Isotopic Evidence for Dietary Variability in the Early Hominin Paranthropus robustus

See allHide authors and affiliations

Science  10 Nov 2006:
Vol. 314, Issue 5801, pp. 980-982
DOI: 10.1126/science.1133827


Traditional methods of dietary reconstruction do not allow the investigation of dietary variability within the lifetimes of individual hominins. However, laser ablation stable isotope analysis reveals that the δ13C values of Paranthropus robustus individuals often changed seasonally and interannually. These data suggest that Paranthropus was not a dietary specialist and that by about 1.8 million years ago, savanna-based foods such as grasses or sedges or animals eating these foods made up an important but highly variable part of its diet.

Both dental microwear texture analysis (1) and stable carbon isotope analysis (25) have demonstrated that the diets of South African australopiths were variable on the whole, but it has not been clear how the diets of individual hominins varied during their lifetimes. Here we provide evidence for short-term (seasonal and interannual) dietary change within the lifetimes of individual hominins, using a laser ablation method for stable isotope analysis (6). This method allows analysis along the growth axis of hominin teeth at submillimeter increments, making it possible to trace an individual's dietary history.

In tropical environments, virtually all trees, bushes, shrubs, and forbs use the C3 photosynthetic pathway, whereas grasses and some sedges use the C4 photosynthetic pathway (7, 8). C3 plants are depleted in 13C[∼–27 per mil (‰)] as compared to C4 plants (∼–12‰). The carbon isotopes in plants are incorporated into the tissues of consumers, with some additional fractionation (9, 10), and consequently carbon isotope ratios of tooth enamel can reveal the degree to which an animal consumed C3 or C4 resources. This allows determination of whether a hominin ate C3 foods, such as the forest fruits and leaves consumed by extant apes, or if they supplemented their diets with savanna-based C4 foods, such as grasses or animals eating those plants (2).

We analyzed the enamel of four permanent teeth of Paranthropus robustus from Swartkrans, South Africa [found in member 1 at the site, dated ∼1.8 million years ago (Ma)] using laser ablation stable isotope analysis (11). We also analyzed enamel of three contemporaneous browsing herbivores (Raphicerus sp.) from Swartkrans to control for postmortem alteration of carbon isotope ratios. We also counted the tooth growth increments (perikymata) that out-cropped to the enamel surface between and adjacent to each ablation track on the hominin specimens to temporally constrain the isotope data where possible (Fig. 1). The number of days represented by perikymata is 9 days in humans and extant apes, with ranges from 6 to 12 days in humans and extant apes (12). Because the actual periodicities of perikymata in fossil teeth cannot be known without sectioning them, we assumed that the periodicity for Paranthropus was 7 days for this study (12).

Fig. 1.

A portion of the imbricational enamel of SKX 5939, on which the total number of perikymata between the first and last ablation samples shown (A to E) is 22, meaning that the interval represented by these samples is approximately 154 days (22 × 7). The day counts are only intended to be rough approximations sufficient to differentiate seasonal from interannual variability. The first visible ablation track is outlined in white. Perikymata are visible as faint horizontal lines across the tooth's surface. Scale bar, 1 mm.

The mean of all carbon isotopic analyses for Raphicerus demonstrates that diagenesis has not obliterated the biogenic carbon isotopic compositions, because it indicates a C3 diet like that of Raphicerus' modern congeners (13) and of other known browsing herbivores from the site (Table 1 and table S1; δ18O values are discussed in fig. S1) (2). Moreover, the expected small range in variation within individual Raphicerus teeth (a maximum of 0.9‰) shows that fossilization has not induced significant carbon isotopic variation at the spatial resolution of our analyses (Fig. 2). In contrast, there is strong variability within individual hominin teeth. The mean range of variation within individual teeth is 3.4‰ for Paranthropus, whereas the mean range for Raphicerus is only 0.7‰ (P < 0.05, Mann-Whitney U test, Table 1), showing that these hominins had more variable diets. In two out of four hominin teeth, the amplitude exceeds 4‰, which, at face value, suggests that their consumption of C4 resources (tropical grasses or sedges or animals eating these foods) varied by ∼40%. However, this isotopic signal is attenuated because of protracted mineral uptake during amelogenesis and our sampling protocol, which required some mixing of enamel layers (14). Thus, a change of 5.2‰, as seen in specimen SKX 5939, probably signifies a change from a diet dominated by C4 resources to one of predominantly C3 foods. Our data do not allow us to determine which C4 resources Paranthropus consumed, although it is likely that grasses (seeds and roots), sedges (tubers and pith), and animal foods were all consumed to varying degrees.

Fig. 2.

δ13C of multiple ablation samples along the growth axes of teeth of the early hominin P. robustus (top) and the browsing bovid Raphicerus sp. (bottom). Precision as gauged by reproducibility of internal enamel and CO2 standards analyzed concurrently with each specimen was found to be 0.2, 1.1, 0.3, and 0.5‰ for SKX 5939 (black circles), SK 24606 (white squares), SK 24605 (white triangles), and SKW 6427 (gray diamonds), respectively. A perikymata count of 50 should be roughly equivalent to 1 year's crown formation. It is also important to note that each sample could incorporate carbon consumed over many months because of protracted mineral uptake during amelogenesis (14). This effectively attenuates the primary dietary signal, meaning that the intratooth variability observed here significantly underestimates actual dietary variability. Determination of the full amplitude of diet change awaits further study of enamel maturation parameters in hominoids [as in (14)]. The lack of variability within the Raphicerus teeth suggests that temporal differences in C3 vegetation δ13C values were very small and would not have contributed significantly to the variability in Paranthropus.

Table 1.

Carbon isotope means, ranges, perikymata counts (PKM), and estimated number of days (perikymata × 7) for the total sampling intervals of specimens in this study. Oxygen isotope compositions are produced in tandem with the carbon isotope data and are further discussed in fig. S1. SD, standard deviation; na, not applicable.

Specimenδ13C (‰)δ13C range (‰)PKM (n)Days (n)
SK 24605 -7.3 1.3 47 329
SK 24606 -6.1 4.4 70 490
SKX 5939 -5.4 5.2 92 644
SKW 6427 -8.6 2.5 63 441
Mean -6.9 3.4 68 467
SD 1.4 1.8 19 131
SKX 14150 -9.9 0.6 na na
SKX 8494 -9.8 0.5 na na
SKX 8535a -8.6 0.9 na na
Mean -9.4 0.7 na na
SD 0.7 0.2 na na

Although all of the Paranthropus specimens show evidence of seasonal variability, there is also evidence of interannual variation that might reflect yearly differences in rainfall-related food availability (Fig. 2). Another possible explanation is that these individuals were migrating between more wooded habitats (favoring C3 food consumption) and more open savannas (favoring C4 resource consumption). Regardless, these results are very unlike what has been observed in our close relative the chimpanzee (Pan troglodytes). Some chimpanzees inhabit savanna woodland environments that are believed to be similar to those inhabited by early hominins [such as Mt. Assirik in Senegal (15)]. However, they do not consume C4 resources to any measurable extent (16, 17), and the carbon isotope compositions of their hair are not known to vary significantly from season to season (fig. S2) (17). Baboons (Papio spp.), in contrast, consume significant quantities of C4 resources such as grass seeds and roots in some regions and some have variable δ13C values (18). Thus, eurytopic Papio might be a more appropriate ecological analog for P. robustus (19).

A dental microwear study of the earlier (3.0 to 3.7 Ma) hominin Australopithecus afarensis found no evidence that its diet changed over time or in different habitats (20). In contrast, stable carbon isotope (3, 4) and dental microwear texture analyses (1) of the slightly younger (∼3.0 to ∼2.4 Ma) hominin A. africanus demonstrated that its diet was far more variable. This suggests the possibility that a major increase in hominin dietary breadth was broadly coincident with the onset of increasing African continental aridity and seasonality after 3 Ma (21, 22) and only shortly antedated the first probable members of the genera Homo and Paranthropus (2325) and the earliest stone tools (26). The undoubted toolmaker Homo is thought to have been a dietary generalist that consumed novel foods such as large ungulate meat and tubers that are abundant in savanna environments (2730). Paranthropus, in contrast, with its extremely large and flat cheek teeth, thick enamel, robust mandible, and heavily buttressed facial architecture, is often portrayed as a dietary specialist (2729). Further, it has been argued that this specialization contributed to its extinction when confronted with increasingly dry and seasonal environments later in the Pleistocene, whereas Homo's generalist adaptation was crucial for its success (28, 29). Our results suggest that Paranthropus had an extremely flexible diet, which may indicate that its derived masticatory morphology signals an increase, rather than a decrease, in its potential foods. Thus, other biological, social, or cultural differences may be needed to explain the different fates of Homo and Paranthropus (31).

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Table S1


References and Notes

Stay Connected to Science

Navigate This Article