Before the Dawn: Recovering the Lost History of Our Ancestors

by Nicholas Wade

Almost all the dogs in Savolainen’s sample fell into three main clusters, suggesting either that they had been domesticated independently three times or that three related wolves from the same litter or pack had been domesticated at the same time and place. Savolainen favors the latter interpretation because it gives a more plausible date for the domestication event—15,000 years ago. (The alternative case of three separate domestica tions implies a date of 40,000 years ago. But an invention as useful as the dog would probably have spread like wildfire, and there is no evidence for dogs for another 26,000 years.)130

  It’s a considerable puzzle to understand how the process of domesticating wolves into dogs got under way. Some species can’t be domesticated at all and with others, many generations of selective breeding are required to produce any results. The difficulties were demonstrated in a remarkable experiment by a Soviet scientist, Dmitri K. Belyaev, who set out to domesticate silver foxes. His theory was that all or most domestic animals had been derived from their wild forebears by the same straightforward criterion, that of tameability. The set of genes required to bring about this profound change of behavior in a wild animal, he believed, also induced the distinctive physical characteristics found in many species of domesticated animal. These include white patches on the pelt, curly hair, shorter tails and floppy ears.

  Belyaev and his successor, Lyudmila N. Trut, selected silver fox puppies on the sole criterion of tameness, choosing only those least hostile to human contact as the parents of the next generation. After 40 years, 45,000 foxes, and 30 to 35 generations of breeding, Trut now has a population of 100 docile, eager-to-please silver foxes, many carrying the white patches that Belyaev had predicted.131

  In the case of dogs, domestication has another ingredient besides tameability, which is the capacity to read human body language. Brian Hare of Harvard University has tested the ability of dogs, wolves and chimps to pick up on cues as to which container holds hidden food. The experimenter would give broad hints, such as tapping the right container, or staring at it. Chimpanzees have a lot more intellectual wattage than dogs, yet very few got the message because they paid no particular attention to what the experimenter was doing. Wolves too are very smart, but did not take the hint. But dogs, and even puppies, picked up instantly on the hint being conveyed. 132 Because even puppies have this ability, it is probably innate and would have been a behavior selected for in the domestication process, Hare concludes, though it may go along with tameability rather than being a separate behavior.

  That still leaves open the question of what humans were hoping to achieve when they set about domesticating wolves, given that the eventual outcome could hardly be foreseen. Ray Coppinger, a dog behavior expert at Hampshire College, believes that people can take little credit for the process; it was wolves who domesticated themselves. Wolves are skillful hunters, but they also scavenge. They would have hung around campsites for scraps, and those that learned to be less afraid of people would have flourished, in his view.

  “It was natural selection—the dogs did it, not people,” he says. “The trouble with the theory that people domesticated dogs is that it requires thousands of dogs, just as Belyaev used thousands of foxes.” From the semi-tame, camp-following wolves, he believes, people may have adopted some cubs into the household and found that they could be trained.133

  Hunter-gathering peoples often bring baby wild animals back to camp and keep them as pets until they become unmanageable. James Serpell, an expert on dog behavior at the University of Pennsylvania, thinks this is a more likely basis for domestication than that people adopted wolves that had taken up life as scavengers. If the wolf was domesticated only once, from a group of related animals, there may have been some special feature of these wolves’ behavior that made them easier to train, Serpell suggests.

  However the bond between man and dog was first forged, it proved unbreakable. The Siberians who first ventured into North America via the lost continent of Beringia, the now sunken lands of the Bering strait between Siberia and Alaska, took their dogs with them. This is a surprising finding since researchers had assumed American Indians would have domesticated their own dogs from North American wolves. But Jennifer Leonard, of the University of California, Los Angeles, extracted ancient mitochondrial DNA from pre-Columbian dog cemeteries in Mexico, Peru and Bolivia. She found they matched the DNA of gray wolves from the Old World and not of wolves from the New World.

  The DNA from the dog cemeteries clustered into five groups, suggesting that five different dogs, or sets of related dogs, entered the New World and were the founding mothers of all pre-Columbian dogs.134 For unknown reasons, these pre-Columbian dog lineages have all disappeared. American Indians seem to have preferred the dogs brought in by Europeans. Breeds of dog that were developed in the New World, such as the Eskimo dog, the Mexican hairless dog and the Chesapeake Bay retriever, are all derived from European dogs.

  The First Discovery of the Americas

  Besides inventing the dog, the Upper Paleolithic people of East Asia made another historical contribution: they discovered and colonized the two major continents of North and South America. Genetic comparison of present day Siberians and American Indians may at last be bringing some resolution to two long running academic disputes about the settling of America. One is linguistic, the other archaeological.

  More than 600 languages are spoken by American Indians and they are so different from each other that most linguists have regarded them as being derived from several different language stocks. In many disciplines there are lumpers, who see patterns and commonalities, and splitters, whose preference is to define differences. In historical linguistics, the splitters have the majority. A leading, and generally lonely, lumper is the late Joseph Greenberg of Stanford, a maverick with a fundamentally different view on how to establish the relationship among groups of languages.

  In 1987 Greenberg caused more than usual distress among his fellow linguists when he announced his finding that all American languages fell into just three major families. There was Eskimo-Aleut, a group of 10 languages spoken by Eskimos and the inhabitants of the Aleutian islands off Alaska. There was Na-Dene, a family of 32 languages spoken only in North America, by the Apache, the Navajo and tribes in Canada and Alaska. And finally there was Amerind, a group to which in his view all 583 other languages of North and South America belonged.

  Greenberg well understood that the specialists in various Indian languages would not embrace the idea that their beloved tongues were all splinters off the same block. “I am therefore well aware that what is attempted in this work runs against the current trends in Amerindian work and will be received in certain quarters with something akin to outrage,” he wrote. “Given the investment in time and energy that has led to results different from mine, such a reaction is wholly understandable.”135 The implication that his opponents’ ardor was more substantial than their acumen reflects the general state of relations between Greenberg and his critics.

  Greenberg broadened his linguistic classification into a sweeping and attractively elegant hypothesis. He suggested that the three language groups he had defined represented three separate waves of migration into the Americas from Siberia. There were independent reasons, he noted, from study of teeth and of immunology, for assuming there had been three distinct waves of immigrants. As might be expected, the three migrations are packed into the Americas in order of arrival. The Amerind-speakers, who reach to the tip of South America, were clearly the first. Greenberg suggested they entered America from Siberia about 12,000 years ago and he linked their arrival with the appearance of what archaeologists call the Clovis culture, the earliest indisputable evidence of human presence in the Americas. The Clovis people lived on the Great Plains and hunted mammoth and bison from 11,500 until about 11,000 years ago. At this date the mammoth and several other larger American species became extinct, a customary indicator of human arrival, although in this instance the ending of the Pleistocene ice age may also have been a fact
or.

  Several thousand years after the Amerind migration came the Na-Dene speakers of northwest North America, Greenberg supposed, and last to arrive were the Eskimo-Aleut of the circum-Arctic.

  The relatively recent date adduced by Greenberg for the first entry to the Americas supported his position that it should be possible still to see links between the various Amerind languages. But archaeologists have long been seriously divided over the question of first entry. No human remains older than those of the Clovis culture have yet been discovered, but there are hints of an earlier presence, notably at the Monte Verde site in southern Chile. One layer of apparent artifacts, mostly plant remains and wooden objects possibly associated with tents, has yielded radiocarbon dates of 12,500 years ago, while a deeper and more doubtful layer has produced dates of 33,000 years ago. After initial rejection and long debate, archaeologists have finally accepted the 12,500-year layer, though not the older stratum, as evidence of a pre-Clovis presence, according to a recent review.136 This gives a date slightly before Clovis, but still leaves the impression, at least on archaeological grounds, that the two continents were empty of people prior to 14,000 years or so.

  When the geneticists first arrived on this particular academic battle-ground, they generally favored the idea of a few migrations, though not necessarily just three. But mitochondrial DNA, the genetic element they were first able to analyze, pointed to much earlier dates for the colonization of the Americas, lending preliminary support to the archaeologists who favored seriously pre-Clovis dates of settlement such as 30,000 years ago.

  There are five groups of mitochondrial DNA lineages in the Americas, the groups known as A, B, C and D, as well as the small lineage X, which has a special history. A, C and D are also found in northern latitudes of Asia and in northeastern Siberia but B has a different distribution, being found in southeastern Siberia. That led Douglas Wallace, of the University of California at Irvine, to suggest that the first entry into the Americas occurred 34,000 years ago and consisted of people carrying lineages A, C and D migrating from Beringia. There was then a second migration, 16,000 to 13,000 years ago, according to mitochondrial DNA evidence, that brought lineage groups B and X to the Americas. The Eskimos and Na-Dene speakers appeared on the scene sometime after 10,000 years ago, Wallace suggested.137

  These dates fitted with those derived by several other research groups working on mitochondrial DNA, which ranged from 10,000 to 40,000 years ago. But it now seems that most of these dates, for reasons that are not wholly clear, may be far too old. A second group of geneticists has now entered the fray on the back of the Y chromosome, and they have made a strong case for much younger dates of entry, broadly coinciding with Greenberg’s original thesis.

  The Y chromosome is usually harder to date than mitochondrial DNA, but in the case of the Americas geneticists have been helped by finding a mutation on the Y that occurred just before the first entrants crossed into Beringia and the Americas. The mutation, known as M242, seems to have cropped up between 15,000 and 18,000 years ago, according to Mark Seielstad and colleagues.138 It occurred just before another mutation on the Y chromosome, called M3, which is found almost exclusively in American Indians. So on this evidence the entry to the Americas could not have occurred earlier than 15,000 to 18,000 years ago.

  The genetic, archaeological and linguistic data have been drawn together in what may prove to be a convincing synthesis by Andres Ruiz-Linares of University College, London. He and colleagues conducted a large survey of Y chromosome variations in Mongolians and in American Indians, mostly from South America. They conclude that two major waves of migration from Siberia account for the origin of American Indians.139 Both waves originated ultimately from the southern latitudes of central Siberia. The first entered North America about 14,000 years ago and spread throughout both hemispheres. The second migration arrived later and remained restricted to North America. This migration may have originated from a region of Siberia occupied by the Kets, whose language Greenberg has suggested is related to Na-Dene.

  Many of the tribes of South America show strong signs of genetic drift, an indication that their populations have bred in isolation for many thousands of years. Ruiz-Linares estimates from a DNA signature found in two tribes, the Ticuna of the upper Amazon and the Wayuu, on the north coast of Colombia, that they have been genetically isolated for some 7,000 to 8,000 years. The finding suggests that tribalization—the division into small, warring populations who each defended a home territory—started soon after the first migrants reached South America.

  Early division and an ancient origin for South America’s tribal populations would go a long way toward explaining why Amerindian languages have grown so different from one another in such a relatively short time. If this genetic interpretation is correct, it would explain why the linguistic splitters are right to point to the large differences between Amerindian languages but also why Greenberg was right in lumping all the languages together in a single family.

  Should the new picture emerging from the Y chromosome be confirmed, it will lend support to Greenberg’s idea of three waves of migration. The first arrivals would have crossed the Beringian land bridge after 14,000 years ago and before 11,000 years ago, when Beringia was submerged by rising sea level. They would have spread quite rapidly southward, perhaps by boat if the 12,500 years ago date for Monte Verde in southern Chile is correct. A second migration brought in the Na-Dene speakers, and sometime later the Eskimo-Aleuts arrived. The mitochondrial DNA lineages are now found in all three language groups, presumably because of subsequent mixing between them, but it seems that A, C and D predominated in the first migration, B and X in the second.

  The mitochondrial lineage known as X has turned up in recent surveys among the Sioux, Navajo, Ojibwa and other tribes of North America. Its discovery at first caused considerable surprise because X is one of the founding lineages of Europe. The finding generated some colorful theories, such as that the women from the Vikings’ unsuccessful Vinland colony in North America had been abducted by the skraelings, as the Vikings called their Indian assailants. But geneticists soon showed that the cluster of X lineages in North America was at least 12,000 years old,140 and ancient mitochondrial DNA of the X lineage was discovered in bones some 1,200 years old from a site in Washington state, far distant from the Vinland colony and somewhat earlier.141

  The explanation must be that X, a daughter lineage of N that reached India, participated in two major migrations. As already noted, some women of the X group of lineages moved westward into Europe while their sisters joined the expansion into the Central Asian steppes and Siberia. Many generations later, the descendants of these daughters of X were among the first discoverers of North America.

  Adapting to Cold with Mitochondrial DNA

  The people who reached the Americas from Siberia may have possessed a special quality. The bridge from Siberia to Alaska was not hard to spot—Beringia was a land mass the size of a continent before it sank. So why did only a handful of groups succeed in making their way across? An obvious possibility is that Siberia and Beringia were cold places in which not everyone could survive. It may have been no accident that people of the mitochondrial lineages A, B, C, D and X were the only ones to reach the Americas. According to a proposal by Wallace, these mitochondria may have conferred a special resistance to cold.

  Human mitochondrial lineages, Wallace has pointed out, are geographically patterned not just by continent but also by latitude. The most ancient lineages, L1, L2 and L3, are specific to sub-Saharan Africa. It was only carriers of L3 who moved northward into northeast Africa, and only L3’s daughter lineages, M and N, that left Africa to colonize temperate zones. Wallace wondered if that distribution might be not just a matter of chance, as generally assumed, but rather of natural selection. Mitochondria produce the body’s energy and heat, and survival in cold and even temperate climates could depend a great deal on which lineage of mitochondria a person inherited.

  Mito
chondria release their output in the form either of heat or of an energy-carrying chemical known as ATP. The balance between heat and ATP production can vary, depending on DNA changes in the mitochondrial genes that operate the energy production system. People living in cold climates would be better off with mitochondria adapted to produce more heat and less chemical energy. If so, Wallace argued, their mitochondrial genes should show signs of having been under pressure from natural selection.f In testing mitochondrial DNAs from around the world, Wallace has found that some do indeed bear the marks of positive selection, particularly those of people who live in Siberia or whose ancestors did, such as most American Indians. The groups of mitochondrial lineages known as A, C, D and G are particularly common among arctic people; 75% of them belong to one of these four groups, but only 14% of Asians living in temperate zones do. Some European lineage groups, such as H, also show signs of adaptation to cold conditions.142

  The adaptation of mitochondria to climate, Wallace believes, could explain why when you look at a map of the world the mitochondrial lineages seem to have a more limited distribution than the Y chromosome lineages. The Y chromosome carries rather few genes, most of them related to male fertility, and there is no reason to suppose it is affected by climate. This might explain why men have ranged farther afield than women, at least on a large scale. On a smaller geographical scale, the genetic evidence shows that women move farther than men, presumably reflecting the fact that most human groups are patrilocal and it is the women who move to other societies to find marriage partners.143

  Adaptation to cold may have affected human populations in other important ways, particularly during the Last Glacial Maximum. In Europe, the ice sheets emptied all northern and central latitudes, which were repopulated many generations later by those who had survived in the southern refuges of Spain or the Ukraine. A similar phenomenon seems to have happened in the eastern half of Eurasia and may be responsible for one of the salient puzzles in human population history, the origin of the mongoloids.