Reconstruction of the band with the animal pattern from the tomb of St Bathilde
Unlike the geometrically patterned band from St Bathilde's tomb and the sleeve trim of St Bertille, the third band from Chelles (France) is seldomly reproduced. (So far I know about the only copy by Marijke van Eppen).
The reason is probably not only its width (4 cm, ca 232 warp ends), but mainly the combination of its weaving techniques. Whereas the lateral parts clearly show the structure of tablet weaving, the central part was woven in tabby, probably double layered. (To know this for sure, seeing the reverse would be necessary.) As the three parts are partly separated in the original band, it is generally believed that the tablet woven strips were additionally sewn to the central band which was woven in another technique.
However, after seeing the photo of the band I have wondered about two things:
1) The gaps between the three parts do not separate tablet-woven trims and the central tabby piece, but they are placed inside the tablet-woven area, between the yellow warp-twined stripe and the pattern in "pebble weave".
2) Even though there is no evident system in the sequence of the motives in the lateral bands, they match exactly on both sides, which indicates that they were woven in the same time.
There are extant textiles from the 1st millenium CE, woven on the vertical warp loom with tablet-woven edges worked at the same time, but those are large mantles. Would it be practical to combine a string heddle (or more string heddles because of the pattern) with tablets to weave such a narrow band, when it is possible to weave a tabby by turning tablets back and forth? Probably not.
However, how to explain those precise tears along the band, if it was woven in one piece? For a double-weave two wefts are necessary, but it is unlikely that they both continued into the lateral bands. Maybe the turning point of the second weft made the band weaker in this place. This second weft would also explain why the edges of the central part are not frayed after the lateral band was torn off.
I decided to prove my ideas by an experiment. To compare the two methods, I wove the central part on the tablets together with one of the lateral bands only, while the other one was added later.
Before I get to the results of my experiment, I should probably explain how to weave a double layered tabby on the tablets:
All 52 tablets of the central pattern were threaded with one red and one yellow thread and all in one direction. (The threading pattern is shown below; nevertheless, it does not matter whether the direction is S or Z.) Unlike in typical tablet weaving, the tablets were positioned on the corner and thus two sheds were formed. By rocking the tablets back and forth, one color stayed on the top and the other below. When the weft of the same color was used in the correct shed, the two-layered unicolor background was woven. For getting the other color to the top, the selected tablets were turned in the same direction for the second time, and then the rocking continued, until another change of colors was required.
In contrast to the complicated description, the method of weaving is quite intuitive. It was enough for me to draw the pattern on the grid paper and then follow the colors, instead of directions of turning. Only for the readers of this blog to better understand my text, I tried to create more usual pattern, but the tablet-weaving draft software is not suitable for this kind of patterns. However, I have included a part of it anyway, because of the threading diagram.
More complicated than turning the tablets proved to be passing of the weft, because in the two-colored rows the weft needed to change the shed several time in one row. It helped me a lot that I was using a vertical loom and could easily flip the whole packet of the tablets backside to the front, when the weft ran mostly in the bottom shed.
After finishing the main part of the band, I wove the narrow lateral band and tried to sew them together. Several problems occurred immediately: 1) Due to fine silk threads it was difficult to pass a needle exactly through the weft loop at the selvedge. 2) Moreover, the weft loops in the lateral band were obscured by long warp floats, as there was no warp-twined edge at the side adjacent to the central band. 3) When the sewing thread was not passed exactly through the weft loops, the stitches became visible on the surface. 4) I did not manage to beat the weft with the same strength through my work and as a result the patterns in both lateral bands did not match exactly.
All of this would support the theory about weaving all three parts together as one piece, but on the other hand, this idea seemed less probable to me due to my weaving experience. The mental switching between two patterns requiring different techniques twice in each row was quite exhausting and time consuming.
Therefore, in the trial no. 2 I threaded weft through a needle and sewed the lateral band to the central part during the weaving. This method eliminated all previous problems.
Based on these experiments, I now believe the band was woven in this way: First the central part was woven, and lateral bands were added later, sewn to it directy during their weaving. The question about the method used for weaving the central pattern I am leaving open. I only proved that it could have been done on tablets. But was it done so? Who knows!
The reason is probably not only its width (4 cm, ca 232 warp ends), but mainly the combination of its weaving techniques. Whereas the lateral parts clearly show the structure of tablet weaving, the central part was woven in tabby, probably double layered. (To know this for sure, seeing the reverse would be necessary.) As the three parts are partly separated in the original band, it is generally believed that the tablet woven strips were additionally sewn to the central band which was woven in another technique.
However, after seeing the photo of the band I have wondered about two things:
1) The gaps between the three parts do not separate tablet-woven trims and the central tabby piece, but they are placed inside the tablet-woven area, between the yellow warp-twined stripe and the pattern in "pebble weave".
2) Even though there is no evident system in the sequence of the motives in the lateral bands, they match exactly on both sides, which indicates that they were woven in the same time.
There are extant textiles from the 1st millenium CE, woven on the vertical warp loom with tablet-woven edges worked at the same time, but those are large mantles. Would it be practical to combine a string heddle (or more string heddles because of the pattern) with tablets to weave such a narrow band, when it is possible to weave a tabby by turning tablets back and forth? Probably not.
However, how to explain those precise tears along the band, if it was woven in one piece? For a double-weave two wefts are necessary, but it is unlikely that they both continued into the lateral bands. Maybe the turning point of the second weft made the band weaker in this place. This second weft would also explain why the edges of the central part are not frayed after the lateral band was torn off.
I decided to prove my ideas by an experiment. To compare the two methods, I wove the central part on the tablets together with one of the lateral bands only, while the other one was added later.
Before I get to the results of my experiment, I should probably explain how to weave a double layered tabby on the tablets:
All 52 tablets of the central pattern were threaded with one red and one yellow thread and all in one direction. (The threading pattern is shown below; nevertheless, it does not matter whether the direction is S or Z.) Unlike in typical tablet weaving, the tablets were positioned on the corner and thus two sheds were formed. By rocking the tablets back and forth, one color stayed on the top and the other below. When the weft of the same color was used in the correct shed, the two-layered unicolor background was woven. For getting the other color to the top, the selected tablets were turned in the same direction for the second time, and then the rocking continued, until another change of colors was required.
In contrast to the complicated description, the method of weaving is quite intuitive. It was enough for me to draw the pattern on the grid paper and then follow the colors, instead of directions of turning. Only for the readers of this blog to better understand my text, I tried to create more usual pattern, but the tablet-weaving draft software is not suitable for this kind of patterns. However, I have included a part of it anyway, because of the threading diagram.
More complicated than turning the tablets proved to be passing of the weft, because in the two-colored rows the weft needed to change the shed several time in one row. It helped me a lot that I was using a vertical loom and could easily flip the whole packet of the tablets backside to the front, when the weft ran mostly in the bottom shed.
After finishing the main part of the band, I wove the narrow lateral band and tried to sew them together. Several problems occurred immediately: 1) Due to fine silk threads it was difficult to pass a needle exactly through the weft loop at the selvedge. 2) Moreover, the weft loops in the lateral band were obscured by long warp floats, as there was no warp-twined edge at the side adjacent to the central band. 3) When the sewing thread was not passed exactly through the weft loops, the stitches became visible on the surface. 4) I did not manage to beat the weft with the same strength through my work and as a result the patterns in both lateral bands did not match exactly.
All of this would support the theory about weaving all three parts together as one piece, but on the other hand, this idea seemed less probable to me due to my weaving experience. The mental switching between two patterns requiring different techniques twice in each row was quite exhausting and time consuming.
Therefore, in the trial no. 2 I threaded weft through a needle and sewed the lateral band to the central part during the weaving. This method eliminated all previous problems.
Based on these experiments, I now believe the band was woven in this way: First the central part was woven, and lateral bands were added later, sewn to it directy during their weaving. The question about the method used for weaving the central pattern I am leaving open. I only proved that it could have been done on tablets. But was it done so? Who knows!