26. Impossible Drains

All journeys start somewhere, even those in the mind, and the mental expedition exploring the uncharted depths of my ignorance that became Theoretical Structural Archaeology started at a site called Orsett ‘Cock’, to the north of the Thames in Essex.

The Orsett ‘Cock’ enclosure, Essex, during the 1976 excavation,
showing the position of buildings S1 and S9, discussed below.

The Orsett ‘Cock’ enclosure had, appropriately for an archaeological site, taken its name from a local hostelry, the ‘Cock’ Inn. The complete liability this would become in a Google search could not have been foreseen when the site was excavated in 1976 (hereafter to be known as the Orsett Enclosure, so as not to attract unwarranted attention). After a brief interim report,[1] the site vanished into the great black hole of unpublished sites known as ‘The Backlog’.

Excavating sites can be interesting, exciting, and urgent. It’s a social activity conducted in the great outdoors, usually near a pub. Writing reports is in many ways the complete opposite: it can be a solitary, boring, and a frustrating task, especially with no decent pub -- as I was to discover when I was given the job of writing the Orsett ’Cock’ excavation report in 1989.

However, Orsett proved to be a very interesting site. By the time of the Roman invasion (Late Iron Age) it was a ‘farmstead’ with a large roundhouse, S9, set in an enclosure of about a third of a hectare, with several other buildings. To the north were an adjoining earlier enclosure and the remains of a Middle Iron Age structure, S1.


To judge by the size of the S9 roundhouse, the site was ‘high status’. As if to prove the point, and almost certainly in response in the Roman invasion, the site was fortified with extra ditches, palisades and a defensive gate. This involved quite a lot of manpower, timber, and general upheaval. It is only with the benefit of hindsight that the futility of trying to defend such an installation against a Roman army that was merrily knocking off hill forts like Maiden Castle becomes apparent -- but an Englishman’s home is his multivallate enclosure.


A tantalising glimpse of what befell the Orsett enclosure was the find of a group of six spear heads at the bottom of one of the ditches, which had been rapidly back-filled; this may suggest that they surrendered, but that's a story for another time.

Nobody, other than a fellow field archaeologist, knows the sense of fear and trepidation that comes over you when you are handed several boxes of 15 year old plans and context cards, complete with perpetrators’ muddy finger prints, and are given the job of explaining what it all meant for posterity.


The process has its own rewards; you make your own discoveries as you excavate the archive, effectively re-excavating the site, and freed from any prior emotional or intellectual involvement, you can take a very dispassionate look at the evidence. Exciting – you bet it is! But it's more like reading a detective story than skiing down the Matterhorn; Orsett made me realise there is a whole library, with hundreds of sites, thousands of features, all yet to be explained. The ‘following up of clues’ is only aspect of the previous analogy that is valid, because ultimately you are the author and you have to solve the crime yourself.


So how do you go about writing an archaeological report? It is quite possible to write a perfectly acceptable archaeological report with a minimum of interpretation, simply presenting the evidence and referencing similar archaeology on appropriate sites. In this way we borrow other people’s interpretations, and subtly avoid any individual responsibility, the principle downside being that a lot of evidence may go interpreted.


However, this is for posterity, so it pays to be conservative, and as we have discussed, there are standard types of structures, such as roundhouses and four-posters made up from standard features that can be universally recognised, which require little further explanation. There is a danger that, like bent coppers checking each others version of events, archaeologists have colluded to create a version of the evidence, complete with artistic reconstructions of the usual suspects.


I had inherited the excavator Hugh Toller's account, written shortly before he left archaeology, in which he duly identified the main structures and major features and put them in ‘phases’. His account still left the majority of the features unexplained, and besides, giving a group of them a label, such as ‘roundhouse’, is a long way short of understanding them.

Plan of the roundhouse S9 at Orsett, defined by 16.5m diameter gully F19

I want to introduce you to an old friend, Orsett S9, a big beast of a roundhouse built late in the Iron Age and standing in AD 43. It is defined by a deep gully, F19, penannular in shape (circle with a bit missing), which enclosed an area of 200m², containing numerous postholes. It was trying get to grips with this structure 20 years ago that made realise I knew nothing about buildings. While I could find at least two ring beams, I could not account for the majority of the postholes. But even more worrying, conventional wisdom at the time suggested that the trench F19 was a drainage feature. So where was the wall?

Sections of the Gully F19 - a drainage feature?

In the Orsett report, my approach to solving these problems and defining structures was to produce scale plan and section drawings of the major structural components as evidenced by the archaeological features, on the basis of clearly defined assumptions. I referred to these as ‘theoretical models’.[2] The models were designed to demonstrate that the structure was viable and how it worked, which in the case of a building, meant showing how the roof was supported. They demonstrate that there is sufficient geometry, proportionality, and relationship between a group of features for them to be considered a structure; that is, they are a best fit, and are usually accompanied by a percentage confidence rating; they are not to be confused with reconstructions. Following on from the previous article, we shall be considering walls, but only in plan.

Building S1 at Orsett, defined by gully F293

My reasoning about S9 was influenced by the Middle Iron Age Structure S1, defined by a shallow gully F293, which, though heavily cut about with later features, was clearly polygonal in shape. F293 was quite variable in depth and section, and crucially, it had gaps in it, which precluded a drainage function. It was clear that builders by this period had started building polygonal structures using either log built or timber frame load-bearing walls, as discussed in the previous article.(25)

A theoretical model plan of S1, showing: A: Polygonal wall foundation; B: possible ring beam;
C: Rafters

In the theoretical model of S1, Gully F293 is interpreted as the foundation of a polygonal wall, I was very confident of this, but its precise shape was speculative and unlikely to be correct. The postholes forming its ring beam were less certain, and the position of the rafters was for illustrative purposes only, but served to demonstrate that S1 was structurally viable.

The original (and erroneous) theoretical model plan of S9 showing: A: Drainage gully;
B: Conjectural wall; C: Main ring beam; D: Inner ring beam

The gully F19, defining S9, was also polygonal, particularly on its inside edge, and I reasoned that this must reflect the shape of the wall. The outer set of the two circles of postholes I interpreted as ring beams had only 10 posts, so the roof would not be that circular. Therefore I set the conjectural polygonal load-bearing wall next to the gully, but far enough away for it not to collapse into the gully. I could relate the wall's corners to the other components geometrically, and create a hypothetical structure that worked. The additional postholes in the interior were interpreted as resulting from alterations or repairs, some being temporary props for the structure, and the position of the rafters was inferred on this basis.


This and many aspects of the report was the product of a year’s analysis, and was as far as I could get in the time available. Much depended on the interpretation of F19 as a drainage feature, but it never quite made sense; the gravel soil at Orsett drained fairly well, so F19 could be a soak-away, but it had a volume of 20m³, enough to hold a month's rainfall running off the roof all at once.


As with my predecessor, Orsett was my last serious contribution to archaeology, and, after a largely unsuccessful attempt to sell the concept of CAD, computing, and structural analysis as a commercial consultancy to an incredulous archaeological world, I decided to sell skills to mammon in 1991. But I took with me the archaeological plans and the reports that had intrigued me, just in case I got bored – there’s no mystery in making money.


When, laden down with years of research into building analysis, I returned to archaeology, it had changed – it was now all commercial consultancies using CAD and computing. However, interpretation and structural understanding had not progressed very far. There was a brief moment of panic at university when a tutor said, “at least now we understand roundhouses” – till I realised they were talking about one of Mike Parker-Pearson’s theories. [3]

The roundhouse from Blyth, Northumberland, defined by a ‘drip gully’

In that strange cyclic way that the universe works sometimes, improbable drainage features came back into my life when a colleague said they had found a roundhouse defined by a ‘drip gully’ at a site in Blyth, Northumberland, [4] (one of over a hundred unearthed by my employer in the county).

A drip gully is defined as “A feature formed by water falling from the roof of a building and creating a linear or curving indentation in the ground” – English Heritage.[5]
The trench was polygonal in plan, and the section showed a roughly semi-circular dark stain, in a feature of variable depth, entirely consistent with a foundation trench for a log-built wall that had rotted in situ. Besides, the feature was cut into boulder clay; and water draining very slowly or evaporating would have left horizontal laminate layers, formed of clay particles. To argue that water dripping off a roof could create a feature penetrating over half a meter into a boulder clay subsoil is frankly preposterous. You don’t really need to have a roof – dig a hole like that in many soils, especially in winter, and it will fill with water of its own accord

My theoretical model plan of the Blyth roundhouse, showing the polygonal wall foundation. The fine blue lines serve only to indicate the geometry inherent in the model.

My superiors were not impressed with my reinterpretation, and if I could not cite similar (peer-reviewed) interpretations, I was wasting my time -- that’s just how archaeological interpretation works for some people.


I am happy to confess that nearly 20 years ago, under the influence of older wiser heads, I mistook the feature F19 at Orsett for a drainage feature, when it was really a trench wall foundation, for which I am very, very, sorry. By way of repentance, and with all the respect due to an old, much loved, but ultimately facile interpretation, I am going kick the concept of ‘drip gully’ as hard as I can in its intellectual nuts, and jump up and down on it till it stops moving.


Water goes downhill; it's done so since God was a boy, and that is how drainage works. You take water, a highly destructive and undesirable agent, away from your built environment under gravity. You do not, unless you are particularly stupid, encourage the water to drain away next to a building with timber foundations and walls largely made of mud. Nor do you allow it to form or dig a trench deeper than your wall foundation right next to your building.

Consider the size of the gutters and downpipes on a modern house: They don't have to be very large, because they take water away. Open drains don't have to be deep; they can be shallow, especially at their point of origin. That way they are easier to maintain, less of a hazard, and the 'drop' is conserved.

Drains, like water, go down hill, not round in a circle -- even in the Iron Age! ‘Drip gullies’ are nonsense, but, beyond confessing my own part in perpetuating this peer-reviewed myth, I have no interest in the extent to which this concept is still used or believed.

My new theoretical model plan of S9 showing: A: Conjectural wall; B: Inner ring beam.

What is far more interesting is that Orsett S9 had a massive load-bearing polygonal wall in a trench, which explains why it has so few posts in the ring beam. It also makes it one of the largest ‘roundhouses’ ever found. My new and improved theoretical model of S9 shows the conjectural wall in its new position, along with the 10 element ring beam. The inner ring beam is now in abeyance, and the rafter positions are still uncertain, but we will return to S9 in a later article.

The geometry of the conjectural wall foundation, and its relation to the ring beam in the new theoretical model of S9

Orsett S9 is very different from many of its processors in that its does not have the axial symmetry or the uneven number of posts typical of earlier roundhouses. [6]
Conclusions
With Orsett S9 we find ourselves very near the end of the roundhouse tradition in southern England, and since we have covered a lot of comparatively new ground in the last few articles, it is worth sketching out some general conclusions.

Graph showing the relationship between the number of posts in a ring beam and its diameter. (Ring beam diameter is used because it should be proportional to the size of the building’s roof, and wall or roof diameter are often archaeologically invisible and not available.)

• In the first millennium BC, post-and=lintel engineering was rational and schematised, as evidenced by similarities in post loading and ring beam design, and amply demonstrated by the graph above.
• Probably in response to a need to increase the functionality and flexibility of roofed interior space, the loading on posts was increased as better perceptions of safe loading were gained through experience (e.g., Little Woodbury & Pimperne Down).
• By the Middle Iron Age, buildings with polygonal load-bearing walls had developed, probably log built, which reduced the need for interior posts still further (Orsett S1/S9).

It should have become apparent that there was a lot more to the prehistoric built environment than mud huts transposed from Africa, and so far we have only looked at the building foundations. However, theoretical structural archaeology is all about understanding foundations in terms of superstructure, and if you think they went to all that trouble to create one room with a fire in the middle, then you are sadly mistaken.


Sources & further reading:

[1] H. Toller, 1980: 'An interim report on the excavation of the Orsett 'Cock' enclosure, Essex: 1976-79', Britannia, 11 pp. 35-42
[2] G. A. Carter, 1998: Excavations at the Orsett ‘Cock’ enclosure, Essex, 1976. East Anglian Archaeology Report No 86.
[3] M. Parker Pearson, 1999: "Food, Sex and Death: Cosmologies in the British Iron Age, with Particular Reference to East Yorkshire." Cambridge Archaeological Journal, 1999 pp 3-69
[4] TWM Archaeology. Unpublished Grey Literature.
[5] http://thesaurus.english-heritage.org.uk/thesaurus_term.asp?thes_no=546&term_no=137725
[6] http://structuralarchaeology.blogspot.com/2009/03/25-world-of-invisible-walls.html
Graeme Guilbert, 1981: "Double-ring roundhouses, probable and possible," in Prehistoric Britain. Proc Prehist Soc 47 & G. Guilbert, 1982: 'Post-ring Symmetry in Roundhouses at Moel y Gaer and some other sites in prehistoric Britain', Structural Reconstruction - Approaches to the interpretation of the excavated remains of buildings, British Archaeological Report 110, BAR, 67-86."