Last fall, on a reporting trip to Johannesburg for a story on the discovery of fossils representing a previously unknown member of human family called Australopithecus sediba, the researchers I met with were buzzing with excitement about, of all things, tartar. That’s right, the crusty deposits that the dentist scrapes off your teeth when you go for a cleaning. Except in this case, it was the tartar on the teeth of the nearly two-million-year-old A. sediba, which has been held up as a candidate ancestor for our genus, Homo. No one had ever found tartar in an early hominin (a creature on the line leading to humans, after the split from the line leading to chimps) before—the oldest samples came from much younger Neandertals and anatomically modern humans. And in analyzing the ancient tartar, the researchers had recovered evidence of what A. sediba ate. It wasn’t at all what they expected.

I had hoped to be able to report on the findings in the cover story of the April issue, but the scientists had yet to publish the results in a peer-reviewed journal and so I had to keep mum until now. In a paper published online by Nature on June 27, Amanda Henry of the Max Planck Institute for Evolutionary Anthropology in Leipzig, Lee Berger of the University of the Witwatersrand in Johannesburg and their colleagues report on the tartar analysis, as well as their analyses of A. sediba’s tooth chemistry and dental wear marks. (Scientific American is part of the Nature Publishing Group.) Their results offer the clearest view yet of what an early human ancestor ate, and deepen the mystery surrounding this species.

The team conducted their studies on the two most complete A. sediba individuals recovered thus far, an adult female and a subadult male found at a site just outside Johannesburg. Analyses of the wear on their molars showed that the two hominins ate hard foods shortly before they died. And their tooth chemistry—specifically the carbon ratios—revealed that, over their lifetime, they dined mostly on so-called C3 foods, which include trees, shrubs, some herbs and the animals that eat those kinds of plants. This is surprising, because other hominins of similar antiquity relied more heavily on C4 foods—most tropical grasses and sedges and the animals that eat those plants. Furthermore, paleoenvironmental evidence from the site that yielded the fossils attests to a setting dominated by C4 plants, not C3 ones. Among early hominins only the much older Ardipithecus ramidus from Ethiopia comes close to A. sediba’s carbon isotope composition; compared to the tooth chemistry of modern creatures, A. sediba’s looks like a savanna chimpanzee’s or a giraffe’s.

Even more startling, when the researchers examined the tartar, they found traces of plant foods no one thought our ancient kin ate, such as bark. The tartar contained silica crystals called phytoliths that plants make as a means of self-defense, some of which the investigators could attribute to particular kinds of plants on the basis of their distinctive shapes. The phytoliths indicate that in addition to bark, A. sediba also probably ate C3 grasses and sedges, as opposed to the more common C4 varieties. According to the authors, together the three lines of evidence suggest that A. sediba foraged for C3 foods in habitats similar to gallery forests surrounded by C4 grasslands.

Berger notes that many primates use bark as a fallback food during times when fruit is hard to come by. He has speculated that the A. sediba individuals whose remains he and his colleagues recovered from what was once a deep underground cave may have ended up there because drought conditions drove the hominins to attempt to access a pool of water inside. The bark finding could bolster that scenario. “We might just be capturing a community living in an environment that is becoming increasingly stressful,” Berger observes.

Viewed in the context of A. sediba's amalgam of autralopithecine-like and Homo-like skeletal traits, the diet findings are even more interesting. Although in many anatomical respects A. sediba is primitive (its tiny brain and long arms, for example), it has a remarkably dexterous hand that may well have been able to make and wield tools and it has small teeth, which are associated with an increase in higher-quality foods such as meat. Conventional wisdom holds that Homo adapted to changing environmental conditions that favored the spread of grasslands by incorporating meat into its diet. The tooth wear and phytolith evidence makes clear that A. sediba was foraging for C3 plants, but might it have also pursued animals that specialized in C3 plants? Says Berger: “With the quality of preservation and type of data we are getting, I think we will reach answers to these questions.”