The "Out of Africa" model of humanity's spread is now about as well-established as anything can be that attempts to describe what happened 60,000 years ago. The fossil evidence is fragmentary, but nonetheless the outlines of our collective past are clear: Human beings, the model posits, evolved in Eastern Africa some 200,000 years ago and then, thousands of years later, a band of intrepid migrants crossed the straits of Bab el Mandeb into what is now Yemen to colonize the rest of the planet.
This does not, though, explain how Africa itself was colonized. In particular, it poses the question of how people crossed from the sub-Saharan part of the continent, where the evidence suggests Homo sapiens originated, across the sweltering Sahara to its northern fringes.
The evidence suggests this internal migration within Africa happened around 125,000 years ago. The most likely scenario is the first migrants to Northern Africa followed rivers thought to have traversed the Sahara at that time, known as the Eemian interglacial.
Then, as now, the ice caps that have dominated Earth's climate for the past 2.6 million years had temporarily shrunk. In the Eemian, the West African monsoon reached about 1,000 kilometres farther north than it does now. It could have fed north-flowing rivers through the desert and, as Jorge Ramirez of Hull University in Britain told the American Geophysical Union's meeting in San Francisco earlier this month, the computer models he and his colleagues have devised suggest this is what happened.
The westernmost of the fossil channels of those streams is called the Irharhar. It runs from the Ahaggar mountains in southern Algeria to the foot of the Atlas mountains. The river probably ran and dried up several times as the glaciers ebbed and flowed, but it is uncertain when this happened.
Ramirez's models suggest the monsoon would, indeed, have reached north of the Ahaggar massif in the Eemian, filling the Irharhar and also two other channels, the Sahabi and the Kufrah, about 1,000 kilometres east, that rise in the Tibesti mountains. They also suggest that, though the Irharhar was ephemeral even when it flowed, dry for nine months because of the seasonality of the monsoon, the climate at its northern end was humid all year round, providing travellers with an attractive destination.
By contrast, even though the easterly rivers flowed permanently, the climate of the area that they led to was dry and uninviting. That would not be a problem for farmers, who could use the water for irrigation, but agriculture still lay more than 100,000 years in the future. For people who depended on hunting and gathering, a wide area of verdant, rain-fed habitat would have been more productive than a continuously flowing river.
That accords with archaeological evidence of human occupation in the region. Signs of such occupation are much more abundant where the Irharhar flowed than they are farther east, a fact that previously had been unexplained.