Transcript
Page 1: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

Research Article

A Geoarchaeological Study of the Middle and Upper PleistoceneLevels at Canteen Kopje, Northern Cape Province, South AfricaMatt G. Lotter,1,* Ryan J. Gibbon,2,7 Kathleen Kuman,1,3 George M. Leader,4 Tim Forssman,1,5

and Darryl E. Granger6

1School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, South Africa2Department of Anthropology, University of New Brunswick, Fredericton, New Brunswick, Canada3Evolutionary Studies Institute, University of the Witwatersrand, Johannesburg, South Africa4Department of Sociology and Anthropology, The College of New Jersey, Ewing, New Jersey, USA5Rock Art Research Institute, University of the Witwatersrand, Johannesburg, South Africa6Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West Lafayette, Indiana, USA7Department of Geological Sciences, University of Cape Town, Cape Town, South Africa

Correspondence*Corresponding author; E-mail:

[email protected]

Received13 February 2015

Revised4 September 2015

Accepted11 September 2015

Scientific editing by Curtis Marean

Published online in Wiley Online Library

(wileyonlinelibrary.com).

doi 10.1002/gea.21541

Canteen Kopje, situated in the Northern Cape Province of South Africa, hastwo main archaeological deposits: alluvial gravels and a mantle of overlyingfine sediments known locally as the “Hutton Sands.” This paper focuses on thefine sediments, the three industries contained within them, and the interfacewith the underlying gravels in an attempt to clarify their formation and trans-formation. A Fauresmith assemblage is found at this interface; it is thus crucialto understand the processes of deposition and modification at this poorly un-derstood boundary. The methods used in this study involved the analysis ofartifact depositional (dip and orientation) and spatial data, artifact condition,raw materials, and assemblage size profiles. Data presented document the mix-ing between the lowest levels of the fine sediments and the underlying allu-vial gravels. This study thus provides important contextual information for theFauresmith industry at Canteen Kopje. C© 2016 Wiley Periodicals, Inc.

INTRODUCTION

Canteen Kopje, situated within the Vaal River Basin ofthe Northern Cape Province, South Africa (Figure 1),has been referred to frequently in literature over thelast century (e.g., Goodwin, 1934; Van Riet Lowe,1937; Partridge & Brink, 1967; Helgren, 1977, 1979).However, only recently has more systematic researchbeen pursued (see De Wit, 1996, 2008; Beaumont &McNabb, 2000; McNabb, 2001; Beaumont, 2004a;Gibbon, Leader, & Kuman, 2008, 2009, Gibbon et al.,2009; Forssman et al., 2010; Sarupen, 2010; McNabb &Beaumont, 2011a,b; Chazan et al., 2013; Leader, 2014).The first controlled excavations at Canteen Kopje wereconducted by P. Beaumont in Areas 1 and 2 (Figure 1)beginning in 1997 (McNabb & Beaumont, 2011a), andrecently in 2007 two new teams began research at thesite. This included a collaborating team from Purdue Uni-versity and the University of the Witwatersrand, startingnew excavations at Pit 6 (Figure 1), and a team led by M.Chazan from the University of Toronto (see Figure 1—siteA). Overall, this recent research has focused on improv-

ing our understanding of the site’s geomorphology, age,archaeology, and site formation history.

Canteen Kopje is unique for an open-air site in south-ern Africa in that it contains a 7 m stratified sequenceof alluvial gravel and fine sediment deposits which pre-serve two early Acheulean assemblages, overlain by Vic-toria West Acheulean and Later Acheulean industries,and then by Middle (MSA) and Later Stone Age (LSA)levels (Figure 2; Gibbon, Leader, & Kuman, 2008, 2009;Forssman et al., 2010; Leader, 2014). The typological,technological and contextual nature of these assemblageshas been well-documented by several studies (Gibbon,Leader, & Kuman, 2008, 2009; Gibbon et al., 2009;Forssman et al., 2010; Sarupen, 2010; McNabb &Beaumont, 2011a,b; Chazan et al., 2013; Leader, 2014),however until now, little work has focused specifically onassessing the degree to which the fine sediments at Can-teen Kopje have been altered and how this has influencedthe assemblages contained within.

In this paper, we address the site formation and trans-formation processes at play in the Later Acheulean toMSA levels that are contained within the fine sediments

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 1

Page 2: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Figure 1 Location of Canteen Kopje and

plan view of site showing locations of the

various excavations (modified from McNabb,

2001). A “koppie” defines a small flat-topped

hill. See Helgren (1979) and de Wit (2008) for

discussions concerning the paleochannel.

and the top of the gravels. In particular, we address thecontext of the Later Acheulean levels identified as a Fau-resmith industry. Canteen Kopje is one of only a fewsites that preserves stratified in situ Fauresmith mate-rial (Underhill, 2011). In previous research (McNabb &Beaumont, 2011a,b; see Area 1 excavation, Figure 1), thecontext of this assemblage was not well defined becauseof its stratigraphic position at the base of the fine sedi-ments and within the contact zone that incorporates thetop of the gravels, as the fine sediment component in thisexcavation had been removed by miners.

The Fauresmith industry has featured prominently insouthern African literature over the last several decades,yet defining the industry has been problematic (see Un-derhill, 2011 for a detailed review). Assemblage descrip-tions through time have been based primarily upon poorcontext surface collections, unexcavated samples, or,on those assemblages which “highlight” certain “Faure-smith” components (i.e., small handaxes, see Fock, 1968

specifically, and see Goodwin, 1927, 1929; Goodwin &van Riet Lowe, 1929; Farnden, 1967; Humphreys, 1970;Mitchell, 1998; Klein, 2000; Porat et al., 2010; Underhill,2011). As a result, reliable details on complete assem-blages have not been available until recently. Tradition-ally, the Fauresmith was considered by most researchersas a transitional industry between the late Acheuleanand the MSA (Goodwin & van Riet Lowe, 1929; seeUnderhill, 2011 for a detailed review). Clark (1959) as-signed it to the onset of regional specializations in Africa,and some researchers have placed it in the early MSA(Herries, 2011). More recently, however, dates of ca. 500ka (thousand years ago) for the Fauresmith at Kathu Pan(Porat et al., 2010) indicate that the industry has somedepth of time in the Later Acheulean. Although there isno single type series which can be used to describe theindustry, the majority of researchers emphasize the co-occurrence of prepared core technology, points, blades,and small handaxes (Kuman, 2001; Porat et al., 2010).

2 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 3: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Figure 2 Stratigraphic profile of the Pit 6 Canteen Kopje deposits (left). Explanation of the mixed contact zone (MCZ) (right).

In addition to Canteen Kopje, sites containing Fau-resmith assemblages include: Wonderwerk Cave, KathuPan, Rooidam, Pniel, Hopefield, and Bundu Farm (Singer& Crawford, 1958; Fock, 1968; Beaumont, 1990b,c,d,2004b,c,d; Kuman, 2001; Beaumont & Richardt, 2004;Beaumont & Vogel, 2006; Kiberd, 2006; Porat et al.,2010). Although these sites have “improved” contexts,their “Fauresmith” assemblage descriptions are still verylimited. However, of all these sites, Kathu Pan providesthe most detailed information on Fauresmith technol-ogy (Porat et al., 2010). The authors also highlight thatthe current radiometric ages provided elsewhere for theFauresmith do not temporally constrain the industry.As a result, there is a great need to obtain radiomet-ric ages from well-defined Fauresmith assemblages ex-cavated from controlled in situ contexts (Porat et al.,2010).

Accordingly, understanding the processes at CanteenKopje that have affected the fine sediments and under-lying gravels is vital. The assemblages at Canteen Kopjeprovide an important record of hominid behaviors andactivities. However, until these assemblages are discussedin light of their spatial, chronological, and functional con-text, our interpretations are meaningless (Morton, 2004).Understanding the formation and transformation of a siteis crucial before behaviors can be inferred from the everchanging archaeological record (Schiffer, 1983; Schick,1991; Karkanas et al., 2000; Stein, 2001; Morton, 2004).Through the investigation of artifact size distribution,condition, deposition, spatial, and raw material data, wediscuss the influence of these processes at play in the Can-teen Kopje deposits. The current study provides new datawhich strengthens the contextual understanding of theFauresmith assemblage at Canteen Kopje.

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 3

Page 4: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

BACKGROUND: THE DEPOSITIONALSEQUENCE AT CANTEEN KOPJE ANDCHRONOLOGY

In Pit 6, the top of the sequence has a 60–70 cm thickcapping of miners’ dump material which contains a rangeof disturbed finds; this is underlain by the fine sediments(these fine sediments are known locally as the “HuttonSands”; however, these will be referred to as “fine sedi-ments” throughout this paper; see Figure 2). The upper-most levels of the fine sediments (70–140 cm) containa late Holocene LSA assemblage (Forssman et al., 2010)and below it (at 140–170 cm) an MSA assemblage withartifacts that are not diagnostic of any particular phase(Sarupen, 2010). Beneath this MSA level, from 170 to195 cm, lies the Fauresmith/Later Acheulean assemblage,followed thereafter by the mixed contact zone (MCZ)from 195 to 230 cm. The boundary between the finesediments and the underlying gravels is unconformable,with the Fauresmith in fine sediments resting on thesurface of an alluvial gravel from 230 cm downwards(Figure 2), with a slight dip to the NE of between 5˚ and10˚. This upper gravel layer, which contains the VictoriaWest prepared core Acheulean industry, has been datedto the early Pleistocene through the application of cos-mogenic nuclide burial dating (Gibbon et al., 2009, 2013;see Leader, 2014 for a detailed report of these underly-ing gravels). Overall, the stratigraphy of the gravels inPit 6 is largely similar to the sequence documented byBeaumont (see McNabb & Beaumont, 2011a,b) in Area 1(Figures 1, 2).

Although our attempts to date the Fauresmith andoverlying MSA levels in Pit 6 were not successful (Evans& Cunningham, 2013), it may be possible to provide anage estimate for these deposits based on the work ofChazan et al. (2013). Their results with Optically Stim-ulated Luminescence (OSL) dating of the sands in theirexcavation (Figure 1—site A) show that the earliest sandgrains began accumulating at ca. 300 ka, which mayprovide a minimum age for the fine sediment depositacross the surrounding area. Thus, in Pit 6, the Fau-resmith assemblage-bearing fine sediments are likely tohave started forming by at least ca. 300 ka. It is clearthat the erosional disconformity between the uppermostgravel layer and the fine sediment capping in Pit 6 rep-resents a gap in time of hundreds of thousands of years.This is confirmed by the earliest reported date for a Fau-resmith assemblage from the nearby site of Kathu Pan, ca.500 ka (potentially as old as 647 ka; Porat et al., 2010).Other Fauresmith assemblages appear to continue in timeto the end of the Earlier Stone Age (Klein, 2000; Herries2011), which today is considered to be ca. 300–250 ka—

namely, Bundu Farm (dated to between 400 and 200 ka),Wonderwerk Cave (dated to between 315 and <195 ka),and Rooidam (dated to >209 ka).

The Fine Sediments

Several studies address the origin and composition ofthese fine sediments. Originally, Helgren (1977, 1979)argued that the formation and deposition of these sed-iments was a result of colluvial processes, namely twospecific intervals of valley colluviation, termed the River-ton III and early Riverton IV (Beaumont, 1990a). Mostrecently Chazan et al. (2013) highlight that these sedi-ments are of aeolian origin and comprise yellow to redsilty sand; the sand component is proposed to have beenintroduced locally from the adjacent Vaal drainage sys-tem (based on modal grain size data). Bioturbation ofthe fine sediments at Canteen Kopje has been widelyproposed (De Wit, 2008; Chazan et al., 2013; Evans &Cunningham, 2013; Leader, 2014).

The influence of bioturbation

De Wit (2008) highlighted the poorly sorted nature ofthe fine sediments and suggested that aeolian processesand bioturbation were essential to their alteration. Morerecently two studies focusing specifically on the datingof these sediments, using OSL techniques, highlight thatbioturbation has played a significant role in the rework-ing of sediments across the site (Chazan et al., 2013;Evans & Cunningham, 2013). Attempts by Chazan et al.(2013) to provide greater chronological resolution for thesite’s late Pleistocene/early Holocene assemblages, per-taining to the late MSA or early LSA, were only ableto provide rough time estimates (due to the large agescatter caused by the mixing of grains). The attempt byEvans and Cunningham (2013) to date the Pit 6 sed-iments containing MSA and Fauresmith artifacts showthat all four OSL dating results fall within the Pleistocene,but post-depositional movement of sand grains has re-sulted in over-dispersion percentages that range between30% and 60%. This is well over the maximum rate of15% expected for reliable results and directly indicatesthe mixing of sand grains through bioturbation, proba-bly primarily through insect activity (as demonstrated byChazan et al., 2013) and potentially also by plant roots.Some dates are also stratigraphically inverted (Evans &Cunningham, 2013).

Bioturbation in these unconsolidated sediments is alsolikely to have created the vertically stretched distributionof lithics in each of the three assemblages—vertical dis-tribution figures cited above range from ca. 70 cm for the

4 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 5: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

LSA, to 30–40 cm for the MSA, and at least 30–40 cm forthe Fauresmith with mixing into the top of the gravels.Stretching of lithic assemblages was also noted by Chazanet al. (2013) for their excavation.

Understanding the influence of bioturbation is there-fore crucial in order to explain the depositional se-quence at Canteen Kopje. The effects of bioturbationhave been addressed in a number of studies (Cahen &Moeyersons, 1977; Moeyersons, 1978; Schiffer, 1983;Stein, 1983; Erlandson, 1984; Hofman, 1986; McBrearty,1990; Vermeersch & Bubel, 1997; Morton, 2004). Ter-mites, plant roots, and burrowing animals are seen asthe primary means through which many assemblagecontexts are disturbed. McBrearty (1990) and Morton(2004) describe that termite activity can cause the fol-lowing features in deposits: relict (often fossil) termite re-mains, no form of microstratigraphy (a single homoge-nous unit), artifact concentrations which have markedupper and lower contacts, and/or a dispersal of artifacts.Artifact distributions can be affected quite considerably,primarily vertically but also horizontally. Plant roots oftenresult in the mixing of soils and can be considerably de-structive to site contexts (Morton, 2004). Personal ob-servations have shown that root systems often entan-gle artifacts, and the toppling of dead trees can leadto an upheaval of roots containing sediments, causingthe local vertical displacement of artifacts. Burrowinganimals can cause artifact displacement both horizon-tally and vertically, leading to the homogenization (ormixing) of site stratigraphy (Erlandson, 1984; Morton,2004).

Models for sediment deposition

The work of Chazan et al. (2013) provides an importantinsight into the depositional events which led to the for-mation of the fine sediments at Canteen Kopje. Theirstudy provides evidence for depositional events at 300and 165 ka; the fine sediments then remained exposedto bioturbation until much later (23–26 ka; Chazan et al.,2013). Most significant is that their data suggest evidencefor a “multiple episode” depositional model (Chazanet al., 2013). If deposition of these sediments occurredas a single event, bioturbation would have been con-strained to only the upper levels within these sediments(Chazan et al., 2013). Chazan et al. (2013) also consid-ered a single depositional event model for the site but didnot favor this explanation. Additional support for “mul-tiple episode” deposition is provided by De Wit (2008),based on the description of several mottled horizons(which are seen as having been deposited at differentperiods) and by Beaumont (2004a), based on palaeosolpreservation.

METHODOLOGY

Excavations through portions of the fine sediments andupper gravels at Canteen Kopje were conducted atPit 6, within an already established radial excavation(Figure 1, see Forssman et al., 2010 for a detailed discus-sion of this excavation). Excavations within the fine sedi-ments proceeded in 5 cm spits, which increased to 10 cmspits in the gravel layer due to the presence of large cob-bles and boulders. All excavated sediment was screenedusing both 4 and 2 mm sieve mesh, and all artifacts�20 mm in size were 3D point-plotted using a total sta-tion. In addition, all artifacts and natural clasts �20 mmwere analyzed with a Brunton compass for dip angles andlong axis orientations (readings relate to magnetic north)to investigate depositional patterns (e.g., those created byalluvial processes).

The analysis of clast orientation is a well-establishedmethod for distinguishing depositional processes. Depo-sitional patterns within alluvial sequences and a rangeof other process-derived deposits have been dealt withby many authors (e.g., Krumbein, 1939; Andrews &Smithson, 1966; Tandon & Kumar, 1981; Butler, 1982;Schiffer, 1983; Yagishita & Jopling, 1983; Petts & Foster,1985; Bluck, 1987; Deitrich, 1987; Naden & Brayshaw,1987; Schick, 1991, 1997; Petraglia & Potts, 1994; Evans,2000; Bridge, 2003; Morton, 2004; Millane, Weir, &Smart, 2006; Kostic & Aigner, 2007; Hodge, Brasington,& Richards, 2009). According to Petts and Foster (1985),these patterns (dip directions/orientations and angles,called imbrication) occur as a result of the final stagesof artifact (or clast) deposition; if flow regimes are suf-ficiently low, imbrication will be absent (Deitrich, 1987).Long axis orientation of clasts is regarded as being theprimary indicator of fluvial flow (Schick, 1991), whereasdip is used to assess the intensity (as fluvial scour, un-der the base of an artifact, will give rise to steeper anglesin higher energy conditions within easily erodible sub-strates) and direction of fluvial flow, and also, to ques-tion the redistribution of assemblage components (suchas small debitage; Schick, 1991; Petraglia & Potts, 1994;Morton, 2004).

In addition, Schick (1984, 1991) highlights the size pat-terns that are created when an assemblage is affected byfluvial flow. Sites affected by moderate-to-intense fluvialflow will exhibit: a loss of assemblage components (espe-cially small flaking debris <20 mm, or SFD), a differentialre-deposition of assemblage components based on size,and the retention of some SFD (except where fluvial flowis intense). Sites exposed to minimal/moderate fluvialflow will preserve most of their original assemblage com-ponents (differentially, based on artifact size and densi-ties), will frequently have portions of the site completely

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 5

Page 6: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

devoid of lithic material, and will have a differential re-location of assemblage components beyond the originalsite boundary (points of reaccumulation where flow isreduced).

Laboratory analysis involved the recording of artifactmaximum lengths for the study of assemblage profiles.Experimental knapping studies by Schick (1987, 1991,1997) and by Kuman and Field (2009) provide modernanalogues to interpret assemblage profiles in archaeolog-ical deposits; their experiments demonstrate what types,sizes, and distributions of flaked material should charac-terize a site of primary knapping activity. As such, assem-blage profiles should be dominated by a high percentage(60–87%) of SFD provided that stone knapping occurredwithin the catchment of the site, that the deposit hasbeen minimally affected by natural processes, and thatthe sieve mesh used in excavations is no larger than 4 mm(Schick, 1997; Kuman & Field, 2009). Deviations in theabove pattern may be the result of lithic influx (transport-ing in of manufactured lithics), removal (through fluvialwinnowing), or re-deposition of sorted material by fluvialforces (Schick, 1991). Additional modification of assem-blage components, beyond fluvial or cultural processes,could include both aeolian (and aeolian-derived sedi-ment processes) and less-intense movement by water,such as sheetwash (Schiffer, 1983; Schick, 1984, 1987,1991, 1997; Kandel, Felix-Henningsen, & Conard, 2003;Morton, 2004).

Additional lab analysis involved the assessment of ar-tifact raw material type (all pieces) and condition (ex-cluding SFD). Artifact condition was established usinga system adapted from Shea (1999). This system wasbest suited to the depositional context of Canteen Kopjeas it allowed for the assessment of both minimally andhighly weathered/abraded pieces. The term weatheringhere refers to the chemical decay/breakdown of raw ma-terials (artifacts), whereas abrasion refers to a “wear-ing down” of artifact surfaces/edges caused specificallyby either the movement of the artifact over an abra-sive surface/medium, or the movement of an abra-sive medium over the artifact surfaces/edges. No at-tempt was made to distinguish between these differentconditions so the following collective types were uti-lized: heavily weathered/abraded pieces (distinct round-ing/modification of edges and ridges), intermediate orslightly weathered/abraded pieces (less-intensive round-ing/modification of edges and ridges), and fresh orunabraded pieces (little to no edge/ridge modification,sharp angular edges). The following raw material typeswere recorded during artifact classification: Ventersdorplava, quartzite, quartz, fine-grained materials (crypto-crystalline silicates, including all cherts, agates and chal-cedonies, and hornfels).

Table I Total assemblage samples, by size, for the excavations at Canteen

Kopje.

Total Sample (n = 3355)

Assemblage Level (cm) <20 mm �20 mm Total

LSA 100–140 637 232 869

MSA 140–170 545 261 806

Fauresmith 170–195 353 250 603

MCZ 195–230 270 673 943

Alluvial gravels 230–240 11 123 134

RESULTS

Excavations through portions of the fine sediments andupper gravels at Canteen Kopje have yielded an assem-blage of 3355 artifacts (Table I). In addition, 274 dip andorientation readings and 292 point plot readings wereobtained during the excavations (only data most per-tinent to the levels in question will be presented). AllMSA data presented here were obtained during a separatestudy (spatial and depositional data were not obtained;Sarupen, 2010).

Assemblage Size Profile Data

Figure 3 illustrates the size distributions of material forthe respective assemblages in Pit 6. To assess the levelof variability in these size distributions, Kruskal–Wallis(KW) tests were performed (summarized in Table IIa andb). There is a difference in the proportions of material be-tween the upper levels with LSA, MSA, and Fauresmith,compared to the lower MCZ and alluvial gravels (TableIIa; n = 3355; P = 0.000). In the LSA levels, over 70%of the entire assemblage falls within the <20 mm cat-egory. Small sizes are typical of LSA industries, but thepresence of lithics <20 mm shows that the respective lev-els have not lost much material. The MSA levels haveonly slightly less than 70% SFD preserved, which showsrelatively good retention of material and minimal distur-bance to the deposit. The Fauresmith levels comprised59% SFD. In contrast to these upper levels, the lowerMCZ and coarse alluvium (230 cm downwards) containgreatly reduced SFD percentages (29% and 8%, respec-tively). In terms of the range of artifact sizes (>20 mm),the LSA levels appear to show the most restricted dis-tribution of larger artifacts (none >160 mm, to be ex-pected), whereas the MSA and Fauresmith levels havea fuller range (up to >180 mm) of pieces >20 mm, al-beit small quantities of each. The coarse alluvium has itslargest quantity of material >20 mm, with many largerpieces reaching >200 mm in length (Figure 3). Table IIbconfirms that there is a statistically significant difference

6 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 7: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Figure 3 Artifact size distribution (profile) for the respective assemblages at Canteen Kopje (n = 3355).

in the size distribution of material between all of theassemblages (n = 3355; P = 0.000).

Artifact Depositional and Spatial Data

Dip and orientation analysis for the LSA, MCZ, andalluvial gravels is shown in Figure 4 (an insufficientnumber of readings were obtained for the Fauresmithlevels). Table III provides a statistical assessment ofthese variables for the three assemblages. A Rayleightest of uniformity was performed to assess whether theassemblage data are uniformly distributed (random)or whether there is a tendency for data to cluster(nonrandom, directed; McPherron, 2005; Bernatchez,2010; Oestmo et al., 2014). A significant value (P <

0.05) indicates a preferential orientation (in one or moredirections); the Z-value indicates how great the clusteringof data is around the mean vector (the larger the value,the more likely the data are nonuniform in distribution).

Long axis orientation readings for the LSA shownortherly and easterly components (Figure 4). The high-est frequency of readings occurs between 320–350º and65–80º. In addition, the lowest frequency of readings oc-curs between 180º and 320º. Table III supports a sig-nificant clustering of the data suggesting a nonrandomdistribution (n = 140; Z = 9.066; P = 0.0001). Fig-ure 4 shows that dip angles are most common be-tween 5º and 15º (>40 readings) with others prevalentbetween 15º and 25º (>20 readings). The Rayleigh P

and Z values for the LSA dip data show a very strong

Table II Artifact size distribution Kruskal–Wallis test statistical data for the upper and lower assemblages (a), and all assemblages (b).

Assemblage Sample Mean Rank Test Statistics Results

(a)

Upper (LSA, MSA and Fauresmith) n = 2278 1423.99 Chi-square (χ2) 593.221

Lower (MCZ and alluvial gravels) n = 1077 2215.26 Degrees of freedom (df) 1

Total n = 3355 Asymptotic significance (P) 0.000

(b)

LSA n = 869 1324.85 Chi-square (χ2) 669.050

MSA n = 806 1415.40 Degrees of freedom (df) 4

Fauresmith n = 603 1578.35 Asymptotic significance (P) 0.000

MCZ n = 943 2146.77

Alluvial gravels n = 134 2697.27

Total n = 3355

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 7

Page 8: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Figure 4 Artifact orientation (left) and dip angle (right) data for the LSA 100–140 cm (n= 140), MCZ 195–230 cm (n= 92), and alluvial gravels>240 cm (n

= 50); insufficient data were obtained for the Fauresmith levels and no data were obtained for the MSA levels. Length of line corresponds to the number

of readings; segments represent 5º intervals.

clustering, suggesting a nonuniform distribution of thedata (Table III).

Long axis orientation readings obtained for the MCZare random (Figure 4). The highest frequency of read-

ings occurs between 285º and 290º (n = 4); the remain-ing readings occupy almost all cardinal directions rangingfrom 0º to 350º. Table III supports the uniform distribu-tion of the MCZ orientation data (n = 92; Z = 0.242;

8 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 9: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Table III Artifact dip and orientation Rayleigh test statistical data for the LSA (n = 140), MCZ (n = 92), and alluvial gravels (AG) beneath 240 cm (n = 50).

Orientation Dip

Assemblage Mean Vector (°) Circular SD (°) Rayleigh Z Rayleigh P Mean Vector (°) Circular SD (°) Rayleigh Z Rayleigh P

LSA (n = 140) 32.323° 94.791° 9.066 0.0001 19.297° 15.543° 130.067 <0.0001

MCZ (n = 92) 193.28° 139.647° 0.242 0.785 16.147° 13.311° 87.166 <0.0001

AG > 240 cm (n = 50) 73.646° 81.343° 6.662 0.001 24.584° 13.369° 47.351 <0.0001

P = 0.785). Additional dip angle data show that thereis a large proportion (40%) of material which has a 0–10º dip, with the remaining pieces (60%) exhibiting dipsbetween 10º and 45º (Figure 4). Table III data show asignificant clustering of the dip readings, suggesting anonuniform distribution of the data (n = 92; Z = 87.166;P < 0.0001).

Depositional data for the alluvial gravels beneath the240 cm level show a strong easterly component in ori-entations (Figure 4). The highest frequency of readings(n = 6) occurs between 45–50º, 110–115º, and 165–170º; a limited number of readings occur between 205ºand 340º. Table III Rayleigh Z and P values support thenonuniform distribution of the data (n = 50; Z = 6.662;P = 0.001). Dip angle data readings are most frequent be-tween 10º and 30º. Table III dip data also support a non-random clustering of these readings.

Artifact spatial data (n = 97) for the fine sedimentand alluvial gravel interface are presented in Figure 5.Figure 5a shows the distribution of artifacts, by size, indi-cating that there is a higher frequency of smaller artifacts(20–59 mm category) within the MCZ than in the lowerlying alluvial gravels. The remaining size categories ap-pear to show a random spatial distribution within boththe MCZ and gravels. Figure 5b shows the distributionof artifacts by condition. For the alluvial gravels from230 cm downwards, there is a clear dominance of heav-ily weathered/abraded pieces (only three pieces occur be-neath the 230 cm level which are fresh/unabraded andslightly weathered/abraded). The MCZ shows a mix ofartifact condition types, which appear to be random indistribution.

Raw Material and Artifact Condition Data

Raw material and artifact condition data are presentedin Figures 6–8 and Tables IV–VII. Figure 6 shows thatthe LSA levels 70–140 cm (data from Forssman et al.,2010) have the highest proportion of fine-grained rawmaterials (56.8%). Ventersdorp lava accounts for 37.1%,followed thereafter by quartz (5.2%). The underlyingMSA levels are also dominated by fine-grained mate-

rials (49.5%), followed thereafter by Ventersdorp lava(31.8%). The greatest quantity of quartz is also foundin these levels (11.5%). The Fauresmith assemblage alsoshows that fine-grained materials are dominant (45.9%).Ventersdorp lava for these depths increases to 41.8%,and quartz and quartzite are similar in their distribu-tions (5.6% and 6.6%, respectively). Table IVa data con-firm that the distribution of raw materials in these up-per assemblages is significantly different (KW; n = 9761;P = 0.000).

For the remaining assemblages (MCZ vs. gravels), theraw material distributions differ when compared withone another (Table IVb; KW; n = 1077; P = 0.003). Ven-tersdorp lava only accounts for 37% within the MCZ,whereas for the gravels it is the dominant raw material(61.2%). Fine-grained materials dominate in the MCZwith 49.5%, and the remaining quartz and quartzite per-centages between these two samples are largely similar.Table IVc data confirm the significant difference in theseraw material distributions, with depth (KW; n = 10,838;P = 0.000).

Artifact condition by raw material for the Faure-smith and MCZ is shown in Tables V and VI. TheFauresmith raw materials show a clear absence ofheavily weathered/abraded pieces (fine-grained materi-als highest at 2.5%). The majority therefore comprisesfresh/unabraded and slightly weathered/abraded mate-rial, collectively 98% for Ventersdorp lava, 97.5% for thefine-grained materials, and 100% for both quartz andquartzite (Table V). From these data, it appears that thereis little difference in artifact condition by raw material.However, for the MCZ, Table VI shows that there aremuch larger proportions of heavily weathered/abradedmaterial, most specifically for Ventersdorp lava (59.9%,with only 11.2% comprising fresh/unabraded). Fine-grained materials are also dominated by heavily weath-ered/abraded pieces, coupled with a similar proportion offresh/unabraded pieces (43.8% and 43.1%, respectively).Quartz and quartzite are the least weathered/abradedpieces for these levels (Table VI).

Artifact condition for the Fauresmith, MCZ and al-luvial gravel assemblages is presented in Figures 7

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 9

Page 10: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Figure 5 Spatial distribution of artifacts within the MCZ and alluvial gravels by size (a) and condition (b) (n = 97).

and 8. Figure 7 illustrates that the Fauresmith has ahigh concentration of fresh/unabraded pieces (56%).Slightly weathered/abraded pieces account for 42% andthe remaining 2% are heavily weathered/abraded. Incontrast, the MCZ is dominated by heavily weath-ered/abraded pieces (51%); fresh/unabraded pieces ac-count for only 25%. Figure 8 illustrates the change inweathered/abraded pieces by depth, most important forthe Fauresmith and MCZ levels. The Fauresmith levelsare clearly dominated by fresh/unabraded and slightlyweathered/abraded pieces (collectively 98.3% for 170–

180 cm and 97.8% for 180–195 cm). For the MCZ, highpercentages (collectively 62%) of fresh/unabraded andslightly weathered/abraded pieces can be found between195 and 210 cm; heavily weathered/abraded pieces ac-count for 38% of the sample for these levels (a largeincrease compared to the 2.2% from 180 to 195 cm).With depth, this percentage of well-preserved materialdecreases considerably. At a final depth of 230–240 cm,76% of the assemblage is heavily weathered/abradedwith only a small percentage of fresh/unabraded mate-rial (<10%). A comparison of artifact condition by depth

10 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 11: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Figure 6 Raw material distributions for the respective assemblages at Canteen Kopje. Data for the LSA levels 70–140 cm were obtained from Forssman

et al. (2010); 11 pieces with unidentifiable raw materials are excluded from this sample.

is summarized in Table VII; these results show that thereis a significant difference in artifact condition with depth(KW; n = 1046; P = 0.000).

DISCUSSION

Assemblage Profiles

Assemblage integrity at Canteen Kopje has been shownto vary with depth (Figure 3; Tables IIa and b). The upperLSA levels have a high portion of small material (SFD >

70%), and orientation data for pieces in these levels showpreferred orientations toward the north and east (dip dataalso show a strong clustering; Table III). High levels ofpreservation are to be expected of a low-energy, aeolianenvironment (where the removal of material is minimal).According to Schick (1987, 1991, 1997) and Kuman andField (2009), this layer would appear minimally affectedby transformation processes, provided knapping occurred

in the area. A reconcentration of this SFD by moderateor intense fluvial flow appears unlikely although there isa clustering in orientation and dip data. However, with-out a more detailed assessment of these dip and orien-tation readings, across a greater area of the excavation,conclusions can only be tentative. The young age of thisassemblage (A.D. 1436–1870) coupled with its microlithiccharacteristic may also account for this high preservation(see Forssman et al., 2010). It is also still possible for theselevels to have been affected by processes which do not re-sult in a complete removal of this smaller material; suchprocesses would most likely also account for the completelack of stratigraphy in the fine sediments. This appears tobe the result of bioturbation, noted by several studies (DeWit, 2008; Forssman et al., 2010; Chazan et al., 2013).

In addition, trampling may also have modified theselayers. Several studies discuss the effect of tram-pling (Schiffer, 1983; Villa & Courtin, 1983; Gonza-les et al., 1985; Nielsen, 1991; Vermeersch & Bubel,

Figure 7 Artifact condition for the Fauresmith and MCZ levels.

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 11

Page 12: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Figure 8 Artifact condition by depth for the Fauresmith, MCZ, and underlying alluvial gravel levels.

1997; McBrearty et al., 1998; Morton, 2004; Forssman &Pargeter, 2014). Experimental studies by McBrearty et al.(1998), Gonzales et al. (1985), Nielsen (1991), and Ver-meersch and Bubel (1997) establish patterns in down-ward artifact movements in sediments; this will vary dueto sediment type and the density, shape and size of theartifacts themselves. A size sorting of materials (as dif-ferent sized pieces are differentially relocated), both ver-tically and horizontally, is possible (Nielsen, 1991); how-ever, trampling seldom results in the removal of material,rather, it merely displaces it (Morton, 2004). For the LSAlevels, future research should focus specifically on iden-tifying any artifact damage which could indicate the in-fluence of trampling; however, it is highly possible thatassemblage components were worked into the surfacesediments by trampling (hence promoting initial down-ward artifact movement/burial).

Assemblage integrity for the MSA levels is high basedon the SFD component, the full range of artifact sizes(which refutes lithic influx), and the variety of artifacttypes (such as core trimming elements; Sarupen, 2010)which indicate primary knapping (Figure 3). If minor re-distribution of the SFD did take place, it would have beenby processes of bioturbation and/or low-energy sheet-

wash (Schiffer, 1983; Morton, 2004). Trampling couldalso have played a role in the initial downward move-ment of artifacts; however, artifact damage studies onthese levels are needed to confirm such findings.

Assemblage integrity for the Fauresmith levels is some-what different from the overlying MSA and LSA assem-blages. Overall, the SFD component is slightly low, at59% (Figure 3); this falls just below the low end of therange established in experiments by Schick (1987) usinga 4 mm sieve mesh (60%) and well below the figure es-tablished by Kuman and Field (2009) using 2 mm mesh(87%). Artifacts for these levels include a high represen-tation of cortex on flakes, as well as core rejuvenationelements (including partial core tablets; Kuman et al., inprep.). Based on these assemblage components, off-sitelithic manufacture does not appear to have been signifi-cant. Possible reasons for a slightly low SFD componentare dilution by mixing of larger pieces from the gravelsbelow, minor loss of SFD by sheetwash, or changes insite function over time. Such changes may relate specifi-cally to the sites use as a knapping (factory) area; duringthe Fauresmith there may have been greater variabilityin such use, leading to a lower SFD distribution for theselevels.

12 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 13: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Table IV Artifact rawmaterial by depth Kruskal–Wallis test statistical data

(test statistic values include chi-square, degrees of freedom, and asymp-

totic significance); upper assemblages comparison (a), lower assemblages

comparison (b), and all assemblages comparison (c).

Assemblage Sample Mean Rank Test Statistics Results

(a)

LSA n = 8352 4919.02 χ2 19.652

MSA n = 806 4797.25 df 2

Fauresmith n = 603 4466.31 P 0.000

Total n = 9761

(b)

MCZ n = 943 548.46 χ2 8.528

Alluvial gravels n = 134 472.44 df 1

Total n = 1077 P 0.003

(c)

LSA n = 8352 5560.58 χ2 155.172

MSA n = 806 5440.94 df 4

Fauresmith n = 603 5062.53 P 0.000

MCZ n = 943 4600.72

Alluvial gravels n = 134 3865.75

Total n = 10,838

Table V Artifact condition by raw material for the Fauresmith levels

(VL= Ventersdorp lava, Qtze= quartzite, Qtz= quartz, FG= fine-grained

materials).

Fauresmith:

170–195 cm

(n = 250)

Fresh/

Unabraded

(%)

Slightly

Weathered/

Abraded

(%)

Heavily

Weathered/

Abraded

(%) Total

VL 59.7 38.3 2.0 100% (n = 154)

Qtze 93.7 6.3 0 100% (n = 16)

Qtz 100 0 0 100% (n = 1)

FG 40.5 57.0 2.5 100% (n = 79)

Table VI Artifact condition by raw material for the MCZ levels (VL = Ven-

tersdorp lava,Qtze=quartzite, Qtz=quartz, FG= fine-grainedmaterials).

MCZ

195–230 cm

(n = 673)

Fresh/

Unabraded

(%)

Slightly

Weathered/

Abraded

(%)

Heavily

Weathered/

Abraded

(%) Total

VL 11.2 28.9 59.9 100% (n = 436)

Qtze 63.3 17.7 19.0 100% (n = 79)

Qtz 100 0 0 100% (n = 5)

FG 43.1 13.1 43.8 100% (n = 153)

The MCZ and alluvial gravels have even more reducedSFD components (Figure 3). Most significant is the almostcomplete absence of SFD in the gravels. This is furthersupported by Leader (2014) in which mention is madeof the significant absence of this smaller material withinthe gravels because the alluvial gravels have been de-posited within a high-energy environment. Interestingly,

TableVII Artifact conditionbydepthKruskal–Wallis test statisticaldata for

the Fauresmith (n= 250), MCZ (n= 673), and alluvial gravels (n= 123; test

statistic values include chi-square, degrees of freedom, and asymptotic

significance).

Assemblage Sample Mean Rank Test Statistics Results

Fauresmith n = 250 299.66 χ2 234.884

MCZ n = 673 570.73 df 2

Alluvial gravels n = 123 720.04 P 0.000

Total n = 1046

the SFD component for the MCZ is higher relative to thealluvial gravels, supporting that the MCZ contains mixedcontents from the Fauresmith above.

Evidence for Bioturbation in the MCZ

Artifact spatial, condition, and raw material data for theFauresmith and MCZ highlight the mixing of some Fau-resmith assemblage components into the MCZ horizon.The Fauresmith extends from 170 cm below datum downto the top of the gravels at 195 cm, and it is mixedinto the top 35 cm of the gravels (down to 230 cm).Fresh/unabraded pieces are more prominent from 195 to210 cm, within the MCZ, when compared to the under-lying 230–240 cm level (Figure 7). The underlying grav-els appear undisturbed by bioturbation. Table V showsthat the level of mixing between the Fauresmith in thefine sediments (170–195 cm) and the lower MCZ is low,due mainly to the limited distribution of highly weath-ered/abraded pieces for these levels, irrespective of rawmaterial type. Table VI, however, shows a very differentpattern. Within the MCZ, the increase in heavily weath-ered/abraded pieces thus highlights the movement andmixing of pieces from the alluvial gravels (as this heavilyweathered/abraded component cannot be derived fromthe overlying fine sediments where fresh material is thenorm). The top 35 cm of the gravels thus include a mixof alluvially deposited artifacts and younger Fauresmithpieces from the aeolian sediments. This is further sup-ported by Figure 5 which indicates a mix of fresh andheavily weathered/abraded pieces within the MCZ; it isclear that the fresh pieces are derived from the upperfine sediments, whereas the weathered/abraded piecesare from the gravels.

We believe the most likely mechanism through whichartifacts would come to rest within the gravels is the de-position of Fauresmith artifacts on the irregular gravelsurface with a shallow sediment covering and biotur-bation, primarily caused by extensive plant root actionand possible tree falls. Bioturbation would only havebeen possible provided the deposition of the fine sedi-ments occurred during multiple events, as proposed by

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 13

Page 14: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Table VIII Abrasion by depth for the 210–250 cm levels.

Abrasion by Depth (n = 15,227)

Level Fresh Slightly Abraded Abraded Very Abraded Heavily Worn Total

210–220 cm 78.3% 67.8% 30.9% 3.0% 0.7% n = 389

220–230 cm 4.3% 27.8% 12.6% 0.9% 0.4% n = 158

230–250 cm 0% 0% 2.3% 1.7% 2.7% n = 368

Total n = 19 (82.6%) n = 86 (95.6%) n = 178 (45.8%) n = 223 (5.6%) n = 409 (3.8%)

Total sample n = 23 (100%) n = 90 (100%) n = 388 (100%) n = 3992 (100%) n = 10,734 (100%)

Reproduced with permission from Leader (2014). Shaded cells highlight notable percentage differences.

Chazan et al. (2013), and that initially a shallow coveringof sediment would have been deposited facilitating thedownward growth of roots into the alluvial gravel surface(which would have resulted in the mixing of the deposit).Sediments would then have been continually reworkedthrough time during subsequent fine sediment deposi-tion, giving rise to the vertical distribution that is evidenttoday (Figure 5). Additional disruption would also havebeen caused by termites (evident at the site today) andanimal burrowing.

The data provided here offer evidence for a multi-ple episode depositional model and suggest that singleepisode deposition of the fine sediments at Canteen Kopjewould have been implausible as it would have preventedthe mixing of Fauresmith artifacts into the top of thegravels. A long time-averaged accumulation for the Fau-resmith is likely, and it is also possible that some Fau-resmith pieces were deposited directly onto the irregulargravel surface, prior to fine sediment deposition. How-ever, it is clear that a large proportion of the Fauresmithassemblage was deposited within and during the depo-sition of the fine sediments as many are fresh in condi-tion (Kuman et al., in prep.). Trampling damage, whichwould be common on pieces exposed at the gravel surface(prior to the onset of fine sediment deposition), is also un-common. This coupled with the position of the unalteredFauresmith levels (higher up between 170 and 195 cm)supports that deposition could have occurred both beforebut definitely during fine sediment accumulation.

Table X Flake platform facets for the 210–250 cm levels.

Flake Platform Facets

Level

(cm)

Plain

(%)

Two Facets

(%)

Multiple Facets

(%) Total

210–220 77 20 3 n = 67 (100%)

220–230 91 4 5 n = 33 (100%)

230–250 93 4 3 n = 96 (100%)

Reproduced with permission from Leader (2014). Shaded cells highlight

notable percentage differences.

Tables VIII–X present original data obtained from a sep-arate study focusing on the 210–250 cm levels in Pit6 (Leader, 2014). Table VIII shows artifact condition bydepth, most interestingly highlighting that for an assem-blage with a total sample of 15,227 pieces, 82.6% ofthe entire fresh component comes from the upper 210–230 cm; slightly abraded pieces account for 95.6% fromthese same levels (combined, 93% of all the freshestpieces occur from 210 to 230 cm; Leader, 2014). Thesepatterns are not dissimilar to those presented in Figures 6and 7, where fresh assemblage components were shownto decrease with depth within the MCZ, dropping off sig-nificantly after 240 cm.

Technological analysis conducted by Leader (2014) onassemblage components is highlighted by Tables IX andX. Flake dorsal scar patterns (modified from McNabb &Beaumont, 2011a) for the upper 210–230 cm levels in-dicate a different distribution compared to those for the

Table IX Flake dorsal scar patterns for the 210–250 cm levels.

Flake Dorsal Scar Patterns

Level 1 2 3 4 5 6 7 8 9 10 11 12 Total

210–220 cm 6.0% 17.9% 7.5% 4.5% 1.5% 4.5% 4.5% 3.0% 32.8% 1.5% 13.4% 3.0% n = 67 (100%)

220–230 cm 3.0% 21.2% 0% 9.1% 0% 0% 3.0% 0% 12.1% 6.1% 42.4% 3.0% n = 33 (100%)

230–250 cm 0% 7.3% 0% 3.1% 0% 0% 1.0% 1.0% 31.3% 1.0% 46.9% 8.3% n = 96 (100%)

Reproduced with permission from Leader (2014). Shaded cells highlight notable percentage differences.

1= Y-pattern, 2= one central ridge, 3= two converging ridge, 4= one offset ridge, 5= double Y, 6= radial, 7= two parallel ridge, 8=multiple parallel

ridge, 9 = all other ridge, 10 = multiple convergent ridge, 11 = cortex, 12 = indeterminate.

14 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 15: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

underlying Victoria West levels (Table IX; an increase in“Y,” one central ridge, two converging ridges, and oneoffset ridge patterns is present); Leader (2014:218; au-thor’s note in parenthesis) provides two possible rea-sons as to why this may have occurred: “First, thereis an advancement in this assemblage, in the knappingtechniques, which would provide more organized dorsalpatterns” and “second . . . admittedly the more likely ex-planation, is the mixing from the upper [fine sediment]levels.” The mixing of Fauresmith assemblage compo-nents into these upper levels is directly responsible forthe increased appearance in these dorsal scar patternsin the MCZ. This is supported furthermore by Table X,highlighting that the greatest number of flakes with twofacets occurs in the uppermost 210–220 cm level. TheFauresmith assemblage at Canteen Kopje is more ad-vanced in preparation and appearance than that whichoccurs lower down within the unmixed alluvial gravels.It is clear from Leader (2014) that on aspects of bothtechnology and artifact condition, mixing of assemblagecomponents has occurred within the upper 210–230 cmlevels.

The influence of fluvial forces on the fine sedimentassemblages at Canteen Kopje appears to be absent.Figure 4 suggests that fluvial forces are not likely tohave altered the MCZ levels significantly, post-deposition,although bioturbation may have disturbed these patternsto some extent. The relatively high proportion of SFD inthe MSA levels and the small size of the LSA material re-flect good site context and the absence of fluvial forces.A more detailed investigation into the preferential orien-tation of pieces in the LSA levels is needed before con-clusions can be made. Overall, there are no dip and ori-entation trends exhibiting flow patterns or velocities andfluvial site modification is unlikely (the fine sedimentsalso show no form of bedding/cross-bedding that wouldbe indicative of fluvial flow).

In contrast, additional research carried out by D.Granger, R. Gibbon, and M. Lotter on the underlying al-luvial deposits (Figure 4) illustrates that there are signif-icant trends in both the dip and orientation of artifactsand natural clasts beneath the 240 cm level (Table III,Figure 4). This finding further supports the argument thatthe interface of the upper gravel and lower fine sedimentshas been bioturbated, disturbing the original depositionalfabric.

Model for the Formation of the Canteen KopjeDeposits

Figure 9 illustrates, in sequence, the series of depo-sitional, formational, and transformational steps whichhave led to the formation of the Canteen Kopje deposits

(Pit 6); erosion and deposition would have continuallyaffected all of these levels throughout their deposition/formation:

� The alluvial gravels are deposited during the earlyPleistocene (Gibbon et al., 2013); clasts and artifactsretain a distinct fabric and the deposit has a slope ofbetween 5º and 10° (Figure 9a).

� Bioturbation and selection by hominids of raw ma-terials for knapping disturbs the upper gravel sur-face (imbrication patterns). Alluvial and/or colluvialsediments may have repeatedly covered and recov-ered these exposed gravels (Figure 9b).

� The fine sediments start to accumulate by at leastca. 300 ka (Chazan et al., 2013); some Fauresmithartifacts may have been deposited upon the irreg-ular gravel surface prior to fine sediment deposi-tion, but most were deposited on top of the finesediments that formed a covering over the gravels.Artifacts then became worked into the sediment.Trampling may have played a role in the initialdownward movement of pieces into the fine sedi-ments as the site is frequented by knappers; sheet-wash may have caused the removal of some of thesmallest assemblage components (Figure 9c).

� Continual bioturbation causes sediment mixing inthese levels, most specifically, through the growthof roots downwards into the upper gravel lay-ers. Over time, Fauresmith artifacts are stretchedthroughout the lower fine sediments. Repeatedmixing leads to the formation of the MCZ (finesediments with gravels). Possible tree falls furtherpromote the mixing of the fine sediments andgravels. The mixing of assemblage componentsalso promotes dilution of the SFD for these levels(Figure 9d).

� Fine sediments are further deposited within whichthe younger MSA material accumulates. Bioturba-tion continues and trampling may have assisted indownward artifact movements (Figure 9e).

� As these sediments are reworked reconsoli-dation occurs, allowing for the stretching ofassemblage components in vertical and horizontalspace (Figure 9f).

� A similar sequence of events can be seen with thedeposition of the LSA assemblage (between A.D.1436 and 1870; Forssman et al., 2010; Figure 9g).

� The assemblage is stretched as bioturbation reworksthe sediments. This extensive bioturbation causessome (minor) mixing between the MSA and LSA as-semblages in the 130–140 cm level (Forssman et al.,2010; Figure 9h).

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 15

Page 16: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Figure 9 Schematic diagram illustrating the depositional sequence and site formation/transformation processes at work on the Canteen Kopje sequence

(image not to scale).

� The final surface overburden layer covering muchof the Canteen Kopje site is deposited through theactions of diamond miners in the late 1800s. Thisoverburden contains a full mix of sediment from theunderlying deposits (Figure 9i).

SUMMARY AND CONCLUSION

The objective of this study was to determine the processesof site formation and disturbance affecting Stone Age as-semblages in the fine sediments which disconformably

16 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 17: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

overlie ancient alluvial gravels at Canteen Kopje. Allthree Stone Age assemblages in the fine sediments haveundergone a degree of vertical movement that we at-tribute largely to bioturbation by roots and insects, butthree distinct industries are nevertheless represented interms of technology and raw material proportions. TheLSA (100–140 cm below datum) is very young, at ca.450 years. It contains the largest proportion of fine-grained raw materials (Forssman et al., 2010). MSA lev-els occur from 140 to 170 cm and are dominated byfresh crypto-crystalline and hornfels lithics, with Venters-dorp lava comprising 31.8% (Sarupen, 2010). The Fau-resmith levels (170–195 cm) contain a higher propor-tion of Ventersdorp lava and a slightly more reduced SFDcomponent.

The Fauresmith levels were found to be mixed throughbioturbation into the top 35 cm of the gravel deposit.This is supported by both the mixed distribution of freshand weathered/abraded artifacts within the upper allu-vial gravel levels, and the presence of dip and orientationreadings which are random regardless of a clearly visi-ble dip in the gravel deposit (coupled with distinct clastimbrication below the 240 cm level). We argue that thiswould only have been possible through extensive root ac-tion in the sediments and surface of the gravels; in addi-tion, this would only have been possible with slow, punc-tuated deposition of the fine sediments across the site.

This research allows us to now isolate the Fauresmithassemblage and describe another variant of this impor-tant industry, which in this case appears to be a formof final Acheulean transitional to the MSA. We providedata to argue that the Fauresmith artifacts do not be-long to the alluvial gravels, which are considerably older,but to the younger fine sediments that lie unconformablyon the surface of the gravels. The detailed typologicaland technological analysis of the Fauresmith assemblagewill be the subject of a separate paper (Kuman et al.,in prep.).

Most important is that new data provided here showconclusively that mixing between these upper fine sedi-ments and underlying alluvial gravels has occurred. Thispaper provides the first ever documentation of mixing be-tween Vaal Gravels, dated to the later early Pleistocene,and fine sediments (Hutton Sands), probably dating toat least ca. 300 ka, in which the Fauresmith industry atCanteen Kopje is contained.

M.G.L. would like to thank the Palaeontological Scientific Trust(PAST) and its Scatterlings of Africa programs, the National Re-search Foundation and the University of the Witwatersrand forfinancial support. We also thank Dr. Alex Sumner, Ashor Saru-pen, Nicole Da Costa Dias, and Andrew Phaswana for field as-sistance. We would also like to thank the reviewers for help-ful comments and suggestions on an earlier draft of this paper.

Many thanks to Dr. Jonah Choiniere for help with image prepa-ration, Shani Reddy for laboratory assistance, Dr. Mathew Caru-ana and Dr. Hao Li for help with statistics. R.J.G. and T.F. wouldlike to thank the Palaeontological Scientific Trust. K.K. thanksthe Palaeontological Scientific Trust for funding since the incep-tion of this project, and the National Research Foundation formultiple grants used for this research, the most recent of whichhas been CPRR Grant No. 81782.

REFERENCES

Andrews, J.T., & Smithson, B.B. (1966). Island, Northwest

Territories, Canada till fabrics of the cross-valley moraines

of North-Central Baffin. Geological Society of America

Bulletin, 77, 271–290.

Beaumont, P.B. (1990a). Canteen Koppie (Klipdrift). In P.B.

Beaumont & D. Morris (Eds.), Guide to archaeological sites

in the Northern Cape (pp. 14–16). Kimberley: McGregor

Museum.

Beaumont, P.B. (1990b). Wonderwerk Cave. In P.B.

Beaumont & D. Morris (Eds.), Guide to archaeological sites

in the Northern Cape (pp. 101–134). Kimberley: McGregor

Museum.

Beaumont, P.B. (1990c). Kathu Pan. In P.B. Beaumont & D.

Morris (Eds.), Guide to archaeological sites in the

Northern Cape (pp. 75–100). Kimberley: McGregor

Museum.

Beaumont, P.B. (1990d). Rooidam 1 & 2. In P.B. Beaumont &

D. Morris (Eds.), Guide to archaeological sites in the

Northern Cape (pp. 19–21). Kimberley: McGregor

Museum.

Beaumont, P.B. (2004a). Canteen Kopje. In D. Morris & P.B.

Beaumont (Eds.), Archaeology in the Northern Cape: Some

key sites (pp. 26–30). Kimberley: McGregor Museum.

Beaumont, P.B. (2004b). Wonderwerk Cave. In D. Morris &

P.B. Beaumont (Eds.), Archaeology in the Northern Cape:

Some key sites (pp. 31–36). Kimberley: McGregor

Museum.

Beaumont, P.B. (2004c). Kathu Pan and Kathu

Townlands/Uitkoms. In D. Morris & P.B. Beaumont (Eds.),

Archaeology in the Northern Cape: Some key sites (pp.

50–53). Kimberley: McGregor Museum.

Beaumont, P.B. (2004d). Pniel 6 (The Bend). In D. Morris &

P.B. Beaumont (Eds.), Archaeology in the Northern Cape:

Some key sites (pp. 61–63). Kimberley: McGregor

Museum.

Beaumont, P.B., & McNabb, J. (2000). Canteen Kopje: The

recent excavations. The Digging Stick, 17, 3–6.

Beaumont, P.B., & Richardt, F. (2004). On the margins of the

Karee Vloer: Rooidam 1 & 2 and Biesiesput. In D. Morris &

P.B. Beaumont (Eds.), Archaeology in the Northern Cape:

Some key sites (pp. 22–25). Kimberley: McGregor

Museum.

Beaumont, P.B., & Vogel, J.C. (2006). On a timescale for the

past million years of human history in central South Africa.

South African Journal of Science, 102, 217–228.

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 17

Page 18: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Bernatchez, J.A. (2010). Taphonomic implications of

orientation of plotted finds from Pinnacle Point 13B

(Mossel Bay, Western Cape Province, South Africa).

Journal of Human Evolution, 59, 274–288.

Bluck, B.J. (1987). Bed forms and clast size changes in

gravel-bed rivers. In K. Richards (Ed.), River channels:

Environment and process (pp. 159–178). UK: Basil

Blackwell.

Bridge, J.S. (2003). Rivers and floodplains: Forms, processes

and sedimentary record. UK: Blackwell Science Ltd.

Butler, R.W.H. (1982). A structural analysis of the Moine

Thrust Zoine between Loch Eriboll and Foinaven, NW

Scotland. Journal of Structural Geology, 4, 19–29.

Cahen, D., & Moeyersons, J. (1977). Subsurface movements

of stone artefacts and their implications for the prehistory

of Central Africa. Nature, 266, 812–815.

Chazan, M., Porat, N., Sumner, T.A., & Horwitz, L.K. (2013).

The use of OSL dating in unstructured sands: The

archaeology and chronology of the Hutton Sands at

Canteen Kopje (Northern Cape Province, South Africa).

Archaeological and Anthropological Sciences, 5, 351–363.

Clark, J.D. (1959). The prehistory of Southern Africa.

London: Penguin Books.

De Wit, M.C.J. (1996). The distribution and stratigraphy of

inland alluvial diamond deposits in South Africa. Africa

Geoscience Review, Special Edition, 3, 19–33.

De Wit, M.C.J. (2008). Canteen Koppie at Barkly West: South

Africa’s first diamond mine. South African Journal of

Geology, 111, 53–66.

Deitrich, W.E. (1987). Mechanics of flow and sediment

transport in river beds. In K. Richards (Ed.), River

channels: Environment and process (pp. 179–227). UK:

Basil Blackwell.

Erlandson, J.M. (1984). A case study in faunalturbation:

Delineating the effects of the burrowing pocket gopher on

the distribution of archaeological materials. American

Antiquity, 49, 785–790.

Evans, D.J.A. (2000). A gravel outwash/deformation till

continuum, Skalafellsjokull, Iceland. Geografiska Annaler.

Series A, Physical Geography, 82, 499–512.

Evans, M., & Cunningham, A. (2013). Optically Stimulated

Luminescence dating report: Canteen Kopje project.

Unpublished manuscript. Johannesburg: University of the

Witwatersrand Archaeology Department.

Farnden, T.H.G. (1967). Preliminary notes on prehistoric

material from the Umgeni Valley near Cato Ridge,

Natal. The South African Archaeological Bulletin, 21,

187–190.

Fock, G.J. (1968). Rooidam, a sealed site of the First

Intermediate. South African Journal of Science, 64,

153–159.

Forssman, T., Kuman, K., Leader, G.M., & Gibbon, R.J.

(2010). A Later Stone Age assemblage from Canteen

Kopje, Northern Cape. South African Archaeological

Bulletin, 192, 204–214.

Forssman, T., & Pargeter, J. (2014). Assessing surface

movement at Stone Age open-air sites: First impressions

from a pilot experiment in northeastern Botswana.

Southern African Humanities, 26, 157–176.

Gibbon, R.J., Leader, G.M., & Kuman, K. (2008). Canteen

Kopje-SAHRA report for 2007–2008. Unpublished

manuscript. Johannesburg: University of the

Witwatersrand Archaeology Department.

Gibbon, R.J., Granger, D.E., Kuman, K., & Partridge, T.C.

(2009). Early Acheulean technology in the Vaal River

gravels, South Africa, dated with cosmogenic nuclides.

Journal of Human Evolution, 56, 152–160.

Gibbon, R.J., Leader, G.M., & Kuman, K. (2009). Canteen

Kopje-SAHRA report for 2008–2009. Unpublished

manuscript. Johannesburg: University of the

Witwatersrand Archaeology Department.

Gibbon, R.J., Granger, D.E., Kuman, K., Leader, G.M.L.,

Lotter, M.G., & Forssman, T. (2013). Isochron burial dating

of the Earlier Stone Age deposits at Canteen Kopje, South

Africa. Paper presented at the Association of Southern

African Professional Archaeologists Biennial Conference,

Gaborone, Botswana.

Gonzales, D.P.G., Damrosch, D.B., Damrosch, D.R., Pryor, J.,

& Thunen, R.L. (1985). The third dimension in site

structure: An experiment in trampling and vertical

dispersal. American Antiquity, 50, 803–818.

Goodwin, A.J.H. (1927). South African archaeology. Man, 27,

29–31.

Goodwin, A.J.H. (1929). The Stone Ages in South Africa.

Africa: Journal of the International African Institute, 2,

174–182.

Goodwin, A.J.H. (1934). Some developments in technique

during the Earlier Stone Age. Transactions of the Royal

Society of South Africa, 21, 109–123.

Goodwin, A.J.H., & van Riet Lowe, C. (1929). The Stone Age

cultures of South Africa. Annals of the South African

Museum, 27, 1–289.

Helgren, D.M. (1977). Environmental stratigraphy of the

relict alluvia and terraces along the lower Vaal River, South

Africa. Palaeoecology of Africa and the Surrounding

Islands, 10, 163–170.

Helgren, D.M. (1979). River of diamonds: An alluvial history

of the lower Vaal Basin, South Africa. Chicago: University

of Chicago.

Herries, A.I.R. (2011). A chronological perspective on the

Acheulean and its transition to the Middle Stone Age in

Southern Africa: The question of the Fauresmith.

International Journal of Evolutionary Biology, 2011,

1–25.

Hodge, R., Brasington, J., & Richards, K. (2009). Analysing

laser-scanned digital terrain models of gravel bed surfaces:

Linking morphology to sediment transport processes and

hydraulics. Sedimentology, 56, 2024–2043.

Hofman, J.L. (1986). Vertical movement of artifacts in alluvial

and stratified deposits. Current Anthropology, 27, 163–171.

18 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.

Page 19: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

LOTTER ET AL. A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA

Humphreys, A.J.B. (1970). The role of raw material and the

concept of the Fauresmith. The South African

Archaeological Bulletin, 25, 139–144.

Kandel, A.W., Felix-Henningsen, P., & Conard, N.J. (2003).

An overview of the spatial archaeology of the Geelbek

Dunes, Western Cape, South Africa. In G. Fu ̈leky (Ed.),

Papers of the 1st International Conference on Archaeology

and Soils (pp. 37–44). Hungary: BAR International Series.

Karkanas, P., Bar-Yosef, O., Goldberg, P., & Weiner, S. (2000).

Diagenesis in prehistoric caves: The use of minerals that

form in situ to assess the completeness of the archaeological

record. Journal of Archaeological Science, 27, 915–929.

Kiberd, P. (2006). Bundu Farm: A report on archaeological

and palaeoenvironmental assemblages from a pan site in

Bushmanland, Northern Cape, South Africa. The South

African Archaeological Bulletin, 61, 189–201.

Klein, R.G. (2000). The Earlier Stone Age of Southern Africa.

The South African Archaeological Bulletin, 55, 107–122.

Kostic, B., & Aigner, T. (2007). Sedimentary architecture and

3D ground-penetrating radar analysis of gravelly

meandering river deposits (Neckar Valley, SW Germany).

Sedimentology, 54, 789–808.

Kuman, K. (2001). An Acheulean factory site with Prepared

Core Technology near Taung, South Africa. The South

African Archaeological Bulletin, 56, 8–22.

Kuman, K., & Field, A.S. (2009). The Oldowan industry from

Sterkfontein Caves, South Africa. In K. Schick & N. Toth

(Eds.), The cutting edge: New approaches to the

archaeology of human origins (pp. 151–169). Indiana:

Stone Age Institute Press.

Kuman, K., Lotter, M.G., Gibbon, R.G., Leader, G.M., Reddy,

S., Forssman, T., & Granger, D. (in prep.). The Fauresmith

at Canteen Kopje, Northern Cape Province: Expanding the

variability of the industry. Johannesburg: University of the

Witwatersrand.

Krumbein, W.C. (1939). Preferred orientation of pebbles in

sedimentary deposits. The Journal of Geology, 47,

673–706.

Leader, G.M.L. (2014). New excavations at Canteen Kopje,

Northern Cape Province, South Africa: A

techno-typological comparison of three earlier Acheulean

assemblages with new interpretations on the Victoria West

phenomenon. Unpublished doctoral dissertation,

University of the Witwatersrand, Johannesburg.

McBrearty, S. (1990). Consider the humble termite: Termites

as agents of post-depositional disturbance at African

archaeological sites. Journal of Archaeological Science, 17,

111–143.

McBrearty, S., Bishop, L., Plummer, T., Dewar, R., & Conard,

N. (1998). Tools underfoot: Human trampling as an agent

of lithic artifact edge modification. American Antiquity, 63,

108–129.

McNabb, J. (2001). The shape of things to come: A

speculative essay on the role of the Victoria West

phenomenon at Canteen Koppie, during the South African

Earlier Stone Age. In S. Milliken & J. Cook (Eds.), A very

remote period indeed (pp. 37–41). Oxford: Oxbow Books.

McNabb, J., & Beaumont, P. (2011a). A report on the

archaeological assemblages from excavations by Peter

Beaumont at Canteen Koppie, Northern Cape, South

Africa. Oxford: BAR International Series.

McNabb, J., & Beaumont, P. (2011b). Excavations in the

Acheulean levels of the Earlier Stone Age site of Canteen

Koppie, Northern Cape Province, South Africa. Proceedings

of the Prehistory Society, 78, 51–71.

McPherron, S.J.P. (2005). Artefact orientations and site

formation processes from total station proveniences.

Journal of Archaeological Science, 32, 1003–1014.

Millane, R.P., Weir, M.I., & Smart, G.M. (2006). Automated

analysis of imbrication and flow direction in alluvial

sediments using laser-scan data. Journal of Sedimentary

Research, 76, 1049–1055.

Mitchell, P. J. (1998). The South African Stone Age in the

collections of the British Museum: Content, history and

significance. The South African Archaeological Bulletin, 53,

26–36.

Moeyersons, J. (1978). The behaviour of stones and stone

implements, buried in consolidating and creeping Kalahari

Sands. Earth Surface Processes, 3, 115–128.

Morton, A.G.T. (2004). Archaeological site formation:

Understanding lake margin contexts. Oxford: BAR

International Series.

Naden, P.S., & Brayshaw, A.C. (1987). Small- and

medium-scale bedforms in gravel bed rivers. In K. Richards

(Ed.), River channels: Environment and process (pp.

249–271). UK: Basil Blackwell.

Nielsen, A.E. (1991). Trampling the archaeological record: An

experimental study. American Antiquity, 56, 483–503.

Oestmo, S., Schoville, B.J., Wilkins, J., & Marean, C.W.

(2014). A Middle Stone Age paleoscape near the Pinnacle

Point caves, Vleesbaai, South Africa. Quaternary

International, 350, 147–168.

Partridge, T.C., & Brink, A.B.A. (1967). Gravels and terraces

of the lower Vaal Basin. South African Geographical

Journal, 49, 21–38.

Petraglia, M.D., & Potts, R. (1994). Water flow and the

formation of Early Pleistocene artifact sites in Olduvai

Gorge, Tanzania. Journal of Anthropological Archaeology,

13, 228–254.

Petts, G., & Foster, I. (1985). Rivers and landscape. London:

Edward Arnold Publishers.

Porat, N., Chazan, M., Gru ̈n, R., Aubert, M., Eisenmann, V., &

Horwitz, L.K. (2010). New radiometric ages for the

Fauresmith industry from Kathu Pan, southern Africa:

Implications for the Earlier to Middle Stone Age transition.

Journal of Archaeological Science, 37, 269–283.

Sarupen, A. (2010). Analysis of Middle Stone Age lithic

artefacts from Canteen Kopje, Northern Cape, South

Africa: Typology and technology. Unpublished honours

report, University of the Witwatersrand, Johannesburg.

Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc. 19

Page 20: A Geoarchaeological Study of the Middle and Upper ......two main archaeological deposits: alluvial gravels and a mantle of overlying fine sediments known locally as the “Hutton

A GEOARCHAEOLOGICAL STUDY AT CANTEEN KOPJE, NORTHERN CAPE, SOUTH AFRICA LOTTER ET AL.

Schick, K.D. (1984). Processes of Palaeolithic site formation:

An experimental study. Unpublished doctoral dissertation,

University of California, Berkeley.

Schick, K.D. (1987). Modelling the formation of Early Stone

Age artifact concentrations. Journal of Human Evolution,

16, 789–807.

Schick, K.D. (1991). On making behavioural inferences from

early archaeological sites. In J.D. Clark (Ed.), Cultural

beginnings (pp. 79–107). Bonn: Dr. Rudolf Habelt GMBH.

Schick, K.D. (1997). Experimental studies of site-formation

processes. In G.I. Isaac & B. Isaac (Eds.), Koobi Fora

Research Project (Volume 5), Plio-Pleistocene Archaeology

(pp. 244–261). Oxford: Clarendon Press.

Schiffer, M.B. (1983). Toward the identification of

formation processes. American Antiquity, 48, 675–

706.

Shea, J.J. (1999). Artifact abrasion, fluvial processes, and

“living floors” from the Early Palaeolithic site of ‘Ubeidiya

(Jordan Valley, Israel). Geoarchaeology: An International

Journal, 14, 191–207.

Singer, R., & Crawford, J.R. (1958). The significance of the

archaeological discoveries at Hopefield, South Africa. The

Journal of the Royal Anthropological Institute of Great

Britain and Ireland, 88, 11–19.

Stein, J.K. (1983). Earthworm activity: A source of potential

disturbance of archaeological sediments. American

Antiquity, 48, 277–289.

Stein, J.K. (2001). A review of site formation processes and

their relevance to geoarchaeology. In P. Goldberg, V.T.

Holliday, & C. Reid Ferring (Eds.), Earth sciences and

archaeology (pp. 37–51). London: Kluwer Academic/

Plenum Publishers.

Tandon, S.K., & Kumar, R. (1981). Gravel fabric in a

sub-Himalayan braided stream. Sedimentary Geology, 28,

133–152.

Underhill, D. (2011). The study of the Fauresmith:

A review. The South African Archaeological Bulletin, 66,

15–26.

Van Riet Lowe, C. (1937). The archaeology of the Vaal River

Basin. In P.G. Sohnge, D.J.L. Visser, & C. van Riet Lowe

(Eds.), The geology and archaeology of the Vaal River

Basin (pp. 61–164). South Africa: Geological Survey

Memoirs.

Vermeersch, P.M., & Bubel S. (1997). Postdepositional

artefact scattering in a podzol: Processes and consequences

for late Palaeolithic and Mesolithic sites. Anthropologie, 2,

119–130.

Villa, P., & Courtin, J. (1983). The interpretation of stratified

sites: A view from underground. Journal of Archaeological

Science, 10, 267–281.

Yagishita, K., & Jopling, A.V. (1983). Grain fabric of planar

cross-bedding formed by lateral accretion, Caledon

outwash, Ontario, Canada. The Journal of Geology, 91,

599–606.

20 Geoarchaeology: An International Journal 00 (2016) 1–20 Copyright C© 2016 Wiley Periodicals, Inc.


Recommended