COLORS: W = white; B = dark indigo blue
Detailed Knitting Directions
The oldest true knitting found so far is from Muslim-period Egypt (Rutt) and dates from between the Fatimid period in the 11th-12th centuries (Tissus d'Égypte) and the Mamluk period, from 1250-1517 (Rutt) - actually the Mamluks continue to the end of SCA-period, since they continued as governors after Ottoman conquest, with little Ottoman cultural influence. It is almost certain that the technique of knitting began earlier, as these Medieval Egyptian knit items show skill and sophistication in their techniques, use of shaping, and complex patterning; however, no earlier pieces have yet been identified.
Several older textiles discovered in archaeological setting in the Near East were falsely identified as knitting. Among them are some small wool fragments from the city of Dura-Europos, Syria (destroyed circa 265 CE). There are a number of woolen ankle-high socks with separate big toes and fragments of them from Roman Egypt (c. 4th-5th centuries CE). A hasty glance suggests that they are knit, and early scholars erroneously claimed that were, which unfortunately is still promulgated over the internet and in print. However, in a definitive article published over 30 years ago, in 1972, Dorothy Burnham examined Romano-Egyptian socks and presented an analysis that shows they were not knit. First, the loops were made in a distinctive way, unlike knitting. Second, there are frequent little tufts inside where the end of one relatively short strand of wool and the beginning of the next were twisted together, which is unnecessary in knitting. Both use of particular technique and short lengths of wool show that the socks were nâlbinded (Burnham)
There are quite a few textiles from Muslim Egypt that are truly knit, shown by a close examination of their structure. Pieces of Medieval Egyptian knitting are in numerous museums. Some of these pieces are of multicolored wool, others of cotton. Among the cotton fragments are a number of individual stockings (I know of no pairs), some complete or nearly so, others fragmentary, many obviously parts of legs or feet (Bellinger, Plate I; Bush, pp. 13 &78; Gibson-Roberts, p. 20; Rutt, p. 35; Tissus d'Égypte, p. 268-269, 272, 273 ). The Textile Museum in Washington DC has over a dozen socks and sock fragments in their collection. I was able to examine them briefly in March 2003, including the sock reproduced here. I was not allowed to take photographs and there is no published catalog of their holdings. The photograph of this sock is from the cover of the museum's 1997 Looping and Knitting exhibition catalog. To chart the patterns of this sock, I relied on the photograph, and to reproduce the sock, on descriptions of techniques from several written sources (Bellinger, Gibson-Roberts, Rutt, Tissus d'Égypte, Zilboorg)
The archaeological context of most Medieval Egyptian knitting is unknown, as little has been found in situ by archaeologists. "Treasure hunters" dug up many pieces in the early 20th century before sites were properly excavated, then sold them to individual collectors, "antiquities" shops, and museums. Thus the exact provenance of much Medieval Egyptian knitting is unknown. More recently knit fragments have been found by archaeologists in Egypt in house ruins and dump sites. Second, it is unknown whether socks were knit by professionals, were a cottage industry, or were a homely art. Based on the variety of socks and sock fragments I examined in The Textile Museum, in March 2003, including this sock, I suspect that there were all three. The pieces I saw showed a wide range of technical accomplishment, but were all made of similar materials - loosely S-plied Z-spun cotton in off-white natural and dyed with indigo, either dark navy or both dark navy and light-medium blue. Third, it is not known if these socks were worn by a particular gender, age group, or social class. However, there exist socks both for adults and children, so at least a range of ages wore them. And finally, it is not known if these were for daily wear, for special occasions, or funerary garments. There is, however, a known history of socks in Egypt for daily wear, albeit nâlbinded, ankle-high split-toe socks in both adult and child sizes. And more conclusively, because those found today by archaeologists were found in house ruins and dumps, and not graves, it is likely that they were for daily or dress wear.
While the sample base of extant socks and identifiable sock fragments is comparatively small, certain patterns recur. This, regardless of who produced them, they seem to reflect a particular design vocabulary. For example, the indigo-and-white S-Z pattern on my socks is found on several other socks and sock fragments in a number of collections (Bellinger, Rutt, Tissus d'Égypte).
It it unknown what other items were knit. Besides socks, there are additional blue and white fragments, some probably from socks, some narrow tubes (garters? belts?), some uncertain, and smaller number of colorful wool fragments with extremely complexly knit complex patterns (Bellinger, Bush, Gibson-Roberts, Rutt, Tissus d'Égypte) which are unlikely to be from socks and may be from cushion covers, coverlets, or draperies of some sort.
II. Reconstructing the Socks
The shape of my socks and the patterning are based on the original, as can be seen from the photographs. The original foot was about 6 inches long; mine is 8 inches long, to fit me. Both the original leg and mine are about 8 inches long.
2. Ply and Twist
Of the archaeological example for which I could find information on the ply, four with patterns and colors somewhat similar to my socks were made of an S-twisted yarn composed of 2 plies of Z-spun cotton (Tissus d'Égypte, figures 168, 169, & 170)
As I do not know how to spin cotton, I purchased "Baby Georgia" made by Crystal Palace Yarns, which is an S-twisted yarn made of 4 plies of Z-spun 100% cotton. It appears to be as fine as the Egyptian cotton yarn. It is more tightly spun and plies than that in the Medieval Egyptian socks. Crystal Palace also makes a two-ply cotton called "Breeze", but it is thicker than the "Baby Georgia" which would affect the gauge, so I chose not to use it.
3. Color and Dye
I purchased commercially dyed yarn. I was thus limited to the colors available commercially. Fortunately I was able to find colors quite close to the original, in "Natural" (off-white) and "Indigo" (dark navy blue).
The tools the Egyptians used to knit their socks are unknown. One can conjecture that, based on the levels of technology and the fineness of much of the knitting, they were made of metal, probably brass. It is highly likely that they used double-pointed needles. Until a sock in progress is found in a dig, however, we are unlikely to know for certain. Rutt conjectures that they may have had a hook on one end like needles used in the Near and Middle East in the 19th and 20th centuries. But there is no evidence for or against this, and there is nothing in the construction of these socks to indicate the use of hooked needles. They can be knit on hooked needles, as I have used them to knit other items in the round.
I used five double-pointed metal needles, size US 1, to achieve a gauge close to the original.
C. Construction and Techniques
1. Basic Sock Shape
Medieval Egyptian socks were knit from the toe up, which can be determined by the direction of the stitches.
These old socks are not tube socks. They have a toe shaped by increases, a straight foot, and a heel shaped by decreases - either a short-row heel as in this example, or an inset/"after-thought"/peasant heel, as in other examples. A close examination of photographs, as well as the description by Bellinger (second page of unnumbered text), shows that the legs of the socks were shaped as they were knit, widened at the calf by the use of increases which were spread around the sock, not concentrated in the center back. Concentrated center rear increases are more typical of later European socks.
Patterns are worked in stranded knitting, sometimes today erroneously called "Fair Isle knitting", entirely in knit stitches. It is possible that they were knit inside out to control the tension of the stranded knitting, as is done in the Andes today. But there is no evidence that this was done, so I knit mine right-side out.
The gauge of other archaeological examples varies. Gibson-Roberts reproduces two socks from the Metropolitan Museum, one of which she says has 12 stitches and 14 rounds per inch, and the other at 10 stitches and 13 rounds per inch (p. 20). Tissus d'Égypte has seven example of Egyptian cotton knitting, although none are identified as being from socks. They are knit at a variety of gauges: (1) about 10 stitches and 13 rows per inch; (2) about 10 stitches and 23 rows per inch; (3) about 13 stitches and 15 rows per inch; and (4) about 15 stitches and 20 rows per inch. (I say "about" because the original gauges are given in centimeters and I've converted them to inches)
I have knit mine at about 10 stitches and 10 rows per inch, which is close to Medieval Egyptian gauge, although it would be closer if I had a few more rounds per inch. For some reason, however, I tend to knit rather "square", having close to the same number of stitches and rows/rounds per inch
Another way to make an increase without using the yarn-over involves knitting a stitch into the back leg and knitting a stitch into the front leg of a single knit stitch. This makes for a tighter fabric (as "make one" can leave a little space), but changes the texture slightly on the face of the fabric.
My examination of existing sock toes, both at The Textile Museum and in photographs, indicated that there were at least three ways to begin a toe, probably more. Because I could not handle the socks at the Textile Museum, much less turn them inside out, I could not determine the methods exactly. It looks as if in some cases a fairly large number of stitches were cast on, possibly around the tail end of the yarn, and then this tail end was pulled to bring the toe stitches close together like a drawstring.
In a second method, the toe was begun with a small number of increases and gradually increased in a regular manner, resulting in a less bulky toe and creating an interesting texture and pattern. I saw a single unpublished example made this way at the Textile Museum.
A third technique involves using two colors to make an interlocking diamond pattern. This is the type of toe used on the original of my sock. I am not exactly certain how this was done, so I used the figure-8 cast-on, augmenting Bellinger's description with instructions from books that made the process clearer to me (Gibson-Roberts, pp. 64-65 and 70; Zilboorg, pp. 19-20).
For the inserted heel, when the foot is long enough, as one knits around, the heel stitches are knit with a piece of waste yarn, which will be removed later. Bellinger explains how this heel is made (unnumbered text page two), including a photo of a modified modernized version of a Medieval Egyptian sock (Plate II). There are photos of two other Medieval Egyptian socks with similar heels (one in both Gibson-Roberts, p. 20, and Rutt, p. 35; and another in both Bellinger, Plate I, Accession number 73.698; and Bush, p. 13). Technical information on making this kind of heel is also in Gibson-Roberts (pp. 74-75). This heel is knit around beginning on 4 needles (with a 5th working needle), regularly decreasing until only very few stitches are left on two needles. These last few stitches are either knit together or grafted to form the heel point. Since I could not examine the insides of any extant socks, exactly how they ended their heels is unclear.
The sock here has the short row heel. I left half the stitches for the front of the sock on their needles but didn't knit them. Then I knitted the heel of the stitches on the two remaining needles. The heel is knit beginning at the foot edge, and is knit back and forth, decreasing regularly at the beginning and the end of each row, until there are two stitches left. When the heel was complete, I continued working the leg, knitting around the stitches for the front of the leg and picking up the edge stitches of each row of the heel.
8. Color Stranded Knitting
Bellinger comments that, because the float threads are not twisted in the wrong side in a manner though could cause the knitter problems, the two yarns must have been held one in each hand. However, I hold both yarns in one hand and don't twist them unless I am careless. I hold my yarns so one is always above the other, except when the top yarn has to be anchored by the other to prevent overly long floats. Carrying the yarns this way, the lower color tends to stand out slightly from the other, creating a slight texture. In these socks, the white was on top and the dark indigo below.
Bellinger also remarked that the sock patterns appear to designed so that there are no exceedingly long floats on the wrong side, and how any places where the floats would be too long (in socks these could be snagged by the toes) are anchored by being knit over. She does not mention how long this would be. I anchor a float when it will extend over 5 stitches or more, by laying the float thread over the working thread, then yarning over and drawing the working thread through the stitch, which anchors the floating thread on the back side.
Bellinger says, "...the final row was bound off by slipping one loop over the next" (second page of unnumbered text), which is a fairly common modern cast-off. This can be seen clearly at the top of the complete sock in Gibson-Roberts (p. 20). This appears to be the technique used in the original of this sock, so even though it is not elastic this is what I have done.
It is assumed that Medieval Egyptian socks must have been held up by some kind of garters (Bellinger, unnumbered text page two), although there is no conclusive evidence for garters. Without garters of some sort, purely stockinet knit socks will fall down. None of the texts I have read discusses what these garters would have been like. They may have also been knitting, as several long narrow knit tubes have been found in Egypt from around the same time period as the socks (Tissus d'Égypte, pp. n). Garters could also have been narrow woven bands, or even torn strips of cloth.
Plain stockinet stitch with a simple bind off tends to roll down the outside, where it could cover and secure a garter to some extent. In Medieval Egyptian knitted bags, a small cord was laced through the stitches themselves (conversation, Chris Laning). It is possible that a small cord was laced through the top of a sock as a garter, however I have not found evidence of this. At the time of this writing I have not worn the socks yet, although I tried them on while I was knitting to make sure they fit. I will wear them with a garter of some sort, although I haven't yet decided what kind.
Rutt includes a photograph of the back side of a knitted wool fragment, in which it appears that ends of colors yarns were just left hanging loosely (Rutt, p. 38). However, these wool knits were most likely not socks, so the hanging threads would probably not be problematic. Gibson-Roberts says that in traditional Eastern knitting (from Eastern Europe, the former Ottoman Empire, and Southwest Asia), as colors are finished and added, the ends are knitted together. And when the sock is finished, the ends are not woven in. (p. 79).
I chose to "weave in" loose ends. This makes for fewer chances of catching a thread when I insert my foot, and potentially damaging the sock or hurting a toe, and fewer chances for the threads to work loose when the socks are washed.