The Scientific Basis for the Reconstruction of Prehistoric and Protohistoric Houses

2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses The scientific basis for the reconstruction of prehistoric and protohistoric houses

The purpose of this paper was al resources, their management, the past with ‚appropriate structures‘ and to explore the scientiﬁc basis of engineering necessities of construc- most unhappily with ‚actors‘ dressed building reconstructions. The tion, environmental observations, for the part. critical issue was to address the functional analysis, structural deg- problems of reconstruction in radation through time and critical- While this is undoubtedly the case, order to specify limits within ly the physical interpretation of ar- the question of responsibility nec- which the reconstruction is of chaeological trace evidence. On the essarily arises. A proportion of that research/educational value and other hand there is the educational responsibility, probably the greater to a set standards which may act motivation which must include both part, must be shouldered by the aca- as guidelines. museological and entertainment el- demic community. Reconstructions ements. The research mode clearly motivated by research intentions contains a major educational museo- have been part of European think- n = Peter J. REYNOLDS logical input in that a three dimen- ing for several decades. Scale mod- Butser Ancient Farm (UK) sional structure is an overt demon- els through one-to-one structures to stration which can be utilised in a complexes of structures can be found In 1966 just outside the boundary of great number of ways that need not in virtually every country. That the a hill fort known as Kemerton Camp conflict with basic research parame- exploitation of this approach has sub- on top of Bredon Hill (Hencken 1939) ters. However, during the last twenty sequently been left to entrepreneuri- in Worcestershire a small round- years, with a marked acceleration in al business can hardly be a charge house was reconstructed, based the last five years,(1) there have been raised against the exploiters. Rather upon the excavations carried out at built a large number of reconstruc- it is a further representative facet of Glastonbury Lake Village some fifty tions throughout Europe which are the malaise of archaeology, then be- years before (Reynolds 1967a, Bullied inspired not by any altruistic educa- ing a basic lack of communication and Grey 1911). A group of students tional or research motive, but rather skills. This is not to impugn the proc- under the guidance of Mr. Philip a thinly disguised commercial venali- ess of academic publication or to en- Barker of Birmingham University, ty. To a great extent these reconstruc- courage popularisation at the cost of carrying out a routine site visit, were tions are to be found in theme parks integrity. It stresses rather an inabil- deeply impressed to come across the and leisure complexes designed to ity to communicate with the public three dimensional reality of some- mimic the Disneyworld Enterprises at an appropriate level and thereafter thing which had been previous- in the United States of America. Such to exploit the undoubted general in- ly discussed in vacuo. In effect the theme parks, a permanent successor terest of the public to the advantage house reconstruction was an ele- to the travelling fair,(2) are designed of archaeology as a whole. That the ment of a small integrated research to entertain the public on a purely public interest exists is abundantly programme examining the domestic commercial basis. It is to be regretted proven by the elements contained and agricultural economy of the Iron that so many of the historic and pre- within theme parks and their like, Age (Reynolds 1967b, 1969). The historic elements have been created, an interest which can arguably be purpose of the house reconstruc- often at great expense, without any identified as the ’heritage syndrome’. tion was entirely oriented towards recourse to professional advice. Had Recently some attempts have been a research mode and fundamental sufficient research been carried out made to rectify this situation, most questions involving material require- by the designers a better end product notably at the Jorvik Viking Centre ments, the effects of the environment could well have resulted, at far less in York where excavation and recon- upon the structure and the effects of cost and to the mutual advantage of struction are presented in dramatic the structure upon its immediate en- all parties, not least the public who style employing Disneyworld tech- vironment. In the event the recon- are most at risk. Educationally, simply niques to the best advantage. Its suc- struction lasted but a short period because of the understandable popu- cess and popularity are undeniable, before it was burned down by van- larity of the ‚modern fairground‘, the its realisation of marriage of invest- dals and consequently only indica- result is quite disastrous. Prejudice ment venture capital and archaeo- tions of the potential of reconstruc- and ignorance are reinforced across logical expertise on a sound busi- tions were realised at that time. This a far greater percentage of the popu- ness footing. Yet even this enterprise episode, nonetheless, focuses exact- lation by default. Unfortunately the denies the presence of any research ly upon the dichotomy of motiva- theme park approach has inspired a element. Rather it is the explanation tions for reconstruction. On the one number of smaller enterprises, sim- of an excavation, the interpretations hand there is the research motivation ilarly motivated by commercial rea- frozen at a moment of time allowing which includes questions of materi- soning, to ‚recreate‘ periods of the for little or no development.

(1) Please remember the original date this text was delivered (ed.) (2) See Comis in this volume (ed.) (3) See Coles in this volume (ed.)

58 euroREA 3/2006 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses ESF workshop 1987 / ITEMS 2

The critical issue is therefore to ad- Similarly reconstructions, to have any state of the archaeological evidence, dress the problems of reconstruction value, must be at a one-to-one scale further empirical trials are needed to in order to specify the limits within with the original evidence. Scaled monitor its life span both as a struc- which reconstruction may be wel- models particularly of wooden structure in terms of its requirements as comed and to set out standards which tures representing vernacular ar- a structure, repair and refurbish- may act as an advisory remit. It is only chitecture are largely unsatisfactory ment for example, and its function by setting out clearly and authorita- from any standpoint. The nature of and the traces that function might tively levels of acceptability against timber, especially in the round, its create. In this particular connection which any product may be measured weight, its size to strength ratios de- function can be further subdivided that the issue may gain any resolution. nies miniaturisation for any empiri- into the human activity for which The purpose of this paper is there- cal purpose.(3) Even in terms of mu- the structure was designed and the fore to explore the scientific basis of seological representation such models effect the structure had upon its im- reconstruction and attempt to iso- are unsatisfactory since the buildings mediate environment in terms of late the pertinent ranges of questions concerned are directly relative to hu- identification by geophysical means. which will not only allow but foster as man experience. Even from the sim- With recent advances in geophysical many variants as possible. To achieve ple visual aspect such models are prospection, particularly the mag- this end a number of case studies in- doomed to be only models. Given the netic susceptibility of the soil (Clark cluding a detailed report and analy- modern skills of architectural model- 1990) there is a clear case for moni- sis of one major reconstruction have ling with contemporary materials of toring within and without the struc- been used. All the examples quoted concrete, brick and steel, for which ture. Recent research indicates that have been the direct responsibility of there is a full mathematical and engi- a structure has an irreversible effect the author and to an extent demon- neering data base, these museological upon the ground surface it covers strate the learning process. In any re- representations of pre- and protohis- (Reynolds 1995). Similarly if a struc- ceived body of information there has toric structures are put into an even ture is given an hypothesised func- to be evidence of progression prima- more inadequate context. tion by the excavator, the replication rily in the exploration of the original of that function empirically may well concept which argued for reconstruc- Empirical trials with models like provide confirmatory evidence. The tion in the first place. those carried out by Sarayader clearest example is the permanent (1983) with Amerindian structures or seasonal housing of livestock and Before approaching any reconstruc- seeking to establish ‚comfort levels‘, the ensuing enhanced phosphate lev- tion it is important to resolve a number similarly beg the question of an ad- els identifiable in the subsoil Pryor( of arguments which are inherent in equate comprehension of materials 1980; Reynolds 1995). Naturally this archaeological methodology. Of these and their interaction both with each approach is significantly useful since, perhaps the paramount one is philo- other and with the environment. Any once an association of cause and ef- sophical, concerning the data recov- reconstruction of a building can only fect with attendant intensity lev- ered by excavation. The basic premise be justified if it is firmly keyed into els has been established, the scien- must be that the data inspire the in- the archaeological process. In prac- tific analysis of site evidence may be terpretation. An interpretation which tical terms an excavation which re- used to provide evidence of function does not agree with the data is en- veals the original presence of a build- where visually none exists. In other tirely unsatisfactory although it may, ing, particularly in ‚dry‘ archaeology, words, the creation of a proven data in fact, prove to be correct. In effect provides evidence of the final state base will enhance archaeological in- this latter approach has gained con- of that building. A pattern of post- terpretation without any further ne- siderable popularity in recent years holes and stake-holes, disturbances, cessity for reconstruction. where ephemeral and postulated data hollows, trenches and the like which have been employed to construct is recognised by the excavator as the The ultimate phase in reconstruction models of social organisation and evidence for a building but needs to has to be, logically, viewed as de- even kinship (Clarke 1978; Hodder, be appreciated as the ‚death‘ state, struction. It is only at this point that Isaac and Hammond 1981; Kimes, a state which not only provides an a real correlation can be made with Haselgrove and Hodder 1982; Renfrew original plan but which also provides the original evidence upon which and Shennan 1982). Essentially the evidence of the life span. Repairs, ex- the reconstruction was created. The data are the hard evidence, particu- tensions, alterations, refurbishments most dramatic and statistically least larly with reference to a structure. are all part of the life of a structure. normal end of any structure has to be While it may be possible, once a suf- The distinction between the original conflagration; a trial carried out but ficiently large database has been ac- building plan and subsequent struc- as yet not fully published by Hansen. quired, to reinstate ‚missing’ elements tural changes which include repairs Probably the most useful data would that database is currently not availa- is critically important. Even in the be forthcoming from abandonment ble. Therefore, a reconstruction virtu- original building plan it is important following a period of use. Such a tri- ally by definition must be specific to to isolate, if possible, constructional al, however, is only possible should a set of excavated data. A generalised and structural elements. In effect a a reconstruction be adjudged to be reconstruction thus relies particularly reconstruction can only be archaeo- sufficiently accurate. This ultimate on a synthesis of data from disparate logically justified if it feeds back into phase, however, begs the real ques- sources and serves only to underpin the excavation positive and nega- tion of the validity of function. prejudices. Inferred hypotheses do tive data about the structure. Ideally, not necessarily advance our compre- since a reconstruction creates an hy- In recent years there has been an hension of the actual problems. pothetical new building, the birth emphasis upon ‚reliving the past’.

3/2006 euroREA 59 2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses

On one occasion a group of people, itial construction stage, the yield of structure, however, since the data suborned by television, attempted to information levelling off significant- yield from the construction is one of live an Iron Age way of life for a year ly after completion and becoming the greatest sections of the learning (Percival 1979). The outcome proved increasingly less after approximately curve, can be thoroughly modern to be a strange style of ‚soap opera‘, fifteen years. Since this is the case, and recognised to be so because the perhaps satisfactory to a viewing au- reconstruction can be categorised experiment is effectively conclud- dience but of no real value to archae- to include those designed for mu- ed and all that is required is exhibi- ology whatsoever. It is the convinced seum purposes, educational motives tion. The same is true for educational opinion of the writer that mankind as well as components of integrated programmes and, indeed, for theme is locked into his own time, the only long-term research programmes. and leisure parks referred to above. escape being into the future. This There is no reason whatsoever why At least the end product has satisfied does not deny the indulgence and any reconstruction should not be a the basic criteria of springing from entertainment of re-enactment, nor part of the excavation/interpretation and therefore elucidating the archae- the potential educational value that process initially, and thereafter fulfil ological evidence. it may have. Rather it points to the alternative requirements. However, impossibility of recreating ‚Everyday this will only happen if the recon- Finally one overriding problem be- Life‘ of any period in the past. struction is site specific and is mo- sets all reconstruction work. Can the Anthropology serves to underline tivated by the requirement for an- time taken to build a structure have the immediate impact of interactive swers from questions raised by the any significant value for the under- presence in remote communities. excavated data. This approach is the standing of prehistoric or historic la- The more so for individuals from a more desirable because of the not in- bour input? The supplementary ques- modern society playing at being pre- considerable costs involved and the tion concerns the tools with which a historic man. Therefore any study of need to maximise a return from the structure is made. Should such tools the function of a structure has to be investment. Further there is a clear be replicas of the originals? To both virtually dehumanised and limited element of responsibility to the pub- principal and supplementary ques- to choiceless action. Within this con- lic at large who invariably, directly or tions the answer has to be a firm straint the area of study is inevitably indirectly, provide the funding both negative. The main question of time much reduced to activities like traffic for excavation and interpretation. A taken to construct a building, espe- wear or erosion through doorways generalised reconstruction without cially since the process is essentially and the effects of domestic fires and any scientific basis is no more than one of trial and error, is entirely su- cooking areas in terms of wood sup- a simple insult. Particularly marked perfluous. Even if one became skilled plies. On the other hand functions are the occasions when the public in- in the construction techniques, com- involving livestock, particularly sea- cludes building engineers, architects parable to a qualified modern brick- sonal indoor maintenance, is per- and craftsmen, all of whom are prob- layer, the question is still meaning- fectly viable as is the examination of ably better qualified to assess the less since the motivation is unknown. specific agricultural and industrial building in question rather than the The analogy of the modern brick- processes and functions. One exam- archaeologist who is straying beyond layer holds good in that, depending ple of a dehumanised empirical func- his brief. Loss of professional cred- upon pay return, two hundred or a tion has been the examination of the ibility can be damaging especially thousand bricks can be laid in a day. so-called Romano-British grain dri- when caused by elementary errors, With regard to tools, the answer can er referred to below (Reynolds and the most common of which is found be found by refining the question. Langley 1979). not so much in physical reconstruc- The object of study is the building it- tion but rather in reconstruction self and given the irrelevance of the Consequently a realistic total test- drawings. Regularly a splendid ar- labour input, modern equivalents ing of a reconstruction, except in chaeological report of an excavation to the original tools are sensible and very particular circumstances, from is marred by a drawing of a struc- adequate. The study of ancient tools building through function, degrada- ture which would defy physical re- is specific to the tools themselves tion and destruction, may be impos- alisation. Such ‚paper tigers‘ are to in that one‘s major interest is the ef- sible not least from the point of view be avoided. Indeed, this aspect alone fect of the work on a replicated tool. of the life expectancy of a structure must be adequate justification for re- Tool manipulation immediately en- which may reasonably be expected construction. croaches upon skill and labour input to extend over many decades if not associated with time expenditure. centuries. On the other hand much Structural reconstruction can quite may be learned, up to and partially properly be part of a museum com- The following case studies are drawn including human activity, of the life plex within the context of empiri- from the experience of the writer span of a structure. Certainly the bal- cism but with a specificcaveat . and are used here to substantiate the ance is in favour of reconstruction in Should the reconstruction prove to elements of the argument presented that the data base can only be en- be in error in part or whole, the error above. No account has been taken of hanced. should be recorded and corrected, other reconstructions, since the na- even if this means a total rebuilding ture of the workshop(4) is to bring The ideal approach outlined above of the structure. The use of modern together interested parties from all presents a full cycle for a reconstruc- materials, the ubiquitous long steel over Europe who will present their tion, as in any ideal it is an extreme nail, to fudge constructional de- own material evidence. With some case. There is no doubt that the tail is of no advantage whatsoever. exceptions a major emphasis is learning curve is strongest at the in- The subsequent maintenance of the placed upon the author‘s research

60 euroREA 3/2006 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses ESF workshop 1987 / ITEMS 2

of structures of the British Iron Age. such reconstructions of simple as op- Detailed analyses are not given for posed to complex roundhouses it is the majority of the cases discussed, necessary to prop the doorposts to only those observations which are counteract the outward thrust of the relevant to the discussion. basketry walls. A central post with a Y fork was set into the centre of the The insular building tradition of floor to support the roof apex, the Iron Age Britain differs from that angle of the roof being at the critical of the continent, with the excep- functional angle of 45° to the hori- tion of northwest Spain, in that the zontal. At angles of less then 45° and normal house was circular in plan. more than 55° thatched roofs leak Thus the problems experienced in (see below Balksbury House). The reconstruction are quite different initial tripod of rafters were laid from those of a rectangular struc- into the Y fork and lashed to the ture. In simple terms a roundhouse side of the wall uprights with raw- n Fig. 1 Maiden Castle house: the completed structure. Set comprises a cylinder surmounted by hide thongs. A supplementary ring on the exposed spur of Little Butser this house has withstood a cone. In building, one is continu- was attached to the tripod one third hurricane force winds with impunity. Aerodynamically its ally exploiting the principle of the down the slant height of the rafters shape is ideal At no time is a flat surface presented against tripod rather than the ‚A‘ frame and and to it were attached the remain- the elements (in: Reynolds 1979, 35). using balanced opposing forces in a ing rafters, the ends of which were three dimensional field. In all cases a similarly lashed to the sides of the fail subsequently. Two structural ele- round house gains strength steadily uprights in the wall. The house was ments were in error. The rafters must throughout the construction period, completed with the application of be set upon the top of the wall posts being strongest only at completion. daub to the walls inside and out and probably with a simple notch, and In effect a roundhouse can be built of the roof was thatched with straw. A secondly the central support pole is slight materials and still achieve re- door of split oak planks was fitted totally unnecessary to the structure. markable strength. In contrast a rec- to the doorframe with pivot hinges In all probability an alternative inter- tangular structure is created with the set in the lintel and a sill set on the pretation for this feature should be in-built strength of each component ground surface. proposed where it will fulfil a non- element. Ultimately the roundhouse structural role. presents the perfect structure since For a period of ten years this structure aerodynamically it offers the least re- fulfilled all expectations. It success- One interesting ancillary result sistance. This factor contributes sig- fully withstood extremes of weather emerged from this structure during nificantly to its life expectancy. including tornado force winds, heavy its life span. The space under the wall snowfalls and torrential rain. After was exploited by rats and mice. They this time, however, the rawhide lash- effectively tunnelled around a third Case Studies ings began to fail and the roof began of the circumference of the build- Each structure referred to below is to move downward roughly at a rate ing at a depth below the base of the allocated the site name from which of 50 mm per annum. Gradually the posts of the wall thus obliterating the the evidence came. wall posts began to protrude through post-hole evidence. In archaeologi- the roof, the brittle wall cladding giv- cal terms the house was part gully, Maiden Castle House ing way under the weight thrust. As part an arc of post-holes. (Wheeler 1943) the roof descended the central support post caused the rafters to pull The further observation this house The plan for this reconstruction inwards at the wall circumference allowed for was the effect of water was drawn from the excavations of in a similar behaviour pattern to an run-off from the roof. Expectation Maiden Castle in Dorset. Specifically umbrella closing. Finally the roof argued for the gradual appearance of the plan for Hut Db was used as the twisted through approximately 15° a drip gully. In fact, the opposite con- basis for this reconstruction, the causing partial wall collapse with the dition occurred. Directly beneath the first to be built at the Butser Ancient original tripod held in place by the eaves a protected habitat was created Farm Research Project on Little central support erupting through the where plants flourished undamaged Butser (Reynolds 1978). Essentially apex of the thatch. In all senses the by the passage of feet. Through time, the plan comprised a circle of post- structure failed. even though these plants, prima- holes with a centrally placed post- rily nettles, weeds and grasses, were hole. It was decided to build this The object lesson from this re- trimmed back, a ridge of humus was structure using the simplest technol- construction underlines the criti- formed. Between it and the house ogy, the only joints being the mortice cal requirement to study a building wall because of the lack of light, and and tenon joints joining the lintel to through time. Fundamental build- therefore an absence of photosyn- the door posts. The walls were in- ing errors will not necessarily be- thesis, shallow erosion occurred to a terwoven with hazel rods creating a come immediately apparent. In this depth of 100 mm. Finally the thresh- very strong basketry frame, the full case the building gained ready ac- old of the house became considera- cylinder being completed with the ceptance as a reasoned projection of bly eroded due to passage of feet es- lintel span across the doorway. In all the archaeological evidence only to pecially in wet weather leading to a

(4) i.e the conference in Århus for which this paper was written (ed.)

3/2006 euroREA 61 2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses

sizeable puddle being created. This stress point serves to hold the roof 0.80 m. This trial begs a number of particular observation allowed the apart counteracting the natural sag questions, not least of which is the author to identify the presence of of timber along the length. In this potential re-utilisation of fallen stone doorways at an excavation of a group structure coincidentally the meas- in local field walls, which are a com- of Iron Age roundhouses at Skipton urement from the apex to the wall at mon feature of the field boundaries in Yorkshire where the only evidence a 45 angle exactly equalled the meas- on the hill. It is likely that any col- comprised a depression at one point urement from the stress point to the lapse material would have been thor- on the outer circumference of each outer lintel of the porch. The angle oughly overgrown had it been left house (pers.comm. A.Aberg). generated at this point was a little be- in situ originally, forming a mound low 40° and was judged at the time of which a large number remain to The Balksbury House to be worth risking, or alternative- this day and to which was ascribed ly, testing. The completed structure local suspicious tradition (Reynolds Excavations at Balksbury, Hamp shire was extremely elegant with sweep- and Lloyd 1967). More importantly, (Wainwright and Davies 1995) re- ing curves in the thatch and an im- the kind of stone used in both the vealed a post-built roundhouse 9.10 pressive porch. It proved extremely house and the field walls of two mil- m in diameter with the post-holes at strong but, in fact, was doomed to lennia later can be found simply by intervals of 1.90 m around the cir- destruction. The roof over the porch, digging into the hillside, especially cumference and an extended porch below the critical functional angle, on the oolitic limestone capping. The or entrance on the eastern side some gradually began to leak, the reed be- large number of quarry pits on the 2.00 m long. A reconstruction was at- coming saturated and steadily in- hill suggest that the field walls were tempted because the wall posts indi- creasing in weight. Within the space created from freshly-quarried stone. cated a different building technique of three years the porch roof was The exercise, therefore, had consid- to interwoven basketry. On this occa- horizontal, the rest of the cone being erable value and allowed one to sug- sion a wall plate was set on top of the dragged down uncompromisingly. gest that the original wall could have posts with mortice and tenon joints The house was dismantled as a fail- been but a little higher than the trial to create the base cylinder. The spac- ure. The failure was generated sim- showed. es were filled in with wattle hurdling ply by disobeying a structural law. which was by definition non-struc- In the event it has been hypothesised As in any reconstruction, all the spe- tural simply serving as a base for the that the original excavation revealed cific data as they survived were care- daub. The roof structure conjectured only part of the building plan of this fully replicated after detailed study. included an integrated porch. In the structure, that being the inner ring of This reconstruction followed the building of a cone shaped roof, as posts. All evidence of the outer wall pattern of all such exercises, in that indeed in any angled roof, the tim- had disappeared (Guilbert 1981), the the major benefit to accrue emerged bers require to be stressed a third original plan being similar to the from those very aspects which were the way down the slant height of the Pimperne House (q.v. below). assumed to be simple, yet when faced roof from the apex. In practice the with the physical problem, proved The Conderton House the most difficult to solve. The original walls were built of the loose up- Excavations at Conderton Camp on per layers of limestone, the general Bredon Hill, Worcestershire (Thomas size being small. It would appear that 2005) revealed the foundations of a the wall was built with two faces and stone-built roundhouse. The struc- then infilled with rubble. This, while ture survived as two courses of dry commonly suggested as a method of stone footings of an Iron Age house building, is an illusion. The advice of with a paved porch. The internal di- a local stonemason, Mr. Hopkins of ameter was 6.09 m, the external di- Tewkesbury, was sought, who, after ameter 7.91 m, the walls averaging some persuasion, taught the author 0.91 m thick. The doorway was quite the rudiments of building dry stone n Fig. 2 The finished Balkesbury house (photo Reynolds). narrow at just 0.90 m in contrast to walls. This examination of the evi- the great majority of timber hous- dence rather disabused the lay con- es. There was evidence of a collapse jecture that one simply built up two of some 2.80 m of the wall around face walls and then filled the interior the circumference in the north-east with rubble. In practical terms such quadrant. During the excavation, an a wall would be inherently unsta- experiment was carried out to at- ble and would readily collapse if any tempt to establish the original wall pressure were exerted on it, espe- height. Stone collapse was carefully cially any lateral thrust such as one gathered from a segment of the ex- would expect from a roof. cavated area centring on the middle of the house and extending to the The original wall was in reality care- limits of the excavation. All the rem- fully laid throughout. The edging n Fig. 3 The framework of the Balkesbury house. Unlike nant stone rubble was gathered and stones were laid first, usually two the Maiden Castle house, its strength is drawn from the rebuilt onto the section of wall con- courses at a time. The interior was continuous wall plate attached to the uprights with sim- tained within the segment. The re- then laid so as to lock in as many of ple mortice and tenon joints (in: Reynolds 1979, 94). sultant height of the wall was about the edging stones as possible. The

62 euroREA 3/2006 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses ESF workshop 1987 / ITEMS 2

whole exercise is like a heavyweight usual five because of the wall rebuild- interlocking jigsaw puzzle. The ma- ing each trial involved. Twelve trials son was quite adamant that ‚rubble- were executed to achieve this an- filled walls‘ was an extremely poor swer. Given the circumference of the description misunderstood by the structure at this position on the wall general public, in that the degree of (21.0 m), the length of 2.40 m neatly skill required and the nature of the allowed eight baulks with a gap left construction was not far removed for the lintel over the doorway. The from the seemingly superior solid ends of the baulks either side of the stone wall. Simple lessons and ad- doorway were built into the wall as it ages were passed on to the author increased in height to accommodate and need to be repeated here. ‚No the doorway. one stone is handled twice - there is always a perfect place for it in the sec- By placing the baulks of timber at n Fig. 4 Conderton House completed (in: Reynolds 1982, tion of wall within reach of the build- this position on the wall, achieved 195). er’. ‚One stone never sits upon one by simple trial and error, the pres- stone‘. ‚Locking stones should be used sure line from the roof passes more as frequently as possible‘. A brief ap- or less diagonally through the wall prenticeship gave some facility to the mass, allowing the outer element author and these adages were found to act as a simple weight buttress. to be fundamental. Similarly, the trials focused attention on the details of the excavated The structure was built at the data and of their value in interpre- Avoncroft Museum of Buildings tation allied to empirical trials. The near Bromsgrove in Worcestershire. baulks themselves were connected The construction of the stone wall with simple half-lap joints. No at- was, in a sense, the easiest element. tempt was made to fasten the tim- It required no less than 50 tonnes of bers together since this would have limestone and was built up to accom- precluded the possibility of the ob- modate a doorway just 1.30 m high. served wall collapse. n Fig. 5 Conderton House as exposed by excavation (in: The major difficulties began when The rafters, on the other hand, were Reynolds 1982, 191). the alliance of disparate materials jointed to the horizontal baulks of had to be forged. It was appreciated timber which effectively formed a problem to emerge again concerned that the roof would exert considera- wall plate. Again the joinery was as the wall. Trials had shown that the ble lateral thrust on a dry stone wall, simple as possible. A prepared face thickness of the wall was necessary which is generally regarded as unsta- was cut into the wall plate at ap- to buttress the weight thrust of the ble in such conditions. Indeed, the proximately 1.50 m intervals, with roof, and that the wall plate needed original showed clear evidence of a the angle of this face at 45° to the to be positioned no more than 0.30 collapse. Mathematically the lateral horizontal. The butt of each rafter m from the inner circumference. pressure exerted on a point of mo- was similarly prepared by a half lap This left a flat surface c. 0.5 m wide ment is least .at 45° and most at 22.5° joint at right angles to the line of the which had to be protected from the and 67.5°, but nonetheless there is wood thus allowing the face to lock weather. Limestone is especially fri- still pressure. A series of trials was by weight alone onto the wall plate. able when subjected to frost action in carried out to seek the best system of The rafters were erected in a precise the horizontal mode. sustaining that pressure. The obvious sequence. Initially a tripod was set in method was to spread the load along place, with a further three rafters set Various means of overcoming this a length of wall rather than allowing into the crosstrees afforded by the tri- problem were considered, especially a pressure focus on the area of the pod. A tripod was utilised because it since the normal system of thatching butt of a rafter – some 0.02 m2. This is the simplest stable structure. Once includes flying rafters to bridge this obvious point was recognised after six rafters were in place the apex of gap. Finally the obvious conclusion the first tripod of rafters caused three the roof was effectively full of timber, was reached from another contempo- points of wall collapse. Thereafter a and a ring beam was inserted to sup- rary experiment examining the crop series of trials was carried out with port the eight subsidiary rafters. This yield of Emmer wheat (Tr. dicoccum). baulks of timber of differing lengths ring beam, of willow withies bound One element of this experiment was being forced against the top of the together to form a circle around the to assess the relative stand heights wall with pressure being exerted at six principal rafters, was set one third of Emmer against the modern hy- 45°. The motive force employed was of the way down the slant height of brid varieties. An average of just over a Land Rover. The objective was to the roof from the apex, as in the 1.10 m was recorded from the crop replicate the 2.80 m wall collapse Pimperne House, to stress the rafters of 1970 after the ears or spikes had of the original excavation. This was outwards. been harvested. When bundles of this achieved with a baulk length of 2.40 straw had been laid flat on top of the m set just 0.30 m from the inside face When all the rafters were in place wil- wall an appreciable eave was created of the wall. This trial was repeated low withies were interwoven between which adequately protected the wall only three times instead of the more them to support the thatch. The next from frost action.

3/2006 euroREA 63 2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses

The system of thatching employed and by the nature of the interweaving hay to eat. This habit necessitated dif- was to cover the roof structure with the opposing tension gives great ri- ficult repair work to the thatch and a thick layer of hay which meshed gidity. This structure with a diameter a protective inner roof skin of hur- into the interwoven willow withies of 5.50 m proved perfectly adequate dles to be inserted. The second habit and formed a support into which the and although it was not complet- developed by the cattle was the rub- straw thatch could be pegged. Over a ed with daub or thatch, the roof was bing of their horns on the daub in- tonne of hay and two tones of straw able to sustain considerably greater side walls, a habit which denuded were used. The total dry weight of weight at 500 kilos per square me- the wickerwork of daub over fifty the roof was in excess of 5.5 tonnes. tre than would have been exerted by percent of the wall area. After five Thus a roof with a free span of over a thatched roof at 50 kilos per square years this structure required com- 6.0 m and weighing almost 6 tonnes metre. Unfortunately this structure plete refurbishment, although the was supported by a dry stone wall. shared the same fate as the Conderton structure itself was perfectly strong. House described above. Phosphate enhancement was iden- Unfortunately long-term study of tifiable but as yet the enhancement this structure was denied because The Moel y Gaer houses on the oth- was minimal. the area was subsequently required er hand have been the subject of for other purposes. The three years long-term study. With virtually the The second Moel y Gaer structure at it stood nonetheless yielded some same base evidence as the Breiddin 7.50 m diameter was the largest stake- valuable observations although the Roundhouse two houses have been built house recovered in the excava- greatest information came from the built. The first with a diameter of tion. This was constructed only in building process. The erosion at the 5.40 m was completed and subse- the timber state without any cladding threshold referred to above (Maiden quently used for over wintering live- whatsoever. The objective was to as- Castle house) occurred quite quickly stock, particularly cattle and sheep, sess how long the timber would re- and clearly explained the paving re- the object was twofold. First to test main viable when totally unprotected. covered in the excavation. Similarly the structure as an agricultural build- In the event the timber became brittle the reverse to the expected drip gul- ing and second to establish whether and useless after only four years. ly occurred with a humic build-up there was a build-up of trace evi- quickly taking place at the base of dence of function by phosphate en- The last stake house examined the stone wall. hancement. A strategy of five years was based upon the excavations at without refurbishment was decided Danebury Hill Fort (Cunliffe 1984). Stake Houses upon with analysis of the structure to The specific structure chosen was take place at this juncture. Analysis CS 20. The stakes were demonstrat- The author has reconstructed of phosphate concentration was set ed to be 30 - 40 mm in diameter four stake houses from respec- for two, four, eight and sixteen years and spaced at intervals averaging tively Breiddin Hill Fort in Wales respectively. The structural analysis 150 mm. The building had a diam- (Musson 1970), Moel y Gaer, Wales after five years showed the building eter of 8.00 m with a doorway 2.00 (Brassil,Guilbert, Livens, stead and to be sound but in need of repair. m wide marked by post-holes and it Bevan-Evans 1982) and Danebury, Cattle proved the most destructive was claimed to be a hut or house. The Hampshire (Cunliffe 1984). Towards on two counts. Most important they evidence was faithfully repeated in the end of the Iron Age post-built preferred to eat the straw from the three dimensions and despite more houses generally gave way to houses thatched roof despite having ad lib than a dozen attempts proved impos- built of closely set stakes. The usual evidence comprises a circular plan of stake holes set some 0.45 m apart with two large post-holes defining the doorway.

The reconstruction programme for the houses set out initially to dis- cover whether such slight timbers could, in fact, sustain the weight of a roof and subsequently what was the greatest diameter such a house could safely reach. The programme developed further to investigate function and function traces.

The Breiddin Roundhouse provided evidence for the stakes to be approximately 75 mm to 80 mm in diameter initially, set 0.45 m apart into a shallow trench with the interwoven willow wands commencing at the base n Fig. 8 The completed Pimperne house: the focal point of the Demonstration of the trench. As stated above the bas- Area, this structutre is the largest reconstruction of a prehistoric roundhouse ketry walls are extremely powerful ever undertaken anywhere (photo Reynolds).

64 euroREA 3/2006 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses ESF workshop 1987 / ITEMS 2

sible to roof let alone apply thatch. In tween structural and constructional all probability it was a compound, evidence from the excavation itself. completely unroofed and may well The greatest difficulty experienced have been used for some animal hus- in the structure was building the bandry purpose. In reality the stakes roof. In all previous reconstructions were simply too slight to sustain the of roundhouses the first phase of lateral weight thrust of a roof. building the roof was the erection of a tripod of rafters which were physi- This last example underlines the val- cally lifted into place on the walls. ue of the empirical approach in that With the scale of this building the it determines the possible and the individual rafters weighed over 170 impossible. Without such testing kilos each and consequently had to misinterpretation can become estab- be raised one at a time. The curved lished and subsequently almost im- scoops referred to above instead of n Fig. 7 The Pimperne house: the six principal rafters in possible to challenge. being rather anomalous features position (in: Reynolds 1982, 184). suddenly became critically impor- The Pimperne House tant. In fact, their location, 1.50 m post-holes show considerable dis- from the wall, with the required turbance, a disturbance which can This is, without doubt, the larg- roof pitch of 45°, allowed the height now be explained as regular post est roundhouse reconstruction un- of the wall to be identified at 1.50 m replacement. These apart, there is dertaken by the author. The de- and the height of the inner ring at no reason why the life expectancy tailed analysis of its reconstruction 3.00 m. Their distribution similarly of such houses should not run into has been published (Reynolds 1982) argued for only six principal rafters hundreds of years rather than dec- with both the excavation report and consequently an hexagonal ring ades. and building study in preparation beam a third the way down the slant (Harding, Blake and Reynolds 1993). height of the roof. This had to be the Similarly the six principal rafters, It is based upon a limited excavation case since all the other rafters were so critical in the preliminary stages of a house plan at Pimperne Down free of the ground and given their of roof construction, have now rot- in Dorset. The plan and photograph weight had to be firmly supported ted at the ground interface. In fact, of the excavated area provide the at wall and inner ring and at ring once the building was completed in primary data. In effect two houses beam height to prevent natural sag. 1977, the ground was dug away from were built utilising the same porch The roundhouse definition of a cone the butts of these rafters in order to n Fig. 6 post-holes for each structure. The set upon a cylinder further served determine if they were still load- Plan of the outer wall comprised a perimeter of to explain why these features were bearing. The fact that they weren‘t, excavated site stake-holes containing an inner ring curved scoops. During the erection further substantiates the claim of at Pimperne of post-holes. The distance between of the main rafters and attachment distinction between constructional Down, Dorset. post-hole ring and stake-hole perim- of the ring beam, the cone of the and structural evidence. The butts of The postholes eter was just 1.50 m. At a distance roof became out of position. By ma- these rafters were left as they were to used in the of 1.50 m beyond the perimeter of noeuvring each of the butts of the test their longevity in an exposed po- reconstruction stake-holes were a number of curved main rafters the cone was recentral- sition. Like the porch posts they were are shown in scoops around the circumference of ised. Final analysis showed remark- badly rotted after a mere eight years heavier outline. the building. Internally there was able correlation between the recon- but only below the limit of the eaves (in: Reynolds evidence for a hearth and a cooking struction evidence and the original of the building. 1979, 96). oven as well as a background scatter archaeological evidence of these of post-holes which seemed to have curved scoops. no structural purpose. The completed structure is quite re- The object of the reconstruction was markable, if only in terms of sheer to build a roundhouse based upon scale and material requirements. one of the plans in order to exam- Over two hundred trees were need- ine the architectural and material reed to build the structure, ten tonnes quirements of such a large structure of daub to plaster the walls and five and thereafter to study the life of the tonnes of thatch to clad the roof. The building for as long as possible. The information yield at construction ﬂint spread construction was completed in July was high. hearth 1977 and has continued to yield valuable information since that time. However, subsequent observations Indeed, it is largely based upon the have proved even more interesting. observations of this building that the After eight years the porch posts, learning curve (q.v. below) is prima- the only ones exposed to the weath- rily formed. er, required replacement. The posts had virtually rotted through at the The construction yielded valuable interface between ground surface interpretative evidence in allow- and air. In all the excavated plans ing differentiation to be made be- of houses of this scale these porch 0 5 m

3/2006 euroREA 65 2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses

of the root fibre originally present idated by experiment, whereas the before the house was built. The re- grain drying capability was totally sults of the trials will be published in disproved. due course (Reynolds 1993). In this case the learning curve is Finally after ten years the house is steep at the construction and imme- being coated with an additional lay- diate testing phase. Subsequent ob- er of thatch. Financial restrictions servation has shown that the build- when the house was built meant ing itself while perfectly adequate for that the original thatch was just 100 the trials was found to be incomplete mm thick instead of the more ac- in the sense that erosion occurred in ceptable 300 mm. Even so the dam- the stoke hole area, while none was age has been minimal, mostly oc- present in the original. casioned by birds robbing straw for nest building and occasional storm The Wroxeter House damage. During the rethatching, inspection showed all the timbers Wroxeter Site 68 Phase Z (Barker to be firm and sound, including 1973). This particular reconstruc- the hazel purlins. Once the thatch tion is based upon the excavations n Fig. 10 The house forms the major focus is completed the roof weight will of Building I of the Baths Basilica Wroxeter. Site of the museum area of the Ancient be increased to over ten tonnes of at Wroxeter, the site of the Roman 68, Phase Z (in: Farm and has been used as an educa- straw dry weight. town of Viriconium in Shropshire. Barker 1977). tional tool from its creation. Despite The building was revealed when the the passage of tens of thousands of The particular importance of this plough soil was removed and the up- feet, wear is minimal. Similarly the study is that it is a long-term re- permost archaeological layers were central hearth has been used regu- search programme. The information cleared simply by removing the hu- larly as well as the reconstructed clay yielded to date is remarkable and in mic soil and leaving all stones, tile, cooking oven. To date some ten years future should become invaluable, clay and pottery in situ. The maxi- later there is minimal sooting in the especially as electronic prospection mum depth of the post sockets was roof and all the timbers are firm and techniques continue to improve. In 20-30 mm so that it was clear that stable. effect the house is a simple labora- none of the posts had been sunk into tory, the history of which is fully re- the ground and that they could not Throughout the experimentation corded. Its value as a source of com- therefore have stood independently. with roundhouses the effect of a cen- parative evidence can only increase The nature of the evidence argued trally placed hearth has been tested with time. that the site had not been damaged exhaustively. The simple conclusion by plough action at all and that the is that there need not be chimney Romano-British building was the last structure on provision of any kind. Should there Grain driers this part of the site before abandon- be a hole in the roof the house im- ment. mediately becomes an elementary Reconstruction can be particular- blast furnace. Sparks would be raised ly valuable in testing the hypothesis The object of the reconstruction from the hearth into the thatch and of a function. On a large number was to test the hypothesis that the within hours conflagration would of Romano-British sites distinctive evidence could actually sustain a quickly ensue. subterranean flues have been recov- structure. To that end the archaeol- ered and identified as grain driers ogy was faithfully replicated, the end The testing of the interior of the and ascribed as an agricultural in- product being a bow-sided hog— house is a continuous programme novation necessitated by the British backed building asymmetric about with special attention to changes in climate on one hand and on the the central latitudinal axis. The evi- the soil matrix of the floor. There is other by improved Roman agricul- dence indicated a framed building an undoubted enhancement of the tural technology. The author was ap- but without sill beams, the walls be- magnetic susceptibility of the soil proached to reconstruct two of these ing presumed to be wattle and daub. within the house brought about by so-called grain driers and test their The end product fully justified the the effects of the fire on one hand and function. Two sites were used, one interpretation and, therefore, the re- on the other by the total degradation from Foxhole Farm in Hertfordshire construction. However, the life of (Partridge 1976), the other from the building itself was chequered. Barton Court Farm in Oxfordshire The first phase, where the wall com- (Miles 1984). The reconstructions prised the uprights bearing the roof replicated the evidence from the two of A frames locked into position by sites, with logical extension to include the purlins and braced by the hipped a necessary superstructure (Reynolds gables, suffered the indignity of be- and Langley 1979). Subsequent testing blown down along its length like n Fig. 11 ing proved incontrovertibly that they a house of cards. It was successful- The framework were not grain driers and were in all ly re-erected and thereafter cross- of the Wroxeter probability malting floors. Or at least braced diagonally across each bay. house. the malting floor hypothesis was val- The result of the construction, there-

66 euroREA 3/2006 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses ESF workshop 1987 / ITEMS 2

fore, amplified the above-ground in- It is in this area of more ephemer- terpretation of the structure. As with al structures, structures which re- the Conderton and Breiddin houses, spond to a functional or economic the structure was not studied beyond need, that the implicit difficulty of the initial stage of timber construc- reconstruction is emphasised and tion. where it must always be stressed that validity is the highest objective, Reconstructions actual prehistoric truth is always be- - postscript yond reach. There has been little or no reference In all of the above cases the objec- covering to the vexed problem of solitary tive has been to present the knowl- shelter post-holes, pairs of post-holes, four- edge gained from the process of post structures, lines of post- and reconstruction and analysis of func- stake holes. All of these patterns, if tion. Full publication of all these wooden they form patterns in the opinion of structures in a ‚corpus’ is forthcom- ﬂor? the excavator, should be the subject ing wherein all details will be made of empirical investigation. However, available. At this point it is perhaps with such structures there is a criti- worth reiterating the nature of the ﬂue cal need for any reconstruction to ‚learning curve‘. Most information, vent. be closely allied to function and, though not necessarily the most therefore, potential function traces valuable information, is achieved which may on the one hand confirm in the construction and immediate the interpretation with the available post-construction phase. Critically, oven data or alternatively provide com- there are secondary and tertiary parative evidence which is yet to be peaks which occur when the struc- recognised and isolated in the ar- ture reaches a stage when repair or chaeology. refurbishment becomes necessary. Indeed, pressure is often brought to n Fig. 9 The experience of the above suggests bear by interested but uninformed Foxholes farm - One particular example, currently that these peaks occur after eight bodies to maintain the newness be- exploded view the subject of a research programme, to ten and twenty years from con- cause it is more impressive. Such of a suggested is the explanation of the ubiquitous struction. Thus given the remit of views need to be challenged on the reconstruction two-post structure. There are a con- the programme, there are clear cut- grounds of the further critical in- of a corndryer siderable number of possible expla- off points. The original building will formation to be gained through 5 (in: Partridge nations but given the requirement yield a dramatic return and may rep- time. In museological and educa- 1989). above of confirmatory evidence the resent a completed phase. However, tional terms it is simple to justi- hypothesis raised is that a pair of particularly from the Maiden Castle fy since the primary purpose is to post-holes may well argue for the House and the Pimperne House, present the normal rather than the presence of a leaf fodder drying rack long-term study and observation abnormal. and imply leaf fodder as a comple- bring more significant and telling ment or alternative to hay. The con- results from the secondary peak. In conclusion there seems little point firmatory evidence sought is the Similarly where function is the sub- in building a reconstruction unless waste product of leaf foddering, the ject of analysis, particularly in terms there is a specific question or ques- inedible twigs. In effect the evidence of mutation of soil nature whether tions which require answers. The has to be in carbonised form and one in phosphate concentration or mag- motivation for reconstruction, how- is seeking an analysis of all carbon- netic susceptibility enhancement, ever, need not be limited to this par- ised twig sections down to species. the long-term approach is the only ticular role since there are undenia- Given a persuasively high percentage viable one where the return is stead- ble benefits from subsequently using of elm (Ulmus sp.) and ash (Fraxinus ily reinforced through time. resultant structures in museological sp.) against other species, the hy- and educational fields. However, to pothesis would gain validity. On consideration of the original build only for those latter options is purposes of European building in to deny the opportunity for scientific The purpose, as with all reconstruc- prehistory and Protohistory, all the investigation with its advantages in tions, is to enhance our understand- evidence points to buildings last- elucidating the potential of the ar- ing of the archaeological data but in ing many decades. They are hardly chaeological evidence. Indeed, giv- this respect the danger of isolated as temporary huts and sheds. There is, en the cost of reconstruction, both opposed to integrated reconstruc- therefore, a danger in accepting the in terms of financial and labour in- tion is underlined. Since the pur- principles of the learning curve/investment, it is profligate to ignore pose of archaeology is to understand formation return and accepting the the scientific basis of reconstruction. man in the landscape, the study of a maximum return in the shortest The rewards in terms of knowledge structure is essentially limited the elapse of time. There is a perfectly on one hand and on the other the in- study of a structure within a con- understandable desire to maintain a tegrity of presentation and therefore text, whether that context compris- reconstruction in a primitive state in education far outweights the extra es process or function or product, the mistaken view that this is what it commitment required in planning has to be more valuable. would have looked like. and execution.

3/2006 euroREA 67 2 ITEMS / ESF workshop 1987 The scientiﬁc basis for the reconstruction of prehistoric and protohistoric houses

Cambridge Univ Press. Forschungsansatz, der Fragen nach Reynolds, Peter J. (1967a): Reconstruction den Materialressourcen, nach ihrem of an Iron Age Hut in: Evesham Hist. Soc. Management, nach den statischen Res. Papers, Volume 1, pp 5-10. United Notwendigkeiten der Konstruktion, Kingdom, Iron Age, construction of nach den paläoökologischen buildings Rahmenbedingungen, nach der Reynolds, Peter J. (1967b): Experiments funktionellen Analyse, nach dem in Iron Age agriculture I, in: Transactions strukturell bedingten Verfall im Laufe of the Bristol and Gloucestershire der Zeit und nach der physischen Archaeological Society (Gloucester a.o.) 86, Interpretation der archäologischen pp 60-73. Befunde stellt. Auf der anderen Seite Reynolds P.J. (1969): A brief survey of gibt es die pädagogische Motivation, agricultural implements and practices of the die sowohl museale als auch prehistoric period, Proc Cotteswold Natur erlebnisorientierte Elemente beinhalten Fld Club, 35(3), 217-21. muss. n Fig. 12 Information return / learning curve. Reynolds, P. J. (1978): Butser Ancient Farm Die Forschungsrichtung enthält Research Project - a Unique Experiment eindeutig einen primär pädagogisch- in World Archaeology, in: Report on the museologisch orientierten Fokus, bei dem Bibliography Conference „Les structures d‘habitat a l‘Age die dreidimensionale Struktur als eine Barker, P.A. (1973): Excavations on the Site du Fer en Europe temperci - l‘evolution Art der Demonstration zu verstehen ist, of the Baths Basilica at Wroxeter 1966-73, d‘habitat en Berry“ (Levroux, France) die in vielfältiger Weise genutzt werden DoE, London. Reynolds, Peter J. (1982): New Approaches kann. Der kritische Punkt ist dabei, Brassil, K.S., Guilbert, G.C., Livens, R.G., to Old Problems, in: Jones, M.K., Dimbleby, hier Problembereiche mit Blick auf die Stead,W.H. & Bevan-Evans, M. (1982): G.W. (editors): The Environment of Man Festlegung von Grenzen zu benennen, Rescue excavations at Moel Hiraddug from the Iron Age to the Anglo-Saxon innerhalb derer die Rekonstruktion von [hillfort] between 1960 and 1980, Flintshire Period, BAR. pädagogischem bzw. von Forschungswert Hist Soc J, 30, 13-88. Reynolds, P.J. (1993): Experimental ist und dabei Standards festzulegen, die Bullied, A. and Gray, H. St. G. (1911): reconstruction, in: Harding, D.W., Blake, als Leitlinien gelten können. Die Absicht The Glastonbury Lake Village, Vol 1., I.M., Reynolds, Peter J.: An Iron age der Abhandlung ist es, auf der Basis von Glastonbury Antiquarian Society. settlement in Dorset, excavation and Fallbeispielen der Hausrekonstruktionen Clarke, D. (ed.)(1978): Spatial archaeology, reconstruction, (Edinburgh, University of von Maiden Castle (Dorset), Balksbury Academic Press. Edinburgh, Department of Archaeology (Hampshire), Conderton Camp (Bredon Clark, A. J. (1990): Seeing beneath the Monograph Series 1) Hill, Worcestershire), Pimperne soil. Prospecting methods in archaeology, Reynolds, P. J. (1995): Rural Life and Down (Dorset) und anderen Orten, die wissenschaftliche Basis für eine B.T.Batsford Ltd. Farming, In: GREEN, Miranda J. (editor): Rekonstruktion herauszuarbeiten Cunliffe, B. & et multi alii (1984): Danebury: The Celtic World London, pp 176-209. und zu versuchen, die verschiedenen, an Iron Age hillfort in Hampshire. Volume Reynolds, P.J. (1993): Experimental zum Themenkomplex gehörenden 1, the excavations, 1969 -1978: the site. CBA reconstruction, in: Harding, D.W., Blake, Fragestellungen voneinander abzugrenzen. Res Rep, 52. I.M., Reynolds, Peter J.: An Iron age Guilbert, G. (1981): Double-ring settlement in Dorset, excavation and Fond scientifique des reconstitutions roundhouses, probable and possible, in reconstruction, (Edinburgh, University of de maisons préhistoriques et prehistoric Britain, Proc Prehist Soc, 47, Edinburgh, Department of Archaeology protohistoriques 299-317. Monograph Series 1) Harding D.W., Blake I. and Reynolds P.J. Reynolds, Peter J., Lloyd, R.H. (1967): Les motivations pour les reconstitutions (1993): Excavations at Pimperne Down, Kemerton Camp in Bredon Hill and its sont doubles. D‘abord, il s‘agit de la Dorset, Monograph1, Edinburgh University Villages, in: Evesham Hist. Soc. Res. Papers, recherche, y compris les questions press. 1, pp 79-83. des ressources de matières et de leur Hencken, T. (1939): The excavation of the Reynolds, P.J., Langley, J.K. (1979): Romano- gestion, des aspects de construction, de iron age camp on Bredon Hill. British corn-drying oven: an experiment, l‘environnement, de l‘analyse fonctionnelle, Hodder, I., Isaac, G., Hammond, N. (ed) Archaeol J, 136, 27-42. de la dégradation structurale en fonction (1981): Pattern of the past: studies in honour Sarayader, E. (1983): Bargaining Power, du temps et, du point de vue critique, les of David Clarke, Cambridge Univ Press. Dissimulation, and the Coase Theorem, questions de l‘interprétation physique des Kimes, T., Haselgrove, C., Hodder, I. (1982): Journal of. Institutional and Theoretical découvertes archéologiques. De l‘autre A method for the identification of the Economics, Vol. 139, pp. 599-611. côté, il y a l‘éducation qui doit impliquer location of regional cultural boundaries, J Thomas, N. (2005): Conderton Camp, des éléments de musée ainsi que ceux Anthropol Archaeol, 1, 113-31. Worcestershire: A small middle Iron Age plus récréatifs. La recherche implique Miles, D (ed.) (1984): Archaeology at Barton hillfort on Bredon Hill, CBA Research sûrement un potentiel considérable Court Farm Abingdon, Oxon, CBA Res Report 143. éducatif et muséologique, puisqu‘une structure à trois dimensions constitue un Rep, 50. Wainwright G.J. and Davies, S.M. (1995): échantillon démonstratif qui se prête à Musson, C. (1970): House-plans and Balksbury Camp, Hampshire: Excavations être utilisé de plusieurs manières. Le point prehistory, Current Archaeol, 2, 267-77. 1973 and 1981, English Heritage critique, c‘est la définition des problèmes Partridge, C. (1976): Lifting a Roman corn- Archaeology Report 4. de reconstitution pour pouvoir délimiter drying oven, Mus J, 76, 69-70. Wheeler, R.E.M (1943): Maiden Castle, un cadre où les reconstitutions assument Percival, A. (1979): Understanding Dorset, Soc. of Ant. Res.Report no.12, une valeur scientifique/éducative ainsi our surroundings: a manual of urban London. qu‘une série de standards qui puisse interpretation, Civic Trust. Summary servir de guide. Cette conférence a pour Pryor, F. (1980): Excavation [of 2nd objectif d‘examiner le fond scientifique millennium BC settlements and enclosures] Zur wissenschaftlichen Basis des reconstitutions et de formuler une at Fengate, Peterborough, England: the third définition du rayon pertinent de questions, report, Northamptonshire Archaeol Soc für die Rekonstruktion ur- und frühgeschichtlicher Häuser tout en s‘appuyant sur des exemples de Monogr, 1 reconstitutions de Maiden Castle (Dorset), Renfrew, C., Shennan, S. (ed.) (1982): Es gibt eine Dichotomie der Motivation Balksbury (Hampshire), Conderton Camp Ranking, resource and exchange; aspects of für die Anlage von Hausrekonstruktionen. (Bredon Hill, Worcestershire), Pimperne the archaeology of early European society, Auf der einen Seite gibt es den Down (Dorset) et d‘autres.

68 euroREA 3/2006