18 February 2015

Warp-Float Weaves with Deflected Wefts: Wabengewebe

© Carolyn Priest-Dorman, 2001

A small group of related weaves survive from the period between the seventh and tenth centuries (Tidow, 133).  The examples (there are, by all accounts, seven or perhaps eight) have been found in England, Sweden, and Germany, and they are believed to be Alamannic or perhaps Frankish in origin.  Hans-Jürgen Hundt, who first took an interest in them after he discovered one, named them Wabengewebe, or honeycomb.

Technical Details

Technical information is only available for six of the pieces.  Two are from Swedish warrior graves of the Vendel period, Valsgärde 8 (the earliest piece) and Valsgärde 13.  One is from an eighth- or ninth-century Slavonic grave at Osmarsleben in Sachsen-Anhalt.  Two are from eighth-century Alamannic warrior graves, Sievern Grave 66 in Lower Saxony and Alladorf Grave 60 in Bavaria.  The last, Jorvík 1336,  is from tenth-century Anglo-Scandinavian levels at York.  The other two known pieces, Vendel 11 and Vendel 12, come from Swedish graves at Vendel and have not been published.

Some pieces used the same grist in warp and weft, while others used a heavier weft than warp yarn.  All were woven using Z-spun singles in warp and weft.  Only one piece (Alladorf 60) has been identified as woven of wool; all the other pieces that were susceptible to fiber type analysis turned out to be linen.

Fig. 1:  Alladorf Grab 60 wabengewebe; 20Z x 16Z wool; 8thC Frankish

Four of the pieces are in the medium fine range, at 20 to 24 ends per cm, with weft counts of 12, 16, and 18 per cm.  One piece is slightly less fine, with 15 ends and picks per cm, and one piece is pretty coarse at 7 ends and 9 picks per cm.  Some pieces were woven with equal numbers of ends and picks per cm, while others were less even.  The largest disparity in thread count comes from the Valsgärde 8 piece, with its 24 ends and 12 picks per cm (Walton 1989, 356).  Although it's unclear what purpose these textiles served, I believe that most of them were used as domestic textiles, e.g., towels or table linens.


The early medieval Wabengewebe is a real puzzler: apparently, no two known examples are identical in structure.  The basic structure is a one-shuttle, two-block weave that resembles both a modern honeycomb weave and a huck weave.  It has groups of warp floats like huck, and it has deflected wefts outlining cells in the ground weave, like modern "honeycomb" weaves.  However, there the resemblances end.  Especially, there are no corresponding long weft floats on the backs of Wabengewebes.  Variations in cell interlacements (tabby or 1/2 twill), numbers of warp floats, and cell sizes make all the known examples different from one another. 

Published photos of textile fragments in this technique are rare, and often it is not possible to determine the weave structure from them.  Since the original fragments are often very small, it is possible to misread a 1/2 twill cell as a tabby cell, or to miscount the number of weft threads in a cell.  The deflected weft sometimes obscures as many as three other wefts, depending on structure, and counting the total number of wefts can therefore be very difficult.

Additionally, not many drawdowns are published, and some of the published drawdowns are incorrect.  The generalized Wabengewebe drawdown published in Bender Jørgensen (Figure 1P, which she labels "honeycomb,") is incorrect: it's for a modern waffle weave.  Penelope Walton twice published a Wabengewebe drawdown that was 90 degrees off true (Walton 1989, 350; Walton 1990, 66). 

Adding to the mystery, Hundt's published schematic for the Sievern 66 fragment is technically impossible: some of the short warp floats abut other long warp floats rather than showing longer continuous warp floats (Hundt 153 and Abb. 4).  During the course of his investigations into Wabengewebes, Hundt consulted a weaving instructor named Frau Kircher.  The sample she wove for Hundt does not match either the photo of the original or Hundt's drawdown for the textile in question (see Hundt 159, Abb. 8-5).  Accordingly, since any reconstruction would be conjectural and without benefit of having consulted the actual textile fragment, a draft for the Sievern piece is not included here.

Two definite examples exist of cell structures in 1/2 twill, those from York (see the draft for Jorvík 1336) and Grave 8 at Valsgärde.  Two shots of 1/2 twill are followed by a shot of tabby in the first block; in the second block the opposite two twill sheds and the opposite tabby shed are used instead.  The reverse of a textile in this technique shows a section of tabby interlacement behind the warps that float on the front; some weft floats are sandwiched between the two layers.

Fig. 2: Jorvik 1336 wabengewebe

According to Walton, the Valsgärde 8 piece has shorter weft floats and longer cells than the Jorvík 1336 piece (Walton 1989, 356).  However, when you consider that Walton believes the textile to have been woven 90 degrees off its true orientation, this means that Valsgärde 8 actually has a larger number of warp floats than Jorvík 1336, and that they are shorter.  This results in wider, flatter cells than Jorvík 1336.

Other examples, more mysterious and less well documented, have been identified by various specialists as consisting of tabby cell structure.  Among these is the Sievern piece, with its impossible drawdown.  Another piece, with small cells definitely woven in tabby, is also woven in a mixture of twill and tabby.  Shots of 2/1 twill alternate with shots of tabby, with the tabby shed changing to mark the change between the two blocks.  Pairs of warp floats on the front alternate with single warp floats on the reverse.  The deflected wefts are of course in twill.  (See the draft for Alladorf Grave 60.)  The Osmarsleben fragment is identified as having tabby cells (Bender Jørgensen 1991, 237, entry Germany V.40.l).  The accompanying photo (Bender Jørgensen 1991, 147, Fig. 181d) appears to have tabby cells; however, since the cells are small I suspect that some interlacement like the Alladorf one may have been used.

Photos and/or drafts are not available for the other three Swedish examples.  The eighth century fragments from Grave 13 at Valsgärde in Sweden were found on a shield boss.  A firm identification of the ground weave is impossible, but it was either "tabby or 2/1 twill" (Bender Jørgensen 1991, 262, entry Sweden IV.17:113).  No further information on the two Vendel fragments is available at the time of writing.


Black, Mary.  The Key to Weaving: A Textbook of Hand Weaving for the Beginning Weaver, Second Revised Edition.  New York:  MacMillan Publishing Company, 1980.

Bender Jørgensen, Lise. North European Textiles until AD 1000. Aarhus, Denmark: Aarhus University Press, 1991.  Huge catalogue of extant early textiles, mostly fragments; exceedingly well documented.

-----.  "The Textiles of the Saxons, Anglo-Saxons and Franks."  Studien zur Sachsenforschung, vol. 7 (1991), pp. 11-23.  Good summary information on textile types particular to these cultures.

Davison, Marguerite Porter.  A Handweaver's Pattern Book, Revised Edition.  Swarthmore, Pennsylvania: Marguerite Porter Davison, 1951 [1944]. 

Hundt, H.-J.  "Textilreste aus dem frühgeschichtlichen Kriegergrab von Sievern, Kr. Wesermünde, 1954."  Studien zur Sachsenforschung, vol. 2 (1980), pp. 151-160.  Write-ups on Sievern 66 and Alladorf 60.
Tidow, Klaus.  "Kleingemusterte Woll- und Leinengewebe aus der Eisenzeit und dem Mittelalter--Herkunft, Herstellung und Verbreitung," pp. 131-137 in Textiles in European Archaeology: Report from the 6th NESAT Symposium, 7-11th May 1996 in Borås, ed. Lise Bender Jørgensen and Christina Rinaldo. GOTARC Series A, Vol. 1. Göteborg: Göteborg University Department of Archaeology, 1998.  Brief summary of information on several interesting early medieval weaves, including Wabengewebes.

Walton, Penelope. Textiles, Cordage and Raw Fibre from 16-22 Coppergate. The Archaeology of York, Volume 17, Fascicule 5. York: York Archaeological Trust and the Council for British Archaeology, 1989.  Write-up on Jorvík 1336 and an excellent comparative section on similar weaves.  [Now available from Pangur Press at this link.]

Walton, Penelope.  "Textile Production at Coppergate, York: Anglo-Saxon or Viking?," pp. 61-72 in Textiles in Northern Archaeology: NESAT III Textile Symposium in York 6-9 May 1987, ed. Penelope Walton and John P. Wild. London:  Archetype Publications, 1990.  Brief mention of Jorvík 1336 and similar weaves.

*  *  *  *  *  *  *  *  *  * 

This article was originally published in Issue 27 (March 2001) of Medieval Textiles, the newsletter of the medieval textiles study group of Complex Weavers.  The draft and drawdown I submitted were created using the freeware drafting program, WinWeave.  Using GIMP, I have turned the original draft into a graphic for republishing here.  I have also restored the annotations to the bibliography; they were omitted for space reasons during publication.

This is a kitchen towel in the Jorvík 1336 weave structure.  It was woven in unmercerized cotton as a present for me by Susan Nalley.  It's been in my normal kitchen towel rotation for maybe seven or eight years now and is holding up beautifully.


I have not pursued the Wabengewebe class of weave structures lately; it may be that other examples of it have turned up in the last 15 years.  If I find any more, I'll post about it with cross-links.


  1. Hello,
    Anita Malmius mentions this weave for Valsgärde 5 and she thinks is edging on a kaftan and on a tunic.
    Best wishes, Linda

    1. Thank you for commenting!

      The float patterning on that textile is different from all the ones mentioned in this article. I haven't figured out how to classify it yet.

  2. This comment has been removed by the author.