An 11th century linen shirt from Viborg

By Mytte Fentz

translated by Maggie Mulvaney

Drawings: P. Nørbo, Jørgen Kragelund
Photos: Ove Madsen
Summary translation: Peter Crabb


References:

This article appeared as "En hørskjorte fra 1000-årenes Viborg" in KUML 1987; Årbog for Jysk Arkælogisk Selskab

See also:

Af skjortens saga, Mytte Fentz. Skalk 1989, nr.1

An 11th Century Shirt from Viborg Søndersø, Denmark. Article by Mytte Fentz in "Archaeological Textiles in Northern Europe - Report from the 4th NESAT Symposium. Ed: Lise Bender Jørgensen and Elisabeth Munksgaard. Tindens Tand 1992, nr. 5 pp 83-92

Vikingeskjorten fra Viborg, Mytte Fentz. Viborg Stiftsmuseum ISBN: 87-87872-46-6


English Summary from the original article:

An 11th century linen shirt from Viborg

A preliminary report

Excavations in 1981 and 1984/85 by Viborg stiftsmuseum have established that the town of Viborg, situated in the centre of Jutland, already existed in the Viking age, the dating being based upon dendrochronology and coins. The site, on the shore of Søndersø, was inhabited from about 1000 until about 1300, when it was submerged in connection with the building of a royal stronghold (Fig.1). Parts of houses, well constructions and fences as well as other archaeological objects, all in a good state of preservation, proved that different artisans had been working and living here; also found was a big lump of textile material.

Analyses at Viborg Amtskonserveringsanstalt of the material, the different fragments (Fig. 2-3, 5) and the seams (Fig. 7-8) revealed that the find was the greater part of a male shirt made of Z/Z-spun, linen tabby (Fig.6). Based on the interpretation a reconstruction was made (Fig.4): A rather slim-fit poncho, without seams on top of the shoulders, the skirt being open on  both sides; the neck lining is continued into two ribbons for tying; the upper part of the shirt has a band, which is radially stitched to the outer garment. It is presumed that the shirt had long sleeves.

The likely utilization of the cloth length (Fig.10) and the relevant etymology of this piece of male garment are discussed.

Most of our knowledge concerning Viking male costume derives from written documents, depictions and surviving textile fragments from archaeological excavations. Men wore tunics reaching down to about the knees, long trousers and a cloak fastened with a buckle on the right shoulder.

The Viborg garment is the only profane shirt/tunic from the Viking period so far found in Scandinavia or in Europe as a whole. Tracing the history of this type of shirt necessitates seeking its roots among the surviving Mediterranean shirts/tunics, especially from the necropoles at Antinoe and other locations in northern Egypt (Fig.12).

These tunics, the cut of which is Persian-inspired, can be followed throughout the Mediterranean countries in the latter half of the first millennium. The slim-fit poncho cut of the Viborg shirt fits into this line of development, and scattered finds in Europe show that this cut continued to be used after the Renaissance.

The cultivation and use of linen, Linum usitatissimum, in Denmark during the Viking and early medieval ages are discussed. Archaeological finds of tiny linen fragments and of pollen and seed/macrofossils are only sparse, perhaps due to generally bad preservation conditions for vegetable fibers and seeds. Certain finds indicate that linen was cultivated mainly for its valuable oils.

The likely place of manufacture of the shirt is discussed. According to experts, linen material woven to medium quality requires a horizontal loom (Fig.14), which was known in Europe at the time and archaeologically documented from the 10th century in Poland and at Haithabu. The Søndersø finds in general attest to professional artisans and some to cultural contacts with West and Central Europe; it is thus likely that the horizontal loom was known in Viborg and that the shirt was made there, possibly from flax grown elsewhere.


Circumstances of the find

The shirt was found during the Viborg Stiftsmuseum’s excavations 1984/85 at Viborg Søndersø, east of the present-day town centre (fig.1). Plans to build a hotel in the area initiated the excavations, led by Hans Krongaard Kristensen (1).

Remains of houses, wells and fences were preserved, and the objects found indicate that a large number of different artisans had lived and worked here. The dating is based on dendrochronology and coins that show a continuous occupation from around 1000 AD to about 1300 AD; remains from the oldest dated house can thus be dated to 1018. Early in the 14th century the watertable in Viborg Søndersø was raised about 2 metres in connection with the royal building projects at the nearby Borgvold, and this probably meant the end of the Søndersø settlement.

A lump of textiles was placed in a pit that apparently was a post hole. Stratigraphically this pit belongs to ceramic layer I, which is roughly dated to the 11th century; the crossover between ceramic layer I and II falls around the year 1100 AD (2).
Fig. 1. Important archaeological excavations elucidating the topographical development of 11th century Viborg.
1. The Søndersø settlement from the 11th century on presumptive Crown lands. 2. The St. Sct. Pederstræde quarter with early Viking Age farmhouses replaced by streets of houses in the first half of the 11th century. 3. The cathedral area with early settlement located near the eastern part of the Romanesque cathedral.


Conditions of preservation, conservation and documentation

In the damp environment of Søndersø the conditions for preservation were favourable to the organic objects found. Textiles of plant material are usually found only as small fragments, so although the linen shirt is not preserved complete, the find must be considered extraordinary (fig. 2-4).

The preservation of the Viborg shirt can be attributed to several factors. The presence of charcoal and bark containing tannin in the immediate vicinity of the linen textile may have created an ideal pH level between 7-9, and additionally the tannin may have aided preservation. The placement of the textile in the silted-up pit could mean that it has been protected in a clay package with high inner moisture levels and low levels of air - a bad environment for the fungi that are the most effective decomposing agents for plant fibres (3).

Fig. 2. The preserved parts of the shirt.
1) The right shoulder apex with a small part of the right front of the shirt body, a tiny part of the back and the preserved part of the front of the right sleeve. 1a) Presumed front-lining with two selvedges. 2) The left shoulder apex with left part of the neck opening, border and left panel piece; plus part of back and front of the shirt. 3) Left back/sleeve and front/sleeve pieces. 4) Flap, right skirt. 5) Flap, left skirt. 6) Flap, left tail, waist and lower part of shirt body, back. 7) Left front and back of shirt body on each side of the lateral seam. Downwards. Flap, upper part of left skirt. 8) Fragment of sleeve gusset.

At the district conservation facility in Skive, where the lump of textile was brought, a lengthy process now ensued: the object was gently washed in distilled water so that soil and roots were removed, and the separate fragments were placed on a light table. Here they were smoothed out so that the direction of warp and weft at right angles to each other was restored. While the fragments were dried they were held with stainless pins; measuring and photo-documentation was ongoing.

Fig. 3. The parts of the shirt laid together in accordance with their relative positions; the position of the seams is shown with an unbroken line.

The weaving structure can be distinguished in most of the remaining cloth (fig. 5), but generally the threads are frail and deteriorated. In some areas the linen cloth’s even surface is tightly closed and has a “waulked” appearance. This is most pronounced in the middle of the shirt’s front skirt; wearing of the fibres can be noted in other areas, and here the weaving is left “open”. Both the “waulking” and the worn fibres could be interpreted as wear, however the areas don’t correspond with the areas where one would expect the most wear: collar, armholes and the belt area.

The textile’s current grey-black colour must be attributed to the placement of the find in a damp environment surrounded by charcoal. Colour analysis showed that the linen was not dyed originally, and blue pigments observed on the surface of the textile was identified as vivianite (4). After the analysis the textile fragments were conserved with Plexisol B 597 dissolved in ethyl acetate.

Fig.4. Reconstruction of the Viborg shirt.


Material, weaving and seams

Fibre and weaving analysis shows that the Viborg shirt was made of single ply Z/Z-spun linen thread (5), woven in tabby weave with a density of about 20/12 per cm2, medium quality. The warp thread is a little thicker and somewhat more tightly spun than the weft thread; under the microscope the warp thread consists of more fibres per length than the weft (fig. 6). These conditions create a surface structure with warp facing, ie. dominating warp thread, but the weft thread can still be seen. The cloth is woven with great evenness, and the preserved remains of selvedges are quite even, as is today known from machine woven cloths.

Fig. 5. The shirt’s left sleeve gusset.

The sewing thread employed is either single-ply Z-spun linen thread or two-ply Z-spun S-plied linen thread. The sewing is carefully and evenly done, and the preserved seams completely flat. The identification of the different seam types (6) and of selvedges (7) in the separate fragments of the shirt were one of the factors that made it possible to collect the pieces into their correct positions relative to each other. The different seams are constructed either from two selvedges, of one selvedge and one folded over cut edge, or of two or more folded over cut edges.

Diagrammatic representation of plain weave. Trend=warp; skud=weft.
Diagram showing S- and Z-spun yarn. Tvunded=twisted.
Amount of twist on the linen thread in relation to the length unit. Tråd=yarn

Fig. 6.

The shirt’s remaining seams were examined under the microscope; the different seams were drawn up and tried, until a copy was achieved which was identical to the original. Despite the very fragmented condition it was possible to discern eight different seam types that were sewn with different stitches: back stitch, whip stitch, running stitch and “through overcast stitching” (8). Examples of combinations of the seam types and variations of stitching are shown in the reconstruction drawings fig. 7-8).


Pattern: interpretation and reconstruction

The reconstruction of the dimensions and the original pattern of the shirt has naturally caused problems due to the lack of material connection between measuring points and the linen’s greater elasticity in the wet condition which was necessary for the smoothing of the fragments. Their relative position is based on factors such as direction of the cloth, seam profiles and the correspondence between torn edges (fig. 2-3).

The upper part or body of the shirt (c/f fig. 9) is slightly fitted, such that the waist is slightly narrower than the shoulder width. The side seams end level with the waist, while the skirts are open at the sides; however they are attached to each other at the top for a short distance, such that the back skirt covers the front skirt by about 5 cm (fig. 4 and 9). The body of the shirt consists of two layers of single cloth. Outer cloth and lining are fixed to each other with quite small running stitches, that form a square and four diagonal lines on front and back respectively. The neck opening appears almost square, and the chest - the area between the neck edge and the square - consist of two panels of single-layered cloth, the outer one formed of the shirt material, the inner of the lining. A band runs along the edge of the neck opening, continuing in two free ends which are passed through a gliding knot each, which presumably were fastened to the corners of the front. It is probable that the bands ended in a stop knot (fig. 4). The sleeves, of which only the upper parts are preserved, were cut along the warp direction on the cloth and sewn onto the body of the shirt along the weft direction. They are probably sewn of two pieces each, since there were remains of two lengthwise seams (fig. 4). The sleeves narrow towards the wrist, but nothing precise can be said about their original length.

The investigation- and reconstruction processes have thus documented that the find is a garment: a shirt or tunic, woven of linen and sewn poncho-wise, ie. without a shoulder seam, probably with long sleeves and having a lined body (fig. 4).
Fig. 7-8

Seam type VII between sleeve and shirt body. At top left the primary sewing of the left sleeve, then its stitching to the body. The arrows show the position of the seam.
Seam type VIII; needle and thread show through over-cast stitchings. The arrows show the course of the seam.

The loomlength’s width and utilisation

Analysis of the linen cloth from all remaining parts of the shirt shows that it most likely comes from the same loomlength; considerations concerning its width and utilisation is of interest. Literary sources and other extant garments show that cloth was woven in limited lengths and used economically for the garment concerned (fig. 10).

A number of factors are given:

Two factors are unknown:

1. Shoulder apex. 2. Longitudinal sleeve seam. 3. Lower sleeve seam. 4. Sleeve gusset. 5. Sliding band. 6. Panel piece, right. 7. Lateral seam in shirt body. 8. Shirt body, front. 9. Flap, right tail. 10. Neck opening. 11. Panel piece, left. 12. Square. 13. Fixation seam. 14. Waist. 15. Flap, left tail. 16. Skirt.

Fig.9. The shirt terminology employed.

Shirt or tunic?

To interpret the written sources and understand the concept of shirt[skjorte] and tunic[kjortel] it is important first to find the synonymous links between Latin, Norse and modern terms.

The Norwegian philologist Hjalmar Falk has analysed garment terms that appear especially in Norwegian and Icelandic saga literature and in [diplomatarier] (9), but whether the Viking Age used these terms cannot be known. Skyrta/scyrte in the Norse literature is the term for a short garment with long sleeves, which could have slit sides. Ullskyrta and silkiskyrta are mentioned. The word derives from the Germanic skurtion, which means a “short garment” (10).

The words skyrta and kyrtill, modern Danish kjortel [kirtle, tunic], have in the saga literature an almost synonymous meaning - however it appears that kyrtill never was worn directly on the body, but over a skyrta of hor[linen] or hampi[hemp]. The kyrtill’s length hardly differed from the skyrta’s, but even if there were other individual differences between these two closely related garments, it is not apparent in the saga literature. Perhaps the biggest difference was in the material and/or the weave, which the two garments were produced from.

This suggested reconstruction shows the poncho and the other shirt parts folded out on the warp. The placement is dictated by the size of the parts and the respective selvedges.

Fig. 10.

In Danish wills from the later 13th century and in inventories ‘skjorte’ appears only twice and then with a low-German term: “...et unum hæmeth melinus...”; in the 14th century “...meo unum hemmethe...” (11). The word hæmeth is believed to be derived from the same old Germanic root, kamita, as the gallic/Roman word for shirt: camisia. ‘Kjortel’ was used by Nørlund for the longer outer garment which the Viking Age kyrtill developed into. This is denoted in the wills by the Latin tunica.

Since under- and overclothes today are seen as functionally completely separate - and the terms therefore cannot be immediately applied to the 11th century - the Viborg find must because of both its close-fitting cut and material be considered to have functioned as a man’s innermost or in some seasons only garment. We can know nothing of the contemporary name, but today shirt would be the most natural. It must have been a man’s shirt, since women wore floor-length serkr.


The history of the shirt

The Viborg find is the only relatively well-preserved shirt from the 11th century in North- and Middle Europe, and hence it cannot immediately be compared with other contemporary finds. It can however be shown that it belongs to a development which can be followed in the Mediterranean cultural sphere, where the poncho cut and a vertical warp direction in the finished garment are the common criteria. Some scientists feel that the construction of the garment is due only to functional, ie. climatic, factors; thus garment types with different histories of development can be found in the same climatic province (12). Nor can the outer form and decoration on their own lead to a garment type history; on the contrary the garment’s historical development is apparent first and foremost in its cut and construction. Gudmund Hatt showed two basic types of poncho, from which other garments in time have developed: “poncho without shoulder seam” and “two-skin poncho” or “shoulder seam poncho” (13).

The Viking Age and older Middle Ages in North and Middle Europe. During the Viking Age and older Middle Ages men wore a shirt, possibly both an under- and an over-shirt/tunic; additionally trousers that could be tight fitting and reach the ankles, or wide and ending underneath the knee, a kind of plus fours with more or less stuffing; for these stockings or hose were used, possibly with garters. Finally a cloak held together with a clasp on the right shoulder. Material, cut and decoration could vary: wool and linen spun to finer or coarser thread, woven in different weaves and tightness were the most common materials, but silk, which was of course imported, was also used. This knowledge is based mainly on finds of textile fragments, especially from the rich Birka burials (14), the man’s grave in Mammen near Viborg (15) and finds from Hedeby harbour (16) where many garment fragments were preserved because they were used, once discarded, in tarring and caulking of the ships. The Viking Age Nordic representations of people in sculpture, on picture stones and tapestries are not very many, and fewer still are naturalistic. The main part of pictorial art from the older Middle Ages is ecclesiastical in its basic theme, and is most often directly descended from foreign examples, so the question is if the displayed fashion was also international (17). The grave finds mainly represent the upper social classes, which prejudices the critical examination of the sources; the literary sources should similarly be viewed with some reservation: are the pictures European stereotypes, and what time is really reflected in the literary sources? (18).

Antiquity and early Middle Ages around the Mediterranean. Archaeological finds from the near Orient and Mediterranean lands have a great influence on the understanding of our oldest textile history: particularly the many and fairly well preserved costume finds from Palmyra in Syria, one of the important trading posts on the Silk Road, and to a lesser extent from the Roman garrison of DuraEuropos on the Eufrat, and Halabiyah further north. But especially important are the many grave finds from the large necropoliae in Northern Egypt: Saqqara, Akmin and Antinoë. The latter contains amongst others tunics from late Antiquity, Coptic and Islamic times (19).

In Diocletian’s edict from AD 301 we find that the finest linen for making shirts/tunics was woven in Asia Minor and Syria (20). The tunic was used both sleeveless, with short and eventually long sleeves. There are signs that before AD 200 the sleeve was woven separately and sewn on afterwards, but after this the cruciform tunic, made in a single weaving process, became predominant (fig.11) (21). Apart from the classical ancient loom, the upright loom, the vertical, two-beamed loom and a horizontal loom were also used: our knowledge of this is based on pictures and preserved woven material as well as a few written sources. Both Ovid and Servius speak of the upright warp-weighted loom as “the old loom” and the two-beamed, vertical loom as “the new loom” (22). Examinations of Egyptian tunic material shows that there is a definite inter-relationship between the new two-beamed loom and the cruciform tunic with decorative tapestry-areas, woven from sleave to sleave, so that the warp direction in the finished garment went across the garment (fig. 11) (23).

Cruciform tunic of linen in plain weave with tapestry woven decoration in wool and linen. Egypt, Byzantine period, 5th century AD

Fig. 11.

The small, horizontal loom was used primarily for silk weaving, and it is naturally connected with a lengthwise utilisation of the finished material; possibly the approximately 70-90 cm wide bands were sewn together (24). The Gallo-Roman tunic from Matres-de-Veyre, Auvergne in France, dated to the 2nd or 3rd century AD, is woven from a rectangular piece [sic!], its selvedges here form the respective top and bottom edges; two smaller pieces of material form the sleeves (25). This tunic immediately appears like a cruciform tunic, and has also been described as one (26), but more recent examinations have brought about the interpretation mentioned above (27).

Some of the preserved garments from Antinoë from the 6-7th century AD were made from wool or linen in a narrow width; the finished garment is folded across the warp direction like a poncho, and have long inset sleeves (fig. 12) (28).

From the 5th century the cruciform tunic loses its dominant role, apparently due to strong influences from Persian garment traditions, which were however known all over the Roman empire earlier; on the Marcus Aurelius column from the 2nd century AD there are representations both of Roman classical costume and of near-eastern costumes. From Egypt the Persian-inspired tunic/shirt spread to the Mediterranean countries, and this originally Oriental style eventually influenced the other European countries too.

Woven linen tunic/shirt. Weaving width c. 80 cm. Egyptian, Antinoë: Byzantine period, 6th-7th century AD

Fig. 12.

Roman and Germanic Iron Age in Northern Europe. The Greek historian Herodote, who lived in the 5th century BC, writes that trousers were worn by Medes, Persians and Scythians. The Scythians as a nomadic people spread over eastern Europe - and gold finds from Denmark’s oldest Iron-age are probably imports from the Scythian cultural sphere (29). This earlier and more direct contact with the eastern equestrian people is also considered documented during the Swedish Vendel age. The art of horse riding and thereby the trousers may have been brought to the North at the beginning of the Iron Age without transmission through the Roman Empire, and the Vendel horseman is placed in an important “intermediary” role between Germanic Iron Age and Viking Age (30). Around the birth of Christ the Greek Strabon (31) describes the Gallic costume: “...they wear tight trousers. Instead of a tunic they use a slit tunic with sleeves, which reaches as far as the crotch.” This description of the gallic tunic corresponds well with the Thorsbjerg-shirt from the weapon-offering found in a South-Schleswig bog, the only completely preserved shirt from the Roman Iron Age (32). The most recent examinations date the find to the first half of the 3rd century (33) and interpret it as Germanic with strong Roman influences. From the 4th century we have the first written example of the use of the word camisia. The Roman Jerome mentions it in a letter where he even explains its meaning: a tight-fitting linen shirt with long, tight sleeves (34). The equivalent of this was the cruciform tunic, which was wide spread in the Roman cultural sphere, and the Reepsholt-kirtle from east Frisia, dated within the first 300 years AD (35), is an excellent example of this. It is however an exceptional find, and it is uncertain if this type of garment has played a large part in the Nordic men’s costume in Roman Iron Age.

a) the Kragelund Kirtle, c. 13th century or later. b) The Bocksten Man’s kirtle, c. 13th century [mistranslated by the original translator; the original says 1300-tallet=14th century]. c) The Skjoldehamn kirtle, c. mid-14th century. [again mistranslated; should be 15th century. The patterns are after Margareta Nockert 1985.]

Fig. 13. Costume in poncho cut.

From the later part of the Germanic Iron Age there are no finds of shirts/tunics, but other sources show that it remained an important part of the men’s costume: a contemporary statuette of the mounted Frankish emperor Charlemagne shows him wearing a shirt, breeches and a cloak. An equivalent description of the emperor’s costume is found in the historian Einhard’s Gesta Caroli Magni, where he tells how the emperor’s shirt and breeches were made from linen; interestingly enough Einhard uses the word camisia for shirt (36).

Costumes from the high- and late Middle Ages. The poncho cut and vertical warp direction in the finished garments are attributes for both the Persian inspired shirts from the Mediterranean countries and the Viborg shirt, but it’s also an attribute of some of the extant, much later Nordic garments from the high and late Middle Ages (fig. 13).

Equally, we can point to a connection between two Mediaeval linen shirts from southern Europe, ie. the French so-called St. Louis shirt from the 13th century, preserved in the treasury of Nôtre Dame in Paris, and the Italian shirt from the 16th century in the Zadar museum, Yugoslavia. Both meet the criteria mentioned, as do the early[sic! should actually be ‘later’, but the original text has the mistake in it] shirts from Vingåker in Sweden (37).


Flax and its use in the Viking Age and older Middle Ages.

The Danish ancient flax is presumed to have been brought from eastern Europe (38). The Latin name of the plant, linum usitatissimum, the very useful, hints at its many possibilities for use; the species appears in two variants: the long stalked spinning flax, var. vulgare, and the short stalked, divided and seed rich oil flax, var. humile. In pre-Roman and older Roman Iron Age oil-flax was grown as a crop plant in Denmark, and finds show that the nutritious seeds must have formed an important part of the diet (39). Seeds and their impressions are found in clay vessels and in the stomach contents of the Tollund and Grauballe men (40).

Some scientists feel that the knowledge in Denmark of flax as a spinning plant is connected to the increasing trade with and therefore influence of the Roman cultural sphere. It is indeed from around the 3rd century AD that the oldest Danish find of linen cloth has survived - from Himlingøje in eastern Zeeland (41).

In later years the total number of Viking Age textiles have increased considerably. Archaeological excavations in S. Onsild, on Fyrkat in Hedeby harbour, the current examination of Hedeby Siedlung 31 and finally a pioneering examination of textile remains from Scandinavian ancient graves should be mentioned as the most important (42).

Bender Jørgensen in her extensive cataloguing of textile finds in graves shows that plainweave fragments can be found in increasing numbers from younger Iron Age to the Viking Age, when the trend stabilises. A comprehensive fibre analysis has not been possible, but Bender Jørgensen opines that about half are linen, and she concludes that the younger Iron Age costume must have differed significantly from the older Iron Age with linen cloth as an important feature of the new fashion (43). Possible sources of error for this observation are both the lack of fibre analysis, and the fact that the percentage of linen in the total quantity of plainweave cloth might not have been constant. The preservation conditions for plant fibre is generally less favourable than for animal fibre in settlements. In graves on the other hand the corrosion of various metal objects can have a preserving effect on textiles. The pre-Christian funerary customs, where costume and grave goods indicated the social position of the dead, additionally causes a social distortion in the picture. Differences between the appearance of grave- and settlement finds therefore do not necessarily reflect the original relationship between plainweave wool and linen (44).

Botanical analysis to show pollen and macro fossils of seed plants contribute to the answer to the question of the extent of flax-growing in the Viking Age and early Middle Ages (45). Macro fossils can however indicate both local production and imports (46), therefore seed analysis should be combined with pollen analysis (47).

a) Vertical loom [Drawing: Jørgen Kragelund.] b) The oldest known representation of a horizontal treadle-loom. 13th century English manuscript.

Fig. 14.

In the immediate surroundings of Hedeby a fairly extensive cultivation of flax for oil production, not spinning flax (48), has been shown, while the excavations from St. Pederstræde in Viborg (49) and Viborg Søndersø (50) don’t distinguish between the two variants of linum usitatissimum.

It has been generally accepted that the Viking age loom was the upright loom with a warp-weighted vertical warp direction (51). Because of this it is of particular interest that among the many wooden remains from Hedeby (52) was a well preserved example of a pulley wheel, which is a necessary part of the treadle loom with horizontally tensioned warp. Similar pulley wheels are known from 10th century Gdansk and Opole (53) and late 12th century Sigtuna and Lödöse (54). In Hedeby warp weights for the upright loom have also been found, so we must assume that both loom types were in use simultaneously (fig. 14).

Trials have shown that the upright loom may be less suitable for the weaving of finer linen, especially since the inelastic warp threads cannot withstand the weight of the warp weights. Material woven on an upright loom is also characterised by an uneven density, while the horizontal loom gives an even, regular result (55). There is clear evidence that the Viking Age pit houses were used as weaving huts, but also that this was not their sole use (56). From the pit house CME Århus Søndervold charred balls of yarn and warp remains have now been safely identified as flax (57), while other charred textile remains are wool. Until now the warp remains EME have been assumed to belonged to an upright loom (58); the flax expert Jette Flemming however states that when wool and linen warp threads are freed from the weight, ie. the fixation, of the loom and return to their original tension, they will wind themselves into bundles; this occurs both for a vertical and a horizontal warp. The bunches formed by the warp threads of the warp remains EME, can therefore not be taken as an absolute indication for this linen fragment having been woven on an upright loom. In analysis of the warp remains EME the plainweave was seen to have a density of around 22/12 per cm2 , like the Viborg shirt, but there are also small remains of both finer and coarser plainweave stuck to the warp remains (59). Warp weights from an upright loom are found in pit house CME, but the question is if the linen plainweave fragments were woven on this loom or rather on a loom with a horizontal warp.

Finds in the burned pit house CME of jewellery, horse gear, craft tools and household goods show both the relative wealth of the occupants and that the house was not exclusively a weaving hut (60).

Neither in St. Pederstræde nor at Viborg Søndersø there were foundations of pit houses or any weaving related tools, apart from spindle weights for hand spinning wool (61). The upright loom’s suitability for finer flax weaving is under renewed trials at the Research centre at Lejre, where the above problem perhaps can be further illuminated (55).


Conclusion

Set in a north-western European context where trading contacts and development went hand in hand, it is hardly unthinkable that the horizontal loom was also known and used in Denmark in the transition period between Viking Age and the early Middle Ages. Thus there is reason to believe that Viborg - if importance and international contacts in the 11th century are proven by historical sources and archaeological finds - had knowledge of this type of loom. Since the Søndersø finds in other areas have shown the existence of professional artisans, it must be considered a possibility that the shirt also was woven and sewn locally - possibly from flax grown elsewhere. The examination of the costume historical placement of the Viborg shirt shows that a widespread tradition and continuity can be discerned from the Mediterranean shirt through the Viborg shirt to some of the high Middle Age kirtles and further to the later shirts from Vingåker in Sweden. This tradition also means that a costume historical placement cannot contribute further to the dating of the shirt. The Viborg shirt is unique as a find and of great importance for understanding the men’s costume of the 11th century, but also provides us with a concrete, material find and equivalent to the uncertain information the pictorial sources usually yield.


Notes:

[1] H. Krongaard Kristensen: Middelalderbyen Viborg (the Medieval town Viborg), 1987. H. Krongaard Kristensen: Vikinge- og middelalderbydelen ved Viborg Søndersø (the Viking and Medieval town area at Viborg Søndersø). In: Bag Moesgaards maske (T. Madsen ed.), Århus 1988, p. 89-94. H. Krongaard Kristensen: Udgravningerne ved Viborg Søndersø (the excavations at Viborg Søndersø). In: Hikuin 14, Viborg 1988, p. 9-22. H. Krongaard Kristensen & J. Vellev: En ikke ringe ære for byen (a not insignificant honour for the town). Skalk 5, 1982, p. 3-9.

[2] H. Krongaard Kristensen assumes that the shirt was put in the supposed post-hole after the post was pulled out. See J. Hjermind: Keramik fra udgravningerne ved Viborg Søndersø 1981-85 (Ceramics from the excavations at Viborg Søndersø). Special 1987. Addendum IV shows field S with the pit SAG in the southern part, where the shirt was placed, and the stratigraphical relation between the pit and the ceramic horizons.

[3] T. Vincents Nissen: Bakterier og svampe (Bacteria and fungi). In: O. Alkærsig et al: Bevaringshåndbogen (preservation handbook). Copenhagen 1986, p. 61-68. Also thanks to Helle Strehle, department of conservation, Moesgård, for information about the conditions of preservation.

[4] Analysis made by textile conservator Penelope Walton, York.

[5] Fibre analysis done at the Viborg regional conservation unit and the National Museum conservation unit both show plant fibres of flax.

[6] ‘seam type’ here denotes the solution which is chosen in each seam to hold two or more pieces of material together.

[7] Selvedge denotes the two side edges of the weave, where the weft turns over the outermost warp threads.

[8] The term ‘through over-cast stitch’ describes the thread’s path through the fabric in the relevant seams. The term is constructed for this shirt-analysis.

[9] H. Falk: Altwestnordische Kleiderkunde. Mit besonderer Berücksichtigung der Terminologie. In: Videnskabs selskapets skrifter II, Hist.-Fil. Klasse no. 3, Kristiania 1919, p. 141-147.

[10] N. Å. Nielsen: Dansk etymologisk ordbog (Danish etymological dictionary), Copenhagen 1982, p. 338: Skjorte.

[11] K. Erslev: Testamenter fra Danmarks Middelalder (Testaments from the Danish Middle ages). Copenhagen 1901, 26,5; 97,1.

[12] G. Hatt: Arktiske Skinddragter I Eurasien og Amerika (arctic leather garments in Eurasia and America). Copenhagen 1914, p. 2, 5 & 58. K. Birket-Smith: Kulturens Veje (the way of culture), Copenhagen 1948, p. 267 f. H. Harald Hansen: Mongol Costumes, Nationalmuseets skrifter (writings of the national museum), etnografisk række III, Copenhagen 1950.

[13] H. Harald Hansen: op. cit. p. 104.

[14] A. Geijer: Birka III. Die Textilfunde aus den Gräbern (the textile finds in the graves). Uppsala 1938.

[15] J.J.A. Worsae: Om Mammenfundet (of the Mammen find). In: Aarbøger for Nordisk Oldkyndighet og Historie (yearbooks for Nordic antiquarian knowledge and history), Copenhagen 1869, p. 203 f. M. Hald: Olddanske textiler (Ancient Danish textiles). Nordiske Fortidsminder V, Copenhagen 1950. E. Munksgaard: Oldtidsdragter (Ancient costumes). Copenhagen 1974. E. Munksgaard: The Embroideries from Bjerringhøj, Mammen. In: Oslo Universitets oldsakssamlings skrifter. Ny rekke nr. 5, Oslo 1984, p. 159-171.

[16] I. Hägg: Die textilfunde aus dem hafen von Haithabu (The textile finds in the harbour of Hedeby). In: Berichte über die Ausgrabungen in Haithabu 20. (K. Shietzel ed.) Neumünster 1984.

[17] N. M. Saxthorp: Kalkmaleriers kildeværdi (the source value of lime paintings). In: Førtid og nutid, bd. 24/3, Copenhagen 1970, p. 211-229.

[18] B. Irgens Larsen: Kongesagaenes skildringer av middelaldersk klædedrakt (The Kings’ sagas’ descriptions of Medieval costume). In: Det kgl. norske Videnskabers Selskabs Skrifter, nr. 4, Trondheim 1937. P. Meulengracht Sørensen: Saga og Samfund (Saga and Society). Copenhagen, 1977. opera cit.

[19] D. K. Burnham: Cut My Cote. Toronto 1973. A-M. Franzén: En koptisk tunika (a Coptic tunic). In: RIG årgang 44.3. Lund 1961, p. 96. V. Gervers: Medieval Garments in the Mediterranean world. In: Cloth and Clothing in Medieval Europe (N.B. Harte et al ed.), London 1983, p. 279-315. D. Renner: Die koptischen Textilien in den vatikanischen Museen (The Coptic textiles in the Vatican museums). In: Monumenti Musei e Gallerie Pontificie Pina coteca Vaticana, Bd. 2, Wiesbaden 1982. M-H. Rutschowscaya: Le vêtement de la période copte (clothing in the Coptic period). In: Exhibition catalogue Musée Archéologique. Guiry-en Vexin 1986.

[20] Diocletian’s edict on Maximum Prices, A.D. 301. cit. Gervers: op.cit. in note 19, p. 290-295.

[21] D. K. Burnham, A-M Franzén, V. Gervers: opera cit. in note 19.

[22] Ovid.1st c BC; Servius 4th century AD; cit. from V. Gervers: op. cit. in note 19, p. 300.

[23] M. Hoffman: Der ungenähte Rock in textilhistorischem Zusammenhang. In: Documenta Textilia, Festschrift für Sigrid Müller-Christensen. München 1981, p. 39. V. Gervers: op. cit. in note 19, p. 300.

[24] V. Gervers. op. cit. in note 19, p. 302.

[25] P.F. Fournier: Patron d’une robe de femme et d’un bas gallo-romains trouvés aux Matres-de-Veyre. In: Bulletin historique et scientifique de l’Auvergne 76. Clermont Ferrard 1956, p. 202 ff. A. Ferdiére: Tissues et vêtements en Gaule. In: Tissu et vêtements. Exhibition catalogue Musée Archéologique. Guiry-en-Vexin 1986. p. 110-111. (The tunic is certainly from a female grave, but both men, women and children wore tunics).

[26] M. Hald: Olddanske textiler (Ancient Danish textiles). Nordiske fortidsminder V. Copenhagen 1950, p. 348. E. Munksgaard: Oldtidsdragter (Ancient costumes). Copenhagen 1974, p. 117.

[27] P.F. Fournier, A. Ferdiére, V. Gervers: opera cit. in notes 25, 25, and 19.

[28] M. Hald: Ancient Textile Techniques in Egypt and Scandinavia. I: Acta Archaeologica XVI. Copenhagen 1945. V. Gervers: op.cit. in note 19, p. 313.

[29] E. Munksgaard: op. cit. in note 26, p. 106 f. S. Rudenko: Frozen Tombs of Siberia. London 1970.

[30] E. Nylén: Vendelryttaren, en länk mellan öst och väst forntid och medeltid (The Vendel horseman, a link between east and west, prehistory and middle ages). In: TOR XIX. 1980-82, p. 163-188.

[31] Strabon, Greek geographer ca 60 BC - 20 AD: Geographica IV: 196.

[32] C. Engelhardt: Thorsbjerg Mosefund (the Bogfind at Thorsbjerg). Copenhagen 1863. M. Hald: op.cit., E. Munksgaard: op.cit.

[33] J. Ilkjær & J. Lønstrup: Interpretation of the Great Votive Deposits of Iron Age Weapons. Journal of Danish Archaeology 1, Copenhagen 1982, p. 95 and 98 f.

[34] Jerome: Epist. LXIV 11, cit. J.P. Wild: Clothing in the North-West Provinces of the Roman Empire. In: Bonner Jahrbücher Bd. 168, Cologne.

[35] H. Potraz: Das Moorgewand von Repsholt (the Bog find at Repsholt). Veröff. d. urgesch. Sammlungen des Landesmuseums Hannover 7. 1942. M. Hald: op.cit. p. 344 f. Hald dates the kirtle through the weaving technique to the 2nd-4th century. E. Munksgaard: op. cit. 1974, p. 131. Munksgaard dates the kirtle on pollen analysis to the 1st - 2nd century.

[36] Einhard: Life of Charlemagne. Cit. from J. Bjernum: Kilder til vikingetidens historie (sources for Viking Age history), Copenhagen, 1972, p. 18 f.

[37] D. K. Burnham: Cut my Cote. Toronto 1973, p. 12 & 15. A-M Nylén: Folklikt dräktskick I västre Vingåker och Österåker (Rustic costumes in west Vingåker and Österåker). Stockholm 1947, p. 40 ff.

[38] H. Helbæk: Notes on the evolution and history of linum. In: Kuml 1959.

[39] G. Hatt: Jernalderens Bopladser I Himmerland (Ironage settlements in Himmerland). Aarbøger for nordisk oldkyndighed og historie. Copenhagen 1938, p. 221 f.

[40] H. Helbæk: Botanical studies of the stomach contents of the Tollund man. In: Aarbøger for nordisk oldkyndighed og historie. Copenhagen 1950. H. Helbæk: Grauballemandens sidste måltid (the Grauballe man’s last meal). In: Kuml. 1958.

[41] E. Munksgaard: Det såkaldte kohorn fra Øksenbjerg, omspundet med hør (the so-called cowhorn from Øksenbjerg, wound with linen). In: Aarbøger for nordisk oldkyndighed og historie. Copenhagen 1979, p. 5-10.

[42] E. Østergaard: Vikingetidstekstilier I Sdr. Onsild (Viking Age textiles in South Onsild). In: Aarbøger for nordisk oldkyndighed og historie. Copenhagen 1976, p. 87-95. E. Østergaard: Textiler. In: E. Rosedahl: Fyrkat. En jysk vikingeborg (Fyrkat: a Jutish Viking fort). Copenhagen 1977, p. 137. I. Hägg: Die textilfunde aus dem Hafen von Haithabu. In: Berichte über die Ausgrabungen in Haithabu 20. (K. Schietzel ed.) Neumünster 1984. I. Hägg: Hedeby Siedlung 31, current examination; personal correspondence. L. Bender Jørgensen: Førhistoriske tekstiler I Skandinavien. Copenhagen 1986.

[43] L. Bender Jørgensen: op.cit. in note 42. p. 164.

[44] See I. Hägg: Textilhistoria, statistik och källkritik (Textile history, statistics and source criticism). In: Tor XX, Uppsala 1985, p. 259-278. I. Hägg: Textilhistoria, statistik och källkritik 2. In Tor XXI. Uppsala 1986, p. 283-296. See also L. Bender Jørgensen: Et svar på debatoplæg (a reply to a debate issue) In: Tor XXI. Uppsala 1986, p. 263-281.

[45] H. A. Jensen: Seeds and other diaspores in medieval layers from Svendborg. In: The Archaeology of Svendborg no. 2 (M. Jansen ed.) Odense 1979. K-E. Behre: Untersuchungen des botanischen Materials der frümittelalterlichen Siedlung Haithabu. In: berichte über die Ausgrabungen in Haithabu 2. (K Schietzel ed.). Neumünster 1969. H.A. Jensen: Seeds and other diaspores in soil samples from Danish town and monastery excavations, dated 700 1536. In: Biologiske skrifter, Kgl. danske videnskabernes selskab. Copenhagen 1986. G. Jørgensen: Medieval plant remains from the settlements in Møllegade 6. In: The Archaeology of Svendborg, Denmark no. 4 (H. M. Jansen ed.). Odense 1986.

[46] H. A. Jensen: op.cit. in note 45, p. 89.

[47] G. Jørgensen.: op.cit. in note 45, p. 46-84.

[48] K-E. Behre: op.cit. in note 45.

[49] H.A. Jensen: op.cit. in note 45 (1986), p. 48f, 50f & 80.

[50] David Robertsson, NM 8 dept.: personal correspondence; pollen analysis planned.

[51] M. Hoffman: The Warp-Weighted Loom. Oslo 1974, p. 321 ff.

[52] K. Schietzel: Hölzerne Kleinfunde aus Haithabu (Ausgrabungen 1963-64). In: ausgrabungen in Haithabu, Bericht 4 (K Schietzel ed.) Neumünster 1970, p. 77-91.

[53] J. Kaminska & A. Nahlik: In: Acta Archaeologica Universitatis Lodziensis 6. Lódz 1958,

[54] A. Geijer & J.E. Andersbjörk: Two textile implements from the early middle ages. In: Folk-Liv III. Stockholm 1939, p. 232.

[55] G. Grenander Nyberg: Trampvävstol I vikingatidens Hedeby (Treadle loom in Viking Age Hedeby). In: Rig. årg. 60. Stockholm 1977, p. 47-48. A. Geijer: Ur textilkonstens historia (from the textile arts history). Lund 1980, p. 89 f. I. Hägg: op.cit. in note 44 (1985), p. 276 footnote 3. Flax expert Jette Flemming, teacher at Aarhus Arts academy: personal correspondence. The textile-historical experimental workshop, Lejre, has on my initiation begun trials with finer linen weaving on the upright loom. A definitive result will be available later in 1989.

[56] I. Stoumann: De, der blev hjemme (those that stayed at home). Esbjerg 1978.

[57] Conservator Jesper Trier, dept. of conservation, Moesgaard, whom I thank for the analysis.

[58] E. Lorenzen: Textilier (Vævninger, net og snor) (Textiles, weaving, nets and yarn). In: H.H. Andersen et al: Århus Søndervold. Aarhus 1971, p. 229 ff.

[59] Analysed by the author at the dept. of conservation, Moesgaard 1988, kindly assisted by conservator Anders Abildgaard.

[60] H.H. Andersen, P.J. Crabb & H.J. Madsen: Århus Søndervold. En byarkæologisk undersøgelse (an urban archaeological examination). Copenhagen 1971, p. 59-61.

[61] H. Krongaard Kristensen: personal correspondence.