As we excavated further, we realized that this was the center of a house measuring about five meters by five meters. The ash layer and the plaster floor were surrounded by a line of holes that had held small upright stakes. These stakeholes showed us where the wattle-and-daub walls of the house had once stood. There were even small pieces of daub surviving on the floor. Between the wall and the edge of the plaster floor, the ashy soil had filled in shallow grooves that had once held horizontally laid logs or planks. These beam slots were all that was left of the furniture—they were the foundations of box beds and storage units, like those crafted in stone in the Orcadian houses. Our 2004 trench had been too small for us to see that these ephemeral traces were the remains of a house. By extending and enlarging the trench we could finally appreciate what we were seeing.
Realizing that a single episode of plowing into the Neolithic ground surface would have virtually destroyed these house floors, we now understood just why so few Neolithic houses have ever been found in England.
Over the next few weeks, we found traces of another four houses within the trench; we knew we couldn’t rush this and would need more time in future summers to excavate each of them to the highest standard possible. Having spent the best part of two decades excavating prehistoric house floors, Colin and I had developed new methods for studying them. As a student I’d listened to an experimental archaeologist, Peter Reynolds, tell us about his reconstruction of an Iron Age roundhouse.1 He’d suggested that the evidence in such houses is so minuscule that archaeologists should dig with teaspoons, not trowels and mattocks, in order to understand how they were lived in. I had roared with laughter at the time: This seemed a preposterously obsessive and time-consuming thing to do.
Years later, though, as I dug my first house floor in the Outer Hebrides, I realized he was very nearly right. I was working with some very talented environmental archaeologists, Helen Smith and Jacqui Mulville, and we worked out that the micro-debris and chemical residues accumulated on the floor during the house’s occupation could tell us a lot about daily domestic tasks and where they were performed—but it was going to be a major job to retrieve the evidence.
Archaeologists can pick larger finds out of the ground they’re working on—those things easily visible in the soil—but it’s usual practice to use a 10-millimeter sieve on site (about the mesh size of a normal garden sieve) to ensure nothing gets missed. Smaller sieves are pretty useless because the soil clogs the mesh so quickly. To retrieve anything smaller than 10 millimeters in diameter, we need to wet-sieve the soil: to wash it through sieve mesh of various sizes in a system of water tanks. It’s a long and dirty job, as the soil from each context has to be bagged up in sacks, labeled in minute detail, and usually taken away from the excavation site to a wet-sieving team working in an area with an ample water supply. It slows down the excavation process enormously, but the results are worth the effort.
The remains of four Late Neolithic houses are visible in this main trench at Durrington Walls. I am standing outside the doorway of one of the houses in an area of midden (heaps of domestic rubbish).
In South Uist we’d carefully wet-sieved the entire occupation layer on top of the floor of a Bronze Age house, which was excavated in half-meter squares. Using a mesh size of just two millimeters, we’d retrieved minute fragments of animal bone, potshards, burned plant remains, and broken artifacts, and were able to identify the areas of the house where cooking and various types of craft-working were carried out. We also took hundreds of small soil samples from the floor, to plot concentrations of chemical elements, such as phosphorus and nitrogen, and many further samples to send to a soil micromorphologist. Micromorphology is a technique of examining sections through the soil under a microscope to establish how the floor layers were formed, of what they were composed, and whether later floors were laid on top of earlier floors.
The Durrington Walls houses needed us to apply these well-honed techniques once again. We knew we had to sample the whole floor of each house in minute detail. We dreaded doing it—we knew how long it would take and how many sacks of soil we’d have to heave back and forth—but we knew the results would make it worthwhile.
Unlike the Hebridean house floors of soft peat and sand, the Durrington floor surfaces were comprised of fairly hard chalk plaster. Every time the Neolithic inhabitants had given their house a good sweeping, they’d sent much of the micro-debris of their lives straight out of the door, so it was more difficult for us to reconstruct activity patterns than it had been on the Scottish sites. Nonetheless, the thin ash layer across the floor gave us a moment frozen in time: the moment of the house’s abandonment. And, as we were to discover later on, the floors held other clues to unraveling the secrets of Neolithic daily life.
The houses were built on a slope but had level floors. This had been achieved by terracing the hillside, stripping the turf from higher up and laying it in a band lower down. Houses higher up were cut into the bare soil while those lower down the slope sat on a leveled platform of turf. This showed an element of organization and planning that went beyond that of the household. Perhaps this new settlement was larger and more organized than we’d at first expected.
This was not the first time that people had lived here. Buried in the turf were flints, including a leaf-shaped arrowhead, indicating that early farmers had lived here at some point during the fourth millennium BC, at least 500–1000 years before the houses were built.
At the bottom of the slope, within the valley that leads from the interior of the henge toward the river, we finally uncovered what we’d set out to find. By extending our excavation trench twenty meters upslope from the eroded area where we’d dug in the valley bottom in 2004, we discovered that a meter-deep layer of colluvium (soil, loosened by plowing, that had washed down from higher up the valley) had settled on top of—and hence protected—the prehistoric ground surface. This surface was a thin layer of remnant turf and topsoil that covered a flat road surface of packed and broken flints. As the roadway had gone out of use, so grass and weeds had sprung up, eventually creating a thin layer of humus worked into soil over decades by earthworms. Carefully plotting the finds within the prehistoric turf layer covering the road, we discovered that this soil had accumulated over several centuries. Mixed in with the oblique arrowheads and Grooved Ware of the Neolithic were the later styles of barbed-and-tanged arrowheads, and pieces of distinctive Beaker pottery from the Early Bronze Age, indicating that there had been activity here long after the Neolithic houses had been abandoned.
Our theory was right—there was an avenue running from Durrington Walls to the river, an avenue so wide that our trench was not big enough for us to see the full width of this Neolithic roadway. In 2005 we found its northern edge, defined by a low chalk bank about five meters wide, and in 2006 we found the parallel bank running along the south side. The flint road surface was 15 meters wide and, in its entirety, the avenue was 30 meters (100 feet) across from the outer edges of its parallel banks. It dwarfs the modern road built in 1968 through the middle of Durrington Walls: That has a roadway only 10 meters wide.
The Neolithic road surface was constructed from hard-packed, natural, broken flint, but it also contained lots of animal bones, pieces of burned and worked flint, and even potshards that had been incorporated when the road’s matrix was laid down. These bits of rubbish mixed into the road construction meant that people were already living here before the surface was laid.
When we stripped off part of the upper surface of the road, we found a lower layer of flints that contained no artifacts at all. Mike Allen and fellow soil specialist Charly French reckoned that this basal deposit was formed by natural agency, a geological deposit of coombe (valley) rock. Before the Neolithic, the bare valley floor had become covered with broken flint eroding out of the valley sides. This natural feature had been exploited and remade by the Neolithic inhabitants of Durrington Walls.
A plan of the main excavation at Durrington Walls showing
the plaster floors of the Neolithic houses (shaded) and their central hearths (black). The other features are pits and stakeholes.
Computer-generated plots showing the relative density in the northern part of the main trench of animal bones (left), worked flints (center) and burned flints (right). The outline plans of the houses are visible, as is the curving line of postholes forming a fence that separated two of the houses.
The avenue had yet more information to give us. Clive Ruggles came to take a good look at the Durrington Avenue while it was being excavated, to record its exact orientation. He found that the avenue’s orientation when looking westward, upslope from the river, was within a degree of the midsummer solstice sunset during the Neolithic. Clive already knew that the Southern Circle inside the henge, partially excavated in 1967, had an entrance facing southeast, to the point at which the sun rose on the midwinter solstice. Since the two directions of midsummer sunset and midwinter sunrise are perpendicular to each other, there is usually no way of being sure whether both directions are significant or only one of them. Our avenue was pretty much aligned with the entrance to the Southern Circle, but was a few degrees off the precise axis of the midwinter sunrise. Which mattered most to the avenue’s builders—was the avenue meant to share the same alignment as the timber circle (toward midwinter sunrise), or was it aligned in the opposite direction, toward midsummer solstice sunset?
Clive worked out that the midsummer solstice sunset was the important direction—the avenue shows a deliberate alignment with the setting sun, not with the axis of the timber circle. He is convinced that the builders made sure that the avenue was a few degrees off the axis of the Southern Circle so that it would align with the midsummer sunset. As a consequence, it has an imperfect alignment with the midwinter sunrise. Had the avenue been constructed simply to lead from the midwinter sunrise to the Southern Circle, it would have missed the alignment with the midsummer sunset. The midsummer solstice alignment is affected by the topography: The Durrington valley rises quite steeply from southeast to northwest. The steepness of the valley means that the sun disappears below the horizon sooner than it would on flatter ground, so the avenue is aligned on the spot where the midsummer sun actually sets.
Full plan of the Southern Circle, combining the 1967 and 2005–2006 excavation and geophysics results. The cone shapes are the postholes and their ramps. Julian Thomas’s excavation trench is marked to the west. The rest of the circle is now buried beneath the modern road.
A plan of the Late Neolithic timber circle of Woodhenge. Today the postholes are marked with small concrete pillars and the bank and ditch are barely visible. The contents of the grave were destroyed during the Blitz but the burial is thought to date to the Early Bronze Age.
Ever since Geoff Wainwright’s discovery of the rings of postholes that held the timbers of the Southern and Northern Circles within Durrington Walls, archaeologists have tried to reconstruct what these timber circles looked like. The Northern Circle consisted of one or perhaps two rings of posts enclosing a rectangular setting of four large posts.2 This circle was approached from the south (from the direction of the Southern Circle) via a post-lined passage that passed through a façade of posts forming a screen on the south side of the rings of timbers. Geoff tentatively ascribed the outer and less convincing of the two post rings (almost 25 meters across) to a first phase of the structure. Since more than half a meter of chalk had gone from the ground surface here by the twentieth century, when he excavated the site, only the deeper features survived. A four-post square setting in the center of the rings appears to have been oriented eastward, toward the midwinter sunrise.
A reconstruction of the timber posts at Woodhenge.
The Southern Circle was much more impressive and better preserved; artifacts lying on the Neolithic ground surface around it were still in position.3 At its center was a rectangular arrangement of posts enclosing a smaller setting of six posts. Geoff interpreted these posts as belonging to the first phase of construction. They were surrounded by two concentric rings of timbers, and approached from the southeast through a screen of posts.
The timbers of this first phase of the Southern Circle had been left to decay in their holes. When the circle was excavated, the outlines of the rotted timbers survived as “post-pipes,” voids left by the rotted-out wood into which soil had slowly trickled. These posts were around 20 centimeters in diameter, slightly thinner than a telegraph pole. Archaeologists can tell from the soil layers if a pit or posthole has been dug into at some point after its first construction. At the center of the Southern Circle, one of the postholes in the rectangular setting showed that the post had been replaced twice; two others had been replaced just once. On the basis of the likely lifespan of these timber uprights, Geoff estimated that the first phase of use of the Southern Circle lasted a minimum of sixty years.
This first phase was replaced by six concentric circles of posts whose uprights ranged in size from around 20 centimeters to over a meter in diameter; the outermost ring of posts measured almost 39 meters across. Among the largest posts were the two marking the southeast entrance. These posts too had decayed in their holes. Geoff reckoned that these huge posts probably survived for the best part of two hundred years.
All of the Southern Circle’s Phase 2 posts were so large and tall that the builders had to cut ramps into the chalk to feed the end of each post into its hole before heaving it to a vertical position. In 2006, Time Team built a replica of the Southern Circle, borrowing twenty soldiers from the Larkhill army barracks to try to raise just one of the smaller-sized posts by muscle power. To everyone’s surprise it was too difficult—they had to resort to twenty-first-century mechanical means to erect it.
In and around the Southern Circle, Geoff found some traces on the surviving Neolithic ground surface of what took place here. A small fireplace was positioned in the center of the circle, and an enormous fire-pit (five meters long) outside the southeast entrance was set into a 15-meter-wide surface of broken flint, adjacent to a platform of chalk blocks. Part of the interior of the circle was surfaced with rammed chalk but otherwise it had no proper floor. On the northeast perimeter of the circle was a post-lined hollow whose shallow filling of pottery, bones, and other rubbish led Geoff to interpret it as a midden.4
Model of Phase 1 of the Southern Circle as a square-shaped arrangement of posts surrounded by two concentric timber circles, with the D-shaped house to the northeast.
Forty years before Geoff Wainwright went to Durrington Walls, Maud Cunnington had excavated a similar timber circle at Woodhenge, situated on the high ground just south of Durrington Walls, overlooking both the River Avon and the small valley in which the large henge lies.5 Sitting on top of a ridge and plowed in recent times, Woodhenge’s original ground surface had disappeared long before Cunnington’s excavations. This circular timber monument had also consisted of six concentric arrangements of posts, except these were laid out as ovals rather than true circles. The posts, of similar diameters to those of the Southern Circle, had been left in position to decay. The whole structure was enclosed within a ditch and external bank. Cunnington noticed that the long axis of the ovals was oriented on both the midsummer sunrise and the midwinter sunset, as is also seen at Stonehenge itself. At the center of the timber circles was a cairn of flint nodules, beneath which was the skeleton of a child.
Model of Phase 1/2 of the Southern Circle. In this phase the builders added a timber portal facing the midwinter sunrise.
Cunnington thought that the child had been sacrificed. She noted that its skull was split in two, perhaps by a stone ax. Curiously, there is no mention of this injury in the pathologist’s report on the skeleton—and we can’t go back and re-examine the original bones because they were destroyed during the bombing of London in the Second World War. Josh Pollard has had another look at Cunnington’s records and he thinks that she may have been mistaken:6 The skull could have collapsed naturally along the bone’s sutures under the pressure of
earth from above, rather than being damaged by a deliberate blow. He also thinks that this is a later burial—that the child was not buried here by the Neolithic builders but was added in the Early Bronze Age, when the wooden monument would have been decayed ruins.
These three timber circles pose difficult questions. Were they built at the same time as Stonehenge or were they erected earlier, some sort of wooden prototype for the ultimate version in stone? Were they roofed? If so, did people live in them? Colin and Julian were also keen to find out more about what was in the postholes—were the animal bones and pottery plain old rubbish, or “structured deposits” of ritual offerings? The question of how these artifacts had fallen into the postholes was also tricky. Geoff had reckoned that the decay of the timbers had created cone-shaped voids (weathering cones) into which material stacked against the posts or lying close to them had fallen. Colin and I were not so sure.
During one of our evenings in the Plume of Feathers in 2003, we’d spent hours looking at the drawings of every posthole from the 1967 excavation. It was a Saturday night and the bar was getting crowded and noisy, so no one seemed to notice us or our Eureka moment. Colin realized that the so-called “weathering cones” were actually pits dug into the tops of the postholes after the wood of the posts had completely decayed. But why should anyone have done this? Why was it so important to dig new pits into old postholes, and there deposit groups of artifacts? Was it something to do with marking change and decay? We really needed to look at some of the postholes that hadn’t been dug out during the 1967 excavation. Geoff’s three-month rescue dig in 1967 had had to stick to the corridor of the proposed new road. Consequently, parts of both timber circles remained unexcavated because their edges lay outside the construction corridor.
Stonehenge—A New Understanding: Solving the Mysteries of the Greatest Stone Age Monument Page 8