Unlocking the Past

by Martin Jones

  By the 1990s, the range of spectrographic methods available had diversified, and Richard Evershed at Bristol had added several refinements. He heated his solutions to high temperatures in order to vaporize any molecular residues. These vaporized molecules could then be sorted by using a combination of gas chromatography (GC) and mass spectrometry (MS), predictably referred to as GC/MS. In general terms, GC/MS can be likened to a molecular racetrack, which ends, not at a neat finishing line, but at a crash barrier. The molecules all start off at the same starting line, but as they get going round the track they separate into a rank order according to variations in structure and mass. That is the GC element. The MS element is the crash barrier. The different molecules are forced to swerve, under the powerful force of a strong magnetic or electric field, and hit the barrier in different places, again according to variations in mass and other structural properties. From the order the molecules move around the track, and the point they hit the barrier, Evershed could work out a great deal about their structure. Furthermore, he could look for a whole series of different molecules by scrutinizing his crash barrier after a single spin of his sample round the track.

  After a number of trials with this highly sensitive method, Evershed realized that chemical residues were not at all uncommon. Something like half the pots recovered from archaeological sites retain some trace of what was inside them. GC/MS methods could describe those chemical residues with remarkable precision. Evershed was keen to push the method as far as it would go. An important point in his favour was the fact that most prehistoric and mediaeval pottery had not been glazed. Such smooth surfaces as existed on ceramic vessels in prehistory and the Roman period were generally achieved through burnishing the still damp clay with a stone, or coating it with a fine clay wash or ‘slip’. Furthermore, the clays were tempered with some non-clay mixers, which could be limestone, dung or some other organic matter, and which enabled them to be fired at relatively low temperatures. Thus the prevalent pottery types, even up to the Industrial Revolution, had a relatively coarse and porous fabric. A second point to consider was that their production, and very often their use, involved exposing them to great heat. Although organic molecules survived, they would be transformed by the heat.

  Bearing these points in mind, it seemed likely that the most useful group of biomolecules to study was the lipid group. By definition they are insoluble in water and so would not get washed in and out of the pot’s pores. In addition, many were reasonably tolerant of heat. Unglazed pottery can act as a kind of sponge to groundwater after a while, and there will be a fair amount of movement of water through buried sherds of pottery. The lipids are tucked away within the microscopic pores of the pot and, being insoluble in water, stay put, however much the water moves back and forth. Moreover, the presence of water does not encourage breakdown, as it does in the case of the long-chain molecules. The side branches of lipid molecules may be lost or transformed, but the original core of the molecule may stay intact for a remarkable period.

  Before long, Evershed turned his attention to another of the key food sources habitually overlooked by prehistorians because of the lack of any obvious trace in the archaeological record–the green vegetable foods. Green tissue, in the form of leaves, buds, stems and flower heads, plays a key role in vegetable diets and is particularly important in relation to vitamins and minerals. Yet we archaeologists know very little about the early use of these foodstuffs. We know virtually nothing about that important group of green vegetables, the brassicas, which includes broccoli, cabbage, cauliflower, kale and kohlrabi, simply because these leafy green tissues are too fragile to survive intact.

  Evershed was aware that, while those large soft expanses of green tissue were about the most vulnerable of all tissues to rapid breakdown, there were certain chemical elements within them that were quite the opposite. The waterproofing molecules that prevent those soft tissues from drying out, such as the cutins and waxes, are among the most obstinate molecules of all. They can persist, in slightly modified form, for a seemingly indefinite period.

  Ordinary cabbage, one of the brassicas, owes its glossy protection to a long-chain molecule, composed of hydrogen and carbon and called ‘nonacosane’, together with a few very close relatives that have a little oxygen within them as well. That little family of molecules hit Evershed’s GC/MS crash barrier fairly close to one another, providing him with a characteristic read-out for the group. He embarked upon a search for the familiar cluster in a group of mediaeval potsherds from West Cotton, a rural settlement in central England, all from unglazed pots used in everyday life. A high proportion of them produced a variety of volatile lipids, separating out within the gas chromatograph to produce a range of peaks. A cluster of those peaks repeatedly matched the nonacosane cluster, and in five of the potsherds they were the only peaks. Those pots had a lipid signature remarkably similar to the cabbage leaf itself. They must have been cooking pots for brassicas, used time and time again for the preparation of those particular foods. The leafy vegetables that completely elude the conventional archaeological record were now showing up, not as an incidental food, but as vegetables so important that some pots were used specifically for them.

  a rich and diverse feast

  By looking at a range of molecular markers, we can place the spread of major world crops in the context of the considerable diversity of prehistoric diets that conventional archaeology has failed to detect. There are many ways of living in and using the natural environment, exploiting countless plants and animals–tubers, roots and leaves as well as meat and grain. The spotlight has also shifted to other parts of the global stage. It has cast its beam on the world’s vast tropical regions, poorly visible to conventional archaeological methods, but now progressively coming into focus, their prehistory just as rich and complex as that of the better known temperate and sub-temperate zones. These same molecular methods allow us to probe further into what early people were actually doing with these foodstuffs, how they were taking fresh plant and animal tissue and forming it into the components of a meal. Take, for example, those cabbage pots that retained the molecular residue of the cabbage leaves’ waxy cuticles. Evershed’s analyses went a step further than simply identifying the food within the pot. The GC/MS methods he employed provide an extremely sensitive route to targeting scarce remains of lipids at particular points, and subjecting them to analysis. Using this technique, he took one of his ancient cabbage pots and searched for the cuticle waxes at different points in the interior. Inside the pot at its base the waxes were present, but at a fairly low level. He took samples again at different levels up the side of the interior. At a certain level, the lipid signal peaked. He had found evidence of boiling, and of the level which the boiling water reached and at which lipids tended to float.

  More evidence of boiling came from other mediaeval coarse pots. These were empty pots blackened by fire on the outside, but solvents applied to the inside picked up a lipid residue. Evershed’s GC/MS analysis was once again targeted upon several points within the pot, from the top, the side and the bottom. This time he found that boiling was not the only aspect of the pot’s history to leave a lipid trace. First, the higher within the vessels they tested, the more lipids they recovered. This, he suggested, came about through boiling water-based meals within the pot, causing fats and other lipids to rise to the surface. Second, the smaller quantities of lipids at the base were of a quite different character from the larger quantities at the top. The ones at the base matched with beeswax, while the lipids higher up matched with animal fats. The beeswax was clearly not associated with the bubbling stew, but was sufficiently well bonded with the pot surface to remain in place throughout the cooking process. Around the world it is still possible to find potters who take a newly fired pot, still warm, and smear the inside with beeswax before anything else has had a chance to seep in. This initial proofing of the mediaeval pots from the West Cotton village improved their performance, as they were used to prepare meat stews
. Once again, looking at molecular patterns in space across the surface of an artefact allows us to construct a sequence through time.

  The analysis of seeds and bones that had expanded in the preceding episode of archaeological science had allowed lists to be assembled of the visible foods of antiquity. Now molecular analyses were going further than simply adding to that list the tubers and vegetables that had hitherto remained invisible, and was moving towards the detection of how those foods were being treated and prepared for the meal. This was opening two further windows on the human past. The first relates to our physiology and the way our digestion works.

  ‘External digestion’ is a catch-all phrase for a range of things we do to get the breakdown of food started before it enters our body–such things as grinding, cooking and so on. It is a highly characteristic feature of our species, and is an important factor in the very diverse diet we adopt. We do not possess a particularly remarkable stomach, compared for example with the multi-chamber affair of a ruminant mammal, and yet our dietary range has spread to encompass some surprising challenges. The initial ingredients for our meals can include things too hard to chew or too intractable to digest, and things that are poisonous, purgative and inflammatory. If we relied on ‘internal digestion’ such a range would be impossible, and the use of grinding, fire and fermentation in external digestion has been a major feature of our ecology since long before agriculture began. The Pacific Island taro traces of 26,000 years ago and the similarly ancient evidence of grass-seed consumption in south-west Asia both indicate the antiquity of such transformations. Both taro and grass seed require careful treatment before consumption.

  Work on ancient seeds and bones had given us a partial list of ingredients. Ancient biomolecules were both adding to those ingredients and shifting our attention from the ingredients to their preparation as a meal. There may be hundreds of ways in which any two or three ingredients can be transformed into a culinary delight and shared within small or large communities of people eating together. How those transformations are carried out tells us a great deal about past societies, the numbers of people who lived together, and the manner in which the passage of time was marked out by feasts.

  Anthropologists have shown great interest in the meal. In an attempt to see some pattern in the great diversity of methods of food preparation Claude Levi-Strauss devised a culinary triangle. At the points of the triangle were the ‘raw’, the ‘cooked’ and the ‘rotten’. The various culinary practices found around the world could then be situated between these points. Patterns within the triangle, and further variations in preparation, notably between roasting and boiling, could in turn be related to social practice and ritual that defined relationships and shaped the world. We have only to think of the status of the roast joint of meat, carved by the head of house, a ritual focus mirrored in different social gatherings in many societies. Boiling can be picked up by lipid analysis, but what about roasting? Evershed examined lipids within sherds that had broken away from a characteristic type of large flat dish found on the same sites as the cabbage pots. The component fatty acids were assayed, and a potential source identified. The suite of fatty acids on the dishes derived from animals rather than plants, and from a quite distinct group of animal fats. They had dripped from a roasting pig. The lipid spectrum provided the link between some rather featureless pieces of broken pot and the spit-roasting of a suckling pig that provided an important social focus for a celebratory meal, its dripping fat caught below in these unglazed dishes.

  The third point of Levi-Strauss’s culinary triangle is ‘rotting’. In biological terms, what that means is some form of digestion by a relatively benign micro-organism. Such external digestion is fundamental to numerous food preparation processes, including the rising of bread, the brewing of beer, the fermentation of poi from taro, the production of yoghurt, bean curd, soy sauce and brown-leafed tea. Detecting it in ancient samples is not easy, for two reasons. First, these micro-organisms are particularly efficient at breaking down substrates to very small, and thus non-persistent, molecules. Second, their actions are similar to decay after disposal in the ground, often enacted by similar, if not the same, micro-organisms. However, a Cambridge research student, Delwen Samuel, managed to pick up traces of fermentation by bringing together some of the approaches above, and looking at starch granules within ancient pots.

  In fact, she found a great deal more than just starch granules coated on the inner surface of her potsherds from ancient Egypt–enough material within the modest smear to instil a little anxiety into generations of archaeological pot-washers. What their scrubbing brushes had cleaned away was transformed, under the microscope, into a rich array of foodstuffs, among them recognizable fragments of cereal grain and other seeds. At a higher magnification the granules of starch came into view, their normally even surfaces pitted by tiny holes. She was observing the direct action of fermentation, with individual yeast cells having digested the surface of the starch granules. Following a range of microscopic and molecular analyses, these seemingly unremarkable residues within ancient pots provided enough information for the bread and beer of the ancient Egyptians to be reconstructed. After a few years of her research, a well-known brewery that was funding her work was ready to re-create the beer the Pharaohs sipped. A limited edition brew went on sale at Harrods in London at £50 a bottle, and speedily sold out.

  sharpening the focus: the isotopic signature

  For molecular archaeology to become a real possibility, the sample requirement had to shrink to within the range of what survives the ravages of time in an archaeological deposit. Like the DNA residues explored in earlier chapters, proteins and lipids can now be detected in quantities of less than a millionth, sometimes less than a billionth of a gramme. With methods of such sensitivity, individual cells, or individual points on the surface of an artefact, can be chemically studied. That sensitivity can take us one step further yet, to examine the atoms from which the ancient biomolecules are built.

  But why should we want to look more closely at the atoms–is there anything further they can tell us about the past? In fact, there is not a great deal of atomic variety in the complex biomolecules featured here. For all their intricate diversity and variation, they are very largely built from the same five elements–carbon, hydrogen, oxygen, nitrogen and sulphur. But each element is not always present in a uniform state, and the different states of the constituent elements, or ‘isotopes’ as they are known, are themselves a mine of information. Carbon, for example, is an element that can occur as one of three isotopes, designated C12, C13, and C14. Those numbers are not arbitrary; they indicate the differences in mass of the atoms. An atom of the lightest isotope, C12, is twelve-thirteenths of the mass of an atom of C13, which in turn is thirteen-fourteenths of the mass of an atom of the heaviest isotope, C14. All three can become incorporated into precisely the same molecules. Generally, they will behave in a similar way, except that the heavier isotopes tend to be more sluggish in certain processes. This causes them to accumulate differentially, and that is when they can capture and retain information useful to archaeologists.

  One way in which this sluggishness of heavy isotopes makes itself known is through the food chain. As carbon passes from the atmosphere into plants, on to herbivores, and from there to carnivores and top carnivores, much of it is burnt off as carbon dioxide from each step in the chain. More of the lightest isotope is released in this form, leaving more of the heavier isotopes to accumulate within the solid tissue. So the proportion of the light carbon isotope drops with each step through the food chain, and what gets incorporated into the proteins within an organism remains as a marker of where those organisms are placed in the food chain. In other words, it becomes a record of past diet. If we compare the collagen of three healthy people, one a vegetarian, one with a mixed diet, and one an avid meat-eater, then the proportion of heavy isotopes grows progressively as we move from vegetarian to avid meat-eater. In ancient bodies, we only use the rati
o between C12 and C13 for dietary analysis. C14 is unstable, which renders it invaluable for one of our most important dating methods, but no good as a dietary signal.

  The stable isotopes of other elements behave in a rather similar manner. Nitrogen is also common in biomolecules, particularly proteins, and occurs as one of two stable isotopes, N15 and N16. In a manner similar to carbon, the heavier isotope is progressively enriched when moving up through the food chain. Even within a single body, the relative fractions of the different isotopes will vary, for example between the bone, soft tissue, milk and any surviving strands of hair. The isotopic patterns of a range of elements come together as a dietary diary, a record of who ate what, and when, in the prehistoric past.

  The virtue of looking at a variety of elements, rather than at any single one, is that each element is subject to other factors, and not just to position on the food chain. Sometimes these other factors are not of direct interest to archaeologists. For example, isotopes of nitrogen are very much affected by the soil micro-organisms that find themselves caught up in the human food chain. But on other occasions these additional factors add to the detail we can reveal about past human diets. Carbon isotopes are a case in point. Certain grasses, for example the cereal maize, carry a particular isotopic signature, which is preserved in the bone collagen of someone who consumes a lot of maize. This particular signature derives from the distinctive way in which many tropical and sub-tropical grasses, including maize, capture the sun’s energy. They employ a distinctive photosynthetic pathway, which incorporates carbon isotopes in a very different ratio from most other plants. While the resulting C12/C13 ratio is fairly common in the tropics and sub-tropics, as maize spread north and south into the temperate zones the ratio became highly distinctive. People consuming a lot of that maize would incorporate the same unusual ratio in their own bones. In fact, maize provides one of the key examples of how isotopic information can reveal a great deal of information about the human past.