Environmental and ecological implications of strontium isotope ratios in mid-Pleistocene fossil teeth from Elandsfontein, South Africa

Accepted Manuscript Environmental and ecological implications of strontium isotope ratios in mid-Pleistocene fossil teeth from Elandsfontein, South Africa

S.B. Lehmann, N.E. Levin, D.R. Braun, D.D. Stynder, M. Zhu, P.J. le Roux, J. Sealy PII: DOI: Reference:

S0031-0182(17)30452-2 doi:10.1016/j.palaeo.2017.10.008 PALAEO 8475

To appear in:

Palaeogeography, Palaeoclimatology, Palaeoecology

Received date: Revised date: Accepted date:

28 April 2017 2 October 2017 9 October 2017

Please cite this article as: S.B. Lehmann, N.E. Levin, D.R. Braun, D.D. Stynder, M. Zhu, P.J. le Roux, J. Sealy , Environmental and ecological implications of strontium isotope ratios in mid-Pleistocene fossil teeth from Elandsfontein, South Africa. The address for the corresponding author was captured as affiliation for all authors. Please check if appropriate. Palaeo(2017), doi:10.1016/j.palaeo.2017.10.008

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Environmental and ecological implications of strontium isotope ratios in mid-Pleistocene fossil teeth from Elandsfontein, South Africa Coauthors: Lehmann, S.B.a, Levin, N.E. b, Braun, D.R.c, d, e, Stynder, D.D.e, Zhu, M.e, le Roux,

IP

T

P.J.e, Sealy, J.e

Affiliations:

University of Pittsburgh, Department of Geology and Environmental Science, 4107 O’Hara Street, SRCC, Room 200, Pittsburgh, PA 15260, USA, [email protected]

CR

a

b

US

University of Michigan, Earth and Environmental Sciences, 2534 C. C. Little Building, 1100 North University Avenue, Ann Arbor, MI 48019, USA, [email protected] George Washington University, Center for the Advanced Study of Human Paleobiology, 800 22nd Street NW, Washington DC, 20052, USA, [email protected]

AN

c

d

M

Max Planck Institute for Evolutionary Anthropology, Department of Human Evolution, Deutscher Platz 6, Leipzig, Germany, [email protected] e

PT

ED

University of Cape Town, Department of Archaeology, Rondebosch 7701, South Africa, [email protected], [email protected], [email protected], [email protected], [email protected]

AC

CE

Journal: Palaeogeography, Palaeoclimatology, Palaeoecology

1

ACCEPTED MANUSCRIPT

ABSTRACT The well-known South African mid-Pleistocene site of Elandsfontein has yielded an abundance of fossil fauna and artifacts (ca. 1 to 0.6 million years ago) and has produced hominin fossils, specifically a calvarium and mandibular fragment often assigned to the species Homo heidelbergensis. Elandsfontein is located within the Greater Cape Floristic Region of South Africa,

IP

T

in an area that is today dry, hot, and covered by marine-derived sand dunes. It is primarily vegetated by woody shrubs and sedges of the fynbos biome with minimal amounts of grass. In

CR

contrast, the Elandsfontein faunal assemblage from the mid-Pleistocene is rich in large mammalian grazing herbivores, indicating that past environment and ecology was substantially different from

US

that of today. It is unknown whether these animals survived on local vegetation, suggesting a substantially different landscape and ecology from today, or if they migrated to nonlocal, nutrient-

AN

rich areas to obtain food, such as to the shale substrates ~ 20 to 30 km inland from Elandsfontein, where palatable grass and shrubs grow today. The strontium isotope ratios (87Sr/86Sr) of fossil teeth

M

provide insights into the ranges of the animals recovered from Elandsfontein because they reflect the 87Sr/86Sr of plants consumed by these animals, which in turn reflect the 87Sr/86Sr of the

ED

substrates (or soils) in which those plants grew. We measured 87Sr/86Sr in 24 fossilized teeth (158 measurements) from Elandsfontein and compared these ratios with the bioavailable 87Sr/86Sr from

PT

major substrates in the region to determine if the large mammalian herbivores obtained sufficient resources locally (i.e., teeth reflect local 87Sr/86Sr) or if they travel inland to find food. The

CE

distribution of bioavailable 87Sr/86Sr for coastal marine sands range from 0.709380 to 0.711690; most of this range is lower than the bioavailable 87Sr/86Sr for granite, sandstone, and shale

AC

substrates in the region which range from 0.711421 to 0.723563. 87Sr/86Sr of fossil teeth of carnivores, large herbivores, and rodents from Elandsfontein fall within the 87Sr/86Sr range of the marine sands. Our results indicate localized habitation patterns for a diversity of large African mammals, including communities of grazing herbivores, and a resource-rich environment that made mid-Pleistocene Elandsfontein and coastal southwestern South Africa attractive to hominins for which there is no modern analogue.

2

ACCEPTED MANUSCRIPT

Keywords: landscape, ecology, herbivore diet, enamel, migration

1. Introduction The southwestern coast of South Africa is located within the Greater Cape Floristic Region

IP

T

(GCFR). This vegetation constitutes one of the most diverse biomes in the world, dating to the Miocene (e.g., Dupont et al., 2011). The composition of the vegetation and its nutritional value to

CR

herbivores varies with the mineralogical composition of substrate (Proche et al., 2006). While the fynbos ecosystem is thought to have evolved substantially since its emplacement, little is known

US

about the variability of vegetation in southwestern South Africa or about its faunal associations prior to the Last Glacial Maximum (LGM), ~21,000 years ago (Stynder, 2009; Rossouw et al.,

AN

2009).

Elandsfontein is an open-air archeological and paleontological site (33º10’S, 18º24’E;

M

Fig. 1) with artifacts and fossils co-occurring on a buried landscape with an age of ca. 1.0 to 0.6 million years ago (Ma), placing it within the mid-Pleistocene (e.g., Klein and Cruz-Uribe, 1991;

ED

Braun et al., 2013). The diversity of the mid-Pleistocene fauna has led to the suggestion that this must have been a productive environment that provided substantial amounts of browse and graze

PT

for herbivores (e.g., Luyt et al., 2000; Klein et al., 2007; Stynder, 2009; Braun et al., 2013). Based on mesowear, Stynder (2009) found the dietary behavior of some faunal species to be different

CE

from expected when compared with what is considered their typical dietary classification using taxonomic analogy. For example, several supposedly grazing ungulates were re-assigned as mixed

AC

feeders and others as browsers. Today Elandsfontein is an arid shrubland with active dunes and is mostly vegetated by sclerophyllous fynbos vegetation growing on nutrient poor marine-derived sands (Mucina and Rutherford, 2006). There are no modern analogues for the mid-Pleistocene ecosystem and environment of Elandsfontein and the current vegetation structure and water budget makes survival for similarly diverse ecosystems difficult to impossible (e.g., Klein et al., 2007; Stynder, 2009). While Elandsfontein cannot support large herbivore communities today, there are more nutrient-rich granite and shale substrates ~ 20 km to 30 km inland (Fig. 1) that weather to fertile

3

ACCEPTED MANUSCRIPT

soils and host more palatable shrubs and a higher proportion of grasses than do the marine-derived sands along the coast (Mucina and Rutherford, 2006). Today these shale substrates are mostly farmland. It remains unknown how large mammals survived at Elandsfontein and elsewhere in coastal southwestern South Africa (i.e., currently nutrient-poor regions) during the mid-Pleistocene. Did animals acquire the nutrition they needed locally near Elandsfontein or by going further afield

T

to access resources growing in the fertile soils associated with shale substrates to the east?

IP

Sealy et al. (1991) have shown that the substrates in the region differ in 87Sr/86Sr, with

CR

Pleistocene coastal sands having lower values than more ancient (and more inland) shales and sandstones, but they did not report 87Sr/86Sr data for granites from the region (Fig. 1). The 87Sr/86Sr

US

of herbivore tissues (e.g., bone and tooth enamel) is the same as that of plants ingested by the animals, which in turn is determined by the 87Sr/86Sr of the soil or rock substrate on which these

AN

plants grew. Since different minerals in the soil or rock can have different 87Sr/86Sr and may weather at different rates, the 87Sr/86Sr in plants and in animals, known as the “bioavailable Sr”,

M

may not be exactly the same as that of the underlying geology or homogeneous across a substrate (e.g., Bentley, 1996). Nevertheless, 87Sr/86Sr data have the potential to indicate where within a

ED

region herbivores ate food. Here we expand upon existing studies of the variation in 87Sr/86Sr from the major substrates in southwestern South Africa to determine whether mid-Pleistocene mammals

PT

from Elandsfontein fed on vegetation in the marine-derived sands near Elandsfontein and were able to obtain adequate nutritional and other resources locally, along the coast, or whether they traveled

CE

to areas with different underlying substrate. As a necessary component of the study, we have developed a map of bioavailable 87Sr/86Sr for each major substrate in the region using modern 87

AC

material, and have used an independent set of fossil and modern teeth to determine whether the Sr/86Sr data from the mid-Pleistocene teeth are primary or if they have been affected by post

depositional alteration.

2. Background 2.1. Study region: climate, vegetation, and geology of southwestern South Africa

4

ACCEPTED MANUSCRIPT

The region of southwestern South Africa lies within the winter rainfall zone, a climate zone encompassing ~ 200 km2 that receives ~ 65% of its annual precipitation between April and September and is distinct from that of the rest of South Africa (e.g., Chase and Meadows, 2007). East of this region, rainfall occurs year-round or during the summer months (> 66% of mean annual precipitation between October and March).

IP

T

A map of southwestern South Africa showing the main geological substrates is presented in Figure 1. The coastal plain in this region is covered by eolian sands of marine provenance,

CR

containing shell fragments that date from the Miocene to the present (Roberts, 2006a,b; Roberts et al., 2009). About twelve granitic plutons that form the Cape Granite suite outcrop in southwestern

US

South Africa and are associated with orogenic events related to the convergence of South American and Namibian cratons during the Precambrian (Rozendaal et al., 1999; Belcher and Kisters, 2003).

AN

These granites, distinguishable from one another by mineralogical and geochemical composition, are scattered throughout the study region. The most dominant granite types are found along the

M

coast (Scheepers, 1995). Inland from the marine sands and coastal granites, Precambrian to Cambrian Malmesbury shales and Cape Fold Mountain Belt complex sandstones and quartzites

ED

forming the Table Mountain/ Witteberg/ Bokkeveld series are the dominant lithologies (see Fig. 1). The winter rainfall zone supports a distinct vegetation type, the Fynbos Biome (Cowling

PT

and Lombard, 2002). The composition of vegetation communities varies across the landscape as a function of geological substrate (Vogel et al., 1978; Cowling, 1997; Cowling et al., 2002). Today,

CE

the region is dominated by nutrient-poor, though diverse fynbos shrubland on coastal sands and on sandstones and quartzite (Rebelo et al., 2006). In contrast, where present, shales support more

AC

nutrient-rich vegetation that include a larger proportion of grasses as well as many evergreen asteracerous shrubs (Cowling, 1997; Goldblatt and Manning; 2002). These are palatable to animals today, as they would have been to the large herbivorous mammals living during the Pleistocene. Granites also produce richer soils than marine sands.

2.2. Elandsfontein Initial site surveys in the 1950s yielded an abundance of exposed fossils and artifacts. Fragments of a hominin calvarium and mandible were also discovered during this initial round of

5

ACCEPTED MANUSCRIPT

research. The calvarium, which became known as the Saldanha or Hopefield calvarium, was subsequently assigned to Homo heidelbergensis. Surveys and excavations carried out in the 1950s and 1960s recovered stone tools in association with large mammalian fossils (e.g., Singer, 1956; Singer and Wymer, 1968; Klein, 1978; Deacon, 1998). Over 50 years later, in situ materials were recovered in new excavations, and it was determined during this most recent round of research that

T

the large mammalian fossils and artifacts were located on a buried landscape that is exposed at

IP

various places within a > 11 km2 area (Braun et al., 2013). Artifacts of the Acheulean Industry,

CR

which have sometimes been considered to have evolved in conjunction with major behavioral and biological changes in human evolution (Lepre et al., 2011; Beyene et al., 2013), are found at

US

Elandsfontein. The large concentrations of these artifacts and the behavioral association with cut marked bones (Forrest et al., in review) indicate that Elandsfontein was frequented by hominins in

AN

the mid-Pleistocene (e.g., Braun et al., 2013).

The faunal assemblage from Elandsfontein indicates that the mid-Pleistocene mammalian

M

community included many large browsing and grazing herbivores, including species within the Bovidae, Rhinocerotidae, Hippopotamidae, Giraffidae, Elephantidae, Equidae and Suidae families

ED

(e.g., Klein and Cruz-Uribe, 1991; Klein et al., 2007; Luyt et al., 2002; Lehmann et al., 2016). The preservation of ancient spring deposits and the presence of Hippopotamidae at Elandsfontein

PT

suggest that there was standing water and that this location may have been a spring-fed environment (e.g., Braun et al., 2013; Lehmann et al., 2016; Patterson et al., 2016). The carbon

CE

isotope compositions (13C values) of fossil teeth from both browsing and grazing species indicate that the mid-Pleistocene landscape was dominated by C3 vegetation (Luyt et al., 2000; Lehmann et

AC

al., 2016). C3 grasses flourish in regions with cool growing seasons such as in the winter-rainfall climate of the south-western Cape today. Elandsfontein is located on the coastal plain of the southwestern coast of South Africa, which is a predominantly eolian landscape, composed of sands originating from marine (biogenic carbonate, including fragments of marine shell) and terrestrial sources (quartz grains) with minimal outcrops of granite and shale (Mabbutt, 1956; Besaans, 1972; Scheepers and Nortje, 2000; Cowling and Lombard, 2002; Cowling et al., 2005; Chase and Thomas, 2007) (Fig. 1). The main vegetation types at Elandsfontein are the Hopefield Sand Fynbos and Strandveld, with Strandveld dunefields

6

ACCEPTED MANUSCRIPT

penetrating > 15 km inland from the coast (Mucina and Rutherford, 2006; Roberts et al., 2009). Both Strandveld and Hopefield Sand Fynbos grow on nutrient-poor marine-derived sands (Rebelo et al., 2006). There is minimal surface water and very little grass; the dominant plant forms are woody shrubs and Restionaceae. The grasses that do exist in these habitats cannot easily sustain

T

herds of large grazing herbivores (Radloff, 2008).

IP

2.3. 87Sr/86Sr of bioapatite and regional vegetation in southwestern South Africa

CR

The 87Sr/86Sr in bioapatite (enamel and bone) can be used as a geochemical fingerprint of the substrates on which animals lived, because the bioavailable 87Sr/86Sr from different substrates is

US

preserved in the teeth and bones of animals eating these plants (e.g., Bentley, 2006). The 87Sr/86Sr ratios of bioapatite have been used to determine migration patterns and land use among ancient

AN

mammals in southern Africa, including herbivores and hominids (Sillen et al., 1998; Sealy et al., 1991; Copeland et al., 2011; Copeland et al., 2016).

M

Previously, three of the four major substrates in southwestern South Africa were characterized in terms of their bioavailable 87Sr/86Sr using bone and teeth of a limited number of

ED

modern animals (Sealy et al., 1991; See Fig.1 and Table 1). Animals that lived on Pleistocene marine-derived sands in the region have bioavailable 87Sr/86Sr that range from 0.70938 to 0.71169

PT

(n = 6), distinct from those that lived on the shales and sandstones with values of 0.71777 and 0.71794 (n = 2) and 0.71543 and 0.71746 (n = 2), respectively. There are currently no published

CE

bioavailable 87Sr/86Sr data from granites in the study region and it is unknown whether the granites are distinct from the other major substrates. However, Copeland et al. (2016) reports the

AC

bioavailable 87Sr/86Sr of the Cape Granite Suite on the southern coast of South Africa, based on plant samples, ranges from 0.7095 to 0.7177. These values overlap with the bioavailable 87Sr/86Sr of major substrates in southwestern South Africa as reported by Sealy et al. (1991). The new data presented in this paper will expand on and improve the characterization of bioavailable 87Sr/86Sr of the substrates in southwestern South Africa. The large mammalian herbivore communities that existed at Elandsfontein during the mid-Pleistocene would have needed substantial quantities of vegetation, both browse and graze (Klein et al., 2007; Braun et al., 2013). Adequate vegetation does not exist at present on the marine-

7

ACCEPTED MANUSCRIPT

derived sands in the immediate vicinity of Elandsfontein and it is unknown whether it existed during the mid-Pleistocene. 2.4. Tooth enamel formation and factors influencing 87Sr/86Sr of bioapatite Previous studies show that it is possible to exploit differences in bioavailable 87Sr/86Sr from

IP

T

the major regional substrates to determine resource- and landscape-use among herbivores in the region (Sealy et al., 1991; Radloff et al., 2010). Tooth enamel bioapatite can be used for this

CR

purpose. Tooth enamel is oldest at the occlusal surface and youngest at the cervix and incorporates the 87Sr/86Sr of the resources consumed over the period it took for the enamel to mineralize. Once

US

formed, enamel is neither added to nor remodeled. By serial sampling the crown along the tooth growth axis (from occlusal surface to the cervix), it is possible to explore animal migration

AN

involving changes in food sources over time. Furthermore, in the case of the large herbivores enamel of a single tooth can take eight months to more than a year to fully mineralize so each

M

sampled portion of enamel from a tooth represents feeding over several months (as reviewed in Kohn and Cerling, 2002).

ED

The primary 87Sr/86Sr of a material can be altered after deposition via leaching or addition of secondary, diagenetic strontium that can mask the primary 87Sr/86Sr (e.g., Bentley, 2006).

PT

Assessment for the presence of diagenetic strontium and its removal when found are important parts of employing 87Sr/86Sr when characterizing samples (e.g., Sealy et al., 1991; Hoppe et al.,

CE

2003; Bentley, 2006). Bone and dentine are more likely to absorb diagenetic strontium from a burial environment than are enamel, since they are porous (Lee-Thorp, 2008). However, the

AC

presence of diagenetic strontium should also be assessed in enamel because (denser than bone and dentine) diagenetic strontium can accumulate at the surfaces and in fractures when enamel is not pristine. In addition, the comparison of dentine and enamel 87Sr/86Sr from the same tooth can help distinguish if the primary enamel has been altered (Copeland et al., 2010). Determining the presence of diagenetic Sr is important especially where bioapatite is not consistently well preserved, like in coastal southwestern South Africa. For example, some enamel and soils at Elandsfontein have experienced post depositional removal of carbonate (Luyt et al., 2000; Braun et al., 2013; Lehmann et al., 2016).

8

ACCEPTED MANUSCRIPT

While primary and diagenetic bioapatite can co-occur, diagenetic bioapatite can be removed from a sample through dissolution. Diagenetic bioapatite is more heavily substituted than primary bioapatite, resulting in a more imperfect bioapatite crystal which tends to be more soluble than the primary bioapatite (Sillen, 1986; Tuross et al., 1989). Thus, when bioapatite is bathed in a weak acid, the secondary mineral dissolves more readily than the primary biogenic mineral (e.g., Sillen,

IP

T

1986; Sealy et al., 1991; Hoppe et al., 2003). Repeated treatment with acid dissolves the most soluble apatite first, then less soluble material (Sealy et al., 1991). In some depositional contexts,

CR

such as those rich in fluoride, diagenetic mineral may be less soluble than biogenic (as reviewed in Sealy et al., 1991), however this is not the case in the Western Cape. Comparison of 87Sr/86Sr of the

US

supernatants from successive acid rinses and the corresponding bioapatite residue that remains at the end of the process have been used to detect the presence of diagenetic strontium and to attempt

AN

to remove it in what is referred to as a “solubility profile” (Sillen, 1986; Sillen and LeGeros, 1991; Sealy et al., 1991). If there are significant changes in 87Sr/86Sr during the profile, diagenetic

M

strontium is likely to be present. Past work with solubility profiles has mostly been done on bone. However, in areas where sediments, bone and enamel carbonate may have experienced post

PT

ED

depositional alteration, such as at Elandsfontein, it is important to determine if 87Sr/86Sr is primary.

3. Methods

CE

3.1. Samples

We collected plant material and bones and teeth of contemporary small herbivores with

AC

small ranges (e.g., rodents and tortoises) from the major substrate in our study region. We mostly collected bones and teeth of small herbivores because their tissue provides the best estimate of the bioavailable 87Sr/86Sr of a substrate (Bentley, 2006). We sampled material from different locations having the same substrate to account for differences in bioavailable 87Sr/86Sr within a substrate. We avoided cultivated areas as the 87Sr/86Sr of fertilizers can mask the primary bioavailable 87Sr/86Sr of a material. The locations of bioavailable 87Sr/86Sr ratios of modern bioapatite and plants are shown in Figure 1 and Table S1.

9

ACCEPTED MANUSCRIPT

Fossil teeth from Elandsfontein were collected as both surface and excavated finds of the West Coast Research Project between 2008 and 2015 (Braun et al., 2013) and are archived in the Department of Archaeology, University of Cape Town. As shown in Table 2 and Table S2, we sampled 24 teeth from the mammalian herbivore families Bovidae (n = 17), Equidae (n = 1) and Elephantidae (n = 2), the rodent families Bathyergidae (n = 1) and Hystricidae (n = 1), and the

IP

T

carnivore family Hyaenidae (n = 2). When possible, we sampled the third molars of browsing and grazing herbivore taxa that were potentially migratory and thus may have traveled for food

CR

(Stynder, 2009; Copeland et al., 2016). We selected teeth that were unworn or lightly worn and so were more likely to preserve variations in 87Sr/86Sr along the growth axis compared with very worn

US

teeth. In addition, we sampled teeth of mid-Pleistocene animals that were likely local to Elandsfontein, such as the rodents (Patterson et al., 2016), and two carnivores, brown hyenas, that

AN

probably traveled long distances in search of food (Klein et al., 2007). We also analyzed a subset of modern tooth enamel (n = 2) and fossil tooth enamel fragments (n = 8) with a range of preservation

M

states to assess the degree of post depositional alteration of the primary 87Sr/86Sr of the fossil teeth at Elandsfontein. Modern samples were collected from various settings and are from animals that

ED

had been exposed long enough to be completely cleared of soft tissue and hair.

PT

3.2. Sample preparation for 87Sr/86Sr determination Samples were prepared and analyzed at the Department of Geological Sciences and the

CE

Department of Archaeology, University of Cape Town. Modern enamel, bone, and plant material were placed in porcelain crucibles with lids and ashed in a muffle furnace (500ºC, overnight) to

AC

remove organics. Each fossil tooth was cleaned of the outermost, exposed enamel surface by abrasion. Serial samples of enamel (~ 10 mg) were collected with a diamond-tip drill bit in a series of approximately horizontal and parallel lines from the base of the tooth up to the occlusal surface. Care was taken during serial sampling to sample enamel in lines close together but without overlap (Fig. 3). To investigate possible diagenesis of primary 87Sr/86Sr, we prepared between 40 and 50 mg of enamel from the subset of modern and fossil enamel collected at Elandsfontein (n = 10). The 87Sr/86Sr of modern and fossil enamel was analyzed using either laser ablation multi-collector-inductively-coupled plasma mass spectroscopy (LA-MC-ICP-MS) or solution

10

ACCEPTED MANUSCRIPT

MC-ICP-MS at the University of Cape Town using the methods described by Copeland et al. (2016) and Copeland et al. (2010), respectively, and the same Nu Instruments NuPlasma HR instrument. Although LA-MC-ICP-MS is less precise than solution MC-ICP-MS (by a factor of 10), it provided a fast and relatively inexpensive method of obtaining 87Sr/86Sr. However, most specimens were studied using solution MC-ICP-MS since the teeth were too large for the LA-MC-

IP

T

ICP-MS cylindrical sample chamber (2.5 cm x 2.5 cm). In addition, at the onset of the study it was not clear whether the fossil teeth had pristine enamel or if they contained diagenetic strontium that

CR

would make them unsuitable for laser ablation.

Teeth studied both by LA-MC-ICP-MS (in situ analysis) and by solution analysis were

US

abraded at the surface with a diamond-tip drill bit to remove the exposed surface enamel. For analyses by solution, ~ 10 mg of enamel powder was rinsed five times in 1 ml of buffered nitric

AN

acid (100 mM, pH 4.5) to remove secondary deposits of strontium from the enamel and dried. Powder was then digested in 2 ml of concentrated (65%) HNO3 in a closed Teflon beaker, heated at

M

140ºC for one hour and dried. The residue was then dissolved in 1.5 ml 2.0 M HNO3 for strontium separation chemistry. For plant samples, between 25 and 50 mg of dried material was digested to

ED

make a solution for strontium separation chemistry using the method outlined in Copeland et al. (2016), following Pin et al. (1994). The strontium-containing fraction of each sample was then

PT

dried, dissolved in 2 ml 0.2% HNO3 and then diluted to a concentration of 200 ppb strontium for analysis by high resolution MC-ICP-MS.

CE

The international standard, NIST SRM987 (87Sr/86Sr = 0.710255), was analyzed between samples. The in-house carbonate standard, NM95, was processed as an unknown with each batch of

AC

nine samples and measured during each run of unknowns, with results in agreement with the long-term average obtained in this laboratory of 0.708907 with two standard deviation (i.e., 2σ) of 0.000031 (n = 24). Unknown samples were analyzed once to determine 87Sr/86Sr of a solution and the error is reported as 2σ internal error.

3.3. Determining the presence of diagenetic strontium To determine if the 87Sr/86Sr of fossil teeth had been significantly changed by environmental strontium after deposition, we analyzed the 87Sr/86Sr of separate set of modern and fossil enamel

11

ACCEPTED MANUSCRIPT

that ranged in quality of preservation (i.e., from well-preserved enamel to enamel that showed alteration of surface morphology that may or may not have been accompanied by chemical changes). Enamel was ground into a powder using a mortar and pestle. Our method for evaluating the presence of diagenetic strontium in an enamel sample (i.e., 87Sr/86Sr different from the primary value) and the removal of the secondary strontium was based on the evaluation of diagenetic

IP

T

strontium in bone in Sealy et al. (1991), in which solubility profiles, or the 87Sr/86Sr from a series of weak acid baths and the remaining, cleaned bone powder, were analyzed. The procedure detects the

CR

presence and subsequent removal of diagenetic strontium in samples from a specific site. Powdered enamel samples were rinsed 24 to 26 times in the buffered nitric acid solution (100 mM, pH 4.5) in

US

the manner described in Section 3.2. Supernatants from every fourth or fifth rinse and the enamel powder that remained at the end (i.e., the residue) were prepared for solution analysis of 87Sr/86Sr

AN

using HR-MC-ICP-MS. Figure 2 is a schematic showing possible scenarios for how the 87Sr/86Sr of supernatant of powdered enamel bathed in weak acid could change with successive rinses for the

M

fossil samples at Elandsfontein.

ED

3.4. Statistical comparison of 87Sr/86Sr values

Comparisons of the 87Sr/86Sr of materials were performed using JMP 11 software, a

PT

statistical analytical program (SAS Institute, Cary, North Carolina, USA). We used the Tukey-Kramer HSD test for these comparisons and, unless otherwise noted, we use the ± symbol

4. Results

AC

CE

throughout the paper to represent one standard deviation from the mean.

4.1. Mapping of bioavailable 87Sr/86Sr ratios Modern bones (n = 15) and teeth (n = 3) from individual animals and seven plant samples were collected from the dominant substrates in the region. Table 1 shows a summary of the 87

Sr/86Sr data for these and the modern bone samples (n = 10) reported by Sealy et al. (1991).

Table S1 provides the full data compilation that is plotted with the new fossil data in Figure 4. We

12

ACCEPTED MANUSCRIPT

have included solution and laser ablation-based data; the difference in 87Sr/86Sr between the two methods based on analysis of three modern enamel samples is < ± 0.000150. The bioavailable 87Sr/86Sr of Pleistocene marine sands ranges from 0.709380 to 0.711690. The bioavailable 87Sr/86Sr of the Table Mountain sandstone complex and the Malmesbury Shale overlap and together these samples range from 0.714092 to 0.720390. The new 87Sr/86Sr data

IP

T

presented here expand upon the range of 87Sr/86Sr for shale and sandstone complexes presented in Sealy et al. (1991) and confirm that the two substrates are not distinct in terms of bioavailable

CR

strontium isotope composition. We have added 87Sr/86Sr data for Cape Granites to the map of bioavailable 87Sr/86Sr of the major substrates in the region and find that animals and plants collected

US

from the Cape Granites have a wide spread of 87Sr/86Sr, spanning from 0.711421 to 0.723563 (Fig. 1). The Cape Granite Suite along the south coast of South Africa has a bioavailable 87Sr/86Sr range

AN

of 0.7095 to 0.7177 and the bioavailable 87Sr/86Sr for the Gamtoos/Malmesbury Group (shale) and the Table Mountain Group (sandstone) range from 0.7095 to 0.7157 and 0.7092 to 0.7169,

M

respectively (Copeland et al., 2016). In general, there is overlap between the bioavailable 87Sr/86Sr of like substrates from the south and southwest coast, however the bioavailable 87Sr/86Sr of

ED

substrates from the southwest coast tend to have more positive values. It is important to note that the bioavailable 87Sr/86Sr ranges of substrates along the south coast are determined from plants,

PT

while they are based mostly on the bioapatite of small herbivores along the southwest coast. Bioapatite can provide a tighter, more averaged range of the bioavailable 87Sr/86Sr of a substrate

CE

than plants (Bentley, 2006).

AC

4.2. Assessment of diagenesis in fossil teeth We measured a subset of 87Sr/86Sr of modern (n = 2) and fossil (n = 8) enamel fragments and the material dissolved in a weak nitric acid supernatant from a series of enamel rinses to evaluate possible diagenesis of the primary isotopic composition of elemental strontium. Fossil fragments were from the families Bovidae and Equidae and from unidentified large mammalian herbivores. We do not observe differences > 0.0006 in 87Sr/86Sr over the series of rinses for any of the solubility profiles (Fig. 5; Table S3). The 87Sr/86Sr from the fossil enamel at Elandsfontein (all rinses and the enamel powder remaining at the end of the process) range from 0.709366 to

13

ACCEPTED MANUSCRIPT

0.710265; these values are within the range of 87Sr/86Sr for marine-derived sand (Fig. 5). The two modern enamel samples have 87Sr/86Sr ranges of 0.0004 and 0.0006 (for all rinses and the enamel powder remaining at the end of the process), which overlaps with the span of 87Sr/86Sr of fossil enamel and their supernatants over a series of ~ 25 rinses (0.00005 to 0.00034). For both modern samples, the first rinse (Rinse 1) have noticeably lower 87Sr/86Sr than the rest of the rinses and the

IP

T

residue enamel powder. In contrast, the fossil samples do not show this pattern. It is possible that modern samples are picking up a labile strontium component soon after deposition whereas

CR

fossilized enamel samples are in some way stabilized. Additionally, the 87Sr/86Sr of the modern alcelaphine enamel (sample 13M-011) is more positive than the 87Sr/86Sr of fossil enamel from

US

Elandsfontein (Fig. 4). Large herbivores currently living at Elandsfontein are brought to the grounds from other locations. It is possible that the bioavailable 87Sr/86Sr range of the substrate

AN

from where the animals originated is different from that of the marine-derived sands at Elandsfontein. Furthermore, animals currently living at Elandsfontein are provided feed because the

M

natural vegetation is not suitable for large herbivores. The origins and 87Sr/86Sr of the feed are currently unknown.

ED

We compared 87Sr/86Sr of modern and fossil enamel with those of their corresponding nitric acid supernatants to see if there were substantial differences between the ranges of 87Sr/86Sr in

PT

pristine, modern enamel (Fig. 5). The change in 87Sr/86Sr of supernatants over a series of ~ 25 rinses is small (~ 0.0004) relative to the range of 87Sr/86Sr for each substrate (> 0.002). The majority of

Table S3).

CE

change in supernatant 87Sr/86Sr for each enamel fragment occurs within the first five rinses (Fig. 5;

AC

Based on the results of this study we conclude that five rinses are sufficient to remove any diagenetic strontium that may be present in the fossil enamel. We therefore carried out five weak nitric acid rinses for each sample of fossil enamel powder used for studies of migration before dissolving the enamel powder for solution chemistry and analyzing the 87Sr/86Sr. Regardless of the condition of fossil enamel, the 87Sr/86Sr of the fossil enamel samples and the series of the supernatants remain within the range of bioavailable 87Sr/86Sr from marine sands, the dominant substrate at Elandsfontein.

14

ACCEPTED MANUSCRIPT

4.3. 87Sr/86Sr of fossil mammalian teeth from Elandsfontein Fossil mammalian tooth enamel from mid-Pleistocene Elandsfontein were serially-sampled along the tooth growth axis whenever possible. We do not observe any significant differences in the 87

Sr/86Sr for different herbivore taxa. The average carnivore 87Sr/86Sr of 0.709507 (n = 2) is not

distinct from other taxa and sits within the range of bioavailable 87Sr/86Sr of marine-derived sands

T

(Fig. 4). All data from the Elandsfontein fossil teeth have 87Sr/86Sr equivalent to bioavailable Sr/86Sr of marine-derived sands and are not within the range of any other substrate, even though

IP

87

CR

there is some overlap between the bioavailable 87Sr/86Sr granite and of marine-derived sands. We observe no behaviorally significant differences between the 87Sr/86Sr of enamel along the growth

US

axes of fossil teeth (i.e., 87Sr/86Sr enamel outside the marine sands range), regardless of the animal

AN

species. However, the ranges of intra-tooth values vary from 0.000049 to 0.001249 (Table S2).

M

5. Discussion

5.1. Distribution of 87Sr/86Sr of regional substrates and determining herbivore migration for food

ED

Determining the distribution of bioavailable 87Sr/86Sr for substrates in southwestern South Africa provides the foundation for understanding how animals utilized landscapes in the

PT

mid-Pleistocene. In addition, this study provides a starting point for future studies of animal and human mobility in this region across a much wider timespan than the mid-Pleistocene.

CE

The values for bioavailable 87Sr/86Sr measured in bone and tooth enamel presented in this study are consistent with those of Sealy et al. (1991), but they expand the amount of data available

AC

for each of the major substrates (i.e., shale, sandstone, and marine sand) (Table S1). From the compilation of new and existing data, we see that the sandstone and shale bioavailable 87Sr/86Sr ranges overlap with one another, with values distinct and more positive than from those of marine sands (Table 1; Fig. 4). Analysis of plants and the bones of modern animals that lived on different Cape Granite outcrops shows that bioavailable 87Sr/86Sr from granite widely varies, encompassing and extending beyond the 87Sr/86Sr ranges of sandstone and shale, and overlapping with the more positive values of marine sands (Fig. 4; Table S1). The lower end of the granite range is surprising, because geologically old granites typically have high 87Sr/86Sr (for review see Faure, 1986; Sposito,

15

ACCEPTED MANUSCRIPT

1989; Capo et al., 1998; Bentley, 2006) as seen at the upper end of the range in this dataset. We note that samples collected closer to the coast (Rondeberg) have lower values than those from localities further inland (Paarl Rock). We suspect that values at the lower end of the range reflect a contribution from marine-derived Sr, perhaps through sea mists, or marine shell fragments included in surrounding sediments. This finding emphasizes the importance of determining bioavailable Sr/86Sr from various locations for studies of this kind.

IP

T

87

CR

5.2. Pretreatment procedure and diagenesis of primary 87Sr/86Sr of fossil enamel Possible diagenesis of the strontium in fossil tooth enamel was tested by analyzing 87Sr/86Sr

US

in solubility profiles (i.e., the supernatants from enamel bathed in a series of weak acid baths and the remaining enamel after ~ 25 baths) of several tooth enamel fragments spanning a range of

AN

preservations. The 87Sr/86Sr along the solubility profiles remain within the range of bioavailable 87

Sr/86Sr of marine sands in the region (See Fig. 2 and Fig. 5). Based on the consistency of 87Sr/86Sr

M

along the solubility profiles, there was no significant diagenesis in any of the specimens. The finding that the primary 87Sr/86Sr of the fossil teeth have been preserved is consistent

ED

with the preservation of biogenic values in the oxygen isotopic composition (i.e., 18O value) of fossil tooth enamel from Elandsfontein. This assessment was based on two observations; first,

PT

taxonomic differences in 18O values of fossil enamel from Elandsfontein are observed and are consistent with modern distributions, where for example hippopotamids have lower enamel 18O

CE

values than the other taxa. Second, the offset between the 18O value of enamel carbonate and that

AC

of enamel phosphate for each sample is as expected in pristine enamel (Lehmann et al., 2016).

5.3. Herbivore migration and environment of southwestern South Africa and Elandsfontein during the mid-Pleistocene

We pretreated each fossil enamel sample in a standard manner (i.e., rinsed five times in weak nitric acid) before measuring its 87Sr/ 86Sr to determine the likely geologic substrate for the vegetation that supported each of the sampled mid-Pleistocene herbivores. This standard

16

ACCEPTED MANUSCRIPT

pretreatment results from our evaluation of solution profiles of modern and fossil teeth at Elandsfontein to assess diagenesis of the bioavailable 87Sr 86Sr of fossil tooth enamel. The 87Sr/86Sr of fossil teeth from Elandsfontein are within the range expected of the marinederived sand substrate found along the coast of southwestern South Africa. These data indicate that browsing and grazing herbivores, rodents, and carnivores from the mid-Pleistocene consistently

IP

T

consumed resources located along the coast and possibly within the area surrounding Elandfontein (Fig. 4; Table 3). Our sample selection included herbivore species that migrate today or that likely

CR

migrated for food in the past (e.g., zebra, buffalo, elephant) to increase the chances of finding 87

Sr/86Sr variation in tooth enamel. However, we did not observe any changes in the 87Sr/86Sr of

US

serial sampled enamel from the fossil teeth that would indicate significant migration for food for any animal (Table 2; Table S2).

AN

Current paleoecological research indicates that the winter rainfall system has been a component of southwestern South Africa for at least the last million years, if not longer (Luyt et al.,

M

2000; Rossouw et al., 2009; Hare and Sealy, 2013; Lehmann et al., 2016). The presence of largebodied herbivores and the mesowear and microwear records from mid-Pleistocene Bovidae teeth

ED

suggest that herbivores had access to both browse and graze (Stynder, 2009). The carbon isotope values of fossil enamel from herbivores at Elandsfontein indicate that grazers ate primarily C3

PT

grasses, grasses adapted to the cool growing seasons, and thus they did not necessarily have to leave southwestern South Africa to find food (Luyt et al., 2000; Lehmann et al., 2016).

CE

The above studies point to differences in the regional environment and ecology of southwestern South Africa during the mid-Pleistocene compared with that of today. However, they

AC

alone cannot provide detail about ecological and environmental variation on the landscape or point to if and or how large browsing and grazing herbivore communities utilized this region for resources and survived ca. 1 – 0.6 Ma. Together with the 87Sr/86Sr of fossil teeth from Elandsfontein presented in this study, these data suggest that herbivores with various diets primarily found food along the coast of southwestern South Africa and possibly, solely in the region of Elandsfontein, suggesting that there must have been a variety of vegetation in this area. Local variation in vegetation and food resources for herbivores is also suggested by the carbon and oxygen isotope datasets for serial samples of tooth enamel from three large grazing mammals (Syncerus antiquus,

17

ACCEPTED MANUSCRIPT

Connochaetes gnou, and Equu capensis) at Elandsfontein that potentially migrated for food (Lehmann et al., 2016). A ≤ 1.1‰ within-tooth variation in 13C values and ≤ 1.0‰ in 18O values, indicating no significant differences in the environment in which the animals fed during the period of tooth mineralization.

T

Marine sands found along the coast of southwestern South Africa and at Elandsfontein may

IP

have had a higher carbonate content during the mid-Pleistocene than they do today due to leaching. If soils had a higher carbonate content, then these soils would have contained more nutrients and

CR

may have supported large amounts of browse and graze (Luyt et al., 2000). In addition, paleospring sediments preserved at Elandsfontein suggest the presence of surface water during the mid-

US

Pleistocene (e.g., Braun et al., 2013). Standing water could have provided adequate water resources to support the growth of grass year round as well as other vegetation like trees, shrubs, and sedges,

AN

all within a winter rainfall zone. The presence of more 13C enriched vegetation, as indicated by the 13C values of micromammal teeth, suggest the possible presence of plants that thrived in a summer

M

growing season (Patterson et al., 2016). However, the enamel 13C values from large herbivore teeth suggest the presence of little to no C4 grasses within the diet of grazers and thus only a minor

ED

contribution of summer rainfall to the growth of grass in southwestern South Africa (Luyt et al., 2000; Lehmann et al., 2016). It is possible that large grazing herbivores incorporated other

PT

vegetation into their diets, as suggested in Stynder (2009), rather than surviving on grass alone. Otherwise, grazers subsisted wholly on a consistent supply of grass from the coastal region (i.e.,

CE

non-seasonal growth).

Overall, Elandsfontein in the mid-Pleistocene would be quite unrecognizable compared to

AC

the current landscape, environment, and ecology. At some point between the mid-Pleistocene and today, southwestern South Africa changed from a lush environment with a diverse animal community of large and very large mammals to an environment that is hot, dry, and hostile to the ecosystems that existed during the mid-Pleistocene. It is entirely plausible that the oscillations of Pleistocene climate resulted in rapid changes in this ecosystem. At different times over the course of the last ~ 2.5 million years, Elandsfontein may have provided a refuge for hominins and other large mammals. Current regional scale paleoecological reconstructions of this region indicate that interglacial periods were likely quite arid. Whereas during times when the southern hemisphere ice

18

ACCEPTED MANUSCRIPT

sheets and concomitant northerly shift of rain-bearing westerlies may have resulted in higher frequencies of rain on the southwestern coast of South Africa (e.g., van Zinderen Bakker, 1976; Stuut et al., 2004; Chase and Meadows, 2007). These fluctuations may explain the great variability

IP

T

in ecosystem recorded in Pleistocene sediments of southwestern South Africa.

6. Conclusions

CR

Our work on modern bone, teeth, and plant samples collected from the major geological substrates in southwestern South Africa forms the foundation of this research and provides the

US

framework to future work that considers 87Sr/86Sr of animals. In addition, our studies of potential diagenetic influences on the 87Sr/86Sr of fossil tooth enamel from Elandsfontein indicate that this is

AN

a faithful proxy of the isotopic values of the plants these animals consumed. As such, 87Sr/86Sr of fossil tooth enamel from Elandsfontein can be used to determine the landscape use of animals

M

during the mid-Pleistocene in southwestern South Africa.

We conclude that herbivores did not migrate from the coastal area of southwestern South

ED

Africa, suggesting that there was enough nutritious vegetation along the coast to feed both browsers and grazers without the need to move inland to find food. These results do not support the

PT

hypothesis that large herbivores undertook long-distance migrations for food or water. Instead, these findings indicate the presence of abundant browse and graze, and by extension, more

CE

abundant water resources on what is now an arid fynbos shrubland. This is consistent with the proposal that the Elandsfontein region had springs during the mid-Pleistocene, as indicated by

AC

previous sedimentological studies. This type of landscape would have supported a large mammal community at Elandsfontein and is fundamentally different from the low carrying capacity environment and impoverished animal community of the present. Further explorations of the nature of the springs in the Elandsfontein deposits will likely help understand how springs modified these landscapes. Additionally, the anthropogenic impact on these ecosystems should be explored further. Locations in southwestern South Africa, such as a spring-fed environment at Elandsfontein, could have formed oases within a relatively drier region, and would have been attractive habitats for animals and hominins alike.

19

ACCEPTED MANUSCRIPT

Acknowledgments This research could not have been completed without the co-operation of the West Coast Fossil Park and the owners and staff of Elandsfontein property (now the Elandsfontein Exploration and Mining). Students from the University of Cape Town, Department of Archaeology (Archaeology in Practice AGE 3013) were essential in sample collection at Elandsfontein (2008 to

IP

T

2015). Sample preparation and measurement was greatly assisted by Kerryn Gray and Fayrooza Rawoot, staff from the Department of Geological Sciences at the University of Cape Town. The

CR

Department of Earth and Planetary Sciences at The Johns Hopkins University greatly assisted in funding for field work and sample collection while SBL was a graduate student at the institution.

US

Analytical costs were supported by the South African Research Chairs initiative of the Department of Science and Technology and the National Research Foundation, and the DST-NRF Centre of

AN

Excellence in Palaeosciences to Judith Sealy at the University of Cape Town. We thank the two anonymous reviewers for their comments and thoughtful suggestions. The opinions, findings and

M

conclusions expressed here are those of the authors and the funding agencies do not accept any liability in this regard. Fieldwork at Elandsfontein was supported by grants from the National

AC

CE

PT

ED

Science Foundation (BCS 1219494 and BCS1219455).

20

ACCEPTED MANUSCRIPT

References Belcher, R.W., Kisters, A.F.M., 2003. Lithostratigraphic correlations in the western branch of the Pan-African Saldania Belt, South Africa: the Malmesbury Group revisited. South African Journal of Geology. 106, 327-342.

IP

T

Bentley, R.A., 2006. Strontium isotopes from the earth to the archaeological skeleton: a review. Journal of Archaeological Method and Theory. 13, 135-187.

CR

Besaans, A.J., 1972. 3217D & 3218 C-st. Helenabaai 3317B & 3318A Saldanhabaai. Geological Survey of South Africa. Department of Mines. Pretoria.

US

Beyene, Y., Katoh, S., WoldeGabriel, G., Hart, W.K., Uto, K., Sudo, M., Kondo, M., Hyodo, M., Renne, P.R., Suwa, G., Asfaw, B., 2013. The characteristics and chronology of the earliest Acheulean at Konso, Ethiopia. Proceedings of the National Academy of Sciences. 110, 1584-1591.

M

AN

Braun, D.R., Levin, N.E., Stynder, D., Herries, A.I.R., Archer, W., Forrest, F., Roberts, D.L., Bishop, L.C., Matthews, T., Lehmann, S.B., Pickering, R., Fitzsimmons, K.E., 2013. Mid-Pleistocene hominin occupation at Elandsfontein, Western Cape, South Africa. Quaternary Science Reviews. 82, 145-166.

ED

Capo, R.C., Stewart, B.W., Chadwick, O.A., 1998. Strontium isotopes as tracers of ecosystem processes: theory and methods. Geoderma. 82, 197-225.

PT

Chase, B. M. and Meadows, M. E., 2007. Late Quaternary dynamics of southern Africa's winter rainfall zone. Earth-Science Reviews. 84, 103-138.

CE

Chase, B.M., Thomas, D.S.G., 2007. Multiphase late Quaternary aeolian sediment accumulation in western South Africa: timing and relationship to palaeoclimatic changes inferred from the marine record. Quaternary International. 166, 29-41.

AC

Copeland, S.R., Cawthra, H.C., Fisher, E.C., Lee-Thorp, J.A., Cowling, R.M., le Roux, P.J., Hodgkins, J., Marean, C.W., 2016. Strontium isotope investigation of ungulate movement patterns on the Pleistocene Paleo-Agulhas Plain of the Greater Cape Floristic Region, South Africa. Quaternary Science Reviews. 141, 65-84. Copeland, S.R., Cawthra, H.C., Fisher, E.C, Lee-Thorp, J.A., Cowling, R.M., le Roux, P.J., Grimes, V., and Richards, M.P., 2011. Strontium isotope evidence for landscape use by early hominins. Nature. 474, 76-100. Copeland, S.R., Sponheimer, M., Lee-Thorp, J.A., le Roux, P.J., de Ruiter, D.J., and Richards, M.P., 2010. Strontium isotope ratios in fossil teeth from South Africa: assessing laser ablation MC ICP-MS analysis and the extent of diagenesis. Journal of Archaeological Science. 37, 1437-1446.

21

ACCEPTED MANUSCRIPT

Cowling, R.M., Lombard, A.T., 2002. Heterogeneity, speciation/extinction history and climate: explaining regional plant diversity patterns in the Cape Floristic Region. Diversity and Distributions. 8, 163-179.

T

Cowling, R.M., Ojeda, F., Lamont, B.B., Rundel, P.W., Lechmere-Oertel, R., 2005. Rainfall reliability, a neglected factor in explaining convergence and divergence of plant traits in fire-prone Mediterranean-climate ecosystems. Global Ecology and Biogeography. 14, 509-519.

IP

Cowling, R.M., Richardson D.M., and Mustart, P.J, 1997. Fynbos: Vegetation of Southern Africa, Cambridge University Press, Cambridge, p. 99-130.

US

CR

Deacon, H.J., 1998. Elandsfontein and Klasies River revisited. In: Ashton, N., Healy, F., Pettitt, P.B. (Eds.), Stone Age Archaeology: Essays in Honour of John Wymer. Oxbow Books. Oxford, England. pp. 23-28. Drennan, M.R., 1953. The Saldanha skull and its associations. Nature. 172, 791-793.

AN

Dupont, L., 2011. Orbital scale vegetation change in Africa. Quaternary Science Reviews. 30, 3589-3602.

M

Ericson, J., 1985. Strontium isotope characterization in the study of prehistoric human-ecology. Journal of Human Evolution. 14, 503-514.

PT

ED

Eze, P.N. and Meadows, M.E., 2014. Multi-proxy palaeosol evidence for late Quaternary (MIS 4) environmental and climate shifts on the coasts of South Africa. Quaternary International. 343, 159-168. Faure, G., 1986. Principles of Isotope Geology. Wiley, New York, 589 pp.

AC

CE

Forrest F.L., Stynder, D.D., Bishop, L.C., Levin, N.E., Lehmann S.B., Patterson, D.B., Matthews, T., Braun D.R., in review, zooarchaeological analysis of newly excavated Middle Pleistocene deposits from Elandsfontein, South Africa. Goldblatt, P., Manning, J.C., 2002. Plant diversity in the Cape Region of Southern Africa. Annals of the Missouri Botanical Garden. 89, 281–302. Hare, V., Sealy, J., 2013. Middle Pleistocene dynamics of southern Africa's winter rainfall zone from 13C and 18O values of Hoedjiespunt faunal enamel. Palaeogeography, Palaeoclimatology, Palaeoecology. 374, 72-80. Hoppe, K.A., Koch, P.L., Furutani, T.T., 2003. Assessing the preservation of biogenic strontium in fossil bones and tooth enamel. International Journal of Osteoarchaeology. 13, 20-28.

22

ACCEPTED MANUSCRIPT

Klein, R. G., 1978. The fauna and overall interpretation of the 'Cutting 10' Acheulean site at Elandsfontein (Hopefield), southwestern Cape Province, South Africa. Quaternary Research. 10, 69-83.

T

Klein, R.G., Avery, G., Cruz-Uribe, K., Steele, T.E., 2007. The mammalian fauna associated with an archaic hominin skullcap at later Acheulean artifacts at Elandsfontein, Western Cape Province, South Africa. Journal of Human Evolution. 52, 164-186.

CR

IP

Klein, R.G., Cruz-Uribe, K., 1991. The bovids from Elandsfontein, South Africa, and their implications for the age, palaeoenvironment, and origins of the site. The African Archaeological Review. 9, 21-79.

US

Kohn, M.J., Cerling, T.E., 2002. Stable isotope compositions of biological apatite. In Kohn, M.J., Rakovan, J., Hughes, J.M., (Eds.). Phosphates: Geochemical, Geobiological and Material Importance. Reviews in Mineralogy and Geochemistry. Mineralogical Society of America, Washington, DC. 48, 455-488.

AN

Lee-Thorp, J.A., van der Merwe, N.J., Brain, C.K., 1989. Isotopic evidence for dietary differences between two extinct baboon species from Swartkrans. Journal of Human Evolution. 18, 183-189.

M

Lee-Thorp, J.A., 2008, On isotopes and old bones. Archaeometry, 50, 925-950.

PT

ED

Lehmann, S.B., Braun, D.R., Dennis, K.J., Patterson, D.B., Stynder, D.D., Bishop, L.C., Forrest, F., Levin, N.E., 2016. Stable isotopic composition of fossil mammal teeth and environmental change in southwestern South Africa during the Pliocene and Pleistocene. Palaeogeography, Palaeoclimatology, Palaeoecology. 457, 396-408.

CE

Lepre, C.J., Roche, H, Kent, D.V., Harmand, S., Quinn, R.L., Brugal, J.P., Texier, P.J., Lenoble, A., and Feibel, C.S., 2011, An earlier origin for the Acheulian. Nature. 477, 82-85.

AC

Luyt, J., Lee-Thorp, J., Avery, G., 2000. New light on Middle Pleistocene west coast environments from Elandsfontein, Western Cape Province, South Africa. South African Journal of Science. 96, 399-403. Mabbutt, J.A., 1956. The physiography and surface geology of the Hopefield fossil site. Transactions of the Royal Society of South Africa. 35, 21-58. Marlow, J., Lange, C., Wefer, G., Rosell-Mele, A., 2000. Upwelling intensification as part of the Pliocene-Pleistocene climate transition. Science. 290, 2288-2291. Mucina, L., Rutherford, M.C., Powrie, L.W., 2006. Vegetation atlas of South Africa, Lesotho and Swaziland. Mucina, L., Rutherford, M.C. (Eds.). The vegetation of South Africa, Lesotho and Swaziland. South African National Biodiversity Institute. Pretoria, South Africa. pp. 748-790.

23

ACCEPTED MANUSCRIPT

Patterson, D.B., Lehmann, S.B., Matthews, T., Levin, N.E., Stynder, D., Braun, D.R., 2016. Stable isotope ecology of Cape dune mole-rats (Bathyergus suillus) from Elandsfontein, South Africa: implications for C4 vegetation and hominin paleoecology in the Cape Floral Region. Palaeogeography, Palaeoclimatology, Palaeoecology. 457, 409-421. Pickering et al., 2016. IGC Abstract

IP

T

Pin, C., Briot, D., Bassin, C., Poitrasson, F., 1994. Concomitant separation of strontium and samarium-neodymium for isotopic analysis in silicate samples, based on specific extraction chromatography. Analytica Chimica Acta. 298, 209-217.

CR

Proches, S., Cowling, R.M., Goldblatt, P., Manning, J.C., Snijman, D.A., 2006. An overview of the Cape geophytes. Biological Journal of the Linnean Society. 87, 27-43.

US

Radloff, F.G.T., Mucina, L., Bond, W.J., le Roux, P. J., 2010. Strontium isotope analyses of large herbivore habitat use in the Cape Fynbos region of South Africa. Oecologia. 164, 567-578.

M

AN

Radloff, F.G.T., 2008. The ecology of large herbivores native to the coastal lowlands of the Fynbos Biome in the Western Cape, South Africa. Doctoral Dissertation, Department of Botany, University of Stellenbosch, South Africa.

ED

Rebelo, A.G., Boucher, C., Helme, N., Mucina, L., and Rutherford, M.C., 2006, Fynbos biome: The vegetation of South Africa, South African National Biodiversity Institute, p. 53-219.

PT

Roberts, D.L., 2006a. Langebaan Formation. In: Johnson, M.R. (Ed.), Catalogue of South African Lithostratigraphic Units, 9-12.

CE

Roberts, D.L., 2006b. Sandveld Group. In: Johnson, M.R. (Ed.), Catalogue of South African Lithostratigraphic Units, 25-26.

AC

Roberts, D.L., Bateman, M.D., Murray-Wallace, C.V., Carr, A.S., Holmes, P.J., 2009. West coast dune plumes: climate driven contrasts in dunefield morphogenesis along the western and southern South African coasts. Palaeogeography, Palaeoclimatology, Palaeoecology. 271, 24-38. Rossouw, L., Stynder, D., Haarhof, P., 2009. Evidence for opal phytolith preservation in the Langebaanweg ’E’ Quarry Varswater Formation and its potential for palaeohabitat reconstruction. South African Journal of Science. 105, 223-227. Rozendaal, A., Gresse, P.G., Scheepers, R. and Le Roux, J.P., 1999. Neoproterozoic to early Cambrian crustal evolution of the Pan-African Saldania Belt, South Africa. Precambrian Research. 97, 303-323.

24

ACCEPTED MANUSCRIPT

Scheepers, R., 1995. Geology, geochemistry and petrogenesis of Late Precambrian S-, I- and A type granitoids in the Saldania Belt, Western Cape Province South Africa. Journal of African Earth Sciences. 2, 35-58. Scheepers, R., Nortje, A.N., 2000. Rhyolitic ignimbrites of the Cape Granite Suite, southwestern Cape Province, South Africa. Journal of African Earth Sciences. 31, 647-656.

IP

T

Scheepers, R., Poujol, M., 2002. U-Pb zircon age of Cape Granite Suite ignimbrites: characteristics of the last phases of the Saldanian magmatism. South African Journal of Geology. 105, 163-178.

CR

Sealy, J., van der Merwe, N., Sillen, A., Kruger, F., Krueger, H., 1991. 87Sr/86Sr as a dietary indicator in modern and archaeological bone. Journal of Archaeological Science. 18, 399-416.

US

Sillen, A., 1986. Biogenic and diagenetic Sr/Ca in Plio-Pleistocene fossils of the Omo Shungura Formation. Paleobiology. 12, 311-323.

AN

Sillen, A., Hall, G., Richardson, S. Armstrong, R., 1998. 87Sr/86Sr ratios in modern and fossil foodwebs of the Sterkfontein Valley: implications for early hominid habitat preference. Geochimica et Cosmochimica Acta. 62, 2463-2473.

M

Sillen, A., LeGeros, R., 1991. Solubility profiles of synthetic apatites and of modern and fossil bones. Journal of Archaeological Science. 18, 385-397.

ED

Singer, R., Wymer, J., 1968. Archaeological investigations at the Saldanha skull site in South Africa. South African Archaeological Bulletin. 23, 63-74.

PT

Sposito, G., 1989. The Chemistry of Soils. Oxford Univ. Press, New York, 277 pp.

CE

Stuut, J.-B.W., Crosta, X., van der Borg, K., Schneider, R.R., 2004. On the relationship between Antarctic sea ice and southwestern African climate during the late Quaternary. Geology. 909-912.

AC

Stynder, D.D., 2009. The diets of ungulates from the hominid fossil-bearing site of Elandsfontein, Western Cape, South Africa. Quaternary Research. 71, 62-70. van Zinderen Bakker, E.M., 1976. The evolution of late Quaternary paleoclimates of Southern Africa. Palaeoecology of Africa. 9, 160-202. Vogel, J.C., Fuls, A., Ellis, R.P., 1978. The geographical distribution of Kranz grasses in South Africa. South African Journal of Science. 74, 209-215.

25

AC

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

Fig. 1: Map of southwestern South Africa outlining the locations of the major substrates in the region and indicating the location of Elandsfontein (black circle). Letters A through H denote the sampling locations for plants, modern bone and enamel collected as part of this study (see Table S1). The range in bioavailable 87Sr/86Sr for each major substrate is reported in the legend and in Table 1. The location of the study region is denoted by a red rectangle on the continent of Africa. The map of major geologic substrates is based on Sealy et al. (1991).

26

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

Fig 2: Scenarios outlining the expected trends of 87Sr/86Sr over a series of supernatants remaining after enamel powder is bathed in weak nitric acid to remove secondary strontium and that of the cleaned enamel. Scenarios A and B are the expected trends in 87Sr/86Sr when there has only been minor alteration of the primary 87Sr/86Sr of a sample. Scenarios C and D (solid black lines) indicate the expected trend in the 87Sr/86Sr when the enamel has not experienced measurable alteration of primary 87Sr/86Sr. Scenario E indicates signifi-cant alteration of a primary enamel 87Sr/86Sr, by which the primary 87Sr/86Sr (reflecting Table Mountain Sandstone) is completely masked by the secondary 87Sr/86Sr (reflecting marine sands).

27

CR

IP

T

ACCEPTED MANUSCRIPT

AC

CE

PT

ED

M

AN

US

Figure 3: Examples of serial sampling tooth enamel (A and B). Arrows indicate youngest (closest to root) to oldest (closest to worn surface) enamel along the tooth growth axis.

28

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

Figure 4: Range of bioavailable 87Sr/86Sr for the major substrates in southwestern South Africa today and the average 87Sr/86Sr for each mid-Pleistocene mammalian tooth from Elandsfontein. The solid black dots represent new bioavailable 87Sr/86Sr data and the gray dots outlined in black represent ratios published by Sealy et al. (1991) and are from bone, enamel and plant samples. The 87 Sr/86Sr for each sample used to determine the range of bioavailable 87Sr/86Sr of substrates can be found in Table S1 and the ranges of these data are presented in Table 1. The 87Sr/86Sr of fossil enamel samples are averaged for a single tooth (i.e., when a tooth was serial sampled) and represent the following mammalian families: Bovidae (◇), Equidae (◻︎), Elephantidae (❍), rodent families Bathyergidae (Δ) and Hystricidae (▽) and feliform family Hyaenidae (✕). The probable dietary behavior of the anamials is indicated by letters and is based on Stynder (2009) and Klein et al. (2007).

29

CE

PT

ED

M

AN

US

CR

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 5: 87Sr/86Sr of the supernatants from a series of rinses of modern and fossil tooth enamel powder bathed in weak nitric acid (gray symbols) and of the remaining insoluble enamel powder (solid black circles). The samples listed below the main graph are described in Table S3. The smaller plot in the upper right hand corner of the figure outlines the range of bioavailable 87Sr/86Sr of the major substrates in southwestern South Africa in comparison with the range of 87Sr/86Sr for the series of supernatants and the remaining powder of enamel samples as presented in Figure 4.

30

ACCEPTED MANUSCRIPT

Table 1. Summary of bioavailable 87Sr/86Sr from bone, enamel, and plants that represent major substrates in southwestern South Africa. a

Pleistocene marine sands

0.710377 ± 0.000786 0.717491 ± 0.004166 0.716704 ± 0.001802 0.718064 ± 0.001632

0.709380 0.711690 0.711421 0.723563 0.714092 0.718546 0.715332 0.720390

IP

Range

US

Cape Granite suite

Average ± Stdev b

CR

Location Substrate

T

87Sr/86Sr

Table Mountain sandstone

AN

Malmesbury shale a

nc

8 12 9 6

ED

M

Data were compiled from the current study (n = 25) and from Sealy et al. (1991) (n = 10). Table S1provides the data and contextual information for each sample included in this compilation. average ± 1 standard deviation

c

number of samples included in the compilation

AC

CE

PT

b

31

ACCEPTED MANUSCRIPT

Tragelaphus strepsiceros Alcelaphini sp. indet. Syncerus antiquus (extinct) Bovid indet. Equus capensis

Elephantidae

Loxodonta africana

Bathyergidae Hystricidae Hyaenidae

Bathyergus suillus Hystrix africaeaustralis Hyaena brunnea

Inclination to migrateb

Greater kudu Giant buffalo Cape zebra African elephant Dune mole rat Porcupine Brown hyena

Browser Grazer Grazer Grazer

Not migratory Potentially migratory Potentially migratory Potentially migratory

Grazer

Potentially migratory

Mixed Mixed Carnivore

Not migratory Not migratory Can travel wide distances

AN

a

Dietary behaviora

T

Bovidae Bovidae Bovidae Bovidae Equidae

Common name

CR

Taxon

US

Mammalian family

IP

Table 2. Diet and migratory behavior of the mid-Pleistocene mammalian herbivores and carnivores from Elandsfontein analyzed for enamel 87Sr/86Sr.

CE

PT

ED

migratory inclination and dietary behavior based on Copeland et al. (2016) and comm. with D.D. Stynder

AC

b

M

dietary behavior based on Stynder (2009) and Klein et al. (2007)

32

ACCEPTED MANUSCRIPT

Table 3. Average 87Sr/86Sr of fossil mammals from mid-Pleistocene Elandsfontein. 87Sr/86Sr

Range

nb

0.709384 ± 0.000031 0.709721 ± 0.000046 0.709688 ± 0.000071 0.709843 ± 0.000058 0.711010 ± 0.000154 0.710293 ± 0.000058 0.709641 ± 0.000241 0.710431 ± 0.000053 0.711047 ± 0.000144 0.710496 ± 0.000142 0.709855 ± 0.000056 0.709876 ± 0.000194 0.709474 0.709640 ± 0.000022 0.710320 ± 0.000433 0.709428 0.710090 ± 0.000080

0.709330 - 0.709428 0.709671 - 0.709780 0.709546 - 0.709754 0.709770 - 0.709980 0.710710 - 0.711120 0.710130 - 0.710570 0.709255 - 0.710078 0.710351 - 0.710488 0.710889 - 0.711171 0.710327 - 0.710835 0.709771 - 0.709888 0.709324 - 0.710141 0.709609 - 0.709658 0.709856 - 0.711105 0.709976 - 0.710181

9 7 10 12 6 12 9 5 3 21 4 12 1 6 12 1 10

0.709705

-

1

Loxodonta africana Loxodonta africana

0.710071 0.710547

-

1 1

Bathyergidae WCRP 32007

Bathyergus suillus

0.709392 ± 0.000083

0.709275 - 0.709489

6

Hystricidae WCRP 32012

Hystrix africaeaustralis

0.709476 ± 0.000066

0.709423 - 0.709587

7

Equus capensis

Elephantidae WCRP 12298 WCRP 6088

IP

Equidae WCRP 2048

CR

Tragelaphus strepsiceros Tragelaphus strepsiceros Tragelaphus strepsiceros (?) Alcelaphini Gen. sp. indet. Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Syncerus antiquus Bovid Gen. sp. indet. Bovid Gen. sp. indet. Bovid Gen. sp. indet.

US

Bovidae WCRP 20443 WCRP 36612 WCRP 9386 WCRP 1358 WCRP 36309 WCRP 9043 WCRP 1666 WCRP 36309 WCRP 8787 WCRP 9046 WCRP 9944 WCRP 46245 WCRP 46251 WCRP 46225 WCRP 32386 WCRP 46257 WCRP 1468

AC

CE

PT

ED

M

AN

Taxon

T

Average ± Stdeva

Sample ID

33

ACCEPTED MANUSCRIPT

Hyaenidae WCRP 32121 WCRP 32442

Hyaena brunnea Hyaena brunnea

0.709393 0.709620

average ± 1 standard deviation

b

number of serial samples averaged from an individual tooth

1 1

AC

CE

PT

ED

M

AN

US

CR

IP

T

a

-

34

ACCEPTED MANUSCRIPT

Highlights Mid-Pleistocene herbivores did not migrate from coastal southwestern South Africa for food



Elandsfontein was fundamentally different from today and supported large herbivores



We provides framework to future studies of bioavailable 87Sr/86Sr the region

AC

CE

PT

ED

M

AN

US

CR

IP

T



35