Contemporary flowstone development links early hominin bearing cave deposits in South Africa

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Contemporary !owstone development links early hominin bearing cave deposits inSouth Africa

Robyn Pickering a,b,!, Jan D. Kramers c,d, Philip John Hancox d, Darryl J. de Ruiter e, Jon D. Woodhead a

a School of Earth Sciences, University of Melbourne, Victoria, 3010, Australiab Institute for Geological Sciences, University of Bern, Bern, 3012, Switzerlandc Department of Geology, University of Johannesburg, Auckland Park 2006, South Africad School of Geosciences, University of the Witwatersrand, South Africae Department of Anthropology, Texas A&M University, College Station, 77843-4352, USA

a b s t r a c ta r t i c l e i n f o

Article history:Received 18 August 2010Received in revised form 2 March 2011Accepted 11 March 2011Available online 17 April 2011

Editor: P. DeMenocal

Keywords:stratigraphycave sedimentsSwartkransSterkfonteinU–Pb datingspeleothemsAustralopithecus africanusParanthropus robustusAustralopithecus sedibaearly Homo

The Cradle of Humankind cave sites in South Africa preserve fossil evidence of four early hominin taxa:Australopithecus africanus, Australopithecus sediba, Paranthropus robustus and early Homo. In order to integratethis record into a pan-African scenario of human evolutionary history it is critical to have reliable dates andtemporal ranges for the southern African hominins. In the past a lack of precise and accurate chronologicaldata has prevented the evaluation of the temporal relationships between the various sites. Here we reportnew uranium–lead (U–Pb) radiometric ages obtained from sheets of calcium carbonate !owstone inter-bedded between clastic cave sediments at the site of Swartkrans, providing bracketing ages for thefossiliferous deposits. The fossil bearing units of Swartkrans, speci"cally the Hanging Remnant and LowerBank of Member 1, are underlain by !owstone layers dated to 2.25±0.05 Ma and 2.25±0.08 Ma and cappedby layers of 1.8±0.01 Ma and 1.7±0.07 Ma. The age bracket of the Member 1 deposits is therefore between2.31 and 1.64 Ma. However, by combining the U–Pb with biostratigraphic data we suggest that this can benarrowed down to between 1.9 and 1.8 Ma. These data can be compared with other recently dated sites and aradiometrically dated U–Pb age sequence formed: Sterkfontein Member 4, Swartkrans Member 1, Malapa,and Cooper's D. From this new U–Pb dataset, a pattern of contemporary !owstone development emerges,with different caves recording the same !owstone-forming event. Speci"cally overlapping !owstoneformation takes place at Swartkrans and Sterkfontein at ~2.29 Ma and ~1.77 Ma, and at Sterkfontein andMalapa at !2.02 Ma. This suggests a regional control over the nature and timing of speleothem developmentin cave deposits and these !owstone layers could assist in future correlation, both internal to speci"c depositsand regionally between sites.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

The single richest source of early hominin fossils and associatedfauna within southern Africa is a series of dolomitic karstic caves,known collectively as the ‘Cradle of Humankind’ World Heritage Site,located 40 km northwest of Johannesburg (Fig. 1). These cavespreserve the fossil remains of at least four early hominin taxa:Australopithecus africanus, Australopithecus sediba, Paranthropusrobustus and early Homo (Berger et al., 2010; Broom, 1938; Clarkeet al., 1970; Dart, 1925). One of themost signi"cant outstanding issuesin South African paleoanthropology is a precise method for dating thefossil remains of our earliest hominin ancestors, as most ageassignments were previously based on biostratigraphic correlationswith the better dated East African sites. An accurate understanding of

the tempo and mode of hominin evolution is critical for determiningthe place of these taxa in our evolutionary history.

The South African cave sites contain two main sediment types:fossil bearing clastic sediments (often referred to as breccia) andspeleothems. The speleothems consist of chemically precipitatedcalcium carbonates, typically occurring as stalagmites or !owstones,which are horizontal layers of calcite and rare aragonite, ofteninterbedded with the clastic sediment. The U–Pb chronology ofsecondary cave carbonates or speleothems is a highly promising newradiometric dating method (Rasbury and Cole, 2009). Following thepioneering work of Smith and Farquhar (1989) and Richards et al.(1998), there are now a growing number of successful studies usingU–Pb dating of speleothems (Cliff et al., 2010; Cole et al., 2005; deRuiter et al., 2009; Lundberg et al., 2000; Pickering et al., 2010; Polyaket al., 2008; Walker et al., 2006; Woodhead et al., 2006, 2010). The!owstones that interdigitate between the cave sediments of the SouthAfrican sites appear to be closed to U mobility after formation(Pickering et al., 2010) and can be directly dated to providemaximum

Earth and Planetary Science Letters 306 (2011) 23–32

! Corresponding author at: Present address: School of Earth Sciences, University ofMelbourne, Victoria 3010, Australia. Tel.: +61 3 8344 6531.

E-mail address: [email protected] (R. Pickering).

0012-821X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.doi:10.1016/j.epsl.2011.03.019

Contents lists available at ScienceDirect

Earth and Planetary Science Letters

j ourna l homepage: www.e lsev ie r.com/ locate /eps l

Author's personal copy

and minimum ages for the fossil faunas preserved in the clasticsedimentary deposits trapped between them.

Under favourable conditions (high U and low initial Pb contents),the U–Pb method is also potentially much more precise than otherexisting dating techniques, such as U–Pb on fossil teeth (Balter et al.,2008), ESR on teeth (Curnoe et al., 2001) and burial dating of thesediments (Partridge et al., 2003), with errors that typically approach,and are sometimes below the 1% level. Here we present a new U–Pbchronology for the fossiliferous cave "lls of Swartkrans and comparethis to the published chronologies for the nearby sites of Sterkfontein(Pickering and Kramers, 2010; Pickering et al., 2010), Cooper's Cave(de Ruiter et al., 2009), and Malapa (Dirks et al., 2010) (Fig. 1). This isthe "rst seriation of the South African hominin cave deposits basedupon direct radiometric dating, allowing us to correlate between thesites using !owstones as markers and to assign ages to fossil bearinglayers sandwiched between the dated !owstones. A similar approachto seriation of the South African cave sites uses the palaeomagneticsignals preserved in the !owstone and sediment layers to produce arelative chronology, complemented by the faunal records of the sites(Adams et al., 2007, 2010; Herries, 2003; Herries et al., 2006a,b, 2009,2010). The direct dating, seriation and possible correlation of the sitesand their associated hominin and other faunal fossils are fundamentalto unravelling many important biogeographic issues. For instance,questions which revolve around the ancestry of P. robustus andwhether it is derived from A. africanus or Paranthropus aethiopicus(Clarke, 2008; Johanson and White, 1979; Kimbel et al., 1988), orperhaps even the origin of the genus Homo (Berger et al., 2010),require an accurate understanding of the relative ages of the fossilsassigned to these taxa.

2. Regional geology and karst relationships

The Bloubank River valley is host to several hominin bearing cavesites, including the Swartkrans, Sterkfontein, Cooper's, Malapa, andGladysdale Caves (Fig. 1). These caves have formed traps for surfacesediments over a period of several million years and, as a result of theabundance of early hominin and other fossils preserved in thesediments, their stratigraphic relationships have been the subject ofmuch investigation (Brain, 1958, 1993; Clarke, 2007; Cooke, 1938; deRuiter et al., 2009; Dirks et al., 2010; Partridge, 1978; Pickering andKramers, 2010; Wilkinson, 1983). The fossil bearing sediments havebeen classi"ed into different members (Brain, 1993; Partridge, 1978),

units (de Ruiter et al., 2009), or facies (Dirks et al., 2010) depending onthe sites. Hereweprovide a brief overviewof Swartkrans, Sterkfontein,Cooper's Cave andMalapa (see Fig. 2), and propose a generalmodel forthe accumulation of the deposits preserved in these caves.

Swartkrans is situated on the north side of the Bloubank Valley,and occurs on the intersection of two fault traces that trend roughlyEast–West and North–South (Brain, 1993). Most of the original roof ofthe cave has been removed by erosion, exposing the fossil bearingsediments to the surface. The stratigraphy is complex, with evidencefor several cycles of deposition and erosion (Brain, 1993, 1995). Fiveremaining members incorporating six discrete fossiliferous depositsare recognised, these being from oldest to youngest: Member 1(consisting of the Hanging Remnant (HR) and Lower Bank (LB)deposits), Member 2, Member 3, Member 4 and Member 5. Theworld's largest sample of P. robustus, consisting of approximately 400fossils from as many as 150 individuals, is derived from Swartkrans, aswell as some 29 fossils assigned to perhaps 10 individuals of earlyHomo (Brain, 1981; Clarke, 1977; de Ruiter, 2001; Grine et al., 1996),making this the "rst South African site recording the co-occurrence ofmultiple hominin species (Broom and Robinson, 1950).

The Sterkfontein Caves occur on the south side of the BloubankValley and constitute one of the richest hominin-bearing sites in theworld, with fossil remains of A. africanus and an as yet unnamedaustralopithecine in Member 4 and the Silberberg Grotto (Clarke,1977, 2008; Dart, 1925; Wood and Richmond, 2000), alongsidechronologically younger traces of early Homo in Member 5 (Kumanand Clarke, 2000; Pickering and Kramers, 2010). The cave system wasdeveloped by a simple solution excavation of the dolomite host rockalong fracture systems, and today fossil-bearing sediments areexposed at surface, as well as underground. The most enduringinterpretive description of the sediments is the member systemproposed by Partridge (Partridge, 1978, 2000; Partridge and Watt,1991), where the deposits are classi"ed into six members, 1 to 6,oldest to youngest, respectively. This systemwas recently reviewed indetail by Pickering and Kramers (2010) who argue that, while themain fossil bearing sediments fall into Members 2, 4 and 5, theMember 3 terminology should be abandoned.

The site of Cooper's Cave is situated about 1 km east ofSterkfontein on the south !ank of the Bloubank Valley and consistsof a collection of three spatially distinct in"lls (Cooper's A, B, and D),all of which preserve fossil-bearing sediments. The Cooper's Ddeposits have been thoroughly excavated, yielding an abundant

Fig. 1. An overviewmap showing the position of the “Cradle of Humankind” UNESCO world heritage site within southern Africa (left) and the relative positions of the sites discussedin the text (right).

24 R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

faunal assemblage (Berger et al., 2003) including P. robustus fossils (deRuiter et al., 2009; Steininger et al., 2008).

The Malapa site was discovered in 2008, and has yielded near-complete skeletons of at least two individuals assigned to the newspecies A. sediba (Berger et al., 2010). The fossils are encased in water-lain, clastic sediments, which were deposited along the lower parts ofa deeply eroded cave system, immediately above a !owstone layer.The two hominin specimens were buried together in a single debris!ow that was lithi"ed soon after deposition in a phreatic environmentinaccessible to scavengers, ensuring their remarkable preservation(Dirks et al., 2010).

Typically these cave sites comprise a basal layer of speleothem,then several units of clastic sediment, separated either by !owstonelayers or erosional hiatuses. The clastic sediments vary in grain sizefrom massive blocks of dolomite to "ne-grained layered mud, withthe bulk of the sediment being reddish-brown sands. We prefer not touse the term ‘breccia’ as this geological term refers speci"cally toangular fragments embedded in a "ne-grained matrix. While truebreccias do occur in these caves, so do other "ner-grained, better-sorted sediments and the term ‘calci"ed clastic sediment’ is a moreaccurate description of the sediments (for further discussion seePickering et al., 2007). All of the sites presented here have undergonemajor surface erosion, with up to 30 m of material removed (Dirkset al., 2010) to expose the fossil-bearing cave deposits at surface,making reconstructing the original cave morphology and cave "llingprocesses challenging. The following simpli"ed model for cave

sedimentation appears to apply generally: initially, basal speleothemsare laid down prior to the caves opening to the surface. After the cavesare opened to surface, the coarser-grained material (the dolomiteblocks) accumulated directly under or close to shaft-like entrancesinto underground chambers, forming cones of poorly sorted sediment.Finer-grained, more mobile material is washed further into the cave,forming more layered deposits. This hydrodynamic sorting thatseparates the "ner-grained sediments also winnows out the morebuoyant fossil bone, which is often concentrated in the "ner, distalportions of the deposits (de Ruiter et al., 2009; Pickering and Kramers,2010; Pickering et al., 2007). Interspersed between the clasticsediments are !owstone layers that formed before detritus waswashed into the cave, and during later breaks in clastic sedimentation.This pattern of an alternating stack of clastic sediments and !owstonelayers is observed at all the sites described here.

3. Swartkrans !owstones: selection and pre-screening

The !owstone layers both above and below the major fossilbearing sediments of Member 1 at Swartkrans were all sampled for U–Pb dating. Member 1 consists of two spatially separate deposits: theLower Bank along the cave's north wall, and the Hanging Remnantexposed along the north-western wall (Fig. 2) The underlying!owstone is clearly visible in both exposures and consists of anundulating layer up to 60 cm thick. Samples SWK7 and SWK12 wereobtained from this layer beneath the Hanging Remnant and Lower

Fig. 2. Schematic summary stratigraphic columns of the sedimentary units or members containing the early hominin fossils and inter-bedded !owstones from Sterkfontein,Swartkrans, Cooper's and Malapa with U–Pb dated speleothem marked with black stars and all scale bars 1 m (Sterkfontein column based on log of Borehole 1 modi"ed fromPickering and Kramers, 2010; Swartkrans sketch redrawn from Brain, 1993; Coopersmodi"ed from de Ruiter et al., 2009; Malapamodi"ed from Dirks et al., 2010. Note Facies A–D forMalapa are different from the Sterkfontein and Cooper's Facies A and B).

25R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

Bank respectively. Both !owstone samples consist of compact, well-preserved calcite with clear, cm-scale layering and are white tocreamy brown in colour (Fig. 3). There are no immediate similaritiesbetween these samples, which is to be expected as !owstones in thesecaves are often restricted in lateral extent surrounding suitable drippoints and can vary hugely in appearance on a metre scale (Pickering,2005; Pickering et al., 2007). The Hanging Remnant is an inaccessiblewall of sediment with several patches of !owstone throughout thesediment pile (Fig. 5). The uppermost layer of !owstone is the largestand caps off the deposits; SWK5 is derived from this layer. Again thesample taken from this !owstone consists of well-preserved, densecalcite, white to creamy-brown in colour, with clear layering on a cm-scale (Fig. 3). The Lower Bank deposits are capped by a well-developed, laterally continuous, layered !owstone, up to 40 cm thick,sampled by SWK9. Sample SWK9 is different from the other!owstones in that it consists of ~1 cm layers of !owstone intercalatedbetween layers of reddish-brown cave earth sediment. The !owstonelayers are white in colour and aremore porous than the other samples(Fig. 3).

A signi"cant challenge of U–Pb dating speleothems is thedistribution of U within samples. Less than around 1 !g/g of U willnot produce suf"cient radiogenic 206Pb to allow isochron construc-tion. Work by others (Cole et al., 2003, 2005; Walker, 2005) andourselves (Pickering et al., 2010) has shown that U is not evenlydistributed throughout speleothem samples and a method to detect‘U-rich’ layers is indispensable. Our "rst samples were pre-screenedusing a FUJI"lm BAS-1800 beta-scanner following the approachpioneered by Cole et al. (2002) and used by Walker (2005) toproduce greyscale image ‘maps’ of the U content (Fig. 3). This methodhas its merits but is time consuming (four–six week exposure times)and so later samples were pre-screened using laser ablationinductively coupled mass spectrometry (ICPMS) traverses.

All the samples present clearly de"ned layers with enhanced U-content, up to 1 cm thick, located near the base of the !owstonehorizons (Fig. 3). This is a typical mode of occurrence for such U-richlayers, and is thought to result from the renewed onset of higherprecipitation after a prolonged dry period, as discussed by Pickeringet al. (2010). Sample SWK9 consists of layers of !owstone inter-

Fig. 3. Flowstone samples from Swartkrans with corresponding beta-scanner images used to detect U-rich layers for dating (images and scans shown to the same scale, scale bar is3 cm; white arrows indicate U-rich layers, black arrows the sample way up).

26 R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

bedded with "ne red cave earth; these sediment grains show up asdarker spots on the beta-scanner image. This latter material is notsuitable for dating, as the U is not hosted in carbonate, and thedetrital particles are rich in common Pb that can be isotopicallyheterogeneous, making isochron dating impossible.

4. U–Pb dating method

Full analytical details of the method used can be found in Pickeringet al. (2010) and Woodhead et al. (2006) and are therefore onlybrie!y recapped here for the sake of clarity. Using the beta-scannerimages and following the approach of Cole et al. (2002), U-rich layersare selected and up to 10 ~0.05 g of blocks are cut from each sample,ideally from within a single growth layer (although this is not alwayspossible) using a dental drill. Samples are spiked with a 236U–202Pb or233U–205Pb tracer and U and Pb separated by conventional ionexchange techniques. A double-focusing Nu Instruments® MC–ICP-MS was used for all isotopic measurements. We used the U standardNIST SRM U050 to tune machine settings and calibrate the gain of theelectron multipliers between runs (every 5th sample). For Pb isotopeanalysis, we have used Tl doping of sample and standard solutions ofNIST SRM 981 as a means to monitor and correct for instrumentalisotope fractionation. While U blanks are negligible, Pb blanks for theSwartkrans samples ranged from 4 to 40 pg, depending uponanalytical session. The blank isotopic compositionwas notmeasurablydifferent from that of the common Pb in the samples, and therefore noblank correction was made.

In the case of an initial (234U/238U) activity ratio of N1 a simpleisochron construction leads to anoverestimate of the sample age becauseit does not take into account 206Pb produced from the decay of excess234U. Largematched (over 0.15 g) unspiked sampleswere therefore usedto constrain the present day 234U/238U disequilibrium and from thesedata a correction can be made for the effects of initial isotopicdisequilibrium in the U decay chain. Ages are calculated either with206Pb/238U ratios generated by Isoplot (Ludwig, 2000) following themethod documented in Pickering et al. (2010) or by using the Tera-Wasserberg isochron construction, employing 238U/206Pb vs 207Pb/206Pbratios (Ludwig, 2000) and following themethodsdescribed inWoodheadet al. (2006).

In general the Swartkrans U–Pb ages are precise with averageerrors of 2.5%, equating to about 50 000 yrs. At best errors can be assmall as 0.3%, equating to an order of magnitude better precision of5000 yrs. Uncertainties up to six times larger (at 12%) than averageare also encountered. These larger errors are particularly associatedwith samples N2 Ma due to (1) the greater relative error in themeasurement of the residual 234U excess over secular equilibrium and(2) the lower levels of precision in the measurements of this sampledue to the very low abundance of radiogenic lead. The application ofU–Pb dating to ‘young’ carbonates of only a few million years old isstill a relatively new technique and development of the procedure tobetter measure the residual 234U excess, and thus more tightlyconstrain ages, is underway.

5. New U–Pb chronology for Swartkrans

Results from the U–Pb dating are given in Tables 1 and 2 andshown in Figs. 4 and 5. Uranium concentrations are ~1 !g/g (parts permillion), while Pb concentrations are on average "fty times lower inthe ng/g (parts per billion) range (Table 1). 238U/204Pb ratios reach upto over 95 000 (Table 1) and display a surprising amount of variationgiven that the sub-samples were taken on a cm-scale. 206Pb/204Pbratios show a corresponding variation, ranging from 18 to over 65.These are in fact some of the most radiogenic samples encountered inspeleothem material from this region. The resulting isochron ages(Fig. 4) are thus correspondingly precise with errors of as little as 0.3%

making this the most precise U–Pb data set for any of the cave sites inthe Cradle of Humankind.

The !owstones underlying the Hanging Remnant and Lower Bankhave U–Pb ages of 2.248±0.052 Ma and 2.249±0.077 Ma respectively.While the two !owstones are not immediately similar in appearanceand occur in separate parts of the cave, their ages are almostindistinguishable from each other, indicating that they formed duringthe same speleothem formation event. Studies at other caves in theCradlehave shown that these speleothemformationevents are linked toshifts in climate, namely wetter episodes (Pickering et al., 2007). Theages also provide a maximum age for the sediments preserved abovethese layers and it is therefore unlikely that there is material older than2.25 Ma preserved at Swartkrans. The !owstone capping the HangingRemnant is dated to 1.800±0.005 Ma, while the Lower Bank depositsare capped by a !owstone dated to 1.706±0.0.069 Ma respectively.These dates are notwithin error of each other. Instead we interpret thisas suggesting a period of speleothem formation between 1.8 and 1.7 Maduringwhichboth theHangingRemnant and Lower Bankdepositswerecapped by a layer of !owstone.

Therefore the Member 1 deposits at Swartkrans are sandwichedbetween two !owstones with dates between 2.25±0.052 and 2.25±0.077 Ma on the one hand and 1.80±0.01 and 1.71±0.0.07 Ma onthe other. Moving from dated !owstones to an age range for thefossiliferous sediments requires some consideration. The U–Pb agesfor these !owstones have modest errors but these must, none-the-less, be taken into account. The most conservative approach (such asin Grün et al., 2011) is to use the oldest possible value for the older age(0.08 on 2.25 Ma) for the lower limit of 2.33 Ma and the youngestpossible value for the youngest age (0.07 on 1.71 Ma) to give theupper limit of 1.64. This gives a generous time range of 69000 yrs forthe accumulation of the Member 1 sediments. It is likely that thesediments accumulated during a much narrower interval within thistime range but at this stage the U–Pb data cannot be used tosubstantiate this.

Until recently, the South African hominin caves could not bedirectly dated, in contrast to the East African fossil sites, where fossil-bearing horizons are bracketed between inter-bedded volcanic tuffsdated accurately by the widely used K–Ar and Ar–Ar techniques(Sarnawojcicki et al., 1985). In the South African cave deposits tufflayers are absent, and this approach is therefore not possible. Insteadages are typically assigned based on biostratigraphic comparisonsbetween the South African and the dated East African faunalassemblages (Berger et al., 2002, 2003; Brain, 1993; Cooke, 1938,1974; de Ruiter, 2003; Delson, 1988; White and Harris, 1977).However, biostratigraphic dating has its limitations, including anassumption of continent-wide, contemporaneous evolutionary eventssuch as "rst and last appearances of taxa (Hill, 1995;White, 1995) andissues involving refugia and relict populations (Reynolds, 2007; Vrba,1988). Magnetostratigraphic approaches have been applied at anumber of sites (Herries, 2003; Herries et al., 2006b, 2009; McFaddenet al., 1979; Partridge, 2000; Partridge et al., 1999) but this is a relativetechnique and has relied on the faunal ages to secure the measuredpattern of reversed and normal signals with the global magnetos-tratigraphy. In these cases more than one "t is often possible (Bergeret al., 2002). Combining the palaeomagnetic data with faunal data andany other age data (where possible U–Pb or ESR) leads to the bestresults and brings us one step closer to chronometric correlationbetween the sites (Adams et al., 2007, 2010; Herries, 2003; Herrieset al., 2006a,b, 2009, 2010). To date, no palaeomagnetic data isavailable for Swartkrans.

Attempts have been made to date fossil teeth with U–Pb (Balteret al., 2008) and ESR (Curnoe et al., 2001), but both these methods aredependent on models of U uptake as a function of time and are proneto high (up to 60%) errors associated with open system behaviour.Another approach is dating the burial of the sediments usingcosmogenic 10Be and 26Al (Partridge et al., 2003) but this method

27R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

has similarly high errors and is yet to produce ages in agreement withothermethods (Pickering et al., 2010;Walker et al., 2006) in the SouthAfrican context.

Currently the only direct radiometric method to produce preciseage estimates for the South African caves is U–Pb dating of the!owstone layers associated with the fossil bearing sediments.Previous biostratigraphic estimates for the Hanging Remnant andLower Bank of Swartkrans typically ranged from 1.9 to 1.5 Ma (Bergeret al., 2002; Brain, 1993; Delson, 1984, 1988; Vrba, 1985). Such an agepartially overlaps with the new U–Pb dates reported here, thus thetwo techniques are not inconsistent. While it is possible that thehominins from Member 1 of Swartkrans are as old as 2.25 Ma, theclose correspondence between the oldest biostratigraphic data andthe youngest U–Pb data suggest that the two can be combined toproduce a more tightly constrained estimate than either alone couldprovide. We therefore conclude that the Hanging Remnant and LowerBank of Swartkrans are best considered to date to between 1.9 and1.8 Ma. The age range of P. robustus can be extended to 1.4 Ma atCooper's Cave (de Ruiter et al., 2009), thus this taxon presently spansa time period from ca. 1.9 to 1.4 Ma, though we still agree with White(1995) that both these "rst and last appearance datums are likely tobe artefacts of an incomplete fossil record.

In addition to P. robustus, the Hanging Remnant and Lower Bank ofSwartkrans also contain remains of early Homo. Although there has

been considerable debate regarding the taxonomic status of theseearly Homo remains (Clarke, 1977; Curnoe, 2006; Grine et al., 1996;Howell, 1978), we concur with Clarke (1977) that these specimens, inparticular the more complete remains such as SK 847, are bestconsidered to belong to Homo erectus. With the new age estimate of1.9–1.8 Ma at Swartkrans, these H. erectus remains are contempora-neous with the earliest appearance of this taxon in East Africa (Wood,1991). This suggests that soon after it appeared H. erectus spreadwidely, extending along (at least) the east coast of Africa from northto south relatively rapidly, and eventually leaving Africa as early as ca.1.8 Ma (Garcia et al., 2010). These data also call into question theassumed origin of H. erectus in east Africa, as the fossils from SouthAfrica now rival the earliest possible appearance of this taxon at KoobiFora (Wood, 1991). At present, despite a relatively good fossil record,or perhaps because of it, the place of origin of H. erectus remainsobscure, though the timing of its origin is reinforced.

6. Age sequence and contemporary !owstone development

The U–Pb dated !owstone layers at Swartkrans can now bedirectly compared to those from nearby Sterkfontein (Pickering andKramers, 2010; Pickering et al., 2010), Cooper's (de Ruiter et al.,2009), and Malapa (Dirks et al., 2010) (Fig. 6; Table 3). The results ofthis comparison are three-fold: "rstly, the sites can now be positioned

Table 1All data needed to calculate U–Pb ages for Swartkrans !owstones.

Sample name Concentrations Inter-elementratios238U/204Pb

2SE 206Pb/204Pb 2SE 238U/206Pb 2SE 207Pb/206Pb 2SE Corr. coef.8/6-7/6

U(ppb)

Pb(ppb)

SWK-5.2 2099 3 86214.6 928.3 68.790 0.647 1253.31 15.70 0.261 0.002 "0.984SWK-5.4 777 3 22268.8 743.2 32.989 0.720 675.04 31.15 0.530 0.014 "0.984SWK-5.5 1562 6 21207.0 891.5 30.505 0.240 695.19 31.16 0.530 0.007 "0.984SWK-5.17 704 56 792.5 6.5 18.264 0.044 43.39 0.35 0.845 0.001 "0.984SWK-5.18 958 42 1455.8 9.3 18.657 0.040 78.03 0.49 0.829 0.001 "0.984SWK-5.19 754 28 1719.5 10.6 18.914 0.045 90.91 0.52 0.824 0.001 "0.984SWK-5.21 685 35 1227.3 7.9 18.707 0.041 65.61 0.42 0.833 0.001 "0.984SWK-7.2 2046 2 95897.9 2841.9 59.312 1.378 1616.84 74.93 0.303 0.009 "0.999SWK-7.3 2355 4 43513.4 935.0 36.630 0.486 1187.90 25.90 0.458 0.004 "0.999SWK-7.4 1634 10 10819.4 90.9 22.876 0.112 472.97 3.91 0.699 0.002 "0.999SWK-7.5 2180 3 81284.9 1723.4 53.254 0.886 1526.37 35.54 0.333 0.005 "0.999SWK-7.6 1808 19 6316.8 47.1 20.658 0.044 305.78 2.27 0.760 0.001 "0.999SWK-7.18 2011 20 6499.9 52.5 20.760 0.066 313.09 2.39 0.755 0.001 "0.999SWK-7.19 638 25 1638.0 8.8 18.717 0.036 87.52 0.44 0.831 0.001 "0.999SWK-9.5 850 26 2160.0 11.4 19.528 0.032 110.61 0.63 0.805 0.002 "0.985SWK-9.6 761 39 1265.9 6.8 19.251 0.038 65.76 0.34 0.814 0.001 "0.985SWK-9.7 883 28 2016.0 12.7 19.482 0.064 103.48 0.59 0.806 0.001 "0.985SWK-9.8 856 73 752.9 3.5 19.041 0.027 39.54 0.18 0.822 0.000 "0.985SWK-9.9 771 12 4255.7 33.5 20.053 0.092 212.22 1.54 0.786 0.002 "0.985SWK-9.10 815 19 2827.3 15.2 19.649 0.043 143.89 0.72 0.794 0.001 "0.985SWK-9.11 903 22 2678.7 17.0 19.687 0.077 136.07 0.84 0.798 0.002 "0.985SWK-9.12 1058 13 5384.4 56.1 20.883 0.119 257.84 3.43 0.764 0.005 "0.985SWK-12.2 1708 7 16768.3 189.6 26.901 0.158 623.32 10.14 0.605 0.006 "0.999SWK-12.3 661 2 27584.6 665.4 33.109 0.641 833.15 21.54 0.482 0.007 "0.999SWK-12.4 206 2 7908.3 247.7 23.053 0.429 343.05 14.10 0.716 0.016 "0.999SWK-12.5 1084 2 41530.7 786.1 41.050 0.525 1011.72 22.95 0.409 0.005 "0.999SWK-12.6 649 1 35873.7 876.9 38.152 0.744 940.29 25.12 0.437 0.007 "0.999SWK-12.20 153 22 438.2 2.3 18.284 0.038 23.96 0.13 0.851 0.001 "0.999SWK-12.21 359 12 1854.7 10.9 18.969 0.046 97.78 0.53 0.824 0.001 "0.999

Table 2Terra-Wasserburg plot and isochron U–Pb ages.

Samplename

Present234U/

238U

± U–Pb (T–W) %error

Initial234U/

238U

± U–Pb (isochron) %error

Initial234U/

238U

±

Age 2SE Age 2SE

SWK-5 1.035207 0.000547 1.800 0.005 0.3 6.756 0.045 1.816 0.041 2.3 6.968 0.678SWK-7 1.002661 0.000612 2.248 0.052 2.3 2.555 0.142 2.285 0.105 4.6 2.670 0.313SWK-9 1.009835 0.000663 1.706 0.069 4.0 2.237 0.220 1.765 0.175 9.9 2.463 0.693SWK-12 1.006527 0.001725 2.249 0.077 3.4 4.824 0.227 2.265 0.160 7.1 4.907 0.500

28 R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

temporally; secondly, the fossil bearing horizons can be assigned U–Pb dated age ranges; and thirdly the age ranges covered by theindividual !owstones can be compared between sites and any periodsof overlap identi"ed.

The oldest !owstone development is recorded at Sterkfontein at2.80±0.28 and 2.65±0.30 Ma from the !owstones bracketingMember 2 (as de"ned by Pickering and Kramers, 2010).The sediments of Sterkfontein Member 4 are sandwiched between2.65±0.30 and 2.01±0.06 Ma !owstone horizons, covering aconservative time range of 2.95–1.95 Ma. This episode is followedand slightly overlapped by SwartkransMember 1 at between 2.31 and1.64 Ma. The !owstone underlying the A. sediba fossils at Malapa(Table 1, Fig. 2) is dated at 2.026±0.021 and the age of fossil bearingsediments of Facies D is constrained to 1.95–1.78 Ma by thepalaeomagnetic signal (Dirks et al., 2010). A late, small !owstoneunit "lling a cavity within a coarse roof-collapse breccia near the topof Sterkfontein Member 4 yielded a U–Pb date of 1.81±0.06 Ma(Pickering and Kramers, 2010) suggesting that theMember 5 depositsare younger than this. Finally, the stalagmite underlying thefossiliferous sediments at Cooper's cave is dated to 1.556±0.088 Maand a small !owstone from the middle of the sequences to ~1.4 Ma(de Ruiter et al., 2009) providing the youngest age range of 1.62–1.4 Ma.

The current data set of U–Pb dated !owstones for the caves in thisregionconsists of the12!owstonesmentionedhere andanother 4U–Pbages from theSilberbergGrottoat Sterkfontein (Walker et al., 2006). The

!owstone underlying the Stw573 fossil is dated to 2.35±0.10 Ma byPickering et al. (2010) (sample SB1) and to 2.25±0.09 byWalker et al.(2006) (sample STA15). These two ages are within error of each otherand, using the same approach asGrün et al. (2011) and aswe did for thepairs of !owstones bracketing the sediments, covering a time range of2.45–2.16 Ma. Similarly, the !owstones from the base of the HangingRemnant and the Lower Bank at Swartkrans cover a time range of 2.33–2.19 Ma. There is a period of 140 000 years of overlap between thesetwo time ranges and the median age for this period of !owstoneformation is 2.29 Ma. The !owstone horizon capping Member 4 atSterkfontein, sample OE-14, is dated to 2.01±0.06 and covers the agerange of 2.07–1.95 Ma. Sample M1 from Malapa is dated at 2.026±0.021 Ma (Dirks et al., 2010), covering a narrower range of 2.05–2.01 Ma, which falls within the time range of OE-14 from Sterkfontein.The median age for this period is 2.02 Ma. Thirdly, sample OE-13 fromSterkfontein is dated 1.81±0.06 Ma (Pickering and Kramers, 2010)with an age range of 1.87–1.75 Ma, which overlaps with the range of1.64–1.81 Ma of upper !owstones from Swartkrans (samples SWK-5and SWK-9). The median age of this time of !owstone formation is1.77 Ma. In summary, three periods of contemporary !owstoneformation are identi"ed between three of the sites discussed here:

(1) ~2.29 Ma, Swartkrans (!owstones underlying Member 1) andSterkfontein (Silberberg Grotto);

(2) ~2.02 Ma, Sterkfontein (!owstone capping Member 4) andMalapa (!owstone underlying Facies D);

Fig. 4. Tera-Wasserburg age plots of the four U–Pb dated Swartkrans !owstones (data-point error ellipses are 2 sigma; arrows point to enlargements of very small data points).

29R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

(3) ~1.77 Ma, Swartkrans (!owstones capping the Member 1) andSterkfontein (isolated !owstone "lling a cavity in Member 4).

The presence of three contemporary periods of !owstoneformation from three sites up to 16 km apart, such as Malapa and

Sterkfontein, suggests that the factors controlling speleothem devel-opment are of a scale somewhat larger than local cave conditions.Regionally concurrent !owstone deposits must have been formed bythe same allocyclic control, and that given that there has been noobvious tectonic uplift, climate must be the driving factor to cross thethreshold required to change from clastic to chemical sedimentationin the caves. Unfortunately even the smaller errors of 20 000 yrs onthe U–Pb ages are too large for the dated speleothems to bemeaningfully compared to variations in orbital parameters, and todate there are no southern hemisphere palaeo-climate recordsspanning this time interval. However, at Gladysdale Cave, situatedbetween Sterkfontein andMalapa, a series of younger !owstones withhigh precision (under 1% error) U–Th ages can be compared to variousrecords of climate change proxies (Pickering et al., 2007). In this casean increase in effective precipitation, either during warm interglacialsor during cool, less evaporative glacials, appears to be the majorcontrol over !owstone formation. We hypothesise that the same istrue for the older deposits, and that the synchronous deposition of!owstone represents a reaction to increased effective precipitationand that different caves several kilometres apart record the same wetphase evidenced by the presence of synchronous !owstone layers.Although clastic (and fossil-bearing) sedimentary units sandwichedbetween coeval !owstones at different cave sites are obviously age-bracketed within the same limits, it does not simply follow that theyare themselves synchronous, given the highly episodic nature of suchdeposits. The hill slopes above the caves probably do degrade due tosimilar environmental conditions producing the potential for concur-rent deposition of clastic sediments, but this is limited by the fact thatcaves need to be open to receive clastic sedimentary input. Local caveconditions, such as trees and roof stability, probably control theopening and closing of caves.

7. Conclusions

Uranium–lead systematics have provided absolute dates for!owstone layers inter-bedded between the cave sediments from

Fig. 6. Flowstones from Sterkfontein, Swartkrans, Malapa and Cooper's with U–Pb dated!owstones (diamonds) with errors (whisker bars), plotted vertically against timeshowing the full age ranges covered by the !owstones and periods of overlap betweendifferent caves with median ages of 2.29 Ma, 2.02 Ma and 1.77 Ma depicted by greyhatched bars. Stratigraphic units sandwiched between !owstones are represented bysolid dark grey bars, the shading represents uncertainty on the age range as thesediments most likely accumulated in a much briefer time interval within this range(published dates taken from de Ruiter et al., 2009; Dirks et al., 2010; Pickering andKramers, 2010; Pickering et al., 2010; Walker et al., 2006).

Table 3U–Pb dates from Swartkrans compared to all the other hominin bearing cave sites withU–Pb dated !owstones in the Cradle of Humankind (dates from: (de Ruiter et al., 2009;Dirks et al., 2010; Pickering and Kramers, 2010; Pickering et al., 2010; Walker et al.,2006). Time ranges for fossil-bearing units sandwiched between dated !owstones werecalculated using the oldest maximum error on the older age and the youngest error onthe younger age.

Site andUnit

Samplename

U–Pbage

±2SE

Position U–Pb agerange (Ma) ofstratigrahicunits

From To

Coopers CDD-3 ~1.4 Middle of Unit 1 1.62 1.40Unit 1 CDD-1 1.526 0.088 Under Unit 1

MalapaFacies D

M-1 2.026 0.021 Under Facies D

Swartkrans SWK-9 1.706 0.069 Above LB 2.31 1.64Member 1 SWK-5 1.800 0.005 Above HR

SWK-7 2.248 0.052 Under HRSWK-12 2.249 0.077 Under LB

Sterkfontein OE-13 1.810 0.060 Below Member 5Member 4 OE-14 2.014 0.055 Above Member 4 2.95 1.95

SB-1 2.347 0.101 Silberberg Grotto,below Stw573

STA15 2.25 0.090 Silberberg Grotto,below Stw573

BH4-9 2.650 0.300 Under Member 4Member 2 BH1-8 2.830 0.344 Above Member 2

BH1-15 2.800 0.280 Under Member 2

Fig. 5. Field photograph of the HR showing the position and U–Pb ages of samples SWK5and 7 (scale bar 1 m).

30 R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

four major hominin sites: Swartkrans, Sterkfontein, Cooper's andMalapa. It is now possible to place these sites into a new, directly U–Pbdated age sequence: Sterkfontein Member 4 (2.95 to 1.95 Ma),Malapa facies D (1.95 to 1.78 Ma), Swartkrans Member 1 (2.31 to1.64 Ma), and "nally Cooper's (1.62 to 1.4 Ma). It must be noted thatdating the !owstones only provides maximum and minimum agelimits on the fossil-bearing sediments sandwiched between them;ideally as much chronological information as possible should begathered and compiled to assign ages to the fossils, as done at Malapa(Dirks et al., 2010). Further, sedimentation was probably sporadic,and the episodes of sediment formation and bone accumulation weremost likely short punctuated events within the given age limits. Muchtime is therefore left unrecorded by the cave deposits. This said, thecaves do appear to record contemporary events in the form of the!owstones with overlapping U–Pb ages, seen twice at Swartkrans andSterkfontein (at ~2.29 and ~1.77 Ma) and once at Sterkfontein andMalapa (at ~2.02 Ma). By analogy with a younger cave site in theregion, Gladysvale, (Pickering et al., 2007) that contains U–Th dated!owstones, it appears that climatic conditions (predominantlyvariation in precipitation) control the presence or absence of!owstones and by extension the fossil bearing sediments, and thusthe cyclycity deposits. The contemporary !owstones, from the oldersites, suggest that here too there is some large-scale control over theformation of these layers, possibly climatic. The !owstone layers canthus be used as marker horizons to link deposits within caves andbetween different cave sites, using U–Pb dating to substantiatecorrelations. The effort at absolute dating of fossil sites in the Cradle ofHumankind heritage site is far from complete, but even at this stagethe results should enable renewed attempts to integrate the SouthAfrican hominin fossil record into pan-African scenarios about humanevolutionary history.

Acknowledgements

The late Tim Partridge is acknowledged for his support of thisresearch. Further grateful acknowledgements to Ron Clarke, StevenMotsumi, Able Molepolle and Lucas Sekowe (Sterkfontein); TravisPickering, Morris Sutton, Jason Heaton, Bob Brain, Kathy Kuman,Stuart Ford, Andrew Phaswana (Swartkrans); Christine Steininger,The Fossil Trackers Team (Cooper's Cave); Lee Berger, Paul Dirks, PeterSchmidt (Malapa); Ingeburga Hebeisen (University of Bern); BencePaul, Alan Grieg, RolandMaas, John Hellstrom, Janet Hergt (Universityof Melbourne). The South African Heritage Resources Agency grantedexcavation permits. BernardWood, Julia Lee-Thorp, Andy Herries, andSarah Feakins are thanked for their encouragement and usefuldiscussions; Simon Pickering for editing. Three anonymous reviewersand Eric Delson are thanked for their insightful, constructive andhelpful reviews, Peter deMenocal for editorial handling. Funding wasreceived from the Swiss National Research Foundation (grants 20-113658 to JDK and PBBEP2-126195 to RP) and the University ofMelbourne (McKenzie Fellowship to RP).

References

Adams, J.W., Herries, A.I.R., Kuykendall, K.L., Conroy, G.C., 2007. Taphonomy of a SouthAfrican cave: geological and hydrological in!uences on the GD 1 fossil assemblageat Gondolin, a Plio–Pleistocene paleocave system in the Northwest Province, SouthAfrica. Quaternary Science Reviews 26, 2526–2543.

Adams, J.W., Herries, A.I.R., Hemingway, J., Kegley, A.D.T., Kgasi, L., Hopley, P., Reade, H.,Potze, S., Thackeray, J.F., 2010. Initial fossil discoveries from Hoogland, a newPliocene primate-bearing karstic system in Gauteng Province, South Africa. Journalof Human Evolution 59, 685–691.

Balter, V., Blichert-Toft, J., Braga, J., Telouk, P., Thackeray, F., Albarede, F., 2008. U–Pbdating of fossil enamel from the Swartkrans Pleistocene hominid site, South Africa.Earth and Planetary Science Letters 267, 236–246.

Berger, L.R., Lacruz, R., de Ruiter, D.J., 2002. Brief communication: revised age estimatesof Australopithecus-bearing deposits at Sterkfontein, South Africa. American Journalof Physical Anthropology 119, 192–197.

Berger, L.R., de Ruiter, D.J., Steininger, C.M., Hancox, J., 2003. Preliminary results ofexcavations at the newly investigated Coopers D deposit, Gauteng, South Africa.South African Journal of Science 99, 276–278.

Berger, L.R., de Ruiter, D.J., Churchill, S.E., Schmid, P., Carlson, K.J., Dirks, P., Kibii, J.M.,2010. Australopithecus sediba: a new species of homo-like Australopith from SouthAfrica. Science 328, 195–204.

Brain, C.K., 1958. The Transvaal Ape-man-bearing cave deposits. Transvaal MuseumMemoir 11.

Brain, C.K., 1981. The hunters or the hunted? An Introduction to African CaveTaphonomy. The University of Chicago Press, Chicago, London.

Brain, C.K., 1993. Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum,Pretoria.

Brain, C.K., 1995. The In!uence of Climatic Changes on the Completeness of the EarlyHominid Record in Southern African Caves, with Particular Reference toSwartkrans.

Broom, R., 1938. The Pleistocene anthropoid apes of South Africa. Nature 142, 377–379.Broom, R., Robinson, J.T., 1950. Man contemporaneous with the Swartkrans ape-man.

American Journal of Physical Anthropology-New Series 8, 151–156.Clarke, R.J., 1977. The Cranium of the Swartkrans Hominid SK 847 and its Relevance to

Human Origins. University of the Witwatersrand.Clarke, R.J., 2007. A deeper understanding of the stratigraphy of Sterkontein fossil

hominid site. Transactions of the Royal Society of South Africa 61, 111–120.Clarke, R.J., 2008. Latest information on Sterkfontein's Australopithecus skeleton and a

new look at Australopithecus. South African Journal of Science 104, 443–449.Clarke, R.J., Howell, F.C., Brain, C.K., 1970. More Evidence of an Advanced Hominid at

Swartkrans Nature 225, 1219–1222.Cliff, R.A., Spoetl, C., Mangini, A., 2010. U–Pb dating of speleothems from Spannagel

Cave, Austrian Alps: a high resolution comparison with U-series ages. QuaternaryGeochronology 5, 452–458.

Cole, J.M., Nienstedt, J., Spataro, G., Rasbury, E.T., Lanzirotti, A., Celestian, A.J., Nilsson,M., Hanson, G.N., 2002. Phosphor imaging as a tool for in situ mapping of ppmlevels of uranium and thorium in rocks and minerals. Chemical Geology 193,127–136.

Cole, J.M., Nienstedt, J., Spataro, G., Rasbury, E.T., Lanzirotti, A., Celestian, A.J., Nilsson,M., Hanson, G.N., 2003. Phosphor imaging as a tool for in situ mapping of ppmlevels of uranium and thorium in rocks and minerals. Chemical Geology 193,127–136.

Cole, J.M., Rasbury, E.T., Hanson, G.N., Montanez, I.P., Pedone, V.A., 2005. Using U–Pbages of Miocene tufa for correlation in a terrestrial succession, Barstow Formation,California. Geological Society of America Bulletin 117, 276–287.

Cooke, H.B.S., 1938. The Sterkfontein bone breccia: a geological note. South AfricanJournal of Science 35, 204–208.

Cooke, H.B.S., 1974. Plio–Pleistocene deposits and mammalian faunas of eastern andsouthern Africa. Proc 5th Congrès Néogène Méditerranéen Memoir 78, 99–108.

Curnoe, D., 2006. Odontometric systematic assessment of the Swartkrans SK 15mandible. Homo 57, 263–294.

Curnoe, D., Gr¸n, R., Taylor, L., Thackeray, F., 2001. Direct ESR dating of a Pliocenehominin from Swartkrans. Journal of Human Evolution 40, 379–391.

Dart, R., 1925. Australopithecus africanus: the man-ape of South Africa. Nature 115,195–199.

de Ruiter, D.J., 2001. A Methodological Analysis of the Relative Abundance of Hominidsand other Macromammals from the site of Swartkrans, South Africa. University ofthe Witwatersrand, Johannesburg.

de Ruiter, D.J., 2003. Revised faunal lists for Members 103 of Swartkrans, South Africa.Annals of the Transvaal Museum 40, 29–41.

de Ruiter, D.J., Pickering, R., Steininger, C.M., Kramers, J.D., Hancox, P.J., Churchill, S.E.,Berger, L.R., Backwell, L., 2009. New Australopithecus robustus fossils and associatedU–Pb dates from Cooper's Cave (Gauteng, South Africa). Journal of HumanEvolution 56, 497–513.

Delson, E., 1984. Cercopithecoid biochronology of the African Plio–Pleistocene:correlation among eastern and southern hominid-bearing localities. CourierForschungs-Institut Senckenberg 69, 199–281.

Delson, E., 1988. Chronology of South African Australopiths site units. In: Grine, F.E.(Ed.), Evolutionary History of the Robust Australopithecines. Aldine de Gruyter,New York, pp. 317–325.

Dirks, P., Kibii, J.M., Kuhn, B.F., Steininger, C., Churchill, S.E., Kramers, J.D., Pickering, R.,Farber, D.L., Meriaux, A.S., Herries, A.I.R., King, G.C.P., Berger, L.R., 2010. Geologicalsetting and age of Australopithecus sediba from Southern Africa. Science 328,205–208.

Garcia, T., Féraud, G., Falguères, C., de Lumley, H., Perrenoud, C., Lordkipanidze, D., 2010.Earliest human remains in Eurasia: new 40Ar/39Ar dating of the Dmanisi hominid-bearing levels, Georgia. Quaternary Geochronology 5, 443–451.

Grine, F.E., Jungers, W.L., Schultz, J., 1996. Phenetic af"nities among early Homo craniafrom East and South Africa. Journal of Human Evolution 30, 189–225.

Grün, R., Spooner, N., Magee, J., Thorne, A., Simpson, J., Yan, G., Mortimer, G., 2011.Stratigraphy and chronology of the WLH 50 human remains, Willandra LakesWorld Heritage Area, Australia. Journal of Human Evolution 60 (5), 597–604.

Herries, A.I.R., 2003. Magnetostratigraphy of the South African hominid palaeocaves.American Journal of Physical Anthropology S36, 113.

Herries, A.I.R., Adams, J.W., Kuykendall, K.L., Shaw, J., 2006a. Speleology andmagnetobiostratigraphic chronology of the GD 2 locality of the Gondolinhominin-bearing paleocave deposits, North West Province, South Africa. Journalof Human Evolution 51, 617–631.

Herries, A.I.R., Reed, K.E., Kuykendall, K.L., Latham, A.G., 2006b. Speleology andmagnetobiostratigraphic chronology of Buffalo Cave fossil site, Makapansgat, SouthAfrica. Quaternary Research 66, 233–245.

31R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32

Author's personal copy

Herries, A.I.R., Curnoe, D., Adams, J.W., 2009. A multi-disciplinary seriation of earlyHomo and Paranthropus bearing palaeocaves in southern Africa. QuaternaryInternational 202, 14–28.

Herries, A.I.R., Hopley, P., Adams, J., Curnoe, D., Maslin, M., 2010. Geochronology andpalaeoenvironments of the South African early hominin bearing sites: a reply to‘Wrangham et al., 2009: Shallow-Water Habitats as Sources of Fallback Foods forHominins’. American Journal of Physical Anthropology 143, 640–646.

Hill, A., 1995. Faunal and environmental change in the Neogene of East Africa. In:Vrba, E.S., Denton, G.H., Partridge, T.C., Burckle, L.H. (Eds.), Paleoclimate andEvolution with Emphasis on Human Origins. Yale University Press, New Haven,pp. 178–193.

Howell, F.C., 1978. Hominidae. In: Maglio, V.J., Cooke, H.B.S. (Eds.), Evolution of AfricanMammals. Harvard University Press, Cambridge, pp. 154–248.

Johanson, D.C., White, T.D., 1979. A systematic assessment of early African hominids.Science 203, 321–330.

Kimbel, W.H., White, T.D., Johanson, D.C., 1988. Implications of KNM-WT 17000 for theevolution of “robust” Australopithecus. In: Grine, F.E. (Ed.), Evolutionary History ofthe “Robust” australopithecines. Aldine de Gruyter, New York, pp. 259–268.

Kuman, K., Clarke, R.J., 2000. Stratigraphy, artefact industries and hominid associationsfor Sterkfontein, Member 5. Journal of Human Evolution 38, 827–847.

Ludwig, K.R., 2000. Isoplot/Ex version 2.2. Berkley Geochronology Centre SpecialPublication 1a.

Lundberg, J., Ford, D.C., Hill, C.A., 2000. A preliminary U–Pb date on cave spar, bigcanyon, Guadalupe Mountains, NewMexico, USA. Journal of Cave and Karst Studies62, 144–148.

McFadden, P.L., Brock, A., Partridge, T.C., 1979. Paleomagnetism and the age of themakapansgat hominid site. Earth and Planetary Science Letters 44, 373–382.

Partridge, T.C., 1978. Re-appraisal of lithostratigraphy of Sterkfontein hominind site.Nature 275, 282–287.

Partridge, T.C., 2000. Hominid-bearing cave and Tufa deposits. In: Partridge, T.C., Maud,R.R. (Eds.), The Cenozoic in Southern Africa. Oxford University Press, pp. 100–125.

Partridge, T.C., Watt, I.B., 1991. The stratigraphy of the Sterkfontein hominid depositand its relationship to the underground cave system. Palaeontologica Africana 28,35–40.

Partridge, T.C., Shaw, J., Heslop, D., Clarke, R.J., 1999. The new hominid skeleton fromSterkfontein, South Africa: age and preliminary assessment. Journal of QuaternaryScience 14, 293–298.

Partridge, T.C., Granger, D.E., Caffee, M.W., Clarke, R.J., 2003. Lower Pliocene hominidremains from Sterkfontein. Science 300, 607–612.

Pickering, R., 2005. The stratigraphy, chronology and palaeoenvironments of thePleistocene cave "ll. Gladysvale Cave, South Africa., School of Geosciences.University of the Witwatersrand, Johannesburg.

Pickering, R., Kramers, J.D., 2010. A re-appraisal of the stratigraphy and new U–Pb datesat the Sterkfontein hominin site, South Africa. Journal of Human Evolution 59,70–86.

Pickering, R., Hancox, P.J., Lee-Thorp, J.A., Grun, R., Mortimer, G.E., McCulloch, M.,Berger, L.R., 2007. Stratigraphy, U–Th chronology, and paleoenvironments at

Gladysvale Cave: insights into the climatic control of South African hominin-bearing cave deposits. Journal of Human Evolution 53, 602–619.

Pickering, R., Kramers, J.D., Partridge, T.C.P., Kodolanyi, J., Pettke, T., 2010. U–Pb Datingof Calcite–Aragonite Layers in Speleothems from Hominin Sites in South Africa byMC–ICP-MS. Quaternary Geochronology.

Polyak, V., Hill, C., Asmerom, Y., 2008. Age and evolution of the grand canyon revealedby U–Pb dating of water table-type speleothems. Science 319, 1377–1380.

Rasbury, E.T., Cole, J.M., 2009. Directly dating geologic events: U–Pb dating ofcarbonates. Reviews of Geophysics 47, RG3001.

Reynolds, S.C., 2007. Mammalian body size changes and Plio–Pleistocene environmen-tal shifts: implications for understanding hominin evolution in eastern andsouthern Africa. Journal of Human Evolution 53, 528–548.

Richards, D.A., Bottrell, S.H., Cliff, R.A., Strohle, K., Rowe, P.J., 1998. U–Pb dating of aspeleothem of Quaternary age. Geochimica Et Cosmochimica Acta 62, 3683–3688.

Sarnawojcicki, A.M., Meyer, C.E., Roth, P.H., Brown, F.H., 1985. Ages of tuff beds at EastAfrican early hominid sites and sediments in the Gulf of Aden. Nature 313, 306–308.

Smith, P.E., Farquhar, R.M., 1989. Direct dating of Phanerozoic sediments by the 238U–206Pb method. Nature 341, 518–521.

Steininger, C., Berger, L.R., Kuhn, B.F., 2008. A partial skull of Paranthropus robustus fromCooper's Cave, South Africa. South African Journal of Science 104, 143–146.

Vrba, E.S., 1985. Early hominids in southern Africa: updated observations onchronological and ecological background. In: Tobias, P.V. (Ed.), Hominid Evolution.Past, Present and Future. Alan R. Liss, New York, pp. 195–200.

Vrba, E.S., 1988. Late Pliocene climatic events and hominid evolution. In: F.E., G. (Ed.),Evolutionary history of the “robust” australopithecines. Aldine de Gruyter, NewYork, pp. 405–426.

Walker, J., 2005. Uranium–Lead dating of Hominid fossil sites in South Africa. Universityof Leeds, Leeds.

Walker, J., Cliff, R.A., Latham, A.G., 2006. U–Pb isotopic age of the StW 573 hominid fromSterkfontein, South Africa. Science 314, 1592–1594.

White, T.D., 1995. African omnivores: global climatic change and Plio–Pleistocenehominids and suids. In: Vrba, E.S., Denton, G.H., Partridge, T.C., Burckle, L.H. (Eds.),Paleoclimate and Evolution, with Emphasis on Human Origins, pp. 369–384.

White, T.D., Harris, J.M., 1977. Suid evolution and correlation of African hominidlocalities. Science 198, 13–21.

Wilkinson, M.J., 1983. Geomorphic perspectives on the Sterkfontein AustralopithecineBreccias. Journal of Archaeological Science 10, 515–529.

Wood, B., 1991. Koobi Fora Research Project, Volume 4: Hominid Cranial Remains.Clarendon Press, Oxford.

Wood, B., Richmond, B.G., 2000. Human evolution: taxonomy and paleobiology. Journalof Anatomy 197, 19–60.

Woodhead, J., Hellstrom, J., Maas, R., Drysdale, R., Zanchetta, G., Devine, P., Taylor, E., 2006.U–Pb geochronology of speleothems by MC–ICPMS. Quaternary Geochronology 1,208–221.

Woodhead, J., Reisz, R., Fox, D., Drysdale, R., Hellstrom, J., Maas, R., Cheng, H., Edwards,R.L., 2010. Speleothem climate records from deep time? Exploring the potentialwith an example from the Permian. Geology 38, 455–458.

32 R. Pickering et al. / Earth and Planetary Science Letters 306 (2011) 23–32