Unlocking the Past

by Martin Jones

  This brief molecular survey has taken us further and further away from the genetic molecule at the heart of life, first to the proteins that preserve much of the DNA’s sequence data, then to carbohydrates and lipids that are several stages removed, and then to biomineral molecules at the interface between the organic and the inorganic. While this progressive movement away from the genetic code diffuses some of the information content of the molecules, it far from erases it. Indeed, it sometimes reaches into a new area of information, passed traditional archaeology and DNA alike. Over and above this, extending the range of our molecular survey opens the way to study molecules with powers of persistence thousands of times greater than the persistence of DNA, retrievable both from more ancient deposits and from far more exposed locations in the archaeological record. We can now return to the archaeological record to see what information they can provide. A good way to start is by returning to one of the ancient farming journeys and looking beyond the DNA of the farmers and the crops they took with them.

  beyond the maize field

  Some time after wild teosinte was transformed to become domesticated maize, its cross-continental journey began. Maize cultivation spread from Central America, both to the north and to the south. Its path can be charted in both directions by finds of ancient maize cobs and maize seeds, often in the caves and rock shelters used by its cultivators. When the conditions of molecular preservation are right, this journey north and south can be related to the DNA both of the farmers themselves and of the maize cobs they left behind. In each case the DNA survives only selectively, within the heart of a seed or a bone that charted a specific course through the hazards of temperature, humidity, decay and fragmentation. But other biomolecules can brave these hazards with far less protection, carrying information that was in any case hard to glean from the DNA.

  In the first few centuries after its domestication, the farming of maize had reached to the south of Panama. Twenty kilometres inland from Parita Bay in central Panama in the heart of the humid American tropics, a rock shelter called ‘Aguadulce’ (‘freshwater’) can be found. Excavations have demonstrated that this shelter was visited by humans 9,000 years ago and intermittently over the following five millennia. In the growing accumulation of debris they left behind there were grinding stones, perhaps for the milling of some kind of flour, possibly maize. This Panamanian rock shelter lies immediately south of the area in which maize agriculture began. Aguadulce is situated 2,500 km to the south-east of the stands of wild maize and climbing bean presumed to have been at the heart of agricultural origins in the New World. It is on a route into South America along which the spread of maize and the other Central American crops can be traced, just as they can be traced to the north. The finding of grinding stones provides a convenient adjunct to that story of a spread of grain crop agriculture sweeping north and south, and laying the foundations for the early civilizations of the New World. But is that story skewed by an overemphasis on the finds of maize? Dolores Piperno and her colleague Irene Holtz tried to find out.

  They did so by examining fine cracks in the surfaces of the grinding tools. Working with a fine needle, they managed to prise out tiny particles of sub-cellular material, trapped within minute crevices in the surface of the stone. Some of these were composed of silica phytoliths, others were granules of starch. Both the silica and the starch particles retained some of their original shape and structure. On grinding tools that were 7,000 years old the characteristic phytoliths of maize could be made out under a microscope. This is an extremely early date for maize, and perhaps a direct marker of that first southward movement of maize farming, as well as confirmation that the production of maize flour was the purpose of the grinding stone. However, the maize phytoliths were not all that was found. Piperno and Holtz also recovered silica particles that came from two of the cucurbits that were so important to early American farmers–bottle gourd and squash–as well as from the lesser known tuber, lleren.

  Higher up the sequence, in deposits laid down 1,000-2,000 years later, another series of grinding stones had entrapped starch granules as well as silica particles. From their size and shape, maize could still be recognized as one of the seeds being ground, alongside arrowroot, manioc and some type of legume. When the maize arrived at Aguadulce cave it was certainly ground up, but cave dwellers did not lose sight of a whole series of other seeds, roots and tubers that served as sources of flour.

  As grinding stones from other sites were examined in a similar way, this picture of dietary range was placed in the context of time. At a shell mound on the coast at Monagrillo, maize was eaten alongside manioc and palm root between 3,000 and 5,000 years ago. In the village of La Mula, 2,000-3,000 years ago, the list had slimmed down to maize and manioc alone. Around 1,300 years ago, at the village of Cerro Juan Diaz, grinding stones were found with traces of maize alone. The genetics of maize and its DNA variations had provided a framework for charting the spread of maize from its area of origin. Now these ancient grinding stones, and more particularly the tiny particles of silica and starch trapped within their surfaces, were charting the actual manner in which maize was incorporated into the lives and the food chains of people along the path of that spread. The new wonder crop did not spread like a flood, submerging everything in its path. It was gradually assimilated by people who were already grinding up a range of seeds, roots and tubers–plant foods they had known and depended upon for generations. Thousands of years were to elapse before the food chain had narrowed down to the kind of grain crop dominance that came to be associated with the monumental civilizations.

  Roots and tubers are among the more elusive foodstuffs in the archaeological records. Seed crops and bones are a lot easier to track down, because of the visible and recognizable fragments they leave behind. We have, for this reason, tended to build our accounts of agricultural beginnings around them, and this has been reflected in the emphasis of DNA research. It is also reflected in a geographical bias in our accounts towards the kinds of seasonal environments where seed plants, open land and grazing herds are prominent features, regions such as the Fertile Crescent of south-west Asia. Our accounts have had a strong bias away from the tropical regions in which this emphasis upon seed foods is not traditionally as great. Here, roots and tubers have been much more prominent in the human food chain in recent history, and they do not persist within the archaeological record in the same ways as seeds. They will mostly be fragmented and disaggregated beyond recognition. Much of the unrecognizable plant tissue within archaeological deposits may be derived from these sources, and Jon Hather at London University has been leading the way in trying to make sense of it. Not many places in the tropics provide the kind of grinding stones for analysis that Piperno was able to examine, their fine fissures so conveniently trapping sub-cellular plant particles. More widespread are the chipped stone tools that are, less fortunately, fashioned to acquire smooth surfaces and sharp edges. But even these smooth surfaces can attract the attachment of plant-based substances–substances that adhere for thousands of years, retaining information about the human past retrievable in no other way.

  back beyond agriculture

  On Buka Island, which lies towards the New Guinea end of the Solomon Island group, Matthew Spriggs and his colleagues from the Australian National University were excavating in a cave called Kilu. They dug down through just over two metres of cave deposit, first of all through scatters of pottery recognizable as just a few centuries old, to much earlier levels containing chipped stone tools rather than pottery. Throughout the deposits, remains of marine shells, fish, mammal and reptile bones were visible. It was clear what foods had been eaten, but the visible plant remains did not penetrate much below the uppermost levels. Further down in the excavation, the deposits and the chipped stone tools with them went back a considerable length of time, over 28,000 years.

  Spriggs placed some of the small stone flakes under a microscope lens. He saw something he thought might be starch grains, not ju
st on the younger material, but even on the oldest flakes recovered from the excavation. Staining with dyes specific to starch, and scrutiny under cross-polarized lights confirmed his suspicion. He took one basalt flake from the oldest levels in the cave and prepared it for scanning electron microscopy. Starch granules were clearly visible on the smooth basalt surface. Not only did the stone flakes have starch adhering to them, but long needles of crystalline calcium oxalate were also visible. These needles, or ‘raphides’ to use the technical term, are abundant in a family of plants called the Araceae. Their presence, together with the size of the starch granules, enabled Spriggs and his colleagues to argue that the flakes had been in contact with the tubers of taro. In modern times, taro is a staple crop of the islands, its tubers being made into puddings and bread, and fermented to produce a food called poi. Taro is not part of the island’s original flora. The tubers may have been brought by sea, before planting, harvesting and scraping with stone tools, followed by cooking to remove the acrid calcium oxalate crystals still visible on the scraping tools. All this was taking place several thousand years before the earliest evidence of wild cereal collection in any part of the world.

  The member of the team least surprised by the adherence and persistence of biological materials on stone tools would have been Matthew Spriggs’s ANU colleague, Thomas Loy. By the time they worked together on the Kilu Cave material, he had spent a decade developing one of the more controversial strands in biomolecular archaeology. He had evidence that a lot more than starch granules adhered to stone tools.

  blood from a stone

  A 1983 article in Science carried a quite remarkable claim. For over a century, stone tools from prehistoric sites around the world had been scrutinized, measured and drawn, leaving us with pages of fine illustrations of the pristine surfaces of fractured stone. The usage of these tools was in large part a matter for speculation. Now a series of 104 such tools had been closely examined and, of these, all but fourteen had remains of their use adhering to them. Those remains were of hair, feathers, tissues and blood. Two years previously Jerold Lowenstein had published his findings of albumin proteins within fossil tissue, but he was working with protein molecules naturally encased and imprisoned while their host was still alive. Thomas Loy’s traces of blood were quite different. They had remained exposed on the surface of artefacts with which they had made only brief contact, thousands of years earlier. The first examples he looked at were relatively young–up to 6,000 years old–and collected from four sites on the coast and in forest regions of Canada. Since then, he has gone on to seek out much older tools, inspecting their surfaces for any trace that could derive from ancient blood. He has had many positive results, and has claimed to be able to identify bloodstains left when a hunter struck his prey 200,000 years or more back in time. In these blood residues he sensed that a whole new world had opened up for prehistoric research. The most eroded parts of the prehistoric record could now be brought to light through molecular analysis.

  The standard techniques of immunology and staining supported the argument that these were in fact blood proteins, but Loy wanted to go further than that and establish the identity of the blood. For this he made use of an elegantly simple property of the proteins, namely the manner in which they formed crystals. Many blood proteins display slight variations in their amino acid sequences, both within and between species. That is why Cavalli-Sforza could draw on them for genetic information. Those same sequence variations also affected the way in which the blood proteins crystallized. Slightly different amino acid sequences generated protein crystals with different shapes. To identify the shape and thus the identity of the ancient blood proteins on his archaeological finds, Loy re-dissolved a small part of the bloodstain from each stone tool under controlled conditions. He transferred the solution so produced to a microscope slide and allowed the blood to re-crystallize. With the minimal quantities used these crystals were only a hundredth of a millimetre in length, but this was large enough to display the kind of variation he was looking for. It enabled him to determine which animal species had come to grief at the hands of the prehistoric Canadians who used the tools. Among their prey were caribou, black-tailed deer, stone mountain sheep, moose, grizzly bear, snowshoe rabbit and sea lion. There was also some human blood.

  In several more recent instances the archaeological record has surprised us with the survival of its complex organic materials, but Loy’s blood residues on stone tools, as a form of molecular survival, have been more of a challenge to credibility. His results were greeted with not a little surprise and scepticism. The considerable persistence of proteins in intimate association with the mineral fractions of bone or shell was well attested–but on the exposed and smooth surface of a stone tool? It was hard to imagine, but there were two reasons to look at it seriously. First, there was clearly something on the surface of the tools. This ‘something’ was smeared across them in a manner suggestive of usage residue. That much could be established just by inspecting the tools by eye. Second, surfaces are remarkable things, and the chemistry at the interface between two surfaces may be quite different from the chemistry of the same substances when separate. This is something we have already seen with proteins clinging to the hydroxy-apatite crystals within bone. Perhaps these stone surfaces provided a similar raft to carry these large organic molecules through the millennia. Blood within a dead body may have no difficulty breaking down, but that does not mean that blood dried on the surface of particular minerals will act in the same way. We now realize that much of the biomolecular survival in bone is to do with the intimate contact between protein and mineral, rather than any intrinsic durability of the protein itself. Perhaps a similar thing might be true of blood on a stone.

  It is probably fair to say that the key word remains ‘perhaps’–we do not yet have a definitive model for this very exposed form of survival. Loy himself has suggested two mechanisms: first, that one group of proteins in the blood–the serum proteins–will, on drying, denature and desiccate to form a matrix which itself protects the coating from attack; and second, that clay minerals, well known to be strong surface actors, bind with the coating as it dries, and contribute to its longevity. Others have suggested that bonding with fatty acids in the soil could be an additional mechanism. If these issues can be sorted out, the potential would be great, as is well illustrated by a piece of work Loy undertook at the Toad River at the southern end of the Yukon Territory in north-west Canada.

  Loy started out with a series of blades, fashioned from a stone called chert, collected from the region, but lacking a secure date or context. It was not clear how old they were, how they had been used, or what they were used on. Very many collections of stone tools that survive from the world’s prehistory suffer this rather mute status. Loy, however, could detect sufficient traces on their surfaces to recapture some of the living world in which they were used. He found traces of conifer wood at the ends of some of the blades. These blades were apparently mounted on to a wooden handle as knives. He also detected repeated smears of blood, overlying each other. His crystallization methods narrowed these down to bison blood. Along with the blood were traces of hair and other tissue, whose presence and orientation on the blade suggested it had been used to cut through hide in the early stages of butchering the animal. There was even enough blood to acquire the fifty micrograms necessary for radiocarbon assay, as a result of which the blades were dated to the first millennium BC, situating them in the context of what was known about the changing environment. It had been assumed that in the first millennium BC, in this region, hunters in small groups had sought their prey from animals dispersed through the woodland. From scrutinizing the residues, Loy could assemble a different picture. The residues indicated that they also gathered together in larger numbers, on a seasonal basis, to exploit herds of bison in more open terrain.

  potted histories

  If this much adheres to the surface of a stone tool, how much more will attach itself to the rough
and porous interior of a ceramic pot? Fired clay vessels have only been a major feature of human life over the last 10,000 years, roughly corresponding to the epoch of agriculture. In that most recent episode of the human story, they have become one of the most familiar archaeological materials alongside the much more ancient corpus of stone tools. As with stone artefacts, traditional approaches to analysis involve describing size and form and ordering into stylistic sequences, rather than establishing what was carried within them. This is understandable, considering that pots are generally recovered as fragments of empty vessels–particularly empty after an army of archaeological pot-washers has scrubbed them clean. At a molecular level, they are not as empty as they at first appear. Because of their porosity and lack of a glaze, they can have an enormous surface area, mostly contained within minuscule pores. A large unglazed storage jar could quite conceivably have a surface area the size of a football pitch.

  Since the 1970s, archaeological scientists have attempted to recover any lingering residues in order to work out how the pots were used and what they had contained. The first attempts involved rinsing the pot fragments with organic solvents and then analysing the resultant solution through infra-red spectroscopy, by then a well established method of identifying organic molecules in very small quantities. With this method it was possible to establish that lipids were preserved, and, for example, to speculate on the various oils, pastes and preserves that had been transported across theRoman Empire. Their lipid residues could be sought within the large, pointy-bottomed amphorae that are a familiar find within Roman shipwrecks, and as prestigious items in high-status graves along the Empire’s periphery.