IN 2009 Richard Wrangham, an anthropologist at Harvard, published an intriguing thesis. He was trying to answer a question that had long puzzled workers in his field: how could the evolution of an organ as energetically expensive to sustain as the human brain have happened?

Before Dr Wrangham’s work the conventional answer was: “meat-eating”. Archaeological evidence such as a lack of tool marks on animal bones suggests humanity’s ancestors, the Australopithecines, were largely vegetarian. By contrast Homo erectus, the first widespread human being (pictured below), also ate meat, which is a more compact source of calories than most plant matter, and might thus have provided the extra brain-food needed.

Dr Wrangham, however, had a different answer: “cooking”. He showed that the ease of digestion and additional nutritional value made available by treating food with fire (which alters starch and protein molecules in ways that make them easier to digest) boosts its calorific value enough for a reasonable daily intake to power both brain and body—so much so that modern humans who attempt to live only on raw foodstuffs (there are a few who try) have great difficulty remaining well-nourished. On top of this, the softening brought about by cooking could explain a second evolutionary trend, that toward smaller teeth and less-powerful jaws.

Just when Homo erectus did start cooking is controversial. The oldest definitive evidence dates back only 500,000 years, though the species evolved 1.9m years ago. But the Wrangham thesis does not depend only on the beginning of heat-treating food. It also includes food preparation using tools to chop or pound meat and vegetables. This presumably makes them easier to digest. It also makes them easier to chew, which might account for the reduction in jaw and tooth size.

A paper published in this week’s Nature by Katherine Zink and Daniel Lieberman, two of Dr Wrangham’s colleagues at Harvard, brings some evidence to bear on these questions, particularly that of chewing. Dr Zink and Dr Lieberman used replicas of the stone tools available to Homo erectus to process food, and looked at the consequences for those who attempted to masticate the result.

The pseudo-Palaeolithic diet the two researchers chose comprised beets, carrots and yams as root vegetables, and goat as meat. They prepared the vegetables four ways: raw and unprocessed; raw and hit six times with a copy of a Palaeolithic hammerstone; raw and cut into small slices; and roasted for 15 minutes. The goat was also served four ways: raw and unprocessed; raw and pounded 50 times by a hammerstone; raw and cut into small slices; and cooked on a grill for 25 minutes. Dr Zink and Dr Lieberman then fed each preparation to a group of volunteers, to see how easy it was to chew.

To measure this, they wired up the skin of their volunteers’ jaws using electrodes which recorded the force a volunteer exerted chewing. Once wired, volunteers were given samples to chew and asked to do so until they felt what they were chewing was ready to swallow. Sometimes the volunteers were then allowed to swallow. On other occasions, though, they were asked to spit the sample out, so that the bits could be analysed. (The raw meat was always spat out, to prevent foodborne illness.)

Dr Zink and Dr Lieberman found, in line with Dr Wrangham’s original thesis, that chewing cooked root vegetables required a third less force than was needed to chew an equivalent amount of raw and unprocessed root. Slicing the vegetables did not provide any benefit, but pounding them reduced the force required to chew by about 9%. Pounding meat, by contrast, brought no benefit, whereas slicing it did. As with cooking the vegetables, it reduced the chewing force needed by around a third. Intriguingly, roasting meat actually increased the masticatory force required.

On top of this, when Dr Zink and Dr Lieberman examined food spat out by their volunteers at the point it was deemed ready to swallow, they found that the unprocessed and the pounded meat usually came back as a single large lump that would be hard for the gut to break down. In contrast, when the meat was sliced or cooked before being chewed, participants were consistently able to chew it into tiny, digestible particles.

Putting all their results together, Dr Zink and Dr Lieberman conclude that a diet of one-third sliced meat and two-thirds pounded vegetables, such as Homo erectus might reasonably have been expected to consume even in the absence of fire, would need 27% less effort to chew than an unpounded all-vegetable diet. Specifically, the inclusion of meat contributed a 15% reduction and the slicing and pounding a 12% reduction, which Dr Lieberman calculates equates to 2.5m fewer chews a year.

That could certainly account for the shrinkage of jaws and teeth undergone by Homo erectus. As to its consequences farther down the digestive tract, those remain the province of further research.