Paranthropus through the looking glass

Most research and public interest in human origins focuses on taxa that are likely to be our ancestors. There must have been genetic continuity between modern humans and the common ancestor we share with chimpanzees and bonobos, and we want to know what each link in this chain looked like and how it behaved. However, the clear evidence for taxic diversity in the human (aka hominin) clade means that we also have close relatives who are not our ancestors (1). Two papers in PNAS focus on the behavior and paleoenvironmental context of Paranthropus boisei, a distinctive and long-extinct nonancestral relative that lived alongside our early Homo ancestors in eastern Africa between just less than 3 Ma and just over 1 Ma. Both papers use stable isotopes to track diet during a largely unknown, but likely crucial, period in our evolutionary history.

The first fossil evidence of P. boisei, two upper milk teeth, a very large molar, and a tiny canine, was discovered in 1955 at Olduvai Gorge, in Tanzania (2). The mystery of the owner of the unusual teeth was solved in 1959 when Mary Leakey recognized fragments of a fossil hominin cranium eroding from a hillside. The Olduvai Hominid (OH) 5 cranium had a small (ca. 500 cm³) brain—not much bigger than that of a gorilla and about a third the size of that of a modern human—a flat and broad face, large attachment areas for chewing muscles, small incisors and canines, and exceptionally large premolar and molar tooth crowns. Louis Leakey proposed a new taxon, Zinjanthropus boisei (3) for OH 5, but within a few years the new genus was dropped in favor of Australopithecus, or Paranthropus; the latter is our preference. More evidence of P. boisei came in 1964 with the discovery at Peninj, just north of Olduvai Gorge, of a large lower jaw with the same unusual relative tooth size relationships seen in OH 5. In addition to the evidence from Olduvai, cranial and mandibular, but mostly dental, remains assigned to P. boisei have been recovered from sites in Kenya and Ethiopia. Konso in Ethiopia is the furthest north the taxon is recorded, and an upper jaw fragment from Malema in Malawi is the southernmost evidence. However, most of what we know about P. boisei comes from fossils from Koobi Fora on the eastern shore of Lake Turkana (4) and from sites in the Nachukui Formation on the western side of the lake (Fig. 1A).

Fig. 1.

Eastern African context of Paranthropus in the LOV. (A) Spatial distribution of Paranthropus-bearing localities in eastern Africa. (B) Vegetation change in eastern Africa over the past 5 My as indicated by fraction woody cover estimates. (C) New LOV hominin δ¹³C values within the context of existing eastern African hominin data. (D) New LOV hominin δ¹³C values in the context of existing Turkana Basin hominin values. (E) Summary of change points in new LOV faunal δ¹³C values.

The cranial and dental morphology of P. boisei is so distinctive its remains are relatively easy to identify (5). Unique features include its flat, wide, and deep face, flexed cranial base, large and thick lower jaw, and small incisors and canines combined with massive chewing teeth. The surface area available for processing food is extended both forward—by having premolar teeth that look like molars—and backward—by the unusually large third molar tooth crowns, all of which are capped by exceptionally thick and fast-forming dental enamel (6). Other early hominins, such as Paranthropus robustus from southern Africa, have similar-looking crania, large postcanine teeth, and thick enamel, but the crania of P. boisei are distinctive, its postcanine teeth are exceptionally large, and the enamel covering them is exceptionally thick. The foramen magnum—where the brain connects with the spinal cord on the underside of the cranium—is situated almost as far forward as it is in modern humans, suggesting that P. boisei was capable of walking erect on its hind legs.

A Dietary Puzzle

Why would a hominin need flared cheek bones, a large mandible, bony sagittal crests, and massive chewing teeth? Several lines of evidence suggest that P. boisei acquired its distinctive morphology within an ecosystem that was trending toward cooler, drier, and more open conditions, which in turn led to an increase in C₄ vegetation on the landscape (Fig. 1B and ref. 7). Stable isotope data extracted from the enamel of post-2-Ma P. boisei fossils from Koobi Fora are consistent with a diet dominated by C₄ foods (Fig. 1 C and D and ref. 8), whereas contemporary evidence of the genus Homo from Koobi Fora (i.e., Homo habilis and Homo rudolfensis) has a more C₃ signal (9). Most C₄ biomass in the tropics are grasses, and the chewing-dominated dentition of P. boisei would have enabled it to process grass, or sedges, which given the strength of the C₄ signal may have been their staple diet (10, 11). However, the same morphology would have also allowed P. boisei to process nuts and hard-shelled seeds. These could have been either additions to their normal diet or they could have been fallback foods that would have kept them going until they could return to eating their preferred food (12). The distinctive craniodental morphology of P. boisei appears as a package that undergoes little change over the course of a million years (13). Whatever niche P. boisei occupied, that niche and P. boisei’s adaptive response to it were remarkably durable. The exception to this morphological conservatism concerns fossils older than 2.3 Ma (14) from two locations—the Shungura Formation in the Lower Omo Valley of Ethiopia and west of Lake Turkana. The best-known of these fossils are a 2.6-My-old lower jaw found in Ethiopia and a ca. 2.5-My-old cranium recovered from the west of Lake Turkana. Many, but not all, researchers put them in a separate species, Paranthropus aethiopicus. Unfortunately, no tooth crowns were preserved in either the lower jaw or the cranium, but the space occupied by the roots of the postcanine teeth suggests that the crowns of the premolars and molars must have been similar in size to those of P. boisei. Compared with P. boisei, P. aethiopicus has a more projecting face, a more ape-like (i.e., less flexed) cranial base, larger incisors and canines, and simpler premolar crowns and roots. The earliest probable evidence for P. aethiopicus, a piece of upper jaw and a shin bone from Laetoli in Tanzania (Fig. 1A), has been dated to ca. 2.66 Ma.

In PNAS, Wynn et al. (15) exploit evidence from the Lower Omo Valley (LOV) to explore the dietary ecology of the P. aethiopicus–P. boisei lineage, and Negash et al. (16) provide a comparative context for these changes by tracking the dietary ecology of the broader herbivore community in the LOV between 3.5 and 2.0 Ma. The LOV, which is one of the few locations in eastern Africa where there is a more-or-less-continuous sedimentary record of this period of human evolutionary history, has been subjected to the type of careful systematic geological (17) and paleontological (18, 19) analysis that provides the necessary context for investigating what preceded the prolonged period of P. boisei stasis that is recorded in younger sediments at sites around Lake Turkana. Both studies use a method called change point detection to search for shifts in stable isotope values. Stable isotope analysis is destructive, but over the years sampling and analytical techniques have been refined in ways that reduce the damage to the fossils, while at the same time increasing precision (20). Thankfully, these developments, and the potential implications of any results, persuaded the curators responsible for these collections that these minimally destructive analyses were justified.

Tracking Diets through Time

Wynn et al. (15) document an isotopic transition in P. aethiopicus that apparently precedes the morphological transition between P. aethiopicus and P. boisei (Fig. 1 C and D). Stable isotopic data collected from enamel record the types of foods an organism was consuming when that particular tooth was forming. Therefore, these data are reflective of behavior, in this case food choice. Wynn et al.’s (15) findings are consistent with the expectation that behavioral change signaled by stable isotope data will precede a morphological response that requires population-level shifts across many generations.

However, changes in hominin diet and morphology in the LOV did not occur within an ecological vacuum. Negash et al. (16) use stable isotope data from the herbivores that lived alongside Paranthropus and our ancestors to better understand the ecological context of hominin dietary change. These data indicate a heterogeneous response to the elevated prevalence of C₄ vegetation in the LOV, with evidence of several dietary shifts between 2.8 and 2.2 Ma (Fig. 1E). Some of the families they investigated (e.g., antelopes and pigs) underwent a series of changes in stable isotope signal, whereas monkeys (including the large-bodied baboon species Theropithecus) only changed their dietary signal once, at approximately the same time as the hominins. These data point to a complex relationship between environmental change and dietary adaptation in the mammal communities that are sampled in the LOV, something that is also seen at other sites in Africa (21), but most sites lack the age control and sample sizes that enabled Negash et al.’s (16) careful analysis.

We do not know what the temporal and geographic ranges of an extinct hominin taxon like P. boisei were. Its presence at a fossil site is evidence that it was living at that time and in that place, but we should not assume that existing site samples circumscribe the temporal or geographic range of a taxon. Each site is like a window that gives us access to what is going on in just one room within a large house, but it is likely that important events in the evolutionary history of a taxon took place in presently windowless rooms. The task of paleoanthropologists is to either find additional sites (i.e., new windows) or improve the view through the existing windows. In Lewis Carroll’s 1871 novel Through the Looking Glass, when Alice decides to climb through a mirror she enters a fantastical world where everything is topsy-turvy. The world of P. boisei is not topsy-turvy, but it is sufficiently unlike contemporary analogs that we have to interpret it cautiously and on its own terms (22, 23). These two contributions substantially improve the view through an important existing window into the paleobiology of P. boisei.