Cosmogenic nuclide burial dating of hominin-bearing Pleistocene cave deposits at Swartkrans, South Africa

Quaternary Geochronology 24 (2014) 10e15

Contents lists available at ScienceDirect

Quaternary Geochronology journal homepage: www.elsevier.com/locate/quageo

Research paper

Cosmogenic nuclide burial dating of hominin-bearing Pleistocene cave deposits at Swartkrans, South Africa Ryan J. Gibbon a, *, Travis Rayne Pickering b, c, d, Morris B. Sutton e, Jason L. Heaton c, d, f, Kathleen Kuman c, e, Ron J. Clarke c, C.K. Brain d, Darryl E. Granger g a

Department of Anthropology, University of New Brunswick, Fredericton, New Brunswick, E3B 5A3, Canada Department of Anthropology, University of Wisconsin-Madison, 1180 Observatory Drive, Madison, WI, 53706, USA Evolutionary Studies Institute, University of the Witwatersrand, WITS 2050, Johannesburg, South Africa d Plio-Pleistocene Palaeontology Section, Department of Vertebrates, Ditsong National Museum of Natural History (Transvaal Museum), Pretoria, 0002, South Africa e School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, WITS 2050, Johannesburg, South Africa f Department of Biology, Birmingham-Southern College, Birmingham, AL, 35254, USA g Department of Earth, Atmospheric, and Planetary Sciences, Purdue University, West Lafayette, IN, 47907, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 June 2014 Received in revised form 21 July 2014 Accepted 24 July 2014 Available online 4 August 2014

Based on the cosmogenic nuclide burial dating technique, we present new radiometric age estimates of 2.19 ± 0.08 and 1.80 ± 0.09 million-years-old (Ma) for Member 1, and 0.96 ± 0.09 Ma for Member 3 of the Swartkrans Formation in South Africa. Our data are consistent with, and expand upon, results from previous radiometric dating techniques used at the site. The burial ages of Member 1 are consistent with the uraniumelead (UePb) age provided by bracketing flowstones (Pickering et al., 2011), while the age of Member 3 is significantly more precise than the large age bracket provided by UePb dating of tooth enamel (Balter et al., 2008) and recently re-evaluated electron spin resonance data (Herries and Adams, 2013). These new dates provide the complete age range for the extinct hominin, Paranthropus robustus, as well as indicate the first appearance of the genus Homo in southern Africa. Our results also indicate: the first, as well as the last, manufacture and use of bone digging tools in South Africa; some of the earliest evidence of stone tool use and large animal butchery in South Africa; and one of the earliest archaeological indications of the domestication of fire in the world. © 2014 Elsevier B.V. All rights reserved.

Keywords: Cosmogenic nuclide burial dating Swartkrans Cave Paranthropus robustus Early Homo Stone and bone tools Fire

1. Introduction Paleoanthropologydthe field of human evolutionary studiesdis not only concerned with tracking the biological and cultural development of our own genus, Homo, but also with that of the genera and species that gave rise to Homo, as well as that of its sympatric, collateral relatives. The latter category includes species of the megadont genus Paranthropus, represented in the Pleistocene fossil record of South Africa by Paranthropus robustus.1 Since its fossils were first discovered at Kromdraai Cave (Broom, 1938), research on P. robustus continues to reveal this extinct hominin2 as a fascinating * Corresponding author. Tel.: þ1 506 458 7998. E-mail address: [email protected] (R.J. Gibbon). 1 TRP and JLH prefer the use of Australopithecus robustus over Paranthropus robustus. However, in order to conform to prevailing usage, we employ the latter designation in this paper. 2 KK and RJC do not recognise any validity to the term hominin and argue for retention of the family name hominid. http://dx.doi.org/10.1016/j.quageo.2014.07.004 1871-1014/© 2014 Elsevier B.V. All rights reserved.

and often unexpected animal. Although its highly derived skull anatomydwith large jaws, a heavily built, crested cranium and thickly enameled postcanine teethdindicates that P. robustus was capable of subsisting on a coarse, low-quality diet (e.g., Robinson, 1962), occlusal microwear (e.g., Scott et al., 2005) and stable carbon isotope results reveal that it actually “had an extremely flexible diet, which may indicate that its derived masticatory morphology signals an increase, rather than decrease, in its potential foods” (Sponheimer et al., 2006, p. 981). P. robustus fossils are also commonly associated spatially with those of early Homo, as well as archaeological indications of tool manufacture and use, the butchery of large animals and the possible control of fire by early hominins (e.g., Brain et al., 1988; Brain, 1993a; Kuman and Clarke, 2000). Attributing these behaviorally important archaeological traces to P. robustus is confounded by its sympatry (Broom and Robinson, 1950) with South African early Homo (Wood and Strait, 2004; Pickering, 2006). We are also left to speculate why P. robustus went extinct in the wake of cofamilial Homo's eventual global ascendancy.

R.J. Gibbon et al. / Quaternary Geochronology 24 (2014) 10e15

Our current multidisciplinary research project at Swartkrans (Fig. 1A)dfollowing in the steps of those conducted there previously by Robert Broom and John Robinson (1948e1949, 1951e1953) and then by C.K. Brain (1965e1986)dis focused on addressing these unresolved issues related to the behavior and evolution of P. robustus and early Homo. Key to shedding light on these issues is construction of an accurate chronological framework for the various fossil-bearing Pleistocene sediments preserved within the cave. To this end, cosmogenic nuclide burial dating (using 26Al and 10Be) was undertaken in order to determine the age of two of the members of the Swartkrans Formation. Dating of early hominin sites in South Africa continues to be a contentious area of research, as often different techniques produce varying age results for the same deposit (e.g. Partridge et al., 2003; Pickering and Kramers, 2010). This variance has usually been attributed to the complexity of multiple episodes of karstification, infilling and erosion at these sites. As numerous dating techniques have been attempted at Swartkrans (see section 3), the site provides an ideal opportunity to compare results from burial dating (a method that is more commonly used in landscape denudation

11

studies) to other established techniques used in research on hominin evolution in the region. Our data confirm, and expand on the results from these previous investigations, demonstrating agreement between several radiometric dating techniques. With this accurate chronological framework, paleoanthropologists and archaeologists now have a sound footing in order to begin to address several key evolutionary issues discussed above. 2. Swartkrans Formation Swartkrans evolved as a phreatic maze cave within the impure dolomitic limestone of the Chunniespoort Group (Palmer, 1991; Brain, 1993b). The cave probably first opened to the ground surface sometime in the early Pleistocene. From that time, it began to admit materials of the Swartkrans Formation (Butzer, 1976; Brain, 1976, 1993b), which comprises five sequential sedimentary members, separated by erosional discontinuities. Because Members 4 and 5 are composed of more recently admitted sedimentsdaccumulated, respectively, 110,980 (Sutton et al., 2009) and ~12,000e9000 years ago (Brain, 1993b; Watson, 1993)dthey are not discussed here.

Fig. 1. (A) Political map of southern Africa (left), with the “Cradle of Humankind” World Heritage area in detail (right); Swartkrans Cave is indicated by a star, with other major Paranthropus robustus sites indicted by black dots. (B) Plan view of Swartkrans, illustrating the major depositional units of the Swartkrans Formation. From oldest to youngest, these are: the Lower Bank (LB) and Hanging Remnant (HR) of Member 1; Member 2 (not figured here in order to clarify the underlying deposits of Member 1); Member 3 (M3); the Member 4 Middle Stone Age colluvium (M4). (C) Schematic section, designated on the plan in (B), running north-south through the filling of the outer portion of Swartkrans Cave (the fabric of the LB deposit, with an inclination of 30 , indicates that clastic infill and associated paleoanthropological materials entered the cave from a location well-above the current cave opening and to its south). Approximate locations of cosmogenic nuclide burial dating samples are indicated on (B) and (C) with stars and their associated ages; all uncertainties are reported at one-sigma.

12

R.J. Gibbon et al. / Quaternary Geochronology 24 (2014) 10e15

Member 1 (see Fig. 1B and C) is composed of three subunits, the Lower Bank (LB), the Lower Bank East Extension (LBEE) and the Hanging Remnant (HR) (Brain, 1993b; Sutton et al., 2009; Pickering et al., 2012). The first two subunits of Member 1 are indistinguishable depositionally and sedimentologically. They form a single infill and are only classified separately because the eastern portion of this depositional body was not recognized until long after the Swartkrans Formation was re-codified by Brain (1993b). The surface-derived soils of the LB/LBEE were admitted into Swartkrans via a shaft (or possibly multiple shafts) that opened to the ground surface above. Eventually this materialdcharacterized by sandy silt, with occasional clasts of subangular dolomite, chert and quartz (4e120 mm in size), as well as numerous stone and bone3 artifacts and fossilsdfilled most of the underground opening, banking against the north wall of the cavern. The LB/LBEE shows various degrees of calcification: most of the deposit has been decalcified or was never calcified to begin with, but there are also portions of it that are lightly, moderately and heavily calcified. The entryway(s) for the LB/LBEE sediments was/were eventually choked by sediment and new shafts then opened to the north, admitting a subsequent cycle of materials. This younger, heavily calcified depositdoriginally referred to as the “pink breccia” but now classified as the HRdconsists of surface-derived soils and sandy sediment, characterized by stony inclusions that include blocks of dolomite roof-spall and chert (Brain, 1958). Fossils are abundant in the HR, but the breccia has yielded no stone tools and only two bone tools (Watson, 1993). The HR sediments eventually completely filled their points of entry. A subsequent period of erosion within the cave opened a gap several meters in width between the base of the HR and the top of the LB. New openings to the ground surface admitted the Member 2 sediments in at least two distinct phases. The first phase filled the gap eroded between the LB and HR, spreading widely across the center of the cave. The second ingression of materials that are included as part of Member 2 in Brain's (1993b) stratigraphic reconstruction occurred through an opening at the extreme northwest corner of the cavern; this material is identifiable as alternating bands of water-lain, fine and coarser-grained sands that adhere to the northwest wall of the cave. Except for a thin remaining skin of these stratified sediments, most of Member 2 was long ago removed by excavations at Swartkrans. Generally, the Member 2 “brown breccia” (Brain, 1958) is easily distinguished by its lack of stony inclusions any bigger than gravel (a few pieces of dolomite roof-spall excepted); fallen blocks of calcified “pink breccia” from the HR are sometimes observed in the Member 2 matrix. Member 2 has also yielded abundant fossils and stone and bone artifacts. The Member 3 (Fig. 1B) sediments occur in a gully that was eroded into Member 1 and 2 deposits along the cave's west wall. Member 3 is composed of heavily calcified sediments grading to lightly calcified in the upper-most exposed levels. The deposit contains predominantly clasts of angular to sub-angular rock, including dolomite (4e120 mm in size). The deeper portions of the gully include a few large to very large clasts of dolomite, some of which may be spall from the original cave roof. Member 3 is rich in fossils, stone and bone tools and burned bones. 3. Previous age determinations There have been several previous attempts to date the fossiliferous Pleistocene deposits of the Swartkrans Formation (Members

1e3). This work yielded only imprecise or very broad age ranges and, consequently, a poorly constrained chronological framework for the evolution of P. robustus and South African early Homo. Initial assignments of age to each of the formation's various members was based solely on biostratigraphic data. Faunal comparisons with absolutely dated East African sites indicated that each of the timesuccessive Members 1e3 is between 1.8 and 1.0 million-years-old (Ma) (Brain, 1993b; de Ruiter, 2003). Some researchers, using bovid and equid data, assigned dates more specifically, with Member 1 at 1.7 Ma, Member 2 at 1.5 Ma, and Member 3 at 1.0 Ma (Vrba, 1985; Churcher and Watson, 1993). Subsequently, the electron spin resonance (ESR) dating technique was used to date bovid enamel from Member 3, yielding a problematically broad age range of >3.0e0.36 Ma for this archaeologically significant deposit (Blackwell, 1994). Another attempt employing ESR produced highly variable results on three tooth enamel samples from the HR depositional subunit (Brain, 1993b) of Member 1, giving an age range of 2.0e1.0 Ma (Curnoe et al., 2001). Balter et al.’s (2008) uraniumelead (UePb) dating of tooth enamel samples from all three early Pleistocene members of the Swartkrans Formation provided ages of 1.83 ± 1.38 Ma for Member 1, 1.36 ± 0.20 Ma for Member 2, and 0.83 ± 0.21 Ma for Member 3. But enamel is an open isotopic system, with the potential for diagenetic gain or loss of U and Pb (Balter et al., 2008). So, although the UePb-derived ages are in the correct stratigraphic order, it is important to validate them employing other methods that are not subject to U migration. Using results from multiple dating studies, Herries et al. (2009) seriated the Swartkrans deposits with deposits from other relevant cave sites in the region, concluding that Swartkrans Member 1 was c. 2.0 Ma, with subsequent phases of deposition over the ensuing million years. Most recently, R. Pickering et al. (2011) employed the more reliable UePb technique to date calcium carbonate flowstones interbedded between clastic cave sediments at two locations: in the northwest corner of the site and on the central portion of its north wall. R. Pickering et al.’s (2011) UePb sample SKW 12 (2.249 ± 0.077 Ma) underlies both the HR and the LB, effectively providing a maximum age for the whole of Member 1. Flowstone at the top of the section provides a minimum age for Member 1 of 1.706 ± 0.069 Ma (UePb uncertainties were reported at twosigma). Although, these results provide precise brackets on the timing of sedimentation, they are still incapable of narrowing the age range of >500,000 years for deposition of Member 1. There is no way to distinguish from the flowstone ages whether Member 1 was deposited early, late or consistently within this excessively long period. In addition, the flowstone ages also provide an upper bound only on the ages of Members 2 and 3. Building on the work of R. Pickering et al. (2011), we used the cosmogenic nuclide burial dating technique (Granger and Muzikar, 2001; Gibbon et al., 2009) to derive radiometric ages for the LB of Member 1 (the oldest known early Pleistocene unit of the Swartkrans Formation), as well as for Member 3 (the most recent early Pleistocene unit of the Swartkrans Formation). We were unable to identify appropriate samples for dating Member 2 at the time of sampling (Member 2 falls chronologically between Members 1 and 3). By dating sediment directly, our burial ages complement the bracketed age of Member 1, based on UePb flowstone analyses (Pickering et al., 2011), and also provide an age that augments previous work dating Member 3 through analysis of UePb in bovid teeth (Balter et al., 2008). 4. Methods

3 KK reserves judgment that bone tool specimens are tools and is currently researching an alternative explanation for their presence.

Four burial dating samples were collected from the Swartkrans Formation (Fig. 1B and C). The first sample, from the Member 1 LB,

R.J. Gibbon et al. / Quaternary Geochronology 24 (2014) 10e15

13

Table 1 Cosmogenic nuclide concentrations and burial ages. Sample ID

Deptha (m)

[10Be] (106 atoms/g)

SK1-Chert SK1-LB SK1-M SK3

8.55 8.55 5.47 7.00

0.032 1.341 1.248 1.424

a b c d

± ± ± ±

0.009 0.038 0.024 0.025

[26Al] (106 atoms/g) 0.089 3.388 2.626 5.528

± ± ± ±

0.128 0.132 0.103 0.240

26

Al/10Be

2.734 2.527 2.104 3.881

± ± ± ±

4.011 0.122 0.092 0.182

True ageb (Ma)

Max agec (Ma)

Min aged (Ma)

NA 1.799 ± 0.088 2.187 ± 0.082 0.958 ± 0.090

NA 1.803 ± 0.088 2.215 ± 0.084 0.960 ± 0.091

NA 1.772 ± 0.086 2.129 ± 0.079 0.946 ± 0.089

Depth below the cave roof. Calculated using a surface erosion rate of 3 m/Ma. Calculated using a surface erosion rate of 0 m/Ma. Calculated using a surface erosion rate of ∞ m/Ma.

is a fluvially rounded quartzite cobble/manuport (SK1-M) that was obtained from the exposed Bloubank River gravels at the bottom of the valley at Swartkrans Hill and then must have been carried upslope to the vicinity of the cave by a Pleistocene hominin toolmaker. The second sample is a fraction of quartz-bearing sediment (SK1-LB) excavated from the LB talus, while the third sample is a chert clast (SK1-Chert) excavated from the immediate vicinity of the sediment sample. The last sample is a fraction of quartzbearing sediment (SK3) excavated from the Member 3 talus. No suitable manuport or chert clast was available from the Member 3 deposit at the time of sampling. Cosmogenic nuclide burial dating was undertaken on the four samples. Details of the burial dating methodology are found in Granger (2014) and Granger and Muzikar (2001). The technique is based on the radioactive decay of 26Al and 10Be that is produced in quartz that has been exposed to secondary cosmic radiation near the ground surface. Once a sample is buried and shielded from cosmic radiation, the relative decay of 26Al (t1/2 ¼ 0.702 Ma) and 10Be (t1/ 2 ¼ 1.39 Ma) within the mineral grains yields an age for when the sample was buried (Granger and Muzikar, 2001; Gibbon et al., 2009). The duration of exposure and the depth of burial are both important for accurate results. For a grain to be dated it first must have been exposed at the ground surface in order to have built up a sufficient inventory of 26Al and 10Be. Then, the grain must have been buried sufficiently deeply to have blocked continued cosmogenic nuclide production. For a slowly eroding landscape, such as found in South Africa, a burial depth of 5e10 m is usually sufficient to block continued cosmogenic nuclide production. A small amount of cosmogenic nuclide production continues to occur due to reactions with muons even after burial; the amount of postburial production can be modeled using known production rate depth profiles. The Swartkrans samples were buried 5e9 m below the surface. This range of depth blocks most cosmogenic nuclide production, but there is a small amount of postburial production by muons that must be factored into analyses. We use a deposit density of 2.1 g/cm3 and the muon profiles recently reported by Balco et al. (2013) and assume that the ground surface at Swartkrans is eroding at 3 m/Ma. That is, we assume that the samples were buried quickly, and that their burial depth has been steadily decreasing at 3 m/Ma since deposition. This value is based on landscape denudation rates of 3e5 m/Ma determined from a nearby location (Dirks et al., 2010), as well as from a modern sediment sample collected from the surface of a hilltop at Sterkfontein Caves ~1 km across the valley from Swartkrans (Partridge et al., 2003). In addition, we calculate maximum and minimum burial ages, as provided in Table 1. The maximum age calculation assumes that the depth of the sample has not changed since it was buried in the cave (i.e. an erosion rate of 0 m/Ma), while the minimum age calculation assumes a high erosion rate of the top of the deposit, such that post-burial production of 26Al and 10Be can be ignored (Gibbon et al., 2009). As neither of these two erosional end members reflects reality, we believe that an erosion rate of 3 m/ Ma most closely reflects the erosional situation recorded at the site. It should be kept in mind that although such variations in the amount

of postburial production lead to some additional uncertainty, for our samples they only affect burial ages by 10e30 thousand years, well within the bounds of analytical error. For burial dating to work, the sediment must have been buried only once in the past ~10 million years, otherwise the burial age reflects a combination of the burial events and will reflect an age older than the most recent burial. It is difficult to know a priori whether a sample has experienced prior burial. If multiple samples are collected from the same site, an isochron method can be used to validate this assumption and identify outliers (e.g., Erlanger et al., 2012). Surface samples can also be collected as a check to confirm that they have a zero burial age. Although we did not collect sediment from the surface at Swartkrans, modern sediment in a similar setting at Sterkfontein Caves (again, only ~1 km from Swartkrans) has no inherited burial signal (Partridge et al., 2003); this supports the assumption that material enters the cave with no prior burial history. Quartz samples were processed following the procedure outlined in Gibbon et al. (2009). 26Al and 10Be were measured by accelerator mass spectrometry (AMS) at PRIME Lab, Purdue University (Indiana, USA), against standards prepared by Nishiizumi (2004); Nishiizumi et al. (2007). All uncertainties are reported at one-sigma. Local production rates of 10.44 and 70.10 atoms/g/year for 10Be and 26Al, respectively, were estimated for a latitude of 26.0172 south and an elevation of 1.48 km, following Stone (2000), revised to match the Nishiizumi et al. (2007) AMS standards. 5. Results 5.1. Member 1 LB Of the three samples collected from the LB, only two (SK1-LB and SK1-M) were suitable for burial dating. The third sample (SK1-Chert), in contrast, contains very low concentrations of 26Al and 10Be, indicating that it had not been exposed at the cave surface prior to incorporation in the deposit (Table 1). Rather, this clast is interpreted to have eroded out of an interior wall of the cave, rendering it unsuitable for burial dating. The low concentrations of the SK1-Chert sample do, however, confirm that postburial production at the sampling site is very minimal and provide an upper limit on postburial production. Samples SK1-LB (sediment sample) and SK1-M (manuport sample) are discussed together as they were located in close stratigraphic proximity and are thus expected to have been deposited nearly simultaneously (Fig. 1B and C). The age of SK1-LB is 1.80 ± 0.09 Ma; the age of SK1-M is 2.19 ± 0.08 Ma (Table 1). The manuport has a burial age slightly older than the sediment, distinguishable at three-sigma. 5.2. Member 3 The single sediment sample from Member 3 (SK3) produced an age of 0.96 ± 0.09 Ma (Table 1), in good agreement with the

14

R.J. Gibbon et al. / Quaternary Geochronology 24 (2014) 10e15

previous determination by UePb on teeth (Balter et al., 2008). The burial age for Member 3 provides the most precise and robust age determination yet for this deposit. 6. Discussion The results from Member 1 indicate depositional ages ranging from 2.19 ± 0.08 Ma (manuport) to 1.80 ± 0.09 Ma (bulk sediment). Although these two samples were collected at different depths within the same unit, they are out of stratigraphic order so the variation in ages requires some explanation. We see three possibilities: (1) the difference could be due to analytical uncertainty, although we can rule this option out with ~99% confidence. In this case, the average age of 1.99 ± 0.19 Ma (±1 s.e.) best represents the age of the deposit. (2) The manuport could have been collected from a previously buried terrace deposit on the surface, making its burial age too old. However, there are no apparent sources that are buried sufficiently deeply. (3) There could be heterogeneity within Member 1. That is, material within an individual talus cone may move downslope discontinuously, being remobilized by mass movements or slumps within the cave. In this case, the ages represent a duration of time during which sediment was introduced into the cave. This third option is not to be confused with mixing between units, which we do not believe to be the case at Swartkrans given the distinct sediment sources for the different units. We cannot choose between the three options at this time. The burial ages of Member 1 could represent a single average age of 1.99 ± 0.19 Ma; or rather a duration of deposition from 2.19 ± 0.08 Ma to 1.80 ± 0.09 Ma; or a depositional age closer to 1.8 Ma (if the manuport was deposited with an inherited burial signal). Whatever the case, the age is fully consistent with the bracketing UePb ages of flowstones bounding Member 1 (Pickering et al., 2011). The close agreement between the two methods supports the accuracy of burial dating at Swartkrans. Our age for Member 3 is consistent with and significantly more precise than the rather large age bracket of 1.3e0.6 Ma provided by UePb dating of enamel and recently re-evaluated ESR data for Member 3 (Balter et al., 2008; Herries and Adams, 2013). Collectively, our results refine the radiometrically derived age range for P. robustus, as well as update the chronological framework for the evolution of South African Homo and major technological innovations, such as bone and stone tools and the control of fire. Our burial ages provide numerical dating of the sedimentary infill at Swartkrans, and are fully consistent with prior work at the site. The agreement between different radiometric dating techniques demonstrates that cave sites with relatively simple site formation histories will produce consistent dating results. Because of the extensive preservation of P. robustus fossils over a great length of time at Swartkrans, the site provides a truly unique situation for chronological research on the age range of this endemic, long-lived and successful South African hominin species. Cosmogenic nuclide burial dating now provides the most accurate and robust age determinations for the first and last appearance of P. robustus in South Africa. The burial dating data are in complete agreement with other chronological data for the first appearance of P. robustus (Herries and Adams, 2013). In contrast, the species' last appearance date has up to now been relatively unclear (Herries and Adams, 2013). 7. Significance Beyond bracketing the age range of P. robustus, the new burial data also place the first appearance of early Homo in South Africa (based on Swartkrans Member 1 fossil specimens such as the crania

SK 847 and SK 27 [Clarke et al., 1970; Clarke, 1977]), as well as the origins of a unique South African bone-digging-tool tradition (Brain et al., 1988; Brain, 1993a). Swartkrans Member 1 also now stands as yielding one of the oldest stone tool assemblages in southern Africa (Clark, 1993; Sutton, 2012), and the subcontinent's oldest butchered animal bones (Pickering et al., 2008). All three classes of early hominin behavioral evidencedbone and stone tools and butchered animal bonesdwere also recovered from Members 2 and 3 of the Swartkrans Formation (Brain et al., 1988; Brain, 1993a; Clark, 1993; Pickering et al., 2008; Sutton, 2012), with Member 3 being the last appearance in the South African archaeological record of early Pleistocene bone tools (Brain et al., 1988; Brain, 1993a). All known South African bone tools are associated stratigraphically with P. robustus (Robinson, 1959; Brain et al., 1988; Brain, 1993a; Backwell and d’Errico, 2008), and conventional wisdom holds that they were used by that hominin (and less likely by sympatric early Homo, also associated stratigraphically with bone tools at some sites) to extract edible underground resources, such as plant roots and tubers (Brain et al., 1988; Brain, 1993a). However, occlusal microwear (Scott et al., 2005) and isotopic (Sponheimer et al., 2006) results suggest that P. robustus had a variable diet, unlikely to have been restricted to the low-quality, tough vegetable foods that are indicated by sole consideration of its highly derived craniodental functional morphology (Robinson, 1962). Thus, the dual disappearance of P. robustus and bone tools from the South African paleoanthropological record, documented in Swartkrans Member 3, might be wholly coincidental rather than causally linked. Swartkrans Member 3 is also well-known for the 270 bone specimens whose heat-altered chemistry and histology suggest that they were burned in fires reaching temperatures around or in excess of 400 e 500  C (Brain and Sillen, 1988; Brain, 1993a; Sillen and Hoering, 1993). The burned bone specimens were excavated from much of the entire 6-m-thickness of Member 3; four of them are also scarred by butchery marks and only two other burned bone specimens are known from the entirety of the rest of the Swartkrans Formation (Brain, 1993a). Based on this context, Brain and Sillen (1988) hypothesized that the Member 3 bones were burned in a series of controlled, intermittent fires maintained by Swartkrans hominins over the thousands of years that it took that depositional stratum to accumulate in the cave. The new age of 0.96 ± 0.09 Ma for Member 3 accords with other well-accepted evidence of hominin-controlled fires from Wonderwerk Cave, South Africa (c. 1.0 Ma) (Berna et al., 2012), and from Gesher Benot Ya'aqov, Israel (c. 0.8 Ma) (Goren-Inbar et al., 2004). Finally, the persistence of P. robustus until c. 1.0 Ma, as evidenced by its fossils recovered from Swartkrans Member 3 (Grine, 1989), would seem to indicate that not only was this species contemporary with H. habilis and H. ergaster at Swartkrans (Clarke, 2012), but also with more evolved forms of our genus at other sites, such as Buia (Abbate et al., 2004) and Daka (De Heinzelin et al., 2000) in East Africa. Acknowledgments The Swartkrans Paleoanthropology Research Project (SPRP) is supported by grants to: T.R. Pickering from the National Science Foundation (USA), the L.S.B. Leakey Foundation, the Palaeontological Scientific Trust (PAST, South Africa), and the Vilas Associates Program (University of WisconsineMadison); Ryan Gibbon from the National Research Foundation (NRF, South Africa) and PAST; Morris Sutton from the NRF and PAST; and Kathleen Kuman from PAST and the NRF. We thank the following individuals, institutions, and companies (in alphabetical order) for their assistance with our project: African Explosives; the Brain/Newman/ Watson family; Laurent Bruxelles; John Cruise; Hendrick

R.J. Gibbon et al. / Quaternary Geochronology 24 (2014) 10e15

Dingiswayo; Manuel Domínguez-Rodrigo; Charles Egeland; the Evolutionary Studies Institute, University of the Witwatersrand; Barry Jacoby; Lazarus Kgasi; Andrea Leenen; Isaac Makhele; Sipho Makhele; Nkwane Molefe; Abel Molepolle; Stephen Motsumi; Andrew Phaswana; Robert Pickering; the T.R. Pickering family; Stephany Potze; the late Dusty van Rooyen; the School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand; Lucas Sekowe; Solomon Seshoene; Francis Thackeray; the Transvaal Museum. This is paper No. 5 in the SPRP publication series. Editorial handling by: D. Bourles

References Abbate, E., Woldehaimanot, B., Libsekal, Y., Tecle, T.M., Rook, L. (Eds.), 2004. A Step Towards Human Origins: the Buia Homo One-million-years Ago in the Eritrean Danakil Depression (East Africa), vol. 110. Revista Italiana di Paleontologia e Stratigrafia, Milan. Supplement. Backwell, L., d'Errico, F., 2008. Early hominid bone tools from Drimolen, South Africa. J. Archaeol. Sci. 35, 2880e2894. Balco, G., Soreghan, G.S., Sweet, D.E., Marra, K.R., Bierman, P.R., 2013. Cosmogenicnuclide burial ages for Pleistocene sedimentary fill in Unaweep Canyon, Colorado, USA. Quat. Geochronol. 18, 149e157. de, F., 2008. UePb Balter, V., Blichert-Toft, J., Braga, J., Telouk, P., Thackeray, F., Albare dating of fossil enamel from the Swartkrans Pleistocene hominid site, South Africa. Earth Planet. Sci. Lett. 267, 236e246. Berna, F., Goldberg, P., Horwitz, L.K., Brink, J., Holt, S., Bamford, M., Chazan, M., 2012. Microstratigraphic evidence of in situ fire in the Acheulean strata of Wonderwerk Cave, Northern Province, South Africa. Proc. Natl. Acad. Sci. U. S. A. 109, E1215eE1220. Blackwell, B.A., 1994. Problems associated with reworked teeth in electron spin resonance (ESR) dating. Quat. Sci. Rev. 13, 651e660. Brain, C.K., 1958. The Transvaal Ape-man-bearing Cave Deposits. Transvaal Museum, Pretoria. Brain, C.K., 1976. A re-interpretation of the Swartkrans site and its remains. S. Afr. J. Sci. 72, 141e146. Brain, C.K. (Ed.), 1993a. Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria. Brain, C.K., 1993b. Structure and stratigraphy of the Swartkrans Cave in light of the new excavations. In: Brain, C.K. (Ed.), Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria, pp. 23e33. Brain, C.K., Churcher, C.S., Clark, J.D., Grine, F.E., Shipman, P., Susman, R.L., Turner, A., Watson, V., 1988. New evidence of early hominids, their culture and environment from the Swartkrans Cave, South Africa. S. Afr. J. Sci. 84, 828e835. Brain, C.K., Sillen, A., 1988. Evidence from Swartkrans Cave for the earliest use of fire. Nature 336, 464e466. Broom, R., 1938. Pleistocene anthropoid apes of South Africa. Nature 142, 377e379. Broom, R., Robinson, J.T., 1950. Man contemporaneous with the Swartkrans apeman. Am. J. Phys. Anthropol. 8, 151e156. Butzer, K.W., 1976. Lithostratigraphy of the Swartkrans formation. S. Afr. J. Sci. 72, 136e141. Churcher, C.S., Watson, V., 1993. Additional fossil Equidae from Swartkrans. In: Brain, C.K. (Ed.), Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria, pp. 137e150. Clark, D., 1993. Stone artefact assemblages from members 1 e 3, Swartkrans Cave. In: Brain, C.K. (Ed.), Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria, pp. 167e194. Clarke, R.J., 1977a. The Cranium of the Swartkrans Hominid, SK 847, and its Relevance to Human Origins (PhD dissertation). University of the Witwatersrand. Clarke, R.J., 1977b. A juvenile cranium and some adult teeth of early Homo from Swartkrans, Transvaal. S. Afr. J. Sci. 3, 46e49. Clarke, R.J., 2012. The history of research in human evolution in Africa and what lessons have been learned. In: Sanz, N. (Ed.), Human Origins Sites and the World Heritage Convention in Africa, World Heritage Papers 33. UNESCO, New York, pp. 44e67. Clarke, R.J., Howell, F.C., Brain, C.K., 1970. More evidence of an advanced hominid at Swartkrans. Nature 225, 1219e1222. Curnoe, D., Grün, R., Taylor, L., Thackeray, F., 2001. Direct ESR dating of a Pliocene hominin from Swartkrans. J. Hum. Evol. 40, 379e391. De Heinzelin, J., Clark, J.D., Schick, K.D., Gilbert, W.H., 2000. The Acheulean and the e Royal de Plio-Pleistocene Deposits of the Middle Awash Valley, Ethiopia. Muse ol. 104. l’Afrique Central, Tervuren. Belg. Ann.-Sci. Ge de Ruiter, D., 2003. Revised faunal lists for members 1 e 3 of Swartkrans, South Africa. Ann. Transvaal Mus. 40, 29e41.

15

Dirks, P.H.G.M., Kibii, J.M., Kuhn, B.F., Steininger, C., Churchill, S.E., Kramers, J.D., riaux, A., Herries, A.I.R., King, G.C.P., Berger, L.R., Pickering, R., Farber, D.L., Me 2010. Geological setting and age of Australopithecus sediba from southern Africa. Science 328, 205e208. Erlanger, E.D., Granger, D.E., Gibbon, R.J., 2012. Rock uplift rates in South Africa from isochron burial dating of fluvial and marine terraces. Geology 40, 1019e1022. Gibbon, R.J., Granger, D.E., Kuman, K., Partridge, T.C., 2009. Early Acheulean technology in the Rietputs Formation, South Africa, dated with cosmogenic nuclides. J. Hum. Evol. 56, 152e160. Goren-Inbar, N., Alperson, N., Kislev, M.E., Simchoni, O., Melamed, Y., Ben-Nun, A., Werker, E., 2004. Evidence of hominin control of fire at Gesher Benot Ya'aqov. Science 304, 725e727. Granger, D.E., 2014. Cosmogenic nuclide burial dating in archaeology and paleoanthropology. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry, second ed., vol. 14. Elsevier, Oxford, pp. 81e97. Granger, D.E., Muzikar, P.F., 2001. Dating sediment burial with in situ-produced cosmogenic nuclides: theory, techniques, and limitations. Earth Planet. Sci. Lett. 188, 269e281. Grine, F.E., 1989. New hominid fossils from the Swartkrans Formation (1979 e 1986 excavations): craniodental specimens. Am. J. Phys. Anthropol. 79, 409e449. Herries, A.I.R., Adams, J.W., 2013. Clarifying the context, dating and age range of the Gondolin hominins and Paranthropus in South Africa. J. Hum. Evol. 65, 676e681. Herries, A.I.R., Curnoe, D., Adams, J.W., 2009. A multi-disciplinary seriation of early Homo and Paranthropus bearing palaeocaves in southern Africa. Quat. Int. 202, 14e28. Kuman, K., Clarke, R.J., 2000. Stratigraphy, artefact industries and hominid associations for Sterkfontein, Member 5. J. Hum. Evol. 38, 827e847. Nishiizumi, K., 2004. Preparation of 26Al AMS standards. Nucl. Instrum. Methods Phys. Res. Sect. B, Beam Interact. Mater. Atoms 223e224, 388e392. Nishiizumi, K., Imamura, M., Caffee, M.W., Southon, J.R., Finkel, R.C., McAninch, J., 2007. Absolute calibration of 10Be AMS standards. Nucl. Instrum. Methods Phys. Res. Sect. B, Beam Interact. Mater. Atoms 258, 403e413. Palmer, A.N., 1991. Origin and morphology of limestone caves. Geol. Soc. Am. Bull. 103, 1e21. Partridge, T.C., Granger, D.E., Caffee, M.W., Clarke, R.J., 2003. Lower Pliocene hominid remains from Sterkfontein. Science 300, 607e612. Pickering, R., Kramers, J.D., 2010. Re-appraisal of the stratigraphy and determination of new U-Pb dates for the Sterkfontein hominin site, South Africa. J. Hum. Evol. 59, 70e86. Pickering, R., Kramers, J.D., Hancox, P.J., de Ruiter, D.J., Woodhead, J.D., 2011. Contemporary flowstone development links early hominin bearing cave deposits in South Africa. Earth Planet. Sci. Lett. 306, 23e32. Pickering, T.R., 2006. Subsistence behaviour of South African Pleistocene hominids. S. Afr. J. Sci. 102, 205e210. Pickering, T.R., Egeland, C.P., Domínguez-Rodrigo, M., Brain, C.K., Schnell, A.G., 2008. Testing the “shift in the balance of power” hypothesis at Swartkrans, South Africa: hominid cave use and subsistence behavior in the Early Pleistocene. J. Anthropol. Archaeol. 27, 30e45. Pickering, T.R., Heaton, J.L., Clarke, R.J., Sutton, M.B., Brain, C.K., Kuman, K., 2012. New hominid fossils from member 1 of the Swartkrans formation, South Africa. J. Hum. Evol. 62, 618e628. Robinson, J.T., 1959. A bone implement from Sterkfontein. Nature 184, 583e585. Robinson, J.T., 1962. The origin and adaptive radiation of the australopithecines. In: Kurth, G. (Ed.), Evolution und Hominisation. Fischer, Stuttgart, pp. 120e140. Scott, R.S., Ungar, P.S., Bergstrom, T.S., Brown, C.A., Grine, F.E., Teaford, M.F., Walker, A., 2005. Dental microwear texture analysis reflects diets of living primates and fossil hominins. Nature 436, 693e695. Sillen, A., Hoering, T., 1993. Chemical characterization of burnt bones from Swartkrans. Swartkrans Cave. In: Brain, C.K. (Ed.), Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria, pp. 243e249. Sponheimer, M., Passey, B., de Ruiter, D., Guatelli-Sternberg, D., Cerling, T., LeeThorp, J., 2006. Isotopic evidence for dietary flexibility in the early hominin Paranthropus robustus. Science 314, 980e982. Stone, J.O., 2000. Air pressure and cosmogenic isotope production. J. Geophys. Res. 105, 23753e23759. Sutton, M.B., 2012. The Archaeology of Swartkrans Cave, Gauteng, South Africa: New Excavations of Members 1 and 4 (PhD dissertation). University of the Witwatersrand. Sutton, M.B., Pickering, T.R., Pickering, R., Brain, C.K., Clarke, R.J., Heaton, J.L., Kuman, K., 2009. Newly discovered fossil- and artifact-bearing deposits, uranium-series ages, and Plio-Pleistocene hominids at Swartkrans Cave, South Africa. J. Hum. Evol. 57, 688e696. Vrba, E.S., 1985. Early hominids in southern Africa: updated observations on chronological and ecological background. In: Tobias, P.V. (Ed.), Hominid Evolution: Past, Present and Future. Alan Liss, New York, pp. 195e200. Watson, V., 1993. Composition of the Swartkrans bone accumulations, in terms of skeletal parts and animals represented. In: Brain, C.K. (Ed.), Swartkrans: a Cave's Chronicle of Early Man. Transvaal Museum, Pretoria, pp. 35e73. Wood, B., Strait, D., 2004. Patterns of resource use in early Homo and Paranthropus. J. Hum. Evol. 46, 119e162.