There's been quite a lot of conjecture, over the years, about what our early ancestors ate. Much of the evidence has been indirect: size of teeth, size of chewing muscles (which can be estimated from measurements of the places where muscles attach to the skull), ridges & crests on the skull, & so on. Teeth are very informative – as well as their size & shape, we can gain information from the tiny patterns of wear left from chewing food. (Or cleaning your teeth – my old dentist used to say he could tell if people followed the recommended brushing pattern by looking at wear on their teeth…)
The paper I've just read (Ungar et al., 2008) looked at the teeth of Paranthropus (aka Australopithecus) boisei – given the nickname of 'nutcracker man' because of its huge grinding teeth. Because of this, and the heavily-built cranium & jaw (assumed to be accompanied by very large chewing muscles), most scientists assumed that boisei must have habitually eaten hard or tough, fibrous food that required a lot of chewing. Others proposed an alternative possibility – that such foods may have been an occasional part of the diet, but not the only thing boisei ate.
How to test this hypothesis? Ungar & his co-workers looked at the tiny wear patterns (dental microwear) on boisei teeth, and compared what they found with tooth wear in living primates whose diets are well known. Apparently microwear forms and changes quickly, throughout an individual's life (there must have been something in what my dentist said!), so studying the boisei teeth could give information on what the owners of the teeth had been eating before they died. The team noted that similar analyses of Paranthropus robustus suggested that this species ate hard foods only sometimes, rather than throughout the year, maybe when softer, more preferred foods were in short supply. This observation fitted with something called Liem's Paradox: specialised morphology can allow for a broader diet wherein a species may actively avoid the very foods to which it is adapted when other, more preferred resources are available (Ungar et al., 2008). Now, boisei seems to be even more specialised for eating hard, tough foods than robustus – what would microwear analysis show?
The research team analysed microwear patterns from 7 boisei teeth. (many more teeth were available, but the patterns weren't clearly visible.) They compared their data with wear patterns from 2 species of monkey that sometimes eat hard foods and 2 species that eat tough food such as plant stems, plus P. robustus & Australopithecus africanus. (One commentator notes that it's perhaps surprising that the team didn't include modern gorillas in their study, given that mountain gorillas, in particular, have adaptations for eating a tough,fibrous diet.)
Despite their apparent dietary specialisations, the data suggested that none of the boisei individuals had been eating particularly hard or tough foods in the days before they died. Surprisingly, it appeared that of the two robust australopiths, robustus ate much harder foods than boisei! The researchers' conclusion: their data suggest that boisei is an example of Liem's paradox: although this species could eat hard or tough foods if there was nothing else on offer, where softer more palatable foods were available it would preferentially choose them. What a neat piece of palaeontological detective work!
P. Ungar, F. Grine, & M. Teaford (2008) Dental microwear and diet of the Plio-Pleistcene hominin Paranthropus boisei. PLoS ONE 3 (4): e2044, doi:10.1371/journal.pone.0002044