were we eating grains 100,000 years ago?

Until fairly recently, it was generally thought that the use of grains for food was a Neolithic innovation, that we only started eating grain after we started farming. But around 2004, analysis of a 23,000 year old site in Israel showed that the inhabitants were eating wheat and barley, as well as small-grained grasses -- and even suggested that they were baking grain-flour dough back that far. That makes breaking bread an ancient tradition indeed.

Now comes evidence suggesting (but by no means proving) that human use of grains for food may go back as far as 105,000 years:

Two years ago, Mercader and colleagues excavated a cave in Mozambique called Ngalue. They uncovered an assortment of stone tools in a layer of sediment deposited on the cave floor 42,000 to 105,000 years ago. The tools can't be directly dated, but Mercader presumes that the ones buried deepest in the layer are at least 100,000 years old. Other researchers had identified tubers as an important food source during the Stone Age, so Mercader decided to check for starch residue on 70 stone tools from the cave, including scrapers, grinders, points, flakes, and drills.

About 80% of the tools had ample starchy residue, Mercader reports today in Science. The starches came from the African wine palm, the false banana, pigeon peas, wild oranges, and the African potato. But the vast majority--89%--came from sorghum, a grass that is still a dietary staple in many parts of Africa.

According to Mercader, the findings suggest that people living in Ngalue routinely brought starchy plants, including sorghum, to their cave. He doesn't have definitive evidence that they ate the grass but says it seems likely. "Why would you be bringing sorghum into the cave unless you are doing something with it?" he asks. "The simplest explanation is that it would be a food item."