pattern

aswirl in spreadsheets by Katherine Hajer

I tried making a swirl jacket last autumn, and summed up my problems with the directions/sizing and decision to rip it all out in a blog post. This past week I've been off whilst recovering from surgery, so I decided to finish unravelling the work, make a new gauge swatch (more on that below), and sit down with the pattern book and a spreadsheet.

What I learned has put me off making any sort of swirl from the yarn I had set aside for it. The reasons why mostly have to do with what I found out from my spreadsheet, but also some "helper" information I found on the author's web site. My main take-away is that it would be very difficult to get predictable results if one went by the book alone. I would go so far as to say that making any adjustments for fit at all would require a virtual redesign without the information on the web site.

I didn't like the fabric hand of the original swirl jacket, so decided to go up to 5.5mm needles to get less springy fabric. The gauge swatch in the photo above is 50 stitches wide by 50 rows high. As you can see, the welted fabric (used on every pattern in the book) makes a horizontal rib. Horizontal ribs are pretty much the same width as a similar piece of stocking stitch, but their length/row gauge is going to change drastically depending on whether the fabric is unstretched, slightly stretched, or as stretched as it can be without distorting the stitches. I couldn't find anywhere in the book which actually says how much you're supposed to stretch the fabric to measure row gauge (there is an explainer on the web site since "knitters" were having trouble with it). In the book there is a lot of detail about making a larger-than-normal swatch, washing and blocking it, etc. etc., but never how to actually measure the thing. That seems like a weird omission for a ribbed fabric. Usually an author will at least put in the standard "slightly stretched" directive.

I took one stitch measurement and three row measurements based on different amounts of stretch applied (without reducing stitch width), and got for 10cm/4":

  • 18 stitches
  • 25 rows when pulled out to "maximum" (matches stocking stitch row height without distorting the stitch width)
  • 30 rows slightly pulled out
  • 45 rows fully scrunched (not pulled out at all)

Large, horizontally-draping parts of the jacket, like the lower back, will stretch from their own weight. Vertically-oriented parts, like the collar lapels, will hardly stretch at all. A big fitting issue I had with the original jacket I made was that while the back and shoulders fit well, the fronts could not be closed except by stretching the fronts to their maxium — not very flattering! It seems telling that most sizing photos show the jackets with the fronts open. If I'm going to make a large, A-line jacket, I want to know it can close. I know plus-sized people are supposed to resign themselves to cold stomachs and chests to look flattering, but I refuse when it comes to custom-made clothing. Especially clothing that claims to flatter a large variety of body types.

Okay, I had my gauge swatch — now what? I went through the pattern book, noting all of the patterns which had a similar gauge, their shape category, and the few measurements provided (just for my yoke size, size 3). The list included:

  • Winter Waves (Centred Circle)
  • Tangerine Rose (Centred Circle)
  • Silken Dreams (Off-centre Circle)
  • Strata Sphere (Off-centre Circle)
  • Shades of Grey (Off-centre Circle)
  • Plum Perfect (Off-centre Circle)
  • Wild Thyme (Off-centre Oval)
  • Coat of Many Colours (Off-centre Oval)
  • Silhouette in the Sun (Off-centre Oval)

Note that none of the Centred Oval patterns were even close to the gauge swatch, because they were all made with much thinner yarns.

Now that I had the required gauge and measurements in a spreadsheet, I started calculating the other measurements I wanted to consider, using the following assumptions/calculations:

  • the cross-back measurement: width of the non-collar stitches right before sleeve/upper back shaping began
  • armhole depth: the number of rows from the first sleeve increase to the welt at the top of the sleeve
  • body circumference: The cross-back measurement plus 2x the row depth of the fronts up to the sleeve/upper back shaping. Because the fronts/lapels hang vertically and therefore are stretched by their own weight less, I did this twice — once assuming slight stretch, and once assuming no stretch.

If my numbers are right, the swirls all seem to top out with an effective body circumference of 44 inches. That explains why the vast majority of photos of plus-sized people wearing them on Ravelry have them open, with the sides hanging well away from centre front.

For my own measurements, that means if the fabric has a loose enough hand, I'll just be able to find a swirl I can close, so long as I stick to shapes with more generous fronts. Therefore I focused on the Off-centre Ovals, since they have wider fronts.

Problem is, the book also says they have "more fitted torsos" and "slim, tapered sleeves". Okay, sleeves are easy to alter, but I wasn't sure what the "torso" part meant. The upper back, which is the only part that has any shaping? My spreadsheet was showing the smallest armhole depth to be a still-generous 14.1", so I really wasn't sure what this meant besides "not dolman sleeves".

And then I read this note on the web site about how gauge and fabric hand work together in a swirl. The 100% wool yarn I had set aside for the project was only useful for two jackets out of the entire book. Sure, I'd noticed there was a lot of cashmere and silk listed in the book, but I'd just shrugged it off — lots of books use luxury fibres in their samples. They photograph beautifully, and suppliers will donate them for the promotion the book will provide. I've made lots of natural-fibre sweaters which cost less than even one skein of a luxury-fibre yarn, though. I can almost justify a silk blend to myself, just because it's so hard-wearing, but cashmere? Nah. There were some mohair and alpaca blends listed as well, which can be more reasonable in cost, but still. This jacket was supposed to be a stash-busting project, not an excuse to get sticker shock over exotic yarn.

This time the silk and cashmere matter, because they're less stretchy than wool. Despite the welted fabric, the idea is to create a fabric that will stretch out and stay that way. Lesson learned: the welts are to:

  1. create a reversible fabric so things like lapel fold depth can remain vague
  2. create extra stretch in certain parts of the jacket (sleeve cross-measurements, those pesky fronts) to justify the hand-waving about "flattering a wide variety of body sizes".

Of the two patterns which were designed with 100% wool in mind, only one was an off-centre oval (Coat of Many Colours), and its gauge was two stitches per 10cm/4" smaller than my swatch: 20 stitches instead of my 18.

Okay, I'd already made a spreadsheet 19 columns wide; may as well see if I could follow a smaller size's directions and get the size I wanted. It's a common adjustment method, and one I've used many times before.

I entered the stitch counts for the longest sides of the body into my spreadsheet, then the goal measurement in inches I needed below that. Then I calculated out how many inches I'd get at 18 stitches to the inch, with the goal of seeing if any numbers in the smaller sizes matched what I needed in size 3.

What I found was a surprise: the size 3 numbers already matched. I figured I must have made a mistake, so I ran another row of the same calculations, using the pattern 20 stitches to 10cm/4". This is what I got.

(All measurements are in inches since that was all that was what was given on the book schematic):

The numbers in the top three rows, in red, come from the book. The first row is how many stitches to cast on for the long side for each size (1, 2, 3). The second row is how many inches the side should be per the schematic, and the third row is how wide the mitre point strip should be (this is the same number of stitches for each size, so never changes). Each side of the body shape starts and ends exactly halfway through a mitre point strip, so the full width of the long side is really:

  • long side length + (one-half left-hand mitre strip + one-half the right-hand mitre strip)

Add one mitre strip to the schematic long side length, and that tells me the length of fabric that should be between section markers. That's the fourth row, labelled "total side width".

Next, I calculated how big the section would be if I just cast on the prescribed number of stitches at the same gauge I got in my swatch. That's the row labelled "18", for 18 stitches per 10cm/4". The row below with the blue text shows the difference between my gauge calculation and the book-prescribed length. They're awfully close — it would be hard to alter to fix a .1" difference at this gauge, in a stretchy fabric like this. Even the .4" of the size 1 calculation isn't that far off.

But wait. This is the calculation from my gauge swatch, which is off from the prescribed gauge by two stitches!

So I did the same calculations over again, using the prescribed gauge (the last two rows in the spreadsheet). Unless I'm making a mistake somewhere, the prescribed gauge cannot make the prescribed measurements. The book is inaccurate with itself. I would be better off working with my bigger-needled, looser gauge swatch than what the book itself recommends.

Given that the prescribed gauge is off by about 2.5" for each size, and given that the mitre strips are 1.25", it seems the gauge was calculated from a finished garment without the mitre strips being included in the measurements. And given that there are 8 mitre strips accounting for a total of 10" of the total circumference of the jacket, that's kind of scary.

At this point, I checked for the errata. This pattern has had different errata for every single one of the four printings the book has had to date (I have a third printing). None of them mention the gauge discrepancy. To me, that raises the possibility that some of the printed errata are fixes for shaping or other structural issues raised by the gauge being off from the pattern. And the way the construction works on swirls, a knitter could be well into the pattern before they realise something is horribly wrong and they have to rip out several rows and do some of their own calculations.

As it stands, I've done more calculations for my swirl jacket than I have for sweaters I've designed myself from scratch, and still there are a lot of unknowns which I'd like to have settled before I started knitting. Like the sleeve shaping. I would have to add about four more columns to my spreadsheet to get the sleeve numbers I'd want before proceeding.

And this is the part where I give up, at least for now. Near as I can figure, swirls work because:

  • Fit is only guaranteed around the yoke/upper bodice — anything else is handwaved with marketing words like "softly", "tapered", and "gently".
  • Because it fits around the yoke, an area where wearers will notice binding or other discomfort, the jackets are seen as "fitting" when they don't actually fit in other areas. I've looked at hundreds of photos of swirls now. There's a lot of photos with cuffed-up sleeves (and I didn't even get to sleeves in this post) and fronts which are worn open because they cannot close comfortably. Swirls get around this by having cutaway-shaped fronts, which encourages people to wear them open.
  • Because the fronts aren't pulled closed, the back drapes over the hips more generously than it normally would. There are a lot of photos where the wearers imitate poses from the book and are photographed from the back, holding up the fronts to show the swirliness of the fabric colour changes. Which is very cool, but no-one walks around holding up the fronts of the jacket.

Someday, I will be in a shop with a great sale on, and I will snag a "sweater's worth" of some handpainted stuff with silk in it because I can't resist the bargain. I will get home, set it out on the coffee table, admire it, and then kick myself because I have no idea what to do with it. Then this book will come to mind, and I will realise it is Time to Make a Swirl.

But it is not this day.

Now, that raises a question. When making the original swirl jacket, I only had to break the yarn once, after the neck divide on the sleeves. The smaller bit of yarn I used for my gauge swatch for this exercise. The larger ball of yarn is about the size of a basketball, because I spit-spliced throughout.  What to do with it?

Something that doesn't mean I have to use a spreadsheet to fill in missing measurements and double-check prescribed gauges. Something that suits the 100% classic worsted wool (it's even called Classic Wool), and works with its many wonderful properties.

Something that makes me happy.