Skip to main content

Environmental and Behavioral Evidence Pertaining to the Evolution of Early Homo


East African paleoenvironmental data increasingly inform an understanding of environmental dynamics. This understanding focuses less on habitat reconstructions at specific sites than on the regional trends, tempo, and amplitudes of climate and habitat change. Sole reliance on any one indicator, such as windblown dust or lake sediments, gives a bias toward strong aridity or high moisture as the driving force behind early human evolution. A synthesis of geological data instead offers a new paleoenvironmental framework in which alternating intervals of high and low climate variability provided the dynamic context in which East African Homo evolved. The Oldowan behavioral record presents further clues about how early Homo and Homo erectus responded to East African environmental change. Shifting conditions of natural selection, which were triggered by climatic variability, helped shape the adaptability of Oldowan hominins. Together, the behavioral and environmental evidence indicates the initial adaptive foundation for the dispersal of H. erectus and the persistence of Homo. In particular, overall dietary expansion made possible by the making and transport of stone tools compensated for increased locomotor and foraging costs and provided effective behavioral-ecological responses to resource instability during the early evolution of Homo.

The interval from ∼3.0 to 1.5 million years ago (Ma) broadly defines when Homo originated and Homo erectus first evolved and dispersed beyond Africa (Antón 2003; Antón and Swisher 2004; Kimbel 2009; Kimbel et al. 1996; Pickering et al. 2011). Despite small samples of hominin fossils, particularly in the interval between 3 and 2 Ma, there is a growing body of African climatic and paleohabitat data relevant to the early evolution of Homo (e.g., Cerling et al. 2011; deMenocal 2011; Potts 2007; Trauth et al. 2007). This time interval in Africa also preserves the oldest definite stone tool flaking and the spread of innovations such as carcass processing, overall dietary expansion, and the transport of resources across ancient landscapes (de Heinzelin et al. 1999; Domínguez-Rodrigo et al. 2005; Plummer 2004; Potts 1991; Semaw et al. 2003). The goals of this paper are (1) to characterize the paleoclimate and overall environmental dynamics in which early Homo evolved, (2) to consider how well the prevailing environmental hypotheses of human evolution explain the adaptations of early Homo, and (3) to examine archaeologically visible behaviors as part of the emerging adaptability of Homo.

Paleoenvironmental research on human evolution has long emphasized the reconstruction of habitat, where the aim is to portray the main type of vegetation or climatic condition in which early hominins lived. However, the leading edge of paleoenvironmental research has changed in recent years toward a focus on environmental dynamics, that is, the nature and tempo of environmental change that resulted from climate variability, specifically the nonlinear interaction of insolation cycles, and the episodic effects of faulting and volcanism (e.g., deMenocal 1995, 2004; Feibel 1997; Potts 1998a, 2007; Trauth et al. 2010). The interaction of these factors is especially apparent in the paleoenvironmental records of the East African Rift System.

Throughout the period of hominin evolution, environmental dynamics inevitably altered the local and regional abundance of water, vegetation, and food sources. This reshaping of the overall landscape and of the time and space distribution of resources had a pervasive influence on ecological opportunities, competition, mortality, and reproductive success. It also stimulated repeated population divergence and coalescence, the subsequent degree of allopatry of populations, and thus eventual speciation in hominins and other organisms (e.g., Potts 1996b, 1998b; Vrba 1985, 1995b).

A rich array of paleoenvironmental data sets offers time sequences of regional resource opportunities and stresses that influenced the prospects for lineage divergence and the benefits and costs of adaptive strategies (ecological, social, developmental, and reproductive) associated with the origin and early evolution of Homo. Although the fossil records of southern and other regions of Africa offer a rich body of evidence of hominin evolution, Plio-Pleistocene climate and vegetation data for these parts of the continent typically offer only short-term snapshots or combine lengthy or unknown periods of time (i.e., they are highly time averaged); in addition, they are patchily distributed in time and space and often less precisely calibrated than those in East Africa. For this combination of reasons, the Plio-Pleistocene environmental synthesis developed here is mainly drawn from and is relevant to the fossil-rich, well-calibrated stratigraphic records of East Africa.

Environmental Dynamics in Which Early Homo Evolved

In this paper, climate and vegetation data for the interval from 3.0 to 1.5 Ma come from several main sources: stable isotopes, eolian dust and plant biomarker records, northeast African sapropels, and East African lake sediments. By combining the variety of environmental data sets, I briefly summarize here four principal developments in the East African and global environmental system during the evolution of early Homo and Homo erectus.

The Gradual Onset of Continental Ice Sheets (∼3.1–2.5 Ma) Denotes the Development of a Periodically Cooler, Drier, and Glaciated Planet

Heating of Earth’s surface, the amount and distribution of moisture, and the strength of ocean currents that redistribute heat are all influenced by solar insolation—the amount of solar radiation reaching marine and land surfaces (Ruddiman 2001). Insolation is regulated by three large-scale orbital cycles: “eccentricity” (the shape of Earth’s orbit around the sun; ∼100,000-yr [100-kyr] and ∼413-kyr periods), “obliquity” (the angle of Earth’s axis of rotation relative to the sun; ∼41-kyr period), and “precession” (an effect of Earth’s axial wobble that creates a progression of the seasons relative to how close the planet is to the sun; ∼19–23-kyr periods). Interactions among these cycles and with millennial-scale and shorter-term sources of variability create a dynamic system of insolation amplification and damping prone to both predictable cycles and nonlinear threshold-type change, especially over the course of the past 3 million years (Myr), since the onset of Northern Hemisphere glaciation.

Measurement of stable oxygen isotopes (δ18O) in ocean benthic foraminifera provides a global record of the trend, amplitude, and periodicity in temperature and glacial ice variability (fig. 1). Because glaciation also lowers sea level worldwide, glacial conditions (high δ18O) can also reduce the moisture (largely originating from the ocean) that reaches continental interiors.

Figure 1. 
Figure 1. 

Oxygen isotope curve (δ18O) for the past 10 Myr (data from Zachos et al. 2001). Arrow 1, beginning around 6 Ma, warm-cold climate fluctuation became more dramatic; compare pre– with post–6 Ma intervals. Arrow 2, beginning 3.0–2.8 Ma, glacial fluctuations strengthened and included the onset of Northern Hemisphere glaciations. Note that the interval 3.0–2.4 Ma was characterized by both overall cooling and an increase in the amplitude of oscillation. Arrow 3, from the mid-Pliocene through the Pleistocene, the range of climate variability increased dramatically; this observation suggests that organismal features that heightened the ability to adjust to ecological dynamics and uncertainty (i.e., adaptability) were at a premium. The genus Homo, and eventually the adaptations characteristic of Homo sapiens, evolved during the strongest fluctuations.

Between 2.8 and 2.4 Ma, an important shift took place in the dominant period of climate oscillation, from the 19–23-kyr mode to the 41-kyr mode, evident in δ18O variability. This shift in the dominant periodicity coincided with an increase in climate variability. Another notable shift in the dominant mode, from 41- to 100-kyr periodicity, occurred around 0.8 Ma, accompanied again by an increase in amplitude. The evolution of mid- and late Pleistocene Homo, including Homo sapiens, was associated with the highest-amplitude fluctuations in marine δ18O (fig. 1).

A Net Increase in Aridity and C4 Grasses Occurred over the Past 4 Myr, with a Large Increase Possibly ∼2.8 Ma and Another ∼1.8 Ma

Continental sedimentary records indicate considerable drying of East Africa and expansion of grasslands around 2.8–2.5 Ma and between 2.0 and 1.7 Ma. These aridity pulses and the timing of substantial increases in grass abundance have best been documented by three types of data: eolian dust obtained from deep-sea cores off the northeast coast of Africa (deMenocal 1995, 2004), molecular plant biomarkers obtained from these same drill cores (Feakins, deMenocal, and Eglinton 2005), and the isotopic composition of East Africa soil carbonates associated with early hominin sites (e.g., Cerling 1992; Cerling et al. 2011).

Drill cores obtained from the Atlantic and northwest Indian Oceans have provided nearly continuous records of the production and atmospheric transport of mineral dust from the African continent. Large rainfall seasonality generates the dust. Summer eolian dust plumes in northeast Africa, which are tied to Indian Ocean monsoonal surface winds, are exported to the Arabian Sea and the Gulf of Aden (Clemens 1998; deMenocal 2004). Measurement of wind-borne dust in drill cores from these areas shows that large-amplitude aridity cycles were associated with the onset of Northern Hemisphere glacial cycles around 2.8 Ma and that major shifts in windblown dust variability occurred at 2.8–2.6 Ma, 1.8–1.6 Ma, and again at 1.0–0.8 Ma (deMenocal 1995, 2004, 2011). On this basis, deMenocal has emphasized the importance of the aridity trend in African climate, which was superimposed on wet-dry cycles. Fossil pollen recovered from terrestrial settings and marine drill cores are consistent with the substantial aridity intervals, especially between 1.8 and 1.6 Ma. Drier vegetation is recorded in Rift Valley lowlands and across northwest Africa by 2.4 Ma, and arid vegetation intensified ∼1.8 Ma (Bonnefille 1995; Leroy and Dupont 1994).

Deep-sea records of African dust are complemented by the analysis of terrestrial plant biomarkers, which are derived from waxy lipids abraded from plant leaf surfaces and transported by wind to marine sediments. Molecular biomarkers in deep-sea cores offer a rich record of terrestrial vegetation. A study by Feakins, deMenocal, and Eglinton (2005) shows that long-chain n-alkanoic acids from the leafy waxes provide a reliable indicator of terrestrial vegetation. Carbon isotopic (δ13C) analysis of these molecules sampled through time from Site 231 in the Gulf of Aden demonstrates the expansion of grasslands across northeast Africa from the Miocene through early Pleistocene. As with the eolian dust record, the analysis of plant biomarkers shows considerable variability in any given time interval; however, an emphasis on the average (fig. 3 in Feakins, deMenocal, and Eglinton 2005) indicates a substantial shift toward C4 grasses relative to C3 woody vegetation between 3.4 and 2.4 Ma, with an ongoing trend toward grass-dominated habitat registered at 1.7 Ma.

Study of paleosols (buried ancient soils) has produced the third type of data set that focuses on East African aridity. Paleosols can preserve organic residues and carbonate deposits that form under certain environmental conditions. The δ13C of these ancient soil components has been used to infer the relative proportions of C3 (wooded) and C4 (grassy) signals of vegetation that grew in a limited area (∼10 m2) averaged over many years (e.g., Ambrose and Sikes 1991; Cerling 1992; Kingston 2007).

Figure 2 illustrates δ13C data over the past 7 Myr, based on the data set compiled by Kingston (2007). An overall increase in C4 biomass occurs within a mixed vegetation setting throughout the time period of human evolution. Compilation of δ13C data from Turkana and Olduvai (Cerling 1992; Cerling and Hay 1986; Wynn 2000), in particular, indicates a long-term trend toward aridity and open habitat in East Africa. However, inspection of broader compilations (e.g., Cerling et al. 2011; Kingston 2007; Levin et al. 2004; see fig. 2) suggests the wide range of vegetation settings that accompanied the aridity trend, even from 3.0 to 1.5 Ma, the period with the strongest turn from wooded to grassland settings. A δ13C analysis of soil carbonates from Gona, Ethiopia, also indicates an overall expansion of C4 grass biomass from 4.5 to 1.5 Ma, although considerable heterogeneity in the vegetation is evident throughout the sequence (Levin et al. 2004). While the onset of a “savanna” ecosystem consisting of more grass than trees is evident in the Middle Awash of Ethiopia by around 2.6 Ma (Quade et al. 2004), the oldest isotopic evidence of open grassland in East Africa is from ∼2.0 Ma at the site of Kanjera South, Kenya (Plummer et al. 2009).

Figure 2. 
Figure 2. 

Stable carbon isotope data (δ13C) for the past 7 Myr, based on the data set compiled by Kingston (2007). An overall increase in C4 biomass occurred. Although some δ13C data sets from Turkana and Olduvai indicate a strong aridity trend and grassland expansion associated with the evolution of early Homo and Homo erectus, a wider data set sampling many East African sites demonstrates the considerable heterogeneity of vegetation throughout the key time interval.

Deep Lakes Formed in East Africa between ∼2.7 and 2.5 Ma and between ∼1.9 and 1.7 Ma as End Members of Strong Moist-Arid Oscillations

Sedimentary sequences preserving evidence of hominins and other fauna include deposits, most notably lake sediments, that are sensitive indicators of past climate. In recent years, Trauth and colleagues have produced an analytical synthesis of lake deposits in East African basins. They conclude that long phases of high moisture characterized East Africa at three important times—2.7–2.5 Ma, 1.9–1.7 Ma, and 1.1–0.9 Ma—and that these times were critical intervals in human evolution (Trauth et al. 2005). In publications since 2005, these researchers have placed greater emphasis on the fact that the prolonged high-moisture phases actually occurred during “periods of extreme climate variability” (Trauth et al. 2007:475) related to Earth’s eccentricity cycle (i.e., its modulation of precession, especially at the ∼413-kyr period; Trauth et al. 2007). In other words, cycles of deep lakes and strong aridity occurred during those ∼200-kyr-long intervals.

The most precisely dated evidence of periodic lake diatomites is from the Tugen Hills, where 40Ar/39Ar dates and orbital tuning of the sedimentary sequence point to five deep-lake cycles over a period of ∼100–115 kyr between ∼2.68 and 2.58 Ma (Deino et al. 2006; Kingston et al. 2007). As to the period between 1.9 and 1.7 Ma, the major lake phase postulated by Trauth et al. (2005, 2007) corresponds to the Lorenyang Lake of the Koobi Fora Formation, Turkana basin, and Beds I and lower II, Olduvai Gorge. Analyses of the Lorenyang Lake (a precursor to the present Lake Turkana) indicate a period of relative stability surrounding a large lake from ∼2.0 to 1.85 Ma followed by stronger fluctuations between low and high lake levels and substantial change in landscape features between 1.85 and 1.7 Ma (Feibel, Harris, and Brown 1991; Joordens et al. 2011; Lepre et al. 2007; Quinn et al. 2007). At Olduvai, Ashley (2007) postulates five episodes of lake expansion and contraction between ∼1.85 and 1.74 Ma as precessionally controlled climate affected the amount of rainfall and the profile of resources available to organisms. Deino (2012) has refined the Bed I Olduvai chronology and shows a more complex picture of wet-dry cycles in relation to orbital precession. These interpretations of strong wet-dry variability are consistent with a synthesis of fossil pollen, stable isotopes, microfauna, and other evidence from this oldest part of the Olduvai sequence (Potts and Teague 2010).

The manifestation of large, deep lakes in East Africa appears at odds with the general aridity trend emphasized by deMenocal (1995, 2004) in his study of the windblown dust record. In fact, an alternative view concerning the timing of East African aridity is offered by Trauth, Larrasoaña, and Mudelsee (2009). Their statistical reanalysis of eolian dust data indicates that significant aridification of East Africa did not begin ∼2.8 Ma. Rather, heightened aridity is evident only in times of highly variable climate (strong wet-dry oscillations) in East Africa, especially ∼1.8 Ma, with a further drying trend that began ∼1.5 Ma and reached a peak starting ∼1.0 Ma, again associated with magnified wet-dry fluctuation (see also Owen, Potts, and Behrensmeyer 2009).

Several Very Long Periods (Lasting ∼130–330 Kyr Each) of Magnified Moist-Arid Variability Occurred in the Period between 3.0 and 1.5 Ma

A strong stepwise increase in monsoonal variability (moist-arid fluctuation) occurred in northeast Africa ∼3 Ma. This increase in moisture variability is recorded in the Mediterranean record of sapropels, which are dark, organic-rich sedimentary layers deposited periodically on the sea floor as the result of heightened runoff of the surface water from the Nile catchment. The succession of Pliocene and Pleistocene sapropels suggests that climate variability in the northeast quadrant of the African continent has been sensitive to orbital precession, with moist-arid cycles recorded nearly every 19–23 kyr (deMenocal 2004; Rossignol-Strick 1983). Sapropel studies have thus documented that African monsoonal climate is driven largely by precession. Geochemical studies indicate that the humid period in each cycle may have ranged from ∼4 to 12 kyr in duration (Wehausen and Brumsack 1999).

Furthermore, according to Mediterranean sapropel records, a relatively low level of monsoonal variability was expressed from ∼4.4 to 3.0 Ma, followed immediately by a sharp expansion in the range of fluctuation. This expansion coincides with the last appearance datum (LAD) of Australopithecus afarensis and the current first appearance data (FAD) for Paranthropus, ∼2.7 Ma, and Homo (sensu stricto), arguably by ∼2.4 Ma.

While the rhythm of low-latitude climate variability is orbital precession with periods of ∼19 and 23 kyr, the amplitude of tropical moist-dry variability is strongly affected by orbital eccentricity with its dual periods of ∼100 and ∼413 kyr. The intersection of the precession and eccentricity curves (essentially the interaction of four sine waves) results in a predicted sequence of high and low climate variability for African low latitudes. This predictive framework of alternating high- and low-amplitude climate variability “packets,” each lasting 104–105 years, is evident in the eolian dust records of deMenocal (1995) and is recognized by other authors (e.g., Campisano and Feibel 2007; Deino et al. 2006; Kingston 2007; Trauth et al. 2007).

The recognition of high- and low-variability intervals offers a novel framework in which to examine East African climate change. Figure 3 shows the time intervals in the period from 3.0 to 1.5 Ma that are defined by high or low eccentricity. During high eccentricity, moist-arid variability is most pronounced (i.e., a high–climate variability interval), whereas relatively stable climate occurs at low eccentricity. High variability dominates this key interval, and three of the most prolonged periods of high climate oscillation in tropical Africa over the past 5 Myr are predicted for this interval, with durations of 326 kyr (∼2.79–2.47 Ma), 288 kyr (∼2.37–2.08 Ma), and 192 kyr (∼1.89–1.69 Ma).

Figure 3. 
Figure 3. 

Alternating high– and low–climate variability intervals from ∼3.2 to 1.4 Ma. Low variability is defined by mean orbital eccentricity or 1 standard deviation (SD) below mean e for the past 5 Myr (R. Potts and P. deMenocal, unpublished data). The time interval, duration, and mean e for each interval are shown. SDs of dust flux measured in Arabian Sea core 721 (deMenocal 1995) test the validity and robustness of this high-/low-variability framework. Asterisks indicate that the direction of change in the SD matches (13 out of 17 pairs of intervals [76%]) the predicted widening and narrowing of climate variability from one interval to the next. Boldface indicates intervals in which low variability dominates for more than 30,000 years or high variability dominates for more than 100,000 years. Note that the critical interval ∼2.85–2.08 Ma (highlighted) exhibits the lengthiest eras of both high and low climate variability, that is, prolonged intervals of wide fluctuation interspersed with relatively stable climate.

Figure 4 shows a broader time perspective by showing the eight longest eras of high climate variability, based on orbital-eccentricity values, during the past 5 Myr, that is, the prediction of when moist-arid variability prevailed for the longest periods in East Africa. Several important FADs and LADs in African human evolutionary history are situated in these intervals. It should be noted, however, that the dates of these first and last appearances are likely to shift as new fossil and archaeological discoveries are made.

Figure 4. 
Figure 4. 

Longest-duration high–climate variability intervals (the longest time intervals of predicted high climate variability in East Africa over the past 5 Myr). The intervals range from 326 to 192 kyr in duration. These are the most prolonged eras of environmental instability in East African hominin evolutionary history. Several of the most prominent events in hominin evolution appear to have occurred in these intervals. FAD = first appearance datum; LAD = last appearance datum.

Figure 5 shows a plot of sapropel spectral reflectance from 2.7 to 1.5 Ma; the intensity of this reflectance records the strengthening and reduction in moist-arid variability. Analysis of sapropel color (spectral reflectance) confirms that monsoonal intensity was magnified in alternating “packets” of high and low climate variability as predicted by eccentricity-modulated precession over long periods of time. The bottom of figure 5 situates the oldest known Oldowan archaeological occurrence (Gona) in a high-variability interval. Oldowan sites throughout the period occurred in both moist and arid environments and in prolonged eras of high or low climate variability.

Figure 5. 
Figure 5. 

Sapropel variability, showing the presence of alternating intervals of high and low climate variability, from 2.7 to 1.5 Ma (right to left). Spectral reflectance is a measure of dark versus light color, with higher values for lighter stratigraphic layers (arid times, prevalent during less variable periods) and lower values for darker strata (moist times, especially prevalent during high moist-arid variability). Variability, noted in the upper right, provides examples of well-defined time periods of high and low climate fluctuation. Malapa, SK, ST, Wonderwerk, and A.Hanech are key southern and northern African Oldowan sites and age estimates. The remaining points and age estimates represent key East African Oldowan archaeological sites, starting with the oldest documented so far (OGS-6/7 at Gona, ∼2.58 Ma). Spectral reflectance data are from ODP Site 967A (Hilgen et al. 1999; P. deMenocal, personal communication).

This synthesis of environmental data sheds light on the larger physical and biotic context of hominin populations in and beyond the localities that have produced fossils so far. Over the past three decades, a small number of key hypotheses have proposed critical linkages between climate, resources, ecological and social interactions, natural selection, and speciation. These hypotheses offer robust and testable ideas regarding the environmental factors involved in the evolution of early Homo.

Environmental Drivers of Evolution in Early Homo: Cooling, Aridity, Moisture, Warmth, or Variability?

The environmental hallmarks and data sets reviewed in the preceding section have led researchers to propose a variety of environmental explanations concerning the origin of Homo and the early evolution of Homo erectus. The following summarizes the most prominent hypotheses.


This idea has been around the longest, probably as a result of the fact that glacial climate has, since the 1800s, been understood as an important context of human evolution. The causal influence of climate cooling on African hominin evolution was formalized by Vrba in her influential turnover-pulse hypothesis (Vrba 1985, 1988, 1995a, 1995b) and habitat theory (Vrba 1992). Vrba proposed that global cooling was a driver of evolutionary change across a wide range of taxa, that is, a concentrated pulse of extinction and speciation events in lineages that, in Africa, favored cool- and arid-adapted taxa at the expense of warm- and moist-adapted lineages, particularly evident in bovids and rodents. Cooling thus set the climatic and ecological contexts in which the genus Homo (as well as Paranthropus) originated, around 2.5 Ma, or more broadly between 2.8 and 2.4 Ma (Vrba 1988). In a more speculative development, Vrba (1994) posited that cooling produced conditions that favored the retention of juvenile neurocranial traits through neotenic processes, which promoted brain enlargement in Homo.


The long-term aridity trend in Africa was established by tectonic uplift, with the formation of the East African Rift System having a dominant effect in that portion of the continent (Sepulchre et al. 2006). Aridity, the expansion of open habitat, and, ultimately, the formation of grasslands were parallel trends. Vrba’s most influential papers seemed to consolidate the savanna hypothesis of early human evolution: global cooling led to African drying and the spread of grass-dominated habitats. The savanna hypothesis is the long-established idea that dry, open, grassy settings provided the stage on which the human evolutionary drama unfolded. In a series of workshops and conferences organized by Vrba (e.g., Vrba et al. 1995), a number of researchers offered paleontological and climate data sets that fell in line with Vrba’s idea that something important happened in African climate and mammalian evolution between 2.8 and 2.4 Ma. Among those data sets was deMenocal’s eminent analysis of eolian dust input from northeast Africa to the Arabian Sea and the Gulf of Aden. His efforts (e.g., deMenocal 1995) drew further attention to the East African drying trend with a proposal that there was a stepwise increase in dust production on the African continent and dust input to nearby deep-sea records ∼2.8–2.5 Ma. The conclusion reached by deMenocal and others is that Homo arose as an integral part of the arid- and grassland-adapted African biota and that the major adaptations of Homo (e.g., stone tools, meat eating, brain size increase, geographic dispersal) were responses to the increase in aridity. The use by deMenocal of Cerling’s East African compilation of paleosol δ13C, which portrays an increase in grass proportions over time, adds evidence to the aridity hypothesis. A comparison of paleosol δ13C for the entire Turkana basin and the lower Awash basin, Ethiopia, by Levin et al. (2011) confirms an overall increase in C4 (grass) vegetation in floodplain settings between 4 million and 700,000 years ago.


In considerable contrast, Trauth, Maslin, and colleagues (Maslin and Trauth 2009; Trauth et al. 2005, 2007, 2010) have argued that lengthy periods of high moisture and climate-driven production of deep lakes were the primary drivers of Pliocene and Pleistocene human evolution in East Africa. They envisioned three main intervals of climatic moisture that were associated with key events in human evolution. Specifically, they considered earliest Paranthropus and Homo to be associated with the first moisture phase, from 2.7 to 2.5 Ma; they placed the earliest H. erectus, its dispersal beyond Africa, and the origin of the Acheulean in the second phase, from 1.9 to 1.7 Ma; and they asserted that the extinction of Paranthropus and further expansion of H. erectus matched up with the third moisture phase, from 1.1 to 0.9 Ma. As noted above, this hypothesis actually emphasizes that lake development was part of a highly variable climate system; thus, these authors actually consider the moisture hypothesis as a subset of the variability selection hypothesis (Trauth et al. 2005, 2007; see “Variability,” below).

Trauth et al. (2010) expanded the moisture hypothesis by combining tectonic evidence for north-south rift formation in East Africa with lake history to postulate that large lakes imposed a barrier to populations on opposite sides (east vs. west) of those lakes, whereas lake contraction and drying of the rift floor created refugia on opposite rift shoulders. This sequence thus promoted vicariance and allopatric speciation. The authors thus consider large lakes as “amplifiers” that drive evolutionary change.


This hypothesis derives largely from Passey et al. (2010), who document what they consider to be extraordinarily persistent high-temperature soils in the Turkana basin, Kenya, over the past 4 Myr. Soil carbonates were analyzed using an innovative method known as the “clumped-isotope thermometer” (Ghosh et al. 2006:1441) to investigate the temperature history of the Turkana basin. This method uses the distribution of 13C-18O bonds in paleosol carbonates as a proxy for soil temperature during carbonate formation. For the Turkana basin, the results indicate persistent hot temperatures above 30°C and often above 35°C for the past 4 Myr, similar to those in the present Turkana area, which is one of the hottest places on Earth. Although Passey et al. (2010) note that these surprising temperature estimates apply only to periods of soil carbonate formation, they claim that such periods make up most of the past several million years in the Turkana basin. The authors link this finding to the evolution of human thermophysiology, specifically, the change in body proportions associated with early H. erectus and “a long-standing human association with marginal environments” (Passey et al. 2010:11245).


While recognizing that environmental variability is more than mere noise in an overall trend, each of the previous four hypotheses emphasizes mainly one dimension of the global or African environmental system and treats it as the dominant adaptive challenge and evolutionary force relevant to the evolution of early Homo and H. erectus. An integrated treatment of the environmental data sets, however, begs the question as to what the principal environmental signal in this dynamic era of Earth’s climate history truly was. Was it cooling or warmth, aridity or monsoons? The variability hypothesis, also known as variability selection, cuts across this possible impasse by highlighting not any one trend, habitat, or extremity of climate oscillation but rather the evolutionary effect of environmental dynamics itself (Potts 1996a, 1996b, 1998a, 1998b, 2007). It is the variability across a large set of environmental axes that led to highly varying adaptive conditions over time and space: warm-cool, wet-dry, high or low resource abundances, concentrated or patchy resource distributions, high or low parasite loads, high or low predation risks, dense or dispersed species populations, and strong or weak interspecies competition.

The variability selection idea points to this spectrum of environmental dynamics in creating a signal that can prompt adaptive change. It seeks to answer how a population of organisms can change over time via a process of adaptation to the variability—that is, to the temporal range—of environmental dynamics in adaptive settings. Furthermore, it posits that adaptation to environmental dynamics fosters plasticity, adaptive versatility, or—perhaps the most encompassing term at many levels of biological organization—adaptability (Potts 1998b, 2002, 2007; see fig. 6).

Figure 6. 
Figure 6. 

Three possible outcomes of population evolution in a time series of environmental dynamics typical of the Plio-Pleistocene. The ability to move and track habitat change geographically (narrow lines) or to expand the degree of adaptive versatility is important for any lineage to persist. Extinction occurs if species populations have specific dietary/habitat adaptations (i.e., a narrow band of “adaptive versatility”; highlighted bands) and cannot relocate to a favored habitat. In the hypothetical situation (rightmost band) where adaptive versatility expands, migration and dispersal may occur independently of the timing and direction of environmental change. The evolution of adaptive versatility is the impetus behind the variability selection idea.

Based on evidence that well-defined eras of pronounced climate variability occurred between 3.0 and 1.5 Ma, coupled with episodic revisions of landscapes due to tectonic events, the idea is that African environmental dynamics created highly diverse conditions of natural selection that led to positive selection for genetic combinations favoring adaptability and thus the ability of certain organisms to adjust to environmental change, move to new habitats, and respond in novel ways to their surroundings. Adaptive change in response to environmental dynamics thus engendered responsiveness at many biological levels—from molecular, cellular, and physiological to developmental, social, and ecological dimensions of life (Potts 2002). The possibility presented by the variability selection hypothesis is that the evolution of early Homo and H. erectus was embedded in environmental instability and can be explained by selection that improved the ability of certain hominin populations (ultimately species) to adjust to variation in their adaptive setting.

Variability selection as a viable process of evolutionary change has recently been tested by Grove (2011), who used a single-locus genetic model originally suggested in Potts (1996a, 1996b, 1998b). In Grove’s simulations, “versatilist” alleles that build genetic combinations favoring plasticity were unable to increase when the fluctuating environment was modeled as a sine wave. However, in an empirical environment based on δ18O for the past 5 Myr, variability selection was inevitable as versatile strategies of adaptation and behavioral plasticity were favored (Grove 2011). This test of variability selection as an evolutionary process aligns with points made by deMenocal (2011): in a situation where seasonal- to orbital-scale fluctuation is regular or even in tempo and amplitude, variability alone is unlikely to have served as a selection agent; however, progressively larger degrees of environmental variability, evident as an overall trend in the δ18O curve (see fig. 1) and in particular intervals during the Pliocene and Pleistocene (figs. 3, 4), may result in new adaptations that promote adaptable behavior, including successful responses and dispersal to novel environments.

Geographic Variation and the Possible Role of Refugia

According to several recent studies, the expression of climatic conditions varied considerably across different areas of East Africa. The Omo-Turkana basin has especially contributed to this emerging picture of geographic variation. In a study of soil δ13C and δ18O by Levin et al. (2011) encompassing ∼5,000 km2, the northern floodplains of the ancestral Omo River between ∼2.9 and 2.0 Ma were dominated by woody C3 vegetation indicative of a seasonally moist, riparian woodland, at the same time as the broader southern floodplains had drier, less productive soils populated by grassy C4 vegetation. The Awash basin, by comparison, was more arid than the Omo-Turkana basin, and it supported a greater proportion of C4 vegetation during this same interval (Levin et al. 2011). These comparisons suggest that different regions, both between and within basins, manifested varied degrees of sensitivity to the large-scale climate shifts that affected East Africa overall during the late Pliocene.

Furthermore, a study by Joordens et al. (2011) introduces the intriguing idea of a habitat refugium as a magnet for hominin populations during the Plio-Pleistocene. The authors develop a new type of climate record, a strontium isotopic (87Sr/86Sr) indicator applied to lake fish fossils. Their study demonstrates, first, that precessional moist-arid variation did indeed affect the lake that existed in the Turkana basin between 2.0 and 1.7 Ma but, second, that the basin retained permanent water and moist wooded habitats throughout the period from ∼2.0 to 1.85 Ma. The authors show that hominin fossils occur during both wet and dry phases of the long-term monsoonal cycles and further suggest that hominins were drawn to continuously well-watered habitats over this long period. In other words, for about 150,000 years the eastern Turkana basin served as a refugium for hominins and other permanent water-dependent fauna. This refugium was buffered from severe lake level fluctuations and drought that affected other regions of East Africa.

The idea that an area on the order of 102–103 km2 was protected over the long term from extreme fluctuations is an important consideration. Could the reliable resources of a refugium have provided a more significant context for the evolution of early Homo or Homo erectus than climate-driven extremes of food and water fluctuation over the broader East African region?

This question cannot yet be answered; however, the moist, wooded refugium suggested by Joordens et al. (2011) appears not to have extended past 1.85 Ma. Isotopic comparison across the Omo-Turkana basin indicates a major restructuring of hydrology and vegetation toward arid C4 habitats beginning no later than 1.9 Ma (Levin et al. 2011). In fact, Quinn et al. (2007) characterize the entire period from 2.0 to 1.75 Ma as one of important vegetational change in the Turkana basin, trending from relatively closed savanna woodland toward open, low-tree shrub savanna. These authors also note that the shift in average floral composition between 2.0 and 1.75 Ma coincides with high species turnover, the principal finding by Behrensmeyer et al. (1997). Furthermore, this finding is consistent with the study by Lepre et al. (2007) at East Turkana, which detects evidence of heightened environmental variability from 1.87 to 1.48 Ma; high wet-dry monsoonal variability is predicted for much of this period (fig. 3).

Do Environmental Drivers Serve as Adequate Evolutionary Explanations?

In brief, the answer to this question is “no” or, at least, “not if any given environmental hypothesis invokes a fairly simplistic notion of correlation.” Tests of correlation seek to determine how key evolutionary events map onto large-scale environmental patterns. The value of correlation hypotheses is that they usually integrate all the evidence from paleoanthropological sites pertinent to a particular evolutionary event and then examine how well the patterning of evolutionary change matches global or continental environmental trends. Information from deep-sea cores has proved especially useful in testing correlation hypotheses. Such tests do not, however, formally establish which particular environments hominins actually encountered in their local settings or the processes by which hominin populations responded to novel survival challenges in those settings.

Two additional factors, discussed below, have become critical in developing environmental explanations for the evolution of early Homo: evidence of evolutionary change in contemporaneous large mammals and the development of compelling, testable models that specify how particular types of environmental change can incite evolutionary change.

Faunal Change Can Test the Causes of Evolutionary Change in Homo

The study of faunal communities can provide an important test of how environmental dynamics influenced the evolution of Homo. Each lineage represents a natural experiment in how external environment may have prompted certain evolutionary changes. Responses across an array of large mammals—for example, the success of arid-adapted taxa over moist-adapted ones, the emergence of dental hypertrophy in multiple animal lineages, an increase in body size across carnivorous taxa, or a widening of geographic ranges across a variety of taxa—can offer clues as to the nature of the adaptive challenges and the conditions that shaped human evolution, including the origin and early evolution of the genus Homo. For these tests to be effective, the relevant faunal lineages must occur in the same times and places in which hominins lived and evolved. This point has been recognized for some time, particularly for the interval between about 3.6 and 1.8 Ma.

Analysis of a wide range of large mammal fossil assemblages from eastern and southern Africa by Reed (1997), for example, showed that grazing adaptations fluctuated within narrow limits (10%–25%) between 3.6 and ∼2.0 Ma. Only by about 1.8 Ma did an increase in the percentage of grazing species occur, exceeding 30%. Reed’s finding corresponds well to information from δ13C (Cerling 1992; Levin et al. 2011), African dust variability (deMenocal 1995, 2004), and evidence of species turnover in the Turkana basin (Behrensmeyer et al. 1997). Reed (1997) concludes that the climatic and ecological transition at ∼1.8 Ma corresponds to the early evolution of Homo erectus, and thus only with this later species of Homo (after 2.0 Ma) do we have a hominin adapted to arid and open landscapes.

A different finding by Bobe and Eck (2001) came from measuring the relative abundances of bovids recovered from the Omo Shungura Formation between 3.4 and 1.9 Ma. The results indicate a climatic shift toward increased aridity beginning ∼2.8 Ma and intensifying by 2.3 Ma, an age for aridification considerably earlier than that reached by Reed (1997). Later analysis of bovids, suids, and cercopithecids in the Turkana basin (Bobe and Behrensmeyer 2004) gave evidence of (1) an initial increase in grassland-adapted mammals at ∼2.5 Ma, (2) fluctuation in the abundance of arid and moist taxa until ∼2.0 Ma, and then (3) a strong excursion toward open-country grazers at ∼1.8 Ma. On the basis of these findings, Bobe and Behrensmeyer (2004) favor a combination of explanations that invoke both aridity and variability selection as evolutionary drivers in the Turkana basin between 2.5 and 1.8 Ma.

In a more refined analysis, Bobe et al. (2007) confirmed that the abundance of grazing bovids underwent an overall increase in the Turkana basin between 3.0 and 1.0 Ma. However, different patterns of fluctuation in the percentage of grazers occurred in different parts of this large basin; in fact, grazing bovids underwent an overall decline in West Turkana between 3.0 and 1.5 Ma. According to the authors, the overall trend nonetheless indicates that East African landscapes inhabited by bovids and hominins became more open, arid, and seasonal; the moist and more vegetated end of the habitat spectrum became particularly limited after 2.0 Ma. They cautioned, however, that different patterns of faunal change in separate parts of the same region point to the difficulty of establishing definitive correlations between climatic and faunal change.

Several other notable studies occur in a compendium focused on the African Pliocene faunal evidence (Bobe, Alemseged, and Behrensmeyer 2007). Frost (2007), for example, examined evolutionary change in the Cercopithecidae and found no support for a turnover pulse from 2.8 to 2.5 Ma, the time range predicted by Vrba (1995b) for speciation and extinction caused by global cooling and the spread of grassland habitats in Africa. Frost did find, however, a clustering of first and last appearances of monkey lineages ∼2.0 Ma, which could be associated with a number of environmental happenings at that time, including East African aridification.

In a study of evolutionary change in Plio-Pleistocene carnivores (members of the Carnivora), Lewis and Werdelin (2007) noted that the earliest known appearance of stone tools ∼2.6 Ma had no obvious effect on the carnivore guild. However, a drop in carnivore speciation rate and a rise in their extinction rate after 1.8 Ma and a pronounced decrease in carnivore lineages after 1.5 Ma could be ascribed to the emergence of H. erectus, climate change, and a drop in overall prey species richness.

From these examples and others (e.g., Behrensmeyer et al. 1997; Reed 2008; Vrba 1995a), it is evident that most of what is termed “hominin paleoecology” to date has involved either the analysis of lineage turnover or habitat reconstruction. Questions about which species commonly occurred together and which taxa had particularly strong associations with hominins have yet to gain much attention, even though such studies would truly reflect “paleoecology” in terms of documenting species co-occurrences and potential interactions. Such studies would offer the strongest clues regarding the adaptive milieu that influenced early hominin populations. Some notable exceptions exist, however, drawn mainly from the late Pliocene Turkana basin and the mid-Pleistocene Olorgesailie basin (Bobe and Behrensmeyer 2004; Bobe, Behrensmeyer, and Chapman 2002; Potts 2007). These studies suggest the fluidity of species composition in faunal communities dated between 2.8 and 1.8 Ma (Turkana basin) and between 1.0 and 0.6 Ma (Olorgesailie basin). These findings are consistent with environmental variability’s role in causing the assembly and disassembly of ecological communities and the continual shifting of the adaptive conditions associated with the evolution of Homo.

An Understanding of Evolutionary Processes Is Also Vital

Any evolutionary explanation must posit explicitly what evolutionary processes were at work in order to evaluate the feasibility of each explanation and to lay out potential tests of the explanation. Vrba (1985, 1992, 1995b) has been explicit on this point by asking what evolutionary mechanisms are involved in translating from external environment to evolutionary responses. She posits, for example, that cooling resulted in directional selection and adaptive evolutionary responses to cooler, drier habitat and in speciation that favored organisms possessing such adaptations. Other researchers implicitly invoke habitat-specific directional selection as the primary adaptive process, although very little attempt is made to state how directional selection related to aridity, moisture, or high temperature translated into evolutionary change at the critical junctures in the environmental system.

The idea of variability selection has followed Vrba’s lead by focusing on an evolutionary process as the foundation for explaining hominin and faunal evolution. As noted above, variability selection makes a specific connection between (1) evidence of increased environmental variability, (2) the effect of this increase on resources and adaptive settings at various temporal scales, and (3) the evolutionary change that may result from the inconsistency of natural selection over time, which is posited to favor adaptive versatility over habitat-specific solutions to survival problems.

A focus on evolutionary processes leads to many questions as to how climate and overall environmental change actually influenced speciation and adaptive shifts in early Homo (deMenocal 2011). The work has hardly begun to relate even the most highly resolved climate records to shifts in seasonality, landscapes, and resource abundances, that is, the things that matter to organisms. So, for example, are evolutionary transitions driven largely by stresses associated with resource scarcity, opportunities related to resource abundance, or uncertainties linked to seasonal unpredictability and longer-term landscape remodeling? Were adaptations in early Homo largely habitat specific (e.g., solutions to a well-defined and consistent set of environmental problems), or did they reflect increases in behavioral plasticity and adaptability (e.g., solutions to highly dynamic settings and environmental novelty)?

Furthermore, it is hard to say whether significant evolutionary shifts were initiated in highly localized settings (e.g., refugia), where intra- and interspecific interactions played a prominent role, or over much of an evolving lineage’s geographic range, where broader climatic and tectonic effects may have been important. Finally, were adaptations in early Homo a response largely to external (climate- and tectonism-mediated) factors or to social and competitive factors that were consistent across a diverse spectrum of habitats? It is counterproductive to frame this last matter as an either/or question; rather, it is better to see environmental, resource, social, and competitive factors as interrelated and potentially reinforcing rather than at odds with one another in explaining how evolutionary change occurred.

Archaeological Behaviors and the Emerging Adaptability of Homo

The behavioral and ecological adaptations of hominins, viewed through the archaeological record, offer their own direct clues as to the nature of responses by early Homo and Homo erectus to environmental challenges and can thus point to whether and how cooling, warming, aridity, moisture, or variability influenced evolutionary change. This section identifies certain key adaptations evident in the behavioral record of artifacts, sites, and archeofaunas and examines them in the light of the changes in the East African environmental system between 3 and 1.5 Ma.

Transporting Rocks and the Oldest Known Stone Toolmaking, ∼2.6–2.0 Ma

During this interval, groups of one or more hominin species began to seek out rocks across distances of up to several kilometers to flake and to assist in processing food. From an adaptive-strategy standpoint, this behavior is more peculiar than is commonly perceived, largely because the costs of learning about and keeping track of good-quality stone, walking considerable distances to get it, and carrying rocks of up to several kilograms across the landscape were likely quite high, especially since rocks by themselves have no caloric or nutritional value. It is not the manipulation and flaking of stone that is unexpected in hominins so much as the dedication to transporting sufficiently large amounts of rock, detectable today as concentrations, over considerable distances from places where that rock naturally occurred.

Between ∼2.6 and 2.3 Ma, the use of flaked stone tools in acquiring food entailed relatively short-distance transport of resources, typically tens to hundreds of meters (e.g., Delagnes and Roche 2005; Goldman-Neuman and Hovers 2009; Semaw et al. 2003). Longer-distance transport is well documented later. By 2.0–1.95 Ma, at the site of Kanjera South (Braun et al. 2008), hominin toolmakers were moving certain types of stone over total distances of at least 12–13 km from their closest rock sources. The energetic and potential social benefits of processing particular foods using stone tools obviously compensated for the costs of rock transport involved in such cumulative distances.

The question, then, is “Under what environmental and selective conditions did this complex of activities—rock transport, precise flaking, and mental mapping of stone and food distributions—catch on and become conspicuous in the repertoire of certain hominin populations?” The oldest known site that preserves stone flaking—Gona, Ethiopia—is associated with δ13C evidence of a prominent grassy component in a mixed vegetation setting (Quade et al. 2004). On this basis, it could be claimed that stone toolmaking was an adaptation associated with aridity and increasingly open vegetation.

However, as noted above, this oldest archaeological site is situated in a period of pronounced moist-arid variability. Figure 5, furthermore, opens the question as to why Oldowan tool behavior eventually spread across East Africa and elsewhere and became an important behavior in the genus Homo. The emergence of the Oldowan and its dispersal in an era of widely shifting environments and diverse food regimes points to an interpretation that contrasts with the long-standing aridity-grassland explanation. It is feasible that stone tools succeeded largely because carrying stone made it possible to bring tools and foods that required tool processing together in the same places—and to do this consistently across diverse habitats, even as the distribution and abundance of food resources varied over time and place. Stone transport not only incurred large energetic costs but also enabled consistent and predictable returns in response to a varying environment. As a result, the effort involved in making sharp edges and using rocks for crushing provided a resilient means of processing a changeable array of foods across a wide range of habitats.

By ∼2.0–1.95 Ma, therefore, we see definitive evidence of persistent Oldowan toolmaking (and stone transport) in an arid grassland setting recorded at Kanjera South (Plummer et al. 2009) and in a nearly contemporaneous moist, wooded habitat recorded at FwJj20, East Turkana (Braun et al. 2010). In this light, the Oldowan repertoire of behavior emerged as a characteristic of the genus Homo (even if adopted initially by non-Homo populations) because it enabled dietary and foraging adaptability in the face of Plio-Pleistocene environmental dynamics.

An important further implication regarding stone transport concerns its effect on the energetic costs of locomotion. This topic has not received much consideration since a “null model” of Oldowan stone and food transport costs was developed in Potts (1988). In light of current ideas about locomotor costs (e.g., Bramble and Leiberman 2004), the point to consider is that the cost of stone transport for particular activities must be factored into interpretations about endurance running versus walking in early Homo.

Based on weights of Oldowan tools from Bed I Olduvai (e.g., Potts 1988, 1991), an estimate of the minimum weight of the basic Oldowan equipment—a single hammerstone plus a number of basalt/trachyte/phonolite cores sufficient to yield enough sharp flakes (minimally 50–100) to cut the hide and disarticulate a fleshed wildebeest-sized ungulate—would come to ∼4–9 kg. (An average of ∼6 flake scars per Oldowan core would require at least 8–16 cores [for ∼50 to ∼100 flakes], assuming that all flake scars define usable flakes, plus one hammerstone, with calculated weights of 375–500 g per core and an average of 0.5–1 kg per hammerstone [Potts 1991].) Repeated trials in an East African field situation (Olorgesailie, Kenya), involving well-conditioned individuals carrying 5–10 kg of rocks in backpacks over uneven terrain from outcrops 5 km away, suggest that running is feasible for 100 m or so before the rate of fatigue strongly urges the desire to walk, which can be accomplished comfortably for the entire trial distance. While this does not cast doubt on whether early Homo or H. erectus ran long distances, it is reasonable to ask what the individual might accomplish after such a run if a minimal butchery tool kit involving 8–16 sizeable rocks were not also transported. Social running is feasible, although it would be speculative at best to wonder whether each individual runner would carry a similarly sized tool kit. Or perhaps running in order to capture and butcher animals was not the point. Adequate quantitative experiments yielding realistic expectations about the addition of stone transport to locomotor costs have yet to be carried out (see Pontzer 2012).

Dietary Expansion Involving Access to Meat/Marrow Resources

Because of the relative rarity and unpredictability of prey species and animal tissues across the landscape relative to plant foods, carnivory typically leads to larger foraging distances and substantially increased energy budgets (Carbone, Teacher, and Rowcliffe 2007; Kelt and Van Vuren 1999; Nagy, Girard, and Brown 1999). That Oldowan hominins coupled these latter costs with the added costs of transporting the rocks needed to process carcasses is reason to suspect that the survival and energetic returns on accessing animal tissues were substantial. Access to nutritious resources in mammalian bones and organs as well as to a wider range of plant foods is likely to have played a fundamental role in offsetting the costs of transporting stone.

The expansion in diet implied by access to animal fat and protein is also thought to have played a role in a number of adaptive changes in the evolution of Homo (Aiello and Wheeler 1995; Antón, Leonard, and Robertson 2002; Plummer 2004; Shipman and Walker 1989). The possible stature increase in early Homo and by 1.9–1.7 Ma in some populations of H. erectus could have had the dual advantage of increasing the foraging range and facilitating the search for carcasses, while animal protein and fatty tissues could have fueled body and brain growth. These foraging strategies appear to have been advantageous in an arid, open habitat. At the same time, the expansion of diet to include animal foods was an advantageous buffer against changing templates of food resources across an array of habitats. That is, the adoption of animal foods likely offered a useful means of ecological and dietary adjustment even if meat/fat acquisition was not the goal of all or most occasions of stone tool use. This approach to buffering environmental instability would have proved equally useful when moving through unfamiliar habitats, thus facilitating range expansion and dispersal (Antón, Leonard, and Robertson 2002).

As hominins ventured farther into the ecological arena of large carnivores, seeking out and carrying portions of animal carcasses had its risks (Blumenschine 1991; Donadio and Buskirk 2006). These risks were conceivably lowered when prey density was high and competition for meat relatively low. Evidence from a variety of sites—Kanjera South, FwJj20 in East Turkana, and FLK Zinj at Olduvai—implies, however, that by 2.0–1.8 Ma, the success of Oldowan toolmakers in competing for animal tissues applied across a broad spectrum of environmental conditions and probably a variety of ecologically competitive situations. This could mean that early Homo and especially H. erectus succeeded in the carnivore end of an omnivorous dietary spectrum not solely in arid, grassland habitats but across the gamut of arid-moist conditions and the changing template of seasonality associated with high climate variability.

Delayed Consumption of Food: An Extraordinary Development in Hominin Behavior with Broad Implications regarding Sociality

Relatively dense concentrations of butchered faunal remains associated with plentiful tools made from rocks 2–13 km from their sources are known by ∼2.0–1.8 Ma (Braun et al. 2008, 2010; Plummer 2004; Plummer et al. 2009). These sites appear to reflect a critical step in foraging and sociality, namely, a delay in eating the marrow- and meat-rich carcass portions that were transported. This behavior is odd, given the emphasis on “eat-as-you-go” foraging in almost all animals except when provisioning young. Although Isaac’s (1984) emphasis on home bases involving male-female division of labor, pair bonding, and other elements of human behavior has been well dissected (Binford 1981; Potts 1988), the repetitive act of carrying and aggregating rich packets of fatty tissue and protein almost certainly had important social implications (see below).

The spread or elaboration of this behavior (the transporting of food) is not well calibrated, but its appearance by ∼2 Ma falls in an era when stable and highly variable climate regimes alternated with one another, evident in the predictive framework of East African monsoonal variability (fig. 3) and in the sapropel record (fig. 5). Ultimately, this strategy of dual transport of stone and food resources was expanded to diverse habitats populated by Oldowan toolmakers, including early Homo and H. erectus, and was elaborated in the relatively variable settings of East Africa between 1.9 and 1.5 Ma as well as across the diverse habitats of northern and southern Africa and from the Caucasus to eastern Eurasia (Potts and Teague 2010).

Oldowan Behavior and Its Implications concerning Mortality in Early Homo

It is commonly assumed that the venture into the ecological domain of large African carnivores, made possible by proficient stone flaking, entailed substantive risks of injury and death and probably led to a marked rise in extrinsic mortality in Oldowan toolmakers. There is evidence, for example in Bed I Olduvai (Domínguez-Rodrigo, Barba, and Egeland 2007; Potts 1988), of substantive carnivore involvement in bone assemblages where aggregations of stone tools also occur, which is indicative of a spatially focused overlap where toolmaking hominins and meat-eating carnivores conducted their business. The overlap of carnivore tooth marks and tool butchery marks, first reported in Bed I Olduvai (Potts and Shipman 1981), might at first suggest a powerful way to investigate whether carnivores or hominins held the upper hand in their competitive interactions over carcasses. Such tooth/cut mark overlaps appear to be quite rare, however, and are unlikely to produce definitive, statistically robust results across a variety of sites. Furthermore, no systematic study has yet been carried out comparing the frequency of carnivore tooth marks on bones of Australopithecus, early Homo, and early H. erectus; the findings of such a study would not necessarily measure carnivore-caused mortality but may imply how commonly carnivores had access to hominin bones and the degree of overlap and potential risk that hominins incurred before and after the emergence of stone tool flaking.

The assumption that stone tool–assisted carnivory incurred higher mortality costs thus has yet to be demonstrated. This point leaves open the reasonable possibility that episodic and eventually persistent stone toolmaking and carnivory occurred only when the mortality costs due to predation had decreased, compared with that for earlier pre-Oldowan hominins (Lewis and Werdelin 2007). We may call this the “decreased-mortality hypothesis” of Oldowan behavior, which sees carnivory as an integral yet variably expressed aspect of the adaptive repertoire in early Homo and H. erectus. Such a decrease could have been sustained by shifts in social grouping, cooperation, vigilance, and signaling, that is, a variety of feasible behavioral changes that could have negated the supposed rise in mortality risk due to predation (Gursky-Doyen and Nekaris 2006). Conditions favoring such social-based strategies likely had to be met as Oldowan hominins carried carcass parts and invested in strategies of food and stone transport that repeatedly concentrated social encounters in certain spots on the landscape, whether through central-place or multiple-place foraging (Isaac 1984; Potts 1988, 1991).

A decrease in extrinsic mortality due to predation could have provided the context in which intrinsic mortality (due to aging) became a more prominent factor in life history (Kuzawa and Bragg 2012). In other words, decreased mortality due to predation served as either a release or the impetus for the prolongation of maturation and an increase in longevity. The latter effects, furthermore, would have improved the opportunities for alloparenting by older siblings and postreproductive adults. This train of cause and effect is currently speculative yet has the following potentially testable implications: (1) taphonomic indicators of decreased hominin-carnivore interaction (at odds with the usual assumption of increased predation risk due to overlapping carnivory); (2) prolongation of dental development/maturation (see Schwartz 2012), correlated with decreased hominin-carnivore interaction; and (3) body size increase, which is predicted to accompany decreased extrinsic mortality (see Kuzawa and Bragg 2012; Migliano and Guillon 2012; Pontzer 2012).

Increase in Body Size

An increase in body size in early H. erectus is documented by fossils such as KNM-WT 15000 (∼1.53 Ma) and KNM-ER 1808 (∼1.7 Ma). The robust innominate KNM-ER 3228 (Ruff et al. 1993) dated ∼1.92 Ma (Joordens et al. 2011) and femora (e.g., KNM-ER 1481) dated ∼1.89 Ma from the Turkana basin (Ruff and Walker 1993) further suggest that body enlargement had begun by at least 1.9 Ma, although the taxon in which this occurred is as yet unknown (see Holliday 2012; Pontzer 2012). It seems unlikely, however, that the body size increase in Plio-Pleistocene Homo was registered across all populations or environmental contexts. Since several intervals of heightened moist-arid fluctuation (with intervening low-variability periods) characterized the time between 2.0 and 1.7 Ma, it is possible that the increase in body size reflects the importance of plasticity in body growth trajectories. This proposed rise in plasticity would have made feasible a broader spectrum of body sizes, from small to large, in response to resource availability and other environmental factors. Evidence of small individuals of early H. erectus (sensu lato) at Dmanisi (Lordkipanidze et al. 2007; Pontzer et al. 2010) may indicate that the apparent evolution of large body size in this species actually reflects the evolution of plasticity in body growth in accord with a variability selection scenario rather than a response solely to arid, hot, open landscapes (see also Antón 2012; Antón and Snodgrass 2012; Kuzawa and Bragg 2012; Migliano and Guillon 2012).

Persistence and Spread of Oldowan Behaviors Beginning by 2.0 Ma

An important shift in the archaeological record of Oldowan behavior appears to be focused in the interval from about 2.0 to 1.8 Ma. Before 2.0 Ma, archaeological sites (defined by clusters of stone artifacts) are distributed over time in an episodic pattern. That is, within a given stratigraphic sequence, sites typically occur within a very confined stratigraphic interval and a narrow and repetitive range of geological settings. By contrast, after 2.0 Ma, Oldowan behavior becomes considerably more persistent, with clusters found more frequently across consecutive layers within a given stratigraphic sequence. In addition, Oldowan tools are found throughout a varied spectrum of habitats and over wider geographic areas both within Africa and, for the first time, in Eurasia.

In the series of 2.58-Ma sites from Gona, Ethiopia, for example, the three most studied lithic assemblages (EG-10, EG-12, and OGS-7) come from the same stratigraphic interval, described by Stout et al. (2010) as within the second fining-upward sequence above the base of the Busidima Formation. Each of about a dozen Gona sites, distributed from ∼2.58 to 2.53 Ma and from ∼2.17 to 2.0 Ma, occurs in fining-upward sediments overlying large cobble conglomerates (Quade et al. 2008). Site location thus seems to have been conditioned by proximity to conglomerates, which were the sources for on-site or very localized (e.g., ∼20-m distance in the case of OGS-7) stone flaking (Stout et al. 2005, 2010).

Oldowan tools associated with the Hata Member at Bouri, dating to ∼2.5 Ma, are described as rare and scattered surface artifacts with no concentrations observed. Several mammalian bones bearing cut marks and hammerstone impact scars have also been described from surface and excavated finds within a single stratigraphic horizon across more than 2 km of outcrop (de Heinzelin et al. 1999).

The next-oldest archaeological sites documented so far—from Hadar and Omo, Ethiopia, and Lokalalei, West Turkana, Kenya—are dated between ∼2.36 and 2.32 Ma. The two Hadar sites known so far (within the Makaamatilu basin), A.L. 894 and A.L. 666, are separated vertically by ∼2 m within the lowermost portion of the Busidima Formation (Goldman-Neuman and Hovers 2011). Five archaeological sites reported in Omo Shungura Member F are distributed though ∼35-m thickness of Member F; here, the repetitive characteristic is that all of the Member F sites occur within small channel deposits or associated floodplain silts of a braided stream (de la Torre 2004; Howell, Haesaerts, and de Heinzelin 1987). In the West Turkana sequence, two sites, the slightly older Lokalalei 1 and the younger Lokalalei 2C, are almost contemporaneous (Delagnes and Roche 2005), and the next Oldowan occurrences reported so far in the basin are more than 300,000 years younger, in the Upper Burgi Member at East Turkana.

There is little doubt that Oldowan assemblages dated between 2.6 and 2.0 Ma provide solid evidence of intentional flaking, selective use of raw materials, and a well-developed sense of fracture mechanics and planning during the process of stone flaking (e.g., Delagnes and Roche 2005; Goldman-Neuman and Hovers 2011; Stout et al. 2005). The boundary defined here beginning ∼2.0 Ma, nonetheless, marks a later interval of Oldowan sites that occur in a wider variety of geological, geographical, and environmental contexts. For example, excavations at Kanjera South, Kenya, have uncovered a continuous stratigraphic distribution of Oldowan artifacts on the order of 102–103 years long, dated between 2.0 and 1.95 Ma, associated with evidence of persistent carnivory (across multiple stratigraphic layers) in the form of cut and percussion marks on small and medium-sized mammalian skeletal remains (Ferraro 2007; Plummer 2004; Plummer et al. 2009). M. D. Leakey’s excavations in Bed I Olduvai, furthermore, documented a nearly continuous distribution of Oldowan stone tools from ∼1.85 to 1.70 Ma (Deino 2012; Hay 1976; Leakey 1971). The temporal distribution of sites in figure 5, moreover, illustrates the expansion of Oldowan sites within and beyond East Africa starting ∼2 Ma, after a noticeable sampling hiatus between roughly 2.3 and 2.0 Ma.

In South Africa, new age estimates and a synthesis of data (Herries, Curnoe, and Adams 2009; Herries and Shaw 2011) indicate that the oldest currently documented Oldowan tools in this part of Africa are between 2.0 and 1.78 Ma, including Sterkfontein M5A (1.8–1.5 Ma), Wonderwerk Cave (1.95–1.78 Ma), Malapa (∼1.98 Ma), and Swartkrans M1 (∼2.0 Ma, although this date is considered relatively poorly constrained). This revised chronology for the oldest archaeological finds in South Africa, along with the presence of a sequence of stone tool layers dated ∼1.8 Ma at Aïn Hanech, Algeria (Sahnouni et al. 2002), further suggests that the making of stone tools had become a regular part of the behavioral repertoire of dispersing populations of Homo by ∼2.0 Ma. The sequence of Oldowan sites now documented at Dmanisi between 1.85 and 1.78 Ma and the archaeological sequence beginning by 1.66 Ma in the Nihewan basin, China, further confirms the persistence of hominin toolmakers across a wide variety of climatic and ecological settings (Ferring et al. 2011; Zhu et al. 2004).

It is unclear whether the stratigraphic persistence of the Oldowan within East Africa beginning roughly 2 Ma and its spread to other regions can be attributed to the emergence of H. erectus or whether these phenomena are the product of two contemporaneous Oldowan toolmaking species, Homo habilis and H. erectus. It does, nonetheless, denote a shift in the regularity of Oldowan toolmaking from ∼2.0 to 1.8 Ma, the persistence of Oldowan hominins across numerous environmental transitions within East Africa, and the spread of these populations into novel environments and diverse climatic regimes. All of this occurred after the early development of the Oldowan within highly dynamic East African environments between 2.6 and 2.0 Ma and appears to have been in place by the time of the aridity pulse in East Africa.

Adaptability as a Framework for the Analysis of Adaptations in Homo

This analysis of the environmental and behavioral adaptations of early Homo and H. erectus point to the importance of adaptability in diet, foraging, and mobility, that is, resilience in the face of moist or arid habitats, abundant or scarce resources, and large lakes or dry landscapes. The picture that emerges is one of shifting variability in the ecological milieu superimposed on an overall drying trend across the time interval from 3 to 1.5 Ma. If this current understanding of environmental dynamics is correct, it suggests that adaptability may be expected in many aspects of the biology of early H. erectus and possibly its immediate precursor, especially a novel degree of developmental and physiological plasticity and a spectrum of life history trajectories that promoted the fine tuning of biological adaptations in Homo to the dynamics of its surroundings (Potts 2002). The idea that particular combinations of genes favoring plasticity can be filtered and selected because of the instability of conditions of natural selection, especially in times of increased seasonal- to orbital-scale climate variability, is consistent with the hypothetical process of variability selection (e.g., deMenocal 2011; Grove 2011; Potts 1996a, 1996b, 1998b; Trauth et al. 2007, 2010).


A few final points arise from this analysis of the East African environmental and behavioral contexts of early Homo and Homo erectus. They are as follows.

Building a Synthesis of Environmental Data

A growing array of environmental indicators—exemplified by δ18O, δ13C, eolian dust, plant biomarkers, lake sediments, and sapropels—offers insights into the evolutionary context of Plio-Pleistocene Africa. On first inspection, these records appear to contradict one another. Eolian dust and δ13C (including its application to plant biomarkers) appear to be most sensitive to the arid aspects of climate variability. This makes sense, given that the most arid times in Africa would correlate with strongest dust productivity and would lead to the precipitation of carbonate nodules in soils, the primary focus of stable carbon isotope analyses. During times of large or deep lakes, when aquatic deposits are dominant, soils and terrestrial fossil animals (whose teeth also provide δ13C data) are often not even recorded in sedimentary exposures. In a parallel vein, analyses that focus on lake sediments are sensitive to the wettest times. Deep-sea δ18O offers a picture of global ocean temperature and tends to highlight global cooling and a trend toward rising amplitude in cold-warm or glacial-interglacial oscillation over the past 6 Myr. Passey et al.’s (2010) paleotemperature proxy based on δ13C of Turkana basin paleosols is at odds with the global picture of overall cooling and pronounced temperature variability. This apparent contradiction may result from the special conditions under which carbonate precipitates in paleosols, and thus Passey’s method may sample primarily the arid and hot end member in the range of environmental variability. Finally, sapropels capture an intriguing pattern in wet-dry climate variability that does not appear in the global ocean record—that is, an alternation between high and low climate variability—rather than a continual rise in variability.

The sapropel record in particular appears to offer a promising way of examining the entire moist-arid spectrum of northeast African climate. The records of sapropels, eolian dust, plant biomarkers, and lake sediments, furthermore, can all be reconciled with one another through the high/low climate variability framework described in this paper. As noted by Feakins, deMenocal, and Eglinton (2005) with regard to biomarker isotopes, the overall aridity trend in Africa seems to be made up of high and low variability intervals; this means that an overall aridity trend can be seen as a matter of sampling: “The data … suggest that changes in northeast African vegetation were primarily related to the amplitude of subtropical orbital insolation variations, since large-amplitude vegetation variability ca. 3.7 and ca. 1.4 Ma coincided with high orbital precession variability, and low variability ca. 2.4 Ma occurred during a precessional minimum” (Feakins, deMenocal, and Eglinton 2005:979–980). Likewise, dust records show that alternating high and low climate variability is a crucial, newly recognized dimension in tropical African environment pertinent to hominin evolution.

The current limitation of the sapropel, dust, and biomarker records as applied to questions of human evolution is that they are focused exclusively in eastern and northeastern Africa. Development of long-term stratigraphic records of climatic, environmental, and ecological dynamics in other regions of Africa remains a critical need.

Evolutionary Interpretations That Emphasize Habitat or Resource Stability Lack Support

At present, the most impressive feature of East African climate records during the evolution of early Homo and H. erectus is the series of lengthy eras of pronounced moist-arid variability. In an era characterized by alternations between high and low variability, the dominance of high climate variability between 3.0 and 1.5 Ma makes it almost inevitable that the most notable first and last appearances in the hominin and faunal records are associated with instability in the conditions of natural selection.

Interpretations that invoke habitat and resource stability as an important factor in the evolution of early Homo or in the origination and dispersal of H. erectus are at odds with this synthetic picture of East African climate variability. The current FADs for these events occur during intervals of strong climate variability, which, along with local tectonic and volcanic events, caused substantial modifications to resource landscapes. The growing body of data from East African hominin-occupied basins confirms that landscapes and resources underwent recurrent, high-amplitude alterations during the focal time interval. Within this dynamic context, of growing interest is the presence of refugia that offered basic needs of water and food options during eras of strong climate variability throughout East Africa.

Evolutionary Interpretations That Emphasize Adaptive Versatility Are Well Supported

The environmental and behavioral evidence summarized here highlights the importance of ecological adaptability and physiological plasticity as an element, perhaps even the central issue, in the evolution of Homo. An evolved responsiveness to environmental variability would seem to have played a central role in the adaptive and phylogenetic history of Plio-Pleistocene Homo. It is important, nonetheless, to recognize that Pleistocene Homo underwent critical behavioral, ecological, and life history transitions over the past 1 Myr. It is thus sensible to avoid the temptation of attributing all that is important in the adaptive history of Homo sapiens to the origin of the genus or of its longest-enduring lineage, H. erectus.

I am grateful to the following colleagues for their collaboration and shared data over the past decade: Kay Behrensmeyer, Chris Campisano, Andy Cohen, Alan Deino, Peter deMenocal, Tim Eglinton, Sarah Feakins, Craig Feibel, Bernie Owen, and Tom Plummer. I also thank Jennifer Clark for assistance with the figures and John Kingston for enabling the use of his data compilation in figure 2. Research reported here was supported by the Peter Buck Fund for Human Origins Research, the Ruth and Vernon Taylor Foundation, National Science Foundation Hominid Program grant BCS-0128511, and the Smithsonian Institution’s Human Origins Program. Logistical support and collaboration with the National Museums of Kenya have provided a strong base for several of the studies reported here. I am grateful to Leslie Aiello and Susan Antón, with assistance from Laurie Obbink, for organizing the Wenner-Gren symposium in Sintra, Portugal, and to the symposium participants for stimulating discussions.

References Cited

  • Aiello, Leslie C., and Peter Wheeler. 1995. The expensive-tissue hypothesis: the brain and the digestive system in human and primate evolution. Current Anthropology 36:199–221.

  • Ambrose, Stanley H., and Nancy E. Sikes. 1991. Soil organic carbon isotopic evidence for vegetation change in the Kenya Rift Valley. Science 253:1402–1405.

  • Antón, Susan C. 2003. Natural history of Homo erectus. Yearbook of Physical Anthropology 46:126–170.

  • ———. 2012. Early Homo: who, when, and where. Current Anthropology 53(suppl. 6):S278–S298.

  • Antón, Susan C., William R. Leonard, and Marcia L. Robertson. 2002. An ecomorphological model of the initial hominid dispersal from Africa. Journal of Human Evolution 43:773–785.

  • Antón, Susan C., and J. Josh Snodgrass. 2012. Origins and evolution of genus Homo: new perspectives. Current Anthropology 53(suppl. 6):S479–S496.

  • Antón, Susan C., and Carl C. Swisher III. 2004. Early dispersals of Homo from Africa. Annual Review of Anthropology 33:271–296.

  • Ashley, Gail M. 2007. Orbital rhythms, monsoons, and playa lake response, Olduvai basin, equatorial East Africa (ca. 1.85–1.74 Ma). Geology 35:1091–1094.

  • Behrensmeyer, Anna K., Nancy E. Todd, Richard Potts, and Geraldine E. McBrinn. 1997. Late Pliocene faunal turnover in the Turkana basin, Kenya and Ethiopia. Science 278:1589–1594.

  • Binford, Lewis R. 1981. Bones: ancient men and modern myths. New York: Academic Press.

  • Blumenschine, Robert J. 1991. Hominid carnivory and foraging strategies, and the socio-economic function of early archaeological sites. Philosophical Transactions of the Royal Society B: Biological Sciences 334:211–221.

  • Bobe, René, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. 2007. Hominin environments in the East African Pliocene: an assessment of the faunal evidence. New York: Springer.

  • Bobe, René, and Anna K. Behrensmeyer. 2004. The expansion of grassland ecosystems in Africa in relation to mammalian evolution and the origin of the genus Homo. Palaeogeography, Palaeoclimatology, Palaeoecology 207:399–420.

  • Bobe, René, Anna K. Behrensmeyer, and Ralph E. Chapman. 2002. Faunal change, environmental variability and late Pliocene hominin evolution. Journal of Human Evolution 42:475–497.

  • Bobe, René, Anna K. Behrensmeyer, Gerald G. Eck, and John M. Harris. 2007. Patterns of abundance and diversity in late Cenozoic bovids from the Turkana and Hadar basins, Kenya and Ethiopia. In Hominin environments in the East African Pliocene: an assessment of the faunal evidence. René Bobe, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 129–157. New York: Springer.

  • Bobe, René, and Gerald G. Eck. 2001. Response of African bovids to Pliocene climatic change. Paleobiology 27(suppl. 2):1–47.

  • Bonnefille, Raymonde. 1995. A reassessment of the Plio-Pleistocene pollen record of East Africa. In Paleoclimate and evolution, with emphasis on human origins. Elizabeth S. Vrba, George H. Denton, Timothy C. Partridge, and Lloyd H. Burckle, eds. Pp. 299–310. New Haven, CT: Yale University Press.

  • Bramble, Dennis M., and Daniel E. Lieberman. 2004. Endurance running and the evolution of Homo. Nature 432:345–352.

  • Braun, David R., John W. K. Harris, Naomi E. Levin, Jack T. McCoy, Andy I. R. Herries, Marion K. Bamford, Laura C. Bishop, Brian G. Richmond, and Mzalendo Kibunjia. 2010. Early hominin diet included diverse terrestrial and aquatic animals 1.95 Ma in East Turkana, Kenya. Proceedings of the National Academy of Sciences of the USA 107:10002–10007.

  • Braun, David R., Thomas Plummer, Peter Ditchfield, Joseph V. Ferraro, David Maina, Laura C. Bishop, and Richard Potts. 2008. Oldowan behavior and raw material transport: perspectives from the Kanjera Formation. Journal of Archaeological Science 35:2329–2345.

  • Campisano, Christopher J., and Craig S. Feibel. 2007. Connecting local environmental sequences to global climate patterns: evidence from the hominin-bearing Hadar Formation, Ethiopia. Journal of Human Evolution 53:515–527.

  • Carbone, Chris, Amber Teacher, and J. Marcus Rowcliffe. 2007. The costs of carnivory. PLoS Biology 5(2):e22, doi:10.1371/journal.pbio.0050022.

  • Cerling, Thure E. 1992. Development of grasslands and savannas in East Africa during the Neogene. Palaeogeography, Palaeoclimatology, Palaeoecology 97:241–247.

  • Cerling, Thure E., and Richard L. Hay. 1986. An isotopic study of paleosol carbonates from Olduvai Gorge. Quaternary Research 25:63–78.

  • Cerling, Thure E., Jonathan G. Wynn, Samuel A. Andanje, Michael I. Bird, David Kimutai Korir, Naomi E. Levin, William Mace, Anthony N. Macharia, Jay Quade, and Christopher H. Remien. 2011. Woody cover and hominin environments in the past 6 million years. Nature 476:51–56.

  • Clemens, Steven C. 1998. Dust response to seasonal atmospheric forcing: proxy evaluation and forcing. Paleoceanography 13:471–490.

  • de Heinzelin, Jean, J. Desmond Clark, Tim White, William Hart, Paul Renne, Giday WoldeGabriel, Yonas Beyene, and Elisabeth Vrba. 1999. Environment and behavior of 2.5-million-year-old Bouri hominids. Science 284:625–629.

  • Deino, Alan L. 2012. 40Ar/39Ar dating of Bed I, Olduvai Gorge, Tanzania, and the chronology of early Pleistocene climate change. Journal of Human Evolution 63:251–273.

  • Deino, Alan L., John D. Kingston, Jonathan M. Glen, Robert K. Edgar, and Andrew Hill. 2006. Precessional forcing of lacustrine sedimentation in the late Cenozoic Chemeron basin, Central Kenya Rift. Earth and Planetary Science Letters 247:41–60.

  • Delagnes, Anne, and Hélène Roche. 2005. Late Pliocene hominid knapping skills: the case of Lokalalei 2C, West Turkana, Kenya. Journal of Human Evolution 48:435–472.

  • de la Torre, Ignacio. 2004. Omo revisited. Current Anthropology 45:439–465.

  • deMenocal, Peter B. 1995. Plio-Pleistocene African climate. Science 270:53–59.

  • ———. 2004. African climate change and faunal evolution during the Pliocene-Pleistocene. Earth and Planetary Science Letters 220:3–24.

  • ———. 2011. Climate and human evolution. Science 331:540–541.

  • Domínguez-Rodrigo, Manuel, Rebeca Barba, and Charles P. Egeland. 2007. Deconstructing Olduvai: a taphonomic study of the Bed I sites. Dordrecht: Springer.

  • Domínguez-Rodrigo, Manuel, Travis Rayne Pickering, Sileshi Semaw, and Michael J. Rogers. 2005. Cutmarked bones from Pliocene archaeological sites at Gona, Ethiopia: implications for the function of the world’s oldest stone tools. Journal of Human Evolution 48:109–121.

  • Donadio, Emiliano, and Steven W. Buskirk. 2006. Diet, morphology, and interspecific killing in carnivora. American Naturalist 167:524–536.

  • Feakins, Sarah J., Peter B. deMenocal, and Timothy I. Eglinton. 2005. Biomarker records of late Neogene changes in northeast African vegetation. Geology 33:977–980.

  • Feibel, Craig S. 1997. Debating the environmental factors in hominid evolution. Geological Society of America Today 7(3):1–7.

  • Feibel, Craig S., John M. Harris, and Francis H. Brown. 1991. Palaeoenvironmental context for the late Neogene of the Turkana basin. In The fossil ungulates: geology, fossil artiodactyls, and palaeoenvironments, vol. 3 of Koobi Fora research project. John M. Harris, ed. Pp. 321–370. Oxford: Clarendon.

  • Ferraro, Joseph V. 2007. Broken bones and smashed stones: on the foraging ecology of Oldowan hominins. PhD dissertation, University of California, Los Angeles.

  • Ferring, Reid, Oriol Oms, Jordi Agustí, Francesco Berna, Medea Nioradze, Teona Shelia, Martha Tappen, Abesalom Vekua, David Zhvania, and David Lordkipanidze. 2011. Earliest human occupations at Dmanisi (Georgian Caucasus) dated to 1.85–1.78 Ma. Proceedings of the National Academy of Sciences of the USA 108:10432–10436.

  • Frost, Stephen R. 2007. African Pliocene and Pleistocene cercopithecid evolution and global climatic change. In Hominin environments in the East African Pliocene: an assessment of the faunal evidence. René Bobe, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 51–76. New York: Springer.

  • Ghosh, Prosenjit, Jess Adkins, Hagit Affek, Brian Balta, Weifu Guo, Edwin A. Schauble, Dan Schrag, and John M. Eiler. 2006. 13C-18O bonds in carbonate minerals: a new kind of paleothermometer. Geochimica et Cosmochimica Acta 70:1439–1456.

  • Goldman-Neuman, Talia, and Erella Hovers. 2009. Methodological issues in the study of Oldowan raw material selectivity: insights from A. L. 894 (Hadar, Ethiopia). In Interdisciplinary approaches to the Oldowan. Erella Hovers and David R. Braun, eds. Pp. 71–84. Dordrecht: Springer.

  • ———. 2011. Raw material selectivity in Late Pliocene Oldowan sites in the Makaamitalu basin, Hadar, Ethiopia. Journal of Human Evolution 62:353–366.

  • Grove, Matt. 2011. Speciation, diversity, and Mode 1 technologies: the impact of variability selection. Journal of Human Evolution 61:306–319.

  • Gursky-Doyen, Sharon, and K. A. I. Nekaris, eds. 2006. Primate anti-predator strategies. New York: Springer.

  • Hay, Richard L. 1976. Geology of the Olduvai Gorge. Berkeley: University of California Press.

  • Herries, Andy I. R., Darren Curnoe, and Justin W. Adams. 2009. A multi-disciplinary seriation of early Homo and Paranthropus bearing palaeocaves in southern Africa. Quaternary International 202:14–28.

  • Herries, Andy I. R., and John Shaw. 2011. Palaeomagnetic analysis of the Sterkfontein palaeocave deposits: age implications for the hominin fossils and stone tool industries. Journal of Human Evolution 60:523–539.

  • Hilgen, Frederik J., Hayfaa Abdul Aziz, Wout Krijgsman, Cor G. Langereis, Lucas J. Lourens, Johan E. Meulenkamp, Isabella Raffi, et al. 1999. Present status of the astronomical (polarity) time-scale for the Mediterranean Late Neogene. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 357:1931–1947.

  • Holliday, Trenton W. 2012. Body size, body shape, and the circumscription of the genus Homo. Current Anthropology 53(suppl. 6):S330–S345.

  • Howell, F. Clark, Paul Haesaerts, and Jean de Heinzelin. 1987. Depositional environments, archeological occurrences and hominids from members E and F of the Shungura Formation (Omo basin, Ethiopia). Journal of Human Evolution 16:665–700.

  • Isaac, Glynn L. 1984. The archaeology of human origins: studies of the Lower Pleistocene in East Africa. Advances in World Archaeology 3:1–87.

  • Joordens, Josephine C. A., Hubert B. Vonhof, Craig S. Feibel, Lucas J. Lourens, Guillaume Dupont-Nivet, Jeroen H. J. L. van der Lubbe, Mark J. Sier, Gareth R. Davies, and Dick Kroon. 2011. An astronomically-tuned climate framework for hominins in the Turkana basin. Earth and Planetary Science Letters 307:1–8.

  • Kelt, Douglas A., and Dirk Van Vuren. 1999. Energetic constraints and the relationship between body size and home range area in mammals. Ecology 80:337–340.

  • Kimbel, William H. 2009. The origin of Homo. In The first humans. Frederick E. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 31–37. New York: Springer.

  • Kimbel, William H., Robert C. Walter, Donald C. Johanson, Kaye E. Reed, James L. Aronson, Zelalem Assefa, Curtis W. Marean, et al. 1996. Late Pliocene Homo and Oldowan tools from the Hadar Formation (Kada Hadar member), Ethiopia. Journal of Human Evolution 31:549–561.

  • Kingston, John D. 2007. Shifting adaptive landscapes: progress and challenges in reconstructing early hominid environments. Yearbook of Physical Anthropology 50:20–58.

  • Kingston, John D., Alan L. Deino, Robert K. Edgar, and Andrew Hill. 2007. Astronomically forced climate change in the Kenyan Rift Valley 2.7–2.55 Ma: implications for the evolution of early hominin ecosystems. Journal of Human Evolution 53:487–503.

  • Kuzawa, Christopher W., and Jared M Bragg. 2012. Plasticity in human life history strategy: implications for contemporary human variation and the evolution of genus Homo. Current Anthropology 53(suppl. 6):S369–S382.

  • Leakey, Mary D. 1971. Olduvai Gorge, vol. 3. Cambridge: Cambridge University Press.

  • Lepre, Christopher J., Rhonda L. Quinn, Josephine C. A. Joordens, Carl C. Swisher III, and Craig S. Feibel. 2007. Plio-Pleistocene facies environments from the KBS member, Koobi Fora Formation: implications for climate controls on the development of lake-margin hominin habitats in the northeast Turkana basin (northwest Kenya). Journal of Human Evolution 53:504–514.

  • Leroy, Suzanne, and Lydie Dupont. 1994. Development of vegetation and continental aridity in northwestern Africa during the Late Pliocene: the pollen record of ODP Site 658. Palaeogeography, Palaeoclimatology, Palaeoecology 109:295–316.

  • Levin, Naomi E., Francis H. Brown, Anna K. Behrensmeyer, René Bobe, and Thure E. Cerling. 2011. Paleosol carbonates from the Omo group: isotopic records of local and regional environmental change in East Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 307:75–89.

  • Levin, Naomi E., Jay Quade, Scott W. Simpson, Sileshi Semaw, and Michael Rogers. 2004. Isotopic evidence for Plio-Pleistocene environmental change at Gona, Ethiopia. Earth and Planetary Science Letters 219:93–110.

  • Lewis, Margaret E., and Lars Werdelin. 2007. Patterns of change in the Plio-Pleistocene carnivorans of eastern Africa. In Hominin environments in the East African Pliocene: an assessment of the faunal evidence. René Bobe, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 77–105. New York: Springer.

  • Lordkipanidze, David, Tea Jashashvili, Abesalom Vekua, Marcia S. Ponce de León, Christoph P. E. Zollikofer, G. Philip Rightmire, Herman Pontzer, et al. 2007. Postcranial evidence from early Homo from Dmanisi, Georgia. Nature 449:305–310.

  • Maslin, Mark A., and Martin H. Trauth. 2009. Plio-Pleistocene East African pulsed climate variability and its influence on early human evolution. In The first humans: origin and early evolution of the genus Homo. Frederick E. Grine, John G. Fleagle, and Richard E. Leakey, eds. Pp. 151–158. New York: Springer.

  • Migliano, Andrea Bamberg, and Myrtille Guillon. 2012. The effects of mortality, subsistence, and ecology on human adult height and implications for Homo evolution. Current Anthropology 53(suppl. 6):S359–S368.

  • Nagy, Ken A., Isabelle A. Girard, and Tracey K. Brown. 1999. Energetics of free-ranging mammals, reptiles, and birds. Annual Review of Nutrition 19:247–277.

  • Owen, R. Bernhart, Richard Potts, and Anna K. Behrensmeyer. 2009. Reply to the comment on “Diatomaceous sediments and environmental change in the Pleistocene Olorgesailie Formation, southern Kenya Rift Valley” by R. Bernhart Owen, Richard Potts, Anna K. Behrensmeyer, and Peter Ditchfield. Palaeogeography, Palaeoclimatology, Palaeoecology 282:147–148.

  • Passey, Benjamin H., Naomi E. Levin, Thure E. Cerling, Francis H. Brown, and John M. Eiler. 2010. High-temperature environments of human evolution in East Africa based on bond ordering in paleosol carbonates. Proceedings of the National Academy of Sciences of the USA 107:11245–11249.

  • Pickering, Robyn, Paul H. G. M. Dirks, Zubair Jinnah, Darryl J. de Ruiter, Steven E. Churchill, Andy I. R. Herries, Jon D. Woodhead, John C. Hellstrom, and Lee R. Berger. 2011. Australopithecus sediba at 1.977 Ma and implications for the origins of the genus Homo. Science 333:1421–1423.

  • Plummer, Thomas W. 2004. Flaked stones and old bones: biological and cultural evolution at the dawn of technology. Yearbook of Physical Anthropology 47:118–164.

  • Plummer, Thomas W., Peter W. Ditchfield, Laura C. Bishop, John D. Kingston, Joseph V. Ferraro, David R. Braun, Fritz Hertel, and Richard Potts. 2009. Oldest evidence of toolmaking hominins in a grassland-dominated ecosystem. PLoS ONE 4(9):e7199, doi:10.1371/journal.pone.0007199.

  • Pontzer, Herman. 2012. Ecological energetics in early Homo. Current Anthropology 53(suppl. 6):S346–S358.

  • Pontzer, Herman, Campbell Rolian, G. Philip Rightmire, Tea Jashashvili, Marcia S. Ponce de León, David Lordkipanidze, and Christoph P. E. Zollikofer. 2010. Locomotor anatomy and biomechanics of the Dmanisi hominins. Journal of Human Evolution 58:492–504.

  • Potts, Richard. 1988. Early hominid activities at Olduvai. New York: Aldine.

  • ———. 1991. Why the Oldowan? Plio-Pleistocene toolmaking and the transport of resources. Journal of Anthropological Research 47:153–176.

  • ———. 1996a. Evolution and climate variability. Science 273:922–923.

  • ———. 1996b. Humanity’s descent: the consequences of ecological instability. New York: Morrow.

  • ———. 1998a. Environmental hypotheses of hominin evolution. Yearbook of Physical Anthropology 41:93–136.

  • ———. 1998b. Variability selection in hominid evolution. Evolutionary Anthropology 7:81–96.

  • ———. 2002. Complexity and adaptability in human evolution. In Probing human origins. Morris Goodman and Anne Moffat, eds. Pp. 33–57. Cambridge, MA: American Academy of Arts and Sciences.

  • ———. 2007. Environmental hypotheses of Pliocene human evolution. In Hominin environments in the East African Pliocene: an assessment of the faunal evidence. René Bobe, Zeresenay Alemseged, and Anna K. Behrensmeyer, eds. Pp. 25–50. New York: Springer.

  • Potts, Richard, and Pat Shipman. 1981. Cutmarks made by stone tools on bones from Olduvai Gorge, Tanzania. Nature 291:577–580.

  • Potts, Richard, and Robin Teague. 2010. Behavioral and environmental background to “Out of Africa I” and the arrival of Homo erectus in East Asia. In Out of Africa I: the first hominin colonization of Eurasia. John G. Fleagle, John J. Shea, Frederick E. Grine, Andrea L. Baden, and Richard E. Leakey, eds. Pp. 67–85. New York: Springer.

  • Quade, Jay, Naomi Levin, Sileshi Semaw, Dietrich Stout, Paul Renne, Michael Rogers, and Scott Simpson. 2004. Paleoenvironments of the earliest stone toolmakers, Gona, Ethiopia. Geological Society of America Bulletin 116:1529–1544.

  • Quade, Jay, Naomi E. Levin, Scott W. Simpson, Robert Butler, William C. McIntosh, Sileshi Semaw, Lynnette Kleinsasser, Guillaume Dupont-Nivet, Paul Renne, and Nelia Dunbar. 2008. The geology of Gona, Afar, Ethiopia. In The geology of early humans in the Horn of Africa. Jay Quade and Jonathan C. Wynn, eds. Special Paper 446. Pp. 1–31. Boulder, CO: Geological Society of America.

  • Quinn, Rhonda L., Christopher J. Lepre, James D. Wright, and Craig S. Feibel. 2007. Paleogeographic variations of pedogenic carbonate δ13C values from Koobi Fora, Kenya: implications for floral compositions of Plio-Pleistocene hominin environments. Journal of Human Evolution 53:560–573.

  • Reed, Kaye E. 1997. Early hominid evolution and ecological change through the African Plio-Pleistocene. Journal of Human Evolution 32:289–322.

  • ———. 2008. Paleoecological patterns at the Hadar hominin site, Afar Regional State, Ethiopia. Journal of Human Evolution 54:743–768.

  • Rossignol-Strick, Martine. 1983. African monsoons, an immediate climate response to orbital insolation. Nature 304:46–49.

  • Ruddiman, William F. 2001. Earth’s climate: past and future. New York: W. H. Freeman.

  • Ruff, Christopher B., Erik Trinkaus, Alan Walker, and Clark Spencer Larsen. 1993. Postcranial robusticity in Homo. 1. Temporal trends and mechanical interpretation. American Journal of Physical Anthropology 91:21–53.

  • Ruff, Christopher B., and Alan Walker. 1993. Body size and body shape. In The Nariokotome Homo erectus skeleton. Alan C. Walker and Richard E. Leakey, eds. Pp. 234–265. Cambridge, MA: Harvard University Press.

  • Sahnouni, Mohamed, Djillali Hadjouis, Jan van der Made, Abd-El-Kader Derradji, Antoni Canals, Mohamed Medig, Hocine Belahrech, Zoheir Harichane, and Merouane Rabhi. 2002. Further research at the Oldowan site of Aïn Hanech, north-eastern Algeria. Journal of Human Evolution 43:925–937.

  • Schwartz, Gary T. 2012. Growth, development, and life history throughout the evolution of Homo. Current Anthropology 53(suppl. 6):S395–S408.

  • Semaw, Sileshi, Michael J. Rogers, Jay Quade, Paul R. Renne, Robert F. Butler, Manuel Domínguez-Rodrigo, Dietrich Stout, William S. Hart, Travis Pickering, and Scott W. Simpson. 2003. 2.6-million-year-old stone tools and associated bones from OGS-6 and OGS-7, Gona, Afar, Ethiopia. Journal of Human Evolution 45:169–177.

  • Sepulchre, Pierre, Gilles Ramstein, Frédéric Fluteau, Mathieu Schuster, Jean-Jacques Tiercelin, and Michel Brunet. 2006. Tectonic uplift and eastern Africa aridification. Science 313:1419–1423.

  • Shipman, Pat, and Alan Walker. 1989. The costs of becoming a predator. Journal of Human Evolution 18:373–392.

  • Stout, Dietrich, Jay Quade, Sileshi Semaw, Michael Rogers, and Naomi E. Levin. 2005. Raw material selectivity of the earliest stone tool makers at Gona, Afar, Ethiopia. Journal of Human Evolution 48(4):365–380.

  • Stout, Dietrich, Sileshi Semaw, Michael J. Rogers, and Dominique Cauche. 2010. Technological variation in the earliest Oldowan from Gona, Afar, Ethiopia. Journal of Human Evolution 58:474–491.

  • Trauth, Martin H., Juan C. Larrasoaña, and Manfred Mudelsee. 2009. Trends, rhythms and events in Plio-Pleistocene African climate. Quaternary Science Reviews 28:399–411.

  • Trauth, Martin H., Mark A. Maslin, Alan L. Deino, Annett Junginger, Moses Lesoloyia, Eric O. Odada, Daniel O. Olago, Lydia A. Olaka, Manfred R. Strecker, and Ralph Tiedemann. 2010. Human evolution in a variable environment: the amplifier lakes of eastern Africa. Quaternary Science Reviews 29:2981–2988.

  • Trauth, Martin H., Mark A. Maslin, Alan L. Deino, and Manfred R. Strecker. 2005. Late Cenozoic moisture history of East Africa. Science 309:2051–2053.

  • Trauth, Martin H., Mark A. Maslin, Alan L. Deino, Manfred R. Strecker, Andreas G. N. Bergner, and Miriam Dühnforth. 2007. High- and low-latitude forcing of Plio-Pleistocene East African climate and human evolution. Journal of Human Evolution 53:475–486.

  • Vrba, Elisabeth S. 1985. Environment and evolution: alternative causes of the temporal distribution of evolutionary events. South African Journal of Science 81:229–236.

  • ———. 1988. Late Pliocene climatic events and hominid evolution. In Evolutionary history of the “robust” australopithecines. Frederick E. Grine, ed. Pp. 405–426. New York: Aldine.

  • ———. 1992. Mammals as a key to evolutionary theory. Journal of Mammalogy 73:1–28.

  • ———. 1994. An hypothesis of heterochrony in response to climatic cooling and its relevance to early hominid evolution. In Integrative paths to the past. Robert S. Corruccini and Russell L. Ciochon, eds. Pp. 345–376. Englewood Cliffs, NJ: Prentice Hall.

  • ———. 1995a. The fossil record of African antelopes (Mammalia, Bovidae) in relation to human evolution and paleoclimate. In Paleoclimate and evolution, with emphasis on human origins. Elizabeth S. Vrba, George H. Denton, Timothy C. Partridge, and Lloyd H. Burckle, eds. Pp. 385–424. New Haven, CT: Yale University Press.

  • ———. 1995b. On the connections between paleoclimate and evolution. In Paleoclimate and evolution, with emphasis on human origins. Elizabeth S. Vrba, George H. Denton, Timothy C. Partridge, and Lloyd H. Burckle, eds. Pp. 24–45. New Haven, CT: Yale University Press.

  • Vrba, Elizabeth S., George H. Denton, Timothy C. Partridge, and Lloyd H. Burckle, eds. 1995. Paleoclimate and evolution, with emphasis on human origins. New Haven, CT: Yale University Press.

  • Wehausen, Rolf, and Hans-Jürgen Brumsack. 1999. Cyclic variations in the chemical composition of eastern Mediterranean Pliocene sediments: a key for understanding sapropel formation. Marine Geology 153:161–176.

  • Wynn, Jonathan G. 2000. Paleosols, stable carbon isotopes, and paleoenvironmental interpretation of Kanapoi, Northern Kenya. Journal of Human Evolution 39:411–432.

  • Zachos, James, Mark Pagani, Lisa Sloan, Ellen Thomas, and Katharina Billups. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292:686–693.

  • Zhu, Rixiang, Richard Potts, Fei Xie, Kenneth A. Hoffman, Chenglong L. Deng, C. D. Shi, Yongxin X. Pan, et al. 2004. New evidence regarding the earliest human presence at high northern latitudes in northeast Asia. Nature 431:559–562.


Richard Potts is Director of the Human Origins Program, National Museum of Natural History, Smithsonian Institution (P.O. Box 37012, Washington, DC 20013-7012, U.S.A. []).