Weaving the Qizilchoqa Tartan

So, finally I come to weave the Qizilchoqa tartan. I'm weaving it now (August 2014) as my Guild is holding an exhibition at 'The Old School House' here in Qualicum Beach, entitled 'Keep Me Warm' and a tartan blanket using the Qizilchoqa pattern would make an interesting exhibit, especially if I can have some historical text beside it.

Choosing the yarn

I'm using Brown Sheep Company's 'NatureSpun' for the blanket mainly because of the close match in their colour range with the background colour. I've used it before and it produces a soft comfy-next-to-the-skin fabric and has a largish range of colours in four weights. I want to have the pattern shown as clearly as possible so I'm using the longer weft threadcount (612 threads) and so to get a couple of repeats of the pattern across a width of approximately 54 inches I will use the fingering weight. This is normally sett at 20 epi for twill but on sampling I found that 24 epi gave a firmer, more blanket-like cloth after fulling.


Choosing the colours

The major problem is finding a colour to match the background red / brown. This has been described to me by Prof. Elizabeth Barber as purply-red-brown. It was achieved originally by dyeing chestnut brown wool with madder. I have tried many suppliers looking for this colour, and the closest match I can find is NatureSpun's 'Brick Road' (225). Given the exhibition title I want to keep the other colours on the warm side, so I'm using:

Red (Originally Madder on white wool) - 'Red Fox' (N46)
Dark Brown (Natural dark brown wool) - 'Storm' (114) - a very dark brown
Dark Blue (Indigo on white wool)- 'Blue Boy' (116) which I find warmer than 'China Blue' (N36)
Light Blue (Indigo on white wool) - 'Blue Marina' (505) which is a shade darker than I would like, but it's warmer than 'Bit of Blue' (115)
White (Natural white wool)- 'Aran' (N91) which is the warmest of the whites.

The Qizilchoqa Tartan.


Designing the Blanket

In order to maximize the warm colours in the blanket I want to have the background squares that have red stripes in them to be central. So in the width the triple red stripe squares will be in the middle, flanked on either side by the empty background squares, and then with as much of the 'red' part as possible at the edges, so:

Br A Br C Bt A Bt C Bt A Br C Br A Br C Bt A Bt C Bt A Br C Br

And the same pattern for the length but with one extra repeat:

Bt C Bt A Br C Br A Br C Bt A Bt C Bt A Br C Br A Br C Bt A Bt C Bt A Br C Br A Br C Bt

Weaving the blanket

I produced a warp of 3 and a half yards to allow for the fringe and loom waste, consisting of 1292 threads, to give a width of just under 54". This was beamed and threaded from the back of the loom, using 8 shafts to reduce the density of the heddles. It was sleyed double in a 12 reed and tied to the cloth rod.

The Qizilchoqa Tartan being beamed.

The blanket was woven in a straight twill for just under 76" (1804 shots), changing weft colour as required. Where possible the background brown weft, and sometinmes the dark blue weft, were carried up the selvedge to minimize tie-ons and tie-offs.

After weaving, the web was cut from the loom, the ends were grouped in 2 cords with 8 ends each and twisted and knotted to produce a stable fringe. The web was then fulled by about 10% to give a final size of 48" x 68".

The Qizilchoqa Tartan Woven.

The Qizilchoqa Photograph - Pattern Analysis

This tartan was found in a grave in South-west China. It is dated to between 1200 - 800 B.C. and is one of the oldest examples of tartan. It is also one of the most complex, requiring over 600 threads to make the pattern.

The photograph of the Qizilchoqa tartan is found in "The Great Treasury of Chinese Fine Arts: Arts and Crafts Volume 6, Printing, Dyeing, Weaving and Embroidery (1)", editor Huang Nengfu and published by the Cultural Relics Publishing House in 1990. It is photograph 3, entitled "A piece of woollen fabric (called ji in Chinese) with coloured check design (Western Zhou period)". The piece was unearthed from the Wubao burial site in Hami, Xinjiang in 1978 and was photographed by Wang Luli. The piece is considered to be over three thousand years old. Unfortunately, due to copyright restrictions I cannot show the photograph, but below is a diagram I made of the photograph.

The Qizilchoqa Tartan.

The piece is 21cm tall by 16cm wide and is made of two pieces of fabric sewn together horizontally with light blue thread and joined together along the right hand side with braiding. Two 'bobbles' sewn in blue thread are attached halfway down the braiding at the join of the two pieces. Although these are two pieces of fabric, I am assuming that they came from the same fabric and hence share the same pattern, if pattern there is, rather than from separate warps with different patterns.

The photograph shows the public side of the fabric, as the bobbles are uppermost, and I am assuming that the warp runs vertically in the photograph and the weft horizontally, although it makes little difference when analysing the pattern. (This orientation show in the photograph is contradicted in the notes. Firstly the notes say that the twill runs "in the right direction" but the photograph shows the twill running to the left. If the photograph is turned by 90o swapping the warp and weft, then the twill runs to the right. The notes also state that the warp is sett closer than the weft (21 warp ends per centimetre, but only 16.5 weft picks per centimetre) which would also imply that the horizontal direction is the warp direction. However it makes more sense in weaving terms to have the warp sett less closely than the weft, that is the warp at 16.5 threads per centimetre and the weft packed more closely at 21 threads per centimetre, which would accord with the photograph's orientation.)

The Qizilchoqa tartan pattern is composed of four symmetrical pattern elements laid on the background brown:

W.darkB.W - two thick white lines enclosing a dark blue stripe,
w.lightB.w - two thin white lines, bordered internally with dark blue,
                    with a central light blue strip, also bordered with dark blue,
B.R.B - two dark blue stripes guarding a red stripe,
B.-.B - two dark blue lines that are filled with background.

Each brown background square is bisected by a thin dark brown stripe.

The warp sequence of these elements in the top piece (from left to right) is

W.darkB.W + B.R.B + w.lightB.w + B.-.B + W.darkB.W + B.-.B + w.lightB.w

And the bottom piece (also from left to right) is

W.darkB.W + B.R.B + w.lightB.w + B.R.B + W.darkB.W + B.-.B


The weft sequence in the top piece (from top to bottom) is

w.lightB.w + B.-.B + W.darkB.W + B.R.B + w.lightB.w (part)

And for the bottom piece (from top to bottom) is

w.lightB.w (part) + B.R.B + W.darkB.W + B.-.B + w.lightB.w

A common feature of all these sequences is the alternation of the w.lightB.w element with the W.darkB.W element.

Also, the top weft sequence is the reverse of the bottom weft sequence, and the two pieces have been sewn together to produce a symmetry with the axis at the join. Given the obvious attempt to make the piece symmetrical across the join, it is surprising that the two pieces weren't joined so that the warp sequences aligned to make a more satisfying piece. Possibly there was insufficient fabric to make this alignment possible and also the longer a pattern sequence, the more difficult it is to make this matching alignment.

As set out above there is no commonality between the warp sequences, so one would have to conclude that either the patterns were laid in randomly, or that they were part of a much larger pattern, of which these are part sequences.

However, if the bottom piece is turned through 180o the following warp sequence is obtained

B.-.B + W.darkB.W + B.R.B + w.lightB.w + B.R.B + W.darkB.W

If the two warp sequences are put together, top first, then the bottom (turned through 180o) and added to the right, the result is

W.darkB.W + B.R.B + w.lightB.w + B.-.B + W.darkB.W + B.-.B + w.lightB.w + B.-.B + W.darkB.W + B.R.B + w.lightB.w + B.R.B + W.darkB.W

which is a long but simple repeating sequence: the white dark blue stripes alternate with the white light blue stripes, and between these elements there is a simple sequence of three dark blue / red stripes followed by three 'empty' dark blue stripes.

Also, if the two warp sequences are put together the other way around, top first, then the bottom (turned through 180o) and added to the left, the result is

B.-.B + W.darkB.W + B.R.B + w.lightB.w + B.R.B + W.darkB.W W.darkB.W + B.R.B + w.lightB.w + B.-.B + W.darkB.W + B.-.B + w.lightB.w

which is the same sequence overlapping on the W.DarkB.W element.

(I also checked two other simple sequences: a symmetrical sequence following the weft pattern and an alternating sequence as above but with only two, rather than three, repeats of the red and empty stripes. The left warp pattern did not fit either sequence.)

The weft sequence, (from left to right) is part of this repeating pattern, spreading across the join of the first (right-hand) warp sequence

W.darkB.W + B.R.B + w.lightB.w + B.-.B + W.darkB.W + B.-.B + w.lightB.w + B.-.B + W.darkB.W + B.R.B + w.lightB.w + B.R.B

Could this have been one wide piece of fabric, cut vertically into two pieces, the left piece becomes the top piece (in the photograph) and the right piece is turned 180o and sewn to the bottom?

However that may be, the pattern above is the simplest that I can obtain from the photograph. If this is right, then I would predict that the warp pattern just outside the left frame of the photograph on both top and bottom pieces of fabric should be a red line guarded with dark blue (B.R.B). The pattern above and below the frame should be two empty dark blue lines (B.-.B).

The threadcount of the elements differ in the warp and the weft, the weft count being roughly twice that of the warp count, except in the brown background areas.
W.darkB.W
  Weft Warp
W 8 3/4
T 6 2
DB 4 2
T 6 2
W 8 3/4
  32 12/14
 
w.lightB.w
  Weft Warp
W 4 2
DB 4 2
T 4 2
DB 4 2
LB 4 2
DB 4 2
T 4 2
DB 4 2
W 4 2
  36 18
 
B.R.B
  Weft Warp
T 12 10
DT 4 2
T 12 10
DB 4 2
R 4 2
DB 4 2
T 12 10
DT 4 2
T 12 10
  68 50
 
B.-.B
  Weft Warp
T 12 10
DT 4 2
T 12 10
DB 4 2
T 4 2
DB 4 2
T 12 10
DT 4 2
T 12 10
  68 50
 

(where T = Brown (background), DT = Dark Brown, R = Red, W = White, LB = Light Blue and DB = Dark Blue.)

Using the longer weft count, the complete threadcount of the pattern is:

Qizilchoqa Tartan Threadcount

W8T6DB4T6W8    T12DT4T12DB4R4DB4T12DT4T12    W4DB4T4DB4LB4DB4T4DB4W4    T12DT4T12DB4T4DB4T12DT4T12    W8T6DB4T6W8    T12DT4T12DB4T4DB4T12DT4T12    W4DB4T4DB4LB4DB4T4DB4W4    T12DT4T12DB4T4DB4T12DT4T12    W8T6DB4T6W8    T12DT4T12DB4R4DB4T12DT4T12    W4DB4T4DB4LB4DB4T4DB4W4    T12DT4T12DB4R4DB4T12DT4T12 (and repeat from the start)

Or more simply: A  Br C  Bt A  Bt C  Bt A  Br C  Br

where A is W.darkB.W, C is w.lightB.w, Br is B.R.B, and Bt is B.-.B

The complete pattern thus requires a total of 612 weft threads, which (at 21 weft threads per cm) gives a width of just over 29 centimetres per pattern repeat. The total warp count is 396 threads for a complete pattern, which gives a length of 24 centimetres (at 16.5 threads per cm) per pattern repeat.


The Qizilchoqa Tartan.

The Qizilchoqa Tartan.