Unlocking the Past

by Martin Jones

  Goloubinoff amplified a 315-base-pair sequence from the target sequence in a sample from each of these ancient specimens, together with modern specimens of both domesticated maize and its wild ancestor, teosinte. All three ancient specimens yielded a DNA product, including the carbonized sample, adding support to Terry Brown’s findings. Two results were of particular interest. Judging from these results, the full range of domesticates did not narrow down to the particular populations of wild teosinte inferred from the isozyme evidence. Second, the ‘speedy clock’ argument did not work. If the molecular clock had speeded up as some had suggested, then the variation among ancient specimens would be considerably less than among modern samples. What Goloubinoff found was the reverse–it was almost 30 percent greater. The molecular clock was not running fast, and the youngest common ancestor for the ancient and modern maize cobs was placed millions rather than thousands of years back in time.

  The results do not actually dispose of the single-origin model. They simply demonstrate that the domestication of maize is not marked by a severe genetic bottleneck. Multiple events of some kind have retained many lines of wild teosinte within the gene pool of domesticated maize. There is a variety of scenarios to explore. The first American farmers could have started with varieties of wild teosinte significantly more diverse than those that survive today. Alternatively, they could have picked up the diversity from wild pollen as they journeyed south, although they would have soon got past the distributions of wild teosinte. The third possibility was that maize was brought into domestication, not once, but many times, and perhaps in more than one place.

  This work was done in the early days of ancient DNA science, and a number of refinements could now be added to narrow clown the possibilities. These days we might tighten up the research design. A further look at geographical variations would help. The wild pollen was a problem, especially as maize and teosinte are cross-pollinating plants that freely release their pollen to the air. It might be interesting to examine the immobile female line through the chloroplast DNA. This has been done on modern maize varieties, and has tended to confirm that the maternal lines in domesticated maize incorporate a wide range of wild teosinte lines. Something we cannot do with maize, but can do with several of the other grass species that have become major cereal resources, is to focus upon self-pollinating plants that release little pollen to the wind. These various refinements have been brought together on another ancient DNA project on the far side of the globe from Mexico.

  another fertile crescent in the east

  Richard MacNeish was guided in his search for early maize cobs by Nicolai Vavilov’s approach, using the distribution of living relatives of the modern crop. This took him to Mexico to search for early sites within the centre of diversity for maize and beans. In the Old World, the equivalent centre of diversity for wheat, barley and a range of other useful plants and animals took Robert Braidwood to the Zagros mountains in the Fertile Crescent of south-west Asia. Further to the east lay another fertile crescent–another centre of crop plant diversity. This ‘Eastern Fertile Crescent’ took in a series of high plateaux at the south-eastern end of the great Himalayan range. Starting in the west around Assam, it stretched through Myanmar (Burma), northern Thailand and Laos, and into the southern provinces of China, Yunnan and Guizhou. This is the realm of that other major world crop, rice. It is here that the wild relatives of rice flourish in their greatest diversity, sometimes in the high plateau marshlands reminiscent of the paddy fields that imitate them.

  The Eastern Fertile Crescent became the target of a number of explorations of the eastern beginnings of agriculture, particularly in India and Thailand, the areas most accessible to archaeologists from the west. Early dates were published, stretching back as far as the seventh millennium BC. Somewhere here, perhaps, the original rice farmers could be tracked down. Rather as with the early American farming sites, the early dates failed to stand up to close scrutiny. In time, they were adjusted or discarded, and the antiquity of rice farming within the Eastern Fertile Crescent has now shrunk to less than 5,000 years. At the same time as these dates were being put in order, news was arriving from China that changed the whole story.

  Much earlier sites, rich in carbonized rice grains, chaff and straw, were coming to light well outside the Eastern Fertile Crescent. In the lower reaches of the Yangtze River, sites such as Luojiajiao and Hemudu were yielding rice that was 7,000 years old. By the late 1980s, even earlier sites had been discovered, with vast quantities of rice preserved in a waterlogged state. Settlements at Pengtoushan and Jaihu were harvesting rice 8,000-9,000 years ago.

  Even further outside the Eastern Fertile Crescent, a rice geneticist working in the city of Shizuoka in Japan had been taking an interest in these early sites from the Yangtze River. Before constraints on movement around China relaxed, Yo-Icho Sato’s main contact with these new discoveries was by word of mouth from Chinese graduate students who had travelled to work with him in Japan. It was still not politically easy for Sato to go to the sites to sample for himself, but his students were able to bring specimens of the ancient rice back to his lab in Shizuoka. After a while he had assembled enough ancient samples to explore their genetic relationship with the better documented material from further south. By the early 1990s, he had read the reports of ancient DNA findings published by Russell Higuchi and Ed Golenberg, and was keen to try them out.

  The first thing he noticed, simply from microscopic examination, was that some of the ancient rice from Hemudu was wild. Vavilov had been building his argument before it was generally realized how much climate, and the vegetation patterns dependent on it, had changed over the millennia. These Hemudu rice grains provided direct evidence of that fact. Wild rices had clearly changed their distribution over the last few thousand years. Careful survey of the modern flora has now yielded wild rice, found growing some way outside the Crescent, though we still have to factor in climate change to push the distribution far enough to the north to encompass the earliest sites.

  Sato turned to the chloroplast genome, which in rice could be split into two groups, according to a stretch of DNA that was sometimes missing. Sixty-nine base-pairs found in the chloroplast sequence in some types of rice were completely absent in others. This mutational accident, known as a ‘deletion’, happened far back in the ancestry of the rice. This is clear from the fact that many wild rices also carry this deletion. It is faithfully reproduced from generation to generation, and so now serves as a lineage marker.

  Sato’s trial run on ancient rice DNA was carried out on a 1,200-year-old sample from Japan. The DNA was remarkably well preserved, with strands of over 1,000 base-pairs amplified, and he was encouraged to move on to his precious collection of very ancient Chinese rice grains. The results he gained from these displayed a clear contrast between the middle Yangtze valley to the north, and the Eastern Fertile Crescent to the south. In the wild and domestic rice alike, the 69-base-pair deletion proved to be absent from the Yangtze specimens. When these sites were occupied, the climate had been a few degrees warmer, enough for perennial rice varieties to spread further north than the area of their current distribution. Perhaps it was on the edge of that distribution that significant changes took place. Here, where the viability of wild rice lessened, so it was brought into domestication, switching from perennial to annual habit. There is no sign of a genetic path leading to the deletion-carrying lineages, which Sato argues derive from the annual wild rices to the south. Here was a well-argued case for multiple domestication of a species, in locations separated by 1,000 kilometres or more.

  a return to the cradle

  Maize and rice are crops of global importance and each has a long prehistory. However, their past in each case remains significantly shorter than in the region to which Childe drew attention, and which has been thought of as the cradle of civilization for many generations. This is the Near East, in the region described as the Fertile Crescent. It may be that future discoveries may take rice agr
iculture that little bit further back in time to make up the difference, but they have not yet done so. For many, the Fertile Crescent 10,000 years ago remains where it all started, and at some point within that region are those original pioneer settlements, whose occupants changed the course of history forever.

  Many excavations have now been conducted along the Fertile Crescent, especially along its western flank. The changing political fortunes of Iraq and Iran have made it difficult to follow up the work of Braidwood’s team along the eastern flank, but early sites are now known from Syria, Turkey, Lebanon, Israel and Jordan. There have recently been two principal contenders for the primary place in the transition to agriculture. Some researchers have focused their attention upon a cluster of four sites in the southern part of the Jordan valley, of which the best known is Jericho. The four sites are very close together, and may possibly have domesticated wheat and barley more than 10,000 years ago. Others have argued that the cereals from that date were wild, and that a group of sites 700 kilometres to the north in south-eastern Turkey have more securely identified early domesticates. What is not in question is a key series of founder crops that were domesticated in that very first episode of agriculture: two species of wheat – einkorn wheat and emmer wheat–one of barley, and a small group of legumes. The key to really getting to grips with the single versus multiple origins debate was how to understand these founder crops in the very first stages of domestication.

  Terry Brown’s group had been working with wheat from the outset of the ancient DNA project. Since their amplification of DNA from the carbonized wheats from Danebury, they had moved on to explore other carbonized deposits of even older dates and from elsewhere in Europe. One of these was a packed grain store in Greece that had burnt down 3,300 years ago, leaving great caches of carbonized grain intact within their storage bins.

  One of their target sequences was a gene group responsible for a series of proteins called the high molecular weight (HMW) glutenins. These are proteins formed in the shape of a flexible spring. Our modern ‘bouncy’ loaves owe their texture to this springy protein. Because of their importance in baking, they had been well studied by modern dieticians, and were known to display the kind of variation that might provide a clue as to their recent evolutionary history. Each species of wheat has a certain number of glutenin gene loci, each occupied by a certain version or ‘allele’ of the gene. Some of these alleles no longer do any useful work. Mutations have disabled them, turning them into ‘pseudogenes’ carried like dead weight from generation to generation. Others have acquired mutations that slightly vary the shape or length of the molecular ‘spring’ they are building, but not in a way that damages the plants. Among these are alleles that make particularly well-formed springs, not in terms of the health of the plant, but of the interests of the baker in making light and springy bread. All this variety is the raw material from which evolutionary stories may be assembled and narrated.

  One interesting finding was of the bread-baking allele. This was the one that optimized the dough’s ability to retain large bubbles of carbon dioxide and produce a light elastic loaf. Modern bakers have certainly selected for this allele, though for how long is uncertain. We are not sure how the artificial selection for any such character was effected prior to the nineteenth century. Carbonized loaves such as those from Roman Pompeii show that leavened bread was made, though not necessarily with specially selected flour. After examining ancient DNA, we can now say with confidence that Greek bakers had already selected for the very same genes that modern bakers favour to produce a light and springy loaf. They had done so well over 1,000 years before Pompeii was buried. Not only were prehistoric communities selecting for bigger and easier harvests, but they were also systematically targeting attributes invisible in the field that required a clear understanding of the notion of a genetic line.

  The glutenin genes had yet more things to tell about the past. One particular aspect of the gene caught the attention of Robin Allaby, one of Brown’s team at Manchester. At one end of the gene, the promoter region, the sequence varied quite significantly. Using the molecular clock estimation, he reasoned that the different lines had diverged from one another 1-2 million years ago, in other words, a very long time before agriculture had begun. This in itself is not surprising–the genome of any taxon preserves quite a number of ancient legacies such as this. Allaby was interested in treating the different phylogenetic branches from the split as markers, to try to make out patterns of evolution and dispersal in the crop. A starting point was to look at a little known wheat, which was once at the heart of Old World agriculture–a species called emmer wheat.

  Although so important to prehistoric farmers, this particular species of wheat is now a rare crop. The last 1,500 years have seen its progressive replacement by modern bread wheat. Twentieth-century agricultural improvement could well have wiped it out, were it not for the recent move to conserve genetic diversity. The scattered emmer wheat populations that survived have been sampled and stored with two major cereal archives at Gatersleben in Germany and Aleppo in Syria. Allaby worked his way through these archives to chart the patterns within the promoter regions of their glutenin genes. A very clear pattern emerged.

  His population of wheats clearly subdivided into two distinct branches, following an evolutionary split some 1-2 million years ago. He labelled the two branches from this split the alpha and beta clades. The alpha clade was the more common and the more widespread, found in emmers throughout the world. By comparison, the beta clade was far more localized. It was encountered in sites in the eastern Mediterranean running into central Europe. When he turned his attention to the wild precursors to domesticated emmer wheat, found today in various parts of the Fertile Crescent, another pattern emerged. The beta clade was quite common towards the Anatolian end of the Crescent, where those early farming sites that attracted Heun’s attention are to be found. However, the alpha clade peaks further to the south, in Israel. This is where a second series of early farming settlements has been found, also with some very early radiocarbon dates for crop domestication. Taking the archaeological and genetic evidence together, it seems that a very similar sequence of events happened at least twice, the same key transition in human ecology, and propelling two very similar spreads of the new crop. In the longer term, one came to predominate across the world, but even so, a substantial genetic trace of the other has survived. There could quite conceivably have been several more such trajectories, one overlaid upon another, of which these two happen to show up on this particular stretch of the DNA sequence.

  an emerging picture

  There have been two ways of seeing and interpreting this major turning point in the human past. The first of these has emphasized isolated revolutionary events, a radical change of direction, rather like a major scientific invention of more recent times. The second has instead emphasized dispersed gradual processes in response to similar global pressures, the convergent evolution of communities steered down the same path of adaptation to variations in nature. Underlying these differences was a wide spectrum of views about the nature of human history and progress. At one end of the spectrum was the view that humans were on a path of improvement by historical action and unique invention. At the other end was the view that we, as much as other species in nature, were subject to the rough and tumble of global fluctuation and natural selection. In that view, the transition to agriculture was just one facet of our adaptation to that process.

  The first interpretation has an affinity with a single isolated origin and an outward spread from there, from time to time triggering secondary domestications. By contrast, the second interpretation would have an affinity with widespread multiple domestication. Such genetic, molecular and archaeological evidence as we now have fits comfortably with neither extreme. The most parsimonious way of accounting for it is as a transition that is rather patchy and bunched, both in space and in time. Some species enter domestication through the kind of tight genetic bottleneck t
hat would suggest a very restricted event, but others do not. In the Neanderthal debate of Chapter 3, the ancient DNA was considered in relation to contrasting hypotheses that were also rooted in deep-seated notions about human history and progress. In that instance, the DNA came down on one side of the argument. With the transition to agriculture, a more diverse picture comes into view. The fine detail of the living past, rather than supporting one or other unequivocal pattern, is pointing up the diversity and contingency of past life. When we project back from the diversity of the present to find its roots in the distant past, the simple explanations are the favoured ones. When, however, we inject a rich and detailed body of evidence from the past itself, perhaps it is not surprising that those simple explanations become less persuasive, and the past becomes as diverse and intricate as the present.

  afterword

  This chapter opened with the question ‘did human evolution stop?’ This was always a rhetorical question, or more a question about why archaeology often narrated the later episodes of our past as if it had stopped. Looking back over the shape of that question and the responses it invoked, it becomes clear how much we relied, as archaeologists, on physical, structural features, both in terms of humans (changes in dentition and the skeleton), and in in terms of the new arena of visible evolution, agriculture (changes in grain size and from brittle to tough rachis forms). Evolution is about those structural traits, but about much more besides, and with biomolecular evidence and genetics, we have become much more able to chart and discuss those other traits.

  These may also be physical in essence, and concern such things as colour, surface texture, and coatings. They can be more to do with chemical characteristics than traits of physical appearance, for example with hormones and the immune response system. Or they can be about behaviour or adaptation to seasonality. As mentioned in the preceding afterword, the emphasis of archaeogenetic work has, over the past two decades, expanded beyond the analysis of phylogenetic trees generated from non-coding regions, in which the nodes become points of ‘origin’ of distinct and internally coherent categories. It has shifted towards studies of the functionality of such coding regions as have been identified, and unfolding into the kinds of traits listed above. The more we become aware of that extensive list, the more ‘origins’ have dispersed, and boundaries become porous. Domestic and wild forms continue, at least to some extent, to interbreed; genetic patterns continue to change, both in the consumed and the consumer. Evolution most certainly carries on.