1855 Warm Undersleeves Project, part 1: Swatcheriffic

First, as some of this is culled from my posts to the HistoricKnit mailing list, I ought to recommend that list to anyone else interested in converting old patterns, reconstructing old techniques, or making reproductions of historic clothing items. It's just chock full of other knitting history geeks, and plenty of good information.

Now that that's done with, on to the quotes and notes and train-of-thought-in-a-blender that is the process of working out this pattern:

Warm Undersleeve (November 1855 Godey's)

Materials: One ounce of white single Berlin Wool; quarter of an ounce of blue wool; pins, no. 14.

Cast on 60 stitches and knit in brioche stitch till the length required, about half a yard, is completed; cast off; join up the sides with a rug needle and wool and knit the frills as follows:
Cast on 90 stitches with white wool and knit three rows before commencing the pattern.
1st row - Slip 1, knit 1, a pearl 1; knit 2 together three times;
repeat from a finishing with knit 2.
2d - Slip 1, knit 1, a pearl 1; knit 12; repeat from a .
3d - like 2d row.
4th - Slip 1, pearl 1, a knit 1; knit 12; repeat from a .

These four rows form the pattern which must be repeated five times with white, then once with blue, and cast off loosely. Two frills are required for each sleeve: The upper is placed about an inch and a half above the under, which is sewed by the edge of the sleeve.

You're probably already thinking "wait, there's something wrong here!" and you'd be right. Keep reading!


First, I found it, at Hope Greenberg's Godey's Lady's book project. Then I went to Knitpicks and bought some Palette in cream, as it looked about right.

I'm planning to make up a readable pattern for this, in three sizes (s,m,l), with variations and a bonus neckpiece, with charts and stuff. Ambitious, no? YES. This is like baking with no measuring implements - you can do it if you already know how, but if you've never done it, it's all experiments until a light goes on. And you get a lot of rejects.

First quote:

I'm working up swatches for a modern translation of the lace-ruffled brioche-stitch woolen undersleeves from the Nov 1855 Godey's Lady's Book. They call for "single Berlin wool" which I believe should be about modern fingering weight, and "size 14 pins" which look to be, from the gauge shown on this page, roughly equivalent to size 2 modern needles.

I got loads of advice, and half of it said I was about right, or maybe they ought to be smaller (like 00), and the other half thought I got the yarn weight and the needle size wrong. My swatches said I was about right, and because I'm a tight knitter (I've tried everything, I just am unless I use a whole size up from what's called for on a given gauge), I opted for using size 2 or 3 needles.

Here's a quote on my reasoning, in which I don't even notice the weight of the yarn called for:

...The confusion may come from the nebulousness of the term I used - hence the annoying, still nebulous, new yarn weight standards - notice there isn't a category of actual lace-weight yarns? So did I. This is the yarn I chose. It's very light and thin, I'd say half as heavy as 4-ply baby yarn, which I've also seen described as fingering weight, but about 2x as heavy as a fluffier laceweight. It looks very similar to the wool my mother uses for her Berlin work, as well, just not quite so tightly spun (and one ply short...).


...I'm using a carrying yarn, 2-ply, fingering weight, and it seems to produce a workable gauge at either size of needles (I've only swatched on 0 and 2 so far). I tend to knit tight, so I may end up going with the larger needles just to make it work for my mother (these undersleeves will be her Christmas gift).

So I figured out that knitting it on smaller needles just gave me a smaller size, as is appropriate, and is not really helpful (Telegram for Captain Obvious!). I think this is the point where I actually settled on size 2 needles as being best for the sleeve itself.

There was some arguing back and forth about yarn and needle sizes until someone pointed out that an ounce of worsted would be unlikely to make anybody a pair of sleeves, let alone with ruffles. Enter my "duh!" moment:

I hadn't even made that connection, as it seemed perfectly natural to choose a lightweight yarn based on all the other factors. Of course the weight gives it away!

[and]

...I cannot imagine knitting a sportweight yarn on 0 or 00 (or even size 2 if one is a tight knitter) needles in brioche stitch and producing a lightweight undersleeve. It would end up being more like a potholder.

Then another lady who'd tried to adapt this pattern pointed out a writing flaw. There are decreases and no increases in the ruffles. This produces an interestingly shaped triangular swatch, but no openwork. I charted my little heart out, and swatched some and frogged some, trying to figure out what was really going on. I figured there were some open increases, somewhere, and went crazy trying stuff. And didn't write down a danged thing once I started swatching. So I had to frog a somewhat successful looking swatch and start over, this time taking notes as I went.

Here's a rundown of that experiment:

Okay, what I've got so far (October 6th):

* As has been mentioned, this pattern must be calling for a light,fingering weight yarn, just from the amount of yarn described in the pattern text. Even for a small person, 1 oz of yarn wouldn't make much of an undersleeve (my mother, the model, is a small person - she now has plump arms, but in her "skinny" youth was able to model actual period clothing without modification).

* The 90 sts cast on [for the ruffles] is WRONG. It doesn't fit the stitch pattern at all when charted, on any of the rows. After fooling around with charts for several hours, I think it's actually 96 sts cast on, as that allows for full repeats.

* The ending sts (the ones required for a clean selvedge and for the smooth repeat of the pattern from row to row) are left off the rows following the first row.

*The rows also seem to need that final knit or purl (before the two ending sts) to close the pattern repeat, though it's not mentioned anywhere. This works well if 96 sts are assumed to be the correct cast on #.

* There are definitely make-ones, yarn-overs, or some other open increase left out of line 1. I suspect copyediting issues, as [...] the ruffle would quickly dwindle to nothing and not have any openwork if worked as printed. That would make a great method for working a shaped frill, though ... with a lot of charting and swatching!

* I am not sure that the rows are actually supposed to go 1,2,3,4 - I think there are more than a few copyediting mistakes here, and its should be 1,2,4,3. The #3 row comes out backwards and disrupts the eyelet pattern if knitted as instructed, but if #3 and #4 are swapped, you get a smooth stockinette ground, with a delicate purl "rib" between rows of eyelets.

* I am swatching to determine whether the open increases should be worked between the k2tog decreases or after, and I've determined that the repeat should be the same size as the other lines pattern repeats (purl, 12 total pattern sts, purl, etc), as this continues the purl line. The two versions I'm trying out (each with a couple of different open increases) are (k2tog = /, purl = - and yarn over = O, and I show 4 repeats for the purposes of diagramming this in text): -/O/O/O/O/O/O-/O/O/O/O/O/O-/O/O/O/O/O/O-/O/O/O/O/O/O- or -///OOO///OOO-///OOO///OOO-///OOO///OOO-///OOO///OOO-

The first looks very like an eyelet stripe motif that my great grandmother worked into my baby sweaters (we still have some for comparison), and I have not yet worked the second, though I've worked lace patterns with large open areas like this before.

Okay, now I have worked a swatch of the possible variations of the ruffle (I only swatched enough of the brioche stitc to figure out how big I need ot make the sleeve for mom and frogged it after taking notes). It's BIG and ruffly:

My swatch went from 24 sts wide to well over 100 (I haven't counted yet), and is twisting up my circulars (I don't have long straights in the size I'm swatching with, but I think they'd be much easier to use for this). I'm at four repeats, and it is certainly ripply. I've determined that the increases are implied in the row 2 instructions (where it tells you to knit 12 over 3 sts on the previous row between purls, I've tried making 3 yo's after each knitted stitch from the previous row), and it works better, giving cleaner, more open eyelets, than adding yo's between the k2tog's in row 1.

What the pattern does NOT allow for is that every other set of rows there will be one incomplete repeat. It's not avoidable, at least in the swatch. The full-length piece may not have this issue.

However, I think you will not acheive /exactly/ what the pattern illustration shows by using the pattern, even if it's redacted. You will acheive a nice ripply frill with an expanding pattern of trios of lace eyelets, not a continuous row of eyelets. To acheive that look, one would do better to work the eyelets without the purls between them. I do put this down to the illustrator and engraver refining the larger "bars" between the sets of eyelets (which the pattern produces) out of the image. There was likely no-one with the job of techincal editor for illustrations then :)

The swatch in question covers two possible variations of the pattern's intended eyelet increases, and I'm happier with the second one (cleaner lines), even if it still doesn't look like the illustration. FOr one, it's not nearly as open, and for another, it's entirely too ruffly. That thing that looks like it's the size of a scrunchie? Yes, that's the 24 stitch swatch!

I may end up trying a small swatch on 4's or 7's to see if that results in anything like the picture. Just for laughs. I may well surprise myself.

Next installment of this pattern will be working toward a modern version, changes being made for usability , and notes about the neckpiece. And, hey, maybe a scrunchie swatch how-to for fun.