INTRODUCTION
King Arthur’s Round Table is the name given to a Neolithic henge monument sited approximately 3 km north-north-west of Penrith (in Cumbria), on a comparatively flat alluvial river terrace a short distance from the confluence of the rivers Eamont and Lowther (latitude: 54.6482 / 54°38’53”N; longitude: 2.7402/2°44’24”W). It is approximately 400m from the Mayburgh henge.
King Arthur’s Round Table attracted early antiquarian attention. In c 1664 William Dugdale sketched the remains showing two opposing entrances and two standing stones, one on either side of the north-western entrance. These stones were no longer there when William Stukeley visited in 1725. In the twentieth century, R G Collingwood conducted a preliminary excavation in 1937; in his interim report, he suggested that there was evidence that the henge had once contained a circle of upright stones or wooden posts. Two years later, Gerhard Bersu investigated the site and disputed Collingwood’s findings.
COLLINGWOOD’S HYPOTHESIS
In July 1936, members of the Cumberland and Westmorland Antiquarian and Archaeological Society (CWAAS) visited the prehistoric monument near Penrith known as King Arthur’s Round Table, ‘a roughly circular earthwork consisting of a central platform surrounded by a ditch: mound and ditch being penetrated by two opposed entrances’.Footnote 1 Their then president, R G Collingwood, noted its similarity to Arbor Low, in Derbyshire, and he spoke in favour of excavation as a means to explore this similarity:
Speaking of the possible purpose for which this great ring-fosse was constructed, Professor Collingwood compared it with Arbor Low in Derbyshire and similar circles in Britain. Merely as a suggestion, he hazarded the theory that we might have here ‘a miniature Avebury’, and that excavation might possibly reveal the former existence upon the site of a stone circle or of a wood circle of the early Bronze Age similar to that at Bleasdale. He strongly urged that this work should be undertaken by our Society.Footnote 2
As Collingwood reiterated in the introduction to his interim report, which was published in 1938:
simply considered as an earthwork, it [the Round Table] showed a very close degree of resemblance to the earthwork at certain other sites, notably Arbor Low, where earthworks of this same type are combined with stone circles. The question therefore naturally suggested itself, whether the Round Table had originally possessed a similar circle, but made of timber instead of stone. It was proposed to excavate the Round Table in order to answer this question.Footnote 3
This was the suggestion that Collingwood set out to test in his excavation in July 1937.
Mr W Cruddas was employed as foreman because his ‘experience on Hadrian’s Wall had familiarized him with modern digging technique in soils of the kind in which we expected to be working, and in particular with the study of post-hole sites in such soils’.Footnote 4 Mr Cruddas had ‘helped to find the post-holes of a timber barrack at High House Milecastle (50TW) on the Turf Wall of Hadrian’.Footnote 5 Collingwood also acknowledged the assistance of Miss K L Hodgson and Dr and Mrs Spence, ‘all of whom had done a good deal of prehistoric excavation in our district’.Footnote 6
His interim report records the confirmation of his suspicions. There was an array of post-holes. No dating evidence was found, but the post-holes divided into three distinct periods. Reproduced in this paper are the plans from Collingwood’s 1938 report. The later second period post-holes, which featured in fig 2 of his 1938 report, were dug into:
a level floor of the hard clayey gravel locally known as sammel. This is the kind of material which is still used for paving footpaths or other surfaces exposed to the weather. It sets very hard and resists weathering almost as well as concrete. Sammel is not natural to the site of the Round Table, where the subsoil is a river-washed gravel, doubtless deposited by the Lowther, lying upon a stratum of clay. It must have been brought from a deposit of boulder-clay and laid artificially on the site. The bluffs immediately to the west of the site are composed of boulder-clay, and might have been the source of this material.Footnote 7
Specimen section, taken through post-holes of ‘Circle B’, west of the entrance; reproduced from Collingwood’s interim report. Diagram: © Cumberland and Westmorland Antiquarian and Archaeological Society Proceedings
The earlier first period post-holes were sealed by the sammel. The post-holes of this, the earliest, phase of the monument are shown in fig 1. He detected indications of three circles of post-holes: A, B and C in fig 1.
First period post-holes; reproduced from Collingwood’s interim report. Diagram: © Cumberland and Westmorland Antiquarian and Archaeological Society Proceedings
After the first season’s digging, Collingwood arrived at the following tentative account of the site’s relative chronology,Footnote 8 this being that there were three periods of use (fig 2), with the first period being the earliest:
The earliest structure on the site was an elliptical wooden building, occupying the entire area of the central platform. On the south and west, its limits coincided with those of the platform; for the north and east this awaits proof. This building was made by erecting concentric rings of posts at intervals of 20 feet. The outermost ring consisted of large posts, mostly over a foot in diameter, standing singly; the other posts were on average much smaller (generally about six inches) and were grouped close together in dense belts rather than rows.Footnote 9
He adds that a fragment of burnt stone is suggestive of a cremation and that the entire monument may have constituted ‘a kind of temple to a dead “hero” or divine king’.Footnote 10
That the structure was roofed is a conclusion that Collingwood arrives at, as he explains:
Anyone erecting a fence in gravel such as exists at the Round Table would drive his four-inch uprights at least 18 inches to two feet into the ground. Yet the post-holes found by us were seldom over a foot deep, and none of them was deep enough to hold a post of the diameter actually involved against the most ordinary types of strain. Moreover, we did not find a single case in which evidence of such strain appeared in the shape of distorted post-holes or disturbed packing-stones. These shallow holes in soft gravel are intelligible only on the hypothesis that the posts which stood in them were connected together above ground in such a way that no lateral strain fell on the holes themselves; these holes in fact must have functioned merely as sockets in which rested the projecting feet of a complex and relatively rigid structure.Footnote 11
The second period of use for the henge could be evidenced, according to Collingwood:
At some period when the original building was completely decayed, but its form still clearly visible, the site was re-used and a quite new set of structures created upon it. The whole of the formerly built-upon area was floored afresh with sammel, well rammed and capable of resisting the weather. At points round its periphery (we know only of one such, but there may have been more) great standing stones were set up in sockets carefully prepared on the site of old post-holes.Footnote 12
As Collingwood makes clear, the existence of standing stones on the site is known from the testimony of antiquaries such as Dugdale. At the centre of the monument was a cremation trench, a bed-like hollow, filled with cracked and split pebbles, oak charcoal and blackened earth (fig 3).Footnote 13
Second period post-holes; reproduced from Collingwood’s interim report. Diagram: © Cumberland and Westmorland Antiquarian and Archaeological Society Proceedings
No plan or section drawing was included to illustrate Collingwood’s projected third period, but Collingwood informs us of a mound of soil and cobbles ‘immediately overlying the second-period floor’.Footnote 14 This third period started:
… by destroying at least one standing stone (possibly others, if there were others) on the edge of the central platform, taking down the central standing stones, and erecting a mound in the centre. The whole was thus converted into something rather like a disk-barrow.Footnote 15
Within the mound, a foot down, was a third-century coin of Gallienus (joint emperor with his father Valerian ad 253–260; ruled alone ad 260–268). Collingwood leaves it as an open question as to whether the coin is evidence of piety or carelessness.
Collingwood also records that overlying the entire central platform are huge quantities of boulders and gravel known, from documentary evidence, to have been brought to the site by William Bushby, the landlord of the nearby Crown Inn, at some date shortly before 1829.
1938: A DELAY IN EXCAVATIONS
There was no second season at the Round Table in 1938 as Collingwood was ill, and, largely because of illness, he was wherever possible attempting to reduce his commitments in archaeology. Thus, for example, in 1938 he resigned from his position as president of the CWAAS.
When an opportunity presented itself to delegate the work on King Arthur’s Round Table in the form of the German archaeologist and political refugee Gerhard Bersu, Collingwood was clearly relieved. This is obvious from a letter written to his wife from the Dutch East Indies:
R.E. Porter thinks that Bersu is available if required to dig the Round Table. I have answered saying grab him. Bersu can do the Round Table if anyone can; and I can; so Bersu can. This is logic. I am glad to think my last (and not least interesting) archaeological child is to be taken care of. I didn’t like the idea of letting the R.T. go undug. But I would much rather it went undug than went ill-dug. And really I couldn’t point to any English archaeologist in whose hands I should like to leave it …Footnote 16
In November 1936,Footnote 17 Bersu had been forced to leave his role as director of the German Archaeological Institute, Frankfurt-am-Main, in Germany, owing to his Jewish ancestry. In 1938 he and his wife were being supported by the British archaeological community.Footnote 18 German archaeology was held in high regard: ‘It is … essential to recognise the respect which German archaeology commanded in the 1930 s.’Footnote 19 It was admired for its professionalism, its state funding, its ambition and its systematic methodology. Furthermore, Bersu, in particular, ‘had been admired in Britain as a prehistorian for some years’.Footnote 20 In Britain the Bersus were supported in particular by Collingwood’s friend, O G S Crawford. Crawford and Bersu had themselves been friends ‘since they met in 1931’.Footnote 21 It was Crawford who found Bersu work directing the excavation at Little Woodbury, in Wiltshire, from 12 June to 18 September 1938.Footnote 22 This was simultaneously a research and training excavation and a showcase for Bersu’s work. Bersu was seen as an ‘outsider’ whom British archaeologists would do well to take notice of.Footnote 23
THE 1939 EXCAVATION: REFUTING THE HYPOTHESIS?
By 1939 the clouds of war may have been influencing Collingwood’s views of German archaeology. In his account of a sailing trip in the Mediterranean he writes damningly of a museum at Olympia:
It is not a nice museum. It combines a truly German pretentiousness with a truly German secretiveness. For evidence about the history of the place one has to search it in vain. It shows not what you want to see but what the Germans who arranged it wanted to show you.Footnote 24
Nonetheless, he was still well-disposed towards Bersu. On 24 July, learning that his wife (who had worked on site for the duration of the 1937 season) was planning to spend a week on the dig, Collingwood replied that he was ‘delighted that you are going to dig the R.T.. Give it my love and give my very kind regards to the Bersus and everyone else’.Footnote 25
However, although at this date Collingwood still held Bersu in high regard, all was about to change. Work resumed at the Round Table on 16 JulyFootnote 26 and ended on 27 August. Before the end of the dig Bersu had come to the conclusion that the ‘sammel’ was natural to the site and he had dismissed all of the post-holes found in 1937 as animal holes.
Bersu’s final report was unequivocal: ‘An artificial spread of “sammel”, as described in the Preliminary Report, could nowhere be observed.’Footnote 27 The supposed post-holes were in fact animal burrows. The excavation ‘leaves no doubt that wooden posts never occupied either the edge or the interior of the central platform. Still less was any certain trace of bedding for a standing-stone discernible’.Footnote 28 The ‘occasional collections of large-sized gravel, which might possibly have been considered as traces of packing for standing-stones, always proved on careful examination to be those occasional packs of larger material which occur at different levels in natural “sammel”‘.Footnote 29 The ‘sammel’ that Collingwood had found surrounded by the ditch was also to be found outside of the monument – it was natural. Bersu also argued that although the cremation trench had once contained cremated remains, the cremation must have taken place elsewhere. In the cremation trench ‘there had never been any fire sufficiently intensive to cremate a corpse’.Footnote 30
He argued that the entire monument was more disturbed than Collingwood had realised. The cremation trench had probably been disturbed when modern tipping had taken place. Moreover, the ditch surrounding the central platform had been entirely re-cut in the modern period, so as to obliterate all trace of the section and fill of the earlier prehistoric ditch. As to the Roman coin discussed by Collingwood, Bersu argued that as evidence it was worthless because it was unstratified. It was only found during the filling-in.Footnote 31
Bersu’s final conclusion was dispiriting:
If we attempt … to relate the monument to similar works, with outer mound, inner ditch, two entrances and an approximately central tomb, there is hardly yet sufficient evidence to place our badly preserved work in any definite group of such sepulchral monuments.Footnote 32
COLLINGWOOD’S RESPONSE
Collingwood’s reputation as an excavator has suffered as a result of Bersu’s judgements. In particular, Collingwood’s pupil, Ian Richmond, shows himself to have been entirely convinced by Bersu. In his ‘Appreciation of R.G. Collingwood as an Archaeologist’, Richmond wrote that at King Arthur’s Round Table ‘he [Collingwood] had made up his mind in advance what he was to find and found it with fatal precision’.Footnote 33 More recently, Ian Hodder and Richard Bradley have also followed Bersu’s line. According to Hodder, Collingwood:
imposed his view that the monument contained wooden circles on the evidence … To demonstrate that there were circles of post-holes within the enclosure he excavated circular trenches, plus some trenches on a radial axis which found entrances into the circle. Thus Collingwood had answered his question. The Round Table contained a prehistoric wooden circle. Collingwood may have been hard-pressed for time, but he made no attempt to consider alternative hypotheses. Indeed his layout of trenches ensured that he found what he wanted to find.Footnote 34
But to be fair, both Collingwood and Bersu dug some trenches that were curved, and in this respect neither conforms to contemporary methods of excavation. Bersu opened up a larger area, but this was because he was able to learn from the first season of excavation in which the removal of the post-medieval stratum was found to be unexpectedly time-consuming.
Bradley is in broad agreement with Hodder: ‘Collingwood was obviously the victim of his own preconceptions.’Footnote 35 One of Bradley’s criticisms is that: ‘So convinced was Collingwood that the enclosure would contain timber buildings that he hired a site foreman, who, he says, had special experience of locating post-holes on sites along Hadrian’s Wall.’Footnote 36 But, it would be more accurate to say that Collingwood was convinced of the possibility of timber buildings, and since these are not always easy to detect, he, reasonably, chose his foreman accordingly. In the years before the Second World War, the role of the site foreman in supervising the day-to-day work of excavation can be underestimated. As in this case, they were often men of great archaeological experience whose role was restricted by constraints of class and education. The suggestion that such an excavator would have been misled by supposed animal holes seems unlikely.Footnote 37
As mentioned, Collingwood’s reputation as a field archaeologist has undoubtedly been tarnished by Bersu’s report. At Little Woodbury, in Wiltshire, Bersu had confirmed his reputation largely by helping to discredit the idea of ‘pit dwellings’ and by detecting post-holes (in chalk natural) that had held wooden posts, so perhaps it was to be expected that Bersu would be believed over Collingwood. But, of course, there is also always the prejudice, in archaeology as elsewhere, that those who discuss theory know little of practice.
However, as the following anecdote illustrates – about Christopher Hawkes’s excavation at Alchester (September 1926) – this seems particularly unfair in Collingwood’s case:
Christopher had one very important visitor. R.G. Collingwood came to see them one afternoon, driven by his wife Ethel. This was a great honour, and Christopher showed him what they thought they had found – a middle-Roman-period occupation, with remains of stone walls reposing on a bed of apparently clean gravel, which he took to be natural.
‘Ye-es …’ said Collingwood, in his funny, high, squeaky voice, taking his penknife from his waistcoat pocket. Very gently and carefully he scraped into the gravel, and at once produced some charcoal. The gravel surface had been purposely laid over an earlier Roman level, which Christopher had failed to reach. He was very chastened by this; and after Collingwood’s visit they removed the gravel and found sherds of very fine early samian, and a number of objects, one of which at least was undoubtedly military, in the early-period level of occupation, well back in the first century.Footnote 38
Until now, it has not been known how Collingwood responded to Bersu’s criticisms of his hypothesis. No reply was made in print. Nonetheless, a reply was made in private. This was in the form of two letters written by Collingwood to his wife in England.Footnote 39 They were not written in response to Bersu’s report, which had yet to be published; but were written in response to the news, received in a letter from his wife, that Bersu had dismissed all of the post-holes as animal holes and that the sammel had been categorised as natural. Here is Collingwood’s first letter:
Hotel Beau-Site, AGAY, Var
20 Aug. 1939
I think you understand that I do not care a button about any theory of my own being disproved. I only did 3 weeks’ work at the R.T. I published it expressly as an interim report, which is as much as to say that everything in it awaited verification by future digging. Like every other decent archaeologist, I have spent my life building up theories about the sites I dig and knocking them down again. It costs me no more anguish to give away my 1937 theory than to give away my 1937 trousers. The situation is now quite different. It is not a question of whether my theory is right, it is a question of whether Bersu’s theory is right.
On that question, I will observe (1) that Bersu was called in to find out what the R.T. was. You say he isn’t at all sure. He has therefore failed (as these German Jews, with the possible exception of Jacobsthal, always fail when you give them a chance to make good).[ Footnote 40 ] (2) that whether or not he is right about mouze-halls,[ Footnote 41 ] he is wrong, and unforgivably wrong, if you quote him correctly, about the ditch. Seventeenth-century antiquaries attest the ditch a hundred years before Bushby. If your diagram is not a libellous misrepresentation of Bersu’s notions, it would seem that he has not even read my interim report, in which these antiquaries are quoted.
… I don’t want to be unkind to Bersu, but I want him not to destroy a possible career as an archaeologist in this country by making a fool of himself in public about the R.T., which is a very tricky site.
… As soon as I feel like travelling at all, I shall get home to see what can be done about this R.T. business.
Later. – I feel better today and think I am perhaps making a Panama Canal out of a mouze-hall. Alternatively, that this is Bersu’s goose and he can cook it any way he likes. But I don’t want him muddling up the minds of my Society, and I still think my points are better raised now than later.
Your R.Footnote 42
And here, in full, is Collingwood’s second letter to his wife:
As from 15 Belbroughton Road, Oxford
20. viii. 39
Your remarks on Bersu’s work at the Round Table are very interesting. Thank you very much for them. You won’t mind my asking a few questions, I know; because they are just the ones you would have asked already and as you say you are convinced by his results the answers must satisfy you. If you want to refer to himself for any answers you may send him this letter.
1. If, as you say, the boulder-clay is full of holes made by mice etc., presumably for domiciliary purposes, can he produce skeletons of occupants? Because boulder-clay is impermeable, and any little animal going to ground in a hole running vertically downwards in a level field of same is going to be drowned next time it rains, which is not likely to be long in Westmorland. Rats and rabbits, in that wet climate, burrow in banks where their holes can run at such an angle as not to be flooded.
1.1. Failing direct proof (skeletons) of the dwelling-hole theory, proof of some other kind will be called for. You don’t write as if there were any.
1.2. Failing that, the opinion of naturalists who have been called in as consulting experts would be useful. I say this more especially as the mouse-theory of the ‘packing-stones’ will, I think, convince nobody without such support. A naturalist who could say authoritatively ‘yes, mice always line their holes with packing-stones’ would be a most useful ally. Myself, I never saw a mouse hole thus lined. Digging animals normally kick stones out behind them.
1.3. What animals? I think the question will have to be definitely answered. You say ‘mice and other animals, rabbits or rats perhaps’. This will satisfy nobody. It doesn’t satisfy me. Such holes have a certain proportion to the size of the occupant, and they vary according to occupant’s habits.
1.4. Rabbits? It is believed that rabbits were introduced in the 10th century. A Bronze Age rabbit-hole, properly certified, would create quite a stir among the naturalists.
1.5. In any case, the dwelling-hole theory doesn’t fit the facts observed by me; viz. it doesn’t fit the fact that we have in a great many places a hard boulder-clay floor, a few inches thick, sealing the mouths of numerous holes beneath it. This clay, on what I take to be B’s theory, is natural. How does he explain its sealing all those front doors? He must have observed such cases for himself.
2. I don’t like your statement that ‘the ditch is entirely Bushby’. I reply, this statement is untrue in fact and scientifically unwarranted.
2.1 It is untrue in fact, because we have several plans and drawings of the R.T. dating up to nearly a century before Bushby’s time, in which the ditch is shown, and shown as the chief feature of the site. I have mentioned these in my report, if B. did not know of them independently. I will mention (a) Marshall’s MS. Plan, about 1660, in the Carlisle cathedral library (b) Dugdale’s MS. description and sketch at the College of Heralds (c) The woodcut plan in Gibson’s Camden (d) Stukeley’s elaborate engraving. I want to know how B. explains this array of testimony; you probably know already.
2.2 It is scientifically unwarrantable, at least I cannot see how it can be otherwise, because when a ditch has been dug in period A and cleaned out and enlarged in period B, which is what I say happened at the R.T., no one expects to find in it silt of period A, and the absence of such silt in no way disproves the theory that there was a period A ditch. And I can think of no conceivable test by which an excavator could prove that none existed. Perhaps you have misquoted Bersu here. If you have not, beg him for his own sake to be careful about the evidence on whose strength he denies the existence of the ditch recorded by the above four authorities.Footnote 43
Unfortunately, we do not have Mrs Collingwood’s letter. We do, however, know that she wrote to her husband on 13 August and that Collingwood received the letter at Antibes, in the south of France, on 19 August. We also know that she spent a week on Bersu’s excavation. Thus, by 13 August at the latest Bersu had already come to the conclusion that Collingwood’s post-holes were animal holes.Footnote 44 He had presumably not yet reached any conclusion about the cremation trench.
There is no evidence that Collingwood’s letter ever reached Bersu, for we may assume that if it had, even if he rejected Collingwood’s arguments, he would have acknowledged the advice in his report. In his report Bersu did acknowledged that the re-cut ditch surrounding the central platform followed the course of an earlier prehistoric ditch; but that alone does not indicate that he received Collingwood’s letter. It is possible that Mrs Collingwood might have misunderstood Bersu on this point; or else Collingwood might have misunderstood his wife. However, we are still left with two related questions. Is the ‘sammel’ natural and are the ‘post-holes’ animal holes?
POST-HOLES?
In relation to these two questions, Collingwood’s points 1 to 1.5 are highly relevant.
1. Collingwood is correct that, especially in a wet climate, animals prefer to dig into a bank rather than vertically down.
1.1 In Bersu’s final report there is no mention that any of the post-holes contained animal bones.
1.2 Collingwood is correct that digging animals ‘normally kick out stones behind them’.
1.3 Some of the post-holes have the diameter of rabbit holes, but
1.4 as Collingwood points out, rabbits were not found in this country in the Bronze Age.
1.5 This is perhaps the most important point. The first period post-holes were sealed by the sammel, which was of such a hardness that no plant roots penetrated it.Footnote 45 The hardness of the sammel is repeatedly emphasised in Collingwood’s interim report. ‘The second-period floor was … extremely hard and could only be penetrated by means of violent picking.’Footnote 46 ‘The second-period floor ran unbroken over it, quite smooth and almost as hard as concrete.’Footnote 47
Collingwood might also have made the point that the fill of the holes was unlike the fill of animal holes. In his report he describes the fill of a first period hole as including a ‘brown powdery soil’ – characteristic of soil that originates from decayed wood.Footnote 48 And, most tellingly, some of the second period holes were ‘full of blackened earth with scraps of charcoal’.Footnote 49 That is not the fill of a typical animal hole.
There remain several points made by Bersu against Collingwood which appear in his final report, but which appear not to have been communicated to Collingwood in France. As we have seen, he doubted that the cremation trench ever held a fire of sufficient intensity to cremate a corpse. He also claimed that the coin of Gallienus was unstratified.Footnote 50
With regard to the cremation trench, it now seems impossible to know whether Bersu or Collingwood is correct. The site was well-recorded for its time, but the records are not good enough to allow this question to be answered with any confidence. Likewise, the status of the coin – stratified or unstratified – must remain a mystery. Of course, in modern practice the coin would have been recorded in situ.
In his final report, Bersu also claimed that the sammel found by Collingwood extends outside of the site.Footnote 51 On this crucial point, a reply is implicit within Collingwood’s interim report: ‘The bluffs immediately to the west of the site are composed of boulder-clay, and might have been the source of this material [the sammel].’ In other words, the clay that Bersu located outside of the site was not the sammel but its source.
In summary, setting aside the questions over the cremation trench and the Roman coin, to the crucial doubts raised by Bersu over the post-holes and the sammel, Collingwood seems to have very strong replies.
CONJECTURES
Collingwood’s summary of his finds at the close of the 1937 season can be read in a notebook in the Bodleian Library. This can be seen as a rough draft of his report published in 1938. Following his usual practice, Collingwood is here organising his thoughts by writing them down. After noting, in strictly descriptive fashion, the discoveries of the 1937 season, Collingwood then sketches a basic history of the site. It is this hypothesis that he would have tested in his next season’s digging, if only his health had permitted.
As there is a fluidity to this account that is missing from his necessarily more technical interim report, the final section of Collingwood’s notebook is quoted at some length here.
Now I stop inferring and begin conjecturing. I suspect that the whole thing began life as tomb-temple: that some great local king was cremated in the middle, & that the whole hut was connected with the worship of his spirit. My reason for this conjecture is that there certainly was a cremation in period B. Why not A too? We haven’t found it yet except for one trace of it, which I recount because it is a curious example of archaeological method.
Near the B cremation trench were various A post-holes. In the bottom of one of these I found a fragment of a cobblestone which had been shattered by heat. It looked exactly like the fragments we found in the B cremation trench. But it could not be one of these because it was right in the bottom of the A post-hole, under the debris of the decayed A post. It had got into that hole before the A post was erected. Therefore the fire which split it (no mean fire) had burnt before the erection of the A hut had begun. I suppose it to have been the A cremation fire; for of course you build the thatched hut after the cremation of the man it commemorates.
The A monument came to an end by natural decay. The posts rotted quietly in situ. Then, when all had rotted, it was decided to re-use the site. The general idea of the new monument was: - (1) a cremation (and hence presumably burial, but we have not found that) in the middle. (2) a small hut, perhaps 50’ in diameter (we haven’t recovered its plan yet) in the middle, over the burial and cremation-trench. (3) a wide open space round that, floored with hard sammel. (4) a ring-ditch & mound outside that with entrances to NNW and SSE. (5) a sort of porch, of thatched wattle-work, at the entrances (we have 8 postholes showing this at the S entrance). (6) a number of big standing stones: a few in the middle. I think 4 inside the hut; a few around the edge of the platform (we have evidence of one); others at the entrances (we know that there were 2 outside the entrance in the 17th century) …
Evidence of date, with one flint flake tells us nothing. But Avebury is early Bronze Age; Arbor Low is Bronze Age; therefore this must surely be Bronze Age too.
The scientific interest of the dig lies in the fact that here for the first time a monument of the Bronze Age temple type, small enough to be manageable and absolutely virgin, has been dug by modern methods. Its resolution into 2 periods is certain: the absence of stone in A and its presence in B is certain; a vast roofed structure in A and a number of smaller ones in B are certain. All this is quite new and throws much light on the very difficult problem of the so called ‘henge monuments’.
For example: whether such monuments were ever, or often or always, wholly or partly roofed is a matter of controversy today. We have the first solid facts to contribute to this controversy. At Stonehenge lots of small post holes have been found but nobody has troubled to follow them up or think about their significance. I suspect that they indicate proto-Stonehenge like our A building. We therefore contribute something of importance to the Stonehenge problem. But it would be tedious to indicate the various ways in which this dig breaks new ground or fertilizes old ground in prehistory.Footnote 52
The difference between Collingwood’s general view of henge monuments and our own is that the origins of these monuments are now usually dated to the Neolithic. Collingwood only mentions the Bronze Age. However, the small post-holes at Stonehenge that Collingwood mentions have now been investigated and the idea of a standing timber structure predating the stone structure is now accepted. Collingwood would not have been surprised.
CONCLUSION
Invariably, even in his notebook, Collingwood is careful to distinguish data from conjecture and hypothesis. Nowhere in Collingwood’s response to Bersu is the point made that the existence of post-holes on the site is likely to be the case because it fits his initial hypothesis. Rightly so: Collingwood does not allow his hypothesis to dictate his findings. The plausibility of those findings, then and now, must be judged independently of any hypothesis. And, just as importantly, data must be separated from interpretation in such a way that when new evidence comes to light, earlier data can be re-interpreted. This is one, fundamental, reason that Pitt-Rivers was a good archaeologist. He recorded data independently of his hypotheses and his interpretation. This enabled Christopher Hawkes, many years later, to revisit Pitt-Rivers’ data and come to new interpretations. It is to Pitt-Rivers’ credit that this was possible.
This separation between data and interpretation is present in Collingwood’s interim report. It allows for re-interpretation. Collingwood realised the importance of this. Hence his comment: ‘It costs me no more anguish to give away my 1937 theory than to give away my 1937 trousers.’ However, so long as these distinctions between hypothesis and data and data and interpretation are maintained, then there is no inherent harm in positioning trenches in such a way that a hypothesis can be answered, one way or another. This amounts to no more than that we come to a site with questions in mind, rather than in the hope that ‘something will turn up’.
Collingwood made very plain the question which he was seeking to answer, but, independently of his hypothesis and his interpretation, his description of the soils, his section drawing and his comments on Bersu’s alternative hypothesis of mouse holes, all argue strongly in favour of redeposited sammel overlying genuine post-holes.
If this is accepted, then Collingwood’s conjectures may be worth revisiting. As Bradley notes:
At Balfarg in south east Scotland a circular enclosure of precisely the same type contained at least one timber circle, and this was apparently replaced by a setting of monoliths. At the centre of the monument was a grave. This was the sequence that Collingwood had worked out in his head over forty years before, but, paradoxically, it was a pattern that he might not have been able to identify on the ground.Footnote 53
But, as this paper has sought to demonstrate, that is very likely to be exactly what Collingwood was able to identify on the ground.
Bradley also notes that the suggestion that the post-holes imply the existence of a ‘huge thatched hut’ is now, in the light of subsequent work on other Bronze Age sites such as Balfarg,Footnote 54 much less surprising than it would have been in the late 1930 s. Other comparable sites include North Mains (Strathallan) and Cairnpapple. North Mains, Cairnpapple and Balfarg are all in Scotland and all originate in the Neolithic. North Mains and Cairnpapple both had twenty-four post circles – moreover, Collingwood’s third period mound is broadly comparable to that at Strathallan.Footnote 55 Cairnpapple was originally thought to have had standing stones, but it is now thought to have upright timber posts.Footnote 56
Admittedly, the excavators of Balfarg, after carefully weighing the evidence, do not themselves accept that that site was roofed; and Collingwood’s argument for a roof at King Arthur’s Round Table seems weak, in that it is dependent on evidence for and against lateral strain within the excavated post-holes,Footnote 57 and such evidence seems inconclusive given the ground conditions.Footnote 58 However, we do not have to accept the conjecture of a roof in order to accept the evidence of a concentric ring of post-holes at King Arthur’s Round Table very similar to that found at Balfarg.Footnote 59 The post-holes may have held timber posts that supported hurdling, which were perhaps, like the bank and ditch, seen as ‘containing’ the spirit at the interior of the monument – in the sense of deterring its escape.Footnote 60 On this interpretation, the interior of the monument would have been rarely visited and a roof might not have been thought necessary.
But though it is doubtful that any part of the site was roofed, that there is a ring of post-holes at King Arthur’s Round Table, similar to those at Balfarg, North Mains and Cairnpapple, but sealed by a layer of sammel ‘so hard that it is quite difficult to dig through with a pick, like road-metal’,Footnote 61 seems highly likely – certainly more likely than that these features are mouse holes.
From the evidence of Collingwood’s plans, the outer ring (A) is most convincing; the middle and inner rings (B and C) are less so. There was insufficient ground opened up to be confident about the inner rings. The inner post-holes may indicate a more complex pattern of features.Footnote 62
In assessing the evidence, we must not be biased by the fact that Collingwood is primarily known as a philosopher and Bersu was primarily known as a field archaeologist. Neither should be seen as infallible.Footnote 63 However, there is no reason – other than the academic pressure to specialise, which Collingwood did his best to resist – why someone cannot be both a philosopher and a field archaeologist. Moreover, as the author hopes to have shown, Collingwood’s reply to Bersu’s criticisms has a strength that will be recognised by field archaeologists.
For Collingwood, the closest parallel to King Arthur’s Round Table was Arbor Low; but, if we take his findings seriously, then, in the light of subsequent work, it may now be worth comparing the King Arthur’s Round Table to Balfarg, North Mains (Strathallan) and Cairnpapple.
On different soil, we might look to geophysics to resolve the problem of the post-holes once and for all, but the conditions at King Arthur’s Round Table are some of the least promising for geophysics. It is therefore very much to be hoped that the current owners of the site, and Historic England, might wish to see the problem resolved by means of a limited excavation, in order to improve the interpretation of the site for their visitors.
ACKNOWLEDGEMENTS
The author wishes to thank the editor of The Antiquaries Journal, the anonymous referees of this paper, and Kenny Brophy and Gordon Barclay.
ABBREVIATIONS AND BIBLIOGRAPHY
Abbreviations
- CWAAS
Cumberland and Westmorland Antiquarian and Archaeological Society
- PLTR
personal letter transcript
- Proc Antiq Scot
Proceedings of the Society of Antiquaries of Scotland
- Proc Brit Acad
Proceedings of the British Academy
- Proc Prehistoric Soc
Proceedings of the Prehistoric Society
- TS/PLTR
Teresa Smith Archive/personal letter transcript
