Our knowledge of domestic living arrangements in the prehistoric period has come a long way in the last 100 years or so. It seems quite incredible to think that, prior to Gerhard Bersu’s excavations of the Iron Age roundhouse at Little Woodbury in the 1930s, the widely held view was that Britain’s pre-Roman inhabitants lived in pits. This myth was finally dispelled by Peter Reynolds (1939-2001), the stalwart pioneer of early experimental archaeology and first director of Butser Ancient Farm, who briefly occupied a pit if only to prove that damp, smoke, and lack of light rendered the space completely uninhabitable. The pits, it transpires, were used for food storage and rubbish disposal. But this was not the end of his quest to establish what the lives and homes of our ancestors might have been like. With the term ‘experimental archaeology’ having been coined in the 1960s to describe efforts to understand ancient artefacts by actually working with them, Reynolds’ investigations continued at Butser Ancient Farm – and 50 years later, staff at the site are still tackling that question by constructing experimental buildings based faithfully on archaeological evidence.
During these projects, modern materials are eschewed in favour of well-recognised, more traditional ones such as earth-fast wooden posts, wattle and daub, clunch, water reed, and straw. We have to run a continual and constant programme of maintenance, doing battle with the elements to keep our buildings safe, sound, and watertight, and occasionally we suffer casualties as a consequence of extreme weather, as was the case in late 2019 when we found ourselves peering disconsolately through the drizzle at a huge crack in the daub wall of one of our Neolithic structures. A thorough post-mortem revealed that gale-force winds had forced what was a very lightweight, somewhat aged building to twist on its axis, causing the daub to be ‘popped’ off the wattle framework underneath. This bout of inclement weather had sounded the death knell for our Neolithic friend. But mourning for our condemned structure was quickly supplanted by the thrilling realisation that ‘what comes down at Butser must go back up’.
A brief was drawn up to ensure a new structure satisfied all the prerequisites of our site. It was a bit of a challenge to find the ‘Goldilocks house’: one that was just the right size to fit in the available space, could accommodate a large school group or a public event, was based on evidence local to Hampshire, was interesting archaeologically, and which posed questions we would aim to answer through an experimental construction process. I am sure Gareth Chaffey of Wessex Archaeology was as surprised as we were when it transpired that a house his team had excavated in 2012 (see CA 292) fitted this rather demanding brief. Excavations at Kingsmead Quarry, Horton, in Berkshire, had proved exceptional, revealing some of the most significant evidence for domesticity yet found from the early Neolithic period (4000-3300 BC) in England. Who knew the prehistoric inhabitants of Berkshire could be so obliging?
At Horton, four structures identified as houses had been uncovered, constructed using two different types of design. Two were post-built, relatively small, and offered limited artefactual and dating evidence, while the remaining two, including that Goldilocks building we now know as the Horton House, had clear foundation trenches, containing evidence indicative of the use of upright wooden posts and planks. Curiously, this 15m × 7.5m building is trapezoidal rather than rectangular, meaning that both ends bowed inwards, most noticeably at the eastern end. Four substantial posts were located in each corner, as well as two placed internally to create a partition that divided the space into two separate rooms. There was no evidence for floor surfaces or a hearth, but a possible entranceway for the building was identified.
The site also offered artefactual evidence, such as early Neolithic pottery, fragments of animal bone, flint-working waste, burnt flint, and a number of worked-flint tools, suggesting that the buildings were used, at least for some of their lifespan, as habitable dwellings. The discovery of charcoal and charred plant remains, including cereal grain and hazelnut shell also hint at domestic activity in and around the site. Interestingly, radiocarbon dating of the features suggests that all the buildings are largely contemporary, possibly spanning around two or three generations sometime between 3800 and 3600 BC. So, how did we set about devising our replica?
The design process
While there is no manual for building a Neolithic house, entering such uncharted territory is all in a day’s work for the Butser Ancient Farm team. Critically, the building was created using only materials and resources available to its prehistoric creators. Structural elements in the Horton House have been fastened using simple lapped notches, drilled and secured with oak pegs, and lashed with cordage. The carpentry was conducted with tools that would have been available during the Neolithic period, and the structure was erected without the aid of any modern machinery. The resulting building is a triumph of both the possibilities of ancient skills and technologies, but also the team’s ingenuity in creating a viable design. The team, which was led by Trevor Creighton and me, is however quick to acknowledge that Neolithic people were likely to have had a far more refined skill-set than we did!
Unlike the well-documented Neolithic housing tradition of the Continent and Ireland, with its distinct and formulaic multiple ‘aisles’ of internal supporting posts, the footprint of our Horton House was stark and minimalist by comparison. The foundation trench, punctuated by probable opposed posts at the corners and mid-point, invited interpretation as a post-in-trench, interrupted sill-beam structure, with unknown intermediate walling; internally, only two post-holes were evident. We still had lots of evidence to work with, though, and a design process of considering not only engineering possibilities but also the properties and limitations of natural materials led to some immediate conclusions. Whatever thatching material we used for the roofing, it would have to sit at an angle of at least 45º in order to shed water effectively. With this knowledge, we calculated a ten-ton load over the 7.5m span of the building. This then led to thoughts of the walls that would have to bear this load. Illustrations created by Wessex Archaeology show several iterations of the design process, including the original concept drawings, as well as computer-generated modelling of later versions with lower walls and a more accurate steeper roof, and ultimately our chosen interpretation where the wall is removed altogether to create a striking load-bearing A-frame design, with earth-fast posts at each corner and at the mid-point of either wall.
Modern health and safety and engineering precautions have necessarily had some influence on the design. Two intermediate roof trusses – not specifically indicated archaeologically, but also not counter-indicated – were included to strengthen the structure. This also influenced the decision away from load-bearing side walls, which would have required a substantial and continuous wall plate, as well as warranting an incredibly tall structure in order to maintain the required pitch of thatched roof. By utilising earth-fast principal rafters, the ground acts as tie beam and wall plate, creating a much more secure structure.
Most profoundly, through following the ground-plan faithfully, and maintaining constant pitch, the roof exhibits a noticeable ‘hog-back’ profile. While this is implicit in the ground-plan, the dramatic effect in the built form emphasises this feature – a known structural strengthening device – perhaps once common among the many prehistoric structures of trapezoidal plan.
The walls that have been created are not structural, but represent three possible interpretations of the original, with planking at the east end, wattling on a sill plate at the west, and low, lateral weather walls on either side under the thatch. These are important variations archaeologically – their residue can be analysed post-demolition to check accordance with the original archaeology, giving the structure an archaeologically significant legacy long beyond the lifespan of the structure itself.
Creating a Neolithic tool kit
With the design agreed by both Butser and Wessex Archaeology, the team set about creating their own Neolithic tool kit, ranging from stone axes, antler picks, and digging sticks, to bone chisels and mallets fashioned from the trunk and branch of a holly tree. The Scots pine selected for the main construction elements of the build was green and therefore in optimum condition for working with Neolithic tools.
Early tools – and, indeed, any tool where the hafting is held in place with little more than beeswax, pitch, and rawhide lashings – should be treated with reverence, as the wooden handle is easily damaged or split. After a few initial trials, a technique of ‘pecking’ at the wood rather than swinging an axe overhead was quickly adopted. From our own experiences both on this project and on others using Bronze Age tools, breakages must have been a major and infuriating occupational hazard. Other tools included the casually acquired ready-formed bone chisel – a chance find from one of our school archaeology digging pits! We tested a number of techniques with this, but one surprising success was the use of the bone as a primitive coring tool, employed like an awl to core through timber for pegging the joints with oak. The preferred and most-successful approach appeared to be simple rotation of the bone in the hole, using the serrated end. The hollow inner of the bone acted as the perfect conduit for storage and removal of the cored woody materials, and it took around 1.5 hours in total to create a hole approximately 2.5cm in diameter penetrating through a 13cm diameter green timber. Not long when you’ve got all the time in the (Neolithic) world.
Andy Sole, of the original excavation team, and the illustrator of those initial sketches, had envisaged a composite roof: essentially a covering made of whatever was available to the builders at the time, which may well have been the case. Based on the location of the original site, this could have included the obvious choices of wheat straw and/or water reed, but also a number of other possibilities such as sedge, bracken, and turf. However, the cost of materials and labour on a building of this size necessitated something reliable, so we selected water reed. It is interesting to note that our team had significant prior experience in thatching experimental buildings (and a lot more by the end of the build!) but still don’t expect their work to last much more than a few years, whereas a modern thatched roof, in the hands of a professional thatcher, could last as long as 50 years.
Anyone who has attempted to construct an experimental building will acknowledge that there will be gaps in the evidence, particularly for those features that are entirely above ground. A large gap above the end gable-wall is a case in point. A window fashioned from animal skin seemed a highly plausible solution.
In fact, the Shetland Islanders had, until the 20th century, used a similar technique of stretching thin rawhide across a frame and coating it in fish oil. The result of our efforts is a translucent, if slightly ‘scented’ window.
Design choices come back to haunt us
Each end of the house has been completed in a different material, to demonstrate the possibilities based on the evidence. One gable-end has been covered with overlapping split and dressed oak planking, and the gaps stuffed with moss and turf, while the other was constructed from wattle and daub. There is plenty of evidence for this method of construction throughout prehistory, and it is a tried- and-tested formula. Or at least that’s what we thought, until we decided to investigate a little variation on the theme.
We treated two large structural uprights within this wall as vertical staves around which to weave the wattle, creating a much thicker wall than is usual practice – thin, flat panels are more normal. We hoped to trial using the deep spaces created by this method as pockets for organic insulating material such as old straw, before plastering over it with an easy, ‘glue-all’ daub coating. However, we had not reckoned on a global pandemic and a lockdown that prevented us from tackling the wall by the prescribed method. You don’t just throw daub at it: you also need to throw a lot of people at it. In fact, excavations carried out at a Bronze Age site in Hungary suggest it was all hands on deck to complete the task within a few short days, with evidence of small, child-sized handfuls of daub present in the archaeology. The extra-thick wall that we had created required more daub than we had ever needed before, which proved an overwhelming task for the skeleton staff still left on site. In the end, the wall was finally finished just two weeks before our public reopening in April (when CA visited the newly completed house).
Above all, this project has highlighted the possibilities of experimental archaeology. A good description of what these experiments can achieve is offered by Dennis Harding, one-time Abercromby Professor of Archaeology at the University of Edinburgh: ‘The basic objectives of scientific experimental archaeology are, first, to test archaeological interpretations by empirical experiment and replication and, second, to conduct such tests in a way that permits feedback into excavation practice, data retrieval, and analysis.’ In other words, erecting the building and the exploration of materials, consumption of resources, and techniques is only the start of the experimental process here at Butser. In order for the project to provide meaningful data, archaeologists at the site will monitor the durability of the chosen materials in the longer term to appreciate the potential lifespan of such constructions, with the aim of feeding back to the wider archaeological community to inform interpretations and identification on the ground.
Speaking of interpretation, one feature of the Horton House that has puzzled both the excavation and the construction team is the existence of what appear to be potentially two closely linked entrances on the long side of the building. Logically, the easiest place for such an entrance would be on the end gable-wall. So, what possible explanation could there be? By actually being able to stand inside the house, it becomes apparent that the right-hand door would offer the opportunity to stall animals in the gable-end of the house, leaving the left-hand door clear for the human occupants. There is no concrete evidence for this kind of dual use, but this idea of being able to engage all our senses within the space as part of the interpretive process is a potent reminder of the value of such experimental constructions.
Experimental archaeology is a wonderful tool for showing us what could have happened, rather than providing definitive answers, and on this basis there are certainly other feasible interpretations of the Horton House which we could have selected to build. However, if we think of archaeology as the study of people from the past, then constructing the spaces they lived in allows us to connect with them in new and interesting ways. We have come a long way from beliefs about living in pits. We hope that further collaborative projects combining archaeological data and experimental constructions will continue to expand the boundaries of our knowledge, revealing further fascinating insights into the domestic world of our Neolithic ancestors.
The team at Butser Ancient Farm extends their heartfelt gratitude for the significant contribution made by Wessex Archaeology towards the success of this project, as well as to the many others who offered their help and expertise in the process of the build – we couldn’t have done it without you.