Renewed Geoarchaeological Investigations of Mwanganda’s Village (Elephant Butchery Site), Karonga,...

Research Article

Renewed Geoarchaeological Investigations of Mwanganda’sVillage (Elephant Butchery Site), Karonga, MalawiDavid K. Wright,1,* Jessica Thompson,2 Alex Mackay,3 Menno Welling,4 Steven L. Forman,5 Gilbert Price,6

Jian-xin Zhao,6 Andrew S. Cohen,7,8 Oris Malijani,9 and Elizabeth Gomani-Chindebvu9

1Department of Archaeology and Art History, Seoul National University, Seoul, Republic of Korea2Archaeology Program, School of Social Science, University of Queensland, Brisbane, Queensland, Australia3Centre for Archaeological Science, School of Earth and Environmental Sciences, University of Wollongong, Wollongong, New South Wales, Australia4African Heritage: Research and Consultancy, Zomba, Malawi5Department of Earth and Environmental Sciences, University of Illinois at Chicago, Chicago, Illinois6School of Earth Sciences, University of Queensland, Brisbane, Queensland, Australia7Department of Geosciences, University of Arizona, Tucson, Arizona8Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona9Ministry of Tourism, Wildlife, and Culture, Tourism House, Lilongwe, Malawi

Correspondence*Corresponding author;

E-mail: [email protected]

Received06 August 2013

Accepted18 December 2013

Scientific editing by Steve Kuhn.

Published online in Wiley Online Library

(wileyonlinelibrary.com).

doi 10.1002/gea.21469

The site of Mwanganda’s Village, located along a paleochannel in northernMalawi, is one of only a few sites that have characterized the Middle StoneAge (MSA) of Malawi for decades (Clark & Haynes, 1970; Clark et al., 1970;Kaufulu, 1990). The Malawi Earlier-Middle Stone Age Project has re-examinedthe site using new mapping and chronometric tools in order to reinterpret thesite’s significance within the context of current debates surrounding humanorigins and the potential role the environment played in shaping human be-havior. The new data do not support the previous hypothesis that the site wasan elephant butchery location (contra Clark & Haynes, 1970; Clark et al., 1970;Kaufulu, 1990). Instead, the evidence shows successive colonization of ripariancorridors by MSA hunter-gatherers focused on exploiting localized resourcesduring periods of generally humid climates while other lakes desiccated acrossAfrica. We challenge the hypothesis that stable and intermediately high lakelevels within the African Rift Valley System (sensu Trauth et al., 2010) cat-alyzed the evolution of regional interaction networks between 42 and 22 ka.Instead, we interpret the evidence to suggest that regional variants of technol-ogy persist into the late MSA as foragers focused on exploiting resources fromlocal catchments. C© 2014 Wiley Periodicals, Inc.

The first regular evidence for traits associated with mod-ern human cognition and behavior has been traced tothe African Middle Stone Age (MSA). During this timeperiod, there is evidence for expanded social networks;increased technological complexity (including tool man-ufacture with multiple materials and components); theevolution of planned, seasonal mobility patterns; and theemergence of symbolic behaviors (McBrearty & Brooks,2000; Brown et al., 2009; Henshilwood, 2009). Althoughsuch traits may have had their origin in the MSA, theirmagnitude has been noted to pale in comparison withLater Stone Age (LSA) behaviors between 30 and 20ka in which people greatly expanded their use of small,composite toolkits; had a more prevalent use of sym-

bolic material culture; and evidence of residential mo-bility is measured in hundreds instead of tens of km peryear (Barut, 1994; Ambrose, 1998; Barut-Kusimba, 1999;Klein, 2000).

Critical to understanding this transition is delineat-ing the temporal and geographic contours of the laterMSA. Until recently, the poor resolution of environmen-tal proxies and temporal controls on constraining ar-chaeological site habitation prevented making meaning-ful inferences about the ecology of early modern humanevolution (Willoughby, 2007; Barham & Mitchell, 2008).Improved use of uranium-based and luminescencedating techniques has allowed researchers more recentlyto explore relationships between the evolution of modern

Geoarchaeology: An International Journal 00 (2014) 1–23 Copyright C© 2014 Wiley Periodicals, Inc. 1

RENEWED INVESTIGATIONS AT MWANGANDA’S VILLAGE MALAWI WRIGHT ET AL.

behavior, changes observed in the archaeological record,and profound changes in past ecosystems (Guerin et al.,1996; Basell, 2008; Bar-Matthews et al., 2010; Blomeet al., 2012). Such data are essential for understandingif the environment placed greater selective pressures onearly modern humans to force a change in cognitive ca-pacity and behavior, or if ancestral human populationsexperienced independent cultural or demographic ex-pansions unrelated to a change in the environment (e.g.,Powell, Shennan, & Thomas, 2009).

Renewed archaeological investigations into MSA de-posits near the town of Karonga have yielded sequencesfrom new sites and new data from previously knownsites (Thompson et al., 2012; Thompson, Welling, &Gomani-Chindebvu, in press), which can be directly re-lated to the nearby paleoclimate and paleoenvironmen-tal records from Lake Malawi (Scholz et al., 2011). Re-search by the Malawi Earlier-Middle Stone Age Project(MEMSAP) in northern Malawi has identified stratifiedarchaeological deposits on alluvial fans located in ri-parian catchments spanning the MSA with a resump-tion of occupation in the historical period (Thompsonet al., 2012; Thompson, Welling, & Gomani-Chinde-bvu, in press). Successive human occupations of the siteof Mwanganda’s Village imply that highly mobile MSAhunter-gatherers exploited fluvial resources within thistransitional region between lowland and upland environ-mental zones. Recent agriculturalists occupied the site ator near the end of a period of alluvial fan deposition andwere less mobile and more tethered to the site than theirpredecessors. Our investigation revises and updates previ-ous site interpretations, using new geoarchaeological andgeochronological data obtained by MEMSAP during field-work in 2010, 2011, and 2012. Herein, we provide newdata about the site itself, with specific reference to howthe evolution of human technological behavior occurredwithin the changing environment of the Late Pleistoceneof Africa.

THE LATER MSA OF NORTHERN MALAWI

Paleogeographic Background

For the most part, the transition from the MSA to theLSA can be viewed as a series of technological and be-havioral adaptations that occurred within highly variableenvironments lasting from 40 to 20 ka throughout Africa.Some technological elements of the LSA such as bladeproduction, hafting, and symbolic expression also occurat earlier sites associated with the MSA technology, par-ticularly in the case of the techno-complex known asthe Howieson’s Poort (Soriano, Villa, & Wadley, 2007;

Backwell, D’Errico, & Wadley, 2008; Henshilwood &Dubreuil, 2011; Charrie-Duhaut et al., 2013), but by 20ka there are no MSA sites unequivocally documented inthe archaeological record (see Lombard, 2012 for a re-cent review of the evidence). Offshore sediment recordsand terrestrial pollen records from across the continentdemonstrate significant longitudinal and temporal vari-ability in climate regimes throughout the Pleistocene, cre-ating an environmental mosaic rife with both ecologicallyproductive and marginal foraging territories (e.g., Schnei-der, Muller, & Acheson, 1999; Stuut et al., 2002; Trauthet al., 2003; Dupont, 2011).

Some of the most critical phases of human biologicaland cultural evolution occurred within the margins ofthe tectonically active African Rift Valley, and the highlyirregular topographic environment combined with ex-treme climatic gradients within relatively short distancesare seen as prime catalysts fostering allopatric specia-tion (King & Bailey, 2006; Trauth et al., 2010; Bailey,Reynolds, & King, 2011; Winder et al., 2013). Within thisenvironment, small changes in global or regional climaticforcing mechanisms are amplified in Rift Valley lakes be-cause they respond with exceptional sensitivity due tothe trough-shaped aspect of the basins in which they sitand contrast between high- and low-precipitation zoneswithin their catchments (Olaka et al., 2010). Accordingto the Hypothesis of Amplifier Lakes, exceptionally high andlow lake levels serve to separate hominin populations liv-ing on opposite sides of the Rift Valley margin becausesuch events create an impenetrable geographic barrier,whereas intermediate lake levels encourage interactionswithin the lacustrine margins (Trauth et al., 2010). Thus,constraining lake levels and the chronometry of tectonicactivity within the Rift Valley provide critical backgroundfor potential drivers of human evolution.

Northern Malawi is a unique landscape to investigatethe role of lacustrine systems and topographic variabilityduring periods of profound climatic change. Lake Malawiprovides a physical barrier along the eastern margin ofthe study area, and has contained water for at least thelast 145 ka (Scholz et al., 2011), although megadroughtssignificantly reduced the water balance to ∼5% of itspresent level as recently as 90 ka (Scholz et al., 2007).Less than 100 km to the west the Nyika Plateau risesfrom 476 masl to nearly 2500 masl, leaving a narrownorth-south strip of land between the two major features(Figure 1). The presence of sharp elevational gradientsand a stable water body since at least 70 ka facilitates ahigh endemic biodiversity and there is a wide range ofpotential ecological niches that could be exploited dur-ing harsh environmental conditions (Mercader, Bennett,& Raja, 2008; Mercader et al., 2013).

2 Geoarchaeology: An International Journal 00 (2014) 1–23 Copyright C© 2014 Wiley Periodicals, Inc.

WRIGHT ET AL. RENEWED INVESTIGATIONS AT MWANGANDA’S VILLAGE MALAWI

Figure 1 Digital elevation model (DEM) of south-central Africa showing

locationsof archaeological sites andpaleoclimatic proxies.Archaeological

sites: (1) Karonga sites, (2) Lake Niassa sites (Mercader, Bennett, & Raja,

2008;Mercader et al., 2012), (3) Ngalue Cave (Mercader et al., 2009). Pale-

oclimatic proxies: (4) Mal05–2A (Scholz et al., 2011), (5) Mal05–1C (Scholz

et al., 2011), (6) M98–1P (Gasse, Barker, & Johnson, 2002; Johnson et al.,

2002), (7) M98–2P (Gasse, Barker, & Johnson, 2002; Johnson et al., 2002),

(8) Lake Masoko (Gibert et al., 2002), (9) MPU-11 and MPU-12 (Vincens

et al., 1993; Bonnefille & Chalie, 2000), (10) KH3 (Tierney et al., 2008), (11)

KH4 (Tierney et al., 2008).

One of the longest and most detailed terrestrial palaeo-climate records in Africa is derived from cores takenin the northern basin of Lake Malawi (Cohen et al.,2007; Scholz et al., 2007, 2011). These records show thatfrom at least 145 to 70 ka, the climate oscillated be-tween cool/dry and warm/wet intervals with several pe-riods of “megadrought” documented in the core record(Cohen et al., 2007; Scholz et al., 2007; Stone, West-over, & Cohen, 2011). The proxy record reflects extremelow stand events between 135 and 126 ka (∼130 ka),117 and 85 ka (∼100 ka), and a less severe episode be-tween 85 and 71 ka (∼75 ka) with intervening highstands comparable to present levels (Scholz et al., 2011;but see Lane, Chorn, & Johnson, 2013 for s slight revi-sion in the age model used to anchor the geochronol-ogy). During some of these episodes the northern basinof Lake Malawi effectively dried up. After 70 ka, thelake transitioned to generally higher conditions punc-tuated by slight regressions between 64–62 and 47–30 ka (Scholz et al., 2011; Stone, Westover, & Cohen,2011).

Geologic Context of the Mwanganda’s VillageSite

The depositional environments of northern Malawi arecomplex due to active rifting processes occurring acrosssteep elevational gradients. Karonga is located in thewestern branch of the East African Rift Valley, anarea of half grabens with alternating polarity (Ebinger,Rosendahl, & Reynolds, 1987; Rosendahl, 1987; Contr-eras, Anders, & Scholz, 2000). The kinematic evolutionof the basin began with block faulting after 8600 ka laterfollowed by normal faulting between <1500 and >200ka (Ring & Betzler, 1995). The shifting tectonic regimecreated asymmetrical rift basins that resulted in the lossof accommodation zones, enhancing the potential for al-luvial fan activation (Ring & Betzler, 1995; Chorowicz,2005).

The underlying geology of the Karonga portion of thebasin is composed of Mesozoic red sandstones, marls, andclays named “Dinosaur Beds” (Ring & Betzler, 1995). Lo-cally, these deposits are overlain by the Chiwondo Beds,thought to date to the late Miocene and early Pliocene(they are called Mikindani Beds on the Mozambican sideof the basin). The Chiwondo/Mikindani Beds are com-posed of bedded alluvium and shallow lacustrine deposits(Kaufulu, Vrba, & White, 1981; Betzler & Ring, 1995;Clark, 1995; Salman & Abdula, 1995). The ChiwondoBeds unconformably overlie the Dinosaur Beds and rep-resent a period when a large lake (or lakes) and wet-lands intermittently occupied a subsiding basin formedfrom normal faulting (Ring & Betzler, 1995; Salman &Abdula, 1995). Downwarping and fault-trough sedimen-tation during the Middle to Late Pleistocene (Ring &Betzler, 1995) combined with oscillating climate regimes(Stone, Westover, & Cohen, 2011) facilitated the ex-tension of alluvial fans that deposited the ChitimweBeds unconformably atop the Chiwondo Beds. On theMozambican side of the basin, a similar complexion ofalluvial fan activation began in the Luchamange Bedsat 42 +77/−15 ka, terminating sometime after 29 +3/−11

ka based on 26Al/10Be cosmogenic nucleotide ages (Mer-cader et al., 2012). The latter age is also associated with anMSA archaeological site with numerous discoidal coresand points, evidence of rare bipolar reduction, no docu-mented blade cores or blades, and an apparent local rawmaterial procurement strategy (Mercader et al., 2012).

The so-called “Elephant Butchery Site” in northernMalawi was discovered at Mwanganda’s Village by J.Desmond Clark and C. Vance Haynes in 1965 (Clark& Haynes, 1970; Clark et al., 1970). Excavations en-sued in 1966 (Clark & Haynes, 1970; Clark et al., 1970),and the site was later explored by Zefe Kaufulu (1983,1990). Mwanganda’s Village is located approximately



Figure 2 Regional DEM and excavated archaeological sites and scattered test pits in the Karonga region from the 2011 and 2012 field seasons.

10 m above the present-day level of Lake Malawi withinthe catchment of the Rukuru River, which drains hun-dreds of kilometers to the west and south along the west-ern wall of the Rift Valley escarpment (Figures 1 and2). Previous geoarchaeological analyses of Mwanganda’sVillage provided evidence of hominin activities from thecontact between the Chiwondo and Chitimwe beds de-marcated by a paleosol that formed on the margin of abraided stream (Clark & Haynes, 1970; Kaufulu, 1990).The majority of artifacts and the elephant fossils were re-covered by Clark and Haynes (1970) from an excavatedarea measuring approximately 11 × 12 m and named“Area 1” (see also Clark et al., 1970). A trench connectedto a second large excavation approximately 5 m to thesouth was named “Area 2,” from which a small numberof additional fossils were recovered (Figure 3).

A Rb/Sr age assayed on soil carbonate that overlies theelephant yielded a minimum constraining age of ca. 300ka associated with artifacts that were tentatively classi-fied as Sangoan (Clark & Haynes, 1970; Kaufulu, 1990;Clark, 1995). However, Rb-Sr ages on soil carbonate maybe spurious because of open system geochemistry andpotential of secondary exchange of carbonate (Bizzarroet al., 2003). The primary focus of hominin activities wasdescribed as occurring on a subtly elevated terrace adja-cent to a stream with maximum flow rates of 80 cm/s(Kaufulu, 1990). Some local fluvial reworking of artifactsas a result of low-energy overbank flooding was inter-preted from morphological data, but, overall, artifacts andfossils recovered from archaeological excavations wereargued to have been located close to their primary depo-

Figure 3 Test units, dGPS-derived DEM and elevational profile of Mwan-

ganda’s Village.

sitional context with little evidence for fluvial winnowingor abrasions resulting from long-distance fluvial trans-port (Clark & Haynes, 1970; Clark et al., 1970; Kaufulu,1990). Kaufulu’s (1983, 1990) description of the sedi-mentology of the site was based on a series of geologicaltrenches emplaced around the site and adjacent to Clark’s“Area 1” (Table I). He reconstructed a main paleochannel



Table I Comparative lithology of Kaufulu (1990) and Clark and Haynes (1970)

Lithology (Kaufulu, 1990)a Designation Lithology (Clark & Haynes, 1970) Designation

Stony soil Unit 8 Light brownish sand Qk

Light red sandstone with basal gravel Unit 7 Light red sand Qct2bSandy pebble gravel Qct2a

Dark brown muddy sandstone Unit 6 Dark brown sand Qct1bCaliche Unit 5 Caliche pebble gravel Qct1aPale grayish orange sandstone Unit 3 Dark grayish brown sand Qco1cGreenish/brownish gray sandy claystone Unit 2 Greenish gray clayey sand Qco1b

Greenish gray sandy clay Qco1a

aUnits not observed to occur at the site but that occur in the region are not included.

flowing SE-NW that would have been responsible for mi-nor fluvial reworking of the elephant fossils and artifacts.In actuality, Kaufulu’s trenches were mistakenly placedadjacent to Clark’s “Area 2,” which was approximately 9m south of the closest elephant fossils. Additionally, thestream he identified (Kaufulu, 1990) was oriented SW-NE, not SE-NW, due to a misplaced north arrow (Thomp-son, Welling, & Gomani-Chindebvu, in press). Thus, thepaleochannel reconstruction cannot apply specifically tothe locus of the elephant skeleton.

METHODS AND TECHNIQUES

Archaeological Excavations

The fact that Mwanganda’s Village was designated as aSangoan elephant butchery site suggested to MEMSAPresearchers that the site had potential to shed light onthe transition to modern human behavior during theearlier part of the MSA. The first full MEMSAP seasontook place in July/August 2010 (Thompson, Mackay, &Welling, 2011) with subsequent field seasons in 2011and 2012 (Thompson, Mackay, & Welling, 2011; Thomp-son, Welling, & Gomani-Chindebvu, 2012). Three testpits totaling 4 m2 were excavated in 2010, with all de-posits sieved through 5-mm mesh and cultural materi-als piece-plotted where they were found in situ. In 2011,13 geologic trenches approximately 1 × 2 m were exca-vated across the site to document the stratigraphy, andthese were not sieved. Optically stimulated luminescence(OSL) samples were collected for analysis from two ofthese trenches to constrain deposition on the site. In 2011and 2012, three main areas were subjected to controlledexcavation by MEMSAP. These are differentiated fromthe original Clark excavations by use of a roman numeral.Area I refers to a 5 × 5 m block emplaced in the depositsupslope and approximately 60 m to the south of the “Ele-phant Butchery Site,” Area II was a 2 × 3 m block approx-imately 3 m south of Clark’s own Area 2, and Area III wasa 4 × 6 m block in the intact deposits under and to the

west of the backdirt from Kaufulu’s (1990) “Trench 1”(Figure 3). A total of 59 m2 were removed as part of con-trolled excavations, which proceeded in natural layers.All finds were piece-plotted and screened through 5-mmmesh. An additional 5 m2 wet-sieved from Area I, 2 m2

from Area II, and all 24 m2 from Area III to recover lithicmicrodebitage and fossil material. A detailed topographicmap was generated of the project area using a Nikon C© C-series total station and Ashtech C© Promark Mobile Mapperdifferential GPS with elevations based on a local elevationbeacon to provide true masl giving centimeter-scale pre-cision topographic maps (Figure 3).

The Area I excavation revealed a lithic assemblageburied under 1.5 m of overburden, which containedmodern pottery and bricks in the uppermost plough zone.Rare faunal material and lithic tools were recovered closeto the modern ground surface from Areas II and III, closerto the reported “Elephant Butchery Site.” These werelikely associated with the originally reported archaeolog-ical finds (Clark & Haynes, 1970).

Dating

Optically stimulated luminescence dating

The geochronology of the site relied primarily on OSLdating of multiple aliquots of buried quartz grains usingthe single aliquot regeneration (SAR) method. Recentadvances in OSL dating (e.g., Rittenour, 2008; Wintle,2008) on carefully selected fluvial sediments, usuallylow-energy deposits, can yield absolute ages for the pastca. 100 ka (see also Sitzia et al., 2012). Dating of fluvialsediments in Africa has been ongoing for over a decade(Woodward, Macklin, & Welsby, 2001; Feathers, 2002;Wright et al., 2007; Rittenour, 2008). In large drainagesystems, OSL is effective in reconstructing large-scalechanges in hydrology and geomorphology related tocontinental-scale shifts in climate (e.g., Williams et al.,2010), whereas in smaller catchments OSL has proven



effective in identifying localized sources of groundwater(Ashley et al., 2011).

Fourteen samples for OSL dating were taken fromnonpedogenically or mildly pedogenically modified sed-iments to specifically avoid bioturbation. Samples weretaken from a freshly cleaned surface using PVC or alu-minum tubes 15 cm long and 2.5 cm in diameter driveninto the pit face with a rubber mallet. The OSL sam-pling was accompanied by the collection of associatedbulk samples for geochemical analysis. Optical ages arereported in years prior to A.D. 2010.

SAR protocols (Murray & Wintle, 2003) were usedin this study to estimate the apparent equivalent doseof the 150–250 μm quartz fraction for 24–30 separatealiquots. Each aliquot contained approximately 100–500quartz grains corresponding to a 1.5–2.0 mm circular di-ameter of grains adhered (with silicon) to a 1 cm diam-eter circular aluminum disc. The quartz fraction was iso-lated by density separations using the heavy liquid Na–polytungstate, and a 40-minute immersion in HF (40%)was applied to etch the outer ∼10 μm of grains, whichis affected by alpha radiation (Mejdahl & Christiansen,1994). Quartz grains were rinsed finally in HCl (10%) toremove any insoluble fluorides. The purity of the quartzseparates were evaluated by petrographic inspection andpoint counting of representative aliquots; this procedurewas repeated if samples contained >1% of nonquartzminerals. The purity of quartz separates was tested byexposing aliquots to infrared excitation (845 ± 4 nm),which preferentially excites feldspar minerals. Samplesmeasured showed weak emissions (<200 counts/s), at orclose to background counts with infrared excitation, andratio of emissions from blue to infrared excitation of >20,indicating a spectrally pure quartz extract (Duller, Bøtter-Jensen, & Murray, 2003).

A series of experiments was performed to evaluate theeffect of preheating at 200, 220, 240, and 260◦C for iso-lating the time-sensitive emissions and assessing thermaltransfer of the regenerative signal prior to the applicationof SAR protocols (see Murray & Wintle, 2003). These ex-periments entailed giving a known dose (20 Gy) and eval-uating which preheat resulted in recovery of this dose.There was concordance with the known dose (20 Gy) forpreheat temperatures above 220◦C with an initial preheattemperature used of 240◦C for 10 seconds in the SARprotocols. A “cut heat” at 240◦C for 10 seconds was ap-plied prior to the measurement of the test dose and a finalheating at 280◦C for 40 seconds was applied to minimizecarry-over of luminescence to the succession of regener-ative doses (Table II). A test for dose reproducibility wasalso performed following procedures of Murray and Win-tle (2003) with the initial and final regenerative dose of

Table II Single aliquot regeneration protocols

Step Treatment

1 Natural dose or give beta dose

2 Preheat 240◦C for 10 seconds

3 Stimulate with blue light (470 nm) for 40 seconds at 125oC

4 Give beta test dose (6.6 gray)

5 Preheat 240◦C for 10 seconds

6 Stimulate with blue light (470 nm) for 40 seconds at 125oC

7 Stimulate with blue light for 40 seconds at 280oC

8 Return to step 1

∼16 Gy yielding concordant luminescence responses (at1-σ error).

Typical OSL shine-down curves for 150–250 μm quartzgrains are shown in Figure 4. The curve shapes showthat OSL signal is probably dominated by a fast compo-nent, with the OSL emission decreasing by 90–95% dur-ing the first 4 seconds of stimulation. The regenerativegrowth curves are modeled by using the exponential pluslinear form. For many aliquots the regenerative growthcurves (Figure 4) show that (1) the recuperation is closeto zero, (2) the recycling ratio is consistent with unity at1-σ , and (3) the natural Lx/Tx ratio is well below 20% ofthe saturated level. The few aliquots removed were be-cause of unacceptable recycling ratio and De values at orclose to saturation with errors of >10%. Error analysis forequivalent dose calculations assumed a measurement er-ror of 1% and Monte Carlo simulation repeats of 2000.Recuperation is lower than 2% for all samples, whichindicates insignificant charge transfer during the mea-surements. These favorable luminescence characteristicsfor a majority of aliquots indicate that credible equivalentdose values for these sediments can be determined by theSAR protocol.

The SAR protocols yielded individual OSL ages by av-eraging 24–30 separate, equivalent doses from respec-tive aliquots of quartz grains (Murray & Wintle, 2003).Equivalent dose distributions were usually log normaland the scatter in the data is quantified with overdisper-sion values (Figure 4). An overdispersion percentage ofa De distribution is an estimate of the relative standarddeviation from a central De value in context of a statis-tical estimate of errors (Galbraith et al., 1999; Galbraith& Roberts, 2012). A zero overdispersion percentage in-dicates high internal consistency in De values with 95%of the De values within 2-σ errors. Overdispersion values≤20% are routinely assessed for quartz grains that arethoroughly reset during exposure to solar rays, like eoliansands (e.g., Olley, Pietsch, & Roberts, 2004; Wright et al.,2011) and this value is considered a threshold metric forcalculation of a De value using the central age model of



Figure 4 Representative regenerative dose growth curves, with inset representative natural shine down curve, and radial plots of equivalent dose values

on small aliquots (2-mm plate of 150–250 μm quartz fraction grains).

Galbraith et al. (1999). Overdispersion values >20%may indicate mixing of grains of various ages or par-tial solar resetting of grains; the minimum age model(three parameters) may be an appropriate statisticaltreatment for such data and effectively weights forthe youngest De distribution. However, some studieshave concluded that overdispersion values between 20%and 32% may reflect a single De population, partic-ularly if the De distribution is symmetrical, with the

dispersion related to variability associated with micro-dosimetry and/or sedimentary processes (e.g., Arnold &Roberts, 2009). Four of the twelve samples (LM2011-03,LM2011-08, LM2011-09, and LM2011-11) have overdis-persion values that are >20% (at 2-σ errors) with neg-atively skewed distributions, thus the minimum agemodel is most appropriate for estimating an equiva-lent dose (Galbraith et al., 1999; Galbraith & Roberts,2012).



Multiple aliquot regenerative dose protocols, similar tothose used in Londono et al. (2012) were employed ontwo samples (UIC3091 and UIC3091) to test the accu-racy of the SAR-based ages. The Multiple Aliquot Regen-eration (MAR) analyses are predicated on different as-sumptions than SAR with resetting of naturals under UVlight for 2 days prior to regenerative dosing. In turn, eachaliquot is used for a single measurement, rather than a se-ries with SAR, which necessitates just one sensitivity cor-rection. The MAR equivalent doses overlap with the cor-responding SAR values, which yields greater confidencein the rendered equivalent doses.

The environmental dose rate is a critical measurementfor calculating a luminescence age. The dose rate is anestimate of the exposure of quartz grains to ionizing ra-diation from the decay of the U and Th series, 40K, andcosmic sources during the burial period. The U, Th, andK concentrations are determined by inductively coupledplasma mass spectrometry (ICP-MS) by Activation Labo-ratory LTD, Ontario, Canada. The beta and gamma doseswere adjusted according to grain diameter to compensatefor mass attenuation for the dose rate (Fain et al., 1999).Beta and gamma attenuation coefficients for 150–250 μmare 0.876 and 0.999, respectively. The U, Th, and K2Ocontent was determined for the bulk sediment to calcu-late the dose rate. A cosmic ray component, taking intoaccount location, elevation, and depth of strata sampledis between 0.17 and 0.20 mGy/yr and is included in theestimated dose rate (Prescott & Hutton, 1994). There isuncertainty in assessing the moisture content of a sam-ple during burial. We estimated moisture contents frompresent values, particle size characteristics, and in refer-ence to the water table.

In the absence of gamma spectrometry, it is unknownif there is disequilibrium in the U and Th decay series. Anumber of samples have relatively high Th values (10–25ppm) with Th to U ratios that exceed 6, which suggestsa granitic source (Van Schmus, 1995), though digeneticprocesses cannot be dismissed for these elevated ratios.

Uranium-thorium dating

Uranium-thorium (U-Th) ages were obtained from threefragments of fossil elephant tusk recovered by Clarkand Haynes (1970). These samples were loaned fromthe Stone Age Institute at the University of Indiana-Bloomington to the University of Queensland (UQ). U-Th dating was conducted at the Radiogenic Isotope Facil-ity at UQ. Analysis was performed using a multi-collectorinductively-coupled mass spectrometer (MC-ICPMS) fol-lowing procedures described in Zhou et al. (2011) andRoff et al. (2013).

U-Th dating is a radiometric dating technique com-monly used to determine the age of carbonate materi-als from sources such as speleothems and corals. The U-Th dating method is based on the decay of 238U (with ahalf-life T1/2 = 4.469 × 109 years) to stable 206Pb via in-termediate daughters such as 234U (T1/2 ∼ 245,000 years)and 230Th (T1/2 ∼ 75,400 years). In this decay series, 238U-234U-230Th disequilibrium occurs when U is differentiatedfrom Th during a particular geological or environmen-tal event or process. In the case of natural aqueous sys-tems, for example, in which U is slightly soluble, but Th ishighly insoluble, carbonate precipitated from the aqueoussystem will contain trace amounts of U (usually 0.01–100ppm), but virtually no Th, leading to excess U in the de-cay chain (i.e., 238U and 234U activities >> 230Th activity).Once disequilibrium is established, it takes about seventimes the half-life of 230Th (∼500 ka) for the system toreturn to near secular equilibrium (i.e., when the activi-ties of the parent and daughter nuclides are equal), or tothe level where the degree of disequilibrium is below thelimit of detection by thermal ionization mass spectrome-try or MC-ICPMS.

Unlike speleothems or corals, teeth and bones of livinganimals contain very little U. Instead, U is taken up fromthe environment during fossilization processes. Thus, theU-Th date of a fossil tooth or ivory sample records themean age of the fossilization process or U uptake history.In ideal cases, U uptake may reach saturation level duringthe early stage of the fossilization process (early uptakemode). In this case, the U-Th date would approximatethe deposition age of the fossil material. In most othercases, U uptake modes might be more complex. The U-Thdates of the fossil are variable but theoretically youngerthan the deposition age of the fossil material. U-Th datesof the fossil may become apparently too old if extensiveU loss through leaching occurred (Pike, Hedges, & VanCalsteren, 2002). However, extensive U loss, sufficientto make the apparent U-Th date older than the true ageof the fossil, is usually extremely rare. In most cases, asdemonstrated in numerous previous studies (e.g., Zhaoet al., 2001; Grun et al., 2010; Price et al., 2011, 2013),the ages returned by the U-Th method provide minimumdates for the fossilization of the bone or molar fragments.

Stone Artifact and Faunal Analysis

A sample of 2363 stone artifacts and specimens of fos-sil bone have been studied from the MEMSAP exca-vations. Attempts have been made to locate the cu-ration site of the elephant remains described by Clarkand Haynes (1970: 393) both in Malawi and in knownplaces to which materials were exported in the 1960s, butthese have proven largely unsuccessful. The Stone Age



Institute in Bloomington, Illinois currently houses themajority of the stone artifacts originally recovered and re-ported by Clark and Haynes (1970), but only has a smallcollection of fossils (N = 156) most likely representingthe fragments found underlying the elephant femur anda small number of fossils from Clark’s Area 2 (Clark &Haynes, 1970).

Analytic techniques for stone artifacts at Mwanganda’sVillage followed those previously employed during workat the nearby Airport Site (Thompson et al., 2012). Allplotted and sieve-recovered artifacts were analyzed. Theanalytical focus was on the metric attributes of artifacts,indicators of manufacturing systems, extent of reduction,and indicators of rock source. Metric data included mea-sures of artifact weight, length, width, and thickness. Forflakes, platform dimensions were also recorded. Indica-tors of manufacturing systems included evidence of lam-inar, disc/discoidal, and Levallois reduction techniquesfrom dorsal and platform scar patterning. Extent of re-duction was assessed using artifact size and percentageof cortex coverage, the latter being of particular in in-terest in this case given potential sourcing of toolstonefrom local gravels. Indicators of rock source included rocktype and cortex types, variance in the latter being takento identify secondary (e.g., cobble) or primary (e.g., out-crop) sources.

An additional component of the analysis concernsevidence for postdepositional fluvial reworking onartifacts—important given the particular interest in siteformation at Mwanganda’s Village. Artifact fluvial re-working was assessed mainly through edge rounding.Four classes of edge rounding were used: 0, no edgerounding; 1, rounding discernable by touch but not vi-sually obvious; 2, visually obvious rounding that may ob-scure some artifact features; and 3, a rock that could stillbe identified as having been an artifact but which wasnow so heavily abraded as to be little more than a fluvialclast.

Although small, the sample of fauna from the StoneAge Institute was the largest available sample of fos-sil fauna from the site. It was examined for taxonomicand taphonomic attributes that could provide informa-tion about site formation at the “Elephant Butchery Site”itself, as no faunal remains were recovered from MEM-SAP’s Area I. Each bone surface was examined under a10–40× binocular zoom microscope with bright incidentlight shining obliquely across the surface. In this way evi-dence of smoothing or polishing was identified that mightindicate fluvial transport or abrasion. Each specimen wasalso examined for potential human modification (suchas cut marks) or carnivore modification (such as toothmarks).

RESULTS

Geologic Evolution of Mwanganda’s Village

There are four primary sedimentary facies found atMwanganda’s Village (Figure 5), which correspond withfour phases of deposition (Figure 6). Complete descrip-tions of the sedimentology and pedology of the excava-tion units are provided in the Supporting InformationTable S1.

The basal, nonarchaeological deposits are composed ofweakly bedded silty clay within a well-developed pale-osol with strong redoximorphic features (Unit 1). Thesedeposits broadly correspond to descriptions of “ChiwondoBeds” or Qco1 made by Clark and Haynes’ (1970; see alsoTable I and Kaufulu, 1990), although there is no directevidence to suggest that these actually are Pliocene lakedeposits. Instead, Unit 1 is consistent with slackwater orwetland sediments and is a localized phenomenon ratherthan a regionally distributed bed.

The second depositional unit identified at Mwan-ganda’s Village is characterized by coarse sands and grav-els with unconformities associated with cut-and-fill de-position separated by argillic and calcic paleosols. Thesedeposits were interpreted by Clark and Haynes (1970)and Kaufulu (1990) as braided stream deposits, whicheroded into parts of the underlying strata (Table I). Kau-fulu (1990) inferred that gentle overbank sedimentationprovided a riparian habitat in which prehistoric peoplebutchered an elephant, with some minimal postdeposi-tional movement of the remains. However, our analysis(based on differential Global Positioning System (dGPS)and OSL reconstruction of the site) shows that the sitedeposition was not as uniform as previously reported.Stream formation and overbank sedimentation was suc-cessive and involved multiple cut-and-fill episodes thatpunctuated periods of argillic carbonate soil (Btk) forma-tion. “Soils on soils” (Bhkm) are identified in Area III,Unit 2 associated with episodic fluvial aggradation withdifferent types of sediments (occurring in different en-vironmental conditions) followed by a period of land-form stability (Figure 5). Each channel formation episodeeroded the upper portion of the calcic soil and cut at leastone terrace into the site. Hiatuses in fluvial depositionlikely occurred during arid conditions based on Stage 2carbonate formation and the accretion of well-developedilluvial lenses of clays (argillans) within the soils. The pe-riod of time that it took to form these soils is uncertain,but was likely on the order of 100s to 1000s of years.U-Th analysis of three elephant tusk samples from thesedeposits returned apparent dates of 228.7 ±2.7, 254.3 ±4.1, and 282.3 ± 2.4 ka (Table III). We argue that 282ka represents the minimum age of the fossilization of



Figure 5 Sedimentology and pedology of Mwanganda’s Village from the 2011 and 2012 field seasons.

the elephant remains (Price et al., 2013). An OSL sam-ple analyzed from the correlated paleosol yielded an ageof 23,613 ± 1900 years (UIC2858) suggesting that theelephant remains were entrained in a lag deposit, onlycoincidentally and are not functionally related to the ar-chaeological artifacts found on the site.

The third depositional unit recorded at the site con-sists of poorly sorted coarse sands and gravels within a

strongly redoximorphic environment, which fine upwardto a massive, well-sorted fine sand capped by a very weakepipedon. This deposit can be correlated with the broaderregional emplacement of Chitimwe Beds across north-ern Malawi. Chitimwe Beds are broadly defined as Pleis-tocene alluvial fan deposits that occurred concomitantto rotational faulting and uplift (Ring & Betzler, 1995).This likely catalyzed erosion of red sandstones within the

Table III U-Th ages from elephant tusk at Mwanganda’s Village, Malawi

Lab Sample Name Date and Time U (ppm) ±2 seconds 232Th (ppb) ±2 seconds

JT-1 #13 September 7, 2010 at 21:21 2.8027 0.0009 2.43 0.025

JT-2 #14 September 7, 2010 at 22:39 2.6222 0.0006 11.22 0.034

JT-3 #15 September 7, 2010 at 23:18 5.2161 0.0019 5.24 0.024

Lab Sample Name (230Th/232Th) ±2 seconds (230Th/238U) ±2 seconds (234U/238U) ±2 seconds

JT-1 #13 3539.31 39.13 1.0149 0.0036 1.1248 0.0011

JT-2 #14 753.21 3.88 1.0640 0.0044 1.1394 0.0013

JT-3 #15 3337.67 16.49 1.1066 0.0019 1.1511 0.0010

Uncorr. 230 Corr. 230 Corr. Initial (234

Lab Sample Name Th Age (ka) ±2 seconds Corr. 230 ±2 seconds U/238U) ±2 seconds

JT-1 #13 228.7 2.7 228.7 2.7 1.2382 0.0023

JT-2 #14 254.4 4.1 254.3 4.1 1.2863 0.0035

JT-3 #15 282.3 2.4 282.3 2.4 1.3356 0.0022



Table

IVOptic

allystim

ulated

luminescenc

e(OSL)age

son

qua

rtzgrains

from

Karong

aDistrict,Malaw

i

Sample

Laboratory

Num

ber

Aliquo

ts

Equivalent

Dose(Gray)

aOverdispersion

(%)b

U(ppm)c

Th(ppm)c

K(%)c

H2O(%)

CosmicDose

(mGray/yr)d

DoseRate

(mGray/yr)

OSL

Age

(yr)e

S683

UIC28

5430

/30

37.99

±2.31

25.2

±3.4

1.4

±0.1

9.4

±0.1

1.41

±0.02

5±2

0.20

±0.02

2.43

±0.12

15,610

±12

80

S684

UIC28

5730

/30

48.58

±2.50

19.7

±2.6

1.6

±0.1

9.6

±0.1

1.54

±0.02

20±5

0.18

±0.02

2.20

±0.11

22,065

±19

20

S685

UIC28

5630

/30

82.70

±4.19

21.4

±2.8

1.8

±0.1

9.6

±0.1

1.39

±0.02

30±5

0.17

±0.02

1.94

±0.10

42,550

±35

50

S186

5UIC28

5828

/40

59.09

±3.53

24.4

±3.9

1.5

±0.1

11.4

±0.1

1.51

±0.02

10±3

0.15

±0.02

3.97

±0.20

23,610

±19

00

LM20

11–0

2UIC30

9728

/30

3.22

±0.20

27.1

±4.0

1.6

±0.1

13.2

±0.1

1.51

±0.02

10±3

0.20

±0.02

2.69

±0.13

1195

±10

0

LM20

11–0

3UIC30

9526

/30

15.80

±0.79

34.6

±4.8

1.1

±0.1

9.6

±0.1

1.39

±0.02

25±5

0.20

±0.02

2.23

±0.11

7065

±55

0f

LM20

11–0

6UIC31

2030

/30

7.74

±0.47

27.3

±4.0

1.5

±0.1

8.8

±0.1

1.85

±0.02

5±2

0.20

±0.02

2.80

±0.14

2760

±23

0

LM20

11–0

7UIC30

9830

/30

14.81

±0.61

11.0

±1.5

1.3

±0.1

7.6

±0.1

1.64

±0.02

7±2

0.18

±0.02

2.42

±0.12

6130

±43

5

LM20

11–0

7UIC30

98m

g14

.09

±0.71

1.3

±0.1

7.6

±0.1

1.64

±0.02

7±2

0.18

±0.02

2.42

±0.12

5820

±39

0

LM20

11–0

8UIC31

1924

/30

4.89

±0.17

39.2

±5.7

1.4

±0.1

10.2

±0.1

1.51

±0.02

5±2

0.21

±0.02

2.62

±0.13

1865

±12

0f

LM20

11–0

9UIC30

9227

/30

32.82

±2.84

37.7

±5.2

1.5

±0.1

13.0

±0.1

1.60

±0.02

5±2

0.20

±0.02

2.88

±0.14

11,380

±11

65f

LM20

11–1

0UIC30

9125

/30

42.93

±2.52

23.7

±3.5

1.4

±0.1

11.0

±0.1

1.57

±0.02

5±2

0.20

±0.02

2.69

±0.13

15,955

±12

85

LM20

11–1

0UIC30

91m

g38

.30

±2.30

1.4

±0.1

11.0

±0.1

1.57

±0.02

5±2

0.20

±0.02

2.69

±0.13

14,245

±82

0

LM20

11–1

1UIC31

3429

/30

0.31

±0.02

59.0

±7.8

3.6

±0.1

21.4

±0.1

2.30

±0.02

5±2

0.20

±0.02

4.59

±0.22

65±5f

aOne

hund

redfifty

to25

0μm

qua

rtzfractio

n(2

mm

plate

area

)ana

lyzedun

der

blue-light

excitatio

n(470

±20

nm)b

ysing

lealiquo

treg

enerationprotoco

l(Murray&Wintle

,200

3).

bValue

srefle

ctprecision

beyon

dinstrumen

talerrors;values

of≤2

0%indicatelowspread

ineq

uivalent

dosevalues

with

aun

imod

aldistribution.

cU,Th,

andK 2Oco

nten

tana

lyzedbyinduc

tivelyco

upledplasm

amassspectrom

etry

analyzed

byActivationLaboratoryLTD,O

ntario,C

anad

a.dFrom

Prescotta

ndHutton(199

4).

eAge

scalculated

usingthecentralage

mod

elof

Galbraith

etal.(19

99).

f Age

scalculated

usingtheminim

umag

emod

elof

Galbraith

etal.(19

99)b

ecau

seof

elevated

overdispersion

values

(>20

%).

gEq

uivalent

dosedetermined

bymultip

lealiquo

treg

enerationprotoco

ls(Jainet

al.,20

03).

Allerrors

areat

1-σan

dag

esfrom

thereferenc

eyear

A.D.2

010.



Figure 6 (A) Aggradational profile of Mwanganda’s Village showing depositional units and terraces; (B) generalized aggradational profile of the Karonga

region; (C) generalized time-staggered model of landscape change in the Karonga region (see also Betzler & Ring, 1995; Ring & Betzler, 1995).

Dinosaur Beds, which exist as remnants in the plateaufoothills ca. 25 km west of the project area. Depositionof the Chitimwe Beds was not a single event, but rathera series of events with different geologic sources, whichis attested by the extremely variable nature of the sedi-ments across the region. At Mwanganda’s Village, basalChitimwe Bed emplacement occurred ca. 42,550 ± 3550years (UIC2856; Table IV), which is statistically coeval tothe OSL age assayed from Unit 2 from Area III. In Area Iof the site, the Chitimwe Beds unconformably overlie de-positional Unit 1 and appear to have buried the paleosolsin a relatively undisturbed state. MnO and Fe2O3 stainingof the sediments in the lower 25 cm of the unit is likelya function of the creation of an aquatard in the argillichorizon that subaerially weathered the sediments in situ.An OSL age assayed from GT9 constrains alluvial fan de-position as ceasing after 2760 ± 230 years (UIC3120).

Following the deposition of the youngest sediments atthe site, a final geomorphic phase occurred marked by aperiod of fluvial incision and terrace formation. Four ter-

races were documented within Mwanganda’s Village andthe northwestern edge of the project area was definedas Chirambiru Creek (Figure 3). Thin lenses of alluviumwere deposited as the stream downcut. The incision eventhas exposed paleosols across the sites, and the water tabledropped significantly from the periods when redoximor-phic features formed. The Area I excavation by MEMSAPwas located on Terrace 3 and Areas II and III (and the“Elephant Butchery Site”) were located on Terrace 2.

STONE ARTIFACT ANALYSIS

A sample of 2363 flaked stone artifacts were analyzedfrom the three excavation areas at Mwanganda’s Village.The assemblage analyzed to date from 2 m2 (of the to-tal 24 m2) at Area III is small and heavily reworked, withhalf of the 12 artifacts showing class 2 edge rounding (theother half show no rounding). Of the flakes with intactplatforms, five have dihedral or faceted platforms sug-gestive of MSA affinity. Two flakes exhibited centripetal



dorsal scar patterning. None of the artifacts in the as-semblage could be classified as characteristic Sangoanimplements, such as picks, core axes, or core scrap-ers (Clark & Haynes, 1970; McBrearty, 1988; Tryon &McBrearty, 2002); bifacial tools were generally absentfrom the Mwanganda’s Village assemblages.

Area II produced a considerably larger and more di-verse assemblage (n = 278). The assemblage includesflakes (78.1%, n = 224), retouched flakes (2.2%, n =6), and cores (2.9%, n = 8), along with other flakingdebris. Slightly over half (50.7%) of the Area II artifactsexhibited cortex, with cobble cortex being the only typepresent, consistent with sourcing of toolstone from localgravels. As with Area III, the evidence from edge round-ing suggests considerable fluvial reworking in the assem-blage. Only ∼40% of the 278 artifacts show no evidenceof abrasion, with class 1 (21.6%), class 2 (21.2%), andclass 3 (14.5%) edge rounding all well represented.

Quartzite is the dominant material within the total as-semblage (54.7%) followed by quartz (39.2%) and crystalquartz (5.8%). With the quartz values combined, thesenumbers are very similar to those in the Clark and Haynessample. The frequency of edge rounding does not dis-criminate the two main rock types. Unrounded piecesaccounted for 38.8% of quartzite artifacts and 42.2% ofquartz artifacts. Heavily rounded pieces (classes 2 and 3)accounted for 40.1% and 33.9% of quartzite and quartzpieces, respectively. Quartz crystal artifacts have consid-erably lower rates of edge rounding, with 68.8% un-rounded.

Flaking patterns weakly attest to the use of discoidaland Levallois methods at the site (Figure 7), though mostflakes appear to be at a relatively early stage of reduc-tion as reflected in the high proportion of cortex notedabove. There was no evidence of laminar working in theexamined samples. However, there are clear differencesin the proportions of platform types in our assemblagecompared to the Clark and Haynes sample. Of the 49complete flakes we recovered from Area II, 24.5% (n =12) had cortical platforms, 34.7% (n = 17) had simpleplatforms (e.g., single scar, no cortex), and 10.2% (n =5) had multiple facets. In the Clark and Haynes (1970)sample of 80 flakes, 59% had cortical platforms, 32.4%had simple platforms (subsuming both plain and singlefaceted in the classification published in Clark & Haynes,1970), and none exhibited multiple facets.

The small core assemblage includes two pieces withscar removals to alternating surfaces around a portion of aworked margin. These may have been early stages in thereduction of discoidal cores though notably in both casesthe selected packages were small quartz crystal pebbles.Two other more typically discoidal cores also occur andthese show some hierarchical arrangement of volumes

with scarring patterns unevenly distributed between sur-faces. Cortex is in both cases restricted to a single sur-face, though in one instance flakes were removed fromboth surfaces while in the other the cortical surface wasentirely unworked (Figure 7). This latter piece may con-ceivably have been classed by Clark and Haynes (1970:Figure 28) as a core axe. The retouched flakes in the as-semblage show only minor and generally unstructuredmarginal flaking; no morphologically regular types wererecorded.

The assemblage from Area I is considerably differentin composition and degree of fluvial reworking from theArea II and Area III assemblages. The analyzed samplecomprises 2073 pieces, and shows similar proportionsof quartzite (59.1%), quartz (32.6%), and quartz crys-tal (7.2%) to Area II. Unlike Areas II and III, however,rounding only occurs on 12.8% of pieces, and heavyrounding (classes 2 and 3) on only 4.2%. That the as-semblage is not substantially reworked is supported byconjoin analysis; 61 conjoin sets were identified compris-ing 161 artifacts, which refit to at least one other artifactor fragment thereof.

The artifacts recovered from Area I include flakes(72%), retouched flakes (2%), and cores (4.5%), as wellas other flaking debris. Formal tools are extremely rare,a single scraper and a single denticulate being the onlytypes identified. Cortex is common, occurring on 53.9%of pieces. This value is as high as 73% if restricted to com-plete flakes. With a single exception, cortex was of cob-ble/pebble form. Again, it seems probable that the mate-rials for artifact manufacture were sourced from locallyavailable gravels.

Like Areas II and III, the Area I assemblage includedMSA markers such as discoidal and Levallois flakes andcores (Figure 7). Though fairly large (200–900 g) discoidalcores are present in the sample, the cores from Area Iwere generally quite small. The smallest core in the as-semblage had a weight of 2.3 g. The Area I sample in-cluded 16 cores characterized by the centripetal reductionof a single surface of a small crystal quartz pebble from acortical platform (Figure 7). Though this was the singlemost common core type at Area I, it was not observed inthe other assemblages. Its abundance at Area I may ac-count for the elevated proportions of cortical platforms inthe flake assemblage, which at 46.5% (n = 180, completeflakes only) is roughly double the proportion at Area II.The proportion of simple platforms (18.6%) at Area I isapproximately half that at Area II.

Bipolar and laminar reduction are also represented atArea I but neither constitutes a significant assemblagecomponent. In spite of the small size of the cores, onlytwo (2.1%) were worked using bipolar techniques. In-deed, the small size of cores is likely a reflection of the



Figure 7 Artifacts from MEMSAP excavations at Mwanganda’s Village:(A) unifacially and (B) bifacially worked cobble cores, Area II; (C) Levallois and (D)

discoid flakes, Area II; flakes with (E) class 2 and (F) class 3 edge rounding; (G–J) pebble cores with unifacial centripetal working, Area I; (K) pebble and (L)

cobble cores with minor working of the cortical surface in addition to centripetal working of the exploitation surface, Area I; (M) discoidal core, Area I;

cobble cores with (N) unifacial and (O) bifacial working of one margin, Area I. All artifacts are shown at the same scale; scale bars = 10 mm increments.

size of the pebbles selected. Nine of the 10 smallest coresin the sample had cortex coverage of 40% or greater,and more than 92% of all complete cores retained somecortex.

Clark and Haynes (1970: 394) report that 99% of thestone tools from their Area 1 excavation were unabraded,although they also report, “Signs of utilization are veryextensive in the form of minute scarring of the edges,crushing, and rubbing” (Clark & Haynes, 1970: 395).Macroscopic signs of tool use were very infrequently ob-served in the MEMSAP sample. Edge damage (percussivedamage as opposed to edge rounding and excluding ex-cavation damage) was noted in 13 cases, all of which oc-curred in the Area I sample. There are, consequently, fewconsistencies between our samples from Areas I, II, andIII and those recovered from Clark and Haynes’s (1970)“Elephant Butchery Site” (see also Clark et al., 1970). Theonly clearly in situ occurrence at Area I was not recoveredfrom the same terrace as the “Elephant Butchery Site”and has ages and a composition that mark it as late MSA,not Sangoan. Area III may have contained more robust

evidence of Sangoan tools given a larger sample. How-ever, this would not explain the high rate of fluvial re-working in our recovered assemblage. The Area II assem-blage exhibits a few potentially Sangoan tools, howeverthe high fluvial reworking rates and the low proportionof cortical platforms render this assemblage an unlikelymatch for the Clark and Haynes sample.

FAUNAL ANALYSIS

Reanalysis of the fossils from Clark and Haynes’ (1970)excavations hosted at Stone Age Institute identified onefragment to be a turtle carapace fragment from the familyPelomedusidae and another was a partial catfish frontalfrom the genus Clarias. Most of the remaining fragmentscome from a very large animal, suggesting they were partof the original elephant skeleton reported from the site(Clark & Haynes, 1970). Of the 156 specimens in the to-tal collection, 108 (or 69.2%) displayed some evidence ofsmoothing. The fluvial context of the sediments makeswaterborne abrasive particles the most parsimonious



candidates for the cause of the smoothing. Of the frag-ments remaining at the Institute only 11/135 (8.1%)were less than 2 cm in the maximum dimension, which isin general agreement with the sizes of the stone artifactsand also suggests fluvial size winnowing (Schick, 1992;Pante & Blumenschine, 2010; Sitzia et al., 2012).

In spite of extensive smoothing, many of the bones stillpreserve clear evidence of carnivore activity (and no evi-dence of hominin butchery). A total of 13% (n = 13) ofthe fossils from Clark’s Area 1 and 14.7% (n = 5) fromClark’s Area 2 preserve tooth marks. Most have mor-phology that involves deep punctures rather than fur-rows and scores, and several are ovate in shape. In somecases the pits are bisected, possibly produced by the bicar-inated teeth of crocodiles rather than mammalian carni-vores (Baquedano, Domınguez-Rodrigo, & Musiba, 2012;Njau & Blumenschine, 2012).

DISCUSSION

Archaeological Habitation of Mwanganda’sVillage

Archaeological occupation of Mwanganda’s Village waspunctuated, but repeated during the Pleistocene andHolocene. The aggradational sediment sequence fromMEMSAP Areas II and III of the site contain MSA artifactsboth in situ and in secondary context, along with the re-mains of turtle, catfish, and elephant faunal remains witha minimum constraining age of fossilization of 282.3 ±2.4 ka (Table III). However, these remains are entrainedin sediments that were last exposed to light 23,613 ±1900 years. Thus, the archaeological and geologic evi-dence indicate that the area from which the elephantwas derived was a depositional context more closely re-sembling a channel-lag deposit rather than a channel-filldeposit (sensu Behrensmeyer, 1988). The presence of apartial elephant skeleton with few other identified verte-brates in association suggests some degree of spatial in-tegrity, but the amount of abrasion, presence of aquaticfauna in nearby Area 2 (Clark & Haynes, 1970; Clarket al., 1970), modification suggestive of crocodile activ-ity, and the fragment size distribution indicates that thefossil assemblage contains allochthonous elements trans-ported by water and not butchered by people. Thus, thereis no current evidence to support the original interpreta-tion of the site as a Sangoan elephant butchery site (Clark& Haynes, 1970; Clark et al., 1970; Kaufulu, 1990).

A later Pleistocene occupation on T-3 of the site isinterpreted as having intact, in situ artifactual depositsand a range of lithic reduction strategies were under-taken within a riparian setting. Diagnostic MSA artifactsare present at MEMSAP Area I from 42,550 ± 3550

years (UIC2856) and are stratified through to 22,065 ±1920 years (UIC2857) with a resumption in occupa-tions occurring by the Iron Working period (>2760 ±230 years, UIC3120) and continuing through the presentday (Table IV). The relatively late age for MSA tech-nology at Mwanganda’s Village is consistent with datafrom most sites in southern and eastern Africa but dif-fers from recent arguments for an MSA/LSA transitionbefore 49 ka at Mumba in eastern Africa (Eren, Diez-Martin, & Dominguez-Rodrigo, 2013). That said, the lateMSA at Mwanganda’s Village appears compositionallydifferent from comparably aged assemblages on the otherside of Lake Malawi at Mvumu and Ngalue Cave (Mer-cader et al., 2012). Those assemblages are overwhelm-ingly quartz dominant (∼95% at Mvumu vs. ∼40% atMwanganda’s Village Area I) and contain numerous uni-facially and bifacially worked flake tools such as scrapersand awls that are effectively absent at Area I. Mercaderet al. (2012) report that retouch occurs on 10% of theMvumu assemblage; at Area I the value is 1.8%. Morestrikingly, formal tools account for 11% of the Mvumuassemblage whereas the value is two orders of magni-tude lower at >0.1% in the Mwanganda’s Village AreaI sample. While classificatory differences might conceiv-ably drive some portion of the typological variability itseems unlikely to account for all of it and would not ex-plain the discrepancy in retouched flake proportions.

Similarly the most common mode of reduction at AreaI, which features unifacial centripetral reduction of smallcrystal quartz pebbles, is not described in the publishedsamples from Niassa (Mercader et al., 2012). While thereis no doubt that the size of the quartz crystal pebbles se-lected influenced the reduction system used, raw materi-als of sizes up to and exceeding 10 cm diameter were alsoavailable in the gravels at Area I and adjacent drainage-ways. Therefore, preferential selection of small quartzcrystal pebbles was a choice not dictated by geologicalconstraints. Overall, while the Mozambican assemblagesand those described here share similarities—notably thepersistence of MSA technologies and the paucity of bipo-lar reduction—the differences at this stage seem consid-erable despite the fact that the radiometric ages for theoccupations of the sites overlap at 1-σ (Mercader et al.,2012).

All phases of occupation at Mwanganda’s Villageinclude artifacts that are likely to have been discardedeither at or near the point of manufacture. This inferenceis based on the presence of local cobble/pebble beds andthe high prevalence of cortex in assemblages recoveredfrom our excavation areas. To that extent the richness ofthe area with respect to archaeological material is likelyin part a reflection of persistent if perhaps opportunisticuse of locally abundant sources of flakeable rock in the



form of gravel beds. These beds would have produceda reliable source of toolstone that would have acted tominimize pressures on curation of transported artifactmaterial, particularly for foragers moving along or inproximity to small watercourses as a means of exploitingriparian resources. That these cobble sources are widelydistributed in the Karonga area may go some way toexplaining the surface and subsurface richness of theregion’s Pleistocene archaeological landscape.

A surface collection of Kisii pottery indicates habitationin historical times. This occupation of agriculturalists ap-pears to have been concentrated on T-3, and postdatedthe onset of fluvial incision. It is impossible to estimatethe distance of the water source to the historical occupa-tion, but pot irrigation and water extraction for consump-tion would have been manageable even from its presentdistance from the locus of historical residence (∼150 m).The ability to store water in vessels and cultivate cas-sava (Manihot esculenta) for off-season consumption is acultural practice that extends to the present day. Mwan-ganda’s Village is currently settled and under cultivationby Ngonde and Tumbuka-speaking people with deep cul-tural connections to the land.

Environmental Context of Occupations atMwanganda’s Village

Basal wetland and soil formation detected from Unit 1of Areas I and III do not conform sensu stricto to PlioceneChiwondo Beds as described by Betzler and Ring (1995),but appear to be part of a broader pluvial period preced-ing Chitimwe Fan activation. There are no archaeologicalartifacts entrained in these sediments, nor has OSL datingbeen successful in determining an age due to saturationof luminescence traps within the assayed quartz miner-als. This phase of site deposition was likely occurring priorto the activation of fault-induced and/or climatically in-duced alluvial fans, and the landscape was more geologi-cally stable and more topographically regular than duringthe Late Pleistocene (Ebinger et al., 1989).

There is an erosional unconformity at the site, whichcompromises the archaeological record until ca. 40 ka,when the emplacement of Chitimwe alluvial fan depositsoccur within a landscape of braided streams (Kaufulu,1990). Initial formation of Chitimwe Beds at Mwan-ganda’s Village are in close chronological agreement withthe formation of depositionally analogous LuchamangeBeds on the east side of the basin (Mercader et al., 2012).Mwanganda’s Village hosts high-energy channel depositswith 2–5 cm rounded to subrounded cobbles as well aslower energy channel deposits composed of sandy clayloam. The diatom assemblage from the Lake Malawi sed-iment core suggests a slight regression in lake levels be-

tween 47 and 30 ka (see also Finney et al., 1996; Scholzet al., 2011; Stone, Westover, & Cohen, 2011). There isa transition to deeper lake conditions (>400 m deep) be-tween 31 and 16 ka with great variability in lake levelsca. 17 ka (Scholz et al., 2011; Stone, Westover, & Cohen,2011). The in situ archaeological occupations documentedat Mwanganda’s Village begin at Area I ∼40–42 ka withthe deposition of a sandy facies that contains minimallyreworked MSA artifacts that retain at least two instancesof flake-to-flake refitting. A fluvial layer truncates this fa-cies and contains heavily rounded artifacts that attest toa separate occupation elsewhere in the catchment thatis unlikely be contemporaneous with the material in thesandy facies. This is overlain by an in situ late, and possiblyterminal, MSA assemblage in fine-grained sediments thatdate to ∼22 ka (Figure 5: Area I). The majority of arti-facts were recovered from this layer, including at least 54instances of knapping refits (e.g., flake-to-flake or flake-to-core). Thus, all lithic debris is located within or onthe margins of fluvial deposits separated by erosional un-conformities. Deposition is discontinuous with numerouscut-and-fill episodes documented across the broader siteas well as Area I.

Between 22 and 7 ka, alluvial fan deposition intensifieson the site, while human activity is poorly represented.There are fluvial deposits identified on T-3 (514 masl) at11 ka, while fluvial deposits also occur on T-1 (507 masl)at 7 ka. A low stand in lake levels centered around theLast Glacial Maximum (LGM, ca. 22–17 ka) inferred fromdiatoms (Gasse, Barker, & Johnson, 2002) is supportedby newer data derived from the 2005 Lake Malawi Sci-entific Drilling Project, indicating that lake levels fell bymuch as 100 m of its preceding and succeeding 700 mdeep level (Scholz et al., 2011; Stone, Westover, & Co-hen, 2011). Beuning et al. (2011) argue that grasses andafromontane vegetation are reflected in the terrigenouspollen assemblages recovered from the drill core duringthe LGM, which is concurrent with drier winters (seealso Ivory et al., 2012). On the other hand, phytolithdata from the east-central portion of the basin suggestthat the lowlands were wooded including high water-consuming panicoids and bambusoids (Mercader et al.,2013). These data suggest a mosaic landscape of wood-lands and grasslands, and hominins were likely focusedon exploiting resources from within the catchment re-gions where the biodiversity potential would have beenat its highest (Mercader et al., 2013).

Site level reconstructions of the paleohydrology atMwanganda’s Village suggests that the site was terracedwith meandering streams descending into the paleo-Rukuru River catchment between 22 and 7 ka (see alsoKaufulu, 1990). Until at least 11 ka, the landform wasgenerally aggrading within active alluvial fan deposition,



but there are carbonate-rich paleosols that developedduring periods of landform stability. After 11 ka, thereappears to have been significant fluvial downcutting con-current to continuing Chitmwe fan activity in the south-east aspect of the site.

During the late Holocene, the rate of Chitimwe fan de-position slowed significantly, and there is a laterite soilthat developed on the site. Colonization of the site by ironworking farmers is recorded from the northeast portion ofthe site, and it is believed that significant erosion of thesite concurrent to the introduction of increasingly inten-sive and modern agricultural techniques has impacted theintegrity of the MSA archaeological occupations. How-ever, the results show a continuation of an ancient prac-tice of riparian resource exploitation even into moderntimes. Although the means of exploitation are differentand the intensity of settlement is significantly higher thanduring the MSA and later phases, the motivation for se-lecting this area appears to have the common denomina-tor of seeking a predictable source of water and possiblyreadily available lithic materials.

Amplifying?

During the Late Pleistocene, Lake Malawi appears to havebeen an attractive environmental feature possibly serv-ing to connect southern and eastern Africa within highlyvariable regional climatic conditions (Beuning et al.,2011; Mercader et al., 2013) and hetereogeneously dis-tributed human (Salas et al., 2002), animal (Cohen et al.,2007), and plant (Cowling et al., 2008) population cen-ters across the African continent. Correlations betweentropical oceanic warming and polar glacial isotopic datasuggest that the latitudinal extents of the intertropicalconvergence zone (ITCZ), which governs the distributionof rainfall on the African continent, is generally in syncwith high-latitude glacial activity, albeit in lag time (Bard,Rostek, & Sonzogni, 1997; Brown et al., 2007; Sche-fuß et al., 2011). The ITCZ is the point of convergenceof northeasterly and southeasterly monsoons, bringingtropical rain where it is located, while adjoining areasreceive little or no moisture (Waliser & Gautier, 1993).The ITCZ is believed to have displaced significantly south-ward during the LGM and Younger Dryas (12.8–11.5 ka),which had profound effects on the distribution of rain-fall and vegetation biomes across the African continent(Adams & Faure, 1997; Garcin et al., 2007; Tierney &Russell, 2007; Gasse et al., 2008; Tierney & deMenocal,2013). Although Lake Malawi desiccated severely dur-ing the Middle Pleistocene, lake levels (reflecting rainfallwithin the catchment) remained relatively high (≥600m) during the Late Pleistocene (Scholz et al., 2011) com-pared to proxy data in southeastern Africa, the Tan-

ganyika Basin, Congo Basin, and East African Rift Valleylakes (Opperman & Heydenrych, 1990; Maley & Brenac,1998; Gibert et al., 2002; Johnson et al., 2002; Schefuß,Schouten, & Schneider, 2005; Gasse et al., 2008; Tierneyet al., 2008). Thus, as many other landscapes in south-ern and eastern Africa were becoming increasingly xeric,mosaic woodland environments favored by MSA foragersremained in the Malawi Basin (Mercader et al., 2013)and in southwestern Africa (Gasse, 2000; Gasse et al.,2008).

Later MSA technology found at the site between ∼42and 22 ka would have occurred during a period of rel-atively stable lake levels and local soil formation, whendifferent MSA technocomplexes are documented fromthe Ngalue Cave and Niassa sites, east side of LakeMalawi (Mercader & Fogelman, 2006; Mercader, Ben-nett, & Raja, 2008; Mercader et al., 2009, 2012, 2013).The limited data suggest that intermediate-high lake lev-els did not foster regional interactions (contra Trauthet al., 2010). If toolkits can be interpreted as cultur-ally mediated responses to ecological conditions (Boyd& Richerson, 1985: 290), the available data here sug-gest a focus on localized exploitation of resources with-out a tremendous degree of sharing technology across thebasin. We recognize that there is a potential lag in oc-cupation between the sites on the east side of the lakeand that of Mwanganda’s Village on the basis of inherentimprecision of the dating methods used. However, thereis a focus on localized raw material exploitation at thetwo locales, the lithic tool production and core reductionsystems employed appear considerably different, and thetemporal overlap of ages overlaps within 1-σ ; therefore,we propose that late MSA technological systems in theMalawi Basin were probably geographically restricted andthe degree of interregional technology sharing concomi-tantly low. The persistence of MSA foragers in portions ofthe Lake Malawi Basin in spite of expanding LSA popu-lations elsewhere across continental Africa may partiallyexplain apparent restricted foraging ranges. Additionally,the extreme topographic variability of the Rift Valley cre-ated a resource mosaic during this critical phase of hu-man evolution, and MSA populations seem focused onexploiting local riparian resources.

CONCLUSION

The site of Mwanganda’s Village in northern Malawidemonstrates recurrent hominin occupation of ripariancorridors dating from the Late Pleistocene through theHolocene. MSA archaeological assemblages (∼42–22 ka)were deposited during periods of relatively high lakelevels in the Malawi Basin (Scholz et al., 2011), which



contrasts with relatively dry conditions throughout muchof southern Africa (Maley & Brenac, 1998; Schefuß,Schouten, & Schneider, 2005; Gasse et al., 2008; Tierneyet al., 2008). Use of riparian corridors by hunter-gatherersis unsurprising (Lourandos, 1997; Stafford, Richards, &Anslinger, 2000; Jayaswal, 2002), and comes during aperiod of active alluvial fan deposition across the west-ern (Ring & Betzler, 1995) and eastern (Mercader et al.,2012) sides of Lake Malawi catalyzed by tectonic activ-ity and longer, more pronounced intra-annual wet-drycycles associated with the movement of the ITCZ (seealso Ivory et al., 2012). Within this environment, fluvialcatchments represented resource caches that could be ac-cessed periodically as part of a residentially mobile forag-ing strategy.

Reanalysis of Mwanganda’s Village has yielded a differ-ent set of interpretations regarding the timing and natureof the archaeological deposits (cf., Clark & Haynes, 1970;Clark et al., 1970; Kaufulu, 1990). Presently, there is lit-tle evidence to support a significant Sangoan occupationof the site. Artifacts recovered from a paleosol interpretedto be the same as that from the Clark Area 1 excavationare highly abraded, deposits are consistent with a channellag, and the fauna from the original assemblage suggeststhe same. The minimum constraining U-Th age from ele-phant tusk assayed from the site is 282 ka, demonstratingthe antiquity of the fossils relative to the dated humanoccupation of the site.

However, repeated use of riparian channels ∼42 kaand until some point after 22 ka coincides with otherlate MSA occupations known elsewhere from the MalawiBasin (Mercader & Fogelman, 2006; Mercader, Bennett,& Raja, 2008; Mercader et al., 2009, 2012, 2013), suggest-ing that hominins were inhabiting predictably resource-rich areas as climates changed toward colder and drierconditions across Africa. The preliminary evidence showsregionalism in the MSA toolkits utilized between theeast and west sides of Lake Malawi, suggesting frag-mentation and lack of interaction between populationsacross the basin. Although more dated archaeologicalhorizons are needed to test these hypotheses, the densityof MSA archaeological deposits across northern Malawicurrently being documented by the MEMSAP team willyield the data needed to understand the complex connec-tions between landscape change, resource use, and hu-man techno-behavioral evolution.

We thank our collaborators at the Malawi Ministry of Tourism,Wildlife and Culture for their assistance and permission in facili-tating this research. Fieldwork and analysis were funded by Na-tional Geographic-Waitt Foundation grant W115-10 (JT), Aus-tralian Research Council Discovery Project DP110101305 (JT),Korean Research Foundation Global Research Network Grant

2012032907 (DW), the UQ Field School, and a generous do-nation from Thomas Jones. An outstanding team of local crewworked on the Mwanganda excavations, with special thanksowed to Liton Adhikari and Gervasio Ngumbira. The 2011 and2012 excavations were partially conducted by students from theUniversity of Queensland and the Catholic University of Malawi.Julia Maskell, Jessica McNeil, Tierney Lu, Casey Frewen-Lord,and Kristina Lee devoted many hours to refitting artifacts fromArea I. Scott Robinson and Marina Bravo Foster were instru-mental in obtaining the dGPS data of the site. Kathy Schickand Nicholas Toth facilitated JT’s access to the assemblages fromMalawi curated at the Stone Age Institute, as well as permis-sion to sample the three tusk fragments for U-Th analysis. KathyStewart identified the catfish fragments and Arun Banerjee andJurg Tuckermann provided the identification of the three fossilsamples from the Stone Age Institute as being derived from fossilelephant ivory. U-Th dating of the elephant tusk was supportedby Australian Research Council LIEF grant LE0989067 to JZ andDiscovery grant DP0881279 to GP. AC thanks the Lake MalawiDrilling Project, NSF-Earth System History Program (NSF-EAR-0602404), DOSECC Inc., and LacCore for support. Two anony-mous reviewers and comments from Jamie Woodward were ofinvaluable assistance in producing this manuscript and we offerour heartfelt appreciation for their time and insights.

REFERENCES

Adams, J.M., & Faure, H. (1997). Preliminary vegetation

maps of the world since the Last Glacial Maximum: An aid

to archaeological understanding. Journal of Archaeological

Science, 24(7), 623–647.

Ambrose, S.H. (1998). Chronology of the Later Stone Age and

food production in East Africa. Journal of Archaeological

Science, 25, 377–392.

Arnold, L.J., & Roberts, R.G. (2009). Stochastic modelling of

multi-grain equivalent dose (De) distributions: Implications

for OSL dating of sediment mixtures. Quaternary

Geochronology, 4(3), 204–230.

Ashley, G.M., Ndiema, E.K., Spencer, J.Q.G., Harris, J.W.K.,

& Kiura, P.W. (2011). Paleoenvironmental context of

archaeological sites, implications for subsistence strategies

under Holocene climate change, northern Kenya.

Geoarchaeology, 26(6), 809–837.

Backwell, L., D’Errico, F., & Wadley, L. (2008). Middle Stone

Age bone tools from the Howiesons Poort layers, Sibudu

Cave, South Africa. Journal of Archaeological Science, 35,

1566–1580.

Bailey, G.N., Reynolds, S.C., & King, G.C.P. (2011).

Landscapes of human evolution: Models and methods of

tectonic geomorphology and the reconstruction of hominin

landscapes. Journal of Human Evolution, 60(3),

257–280.

Baquedano, E., Domınguez-Rodrigo, M., & Musiba, C. (2012).

An experimental study of large mammal bone modification

by crocodiles and its bearing on the interpretation of

crocodile predation at FLK Zinj and FLK NN3. Journal of

Archaeological Science, 39(6), 1728–1737.



Bar-Matthews, M., Marean, C.W., Jacobs, Z., Karkanas, P.,

Fisher, E.C., Herries, A.I.R., Brown, K., Williams, H.M.,

Bernatchez, J., Ayalon, A., & Nilssen, P.J. (2010). A high

resolution and continuous isotopic speleothem record of

paleoclimate and paleoenvironment from 90 to 53 ka from

Pinnacle Point on the south coast of South Africa.

Quaternary Science Reviews, 29(17–18), 2131–2145.

Bard, E., Rostek, F., & Sonzogni, C. (1997). Interhemispheric

synchrony of the last deglaciation inferred from alkenone

palaeothermometry. Nature, 385(6618), 707–710.

Barham, L., & Mitchell, P. (2008). The first Africans: African

archaeology from the earliest toolmakers to the most

recent foragers. Cambridge: Cambridge University Press.

Barut, S. (1994). Middle and Later Stone Age lithic

technology and land use in east African savannas. African

Archaeological Review, 12, 43–72.

Barut-Kusimba, S. (1999). Hunter-gatherer land use patterns

in later Stone Age East Africa. Journal of Anthropological

Archaeology, 18, 165–200.

Basell, L.S. (2008). Middle Stone Age (MSA) site distributions

in eastern Africa and their relationship to quaternary

environmental change, refugia and the evolution of Homo

sapiens. Quaternary Science Reviews, 27, 2484–2498.

Behrensmeyer, A.K. (1988). Vertebrate preservation in fluvial

channels. Palaeogeography, Palaeoclimatology,

Palaeoecology, 63, 183–199.

Betzler, C., & Ring, U. (1995). Sedimentology of the Malawi

Rift: Facies and stratigraphy of the Chiwondo Beds,

northern Malawi. Journal of Human Evolution, 28, 23–35.

Beuning, K.R.M., Zimmerman, K.A., Ivory, S.J., & Cohen,

A.S. (2011). Vegetation response to glacial–interglacial

climate variability near Lake Malawi in the southern

African tropics. Palaeogeography, Palaeoclimatology,


Bizzarro, M., Simonetti, A., Stevenson, R.K., & Kurszlaukis, S.

(2003). In situ 87Sr/86Sr investigation of igneous apatites

and carbonates using laser-ablation MC-ICP-MS.

Geochimica et Cosmochimica Acta, 67(2), 289–302.

Blome, M.W., Cohen, A.S., Tryon, C.A., Brooks, A.S., &

Russell, J. (2012). The environmental context for the

origins of modern human diversity: A synthesis of regional

variability in African climate 150,000–30,000 years ago.

Journal of Human Evolution, 62(5), 563–592.

Bonnefille, R., & Chalie, F. (2000). Pollen-inferred

precipitation time-series from equatorial mountains, Africa,

the last 40 kyr BP. Global and Planetary Change, 26, 25–50.

Boyd, R., & Richerson, P.J. (1985). Culture and the

evolutionary process. Chicago: University of Chicago Press.

Brown, E.T., Johnson, T.C., Scholz, C.A., Cohen, A.S., &

King, J.W. (2007). Abrupt change in tropical African

climate linked to the bipolar seesaw over the past 55,000

years. Geophysical Research Letters, 34, L20702,

doi:10.1029/2007GL031240.

Brown, K.S., Marean, C.W., Herries, A.I.R., Jacobs, Z.,

Tribolo, C., Braun, D., Roberts, D.L., Meyer, M.C., &

Bernatchez, J. (2009). Fire as an engineering tool of early

modern humans. Science, 325, 859–862.

Charrie-Duhaut, A., Porraz, G., Cartwright, C.R., Igreja, M.,

Connan, J., Poggenpoel, C., & Texier, P.-J. (2013). First

molecular identification of a hafting adhesive in the Late

Howiesons Poort at Diepkloof Rock Shelter (Western Cape,

South Africa). Journal of Archaeological Science, 40(9),

3506–3518.

Chorowicz, J. (2005). The East African rift system. Journal of

African Earth Sciences, 43(1–3), 379–410.

Clark, J.D. (1995). Introduction to research on the Chiwondo

Beds, northern Malawi. Journal of Human Evolution, 28,

3–5.

Clark, J.D., & Haynes, C.V. (1970). An elephant butchery site

at Mwanganda’s Village, Karonga, Malawi, and its

relevance for Palaeolithic archaeology. World Archaeology,

1, 390–411.

Clark, J.D., Haynes, C.V., Mawby, J.E., & Gautier, A. (1970).

Interim report on palaeoanthropological investigations in

the Lake Malawi Rift. Quaternaria, 13,

305–354.

Cohen, A., Stone, J.R., Beuning, K.R.M., Park, L.E., Rainthal,

P.N., Dettman, D., Scholz, C.A., Johnson, T.C., King, J.W.,

Talbot, M.R., Brown, E.T., & Ivory, S.J. (2007). Ecological

consequences of early Late Pleistocene megadroughts in

tropical Africa. Proceedings of the National Academy of

Science, 104, 16422–16427.

Contreras, J., Anders, M.H., & Scholz, C.H. (2000). Growth of

a normal fault system: observations from the Lake Malawi

basin of the east African rift. Journal of Structural Geology,

22(2), 159–168.

Cowling, S.A., Cox, P.M., Jones, C.D., Maslin, M.A., Peros,

M., & Spall, S.A. (2008). Simulated glacial and interglacial

vegetation across Africa: Implications for species

phylogenies and trans-African migration of plants and

animals. Global Change Biology, 14(4), 827–840.

Duller, G.A.T., Bøtter-Jensen, L., & Murray, A.S. (2003).

Combining infrared- and green-laser stimulation sources in

single-grain luminescence measurements of feldspar and

quartz. Radiation Measurements, 37(4), 543–550.

Dupont, L. (2011). Orbital scale vegetation change in Africa.


Ebinger, C.J., Rosendahl, B.R., & Reynolds, D.J. (1987).

Tectonic model of the Malawi Rift, Africa. Tectonophysics,

141(1–3), 215–235.

Ebinger, C.J., Deino, A.L., Drake, R.E., & Tesha, A.L. (1989).

Chronology of volcanism and rift basin

propagation—Rungwe Volcanic Province. East Africa

Journal of Geophysical Research, 94, 15785–15803.

Eren, M.I., Diez-Martin, F., & Dominguez-Rodrigo, M.

(2013). An empirical test of the relative frequency of

bipolar reduction in Beds VI, V, and III at Mumba

Rockshelter, Tanzania: Implications for the East African

Middle to Late Stone Age transition. Journal of

Archaeological Science, 40(1), 248–256.



Fain, J., Soumana, S., Montret, M., Miallier, D., Pilleyre, T., &

Sanzelle, S. (1999). Luminescence and ESR dating

beta-dose attenuation for various grain shapes calculated

by a Monte-Carlo method. Quaternary Science Reviews,

18(2), 231–234.

Feathers, J.K. (2002). Luminescence dating in less than ideal

conditions: Case studies from Klasies River Main Site and

Duinefontein, South Africa. Journal of Archaeological

Science, 29(2), 177–194.

Finney, B.P., Scholz, C.A., Johnson, T.C., & Trumbore, S.

(1996). Late quaternary lake-level changes of Lake Malawi.

In T.C. Johnson & E.O. Odada (Eds.), The limnology,

climatology, and paleoclimatology of the East African lakes

(pp. 495–508). Amsterdam: Gordon and Breach Publishers.

Galbraith, R.F., & Roberts, R.G. (2012). Statistical aspects of

equivalent dose and error calculation and display in OSL

dating: An overview and some recommendations.

Quaternary Geochronology, 11, 1–27.

Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., &

Olley, J.M. (1999). Optical dating of single and multiple

grains of quartz from jinmium rock shelter, northern

Australia, part 1, Experimental design and statistical

models. Archaeometry, 41(2), 339–364.

Garcin, Y., Vincens, A., Williamson, D., Buchet, G., & Guiot,

J. (2007). Abrupt resumption of the African Monsoon at

the Younger Dryas—Holocene climatic transition.

Quaternary Science Reviews, 26, 690–704.

Gasse, F. (2000). Hydrological changes in the African tropics

since the Last Glacial Maximum. Quaternary Science

Reviews, 19(1–5), 189–211.

Gasse, F., Barker, P., & Johnson, T.C. (2002). A 24,000 yr

diatom record from the northern basin of Lake Malawi. In

E.O. Odada & D.O. Olago (Eds.), The East African great

lakes: Limnology, paleolimnology and biodiversity (pp.

393–414). Dordrecht, The Netherlands: Kluwer Academic

Publishers.

Gasse, F., Chalie, F., Vincens, A., Williams, M.A.J., &

Williamson, D. (2008). Climatic patterns in equatorial and

southern Africa from 30,000 to 10,000 years ago

reconstructed from terrestrial and near-shore proxy data.


Gibert, E., Bergozoni, L., Massault, M., & Williamson, D.

(2002). AMS-C-14 chronology of 40.0 cal ka BP

continuous deposits from a crater lake (Lake Massoko,

Tanzania)—Modern water balance and environmental

implications. Palaeogeography, Palaeoclimatology,


Grun, R., Eggins, S., Aubert, M., Spooner, N., Pike, A.W.G., &

Muller, W. (2010). ESR and U-series analyses of faunal

material from Cuddie Springs, NSW, Australia: Implications

for the timing of the extinction of the Australian

megafauna. Quaternary Science Reviews, 29(5–6),

596–610.

Guerin, C., Bar-Yosef, O., Debard, E., Faure, M., Shea, J.J., &

Tchernov, E. (1996). Mission archeologique et

paleontologique dans le Pleistocene ancien d’Oubeidiyeh

(Israel): Resultats 1992–1994. Academie des Sciences, 322,

709–712.

Henshilwood, C.S. (2009). The origins of symbolism,

spirituality & shamans: Exploring Middle Stone Age

material culture in South Africa. In C. Renfrew & I. Morley

(Eds.), Becoming human: Innovation in prehistoric

material and spiritual cultures (pp. 29–49). Cambridge:

Cambridge University Press.

Henshilwood, C.S., & Dubreuil, B. (2011). The Still Bay and

Howiesons Poort, 77–59 ka: Perspective-taking and the

evolution of the modern human mind during the African

Middle Stone Age. Current Anthropology, 52(3), 361–400.

Ivory, S.J., Lezine, A.-M., Vincens, A., & Cohen, A.S. (2012).

Effect of aridity and rainfall seasonality on vegetation in

the southern tropics of East Africa during the

Pleistocene/Holocene transition. Quaternary Research,

77(1), 77–86.

Jain, M., Bøtter-Jensen, L., & Singhvi, A.K. (2003). Dose

evaluation using multiple aliquot quartz OSL: test of

methods and a new protocol for improved accuracy and

precision. Radiation Measurements, 37, 67–80.

Jayaswal, V. (2002). South Asian Upper Paleolithic. In P.N.

Peregrine & M. Ember (Eds.), Encyclopedia of prehistory

(pp. 326–343). New York: Springer US.

Johnson, T.C., Brown, E.T., McManus, J., Barry, S., Barker,

P., & Gasse, F. (2002). A high-resolution paleoclimate

record spanning the past 25,000 years in Southern East

Africa. Science, 296(5565), 113–132.

Kaufulu, Z.M. (1983). The geological context of some early

archaeological sites in Kenya, Malawi and Tanzania:

Microstratigraphy, site formation and interpretation.

Berkeley: University of California.

Kaufulu, Z.M. (1990). Sedimentary environments at the

Mwanganda site, Malawi. Geoarchaeology, 5(1), 15–27.

Kaufulu, Z.M., Vrba, E.S., & White, T.D. (1981). Age of the

Chiwondo Beds, Northern Malawi. Annals of the Transvaal

Museum, 33(1), 1–8.

King, G.C.P., & Bailey, G.N. (2006). Tectonics and human

evolution. Antiquity, 80, 265–286.

Klein, R.G. (2000). Archeology and the evolution of human

behavior. Evolutionary Anthropology, 9(1), 17–36.

Lane, C.S., Chorn, B.T., & Johnson, T.C. (2013). Ash from the

Toba supereruption in Lake Malawi shows no volcanic

winter in East Africa at 75 ka. Proceedings of the National

Academy of Sciences, 110(20), 8025–8029.

Lombard, M. (2012). Thinking through the Middle Stone Age

of sub-Saharan Africa. Quaternary International, 270,

140–155.

Londono, A.C., Forman, S.L., Eichler, T., & Pierson, J. (2012).

Episodic eolian deposition in the past ca. 50,000 years in

the Alto Ilo dune field, southern Peru. Palaeogeography,

Palaeoclimatology, Palaeoecology, 346–347, 12–24.

Lourandos, H. (1997). Continent of hunter-gatherers. New

York: Cambridge University Press.



Maley, J., & Brenac, P. (1998). Vegetation dynamics,

palaeoenvironments and climatic changes in the forests of

western Cameroon during the last 28,000 years B.P.

Review of Palaeobotany and Palynology, 99(2),

157–187.

McBrearty, S. (1988). The Sangoan-Lupemban and Middle

Stone Age sequence at the Muguruk site, western Kenya.

World Archaeology, 19(3), 388–420.

McBrearty, S., & Brooks, A.S. (2000). The revolution that

wasn’t: A new interpretation of the origin of modern

human behavior. Journal of Human Evolution, 39,

453–563.

Mejdahl, V., & Christiansen, H.H. (1994). Procedures used for

luminescence dating of sediments. Quaternary Science

Reviews, 13(5–7), 403–406.

Mercader, J., Bennett, T., & Raja, M. (2008). Middle Stone

Age starch acquisition in the Niassa Rift, Mozambique.

Quaternary Research, 70(2), 283–300.

Mercader, J., & Fogelman, A. (2006). Between East and

South: First Middle Stone Age site in Northern

Mozambique, Paleoanthropology Society. San Juan, Puerto

Rico.

Mercader, J., Asmeromb, Y., Bennett, T., Raja, M., & Skinner,

A. (2009). Initial excavation and dating of Ngalue Cave: A

Middle Stone Age site along the Niassa Rift, Mozambique.

Journal of Human Evolution, 57, 63–74.

Mercader, J., Gosse, J.C., Bennett, T., Hidy, A.J., & Rood, D.H.

(2012). Cosmogenic nuclide age constraints on Middle

Stone Age lithics from Niassa, Mozambique. Quaternary

Science Reviews, 47, 116–130.

Mercader, J., Bennett, T., Esselmont, C., Simpson, S., &

Walde, D. (2013). Phytoliths from Middle Stone Age

habitats in the Mozambican Rift (105–29 ka). Journal of

Human Evolution, 64(5), 328–336.

Murray, A.S., & Wintle, A.G. (2003). The single aliquot

regenerative dose protocol: Potential for improvements in

reliability. Radiation Measurements, 37(4–5), 377–381.

Njau, J., & Blumenschine, R.J. (2012). Crocodilian and

mammalian carnivore feeding traces on hominin fossils

from FLK Zinjanthropus and FLK NN Level 3, Late

Pliocene, Olduvai gorge, Tanzania. Journal of Human

Evolution, 63(2), 408–417.

Olaka, L., Odada, E., Trauth, M., & Olago, D. (2010). The

sensitivity of East African rift lakes to climate fluctuations.

Journal of Paleolimnology, 44(2), 629–644.

Olley, J.M., Pietsch, T., & Roberts, R.G. (2004). Optical dating

of Holocene sediments from a variety of geomorphic

settings using single grains of quartz. Geomorphology,

60(3–4), 337–358.

Opperman, H., & Heydenrych, B. (1990). A 22,000 year-old

Middle Stone Age camp site with plant remains from the

north-eastern Cape. South African Archaeological Bulletin,

45, 93–99.

Pante, M.C., & Blumenschine, R.J. (2010). Fluvial transport

of bovid long bones fragmented by the feeding activities of

hominins and carnivores. Journal of Archaeological

Science, 37(4), 846–854.

Pike, A.W.G., Hedges, R.E.M., & Van Calsteren, P. (2002).

U-series dating of bone using the diffusion-adsorption

model. Geochimica et Cosmochimica Acta, 66(24),

4273–4286.

Powell, A., Shennan, S., & Thomas, M.G. (2009). Late

Pleistocene demography and the appearance of modern

human behavior. Science, 324(5932), 1298–1301.

Prescott, J.R., & Hutton, J.T. (1994). Cosmic ray contributions

to dose rates for luminescence and ESR dating: Large

depths and long-term time variations. Radiation

Measurements, 23(2–3), 497–500.

Price, G.J., Webb, G.E., Zhao, J.-x., Feng, Y.-x., Murray, A.S.,

Cooke, B.N., Hocknull, S.A., & Sobbe, I.H. (2011). Dating

megafaunal extinction on the Pleistocene Darling Downs,

eastern Australia: The promise and pitfalls of dating as a

test of extinction hypotheses. Quaternary Science Reviews,

30(7–8), 899–914.

Price, G.J., Feng, Y.-x., Zhao, J.-x., & Webb, G.E. (2013).

Direct U-Th dating of vertebrate fossils with minimum

sampling destruction and application to museum

specimens. Quaternary Geochronology, 18, 1–8.

Ring, U., & Betzler, C. (1995). Geology of the Malawi Rift:

Kinematic and technosedimentary background to the

Chiwondo Beds, northern Malawi. Journal of Human

Evolution, 28, 7–21.

Rittenour, T.M. (2008). Luminescence dating of fluvial

deposits: Applications to geomorphic, palaeoseismic and

archaeological research. Boreas, 37(4), 613–635.

Roff, G., Clark, T.R., Reymond, C.E., Zhao, J.-x., Feng, Y.,

McCook, L.J., Done, T.J., & Pandolfi, J.M. (2013).

Palaeoecological evidence of a historical collapse of corals

at Pelorus Island, inshore Great Barrier Reef, following

European settlement. Proceedings of the Royal Society B:

Biological Sciences, 280(1750), 20122100.

Rosendahl, B.R. (1987). Architecture of continental rifts with

special reference to East Africa. Annual Review of Earth

and Planetary Sciences, 15, 445–494.

Salas, A., Richards, M., De la Fe, T., Lareu, M.-V., Sobrino, B.,

Sanchez-Diz, P., Macaulay, V., & Carracedo, A. (2002). The

making of the African mtDNA landscape. American

Journal of Human Genetics, 71(5), 1082–1111.

Salman, G., & Abdula, I. (1995). Development of the

Mozambique and Ruvuma sedimentary basins, offshore

Mozambique. Sedimentary Geology, 96(1–2), 7–41.

Schefuß, E., Schouten, S., & Schneider, R.R. (2005). Climatic

controls on central African hydrology during the past

20,000 years. Nature, 437(7061), 1003–1006.

Schefuß, E., Kuhlmann, H., Mollenhauer, G., Prange, M., &

Patzold, J. (2011). Forcing of wet phases in southeast

Africa over the past 17,000 years. Nature, 480, 509–512.

Schick, K.D. (1992). Geolarchaeological analysis of an

Acheulean site at Kalambo Falls, Zambia. Geoarchaeology,

7, 1–26.



Schneider, R.R., Muller, P.J., & Acheson, R. (1999). Atlantic

alkenone sea-surface temperature records. In F. Abrantes &

A.C. Mix (Eds.), Reconstructing ocean history (pp. 33–55).

New York: Springer US.

Scholz, C.A., Johnson, T.C., Cohen, A.S., King, J.W., Peck,

J.A., Overpeck, J.T., Talbot, M.R., Brown, E.T.,

Kalindekafe, L., Amoako, P.Y.O., Lyons, R.P., Shanahan,

T.M., Castaneda, I.S., Heil, C.W., Forman, S.L., McHargue,

L.R., Beuning, K.R., Gomez, J., & Pierson, J. (2007). East

African megadroughts between 135 and 75 thousand years

ago and bearing on early-modern human origins.

Proceedings of the National Academy of Science, 104,

16416–16421.

Scholz, C.A., Cohen, A.S., Johnson, T.C., King, J., Talbot,

M.R., & Brown, E.T. (2011). Scientific drilling in the Great

Rift Valley: The 2005 Lake Malawi Scientific Drilling

Project—An overview of the past 145,000 years of climate

variability in Southern Hemisphere East Africa.

Palaeogeography, Palaeoclimatology, Palaeoecology, 303,

3–19.

Sitzia, L., Bertran, P., Boulogne, S., Brenet, M., Crassard, R.,

Delagnes, A., Frouin, M., Hatte, C., Jaubert, J., Khalidi, L.,

Messager, E., Mercier, N., Meunier, A., Peigne, S.,

Queffelec, A., Tribolo, C., & Macchiarelli, R. (2012). The

paleoenvironment and lithic taphonomy of Shi’Bat Dihya

1, a Middle Paleolithic site in Wadi Surdud, Yemen.

Geoarchaeology, 27(6), 471–491.

Soriano, S., Villa, P., & Wadley, L. (2007). Blade technology

and tool forms in the Middle Stone Age of South Africa:

The Howiesons Poort and post-Howiesons Poort at Rose

Cottage Cave. Journal of Archaeological Science, 34(5),

681–703.

Stafford, C.R., Richards, R.L., & Anslinger, C.M. (2000). The

bluegrass fauna and changes in Middle Holocene

hunter-gatherer foraging in the Southern Midwest.

American Antiquity, 65(2), 317–336.

Stone, J.R., Westover, K.S., & Cohen, A.S. (2011). Late

Pleistocene paleohydrography and diatom paleoecology of

the central basin of Lake Malawi, Africa. Palaeogeography,

Palaeoclimatology, Palaeoecology, 303(1–4), 51–70.

Stuut, J.-B.W., Prins, M.A., Schneider, R.R., Weltje, G.J.,

Jansen, J.H.F., & Postma, G. (2002). A 300-kyr record of

aridity and wind strength in southwestern Africa:

Inferences from grain-size distributions of sediments on

Walvis Ridge, SE Atlantic. Marine Geology, 180(1–4),

221–233.

Thompson, J.C., Mackay, A., & Welling, M. (2011). Malawi

Earlier-Middle Stone Age Project—2010 Project Report.

Lilongwe: Malawi Department of Antiquities.

Thompson, J.C., Mackay, A., Wright, D.K., Welling, M.,

Greaves, A., Gomani-Chindebvu, E., & Simengewa, D.

(2012). Renewed investigations into the Middle Stone Age

of northern Malawi. Quaternary International, 270,

129–139.

Thompson, J.C., Welling, M., & Gomani-Chindebvu, E.

(2012). Malawi Earlier-Middle Stone Age Project—2011

Preliminary Project Report. Lilongwe: Malawi Department

of Antiquities.

Thompson, J.C., Welling, M., & Gomani-Chindebvu, E. (in

press). Using GIS to integrate old and new archaeological

data from Stone Age deposits in Karonga, Malawi.

International Journal of Heritage in the Digital Era, 2(4).

Tierney, J.E., & deMenocal, P.B. (2013). Abrupt shifts in horn

of Africa hydroclimate since the Last Glacial Maximum.

Science, 342(6160), 843–846.

Tierney, J.E., & Russell, J.M. (2007). Abrupt climate change

in southeast tropical Africa influenced by Indian monsoon

variability and ITCZ migration. Geophysical Research

Letters, 34(15), L15709.

Tierney, J.E., Russell, J.M., Huang, Y., Damste, J.S.S.,

Hopmans, E.C., & Cohen, A.S. (2008). Northern

Hemisphere controls on tropical southeast African climate

during the past 60,000 years. Science, 322(5899),

252–255.

Trauth, M.H., Deino, A.L., Bergner, A.G.N., & Strecker, M.R.

(2003). East African climate change and orbital forcing

during the last 175 kyr BP. Earth and Planetary Science

Letters, 206(3–4), 297–313.

Trauth, M.H., Maslin, M.A., Deino, A.L., Junginger, A.,

Lesoloyia, M., Odada, E.O., Olago, D.O., Olaka, L.A.,

Strecker, M.R., & Tiedemann, R. (2010). Human evolution

in a variable environment: The amplifier lakes of Eastern

Africa. Quaternary Science Reviews, 29(23–24),

2981–2988.

Tryon, C.A., & McBrearty, S. (2002). Tephrostratigraphy and

the Acheulian to Middle Stone Age transition in the

Kapthurin Formation, Kenya. Journal of Human

Evolution, 42(1/2), 211–236.

Van Schmus, W.R. (1995). Natural radioactivity of the crust

and mantle. In T.J. Ahrens (Ed.), Global earth physics: A

handbook of physical constants (Vol. 1, pp. 283–291).

Washington, DC: American Geophysical Union.

Vincens, A., Chalie, F., Bonnefille, R., & Tiercelin, J. (1993).

Pollen-derived rainfall and temperature estimates for Lake

Tanganyika and their implication for Late Pleistocene

water levels. Quaternary Research, 40, 343–350.

Waliser, D.E., & Gautier, C. (1993). A satellite-derived

climatology of the ITCZ. Journal of Climate, 6(11),

2162–2174.

Williams, M.A.J., Williams, F.M., Duller, G.A.T., Munro, R.N.,

El Tom, O.A.M., Barrows, T.T., Macklin, M.G., Woodward,

J.C., Talbot, M.R., Haberlah, D., & Fluin, J. (2010). Late

Quaternary floods and droughts in the Nile valley, Sudan:

New evidence from optically stimulated luminescence and

AMS radiocarbon dating. Quaternary Science Reviews,

29(9–10), 1116–1137.

Willoughby, P. (2007). The evolution of modern humans in

Africa: A comprehensive guide. Lanham, Maryland: Alta

Mira Press.

Winder, I.C., King, G.C.P., Deves, M., & Bailey, G.N. (2013).

Complex topography and human evolution: The missing

link. Antiquity, 87(336), 333–349.



Wintle, A.G. (2008). Luminescence dating: Where it has been

and where it is going. Boreas, 37(4), 471–482.

Woodward, J.C., Macklin, M.G., & Welsby, D.A. (2001).

Holocene alluvial history and the palaeochannels of the

River Nile in the Northern Dongola Reach. In D.M. Maddy,

M.G. Macklin, & J.C. Woodward (Eds.), River basin

sediment systems: Archives of environmental change (pp.

327–356). Rotterdam: A. A. Balkema.

Wright, D.K., Forman, S.L., Kusimba, C.M., Pierson, J.,

Gomez, J., & Tattersfield, P. (2007). Stratigraphic and

geochronological context of human habitation along the

Galana River, Kenya. Geoarchaelogy, 22(7), 709–728.

Wright, D.K., Forman, S.L., Waters, M.R., & Ravesloot, J.C.

(2011). Holocene eolian activation as a proxy for

broad-scale landscape change on the Gila River Indian

Community, Arizona. Quaternary Research, 76(1), 10–21.

Zhao, J.-x., Hu, K., Collerson, K.D., Xu, H.-k., & Falgueres, C.

(2001). Thermal ionization mass spectrometry U-series

dating of a hominid site near Nanjing, China. Geology,

29(1), 27–30.

Zhou, H., Zhao, J., Qing, W., Feng, Y., & Tang, J. (2011).

Speleothem-derived Asian summer monsoon variations in

Central China, 54–46 ka. Journal of Quaternary Science,

26(8), 781–790.


Renewed Geoarchaeological Investigations of Mwanganda’s Village (Elephant Butchery Site), Karonga,...

Documents

Transcript of Renewed Geoarchaeological Investigations of Mwanganda’s Village (Elephant Butchery Site), Karonga,...