Unlocking the Past

by Martin Jones

  The Ocampo caves, high up in the Sierra Madre Oriental mountains of northeastern Mexico have preserved in their sediments many of America’s early domesticated plants. Maize cobs can be recovered from various sediments ranging between 2,300 and 4,300 years old. Just like the maize cobs Goloubinoff had earlier studied, dry conditions greatly aided their molecular integrity. A younger set of cobs were to be found in a separate cave in the Mogollon highlands of New Mexico. Together with a series of modern plants, these provided Viviane Jaenicke-Despres a time series of ancient and modern maize cobs spanning over four millennia. By targeting different kinds of expressed gene, she was able to comment directly on that ongoing journey of domestication, commenting upon which transformations were fast, and which were slow.

  Not surprisingly, a gene that controlled one of the most conspicuous visible attributes of domestication, the very special branching that transformed the wild ancestral fruiting body into something resembling a golden lollipop rather than a more typical grass-head, was there from the earliest cobs; it confirms the pace of architectural transformation visible in the physical cobs themselves. Moving from the physical to the chemical makeup, a gene that managed the storage proteins within the kernel also proved to be an early target for selection. A third gene, also involved in chemical makeup, displayed a different pattern through time. It is a gene involved in the production of starch, and provides an opportunity to go beyond one single species in the food chain, and to explore the co-evolution between consumer and consumed.

  There are two types of starch in plants. They both serve as chemical stores of plant energy. They both form energy rich foods for the human consumers. In many contexts, they may be regarded as interchangeable, but not in the kitchen. The balance between the two forms (amylose and amylopectin) can make the difference between a foodstuff that is sticky and gluey, and one that is crisp and crumbly. In a very functional view of the human world, cuisine may not seem to be a high priority, but archaeological evidence proves otherwise. We had already seen from Terry Brown’s studies of the glutenin how such features as texture and consistency of a meal were valued. In the case of maize, the genetic control of the cob’s starch profile was clearly a target for selection, but along a much more gradual timescale, with incremental changes over thousands of years.

  Switching attention from the consumed to the consumer, starch is not the easiest energy molecule for the primate gut to absorb. Much of our own species’ culinary practice involves predigesting it into small sugary molecules. Primates do, however, carry a gene, or genes, that produce amylase, the enzyme needed to break up starch, to some extent in the saliva and then in the gut. Like many genes actually generating proteins, they may occur in multiple copy numbers. The well-studied amylase gene in chimpanzees, labelled AMY1 has an average copy number of two, and other great apes are broadly similar. Humans, however, have an average copy number of six for the same gene, some individuals less, and some more, even as high as eighteen copies. We can easily make the connection between that contrast and the high consumption by humans of starchy grains such as maize.

  So when did that human response to starchy diets take place? Did the multiplication from two to eighteen copies of the gene happen in the short (in evolutionary time) period since the beginnings of agriculture, or does it have a deeper, more gradual history? Analysis of living humans does display the kind of regional variation to suggest the selection is still active, but much of that evidence comes from grain-fed societies rather than hunter-gatherers, which today are a much diminished group.

  Seven millennia back, Europe had significant populations of both farmers and hunter-gatherers, providing Iosif Lazaridis with the opportunity to compare such genetic traits as amylase production, and AMY1 copy number. He examined the DNA from an individual found near Stuttgart of the Linear Bandkeramic culture (the group of pioneer farmers that Bryan Sykes had earlier sampled to better understand early human journeys), and found a copy number of sixteen. We know enough about the wheat and barley farming of this particular society not to be surprised. To place this within context, Lazaridis identified a series of hunter-gatherers, identified by their distinctive toolkits, more or less coeval with this farmer. These included skeletons from the Loschbour rockshelter in Heffingen, Luxembourg, and the cold mountain cave of La Brana, in northern Spain. Alongside these were seven skulls from the edge of a lake at the Swedish site of Motala.

  The La Brana and Motala skulls produced copy numbers of five to six, still higher than chimpanzees, but only a third of the contemporary farmer. The man from Loschbour stands out from these; his copy number is thirteen, an extreme figure for an individual not consuming much starch.

  Perhaps he just underlines how much we still have to find out about pre-agricultural diets. The rockshelter beneath which he was laid to rest is indeed a good look-out position for migrating game, and an easy access to river fish. As he approached those rivers, he would have the opportunity to uproot tubers and corms, and looking back to the bluff above him, would have seen woodlands bearing nuts in season, with tuberous ferns beneath them. The very diverse diets of hunter-gatherers around the time of the dawn of agriculture could well have included enough starch for a high copy number of AMY genes to be advantageous.

  If we shift the focus from cereal grains to starchy underground storage organs, then the evolutionary timescale of amylase evolution could be longer still. An understanding of that longer timescale has emerged from one of the many outputs of a complete genome study published by Kay Prufer and colleagues of an Altai Mountains Neanderthal. Placing that individual in the context both of great ape data, and the data arising from the 1000 Genome Project, allows us to look in more detail at AMY evolution. What is seen from their work is that the low copy numbers of great apes is shared both by Neanderthals and Denisovans. The high copy numbers relate to our species alone. Communities from Africa, Europe, and Asia all include individuals with double figures. Whether or not this multiplication tracks back to early modern humans in sub-Saharan Africa remains a matter for speculation, but underground storage organs have long been suggested to be a component of their diet. Whatever the case, the diversity between and within communities evident from the 1000 genome project underlines, in relation to this gene, how active evolution still is.

  I have taken the example of digestive chemistry to illustrate ongoing co-evolution between predator and prey. Similar illustrations may be made of other interfaces, for example in the protective surfaces of organisms, and particularly in their immune systems. In Svante Pääbo’s 2014 review, he highlights these three themes. He also reviews our knowledge of the porosity of the human evolutionary family. Just as that porosity applies to the different inferred species of Homo, so Iosif Lazaridis’s study indicates it applies within our own species, between hunting-gathering and farming communities. The family tree he put forward is rather more complicated and more reticulate than its predecessors. Early European farmers remain key contributors to modern European populations, but two other groups, identified as ‘Western Hunter-Gatherers’ and ‘Ancient North Eurasians’ make very significant contributions. Their mixed ancestry of the majority of contemporary European populations is clear.

  Within this co-evolutionary tale, a similar trend is repeatedly seen in the species of food consumed. Reproductive isolation is more often the consequence of the considerable shrinkage or complete disappearance of the non-domesticated form, as has happened with many wild ancestors of domesticated plants and animals, just as it has happened with extant hunter-gatherer communities. There are, however, some domesticates, such as barley, sheep, and foxtail millet, for which wild relatives and possible ancestors remain reasonably widely dispersed, and our understanding of their continued evolution may retain a theme of sustained porosity.

  The way we have thought about agriculture having a simple ‘origin’ is a little bit similar to the way we once thought about the world having a simple ‘origin’. I am reminded of the shrewd figure of James Hu
tton, peering at the fine detail of geological stratigraphy, newly exposed by eighteenth century industrial works. Realizing a simple story was illusory, he instead inferred a world in which ‘we find no vestige of a beginning—no prospect of an end’. As we observe the emerging co-evolution between plants generating large bodies of starch, human digestive systems adapting to starch consumption, and human minds creatively exploring the never ending culinary possibilities of starch chemistry, we see something rather similar.

  6

  ending the chase

  the roots of captivity

  The stretch of land that came to be known as the Fertile Crescent had a double attraction for those in search of agricultural origins. The arc of land that reached from Israel in the west, through south-east Turkey and Syria to the north, and on to the slopes of the Zagros mountains in Iran to the east, was not just the home of the wild ancestors of so many major cereals and legumes. Roaming through the woodlands in the north and west of the crescent were the wild boar and the massive aurochsen that gave rise to domestic pigs and cattle. Higher in the foothills and mountains to the east, two agile grazers, the mouflon and bezoar, dodged their predators. These were the wild ancestors of our modern sheep and goats. Domestication was not just about plants, but about animals too. When, 10,000 years ago, hunter-gatherers made their way across the Fertile Crescent, they came across both the plants and the animals, living in the wild, that would subsequently feed the greater part of the world’s population.

  It was the animals in particular that brought that pioneer seeker of agricultural origins, Robert Braidwood, to the eastern stretch of the Fertile Crescent. He searched the Zagros foothills for pioneer farming sites immediately below the places where mouflons and bezoar roam today. At the 9,000-year-old settlement at Jarmo he was rewarded, and not just by bones of the world’s earliest domesticated goats alongside some very early cereals. The site also yielded numerous clay models of this animal that was clearly so important to this ancient community. While Braidwood’s animal evidence was graphically enriching the story of a localized origin, animal bones were also to prove pivotal in the arguments favouring a more dispersed, evolutionary path to agriculture.

  In Baluchistan, 1,500 km further east from the easternmost part of the Fertile Crescent in the Zagros mountains, beyond the Iranian Plain, the collapsed remains of ancient red mud brick dwellings spread out over twelve hectares of a hillside, perched up above a valley leading down into the Indus Basin. Those mud bricks were laid over 8,000 years ago, and beneath them are the remains of settlements stretching back a further millennium at least. Excavations at this site of Mehrgarh revealed huts and granaries, as well as the remains of some very early barley and wheat. There were also many animal bones, which have been subjected to intensive study by Harvard archaeologist Richard Meadow. In the site’s early levels he found the bones of many of the wild forms of many domestic species, including sheep, goat, pig, horse and cow. In higher, more recent levels, those same species appeared in domesticated form. It seemed that the realm of pioneer domestication might have extended far beyond the Fertile Crescent to the south and east. In parts of the world even further from the Fertile Crescent, the geographical match between domesticates and their wild progenitors was still less clear.

  The animal evidence was central to the arguments of Eric Higgs’s group at Cambridge against a revolution that was restricted in space and time. They argued that the beginnings of animal management were not neatly divided into discrete episodes within centres of origin. Those beginnings were instead widely distributed. They were still proceeding, and had been proceeding well before the visible transformation of wild cereals and legumes in the Fertile Crescent. The group toyed with the idea that, even in previous interglacials, human management of reindeers and other herbivores was on a par with more recent management of domestic animals. Whether or not that speculative idea is persuasive, there is little doubt that the earliest instance of an animal modified by domestication has little to do either with the Fertile Crescent or with agriculture.

  an ancient friend

  The shorter snout and modified teeth that distinguish the domestic dog from the wild wolf have been identified in pre-agricultural sites as far afield as north-west Europe and Japan. Dogs and humans have to some extent domesticated each other, forging a hunting partnership advantageous to both. How ancient that partnership is we do not know, but we can try to find out by using the same evolutionary stopwatch on dogs as we have done on humans, the mitochondrial control region.

  In fact, it is not exactly the same stopwatch. Because the control region evolves so fast, it varies a great deal between different types of mammal. Such variations are one of the main reasons why different species’ mitochondrial DNA varies in length as well as in sequence. Nevertheless, the principle of using the control region as a measure of close evolutionary relationships is much the same in all these species.

  At the University of Los Angeles, Carles Vila embarked on a molecular survey similar to the one conducted on humans in the Mitochondrial Eve project. He took samples from sixty-seven breeds of modern dog, selected from all around the world. They ranged from Australian dingoes to Alaskan huskies, from Afghan hounds to English sheepdogs to New Guinea singing dogs. He assembled a parallel range of wolf samples, from localities scattered throughout the world. To root his phylogenetic tree, he also included coyotes and jackals in the survey.

  The data from these wolves and dogs generated a family tree with a broad spreading crown. Across the breadth of this crown, the dogs clustered into four lineages, each of them with a considerable degree of internal variation. Individual breeds of dog did not match on to particular haplotypes, and a single breed might include several haplotypes. One of the lineages was of particular interest, as it included only domestic dogs, and none of the wild wolves. This looked like a lineage the story of which was completely contained within the history of that most ancient of partnerships, the domestication of the dog. The haplotype diversity was the key.

  So far as we know from fossil evidence, the wolf and coyote diverged about 1 million years ago. The sequence divergence between them, which these data placed at around 7.5 percent, could therefore form the basis of a molecular clock. If we compare that figure with its equivalent measured from the single lineage composed only of domesticated dogs, then some sense of the antiquity of the partnership can be gained. That second measure of diversity is around 1 percent. If 7.5 percent is what you get after 1 million years, then 1 percent is what you might expect after 135,000 years, but let us not be deluded into introducing too much precision. We are, after all, relying here on that uncertain timepiece, the molecular clock. Nevertheless, whatever the imprecision of that estimate of 135,000 years, we would be hard pushed indeed to compress it within the 10,000-year period conventionally associated with farming. Our first domestication is much more ancient than that, and, judging from the spread and diversity of haplotypes, could well have happened at various times and places.

  Soon after the publication of this result, one of Vila’s students at Los Angeles, Jennifer Leonard, set about placing this tree within its archaeological context. She made use of a piece of recent geological history that has been central to the conception and design of a number of studies in molecular archaeology. The two great landmasses on which most humans live, the Old World and the New, have on several occasions been linked by dry land between Siberia and Alaska. This has happened during each cold phase of the last 2 million years, as a result of the fall in sea level that accompanied expansion of the ice-cap. There was a series of discrete episodes when land animals could freely migrate from one land mass to the other. For this to be possible, it had to be cold enough for the sea level to drop and a land bridge to form, but warm enough for such a migration to be feasible. By the time the most recent of these episodes was coming to an end, around 14,000 years ago, our own species had made the journey from the Old World to the New, and so had the wolf or dog, in some form.


  The Native Americans whom Europeans first encountered commonly raised dogs, and some of their breeds can still be found today. There is much we do not know about such ancient American breeds as the Alaskan Malamute, and we cannot even be sure that their ancestry is truly, or fully, American. After all, the European horse was speedily assimilated by the High Plains Indian. It could easily be that European dogs came over in ships and were also drawn into the existing gene pool. To find out more about Native American dogs and their ancestry, Jennifer Leonard sought out dog bones that had been excavated from archaeological deposits that pre-dated Columbus. She managed to track down seven different pre-Columbian dogs from Bolivia and Mexico, and to amplify their DNA. They were quite diverse; the seven dogs generated six haplotypes which could then be positioned on the evolutionary bush that Carles Vila had published a few years earlier. They gave a clear and significant result.

  The pre-Columbian dogs all clustered together with their domesticated Old World relatives, and evidently shared a common ancestry. What this means is that the ancestors of these particular dogs did not make the long journey from Asia to America as wolves, but came with their owners as domesticated dogs. Moreover, they did so several thousand years before the term ‘domestication’ could be applied to any other species, in any part of the world. It is estimated that 14,000 years ago is about the latest that it could have occurred. Many believe that the journey happened much earlier, and certainly the journey from the heart of Asia was much earlier. Vila’s first estimate of over 100,000 years may well remain a reasonable time frame for the most ancient domestication of all.