The Chacma Baboon (Papio Ursinus) Through Time: a Model of Potential Core Habitat Regions During a Glacial–Interglacial Cycle

Evol Ecol DOI 10.1007/s10682-016-9833-8

ORIGINAL PAPER

The chacma baboon (Papio ursinus) through time: a model of potential core habitat regions during a glacial–interglacial cycle

1 2,3 4,5 Olivia M. L. Stone • Andy I. R. Herries • James S. Brink • Shawn W. Laffan1

Received: 10 November 2015 / Accepted: 11 April 2016 Ó Springer International Publishing Switzerland 2016

Abstract The aim of this research is to understand changes in the biogeography of the chacma baboon (Papio ursinus) through time, by modelling potential habitat changes through the last glacial–interglacial cycle (last interglacial, glacial maximum and current conditions). An environmental envelope model in a geographic information system is used to produce a range of habitat distribution models for the chacma baboon. Initially it is modelled as a single taxon, following which the data are further divided and explored to model predicted habitats for the chacma clades during different stages of the glacial– interglacial cycle. An area of approximately 1,044,000 km2 was identiﬁed as potentially having been within the environmental core habitat at some stage of the glacial–interglacial cycle. Of this, 63,700 km2 of land was predicted to have been core habitat regardless of the stage of the glacial–interglacial cycle. Additionally, rainfall appears to be the environmental variable with the most limiting effect on habitat size. This is true for the current, last glacial maximum and last interglacial environmental conditions. The largest area that remains within the core habitat throughout the last glacial–interglacial cycle is found in the northern provinces of South Africa. The chacma clades appear to interact in a cyclic pattern of expansion and contraction, each clade being more prominent under different environmental conditions. It appears that grey footed chacma habitat periodically extends

Electronic supplementary material The online version of this article (doi:10.1007/s10682-016-9833-8) contains supplementary material, which is available to authorized users.

& Olivia M. L. Stone [email protected]

1 School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW 2052, Australia 2 Palaeoscience Laboratories, Department of Archaeology and History, La Trobe University, Melbourne Campus, Bundoora, Melbourne, VIC 3086, Australia 3 Centre for Anthropological Research, University of Johannesburg, Gauteng, South Africa 4 Florisbad Quaternary Research Department, National Museum, Bloemfontein, South Africa 5 Centre for Environmental Management, University of the Free State, Bloemfontein, South Africa 123 Evol Ecol northward towards the yellow baboon. These ﬁndings suggest dynamic and varied progression of the chacma baboon palaeo-distributions.

Keywords Papio ursinus Á Palaeobiogeography Á GIS Á Biogeography Á Core habitat Á Refugia

Introduction

Quantifying changes in spatial distribution through time allows us to infer estimates of population expansion, contraction and interactions, and allows us to estimate current population trends and/or patterns (Stigall and Lieberman 2006; Stigall Rode 2005). These predicted species palaeo-distributions and palaeo-habitats provide information otherwise not available. They can lead to estimates of diversiﬁcation and speciation or suggest dispersal as habitats shift (Stigall and Lieberman 2006). Ultimately, ancient species distributions and habitat predictions can help us understand species dynamics through time and how distributions become what we see today. They can even allow us to track potential changes resulting from the effects of time or climate change and more fully assess palaeobiogeography of particular species more effectively than simple fossil distribution maps. Geographic information system (GIS) modelling, including environmental envelope models (EEMs), can be an effective way to assess habitat and by inference species distributions in differing environmental conditions (Chatterjee et al. 2012; Hijmans and Graham 2006). The repeatability of such methodologies means it is easy to integrate new data as it becomes available. This is valuable in ﬁelds such as palaoebiogeography, where knowledge is constantly expanding and evolving. GIS has been used in a wide range of research in primatology, ranging from small localised analyses (Guy et al. 2012; Hoffman and O’Riain 2011; Ostro et al. 1999) to larger national or continental scales (Chatterjee et al. 2012; Junker et al. 2012; Sleeman 2005). GIS based modelling has already been successfully used to model and assess primate biogeography and palaeobiogeography (Chatterjee et al. 2012; Stone et al. 2012, 2013). The ability to assess species dynamics through time can be invaluable when trying to predict the effects of and prepare for future environmental and climatic change. However, palaeobiogeographic predictions are fraught with problems. Fossil data are often too sparsely distributed to reliably calibrate predictive models, leaving modern data as the best available source. In the case of fossil Papio, specimens have been recovered from a range of fossil sites in southern Africa (Delson 1984; Freedman 1970; Gilbert 2013). However, the majority of these have come from a restricted area of karst (480 km2) near Johan- nesburg, South Africa and satellite sites near Taung (Buxton-Norlim Limeworks) and at Makapansgat near Mokopane (designated as the UNESCO World Heritage ‘Fossil Hominid Sites of South Africa’) (Delson 1984; Freedman 1976; Jablonski and Frost 2010). There is a major geological bias evident in southern Africa, with other early fossil primates also being found in caves in restricted karst areas in northern Namibia, southern Angola and western Botswana (Pickford et al. 1994). Thus data for the inhabited palaeo landscapes and environments is similarly restricted and remains largely unknown beyond basic models from faunal data (Herries et al. 2010). The limited data that exists for the Plio-Pleistocene suggests increasing aridity from *2 million years ago in southern Africa (Doran et al.

123 Evol Ecol

2015; Dupont et al. 2005; Hopley et al. 2007), and major faunal turnover (Herries et al. 2010). However, detailed palaeoenvironmental records do not exist for the region at this time. Moreover, over the last 4.0–1.5 million years, since the fossils were deposited, significant erosion has taken place (Dirks et al. 2010). Thus the paucity of fossil data means predictions based on fossils alone are problematic. This research aims to predict the potential distribution of fossil southern African Papio by extrapolating suitable past environmental conditions (thus suitable habitat) derived from current and historic chacma baboon (Papio ursinus) observation data. The modern chacma baboon is the best available analogue for southern African Papio, including its extinct lineages. This is because the Papio genus is believed to have a southern African origin (Newman et al. 2004; Zinner et al. 2009) and the chacma baboon is likely to be the oldest clade within Papio (Newman et al. 2004; Sithaldeen et al. 2009). As shown by the fossil record, the chacma baboon has evolved as a distinct taxon within the geographic region of its ancestral populations (Jablonski and Frost 2010). This, in addition to the likelihood of shared evolutionary traits, means that both the current and ancestral populations are most likely to survive within similar environmental conditions. Understandably, there are numerous issues with using modern population data to predict an ancient biogeography. These include: 1. The improbability that two species, separated by geological time, will have identical niche requirements. Assemblages attributed to a single species of Papio (P. angusticeps) have been shown to have wide ranging isotopic values (Adams et al. 2013). This suggests niche separation due to time, or one species with wide ranging diets across C3 and C4 environments, or two separate species that have not previously been identified. 2. The probability that current primate distributions will have been affected by anthropogenic alteration of the landscape. 3. Past environmental conditions are, at best, only estimates. Such issues must be considered, however, as the extinct species were within the same geographic area, it is also highly unlikely that the descendant species will have evolved to require and inhabit an entirely different environment. More likely, there would be some form of niche differentiation, such as variation in habitat types (i.e. forest cover, savannah etc.) within similar environmental bounds. Therefore, we would expect to observe some degree of habitat overlap. We do not know the direction of any likely shifts, expansions or reductions that would have occurred to this habitat, but the central part of the distribution is likely to be the least affected. For this reason the central niche, used to predict the environmental core habitat, becomes of particular interest. The central niche is the environmental zone within an animal’s environmental niche that encompasses the central 50 % of observations for each environmental variable that affects that specific animal’s distribution (Stone et al. 2015). For example, if rainfall was deemed influential then it would include the areas that fell within the central part of that species’ rainfall distribution. The central niche is the most moderate part of an animal’s environmental range; as such, it affords less environmental stress and is likely to be stable given changes in the environmental limits between extinct and extant populations. Such environmental core habitat areas have the best potential to form refugia during climatic fluctuations which support the persistence of species and subspecies through time (Rosauer et al. 2015). The modern location data used in this process may have been subject to anthropogenic influence. However, the environmental limits within which the species can survive will not 123 Evol Ecol be affected. These limits can be found using modern habitat and historical data and can be applied to different environmental scenarios to predict areas of likely habitat in different climatic environments. This research focuses on three different climatic environments, the last interglacial (LIG) [130–116 ka (Kukla et al. 2002)], the last glacial maximum (LGM) [26.5–19 ka (Clark et al. 2009)] and the current interglacial Holocene conditions (HOL) [11.7 ka– present (Walker et al. 2009)]. By using two extremes and an intermediate phase (HOL) of a glacial–interglacial cycle in our model, we hope to predict the extremes of probable changes in habitat distribution though time. Areas that are predicted to have remained within the central niche (environmental core habitat) throughout the last glacial cycle may be areas that are of heightened importance to the species. These areas may have provided source populations during times of habitat expansion, or refugia during habitat recession. We use an EEM to define and explore the central niche for the HOL, LIG and the LGM. To further understand the dynamics of the chacma baboon during the last glacial–interglacial cycle, we also modelled the potential habitat changes through time for chacma baboon clades. This allowed us to investigate possible interactions between clades and individual clade dynamics as well as consider the effects of future environmental and climatic changes separately for the different populations. This study explores palaeobiogeography for the chacma baboon at both the species and subspecies level. The taxonomy within Papio is constantly changing, as are the taxonomic definitions within P. ursinus. The IUCN/SSC Primate Specialist Group (Grubb et al. 2003) identified two chacma clades, P. ursinus ursinus (the southern chacma, this included the P. ruacana population) and P. ursinus gripseises (the grey footed chacma). Later Groves (2005) acknowledged P. ursinus ruacana, raising it to equal footing with P. ursinus ursinus and P. ursinus gripseises. Although the taxonomic levels and even precise definitions of these populations are still somewhat dynamic and debated, during the previous few glacial–interglacial cycles, three distinct populations have existed. Thus this manuscript divides the chacma baboon into three separate clades.

Materials and methods

Our initial model required central niche models to be predicted for three different environmental conditions (LIG, LGM and HOL) for the entire chacma baboon (P. ursinus) species. These predictions are achieved by identifying the range inhabited by the central 50 % of the population (half the location data), for each environmental variable that inﬂu- ences the chacma baboon distribution. When modeling the chacma baboon in HOL we used the central niche prediction from Stone et al. (2015). Then, following the same methodology (using annual rainfall, maximum temperature of the warmest month and minimum temperature of the coldest month as the environmental variables) we formed two further models for the LIG and LGM. Our models were formed using the best available predicted environmental data. The environmental data sets utilised were extracted from the Worldclim data set for climatic variables (Hijmans et al. 2005). The climate data used for the LIG and LGM are from the Worldclim past climate reconstructions, which in turn are derived from the Paleoclimate Modelling Intercomparison Project Phase II climatic variables (Braconnot et al. 2007), and the climatic variables for the last interglacial (Otto-Bliesner et al. 2006). Environmental data were collected by sampling 459 locations and the adjacent surrounding areas (each datum was enclosed by a 2.5 km buffer) (Fig. 1). Location data

123 Evol Ecol

Fig. 1 Sample locations used to extract the environmental variable data. These were used to calibrate the model parameters for the central niche consisted of modern and historic chacma baboon sightings, including locations from the published literature. These location data were thinned to a more even spatial density for sampling, to reduce any bias to sites with high data density (see Stone et al. 2013) This provided us with the total range of environmental conditions inhabited by chacma baboons. The central niche models identify land areas that satisfy certain criteria. These criteria (a range of values) are set for each individual environmental variable, and locations that fall within the range are selected. The criteria (or parameters) that limit the model to select the conditions experienced by the central 50 % of the population, were taken from Stone et al. (2015) (rainfall: 409–788 mm; maximum temperature: 25.6–31.5 °C; minimum temperature: 0.0–5.9 °C). The selected locations are then combined and the result depicts the geographic areas that are within the central 50 % of the range for all the environmental variables. These resulting areas are deﬁned as the predicted central niche and are the environmental core habitat. This is then repeated for each environmental scenario (LIG, LGM and HOL). The core habitat predictions for the LGM and LIG were then summed together with the core habitat for the HOL. This produced a single map detailing the environmental core habitat that persisted during the maximum, minimum and interglacial conditions. This is the baboon central niche glacial–interglacial cycle model. To gain a clearer understanding of population dynamics, including potential areas for dispersal corridors, we also modelled the 95 % environmental range. The 95 % habitat distributions for the LIG, LGM and HOL were likely to be inhabited by the majority of the 123 Evol Ecol baboon population. Despite not being within the more moderate core habitat regions, these were potentially inhabitable areas and had the potential to act as migration corridors. This is especially so if some environmental variables were within the central 50 % bounds. We deﬁned preferred migration corridors as the areas where two out of three variables were within the central 50 % and the third variable was within the central 95 % of the range. The parameters for this process and the model for current (HOL) conditions were taken from Stone et al. (2013). The northern boundary of the study region was set to 1° South to encompass the entirety of southern Africa and the entire process was completed using ArcGIS version 10. Parameter data were analysed in R 2.15 and Microsoft Excel 2012. The glacial–interglacial model is based on data sampled using baboon locations in the current climatic conditions. To assess if the model was representative of Papio occurring during the LIG and LGM, southern African fossil Papio and Parapapio data, (including a 2.5 km buffer), were overlaid onto and compared with the glacial–interglacial cycle model. To further explore the potential palaeobiogeography of the chacma baboon, location data were assigned by best estimate into subgroups (Fig. 2). Initially the data were divided into southern chacma (n = 403) and northern grey footed chacma (n = 56). However, the resulting predictions appeared more congruent with our current understanding of chacma baboon distribution when the three separate populations were acknowledged. This is potentially due to the very different conditions inhabited by the western population. Thus, the southern clade was divided into the Cape chacma (n = 356) and the Ruacana baboon

Fig. 2 Raw location data for the chacma baboon genetic clades, overlaid onto a ﬁgure estimating the distribution of the chacma baboon morphotypes, adapted from Sithaldeen et al. (2009) 123 Evol Ecol

(n = 47). This division followed the work of Sithaldeen et al. (2009) which demonstrated a divergence resulting in the grey footed chacma in the north and a southern clade at approximately 1.84 Ma and the emergence of the Ruacana baboon as a monophyletic group at approximately 0.68 Ma. We used location data that had been previously been assigned to different mtDNA clades (Sithaldeen et al. 2009) to guide this process. Unlike the modelling process for the entire chacma species, the clade level models also included altitude due to its regional scale effects. Thus the modelling methods are the same as those above but with the addition of altitude. The altitude ranges for the clades were calculated from the 3 arc second resolution NASA Shuttle Radar Topography Mission Digital Elevation Model (SRTM DEM) (Jarvis et al. 2006) using a 2.5 km buffer surrounding the location points. The data were analysed using R software, version 3.0.2. New parameters [that were based on a summary of the variable data extracted from the buffered locations allocated to each clade (Fig. 2)], were used to identify the central 50 % (Table 1). This produced a central niche model for all the clades [grey footed, cape, Ruacana and the combined (cape and Ruacana together) southern clade], during the LIG, LGM, and HOL. In addition we measured the total area of each central niche (environmental core habitat) prediction to compare sizes during the different stages of the glacial–interglacial cycle and form a comparison between the clades.

Table 1 The parameters used to deﬁne the model for the genetic clades within the chacma baboon Environmental variable Units Central Niche parameters Habitat parameters

Lower Upper Lower Upper

Clades Southern clade Annual rainfall mm/year 389 783 89 1055 Maximum temperature °C 25.3 30.8 21.9 34.4 Minimum temperature °C -0.2 5.3 -3.6 10.3 Altitude m 815 1608 51 2276 Northern clade (grey footed) Annual rainfall mm/year 458 821 324 1288 Maximum temperature °C 30.8 34.6 21.9 36.6 Minimum temperature °C 5.1 8.2 2.6 13.5 Altitude m 788 1301 37 1933 Southern clade lineages Cape Annual rainfall mm/year 463 810 201 1063 Maximum temperature °C 24.9 30.1 21.7 34.2 Minimum temperature °C -0.5 4.7 -3.9 9.3 Altitude m 856 1618 37 2325 Ruacana Annual rainfall mm/year 108 322 27 439 Maximum temperature °C 29.4 32.9 24.4 35.3 Minimum temperature °C 4.0 8.4 0.8 10.9 Altitude m 736 1527 245 1984

123 Evol Ecol

Results

Entire chacma baboon species

The baboon central niche modelled through time is a non-uniform patchy distribution stretching across southern Angola through to central Zimbabwe, south through central- eastern South Africa to more southern areas of the Eastern and Western Cape provinces. This area includes approximately 1,044,000 km2 of land. Approximately 6 % (*63,700 km2) of this is predicted as persistent core habitat throughout the glacial–interglacial cycle (Fig. 3), with the largest area of this located *60 km north of Johan- nesburg, in northern South Africa. The largest amount of core habitat is predicted during the LGM with 684,000 km2 of land. Our data suggests this results from wetter conditions in Botswana coupled with generally cooler temperatures. At 47 % of the land area of the LGM, the smallest amount of predicted environmental core habitat is found in the LIG, totalling 324,000 km2. As current Holocene environmental conditions fall somewhere between these two scenarios, the 389,000 km2 of predicted habitat is unsurprising. The fact that the current area prediction is closer in size to the LIG is equally unsurprising as both are interglacial periods, although unlike the LIG period the HOL period is post the warmest part of the interglacial. Annual rainfall appears to be the factor most limiting predicted central niche (environmental core habitat) in all three climatic scenarios (Fig. 4). After rainfall is considered, the increase in predicted core habitat during the LGM appears to result from the substantial increase in land satisfying the maximum temperature bounds. During the LGM there was an 89 % increase (from current conditions) in land satisfying the maximum temperature parameters. Notably, there was also a 45 % increase in land satisfying the minimum temperature bounds. During the LGM, the majority (62 %) of the predicted core habitat occurred in Angola and Botswana. This was markedly different to the LIG and HOL where the majority of land (72 and 78 %, respectively) was further east within South Africa and Zimbabwe. Addi- tionally, Namibia, Zambia, Lesotho, Swaziland, Mozambique and the Democratic Republic of the Congo contained predicted core habitat at some stage during the glacial– interglacial cycle. South Africa, Zimbabwe, Botswana and Namibia contained the only areas predicted as core habitat in all three environmental scenarios. The majority of this was in central eastern southern Africa, with 86 % of this within South Africa.

Chacma clades

In support of the southern chacma baboons’ preference for more montane habitat (see ‘‘Discussion’’ section), both the altitude ranges for the two southern chacma clades (the Cape chacma and the Ruacana) were found to be signiﬁcantly different to the altitude range of the more northern clade [Mann–Whitney–Wilcoxon test; W = 13,206, p \ 9.273e-5 (Cape chacma, n = 651,400; grey footed, n = 108,850) and W = 1623.5, p = 0.042 (Ruacana, n = 101,224; grey footed, n = 108,850)]. In current HOL conditions the predicted central niche for the combined southern chacma clade (including both the Cape chacma and the Ruacana chacma) is found predominantly in central eastern southern Africa (Fig. 5). Although this core habitat is predicted in the South African coastal provinces of Western Cape, Eastern Cape and

123 Evol Ecol

Fig. 3 a Chacma baboon central niche through a glacial–interglacial cycle. b Predicted current central niche adapted from Stone et al. (2015). c The predicted central niche during the last glacial maximum. d The predicted central niche during the last interglacial

KwaZulu-Natal, the majority is in the northern provinces of South Africa. Sizable pockets are noted in Botswana and Zimbabwe. During the LIG there appears to be a subtle eastern shift in the more southern predicted areas with slight north western shift in the northern regions. There is limited predicted habitat in Botswana, and small pockets are now predicted in Zambia and Angola with the habitat in Namibia increasing in size. The LGM shows the largest area of predicted core habitat with the majority of this land found centrally, within Botswana. Large amounts of habitat are predicted in Zimbabwe, Angola and Namibia with South Africa’s portion becoming smaller than in other time periods. This

123 Evol Ecol

Fig. 4 The areas that satisfy the annual rainfall, minimum temperature and maximum temperature parameters overlaid with the predicted central niche predicted central niche for the combined southern clade did not appear to ﬁt as well with known Ruacana locations, emphasising the need for separate analyses for the Cape and Ruacana clades. The grey footed chacma has the largest amount of core habitat in current conditions. This is predicted to extend from southern Angola and northern Namibia in the west, and eastward to eastern Zimbabwe. This area appears to recede west during the LIG to western Zimbabwe, and then reduce considerably to a very small and isolated pocket predicted in the south of the Democratic Republic of the Congo in the LGM (Fig. 5). The Cape baboon core habitat (Fig. 5), much like the habitat for the greater southern clade, has a predominantly eastern distribution in both the current conditions and the LIG. This is primarily located in South Africa with small regions present in Botswana and Zimbabwe in current conditions and Zimbabwe and Zambia in the LIG. It appears that the habitat for the Cape baboon is most limited in warmer conditions, as it shows a northwest 123 Evol Ecol

Fig. 5 The predicted central niche for the grey footed chacma, the southern chacma, the Cape chacma and the Ruacana baboon, during the current conditions, the last glacial maximum (LGM) and the last interglacial (LIG) expansion during the coldest period (LGM), stretching to western Angola and a large expanse of habitat in northern Botswana. The habitat predicted for the Ruacana clade (Fig. 5) is markedly different in that it appears largest during the LIG. The core habitat extends eastward into the northern most tip of South Africa with the majority of land located in Botswana. This appears to result

123 Evol Ecol from decreased rainfall and warmer minimum temperatures. Under current conditions and in the LGM the core habitat is mostly restricted to a narrow north–south strip along the west coast. In current conditions this fragmented habitat extends from northern Namibia into southern South Africa. During the LGM the habitat becomes more connected and is located further north extending from Northern Namibia to southern Angola. It appears that the environmental core habitat of the three clades (grey footed, Cape and Ruacana) expand and contract in a reciprocal pattern. When one clade expands, another contracts, then the third clade expands and the ﬁrst clade retracts. This appears to be a cyclic pattern with each clade expanding, retracting and reaching an intermediate state under different environmental scenarios. The clade that achieves the greatest predicted environmental core habitat area is the grey footed chacma in current conditions. Their predicted habitat is approximately 445,000 km2, an increase from its smallest size of only 1950 km2 during the LGM. The Cape baboon core habitat is predicted to be largest, at 364,000 km2, during the LGM reducing by more than half to its smallest size of 150,000 km2 during the LIG. Alternatively the Ruacana baboon experiences its largest expansion during the LIG with an estimated core habitat covering 167,000 km2 whilst its smallest predicted area of 42,000 km2 occurs in current HOL conditions.

Fossil Papio

Figure 6 depicts the glacial–interglacial cycle model overlaid with fossil Papio locations (Table 2) from southern Africa. Of the 41 fossil Papio sites identiﬁed for the analyses, 41 % were predicted within core habitat at some stage of the glacial interglacial cycle. 80 % were found within core habitat and possible dispersal corridors combined (Fig. 6). Of most interest, 32 % (13 sites) were within 20 km of land that was predicted to be per- sistently core habitat, and 98 % (40 sites) were within 20 km of areas predicted to be preferred migration corridors. When the predicted core habitat was compared with fossil Papio locations from the last approximately 125,000 years (n = 27), 19 % were within the core and 74 % were within the combined environmental core habitat dispersal corridor area. In HOL conditions, 39 % of all fossil locations were within the core region, rising to 73 % with the addition of possible corridors. LIG conditions predicted that the core habitat contained 32 %, increasing to 61 % when combined with the corridors. Nonetheless, during the LGM only 2 % of fossil locations were within the predicted environmental core habitat. This is low, although 61 % were within the area containing the combined core habitat and predicted corridors.

Discussion

Modelling

The modelling process is limited by the available data. Nonetheless, we can combine current knowledge and data to provide the best available estimate at the time. Although the current Holocene environmental conditions are not at the highpoint of the cycle [they are cooling from an interglacial maximum, the Holocene Climatic Optimum some 7–5 kya (Kaufman et al. 2004)], similar conditions would have occurred at various stages during the numerous glacial–interglacial cycles of the Quaternary (DeMenocal 1995). While we are limited by the

123 Evol Ecol

Fig. 6 The central niche and potential dispersal corridors overlaid with southern African Papio fossil sites. a Depicted through a glacial–interglacial cycle. b Under current conditions. c During the last glacial maximum. d During the last interglacial availability of reliable datasets representing the last 125 ka, much of this period would be represented by environmental conditions that would be intermediate between the LIG and LGM. In fact, the last glacial–interglacial cycle shows more extreme conditions than most of the previous cycles for the Quaternary (DeMenocal 1995). Since predicted environmental data is available for the LIG and LGM, we can assess the predicted extremes of the last glacial–interglacial cycle and theorise that these are likely to represent some of the more extreme conditions that have occurred during the Quaternary. The majority of climatic conditions during this time will also be somewhat intermediate to those model extremes.

123 123 Table 2 Southern African excavation sites known to contain Papio fossils Fossil Papio Approximate age Max temp. (°C) Min temp. (°C) Annual rainfall (mm) Speciesa References site Current LGM LIG Current LGM LIG Current LGM LIG

Langebaanweg 5.2–5.0 Ma 28 26.1 26.1 8.3 6.4 7.5 274 248 230 Parapapio sp. Grine and Hendey (1981) and Roberts et al. (2011) Bolts Farm Faunally dated 27.6 26.7 27.2 1.2 -0.9 0.7 691 726 564 Parapapio Freedman (1976), Jablonski (2002), from the mid to broomi; Heaton (2006), Jablonski and Frost late Pliocene to Parapapio (2010) and Gilbert et al. (2015) early whitei; Papio Pleistocene. robinsoni Multiple sites Makapansgat 3.0–2.6 Ma 26.6 21.9 25.3 3.4 0.5 3.6 599 665 485 Parapapio jonesi; Freedman (1976), McKee (1995), Limeworks Parapapio Jablonski and Frost (2010) and Members 3 broomi; Herries et al. (2013) and 4 Parapapio whitei Taung type 3.0–2.0 Ma 32.7 30.4 32.1 1 -1.6 0 447 362 311 Parapapio McKee (1993, 1995), Gilbert (2007), site broomi; Papio Jablonski (2002), Jablonski and izodi; Frost (2010), Herries et al. (2013) Parapapio and Hopley et al. (2013) jonesi Sterkfontein 27.8 26.7 27.4 1.3 -0.9 0.7 682 726 556 Member 4, M4 2.6–2.0 Ma Papio izodi; Broom and Schepers (1946), McKee Member 2 Parapapio (1995), Herries et al. (2010), and jonesi; Herries and Shaw (2011), Jablonski Jacovec Parapapio (2002), Kibii (2007), Jablonski and Cavern broomi; Frost (2010), Pickering et al. (2010) Parapapio and Reynolds and Kibii (2011) whitei vlEcol Evol Table 2 continued Ecol Evol

Fossil Papio Approximate age Max temp. (°C) Min temp. (°C) Annual rainfall (mm) Speciesa References site Current LGM LIG Current LGM LIG Current LGM LIG

Member 5 M5 1.8–1.1 Ma Papio robinsoni Broom and Schepers (1946), Herries et al. (2010), Herries and Shaw (2011), Jablonski (2002), Jablonski and Frost (2010), Pickering et al. (2010) and Reynolds and Kibii (2011) Koanaka Not dated, Plio- 34 29.3 31.2 5.6 2.2 5.9 427 455 331 Parapapio sp. Pickford (1990, Pers. Comm.) North Pleistocene Bone Cave C317 ± 114 ka Papio botswanae Williams et al. (2012) Minaars Cave, Not dated, early 27.9 26.8 27.5 1.5 -0.7 0.9 683 725 558 Papio izodi Brain (1981), Freedman (1970) and Hadeco Pleistocene Gommery et al. (2012) Gladysvale Int. 2.4 Ma–780 28.1 27 27.7 1.6 -0.6 1 683 719 558 Papio cf. izodi; Berger et al. (1993), Herries et al. ka Papio cf. (2013) and Lacruz et al. (2002) Ext. 780–578 ka robinsoni; Papio angusticeps Haasgat 2.3–1.8 Ma 27.4 26.7 27 2 -0.2 1.5 719 737 595 Papio angusticeps Adams (2012) Malapa *1.98 Ma 28 26.6 27.6 1.6 -0.4 1 683 740 558 Papio angusticeps Gilbert et al. (2015) Swartkrans M1 2.0–1.8 Ma 27.8 26.7 27.4 1.2 -0.9 0.7 680 726 554 Papio robinsoni; Broom and Schepers (1946), de M2 1.7–1.1 Ma Parapapio Ruiter (2003), Freedman (1957), M3 1.3–0.6 Ma jonesi McKee (1995), Herries and Adams (2013), Herries et al. (2009), Jablonski and Frost (2010), Pickering et al. (2011) and Gilbert et al. (2015)

123 Drimolen 2.0–1.4 Ma 27.2 26.8 26.8 1.7 -0.7 1.2 717 725 590 Papio robinsoni Herries and Adams (2013), Keyser Main Quarry et al. (2000) and Gilbert et al. (2015) 123 Table 2 continued

Fossil Papio Approximate age Max temp. (°C) Min temp. (°C) Annual rainfall (mm) Speciesa References site Current LGM LIG Current LGM LIG Current LGM LIG

Kromdraai A 1.8–1.6 Ma (KB) 27.6 26.7 27.2 1.4 -0.8 0.8 694 733 568 Parapapio jonesi; Brain (1981), Broom and Schepers and B KA not dated Papio izodi; (1946), Freedman (1957), McKee Papio (1995), Herries et al. (2009), angusticeps; Jablonski and Frost (2010) and Papio robinsoni Gilbert (2013) Coopers D 1.6–\1.4 Ma 27.8 26.7 27.4 1.4 -0.9 0.8 683 726 558 Papio robinsoni; de Ruiter et al. (2009) and Gilbert (coopers Papio (2013) other) angusticeps Plovers Lake 27.9 26.8 27.5 1.4 -0.7 0.8 681 725 557 Ext. *1Ma Papio robinsoni Thackeray and Watson (1994) Int. 89–63 ka Papio ursinus de Ruiter et al. (2008) Cave of 780 ka to present 26.1 21.9 24.8 3.2 0.5 3.4 620 665 505 Papio sp. McKee (1995) and Herries (2011) Hearths Border Cave Mid-Late 27 27.7 26.9 7.6 6.4 7 864 677 748 Papio ursinus Klein (1977) Pleistocene *200 ka to present Klasies River 110–40 ka 23.7 23.3 22.8 8.8 4.4 8.4 845 687 777 Papio ursinus Klein (1976) and McKee (1995) Mouth 5 ka to present Herolds Bay *80 ka 25.1 23.2 24.3 8.4 4.8 7.9 573 517 519 Papio ursinus Brink and Deacon (1982) Cave Boomplaas 70–2 ka 28.3 26.8 26.2 3 0.5 2.5 363 290 299 Papio ursinus Klein (1978b) Cave Blombos Cave *70 ka 27.3 25.1 25.3 7.1 3.8 6.5 511 452 475 Papio ursinus Henshilwood et al. (2001) and Jacobs et al. (2006) Equus Cave \30 ka 32.9 30.6 32.2 0.9 -1.8 -0.1 446 369 316 Papio ursinus Johnson et al. (1997) and Klein et al. Ecol Evol (1991) Table 2 continued Ecol Evol

Fossil Papio Approximate age Max temp. (°C) Min temp. (°C) Annual rainfall (mm) Speciesa References site Current LGM LIG Current LGM LIG Current LGM LIG

Buffelskloof HE 22–8 ka 30.2 28.2 28.3 3.6 0.8 3 232 190 173 Papio ursinus Klein (1978a) Rock Shelter MDA 8–3 ka BOL 5–3 ka Besaansklip 20–16 ka 26.3 24.5 24.4 8.8 6.7 8 304 264 260 Papio ursinus Brink and Gru¨n Pers. Comm. Rose Cottage 18–\2 ka 28.2 27.4 27.3 -0.8 -3.2 -1.6 689 626 557 Papio ursinus Plug and Engela (1992) Cave Nelson Bay 18–5 ka 27.3 25.1 25.3 7.1 3.8 6.5 511 452 475 Papio ursinus Klein (1972) Cave Nooitgedacht *16–15.3 ka 29.9 26.8 27.6 3.7 0.5 3.2 336 290 273 Papio ursinus Brink (1999) and Brink and Herries No. 3 Pers. Comm. Bushman’s Level 30.2 28.4 28.4 -4.2 -6.9 -4.9 346 290 215 Papio ursinus Plug (1981) Rock Shelter 16–12.5 Ka Level 8–10 Ka Nooitgedacht *15.3–14 ka 29.9 26.8 27.6 3.7 0.5 3.2 336 290 273 Papio ursinus Brink (1999) and Brink Pers. Comm. No. 1 Die Kelders Holocene and 23.1 21.1 20.8 9.2 7 8.5 543 470 516 Papio ursinus Klein and Cruz-Uribe (2000) Cave 80–50 ka Wonder Cave Holocene 27.5 26.8 27.1 1.6 -0.7 1.1 702 725 577 Papio ursinus Brain (1981) and Herries Pers. Comm. Byeneskrans Holocene 23.4 21.3 21.2 9 6.8 8.3 530 460 505 Papio ursinus Klein (1981) Powerhouse 4–2 ka 32.4 29.7 31.8 0.7 -1.9 -0.3 459 381 322 Papio ursinus Klein (1979) Cave Lithakong 5–0.5 ka 22.3 23.6 21.8 -3.5 -4.5 -4.4 766 658 636 Papio ursinus Kaplan and Mitchell (2012) and Brink (2012) 123 Blydefontein Level III *2 ka 27.3 25.5 25 -3.8 -6.1 -4.4 471 399 339 Papio ursinus Klein (1979) Rock Shelter Witkrans Cave Upper 32.3 29.8 31.7 0.7 -1.8 -0.3 460 375 321 Papio ursinus Freedman (1965) Pleistocene 123 Table 2 continued

Fossil Papio Approximate age Max temp. (°C) Min temp. (°C) Annual rainfall (mm) Speciesa References site Current LGM LIG Current LGM LIG Current LGM LIG

Dikbosch \2 ka 33.6 31.1 32.6 0.8 -2.1 -0.2 371 308 234 Papio ursinus Klein (1979) Rock Shelter Doornfontein \2 ka 31.5 28.9 30.7 1 -1.5 0.1 369 302 238 Papio ursinus Klein (1979) Glen Elliott Level IV \2 ka 31.8 30.1 29.2 0.9 -1.7 0.2 415 355 279 Papio ursinus Klein (1979) Rock Shelter Riversmead Level II \2 ka 31.4 30 29.1 1.1 -1.2 0.3 423 358 287 Papio ursinus Klein (1979) Rock Shelter Abbott’s Cave 0.3–0.7 ka 29.2 27 26.7 -1.2 -3.4 -2 385 319 259 Papio ursinus Plug (1993) a We acknowledge there are varying views on fossil baboon taxonomy. These species designations follow that of Gilbert (2013), Gilbert et al. (2015) and Butynski et al. (2013) vlEcol Evol Evol Ecol

Chacma clade location data

There are always areas of overlap in both genetics and morphology when defining the populations. This is inevitable in any climatically changing environment, as population expansions, regressions and introgressions are natural occurrences. We accept that as more detailed genetic data becomes available, the areas we used to define these three chacma clades may become more defined, however the overall predictions will not be greatly affected by changes in fringe data.

Model validation with fossil Papio

Only 2 % of fossil sites were within the predicted environmental core habitat during the LGM. It may be that the current population (from which the data to produce the models were sourced) is living within conditions different enough to affect predictions in the past. It could also represent a limitation in the fossil data, as only 41 sites were identified for the data set. Finding 2 % of sites within the predicted habitat during the LGM may suggest a sampling bias. It is possible that a high percentage of fossil sites in the region were formed during interglacial or at least warmer glacial conditions, and this could imply a larger or denser population at these times or a longer time period for deposition. This is in addition to the already known and potentially large geological bias that occurs in the fossil record due to the fact Papio fossils have been primarily recovered from restricted karst landscapes and that this bias increases further back in time (Herries et al. 2010; Herries et al. in review). The future identification of many more fossil sites to include in the model will help clarify these possibilities. There is no fossil evidence for baboons in the South African central interior (Fig. 6) and modern location data in this area is sparse (Fig. 1). The paucity of such data is unsurprising as, despite having sufficient rainfall, the area is excluded from core habitat due to the temperature variables and the grassland biome offers little cover and minimal sleeping sites. The dating and or palaeoreconstruction that has been attempted to some degree at most fossil Papio/Parapapio sites (as can be seen in Table 2), is too general to allow for statistical or GIS analysis between locations (see preliminary divisions into chronological categories [Holocene, Pleistocene, Pliocene sites with fossils that may have occurred during the LGM and sites with fossils that may have occurred during the LIG] in Supp. Fig. 1). We cannot guarantee the accuracy of the dating, nor can we confirm that all sites that fit into each category are included. Similarly, uncertainty with taxonomy and paucity of data inhibits analyses of taxonomic patterns. Preliminary taxonomic divisions can be seen in Supp. Fig. 2. The divisions of data are based on dates and species names recorded in Table 2.

Southern Papio core habitat through a glacial–interglacial cycle

The environmental core habitat predicted for the entire chacma species during the LGM is approximately twice the size of the predicted core habitat during the LIG. This appears to result from cooler temperatures in the LGM that allowed more northern land to fall within the bounds. Potentially this could have enabled more northward migration. However, even with the considerable increase in potential core habitat, it should be noted that more southern core areas became limited and therefore the southern population may not have

123 Evol Ecol had access to the majority of this northerly placed core habitat expansion. Despite the 63 % increase in the land area within the maximum temperature envelope during the LIG, predicted habitat was still predominantly limited by rainfall. This was because the majority of additional land area was outside the modelled rainfall envelope. This reflects the lower annual rainfall in areas of the south west including Botswana, Namibia and South Africa during this period. There appears to be a westward shift in core habitat with the LGM environmental conditions. During the LGM the core habitat was primarily within Angola and Botswana. This is markedly different from the other two environmental scenarios (LIG and current conditions), where South Africa and Zimbabwe together contained the majority. This variation is most prominent in Botswana, increasing from 1 % of predicted habitat in the LIG, to 40 % in the LGM. Namibia also contained more predicted habitat during the LGM (*8 %) than during current and LIG conditions. Angola was predicted to contain 23 % of the total predicted core habitat for the LGM. This then changes to approximately 11 % during the LIG. However, during the more intermediate conditions of the current climate, Angola’s predicted core area was\1 %. This was shown to be a direct result of wetter and warmer conditions. Lesotho contains small amounts of predicted core habitat during the LIG and current conditions, where Swaziland contains predicted habitat during the LIG and LGM. To the north, Zambia contains small amounts of predicted core habitat for the LGM and current conditions but has a significant proportion (12 %) during the LIG. Small areas of environmental core habitat were seen during the LGM along the western border of Mozambique. Interestingly, these core habitat areas located in Mozambique during the LGM could potentially explain the current chacma baboon population residing in that area. Stone et al. (2013) note that Mozambique contains a large amount of marginal habitat in current conditions. Palaeooclimatic data suggest a dynamic cyclic pattern over time. In accordance with the palaoeclimatic predictions by DeMenocal (1995) the predicted central niche is likely to have been viable throughout the Quaternary. Prior to the Quaternary glacial–interglacial cycling and intensity is less pronounced and the climate was consistently warmer (DeMenocal 1995). Although these environmental conditions are less suited to the scope of this model, the central niche prior to the Quaternary more likely reflects environmental conditions closer to or more extreme than those of the LIG (Fig. 3d).

Altitude and the mountain baboon

Altitude has previously been included when modelling for the total habitat of the chacma baboon (Stone et al. 2013). However, the methods for the central niche through time model were consistent with Stone et al. (2015), who excluded altitude from central niche modelling as its effect was found to be negligible. However, unlike the entire chacma species models, the clades were affected by the exclusion of altitude. This difference results from the chacma species encompassing the entire altitude range of southern Africa, in comparison to the chacma clades which are limited to more speciﬁc geographic regions, thus altitude was appropriate. It appears that the inclusion of altitude acts as a proxy for selection of habitat type. This is clearly visible with the southern clades (Cape chacma and Ruacana), for which the core habitat is located in montane regions (Fig. 5). Altitude is more limiting for the northern clade (Supp. Fig. 3). The southern–northern altitudinal split is perhaps unsurprising

123 Evol Ecol considering the 1.84 Ma divergence of the grey footed chacma baboon from the southern chacma baboons (Sithaldeen et al. 2009). Some southern chacma populations are often referred to as mountain baboons (An- derson 1990; Henzi and Lycett 1995; Whiten et al. 1987) and their reliance on montane habitats appears to be reﬂected in their sleeping site selection. The majority of studies focusing on the southern chacma [87 %, n = 28 (Stone 2014)] report the use of geological sleeping sites and a preference for these sites even when suitable sleeping trees are available (Anderson 1982; Hamilton et al. 1976; Henzi et al. 1992). Chacma baboon populations that do use sleeping trees appear to be in the more northern warmer regions (Hall 1963; Hamilton et al. 1976; Stoltz and Saayman 1970) or include exotic tree species and/or human modiﬁed environments (Hoffman and O’Riain 2011; Pebsworth et al. 2012).

The effect of environmental conditions on the chacma baboon clades

The considerable change in both latitude and the size of the predicted central niche for the northern clade during the LGM was further explored by depicting both the northern and combined southern clades together, along with their 95 % environmental range (potential total habitat). Figure 7 shows the predicted core habitat for the grey footed chacma

Fig. 7 The predicted central niche displayed with the potential total habitat (which includes areas excluded from the central niche to counter for shifts, contractions and expansions in range, outliers and extreme environmental conditions) for both the southern chacma and the grey footed chacma during the last glacial maximum 123 Evol Ecol receding northward during the LGM. Potentially, the grey footed chacma population may have been yielding to competition and retreating north, due to the northward movement from the potentially larger clade of southern baboons. The grey footed chacma core habitat is almost surrounded by land predicted as core habitat for the southern clade. It can be seen that both of these predicted areas are within the total ranges predicted for each clade. This could imply considerable habitat competition between clades at this time. Despite being unable to model for a possible southward push resulting from inter- speciﬁc competition with the yellow baboon, it is possible that the environmental changes, coupled with the northward movement of the combined southern clade would force a northward progression for the grey footed clade. This is of particular interest following the suggested introgression of the chacma baboon into parts of the southern yellow baboon population (Zinner et al. 2009). This model would support such a dynamic, with a cyclic process of progression and regression during the glacial–interglacial cycle. Alternatively it could be argued that such a large ﬂuctuation in core habitat area for the northern clade (grey footed) could have resulted from limited data, however, 56 points were allocated to the northern chacma clade and the surrounding landscape was sampled. Despite the large amount of predicted core habitat for the combined southern clade during the LGM (Fig. 7), the considerable northward shift could potentially be a limiting factor for the population. The limited southern areas may well have forced a northward migration to access the core habitat, but even so it is possible that the majority of this land was inaccessible or accessible to only the northern fringes of the southern clade, due to the northern locality. This, coupled with the northward migration of the grey footed chacma, potentially supports the period of isolation between the northern and southern clades, as suggested by Sithaldeen et al. (2015). The southernmost population, the Cape baboon, is most likely to have been affected by this. The dynamics of the core habitat expansion and contraction for the three chacma clades (northern grey footed and two southern clades) demonstrates that each clade responds differently to the environmental changes (Fig. 8). During the LGM the Cape chacma core habitat, and by inference the population, is at its largest, whilst the grey footed chacma population is at its smallest and likely most northerly position. In current conditions as the grey footed chacma population appears to expand, the Ruacana chacma recedes to the south west coast; its core habitat, and by inference its population, appears to be in full regression. However, this considerable reduction follows an extreme expansion decreasing from an area that was 390 % larger during the LIG. Lastly, during the LIG when the Ruacana baboon core habitat was at its maximum, the Cape baboon habitat was at its smallest, \80 % of what it is today. The LIG conditions are unique in that the core habitat of all three clades was within relatively close proximity. Close proximity of core regions suggests increased potential for introgression, because core regions are within total range, thus range overlap was potentially most prominent during these conditions. Cape baboon core habitat can be seen within 5 km of core habitat for the two remaining clades and grey footed and Ruacana baboon core habitat can be found less than 90 km apart. During the LGM Cape chacma habitat can be seen to divide into four major areas; eastern coastal South Africa, northern South Africa, an area centred over northern Bots- wana and an area in central eastern Angola. The two most northern patches make up the majority of the predicted core habitat (74 %) and are well beyond current Cape chacma range. In a scenario where the Cape chacma was unable to access these two areas, the remaining habitat still equates to the largest core area for any baboon clade during the LGM. Thus the population is still likely to have suffered the least decline, if any. In 123 Evol Ecol

Fig. 8 Predicted progression of the central niche for the grey footed, Ruacana and Cape chacma baboons proportion to the decline in the other clades, the Cape chacma may have had little com- petitive resistance to some degree of northward expansion. The core habitat expansion and contraction dynamics show that the predicted area for each clade appears to fluctuate under different conditions. This difference could result from the fluctuations of the glacial–interglacial cycle, driving genetic variation to benefit each clade under different conditions. The percentage changes in the core area experienced by 123 Evol Ecol the grey footed clade suggest these animals are more susceptible to climate influence. The hardier baboon appears to be the Cape chacma, experiencing the smaller percentage change when comparing estimated largest prediction to smallest prediction. This may have interesting implications with regard to climate change. The clade predicted to be most negatively affected by an increase in temperature is the Cape baboon. Although the Cape baboon is possibly the hardiest of the baboons, the current population is already suffering the effects of anthropogenic influence, with some local populations endangered or at risk of extinction (Uys 2011a, b, 2012). As a result, the Cape baboon may be adversely affected by rising temperatures. Conversely, the Ruacana baboon, predicted to currently occupy the smallest amount of core habitat, may find rising temperatures beneficial. This research is an initial step towards a better understanding of the palaeobiogeography of the chacma baboon. Core habitat areas were predicted in every southern Africa country during some stage of glacial–interglacial cycle. The areas that were always within the core habitat were focused in the south east, with the majority located in South Africa. This modelling process has provided a unique view of a possible expansion and contraction cycle that may well have led to introgression, diversification and even speciation events. The Cape chacma baboon appears to be the environmentally hardier baboon, though; in current conditions, the grey footed chacma has the largest core habitat. These preliminary palaeobiogeographic estimates can be assessed and remodelled as data allows, whilst the method allows us to gauge species and clade dynamics including possible outcomes following future climate change.

Acknowledgments We would like to thank Amathole Museum, Ezemvelo Wildlife, Iziko South African Museum, National Museum Bloemfontein, and the Ditsong National Museum of Natural History formerly the Transvaal Museum for the provision of data. In addition we would like to thank Ezemvelo KZN wildlife staff and residents of KwaZulu-Natal who assisted us in gathering baboon location data and Justin Adams for discussions on fossil primate taxonomy. AIRH acknowledges support from the Australian Research Council Future Fellowship FT 120100399. Thank you to the anonymous reviewers whose comments improved the manuscript.

References

Adams JW (2012) A revised listing of fossil mammals from the Haasgat cave system ex situ deposits (HGD), South Africa. Palaeontol Electron 15:29A Adams JW, Kegley AD, Krigbaum J (2013) New faunal stable carbon isotope data from the Haasgat HGD assemblage, South Africa, including the first reported values for Papio angusticeps and Cercopithe- coides haasgati. J Hum Evol 64:693–698 Anderson CM (1982) Baboons below the tropic of Capricorn. J Hum Evol 11:205–217 Anderson CM (1990) Desert, mountain and savanna baboons: a comparison with special reference to the Suikerbosrand population. In: Thiago de Mello M, Whiten A, Byrne RW (eds) Baboons: behaviour and ecology. Use and care. Selected proceedings of the XIIth congress of the international primatological society, Brasil, Brasilia, pp 89–103 Berger LR, Keyser AW, Tobias PV (1993) Gladysvale: first early hominid site discovered in South Africa since 1948. Am J Phys Anthropol 92:107–111 Braconnot P, Otto-Bliesner B, Harrison S, Joussaume S, Peterchmitt JY, Abe-Ouchi A, Crucifix M, Driesschaert E, Fichefet T, Hewitt CD, Kageyama M, Kitoh A, Laıˆne´ A, Loutre MF, Marti O, Merkel U, Ramstein G, Valdes P, Weber SL, Yu Y, Zhao Y (2007) Results of PMIP2 coupled simulations of the Mid-Holocene and Last Glacial Maximum—part 1: experiments and large-scale features. Clim Past 3:261–277 Brain CK (1981) The hunters or the hunted? An introduction to African cave taphonomy. University of Chicago Press, Chicago Brink J (1999) Preliminary report on a caprine from the Cape mountains, South Africa. Archaeozoologia 10:11–25 123 Evol Ecol

Brink JS (2012) Appendix: The fauna from Lithakong. S Afr Humanit 24:30–32 Brink JS, Deacon HJ (1982) A study of a last interglacial shell midden and bone accumulation at Herolds Bay, Cape Province, South Africa. Palaeoecol Afr 15:31–39 Broom R, Schepers GWH (1946) The South African fossil ape-man: the Australopithecinae. Transvaal Museum, Pretoria Butynski T, Kingdon J, Kalina J (2013) Mammals of Africa volume II primates. Bloomsbury Publishing, London Chatterjee H, Tse J, Turvey S (2012) Using ecological niche modelling to predict spatial and temporal distribution patterns in Chinese gibbons: lessons from the present and the past. Folia Primatol 83:85–99 Clark PU, Dyke AS, Shakun JD, Carlson AE, Clark J, Wohlfarth B, Mitrovica JX, Hostetler SW, McCabe AM (2009) The last glacial maximum. Science 325:710–714 de Ruiter D (2003) Revised faunal lists for Members 1–3 of Swartkrans, South Africa. Ann Transvaal Museum 40:29–41 de Ruiter DJ, Brophy JK, Lewis PJ, Churchill SE, Berger LR (2008) Faunal assemblage composition and paleoenvironment of Plovers Lake, a Middle Stone Age locality in Gauteng Province, South Africa. J Hum Evol 55:1102–1117 de Ruiter DJ, Pickering R, Steininger CM, Kramers JD, Hancox PJ, Churchill SE, Berger LR, Backwell L (2009) New Australopithecus robustus fossils and associated U–Pb dates from Cooper’s Cave (Gauteng, South Africa). J Hum Evol 56:497–513 Delson E (1984) Cercopithecoid biochronology of the African Plio-Pleistocene: correlation among eastern and southern hominid-bearing localities. Courier Forschungsinstitut Senckenberg 69:199–218 DeMenocal PB (1995) Plio-Pleistocene African climate. Science 270:53 Dirks PH, Kibii JM, Kuhn BF, Steininger C, Churchill SE, Kramers JD, Pickering R, Farber DL, Me´riaux A-S, Herries AI, King GC, Berger LR (2010) Geological setting and age of Australopithecus sediba from southern Africa. Science 328:205–208 Doran TL, Herries AI, Hopley PJ, Sombroek H, Hellstrom J, Hodge E, Kuhn BF (2015) Assessing the paleoenvironmental potential of Pliocene to Holocene tufa deposits along the Ghaap Plateau escarp- ment (South Africa) using stable isotopes. Quatern Res 84:133–143 Dupont LM, Donner B, Vidal L, Pe´rez EM, Wefer G (2005) Linking desert evolution and coastal upwelling: Pliocene climate change in Namibia. Geology 33:461–464 Freedman L (1957) The fossil cercopithecoidea of South Africa. Ann Transvaal Mus 23:121–262 Freedman L (1965) Fossil and subfossil primates from the limestone deposits at Taung, Bolt’s Farm and Witkrans, South Africa. Palaeontol Africana 9:19–48 Freedman L (1970) A new check-list of fossil cercopithecoidea of South Africa. Palaeontol Africana 13:109–110 Freedman L (1976) South African fossil Cercopithecoidea: a re-assessment including a description of new material from Makapansgat, Sterkfontein and Taung. J Hum Evol 5:297–315 Gilbert CC (2007) Craniomandibular morphology supporting the diphyletic origin of mangabeys and a new genus of the Cercocebus/Mandrillus clade, Procercocebus. J Hum Evol 53:69–102 Gilbert CC (2013) Cladistic analysis of extant and fossil African papionins using craniodental data. J Hum Evol 64:399–433 Gilbert CC, Steininger CM, Kibii JM, Berger LR (2015) Papio cranium from the hominin-bearing site of Malapa: implications for the evolution of modern baboon cranial morphology and South African Plio- Pleistocene biochronology. PLoS One 10:e0133361 Gommery D, Badenhorst S, Se´ne´gas F, Potze S, Kgasi L (2012) Minnaar’Cave: a Plio-Pleistocene site in the Cradle of Humankind, South Africa: its history, location, and fauna. Ann Ditsong Natl Mus Nat Hist 2:19–31 Grine FE, Hendey QB (1981) Earliest Primate Remains from South Africa. S Afr J Sci 77:374–376 Groves CP (2005) Order primates. In: Wilson DE, Reeder DM (eds) Mammal species of the world. The Johns Hopkins University Press, Baltimore, pp 111–184 Grubb P, Butynski TM, Oates JF, Bearder SK, Disotell TR, Groves CP, Struhsaker TT (2003) Assessment of the diversity of African Primates. Int J Primatol 24:1301–1357 Guy AJ, Stone OML, Curnoe D (2012) Assessment of the release of rehabilitated vervet monkeys into the Ntendeka Wilderness Area, KwaZulu-Natal, South Africa: a case study. Primates 53:171–179 Hall KRL (1963) Variations in the ecology of the chacma baboon, Papio ursinus. Symp Zool Soc Lond 10:1–28 Hamilton WJ, Buskirk RE, Buskirk WH (1976) Defense of space and resources by Chacma (Papio ursinus) Baboon Troops in an African Desert and Swamp. Ecology 57:1264–1272 Heaton JL (2006) Taxonomy of the Sterkfontein fossil Cercopithecinae: the papionini of members 2 and 4 (Gauteng, South Africa). Doctoral thesis, Indiana University

123 Evol Ecol

Henshilwood CS, Sealy JC, Yates R, Cruz-Uribe K, Goldberg P, Grine FE, Klein RG, Poggenpoel C, van Niekerk K, Watts I (2001) Blombos Cave, southern Cape, South Africa: preliminary report on the 1992–1999 excavations of the Middle Stone Age levels. J Archaeol Sci 28:421–448 Henzi SP, Lycett JE (1995) Population structure, demography, and dynamics of mountain baboons: an interim report. Am J Primatol 35:155–163 Henzi S, Byrne R, Whiten A (1992) Patterns of movement by baboons in the Drakensberg mountains: primary responses to the environment. Int J Primatol 13:601–629 Herries AIR (2011) A chronological perspective on the acheulian and its transition to the middle stone age in southern Africa: the question of the Fauresmith. Int J Evolut Biol. doi:10.4061/2011/961401 Herries AIR, Adams JW (2013) Clarifying the context, dating and age range of the Gondolin hominins and Paranthropus in South Africa. J Hum Evol 65:676–681 Herries AIR, Shaw J (2011) Palaeomagnetic analysis of the Sterkfontein palaeocave deposits; age implications for the hominin fossils and stone tool industries. J Hum Evol 60:523–539 Herries AIR, Curnoe D, Adams JW (2009) A multi-disciplinary seriation of early Homo and Paranthropus bearing palaeocaves in southern Africa. Quatern Int 202:14–28 Herries AIR, Hopley P, Adams J, Curnoe D, Maslin M (2010) Geochronology and palaeoenvironments of the South African early hominin bearing sites: a reply to ‘Wrangham et al., 2009: Shallow-Water Habitats as Sources of Fallback Foods for Hominins’. Am J Phys Anthropol 143:640–646 Herries AIR, Pickering R, Adams JW, Curnoe D, Warr G, Latham AG, Shaw J (2013) A multi-disciplinary perspective on the age of Australopithecus in southern Africa. In: Reed KE, Fleagle JG, Leakey R (eds) Paleobiology of Australopithecus. Vertebrate paleobiology and paleoanthropology series. Published March 31st 2013 Herries AIR, Rovinsky D, Stone OML, Adams JW (in review) Lost between Langebaanweg and Maka- pansgat: a review of the geochronology and biochronology of the Pliocene fossil record of South Africa and its implications for understanding the hominin occupation of southern Africa. Earth-Sci Rev Hijmans RJ, Graham CH (2006) The ability of climate envelope models to predict the effect of climate change on species distributions. Glob Change Biol 12:2272–2281 Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. Int J Climatol 25:1965–1978 Hoffman T, O’Riain MJ (2011) The spatial ecology of chacma baboons (Papio ursinus) in a human- modiﬁed environment. Int J Primatol 32:308–328 Hopley PJ, Marshall JD, Weedon GP, Latham AG, Herries AIR, Kuykendall KL (2007) Orbital forcing and the spread of C4 grasses in the late Neogene: stable isotope evidence from South African speleothems. J Hum Evol 53:620–634 Hopley PJ, Herries AI, Baker SE, Kuhn BF, Menter CG (2013) Brief communication: beyond the South African cave paradigm—Australopithecus africanus from Plio-Pleistocene paleosol deposits at Taung. Am J Phys Anthropol 151:316–324 Jablonski NG (2002) Fossil Old World monkeys: the late Neogene radiation. In: Hartwig WC (ed) The primate fossil record. Cambridge University Press, Cambridge, pp 255–299 Jablonski N, Frost S (2010) Cercopithecoidea. In: Werdelin L, Sanders W (eds) Cenozoic mammals of Africa. University of California Press, Berkeley, pp 393–428 Jacobs Z, Duller GA, Wintle AG, Henshilwood CS (2006) Extending the chronology of deposits at Blombos Cave, South Africa, back to 140 ka using optical dating of single and multiple grains of quartz. J Hum Evol 51:255–273 Jarvis A, Reuter HI, Nelson A, Guevara E (2006) Hole-ﬁlled seamless SRTM data V3, International Centre for Tropical Agriculture (CIAT). Available from http://srtm.csi.cgiar.org Johnson BJ, Miller GH, Fogel ML, Beaumont PB (1997) The determination of late Quaternary paleoen- vironments at Equus Cave, South Africa, using stable isotopes and amino acid racemization in ostrich eggshell. Palaeogeogr Palaeoclimatol Palaeoecol 136:121–137 Junker J, Blake S, Boesch C, Campbell G, Toit LD, Duvall C, Ekobo A, Etoga G, Galat-Luong A, Gamys J, Ganas-Swaray J, Gatti S, Ghiurghi A, Granier N, Hart J, Head J, Herbinger I, Hicks TC, Huijbregts B, Imong IS, Kuempel N, Lahm S, Lindsell J, Maisels F, McLennan M, Martinez L, Morgan B, Morgan D, Mulindahabi F, Mundry R, N’Goran KP, Normand E, Ntongho A, Okon DT, Petre C-A, Plumptre A, Rainey H, Regnaut S, Sanz C, Stokes E, Tondossama A, Tranquilli S, Sunderland-Groves J, Walsh P, Warren Y, Williamson EA, Kuehl HS (2012) Recent decline in suitable environmental conditions for African great apes. Divers Distrib. doi:10.1111/ddi.12005 Kaplan J, Mitchell P (2012) The archaeology of the Lesotho Highlands Water Project Phases IA and IB. S Afr Humanit 24:1–32 Kaufman D, Ager T, Anderson N, Anderson P, Andrews J, Bartlein P, Brubaker L, Coats L, Cwynar LC, Duvall M, Dyke AS, Edwards ME, Eisner WR, Gajewski K, Geirsdottir A, Hu FS, Jennings AE,

123 Evol Ecol

Kaplan MR, Kerwin MW, Lozhkin AV, MacDonald GM, Miller GH, Mock CJ, Oswald WW, Otto- Bliesner BL, Porinchu DF, Ruhland K, Smol JP, Steig EJ, Wolfe BB (2004) Holocene thermal maximum in the western Arctic (0–180 W). Quatern Sci Rev 23:529–560 Keyser A, Menter CG, Moggi-Cecchi J, Pickering TR, Berger L (2000) Drimolen: a new hominid-bearing site in Gauteng, South Africa. S Afr J Sci 96:193 Kibii JM (2007) Taxonomy, taphonomy and palaeoenvironment of hominid and non-hominid primates from the Jacovec Cavern, Sterkfontein. S Afr Archaeol Bull 62(186):90–97 Klein RG (1972) The late Quaternary mammalian fauna of Nelson Bay Cave (Cape Province, South Africa): its implications for megafaunal extinctions and environmental and cultural change. Quatern Res 2:135–142 Klein RG (1976) The mammalian fauna of the Klasies River mouth sites, southern Cape Province, South Africa. S Afr Archaeol Bull 31:75–98 Klein RG (1977) The mammalian fauna from the middle and later stone age (later Pleistocene) levels of Border Cave, Natal Province, South Africa. S Afr Archaeol Bull 32:14–27 Klein R (1978a) Vertebrate fauna from the Buffelskloof Rock Shelter. S Afr Archaeol Bull 33:35–38 Klein RG (1978b) A preliminary report on the larger mammals from the Boomplaas Stone Age cave site, Cango Valley, Oudtshoorn District, South Africa. S Afr Archaeol Bull 33:66–75 Klein RG (1979) Paleoenvironmental and cultural implications of late Holocene archaeological faunas from the Orange Free State and north-central Cape Province, South Africa. S Afr Archaeol Bull 34:34–49 Klein R (1981) Archaeological implications of the mammalian fauna from the Later Stone Age cave site of Byneskranskop, southern Cape Province, South Africa. In: Harding RSO, Teleki G (eds) Omnivorous primates: gathering and hunting in human evolution. Columbia University Press, New York, pp 166–190 Klein RG, Cruz-Uribe K (2000) Middle and later stone age large mammal and tortoise remains from Die Kelders Cave 1, Western Cape Province, South Africa. J Hum Evol 38:169–195 Klein RG, Cruz-Uribe K, Beaumont PB (1991) Environmental, ecological, and paleoanthropological implications of the late Pleistocene mammalian fauna from Equus Cave, Northern Cape Province, South Africa. Quatern Res 36:94–119 Kukla GJ, Bender ML, de Beaulieu J-L, Bond G, Broecker WS, Cleveringa P, Gavin JE, Herbert TD, Imbrie J, Jouzel J (2002) Last interglacial climates. Quatern Res 58:2–13 Lacruz R, Brink J, Hancox P, Skinner A, Herries A, Schmid P, Berger L (2002) Palaeontology and geological context of a middle Pleistocene faunal assemblage from the Gladysvale Cave, South Africa. Palaeontol Afr 38:99–114 McKee JK (1993) Faunal dating of the Taung hominid fossil deposit. J Hum Evol 25:363–376 McKee JK (1995) Further chronological seriations of southern African Pliocene and Pleistocene mammalian faunal assemblages. Palaeontol Afr 32:11–16 Newman TK, Jolly CJ, Rogers J (2004) Mitochondrial phylogeny and systematics of baboons (Papio). Am J Phys Anthropol 124:17–27 Ostro LET, Young TP, Silver SC, Koontz FW (1999) A geographic information system method for estimating home range size. J Wildl Manag 63:748–755 Otto-Bliesner BL, Marshall SJ, Overpeck JT, Miller GH, Hu A (2006) Simulating arctic climate warmth and icefield retreat in the last interglaciation. Science 311:1751–1753 Pebsworth PA, MacIntosh AJJ, Morgan HR, Huffman MA (2012) Factors influencing the ranging behavior of chacma baboons (Papio hamadryas ursinus) living in a human-modified habitat. Int J Primatol 33:872–887 Pickering R, Kramers JD, Partridge T, Kodolanyi J, Pettke T (2010) U–Pb dating of calcite–aragonite layers in speleothems from hominin sites in South Africa by MC–ICP–MS. Quat Geochronol 5:544–558 Pickering R, Dirks P, Jinnah Z, de Ruiter DJ, Churchill SE, Herries AIR, Woodhead J, Hellstrom J, Berger LR (2011) Australopithecus sediba at 1.977 Ma and implications for the origins of the genus Homo. Science 333:1421–1423 Pickford M (1990) Some fossiliferous Plio-Pleistocene cave systems of Ngamiland, Botswana. Botsw Notes Rec 22:1–15 Pickford M, Mein P, Senut B (1994) Fossiliferous Neogene karst fillings in Angola, Botswana and Namibia. S Afr J Sci 90:227 Plug I (1981) Some research results on the late Pleistocene and early Holocene deposits of Bushman Rock Shelter, eastern Transvaal. S Afr Archaeol Bull 36:14–21 Plug I (1993) The macrofaunal remains of wild animals from Abbot’s Cave and Lame Sheep Shelter, Seacow Valley, Cape. Koedoe 36:15–26 Plug I, Engela R (1992) The macrofaunal remains from recent excavations at Rose Cottage Cave, Orange Free State. S Afr Archaeol Bull 47:16–25

123 Evol Ecol

Reynolds S, Kibii J (2011) Sterkfontein at 75: Review of palaeoenvironments, fauna and archaeology from the hominin site of Sterkfontein (Gauteng Province, South Africa). Palaeontol Afr 46:59–88 Roberts DL, Matthews T, Herries AIR, Boulter C, Scott L, Dondo C, Mtembi P, Browning C, Smith RMH, Haarhoff P (2011) Regional and global context of the Late Cenozoic Langebaanweg (LBW) palaeontological site: West Coast of South Africa. Earth Sci Rev 106:191–214 Rosauer DF, Catullo RA, VanDerWal J, Moussalli A, Moritz C (2015) Lineage range estimation method reveals fine-scale endemism linked to pleistocene stability in Australian rainforest herpetofauna. PLoS One 10:e0126274 Sithaldeen R, Bishop JM, Ackermann RR (2009) Mitochondrial DNA analysis reveals Plio-Pleistocene diversification within the chacma baboon. Mol Phylogenet Evol 53:1042–1048 Sithaldeen R, Ackermann RR, Bishop JM (2015) Pleistocene aridification cycles shaped the contemporary genetic architecture of southern African Baboons. PLoS One 10:e0123207 Sleeman JM (2005) Disease risk assessment in African great apes using geographic information systems. EcoHealth 2:222–227 Stigall AL, Lieberman BS (2006) Quantitative palaeobiogeography: GIS, phylogenetic biogeographical analysis, and conservation insights. J Biogeogr 33:2051–2060 Stigall Rode AL (2005) The application of geographic information systems to paleobiogeography: implications for the study of invasions and mass extinctions. Paleontol Soc Pap 11:77–88 Stoltz LP, Saayman GS (1970) Ecology and behaviour of Baboons in the Northern Transvaal. Ann Transvaal Mus 26:99–143 Stone OML (2014) The biogeography and palaeobiogeography of the chacma baboon (Papio ursinus). University of New South Wales (UNSW), PhD Stone OML, Laffan SW, Curnoe D, Rushworth I, Herries AIR (2012) Distribution and population estimate for the chacma baboon (Papio ursinus) in KwaZulu-Natal, South Africa. Primates 53:337–344 Stone OML, Laffan S, Curnoe D, Herries AIR (2013) The spatial distribution of chacma baboon (Papio ursinus) habitat based on an environmental envelope model. Int J Primatol 34:407–422. doi:10.1007/ s10764-10013-19669-10769 Stone OML, Laffan SW, Curnoe D, Herries AIR (2015) Potential human impact on the environmental central niche of the chacma baboon. S Afr J Sci. doi:10.17159/sajs.2015/20140279 Thackeray J, Watson V (1994) A preliminary account of faunal remains from Plovers Lake. S Afr J Sci 90:231–232 Uys R (2011a) 2011 Oribi Gorge baboon count. Ezemvelo KZN Wildlife, Durban Uys R (2011b) 2011 Umtamvuna baboon count. Ezemvelo KZN Wildlife, Durban Uys R (2012) 2012 Ongoye baboon count. Ezemvelo KZN Wildlife, Durban Walker M, Johnsen S, Rasmussen SO, Popp T, Steffensen JP, Gibbard P, Hoek W, Lowe J, Andrews J, Bjo¨rck S (2009) Formal definition and dating of the GSSP (Global Stratotype section and point) for the base of the Holocene using the Greenland NGRIP ice core, and selected auxiliary records. J Quat Sci 24:3–17 Whiten A, Byrne R, Henzi S (1987) The behavioral ecology of mountain baboons. Int J Primatol 8:367–388 Williams BA, Ross CF, Frost SR, Waddle DM, Gabadirwe M, Brook GA (2012) Fossil Papio cranium from !Ncumtsa (Koanaka) Hills, western Ngamiland, Botswana. Am J Phys Anthropol 149:1–17 Zinner D, Groeneveld L, Keller C, Roos C (2009) Mitochondrial phylogeography of baboons (Papio spp.)— indication for introgressive hybridization? BMC Evol Biol 9:83

123