stitch by stitch

I'm a self-taught knitter and knitwear designer.

When I was 24, I had my first opportunity to serve on a jury. We were empaneled for 2 weeks. During that time, I came to deeply appreciate the American justice system. The woman sitting next to me, in addition to performing her civic duty, also knit a vest.

I decided then and there to learn to knit. I went to my local yarn store, Nimble Fingers, and emerged with yarn, needles, and Maggie Righetti's book _Knitting in Plain English_.

That book had a profound effect on me. I started knitting, worked the easy projects in the book, and branched out to other easy patterns.

I wasn't very satisfied with the fit of most of the easy projects. Knitting in Plain English gave me some idea of why that was, and suggested that I could learn more from Sweater Design in Plain English. Within a few years, I started designing my own garments.

I read what I could find about knitwear design. Maggie Righetti was a proponent of top-down design, so I branched out into Barbara Walker's Knitting From the Top. I found Sweater 101 and a Swedish Leisure Arts brochure on raglans that handled the basic math for the very useful raglan sweaters. June Hemmings Hiatt's The Principles of Knitting was my bible.

Ann Budd and Priscilla Gibson-Roberts followed later, as did the excellent 

Big Girl Knits.

I discovered Elizabeth Zimmerman after I was already an experienced designer. I appreciated her free and zany approach to yarn and needles. I loved her stories about knitting under challenging conditions. Her sweater design principles were refreshingly simple, but I found I wanted a more refined knit.

I have a generous hourglass figure. Straight up-and-down designs hang like a circus tent off my bust tips. Tighter sweaters pull against the hips. I do better with explicit waist-shaping, high hip lengths, and hemlines that drape over the hips instead of cutting a straight line across them.

It takes a lot of time, effort, and energy to knit a garment, If I'm going to put that much energy into something, I want to feel good about it. I like wearing garments that fit well, and I like seeing good-fitting garments on other people.

Designing your own garments takes a lot of math. You might was well do the math based on your real body size, and use styles and techniques that yield a comfortable, flattering fit.

I had some amazing guides about the work of designing good-fitting knitwear, but I still had to do a lot of legwork and synthesis to figure out what knitting math to apply to different yarns, different designs, and different fitting issues. 

I'm a programmer by trade. I started working on a digital assistant to do the preliminary knitting math in the mid-1990s. The first digital assistant, written in perl, took a keyword-value input file and generated a plain-text pattern. I then edited the pattern in a word processor to add the details the program didn't include.

I added more yoke types and specifications over time. A few years later, I rewrote the whole program in Python and started looking for a way to give it a user-friendly front end. My program worked great for me, but it was way too barebones for most knitters.

Top-down design fascinates me. With two sticks, a knitter can turn one-dimensional yarn into a real 3D object.  Using the magic of increases, decreases, and short rows, a knitter can sculpt the shape of an object right on the needles. Most knitwear designs create flat pattern pieces to be sewn together, just like fashion designers who work with woven cloth.

Knit fabric has more personality than woven fabric. Knitted fabric is stretchier and more resilient than woven fabric. Woven fabric resists stretching lengthwise and widthwise, being stretchiest on the bias. Depending on the pattern stitch, knit fabric can be expansive lengthwise or widthwise, but rarely on the bias.

I have never liked the way seams look or feel in hand-knit fabric. Seams form bulky, non-resilient ridges that don't flow with the fabric. Some knitters like the extra structure provided by seams, but I prefer working with the knit fabric directly to get the structure I want in a garment. On the other hand, a zipper is just about my favorite way to close a garment. Talk about seams! Talk about added stability! Talk about non-resilient!

Knitting is a very personal thing. We put a lot of ourselves into the garments we make. One of my goals in sharing the KnitFitter is to help other people find their knitting voices, to share a basic system of knitting that can be personalized and customized to fit the wearer, the wearer's lifestyle, and the knitter's way of working.

If you don't like something in one of my designs, feel free to change it. Choose different increases and decreases. Change the borders or replace them with turned-in hems. Choose the length the feels best to you, add or subtract pattern elements, experiment with different ways of working.

Ear Cozy Cap TestFlight

My first knitting app for iOS has been approved for beta testing through Apple's TestFlight program.




Ear Cozy Cap takes a size and a yarn weight and generates a custom cap pattern. I've been working on backend code to generate knitting patterns from raw numbers for the past couple of decades. The current app is a front end for a very small part of that backend engine. From the point of view of the knitter, this app replaces a single knitting pattern.

The app is much more flexible than any knitting pattern on the market. Traditional knitting patterns are tied to a specific yarn and support a small range of sizes. With the app, the knitter has a wider range of sizes and completely free yarn choice.

I am ready to invite potential beta testers to try Ear Cozy Cap on iPad or iPhone. If you or anyone you know both knits and uses an iOS [0] device, please email me at KnitFitter@gmail.com and offer to be included in the beta.

I also welcome folks who don't knit but are willing to put the program through its paces on whatever devices they have.

stop flipping out

Sometimes, borders just flip me out.

See this? The edge of an otherwise beautifully knit sweater stubbornly insists on flipping out.

I've had this problem more times than I can count. It usually happens with bouncy yarn knit at a fairly firm tension. Single ply yarns seem to be the worst offenders. The very worst offender I had, though, was this baby sweater out of superwash in a slip-stitch pattern with a garter stitch border.

What is really happening here is that the stockinette is exhibiting its tendency to curl and the ribbing is just along for the ride. So changing to garter or seed stitch won't fix the problem, because it's the stockinette causing the problem and not the edge stitch. Slip-stitch patterns have an even more pronounced tendency to curl, and that's what made the baby sweater such a problem.

Ideas that sometimes help:

1. Reduce the size of the needles for the border. Going down two sizes usually helps a lot, unless the garment was knit at a very tight tension. In my experience, the hems of tightly-knit garments just tend to roll.
2. Reduce the number of stitches 5-20% on the first row of the border.
3. Knit a row of purl before beginning the border. This will encourage the border to roll towards the body instead of away from it. This is a weak remedy, but it's fixed some borderline cases for me.
4. Increase the depth of the border. Sometimes, the garment gauge distorts the border gauge, and the border needs an extra inch or so to settle into shape.

If the border draws in narrower than the main body of the sweater, it will cling to the body and help keep the border in place. A nice deep border worked on needles several sizes smaller and up to 20% fewer stitches can help dramatically, particularly if you toss in a purl row before the border.

The tendency of stockinette to flip up is so pronounced that you can use a single row of purl stitches anywhere in any knit piece to make a crisp fold line. I do this anywhere I want the fabric to fold back on itself (hems, turn-back cuffs, etc). You can also use a single row of purl right before starting a border to encourage the border to turn towards the body instead of flipping away from it. (This doesn't work with garter stitch borders, though.)

Ribbing used as a border at the cuffs or body of the sweater really needs to be about 3" deep. It takes about 6 rows (1" or so in most handknit gauges) for ribbing to establish itself and really draw in. So during the first inch of transition from stockinette, the ribbing isn't really free to assert itself, and in the last inch in the transition to the air, the ribbing is a bit insecure. So give it enough depth so it can really be ribbing.

My personal experience is that ribbing is a great finish for men's and children's sweaters, but it often doesn't work so well on women. We have hips, and the flare of the hip encourages borders to misbehave. For women's sweaters, I use different kinds of edge treatments to tame the stockinette curl.

One of my favourites is a turned knit hem. Make the sweater as long as you want it, purl one row, and then knit at least another inch and a half. Gently tack the live stitches down to the wrong side of the sweater, making sure to allow enough give in your stitching for the sweater to stretch any way it wants to. The resulting edge is crisp, tailored, and unobtrusive. This will always counteract the flip if the hem is deep enough and you sew it down loosely.

Borders worked sideways often, but not always, counteract the flip. One that I think often looks classy is any cable knit sideways along the edge of a cardigan. A bonus with a sideways border is that you can match the row gauge of the border with the stitch gauge of the edge as you go. There's no need to figure out how many stitches to decrease to get a gauge that will suit the edge.

A lace edging often works beautifully, particularly one with an undulating edge. Typically, I start working these just below the waist because the more beautiful ones are 4-6" long when worked at a worsted gauge. I've worked these top-down, but often choose to work the edging bottom-up and then graft it in place. Sideways lace edgings work as well.

Lining or facing the sweater can help with slightly misbehaving edges. You see this a lot in vintage knits, edges that are tamed by sewing in a lining or facing. A wide grosgrain ribbon can help stabilize a button band or horizontal edge. I've used this to good effect in children's sweaters, but I don't think a ribbon would have enough oomph to stabilize the body border in an adult sweater. You'd need a deep facing to make an adult border behave, and in the process, you would compromise the elasticity of the edge. This is a good solution for some garments, but you rarely see it in modern knits.

Knitting Deceivers: Blocking

When I read my very first knitting book 28 years ago, the author talked about how blocking is not a knitting cure-all. It will not make a too big or too small sweater fit. It will not allow you to shift a huge amount of length into width or vice versa. It will not fix edges that misbehave.

There are exactly three (3) things that blocking does well:

1. Stretch out a gossamer piece of lace so that it struts its lacy stuff.
2. Allow the yarn to settle into its new shape gracefully.
3. Smooth out small irregularities.

Blocking works very well for lace knit on fine yarn. Lace stitches like to be open, and blocking can help them do this. In a fine yarn, blocked lace can hold its shape well for a long time.

In most other situations, however, blocking can't do anything that dramatic. It can help shift yarn to neighboring stitches so that a piece looks more beautiful and finished, but it can't drastically alter the shape of a garment.

In laceweight yarns knit on relatively large needles, the knitted item doesn't have enough shape to combat the blocking. You block the piece, and there is not much yarn to pull back against your blocking. If you work with worsted weight yarn, though, the resilience of the yarn will pull back against the blocking and the item will re-assume its natural shape.

With a heavier yarn, therefore, you might be able to block a piece into shape, but it won't hold that shape very long. When you put it on a moving human body, the yarn will shift back around until it gets comfortable. Gravity will take its toll. Quite soon, the knit item will assume the shape it wants to be instead of the shape you blocked it into.

In some cases, such as flipping borders, a piece won't even hold its blocking long enough to dry. As it dries, the character of the knit will reassert itself.

Often, people suggest blocking a piece of knitting to change its size or correct a major problem. I don't know of any situation where this has actually worked. A person can confidently block the devil out of a piece. They wear it once, it assumes its real shape, and it gets stuffed in the back of a drawer forever.

If you have a piece of knitting that is:

• Too big,
• Too small,
• Too wide and too short,
• Too narrow and too long,
• Has a border that flips out or in,
• Has edge curl,
• Has a collar or trim section that won't lie flat,
• Biases,

Then blocking won't fix the real problem. Blocking can't change the character of knitting.

Knitted items have their shape because of the properties of the yarn and the knit stitches used. Knitting is inherently three-dimensional. We use this to advantage when shaping hats and sock heels. It comes into play when we use pattern stitches as well, even when we don't desire the three-dimensional effects.

If you knit a plain piece of stockinette fabric, it naturally curls toward the knit side on the top and bottom edges and towards the purl side on the vertical edges. The natural shape of a flat piece of stockinette fabric is not flat — it's more like a potato chip with rolled edges.

Experienced knitters know that stockinette is not a good stitch pattern for flat items, nor is any stitch pattern based on stockinette. For flat items, you need a balanced stitch pattern. The knit and purl stitches need to be arranged in such a way that they counteract one another's tendencies to curl.

Stockinette is great for items that encircle a body. The vertical tendency to curl inwards helps the item hug the body. Anywhere you have an open edge in stockinette, though, you need to do something to counter the fact that stockinette edges curl. If you do this properly, the item will hold the shape you want it to. If the edge is not strong enough to counter the curl, however, you will have borders that flip or otherwise misbehave.

Other stitch patterns have other tendencies. Ribbing is elastic widthwise (and thus draws in widthwise) and not lengthwise. Garter stitch compresses lengthwise and spreads out widthwise. At one time, these two stitch patterns were often used as complements -- garter stitch on vertical borders and ribbing on horizontal ones.

Moss stitch, seed stitch, and all members of that stitch family have a strong tendency to bias, particularly when they are knit from singles. These stitch patterns are also wildly expansive. They can grow lengthwise and widthwise with wear.

Understanding the way different yarns and stitch patterns work can help you plan garments that work. When you want to use a new yarn or stitch pattern, make a nice big swatch. When you're finished, let it sit for a day. Wash it, smooth it out, and let it dry flat. Play with it. See what directions it stretches in. Notice whether it has biased at all. Thread a knitting needle through the top and bottom edges, hang it up, and hang a little ball of yarn from the bottom edge. Does it stretch much after an hour or two?

If you end up with a piece of knitwear that has a problem, you can try blocking and see if it fixes the problem. In many cases, though, you might not be happy with the results. In that case, you can give the item away or you can rip out the parts that don't work and make something you will be happy with.

The ability of knits to be ripped out and re-knit is one of their big strengths. 

Habit Formation

Secret confession time:

I have a thing for guys who knit. +1 if they knit in public. +1 if they hang out with other knitters. +1 if they knit lace. +2 if they have beards. +3 if they knit lace in public. +5 for knitting lace in public while bearded. +7 if they have ever knit their beards into their lace. 

It should come as small surprise that I have a thing for Franklin Habit. He doesn't have a beard, but in every other way he is my ideal male knitter. (Franklin, if you ever want to seriously spin my fleece, grow a beard and knit it into your lace in public, with photographic evidence, and blog about it afterwards.)

I knit not one, but TWO of Franklin's retro Ladies' Travelling Caps. I knit the first one for my teenaged daughter, who wears her waist-length hair in a bun. She wore it almost daily with delight. Wherever she went, people asked if she was Amish. I made her a second one, with a matching scarf, in plum pudding colours.

I will probably have to make her another one every couple of years for the rest of my life.

As a knitter, I don't like to follow other people's patterns. I prefer to design my own knitwear. I like the number crunching to find the perfect fit and the joy of working something out on the needles until it is just right. I can count on one hand the number of other people's patterns I have knit in the past 10 years. 

I was emerging from the deep fog of bronchitis when Franklin came out with the pineapple. I had been in that state of sickness where you have been sick enough for long enough that you don't really believe you will ever be able to do normal things again, like knit ginormous pineapples.

Oh wait. That doesn't sound quite right. I had been deeply sick enough that I believed the rest of my life would consist of coughing fits interspersed with naps, with no time or energy for normal activities, let alone deranged ones like knitting ginormous pineapples.

The moment I saw that ginormous pineapple, I wanted to knit one. Evening bag sized. In a luxury yarn. With beads on. For my mom to take to the opera. I wanted to live in a world where knitted evening bags in the shape of pineapples make sense.

This desire seized me. My family, who had become so used to my wan convalescent state that they had forgotten what I am like when I get into the grips of an enthusiasm, was a little alarmed.

“You,” they said slowly, “want to knit a pineapple evening bag? Whatever would make you want to do such a thing?”

No worries. I'd already ordered the yarn. I'd started by looking for something that whispered elegance. On the way, I'd become mesmerized by a wild strain of Koigu sock yarn, one that insisted it was deeply luxurious and would make a pineapple that no one would ever forget. I believed it, and tossed in some Bretton sock yarn in sedate hues that might make a pineapple a person could, in fact, wear to the opera.

I now had enough yarn to make TWO pineapples.

I had had the image of twinkling little crystal beads, but a trip to the bead store yielded a stash of little tiger's eye chips that harmonized with the Koigu and looked like they would make wicked spines for a pineapple.

The Koigu pineapple was becoming distinctly more operatic, and yet less suited for opera wear, with every decision I made. This happens when a knitter falls into the creative process.

I took a look at Franklin's patterns, fixed the parts that I was sure were broken, and considered the size of the pineapple, the number of stitches, and my probable gauge.

I am a loose knitter, especially on small needles. I usually work socks on size 0 needles, and often work sock ribbing on size 000 needles. With some patterns, I have been unable to get some sock yarns to a fine enough gauge even with 000 needles.

I judged that the 320 stitches Franklin suggested I cast on would make a pineapple a trifle on the large size. I decided to whittle the cast-on down to 240 stitches (still a multiple of 16, as suggested briefly in this blog post about the pineapple's bottom). Although I was beading, I was damned well not going to bead each one of the 240 cast-on stitches. I would begin the beading later.

As I had not yet bought the beads, this meant that I could begin knitting right away. I congratulated myself on my knitting acumen, and picked up my needles.

I started knitting the leaves. After few inches, I noticed two things. The first was that the leaf lace pattern was unbalanced, leading to radically different-sized holes. This could be fixed by changing the lace pattern into a balanced lace pattern, which I would do on the next knit.

The second was that this pineapple was turning out to be a bit on the ginormous side. If I continued to knit it, I would have a beachball-sized pineapple rather than the elegant little evening bag that still inexplicably haunted my knitterly dreams. Moreover, my inner knitterly calculator deduced that I would need far more yarn than I had to finish a pineapple of these proportions.

I ripped out the leaves and cast on a much more sensible 160 stitches. The leaves went swimmingly and I soon got to the point where the knitter performs a little magic step of turning the pineapple outside in and changing to the main yarn.

Here I stopped for a while to thread the beads onto the main yarn. This proved to be challenging. The wicked little tiger's eye spines had been drilled with holes too small to easily fit over even such a slender yarn as Koigu sock yarn. A good third of the beads had holes that were too small, even after I threaded the beads first to a doubled piece of nylon thread that had the koigu threaded through it.

From Franklin's instructions, I had a clear idea how many beads I needed. Since I was losing a third of the beads, I had nowhere near enough.

I bought more beads. A week and much creative language later, I finally had all the kerfaluting beads I would need strung on the koigu.

Unfortunately, the beads were still too tight to move easily on the yarn. I decided therefore to work the plain rows from the unbeaded end of the ball and the beaded rows from the beaded end of the yarn. This decision saved me a great deal of creative language during the knitting of the pineapple.

With the first beading row, I had two awful realizations. The first was that beading ideally should have begun on the setup row after the colour change. The first pineapple spines are formed on that row, and beading that row would result in a more consistent pineapple. If a person wished to do this, they would bead the 8th stitch and every 16th stitch thereafter.

The second was that the beads are to be placed, not every 8th stitch, but every 16th stitch. This mean that I had strung over twice as many beads as I could possibly need. Given the agony of stringing those beads in the first place, this led to a little more creative language directed Franklin-wards.  I also might have cried just a little.

Once these little ripples were out of the way, I set in to knitting the pineapple with happy fascination. It is a beguilingly simple and beautiful pattern. I loved it. As it grew, it made me smile, and giggle, and occasionally laugh out loud.

Despite the severely constrained dimensions of my pineapple, I could see that yarn was going to be tight. I decided to take a leaf from the original pattern and transition back to the green yarn as soon as I started to run out of the main colour. I tossed in one round of green for every 4 rounds of main colour, then every 3, then every 2, then every other. Then I threw in a round of main colour after 2 rounds of green, then 3, then 4. At that point, I was close to out of the main colour, so I started in on the bottom.

I beaded right up until I started the bottom. This is contrary to Franklin's instructions, but I was in love with the wicked little spines and no longer cared whether Franklin approved of the liberties I was taking with his pineapple.

I had read Franklin's blog post on the design of the pineapple bottom with a level of fascination and appreciation that only someone who has designed a lot of knitwear could understand. This led to a deep enjoyment in knitting Franklin's elegant little bottom.

I finished the pineapple, braided a cord (monk's cord would be another wonderful and period-appropriate way to make cord for this bag, just saying), wove the cord through the first row of leaf stitches, and dangled the pineapple for inspection. I grinned, then went to my sock drawer and grabbed a fat, cushy pair of worsted socks. These I stuffed into the pineapple.

This was not an elegant pineapple. It was a wild and rather dangerous-looking fruit, and it was mine, all mine. I squeezed it in my hand. The spines, complete with their tiger's eye points, gave a deep tactile satisfaction.

I threw the pineapple to my daughter.

She held it aloft and remarked, “Like all healthy fruits, pineapples explode shortly after being placed.”

She was right. The pineapple did rather resemble a pineapple grenade from Ninja Kiwi's Bloons Tower Defense games. One of Franklin's pineapples, stuffed and outfitted with a kitchen timer, would make an excellent prop for a children's game of hot potato.

I took the pineapple to Quaker Meeting the next Sunday to show my Friends. As I was explaining about the 1840s vintage of the original pattern, I remarked that I did not know why knitted pineapples were popular in that era.

A bearded Friend immediately said, “I know why.”

In the early nineteenth century Britain and America, it was a common investment strategy to buy shares in a ship that would ply the world in search of profitable items to sell. When the ships ported, runners carried a pineapple to each of the backers to let them know that their ships had come in. The ship's captain stuck a pineapple on a post near his front door as a sign that he had returned home safely and profitably.

Pineapples were thus seen as a sign of good fortune. They were worked into crocheted bedspreads, tablecloths, and edgings. They were carved into bedposts and newel posts. They were also, apparently, knitted into bag form.

In ending my bout with bronchitis by knitting a wild pineapple, I was thus engaging in an old form of sympathetic magic. I was affirming not only a future where knitted evening bags in the shape of pineapples made sense, but also my own future good fortune.

With all of this under my belt, it was time to knit the second pineapple.

One thing that I had meditated on whilst knitting and enjoying the first pineapple was the fact that downsizing the pineapple so much had made the spines obscenely large. The resulting pineapple, while recognizable as a fruit that might potentially explode, also resembled a sort of many-nippled fertility fetish. While this was a fine, perhaps even a desirable, trait in a wild pineapple, it was not suitable for the model of elegant pineapplyness that had been my original intention.

I loved the wild pineapple deeply. I loved the discovery that went into its making and I loved the results that bristled from its uncompromising pineapply character. It was not, however, a pineapple that I could offer my mother on a suitable holiday with the comment that it was intended to be worn to the opera.

I went to a different bead store. Once again I had the image of glittering crystal beads to festoon the ends of the spines. Once again, I emerged with something completely different. This time the beads were coral beads. I also emerged with several different potential solution to the problem of stringing the tiny-holed beads on the slightly fatter yarn.

The coral beads wouldn't thread on the yarn at all. No how, no way. After breaking several beads, three beading needles, and fraying a few feet of yarn, I accepted this fact. These beads just weren't going to thread on anything much fatter than the nylon beading thread I had.

So I resolved to thread the beads on the nylon beading thread. I would work that thread double with my yarn on the beaded rows only.

This turned out to be a wonderful solution. I wished I had used it on the wild pineapple.

The next issue was the size of the spines. On some reflection, I decided to change the basic pattern stitch to (K4, yo, k1, yo, k4, sl1-k2tog-psso). This made the multiple of the pattern 12 stitches, including the distance between beads. which also required some minor finagling of Franklin's bottom. If a person was beading according to my design, they would bead the 6th stitch and every 12th stitch on the setup round.

I cast on 120 stitches. In the version of the pattern that I have, there was an error in the pattern that has since been corrected on the Knitty site. My pattern has the leaf rounds at (k6, sl1-k2tog-psso, yo, k1, yo) which is a multiple of 10 rather than 16. This caused me additional dancing in trying to get the downsizing to work out right, in addition to being the likely cause of the yarnover imbalance I mentioned earlier. If you attempt a downsized pineapple, you will not have either of these problems.

To finagle the bottom, I saw what Franklin had done when I knit the wild pineapple and decided to make things easy on myself by just faking it with the second. I did something like this:

Heather's Bottom

Round 1: *K4, sl1-k2tog-psso, k5 * (100 stitches)
Round 2: *K3, sl1-k2tog-psso, k4* (80 stitches)
Round 3: *K2, sl1-k2tog-psso, k3* (60 stitches)
Round 4: *k1, sl1-k2tog-psso, k2* (40 stitches)
Round 5: *sl1-k2tog-psso, k1* (20 stitches)
Round 6: *k2tog* (10 stitches)

I didn't count or anything, though. I just kept knitting the pineapple, lining up the decrease so that it centered below the last bead and omitting the yarnovers. I judge my bottom to be slightly less elegant and a wee bit more pineappley than Franklin's, but I understand that this is getting a little too personal so we will move on.

I found some delicious gold silk fabric and hand-stitched linings for the two pineapples.

With both pineapples, I used every bit of the single ball of sock yarn that I bought to use as the main colour. I would suggest the other people might want to get two balls of the gold yarn and one of the green to give them a bit more freedom in how long to make their pineapples.

At the end of all this, a salute to Franklin as a knitting designer. His patterns are well-worked-out, not overly complex, and quite charming. It is a tribute to him that I have worked not just one, but two of his patterns, and a further tribute that I have seen fit to finish both items and knit a second of each.

And do you know what would be just great? A pineapple hat pattern where the body of the hat is knit from the main pineapple stitch pattern and the hat has a topknot of suitably wild and properly shaped bromeliad leaves.

knit pleats that work

I tried to put pleats into Matisse's 1940s Peplum Sweater, but the results ended up being more ruffled than pleated:

I got the basic shape of the pleat right, but the inner fold didn't work as I had anticipated. It's a handsome sweater, with princess shaping and the flouncy peplum, but the end result is softer and not as tailored as I had hoped.

I knew I would have to pleat again. I read up on pleats in both woven fabric and in knits, and I felt confident that I could design knitted pleats that worked like box pleats in woven fabric.

I was working on a sweater in the oh-so-delicious undyed organic Inca Cotton to replace the two Inca Cotton sweaters that flitted out of my life.

The first Inca Cotton sweater, in Quaker rib with a zipper, stopped fitting when I lost weight, so I gave it away. The second one, in Hunter rib with a better zipper, was stolen one morning when I was dancing. I've been expecting to see that sweater about town, and looked forward to confronting the wearer with the fact that I created that sweater and it was cruelly stolen from me.

Oh well.

I considered letting the whole idea of Inca Cotton go, but it's the softest, cushiest, comfiest cotton I have ever worn. It's just the thing to brighten a gray day, warm a cold, or soften the edges of a harsh comment from a relative.

This new one is in Cartridge Belt rib with buttons, and I wanted something a little more formal than the previous zippered sweatshirts. Something with a little panache. Something that would whisper “I'm a luscious crème caramel and I'm worth every bite.”

As I was finishing the bodice and getting ready to make the first buttonhole, the slipped-stitch ribs in the Cartridge Belt pattern suggested pleats to me. This jacket, they seemed to say, would look ever-so-much-more appropriate out on the town or at a business event if it had a peplum with pleats.

And, the ribs went on, the ribs supposed they could be the focii of princess shaping to set up a curvy, 1940s-style silhouette that would go great with a pleated peplum.

A vertical line of slipped stitches bends inward along the slipped yarn. Cartridge belt rib uses this effect to create a softly fluted fabric with slipped ribs. I would extend this effect to make three box pleats. The outer edges of pleats would be formed by the ribs on the right side of the fabric and the inner fold line of the pleats would be formed by ribs on the right side of the fabric. The ground of the pleats would be garter stitch, in keeping with the basic pattern of Cartridge Belt rib.

The line in the photo above shows a pleat that has been opened. The inner fold line is faintly visible just to the left of the black line in the photo. The two outer fold lines are clearly visible as stockinette ribs.

The pleat is fan-shaped through the use of increases.

The pleat starts on 1 knit stitch (a slipped rib stitch in this pattern). All slipped stitches are slipped with the yarn in front. I used a knit-front-and-back for the increase, but a make 1 or knit-below would work just as well.

Here's the basic pleat formula I used:

Row 1 (right side): k1, yo, k1 in the one knit stitch

Row 2: sl 1, k1, sl 1

Row 3: k1, (k1, yo, k1) in next stitch, k1

Row 4: sl 1, k3, sl 1

Row 5: k1, (k1, yo, k1) in each of next 3 stitches, k1

Row 6 and all even rows from here on: sl 1, knit to last stitch, sl 1

Row 7: k2, place marker, sl 1, k5, sl 1, place marker, k2

Row 9, 17, 25, 33, 41, 49: k to marker, sl 1, knit in front and back of next stitch (kfb), knit to two stitches in front of next marker, kfb, sl 1, knit to last stitch, sl 1

Row 11, 15, 19, 23, 27, 31, 35, 39, 43, 47, 51: k to marker, sl 1, k to 1 stitch before marker, sl 1, k to end

Row 13, 21, 29, 37, 45: k1, kfb, k to marker, sl 1, kfb, k to 2 stitches in front of next marker, kfb, sl 1, k to 2 stitches before end, kfb, k1

Work 50 rows.

These pleats will also work for stockinette by purling all the wrong side worked stitches except for the stitches on the inside of the markers, which should be knitted.

Top-Down Design Tutorial 11: Figuring the Standard Yoke

Standard and saddle-shoulder yokes are relatively straightforward to calculate.

First, you need the back neck measurement, the shoulder measurement, and the sleeve cap measurement for the size you're making.

You'll start by making front right and left shoulders. Multiply your shoulder measurement by your stitch gauge. Cast on that number of stitches using a removable cast on for each shoulder. Knit until your front shoulders measure about one-third of your sleeve cap measurement, usually an inch or an inch and a half.

Leave your front shoulder stitches on a holder. Pick up and knit the right shoulder stitches from the cast-on edge. Multiply your back neck measurement by your stitch gauge and use the long-tail cast-on to cast on that number of stitches for the back neck. Pick up and knit the left shoulder stitches from the cast-on edge. These stitches together form the back of the sweater.

Work the wrong side row.

Work 1-3" of short rows across the shoulders and back to shape the shoulder slope. This puts more rows in the middle of the back than along the edges of the shoulders.

After you've finished shaping the shoulders with short rows, work even until the shoulder edge is the size of your sleeve cap measurement.

Your sweater will look like this:

The next step is to pick up the sleeves along the shoulder edge. Multiply your sleeve cap measurement by your stitch gauge. Pick up that number of stitches along each shoulder. Place markers to separate the sleeve sections from the body sections.

The corners at the edges of the sleeves are tight for the first several rows. You might need to pull your cable through at those spots in order to make that turn.

Next you'll need to figure out your arm increases. To find how many sleeve increases you'll need, take your desired sleeve width at the bicep and multiply by your stitch gauge. Add together the number of stitches you picked up for the sleeve cap, your underarm cast-on, and 4. Subtract that number from your desired sleeve stitches to get the number of sleeve increases you'll need. You'll do two sleeve increases on every increase row, so divide your sleeve increases by two to get the number of increase rows you need.

Subtract 2 from your underarm depth and multiply the result by your row gauge. This is the number of rows you have to do your sleeve and body increases. We'll call this number of rows the raglanline.

Divide your raglanline by your increase rows. Round up to the nearest even number.

That's your increase frequency. Now multiply your increase frequency by the number of increase rows and subtract that number from your raglanline to get the number of plain rows to be worked.

Subtract 5 from your sleeve increases. Work that number of sleeve increases at your calculated frequency. After that, you'll work the number of plain rows you calculated. Finally, you'll work 5 more sleeve increases at your calculated frequency.

Having figured the sleeve increases, you now calculate the body increases. Multiply your upper chest measurement by your stitch gauge. Subtract your back stitches from the desired upper chest stitches to find your upper chest increases. Divide by two to get the number of upper chest increase rows. Make those increases every other row at the beginning and end of the back and front sections of your sweater.

You'll work the back and front even (once you've filled in the front neck line) until just before the underarm. Take the desired full chest measurement, divide by two, and subtract your upper chest stitches from that to get the number of full chest increases you'll need. Divide by two to get the number of full chest increase rows.

Add the upper chest increase rows to the full chest increase rows to get the total increase rows you need. Subtract this number from the raglanline to get the number of rows to knit plain.

So, like with the sleeves, you'll first do the upper chest increases every other row. Next, you'll knit the plain rows even. Then, you'll work the full chest increases to the underarm.

As you're shaping the body, you're also shaping the sleeves and working the front neck:

See how the sweater is shaping up? Once you've joined the neck, you've done all the tricky knitting in the sweater.

And here's the neck being joined up:

And the sweater with half the body knit:

All Sorts of Excitement

I've made some substantial improvements to the KnitFitter. Go over and check it out! (See sidebar.)

I've also used the KnitFitter to write patterns for all sizes of the poncho sweater for children. It's a plain text file with all of the patterns included.

Enjoy!

Using the KnitFitter

If you're following my top-down design tutorial (see sidebar), you can use my open source KnitFitter program to do the math and write the knitting instructions for you.

The KnitFitter is a Python program. It takes an input file containing the following sweater fields and outputs the instruction for knitting the sweater. To run the KnitFitter program with the input file Mysweater.txt, type the following at a command prompt:

python KnitFitter.py Mysweater.txt

The input file contains parameters to specify the design and size of the sweater. Each parameter is specified by a keyword followed by a colon(:) followed by the value of the parameter. If you do not specify a parameter in the input file, the KnitFitter will use a default value.

The current input parameters are:

Patternname: {Name of Pattern}

The Pattername parameter is just a name for this particular sweater. It can be any string.

Sweatertype: [raglan, poncho, round yoke, saddle shoulder, standard]

The Sweatertype specifies the basic construction type of the sweater. It can be one of raglan, poncho, round yoke, saddle shoulder, or standard. The The default is a raglan sweater.

Style: [cardigan, pullover]

The style indicates whether the sweater opens in front or not. It can be either cardigan or pullover. The default is a pullover sweater.

Neckline: [crew, scoop, vee, shawl]

The Neckline specifies the type of neckline for the sweater. It can be one of crew, scoop, vee, or shawl. The default for poncho sweaters is a vee; the default for all other sweater types is a crew neck.

Neckdepth: {Neck depth}

The neck depth indicates the depth, in inches, of the front neck. This applies only to vee and scoop necks.

Sleeves: [none, tapered, rectangular, bell, puffed]

The Sleeves can be none, tapered, rectangular, bell or puffed. Currently, only tapered sleeves are implemented. Short sleeves can be specified by modifying the Sleevelength parameter in the sizing portion of the input file.

Mainstripe: {Main stripe}

Accentstripe: {Accent stripe}

Mainstripe and Accentstripe apply to poncho sweaters only. They indicate the length, in rounds, of the main and accent stripes on the sweater. When these numbers are specified, the KnitFitter uses them to fill in the back neck with an even stripe pattern.

Stitchgauge: {Stitch gauge}

Rowgauge: {Row gauge}

Diagonalgauge: {Diagonal gauge}

These three parameter specify the gauge at which the garment will be knit. The first two are the standard stitches and rows per inch from the knitter's gauge swatch. The Diagonalgauge is used only for poncho sweaters. It's needed because poncho sweaters are knit on the bias.

Sizingclass: [Men's, Women's, Child's]

The Sizingclass indicates which sizing table the KnitFitter should use to find standard sizes. The default is Women's.

Size: {Size}

The Size indicates which standard size to use. Children's sizes are 6 months, 1, 2, 3, 4, 5, 6, 8, 10, and 12. Women's sizes are even sizes based on chest measurements, 30-50. Men's sizes are even sizes based on chest measurements, 34-50. If you specify a standard size, all measurements are taken from the standard size chart.

The remaining parameters all specify custom sizing. They are only processed if the input file does not include a standard size. All measurements are in inches. The custom sizing parameters are as follows (shown here with their default values):

Backneck: 5

Underarmdepth: 10

Upperchest: 15

Chest: 40

Chestdepth: 3

Bodylength: 20

Sleevelength: 18

Upperarm: 16

Wrist: 8

Shoulder: 4

Sleevecap: 4

Top-Down Design: Gloves without a Pattern

Amazing!

Death defying!

Watch as the intrepid knitter negotiates the tubes and bends of finger-down gloves without a pattern!

Or even a gauge swatch.

Gloves are one of the trickiest knits to fit well. They need to fit snugly, but not too tightly. The fingers need to be long enough but not so long that they're awkward at the fingertips. Those little tubes have a small tolerance for error.

Yarn for gloves should be smooth and strong and soft. Sock yarn works pretty well, but I usually prefer sport or DK weight. I like wool/silk blends for their combination of warm, lightness, and strength.

Gloves should be knit more tightly than hats or sweaters, but perhaps not quite as tightly as socks. If you're knitting with sock yarn, choose the recommended needle size or the same size needle you would use to knit socks. If you're knitting with DK or sport yarn, choose a needle a few sizes smaller than the one recommended on the label.

Cast on one stitch for the first finger. I like to knit my fingers in pairs using magic loop or two circs, but you can also knit them singly if you prefer.

Row 1: Knit in the front and back of your single stitch twice for a total of 4 stitches. Divide the four stitches for circular knitting.

Round 2: Knit in the front and back of each stitch for a total of 8 stitches.

Round 3: Knit

The next step is to guess how many stitches around you'll need for your index or ring finger. The aim here is to get glove fingers to fit each of your fingers. The index and ring fingers tend to be about the same size, smaller than the thumb or middle finger but bigger than the pinkie. By knitting those fingers first, you have a good chance of hitting the right diameter for at least some of your fingers.

First, guess your gauge (5, 6, 7, 8, or 9 stitches per inch). If you have small hands, your target (depending on your gauge) will be (10, 12, 14, 16, 18) stitches. If you have medium hands, your target will be (12, 14, 16, 18, 20) stitches. If you have large hands, your target will be (14, 16, 18, 20, 22) stitches.

Round 4: Increase evenly across the round to get your target number of stitches.

Knit a couple of inches and then try the tubelet on the fingers of the intended wearer.

If the tubelet fits any of the fingers snugly but comfortably, eureka! you have a finger.

If the tubelet is too big or too small for any of your fingers (this has never yet happened to me, but it could), guess how much bigger or smaller it would need to be to fit your index or ring finger. Rip out the tubelet and repeat the process until you have a finger.

Now keep knitting the finger until it is long enough. It should be a little (maybe a quarter of an inch) too short because the finger will stretch part of the hand to accommodate itself. For the thumb, measure to the top part of the hand where the thumb joins.

When you have your first pair of fingers, select a different target finger. The middle finger is about 20% bigger than the index and ring fingers, the thumb about 25% bigger, and the little finger about 20% smaller. Give or take, depending on the hand of the person you're knitting for.

Put your finished fingers on holders and keep making tubelets until you have snug homes for all your fingers:

When you've finished all your fingers, string an index finger, a middle finger, and a ring finger together onto your needles like this:

The fingers need to be joined together where they meet. Some people suggest that you knit the edge stitches of adjacent fingers together. I find that this leaves a hole, which I don't like. I take the four edge stitches of each finger (two on the front needle, two on the back) and graft them to the four edge stitches of the adjacent finger, leaving all stitches on the working needles.

The grafting is fiddly and time-consuming, satisfying to the knitting perfectionist, but probably onerous to everyone else. If you don't want to graft, just tuck the loose ends of yarn inside the fingers. When you've finished the gloves, you can go in and sew up any apparent holes.

When you have the three fingers together, knit a couple of rounds, a quarter inch or so, and then join the pinkies to the other fingers.

Knit one round with all four fingers.

Next round: *K3, k2 tog*

This will eliminate about 20% of the stitches. Knit a couple more rounds, then try the gloves on. If they look like they're going to be at all loose around the palms, do some more decreases. Keep decreasing, trying on, and knitting plain until you have a nice fit for the palm.

Note that I put the pinkies on the inside in this photo. This is the wrong way to do it. Put the pinkies on the outside and the thumbs on the inside. It makes the gloves easier to try on.

Knit straight until the palm of the glove comes down to the spot where the thumb joins.

Join the thumb in the exact way that you joined the other fingers. Place markers one stitch on the hand side of the former thumb stitches. You will use these stitches to make matched decreases every 3 rounds or so until you have eliminated all the former thumb stitches.

Keep knitting until all of the thumb stitches are gone. Try on the glove. With luck, it should be about wrist length by now. If not, continue even until it is.

Next round: *K3, k2tog* cheating enough so that your total number of stitches on each glove is an even multiple of 4.

Change to needles 2 or 3 sizes smaller than your working needles. Finish with a couple of inches of k2, p2 ribbing.

Twisted Yarn

I've been knitting a pair of gloves with two colors, juggling four balls of yarn to simultaneously knit the two gloves on my circular needles.

I did my usual knitting-two-objects-at-once trick of putting each glove's yarn supply on a different side of me. By alternating whether I rotate the needles left or right, I can keep these two supplies from tangling.

The two colors used to knit each glove, however, were tangling badly. The yarn is a wool/silk blend that is both slippery and has a tendency to felt. I stopped every few inches to untangle the yarn, wishing there was an easier way to do this.

The slipperiness of the yarn was also causing it to unball somewhat, so I stuffed each ball in its own ziploc bag. That helped with the tendency of the balls to tangle, but the strands running from the balls to the gloves were still twisting and tangling.

One of the little plastic zippers accidentally closed around the yarn, bringing me up short.

It also gave me an idea.

If I stuffed as much of one strand of yarn as I could in the ziploc bag and zipped it shut, could I just dangle the ball and encourage it to unwind.

Yup. Who woulda thunk.

Unvented: Underarm Gusset

On the Broad Spiral Rib cardigan, I had an attractive Knotted Rib inset into the raglan line. As I was thinking about writing up the way I divide the arms from the body, I had an idea.

“Wouldn't it be handsome,” thought I, “if I continued the Knotted Rib through the underarm area and grafted it to the other side of the underarm? Then I could pick up along both edges of the Knotted Rib strip and have a little underarm gusset. It would probably help with the stretched stitches that I sometimes get in that area.”

The more I thought about this idea, the more I liked it.

When I got to the spot where I ordinarily do the underarm cast-on, I continued the Knotted Rib strip instead:

When I got to the end of the little strip, I grafted the end of the gusset to the Knotted Rib on the back raglan line:

At about this point, I briefly wondered if I was a little obsessed, knitting that little tiny strip and then grafting it to the other side. Could it possibly be worth the effort it was taking?

I continued knitting across the back, and then knitted and grafted the other little strip on the other underarm.

At the end of the row, I had underarms that looked like this:

Next, I picked up along the body edge of the gusset, muttering more things about how obsessive I can be in the knitting of the perfect sweater:

I'm very pleased with the results.

Top-Down Design Tutorial 10: Dividing the Arms from the Body

In all kinds of top-down sweaters, the yoke and arms are knit as a continuous unit to the underarm.

The yoke of the sweater has a cape form when you get to the underarm:

And here's the yoke open so you can see the shawl neck shaping:

When you get to the underarm, you knit the sleeve stitches to holders, dividing the arms from the body (oh, how gruesome):

The sleeve stitches on their holders:

The sleeve stitches stay on their holders while you knit the body.

On the first row after dividing the sleeves from the body, you cast on a few stitches at each underarm and then knit the body on down to the hem:

When the body is complete, you pick up along the underarm cast-on and knit the sleeves down in the round. Pick up the cast-on stitches and add two transition stitches at each end of the cast-on. If you don't pick up transition stitches, the corner stitches along the underarm will be stretched, leaving unsightly holes at the underarm. The transition stitches close the gap and make the fabric of the underarm smooth.

The steps for dividing the sleeves from the body:

1. Knit one row, knitting the sleeve stitches onto holders.
2. Knit back across body stitches only, casting on an inch or two of stitches at each underarm.
3. Finish knitting the body of the sweater.
4. Pick up the cast-on stitches along the sleeve edges of the underarms, adding 2 transition stitches at each end of the cast-on stitches.
5. Knit the arms down in the round.

Top-Down Design Tutorial 9: Figuring the Raglan Yoke

I recently updated my open-source KnitFitter project to do the calculations for round-yoke and raglan sweaters as well as for poncho sweaters.

I'd worked up a Broad Spiral Rib pattern in a lovely Peace Fleece yarn that I wanted to make into a zippered sweater. I put the gauge and my dimensions into the KnitFitter and tinkered with the raglan figuring until the program gave me a sensible answer.

This took a lot longer than I expected.

Raglans were the first sweaters I designed. I knit them exclusively for years. They're fairly straightforward to knit. They fit a wide variety of bodies nicely and work with a wide range of designs.

As I tinkered with my program, I discovered that getting the numbers right for raglan yokes is more complicated than I had realized. The logic for round-yoke sweaters is 15 lines long. The logic for raglans and poncho sweaters is 80 lines long, and it only covers the basics.

The good news is that figuring raglans is easier for humans than it is for computers. You just juggle things until they're right. You have a lot of options, a lot of factors you can fudge to get the numbers to come out right. Computers can't do that; they need to do things the same way every time.

A raglan has four or five distinct sections: the front (or left front and right front, if the sweater is a cardigan), the back, the left sleeve, and the right sleeve. Each section is separated by a raglan marker and one or more raglan line stitches. The increases for each section are done before and after the raglan markers.

In the Broad Spiral Rib Jacket, the five sections of the sweater are separated by a single Knotted Rib that emphasizes the raglan line.

Each section of the sweater grows from its initial width at the neck to the desired width at the underarm. Sleeve increases and body increases can be independent of one another; front and back increases can also be independent of one another.

The standard raglan instructions say to increase before and after each raglan marker until the sweater is wide enough and then to knit straight until the desired underarm depth is reached. For a lot of sweaters, this leads to a wide-enough sweater with appropriately-sized sleeves and an underarm somewhere in the neighborhood of the wearer's underarm. I find that this basic formula works well for children's sweaters worked in stockinette. For other stitch patterns or other sizes, it doesn't work well at all.

Okay, so a few numbers that we'll need to figure the raglan yoke:

Raglan rows: multiply the underarm depth of the wearer by the row gauge of the stitch pattern. This number tells us how many rows we have to make the raglan increases.

Sleeve stitches: multiply the desired sleeve width at the bicep by the stitch gauge.

Body stitches: multiply the chest measurement by the stitch gauge.

Neck stitches: the number of stitches cast on for the neck. In a basic crew-necked raglan, this is ((2 * back-neck) + 2) * stitch-gauge.

Underarm cast-on: A small number of stitches cast on at the underarm to form a gusset between the sleeves and the body and allow freedom of movement. In standard raglan sizes, this is usually an inch or two. I usually allow 1" for children's sizes and 1.5" for adult sizes. Multiply that number by the stitch gauge to get the underarm cast-on.

You have some variables to play with. You have a fair amount of wiggle room in dividing the neck stitches between the sleeves and body. You can adjust the underarm cast-on by a few stitches. You can undershoot the raglan rows by an inch or overshoot it by two inches.

Look at the numbers for neck stitches, body stitches, and sleeve stitches. About an inch of the neck is usually apportioned to each sleeve. So start by assigning an inch of stitches to each arm and divide the rest between the front and the back. Find the number of increases for each section by subtracting the neck stitches for that section and the underarm cast-on from the total number of stitches needed for that section.

Are the body and sleeve increases close to the same? If so, you might be able to juggle the neck stitches so that you can do the same number of increases on the sleeves as you do on the body. If not, the sleeves and the body will have to grow at a different rate.

Once you've juggled the neck stitches, figure the rate of increase for each section. Divide the number of increases you need to do by 2 to get the number of increase rows you need. Next, divide the raglan rows by the increase rows to get the increase rate. The increase rate tells you how often you'll need to do raglan increases for that section. If the increase rate is close to a whole number, your increase instructions for that section are simply Increase at beginning and end of section every [increase rate] rows until section is wide enough.

Now, if the increase rate is not close to a whole number, you'll have to do more fancy figuring to get those increases in. There are also stitch patterns where you don't want to increase on wrong side rows, so you might want to do more fancy figuring in those cases as well.

The simplest kind of fancy figuring is to round down to the nearest whole number, stop increasing when you have enough increases, and then work straight to the underarm depth. This conforms pretty well to the shape of the human body. The shoulders widen quickly, then the torso continues straight until the underarm. At the underarm, the sleeves and body of the sweater need to expand to form tubes around the arms and body. Ideal shaping for a sweater looks something like this:

Another option is to build a complex-angle raglan which conforms to the shape of the ideal sweater. This will fit better than a simple raglan. To build a complex angle raglan, you decide that you need to work so many increases, say, every row, and so many other increases, say, every other row. You put the more frequent increases towards the neck, saving an inch or two for just before the underarm and do the less frequent increases on the chest section where the body isn't changing.

Arms, luckily enough, follow a similar pattern. The sleeve cap can use a little extra ease, and the sleeve also needs extra width at the underarm.

Once you've figured out the rate of increases for each body section, you can write your pattern to the underarm.