The Woollen Industry in the North-West Roman Provinces The character of the woollen industry in the western Roman provinces – as in the eastern – was determined primarily by the region’s geography and natural resources and secondarily by how man had responded to, and adapted, those factors for his own purposes. The western Roman textile industry consequently has more in common with what preceded it in the late Iron Age and followed it in the early Middle Ages than it has with the contemporary industry of the Roman East. For our knowledge of the woollen industry in the north-west Roman provinces we can draw upon information from a number of sources; the surviving wool textiles themselves and textile implements, occasional references in Roman literature, and representations of clothing, textile crafts and craftspeople in Roman art. Wool was the principal fibre used in north-west Europe, and that impression is strengthened by an accident of survival: the unusual wet anaerobic conditions on a few sites like the riverside areas at Mainz allow wool to be preserved but not the cellulosic fibres like flax and hemp, which are therefore under-represented in the archaeological record. 1. Sheep, fleece and raw wool Caesar frequently refers to flocks of sheep in northern Gaul, and some years later the geographer Strabo comments that the Belgae of northern Gaul kept hairy sheep, but that the Romans reared ‘jacketed’ sheep with good quality wool. By the late third century (and probably earlier) wool textiles produced by the Ambiani, Atrebates, Nervii and Treveri had considerable economic importance in the Roman Empire. The survival of wool textiles enabled Michael Ryder to pioneer the archaeological study of Roman wool fleece types. By sampling wool yarn and measuring the diameter of 100 fibres in each sample he was able to reconstruct the character of the fleece from which the yarn was spun and enable us visualise the kinds of sheep which populated the north-west provinces. While the fleece categories which he defined have stood the test of time, we now recognise that some selection, by hand, of the finer fibres in a fleece may distort the picture. The range of fleece types evident from the wool textiles is confirmed by another source, breeds of sheep which are feral or primitive domesticates. The small brown Soay sheep of St Kilda in the Outer Hebrides (North-West Scotland) which have roamed free since at least Viking times are a ‘type fossil’ for the sheep of Bronze Age and Iron-Age Europe, while the slightly less primitive sheep of Orkney (North-West Scotland) resemble those of the everyday provincial Roman flock. Fleece types within the Orkney flock vary, as can be seen in the wool yarns sampled from the Roman fort of Vindolanda near Hadrian’s Wall and from rubbish dumps near the legionary fortress at Mainz. The garrison’s clothing at Vindolanda was woven predominantly from hairy medium (34%) and generalised medium (51%) wool fleeces, although fine (9%), medium (2%) and semi-fine (4%) wools were also present. At Mainz more fine (17%) and semi-fine (13%) fleece types are represented, probably because more fine-quality textiles are preserved there than at Vindolanda. The spectrum of colours in the Orkney breed has developed from the predominantly brown of the Soay to black and dark brown, moorit (red-brown), grey and white (the grey in fact is a mixture of black and white fibres). Some 40% of the wool at Vindolanda was white, and that highlights the gradual loss of pigment in the fibres which is a result of selective breeding by man for an improved, ie finer, more uniform, fleece. White fibres can also be dyed more easily, and that opens up new perspectives for clothing decoration. Nonetheless, natural colours of wool chosen to contrast with one another in the same textile were also used for decorative effect. The leading farmers of Italy were aware of the principles of selective breeding. Columella records that his uncle in Spain crossed African wild rams with his own fine-wool sheep, and crossed the progeny again with fine-wool ewes, which gave birth to fine-wool lambs with traces of the reddish-brown colouring of their forbears. It is probable that fleece improvement was also practised in northern Europe in the Iron Age and Roman period. The key to success was the ability to control mating, and that implies segregated, not entirely free-range, flocks. It is likely that some enterprising farmers in Britain and Gaul introduced new breeds of sheep from the Mediterranean, where animals (under the name ‘Tarentine’) were covered in jackets of skin to protect their wool. ‘Jacketed’ sheep, however, were recognised as a rich farmer’s interest. Skeletal evidence from archaeological sites does not unfortunately shed much light on wool character: there is a hint of increasing body size with time and hence a greater fleece weight ( perhaps above 1.5kg); but there is no direct correlation between skeleton and fleece type. Most peasant families probably kept a few sheep, for wool, milk, cheese and meat. Sheep farming on a larger scale is indicated by documentary sources in parts of Lower Germany and Gallia Belgica. For Britain the distribution of finds of woolcombs suggest that the heathlands of East Anglia and the chalk downlands of southern England were wool-producing areas. Within the Empire long-distance trade in wool is attested, but mainly for the most famous fine wools. Nonetheless wool fleeces were delivered to the fort at Vindolanda, and there must have been seasonal wool sales at local markets. 2. Wool harvesting and preparation A sheep in the wild sheds its winter coat in the spring, and grows a new one for the autumn. Early farmers took advantage of this natural moult to pluck the wool from the sheep before it was shed, and this practice continued in the Roman period, as the writers Pliny and Varro confirm. Plucking has the advantage of taking those fibres which are already loose, the finer fibres of the undercoat, while leaving behind the coarser kemp and hair which falls out later. It is an inefficient method of wool harvesting, however, and the introduction of iron shears enabled the shepherd to remove the whole fleece at once. Roman sheep shears – two opposed cutting blades linked by a U-shaped spring and handle – were much like those still in use today. The wool is removed by undercutting the fleece at the base of the old fibres close to the skin, and the whole fleece is stripped off in one piece for storage and transport. There were professional sheep-shearers in Roman Egypt, but shearing was probably a part-time occupation in the North. Fleeces would be washed to release much of the lanolin and dirt, especially if it were to be dyed. (Most wool was dyed ‘in the fleece’ at this point in the production sequence rather than as a finished textile; but because dyeing is a separate specialised craft, it is discussed here later.) Wool can be spun (the next stage) direct from the fleece – provided the fibres are kept roughly aligned – and this seems to have been normal Roman practice. There is evidence, however, for professional woolcombers, whose job it was to separate by combing the long from the short fibres, each of which had its own uses: short fibres were used for soft woollens and felt, long for finer long-stapled yarns. Flat iron woolcombs – forerunners of the Medieval combs with wooden handles – have been found throughout the north-west provinces, their distribution presumably revealing the foci of the regional woollen industries. Gallic woolcombs are rectangular in outline, with teeth cut or welded into the two short axes; the British examples have a single row of teeth at one end, balanced by two broad prongs at the other which were presumably mounted in a combing-post. The Gallic combing-posts must have been provided with a device to clamp the centre of the comb and hold it in position. A funerary relief from Italy shows a woolcomber at work, drawing the wool fibres through the teeth of a comb projecting from the top of a combing-post. The relief also indicates that the woolcomber formed the combed wool into rolls or rovings, wound into a ball, and that he sold it by weight. 3. Spinning Spinning yarn was the foremost domestic craft carried out by the women of every household – except that the wealthy had ‘spinning girls’ to do the work. The quality of the surviving yarn proves the generally high level of competence. The spinner’s toolkit consisted of spindles and a distaff. The spindle was a short rod of wood or bone, the (usually) lower end of which was wedged into the central perforation of a circular spindle whorl, a symmetrical weight of clay or a cut-down potsherd, to give momentum to the spindle’s rotation. The distaff was essentially another short rod, the lower shaft of which was grasped in the spinner’s left hand: it might be of wood, bone, jet or amber segments. The wool rovings were fastened to the top of the distaff, and groups of fibres drawn off the mass by the spinner’s right thumb and fingers and attached to the tip of the spindle. The latter, when set rotating and allowed to drop, drew out the fibres and twisted them together, giving the nascent yarn strength. To accrete enough yarn to supply the next production stage, weaving, required many woman-hours of spinning; it was frequently carried out at the same time as other less demanding domestic tasks. The direction of the spindle’s rotation was governed by local convention. In the Roman North-West it was rotated clockwise, giving the yarn a Z-twist – the lie of the fibre corresponding to the central stroke of the letter Z. No two spinners have exactly the same style, and by applying to yarn the new techniques of digital image analysis, it is now possible to identify the work of individual spinners. There are exceptions to the Z-spin convention. When complex diamond twill was woven, warp was spun in Z-direction, weft in S-direction: this helps the yarns to bond effectively. Even the plainest cloth woven from warp and weft which both contain groups of Z- and S-spun yarns showed a pattern created by light reflected differentially by the opposing spin directions. 4. Looms and weaving The archetypal loom of the Iron Age and the north-west provinces was the vertical warp-weighted loom. It consisted of two timber uprights, an upper cross-beam between the two from which the warp threads were suspended and a row of loomweights across the bottom attached to the warp to keep it under tension. Not strictly vertical, it leaned against a wall. Half of the warp sheet hung straight down from the cross-beam, but the other half lay parallel to the slanting uprights because it was passed over a fixed shed rod which spanned the uprights. The gap between the two warp sheets formed the first shed – through which the weaver passed the weft yarn horizontally, the first act of weaving. A second shed was formed when the rear warp threads were drawn forward to a position in front of the front warp sheet by means of a row of string loops (heddle loops) attached to a movable heddle-rod that rested on brackets, one projecting from each upright. The weft was passed back through this ‘artificial’ shed, and the movable warp sheet returned to its vertical position at the back, and the first shed re-opened so that the whole process could begin again. The weft was beaten upwards to make a firm fabric with the aid of a wooden weaving sword, like a large paper knife, thrust into the shed. No Roman loom or recognisable loom parts have survived as archaeological artefacts; but finds of sets of terracotta loomweights prove the use of the warp-weighted loom, as do the preserved wool textiles with a distinctive flat-woven ‘starting border’, woven in a preliminary operation and secured along the upper cross-beam on the loom to anchor the main warp. We have some idea of how weavers on this type of loom operated because, by a minor miracle, the warp-weighted loom has continued in use in remote parts of Norway. Twill weave, a structure which results in oblique lines or even diamonds visible across the cloth face was traditionally associated with the warp-weighted loom in the North. It was a very popular weave in the Iron Age and early Roman period, quite familiar to the Roman army on the northern frontiers who depended on it for many of their outer garments. Thanks to influences from the Mediterranean a new loom, the two-beam vertical loom, appeared in the North in the course of the Roman period. It was a free-standing loom, and the loomweights were replaced by a lower cross-beam linking the uprights. The warp was spanned between the upper and lower cross-beams; the weaver sat at her work (as opposed to standing – which the old warp-weighted loom had required) and used a comb to beat the weft downwards. The principal role of the two-beam loom was in weaving garments which incorporated in-woven bands or patterned roundels in coloured wool yarns. There are numerous textile fragments from Mainz, probably from tunics, which carried such bands in red or ‘purple’. 5. Cloth finishing Metropolitan Romans were sensitive to the impression which their clothing presented: garments had to be clean, bright and neatly pressed. In the North a more relaxed provincial attitude may have prevailed. Nonetheless, woollen garments straight from the loom were regularly entrusted to a fuller, a craftsman who cleaned and finished new clothing, as well as acting as laundryman for soiled items. In the towns of Italy fullers’ workshops can be recognised by their characteristic fittings: rows of basins against the wall for treading the cloth and shallow tanks of water for rinsing operations. Cloth was pounded underfoot in the basins in a solution of water and fuller’s earth or decayed urine, an effective if smelly cleaning agent. Once rinsed and dried, the garments might be whitened by being spread over a cage containing burning sulphur. They might also be given a special soft feel by being worked over with an aena, a board set with spikes to raise the fibres, the nap. The nap was then trimmed with a pair of giant cropping shears, attested both in funerary art and by actual finds. Many of the surviving textile finds come from clothing which once had a soft nap. Tunics in particular might be woven from a weak-spun yarn which would lend itself readily to being raised with an aena. Worn or discoloured clothing might be entrusted to a colorator who would treat it with special ‘earths’ to disguise the deficiencies. Evidence for fulling in northern Gaul is meagre. No remains of the fixed plant of a fullery have been recognised, but it should be noted that portable wooden tanks could have been used for treading cloth, as seen on a fuller’s tombstone at Sens (Yonne). An inscription from Köln mentions a master fuller, and another from Alzey is a fuller’s dedication to Minerva, patroness of craftsmen. Fulling was both a necessary and a well known occupation in centres of population. 6. Clothing and textiles Roman garments were woven on the loom in one piece (or sometimes two); the weaver knew at the outset what the final product was to be and planned it in detail. A short-sleeved tunic, for example, had a cruciform outline on the loom. When finished, its sides and sleeves had to be sewn up, but it was ready to wear with minimal tailoring. The same principles applied to the weaving of cloaks and even hooded capes. This tradition is in striking contrast to medieval and modern practice where garments are assembled from components cut from a roll of cloth – and it is the roll of cloth which is woven on the loom and traded. Soldiers’ clothing – mostly of wool on the northern frontiers – was distinct from that worn by civilian Gallic males. The main garment was a short full-bodied sleeveless tunic, worn with a girdle, sometimes made from two separate sheets, front and back, sewn together across the shoulders and down the sides. Over it the common soldier wore a sagum, a rectangular cloak fastened on the right shoulder with a brooch: offices often wore a large semi-circular cloak (paludamentum) over their tunic. The Vindolanda tablets list many other garments which were bought by the garrison: hooded capes (paenulae), scarves, leggings and various types of underwear. The commanding officer possessed a chest of formal clothes for dining and hunting. Living accommodation was sparsely furnished. There would be blankets, sleeping mats, mattresses and cushions, and the praetorium had long bright curtains or hangings. Styles changed with time. In the second century soldiers started to wear long-sleeved tunics, the native dress of the Celtic and Germanic auxiliary troops. In the course of the third century clothing, even for soldiers, became more ornate. The simple red or ‘purple’ bands (clavi), one down each side of a tunic, front and back, were augmented by intricately woven roundels on the shoulders and at the knee. Techniques of decorative weaving, once regarded as too expensive or luxurious, became almost everyday, particularly among the officer class. Male civilians along the northern frontiers wore a wide enveloping sleeved or sleeveless tunic (‘Gallic coat’), probably inherited from the Iron Age, which reached to below the knee: it was not girdled or decorated. Over it was worn a hooded cape, of which there were several variant forms. Women, particularly in the first century, continued to wear native costume: a sleeveless tunic with shoulder brooches over a long-sleeved under-tunic, topped by a cloak. Later, they wore a longer version of the man’s Gallic coat. Very rarely is it possible to name the specific garment to which surviving archaeological textile fragments belonged. A wool sock, a shoe insole and the corner of a cloak have been identified at Vindolanda; the numerous pieces of woollen band found at Mainz were probably leg-wrappings, and the scraps of cloth with red bands were probably once tunics. Tombstones give a better impression of well dressed soldiers in uniform or undress uniform. What they do not show is how heavily patched the clothing must once have been, to judge by the patches, rags and bits of sewing among the Mainz textile fragments. 7. Dyeing Dyeing was a branch of empirical chemistry, and in that sense is difficult to position within the chain of textile processes. The art was at its most advanced in the hands of the professional dyers of Syria and Egypt; but at a simple, non-professional, level it was not beyond the capabilities or resources of urban householders and rural peasantry in the north-west provinces. No identifiable plant for dyeing, such as heavy cauldrons and fixed boilers, has yet been found in northern Gaul. There is a reference to a dyer at Planig near Kreuznach, and to a guild of dyers, probably at Trier. Analysis of the dyestuffs still present in extant Roman textiles offers the best route to greater knowledge, and work has begun on the textiles from Mainz. At Vindolanda, from a small sample, there are eight instances of clothing woven from wool dyed with the root of the madder plant, which gives shades of red; a locally occurring bedstraw (galium verum) may also have been used for red. A purple was obtained from a local lichen. The literature indicates that woad would have been used for blue, but there is no sign of it yet at Vindolanda. The most prized ancient dyestuff, purple, was obtained from one of three types of Mediterranean whelk. Tyre is the best known production centre, but purple dyeing based on the whelk was carried out in many places around the Mediterranean coast. While the dyestuff might have been traded in a dehydrated form, most trade would have been in ready dyed wool and yarn: a prominent purple merchant was commemorated at Augsburg. Dyestuff analysis on textiles shows that true purple wool bands were incorporated into some of the expensive clothing worn by the upper classes in northern Gaul in the later Roman period. Most ‘purple’, however, was achieved more cheaply by over-dyeing woad-dyed wool with madder. 8. Industrial organisation and the customer The presence of a standing army along the Rhine frontier provided a major stimulus to the economy of Gaul, not always entirely welcome. Soldiers had to be fed and clothed, and the government recognised its responsibilities in this respect – even if compulsory deductions for the cost of basic clothing were made when the troops were paid. Clothing supplies were secured by army contracts with textile merchants, and sometimes direct with the weavers, who were paid for all – or a proportion – of the costs of labour and materials: clothing levies gradually became a tax in kind. When that failed, Diocletian established strategically placed gynaecea, ‘factories’, for the direct production of army uniforms. Soldiers, however, commonly supplemented official channels by buying clothing privately, as is revealed in the Vindolanda tablets, or by begging from their relatives. A high proportion of the textile assemblage at Vindolanda, and a lesser proportion at Mainz, is woven in diamond twill, a weave popular in north-west Europe in the late Iron Age. The weavers supplying the army were still working in the Iron Age tradition, not that of the Mediterranean for the most part, and all the evidence suggests that military clothing was made and acquired as close to the frontiers as possible. Fabric quality was excellent: soldiers expected, and had the money to buy, the best, as can also be seen in the pottery they used. How the textile industry in the north-west provinces was organised is much debated. Farming and textile production, however, were closely linked. The scenes on the funerary pillar of the Secundinii at Igel near Trier suggests that the large landowners played a significant role in the industry, selling the products of their estate and estate workers – and possibly others – to important customers like the army. The Edict of Diocletian (AD 301) mentions the good-quality (cloaks) saga made by the Ambiani of the Somme Valley, where villa estates were thickly clustered, and the even higher quality capes (birri) of the Nervii, in whose territory far fewer villas are known. But few communities can have escaped the obligation to weave for the army: particularly successful producers like the Nervii and the Ambiani gained an empire-wide reputation. John Peter Wild <j.p.wild.58@cantab.net>