by Martin Jones
The existing diffusionary journeys had been dispatched, but a new set of journeys was on the horizon, based on fairly new scientific data that were still only sketchily understood. None the less, both data-sets were about to expand rapidly. The number of carbon dates was still increasing, and Cavalli-Sforza was persuaded the time was ripe to push ahead with collating the genetic data. In 1978 he embarked on an eight-year collation of molecular genetic data from living humans, at the end of which time he had persuaded Colin Renfrew and others that his genetically traced journeys were well worth taking seriously.
prehistory from gene maps
Before DNA science fully came into its own, some of the most reliable information about genetic groupings of humans came from blood groups. These groups relate to slight variations in the protein structure of certain components of blood, which in turn arise directly from genetic variation. They range from the well-known ABO groups, through variations in an important group of proteins known as the immunoglobulins, to some obscure variations little known beyond the medical world. By the time Cavalli-Sforza set to work compiling and collating information on human genetic diversity, there were data on numerous protein groups displaying genetic variation, for which medical and anthropological records allowed a global survey. With his colleagues, Paolo Menozzi and Alberto Piazza, he set to work gathering data for all the gene loci which had been studied and charted in sufficient detail for global comparisons to be made. Each of these coded for a particular blood protein or enzyme, which varied between different people.
An important aspect of such variation is that it doesn’t disrupt bodily function. Three people in the same community could quite easily be of A, B and O groups respectively without knowing they were different in any way. This apparently non-functional nature of the variation renders it an ideal tracker of lineage, undeflected by ecology and selection. The geneticists gathered information on protein variations from about 500 different human populations around the world, and set about building trees and constructing maps to represent the global origins of human genetic diversity. What emerged was a masterpiece of genetic analysis, placed in the context of a meticulous survey of human archaeology and language history. It provided a synthesis of great journeys in the human past that matched the scale of Smith’s thinking, but with one key advantage over that earlier author’s argument. This synthesis was in keeping with the scientific evidence for the dating of the human past.
An example of what they achieved is provided by their new enhanced analysis of Europe. They now had information on almost 100 gene loci for the continent, all varying in slightly different ways. To present all this variation would require a diagram of around 100 dimensions. It simply could not be displayed on a two-dimensional map. Fortunately, a statistical procedure exists that can flatten such multidimensional monsters into a two-dimensional shape. It is called Principal Components Analysis (PCA), and the flattened pattern so produced is called the ‘first principal component’. The sheer process of flattening not surprisingly squeezes out a great deal of the multidimensional information. Indeed, quite commonly more is squeezed out than is left in. The PCA can do something with that, however, by going for a ‘second pressing’ to produce a ‘second principal component’ that accounts for another portion of the information. The sequential pressing can go on until it is judged that little information remains to be squeezed out. What sounds like olive oil production is, of course, a number-crunching exercise proceeding within a computer. Like the series that ranges from a first cold pressing of ‘extra virgin’ oil through to a series of cruder, less expensive oils, PCA produces a series of principal components, each successive one accounting for a little less of the desired information. On a map of Europe, the first principal component of genetic variation forms a pattern resonating with Cavalli-Sforza’s earlier result with Albert Ammerman. A series of bands runs from south-west Asia to north-west Europe. The idea of an expanding wave of farmers moving across Europe was holding up. The second principal component highlighted a community on the fringes of this wave, and not actually a part of it. The reindeer-herding Saami from northern Scandinavia (formerly known as the Lapps) had followed a separate path from north-western Asia. The third principal component seemed to the authors to mirror another popular expansion, that of Maria Gimbutas’ horse-riders of the steppes. Each principal component was elegantly pointing to a particular expansion, migration, or set of journeys.
Each of the other continents in the world was mapped in a similar way. Genetic variation was broken down into a series of maps displaying the first, second, third principal component and so on. Each of these twenty or so maps relayed a story of one or more movements of ancient peoples, carrying with them genes, language and culture across the land. Two major types of movement stand out in their interpretation of these maps. The first is the series of population movements that first brought anatomically modern humans to particular landmasses. The second is the series of population movements that brought pioneer farmers to take their ecological advantage to new niches, often replacing hunters and gatherers in their path. In terms of the human past, twentieth-century archaeologists ended the century as they had begun, by explaining the global diversity of human societies in terms of a simplified model of population movements and transfer of ideas. At the beginning the explanation was based on skulls and objects; at the end it was based on molecular genetics and radiocarbon dates. Today we can also reach back into the molecular genetics of the past and ascertain whether the ancient travellers themselves conform to that new global story.
first steps across america
One of these journeys had been a source of fascination long before Grafton Elliott Smith. The contact that followed Columbus’s discovery of the New World opened European eyes to the fact that not only was there another landmass beyond the Old World, but somehow people had arrived there before Europeans had conquered ocean travel. Several centuries after Columbus set sail, we now know that America was not always encircled by water. In each cold phase of the last 2 million years, Siberia and Alaska were linked by a considerable tract of dry land, exposed by a drop in the level of the sea. Nevertheless, no hominid species apart from our own made the crossing, and it was a late crossing at that. The earliest dates for evidence of humans in the New World go back around 30,000 years, and even these are contested. Some would argue that the record is only half that old, at most. Unlike the other journeyers who feature in this chapter, the first people in the New World had not yet turned to agriculture, but we now know that they did have one domesticated species with them, the dog. That much is clear from the DNA evidence considered in the previous chapter. This most ancient of domestications may have played a key part in enabling those ancient people, more accustomed to the middle latitudes, to survive in the frozen, windswept landscapes of the northern crossing. Their dogs could pull heavy burdens on sleds as they moved across hazardous terrain, and would have been vital in helping humans cull the migrating herbivores that grazed upon the Arctic sward. Whenever the first crossing was made, it appears that by around 12,000 years ago, their descendants had reached America’s southern tip. Now modern humans occupied all the world’s major habitable landmasses.
America’s first inhabitants did not leave an elaborate material trace behind them. It is not really possible to construct a detailed story of migration and colonization from their artefacts alone. Scholars have turned to other sorts of patterns, including those arising from language and genetics. What they have looked for is a set of wave fronts, something akin to the traces the tide leaves on a beach. If we carefully mapped the arcs of flotsam described upon the sand, then with a bit of luck the story of each successive tide could be charted. That is what Christy Turner thought when looking for variations in the teeth of living Native Americans. The teeth are clearly the most accessible part of the skeleton in living people, and are presumed to contain a range of genetic signatures. Turner arrived at a three-way classification that might correspond to ‘arc
s in the sand’. The arc closest to the northern crossing encompassed two northerly groups: the Inuit and the Aleut. The latter were the inhabitants of the final remnants of the submerged land bridge, the Aleutian islands that reach out from Alaska in the direction of Siberia. Further to the south, a second arc encompassed inhabitants of the north-west coast. The third and final arc drew in the rest of the New World.
Turner’s pattern caught the attention of a historical linguist, Joseph Greenberg, who saw a similar pattern in Native American languages. He grouped those numerous languages into three ‘super-families’ that displayed a rough match with the dental groups of Turner. In the north were the Inuit and Aleut. In the north-west and in a few places further south, Greenberg recognized a group of ‘Na-Dene’ languages. The rest of North America, and all of South America, was rather contentiously lumped together to match Turner’s third arc, and referred to as the ‘Palaeo-Indian’ group. At first, these two patterns shared a fairly uncertain relationship with genetics, but as the protein and DNA evidence was brought into consideration, some could see a match. Certain blood proteins mirrored the division, and the DNA displayed a north-south banding of some sort which, with a little bit of gentle persuasion, could be coaxed into the three-wave pattern. Even before that DNA evidence was available, Greenberg could speculate on the actual pioneer journeys into America.
One of these journeys began from one of the north-flowing rivers that drain from the Siberian mountains into the Arctic Sea. From the River Lena, this journey passed through the heart of ancient ‘Beringia’, the name given to the vast expanse of land exposed by the lowering of the sea. Occasional herds of large game, upon which the travellers preyed, crossed this region of arid Arctic steppe. Those travellers would be the ancestors of the ‘Palaeo-Indians’.
Another journey began from the regions between these two great rivers, the Lena in the north and the Amur in the south. Here, foraging communities lived among the cold valley woodlands, very different from the distant Beringian steppe. However, as the climate warmed and ancient Beringia began to drown, the steppe gave way in many places to a mosaic of wooded and watery land, a new kind of corridor to which the valley woodland peoples were already acclimatized. These would be the ancestors of the Na-Dene speakers, who established themselves at various points in America’s north-west.
A third journey started in the lower reaches of the Amur River, which separates modern-day Manchuria and Siberia. The fisher-foragers from this region made their way along the land ridge that separated the Sea of Okhotsk from the North Pacific, now a string of islands but once a substantial land corridor. From there, they continued to hug the south coast of ancient Beringia on their way to the New World, sustaining their fishing and foraging traditions throughout. These would be the forefathers of the Aleut and Inuit.
The new evidence coming from mitochondrial DNA then bolstered Greenberg’s account. The global mitochondrial surveys had indicated that just a few branches of the human evolutionary bush had made their way into the New World. There had clearly been a genetic bottleneck, and it seemed reasonable to assume this related to the difficult passage across the frozen north, those pioneering journeys that Green berg had described, across the land exposed during episodes of low sea level. This bottleneck allowed only a small cluster of lineages through from Asia. The geneticist Douglas Wallace identified four mitochondrial lineages, which he labelled A, B, C and D. They were distinguished by a small number of recognition sites around the mitochondrial genome, and in one case by the presence or absence of a ‘deletion’, a mutational loss of nine base-pairs at one point along its length. Within America, there was a clear trend towards a north-south banding of these haplotypes, with A most prominent in the north, B in the centre and south, and C and D becoming more common towards the south. All four lineages were found to be present within the most diverse group, the ‘Palaeo-Indian’ speakers, while the Na-Dene speakers were all of lineage A, and the Inuit and Aleut were either A or D. It was not a simple match, but it none the less seemed to offer support to Greenberg’s three journeys. Now that DNA had been brought in, the molecular clock might also date those separate journeys. Wallace dated the first journey between 26,000 and 24,000 years ago, the second journey between 15,000 and 12,000 years ago, and the third journey at between 9,000 and 7,000 years ago. An engaging account, but not everyone was convinced. A number of counterarguments were arising from a closer examination of DNA variation.
Among the critics of this account was one of Douglas Wallace’s students who, like a good student does, went on to challenge his mentor’s ideas. Andy Merriwether shifted his attention from the Native American communities in the front line of European contact, to a tribe which had managed to survive in the remote depths of South America’s Amazon forest. There is a group of Indians living around the northernmost reaches of the Amazon River in Brazil known as the Yanomami. Merriwether checked their mitochondrial lineages to see whether this remote community too could be brought into the four-lineage pattern. After sequencing a number of individuals, it became clear that each of the four lineages, A-D, was well represented. However, he was not convinced there were only four. It seemed there were seven such lineages, perhaps more. While such a remote people might be expected to harbour the occasional rare lineage, a closer inspection of other Native American groups was revealing other occasional deviations from the four-lineage account. Looking across the continental data, Merriwether arrived at the observation that all four lineages were widespread, but peppered across the map were instances of rarer lineages. It seemed to him that there was a much more straightforward explanation than Greenberg’s and Wallace’s three journeys. A single journey could account for the whole picture, the geographical variations being simply the consequence of sampling. Imagine a pack of fifty-two cards, each with the rather unusual ability to reproduce. Scatter those cards across the Bering Straits into America, and let them fall into clusters or ‘hands’. One hand may contain rather a lot of hearts, another have no kings, and so on. If these hands reproduce mainly locally, then those founding biases will be retained, and that is how Merriwether argued that a single entry could produce the kind of variation recorded in the living population.
Meanwhile, a young molecular biologist working in the German University of Hamburg believed he could get closer to the issue by shifting the emphasis from the handful of DNA marker sites used in Wallace’s argument to one of the most informative regions of the mitochondrial genome. Peter Forster turned his attention to the DNA sequence that has been key to so many of the projects discussed here, the first hypervariable segment of the mitochondrial control region.
Small variations within this fast evolving segment allowed Forster both to separate the founder lineages and to explore the details of their evolution. Models could be built of how they were related, and what had happened to them, before and after passing through the genetic bottleneck. These models were presented as networks of evolutionary pathways, and the networks so produced tallied better with some migrational stories than with others. Forster scrutinized his networks carefully, and concluded that all known founder lineages arrived as part of the same episode. He employed the molecular clock to place that episode around 20,000-25,000 years ago, spreading to the tip of South America by 13,000 years ago.
While this core network accounted for all founder lineages, it appeared to be overlaid by a second network, which could best be explained by a similar episode, but a more recent one, half its age or less. These Forster attributed to the Inuit and Na-Dene populations. However, rather than returning to Greenberg’s three discrete journeys, Forster saw this double episode as part of a global phenomenon, nested within the earth’s climatic history. His two episodes lay either side of the considerable expansion of ice from the poles. Like other northerly peoples, the populations of ancient Beringia expanded when the climate allowed it, and contracted when it did not. Furthermore, Forster’s climatically driven account cast a new light on the genetic ‘arcs in
the sand’.
waves of migration or depletion?
The wave-front model of migration and its relation to ‘arcs in the sand’ would work better if particular genetic types were found to be absent beyond the arc. That would indicate that a particular lineage only got so far. What we find instead is that lineages may be rarer or absent in the lee of the arc. The C lineage is less frequent in the north, and two lineages seem to be completely absent along the presumed pathway from Asia into America. One of these is lineage B; the other, known as lineage X, is one of the rarer lineages that had begun to turn up. In both cases, some had speculated on quite separate journeys to account for their complete absence from the northern land route. Could lineages B and X have nothing to do with the northern bridge and instead have reached America by sea?
The first ancient DNA data from pre-Columbian America could in fact have supported the idea of arrival in part by sea. These early pre-Columbian amplifications yielded only A, C and D lineages. When Erika Hagelberg’s study of South American mummies finally came up with the nine base-pair deletion characteristic of lineage B, a short-lived but romantic notion of ancient Pacific island journeyers began to recede from the account of American settlement. Since then, a significant proportion of the 500 or so pre-Columbian individuals examined have proven to be of lineage B, including an 8,000-year-old hunter stranded 3,000 metres up in the Rockies. The B lineage was clearly an early overland arrival, much as Forster and others argued, which had disappeared from the path of its entry. This left the yet more enigmatic lineage X.