Chemistry and early hominid diets

The chemical analysis of bones to interpret diet rests on the observation that different foods vary in the composition of different chemical elements or isotopes. Isotopes are different forms of an element that have different numbers of neutrons in their atomic nuclei (if they had different numbers of protons, they would be different elements). The number of neutrons in the nucleus affects the atomic weight of the isotope, and for this reason different isotopes may be taken up differently by different kinds of plants or animals, or they may be more or less abundant from different natural sources, such as the different mineral compositions of local soils. The chemical composition of an animal depends on the foods that it has eaten over the course of its life. Therefore, different kinds of animals may have different chemical signatures based on their preferred diets. If these isotopes are stable, then they may be preserved in fossil remains long after the death of the individual, and paleontologists may be able to access these ratios and make interpretations about the diets of ancient species. The amount of preserved material varies depending on the type of tissue examined, so that chemical analyses are usually expressed in terms of the ratio or proportion of one isotope or element to another. The major stable isotopes that have been examined in fossil hominid remains include the ratio of strontium (Sr) to calcium (Ca) and the ratio of carbon-13 (13C) to carbon-12 (12C).

Strontium/calcium ratios

Strontium and calcium are chemically similar elements that occupy the same column on the periodic table. For this reason, strontium can be taken up by plants in the place of calcium, and the two form a ratio that depends on the environmental abundance of strontium. When herbivores eat the plants, their bodies preferentially incorporate calcium instead of strontium, so that their Sr/Ca ratio is lower than that of the plants. And when carnivores eat the herbivores, they again incorporate more calcium than strontium, so that their Sr/Ca ratio is lower than the herbivores. Taken together, this means that we could infer the general diet composition (trophic level) of a fossil hominid if we knew its ratio of strontium to calcium. This wouldn't tell everything about the diet, and in fact it leaves many blanks. But with respect to the question of whether early humans were significant meat eaters, and whether robust australopithecines and other early hominids significantly differed in diet, this technique has great potential to inform.

One thing that is worth noting about these kinds of chemical ratios is that they reflect the average diet of ancient hominids across a large part of their lifespan. This time probably varies with circumstances, but it must always have included multiple years of dietary intake. This means that these ratios may respond to different aspects of the diet than the anatomy and size of the teeth, especially if the teeth were significantly adapted to fallback foods that did not make up the majority of the dietary intake of the animal.

Analysis of strontium and calcium in fossil bones requires some background work. The amount of strontium available to the food web depends on the local soil composition, so the Sr/Ca ratio may vary among samples of the same kind of animal taken at different sites. This means that different kinds of herbivores, carnivores, and omnivores must be sampled at the same location in order to interpret their Sr/Ca ratios. Additionally, it appears that fossil bone and fossil teeth may vary in their preservation of Sr/Ca ratios. The initial work on early hominid Sr/Ca ratios was done by Andy Sillen (1992) on fossil bone from Swartkrans. But Sillen and others have shown that the process of fossilization alters composition of bones in ways that may skew or erase the endogenous ratio of strontium to calcium.

Sponheimer and colleagues (2005b) examine the ratio of strontium to calcium in the tooth enamel of fossil hominids from South African sites. Enamel is less susceptible to diagenesis (change over time) than bone, and should preserve more accurate estimates of Sr/Ca ratios. A possible issue is that enamel is mainly deposited early in life, and therefore reflects preweaning or early juvenile diets that may not be fully representative of the dietary repertoire of the animal. To examine this, Sponheimer et al. examined comparative samples of mammals from the fossil localities and from recent contexts, finding that the Sr/Ca ratios did differentiate browsers, grazers, and carnivores from each other. The differences between these animal groups are that the grazers have the highest Sr/Ca ratios, and the carnivores and browsers. Browsers eat a high proportion of leafy species that tend to have lower Sr/Ca ratios, and as a consequence their Sr/Ca ratios tend to be slightly lower than those of the carnivores.

The analysis found that the remains of A. africanus from Sterkfontein Member 4 had relatively high Sr/Ca ratios, easily within the range of or even exceeding those of the grazers. A. robustus from Swartkrans Member 1 had substantially lower Sr/Ca ratios than A. africanus, but these were within the range of all the other animals, including browsers, grazers, and carnivores.

Sponheimer and colleagues (2005b) note that the results here were different from those of Sillen (1992), who showed the robust australopithecine bones to have rather low Sr/Ca ratios. Sillen suggested that this meant that the robust australopithecines was significantly omnivorous. The tooth enamel is consistent with a broad range of diets, so it does not disprove the hypothesis that robust australopithecines were omnivores, but it does not specifically disprove the notion that they were exclusive herbivores either.

The bottom line is that it is very difficult to differentiate diets with this kind of information. One problem is the nature of the overlap among the comparison samples. The whisker plots overlap substantially among these, and since they show the 10th and 90th percentiles, the extent of overlap may have been almost complete. This is not to say that the distributions are the same, but that individual fossils that are in the area of overlap (which would include most of the robust specimens and many of the A. africanus specimens) may not be diagnosed. Fortunately the study included a large number of teeth, so that samples may be compared to each other, and these samples are significantly different from each other. Assuming that the teeth have been assigned correctly to samples, this provides some confidence in the idea of a dietary difference between these samples.

A more important problem is that very different dietary compositions may have the same Sr/Ca signature. For example, a leaf browser that included some grass seeds in its diet might have the same Sr/Ca ratio as a fruit eater that included significant meat. And the Sr/Ca ratio does not give any indication of seasonal variations that might have ecological importance.

Carbon stable isotopes

Not all plants photosynthesize in the same way. The majority of plant species use a three carbon photosynthetic pathway. These are called C3 plants. But some plants use instead a four carbon pathway, and these are called C4 plants. The C4 plants are a minority, but include a large proportion of grasses and sedges, and a few other kinds of plants.

There are two stable isotopes of carbon in nature. Most of this carbon has six neutrons, resulting in an atomic weight of 12. But a minority of carbon has seven neutrons, with an atomic weight of 13 (an additional small proportion is the radioactive carbon 14). The C3 photosynthetic pathway preferentially includes carbon 12 (12C), so that C3 plants have a ratio of 13C to 12C that is substantially lower than the 13C/12C ratio in nature. For C4 plants, this discrimination is not as great, so that C3 plants and C4 plants differ in their 13C/12C ratios. Animals obtain their carbon from the foods they eat, so that the 13C/12C ratio of a herbivore marks the proportion of C3 and C4 plants in its diet. Likewise, the 13C/12C ratio of a carnivore reflects the plant diets of its prey species.

For example, grazers tend to eat a high proportion of grasses, which in Africa are predominantly C4 plants. This means that grazers have a relatively high 13C/12C ratio compared to other herbivores. It also means that carnivores who focus on grazers as prey species also have a high 13C/12C ratio. As noted by Sponheimer et al. (2005a:302): "the tissues of zebra, which eat C4 grass, are more enriched in 13C than the tissues of giraffe, which eat leaves from C3 trees."

A number of studies have examined the 13C/12C ratio in early hominid remains, focusing on those from the South African caves (Lee-Thorp et al. 1994; Sponheimer and Lee-Thorp 1999; van der Merwe et al. 2003). These studies are reviewed along with new results by Sponheimer and colleagues (2005a). The two basic results are that A. africanus and A. robustus are indistinguishable from their 13C/12C ratios, and that both australopithecine species have 13C/12C ratios that are elevated compared to C3 consumers and intermediate between them and C4 grazers. The fact that their ratios are lower than C4 grazers is not surprising, since australopithecines clearly did not eat grass. But if they depended largely on fruits, nuts, or other C3 foods, then it is difficult to explain why they should have stable isotope ratios that reflect a partial consumption of C4 foods.

Several hypotheses might explain this observation:

  1. Australopithecines may have eaten underground storage organs of C4 plants, such as grass corms or tubers of certain sedges.
  2. They may have eaten seeds from C4 grasses.
  3. They may have eaten the meat from grazing species.
  4. They may have eaten termites that relied on grasses and other C4 species.
  5. Diagenesis of 13C/12C ratios in fossils may have altered the isotopic signature, which actually may have been the same as that of C3 consumers (Schoeninger et al. 2001).

Sponheimer and colleagues (2005a) address the last hypothesis by testing a greater number of australopithecine teeth, finding results consistent with earlier findings. It is not obvious that this eliminates doubt entirely, but more samples provide more confidence that a real phenomenon has been observed. A comparison of all the hominids with recent C3 consumers shows clearly that they are significantly different, with relatively little overlap. They are also very different from the fossil C3 consumers preserved at the same sites (Sterkfontein and Swartkrans), which include browsing antelopes and giraffids. They are also distinct from C4 grazers in having a lower 13C level. From these values, Sponheimer and colleagues (2005a:305, emphasis in original) write:

[T]he data suggest that Australopithecus and Paranthropus ate about 40% and 35% C4-derived foods respectively. Such a significant C4 contribution, whatever its origin, is very distinct from what has been observed for modern chimpanzees (Pan troglodytes). Schoeninger et al. (1999) found no evidence of C4 foods in chimpanzee diets even in open environments with abundant C4-grass cover.

With respect to the termites and sedges, Sponheimer and colleauges (2005a) found that termites in open environments do have a high C4 proportion, while South African sedge species were found to be predominantly C3 plants. This means that termites might have provided part of the C4 component of early hominid diets, but the underground storage organs of sedges most likely did not. This does not mean that hominids may not have used sedges as a resource, but instead that such use would not explain their relatively high C4 proportion. And a diet of 35 to 40 percent termites seems quite high, so even if these were included in the diet, there were likely other C4 sources for the early hominids.

Another factor of the observations is that A. africanus teeth were quite variable in their 13C levels. Sponheimer and colleagues (2005a) suggest that one hypothesis to explain this variability would be if the sample changed over time--for example, in response to environmental change toward more open environments. Such changes in environment may be evidenced by a difference in the stable oxygen isotope ratios of the Sterkfontein and Swartkrans hominids. But when the ages of fossils were compared to 13C/12C ratios, there was no change over time, indicating that whatever dietary changes may have occurred, they evidently did not greatly affect the C4 proportion in the diets of the australopithecines. Sponheimer and colleagues (2005a:308) conclude that the australopithecines with high variability may simply have been "extremely opportunistic primates with wide habitat tolerances that always inhabited a similarly wide range of microhabitats regardless of broad-scale environmental flux."

Combining the data

Can these different sources of evidence be put together into a single picture of ancient hominid diets? The answer is yes, but unfortunately there is more than one hypothesis that may fit the bill. The facts that must be explained are as follows:

  1. High Sr/Ca ratios in A. africanus
  2. Moderate Sr/Ca ratios in A. robustus
  3. High proportion of C4 sources in both A. africanus and A. robustus
  4. Dental anatomy unsuited to leaf or grass eating in either species
  5. Tooth wear and anatomy reflecting hard, brittle food consumption by robust australopithecines (Grine 1986; Grine and Kay 1988), and possibly similar but to a lesser degree in other early hominids (Ungar 2004).

Sponheimer et al. (2005b) treat the first of these observations as the most problematic, and try to account for it with hypotheses that are consistent with the other observations. One hypothesis is that early hominids were insectivorous. They indicate that modern insectivores do have higher Sr/Ca ratios than other faunivores (153). In combination with possible evidence for termite digging at Swartkrans (Backwell and d'Errico 2001), this observation might suggest that early hominids used termites and other insects as a significant food source, even moreso than living chimpanzees. Sponheimer and colleagues judge this hypothesis as problematic because the fossil hominids differ from recent insectivores in having a low ratio of barium to calcium (154). One may also add that the robust australopithecines from Swartkrans did not have especially high Sr/Ca ratios, while there has not yet been evidence of termite digging for earlier hominids.

A second hypothesis is described as follows:

We have noticed that among the modern fauna that have the unusual combination of high Sr/Ca and low Ba/Ca are warthogs (Phacochoerus africanus) and mole rats (Cryptomys hottentotus (Sponheimer, unpublished data), both of which eat diets rich in underground resources such as roots and rhizomes. Thus, the possibility of greater exploitation of underground resources by Australopithecus compared to Paranthropus requires consideration. In addition, the slightly enriched Sr/Ca of Paranthropus compared to papionins might also be evidence of increased utilization of underground resources.

This last point about A. robustus may be reaching. On the other hand, this leaves the dietary mix issue somewhat unsettled. For example, what if both A. africanus and A. robustus ate underground resources, but A. robustus also ate meat? Or if A. africanus also ate insects. And so on. It seems unclear that anything short of a clear identity of diet between an early hominid and a modern analog in the African fauna will leave the possibility of exotic mixes.

This leaves us to reflect on the full pattern of evidence more closely. How is the hard, brittle diet inferred from dental anatomy and wear reconcilable with the hypothesis that australopithecines were eating the tough, fibrous underground storage organs of C4 plants?

Peters and Vogel (2005) address the issue of C4 diet proportion by examining the range of C4 plants that may have been available to early hominids. They make a number of observations:

  1. C4 sedges that produce edible roots, tubers, or stems are water-reliant, and do not compete with grasses in areas where drought occurs seasonally. They are therefore limited to relatively permanent watercourses including areas that are seasonally inundated with water. The South African sites do not represent such wetlands.
  2. Interestingly, C4 grasses have an evolutionary origin in the late Middle Miocene, and had increased in abundance in the African flora by the origin of the hominids.
  3. A majority C4 food intake by early hominids seems unlikely because of the wide availability of hominid-edible C3 foods in areas where the relatively rarer C4 hominid-edible plants also exist.
  4. Mature tubers of C4 sedges appear to have toxicity that may have impeded their edibility by early hominids, and they could probably have been consumed only in very small amounts.
  5. A number of potential animals may have provided a C4 component for early hominids, beyond the relatively large C4 grazing ungulates. These would include reptiles, birds, and rodents as well as insects. Early hominids would not have competed with other large carnivores for these small animals.

This brings us to a third hypothesis for early hominid diets. Peters and Vogel (2005:232-233) support an interpretation of omnivory for the early hominids, giving the following scenario:

As a starting point we can offer the following theoretical formulation of possibilities for a 30% C4 contribution to a subadult hominid diet based on minor potential C4 food categories:
  • 5% C4 input from sedge stem/rootstock, green grass seed, and forb leaves
  • 5% C4 input from invertebrates
  • 5% C4 input from bird eggs and nestlings
  • 5% C4 input from reptiles and micromammals
  • 5% C4 input from small ungulates
  • 5% C4 input from medium and large ungulates
This type of formulation maximizes the diversity of food species, i.e., both food-species-richness and evenness of contribution. The exact numbers are not as important as the species richness of the formulation.

A couple of things can be noted from this scenario. First, the predominant part of the C4 contribution comes from animal resources rather than plants. This conforms with Peters and Vogel's (2005) examination of C4 plant resources, only few of which are both edible by hominids and potentially available in quantity in their apparent paleoenvironments. However, this dietary component does not explain the masticatory adaptations of the australopithecines (indeed, it is not intended to explain them, since early Homo does not differ in dietary C4 contribution from earlier hominids). There is certainly nothing about the dentitions and jaw musculature of robust australopithecines to preclude an omnivorous diet of this type, but that invites the question of what fallback foods may explain that adaptation, or may explain the difference between robust and other australopithecines in that respect.


None of these three hypotheses really accounts for the full pattern of evidence about early hominid diets. The consensus so far appears to be that the chemical characteristics of the bones and teeth of early hominids reflect a majority diet that did not require a specialized dental adaptation. Therefore, the dental specializations of early hominids, in particular the enlargement of the postcanine dentition, reduction of the incisors and canines, and the low crowns of the molar teeth probably were adaptations to a minority of dietary intake that nevertheless was extremely important in selective terms. This would be characteristic of fallback foods eaten at times of resource scarcity, and would evidently have consisted of hard, brittle food items that could be effectively pulverized and ground by low-crowned teeth with large surface areas and thick enamel. This interpretation is supported by Ungar (2004) in an analysis of dental topography in early hominids and living hominoids (discussed in another article).

There are some remaining mysteries:

  1. If australopithecines had basically similar C4 dietary proportions, then what accounts for their differences in Sr/Ca ratios?
  2. Did any A. africanus-like hominids ever coexist with a robust australopithecine species?
  3. If early Homo had a C4 proportion that came in large part from hunting or scavenging grazing species, a hypothesis also supported by their dental anatomy (Ungar 2004), then did they abandon any of the C3 resources used by australopithecines?
  4. If australopithecines were opportunistic omnivores, were there important regional differences in their dietary composition?

These questions and others might be addressed with further sampling of dental chemistry.


Backwell LR, d'Errico F. 2001. Evidence of termite foraging by Swartkrans early hominids. Proc Natl Acad Sci U S A 98:1358-1363.

Grine FE. 1986. Dental evidence for dietary differences in Australopithecus and Paranthropus: a quantitative analysis of permanent molar microwear. J Hum Evol 15:783-822.

Grine FE, Kay RF. 1988. Early hominid diets from quantitative image analysis of dental microwear. Nature 333:765-768.

Peters CR, Vogel JC. 2005. Africa's wild C4 plant foods and possible early hominid diets. J Hum Evol 48:219-236.

Schoeninger MJ, Bunn HT, Murray S, Pickering T, Moore J. 2001. Meat-eating by the fourth African ape. In: Stanford CB, Bunn HT, editors, Meat-eating and human evolution. Oxford, UK: Oxford University Press. p 179-195.

Schoeninger MJ, Moore J, Sept JM. 1999. Subsistence strategies of two savanna chimpanzee populations: The stable isotope evidence. Am J Primatol 49:297-314.

Sillen A. 1992. Strontium-calcium ratios (Sr/Ca) of Australopithecus robustus and associated fauna from Swartkrans. J Hum Evol 23:495-516.

Sponheimer M, de Ruiter D, Lee-Thorp J, Sp├Ąth A. 2005b. Sr/Ca and early hominin diets revisited: New data from modern and fossil tooth enamel. J Hum Evol 48:147-156.

Sponheimer M, Lee-Thorp J, de Ruiter D, Codron D, Codron J, Baugh AT, Thackeray F. 2005a. Hominins, sedges, and termites: New carbon isotope data from the Sterkfontein valley and Kruger National Park. J Hum Evol 48:301-312.

Sponheimer M, Lee-Thorp JA. 1999. Isotopic evidence for the diet of an early hominid, Australopithecus africanus. Science 283:368-370.

Ungar P. 2004. Dental topography and diets of Australopithecus afarensis and early Homo. J Hum Evol 46:605-622.

van der Merwe NJ, Thackeray JF, Lee-Thorp JA, Luyt J. 2003. The carbon isotope ecology and diet of Australopithecus africanus at Sterkfontein, South Africa. J Hum Evol 44:581-597.