-
erties of CCS as compared to quartz.
A significant increase in the use of CCS (cryptocrystalline silicates) from 64%
of artifacts in contexts 6–8 to 78.6% of artifacts in context 5 is consistent with a
greater use of fine-grained raw materials, as expected during the Robberg (p <
0.001, 95% CI [0.133, 0.159]; Fig.5; OR2). Hornfels is the second most common
material (18.2% and 13%), followed by sandstone (15.8% and 7.3%). There are
minute quantities of quartz, mudstone, and unknown materials (1.9% and 1%).
Fig. 5 Raw materials by context

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 13 of 35 24
Flakes
Differences in blank dimensions and attributes between the contexts are subtle,
but not insignificant: while context 5 does show a more typical Robberg profile,
the foundations of bladelet technology are clearly already in place in contexts
6-8 (Fig.6). Blanks in Melikane’s early MIS 2 assemblages are small and irregu-
lar, with mean absolute lengths of 19.2 mm and 16.7 mm in contexts 6–8 and
5, respectively. Only the percussive (p = 0.01, 95% CI [−5.6, −0.8]) and abso-
lute lengths (p = 0.02, 95% CI [−5.0, −0.4]) of CCS blanks differ significantly
between contexts (see OR2 for significance tests of all dimensions), although 95%
confidence intervals indicate the real differences in means may be relatively mini-
mal (<1 mm). Although CCS blanks are slightly shorter in context 5, as expected
in a more microlithic Robberg, consistency in other dimensions and among the
other raw materials does not indicate significant change.
Blanks from all contexts typically have unprepared platforms, exterior platform
angles (EPAs) ~90°, parallel or multidirectional dorsal scars, straight profiles, and
triangular cross-sections. Attributes remain relatively stable between contexts, but
the few changes that do occur in context 5 align with our expectations for Robberg
flaking. First, platforms <3 mm in maximum dimension represent a significantly
greater proportion of platforms in context 5, increasing from 5.9 to 13.9% of proxi-
mal and whole flakes (p < .001, 95% CI [0.034, 0.126]; OR2). Secondly, a row-
wise test for proportions of termination types (feather, hinge, plunging, axial, and
step) was significant for plunging (p = .005, 95% CI [0.023, 0.078]) and step (p <
Fig. 6 a–h CCS debitage and bladelets, context 5; i, j, o hornfels debitage, contexts 6–8; k–n CCS debit-
age and bladelets, contexts 6–8

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 14 of 35
.001, 95% CI [−.120, −0.047]) terminations. The latter comprised 10% of termina-
tions in contexts 6–8 but only 1.8% in context 5 (see OR2 for details). Step termina-
tions can form due to raw material irregularities or by “buckling” — when a (usu-
ally relatively thin) flake snaps as it detaches from a core (Cotterell & Kamminga,
1987, p. 700). Fewer step terminations in context 5 are consistent with the use of
higher quality raw materials, as expected in the Robberg, or refinement of the reduc-
tion process required to make thin flakes (Cotterell & Kamminga, 1987). However,
plunging terminations, also considered a knapping “mistake,” increase from 1.5% of
terminations in contexts 6–8 to 6.5% in context 5. These are particularly common
on blade cores with a sharp angle at their base or when the flaking surface becomes
too narrow in relation to the desired product (Bradley & Giria, 1996; Cotterell &
Kamminga, 1987). While inconsistent with the refinement anticipated in Robberg
assemblages, this increase does align with more frequent blade production, freehand
reduction, and smaller core sizes.
Despite a more Robberg-like distribution of flake attributes, bladelets (defined as
flakes <25mm in absolute length with a length/width ratio >2) are equally common
in contexts 6–8 and 5, comprising 7.9% of the unmodified and utilized whole flakes
in each context. Equal representation is problematic for a rapid replacement event,
in which a bladelet-poor ELSA and bladelet-rich Robberg should be easily distin-
guishable. Bladelets are proportionally more common among unmodified blanks
than they are among utilized flakes, consistent with patterning at nearby Sehong-
hong Rockshelter (Mitchell, 1995). In context 5, they encompass 11.3% of blanks
but only 4.5% of utilized flakes. Mitchell (1988a) suggests this pattern resulted from
the expedient disposal and replacement of bladelets on multi-component tools while
completing tasks away from the site. The presence of mastic on the lateral margin of
a bladelet from context 5 suggests that it was hafted longitudinally, like later Rob-
berg bladelets from Sehonghong and Rose Cottage Cave (Binneman, 1997; Binne-
man & Mitchell, 1997).
Cores
935 whole cores were identified from contexts 6–8 and 5 (Fig.7; Table5). CCS is
overwhelmingly the most common raw material for cores in both contexts (83.8% in
contexts 6–8 and 89.9% in context 5). It is disproportionately represented in com-
parison to its frequency in the rest of the assemblage (p < 0.001, 95% CI [0.143,
0.190]; OR2). This mirrors the pattern seen in the Melikane MIS 5a assemblage,
which was interpreted as the result of different provisioning strategies for CCS
and other materials (Pazan et al., 2022). Because sandstone and hornfels can be
exploited in larger nodules, core reduction is more likely to occur at the source,
minimizing transport costs. Hyperlocal outcrops, possibly including that of lydianite
photographed by Carter (1978), may even have served as extensions of the shelter.
As a result, fewer cores made with these materials would be recovered within the
shelter itself. The continuation of this pattern into MIS 2 suggests that early LGM
foragers at Melikaneapplied a similar logic to the raw material landscape, even if
their preferences(and likely also specific source locations) had shifted.

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 15 of 35 24
Core types
Bipolar reduction is defined in its broad sense, “as a percussion technique in which a
stone core is placed on an anvil and struck with a hammer to produce flakes” (Gur-
tov & Eren, 2014, p. 285). The majority of Melikane’s bipolar cores are blocky, with
extensive battering/crushing on >2 surfaces and disorganized flake scars, sometimes
with feather/freehand-like terminations (Fig. 7a). These are interpreted as cores
rotated repeatedly on an anvil during the reduction process and reduced opportun-
istically, incorporating elements of the “horizonal axial” technique (Diez-Martín
etal., 2011, p. 692), “oblique bipolar flaking” (van der Drift, 2012, p. 160), in which
the core is placed on a surface but struck along an axis that does not terminate with
ground contact, and anvil percussion (Callahan, 1987, p. 15 fig.2,p. 20; Diez-Mar-
tín etal., 2011). Experimental studies are lacking, but it seems likely that this reduc-
tion strategy would have allowed quick, maximal core mass reduction on small nod-
ules while providing an abundance of irregularly shaped flakes.
Possible anvil-assisted flaking was also noted on some platform cores, evidenced
by flattening and abrasion at the base (see Callahan, 1987;Fig.7c). A similar phe-
nomenon was noted by Pargeter (2017) at Sehonghong and Boomplaas, where small
freehand cores were stabilized on an anvil (see Pargeter, 2017, p. 333 fig.7.23), and
by Schmid etal. (2022) at Umhlatuzana. However, these cores at Melikane were not
Fig. 7 CCS cores. aBipolar core, context 5 (blue and green dotted areas: initial platforms/anvil contact
points, red dotted area: secondary anvil contact point); b, d high-backed bladelet cores, context 5; c sin-
gle-platform bladelet core (anvil-assisted?), contexts 6–8; e bipolar core, contexts 6–8 (visible flake scars
outlined; note extreme crushing on all surfaces, especially the platforms/bases); f single-platform core,
contexts 6–8; g core-reduced piece, contexts 6–8

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 16 of 35
rotated, have otherwise freehand core characteristics, and exhibit organized removal
surfaces. They were thus classified as platform cores (cf. Conard etal., 2004).
We expected that rapid replacement should produce sudden turnover in core
reduction strategies. However, only the frequency of core-reduced pieces (CRPs,
interpreted as exhausted bipolar cores) differs significantly between the contexts (p
< .001; see OR2), dropping from 36% in contexts 6–8 to 23.5% in overlying context
5. When broken down by spit, it becomes apparent that this is not a sudden change:
evidence for bipolar reduction is heaviest in 6–8:2 (61% of cores), slightly lighter in
6–8:1 (49.6%), and lighter still in 5:5– (40.5%;Fig.8).
Platform cores, associated with Robberg bladelet production except in quartz-
dominated assemblages, follow an inverse pattern (Binneman & Mitchell, 1997;
Deacon, 1978, 1984a; Mitchell, 1995). A trend rather than a sudden reversal of the
two strategies is inconsistent with rapid replacement and speaks to a technological
shift that occurred over a more extended period. Interestingly, it also correlates with
changes in landscape structure visible in the MLK4 phytolith sample. Regardless, a
preference for a more organized removal strategy is a marked change.
Bladelet Cores
Although bladelet frequencies are identical in contexts 6–8 and 5, bladelet cores
(defined as whole cores, core tools, and core fragments with ≥1 bladelet removal)
are significantly more common in the latter context (p <.001, 95% CI [0.058,
0.135]; Table6). Disjuncture between change in the two artifact types is inconsistent
with rapid replacement, which would produce change in both simultaneously. Lami-
nar removals exist on a variety of core types, including bipolar, multidirectional,
platform, and parallel cores. This is consistent with Pargeter and Redondo’s (2016)
analysis of Sehonghong bladelets, which determined that the costs of bladelet pro-
duction at that site were also distributed across diverse reduction techniques.
Some bladelet cores exist on a continuum with nosed scrapers and endscrapers,
a trend similarly noted by J. Deacon (1978) at Nelson Bay Cave (NBC). Deacon
Table 5 Core type by context
*Significant in row-wise test for proportions. See OR2 for details
Core type Context 5 Contexts 6–8 Total
Frequency %Frequency %Frequency %
Bipolar 95 20.9 89 18.5 184 19.7
Inclined 13 2.9 10 2.1 23 2.5
Initial 39 8.6 26 5.4 65 7.0
Multidirectional 133 29.2 115 24.0 248 26.5
Platform 45 9.9 25 5.2 70 7.5
Parallel 23 5.1 42 8.8 65 7.0
CRP* 107 23.5 173 36.0 280 29.9
Total 455 480 935

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 17 of 35 24
(1978, p. 97) initially termed these cores “high-backed scrapers” because of crush-
ing on the striking platform resembling scraper-like use. She later reclassified them
as cores after determining that their primary function was for bladelet production,
acknowledging ambiguity and continuity between the types. “High-backed” or
“wedge-shaped” bladelet cores are also diagnostic of the early Robberg assemblages
at Boomplaas and Sehonghong (Deacon, 1978, 1984b; Mitchell, 1988a, 1995).
At Melikane, we attempted to separate high-backed cores resembling drawings in
Mitchell (1995), Carter etal. (1988), and Deacon (1978, 1984b) from more ambigu-
ous, multipurpose, scraper-like forms. True high-backed bladelet cores were found
primarily in context 5 (n = 17), although two were found in a single quadrant (0.5 ×
0.5 m2) of contexts 6–8:1. The presence of only two high-backed cores in the upper-
most spit of contexts 6–8, in close proximity with each other and near the interface
with context 5, suggests either slight mixing or the introduction of a new reduction
sequence late in the context. Other bladelet core subtypes (flat, small, and prismatic)
were found in small numbers in both contexts. Flat bladelet cores, often associated
with bipolar reduction on quartz, are less common in context 5, consistent with
observations at Boomplaas and Sehonghong (J. Deacon, 1982, 1984b; Mitchell,
1988b).
Fig. 8 Core type proportions by context and spit (6–8:3 and surface scrapes eliminated due to low sam-
ple sizes)

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 18 of 35
Tools
We expected that adaptive challenges strong enough to accelerate the adoption of
bladelet technology would also affect mobility and site use strategies, evidenced
through changes in retouch frequency and tool type representation. We also expected
that assemblages belonging to the ELSA or Robberg would lack stereotypically
MSA tool types (e.g., denticulates, sidescrapers, points). A total of 5865 retouched
artifacts and one hammerstone were recorded in contexts 6–8 and 5 (Table7; Fig.9).
As previously discussed, more common retouch and higher artifact densities in con-
texts 6–8 suggest longer periods of site occupation than in context 5 (see “artifact
density and raw materials”). This does not appear to be driven by shifts in raw mate-
rial representation. Although hornfels is both more common in contexts 6–8 and
retouched more frequently than CCS (p < .001; OR2), CCS alone is still retouched
more frequently in the latter context (p < 0.001; OR 2).
Higher subsistence risk, expected closer to the LGM, should correlate to
“resource-maximization tactics”: more frequent disposal of extractive tools (those
used for procuring resources, e.g., composite spears), but more intensive curation
of maintenance tools (those used for processing or making tools, e.g., endscrap-
ers) (Bousman, 2005, p. 219). It is unclear how the Melikane data fit into this
Table 6 Bladelet core type by
context
*Proportion of bladelet cores significantly different between con-
texts. See OR2
Core type Context 5 Contexts 6–8
Frequency %Frequency %
Irregular 241 87.3 145 85.3
Flat 8 2.9 16 9.4
High-backed 17 6.2 21.2
Small 9 3.3 52.9
Prismatic 1 <1 21.2
Total bladelet cores* 276 170
Total cores 983 923
% bladelet cores 28.1% 18.4%
Table 7 Retouch frequency by raw material and context
*Complete results for pairwise tests for equality of proportions by raw material within and between con-
texts can be found in OR2
Context 5 Contexts 6–8
Tools All artifacts % Tools All artifacts %
CCS 1758 6572 26.7 2042 5891 34.7
Hornfels 462 1086 42.5 832 1674 49.7
Sandstone 201 611 33.4 537 1454 36.9
Other 9 88 10.2 25 179 14.0
Total 2430 8357 29.1% 3436 9198 37.4%

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 19 of 35 24
model: although none of Melikane’s common tool types are obviously extractive
(e.g., points), lack of usewear prevents us from making a definitive statement on
the matter. In any event, tool type representation shifts only slightly between the
contexts (Table8; OR2). In both contexts, borers were the most common tool type,
becoming more prominent overall in context 5 (p < .001, 95% CI [0.061, 0.112]),
whereas miscellaneous retouched pieces (MRPs) become slightly rarer (p < .001,
95% CI [−0.085, −0.043]). In neither context do convergent flakes (points) form a
Fig. 9 Tools. a CCS burin, context 5 (single burination indicated); b hornfels borer, context 5 (burination
plus additional retouch); c CCS borer, context 5 (burination plus retouch); d CCS nosed scraper, context
5 (noses indicated); e hornfels MRP, context 5 (retouch indicated); f CCS side/endscraper, contexts 6–8
(retouched margins indicated); g CCS MRP, contexts 6–8; h, j CCS borers, contexts 6–8 (removals indi-
cated); i CCS borer with possible bladelet removals (indicated), contexts 6–8; k CCS nosed scraper with
possible bladelet removals, contexts 6–8 (removals in call-out)

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 20 of 35
significant portion of the tool assemblage. In both layers combined, they comprise
only 0.1% (n = 7) of formal tools, and none have the convergent dorsal scars sug-
gestive of Levallois reduction. In contrast, they comprise 7.4% of the total toolkit
in MIS 5a and 23.3% of tools in MIS 4 (context 28). Despite frequent retouch, the
rarity of these and other MSA-type implements relative to earlier contexts is consist-
ent with expectations for an ELSA or Robberg assemblage (Pazan etal., 2022). The
continued presence of borers, well-represented in the Melikane MSA, is also prob-
lematic for a rapid replacement event.
Discussion
The Melikane early MIS 2 lithic assemblages depict the gradual transformation of
an early microlithic technocomplex (contexts 6–8) into a fully developed Robberg
industry (context 5). There is clear intrasite continuity, with minor technological
changes accumulating gradually and asynchronously over time. We have attempted
to limit the effects of time-averaging by breaking our contexts into spits, but concede
that arbitrary levels cannot discriminate between occupation events. It is within the
realm of possibility that our data are simply too granular to detect a rapid replace-
ment event and that the gradual changes we detect are really a mix of “before” and
“after” technologies. However, technological trends at Melikane are so long-lived
that we believe this to be extremely unlikely.
Because our data support neither sudden turnover nor technological stasis, they
do not support the expectations of Bousman and Brink’s (2018) rapid replacement
Table 8 Tool type by context
*Significant in row-wise test for proportions. See OR2 for details
Tool type Context 5 Contexts 6–8
Frequency %Frequency %
Point 3 .1 4.1
Hammerstone 0 01<.1
Backed 56 2.3 102 3.0
Borer* 1012 41.6 1135 33.0
Burin 21 .9 51 1.5
Denticulate 5 0.2 11 0.3
MRP* 416 17.1 808 23.5
Nat. backed 2 <0.1 50.1
Notched 66 2.7 96 2.8
Convergent scraper 35 1.4 52 1.5
Endscraper 134 5.5 166 4.8
Nosed scraper 480 19.8 682 19.8
Sidescraper 46 1.9 62 1.8
Side/endscraper 154 6.3 261 7.6
Total 2430 3436

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 21 of 35 24
hypothesis. Rather, they match our expectations for the alternative, the LGM accel-
eration hypothesis. By focusing on a reliable yet flexible technology, Melikane for-
agers were able to adapt to a variety of circumstances and take advantage of increas-
ingly rare and valuable subsistence opportunities. Shifts in the relative proportions
of technologies in conjunction with environmental change and the continuation of
intrasite trends signify that the LSA, rather than being an invention introduced from
elsewhere, was simply the acceleration of technological trends already in progress
during contexts 6–8 and — as discussed below — the MSA. Yet, certain regional
patterns and similarities (as expected in a regional technocomplex) still indicate a
degree of population connectivity.
From Bipolar toBladelets
The most significant differences between the contexts concern bipolar reduction and
bladelet production. As a percentage of whole blanks, bladelets are equally com-
mon in contexts 6–8 and 5, but the flaking systems responsible for their production
were undergoing continued refinement. Bipolar reduction, heaviest in the heart of
contexts 6–8, becomes less frequent with the introduction of high-backed bladelet
cores and rising overall bladelet core frequencies. This asynchrony suggests that the
reduction strategy typified in quartz-poor Robberg assemblages — freehand reduc-
tion on single-platform cores (cf. Binneman & Mitchell, 1997; Deacon, 1984b;
Wadley, 1996) — did not arrive with bladelets, as would be expected with a replace-
ment event, but formed around them.
This shift in flaking systems may have been linked to raw material economy. On
chert, flint, and other cryptocrystalline silicates (CCS), freehand reduction produces
less shatter, fewer knapping errors, more usable flakes, and the same or more cut-
ting edge per unit of core mass (Morgan etal., 2015; Muller & Clarkson, 2016;
Pargeter & Eren, 2017). Bipolar reduction may have been an economical strategy
if both bladelets and “bipolar waste” (i.e., microflakes, shatter) were used on mul-
ticomponent tools (e.g., Davidson, 1934; Flood, 1995; Knutsson etal., 2016), but
freehand reduction would be more economical if bladelets were the only desired
product. Fewer step terminations in context 5 also suggest that reduction strategies
became more polished (thin flakes are particularly susceptible) or that raw mate-
rial choice became more deliberate (Cotterell & Kamminga, 1987). Cochrane (2008)
also noted that Robberg bladelets were more refined than those from the Howiesons
Poort, showing more control over flaking. Rather than citing raw material economy
as the driver, Cochrane (2008, p. 445) argues that Robberg groups dealing with high
levels of environmental risk may have held technological skill in high esteem, pro-
moting “more meticulous” bladelet production.
But why would foragers transition from a generically microlithic technological
system to one oriented specifically around bladelets? While microlithization in general
and backed microliths in particular have been studied in relation to social currency
(Ambrose, 2002; Deacon, 1992, 1995; Deacon & Wurz, 1996; but see Hiscock etal.,
2011), raw material economy (Ambrose, 2002; Mitchell, 1988a; Muller & Clarkson,
2016; Pargeter & Faith, 2020; Tryon & Faith, 2016), and risk reduction through

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 22 of 35
various mechanisms (Bleed, 1986; Clarke, 1978; Clarkson et al., 2018; Hiscock,
2002; Hiscock etal., 2011), the advantages of unbacked bladelets specifically have
gone largely unexplored(but see Mitchell, 1988a, 1988b). We propose that bladelet
technology was favored by early LGM foragers at Melikane because of its superior
versatility and effectiveness, allowing for maximum flexibility while increasing
reliability during a time of resource shortage and scheduling challenges.
Although the archaeological and ethnographic records preserve examples of
effective multicomponent weapons equipped with irregular microliths (eg. David-
son, 1934; Flood, 1995; Knutsson et al., 2016), experimental data do show that
longitudinally mounted bladelets are more lethal and reliable on projectiles than
other types or arrangements of microlithic inserts (Lombard & Parsons, 2008; Yaro-
shevich etal., 2010). Yaroshevich etal. (2010) compared different microlithic pro-
jectiles and discovered that self-pointed arrows using longitudinally hafted blade
fragments (Fig.10b) combined great penetrating power with exceptional durability.
In contrast, points equipped with microflake barbs were more prone to breakage and
had less penetrating power (Fig.10a, c). By using longitudinally hafted bladelets
instead of microflake barbs, Melikane forgers would have increased the reliability of
their weapons and decreased the time spent on repairs. Mastic on the lateral margin
of a bladelet from context 5 and usewear studies on Robberg bladelets from Sehong-
hong and Rose Cottage Cave are consistent with longitudinal orientation (Binne-
man, 1997; Binneman & Mitchell, 1997; Wadley, 1996).
The versatility of such inserts is also of obvious economic advantage and
would override the limitations of using a curated technology (Lombard & Parsons,
2008; Nelson, 1991; Vaquero & Romagnoli, 2018). Preplanned, premade curated
Fig. 10 Potential arrangements
of hafted microlithsbased on
imageand captiondescriptions
in Yaroshevich etal. (2010, p.
370 fig.1b): “oblique point”
with one barb (a), “self-pointed
arrow” with longitudinally-
mounted blades (b), “arrow with
straight point and fouroblique
barbs” (c). Line art by Bruce
Worden

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 23 of 35 24
technologies are advantageous when resources are limited, but predictable in time
and space, enabling foragers to maximize resource collection in predetermined
settings (Bleed, 1986; Bousman, 1993; Torrence, 1983; Vaquero & Romagnoli,
2018). However, users are left vulnerable when subsistence opportunities become
erratic (Nelson, 1991; Vaquero & Romagnoli, 2018). This was almost certainly the
case at Melikane: at ~24 ka, the previously productive landscape around the site
likelybecame only seasonally attractive to migratory fauna. Short-duration climate
events could also have impacted productivity. For example, heavy snow in 1996
resulted in less intense winter use of grazing posts in valleys near the Escarpment
(Nüsser, 2002).
However, by using identical bladelets to repair and refit tools with very differ-
ent purposes, foragers could adapt to unpredictable circumstances using a rela-
tively limited toolkit (Clarke, 1978; Elston & Brantingham, 2002; Mitchell, 1988a).
Regularly shaped inserts would maximize ease of maintenance — no modification
of a pre-slotted haft or point would be necessary. They would also accommodate
a more flexible mobility strategy by shrinking the size of repair kits (Kuhn, 1994)
and increasing the maintainability of all composite tools, not just projectiles (Bleed,
1986). Technological systems in the European Mesolithic exemplify the versatility
of such a system (Manninen etal., 2021). Mesolithic slotted bone tools including
inset points, daggers, and knives had myriad uses — some as projectiles, others as
harvesting implements (Bjørnevad et al., 2019; Knutsson et al., 2016; Manninen
etal., 2018, 2021). Similarly, Robberg bladelets from Sehonghong and Rose Cot-
tage Cave have usewear traces not only from hunting tasks, but also from cutting and
processing activities (Binneman, 1997; Binneman & Mitchell, 1997; Wadley, 1996).
The flexibility of this system also explains why Robberg bladelets were adopted in a
variety of environments by foragers experiencing a variety of challenges — although
bladelet technology may have mitigated risk at Melikane, its uptake elsewhere may
have been linked to different drivers (also see Hiscock etal., 2011; Manninen etal.,
2021; Pargeter & Faith, 2020).
Other differences between contexts 6–8 and 5 can be attributed to shifts in mobil-
ity patterns, as expected during a period of resource restructuring. Contexts 6–8,
with higher artifact densities, higher retouch frequencies, more bipolar reduc-
tion (especially in 6–8:2), and hyperlocal raw materials, suggest less frequent res-
idential and logistical movement (Hiscock, 1996; Parry & Kelly, 1987; Schoville
etal., 2021). Schoville etal. (2021, p. 10) suggest that residentially mobile forag-
ers employing individual provisioning can move infrequently only if “people are
essentially located on top of resources.” This may have been possible in the Leso-
tho Highlands ~27 ka, when highly productive grasslands dominated the landscape
around the shelter (Stewart & Mitchell, 2018). In contrast, lower artifact densities
and retouch frequencies, less bipolar reduction, and greater CCS use in overlying
context 5 indicate more frequent residential movement and shorter occupation dura-
tions, possibly reflecting the adoption of a seasonal mobility pattern.
Higher residential mobility in context 5 may represent the implosion of the rela-
tively sedentary strategy used by the earlier foragers. Dense, reliable resources in the
Melikane River Valley ~27 ka may have encouraged population growth and year-
round occupation, restricting movement and territory size as groups laid claim to

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 24 of 35
adjacent areas. As resources became more seasonal, groups may have faced difficult
choices about granting reciprocal access to resources, catalyzing territory defense
(Wiessner, 1982, 1994, 1986). Lowered snowlines may have exacerbated the situa-
tion by imposing natural upward boundaries. Eventually, environmental productivity
likely declined to the point where population size would have outpaced the land-
scape’s carrying capacity and forced changes to mobility patterns, potentially leav-
ing a less visible archaeological signature (Carter, 1976).
Intersite Comparisons
Penecontemporary assemblages from the broader region show similar trajectories
of gradual, intrasite change, yet still support a degree of population connectivity
(Table9). They also provide a framework in which to define the Melikane assem-
blages. Bladelet technology was also adopted incrementally at Sehonghong, a mere
24 km away, and Boomplaas, in the Cape Fold Mountain Belt. Both sites have tran-
sitional/ELSA and LGM-aged Robberg assemblages. Like contexts 6–8, the transi-
tional/ELSA contexts have more evidence for bipolar reduction, more formal tool
production, and lower bladelet core frequencies than the Robberg layers overlying
them (Mitchell, 1994, 1995, 1996; Pargeter etal., 2018; Pargeter & Faith, 2020).
Because of these parallels, we are confident that context 5 represents a fully devel-
oped expression of the Robberg industry, and contexts 6–8 represent an incipient,
or developing, Robberg technocomplex. We prefer “incipient Robberg” because we
are confident that despite surface similarities, contexts 6–8 are not part of the Bor-
der Cave ELSA technocomplex identified by Bousman and Brink (2018). They may
share a microlithic, bipolar-focused flaking system with the Border Cave ELSA, but
as Bader etal. (2022) point out for Sibebe in eSwatini, there is little evidence that
such an industry persisted over the 16,000 years separating the occupation pulses
at Border Cave and Melikane. Bipolar reduction is also an insufficient criterion for
defining an industry. Mitchell (1988a) has previously argued that high frequencies of
bipolar reduction are more related to raw material availability than cultural prefer-
ences. We likewise agree that it should not be used to link ELSA assemblages across
the subcontinent, both for this reason and because of its great antiquity in southern
Africa reaching back to the Early Stone Age (Barham, 1987).
Furthermore, numerous idiosyncrasies emphasize that contexts 6–8 and 5 are
more similar than they are different. High frequencies of formal tools distinguish the
Melikane assemblages from those at other late Pleistocene southern African sites
and run counter to definitions of ELSA and Robberg assemblages (J. Deacon, 1984a,
1984b). Importantly, however, retouch has not excluded other assemblages from
membership in either industry. For example, the Robberg at Boomplaas is nearly 10%
retouched (Deacon, 1984b), and the “truncated flake” is a retouched type peculiar to
the Robberg at Sehonghong (Mitchell, 1995 p. 33). Variation likely represents differ-
ences in mobility, site use, and provisioning patterns, all contingent on local environ-
ments and landscape structures (cf. Marean, 2016; Schoville etal., 2021), as well
as taphonomic processes biasing the assemblage towards larger artifacts. Retouch
frequencies aside, borers are site-specific and clearly comprise a critical part of the

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 25 of 35 24
Table 9 Intersite comparisons for Sehonghong, Boomplaas, and Melikane (Deacon, 1984b; Mitchell, 1994, 1995, 1996; Pargeter etal., 2018; Pargeter & Faith, 2020)
Site Level Dates Industry Bladelet % Bladelet core %
Sehonghong RFS, MOS, OS ~31–24 kcal BP MSA/LSA transition ~3.5–4.7 13.8–15 No More
BAS 24.3–23.1 kcal BP Early Robberg 13.7 24.4 Yes Less
Boomplaas LPC 26.4–24.3 kcal BP “Earliest LSA”/ELSA/Robberg Not available 12.9 (quartz only) Yes 45.2 (quartz only)
LP 23.5–21.4 kcal BP “Earliest LSA”/ELSA/Robberg 21 23.8 (quartz only) No 90
GWA/HCA 21.9–20.5 kcal BP Robberg 40 26.7 (quartz only) Yes 62
Melikane Contexts 6-8 27.1 ± 1.8 ka (spits 2–3,
OSL); 24.2–23.6 kcal BP
(spit 1)
Transitional/incipient Robberg 7.9 18.4 Yes, in spit
1 only
54.5
Context 5 < 23.8–23.3 kcal BP Robberg 7.9 28.1 Yes 44.4

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 26 of 35
Melikane toolkit (Pazan etal., 2022). A significant component of the MIS 5a, MIS
4, and Howiesons Poort assemblages, they are uncommon at other southern African
MSA and LSA sites and may be an adaptation specific to highland environments.
Despite these idiosyncrasies and lack of evidence for rapid replacement, a sim-
ilar trend towards more intentional bladelet production at penecontemporary sites
— and the presence of high-backed cores in particular — suggests that there was
some degree of population connectivity allowing for the transfer of flaking systems,
at least after ~24 kcal BP. High-backed cores are distinctive and unique, making
independent invention unlikely. Yet, they do not appear at all Robberg sites, and are
less common in post-LGM Robberg assemblages (i.e. Mitchell, 1995). Their patchy
appearance suggests that population connectivity was not subcontinent-wide, but
confined to specific regions at certain times. LGM environments were not uniform,
stochastically creating and lifting barriers to movement between the highlands and
lowlands. Periods of isolation and decreased mobility (as hypothesized for con-
texts 6–8) may have encouraged local adaptation and intrasite continuity, but epi-
sodic connectivity and greater mobility could have facilitated the transfer of flaking
systems (Mackay et al., 2014). Populations with intermediate levels of connectiv-
ity create the most innovative and complex technological solutions (Creanza etal.,
2017; Derex etal., 2018). In this case, fluctuating environments may have helped
the development of this balance — and with it the Robberg industry — in southern
Africa. Additional lines of evidence, such as beads or stable isotopes, are needed to
fully test this hypothesis.
Conclusion
Significant intrasite continuities and gradual technological change at Melikane are
inconsistent with Bousman and Brink’s (2018) rapid replacement hypothesis. We
do not intend to say that the Robberg began at Melikane, nor are we implying that
the adoption of bladelet technology in southern Africa was universally a response to
environmental change. Rather, we suggest that the subcontinent-wide fluorescence
of the Robberg was the product of sporadic information exchange and the extraordi-
nary versatility of bladelet technology, whilst its adoption at Melikane was acceler-
ated by the effects of the LGM on a vulnerable highland environment. At Melikane,
the Robberg simply served as an available and adequate solution to subsistence risk.
Considering the intrasite continuities noted at Melikane and at penecontemporary
sites, we argue that population replacement should be discarded as an explanation
for technological change across the MSA/LSA transition. In consensus with Bader
etal. (2022), we also recommend re-evaluating Border Cave as the source region
for the ELSA technocomplex. Although the lithic assemblages described therein do
contain aspects of LSA technologies, including microliths and bipolar flaking (Villa
etal., 2012), these were present at Melikane as early as MIS 5a (Pazan etal., 2022),
and the persistence of radial cores and MSA-type flakes indicates that technologi-
cal change was incomplete (Backwell etal., 2018; Beaumont, 1978). Furthermore,
Border Cave’s abandonment until the Iron Age prevents a full understanding of the
direction of technological change. The southern African MSA is full of false starts

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 27 of 35 24
towards microlithization (i.e., the Howiesons Poort), leaving the significance of
Border Cave’s abbreviated technological sequence unresolved.In this vein, Bader
etal.’s (2022) suggestion that the Border Cave “ELSA” was simply a variant of the
final MSA serves as a viable null hypothesis.
Accurately describing southern African lithic technologies from early MIS 2
remains a challenge. A number of archaeologists (Barham, 1989; Carter etal., 1988;
Clark, 1997; Kaplan, 1989; Mitchell, 1994; Mitchell & Vogel, 1994) have advocated
discarding “ELSA” and in favor of the term “transitional,” whereas Bader et al.
(2022) eschew “transitional” and group these assemblages (including those from
Melikane, Sehonghong, and RCC) with the final MSA. However, we argue that even
“transitional” – and certainly not final MSA – accurately describes assemblages
dating to ~29-24 kcal BP. The pre-LGM assemblages from Melikane, Boomplaas,
and Sehonghong are all tightly linked to the fully established Robberg assemblages
that post-date them, but are incipient in their levels of development. A clear focus
on bladelet technology and loss of MSA-type artifacts distinguishes them from
assemblages we would argue are more justifiably termed transitional, such as the
late MIS 3 assemblages at Umhlatuzana and Rose Cottage Cave, which retain con-
siderable numbers of MSA tool types and prepared core technology (Clark, 1997;
Kaplan, 1990; Loftus etal., 2019). Mitchell (1988a, 1988b) has also proposed the
term “early microlithic” to refer to ELSA and Robberg assemblages from the Late
Pleistocene, a term which we agree accurately describes contexts 6–8 at Melikane.
However, rather than implementing new terminology, we encourage the continued
exploration of long-term trends and recognition of intrasite continuities, as well as
an expanded concept of what “Robberg” looks like. The technocomplex endured
for over 10,000 years in a diversity of southern African environments — we should
expect to see temporal change and spatial heterogeneity, even during times of popu-
lation connectivity.
Supplementary Information The online version contains supplementary material available at https:// doi.
org/ 10. 1007/ s41982- 023- 00150-2.
Acknowledgements Melikane was excavated under permits granted by the Ministry of Technology,
Environment and Culture (MTEC), Kingdom of Lesotho. We extend thanks to MTEC, and particularly
bo-Mme Matsosane Molibeli, Moitheri Molibeli, Moliehi Ntene, Tsepang Shano, Puseletso Moremi, and
the late Ntsema Khitsane, for their continued support. We are grateful for the hospitality of the Melikane
villagers, especially Morena Pule Machine. We are indebted to Mme Refiloe Okello for her friendship
and support. We appreciatively acknowledge the assistance of Jim Moss and John Klausmeyer of the
University of Michigan Museum of Anthropological Archaeology (UMMAA) for facilitating access to
photographic equipment. We are extremely grateful to Bruce Worden, UMMAA’s illustrator, for his won-
derful work on Fig.10.
Author Contribution K. Pazan conducted the lithic analysis, statistical analyses, and wrote the manu-
script, with contributions from B. A. Stewart. B. A. Stewart and G. Dewar secured funding, conducted
excavations, and obtained dates and paleoenvironment data. All authors reviewed the manuscript.
Funding Fieldwork at Melikane was supported by generous grants from the Wenner-Gren Foundation,
the British Academy, the Prehistoric Society, and the McDonald Institute for Archaeological Research
and the Centre of African Studies (University of Cambridge).
Data Availability The data that support the findings of this study are available from the corresponding
authors upon request.

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 28 of 35
Declarations
Ethics Approval Not applicable.
Competing Interests The authors declare no competing interests.
References
Ambrose, S. H. (2002). Small things remembered: Origins of early microlithic industries in Sub-Saharan
Africa. Archaeological Papers of the American Anthropological Association, 12(1), 9–29. https://
doi. org/ 10. 1525/ ap3a. 2002. 12.1.9
Backwell, L. R., d’Errico, F., Banks, W. E., de la Peña, P., Sievers, C., Stratford, D., Lennox, S. J.,
Wojcieszak, M., Bordy, E. M., Bradfield, J., & Wadley, L. (2018). New excavations at Border Cave,
KwaZulu-Natal, South Africa. Journal of Field Archaeology, 43(6), 417–436. https:// doi. org/ 10.
1080/ 00934 690. 2018. 15045 44
Bader, G. D., Mabuza, A., Price Williams, D., & Will, M. (2022). Rethinking the middle to later stone
age transition in southern Africa—A perspective from the highveld of Eswatini. Quaternary Sci-
ence Reviews, 286, 107540. https:// doi. org/ 10. 1016/j. quasc irev. 2022. 107540
Barham, L. S. (1987). The bipolar technique in Southern Africa: A replication experiment. The South
African Archaeological Bulletin, 42(145), 45–50. https:// doi. org/ 10. 2307/ 38877 73
Barham, L. S. (1989). A preliminary report on the Later Stone Age artefacts from Siphiso Shelter in Swa-
ziland. The South African Archaeological Bulletin, 33–43.
Beaumont, P. B. (1978). Border Cave [Unpublished M.A. thesis]. University of Cape Town.
Behar, D. M., van Oven, M., Rosset, S., Metspalu, M., Loogväli, E.-L., Silva, N. M., Kivisild, T., Torroni,
A., & Villems, R. (2012). A “Copernican” reassessment of the human mitochondrial DNA tree
from its root. The American Journal of Human Genetics, 90(4), 675–684. https:// doi. org/ 10. 1016/j.
ajhg. 2012. 03. 002
Binneman, J. N. F. (1997). Usewear traces on Robberg bladelets from Rose Cottage Cave. South African
Journal of Science, 93(10), 479–481.
Binneman, J. N. F., & Mitchell, P. J. (1997). Usewear analysis of Robberg bladelets from Sehonghong
shelter, Lesotho. Southern African Field Archaeology, 6, 42–49.
Bjørnevad, M., Jonuks, T., Bye-Jensen, P., Manninen, M. A., Oras, E., Vahur, S., & Riede, F. (2019). The
life and times of an Estonian Mesolithic slotted bone “dagger”: Extended object biographies for
legacy objects. Estonian Journal of Archaeology, 23(2), 103. https:// doi. org/ 10. 3176/ arch. 2019.2.
02
Bleed, P. (1986). The optimal design of hunting weapons: Maintainability or reliability. American Antiq-
uity, 51(4), 737–747. https:// doi. org/ 10. 2307/ 280862
Bordy, E. M., & Head, H. V. (2018). Lithostratigraphy of the Clarens formation (Stormberg Group,
Karoo Supergroup), South Africa. South African Journal of Geology, 121(1), 119–130. https:// doi.
org/ 10. 25131/ sajg. 121. 0009
Bousman, C. B. (1993). Hunter-gatherer adaptations, economic risk and tool design. Lithic Technology,
18(1–2), 59–86. https:// doi. org/ 10. 1080/ 01977 261. 1993. 11720 897
Bousman, C. B. (2005). Coping with risk: Later stone age technological strategies at Blydefontein Rock
Shelter, South Africa. Journal of Anthropological Archaeology, 24(3), 193–226. https:// doi. org/ 10.
1016/j. jaa. 2005. 05. 001
Bousman, C. B., & Brink, J. S. (2018). The emergence, spread, and termination of the Early Later Stone
Age event in South Africa and southern Namibia. Quaternary International, 495, 116–135. https://
doi. org/ 10. 1016/j. quaint. 2017. 11. 033
Bradley, B. A., & Giria, Y. (1996). Concepts of the technological analysis of flaked stone: A case study
from the High Arctic. Lithic Technology, 21(1), 23–39.
Bregman, A., & Knight, J. (2022). Analysis of a blockstream in the northern Lesotho Drakensberg, south-
ern Africa. Quaternary International, 611, 41–54.
Brock, F., Higham, T., Ditchfield, P., & Ramsey, C. B. (2010). Current pretreatment methods for AMS
radiocarbon dating at the Oxford radiocarbon accelerator Unit (Orau). Radiocarbon, 52(1), 103–
112. https:// doi. org/ 10. 1017/ S0033 82220 00450 69

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 29 of 35 24
Callahan, E. (1987). An evaluation of the lithic technology in Middle Sweden during the Mesolithic and
Neolithic. Societas Archaeologica Upsaliensis.
Carter, P. L. (1976). The effects of climatic change on settlement in Eastern Lesotho during the Middle
and Later Stone Age. World Archaeology, 8(2), 197–206.
Carter, P. L. (1978). The prehistory of Eastern Lesotho. [Ph.D. thesis].University of Cambridge.
Carter, P. L., Mitchell, P. J., & Vinnicombe, P. (1988). Sehonghong: The Middle and Later Stone Age
industrial sequence at a Lesotho Rockshelter. BAR Publishing.
Chan, E. K. F., Hardie, R.-A., Petersen, D. C., Beeson, K., Bornman, R. M. S., Smith, A. B., & Hayes, V.
M. (2015). Revised timeline and distribution of the earliest diverged human maternal lineages in
Southern Africa. PLOS ONE, 10(3), e0121223. https:// doi. org/ 10. 1371/ journ al. pone. 01212 23
Clark, A. M. B. (1997). The MSA/LSA transition in Southern Africa: New technological evidence from
Rose Cottage Cave. The South African Archaeological Bulletin, 52(166), 113–121. https:// doi. org/
10. 2307/ 38890 76
Clark, J. D. (1959). The prehistory of Southern Africa. Penguin Books.
Clark, J. D. (1974). Kalambo Falls prehistoric site. Vol. II: The late prehistoric remains. Cambridge Uni-
versity Press.
Clark, P. U., Dyke, A. S., Shakun, J. D., Carlson, A. E., Clark, J., Wohlfarth, B., Mitrovica, J. X.,
Hostetler, S. W., & McCabe, A. M. (2009). The Last Glacial Maximum. Science, 325(5941), 710–
714. https:// doi. org/ 10. 1126/ scien ce. 11728 73
Clark, P. U., & Mix, A. C. (2002). Ice sheets and sea level of the Last Glacial Maximum. Quaternary Sci-
ence Reviews, 21(1), 1–7. https:// doi. org/ 10. 1016/ S0277- 3791(01) 00118-4
Clarke, D. L. (1978). Mesolithic Europe: The economic basis. Duckworth.
Clarkson, C., Hiscock, P., Mackay, A., & Shipton, C. (2018). Small, sharp, and standardized: Global
convergence in backed-microlith technology. In M. J. O’Brien, B. Buchanan, & M. I. Eren (Eds.),
Convergent evolution in stone tool technology (pp. 175–200). MIT Press.
Cochrane, G. W. G. (2008). A comparison of Middle Stone Age and Later Stone Age blades from South
Africa. Journal of Field Archaeology, 33(4), 429–448. https:// doi. org/ 10. 1179/ 00934 69087 91071
132
Conard, N. J., Soressi, M., Parkington, J. E., Wurz, S., & Yates, R. (2004). A unified lithic taxonomy
based on patterns of core reduction. The South African Archaeological Bulletin, 59(179), 12.
https:// doi. org/ 10. 2307/ 38893 18
Cotterell, B., & Kamminga, J. (1987). The formation of flakes. American Antiquity, 52(4), 675–708.
https:// doi. org/ 10. 2307/ 281378
Creanza, N., Kolodny, O., & Feldman, M. W. (2017). Greater than the sum of its parts? Modelling
population contact and interaction of cultural repertoires. Journal of The Royal Society Interface,
14(130), 20170171. https:// doi. org/ 10. 1098/ rsif. 2017. 0171
d’Errico, F., Backwell, L., Villa, P., Degano, I., Lucejko, J. J., Bamford, M. K., Higham, T. F. G., Colom-
bini, M. P., & Beaumont, P. B. (2012). Early evidence of San material culture represented by
organic artifacts from Border Cave, South Africa. Proceedings of the National Academy of Sci-
ences, 109(33), 13214–13219. https:// doi. org/ 10. 1073/ pnas. 12042 13109
Davidson, D. S. (1934). Australian spear-traits and their derivations (Continued). The Journal of the Poly-
nesian Society, 43(3(171)), 143–162.
Deacon, H. J. (1976). Where hunters gathered: A study of Holocene Later Stone Age people in the eastern
Cape. South African Archaeological Society.
Deacon, H. J. (1992). Southern Africa and modern human origins. Philosophical Transactions of the
Royal Society of London. Series B: Biological Sciences, 337(1280), 177–183.
Deacon, H. J. (1995). Two Late Pleistocene-Holocene archaeological depositories from the Southern
Cape, South Africa. The South African Archaeological Bulletin, 50(162), 121–131. https:// doi. org/
10. 2307/ 38890 61
Deacon, H. J., & Wurz, S. (1996). Klasies River main site, cave 2: A Howiesons Poort occurrence.
Aspects of African Archaeology (pp. 213–218). University of Zimbabwe Publications.
Deacon, J. (1978). Changing patterns in the late Pleistocene/early Holocene prehistory of Southern Africa
as seen from the Nelson Bay Cave Stone artifact sequence. Quaternary Research, 10(1), 84–111.
https:// doi. org/ 10. 1016/ 0033- 5894(78) 90015-7
Deacon, J. (1982). The Later Stone Age in Southern Cape, South Africa. [Ph.D. thesis]. University of
Cape Town.
Deacon, J. (1984a). Southern African Prehistory and Paleoenvironments. In R. G. Klein (Ed.), Later
Stone Age people and their descendants in southern Africa (pp. 221–328). AA Balkema.

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 30 of 35
Deacon, J. (1984b). The Later Stone Age of Southernmost Africa. BAR Publishing.
Derex, M., Perreault, C., & Boyd, R. (2018). Divide and conquer: Intermediate levels of population frag-
mentation maximize cultural accumulation. Philosophical Transactions of the Royal Society B:
Biological Sciences, 373(1743), 20170062. https:// doi. org/ 10. 1098/ rstb. 2017. 0062
Diez-Martín, F., Yustos, P. S., Domínguez-Rodrigo, M., & Prendergast, M. E. (2011). An experimental
study of bipolar and freehand knapping of Naibor Soit Quartz from Olduvai Gorge (Tanzania).
American Antiquity, 76(4), 690–708. https:// doi. org/ 10. 7183/ 0002- 7316. 76.4. 690
Doelman, T. (2008). Flexibility and creativity in microblade core manufacture in Southern Primorye, Far
East Russia. Asian Perspectives, 47(2), 352–370. https:// doi. org/ 10. 1353/ asi.0. 0006
Ehleringer, J. R., Cerling, T. E., & Dearing, M. D. (2002). Atmospheric CO2 as a global change driver
influencing plant-animal interactions. Integrative and Comparative Biology, 42(3), 424–430.
Ehleringer, J. R., Cerling, T. E., & Helliker, B. R. (1997). C4 photosynthesis, atmospheric CO2, and cli-
mate. Oecologia, 112(3), 285–299. https:// doi. org/ 10. 1007/ s0044 20050 311
Elston, R. G., & Brantingham, P. J. (2002). Microlithic technology in Northern Asia: A risk-minimiz-
ing strategy of the Late Paleolithic and Early Holocene. Archaeological Papers of the American
Anthropological Association, 12(1), 103–116. https:// doi. org/ 10. 1525/ ap3a. 2002. 12.1. 103
Flood, J. (1995). Archaeology of the dreamtime: The story of prehistoric Australia and its people.
HarperCollins.
Fritzen, D. K. (1959). The rock-hunter’s field manual: A guide to identification of rocks and minerals (1st
ed.) http:// hdl. handle. net/ 2027/ mdp. 39015 00684 1707
Gliozzo, E., Cairncross, B., & Vennemann, T. (2019). A geochemical and micro-textural comparison of
basalt-hosted chalcedony from the Jurassic Drakensberg and Neoarchean Ventersdorp Supergroup
(Vaal River alluvial gravels), South Africa. International Journal of Earth Sciences, 108(6), 1857–
1877. https:// doi. org/ 10. 1007/ s00531- 019- 01737-3
Goebel, T. (2002). The “microblade adaptation” and recolonization of Siberia during the Late Upper
Pleistocene. Archaeological Papers of the American Anthropological Association, 12(1), 117–131.
https:// doi. org/ 10. 1525/ ap3a. 2002. 12.1. 117
Goodwin, A. J. H., & Van Riet Lowe, C. (1929). The Stone Age cultures of South Africa. Neill.
Grab, S., Knight, J., Mol, L., Botha, T., Carbutt, C., & Woodborne, S. (2021). Periglacial landforms in the
high Drakensberg, Southern Africa: Morphogenetic associations with rock weathering rinds and
shrub growth patterns. Geografiska Annaler: Series A, Physical Geography, 1–24. https:// doi. org/
10. 1080/ 04353 676. 2020. 18566 25
Grab, S. W. (1996). Debris deposits in the high Drakensberg, South Africa: Possible indicators for pla-
teau, niche and cirque glaciation. Zeitschrift Für Geomorphologie. Supplementband, 103, 389–403.
Grab, S. W. (1997). Thermal regime for a thufa apex and its adjoining depression, Mashai Valley, Leso-
tho. Permafrost and Periglacial Processes, 8(4), 437–445. https:// doi. org/ 10. 1002/ (SICI) 1099-
1530(199710/ 12)8: 4% 3C437:: AID- PPP264% 3E3.0. CO;2-O
Gurtov, A. N., & Eren, M. I. (2014). Lower Paleolithic bipolar reduction and hominin selection of quartz
at Olduvai Gorge, Tanzania: What’s the connection? Quaternary International, 322–323, 285–291.
https:// doi. org/ 10. 1016/j. quaint. 2013. 08. 010
Hiscock, P. (1996). Mobility and technology in the Kakadu coastal wetlands. Bulletin of the Indo-Pacific
Prehistory Association, 15, 151–157.
Hiscock, P. (2002). Pattern and context in the Holocene Proliferation of backed artifacts in Australia.
Archaeological Papers of the American Anthropological Association, 12(1), 163–177. https:// doi.
org/ 10. 1525/ ap3a. 2002. 12.1. 163
Hiscock, P., Clarkson, C., & Mackay, A. (2011). Big debates over little tools: Ongoing disputes over
microliths on three continents. World Archaeology, 43(4), 653–664.
Holmgren, K., Lee-Thorp, J. A., Cooper, G. R. J., Lundblad, K., Partridge, T. C., Scott, L., Sithaldeen, R.,
Siep Talma, A., & Tyson, P. D. (2003). Persistent millennial-scale climatic variability over the past
25,000 years in Southern Africa. Quaternary Science Reviews, 22(21), 2311–2326. https:// doi. org/
10. 1016/ S0277- 3791(03) 00204-X
Humphreys, A. J. B. (1972). Comments on aspects of raw material usage in the Later Stone Age of the
Middle Orange River area. Goodwin Series, 1, 46–53. https:// doi. org/ 10. 2307/ 38580 92
Iwase, A. (2016). A functional analysis of the LGM microblade assemblage in Hokkaido, northern Japan:
A case study of Kashiwadai 1. Quaternary International, 425, 140–157. https:// doi. org/ 10. 1016/j.
quaint. 2016. 04. 008
Jacobs, Z., Roberts, R. G., Galbraith, R. F., Deacon, H. J., Grun, R., Mackay, A., Mitchell, P., Vogel-
sang, R., & Wadley, L. (2008). Ages for the Middle Stone Age of Southern Africa: Implications

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 31 of 35 24
for human behavior and dispersal. Science, 322(5902), 733–735. https:// doi. org/ 10. 1126/ scien ce.
11622 19
Kaplan, J. (1990). The Umhlatuzana Rock Shelter sequence: 100 000 years of Stone Age history. South-
ern African Humanities, 2(11), 1–94.
Kaplan, J. M. (1989). 45000 years of hunter-gatherer history in Natal as seen from Umhlatuzana Rock
Shelter. Goodwin Series, 6, 7–16. https:// doi. org/ 10. 2307/ 38581 28
Killick, D. J. B. (1978). Further data on the climate of the Alpine Vegetation Belt of eastern Lesotho.
Bothalia, 12, 567–572.
Knutsson, H., Knutsson, K., Molin, F., & Zetterlund, P. (2016). From flint to quartz: Organization of
lithic technology in relation to raw material availability during the pioneer process of Scandinavia.
Quaternary International, 424, 32–57. https:// doi. org/ 10. 1016/j. quaint. 2015. 10. 062
Kuhn, S. L. (1994). A formal approach to the design and assembly of mobile toolkits. American Antiq-
uity, 59(3), 426–442. https:// doi. org/ 10. 2307/ 282456
Kuhn, S. L., & Elston, R. G. (2002). Introduction: Thinking small globally. Archaeological Papers of the
American Anthropological Association, 12(1), 1–7. https:// doi. org/ 10. 1525/ ap3a. 2002. 12.1.1
Kulongoski, J. T., Hilton, D. R., & Selaolo, E. T. (2004). Climate variability in the Botswana Kalahari
from the late Pleistocene to the present day. Geophysical Research Letters, 31(10). https:// doi. org/
10. 1029/ 2003G L0192 38
Loftus, E., Pargeter, J., Mackay, A., Stewart, B. A., & Mitchell, P. (2019). Late Pleistocene human occu-
pation in the Maloti-Drakensberg region of southern Africa: New radiocarbon dates from Rose
Cottage Cave and inter-site comparisons. Journal of Anthropological Archaeology, 56, 101117.
https:// doi. org/ 10. 1016/j. jaa. 2019. 101117
Lombard, M., Bradfield, J., Caruana, M. V., Makhubela, T. V., Dusseldorp, G. L., Kramers, J. D., &
Wurz, S. (2022). The South African Stone Age sequence updated (II). South African Archaeologi-
cal Bulletin, 77(217), 172–212.
Lombard, M., & Parsons, I. (2008). Blade and bladelet function and variability in risk management dur-
ing the last 2000 years in the Northern Cape. The South African Archaeological Bulletin, 63(187),
18–27. https:// doi. org/ 10. 2307/ 20474 988
Lombard, M., Wadley, L., Deacon, J., Wurz, S., Parsons, I., Mohapi, M., Swart, J., & Mitchell, P. (2012).
South African and Lesotho Stone Age sequence updated. The South African Archaeological Bul-
letin, 67(195), 123–144.
Mackay, A., Stewart, B. A., & Chase, B. M. (2014). Coalescence and fragmentation in the late Pleisto-
cene archaeology of southernmost Africa. Journal of Human Evolution, 72, 26–51. https:// doi. org/
10. 1016/j. jhevol. 2014. 03. 003
Manninen, M. A., Asheichyk, V., Jonuks, T., Kriiska, A., Osipowicz, G., Sorokin, A. N., Vashanau,
A., Riede, F., & Persson, P. (2021). Using radiocarbon dates and tool design principles to assess
the role of composite slotted bone tool technology at the intersection of adaptation and culture-
history. Journal of Archaeological Method & Theory, 28(3), 845–870. https:// doi. org/ 10. 1007/
s10816- 021- 09517-7
Manninen, M. A., Hertell, E. J., Pesonen, P. A. P., & Tallavaara, M. (2018). Postglacial pioneer colonisa-
tion of Eastern Fennoscandia: Modeling technological change. In K. Knutsson, H. Knutsson, J.
Apel, & H. Glørstad (Eds.), Technology of early settlement in Northern Europe—Transmission of
knowledge and culture (Vol. 2, pp. 23–46). Equinox Publishing. https:// doi. org/ 10. 1558/ equin ox.
30723
Marean, C. W. (2016). The transition to foraging for dense and predictable resources and its impact on the
evolution of modern humans. Philosophical Transactions of the Royal Society B: Biological Sci-
ences, 371(1698), 20150239. https:// doi. org/ 10. 1098/ rstb. 2015. 0239
Mills, S. C., & Grab, S. W. (2005). Debris ridges along the southern Drakensberg escarpment as evidence
for Quaternary glaciation in southern Africa. Quaternary International, 129(1), 61–73. https:// doi.
org/ 10. 1016/j. quaint. 2004. 04. 007
Mitchell, P. J. (1988a). The Early Microlithic assemblages of Southern Africa. BAR Publishing. https://
doi. org/ 10. 30861/ 97808 60545 026
Mitchell, P. J. (1988b). The late Pleistocene early microlithic assemblages of southern Africa. World
Archaeology, 20(1), 27–39. https:// doi. org/ 10. 1080/ 00438 243. 1988. 99800 54
Mitchell, P. J. (1994). Understanding the MSA/LSA Transition: The pre-20,000 BP assemblages from
new excavations at Sehonghong Rock Shelter, Lesotho. Southern African Field Archaeology, 3,
15–25.

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 32 of 35
Mitchell, P. J. (1995). Revisiting the Robberg: New results and a revision of old ideas at Sehonghong
Rock Shelter. Lesotho. The South African Archaeological Bulletin, 50(161), 28–38. https:// doi. org/
10. 2307/ 38892 72
Mitchell, P. J. (1996). The late Quaternary landscape at Sehonghong in the Lesotho highlands, southern
Africa. Antiquity, 70(269), 623–638. https:// doi. org/ 10. 1017/ S0003 598X0 00837 57
Mitchell, P. J., & Vogel, J. C. (1994). New radiocarbon dates from Sehonghong rock shelter, Lesotho.
South African Journal of Science, 90(5), 284–288.
Morgan, B., Eren, M. I., Khreisheh, N., Hill, G., Bradley, B., Jennings, T., & Smallwood, A. (2015). Clo-
vis bipolar lithic reduction at Paleo Crossing, Ohio: A reinterpretation based on the examination of
experimental replications. In Clovis: On the Edge of a New Understanding (pp. 121–143). Texas
A&M Press.
Mucina, L., & Rutherford, M. C. (Eds.). (2006). The vegetation of South Africa, Lesotho and Swaziland.
South African National Biodiversity Institute.
Muller, A., & Clarkson, C. (2016). Identifying major transitions in the evolution of lithic cutting edge
production rates. Plos One, 11(12), e0167244. https:// doi. org/ 10. 1371/ journ al. pone. 01672 44
Nelson, M. C. (1991). The study of technological organization. Archaeological Method and Theory, 3,
57–100.
Nüsser, M. (2002). Pastoral utilization and land cover change: A case study from the Sanqebethu Valley,
Eastern Lesotho Weidenutzung und Landschaftsveränderungen: Ein Fallbeispiel aus dem Sanqe-
bethu-Tal im östlichen Lesotho. Erdkunde, 56(2), 207–221.
Orlandi, F., Vazquez, L. M., Ruga, L., Bonofiglio, T., Fornaciari, M., Garcia-Mozo, H., Domínguez, E.,
Romano, B., & Galan, C. (2005). Bioclimatic requirements for olive flowering in two Mediter-
ranean regions located at the same latitude (Andalucia, Spain and Sicily, Italy). Annals of Agricul-
tural and Environmental Medicine, 12(1), 47–52.
Otsuka, Y. (2017). The background of transitions in microblade industries in Hokkaido, northern
Japan. Quaternary International, 442, 33–42. https:// doi. org/ 10. 1016/j. quaint. 2016. 07. 023
Otunga, C., Odindi, J., Mutanga, O., Adjorlolo, C., & Botha, C. (2016). Predicting the distribution
of C3 (Festuca spp.) grass species using topographic variables and binary logistic regression
model. Geocarto International, 1–36. https:// doi. org/ 10. 1080/ 10106 049. 2016. 12655 98
Pargeter, J. (2017). Lithic miniaturization in Late Pleistocene Southern Africa. Ph.D. thesis, State
University of New York at Stony Brook http:// search. proqu est. com/ docvi ew/ 19329 98079/ abstr
act/ 994B7 CA621 87490 EPQ/1
Pargeter, J., & Eren, M. I. (2017). Quantifying and comparing bipolar versus freehand flake mor-
phologies, production currencies, and reduction energetics during lithic miniaturization. Lithic
Technology, 42(2–3), 90–108. https:// doi. org/ 10. 1080/ 01977 261. 2017. 13454 42
Pargeter, J., & Faith, J. T. (2020). Lithic miniaturization as adaptive strategy: A case study from
Boomplaas Cave. South Africa. Archaeological and Anthropological Sciences, 12(9), 1–13.
Pargeter, J., Loftus, E., Mackay, A., Mitchell, P., & Stewart, B. A. (2018). New ages from Boomplaas
Cave, South Africa, provide increased resolution on late/terminal Pleistocene human behav-
ioural variability. Azania: Archaeological Research in Africa, 53(2), 156–184. https:// doi. org/
10. 1080/ 00672 70X. 2018. 14367 40
Pargeter, J., Loftus, E., & Mitchell, P. (2017). New ages from Sehonghong rock shelter: Implications
for the late Pleistocene occupation of highland Lesotho. Journal of Archaeological Science:
Reports, 12, 307–315. https:// doi. org/ 10. 1016/j. jasrep. 2017. 01. 027
Pargeter, J., & Redondo, M. (2016). Contextual approaches to studying unretouched bladelets: A late
Pleistocene case study at Sehonghong Rockshelter, Lesotho. Quaternary International, 404,
30–43. https:// doi. org/ 10. 1016/j. quaint. 2015. 08. 038
Parker, A. G., Lee-Thorp, J., & Mitchell, P. J. (2011). Late Holocene Neoglacial conditions from the
Lesotho highlands, southern Africa: Phytolith and stable carbon isotope evidence from the
archaeological site of Likoaeng. Proceedings of the Geologists’ Association, 122(1), 201–211.
https:// doi. org/ 10. 1016/j. pgeola. 2010. 09. 005
Parry, W. J., & Kelly, R. I. (1987). In J. Johnson & C. Morrow (Eds.), Expedient core technology and
sedentism (pp. 285–304). The organization of core technology.
Partridge, T. C., Scott, L., & Hamilton, J. E. (1999). Synthetic reconstructions of southern African
environments during the Last Glacial Maximum (21–18kyr) and the Holocene Altithermal
(8–6kyr). Quaternary International, 57–58, 207–214. https:// doi. org/ 10. 1016/ S1040- 6182(98)
00061-5

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 33 of 35 24
Patalano, R., Arthur, C., Carleton, W. C., Challis, S., Dewar, G., Gayantha, K., Gleixner, G., Ilgner,
J., Lucas, M., Marzo, S., Mokhachane, R., Pazan, K., Spurite, D., Morley, M. W., Parker, A.,
Mitchell, P., Stewart, B. A., & Roberts, P. (2023). Ecological stability of Late Pleistocene-
to-Holocene Lesotho, southern Africa, facilitated human upland habitation. Communications
Earth & Environment, 4(1). https:// doi. org/ 10. 1038/ s43247- 023- 00784-8
Pazan, K. R., Dewar, G., & Stewart, B. A. (2022). The MIS 5a (~80 ka) Middle Stone Age lithic assem-
blages from Melikane Rockshelter, Lesotho: Highland adaptation and social fragmentation. Qua-
ternary International, 611–612, 119–137. https:// doi. org/ 10. 1016/j. quaint. 2020. 11. 046
Phillipson, D. W. (1976). The prehistory of Eastern Zambia. Memoirs of the British Institute of East-
ern Africa, 6, 1–229.
Phillipson, D. W. (1977). The Later Prehistory of Eastern and Southern Africa. Heinemann.
Polanská, M., Hromadová, B., & Sázelová, S. (2021). The Upper and Final Gravettian in Western Slo-
vakia and Moravia. Different approaches, new questions. Quaternary International, 581–582,
205–224. https:// doi. org/ 10. 1016/j. quaint. 2020. 08. 004
Porraz, G., Igreja, M., Schmidt, P., & Parkington, J. (2016). A shape to the microlithic Robberg from
Elands Bay Cave (South Africa). Southern African Humanities, 29, 203–247.
Ramsey, C. B. (2009). Bayesian analysis of radiocarbon dates. Radiocarbon, 51(1), 337–360.
R Core Team. (2021). R: A language and environment for statistical computing (4.0.5). Foundation for
Statistical Computing. https:// www.R- proje ct. org/
Reimer, P. J., Baillie, M. G., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Ramsey, C. B., Buck,
C. E., Burr, G. S., & Edwards, R. L. (2009). IntCal09 and Marine09 radiocarbon age calibration
curves, 0–50,000 years cal BP. Radiocarbon, 51(4), 1111–1150.
Roffe, S. J., Fitchett, J. M., & Curtis, C. J. (2019). Classifying and mapping rainfall seasonality in South
Africa: A review. South African Geographical Journal, 101(2), 158–174. https:// doi. org/ 10. 1080/
03736 245. 2019. 15731 51
Sampson, C. G. (1968). Middle Stone Age industries of the Orange River scheme area (Vol. 4). National
Museum.
Sampson, C. G. (1974). The Stone Age archaeology of southern Africa. Academic Press.
Sampson, C. G., & Sampson, M. (1967). Riversmead Shelter: Excavations and Analysis. National Museum.
Schmid, V. C., Roussel, M., Abruzzese, T., Sifogeorgaki, I., & Dusseldorp, G. L. (2022). The role of anvil
and bipolar flaking on quartz in the South African Robberg. In 6th International Conference of
Experimental Archaeology. Pézénas.
Schoville, B. J., Brown, K. S., & Wilkins, J. (2021). A lithic provisioning model as a proxy for landscape
mobility in the Southern and Middle Kalahari. Journal of Archaeological Method and Theory.
https:// doi. org/ 10. 1007/ s10816- 021- 09507-9
Seltzer, A. M., Ng, J., Aeschbach, W., Kipfer, R., Kulongoski, J. T., Severinghaus, J. P., & Stute, M.
(2021). Widespread six degrees Celsius cooling on land during the Last Glacial Maximum. Nature,
593(7858), 228–232. https:// doi. org/ 10. 1038/ s41586- 021- 03467-6
Soffer, O., & Gamble, C. (Eds.). (1990a). The world at 18 000 BP: Vol. 1. High Latitudes. Unwin Hyman.
Soffer, O., & Gamble, C. (Eds.). (1990b). The world at 18 000 BP: Vol. 2. Low Latitudes. Unwin Hyman.
Stewart, B. A., & Mitchell, P. (2018). Beyond the shadow of a desert: aquatic resource intensification on
the roof of southern Africa. In A. K. Lemke (Ed.), Foraging in the past: archaeological studies
of hunter-gatherer diversity (pp. 159–208). University Press of Colorado. https:// doi. org/ 10. 5876/
97816 07327 745. c006
Stewart, B. A., Dewar, G. I., Morley, M. W., Inglis, R. H., Wheeler, M., Jacobs, Z., & Roberts, R. G.
(2012). Afromontane foragers of the Late Pleistocene: Site formation, chronology and occupational
pulsing at Melikane Rockshelter, Lesotho. Quaternary International, 270, 40–60. https:// doi. org/
10. 1016/j. quaint. 2011. 11. 028
Stewart, B. A., Parker, A. G., Dewar, G., Morley, M. W., & Allott, L. F. (2016). Follow the Senqu:
Maloti-Drakensberg paleoenvironments and implications for early human dispersals into mountain
systems. In S. C. Jones & B. A. Stewart (Eds.), Africa from MIS 6-2: Population dynamics and
paleoenvironments (pp. 247–271). Springer. https:// doi. org/ 10. 1007/ 978- 94- 017- 7520-5_ 14
Straus, L. G. (2015). The human occupation of Southwestern Europe during the Last Glacial Maximum:
Solutrean cultural adaptations in France and Iberia. Journal of Anthropological Research, 71(4),
465–492. https:// doi. org/ 10. 3998/ jar. 05210 04. 0071. 401
Straus, L. G. (2016). Humans confront the Last Glacial Maximum in Western Europe: Reflections on
the Solutrean weaponry phenomenon in the broader contexts of technological change and cultural
adaptation. Quaternary International, 425, 62–68. https:// doi. org/ 10. 1016/j. quaint. 2016. 01. 059

Journal of Paleolithic Archaeology (2023) 6:24
1 3
24 Page 34 of 35
Stute, M., & Talma, A. S. (1998). Glacial temperatures and moisture transport regimes reconstructed
from noble gases and delta-O-18, Stampriet aquifer, Namibia. In International symposium on iso-
tope techniques in the study of past and current environmental changes in the hydrosphere and the
atmosphere https:// www. osti. gov/ etdew eb/ biblio/ 651671
Tomasso, A., Rots, V., Purdue, L., Beyries, S., Buckley, M., Cheval, C., Cnuts, D., Coppe, J., Julien, M.-A.,
Grenet, M., Lepers, C., & M’hamdi, M., Simon, P., Sorin, S., & Porraz, G. (2018). Gravettian weap-
onry: 23,500-year-old evidence of a composite barbed point from Les Prés de Laure (France). Jour-
nal of Archaeological Science, 100, 158–175. https:// doi. org/ 10. 1016/j. jas. 2018. 05. 003
Torrence, R. (1983). Time budgeting and hunter-gatherer technology. In G. Bailey (Ed.), Hunter-gatherer
economy in Prehistory: A European perspective (pp. 11–22). Cambridge University Press.
Tryon, C. A., & Faith, J. T. (2016). A demographic perspective on the Middle to Later Stone Age transi-
tion from Nasera rockshelter, Tanzania. Philosophical Transactions of the Royal Society B: Bio-
logical Sciences, 371(1698), 20150238. https:// doi. org/ 10. 1098/ rstb. 2015. 0238
Twiss, P. C. (1992). Predicted world distribution of C3 and C4 grass phytoliths. In G. Rapp & S. C. Mul-
holland (Eds.), Phytolith systematics: Emerging issues (pp. 113–128). Springer US. https:// doi. org/
10. 1007/ 978-1- 4899- 1155-1_6
van der Drift, J. W. P. (2012). Oblique bipolar flaking, the new interpretation of mode-1. Notae Praehis-
toricae, 32, 159–164.
Vaquero, M., & Romagnoli, F. (2018). Searching for lazy people: The significance of expedient behavior
in the interpretation of Paleolithic assemblages. Journal of Archaeological Method and Theory,
25(2), 334–367. https:// doi. org/ 10. 1007/ s10816- 017- 9339-x
Villa, P., Soriano, S., Tsanova, T., Degano, I., Higham, T. F. G., d’Errico, F., Backwell, L., Lucejko, J.
J., Colombini, M. P., & Beaumont, P. B. (2012). Border Cave and the beginning of the Later Stone
Age in South Africa. Proceedings of the National Academy of Sciences, 109(33), 13208–13213.
https:// doi. org/ 10. 1073/ pnas. 12026 29109
Visser, J. N. J. (1984). A review of the Stormberg Group and Drakensberg volcanics in southern Africa.
Palaeontologica Africana, 25 http:// wired space. wits. ac. za/ handle/ 10539/ 16136
Vogel, J. C. (1983). Isotopic evidence for the past climates and vegetation of southern Africa. Bothalia,
14(3/4). https:// doi. org/ 10. 4102/ abc. v14i3/4. 1183
Vogel, J. C., Fuls, A., & Visser, E. (1986). Pretoria radiocarbon dates III. Radiocarbon, 28(3), 1133–
1172. https:// doi. org/ 10. 1017/ S0033 82220 00201 8X
Wadley, L. (1993). The Pleistocene later stone age south of the Limpopo river. Journal of World Prehis-
tory, 7(3), 243–296.
Wadley, L. (1996). The Robberg industry of Rose Cottage Cave, Eastern Free State: The technology, spa-
tial patterns and environment. The South African Archaeological Bulletin, 51(164), 64–74. https://
doi. org/ 10. 2307/ 38888 41
Wang, B., & French, H. M. (1995). Permafrost on the Tibet Plateau, China. Quaternary Science Reviews,
14(3), 255–274. https:// doi. org/ 10. 1016/ 0277- 3791(95) 00006-B
Wiessner, P. (1982). Risk, reciprocity and social influences on!Kung San economies. In E. Leacock &
R. B. Lee (Eds.), Politics and history in band societies (pp. 61–84). Cambridge University Press.
Wiessner, P. (1986). !Kung San networks in a generational perspective. In M. Biesele, R. Gordon, & R.
B. Lee (Eds.), The past and future of!Kung ethnography: Critical reflections and symbolic per-
spectives (pp. 103–129). Helmut Buske Verlag.
Wiessner, P. W. (1994). The pathways of the past: !Kung San hxaro exchange and history. In M. Bollig &
F. Klees (Eds.), Überlebensstrategien in Afrika (pp. 101–124). Heinrich-Barth Institut.
Yaroshevich, A., Kaufman, D., Nuzhnyy, D., Bar-Yosef, O., & Weinstein-Evron, M. (2010). Design and
performance of microlith implemented projectiles during the Middle and the Late Epipaleolithic of
the Levant: Experimental and archaeological evidence. Journal of Archaeological Science, 37(2),
368–388. https:// doi. org/ 10. 1016/j. jas. 2009. 09. 050
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under
a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted
manuscript version of this article is solely governed by the terms of such publishing agreement and
applicable law.

1 3
Journal of Paleolithic Archaeology (2023) 6:24 Page 35 of 35 24
Authors and Aliations
KyraPazan1 · BrianA.Stewart2,3 · GenevieveDewar3,4
* Kyra Pazan
kpazan@csustan.edu
* Brian A. Stewart
bastew@umich.edu
1 Department ofAnthropology andGeography/Environmental Resources, California State
University, Stanislaus, 1 University Circle, Turlock, CA95382, USA
2 Museum ofAnthropological Archaeology andDepartment ofAnthropology, University
ofMichigan, 3010 School of Education Building, AnnArbor, MI48109, USA
3 Rock Art Research Institute, University oftheWitwatersrand, Private Bag X3, Wits 2050,
Braamfontein,Johannesburg, Gauteng, SouthAfrica
4 Department ofAnthropology, University ofToronto Scarborough, TorontoM1C1A4, Canada