A new cap

by maggie


A new hat. It’s very warm dispute the airy appearance. I like the little peak, very pixie. Here follows a technical discussion for those interested.

The yarn is Regia sock yarn. 25g ± 105m ( ball for baby socks ).

The warp length is 1/3 metre. Which should give 315 ends (78 warp pairs on each face).

I made a figure of eight warp on the frame. I’m still having trouble keeping the warp tension even. I’m used to much longer warps. Also having to release the yarn ball for each pass of the frame gives a highly variable tension delivery. While still on the frame I separated the warp into odd and even pairs. This allows a double layer of sprang which is still attached at the bottom. I then transferred the warp to these smaller diameter metal rods on the frame.

In the photo you can see two rods at the top for the front and back face of the tube and just one rod at the bottom. This rod has twice as many pairs so the  yarns are more densely packed.


The spranging is all the same, all the way around and down. It’s describe on pages 125-7 in Collingwood. As you might imagine the last interlinking of the row is the most fiddly and confusing but take heart. The yarn pair gradually migrate around the  work in the direction of the twist. Which can mean that the edges of front and back fabrics are not adjacent.

The yarn is variegated with the stripes appearing on both faces which gives the finished cap a certain symmetry.


So the finished item. (Actual number of ends 67 pairs each face.)

The natural warp density is ± 8 ends per cm

By the end of the working the resting width was 10cm and length 31cm under tension. Each row is about 5mm high. There are thirty rows to the centre. As you can see from the photo the twist is reversed for part of the interlinking. This is because I started working upside down and didn’t realise immediately. I realised something was different, but not quite what.

Because it a tube i don’t think you an use the triangle method to work the tight space at the centre of the warp. I gradually reduced the number of knitting needles controlling the tension and then finally used two needles to pick up and drop off the warp ends. the gap i worked the the chain in was 15mm.


When I turned the frame around again to get the original twist direction. The whole row was accidentally 3B/F. This is why the strip does not match up with he outer edges of the warp.

Of the frame the sprang immediately relaxed to 13cm wide and 27cm long. I chained the loops on the top and bottom of the warp, with two warp loops per chain. If I used more loops in the sealed bottom then the pixie top would be less obvious.

Now that i have been wearing the cap the width at the mouth is 20cm and the length 17cm at rest.


Here the cap is ‘inside out’. The chain at the centre is inside, all you can see is a ridge where the twist changes direction. there is also movement in the fabric where the twist changes direction within the face of the fabric. The ‘joining’ of the front and back of the fabric occurs at the middle of the forehead (and back of neck). The fabric relaxes in on the top of the head, but is then forced out again by the chained row / doubling of number of warp ends at the ‘bottom’ of the warp. this is what makes the  pixie peak. If you start the chain with four warp loops instead of two you could minimise the second peak at the back of the head.

So does it look a bit like a Phrygain Cap?

next warp more, bigger…