On-line version ISSN 1996-7489
Print version ISSN 0038-2353
S. Afr. j. sci. vol.109 n.11-12 Pretoria Jan. 2013
Nikolaas J. van der Merwe
Department of Archaeology, University of Cape Town, Cape Town, South Africa
Light stable isotope ratios (δ13C and δ18O) of tooth enamel have been widely used to determine the diets and water sources of fossil fauna. The carbon isotope ratios indicate whether the plants at the base of the food web used C3 or C4 photosynthetic pathways, while the oxygen isotope ratios indicate the composition of the local rainfall and whether the animals drank water or obtained it from plants. The contrasting diets of two early hominin species - Homo habilis and Paranthropus boisei - of ca 1.8 Ma (million years ago) in Tanzania were determined by means of stable carbon isotope analysis of their tooth enamel in a previous study. The diets of two specimens of P boisei, from Olduvai and Peninj, proved to be particularly unusual, because 80% of their carbon was derived from C4 plants. It was suggested that their diet consisted primarily of plants, with particular emphasis on papyrus, a C4 sedge. The dominance of C4 plants in the diet of P boisei is a finding supported in another study of 22 specimens from Kenya. The isotopic ecology and diets of fossil fauna that were present at the same time as the two fossil hominin species are described here, in order to provide a fuller understanding of their contrasting diets and of the moisture sources of their water intake. This information was then compared with the isotopic composition of modern fauna from the same region of Tanzania. The carbon isotope ratios for both fossil and modern specimens show that the habitats in which these faunal populations lived were quite similar - grassland or wooded grassland. They had enough bushes and trees to support a few species of browsers, but most of the animals were grazers or mixed feeders. The oxygen isotope ratios of the fossil and modern fauna were, however, very different, suggesting strongly that the source of moisture for the rain in the Olduvai region has changed during the past 1.8 million years.
Keywords: carbon isotopes; oxygen isotopes; Olduvai fossils; moisture sources; vertebrate ecology
The purpose of this article is to provide isotopic information on the ecology and diets of fossil fauna from Olduvai Gorge Beds I and II at ca 1.8 Ma, in order to provide an ecological context for the diets of two species of early hominins. Recent isotopic analyses of the tooth enamel of Homo habilis and Paranthropus boisei (formerly Zinjanthropus boisei, also Australopithecus boisei) showed that the two species had distinctly different diets.13
Three specimens of Homo habilis from Olduvai had, respectively, 23%, 27% and 49% of carbon derived from C4 plants, much like early hominins from South Africa4-6; their diets probably included grass-eating animals and/or insects. However, two specimens of P. boisei from Olduvai and Peninj had 77% and 81% of carbon derived from C4 plants. Because modern humans are limited to about 20-50% protein-rich foods for their energy requirements,7 it was suggested that the diet of P. boisei included a large component of C4 plants. As grasses, especially edible seeds, are highly seasonal at latitudes applicable to Olduvai, it was also suggested that the C4 sedge Cyperus papyrus, which was presumably available in the freshwater swamps at Lake Olduvai and Lake Natron (Peninj), may have been a major component in the diet of P. boisei. The very high contribution of C4 plants to the diet of this hominin has been confirmed by the isotopic analysis of 22 individuals from sites in Kenya that stretch over 700 km of the Rift Valley.3 The authors of the latter study suggested that P. boisei had a diet that comprised mostly C4 plants, without specifying whether these were grasses or sedges. An online comment on their article by Lee-Thorp8 supports the opinion that sedges were predominant in the diet. Recent publications provide further evidence to put these early hominin diets in context.9,10
An assessment of the isotopic ecology at Peninj and Olduvai during the presence of P. boisei cannot settle the argument about grasses and sedges in P. boisei's diet, but it can illuminate whether the environment was dominated by C4 plants and their consumers. This assessment is of particular importance at Olduvai around 1.8 Ma, for which the isotope values of the two hominins are available. In East Africa, C4 plants appeared in the Late Miocene and continued through the Pliocene,11-15 but did not become dominant until ca 1.8 Ma.16 At Olduvai, the carbon isotopes in palaeosol carbonates indicate that C4 plants made up about 40-60% of the biomass during the time of Beds I and II,17 but a preliminary study of the carbon isotopes in tooth enamel of fossil fauna indicate a much higher C4 component.16,18 An assessment of the isotopic ecology at Peninj was undertaken by measuring the carbon and oxygen isotopes in the tooth enamel of 40 specimens of fossil fauna from the Maritinane Type Section.2 These fossils were all grazers and the carbon isotopes were closely similar to those of modern fauna of related species from the modern Serengeti. The environment was essentially open grassland with very few trees. However, the fossil specimens were some 1.3 million years old, so the results do not provide information about the ecology of Olduvai at ca 1.8 Ma. It was noted that the 518O values of the Peninj fossils were distinctly negative, relative to the Vienna Pee Dee Belemnite (VPDB) standard, while those of the modern fauna were all positive. It was suggested that the moisture source of rain in this region had presumably changed during the past 1.3 million years; this suggestion is addressed in detail here.
For this purpose, tooth enamel samples were obtained from 145 specimens of fossil fauna from Olduvai Gorge Middle Bed I (ca 1.785-1.83 Ma) and Lowermost Bed II (ca 1.75-1.83 Ma) and their 513C and 518O values were measured. These values were then compared with those of 77 modern animals from six wildlife reserves, primarily Serengeti and Maswa, but also Lobo, Lukwati, Ugalla and Selous (Figure 1). Their carbon and oxygen isotope ratios were measured to illuminate their diets and water sources.
Tooth enamel and stable light isotopes
The assessment of prehistoric diets and environments by means of isotopic analysis of bone has been developed over the past 30 years and is widely used in archaeology19-21 (for review see Van der Merwe21). In the early development of this method, the major interest involved 13C/12C ratios (513C values) in bone collagen, which provide a measure of the proportion of C3 and C4 plants at the base of the food web. However, collagen in bone has a limited lifetime, especially in hot and humid environments where organic materials deteriorate rapidly. The oldest hominin collagen specimens that have been analysed isotopically were those of Neanderthals from cold, dry caves.22,23 The fossilised faunal specimens from Olduvai contain no collagen, but carbon and oxygen isotopes can be measured in the mineral phase of their skeletons. The earliest measurements on bone apatite were done on fossil bone,24,25 but tooth enamel has proved to be the most dependable material for isotopic analysis.26,27 Tooth enamel is highly crystalline and resists alteration by carbonates in groundwater. Contaminating carbonates may precipitate in cracks in the tooth enamel, but this material is much more soluble than tooth enamel and is readily removed with dilute acetic acid.
Tooth enamel is a biological apatite with about 3% carbonate. With appropriate pretreatment, the carbon and oxygen isotope ratios of the carbonate can be measured with confidence to provide dietary information. The carbon isotope ratio is an average of the plants at the base of the food web, acquired by herbivores from eating the plants and passed along the food chain to omnivores and carnivores, with some digestive differences between different species.28 The oxygen isotope ratio is a measure of the body water of an animal, acquired from plant water or from drinking water and is altered by the thermophysiology of the animal. The oxygen isotope ratio can contribute to dietary and environmental reconstructions9,29,30 and is particularly indicative of changes in humidity and aridity.31-33
To reach dietary conclusions based on carbon isotope ratios, it is necessary to determine the C3 and C4 end members for a given time and place. This determination is most readily achieved by measuring the tooth enamel δ13C values of dedicated browsers and grazers. Giraffes, for example, rarely eat anything but C3 leaves of trees and shrubs, while alcelaphine antelope like the wildebeest concentrate on C4 grasses. At this point in time, dedicated browsing herbivores of the South African interior (whose tooth enamel the Cape Town laboratory has often analysed) have mean δ13C values of -14.5‰, while dedicated grazers have mean values of -0.5‰. These values can be altered by changes in climate and in the atmosphere. The 'industrial effect' of the past 200 years, for example, substantially increased the CO2 content of the atmosphere, as a result of fossil fuel burning, resulting in the δ13C value of the atmosphere (and that of all plants) becoming more negative by 1.5‰. The δ13C values of all modern specimens must consequently be corrected by adding 1.5‰ to the measured number. Climatic changes that alter δ13C values of plants include increased humidity, which may make the δ13C values of C3 plants (but not C4) more negative by 2‰.34 Tropical forests provide an extreme example in this regard - their C3 plants (they have no C4 plants) have δ13C values as much as 10‰ more negative than those of plants growing in the open; this difference results from high humidity, low light and, especially, from recycled CO2 produced by rotting leaf litter on the forest floor.35 In contrast, the δ13C values may become more positive as a result of aridity and bright sunlight,36,37 while C4 plants may respond with slightly more negative values, as a result of the increased occurrence of enzymatic C4 subtypes that are adapted to such conditions.
The approximate ratio of C3 and C4 plants in the diet of herbivores can be arrived at by interpolating between the C3 end member (100% C3 diet) and C4 end member (100% C4 diet) of the ecosystem under study, although there are some slight differences between different species that have the same diets.
Oxygen isotope ratios in tooth enamel can help to understand certain elements of a local habitat. The primary source of oxygen in biological apatite is from water (drinking water or water in food) and from oxygen bound in food. The light 16O isotope evaporates more readily than the heavy 18O isotope, while the latter precipitates more readily. These characteristics have produced such localised effects as the water in modern East African lakes being enriched by 5-10‰ compared to the waters flowing into them.17 Ultimately, the 18O content of water in an area depends on precipitation - or meteoric water (δ18Omw) - which decreases with distance from the source of moisture (an ocean or lake), increasing latitude, increasing altitude and decreasing temperature.38-42 An example of the different moisture sources can be seen in East Africa, where δ18O values of waters in Kenya and Ethiopia differ by about 2-3‰.43 The δ18O value of leaf water derived from the same water source also varies with aridity, because of evapotranspiration.
The integrity of δ13C and δ18O values in tooth enamel for a given time and place can best be judged by observing the pattern of results for all the animal species in the biome. A considerable database for the Plio-Pleistocene of Africa is available by now and can be used to judge the results of a given study. Browsing animals (C3 plant feeders) will have distinctly more negative δ13C values than grazers (C4 plant feeders), with the most dedicated browsers and grazers differing by about 14‰. There is also a distinct pattern between different species of grazers, depending on the different amounts of C3 plants like forbs that are included in their diets. Wherever they have been compared, specimens of Damaliscus spp. (e.g. the tsessebe and topi) have more positive δ13C values than Connochaetes spp. (wildebeest), which, in turn, have more positive values than Equus spp. (zebra). For oxygen isotopes, δ18O values of hippos are usually more negative than those of equids, because they drink surface water that is closest to meteoric water and feed at night, when humidity increases and plant δ18O decreases.44
Materials and methods
Samples of tooth enamel used for isotopic analysis in this study were removed from the tooth specimens at their location of storage. In the case of Olduvai Beds I and II, this sampling was done at the National Museum of Natural History in Arusha and at the Olduvai field station (Mary Leakey's old field camp, which has been expanded). Most of the teeth from Olduvai Beds I and II had been excavated by the Olduvai Landscape Palaeoanthropology Project (OLAPP) - an international group of researchers who have been working at Olduvai during the midyear field season since 1989.45 Specimens from the OLAPP collections can be identified in Tables 1 and 2 on the basis of their catalogue numbers, which are in the format 95/43:156 (excavated 1995, trench 43, number 156). Some specimens from Mary Leakey's collection, which is stored at the Olduvai field station, were also sampled. These are from Bed II (Table 2), excavation sites FLK N and HWKE. In total, 145 faunal specimens were sampled from the collections of Olduvai Middle Bed I (ca 1.785-1.83 Ma) and Lowermost Bed II (ca 1.75-1.78 Ma).
The modern samples from the Serengeti National Park, Maswa Game Reserve (southern Serengeti), Lobo (northeast Serengeti) and Selous Game Reserve (southeast Tanzania) were obtained from collections at Seronera, the headquarters of Serengeti National Park, and from some of the other ranger stations in the park. A few specimens from Lukwati and Ugalla Game Reserves were sampled in Arusha at a taxidermy shop and at the Rock Art Centre. A collection of 77 modern bone specimens, primarily from Serengeti National Park which adjoins Olduvai, was also assembled (120 measurements).
Samples were removed from tooth specimens by using a variable speed drill with a dental diamond drill tip of 0.5-mm diameter. A sample of 3 mg enamel powder is sufficient for at least two isotopic analyses. This amount is the equivalent of about two sugar grains; if it is removed from a broken enamel edge, the sampling scar is usually invisible. By moving the drill tip along an enamel edge, broken along the length of a tooth, seasonal variations (particularly in 18O) are averaged. The enamel powder was gathered on smooth weighing paper and poured into a small centrifuge vial with snap lid, in which all the subsequent chemical pretreatment was carried out at the Stable Light Isotope Facility of the University of Cape Town. The powder was pretreated with 1.5-2.0% sodium hypochlorite (to remove organic materials and humic acids), rinsed with water, and then reacted for 15 min with 0.1 M acetic acid (to remove readily dissolved carbonates).
After washing and drying, 1 mg of the enamel was weighed into an individual reaction vessel of a Kiel II autocarbonate device (Thermo Electron Corporation, Bremen, Germany). The powder sample was reacted at 70°C with 100% phosphoric acid, which produces CO2 from the carbonate in the enamel. This sample gas was cryogenically distilled and its isotope ratios were measured in a Finnegan MaT 252 ratio mass spectrometer (Thermo Electron Corporation, Bremen, Germany). The δ13C and δ18O were calibrated against the international VPDB carbonate standard by using a calibration curve established from National Bureau of Standards standards 18 and 19, as well as by inserting internal laboratory standards of 'Lincoln limestone' and 'Carrara 2 marble' at regular intervals in the autocarbonate device. The precision of replicate analyses was better than 0.1 ‰.
The δ13C and δ18O values for fossil fauna from Olduvai West Middle Bed I (Table 1) and Olduvai East Lowermost Bed II (Table 2) are combined in Figure 2 to illustrate the faunal community in the Olduvai region at ca 1.8 Ma. These values can be compared with the isotope ratios of modern fauna from the adjoining Serengeti National Park, plus some specimens from Maswa, Lobo, Ugalla, Lukwati and Selous wildlife reserves (Table 3, Figure 3). The δ13C and δ18O values of the fossil and modern fauna are compared in Figures 4 and 5, respectively.
In general, the carbon isotope ratios of the same or related species have stayed remarkably similar over 1.8 million years. The available plant foods have obviously stayed much the same. The oxygen isotope ratios, in contrast, have become enriched by some 6‰. One obvious difference can be found in the case of hippos: the 11 specimens of Hippopotamus gorgops from Beds I and II were dedicated grazers, while the single modern specimen of Hippopotamus amphibious from the Serengeti was a mixed feeder. Hippos are generally regarded as dedicated grazers, but this assumption is not correct44 - they will eat C3 plants if grasses are not available nearby.
The collection from Olduvai West Bed I is somewhat limited: the 27 specimens comprise 1 black-backed jackal, 6 crocodiles and 20 grazing animals (suids and bovids). No browsers are included, which makes it impossible to determine the C3 end member of this faunal community. At the C4 end of the collection, the most positive δ13C value is that of + 1.1‰ for two specimens of Parmularius (an alcelaphine). The δ13C values from the Bed I specimens are very similar to those from the larger collection of 118 specimens from Bed II, which includes canids, suids, hippotragines, antelopines, alcelophines and crocodylids. The values for Beds I and II have been added together in the figures to represent the fossil fauna.
Olduvai Bed II is represented by 118 specimens, which provide a comprehensive view of the community at ca 1.8 Ma. The browsing end of the spectrum is represented by two specimens of Deinotherium bozasi (-11.05; -10.32), a giant browsing elephant with tusks growing from its lower jaw and bent downwards towards the ground. A specimen of giraffe, Giraffa sp. (probably gracilis), has a less negative δ13C value of -9.1‰. It is of interest that Elephas recki, an elephant species that occurred widely from East to southern Africa at this time, was clearly a grazer (δ13C = + 1.3), while Sivatherium sp. (also Libytherium sp.), a short-necked grazing giraffe (δ13C = -1.5), was also present during Bed II times. These species of elephant and giraffe were later replaced by the modern dedicated browsing elephant, Loxodonta africana, and giraffe, Giraffa camelopardis. The grazing end of the spectrum at Olduvai in Bed II times was represented by alcelaphines of the genera Beatragus/ Connochaetes (δ13C = +2.1), Parmularius (+1.9) and Megalotragus (+2.7). Based on these values and that of Deinotherium bozasi, one can estimate the δ13C values for the C3 and C4 end members for Bed II at about -11.5‰ and +2.5‰, respectively.
It should be noted that the Bed II specimens are clustered toward the grazing end of the spectrum with few browser representatives. This pattern is especially evident when the Bed II assemblage is compared with that of the modern Serengeti, which has a similar scarcity of browsers. It is noteworthy that the tragelaphines (kudu, bushbuck and eland) from Bed II have a mean δ13C value of -5.0‰. They were evidently mixed feeders, in contrast to the modern kudu and bushbuck from Maswa, which are dedicated browsers with a mean δ13C value of -14.7‰.
The δ13C values for modern fauna (Table 3) have not been corrected for the industrial effect. In order to compare them with the δ13C values for Olduvai Beds I and II, it is necessary to add 1.5‰. Figure 4 shows the corrected values.
The δ13C values of modern animals (Table 3) represent several different biomes in Tanzania. This multi-representation is immediately obvious when one compares the δ13C values for Equus burchelli (plains zebra) from Serengeti National Park (+0.8) with those from Lukwati (-2.0). Lukwati is located near Lake Rukwa, between Lakes Tanganyika and Nyasa, and its grazers clearly have more C3 plants in their diet. Similarly, the browsers from the Serengeti include the modern giraffe (Giraffa camelopardis, δ13C = -12.7), elephant (Loxodonta africana, δ13C = -11.9) and black rhino (Diceros bicornis, δ13C = -13.9). Maswa Game Reserve, which borders Serengeti on the southwest side but is much more wooded, is represented among the browsers by three tragelaphines: the greater kudu (Tragelaphus strepsiceros, δ13C = -15.4), lesser kudu (T. imberbes, δ13C = -12.6) and bushbuck (T. scriptus, δ13C = -15.1). The grazing end of the Serengeti spectrum is represented by Hippotragus equinus (roan antelope, δ13C = +1.9), Connochaetes gnu (black wildebeest, δ13C = +1.7) and Alcelaphus lichtensteini (Lichtenstein's hartebeest, δ13C = +2.0). In calculating the C3 and C4 end members, the Serengeti specimens have been emphasised, yielding approximately -12.5 and +1.5. When corrected for the industrial effect, the end members are about -11.0 and +3.0. This result is similar to that from Olduvai Bed II; however, similarity does not mean that the plant communities of the two biomes were the same, only that the availability of C3 and C4 plants for browsers and grazers were similar.
In Figure 1, a selection of tooth enamel δ13C values from Olduvai Bed II and the modern Serengeti are compared; the modern values have been corrected for the industrial effect. The results are extraordinarily similar, suggesting two similar landscapes of wooded grassland. The exception is a single δ13C value for a modern hippo from the Serengeti, which is evidently a mixed feeder.
From the δ18O values reported in Tables 1, 2 and 3, it is clear that oxygen isotope ratios vary substantially with time and place. Among 27 measurements for Olduvai West Bed I (27 animals, Table 1) there is only a single positive value for δ18O. Among 124 measurements for Olduvai East Bed II (111 animals, Table 2) there are only 10 positive δ18O values. Among 77 measurements for δ18O in the tooth enamel of modern animals (74 animals, Table 3), 56 values are positive. This contrast is well illustrated in Figure 5, which shows that modern animals have δ18O values that are more positive than those of the fossil specimens by about 6‰. To be precise, the 50 animals from the Serengeti are 6.24‰ more positive than 27 animals from Bed I and 5.56 more positive than those from Bed II (112 animals). Of the 21 negative δ18O values for modern animals, 9 are for animals from the Serengeti (out of 53 animals, exact location unknown), 8 are for animals from Maswa (out of 13 animals) and 2 are for animals from Lukwati (out of 4 animals).
As the body water of animals is largely controlled by the rain in the area, it is suggested that the source(s) of rain in the Serengeti, Maswa and Lukwati are different. This hypothesis is under investigation, with the aim of establishing when the change in the rain source in the Olduvai/ Serengeti area occurred.
The stable carbon and oxygen isotope ratios of faunal tooth enamel at Olduvai and in the adjoining Serengeti National Park in Tanzania provide valuable information about the environment in that region at ca 1.8 Ma and in modern times. In general, the carbon isotope ratios indicate that the environment of Olduvai Bed I and II was very similar to that of the modern Serengeti - that is, 'savannah grassland with scrub and bush' (Gentry and Gentry, quoted by Cerling and Hay17). In terms of the UNESCO definition,46 it was a wooded grassland with 10-40% woody plant cover or a grassland with less than 10% woody plants. When the carbon isotope ratios of modern fauna have been corrected for the industrial effect in the modern atmosphere, the values of the fossil fauna are essentially the same as those of their modern counterparts. There is, however, a contrast between the fossil and modern tragelaphines. The woody plants growing at Olduvai at ca 1.8 Ma were evidently insufficient for these animals to be dedicated browsers.
A major change in the environment at Olduvai obviously occurred when Lake Olduvai dried up when the Gorge was formed (later than Bed II). This change removed a number of water-related plant species and freshwater fauna from the environment, but did not result in major changes in the diets of browsing and grazing animals, or in their relative prevalence in the landscape. However, if the major component of the diet of P. boisei was papyrus, then their staple food disappeared.
In contrast to the lack of difference in carbon isotope signatures in the fauna at Olduvai between fossil and modern specimens, their oxygen isotope signatures are distinctly different. The δ18O ratios of the modern faunal community are about 6‰ more positive than those of the fossil fauna from Beds 1 and II. A change of this magnitude has been observed in palaeosol carbonates in Olduvai Gorge by Cerling and Hay17, with the major change taking place at about 0.5 Ma.
A ready explanation for such a change in δ18O ratios is that there was a major change in temperature and humidity at some point during the past 1.8 Ma. This environmental change would have caused a major change in the rate of leaf water evaporation and would have increased the difference in enamel δ18O ratios between water-independent (evaporation-sensitive) Giraffidae and water-dependent (evaporation-insensitive) Hippopotimidae. Such an increase is not observable in this case. The δ18O ratios of Giraffa sp. (-2.2) and Hippopotamus gorgops (-5.5) from Lower Bed II differ by 3.3. Among the modern fauna, the δ18O ratios of Giraffa camelopardis (-3.6) and Hippopotamus amphibious (-0.2) differ by 3.4. Therefore the difference did not change, indicating that the humidity and temperature did not change. The relatively small difference between the δ18O ratios of water-dependent and water-independent animals also indicates that the environment in each case was not very dry and that water was permanently available.
An alternative explanation for the change in enamel δ18O ratios is that the primary source of meteoric water changed at some point during the past 1.8 Ma. It is suggested that this change took place with the introduction of the Indian Ocean monsoon to this part of East Africa. Currently, the major source of rain is the Indian Ocean monsoon which brings the 'long rains' at mid-year, while the Atlantic Ocean provides the 'short rains' at the end of the year. Given the long distance from the Atlantic Ocean, across Angola and Tanzania, the heavy isotopes of hydrogen and oxygen are substantially rained out along the way. This phenomenon has been documented for the rainfall at the research station at Seronera in the Serengeti National Park. A research project, as yet unfinished, has been launched by the author in collaboration with Cassian Mumbi of TAWIRI (Tanzania Wildlife Research Institute) and staff members at Seronera, to measure the oxygen isotopes in samples from the rain gauge over nearly 2 years. The oxygen isotope values of rain from the two oceanic moisture sources differ by as much as 10‰. It is also possible that nearby Lake Victoria, which was only formed around 50 ka, could be adding 18O to the local rainfall.
It is noteworthy that the same isotopic phenomenon (no change in δ13C, but substantial enrichment in δ18O) can be observed at Lake Natron between ca 1.5 Ma and modern times.2 As more isotopic analyses are done on fossil tooth enamel from northwest Tanzania, it is likely that the timing of this change will be established.
The fossil specimens from Olduvai were obtained from the collections of the Olduvai Landscape Archaeology and Palaeoanthropology Project (OLAPP), curated at the National History Museum in Arusha and at the Olduvai Field Camp. Rutgers graduate student Amy Cushing identified the fossil specimens and helped with sampling. The modern specimens from Serengeti National Park were obtained from the Research Station at Seronera. Emily Kisamo, the acting chief warden of Serengeti National Park at the time, was instrumental in providing sampling access. The isotopic ratios of all the specimens were measured at the Stable Light Isotope Facility of the Archaeology Department at the University of Cape Town by Ian Newton and John Lanham. Cape Town graduate students helped with the production of the manuscript: Mariagrazia Galimberti produced the figures, while Kerryn Warren typed the manuscript and checked the statistical calculations. I thank all of these individuals most heartily. Funds for the research were provided by the National Research Foundation of South Africa, the University of Cape Town, and the Landon T. Clay Fund of Harvard University.
1. Van der Merwe NJ, Masao FT, Bamford MK. Isotopic evidence for contrasting diets of early hominins Homo habilis and Australopithecus boisei of Tanzania. S Afr J Sci. 2008;104:153-156. [ Links ]
2. Van der Merwe NJ. Isotopic ecology and diets of fossil fauna from the T-1 (Type Section, Maritanane) paleosurface. In: Dominguez-Rodrigo M, Alcala L, Luque L, editors. Peninj, a research project on human origins (1995-2005). Oxford: Oxbow Books; 2009. p. 109-114. [ Links ]
3. Cerling TE, Mbua E, Kirera FM, Manthi FK, Grine FE, Leakey MG, et al. Diet of Paranthropus boisei in the early Pleistocene of East Africa. Proc Natl Acad Sci USA. 2011;108:9337-9341. http://dx.doi.org/10.1073/pnas.1104627108 [ Links ]
5. Lee-Thorp JA, Sponheimer M, Van der Merwe NJ. What do stable isotopes tell us about the hominid dietary and ecological niches in the Pliocene? Int J Osteoarchaeol. 2003;13:104-113. http://dx.doi.org/10.1002/oa.659 [ Links ]
6. Van der Merwe NJ, Thackeray JF, Lee-Thorp JA, Luyt J. The carbon isotope ecology and diet of Australopithecus africanus at Sterkfontein, South Africa. J Hum Evol. 2003;44:581-597. http://dx.doi.org/10.1016/S0047-2484(03)00050-2 [ Links ]
9. Cerling TE, Manthi FK, Mbua EN, Leakey LN, Leakey MG, Leakey RE, et al. Stable isotope-based diet reconstructions of Turkana Basin hominids. Proc Natl Acad Sci USA. 2013;110(26):10501 -10506. http://dx.doi.org/10.1073/pnas.1222568110 [ Links ]
10. Sponnheimer M, Alamseged Z, Cerling TE, Grine FE, Kimbel WH, Leakey MG, et al. Isotopic evidence of early hominin diets. Proc Natl Acad Sci USA. 2013;110:10513-10518. http://dx.doi.org/10.1073/pnas.1222579110 [ Links ]
11. Cerling TE, Bowman JC, O'Neill JR. An isotopic study of a fluvial-lacustrine sequence: The Plio-Pleistocene Koobi for a sequence, East Africa. Palaeogeogr Palaeoclimatol Palaeoecol. 1988;63:335-356. http://dx.doi.org/10.1016/0031-0182(88)90104-6 [ Links ]
12. Cerling TE. Development of grasslands and savannas in East Africa during the Neogene. Palaeogeogr Palaeoclimatol Palaeoecol. 1992;97:241-247. http://dx.doi.org/10.1016/0031-0182(92)90211-M [ Links ]
13. Kingston JD, Marina BD, Hill A. Isotopic evidence for Neogene hominid paleoenvironments in the Kenya Rift Valley. Science. 1994;264:955-959. http://dx.doi.org/10.1126/science.264.5161.955 [ Links ]
14. Wolde-Gabriel G, Haile-Selassie Y, Rennie PR, Hart WK, Ambrose SH, Asfaw B, et al. Geology and palaeontology of the Late Miocene Middle Awash Valley, Afar Rift, Ethiopia. Nature. 2001;412:175-178. http://dx.doi.org/10.1038/35084058 [ Links ]
15. Wynn JG. Influence of Plio-Pleistocene aridification on human evolution: Evidence from paleosols of the Turkana Basin, Kenya. Am J Phys Anthropol. 2004;123:106-118. http://dx.doi.org/10.1002/ajpa.10317 [ Links ]
16. Lee-Thorp JA, Sponnheimer M. Contribution of stable light isotopes to paleoenvironmental reconstruction. In: Henke W, Tattersall TH, editors. Handbook of paleoanthropology. vol.1. Bremen: Springer Verlag; 2007. p. 289-310. http://dx.doi.org/10.1007/978-3-540-33761-4_9 [ Links ]
20. Van der Merwe NJ, Vogel JC. 13C content of human collagen as a measure of prehistoric diet in woodland North America. Nature. 1978;276:815-816. [ Links ]
21. Van der Merwe NJ. Carbon isotopes, photosynthesis, and archaeology. Am Sci. 1982;70:596-606. [ Links ]
22. Bocherens H, Billiou D, Mariotti A, Pathou-Mathi M, Otte M, Bojean D, et al. Palaeoenvironmental and palaeodietary implications of isotopic biogeochemistry of last interglacial Neanderthal and mammal bones in Scladina Cave (Belgium). J Archaeol Sci. 1999;26:599-607. http://dx.doi.org/10.1006/jasc.1998.0377 [ Links ]
23. Richards MP, Pettitt PB, Trinkhaus E, Smith FH, Paunovic M, Karavanic I. Neanderthal diet at Vindija and Neanderthal predation: The evidence from stable isotopes. Proc Natl Acad Sci. 2000;97:7663-7666. http://dx.doi.org/10.1073/pnas.120178997 [ Links ]
26. Lee-Thorp JA, Van der Merwe NJ. Carbon isotope analysis of fossil bone apatite. S Afr J Sci. 1987;83:712-715. [ Links ]
28. Passey BH, Robinson TF, Ayliffe LK, Cerling TE, Sponnheimer M, Dearing MD, et al. Carbon isotope fractionation between diet, breath, and bioapatite in different mammals. J Archaeol Sci. 2005;32:1459-1470. http://dx.doi.org/10.1016/j.jas.2005.03.015 [ Links ]
29. Quade J, Cerling TE, Barry JC, Morgan ME, Pilbeam D, Chivas AR, et al. A 16- Ma record of paleodiet using carbon and oxygen isotopes in fossil teeth from Pakistan. Chem Geol (Isotope Geosc). 1992;94:183-192. http://dx.doi.org/10.1016/0168-9622(92)90011-X [ Links ]
32. Levin NE, Cerling TE, Passey BH, Harris JM, Ehleringer JR. A stable isotope aridity index for terrestrial environments. Proc Natl Acad Sci. 2006;103:11201-11205. http://dx.doi.org/10.1073/pnas.0604719103 [ Links ]
33. White TD, Ambrose SH, Suwa G, Su DF, DeGusta D, Bernot RL, et.al. Macrovertebrate paleontology and the Pliocene habitat of Ardipithecus ramidus. Science. 2009;326:67-93. http://dx.doi.org/10.1126/science.1175822 [ Links ]
34. Tieszen LL. Natural variation in the carbon isotope values of plants: Implications for archaeology, ecology and palaeoecology. J Archaeol Sci. 1991;18:227-248. http://dx.doi.org/10.1016/0305-4403(91)90063-U [ Links ]
36. Ehleringer JR, Field CB, Lin ZF, Kuo CY. Leaf carbon isotope and mineral composition in subtropical plants along an irradiance cline. Oecologia. 1986;70:520-526. http://dx.doi.org/10.1007/BF00379898 [ Links ]
37. Ehleringer JR, Cooper TA. Correlations between carbon isotope ratio and microhabitat in desert plants. Oecologia. 1988;76:562-566. [ Links ]
38. Bowen GJ, Wilkinson B. Spatial distribution of δ18O in meteoric precipitation. Geology. 2002;30:315-318. http://dx.doi.org/10.1130/0091-7613(2002)030<0315:SDOOIM>2.0.CO;2 [ Links ]
39. Dansgaard W. Stable isotopes in precipitation. Tellus. 1964;16:436-438. http://dx.doi.org/10.1111/j.2153-3490.1964.tb00181.x [ Links ]
43. Levin N, Zipser E, Cerling TE. Isotopic composition of waters from Ethiopia and Kenya: Insights into moisture sources from Eastern Africa. J Geophys Res. 2009;114:D23306. http://dx.doi.org/10.1029/2009JD012166 [ Links ]
44. Cerling TE, Harris JM, Hart JA, Kaleme P Klingel H, Leakey MG, et.al. Stable isotope ecology of the common hippopotamus. J Zool. 2008;276:204-212. http://dx.doi.org/10.1111/j.1469-7998.2008.00450.x [ Links ]
45. Blumenschine RJ, Peters CR, Masao FT, Clarke RJ, Deino AL, Hay RL, et al. Late Pliocene Homo and hominid land use from western Olduvai Gorge, Tanzania. Science. 2003;299:1217-1221. http://dx.doi.org/10.1126/science.1075374 [ Links ]
46. White F. The vegetation of Africa. Natural Resources Research. Vol. 20. Paris:UNESCO; 1983. [ Links ]
Nikolaas van der Merwe
Department of Archaeology
University of Cape Town
Rondebosch 7701, South Africa
Received: 12 Apr. 2013
Revised: 24 Jul. 2013
Accepted: 22 Aug. 2013