Humans left Africa with a meagre tool kits at a remarkably early date, possibly around 1.9 Ma from finds of primitive stone tools in Pakistan and Central China, and certainly before 1.7 Ma in the case of the now celebrated human remains at Dmanisi in Georgia and in Java. Around 1.7 Ma sites with evidence for human occupation extend from southern to north-western Africa and over 2/3 of the width of southern Eurasia. Despite the increased chances of preservation in later times, such a wide-ranging expansion seems not to have recurred until the fully modern human diaspora from Africa that began around 70 to 100 ka. Fossil evidence suggests that descendants of these earliest known migrants thrived until as recently as 20 ka in south-east Asia, and perhaps longer, if tiny Homo floresiensis prove to be other than symptomatic of congenital dwarfism. They represent a puzzle, and absence of evidence has deterred palaeoanthropologists from sticking out their necks, until a recent review of possibilities (Dennell, R. & Roebroeks, W. 2005. An Asian perspective on early human dispersal from Africa. Nature, v. 438, p. 1099-1104).

For a long time all human remains dated between 1 and 1.9 Ma were ascribed to H. erectus, whose type specimen hails from Java, not Africa. Anatomical re-evaluation of specimens from Africa, notably the famous, 1.6 Ma old Turkana Boy from Kenya, shows that they are sufficiently different from Eugéne Dubois’s Javan H. erectus type specimen to warrant a different species name – ‘Action Man’ or H. ergaster. The Dmanisi humans have close affinities, but are older. Therein lies one puzzle: apart from the very much more primitive (and very rare) H. habilis of east Africa, there is no obvious African candidate as an ancestor for H. ergaster there. Dennell and Roebroeks speculate that they migrated back to Africa after evolving there from some unknown earlier species. Another puzzle centres on the tools carried by the early migrants from Africa.

Simple chopper and rough flake tools first appear in north-east Ethiopia at 2.6 Ma, but with no clear sign of who made them. The first discovery of the earliest known tool kit was at Olduvai Gorge in Tanzania – hence their name, Oldowan. They are associated with remains of the earliest known human species H. habilis, but date only to 1.8 Ma. Since Oldowan tool use is now known to have extended over a huge range of Africa and Eurasia at that time, the original emigrants must have carried the culture with them sometime after its first appearance in Ethiopia at 2.6 Ma. The emblematic artefact of ‘H. erectus’ is the beautiful pear-shaped biface axe, yet it first appeared at 1.5 Ma in Africa, and did not make an appearance outside the continent until about 700 ka and never made it to east Asia until carried their by fully modern humans: it was an African invention. Oddly, these highly crafted tools are often found with little sign of wear, and indeed opinion about what they were for is divided.

The great problem in palaeoanthropology is absence of fossils, which is hardly surprising. Dennell and Roebroeks comment that most Late Pliocene to Early Pleistocene terrestrial faunas are nearly always of large, robust animals, and even they are uncommon. The ravages of erosion and transportation also make it difficult to date finds of stone tools, as they may have been mixed with younger dateable materials. With confidence, they rely on the old adage (not well liked by the Popperian school of scientific methodology) that, ‘Absence of evidence is not evidence of absence’, and also that the earliest evidence for a new migrant is bound to be younger than its first presence. They look to the palaeoecological record of the period, which suggests a vast extent of open savannah covering much of Africa and southern Asia in the period when the climatic effects of glacial-interglacial cycles had not gripped low latitudes to create the desert barriers of later Pleistocene times. For species adapted to savannah life there was little to prevent their very wide migration, indeed simple diffusion would have moved them across the entire savannah range. Once thought to be confined to the East African Rift, australopithecines have turned up as far afield as modern Chad, 2500 km away, and as long ago as 3.5 Ma. If such diminutive creatures with no tools could diffuse so far, then what might have been the geographic limitation to the earliest tool users? Moreover, diffusion has no direction in the area that presents its possibility: movement could have been back and forth. An intriguing point made by Dennell and Roebroeks is that climatic instability first appeared around 2.6 Ma in Central China, so any emigrants moving north would have been subject to greater evolutionary-selective pressures for longer. Homo ergaster might have evolved in Asia and returned to Africa in the face of worsening conditions. This approach raises as many plausible hypotheses as a stick can be poked at, and should re-vitalise palaeoanthropological research outside Africa as a means of testing them.

Dee also: Kohn, M. 2006. Made in Savannahstan. New Scientist, v. 191 (1 July 2006 issue), p. 34-39.