by Stephen Oppenheimer

As with all previous explorations, by both modern and archaic humans, geography and climate decided the newly arrived occupants of Asia where to go next. The rules would have been simple: stay near water, and near reliable rainfall; when moving, avoid deserts and high mountains and follow the game and the rivers. We have seen circumstantial evidence that the beachcombing route round the coast of the Indian Ocean to Australia was the easiest and earliest option. Why should this have been? It was not that easy: For a start, every few hundred kilometres our explorers would have had to ford a great river at its mouth. Yet this is just what they must have done to get to Australia, so it is possible they did the same along the East Asian coast. At each river there was the option for some people to turn left and head inland, harvesting river produce and game as they went.

As one of the earliest European explorers, Marco Polo, found out, mountains and deserts present formidable barriers to those trying to gain access to Central Asia; apart from a few trails, the only routes of entry are along the river valleys. We have seen that our first successful exodus from Africa took the ancestors of all non-Africans south along the Indian Ocean coast perhaps as long ago as 75,000 years. They may also have beachcombed as far as eastern China and Japan rather early on. They would thus have skirted the whole of the Central Asian region. They could have tried to head upriver and inland at any point on their journey.

North of India, with the Himalayas in the way, it was not as straightforward as that. The raised folds of mountains caused by India's ancient tectonic collision with Asia extend either side of Nepal and Tibet well beyond the highest Himalayas. A vast band of mountains, all over 3,000 metres (10,000 feet), blocks Central Asia to access from the Indian Ocean coast for a distance of 6,500 km (4,000 miles) from Afghanistan in the west to Chengdu, in China, to the east. This band is rucked up like a carpet in the east, thus extending the mountain barrier south as a series of north-south ridges over a distance of about 2,500 km (1,500 miles) from the beginning of the Silk Road in northern China south to Thailand.

The Silk Road, first made famous in the West by Marco Polo, is a long trading route, parallel to and to the north of the Himalayas, connecting West with East. It passes right through Central Asia, directly along the southern edge of Guthrie's Mammoth Steppe heartland. The Silk Road was then, as it is now, one of the few links between China and the West, if the long coastal route round south via Singapore was to be avoided.

Today the Silk Road skirts both the southern and northern edges of the Taklamakan Desert of Singkiang. During the Palaeolithic, what is now desert was mostly lush grassland, and farther north a series of waterways, including the Tarim and Dzungaria rivers, provided easy west-east access for hunters from the western Central Asian regions of Tajikistan, Uzbekistan, Kirghistan, and Kazakhstan into Singkiang and Mongolia. These waterways may have been used by earlier humans to get to Central Asia.

To look at Stone Age practicalities, let us take the first access route to Central Asia up the Indus, 8,000 km (5,000 miles) to the west of China at the western end of the Silk Road. Assuming for the moment that we are talking about an offshoot of the first Indian beachcombers, their first task after moving up the Indus would have been to negotiate the mountain barriers to the north of India and Pakistan. These extend as far west as Afghanistan. Bypassing the mountains and moving through Afghanistan too far to the west would have been difficult if not impossible, since it was near-desert. Marco Polo crossed these deserts, leaving from Hormuz at the mouth of the Gulf and passing through Afghanistan to Kashmir, crossing a high pass directly into China and the city of Kashgar, to arrive well along the Silk Road and directly in the heartland of the former Mammoth Steppe.

Marco Polo could have followed a much easier route to Kashmir, however. From the coast of Pakistan a little further to the east, the great Indus snakes northward to a point where there is a water connection through to Kashmir. Another lower-altitude route into Central Asia, also via the headwaters of the Indus, would have been to cross the Khyber Pass to Kabul, and thence to Uzbekistan and Kazakhstan, and then east towards Singkiang.

The second access route to the Mammoth Steppe, not much used by traders today, is just to the east of the Himalayas. The eastern edge of the Himalayas consists of multiple folds where the edge of the Indian plate rucked up the Asian continent on collision. These rucks are the conduits for most of the great rivers of South and Southeast Asia. From west to east, these are the Brahmaputra, which flows into Bangladesh, the Salween, which flows into Burma, the Mekong, which flows into Vietnam, and the Yangtzi, which flows into southern China. As they flow out of south-east Tibet these four major rivers run parallel for about 150 km (around 100 miles), separated from one another by only a few kilometres. The last of the four, the Yangtzi, originates in north-east Tibet near the northern edge of the plateau and at the beginning of the Mammoth Steppe.These multiple river routes are important not because many legitimate traders use them today, but because they allow direct access to Tibet, Mongolia, and Central Asia from any of four widely separated river mouths in Southeast and East Asia discharging on to the Indo-Pacific coast. Also, Tibet shares much genetically, with Indo-China and Southeast Asia.

Equally, during all of human history, the Silk Road has been the only route west from China into Central Asia. So, an alternative route into the Mammoth Steppe would have been from the East Asian Pacific coast. The vanguard of the earliest beachcombers could have gone all the way to China and then moved west from northern China along the Silk Road into Mongolia, Sinkiang, and southern Siberia.

Finally, there is another, more northerly route of migration from the West into East Asia to be considered: via Asian Russia, known as the Russian Altai. The easiest direct land access from the Russian Altai to Central Asia during the milder parts of the Late Stone Age 30,000-50,000 years ago would have been to cross the steppe directly. Travelling east through southern Siberia via a series of lakes and waterways, our ancient explorers could have reached the Lake Baikal region by a route passing north of Singkiang and Mongolia. At that time the steppe covered the whole region in greensward and open woodland. Clearly, for modern humans to have taken this route they must have got to the Russian Altai in the first place; they had reached both the Russian Altai and Lake Baikal in southern Siberia by 40,000 years ago.

To summarize, we have now seen that there are four possible access routes into Central Asia: three from the Indo-Pacific coast (west, south, and east) and one from Russia (north-west). Once in Central Asia, there were three parallel routes along water bodies between East and West Asia which the pioneers could have followed: two southern ones, through Singkiang and Mongolia, and a northern one across Southern Siberia. The northern route would have been accessible only during the milder periods of the Palaeolithic 30,000-50,000 years ago, during the interstadials.