Birka silk translation

Translation of Chapter 5, Birka III by Agnes Geijer by Katherine Barich

Silk fabrics

Fragments of silk fabrics are found in a fairly large scale – in about 45 of the graves.  These remnants, each considered on their own are certainly quite unassuming and damages, but there are enough of them to provide a sample card of a not insignificant import.   No doubt can prevail that the represented silks were entirely imported.

The most common silk fabric, S 4 - which one is tempted to call the standard quality – was an ordinary type in the Mediterranean area at that time, defined as a 3-binding double twill with a binding warp and a pattern forming filling weft.  It is found in finer and coarser qualities, more or less evenly woven and finely executed, plain or with traces of pattern, but in technical terms they are so consistent that one can hardly doubt that these wares, by and large, came from the same site of manufacture.  Technically speaking the thin striped silk fabric from grave 944 S 4 is included in this category, although the first glimpse of it gives a different impression.

The twill weave is shown in figure 16.  It has a double warp system.  The top of the fabric is similar to a true twill (3-binding weft faced twill), the bottom side is different, see Plate 12:2.  For the patterns, the effects are achieved by the different colors of the weft yarns, which swap position; the surface is always consistent.  The two warps perform a different function:  the binding warp binds the weft and is visible from the right side, while the pattern forming filling warp serves as a support for the weft threads inside the weave and at the selvedges.  At the point or the row where the color change takes place non-uniformity, or furrow occurs, which is the only indication that a silk was patterned where the fabric has completely lost its color. [1]

The material in this twill fabric is true silk that is more or less completely scoured of its sericin.  Unlike a number of other examples of the silk fabric, the Birka fragments with the exception of S 5 always have a double filling warp, between each binding thread there are always two filling threads.  In the Oseberg finds are similar quantities of strips with patterned double twill, however, in many cases the filling warp is only a single thread.

Found not nearly as often as twill, only in 5 – 6 occasions, is a very well woven unpatterned taffeta weave silk, the threads of which by micro-chemical analysis are shown to be raw silk. The fine cocoon threads have not had the sericin removed.  Although raw silk is generally less highly regards, we have reason to believe here that this fabric was consider a rare and costly here.  It was always used with remarkable economy and in the context of the best ornaments, as a background for the finest embroidery (such as the golden stag) or in the millimeter small strips as the edge of gold bands, in the latter case, for the purpose of providing a special contrast to the gold bands.  Unfortunately we cannot say with certainty what color it was because all chemically reactive dyes are now gone.  On is inclined to believe that there was a color that contracted with the gold, perhaps red or maybe blue or green, but not yellow, as was proposed by Ing. Ljungh (Supplement I).   The data also shows that in textiles especially the color red is related to brown and yellow.  

Illustration 16 – Diagram of Double Twill S 4.

Left unpatterned, Right the thread pattern for color change


The material in an interesting patterned fabric (S 3) is consistent with this smooth taffeta.  The fabric is monochromatic and the pattern is achieved by binding effects.   The ground weave is tabby binding, from which the geometric pattern shape stands out, where the threads are unbound in each thread system – in the warp on one side and in the weft on the other -   thereby forming a glossy surface.   The fabric also has no definite top or bottom.

Compared with the uniform texture of the twill fabrics, where patterning is achieved by different colors of the weft, here we encounter an entirely different principle, which is based on the sheen of the surface floats, in contrast to the tightly bound tabby weave.  Fabric of this kind should probably be regarded as a precursor to real damask.  From this there can be the possibility a direct relationship to the development of such technology from the expensive as well as strong and shiny silken material.

Although the fabric analyzed from S2 has a plain weave and is made from raw silk, it makes a curious impression of being strikingly different from the rest.  To the naked it eye it shows little similarity to the silks in general.  The fabric is not only very thin but also dull and smooth as thin silk paper, a condition which must be due to a kind of artificial compression or finishing.    It is noteworthy that this treatment was not rendered ineffective by exposure to moisture over the centuries.

Illustration 17

Completion of the pattern of S3

The study of the silk S 3 brought an interesting surprise.  Under the microscope small gold traces became apparent; they must necessarily be the result of painting or printing a pattern on the fabric.  The fabric came from under the sewn on silver trim, which could not have produced the gold color, and it would have been completely unaffected by spots from water.

S 1 a. Grave 824, Table 13:3 – Silk Taffeta – Very smooth but thick and tightly woven.  The same number of threads – 28 per cm – and type for both directions.  According to the study material (See Suppl. 1) the silk thread is very even and consists of nearly 150 cocoon filaments, which are only slightly twisted.  The silk is not scoured – the cocoon filaments still possess the protective silk gum (sericin).   The fact that the cocoon fragments need not be spun together because of their remarkable evenness indicates skill in the treatment and sorting of the silk fibers.  In the analysis, the assumption is expressed that the original color was yellow – a view with which the writer can hardly agree.  The above described silk taffeta is found in the form of a 2 – 3 mm wide edge around a piece of a band shot with gold about 1 dm long.

In grave 731 and 750 there was taffeta with similar usage.  It also formed the background to two embroideries St 16 (the “golden deer”) and St 26.

S 1 b. Grab 731 On the chain approach on the back of one small round silver clasp with filigree decoration of Swedish production is sewn a small strap in fine light-brown taffeta.  The fabric is folded several times and sewn together with tight silk stitches.

S 2. Grave 660 Table 13:2  Thin, relatively sparse 2-bind weave.  In one thread system 19 threads per inch, the other, non-uniform, 26 – 28 per cm.  Which one forms the warp is uncertain, but it is probably the less dense and even sized.  The thread is dull and with ordinary examination with normal magnification, it does not resemble silk.  The small fabric fragments are thin, flat and very brittle; in appearance they rather resemble thin silk paper.  The chemical analysis (Suppl. 1) however revealed that the material is raw silk.  Each thread in the fabric consists of an equal number of cocoon filaments which have been fused together and woven without further twist.   For the curious appearance of the fabric, the analysis provides no explanation.  One must assume, however, that any artificial pressure or finishing had been subjected to a centuries long soaking which remarkably did not render the treatment ineffective. 

The above described fabric fragment was found together with silk strips (S 4) and were partially folded several times.

S 3. Grave 944. Table 13:4 and Figure 17.  Thin silk, patterned by means of binding effects.  The pattern repeat is not completely preserved, but it can be reconstructed with great probability as shown in Figure 17; alternating small and large squares arranged stepwise, the centers filled with small dots or stars.   The ground is a tabby weave of fairly even weight.   The patterned is formed by the threads (on one side the weft, on the other the warp) 3, 5 or 7 arranged crossing threads remain unbound (“floating”).   The transition from one floating system to another naturally produces diagonal pattern lines.  As mentioned above, this technique is based on the difference in the effect of the matte surface of the tabby weave and the luster of the floats; this is especially true when the fabric is dense.  See Table 38: 6 – 7.

With fabric transparent as this one present, a further effect is produced, which for the most part isn’t apparent to the eye when examining this old material.  The floating threads have a tendency to bunch together and compress when they cross, thereby resulting in some areas which show partly closed, partly open to a certain degree in relationship with the base fabric.    It calls to mind a gauze weave.

The fabric has no definite top or bottom, although that pattern may stand out more clearly on one side depending on the warp and weft position.    A difference is noted, that in the production of this fabric the threads are bound by each warp thread, while the Palmyra fabric (see page 54) has the thread only bound by the second warp to produce the tabby binding.  The thread density is 46 per cm in one direction and 54 the other.  Which of the two thread systems functioned as the warp cannot be determined with certainty, but the greatest probability is that the denser is the weft.  This orientation also forms the top side as judged by the seam and placing of the silver band, and shows the pattern most clearly.

The micro-chemical analysis (Suppl. 1) shows that the material is raw silk.  In addition on one of the thread systems, probably the weft, a pale blue color can be detected, which one can assume originated from indigo.  The fabric would therefore have been in two colors, so that the pattern would have emerged more clearly.  It is, however, somewhat uncertain, as one can also imagine that the blue could also originated from the copper in the silver alloy.  Now the color is rather dark brown; however, any color or shade differentiation cannot be observed either with the naked eye even at normal magnification (2 or 3 X).

The shape of the conserved fragment was apparently caused by the preservative chemical precipitation for the silver thread that is sewn onto it.  Note the holes left by the stitches!  Only after the silver was lifted and the fabric illuminated was the pattern discovered, because lighted from one side the pattern is difficult to ascertain.  The silver trim concealed a seam, neatly executed, folded and stitched through on both sides, as a so-called felled seam.  The shape of these two joined pieces of material were curved, the one a bit, the other more so.  At one end where the trim is narrower and probably terminates – the fabric is cut straight (where the threads run parallel or perpendicular to the seam), while further up it is gradually cut diagonally across the threads.  Further, the narrow end of the braid is rapped by the fabric, seemingly hemmed by the fabric.  Could it be some kind of cap, with a seam across the head and ending at the ears with pleats?

To the naked eye are hardly recognizable forms here and there that under magnification are visible flecks of gold on the fabric.  The gold dust is clearly on the surface and must have been painted or possibly printed by using some applied binder.  The fragments, however, are much too small to subject to a chemical analysis that could give information on the procedure.

S 4 a-d Plates 12 and 16 double weft silk twill.  The binding is evident from the schematic above.  The binding warp thread binds with any of 3 weft threads, giving the surface the appearance of a 3-bind weft twill.  Only every second weft thread is visible on the top surface, while the other is only visible on the underside.  To change colors the two weft threads could trade places, as shown on the figure on the right.  Two threads of the same color are situated – as is shown – they change colors next to each other suggesting that the same shuttle throw returned them, before picking up another color.  The visible thread is for the shiny weave the same, although I have drawn the hidden thread in between the weft threads on the left of the diagram. 

Only interruptions in the texture, sometimes in association with a change of brown shades let us suspect  from these instances that the fabric was originally woven in several colors. [2]

The filaments are fairly uniform in all the pieces.  The warp in relation to the weft is of softer and firmer thread, not spun too strongly to the left. 

The weft yarn is always coarser than the warp yarns and also can vary quite a bit in strength – it is spun so loosely that the spin direction cannot be observed.   The weft is so dense, that one can measure a good double the size of the warp at the point of crossing.   Since one half of the weft threads are always covered by the other, the double number of weft yarn corresponds to the number visible on the top crossing points. 

The above-described type of silk twill exists in more or less qualities of fineness, of which the following examples demonstrate the various degrees.

S 4 2. Grave 735.  Table 15.  The largest remaining fragment of silk fabric of the Birka material, about 19 x 12 cm.  At places a few interruptions in weave can be distinguished, an indication that the fabric had some pattern.  The fabric is, in terms of both design and material, very uneven and rough.   The warp can be counted with at least 10 – 12, sometimes up to 14, per cm.  The weft has about 30 binding points per cm, corresponding to a fill density of about 60 threads per cm. 

S 4 b. Grave 944.  Weft very unequal, sometimes with clumps.  Faint traces of a pattern.  In the warp 12 – 14 binding threads per cm.

S 4 c. Grave 524.  Evenly woven, middle-fine quality.  Uniform warp with 15 binding threads per cm, the weft is less dense than usual: about 23-25 binding points per cm.  Indications of irregularity in the structure possibly indicate changes in color, but now no differences in nuances are noticeable.  The fabric was examined (Suppl. I) and determined to be a true silk, which was apparently scoured.  The investigator makes a dubious assumption in the examination report about the original color of the silk fabric (yellow?).

S 4 d. Grave 944.  Finest quality with apparent color changes.  In the warp 16-17 binding threads per cm.  The weft count is 42 binding points per cm.  The fabric is evenly proportioned, the silk unusually shiny.  The color change becomes known by apparent nuances in two brown shades.  The fragment is too small to be able to provide information about the character of the pattern.

S 5. Grave 944. Table 13:1.  Thin striped silk twill with a distinct pattern.  The binding system is the same as those described above except that there is only a single weft thread.  Both the warp and the weft are much sparser, so that the texture appears obscure.  The warp consists of fine, sharply spun silk threads, which number 32 per 1 cm (16 binding points).  The weft is coarse, especially in the pattern areas, and virtually unspun.  The number of weft threads per cm is quite variable. In the widest part of the fragment can be distinguished clearly a 13 mm wide band, in which 2 similar figures emerge, whose exact form however cannot be determined with certainty – the largest piece consists of 3 parallel bars, which may possibly be connected to form a simplified palmette or similar form.  The border pattern runs between narrow stripes with a quite delicate weft, so that the fabric is thinner and less glossy.  Then follows a broad stripe of thick and glossy weft, in which is found no discernible pattern.  A thicker area follows after another thin, narrow stripe, where darker spots may indicate a pattern. 

The fragment consists of an uneven, diagonally cut piece approximately 3 x 10 cm, having on one end a 5 mm wide seam running the length of the fabric, however it is frayed on the edges.   It gives the impression as if it had been heavily pleated (wrinkled?)  There is an additional small piece of cloth about 3 cm. long, which consists of a very narrow seam, which is worked with a rather coarse silk thread.  Judging by its form, the silk is obviously from an independent piece of clothing, because it is not, as found on the other examples, sewn on to another fabric as decoration.

 That silk was a precious commodity can be seen from the fact that it seemingly was used mainly in small quantities, in the form of narrow strips or small pieces of fabric sewn onto other fabric as ornamentation.  See Chapter XII where clothing is discussed in particular.  The silk twill was primarily used for this purpose and the taffeta more seldom and sparingly.

In one case such a large piece of twill fabric remains that a reasonable suggestion of a larger piece of silk clothing can be made.  The two fragments of thin patterned silks, S3 and S5 might even have been made into whole garments, perhaps a head covering.  This is indicated by the seam on S3 which was covered by a silver trim.  The same quality silk was used for the tablet-woven bands.

However in the thin, taffeta-like silk varieties as well as in the patterned effects of S1, S2 and S3, the cocoon filament is not scoured, it is in other words raw silk, but in general of very fine quality.  The fact that the trouble was not take to spin the cocoon threads together, but rather just let them stick together (by means of the sericin), suggest that the place where the silk was reeled allowed this so that they could survive a long transport and still be good to use as warp threads.  For the weft threads there is not a great a demand for uniformity or strength of the yarn.  Raw silk generally has a lower gloss and is therefore less valued than the de-gummed silk, despite it being very strong.  The quality of the exceptionally fine and well-crafted taffeta, and especially the patterned silk, suggests that the best and finest material was used, which in the place where the silk was produced, was made to requirements.   To weave fabric this thin, the thread should probably be regarded as almost perfect, and understandably the worker required great skill and execution.  In comparison more western twill fabrics, for the most part the wefts of similar workmanship and their warps are often plied. [3]

Regarding the quality of the silk, it is also found in other Birka material, namely the bands.  In the tablet woven bands the thread is usually of degummed true silk, which is carefully spun and plied.  In the analysis of the wide band B 1 was found a unique type of spun silk.  The warp here is also de-gummed silk, while the coarse, multicolored silk yarn that was used for the pattern was made of a material now called Florette silk, i.e. waste silk that was not de-gummed.  This silk cannot be reeled, and therefore can only be processed by carding and spinning in the same way wool is.  The fact that the quality of the cocoon threads is seemingly coarse suggests that only the outer layers – what wasn’t able to be reeled - were carded. 

The predominant quantity of silks that are characterized as de-gummed silk twill, are more decidedly western than other materials.   The twill technique at this time is particularly characteristic for the Mediterranean countries as well as Persia.  [4]  However, because the shapes of the patterning are not preserved, it is impossible to conduct a detailed systematic study that would provide information in respect to technical variations that might exist among the numerous materials here.  One of such technical differences, a single filling warp from the same period’s Oseberg finds differed from Birka’s double warp described above.  The same difference was found when compared to the silks from the Danish shrine, one datable to the 11th century and the other probably much older.[5]  In any case, the uniform quality of the silk twills, at least in the type S 4 – suggests that they originate from the same production area.  It was probably Byzantine, transported over the length of Russia.  The Russian chroniclers often report on the purchase of Byzantine silk wares.    IbnFoszlan made mention of gold brocades, which he saw on the Volga-Bulgarian and Russians, came from the Byzantine Empire and the Russian chroniclers often mention purchase from there.  An indication of the popularity of Byzantine silks is the stipulation in a contract with Prince Igor that the Russian-Scandinavian merchants should not be allowed to buy the most expensive weaves. [6]

As for the silk in the tablet woven bands, what we think is likely (cf. pp. 90 ff) were produced in the north, and also passed to the Eastern route as well as into the more western areas of trade in the Rhine area. [7]  Certain the import of oriental silks after the Carolingian empire was by no means insignificant.[8] 

Compared with the other silks there is second group of rarer examples:  there are only 5 or 6 examples.  The material’s extremely delicate almost unspun silk gives reason to believe that the finished fabrics come from the Far East.  At this point of time the Chinese had achieved the upper hand in sericulture.  This technology and patterning of fabric is found only in the group of silks in S 3. 

From the pictures shown from the Lou-Lan stone in Chinese Turkestan (Table 38: 6-7), it appears that the patterns and technique agree well with some of the Han period Chinese silk. [9]   Of approximately the same age, the thin crepe silk fabrics found by Kozlow’s expedition to the Noin-Ula Mountains in northern Mongolia is the same type.[10]  The same is also true from some fabrics found in Greek tombs in Crimea, dated to the 3rd century before Christ, allocating them as the earliest Chinese imports.[11]   The beautiful, large-patterned silks of China, which were found in Palmyra and published in an exemplary manner by Pfister[12], dated from before AD 273, when the city was abandoned.  The latter represents a more advanced stage of technology, however the quality of the spun silk yarn is apparently quite the same as in the small Birka fragment.

In an essay Vivi Sylwan showed that a previously regarded 2-tone silk regarded as Occidental was Chinese; she also correctly pointed out that the possibility of European church treasures from the early Middle Ages as being of Chinese origin.[13]  Although it is possible that silk fabrics similar to S 3 might have made their way to Scandinavia through northwestern European trade, it is surely more probable that it came from the East.  Some of the Birka bronze fitting finds show indications that the Swedes had relations with the Khazars.  The Arabs from the southern Caspian Sea region could certainly have easily integrated the wares of the Khazars from the trans-Asiatic caravans.







[1] The use of the term “Polymita for this ribbed or twill fabric was first proposed by VIVI SYLWAN, Studier i senantik textilkonst, Några skaftvävnader, Rig 1923, and was recorded as used by Pliny and others.

[2] It is easy to see the Birka fragments are far too ambiguous to serve as a basis for critical examination of the thread system.  The diagram in figure 16 shows as its basis the red-blue eagle silk from Saint Knut’s shrine in cathedral of Odense (investigated by A. Geijer in Aarböger 1935) which fabric matches the rest of the structure in the Birka fabrics.


[3] In order to refute any objections it may be good to add the following statement.  The method of not de-gumming the silk does not imply, in and of itself, a more primitive method.  Pfister’s investigation (see below) shows, for example, that degumming along with all intended dyeing was postponed until the weaving was finished.  Then the stresses on the strength of the single thread in this state are less great as compared to when the thread is on the loom.  In addition was the fact that two fused cocoon filaments with sericin were significantly stronger than a de-gummed single cocoon thread.

[4] O. v. Falke, Kunstgeschichte der Seidenweberei, see Illustrations 103, 140, 143, 211, 241.

[5] Agnes Geijer, Sidenvävnaderna i Helige Knuts helgonskrin in Odense domkyrka, Aarböger 1935. (Silk weaves in Saint Knut’s Shrine in Odense Cathedral)

[6] W. Heyd, Histoire du commerce du Levant, I, page 64 and 72.

[7] H. Arbman, a.a.O., Chapter 1.

[8] See, for example, B. Francois L. Ganshof, Note sur un passage de la vie de Saint Géraud d’Aurillac, Mélange Iorga 1935.

[9] Sir Aurel Stein, Innermost Asia, Oxford 1928. 

[10] Veroff. Auf Russisch von der Akademie der Wissenschaften in Leningrad 1925, as well as Camilla Trever, Excavations in Northern Mongolia, Leningrad 1932, Table 21.

[11] Compt-rendu de la commission imp. Archéol. 1878-1897, Table V, 3.

[12] R. Pfister, Textiles de Palmyre, Paris 1934

[13] Vivi Sylwan, Eine chinesische Seide mit spätgriechischem Muster aus dem 5. Und 5. Jahrhundert.  Mit we-technischer Untersuchung von Kurt Hentschel.  Ostasitische Zeitschrift, Neue Folge XI, Heft I, also B. W. F. Valback, Spätantike und frühmittelalterliche Stoffe, Katalog des Römischen-Germanischen Zentralmuseums Nr. 10, Table 7, catalog # 276. [A Chinese Silk with late Grecian pattern from the 5th and 6th Century.  With Technical Weaving Examination  by Kurt Hentschel; and Late Antique and Early Medieval Fabrics, Catalog of the Roman-German Central Museum.