England: The Butser Experiment
Second thoughts about the Iron Age and the rubbish heaps of history
THERE IS A stretch of highway over the chalk downs of Hampshire, a few miles south of Petersfield, where the road arcs gently down and east and then curls steeply up to the west. Snug in the dell, at the base of Butser Hill, is a farmstead.
It lies in a part of Britain that is known to have been settled and farmed intensively for at least three thousand years. Because the soil cover is so thin, excavations on the downs frequently uncover long, narrow furrows in the underlying chalk—etched by wooden ards, ancestors of the modern plow, as they scored the ancient ground. Here and there the contours of prehistoric fields are suggested still by lynchets—the banks formed by erosion on the lower edges of sloping plots. When the sun is low, the patterns of Celtic tillage are limned in sharp relief. Barrows dot the downland landscape, sheltering the remains of men whose fields these may have been.
Such folk might well have occupied manors like the one at Butser Hill. Erase the electric-power lines strung above the site, erase the A3 motorway, and what remains is a landscape as it could have looked around the year 300 B.C.—at the heart, chronologically, of Britain’s Iron Age. Amid five acres of paddocks, pens, and fields stands a sturdy roundhouse, more than forty feet in diameter and thirty feet high, its basketwork walls plastered with daub. The thatched, conical roof, where swallows nest, protects ovens and querns and crockery. Immediately outside are several haystacks and a byre. A low bank and a shallow dike enclose the central compound. Beyond lie fields of wheat, barley, beans, and flax. In outlying pastures livestock graze—unusual breeds of sheep and cattle, gamy and hirsute.
Elsewhere on the farm are hives, kilns, vineyards, and a shop for working metal. Great mounds of earth-covered timbers—charcoal clamps—await firing, to produce fuel for the forge. There are terraced beds thick with herbs: woad and meadowsweet, cowslip and eyebright, elder, valerian, dog’s mercury, field pennycress, and a hundred species more.
This tract is the demonstration area of Butser Ancient Farm. It is not a theme park or a tourist trap, and the men and women who labor here do not wear bearskins and pose for pictures. The demonstration area is, rather, the public side of a serious experiment that archaeologists have been conducting for more than a decade and will continue to conduct for at least a decade more. The aim of the experiment is simple: to take what scholars have inferred over the years about Iron Age agriculture—about its economics and techniques—and put those inferences to the test by actually creating and operating an Iron Age farm. An implement that museums label “primitive sickle”: is it capable, in fact, of mowing wheat? Does the “grain dryer” dry grain? Are “rubbish pits” really rubbish pits? More broadly, how did the Iron Age farmer prepare his soil, manage his livestock, build his house? How efficient was his technology? And just how productive might one of his farms have been?
These are the kinds of questions that Peter J. Reynolds, the director of the Butser Ancient Farm Research Project Trust, has been asking himself since the Butser experiment got under way, in 1972. Some of the answers that he and his colleagues have arrived at are challenging conventional views of Iron Age man and society.
I SPENT SEVERAL days with Peter Reynolds last fall, visiting both the demonstration area, which is open to the public, and the research farm itself, high above on Butser Hill, which is ordinarily off-limits to all but project staff. Reynolds is tall and trim, with a dark beard and moustache. He holds a degree in classics from Trinity College, Dublin, and has been pursuing experimental archaeology, as his craft is called, for some twenty years. Experimental archaeology is one manifestation of the so-called new archaeology that emerged during the 1960s—an archaeology focused more on understanding cultural processes and systems than on establishing cultural chronology. The work is not sedentary. When I first met Reynolds, in his office, he had just returned from the research farm “up top.”His jeans were soiled and his heavy boots were caked with mud.
“We have been trying to confine ourselves to the basics, making sure the bricks are sound before the edifice is built,” Reynolds said. “In terms of farming, we want to find out what actually happens if we do A or B or C. The very process of putting an Iron Age farm together raises vexing questions. We are dealing with many elements—physical structures, fields, livestock—which in turn comprise many subsystems. All of these various systems are interacting, one with another. To keep animals you need fences, and for fences you need hazel rods, and for hazel rods you must coppice woodland so that the stalks grow straight and tall from the stumps. Our aim is to create an integrated system, one that we can actually see in operation.”
“Our starting point,” Reynolds continued, “is almost always the material evidence uncovered by excavation—a pattern of post holes, a charred pit, some carbonized seeds. Some implements have been preserved in bogs, but generally the Iron Age doesn’t offer very much. We also make use, cautiously, of comparative ethnology, looking at how primitive societies today may have solved a problem that is troubling us. To some degree, what we are engaged in is not only working forward from the evidence but, by experiment, working backward to the evidence, trying to establish what the physical record, after the passage of many years, is supposed to look like. In other words, what kinds of human behavior in the past left behind the kinds of remains we have found?
“The whole point of experimental archaeology is to subject theory to scientific scrutiny and when an interpretation is found wanting, to develop alternative hypotheses and put them to the test. In no sense are we play-acting. The research farm and the demonstration area are vast, open-air laboratories. We have nine fields under the plow and we take 60,000 measurements of various kinds on each one of them every year. Our photographic collection now exceeds 12,000 slides. We analyze our quantitative data with computers. The power lines that run across the demonstration area arc in a way appropriate, because they remind us that far from pretending to be Iron Are people, which we could never do successfully, we are twentiethcentury scientists using every ounce of microchip technology.”
Reynolds’s office occupies a corner of what was formerly a Little Chef restaurant, part of a British fast-food chain. It is cluttered with computer printouts, offprints of journal articles, stalks of wheat, samples of Celtic beans. A photograph of a ditch in cross section shows the layers of silt that build up before the banks can be stabilized by vegetation. Shelves covering two walls hold books like Elsdon Best’s Stone Implements of the Maori, M. C. Ryder’s Sheep and Man, and Charles Darwin’s Vegetable Mould and Earthworms (Reynolds is interested in earthworms because, as he has written, “certain varieties . . . can effectively invert the soil cover,” thereby displacing potsherds and other artifacts). Near the door a makeshift calendar traces the schedule of farm activities during the coming year. The rest of the building, where lunch counters used to be, is part warehouse, part laboratory, part classroom, and part museum. Reynolds takes in as many as ten university students during the summer. Their fees, together with receipts from the demonstration area (admission is ninety pence for adults, fifty pence for children) and modest subventions from private foundations, support the Butser experiment and pay the salaries of three full-time professional staff members.
CREATING THE research farm, Reynolds explained, had been an enormous exercise in both addition and subtraction—addition of much that time had subtracted and subtraction of much it had added. To begin with, Reynolds needed a site that was free of contamination by artificial pesticides, herbicides, and fertilizers. One of the spurs radiating from Butser Hill, known as Little Butser, met that criterion, and Reynolds was able to persuade local farmers not to spray their crops when the wind was blowing toward it. Little Butser possessed an additional advantage: it is marginal land, 800 feet above sea level, with a soil cover less than half a foot deep on a base of hard chalk. These adverse conditions guaranteed that the farm’s performance would not be enhanced by accidents of environment.
Farm buildings had to be erected. The first of these, completed in 1973, was based on a pattern of post holes brought to light by Sir Mortimer Wheeler during his excavations at Maiden Castle, in Dorset, in the 1930s. Reconstruction proceeded by trial and error, with a few crucial hints, by way of descriptions, from Pytheas and Tacitus. The finished structure, known as the Maiden Castle house, represents only one possible interpretation of the archaeological record. And in any number of ways—some of them to be exposed, perhaps, by the passage of time—the reconstruction may be flawed. Experimental archaeologists deal not in certainties but in probabilities.
Before planting his fields Reynolds needed livestock, both for traction and for manure. Because many Iron Age species, such as Celtic shorthorn cattle, are extinct, and others, such as Old English goats, are nearly so, Reynolds had to search for close approximations—substituting Exmoor ponies, say, for the horses of Iron Age times, or mating wild boars and Tamworth pigs in an effort to “back-breed” something like the prehistoric porker. He also had to acquire the appropriate technology and to become proficient in its use. Often working from scanty evidence, perhaps a fragment of rock carving, Reynolds fashioned mattock hoes and digging sticks and wooden ards to stir the soil. He crafted neck yokes and horn yokes, and trained cattle to them both.
For sowing Reynolds had to procure not only the right kinds of food crops but also the right kinds of weeds. Paleobotanists working with carbonized seeds or with seed impressions fired into potters had already identified the major varieties of wheat grown in pre-Roman Britain, notably emmer and spelt, now cultivated only in remote regions of Asia Minor. Reynolds obtained samples of these rare cereals, which are cultivated at Butser under a variety of experimental regimes: manured and unfertilized, sown in spring and in autumn, raised on plots cleared variously by ax and fire. In certain fields weeds are permitted to grow naturally; in others special varieties, long extinct in the British countryside, are deliberately introduced. (One field at the farm is reserved exclusively for the propagation of unusual weeds.) The purpose of this diversity is to evaluate the degree to which the variables tested affect yields. Most of the experiments have been planned to last for at least twenty seasons, to ensure a valid annual average. Some of the trials have by now been under way for eleven or twelve years.
The preliminary results have surprised both archaeologists and historians. Though Butser’s crops have been planted on unimproved ground and tended with primitive tools, Reynolds has been getting enormous yields—larger, for example, than any yields indicated by written records from medieval times, and almost as large as those that British farmers were achieving on the eve of the First World War. The Iron Age Briton, it would appear, was no subsistence farmer.
Consider the spelt wheat sown every October in Field No. 2. Reynolds has never added manure to this field and never allowed it to lie fallow. He has not rotated the grain with Celtic beans, which fix nitrogen in the soil. Cultivation to control weeds has deliberately been kept to a minimum. Apart from planting seed in soil that ards have stirred and scored, virtually nothing has been done in Field No. 2 to enhance the viability of its crop. And yet the field has been producing, on average, more than three quarters of a ton of spelt wheat per acre every year. Wheat farmers in the United States last year did not produce much more: a little over a ton per acre. The seed-to-yield ratio on Butser’s Field No. 2—the volume of seed wheat planted to volume of wheat harvested—has averaged 1:28. That by far exceeds the 1:3 or 1:6 that medieval enumerators, according to the historian Marc Bloch, deemed a “reasonable” yield. It is comparable to the ratios achieved by modern American farmers. And it was produced on an inhospitable plot of land where modern wheat strains do not prosper.
REYNOLDS DROVE ME up to the research farm, of which Field No. 2 is a part, on a wild Friday morning, amid thirty-mile-an-hour winds and a fine, biting rain that etched color in the cheeks. In a few weeks, with the onset of winter, mud on the steep roads would make the site inaccessible to motor vehicles. For now, however, the route was clear. We climbed to the summit, our progress acknowledged with tweedy benevolence by the rare pedestrian, out for a bracing stroll. At last we emerged onto a broad, bald plateau. Butser Hill is the highest point in Hampshire, and in good weather it affords a splendid view of the South Downs. On this late autumn day the view consisted of low, gray clouds, relentlessly chased eastward by a constant gale.
It was to the South Downs that Sherlock Holmes repaired in his retirement, to keep bees, and Reynolds had a bit of Holmes about him as he headed toward the farm, interpreting the terrain along the way. “What we’re walking on now is a Neolithic track,” he explained, indicating a faint, grassy depression. “There’s another, over there. The rutted path this one merges into was a medieval road, once the main route from London to Portsmouth. They didn’t have much choice about running roads up hills back then. Nowadays, of course, they can slice right through, like the A3 does down there.”In isolated moments, when the weather caught its breath, we could hear the traffic speeding far below. “The cross-dike intersecting here predates the road,” Reynolds continued. “I doubt it was meant to serve as a fortification. It was probably a simple boundary marker to divide the spur from the rest of the hill. Our farm, as you can see, is on the spur, which we know was occupied during the Iron Age. Let’s go down.”
Little Butser extends outward from Butser Hill like an accusing finger. A path runs down its middle for perhaps a hundred yards, between snaking fences of latticed hazel rods. Pens and fields lie on either side. At the center of the site stand two thatched roundhouses, of different sizes and designs. A modern wiremesh fence surrounds the entire farm to keep out rabbits, which did not inhabit Celtic Britain. Rabbits, probably introduced in Norman times, are what Reynolds calls an “unwanted variable.”
The fence is not the only twentiethcentury intrusion. Little Butser has been painstakingly surveyed, and metal markers are embedded in the ground as reference points. There is also a compact weather station to monitor wind velocity and direction, rainfall, air temperature, and the temperature of the soil at six different depths. Someone on the small Butser staff must collect the data every day, even if that means, as it does for six months of the year, hiking half a mile up to the spur.
“Because we’re dealing with plants and animals, we have to keep detailed records on the weather,” Reynolds said. “And not just ‘weather’ but the microclimate of the spur itself. We probably just went through three different microclimates on our walk from the top. As you might imagine, conditions up here become marginal occasionally. I’ve been picked up by the winds and carried three or four yards. It was like hang-gliding without the glider. But I wanted the most hostile landscape you could grow something on.”
Reynolds unlocked the gate to the compound and we walked down the path between a herd of Dexter cattle and a flock of Soay sheep. Dexter and Highland cattle are Reynolds’s stand-ins for Bos taurus—the extinct Celtic shorthorn. The Soavs, however, are the genuine article. On Scotland’s windswept St. Kilda Islands, Soay sheep have thriven since prehistoric times, preserved by their isolation. They are a boisterous breed. It was shortly before the mating season when I visited the research farm, and rutting Soay rams were vying to establish dominance. One hapless pair locked horns and could not disengage. Reynolds entered the fold and eased apart the tangled spirals.
The first foray at Butser into Iron Age architecture, the Maiden Castle house, occupies the top of the spur. It has stood for almost a decade, through snow and rain and frost. Winds of hurricane force have done it little harm. It has never leaked. But the structure has begun to sag visibly, the result of an error in construction made when the rafters were fastened to the uprights. Years of stress have called attention to the mistake. The rakish slant of the building’s roof serves less as a rebuke than as a reminder that the key ingredient in any Butser experiment is time.
Parcels of woodland are put to the torch and then observed over time as plant life returns. Frenches are dug and earthworks erected; then they are left alone and allowed to erode. Reynolds is obsessed with what, in his own words, “shows up in the long term.” That is one reason why he experiments continually with pits, which are found by the hundreds, in a variety of shapes and sizes, whenever an Iron Age site is excavated.
Pits: if not for rubbish, as is often supposed, then what were they for? Reynolds has established, among other things, that a chalk-lined hole five feet deep and four feet wide, when sealed airtight with moist clay and topsoil, can safely preserve a ton of grain for years. Pits may also have been used for tanning leather and storing water, for making silage and salting meat. Perhaps some harbored potter’s clay. Others may have served as larders. Beside the Maiden Castle house are sunk half a dozen pits, each the subject of a different experiment in pit technology. A few of these sprout tubes and gauges, for tracking temperatures and gas concentrations in the chambers below.
BETWEEN LITTLE Butser’s pits and Little Butser’s prow lie Little Butser’s prodigious fields. Each is the size of an average Celtic plot—a quarter of an acre or so. Along the edges of some of the fields, lynchets have begun to form. Winter sowing was set to commence within a few weeks, and several farmhands were at work, courtesy of Britain’s Youth draining Scheme and Community Manpower Program.
“The yields we’ve been getting lend credence to the claims of classical writers like Strabo and Caesar, otherwise hard to accept, that Britain was a significant exporter of grain,” Reynolds said. “In fact, Britain may well have been the breadbasket of Northern Europe. If this was the case, then the suggestion that many of our ‘rubbish pits’ were really granaries becomes rather persuasive. The size of our yields has other implications. For example, the population density of Iron Age Britain was probably far higher than we have supposed, particularly when you consider the nutritional quality of emmer and spelt. They’re extremely hard to thresh, but they have about double the protein of modern bread wheats.”
As population increased, so did competition for arable land. Reynolds believes that considerably more of the British countryside was cultivated during the Iron Age than is cultivated now. Because Celtic field systems are visible almost exclusively on hillsides, it used to be thought that Iron Age Britons lacked the technology to cope with the heavier clay of the valleys. But Reynolds and others have shown that the early farmers exploited the valleys fully. The movement of fields ever farther up the hillsides represents a remarkable agricultural expansion. Only after the Romans came were marginal lands like Butser Hill turned over to grazing. It is because they have been grazed for millennia rather than plowed for millennia that Britain’s hills today display the imprint of the Celtic farmer. In the valleys few’ traces of his fields survive.
“An agricultural surplus indicates something else about Iron Age Britain,” Reynolds said. “Celtic society w-as initially one of farmers, wdth a warrior elite and a small class of priests and artisans. Surpluses of food would have freed large numbers of people from dependence on the land. We see glimmers of this whenever we excavate a so-called hill-fort: there are signs of urbanization, of economic specialization. After a point in the Iron Age you probably didn’t thatch your own house; you hired a thatcher. You didn’t make your own pottery; you went to a potter. I think that in 300 B.C. we are dealing with a much more complex, sophisticated, organized, and stable society than we tend to give it credit for.”
In a famous article published nearly half a century ago the historian Lynn White, Jr., lamented the reluctance of his colleagues to analyze what was mundane about the past. He wrote, “It is startling to reflect that e have, as a rule, only the vaguest notion of how the men of the Middle Ages actually did things, and how’, from time to time, they learned to do them better.” What held true for medieval times holds doubly true today for the Iron Age. Much as the advent of social history has helped redress the balance in historical studies, so experimental archaeology has begun to make a powerful contribution to our knowledge of the prehistoric.
A great deal, of course, remains to be done—and, at Butser, redone. Reynolds’s reconstructed “sickle” proved a useless tool at harvest and was readily discarded by the reapers. (It was excellent, however, for splitting hazel rods.) His reconstructed “grain dryer” was not efficient at drying grain but turned out to be ideal for malting barley. (The malt produced was roughly comparable to that used by Guinness.) The archaeologists at Butser Hill are still trying to figure out how the woodlands would have been managed. They also haven’t yet calculated the total manpower needs of their Iron Age farm. Important investigations continue into “data recovery”: establishing not only what archaeological evidence is supposed to look like but also how it can be retrieved. After the passage of thousands of years, what minute chemical traces may remain of a certain human activity? How does one reconstruct plant populations from the pattern of pollen distribution in the soil? How deep underground might earthworm activity have deposited debris from the surface?
Reynolds emphasized that his work at Butser, like the experimental work of other archaeologists on other questions, is not an entirely novel approach to the study of the past. Rather, in his view experimental archaeology is bringing to the larger discipline some very old perspectives. “I don’t think that what were doing here really represents any new branch of archaeology, or even a twig,” he said. “Rather, it’s part of the main trunk. At least it should be. The only way to refine archaeology as a discipline is to conduct experimental trials. That is not a new idea. The original British antiquarians did many of the kinds of things we’re doing at Butser. The vicar who discovered a hand-ax—the first thing he would do was conduct a little experiment, to see if he could chop down a tree with it. Archaeology fora time tended to turn away from all that. We’re just turning back.”
—Cullen Murphy