Tuesday, January 29, 2013

The Puzzle of Bias


I have a vintage treasure:  an old knitting magazine that I got from some long ago rummage sale. The cover of the magazine is missing. There is no indication of publication date, but I suspect from the designs that it was published in the 1940’s. I nearly gave this magazine away in a fit of shelf cleaning, but then I noticed the section in the back containing a dozen or so stitch patterns. After a quick perusal I realized that this small collection contained several unusual stitch patterns that I’d never seen before.

In 2011, I used one of the patterns, “Cayce”, for the edges and collar of my Tattoo jacket. 

Last month, I swatched a number of other patterns from this collection.  One was “Cross Stitch Mesh”.

Cross Stitch Mesh
(multiples of 3 sts)

Row 1 (RS):  k2, * yo, k3, pass first k st over 2 k sts, rep from * end k1.
Row 2:  p
Row 3:  k1, * yo, k3, pass first k st over 2 k sts, rep from * end k2.
Row 4:  p

Here’s a picture of the swatch.


See how it biases upward and to the right? I’d experienced something similar several years ago in a shawl I’d designed with a very similar stitch pattern. I wanted to know why these patterns biased. So, I went on a theme-and-variation swatch-a-thon.

After several variations, I was convinced that it was the “pass stitch over” maneuver that caused the bias. To test this hypothesis, I knitted up a few other stitch patterns that included a “pass stitch over”.  Here are three:

Camulo (also from the vintage magazine)
(multiples of 9 sts plus 3)

Row 1 (RS):  p3, * yo, k3, sl the yo over the 3 sts, k3, p3, rep from * across row.
Rows 2, 4, 6, 8:  k3, *p6, k3, rep from * across row.
Row 3:  p3, * k1, yo, k3, slip the yo over the 3 sts, k2, p3, rep from * across row.
Row 5:  p3, * k2, yo, k3, slip the yo over the 3 sts, k1, p3, rep from * across row.
Row 7:  p3, * k3, yo, k3, slip the yo over the 3 sts, p3, rep from * across row.

Here’s a picture of the swatch:


A cool pattern I expect to use in the future.  But it did NOT bias.

I found two similar and interesting patterns in an old knitting booklet on my shelf (Nomis Knitting Primer, published in 1988 by the Nomis Yarn Co, Stoughton, MA)

Nomis Booklet, Pattern #12
(multiples of 2 sts)

Row 1 (RS):  k1, * yo, k2, pass the yo over the k2, rep from *, end k1.
Row 2:  p

Here’s the swatch, with no bias.


Nomis Booklet, Pattern # 90
(multiples of 2 sts)

Row 1 (RS):  k1, * sl1, k1, yo, pass sl st over the k1 and yo, rep from *, end k1.
Row 2:  p
Row 3:  k2, * sl1, k1, yo, pass sl st over the k1 and yo, rep from *, end k2.
Row 4:  p

And here’s this swatch. No bias here either.


So, I concluded the “pass stitch over” was not responsible for the biasing effect.  And it didn’t seem to be the number of stitches that were passed over.  So, I tried more variations of Cross Stitch Mesh to find out the real reason.

I noticed that in the patterns that did not bias, the yo was incorporated into the pass stitch over, while in the patterns that did bias, the yo was separate from the pass st over. I got to thinking that perhaps the biasing was just due to the distance between the increase (yo) and the decrease (pass stitch over).

After swatching several more variations, it was clear to me that the separation of the yo and the pass stitch over was indeed responsible for the bias. Think of it this way: if you want to create a biasing fabric, you put an increase on one edge and a decrease on the other. That’s what’s happening with the CSM pattern.

Here’s some of the “evidence”. For CSM Variation #9, I reversed the order of the yo and the pass stitch over, expecting to see the fabric bias in the opposite direction.

Cross Stitch Mesh, Variation #9
(multiples of 3 plus 2)

Row 1 (RS):  k1, * sl1, k2, pass the sl st over the k2, yo, rep from *, end k1.
Row 2:  p

Yup! It biased upward and to the left.


More evidence:

Cross Stitch Mesh, Variation #10:  it seemed that it was really only the far right increase and the far left decrease that caused the bias, so I tried a swatch that was mostly st st, but with one patt repeat.
(any number of sts more than 5)

Row 1 (RS):  k1, yo, k to 4 sts from end, sl1, k2, pass sl st over the k2, k1.
Row 2:  p

This swatch, with only one repeat of the stitch pattern, still biases upward and to the right.


For the last in the series of swatches, I wanted to verify that the specific stitch decrease you use doesn’t change the biasing. I used a SSK instead of the pass stitch over.

Cross Stitch Mesh Variation #12.
(multiples of 3 sts plus 2)

Row 1 (RS):  k1, * yo, k1, ssk, rep from *, end k1.
Row 2:  p

The swatch still biased.


I had (and still have) the urge to swatch more variations, but I stopped at 12 variations. I did play with needle size on some other swatches of CSM, only to find that needle size had essentially no effect on the magnitude (angle of slant) of the bias.

This was a compelling series for me. And I do plan to re-knit the biasing shawl that I’d made several years ago. I have been “encouraged” to have that pattern ready by May. I’ll do my best.

One other thing: when you see pictures of these biasing patterns in stitch pattern collections, they do not show you that the stitch biases! It was an interesting lesson for me to learn.

7 comments:

  1. In variation 10, the fabric isn't actually leaning, or biasing. You are decreasing on the left and increasing on the right. The st stitches are vertical. If the fabric itself were biasing, they would be leaning.

    ReplyDelete
  2. You're right, but the ultimate effect is the same; it creates a parallelogram. That's why I showed this particular variation. As I was falling asleep one night, I had an image of CSM as half a stitch on the far right (yo) and half a stitch on the far left (pass st over), and all the rest of the stitches in between have no effect.

    ReplyDelete
  3. In the cross mesh swatch, the stitch that is passed over is already knitted (tighter) than a slip st or a yo. And it is passed over two sts. Maybe those qualities make the difference.

    ReplyDelete
    Replies
    1. Oops, I meant in that first swatch, the orange one.

      Delete
    2. Hi Vicki! I didn't show all the variations that I tried, but I did change the k to a sl1 and it didn't make any difference (except that the sl1 was easier to perform than the k). I'm delighted that this has caught your attention!

      Delete
  4. Hmmm. I'll bet if you pass a yo over it won't bias. I'm betting it is the tightness of either a slip or a knit st that pulls the swatch to bias.

    ReplyDelete
    Replies
    1. You're right. See "Camulo" and "Nomis patt 12" above. Both pass the yo over and neither biases. I'm pretty sure it's because the yo (the increase) is incorporated into the pass st over (the decrease), thereby cancelling each other out. In all the patts that bias, the yo is separate from the pass st over.

      Delete