Crochet · Early instructions · Structures · Techniques

Twists and turns in the development of crochet — Part 2

This is a direct continuation of the preceding post. I’ve also tweaked its initial version to mesh better with the following text and even readers who have already seen it may find it worth reviewing before proceeding.

The first installment presented two source documents from 1846 and 1842 in reverse chronological order. This one works forward to 1842 from the first English language text that explicitly describes crochet — The lady’s assistant for executing useful and fancy designs in knitting, netting, and crotchet work by Jane Gaugain, from 1840.

This differs from the other documents under consideration by using the term “tambour” to designate a crochet stitch. This was taken from tambour embroidery, which together with shepherd’s knitting were the two immediate precursors of the craft all the authors term crochet.

Gaugain presents the chain as the basic tambour stitch and doesn’t differentiate between it and the slip stitch. Her compacted definition covers both the free chains of openwork mesh and the kinds of objects other authors say are typical for “Shepherd or Single Crochet.”


This is worked by drawing one loop through the other; it is seldom used save for open purses, and sometimes for muffattees, shoes, &c. &c.

The instruction that follows is for a “LONG PURSE OF OPEN STITCH OF SINGLE TAMBOUR” in the same arched mesh that appeared in the Dutch instruction from 1823 discussed in an earlier post. The next two instructions describe working double crochet with the alternate techniques of turning the fabric at the end of a completed row, and cutting the yarn at the end of every row and beginning all new ones from the same edge.


The purse is alternately worked on the right and wrong side… Cast on 100 loops in single chain stitch, having the last of the cast-on loops on the needle. 2d row, insert the needle in the first loop, and catch the silk from behind; pull it through the loop. You now have 2 loops on the needle, then catch the thread, and pull it through the two loops; this forms one stitch…


This is…all worked on one side. When you come to end of the row, cut off the thread, and draw it through the last loop, which fastens it. 2nd row, commence at the same stitch which you began the last row on…

Gaugain added a second volume to the 1842 edition of her text. This includes a detailed explanation of the turning chain.

Edge Stitch. — This stitch is worked by drawing a loop through the first loop or stitch on the row or round, then another loop through the one just made. This forms the edge stitch; then work on through the pattern. If the edge stitch of every row were not worked in this way, you would lose a stitch each row. It is not necessary to work an edge stitch on a round, but only where the work requires to be turned to the wrong side, in order to work round the other way.

A footnote (on p. 279) mentions the need to match the height of the turning chain to the stitches in the row it commences.

Again turn the work, and work back this round, and make an edge stitch; but instead of making one loop or chain stitch, make two, as this open stitch is higher than the plain rows.

Gaugain describes open stitches of two different lengths. One is now called half treble crochet (UK) and the other extended treble crochet (with an extra chain at the top of the post). She also modifies her terminology midway through the added volume, shifting from referring to a stitch as a “tambour” to the synonymous term “crochet,”

Worked in double tambour or crochet stitch, as described in the 125th Receipt of Volume First.

The first volume of the 1842 edition was unchanged from 1840 and the 125th instruction is excerpted above. Despite the statement about the limited use of the single tambour, the second volume includes several instructions for it. As with the statement about using an edge stitch for turning work to the wrong side, at least one of Lambert’s instructions for what she now consistently terms “single crochet” also moves the back of the fabric to its public face.


This Boot is worked exactly as the long and short Mittens for children, pages 317—323…[however] when finished it is turned out…

(Mittens, bootees, and the public display of the back side of the fabric will all figure prominently in the discussion of pre-19th-century shepherd’s knitting in subsequent posts.)

Cornelia Mee describes the same two basic flatwork methods in A Manual of Knitting, Netting, and Crochet Work from 1842. She doesn’t ascribe stitch status to the chain and presents the slip stitch as the:

Elementary stitch of crochet

First, make a chain, by making a loop and drawing one loop through the other, till it is of sufficient length; this forms a foundation; and all other crochet work must be begun in this way; then pass your crochet through the end loop, and taking up the wool or silk, draw it through; repeat this in every successive loop to the end of the row; then turn it, taking the under loop, and continue backwards and forward in the same way: this forms a ribbed kind of crochet, and is the most simple to begin with; the appearance of both sides is the same.

She then enters the familiar numbered sequence with:

Double Crochet

This may be worked round and round or in rows; if the latter, you must always break off at the end of every row, as it cannot well be done backwards and forwards, especially if intended to have a pattern on it. After the foundation is made, you will have one loop on your needle, insert the crochet through the next, and then draw your wool or silk through both; this still leaves a loop on your crochet, keep this on it, and draw your wool through the next loop, and then through both. When you come to the end of a row, draw the wool through the last loop, and cut it off, leaving an end of about three inches.

There is a key difference in visual effect between flatwork crochet turned after every row, and with all rows starting at the same edge. This is further influenced by other structural details that the Victorian texts leave almost without comment. I’ll go further into this in the next installment.

Crochet · Early instructions · Structures · Techniques

Twists and turns in the development of crochet — Part 1

Instructions for crochet began to appear in the Victorian fancywork press in 1840, presenting it as a more elaborate and fashionable successor to the long-practiced shepherd’s knitting — the slip-stitch fabric made on a flat hook discussed in several previous posts. That stitch appeared in instructions for the new craft but opinions about its utility varied and its role at the urban worktable began to shift from the production of fabric to an adjunct technique.

The double crochet (UK) supplanted the single crochet (later aka slip stitch) as the workhorse for closed work. However, in contrast to present-day practice where a new such stitch is normally worked under both legs of the chain at the top of the stitch it is anchored to in the preceding row, the Victorian instructions prescribe working through only one leg of that loop. (The current terms “front loop” and “back loop” refer to the respective legs and are better read as “front leg of the loop” and “back leg of the loop.”)

Another difference between early Victorian and current practice is that fabric worked flat was not commonly turned at the end of each row. This technique does appear in the first published instructions (discussed in Part 2) but the alternative of cutting the yarn or thread at the end of each row and starting all new rows from the same edge was generally preferred. The further option of crocheting in alternating directions without turning the work was also described.

I’m going to proceed by discussing such aspects of Victorian crochet and then move back to details of shepherd’s knitting that have long since disappeared from mainstream practice. Beyond the obvious purpose of describing procedures that have come and gone during the technical development of the craft, I want to call attention to crocheted structures that might not be recognized as such when assessing the method of production of non-woven fabric with an unfamiliar appearance. There are probable instances of this having happened during the analysis of some objects published in the archaeological literature, which I will also go into in detail.

In order to provide an easy path into unaccustomed terminology, I’m going to begin where the current glossary has its clearest roots and work backward from there. Two relevant texts are covered below and others will be considered in the following post.

In her Knitting, crochet, and netting from 1846, Mme. Riego issues the general instruction that:

In crochet that is worked square, at the end of a row, cut the wool off, and draw it through to fasten it; begin at the other end.

The first of the following stitch descriptions is for:

Shepherd or Single Crochet.
This stitch is usually worked round, for Cuffs,
Mufatees, Boots, &c. &c

Make a chain, join it, keep the loop on the needle.
1st round — Put the needle in the 1st chain stitch, draw the wool through; there will now be 2 loops on the needle; draw the last loop through the 1st.
In the 2nd and following rounds, take the 1st part of the chain on the needle.

Riego continues with “Plain, Double, or French Crochet” and then “Treble Crochet.” Despite the familiar labels, there are two significant differences between her single, double, and treble crochet fabrics and their contemporary counterparts. These were discussed above and are the cutting of the yarn at the end of each row worked flat, and routinely crocheting each stitch through only one leg of the loop in the chain along the top of the preceding row — i.e., into the loop and not under it.

The latter procedure produced a structure that was described with a name of its own by Frances Lambert in The Handbook of Needlework from 1842.

Double stitch crochet.—In this, both meshes of the chain are taken. It is principally employed for the soles of shoes, and where extra thickness is required, but it is not suitable for working patterns.

Lambert uses the term plain crochet or plain single crochet to designate a slip stitch, and plain double crochet for what became the familiar double crochet. She notes of it that:

This is the crochet stitch generally practised, and that used for working table-covers, etc.

She then describes rows of both the single and double forms of plain crochet being worked in two directions using yet another term that was not taken into common usage.

Plain stitch elastic crochet—is worked alternately in rows backwards and forwards, first taking the upper, and then the under mesh of the chain.

Both this and the following descriptions of that process can be read as reversing the direction of the crocheting at the end of each row without turning the work. However, the indications of direction may also be relative to the front side of the fabric, facing toward and away from the worker in alternate rows (as in contemporary knitting charts).

…a chain of sufficient length is made to serve as the foundation for the article it is intended to make. Pass the needle through the last made loop of this foundation, and, catching the silk, draw it through, repeating the same at every successive loop; then returning along this row, repeat the same to form a second. A repetition of which, alternately backwards and forwards, from right to left, and from left to right, will give the first and easiest lesson.

Lambert abandons working in two directions (however it was executed) in the immediately following first complete instruction, for “a sofa pillow or table cover,” and all others in the same section.

This pattern, be it understood, is merely given as the first and easiest pattern in crochet, for the purpose of teaching the stitch… Instead of working the rows backwards and forwards, as before described, begin each row separately at the same end. When the last stitch of each row is finished, draw the wool through, and cut it off, leaving an end of three or four inches.

There is similar ambiguity in an instruction under a later heading. This indicates that cutting the yarn at the end of every row is the ordinary method. It therefore seems likely that the following snippet should end with a reference to “plain stitch elastic crochet” (as presented above) and not the contradictory “plain crochet.”

Raised, or ribbed crochet is worked in rows from right to left, according to the ordinary method; but the side of the work is reversed at every alternate row, as in plain [ribbed] crochet.

Lambert also recognizes the significance of what is now termed a turning chain.

To make a stitch — at the commencement and end of a row, is to make one stitch of a chain before the first stitch, and after the last, which in the next row are to be crocheted.

This is described in greater detail as an “edge stitch” in another text from 1842 and will be considered further in the next installment.

Crochet · Knitting · Structures · Tricot · Tunisian crochet

The chain gang

The preceding two posts discussed inconsistencies in some of Irene Emery’s remarks about the fabric structures she calls plain knitting (here) and plain crochet (here). Her narrative continues with the observations:

In plain knitting all the loops in one row are on the same face of the fabric… One face has a smooth surface with an appearance of vertical chaining which emphasizes the vertical alignment of the interlooping


Crochet is really a doubly interlooped structure (made with a hooked implement)… It is basically a kind of chaining… In the simplest stitch — plain crochet — each loop is drawn through two previous loops, the corresponding one in the previous row and the previous one in the same row.

The first statement says the vertical chaining on one face of plain knitted fabric is not a structural attribute but only has the appearance of one. Nonetheless, the chains that create that appearance have a demonstrable structural coherence. This is seen when the loop at the top of a chain is dropped from its support and then runs through the stitches below it, forming a “ladder” of horizontal bars between the corresponding stitches in the adjacent intact chains. The vertical structure is commonly restored by using a crochet hook to loop each bar in the ladder around the bar above it in a process that can as reasonably be termed chaining as can the mechanical action fundamental to crochet.

The second statement presents two requirements for fabric to qualify as crochet. An individual crochet-type chain intrinsically fulfills one of them; each loop is drawn through the previous loop in the same row. However, a chain does not meet the additional condition of also being drawn through the corresponding loop in the preceding row. A row of chains is therefore not crochet in itself and only becomes so when it is worked into an adjacent row of chains.

The challenge addressed by the present series of posts is devising a descriptive framework that includes not only plain crochet and plain knitting, but also plain tricot (Tunisian simple stitch). This is procedurally more closely affiliated with ordinary crochet than it is with knitting but the same cannot be said about the fabric’s structure.

One of the definitive attributes of tricot is that a complete row of stitches requires two passes across the fabric. The forward pass resembles knitting in both procedural and structural regards. The return pass is similarly akin to crocheted chaining. However, its purpose is to return the tip of the hook to the starting edge of the fabric and does not necessarily add a crocheted structure to it.

This is clearly illustrated with plain tricot. The return chains can be removed from such fabric entirely, transforming it into plain knitting. If the vertical knitted structure is removed from plain tricot, the fabric simply disintegrates. However, there are tricot stitches where the forward row is worked directly into the preceding return chain. This interlooping qualifies them as crochet (again, by Emery’s definition) and there is an array of intermediate states.

Since both the horizontal and vertical structures in plain tricot can be described as chains, as can plain crochet and plain knitting, an obvious next step is to consider whether the chain can serve as the modulating factor in an overarching categorization scheme. The difference between its two forms is that a crochet-type chain is made by placing a single loop on a support and working a new loop through it in immediate succession. Although such a chain has no inherent directionality, when interlooped into fabric it is positioned horizontally. Knit-type chains are made by first placing a row of loops on a support and then successively working a new loop into each of the preceding ones. This forms a vertical structure with as many parallel chains as there are loops on the support.

The loops in the two types of chains can themselves be differentiated in terms native to industrial knitting. A distinction is made there between symmetrical and asymmetrical loops.


It can readily be seen how the symmetrical form relates to plain knitting.


The asymmetrical form (fundamental to warp knitting) only needs to be envisioned horizontally for its relationship to plain crochet to be similarly apparent.


The height of both knitted and crocheted fabric grows as new chains are added to it. However, since chains of asymmetrical loops lie flat across the working edge of the fabric, the vertical increments are smaller than those of chains made from symmetrical loops of the same diameter. That difference is recovered and can be dramatically exceeded by a structural device specific to crochet. This adds height to a row of what would otherwise be plain crochet by interposing a “post” between each new chain and the chain capping the corresponding stitch in the preceding row.

The length of a post is determined by the number of times the yarn is wrapped around the hook during its production, and the position of each “yarnover” relative to the loop(s) in the preexisting fabric into which the hook is inserted. This variation is labeled with the familiar (albeit ambiguous) sequence double crochet, triple crochet, etc., each with an extended form that adds a vertically oriented chain to the post. It is also possible to insert the hook into the post in a preceding stitch, or through a combination of loops and posts, adding further variation to the repertoire.

Crocheted posts are structural elements of tricot, together with crochet-type and knit-type chains. As noted, many tricot stitches combine them in manners that permit the resulting fabric to be characterized as a structural hybrid of knitting and crochet. However, other tricot stitches cannot be described adequately in terms of either. The remaining tricot-specific structural attribute(s) will be considered in the next installment in this series (after revisiting a few other topics).

Crochet · Structures · Systematics · Terminology

Crochet plain and simple

Irene Emery takes care to distinguish between procedural and structural detail when describing the primary fabric structures included in her classification system. (This post is a direct continuation of the previous one, which includes additional information about the source references.) She also separates the core definitions of categorized structures from contextual discussions of selected aspects of their manufacture and application.

Those narratives include statements of Emery’s personal opinion about systematics and terminology in prior literature. Her remarks are more likely to address misunderstandings about knitting than anything related to crochet but she does emphasize the incorrectness of the routine identification of “fabric…as crochet simply because hooked needles were used to construct it.” The defining and differentiating attributes of the two structures are:

Knitting and crochet represent the two major types of interlooping. In knitting the interlooping is vertical and the loops are vertically aligned, each loop securing the corresponding one in the previous row. Crochet is characterized by both vertical and lateral interworking of loops, and each new loop (or the series of loops constituting the new stitch) secures the one before it in the same row.

With allowance for the differences in the way Emery relates closed-loop and open-loop knitting to the parent category of looping (discussed in the preceding post) her remarks about knitted fabric and illustrations of it are consistent with corresponding material in the craft literature. The same cannot be said of the way she treats crochet.

Emery uses familiar UK terminology for crochet stitches, with the exception of the slip stitch or single crochet (synonymous in UK, different in US; she says nothing about the separate glossaries or her departure from them). She terms this a plain crochet, seemingly coopting a term that appeared in the Victorian fancywork press. However, in that context plain basically meant what is now termed front loop only (FLO) and applied to both single and double crochet.

Emery also uses the name “simple stitch” without italics in narrative text and doesn’t indicate if it is a synonym for plain crochet. A description of how the “simple stitch can be elaborated” seems to use it in that sense.

If a second loop is added before a new [simple] stitch is complete, the stitch may be called ‘double,’ if three are made, ‘treble,’ and so on.

The same clarity does not apply to the caption under a photograph of square filet mesh where both the horizontal chains and the treble crochet (UK) vertical separators are described as “Simple crochet stitches combined to form open meshes and solid areas.”


Anyone reading that caption without having studied the narrative text would likely have difficulty understanding the labeling of the stitches. Even at that, if simple and plain are properly read as synonyms, the caption does not correctly describe the treble crochet in the mesh. By Emery’s definition this stitch is made by adding two extra loops to a simple crochet — i.e., slip stitch — before the stitch is completed. This is technically correct but arguably an oversimplification. It is not equivalent to the erroneous caption description of a treble crochet as a combination of simple stitches.

Seiler-Baldinger apparently attempts to disambiguate the terminology by using Emery’s “simple crochet stitch” to designate a chain. However, by Emery’s definition crochet is interworked both laterally and vertically. Since a chain is worked without running anchorage in any adjacent structure, it does not qualify structurally as crochet in itself, even if chains are fundamental elements of crocheted fabric.

Adopting Seiler-Baldinger’s definition, instead, and accepting the chain as crochet would have staggering implications for the history of crochet. Chains are encountered in many other contexts and are counted among the universal constructs that date back to early stages of human invention — devised independently at an indeterminable number of times and places. If a chain is crochet, the craft did not originate in the late-18th-century but in deep prehistory. I don’t believe that Seiler-Baldinger’s intended any such implication. A chain is an “air stitch” in German – Luftmasche – which I suspect was conflated in back-and-forth translation with the German word for slip stitch – Kettenmasche – literally meaning “chain stitch.”

Both authors do use “plain crochet stitch” to designate a slip stitch, which is the smallest structure that meets Emery’s definition. There is nothing to be gained in current discourse by substituting the unfamiliar label plain crochet stitch for the established slip stitch, which has the advantage of being the only label that designates the same stitch in both the UK and US glossaries. However, plain knitting frequently appears in the comparative discussion of textile structures and it is useful to have a corresponding plain crochet.

The series of posts to which the present one belongs has the goal of defining Tunisian crochet in terms that Emery or Seiler-Baldinger might plausibly have used had either of them chosen to cover it. Although this is a potentially intricate hybrid of crochet and knitting, there is a ubiquitous basic form of such fabric that can be compared nominally and structurally to plain crochet and plain knitting.

It was known as “plain tricot” in the Victorian literature, where it was first described in 1858 (in three separate publications), and only contains structural elements that are found in plain crochet or plain knitting. The next post in this series will consider ways of clarifying Emery’s definitions of both crochet and knitting so that they lead more directly to an equivalent statement about the composite fabric structure.

History · Knitting · Musings · Structures

First impressions of knitting

I’ve been diverting a fair amount of time that would otherwise have gone into blogging, to the preparation of a paper for the In the Loop at 10 conference at the University of Southhampton at the end of next week. Its title is Taking a Loupe to the Loop and it reviews some of the topics I’ve covered here, as well as providing fuel for coming posts.

Just as I was beginning to feel that I had the bulk of the work behind me, an article by Claudia Sagona was published on 25 June in the Oxford Journal of Archaeology, titled “Two-Needle Knitting and Cross-Knit Looping: Early Bronze Age Pottery Imprints from Anatolia and the Caucasus.” There is nothing unusual about an archaeological study of textile impressions found on ceramics — such things vastly outnumber prehistoric specimens of worked fiber — but this one claims to have found evidence of open-stitch stockinette knitting significantly earlier than had previously been attested.

That made it obligatory reading for me and it had also caught the attention of others. Its discussion was on the program of the Knitting in Early Modern Europe (KEME) seminar in Copenhagen on 7 July. I wasn’t there but understand that the participants were highly critical of the article’s methodology and conclusions.

I’ll be taking two of the drums that I’ve been beating on this blog with me to England. One calls attention to the limitations of our ability to identify the tools and techniques used to produce a piece of looped fabric on the basis of the object itself — a concern that only increases when assessing an impression of its surface detail. The second one signals that the earliest known specimens of knitted (as opposed to cross-knit looped) fabric have an open-loop structure, and not the twisted stitches that are frequently said to be the initial form.

Assuming that Sagona has plausibly identified similarities between the structures imprinted on the pottery fragments and those that could be left by looped textiles, it does not follow that that is how they were made. Her further association of those impressions with specific production techniques is entirely unsupported. She doesn’t explain why the fabric in the open-knit impressions would have been made with two knitting needles. The same structure can be made on a peg loom, which other authors have suggested was the first implement used for that purpose, to say nothing of other methods that may have been known six thousand years ago that we have yet to recognize. Similarly, the basic structure of cross-knit looping is identical to that of twisted-stitch knitting — again makable on both a peg loom or with knitting needles — and the impressions of that structure do not reveal the secondary detail necessary to differentiate between the two production methods.

The characterization of open-stitch stockinette knitting as ‘two-needle knitting’ implies fabric being worked flat with alternating rows of the knit and purl forms of the stitch. This is counter to the evidence of yarn having been knitted exclusively in the round during the first several centuries of the craft’s practice but Sagona explicitly states her belief that the impressions can be of fabric knitted flat.

Notwithstanding, moving the earliest direct evidence of the open-loop structure from the mid-first-millennium CE to possibly as early as the mid-fourth-millenium BCE would be of blockbuster consequence and therefore requires more rigorous substantiation than the article provides. (The earliest known specimen of cross-knit looping is from the mid-sixth-millennium BCE, found at a site near the Dead Sea, so there are no surprises there.) This wouldn’t alter the chronological order in which cross-knit looping and open-stitch knitting appeared. However, since the latter structure is a hallmark of true knitting, pending the unexpected appearance of corroborating evidence during whatever debate Sagona’s article triggers, it may be that the craft can be traced farther back than anyone else has yet suggested.


History · Knitting · Structures · Terminology

Who said knitting started with twisted stitches and hooked needles?

Several previous posts refer to generally held beliefs about the earliest knitters in Egypt using needles with hooked tips to make twisted-stitch stockinette fabric. More recent scientific examination of archaeologically recovered knitted fabric has radiocarbon dated the oldest known specimen of true knitting to the interval 425–594 CE. Counter to what the established tenet leads us to expect, this has an open-loop structure. Additional knitted objects through to the early 2nd millennium CE, found (but not necessarily made) in Egypt, have undergone similar examination and images in the published reports suggest that the open-knit form was by far the predominant one.

The questioned notion about the developmental sequence was fostered by Fritz Iklé in an article titled Über das Stricken (About Knitting), published in 1936 in the Schweizerische Arbeitslehrerinnen-Zeitung (Swiss Trade Teachers Journal, vol. 19, no. 8).  He discusses the earlier conflation of looped fabric with a cross-knit structure made with a single eyed needle, now generally regarded as a form of nalbinding, and true knitting. However, he characterizes the earliest knitted material as having a twisted-loop structure. The article includes a section on knitting with hook-tipped needles and he draws the conclusion that the use of such tools to produce twisted-knit stockinette was “apparently the form of knitting that preceded our customary knitting.”

Iklé then discusses later regional schools of knitting that employ hooked needles, noting that they are also used for open-knit stockinette. He illustrates this with a photograph of an unfinished sock.


“The beginning of knitted work from Turkey shows us that hooked needles can also be used to knit open stitches, for which we also have evidence from Arabic graves from the 9th to the 12th centuries…”

Iklé cites the work of Luise Schinnerer during the 1890s (discussed in detail in the following post), who was the immediate source for several of the ideas that he propagated. Their conclusions would less likely have been reflected in the English-language literature if Mary Thomas had not picked up on Iklé’s article in the preparation of Mary Thomas’s Knitting Book, from 1938. She explicitly credits him as one her informants and appears to have paraphrased the caption of the preceding photograph but misread a pivotal detail in the original German.

“…a half-finished sock of the 12th century found in a Turkish tomb reveals that the knitter was then working with five hooked needles…”

Either way, Thomas does not identify the Turkish tomb to which she is referring or substantiate her statement in any other manner. It should also be noted that if the Turkish sock illustrated by Iklé were indeed from the 12th century, despite his saying nothing about its age, its ribbing would provide incontrovertible evidence of knitting and purling side-by-side at a significantly earlier date than can otherwise be attested. Finding needles in place in work of that age would also be quite sensational.

Thomas bases what is now an entrenched dichotomy between “Eastern” and “Western” knitting on the assumption that fabric produced in the corresponding areas of Europe can be characterized by preferential knitting with twisted or open loops. Although the geographic labels are of mnemonic utility when discussing craft practice (for a right-handed knitter, loops with an Eastern mount face to the east and those with a Western mount to the west) the derivation of her nomenclature is gainsaid by the open-knit Turkish sock.

Thomas discusses and illustrates another regional form of knitting with hooked needles practiced in Landes, on the Atlantic coast of southern France. This is also mentioned briefly by Iklé with details that Thomas includes in her own description. She says that “the fabric is Crossed Stocking Stitch, knitted in the Eastern way” again contradicting the geographic basis for her classification of stitch structures. She resolves this by permitting both the Eastern and Western forms to be “crossed” or “uncrossed,” further treating knitted and purled stitches as separate constructs. Although again useful in knitting pedagogy, that model occludes rather than clarifies historical and structural relationships between the various forms, as well as the differentiation of the techniques used for their production.

Whatever the extent of Thomas’s reliance on Iklé may have been, he provided her with at least one item that is not described in his own text — a knitted fragment in his collection.

Iklé fragment

Thomas calls it a “magnificent example of Arabian color knitting of the 7th to 9th centuries…found in Fostat…and knitted in Crossed Stocking Stitch (Eastern)….” This dating is consistent with Iklé’s general appraisal of such material. Thomas notes that the decoration was knitted upside-down and that she aligned the photo with the direction of the stitching. This apparently assumes that the cuff-down working of the Turkish sock was normative for early knitting, rather than taking the orientation of the pattern to indicate what, in reference to socks, would be toe up.

When inverted, the pattern can be compared directly with the appreciable amount of decorated Islamic knitting from the Fatamid Period (969-1171 CE) for which descriptions have since been published. Several commentators have suggested on this basis that the Iklé fragment is also correctly dated to that period. The current location of the fragment is unknown (if it still exists) and its age cannot be determined more precisely. (Nor can the assertion that it has a twisted-stitch structure be confirmed; the photo does not display that level of detail.) For as long as it was considered illustrative of the earliest form of true knitting, the photograph in Thomas’s book was regarded as particularly valuable documentation.

One of the more rigorous recent discussions of Egyptian textiles is found in a presentation of selected objects from the Katoen Natie collection in Antwerp, written by Antoine De Moor, Chris Verhecken-Lammens, and André Verhecken, titled 3500 Years of Textile Art, and published in 2008. This includes photographs of a knitted stocking and fragments of three others that were all radiocarbon dated to the Fatamid period. The four photos are detailed enough to show open-knit structures, as can be seen in the full stocking here. The decorative pattern of one of the fragments closely resembles that of the Iklé fragment and a close-up detail shows its stitch structure with particular clarity (here in a detail of the detail).


This corroborates Iklé’s report of open-loop knitting being evidenced by material found in Arabic graves from the 9th to the 12th centuries, even if he estimated an earlier date for the fragment from his own collection. It seems likely that he based that assessment on the presumption that twisted-loop knitting was the older practice. However, the bulk of evidence now available does not support either the chronology or distribution statistics he described and Thomas then injected into the mainstream craft literature.

History · Knitting · Sprang · Structures · Tools

Early knitting with hook-tipped needles

Texts about knitting often present needles with hooked tips as limited regional preferences to the commonplace smooth-tipped needles. The hooked form is considered to be the older of the two and initially used in Egypt where the craft is also believed to have originated. Such needles have been documented in Eastern Europe, Portugal, and Southern France — all on or near routes along which knitting would have been carried into Europe from Egypt before diffusing more widely.

One problem with taking this as an indication of hook-tipped needles in prior Egyptian practice is that the presumed origin of knitting there and the more recently observed distribution of hook-tipped needles, are circularly taken to provide evidence of each other. Another problem is that there is no way to determine if the local European hooked-needle traditions all date back to the period when the Egyptian technology would have been transferred, or if they reflect later intra-European cultural exchange or even independent development. (This becomes particularly interesting when considering the relationship between the knitting carried to South America by Iberian colonialists and the indigenous looping traditions, which were often of breathtaking intricacy, in practice before European contact.)

Needles with smooth tips were illustrated in European portraits of “Knitting Madonnas” not all that long after the craft would have arrived in the locations where the paintings were made. The earliest is Italian from ca. 1345. Perhaps the best known of the genre, and certainly the one with the clearest detail, appears on the Buxtehude Altar from ca. 1400. (A set of five needles is shown in an illustration of a sock being knitted from ca. 1450.)


This either gainsays the northward importation (or at least its dating) or indicates that the hooks were dropped in rapid order thereafter — assuming that they were present in the first place. Assessing this also requires account to be taken of the Celtic and Viking knitting in wire prior to the putative arrival of knitting with non-rigid fibers via southern Europe.

It is also often suggested that the first device used for narrow tubular knitting in Egypt was a peg loom. This is based on the relative ease with which such implements can be used to knit all of the stitch structures seen in extant material, with a single hook. If that is accepted as correct, it could then be posited that needle knitting was spawned by the realization that a number of identical hooks could be used in a manner that obviated need for the loom. However, it could be suggested conversely that loom knitting is the younger technique, devised to provide a more easily mastered alternative to multi-needle knitting, also reducing the battery of needles — whether smooth or hook tipped — to a single hook.

There is no conclusive way of determining if loom knitting and needle knitting appeared sequentially in the same lineage or developed independently, perhaps even before either was introduced into Coptic Egypt. Nor is there any reason to assume that both methods were practiced there. Turning the perspective instead to structural detail, the simple and compound open stocking stitch observed in early tubular knitting can be characterized by the parallel chains that are formed at right angles to the path of the yarn as each row is worked into the preceding one.



Such chains also appear in the significantly older plaited fabric widely termed sprang, which is worked in two symmetrical halves toward a central “meeting line.” If this is retained in the final object (which is otherwise split into separate halves), it is secured by one of several methods that include looping a row of chains across the fabric. The plaiting is normally done with bare hands but the room for maneuvering decreases as the two halves approach each other, and the chained form of the meeting line is therefore made with  a hook. (I’ll revisit this all in greater historical and technical detail in separate posts and, as sprang is a Swedish word, will also examine relevant documents in which it appears.)

Here is a detail of the simple chain meeting line in the oldest known exemplar of sprang — a cap found in a Danish grave from ca. 1300 BCE,


and another of a compound chain meeting line in a hairnet found at a Danish site from ca. 490 BCE.


Chains also appear in multiple adjacent rows at meeting lines and as decorative ridges in other positions. This structure is properly classified as knitting-type interlooping and is an element of sprang even if not primarily identified with it. As with the early contemporaneous manifestations of remarkably similar forms of needle binding and tubular knitting in Northern Europe and Egypt, sprang also appears in both regions. (Extensive additional material about Nordic and Coptic sprang are found in the monumental Ancient Danish Textiles from Bogs and Burials by Margrethe Hald, who we met in an earlier post about nalbinding and whose illustrations of meeting lines appear above.)

Coptic sprang survives from the 4th century CE and the craft was practiced in Egypt through and beyond the entire period during which knitting developed there. It thus seems reasonable at least to wonder if the simple and compound knitted structures found in sprang, and their production with a hooked needle, provided impetus to the development of true knitting as initially seen in the tubular pieces that began to appear during the 5th century CE.

Knitting · Structures · Techniques · Tools

More about the structure of early Egyptian knitting

A previous post discussed several pieces of tubular knitting reported to have been made in Egypt during the 1st millennium CE. (Thanks to Matthew Pius for spotting the earlier studies and guest blogging their central details, summarized and commented on below.) One of the tubes had been radiocarbon dated to the interval 425–594 CE (in this test report) and is thus the oldest such object of substantiated age that has yet come to light. It is described by Dominique Bénazeth in an article titled “Accessoires vestimentaires dans la collection de textiles coptes du musée du Louvre” in a report from a conference in 2009, published in 2013 as Drawing the threads together; Dress accessories of the 1st millennium AD from Egypt.

The direct indications of this item having been made in Egypt are too weak to establish its provenance. However, its similarity to other knitted tubes that have a better-established Egyptian nexus — particularly one that is currently at the Museum of Byzantine Art in Berlin — was accepted as sufficient for grouping the Louvre piece with them. The Berlin tube is not clearly dated but is of certain geographic provenance and has been examined in greater structural detail. A report about it appears in an exhibition catalog from 2010 prepared by Klaus Finneiser, Petra Linscheid, and Meliné Pehlivanian, titled Georg Schweinfurth; Pionier der Textilforschung und Afrikaforscher. Their structural analysis concluded that it was likely to have been knitted on a peg loom, which is generally an efficient production device for tubular knitting. Since the Paris tube is both longer and narrower, the same method of production was assumed for it, as well.

The fabric structure of the Berlin tube is what a present-day knitter would call open (or uncrossed or untwisted) stocking stitch, albeit in a compound variant where each stitch is worked over the corresponding stitches in both of the two preceding courses (discussed and illustrated here). Loom knitters call this 1-over-2 (or double stitch) knitting. Its twisted “e-wrap” form commonly appears in stitch dictionaries of loom knitting, but descriptions of an open 1-over-2 structure produced with open “u-wrap” or “true” knitted stitches do not.

Compound knitting isn’t found in the established glossary of contemporary needle knitting but the structure appears in single-wale “false seams” worked with a crochet hook, such as Elizabeth Zimmerman’s phony seam (alternating 1-over-2 and 2-over-1). However, there is no particular difficulty in making compound stitches with ordinary knitting needles that justifies, much less necessitates, positing the greater utility of a knitting loom. There is no unequivocal physical or iconographic evidence of any of the candidate implements until significantly later than the production date of the Paris tube. There is also some basis for adding a circular needle to the speculative list of applicable tools.

An illustrated description of the needle-based technique was published in 1997 in a catalog prepared by Marianne Erikson titled Textiles in Egypt 200–1500 AD. It is part of the analysis of a fragment from ca. 1000 CE presented as flatwork but plausibly a segment of a piece worked in the round. That method of compound knitting was considered in further detail in a series of tutorial blog posts by Katrin Kania, in turn based on a workshop conducted by Petra Linscheid that examined the spool-based approach. A book recently published by Assia Brill titled Distitch, describes several needle-based techniques from the perspective of their contemporary application.

The photograph included in the Berlin report shows the compound structure from both the front and back of the fabric. The tube is not intact and several courses can be seen from the inside. The photograph in the report on the piece at the Louvre shows similar damage and reveals additional structural detail that might otherwise go unnoticed. Both tubes display the notable vertical compression of the individual stitches that is typical of compound knitting.

However, the second yellow band from the left in the lower segment in the following photo of the Paris tube has several stitches that are strained open and appear as though they might be looped into the immediately adjacent course only. The entire piece is close to two meters long and the full published photograph shows it in a serpentine position. The fineness of the stitching can be understood by visualizing the segments shown here as parts of a continuous tube of such length. If it is accepted that mistakes in this form of compound knitting can appear as ordinary knit stitches, there is no need for further explanation of the deviation in the lower segment.


The Berlin report references additional examples of Coptic tubular knitting and other sources include still more. Both simple and compound stocking stitch are represented and the detailed examination of a larger number of samples would permit a more certain determination of the balance between the two types. Some of the tubes are also closed at one end either with a drawstring or decreases in the final few rounds of stitches. The analysis of those structures might also provide more specific detail about the technique(s) applied to their production.

History · Looping · Structures · Tools

Crochet isn’t for the birds

Hooks and needles have been around immeasurably longer than any evidence of either being used in the production of looped fabric, and looping without tools all but certainly predates the use of any such implements for that craft. In fact, there’s no way even to determine if our species was the first to figure out how to do any of this.

Tailorbirds know how to draw loose fiber into thread and sew with it.

Weaverbirds know how to gather uniform strips of vegetable fiber and loop and weave them into shaped structures.

A slender pointed beak that can both pierce and grasp is remarkably well suited to the sewing and looping seen here. It is also useful for converting a needle into a hook when needed.

From the perspective of sewing-tool design, the dual-purpose awl and tweezers provided by a beak is a more precise implement than the functionally equivalent index finger and opposable thumb on a human hand. There is no reason to assume that these birds developed their needlecraft and toolmaking skills before we started doing similar things but Homo sapiens can’t even claim credit for the oldest known eyed needle.

It’s a fair guess that this needle was used to pierce fabric and pull some kind of fiber through the resulting hole but that doesn’t preclude similar action at the site of the needle’s manufacture or elsewhere with other tools. It would be more of a stretch to see the needle as evidence of tool-based looping but at some point eyed needles clearly did come into use for that purpose.

There is an obvious upshot to all this. Just as there are several basic looped structures that can readily be made without tools and are seen as universal constructs, appearing independently at uncountable times and places, a battery of ubiquitous tools has been available throughout to ease and extend the production and design of looped fabric. However much local applications might diverge and whatever degree of specialized complexity a craft might ultimately acquire, at least in principle, all can be traced back to one or more elements of the same initial set of structures and tools.

In the largest number of cases, the appearance of similar techniques at widely separated times and locations is therefore best treated as coincidental. However, there are situations where corroborating evidence indicates cultural cross-pollination, if not the outright transfer of technology. One possible such occurrence can be seen in previous posts about the parallels between Egyptian yarncraft and Viking wirecraft, with tubular knitting appearing in each. A similar parallel is found in the looped yarncraft of the two regions.

The eyed needle was used by both communities when working with yarn but the only evidence of a hook used with that material is Egyptian, assuming the practice of  knitting on a peg loom. The Vikings similarly appear likely to have used a hook for knitting wire but, again, there is no indication of their having used such tools to work yarn. That technique does appear in Northern Europe far later as “shepherd’s knitting” (the basic form of what is now called slip stitch crochet), where it is believed to have developed over an indeterminate period for producing the types of warm utilitarian garments that were also made in the same region by nalbinding.

However, there is at least one instance of shepherd’s knitting in archaeologically recovered material claimed to be from Coptic Egypt. If that provenance is accepted, it raises a few interesting questions. It is entirely possible that the object is simply an isolated occurrence of a fairly obvious method of looping that would become commonplace at a significantly later date, both in North Africa and Northern Europe. On the other hand, if the parallels between the looping techniques in those two regions are a result of cultural interaction — of which there is otherwise ample evidence — it may be possible that the use of a hook to make crochet-type structures was communicated via those channels. I’ll get into the details of this by describing the Coptic object in a subsequent post.

History · Knitting · Structures

Early Irish knitting

Knitted silver wire of the type found in the 9th- through 11th-century Viking hoards in England and Scotland (discussed in a preceding post) and at other Viking sites, has also been found in 9th-century CE Irish hoards of altar vessels. Two such objects are of particular interest. The Derrynaflan Paten is decorated around its circumference with three knitted tubes (seen in greater detail here).


The tubes appear to consist of shorter segments with slightly different appearances (and perhaps structures) joined together. One of the alternating elements was created by a gently spiraling band of silver wrapped tightly enough around the tube to press into it, forming a secondary decoration.

The paten is described in great detail by Michael Ryan in The Derrynaflan Hoard and Early Irish Art, published in the October 1997 issue of Speculum. He notes that the knitting around its outer circumference is soft enough to risk being deformed during normal handling (for which reason the hoard includes a separate support for the paten). He also states that “nothing later than the ninth century occurs in the find” but does not provide a more specific date for the paten.

It is currently in the collections of the National Museum of Ireland, which dates it to the 8th century, together with the more widely known Ardagh Chalice. On the basis of the shared intricacy of their decorative detail, the museum believes that both can have been made in the same workshop, which was presumably near the sites where they were recovered. Unlike the jewelry found at the Viking sites, where the knitting was the centerpiece, the Irish knitting considered here was a decorative component of larger elaborate objects that would in all likelihood have been commissioned by the churches where they were found.

The following detail from one of the two strips of knitting on the chalice shows the same compound structure as that seen on the Viking and Egyptian knitted tubes. It is formed by alternating rows of different-colored wires worked over each other. This is placed alongside plain open-loop knitting. The most noteworthy difference between the knitting on the chalice and all the other early knitting examined here thus far is that it was worked flat.


This upsets the established notion of knitting first having arrived in Europe at some time after 1000 CE, from its Egyptian origin, via Islamic Spain. There is convincing evidence of knitting in yarn having traversed that path, nonetheless, but knitting in wire was well established in northern Europe by the time it did.

There is no way to determine if the contemporaneous appearance of tubular knitting in Egypt and the northern European sites is coincidental, or provides yet another indication of cultural exchange between those communities. (Ryan includes a footnoted reference to the silver used for decorating Viking objects seeming “to have been derived in the main from Islamic dirhams with inscriptions in Kufic characters…brought along Viking trade routes to the East.”) Similarly, there is no way to tell if the Vikings introduced the craft to Ireland or picked it up there — assuming that’s the way it happened, at all.

The knitting found at all sites displays the same fundamental structure. The yarn used in the one region and the wire used in the other can be knitted with the same tools, although the working properties of the respective materials make some implements and methods better suited to the one than the other.

The Ardagh Chalice was examined in minute detail by the Conservation Laboratory at the British Museum during the late 1960s. A report about it was presented by Robert Organ at a seminar on the Application of science in examination of works of art, in 1970. This report was included in the published proceedings of the seminar three years later, titled Examination of the Ardagh Chalice — A Case History.

The preceding photograph shows a detail of one of the decorative elements on the chalice presented in that report. It is the less complex of the two strips of knitted wire. The other is significantly more intricate, with an additional structure woven into the knitting and the legs of some of the knitted stitches tracing an elaborate path through the work. This parallels the secondary device worked into half of the knitting on the Derrynaflan Paten. I’ll return to it when I feel surer about my ability to provide an accurate description. Puzzling that out will likely reveal more about the method of production, the basics of which I intend to discuss next.