One of the nice parts about using a blog to present the results of ongoing historical research, is the ease with which a report of “the earliest evidence [of whatever] that has yet come to light” can be amended when even older evidence is uncovered. Since such work constantly strives to extend the resulting timeline, every new success paradoxically risks invalidating a previous one. The corresponding revision of the broader narrative may entail nothing more than noting that something is a few years older than previously believed. However, things such as the radiocarbon dating of a questioned fragment of fabric can necessitate a fundamental re-contextualization of previous documentation. This in turn can effect a major change in our understanding of, say, the origin of a given mode of looped fabric production.
I tacitly tweak posts on this blog to reflect subsequent insight without calling attention to such revision. However, there have been a few stop-press situations where the retroactive editing has been paired with a new post about the details of the more recently uncovered material. The last such case (reported here) arose from my having overlooked the first attested mention of Tunisian crochet — in a Swedish publication that I had in fact examined. It appeared one year earlier than the German source I had cited.
This post is an updated replacement for an earlier one titled Scottish and shepherd’s knitting revisited that I took offline before preparing an article on the underlying topic for publication. New questions about shepherd’s knitting and its relationship to crochet have arisen in the interim and a book that was central to the initial post sheds quite a bit of light on them. It was published in Dublin in 1835, with the title page:
SIMPLE DIRECTIONS IN NEEDLE-WORK AND CUTTING OUT INTENDED FOR THE USE OF THE NATIONAL FEMALE SCHOOLS OF IRELAND. TO WHICH ARE ADDED, SPECIMENS OF WORK Executed by the Pupils of THE NATIONAL MODEL FEMALE SCHOOL
The first tutorial text about crochet written entirely in English was published in 1840 by Jane Gaugain, in The Lady’s Assistant for Executing Useful and Fancy Designs in Knitting, Netting and Crotchet Work. She uses the French loanword (alternating between the spelling in the title and the native one) to designate the craft but not the individual stitches that it comprises. Each is labeled a “tambour” and the action of their production is “tambouring,” without any reference to crochet in the instructions. She settled on the now standard spelling in subsequent texts but left the substantive presentation of the craft unchanged in the enlarged 1847 edition of The Lady’s Assistant, despite the different nomenclature her colleagues had begun to apply in similar presentations starting in 1842.
This strongly suggests that Gaugain took tambour embroidery to be the sole point of departure for the new craft. Other authors saw tambour embroidery as having contributed elements that were merged with the older Scottish shepherd’s knitting, which they incorporated into the new stitch repertoire as single crochet (later aka slip stitch crochet). Gaugain was also the only one who placed the elemental chain stitch in the ordered sequence that extended to double and treble crochet.
There are two well-established glossaries used to describe crochet in the English language. They are referred to as “US” and “UK,” with other anglophone countries using the one or the other. Both include the same terms and present the same stitches, but associate the labels with the structures differently. A “single crochet (US)” is a “double crochet (UK)” and a “double crochet (US)” is a “treble crochet (UK).” A “slip stitch” is now the same in both but was a “single crochet” on, and for a long time after its first appearance in the UK terminology. This was the earlier of the two to develop and is used in the following discussion unless otherwise noted.
Frances Lambert published an ordered set of definitions for crochet stitches in 1844, in My Crochet Sampler. A “plain single crochet” starts a counting sequence that continues with a “plain double crochet.” However, there is a confusingly similar “double stitch crochet” that designates a stitch made by pulling its initial loop under both legs of the loop to which it is anchored.
The first known German instructions for Tunisian crochet are for an ornate shawl, published in the 9 January 1858 issue of the German publication Der Bazar. They are accompanied by four illustrations, of which the third shows the front of the garment and the fourth is a thumbnail representation of its back.
Figs. 3 & 4.
The first and second illustrations are ostensibly drawn at full scale to indicate the gauges of the stitching and hook. However, the rows are not the same height in both. They appear together on the same page and the difference is not an artifact of the printing. The original objects from which the two drawings were prepared also appear to have been made by different people, one left-handed and the other right-handed, as indicated by the opposite slant of the vertical loops. Continue reading “From grey shawl to pink mantle in 10 months and 14 rows”→
The 15 December 1857 issue of the Swedish journal Penelope, includes instructions for a child’s upper-body garment made with a crochet stitch that had been described without a name in instructions for a window shade in the 1 January 1856 issue of that publication. In the 1857 instructions, the same anonymous author retrospectively labels it the Tunisian crochet stitch — the earliest attested use of the term that has yet come to light. The preceding post includes a translation of the instructions for the window shade. The ones for the child’s garment are translated below.
In the 1856 description, the author notes of the stitch:
“Although somewhat awkward to describe…I hope that I have expressed myself tolerably well.”
That goal was reasonably well met in the text it prefaced but the adequacy of the description of the more complex 1857 garment is not as immediately apparent. It omits key procedural details from the text and the accompanying illustration does not accurately reflect the prescribed stitches counts. The need for interpretation and interpolation makes it difficult for a translation both to be faithful to the original and provide a sufficient basis for making the object. The readers of the initial document would, of course, have been addressing the same issues. This raises the equally important matter of the familiarity with crochet techniques that the author can reasonably have expected them to bring to the task. Continue reading “Tunisian crochet in Sweden in the 1850s”→
Several posts during the first months of this blog provide translations of Dutch instructions from 1823 for a number of purses made with different looping techniques. They include three that mark the first use of the word crochet in the fancywork press to designate the craft now widely known by that name. The label appears in the same context in English language publication in the 1830s and its German equivalent — häkeln — began to appear at the end of the first decade of that century.
The first instructions published in France for a craft named with the word crochet (“hook”) are somewhat surprisingly for loom knitted purses (bourses au crochet). They were presented by Élisabeth Bayle-Mouillard (writing as Élisabeth Celnart) in 1826 and are discussed and translated here.
The first attested occurrence of the word crochet in its modern sense is found in French instructions for yet another purse. They are incongruously embedded in an anonymous compilation of knitting instructions from 1837. These are otherwise entirely in English (seen unaltered in the 2nd ed. from 1838; the 5th ed. from 1840 names the “compiler” as Miss Watts).
One of the recurring topics in the discussion of Tunisian crochet is whether fabric produced with a double-ended hook should be regarded as a variant form of ordinary Tunisian crochet or as an entity of its own. The earliest instructions calling for that tool that I have been able to locate so far are in the 10 February 1896 issue of the German craft periodical Der Bazar. This is eight years before similar instructions appeared in the US publication Columbia Book of the Use of Yarns, detailed in a previous post.
The stitch dictionary in the 8th edition of the Columbia Book, from 1907, includes a Double Hook Afghan Stitch.
“Wind the yarn into 2 balls, as 1 ball is used at each end of the work. Make a chain the desired length, take up each stitch as in afghan stitch, retaining the stitches on the needle. Turn the work (fasten the other ball of yarn to the other end of the work), take the stitches off the needle with the other hook.
Third Row—With the same hook take the stitches up again.
Fourth Row—Turn the work, taking them off with the other hook. Repeat 3d and 4th rows alternately for all the work.”
With the exception of the hook and fabric being turned at the end of each forward pass, with a resulting need for a second strand of yarn, this stitch is worked the same way the Tunisian simple stitch (TSS) is. Both strands can also be taken from the opposite ends of a single ball of yarn (as done in twined knitting), just as the two hooks are on opposite ends of the same shaft. Using separate yarn sources additionally enables colorwork but that is also true of single-hook TSS.
The appearance and texture of the ribbed double-hook fabric differ markedly from ordinary TSS. The prominent horizontal chains characteristic of the latter are concealed entirely, although again, the same applies to many single-hook variants. There are additional double-hook variants that relate similarly to other single-hook forms. If a categorical distinction is to be made between them, the double-hook qualifier first noted in the 1907 pattern provides a good basis for it.
The double hook also makes it possible to work TSS in the round. Here, the ends of a foundation chain are joined and a new loop is worked into each of the loops in that chain for as long as the straight hook comfortably permits. The work is then turned and the other end of the hook used to return a chain through all but the last few of the pending stitches, using a separate strand of yarn. The work is turned again and the forward round is resumed.
The enclosed chains are oriented in the opposite direction from those in flatwork TSS but the two forms are structurally identical. The details of the loops in a TSS chain that indicate the direction in which it was worked are often obscured by the vertical loops, with the exposed edges of the chain loops appearing as parallel horizontal lines in both cases. The difference can be seen on closer examination, but whatever significance it may have, it provides little justification for a categorical distinction between chained-toward-the-right and chained-toward-the-left TSS variants. Since Tunisian crochet stitches can only be worked in the round with a double-ended hook, the single/double hook attribute is superfluous in the labeling of such fabric.
There is an intermediate aspect to the instructions for producing the Tunisian stitch variant using two separate hooks that appeared in Der Bazar in 1862 (discussed here). Taken with the earliest mention of any form of Tunisian crochet being from 1858, it seems unlikely that references to the double hook before the final decade or two of the 19th century remain to be discovered in the fancywork literature.
This still leaves a question about when in the 20th century (assuming no surprises) the first instructions for using that tool for work in the round appeared. The Columbia Yarn Company added celluloid to the materials in their listing of hooks and needles in 1908, also offering double-ended crochet hooks under their own heading for the first time.
However, the 1904 instructions linked to above explicitly prescribe a “Wooden Double End Crochet Hook, 20 in., No. 13.” The industrial production of such tools had therefore commenced prior to their availability in celluloid. Nonetheless, as of the 19th ed. from 1918, where the Double Hook Afghan Stitch also appears, the Columbia Book still only uses the requisite hook for flatwork.
Although this suggests a later advent of TSS worked in the round, countless similar publications remain to be examined before any date can safely be placed on its first appearance in print. Present-day Tunisian crocheters are familiar with its comparatively recent proprietary manifestations but the initial use of the double hook is not part of the general craft lore.
The 10 February 1896 issue of Der Bazar includes instructions for a child’s cradle cover made in alternating bands of shell stitch (Muschenstreifen) in ordinary crochet and TSS ribbing (Rippenstreifen). Both are worked with the same double-ended hook, as is the elaborate border. The ribbing has a more complex form than that discussed above, and uses different color yarns to good effect. (Fabric worked entirely with it is also fully reversible.)
The illustration is presented in the immaculate detail that is a hallmark of Der Bazar (where the first explicit reference to the “Tunisian crochet stitch” appeared, as discussed here.) The written instructions, however, are atypically difficult to follow. Those for the ribbing are clear enough and describe a variant that remains in the current repertoire. The instructions for the shell stitching are more opaque and I haven’t quite puzzled out how they lead to the illustrated structure well enough to be able to provide a useful but objective translation. In any case, these shells are not intrinsically dependent on a long hook, much less a double-ended one, and lie outside the scope of the present post anyway. The following text therefore stops at the end of the instructions for the ribbing.
Part of a crocheted cover for a child’s cradle
“The pretty cover is made with white and blue woolen yarn singles [Dochtwolle] in a variant of the Tunisian crochet stitch, together with a shell pattern, using a heavy wooden needle that has a hook at both ends. It is worked on a white foundation row of appropriate length and an even number of stitches, as follows:
1st pattern row: forward. With a loop of the blue yarn around the hook, skip the first chain and draw one loop through each of the remaining chains. Turn the work and return using the other end of the hook with the white yarn to close the stitches one after the other.
2nd pattern row: forward. With the blue yarn (the active yarn is always led through the first loop) skip over a stitch and draw a loop through both the next vertical bar and horizontal chain together. Having worked through each stitch in this manner, return as in the first pattern row.”
The use of a long double-ended hook both for stitches that require one, and those that could as easily be worked on a conventional short crochet hook, is interesting in itself. There is a surprising amount of additional evidence of the mid-19th-century use of long hooks for ordinary crochet. Such hooks were manufactured with both cylindrical and tapered shafts, so there was more to it than a simple matter of some workers preferring Tunisian hooks where they were not a necessity.
The entire discussion here is restricted to evidence found in the context of fancywork. An earlier post considered the relationship between Tunisian crochet as an urban practice in Sweden and its rural counterpart krokning (hooking). It remains unclear whether this is just a matter of alternate nomenclature or if krokning has roots that predate the emergence of Tunisian crochet in the craft literature. For now, it is sufficient to note that the use of a double-ended hook for TSS in the round is a mainstay of the former craft as it is currently practiced.
In January 1864, a Swedish monthly publication for fashion and fancywork commenced publication with the title Iduna, a Norse goddess associated with femininity and knowledge.
Iduna,
Journal for the Tasks and Concerns of Women,
with Supplementary Patterns for Counted-Thread and Free Embroidery,
Crochet, and Knitting,
plus Fashion Plates.
It was the latest in a series of similarly oriented publications that changed titles a few times over the decades. There was often some editorial continuity between the final issues of a series and the first ones of its successor. The incarnation that immediately preceded Iduna was Penelope. It first appeared in 1854 with essentially the same subtitle: Album för qvinliga arbeten och moder, med bilagor af tapisseri- broderi- virk- och stickmönster, samt modeplancher (Album for the tasks and concerns of women, with supplementary patterns for tapestry embroidery, free embroidery, crochet, and knitting, plus fashion plates). Its final volume was published in 1863.
The period spanned by the production of these two publications included the interval during which the “Tunisian crochet stitch” made its meteoric spread through the European fancywork press. This began with a description of the stitch itself in 1856, unnamed, with the Tunisian label retrospectively applied to it in 1857. When writing the initial version of the present post, I was unaware of the earliest attested instances being in Penelope and have provided further details about this in two subsequent posts, describing the first published appearance (as far as I have yet been able to determine) of the unnamed stitch in 1856 here, and the instructions from 1857 that label it the Tunisian crochet stitch here.
The first of many instructions for Tunisian crochet that appear in Iduna are found on page 4 of its inaugural issue. They are for a “crocheted pelerine” (wirkad pelerine) in “the ordinary Tunisian crochet” (den wanliga tunesiska wirkningen). This is what we now call the Tunisian simple stitch and it is clear from the text that readers are expected to be familiar with the technique. A repeated reminder in subsequent instructions about a row of Tunisian crochet being counted as a forward and return pass together may indicate that its written description was unfamiliar to a Swedish readership. However, there are no remarks about the stitch itself, which also appears in the traditional Swedish craft repertoire as krokning (hooking). The first attested use of that term is presented below. It does not appear to be a coinage, leaving an open and important question about whether it was taken from an older orally transmitted glossary.
The supplementary illustration of the pelerine is missing from the copy of Iduna that I examined but the same issue includes complete instructions for a child’s dress.
Tunisian crochet and krokning are generally regarded as synonymous in present-day Swedish usage. However, a distinction is made between the two terms in a four-part Handbook of Women’s Handicraft (Handbok i fruntimmers-handarbeten) published by Hedvig Berg in 1873-74. This is where krokning is first attested, as a collective designation for all crochet stitch variants that are worked row-by-row on a long hook. The alternative name Tunisian crochet is again restricted to the simple stitch. However, “almost all crochet known under the name krokning is just a variation of it.”
The chapter on crochet has four sections, of which the last is headed “Crochet with wool yarn” (Virkning med ullgarn). The chapter begins:
“Since crochet is now so widely known and practiced that there is no need for a fundamental description, instructions are only given here for the lesser-know Irish guipure crochet, for which many instructions are provided, together with some more unusual mignardise lace [ganzspetsar], as well as a few crochet stitches or so-called krokningar [hookings] with wool yarn, the latter also with some patterns.”
The section on yarn crochet has six subsections, each for a specific stitch. The first five are labeled as krokning but the sixth is for “axvirking.” An ax is a head of grain and the corresponding German name — Aehren-Häkelstich — as well as a drawing on which the one shown below is clearly based, appear in an 1865 issue of Der Blatt.
The important point is that Berg regards it as a lesser known stitch worth describing in terms of Tunisian crochet, which the German text does not. She doesn’t classify it as krokning because it isn’t worked along the entire length of hook in continuous forward and return passes. However, it does apply the same basic technique repeatedly to a smaller stitch cluster that was apparently presented under the same heading for this reason.
“Take up three loops through three of the foundation chains just as with Tunisian crochet, then pull one loop through all three at once forming a small group of stitches, crochet one chain and repeat with three loops again all drawn through together…”
This Tunisian component and it being the only one not labeled as such, among what are otherwise all proper Tunisian crochet stitches, may indicate that the latter were generally preferred to ordinary crochet for work in yarn. This is reinforced by a text from 1884 by Mathilda Langlet, titled “The Homemaker in the City and in the Country” (Husmodern i staden och på landet). Her chapter on crochet begins:
“This form of handicraft has now become so widespread and developed that commenting on it seems superfluous. However, for the sake of completeness we do not feel that we should exclude it.
Vests, sweaters, and shawls ought these days to be ‘hooked’ (krokas) or knitted (stickas), rather than crocheted (virkas). Crochet with wool yarn is normally restricted to so-called slipper crochet (toffelvirkning), which can be compared approximately with knitting, all plain, or garter stitch (strumpebandstickning) as it is also called. One crochets back and forth but pulls the needle not through the closest loop but the one behind it. This gives the crochet a distinctive striped appearance.”
This is followed by instructions for a slipper and a note that “cuffs are also worked in the same manner.” It is not clear if this ribbing is worked with slip stitch or double crochet (UK) and Langlet also describes it being done with “fine thread…for cuffs and collars.” Given the long-standing association between footwear and slip stitch crochet it may be safe to assume she is prescribing that technique. In either case, ‘BLO slip stitch ribbing’ fashioned in the described manner remains a part of the crocheter’s repertoire.
There are also instructions for a “Crocheted child’s shoe…in Tunisian crochet” in the third issue of Iduna (March 1864).
The cuff is “pineapple crochet” (ananasvirkning), which had previously been described as a “velour” crochet stitch in Der Blatt. One of its repetitive elements is:
* wrap the yarn around the needle four times from back to front and insert the needle into the following stitch, wrap the yarn around the needle as usual and pull it through all the loops at once; make one chain *
In both this stitch and Berg’s axvirkning, the hook is inserted into the chain space between stitch clusters in subsequent rounds. Their primary shared element is the pulling of a single loop through a number of loops that are already on the hook. One of the ways for placing them there can be compared directly with the forward pass in Tunisian crochet. The return loop pulled through the entire sequence on the hook similarly resembles crochet tatting, possibly explaining why Berg includes that craft in her chapter on crochet, in the section immediately preceding the one with the Tunisian and Tunisian-like crochet stitches.
Further variants of this type of hybrid crochet stitch have been found in Swedish mittens and seen as an attempt to mimic the appearance of nalbinding. I’ll take a closer look at them in a separate post.
The earliest text describing and comparing various forms of hand knitting that has yet come to light is “The art of knitting in its full extent” (Die Kunst zu stricken in ihrem ganzen Umfange), published in 1800 by Johann Friedrich Netto and Friedrich Leonhard Lehmann. An earlier post covered their description of work with hook tipped knitting needles. The following material extends this with their discussion of knitting with a single hook-tipped tool, identical to the one from 1785 illustrated in the preceding post. The reference date for their introductory mention of it as “not an entirely new invention” was 1780 but they don’t indicate how long before then the described technique was in practice.
Chapter 9: Knitting…with needles and hooks
§ 39
“The hook knitting mentioned previously is not an entirely new invention but was once limited to winter shoes and boots. Primarily in Leipzig, the uppers of many hundreds of pairs of shoes were knit from ordinary white sheep wool yarn with a hook that was commonly made from the handle of a spoon to save expense. (A more exact description of this follows below in the explanation of the plates.) The knitted shoe uppers were then dyed black, fitted with soles and heels, lined with coarse felt, and shipped far and wide.
The simplicity and rapidity of this hook knitting caught our attention and we were delighted by the complete success of our attempts at using it for other forms of knitting. One can knit all sorts of colored borders, flowers, arabesques, etc., with the hook, just as with knitting needles.
The hook used for this is illustrated on Plate 15, Images 1, 2, and 3. First one wraps a single simple loop around it (see the small number 1 on the hook in image 1) and holds it in the left hand exactly as illustrated, so that the notch in the tip of the hook is on the left side. The thread is then led under the shaft as shown at number two (at the notch on image 1) and positioned in the notch above toward the right side. From here, the right hand draws the hook under the left, whereby the left hand slides the loop over both the newly formed stitch and the tip of the hook. The thread is placed repeatedly under the shaft and over the tip of the hook and the new loop is drawn through the old.
When the first row is finished and one wishes to knit back again from the right hand side, as with a blanket, the hook is inserted leftward into the last stitch (one remains on the hook) from below. Working in the same direction as just taught, the thread is again led under the hook and up over the notch, as shown on the plate. The left hand then slides off two stitches (see the hook in image 3), namely the one that was on the hook at number 2, and the one newly formed by inserting through 1 over 3. This continues by repeatedly inserting the hook to the left into the adjacent stitch.
Plate 15
The hooks for knitting on Plate 15 must be made in the following manner. The shaft may not be over one sixth of an inch thick [Leipzig standard ≈ 4 mm], at most, i.e. like a medium knife blade. The shape of the hook is conical to make it all the more convenient to slide off the stitches and can be half an inch wide at the bottom. The point of the hook must turn inwards, so that it does not snag when pulling through. No. 1 has a wooden handle but that can also be done without.
§ 40
If flowers and the like are to be added to the knitting, colored yarn is pulled into the first or second preceding stitch with a knitting needle that has a hook at its tip. Placing the flower on it, the colored thread is looped around the hook and worked into a stitch.
Hook knitting can also be worked from the left as with ordinary knitting. The only difference is the positioning of the thread. Instead of leading it under the shaft as usual, it is first passed over the shaft and then led under it. It is then drawn upward into the notch as seen on hook no. 2 [incorrectly ‘3’ in the original].
§ 41
Round pieces, for example oval rugs for covered floors, are also amenable to knitting with the hook. First a row of stitches is made about two feet long [565 mm] always knitting toward the right side. Then going diagonally down and back two stitches, oval rounds are knit to the left until the middle section reaches the intended size. A light-colored border can then be knit around it a half foot wide, and it can be decorated in the middle with arabesques or other leaf work in indigo colored yarn. When the border is done, another half foot is knit around it in the main color. Finally, the stitches are looped tightly together at the edge with the thread, one after the other.
§ 42
Stockings can also be knit with the hook. A row of stitches is made of a length that, when held together at its ends, has the width of the intended stocking. The hook is inserted into the first stitch that was made and a new one pulled through it so that the entire row forms a circle which is then knit continuously in the round. A stitch is increased by making a new loop as at the outset and pulling it again through the same stitch. A decrease, on the other hand, is worked into two adjacent stitches simultaneously.
§ 43
It need not be pointed out that the entire bodies of vests and leggings can be knit in this manner with the hook. For plush knitting, the thread on the second ball from which the pile is to be made (§ 10) is first placed in the notch and then pulled through in a loop. However, since the hook will be inserted deeply into the stitch, the thread from the pile ball needs to be pulled back a bit because the loop will otherwise be too long.
§ 44
Hook knitting is recommended above all for purses. The stitches for this are created by wrapping the thread as already described for hook no. 3. Picots are formed by wrapping the thread around the hook twice as on no. 2, before positioning it in the mouth of the hook, thereby forming a long stitch. Once the hook is pulled through, these long stitches need to be placed around a small stick or heavy needle, as with netting, until the hook comes around again and completes the proper stitch.
This hook knitting resembles the manner in which Dübois used his hooked knitting needles to knit with such amazing rapidity. We are convinced that many of our lovely readers will apply their special womanly artistry to this knitting and soon bring it to the highest perfection.”
The basic descriptions of the flat hook and its vertical orientation are essentially the same as those in de la Platière’s text from 1785. However, Netto and Lehmann provide more detail about changing the direction of the stitching at the end of a row worked flat, as with the blanket described at the end of §39.
The 1785 instructions simply note that a new stitch can be worked into the preceding one either from the right or from the left. One obvious way for changing the working direction would be to turn the notched-side of the hook accordingly. The 1800 instructions are more specific about this and leave the hook in the same position throughout. Instead, the direction of the work determines whether the yarn is wrapped around the hook yarn-under as is usual for German knitting and shown in Figs. 1 and 3, or the later crochet standard yarn-over shown in Fig. 2.
A cylindrical hook-tipped knitting needle is presumably better suited for holding a flower as described in §40 than a tapered flat hook would be. Given that Netto had both tools at hand, a question remains about why he neither mentions, nor perhaps even considered, the use of the cylindrical tool for hook knitting. That craft was on the verge of a major step toward “highest perfection” as the crochet that would be described only two decades later in Holland. It is reasonable to suspect that the pivotal coopting of the similarly cylindrical tambour needle — a tool with which Netto was otherwise fully familiar — had taken place by the time of his writing about hook knitting.
The German term for the plush knitting discussed in §43 is Felbel (possibly a cognate of velvet). The parenthetical reference to §10 leads to a description of making warm fabric by pulling a short loop of thinner yarn through every sixth stitch in the base fabric, or if greater warmth is desired, through every fourth stitch.
