So it’s been a while since I did a temari post! Sorry about that. There’s a long story about how I had all these photos on my camera that I didn’t want to download onto the new computer when the old one stopped working, because I’d already downloaded and labeled them all, so I was waiting to find a way to transfer the data and keeping everything on the camera as backup, but the short version is that tonight I finally downloaded everything and was able to use the camera again. Hooray!

This temari, which I’ve named “Dawn & Dusk” because I think it is so awesome it deserves a pretentious, er, *artistic* name, is one that I made up on the fly one day while playing around. There’s this photo set on Flickr by Nana Akua of her grandmother’s temari that is a great place to go for inspiration, and I had been looking at this variation of the rose garden pattern. If you look at it, you’ll see that the alternating bands of light and dark color seem to make a swirl around the center, and I wanted to see if I could figure out how, (except without having to go to the trouble of marking a C10 for a true recreation of that pattern.) It appeared that she had done pentagons on a division of 20, rotating each subsequent pentagon over one row, so I figured I’d try it on a simpler division and test some theories.

I did the Dawn side first:

I divided the ball into an S12 and started with triangles, on the theory that it was the multiple of 4 division that had allowed the swirl effect. As you can see, this is not the case. Still a very pretty rose variation, but no swirling.

I pondered doing the other side of the ball the same way, for symmetry, which is what I usually do, but rose garden patterns are thread eaters and it didn’t look like I’d have enough of the same colors to do that. So, since I was going to have to change colors anyway, I figured why not go ahead and do something entirely different on the other side to complete the experimental nature of the ball. Here’s the Dusk side:

Swirls! They appeared when I used squares instead of triangles. My theory now is that the number of sides on the shape being stitched has to be 1 larger than the multiple for the division. Ex: the pentagon on a 20-division (5 multiplied by 4), the square on a 12-division (4 multiplied by 3.) Except I don’t think this will ever work for triangles… Some mathematically inclined person out there want to figure out what the actual rule governing this is?

Beyond the mathematical interest, though, I thought the ball turned out quite pretty! It did need something in the middle, though, so I did a quick leaf-like obi:

I kind of wanted the points on the zigzags to be bigger, but by the time I decided that, I would have had to rip it out and do it over again, and I wasn’t dissatisfied with how it looked to warrant all that. The reason the ball may look a little lopsided in this view is because of the way the layers overlapped on the triangles (pink) vs. the squares (blue.) Because the squares have so much more of an open middle to them, it spreads the thread out more evenly, whereas the triangles really stack it up.

So there’s the tale of the ball I decided was worthy of an actual name that started off as an idle experiment!

on November 3, 2011 at 1:08 pm |AliI suspect this has to do with rotational symmetry. The triangle, being equilateral, has rotational symmetry every 60 degrees. On an S12, each of your divisions has an arc length of 30 degrees. Since this is exactly half of the measure of rotational symmetry, you’ll line up every other row, producing that alternating pattern. A square, though, has rotational symmetry every 90 degrees, so those same 30 degree divisions will only cause you to line up every 3 rows, which yields the spiral effect. I suspect this would work with any odd division of whatever the measure of rotational symmetry is. That starts rapidly getting you into messy divisions rather quickly, though.