2 ITEMS / ESF workshop 1987 Ancient wood, woodworking and wooden houses Ancient wood, woodworking and wooden houses

The article introduces a record their houses, barns, fences and other wood has been maintained in a good on the management and use of equipment, archaeologists often have condition, archaeologists have an ob- prehistoric woodland gained to rely upon a record that has bare- ligation to extract as much useful in- from the research ly survived time‘s of destruc- formation as possible, as their results of the Levels. tion. The raw material most consist- may be relevant to other archaeolog- ently and abundantly used in ancient ical situations where conditions of n J. M. COLES (UK) times, and in fact into the 20th cen- preservation are not as good. It is, of tury in some areas of the world, has course, important to remember that 1. The not often been preserved for detailed the woods available and used in cer- study, and even where traces survive, tain areas, where they have survived, In efforts to understand how ancient their quality may be unsuitable for may not be exactly comparable to people obtained supplies of wood for scientific research. Where ancient those of other areas where survival is poor, and care is thus needed in ex- trapolation of results. Nevertheless,

llex there are important reasons why ar- Acer Salix Pinus Taxus Alnus Fagus Malus Betula Ulmus chaeological sites and regions con- Sorbus Cornus Hedera Corylus Quercus Clematis Frangula Carpinus Rhamnus

Viburnum taining abundant preserved wood Myrica gale can be used in the construction of a Prunus avium Prunus Prunus spinosa Prunus general assessment of ancient wood- l l l l l l l l l l l l l l l bc Sweet ? ? lands and wood , so long l l c3000 Bisgrove as care is taken in utilising the re- l l Honeygore sults. In this paper, one area where l l Chilton preservation has been exceptionally l l Walton 83 good will be used to show the char- l l l l l l Baker acter and quality of evidence that l l l l Bell A can survive, and this will be used as a l l l Bell B guide to the woods used in the con- l Honeybee struction of ancient buildings. Honeycat l l l Honeydew The Somerset Levels are an area Jones l l l c2500 where conditions for the preserva- Burtle Br. l tion of the environmental and cul- Garvin‘s l l l l tural records allow some degree Walton l l l l l l l l of generalisation about prehistoric Rowland‘s l l l woodlands and woodworking. Due Blakeway l l to extensive formation over Signal Pole l 5,000 years (4,500 BC - AD 500), the Abbot‘s Way l l l l l l l l c2000 peat buried hundreds of ancient East Moors l structures, and because the area is Eclipse l l l l l l l l? c1500 low and flat, silts and clays have also Meare Heath l l l l l l l l l l l l l helped seal many sites from decay Tinney‘s l l l l l l l l l l (Coles and Coles 1986). The advan- Godwin‘s l l l l c1000 tage of a such as the Levels Stileway l l l l is that it contains within its not l only the wooden elements of track- Withy Bed l l l l l l l ways, platforms and other struc- Viper‘s l l l tures, including houses, but also the Nidon‘s l? l record of the woodlands growing in Tollgate l and around the area. Pollen analysis, Platform l l allied to macroscopic identification Skinner‘s l l l l of leaves, seeds, bark and charcoal, Shapwick l l l l l l c500 Meare West l l l l l l l l l l l l and the wood of structures, provide an important record of the chang- l l l l l l l l l bc/ad Difford‘s l l l l l l l l ing use of primary and secondary forests, the development of wood- n Table 1 Species of wood identifications from prehistoric structures in the land management practices, and the Somerset Levels, circa 3200 BC – circa 100 AD. exploitation of preferred species for l - dominant; l - common; l - occasional; l - rare particular purposes. Allied to exper-

50 euroREA 3/2006 Ancient wood, woodworking and wooden houses ESF workshop 1987 / ITEMS 2 imental work in the Levels, we can 2. Classifications Many pieces are found broken and so now make some suggestions about such definitions are often difficult. ancient wood , from (See Rackham 1980) species selection to felling, splitting 3. Felling and and working, and also about the like- Timber is sometimes used to de- ly yield from various kinds of wood- scribe trees large enough to make The evidence surviving from the lands of the timber and roundwood planks and beams. In this case, the Levels allows some observations to needed for structures. word wood is reserved for small- be made about the felling of trees, er trees suitable for poles, rods and and in the use of different axes of Among the many archaeological other small-diameter stems. Most stone, bronze and iron. It is very unu- structures and other discoveries in wood consists of underwood, grown sual to recover any direct evidence of n Fig. 2 the Levels, 36 sites can be extracted by a coppice system, in which a tree the felling operation itself. However, A roundwood which provide a useful guide to the is felled and the stump thus encour- tree-stumps occasionally survive in peg of woods selected over circa 4,500 years aged to send up shoots or stems; in the Levels‘ peats, and show the pre- from the Early (circa 4,000 BC - AD 500) (Table 1). time the stump or stool yields many cise method of felling, e.g. very acute of From these sites we can identify 26 crops of poles and rods, felled at in- chopping, around the whole stem, the Somerset species of wood actually used in tervals dependent on the length and marks the felling operation on yew Levels. Note the the constructions (Orme and Coles diameter of the poles or rods re- circa 1,500 BC. The chopped ends of clear facets left 1985). The age, size and condition quired. Branch wood is mostly un- stems and branches are less reliable by a narrow- of the different species can be meas- suitable for building purposes, and is as they may have been re-worked af- bladed stone ured, and other studies can indicate one of the principal sources of fire- ter felling. Timber burnt at one end . (photo the nature of woodworking wood (see below). may suggest the use of fire to assist J. M. Coles) used and the quality of woodwork- ing. Primary forests, previously un- In a prehistoric woodland where touched by man, are also indicated some form of control was exercised in the record of pollen preserved in there would be both timber, or tim- the peat beneath the earliest struc- ber-bearing, trees and wood, or un- tures. It is apparent that a wide range derwood or roundwood, of more of tall and ancient trees, such as , slender dimensions, all growing to- ash and lime, existed in this ‚wild- gether in a system sometimes de- wood’ for utilisation circa 4,000 BC. scribed as coppice-with-standard. By circa 3,000 BC, managed wood- The underwood in England might be lands containing coppiced hazel had elm, hazel, ash or lime, and timber been long established, and these trees would be oak; conifers would be were still in existence by circa 2,000 represented in Alpine regions, AD. Regenerated forests were heavily was important in parts of northern exploited circa 1,500 BC, but yield- Europe, and cedar was pre-eminent ed timber mostly of inferior size and in northwest coastal America. quality to those of the primary for- ests. After circa 500 BC, oak wood- In the Somerset Levels, we accept lands were diminished and a wider these basic definitions, but for the re- range of trees was being used. cording of pieces of prehistoric wood found in the trackways, platforms and Such differing varieties and qualities collapsed houses we use the terms of wood have important implications timber and roundwood. Timber re- for the character of ancient struc- fers to split wood, as distinct from tures, just as has the development roundwood which is an unsplit stem and use of different tools for work- or branch. The division between the ing of wood. The felling, splitting and two categories may be blurred when working of a variety of woods, and dealing with half or quarter trunks; the use of both roundwood and tim- in the Levels these are described as n Fig.1 ber, is represented in the Levels by split roundwood, and timber is re- Timber (oak 100,000 recorded Neolithic, Bronze served for pieces which have little or planks, boards Age and Iron Age wooden artefacts no outer surface remaining. Timber and slats) from the trackways, platforms, hous- may be further classified on site as of the Early es and dumps so far identified, and planks (length circa 1.0m - 10.0 m Neolithic from by several million pieces of wood ob- or more; width circa 0.15 - 0.50 m the Somerset served over the past 25 years. Some or more; thickness < 0.15 m), boards Levels. Note the of the archaeological evidence for (length circa 0.30 - 1.0 m; width circa tapered edges, primary woodworking on a variety 0.15 - 0.50 m) and slats (length circa flat surfaces, of trees through time will be exam- 0.1 - 1.0 m; width < 0.15 m, thick- perforations ined in the paper, and suggestions ness < 0.05 m) (fig. ).1 Smaller pieces and notches. will be made about the implications of timber are silvers or chips. Beams (photo J. M. for the ‚wildwood‘ through such ac- are heavy pieces of timber of plank Coles) tivities. or board length and > 0.15 m thick.

3/2006 euroREA 51 2 ITEMS / ESF workshop 1987 Ancient wood, woodworking and wooden houses

in felling, but only if, as in the Levels’ evidence for the shape of the pieces from the oak forest of ca evidence, the wood was not being re- used (fig. ).2 Where the blow carried 1,500 BC were distinctly twisted. used from earlier structures. Because through, a facet is left. Experimental trees once felled were most often cut work has shown that the number The timber recovered from the Levels to shape, it is only the rather rough of facets on a piece bears little rela- consists predominantly of planks substructural components that pre- tion to the number of blows used in and stakes, with some Iron Age serve traces of the felling operation. working the wood. In recent experi- beams, and in discussing the con- ments, as many as 50 or more accu- version of trunks the production of The size of trees felled in different pe- rate chops with a stone or bronze axe planks will be considered first. From riods may be compared, taking evi- would produce a sharpened stake the Neolithic, planks up to 5 m long dence mainly from the Early Neolithic but the artefact would bear only 5 are known; these are of ash and lime. and the Bronze Age. The evidence is or 6 identifiable facets or axe-scars, The oak planks occasionally exceed largely derived from estimates of total the others truncated or lost by the 3 m. Bronze Age planks have been tree diameter based on examination sequence of action. This in a way is rarely preserved over 2 m in length of the tree-rings in the cross-section the same as retouch on flint, where but 5 m lengths exist and there are of planks. These indicate that stone many of the first flake scars are lost. records of timber of ca 10m length axes could be used to fell trees at least Particularly useful are occasions (Orme and Coles 1983). as large as those exploited in later pe- where the axe head bit into the wood riods. Neolithic stone axes were ad- and had to be pulled clear, without 6. Wedges and beetles equate for clearing primary and sec- necessarily detaching a chip, leaving ondary deciduous forest, including a mark of the cutting edge which may (Darrah 1982) oak trees up to one metre thick and survive and identify the blade. Clear 350/400 years old. Most of the Bronze marks of this sort are often present The tools needed to convert a fallen Age observed were only 30-50 on Bronze Age wood but not on the tree into timber are wooden wedges cm in diameter, but that is a reflec- Neolithic. And sometimes where the and a mallet or beetle. The process tion of the woodlands in the Levels blade has been damaged in use or in of such timber production is simple a rather than the tools. The gross size casting, it leaves a ‚signature‘ on the and -understood, but in prac- of woods used in the structures gives wood, generally a small ridge or set tice it is often different from its the- a general impression of the nature of of ridges on each facet produced by ory. By driving wedges to exploit the the wood species available, although that blade; the potential correlation basic lines of wood structure, stems tree-ring analyses are also required to of worked pieces is obvious. can be split, and some woods such as obtain a better picture of the quality oak and cedar are thus easily reduced b of trees. 5. Timber (fig. ),3 other woods are more intrac- table. Where knots and branches ex- 4. Roundwood It is likely that the bulk of trees were ist on a stem, the wedge and beetle processed soon after felling and not technique will be less easily applied. Roundwood is an easy term to appre- left to season. The rate at which a tree Our experimental work, however, ciate, as it is the basic state which is en- dries, or seasons, depends partly on has demonstrated the effectiveness countered. For archaeological record- species; of the timber trees discussed of wedge-splitting on a variety of

c ing of roundwood in a structure we use here, alder is relatively quick and a woods in different conditions. our descriptive terms carefully while tree felled for six months is notice- admitting that wood, being a flexible ably different to work to one that has Seasoned oak chips make suitable medium, allows infinite variation in just been cut down. Oak is slow to wedges, but unless they are well-pre- the ways by which it may be worked. dry, and may be worked ‘green’ for served on sites, they may not be dis- For roundwood there are seven obvi- up to two years. Ash is intermediate, tinguishable from ordinary chips de- ous elements to consider; species of perhaps closer to alder than to oak. tached in preparing wood. Wedges wood, whether green or seasoned, its No information is available to us for heavily used in splitting will often d growth rate, the type of , angle of lime owing to its present scarcity, bear clear signs of such use, in com- cut, shape of product and character of and elm of recent years has all been pression, curvature, splitting and facets (Coles and Orme 1985). well-dead, thanks to disease. burring of the head. Due to the great weight of freshly-felled stems, split- Many thousands of pieces of round- Side branches would be trimmed ting would be logically carried out in e wood from the Levels bear clear from the felled trunk before split- the woodlands, and so wedges may axe-marks, or facets, and some have ting, and topwood removed to leave well not be in great abundance on the bole clean. In the finished timber the sites of structures themselves; n Fig. 3 The use of wedge and mallet to split a cedar tree. occasional knots and quirks in the nonetheless, wedges for secondary Although the tools shown are from northwest coastal grain indicate where side branches splitting may well be present. America, the basic principles for working European once grew. An overall impression oak, pine and other large trees are the same. Cedar was of the prehistoric timber from the Mallets are tools requiring more particularly easy to split and the European trees may have Levels suggests that there were few preparation, and represent fine ex- been somewhat more difficult to work. a marker wedge to side branches on the trunks felled for amples of the woodworker‘s craft. define the line of split; b wedges driven to start the split; timber, indicating that the specimens Made from stem and branch inter- c wedges driven in along both sides; d a spreading stick had grown in close competition with section in one piece, or from the pounded on both ends to spread the split; e the process other trees, probably in a forest, al- insertion of a straight handle into a repeated to make planks. (based on Stewart 1984) though a considerable number of heavy chunk, they should survive in

52 euroREA 3/2006 Ancient wood, woodworking and wooden houses ESF workshop 1987 / ITEMS 2

waterlogged conditions. From the of the stem. In this work, it is notice- Somerset Levels, only one Neolithic able that when completed, the faces mallet has been found, but there are of the plank will bear traces of clean several later and smaller specimens; splitting of the wood interrupted by the Neolithic mallet was made of burring of the faces where the wedges yew, one of the toughest native spe- were forced in and often the wedges cies of the British Isles. This mallet will remain stuck between the main showed its use as a heavy blunt in- face and a slender but resistant sliver strument, in the burring down of of wood stubbornly attached to the the inner parts of the rounded ends. wrong side. In such cases the sliver The use of such a tool was either to would be axed free. Such slivers have drive pegs or piles, or to force wedg- been observed on Levels‘ sites. es into tree stems in the early stag- es of timber or plank production. Cross-sections of oak timbers illus- Experimental work shows how eas- trate that in practice there were many ily fresh oak, and ash, can be split variations on the basic theme of ra- with wooden wedges and beetle, and dial and tangential splits (Orme and our work suggests that a heavy oak Coles 1983) (fig. ).5 The Neolithic tree once felled could be reduced to pieces are mostly variations on a radi- 8ths or 16ths within an hour or so, al split, and those which deviate only granted a stem without major imper- slightly probably do so because of the fections. Prehistoric timber from the grain of the wood rather than any de- n Fig. 7 Finished Levels occasionally carries the traces liberate intention of the woodworker. planks of oak of wedge ‚burring‘ (Orme and Coles There is a preponderance of true ra- of the Bronze 1983). Normally however it would be dials. Redirected radials occur where Age from the expected that the further trimming the wedge has been re-positioned in Somerset of split wood, or its use in structures, the course of a split, to change the line Levels. Note the would remove or disguise such pri- slightly. This was usually done for one smooth split mary signs. Nonetheless such traces surface of the plank only and the pur- and axed sur- should always be sought. pose would appear to be to achieve a faces, and the flatter plank than a true radial. It has vertical-sided From the start of our record both ra- been a frequent practice in our exper- hole and notch. dial and tangential planks are found, imental work. There are also some ex- (photo J. M. with a few split diagonal to the rays amples of tangential splits. Coles) and some straight through the mid- dle of the tree (fig. ).4 Different ap- The Bronze Age examples range from detaching timber from trunk, and proaches to splitting the trunk are redirected radials to true tangentials. they include results best described as required for these different types. The pieces intermediate between ra- parallel-ray, diagonal-ray, or tangen- Radials, the most common type of dial and tangential are split across the tial with extra splitting. oak plank, can be produced by split- primary natural planes of weakness in ting a trunk in half and half again and oak, namely the rays, and across the Although there are relatively few then splitting ‚slices‘ off each quarter. growth rings which can also allow easy planks and beams from the period c Tangentials require carefully control- splitting. Theoretically, these exam- 500 BC - AD 500, those that survive led splitting across the natural planes ples are the most difficult to produce, from the Iron Age settlements reach of weakness in oak, and are perhaps yet they are found in both Neolithic 3 m or more in length and often bear easier to produce in ash which has and Bronze Age structures. The true signs of axe or on their surfaces; no rays and splits more readily along tangentials are relatively small tim- they were presumably split radials lat- the growth rings (fig. ).4 A plank bers, and may have been produced by er truncated. Of particular note are the split across the centre of a trunk is ef- secondary splitting of a hefty radial. slats or slender planks interpreted as n Fig. 4 (top) fectively what is left when a tangen- Lime tree-ring samples provide some loom parts, from Glastonbury; these A large tree tial plank has been taken off oppos- information about the working of this were mostly of ash and some were said stem split radi- ing sides of the trunk. species in the Early Neolithic; they to have been adzed. The thinness of ally to make appear to be rather thick tangentials some of these slats, down to 25 mm, planks of maxi- In experimental work, oak can be split in two down the centre, and true suggests a high degree of skill and ex- mum width shown to split easily and cleanly, tangentials have been recorded. perience in working with wedge and equal to the but ash is generally a more difficult axe or adze. Indeed, a few smaller piec- radius of the proposition, particularly if many 7. Secondary splitting es were split to under 5 mm thickness. tree. (bottom) side branches and knots occur. Alder A small tree split is relatively easy to split but birch in Four categories of secondary split- 8. Sapwood tangentially to contrast is more resistant and sappy. ting have been distinguished (fig. ),6 make planks In all these operations it is not force one of which, the radial with extra In the majority of oak timbers stud- of maximum that is required, but the eye of the splitting, may overlap with the redi- ied recently from the Levels, the sap- width equal to worker in turning the wedge to fol- rected radials discussed above. Some wood was trimmed off. The discovery the approxi- low the line, and to bypass or force pieces, however, cannot have been of chips of sapwood indicates that this mate diameter through the more convoluted parts produced in the single operation of stage in the wood-working was some- of the tree.

3/2006 euroREA 53 2 ITEMS / ESF workshop 1987 Ancient wood, woodworking and wooden houses

a times carried out at the place of con- disparity to a minimal and unim- suitable species, can be split radially struction. Sapwood chips have rarely portant level. The conclusion of this to make at least 16 planks, each circa been identified, though, and it is likely work is that the quality of work does 13 m long, circa 40 – 50 cm wide and b that the planks were generally trimmed not seem to depend upon the tool, 12 – 15 cm thick near the (trimmed soon after splitting, before being taken but upon its user, whether stone or off) sapwood. Tree A thus yields 16 × to where the track was to be built. For metal is used, and that a 13 × 0.45 = circa 93 square metres of example, of 115 pieces of timber used can produce comparable effects to planking. Such a tree might also be c to build up a Bronze Age chronology those made slightly more easily with split radially into more than 16 thin eight retained sapwood and the rest a metal tool. planks, given an evenly-grained tree had been trimmed, or originated in and a skilful worker. We will calcu- the inner part of the trunk. Very oc- 10. Building late below on the basis of circa 90 d casionally, axing has been evident square metres of planking from an down the side of a plank. Sapwood Although there are many thousands A tree. The stem could be used as a rots more quickly than hardwood and of planks and posts known from the heavy roundwood post, but branch 0 cm 5 its removal delays decay of the tim- prehistoric structures in the Levels, wood was restricted in useful lengths n Fig. 6 ber. Oak has a marked sapwood zone there is still relatively little evidence to perhaps 50 m in total. Sections of oak and elm has approximately 20-30 sap- of Neolithic and Bronze Age build- timber from wood rings but lime is less clear. It is ings. Ash, oak and lime planks, un- Tree B existed in secondary wood- the Somerset not surprising that even in the Early needed for the contemporary build- land, previously cut-over and opened Levels to show Neolithic, sapwood was almost always ings of the Early Neolithic, came to out, so clear stem length is reduced; if secondary removed from timber; the structures rest in the marsh, and they show that oak, it was 100 years old, with a stem splitting built involved considerable pressure structures of the period could have diameter of 40 - 50 cm and a usable methods. on plank edges, and a sapwood edge contained planks 5 m long, and up stem length of 10 - 12 m. A tree of this Neolithic and would rapidly collapse. to 40 cm wide, split and worked into width can only be split tangentially if Bronze Age. remarkably uniform timber. Bronze relatively wide planks are needed; it a tangential 9. Finishing Age timber includes planks of oak up would yield four planks, each 10 m with extra split; to 10 m long, as well as heavy beams. long, circa 35 – 45 cm wide and cir- b parallel-ray The degree of axe work in the prepa- All of these planks could be perforat- ca 8 cm thick. Tree B thus yields 4 × stake from ration of plank surfaces, subsequent ed to take pegs or posts, or ties. For 10 × 0.4 = 16 square metres of us- tangential split; to splitting, is often difficult to cal- the Iron Age of the Levels there is am- able planking. If radially split, into c radial with culate. A plank split to its requisite ple evidence for buildings at the set- eight planks each only 25 cm wide, extra split; thinness may well be considered tlements of Glastonbury and Meare. the yield is about 20 square metres. d diagonal-ray adequate for the purpose of, e.g., a The advantage of the Somerset Levels Otherwise, the roundwood yield strake. track or house foundation. For wall- is that it shows not only the charac- is 10 m of thick stem, and perhaps ing and flooring however, or for a ter and quality of woodworking of 100m of useful branch roundwood raft or other fitted piece, some fin- the Neolithic, Bronze and Iron Ages, in short lengths. a ishing of the split timber would be but it also demonstrates the nature of necessary. This would involve axing the woodlands from which the raw Tree C was a younger stem, again in the plank along its two faces, trim- material for structures was obtained. secondary woodland, possibly cop- b ming off splinters and other imper- These factors can be used when we piced oak, ash, lime, or other species fections, and creating a straight face turn from the Levels to the wider such as pine; if oak it was 40 - 50 years by eye or line through the careful field of ancient buildings and the old, if pine a younger tree, with a stem c axing of the entire surfaces (fig. 7). character of the wood used in them. diameter of 20 – 30 cm, and a length In experiments, this work on oak We will approach this in two ways, of 3 – 5 m or so. It would be split tan- d planks has been carried out with the first concerning the size of trees, gentially for planks 20 cm wide, or stone axe, flat bronze axe and iron and their yield of timber and round- radially for 10 - 15 cm wide timber, axe. It is important to note that the wood. The second approach will be and the yield was about 4 square me- e finished surfaces of timber worked directed towards particular types of tres per tree. Note that it is the size of with iron and bronze axes were structure, in an attempt to estimate trees that we are dealing with, not the barely distinguishable, and that the the quantity of wood needed in their species. The stem could also serve as a 0 cm 5 stone axe created a rougher sur- manufacture. heavy roundwood post; branch wood n Fig. 5 face with many small rather bruised would be rather weak. Sections of oak splinters still partly attached to If we take three sizes of trees such as timber from the surface; after a few months of are known from the Somerset Levels, As noted, these trees could yield the Somerset weather and/or use, all of these trac- we can calculate their yield of round- not only stems for timber, but also Levels to es would vanish through wear and wood and timber for house posts, branches and topstems for round- show different weathering, leaving a smooth face walls and floors. Tree A, the largest, wood. But the main source of round- splitting hardly distinguishable from that existed in primary or long-estab- wood for posts, pegs and rods is methods. worked by metal axes. The crea- lished woodlands; it had a stem di- likely to have been younger trees of Neolithic and tion of half-lap joints with the same ameter approaching 100 cm, and a fast-growing species such as hazel, Bronze Age. stone axe was equally comparable to clear stem length of 12 – 15 m. before alder, ash and pine. These, as do oth- Radials a and those produced by iron axes, the dif- major branching. If of oak, it could er trees including oak, respond well b; c and d ference being in the time taken to be circa 400 years old, but other spe- to a management system involving tangentials; complete the work. Familiarity with cies such as pine could be less old. A regular clearance down to the stool, e intermediate. the stone axe might reduce the time tree of this width, if of oak or other to encourage new vigorous growth

54 euroREA 3/2006 Ancient wood, woodworking and wooden houses ESF workshop 1987 / ITEMS 2

upwards (fig. ).8 It is not often real- Although our interest here is in the ized that coppiced or pollarded spe- wood for buildings, we should not cies can put on enormous rates of forget the need by ancient farmers growth. An old ash stool, for exam- and others for wood for fences, for ple, if cut back properly will create equipment and above all for fire- in one year 20-30 shoots each 2-3 m wood. Many records exist for the long and 1-2 cm diameter. With nat- historic period which speak of the ural pruning by weather and nutri- sole reason for the abandonment of a a ent supplies, an ash stool 8-10 years settlement as being the exhaustion of old may have about 10-15 shoots left, wood for fires. each 3 m long and 5-8 cm diameter, excluding the topmost very slender 11. Houses shoots; that is 30 m of useful rods. At 15 years, the same species may have Let us now consider the timber re- 8-15 shoots, each 4-6 m long and quirements for several different b 6-10 cm diameter; that is up to 90 m kinds of houses such as are known of thick rods. These figures are based from the prehistoric and early his- on actual observations in a well-wa- toric periods in Europe. First, how- tered position where stools were ever, there is one obvious remark closely-packed, thus forcing the trees to make. It will be obvious that dif- to reach upwards to the light. ferent wood species have different characteristics, whether roundwood For more open positions, on a or timber is created. For example, stream bank for example, coppiced in its ability to bend, and tenden- alder can provide equally rich yields cy to split, break or erode, a pine c of useful roundwood poles and log will differ from an oak log, and rods. At age 3-4 years, a small stool an oak from an ash. Even within a may have 6-10 stems 3-4 m long single species, there can be consid- and 5-6 cm diameter. At 8 years, the erable variation in growth rate and stool may have 8 stems 4-5 m long, therefore density and strength. For 6-8 cm diameter, as well as sever- example, an oak slow-grown in a al smaller shoots. At 10-12 years, a dense forest may achieve a diameter stool may carry 12 stems 4 m long of 35 cm in 100 years; as a coppiced and 10-12 cm diameter, plus useful oak, in a favoured environment, it d smaller shoots as well as the thinner could reach the same diameter in upper parts of the main stems; the 35 years. But although of the same yield is 48 m of heavy rods. At 15-20 size, its strength will be different, as 100-year old B trees, radially split into n Fig. 9 (top) years, an alder stool may carry from a slow-grown more densely-ringed 1/8, 1/12 or 1/16s. Structures built 15 a Framework 6-20 stems, 6-10 m long, with diam- tree will be in general stronger than - 20 years later used a mix of B and of a Tsimshian eters 15-20 cm; in addition, branch a fast-grown less densely-ringed tree. C trees of 50 - 80 years age, many of house of the wood of exposed trees will also have But there can be exceptions to this in smaller diameter; these were split northwest developed from the larger shoots, terms of, e.g. the degree of bending into 1/2, 1/4 or 1/8s. Thirty years lat- coast of providing slender rods. The yield which a piece will sustain, and this er a further group of even younger C America, with from one such tree could be 20 m alone is often the most cru- trees were used, from a regenerated (below) the of heavy rods, more like small posts, cial in selection of wood for struc- woodland started at the time of fell- horizontal and 40 m of slender rods; from an- tures. In addition, different parts of ing of the first episodes; these trees wall planks other tree of the same age, up to the same tree will yield wood of dif- were only 40 years old and of small fitted into the 160m of rods of 5-15 cm diameter; ferent densities, and hence its abil- diameter, split into 1/2 or 1/4s. The grooved wall from a third, 60 m of rods of uni- ity to be worked and to serve as a final episode of building, some 15 posts. centre. form 15 cm diameter. load-bearing element; branch wood and 30 years later, used B trees from Framework of tends to be more dense than stem an old stand of oak, over 100 years a Coast Salish The archaeological record from wood, and rootwood the least dense, old, and the wood was split into 1/12 shedroof the Somerset Levels and elsewhere although there are exceptions. or 1/16s. This final wood may have house, with speaks of managed woodlands from been standing in a ‚standard- width (below) the the Early Neolithic onwards, and of The age of the wood used is an im- coppice‘ woodland but is believed to wall planks set the fact that many trees were cop- portant element in any consideration be a different old woodland than the between pairs piced. It is not only ash, alder and of ancient buildings and modern re- first which had been felled 80 years of poles and hazel that were controlled in this constructions. At the Neolithic set- previously. tied in place. way; oak was too. The point to make tlement of Hornstadt-Hörnle, exten- (after Stewart is that managed woodlands could sive tree-ring analyses have shown There are, broadly speaking, two dif- 1984) and did yield vast quantities of use- graphically the variety in age and size ferent ways of using wood in the ful wood for all the structures that of oak timber used in the house con- construction of ancient houses. The have survived into the archaeologi- structions over about 80 years (3586 first is by utilising the properties of cal record, and for most of those that - 3507 BC) (Billamboz 1985). The oak and other trees of large diameter did not as well. earliest structures were made with to make planks. The other is to use

3/2006 euroREA 55 2 ITEMS / ESF workshop 1987 Ancient wood, woodworking and wooden houses

a the round stems, or large branches, tial house. Another type of house was cle, that is, 40 - 50 m of rods. Thus a direct without splitting. In both cas- also of timber, and one described by panel is the product of 6 - 7 stools, es, wall infilling may be by split tim- a European explorer was 200 m long and about one hour‘s work (Coles and ber or, more often, by woven and 20m wide, with a front wall 6 m Darrah 1977). House 4 had 70 stake and daub. We will look now at sev- high and slanted roof (fig. );9 it was or sail holes, arranged to suggest 11 eral different kinds of houses to es- built of cedar planks (6500 square panels each about 2 m wide; at 1.5 m timate their requirements for timber metres) and heavy support planks high, that is 600 rods and sails, 1200 m or roundwood; most of these house and posts set along the walls as well of wood and 25 - 40 stools cleared, as forms are based on actual excavat- as in two rows along the interior. This well as the 1-2 B trees felled and split. ed examples, and we concentrate on two-aisled ‘‘ provided shel- A standard-with-coppice system in an floors and walls rather than the more ter for many families; its construc- area of woodland of under l00 square problematic roofs. tion again was no drain on the forests, metres would hold all these trees. as one cedar tree might yield almost House 1 is 40 m long and 10 m wide 1000 square metres of planking. House 1, a long house, with wattle b with plank walls 2 m high, and it has Planks 12 m long and one metre wide and daub walling rather than plank a plank floor. Thus, 600 square me- are recorded, and their production in- walling would need 100 m of panels, tres of planking is required. That is volved only mallets and wedges. say 35 panels from 210 – 225 stools. 6-7 A trees at 16 radials per tree, or With timber panel-supports and oth- 30 B trees. The timber and poles for House 3 is smaller, 10 × 6 m, with er structural pieces, 2 - 3 A trees or the roof of this house might require front and back walls 2 m high, and 10 - 15 B trees would also be needed. 4 more A trees or 20 B trees. 4 m at the ends, and a plank floor; this A planked floor would also put de- is like an Anglo-Saxon house in some mands on timber production. There seems little point in estimat- models. The house would require 2 A ing the time needed to fell and split trees or 44 B trees or 45 C trees. Many houses of the prehistoric and this wood, as efficiency of ancient early historic periods in Europe tools and men is unknown, but ex- House 4 is an Iron Age round house are not plank built, but rather are perimental work suggests an A tree and is based on a well-documented formed by the skilful relationship of might take one hour to fell and one house from the Glastonbury lake vil- vertical and horizontal roundwood hour to split into 16 radials. A B tree lage (fig. 10). It is 6 m diameter with posts and poles, and this has impor- might take 20 minutes to fell, and a planked floor area of 28 square me- tant implications for the character one hour to complete 5 tangential tres The floor would require 2 B or of ancient woodlands. These houses splits, or 20 minutes to make 8 radi- 10 C trees which were growing near often have walls of , al splits. This is 20 man-hours on A the site circa 100 BC; A trees were and others have infilled horizontal trees, or 120 man-hours on B trees, not available. The plans of house planks set between the major up- for this house. 74 at Glastonbury show the ancient rights. Sometimes the wall construc- planks clearly, and they fall into three tion is not uniform, and a major part n Fig. 8 Sketch House 2 is a long house, 60 m long distinct groups: 14 planks 0.5 - 1.5 m of the house may be of post and wat- of a coppice and 10m wide with plank walls 2.5 m long, 17 planks 3.0 - 4.5 m long, and tle construction, with an end portion and pollard high and an earth floor, like a Viking 12 planks 5.0 - 6.5 m long. They av- of plank construction. Houses of the system. Top, a Age house. That is 350 square metres erage 10cm wide. This is 1-2 B trees, Bandceramic tradition are some- hazel coppiced of timber from 4-5 A trees or 17 B yielding about 30 6 m planks, one set times of this type. House 5, based stool with trees or 100 C trees, or any suitable cut into unequal segments. In addi- on this model, is 30 m long and 8 m several years‘ combination. With buttresses (100 tion, the house had wattle and daub wide, and would need 40 wall posts growth of rods; square metres) and upper members walls, and this introduces the sec- 2 m long, and 56 planks each 50 cm middle, clear- (4 A trees or 20 B trees), the totals ond element in house construction wide; thus perhaps 10 B trees and 4 felled of all are approximately 10 A trees or 40 for long, square and round houses of A trees. The internal posts for such rods; bottom, B trees. This is not far off an earlier many periods. houses are of larger diameter than one year‘s estimate made for the Fyrkat Viking the wall posts and this might mean rapid growth of Age house of 50 trees, presumably of The use of woven panels (hurdles, 30 heavy internal posts, 10 of which straight rods. both A and B types. wattles) for house walls, as well as in were 5 - 6 m long from 5 A trees, and The pollard roofing, is well attested and there are 20 of 4 m length from 5 A/B trees. system works As a comparison, the houses of ce- Neolithic, Bronze Age, Iron Age and The major horizontal roofing mem- the same way, dar from the north-west coast of early historic panels surviving in wa- bers would total 90 m, from 6 B trees. but the stem America may be briefly noted here terlogged settlements and trackways The overall sum is thus 9A, 5 A/B, 19 is cut above (Stewart 1984). Houses were regu- (fig.11). A convenient size is a pan- B trees. The wall infilling is of course the height of larly 30 m long and 20 m wide, with el 3 m wide and 1.5 m high, and this extra, about 5,000 m of rods from grazing cattle corner posts 6 m high; most of the requires about 100 rods 3 m long, 250 coppiced stools; this all could and deer. framework was roundwood cedar and 10 sails 2 m long; that is 320 m have been growing in about 1,000 stems of a diameter of 1.5m. Some of rods. The rods might be of hazel, square metres of managed woodland of the house posts were larger still. aged 4-9 years and 20 - 30 mm diam- with both standard trees and cop- These houses were plank walled and eter. They were obtained from hazel piced hazel or other species. roofed. Yet the native cedar trees or other species, grown from stools available were so large that only in a coppiced system of woodland House 6 is a long house of the early his- two or three could furnish enough management. Each stool could yield toric period in northern Europe and posts and planking for this substan- 15 - 20 rods 2 - 3 m long in every cy- might be 35 m long, 5 m wide; such a

56 euroREA 3/2006 Ancient wood, woodworking and wooden houses ESF workshop 1987 / ITEMS 2 house will have 90 wall posts of 15 cm References diameter and 2 m long, and 30 internal Mound 74 Billamboz, A. 1985: Premieres posts for the roof of 20-25 cm diameter investigations archeodendrologiques dans and 3 - 4 m long. This is about 300 m of le champ de pieux de la station Hornstaad- B tree 30 stems, say B trees. The panel Hornle I sur les bords du lac de Constance. infilling for 80 m would require almost Berichte zu Ufer- und Moorsiedlungen Entrance 10,000 m (about one tonne) of cop- Sudwestdeutschlands 2, 125-147. Floor piced rods and sails, from 500 stools. Coles, B. and J. 1986: Sweet Track to Glastonbury. Thames and Hudson, Extent of Floor House 7 is a log and plank built London. cabin from the Iron Age settlement Coles, J. M. and Orme, B. J. 1985: of Biskupin. There were over 100 Prehistoric woodworking from the Wall Post Somerset Levels:3. Roundwood. Somerset of these structures at the site. Each Levels Papers 11, 25-50. required about 10 B trees and 35 Coles, J. M. and Darrah, R. J. 1977: C trees, much of it from managed Experimental investigations in hurdle- West East woodland. The whole site, with its making. Somerset Levels Papers 3, 32-38. palisade, roadways, gateways and Darrah, R. 1982: Working unseasoned outworks would have required a very oak. In S McGrail (ed) Woodworking, large amount of wood from a wide techniques before AD 1500, 219-229. Peat Peat British Archaeological Reports S129. area of organised forest. In compar- Topsoil Brushwood ison, several medieval or later tim- Orme, B.J. and Coles, J.M. 1983: Prehistoric woodworking from the Somerset Levels. 0 m 3 ber built houses have been assessed ClayFloors Timber Timber. Somerset Levels Papers 9, 19-43. for their requirements. A house Orme, B.J. and Coles, J.M. 1985: Prehistoric 12 × 5 m would have needed about woodworking from the Somerset Levels: boten, nicht zu vergessen dabei der Bedarf n Fig. 10 80 C trees, and one 20 × 6 m would 2. Species selection and prehistoric der frühen Bauern an hölzernem Material Plan and require over 300 trees, mostly of C woodlands. Somerset Levels Papers 11, für Zäune, Geräte und vor allem auch section of type. A cathedral with a wide roof 7-24. Brennholz. an Iron Age Rackham, O. 1980: Ancient Woodland. Its span would involve A or A/B trees; house from the Norwich Cathedral is estimated to history, vegetation and uses in England. Bois préhistoriques, travail du bois et Arnold, London. maisons en bois Glastonbury have required almost 700 A/B trees. lake village, But an ordinary large three-decker Stewart, H. 1984:. Cedar. Tree of life to Des tourbières, telle Somerset Levels, cachent non seulement des restes de Somerset. ship of the line in the 17th century the northwest Coast Indians Douglas and McIntyre, Vancouver santiers de caillebotis, de plate-formes de needed 3,500 A/B trees, from about bois et d‘autres structures, y compris des 900 acres of woodland. maisons, mais encore un enregistrement du Summary développement des forêts dans la région, Some may question the precise quan- elles rendent des témoignages importants Altes Holz, Holzarbeiten und hölzerne tities of wood for these houses, and I de changements d‘exploitation des forêts Häuser primaires et secondaires, de l‘évolution de have paid little attention to roofing Feuchtgebiete wie die Somerset Levels la gestion des forêts et d‘une exploitation roundwood, but it is surely of impor- bewahren in ihren Mooren nicht nur sélective de certaines essences. tance that the large amount of straight die hölzernen Elemente von Wegen, stems, fine timber, and the enormous Plattformen oder anderen Strukturen n Fig. 11 quantities of rods and slender poles wie Häusern, sondern sind gleichzeitig Neolithic signify the existence of well-man- auch eine Quelle für die Waldbereiche, hurdles from aged woodlands with stands of cop- die in der näheren und ferneren the Somerset Umgebung existiert haben. Sie stellen piced trees and tall standards. Such damit ein wichtiges Archiv der im Laufe Levels. These woodlands were in existence at least der ur- und frühgeschichtlichen Zeit were laid on 6,000 years ago, and probably much wechselnden Nutzung von Primär- und the marsh to earlier. Man‘s reliance on wood for Sekundärwäldern, zur Entwicklung der form a footpath more than houses must mean that damaligen Waldnutzungsmethoden but are care and attention was given to the und zur Gewinnung von bevorzugten otherwise of Holzarten für spezielle Bedürfnisse des the type used control and management of the trees Menschen dar. in his territory. Cyclical felling, and Des documentations archéologiques extensively planting of whips and shoots, must Die archäologische Überlieferung der établies à Somerset Levels et sur d‘autres in wattle and have been a feature of life, and the Somerset Levels und anderer Feuchtgebiete sites révèlent la gestion des forêts dès le daub house zeigt die Existenz von systematisch début du Néolithique, un rajeunissement woodlands thus in a sense owned genutzten und gepflegten Wäldern de nombreux arbres. Il ne s‘agit pas construction. and controlled. The very great value seit dem frühen Neolithikum an, viele seulement du frêne, aulne ou noisetier, (photo J. M. of A trees, and the need to ensure a Bäume wurden dabei einer Schneitelung le chêne a été soumis, lui aussi, à cette Coles) good supply of B and C trees, as well unterzogen. Es handelt sich dabei nicht gestion. Il faut remarquer que les forêts as the continuing maintenance of the nur um Esche, Erle und Hasel, die auf gérées étaient en mesure de produire et, coppice cycles, all suggest an interest diese Weise behandelt wurden, sondern en fait, elles produisaient une quantité importante de bois d‘œuvre; il ne faut on the part of early communities in eine systematische Nutzung lässt sich auch für die Eiche feststellen. Es ist pas oublier que les agriculteurs/éleveurs their woodland resources that was dabei hervorzuheben, dass entsprechend préhistoriques en avaient besoin beaucoup more organised than the term ‚wild- kontrollierte Waldbereiche eine große pour pouvoir faire leurs habitations, des wood’ implies. It was wildwood in Menge an Nutzholz für alle möglichen clôtures, du mobilier et surtout pour part, but domesticated as well. Zwecke bieten konnten und sicher auch chauffer.

3/2006 euroREA 57