Project 3

Cognition in Marginal Habitats

The challenge of marginal habitats in understanding complex behavior in the Pleistocene
Fynbos biome in South Africa. Credit: Chris Eason/Wikimedia

Principal Investigators

Project 3 takes the link between cognition and environmental variation forward in time by probing the relationship between cognitive evolution, as read through lithic technology, and the exploitation of resources in a marginal habitat called fynbos in the middle Pleistocene of southern Africa.

In exploring the challenges of occupying marginal habitats in the early-middle Pleistocene, this project addresses questions concerning of the relationship between cognitive development in the early record of the genus Homo and its links to cumulative cultural evolution.

Detailed Project Description

While the Pleistocene increase in brain size is one of the most well documented phenomena in the evolution of our own genus, Homo, our understanding of the ecological and behavioral/cognitive factors underlying this adaptive trend remains abstract. Reallocation of hominin energy budgets associated with the shift to a high meat-protein diet, presumably via hunting, have been implicated in the apparent jump in brain size between “early” Homo and H. erectus. Environmental reconstructions, however, indicate similar woodland-grassland savanna conditions at H. erectus sites in eastern (e.g., Turkana Basin, Olduvai Gorge) and southern Africa (e.g., Swartkrans). This implies highly productive habitats with abundant prey species and opportunities for competition with other predators—which would have placed a premium on the cognitive skills of early foraging populations. However, we know very little about the distribution of early Homo populations across the range of available early-middle Pleistocene African habitats or how habitat heterogeneity may have influenced their adaptations.

The research proposed in this project aims to address this gap in knowledge of early Homo habitat use by investigating whether and how early Homo species occupied environments that can be described as “marginal” or “low quality” with respect to the species’ core resource extraction strategies. The selection for novel planning strategies associated with expansion into such challenging habitats has been suggested as one factor potentially associated with increases in brain size. Thus, our research may have implications for the timing and context of cumulative cultural evolution in Pleistocene Homo, which predicts that transgenerational transmission of knowledge about the complex behaviors expected in marginal habitats should lead to locally rapid change and ecological differentiation of technology and other archaeological proxies.

In connection with expanded field work in the later Pleistocene of southern Africa (Project 5), we propose to address the issue of marginal habitats of Pleistocene hominin occupation through 1) new paleoanthropological field work on the southwestern coast of South Africa, and 2) comparison of our results with data from known early Homo sites in eastern and southern Africa.

The fynbos habitat of southern Africa

Today, the southwestern coast of South Africa is home to a unique vegetation community called fynbos. Fynbos is characterized by a (southern hemisphere) winter rather than summer rainfall and rare endemic succulent plants that support few large mammals; nutritious parts of most plants are mainly underground storage organs. The fynbos habitat would have presented a behavioral challenge to hominins whose main resource adaptation was to large game animals and above-ground plant resources. Fynbos is proposed to have been in place from the beginning of the Pliocene, at ~5 Ma, but at various times, including the late Pleistocene, low sea levels and greater summer rainfall caused the spread of grasslands along coastal regions where fynbos now resides.

In 2009, an IHO field survey discovered Acheulean tools and fossil fauna at Cliff Point, along the western coast of South Africa, in approximately 30 open-air localities, ranging in age from ~1.9 to ~0.1 Ma. This time period overlaps H. erectus-bearing sites in other parts of Africa. Today, the Cliff Point Region is firmly within the winter rainfall regime and has nothing but fynbos vegetation. Several lines of evidence can be used to reconstruct the vegetation structure of the past including fauna, pollen, and carbon isotopes. Through integrated analyses of our preliminary survey data—consisting of archeological, geological, and paleontological evidence—we can determine the habitat with which these hominins were most closely associated. Many of the sites that we surveyed are above or below a calcrete soil layer (evidence of a drying event) and within paleosols that contain root casts of much larger plants than exist in the region today. These paleosols contain fossilized termite mounds, which are indicative of non-fynbos habitats, and the fossils that we have noted so far (no collection was made during the survey) includes mammalian taxa quite different from those found in the region today. This suggests the presence of variable habitats throughout the early to middle Pleistocene of the southwestern coast.

The sites in the Cliff Point area may provide evidence on whether Pleistocene hominin populations were consistently associated with wetter, more productive woodland-savannah habitats or were also able to exploit the marginal resources associated with the unusual fynbos vegetation community. Focused excavations are necessary to provide information on the habitat and resource utilization of the resident hominin populations.


Field work will focus on understanding associations among fossils, lithic artifacts, and the geological history of the sediments in the Cliff Point area. We will begin with a detailed examination of two sites in the northern part of the survey region.

We will also gather data on artifacts, isotopes, pollen, and other habitat proxies from other early Homo localities. These data will be compiled in a relational database with which we can compare with the results of field-collections.

The first locality (CP-510 – 31°30'42.13"S, 18° 2'56.56"E) contains in situ fossils and termite mounds, which are present above a calcrete layer. Magnetostratigraphy—the sediments here yield a reversed polarity signal—and fauna are consistent with an age of >0.78 Ma. Acheulean bifaces and flakes are visible on the surface and appear to be in situ. At CP-510, excavations will follow protocols outlined by McPherron and colleagues to accurately collect 3D position as well as fabric analysis data (slope, orientation) on all specimens larger than 1 cm. These data will be used to understand whether fossils and artifacts have undergone deflationary processes. All specimens will be collected and assigned preprinted barcodes that will be entered using a handheld bar code scanner; this will drastically reduce database errors and provide the highest resolution of data. The excavated material will be integrated into a geospatial database as their precise locations will be rendered in high accuracy UTM coordinates. Fossil dentitions will be used to reconstruct diets of the animals using both functional traits and isotopes. Functional analyses of postcranial articular elements will be used in habitat reconstructions, and the macrofauna will be used to construct a community-level profile of the local habitats. Excavations at CP-510 will be directed by A. Herries, W. Archer, and D. Braun.

The second locality (CP-537 – 31°30'44.85"S, 18° 2'55.68"E) is a channel fill situated above a calcrete. Here the calcrete has normal polarity, so this site could range either from 1.0-0.9 Ma or 2.0-1.8 Ma, both of which correspond to the presence of early Homo, including H. erectus, in other parts of Africa. At CP-537, we will dry sieve the channel fill with .25 inch mesh screens. All macro- and microfossils will be collected. This includes long bone shafts and other fragments that are not usually collected, as we will examine them for cut marks, carnivore damage, etc. Each fossil will be identified to the lowest taxonomic group possible. We will also sample the calcrete and termite mounds for carbon isotope analysis, as these will provide data on: a) the dominant vegetation (from isotopes in the calcrete); and b) the dominant grass type eaten by the termites (from isotopes in the termite mound). A predominantly C4 isotopic signal would suggest the presence grasses relying on a summer rainfall regime. All analyses listed above for fossils at CP-510 will be followed in the work at CP-537. Fieldwork and analyses at CP-537 will be led by K. Reed and G. Schwartz.

At both sites, R. Pickering will attempt uranium-lead isotope dating on samples of paleosol and termite calcretes. She has previously analyzed these types of materials in the region with some success.