The cloth of the hood was a 2/2 herringbone twill weave with very erratic widths of the chevron stripes. These widths ranged from 18 warp threads per stripe to 88 warp threads. This is very unusual as the chevron stripes in 2/2 twill weaves in the iron age period are customarily even. In a study of woven fabrics and their construction in Danish Iron Age textiles Margaret Hald observes:

‘ 2/2 twill is the most usual form of twill. A weft thread passes alternately over two and under two warp threads, with the result that the right side and the reverse side of the fabric is similar. The points of intersection are not as close as tabby, therefore a twill fabric is softer and more comfortable - an advantage which probably contributed towards the change in the Bronze Age from Tabby to twill fabrics…. The simpler variations of twill are made by changing the direction of the diagonals. Shifting the points of intersection so that horizontal zig zag lines are made produces the weave called pointed twill.’(Hald 1980, 148)

Hald goes on to describe twill pattern variations but in all instances the zig zag patterns are symmetrical. Henshall surmised in her research the warp threads were single threads and the weft threads were fine and used double, in other words each horizontal thread was made up of two fine threads used together. Also she suggests that the chevrons ran horizontally across the loom in a very uneven zig zag pattern.  

I began to thread the loom using this assumption. A 2/2 twill weave requires four different sheds to form the twill weave pattern. Therefore I had to calculate which warp threads should be attached to the four sheds incorporating the uneven chevrons into the calculations. I really did not think this would be a problem until I actually tried to do it. It is important I feel at this stage of my report to detail just how uneven the chevron stripes were. These numbers are the warp threads per stripe each time they changed the direction of the stripe: 18, 18, 28, 24, 18, 38, 42, 26, 22, 88, 26, 26, and 18. It took about two days or 16 hours to thread the loom before I could start weaving. After twelve rows of weaving the pattern appeared to be working, until I inspected the reverse side of the cloth and discovered innumerable loose threads forming loops at the change in the chevron stripes. So I unpicked the weaving and started to re-thread the sheds again because the cloth was a 2/2 twill weave the whole loom had to be re threaded again another two days work. I did in fact re thread and start weaving the loom for 48 hours trying every conceivable method I could think of to accommodate the uneven chevron stripes. 

The only feasible conclusion I came to was that it could not be done with the zig zag stripes horizontal. The only other way to try it was to treat the single warp threads as weft threads and the double fine weft threads as warps. So I re threaded the loom again using double fine threads as one, which worked very well although the two threads were not plyed together, they were just a strong as if they were plied. The loom was threaded again and this time I was sure it would work as the twill weave went as a block in one direction until the number of rows were woven such as 18 rows, 18 rows and 28 rows etc. While studying the hood at the NMS I noticed that at each change in the chevron direction there were an indented strips at each change. After examining some of the close up photographs taken during its conservation in 1981 I noticed that there seemed to be three treads picked up along the strip instead of the two needed for the twill weave.

A simple solution to this anomaly occurred to me rather if I used a bone needle I could catch three warps treads and leaving one back and on the next row I could catch one thread with the needle and leave three back ( illus 3 ) this created a similar indenting that was evident in the original. I continued to weave the fabric in this way, there were no looped threads and the weaving looked identical to my photographs of the original. However when the weaving was complete and I measured the fabric and it was at least 20 cm too long! This was turning out to be a very challenging project. I had to go back to my measurements and re- think where I had gone wrong. It had to be something to do with the thickness of the weft threads as the width was the right measurement. I had wrongly assumed that all the weft threads in the fabric were the same thickness I found when I calculated the widths of the chevrons to the number of rows that there were in fact four very different thickness of yarn used for the weft. This was very noticeable with hindsight the 42 row band measured 4 cm whereas the 38 row band measured 5 cm. I ascertained that the four distinct yarns were as follows 7 rows per cm, 8 rows per cm, 9 rows per cm and 10.5 rows per cm. These different thicknesses of yarn were erratically distributed throughout the fabric.

Having had some experience teaching groups of people how to spin on a spindle whorl I have found that people find their own thickness of yarn that they find easy to spin. Most students acquiring spinning skills find that they can easily spin an even yarn at their own personal thickness, some very fine yarn and some much thicker. Whereas a skilled spinner can spin any yarn thickness to order the average spinner tends to spin always at the same thickness. I suggest that therefore that there were four distinct spinners making the yarn for the hood. This would account for the uneven chevrons of the pattern. If a fine thread was added after a thick one this would form a ridge in the weaving and be noticeable, but if one always changed the direction of the chevron when a new yarn was added the difference is unnoticeable, as I found to my cost when making my first replication of the weaving. I continued to re weave the cloth, which I was content, was as near a reproduction of the original as possible. I made some very tentative calculations about the possible size of the original cloth, which I must emphasize, are pure conjecture. By calculating how many rows a full spindle of the various thicknesses could weave I deduced that the original cloth that the hood fabric might have been cut from could have been over one metre wide.