July 1, 2017

For the last decade or so, I’ve been participating in international braid swaps but on the occasion of the 70th anniversary of my local handweavers and spinners guild, I organised a braid swap among members of my own guild here in Sydney, Australia.

Happily, there were nine other participants and having made six samples, we got five back with one being retained by the Guild for archive purposes. The braids were done in a range of materials and a range of braiding traditions including ply-splitting, lucet, Sami band weaving (done on a Gilmore Mini Wave loom), Lithuanian band weaving (done on inkle loom), Amerindian finger weaving (adapted to Japanese marudai) with some kumihimo on marudai and takadai, traditional and contemporary. I’m inspired to have a go at copying all of them!

My kainokuchi was inspired by a mimi-ito on a set of 18th century samurai armour held in the collection of the Powerhouse Museum, Sydney. I used 2/20 tencel, four strands per bobbin. I made the sample extra long so I could cut off small pieces for colleagues at the Powerhouse who regularly demonstrate lace making with me there at the museum’s Lace Study Centre.

My three-three twill needed no tou or sword to tension the threads while working. I was constantly checking my work as I went because the last time I did kainokuchi I got one stitch out of sequence early on meaning the “shell-fish mouth” was out of order through most of the braid. My reference for kainokuchi was Makiko Tada’s first book of takadai braids. Sometime I want to reproduce the Powerhouse armour braid more closely, with its five stitches instead of my twelve, in white, brown and pale green.



I’ve explored some possibilities of a design by Nana Akua, experimenting with ball size, colour palette and thread size along the way.

From left: 28cm circumference, pearl cotton 8 in white and orange and a tan ball; 26.5cm circumference, pearl cotton 8 in white and pale green on a bright yellow ball; 34cm circumference, pearl cotton 5 in purple and white on a burgundy ball and a 41.5cm circumference, pearl cotton 8 in orange and white on a red ball.

The most obvious feature to note is tonality and the need to get sufficient contrast between the background colour and the white thread. The yellow ball fails in this regard, but in the others the white stands out sufficiently. What I found interesting is the need in the red multi-centre ball to use very bright orange and very bright red threads, far stronger colour than I thought necessary. Optically, the orange and red in the border mix to form a general mid-tone orange, which fits the bill tonally.

To make my point about tonality, look what happens when I take the colour out. The contrast between light, mid-tone and dark tones works well with the multicenter; the others are more distinctly dark and light with no contrasting mid-tones.

Of course, if I was submitting any ball for JTA certification, I’d want to stitch it and take a desaturated, black-and-white photo first to check tonality was the best it could be!


One not so obvious feature is the sequence of stitching. After experimenting with stitching the coloured borders first, I found it easier to stitch all the white thread flax-leaf first. That involved stitching the flax-leaf motifs within each boundary, then joining the motifs via a series of triangles and diamonds. Most often, the stitching of the triangles and diamonds sorts out any geometrical issues with the motifs, but (as you can see the photo of my multi-center) care needs to be taken with the hexagons (!).

The next fun part of this journey with this Nana Akua design is an online stitchalong with TemariChallenge. I look forward to seeing what my colleagues come up with!






Here are some photos showing the stitching sequence of the asanoha (hemp leaf) pattern on the Complex 10 division temari ball, after Nana Akua. The beauty of stitching asanoha is that with each stitch sequence, the geometrical appearance of the pattern improves ever so slightly. This is because each sequence magically pulls against the previous stitch.

I’ve worked pearl cotton 5 on this 33cm circumference ball. I’m doing all the ‘white thread’ pattern before adding the coloured interlocked border. Doing it this way, and because of the size of the ball, I won’t be able to fit in three full triangles inside my border. At most, I’ll only be able to fit three colours of two layers each.

You’ll also notice that I’ve tacked each of the points where the division lines cross in burgundy-coloured thread. With a large ball, tacking is necessary.

When stitching with a light-coloured thread on a very dark background, care must be taken to bury the white threads judiciously, preferably under where the border will sit!

Photo 1. The inner pentagon inside the C10 pentagon divisions. The pins for the inner pentagon are two-thirds the distance along the division out from the pole.


Photo 2. The sunburst or matsuba (pine needles tied in the middle) inside each inner pentagon. Try to get each stitch in the centre of each triangle (think of the centre of a 90-degree angle from all three sides) but don’t stress – the following stitch sequence will equalise things visually!


Photo 3. This is the asanoha design proper. What you’re aiming for is three spokes inside each of the ten triangles, through careful nudging-and-fudging.


Photos 4 and 5. It is possible now to start work on the border and add more of the white thread stitching later on, burying the stitches under the border as necessary. For the purposes of this exercise however, I’ve stitched all the white thread before starting on the border. I have “extended” the previous lines of stitching by using a series of triangles and diamonds, taking care to always operate inside where the previous stitches meet. This helps enormously with providing the required tension and pleasing geometry.



I’m now ready to add an interlocked border around each of the original C10 pentagons. The corners of the border will fit inside the small diamond shapes. Space however will be limited! I will only be able to fit in three colours, from light to dark, with two layers each. See my next post for the result!


Here’s a relatively straightforward pattern by the renowned Japanese temari stitcher, Nana Akua. I’m replicating her original C10 design, as well as considering simpler C8 variations.

Complex 10

The original is in a C10 format and my work-in-progress is the one pictured on the far left. I’ve chosen a dark burgundy red because a strong contrast between the background and the white thread of the hemp leaf pattern is a must. You can see from the yellow C8 version pictured far right how important this tonal contrast is – the white thread won’t stand out if the background is a light or mid-tone colour.

I need to proceed carefully with this size ball, 34cm circumference, because a border involving three diamonds done in pearl cotton 5 can mean the small triangles surrounding the big squares become too tiny, resulting in a “choked” look.

Colour choices

You’ll notice how the colours are adjacent on the colour wheel. Nana Akua’s original is white and orange on a red background, so I’ve replicated this colour format with my yellow/green/white and red-brown/orange/white. I haven’t chosen dark blues or greens because flowers are rarely blue or green. The burgundy ball will be worked in very light purple (instead of a bright white) and a mid-purple (too dark a purple will blend in to the dark background).

Interlocking boundary – the flower & foliage

Nana Akua references the traditional flower temari with her large flower (white stitching on a dark background) surrounded by foliage/leafs/vegetation (two-colour border). I’ve experimented with a boundary made up of either two and three triangles (I steer clear of four because of link to death in Japanese numerology) and in one and two layers. The green-and-white single layer border looks too “thin”.

Asanoha hemp-leaf pattern

The two smaller C8 ones are done in pearl cotton 8 because of the ball size – 26 and 28cm circumference. The yellow one involves single layers of green and white; the brown one involves two layers each of orange and white. The key factor here is keeping a nice balance between the big squares and the surrounding tiny triangles. In working the traditional asanoha or hemp-leaf stitching pattern, two things are important: the inner white square needs to be at 90 degrees to the C8 division lines and two-thirds the distance out from the pole to the C8 division lines forming each large square.

Method of working

I’ve been experimenting with different ways of working. My preferred sequence of stitching is as follows:

  • division lines (either C10 or C8);
  • inner square in each of the six big squares (see picture below – the North Pole is at the top of the photo);
  • starburst inside each inner square – aiming for the dead centre of each of the eight triangles, using pins if necessary, and aim for a symmetrical hemp leaf “square” inside a square (later stitching will correct any mistakes here!) – see picture below;
  • stitching connecting the starburst with the inner square outline – start with the inner square outline and “collect” each starburst stitch as you go (this will enhance the starburst stitches;
  • stitch the interlocking two-colour ‘boundary’ (starting with the middle of each large square) making sure you go over-and-under to get the best possible interlocking in each of the three triangles (two triangles doesn’t look as effective);
  • extend each of the asanoha stitches, passing beneath the two-colour boundary. This magically “pulls up” the previous stitches.




Exif_JPEG_PICTUREExif_JPEG_PICTUREs02 orange.jpg

A student of mine sought my advice about making a small kiku herringbone temari as a special gift, with particular colours in mind. I gave her the following advice, complete with a 21.5cm circumference sample:

  1. Stitch three balls: the first to revise/consolidate the stitching, the second to think about the number of layers/rows and colours and the third to give away.
  2. The number of layers/rows for each colour is significant: a single row of colour can too often be “lost” in its surroundings, two rows makes a solid statement and three becomes a “band” of colour.
  3. The combination of colours is significant. I urge my students to work with ungen or shaded colour from light to dark in the Japanese tradition. Often the final layer can be significant, as shown in the example I provided: one hemisphere has a final row of dark blue (which seems to “lock in” the petals and reinforces the strong contrast between ‘warm’ and ‘cool’ colours) and the other hemisphere has a final row of orange which moves the colour palette more to ‘warm’ colours.
  4. The overall size of the ball can be significant (because the bigger the ball, the more the layers/rows), so I gave her two divided blank balls to consider, one at 27cm and the other at 30cm circumference.
  5. I like to keep the obi a green colour, in reference to the vegetation surrounding the flowers.

I’m pleased to report the ball my student stitched was exquisitely done!

I’ve spent the last year or so teaching temari ball stitching to beginners but have recently taken time out to think about new challenges and opportunities. New directions have developed out of having had the pleasure of meeting other temari stitches in my local area. I’d never have imagined it, but local lace makers are very proficient and experienced in stitching temari – a new source of inspiration, far more motivating than interacting with a virtual community of stitchers online with its intrinsic distance and indifference!


When I first started stitching Intermediate Level temari, I adopted the same approach I’d been using with Beginner temari: look at the pattern, visualise where threads are stitched and then jump in.

This approach didn’t work and proved to be very dispiriting.

Since then, I’ve realised there are two options available. The first involves computer technology and specifically software which allows the stitcher to plot stitches on a blank virtual ball. The process is slow, but the chief advantage is that you allow a machine to do the imagining instead of the mind.

I’ve hit upon a second option, which involves the “trial” stitching of an entire ball. The chief advantage is that a “rough” attempt at an intermediate ball can be knocked out in a relatively short period of time – one to three days working part-time. The downside is that it looks rough, is less than perfect and, unless it’s to be used later for teaching purposes, needs to be thrown away or stitched over. Yes, there is ‘wasted’ thread involved. But one can test colours more thoroughly and certainly assess spacing and numbers of threads required with much greater accuracy.

My source for my first two experiments has been Yukio Hisayama and Hiroko Takimoto’s book, Kii temari (Macaw, 1993, ISBN 4-8377-0933-3).


Test ball #1 – Kii temari #7, p.5 and p.44

I followed the recommended size of 36cm circumference, but after that I was on my own. The pattern inside the small squares is standard over-under interlocking, but certainly the stitching inside the hexagons is not standard over-under interlocking and, as shown by posts to the Temari Challenge Groups.io website, the hexagon interlocking requires more appreciation and understanding than I’m capable of.

What I did do, however, was use a standard over-under interlocking in both squares and hexagons. Standard interlocking inside the hexagons is used by the book’s authors in ball #4 on the same page 5, so the experience on my part wasn’t entirely wasted.

I abandoned the ball late in the piece mainly because the division lines were less than 1/32″ in accuracy. And frankly, if your division lines are 1/32″ out, one is are wasting one’s time stitching any “all-over” design and expecting perfection.

The good news is that I feel very confident about tackling ball #4 on page 5 of Kii temari!


Test ball #2 – Kii temari #3, p.8 and p,.32

Again, I followed the recommended size of 30cm circumference, but was left to my own devices after that. I had stitched a similar ball years ago.

I stuck with the same dark green mari, full thickness of DMC gold metallic thread for the C8 divisions, a pure optic white and a slightly darker red than Chinese Red.

What I’ve learned from an afternoon stitching this “trial” ball is as follows:

  • the ball has to be divided accurately within 1/32″ accuracy, anything more than this is fatal;
  • somehow or other the red triangles have to be stitched so they are parallel to all sides at all times;
  • I have doubts about the 1/2 marking required for the initial red stitches because the authors stitch 6 strips of 4 rows each and I only managed three – either that, or the thread is pearl cotton 8 not 5.
  • don’t interweave the strips of red in the ‘carnation petals’
  • the tone of green going around the outside of each square is wrong (and obviously needs to be wider – probably eight rows wide).


I just may have solved the problem of my tackling Intermediate Level balls. It’s just a pity it’s taken me six or seven years to come this conclusion!

I took a 17cm-circumference styrofoam ball and prepared a mari 24cm-circumference with Simple 8 (S8) divisions on a dark green background.

I stitched the “basket” first in one hemisphere of the ball. It will be made up of nine rows of stitches in three shades of green. I will add the “flower” in the other hemisphere of the ball later on.

Photo 1. This shows the equator. Adding pins 1/3 up from the equator on all eight division lines, I added extra division lines between all the pins. If you like, you can alter the distance of the pins from the equator as follows: 1cm = four rows of stitching.

Photo 2. This is a shot of the “basket” hemisphere head-on. The orange pin is my starting point for my stitching; I move to the  red pin, keeping at all times to the left of the division line. I continue the stitch around the ball till I end up at the orange pin where I started. Half the round is one side of the division line; the other half of the round is on the other side of the division line.

Photo 3. This shows the ‘basket’ hemisphere after stitching four rows.

Photo 4. This shows the ‘basket’ hemisphere after stitching nine rows.

My next step is to do the same stitching on the remaining four divisions, interleaving each stitch at the equator, so the points at the top of the basket will be overlapping.

I’ve paused momentarily in braiding the embroidery braids in Jacqui Carey’s book, Chinese Braid Embroidery, just to consolidate what I’ve been learning.

I’m finding the braids look better with the high tension that comes from the bobbins standing high up from the platform – that is, just at that point when extra thread has to be released. It goes without saying that for the first braids in Carey’s book, her advice about keeping the far bobbins as close as possible to the back uprights is excellent!

Braid JC2. I wasn’t happy with my first attempt at the purple-and-white chevrons, so I had another go. As with all braiding, I need to keep my eyes both on the bobbin movements as well as what’s happening at the braiding point.

Braid JC5. This is the green-and-white braid, reproducing one held in the British Museum. Perhaps it was the strong contrast between the very dark green and white, but (in pearl cotton 5) it has come out the best-looking braid so far. For some unknown reason, it does require ‘tweaking’ at the braiding point. As with the earlier braids in two colours, it’s very easy to mislay the bobbins and end up with a mistake, especially when working at speed; I’m getting used to “reverse braiding” and undoing any mistakes.

Braid JC1. I came across some very thin, very glossy silk-like synthetic which is two-ply. The braid ends up being 2mm wide. There appear to be inconsistencies in the way the ridges fall, but frankly I’d need a magnifying glass to analyse them; I think it all comes down to tweaking anything vaguely awry at the braiding point as soon as you detect a problem.

Importantly, this is my first attempt at joining threads as Miao braiders themselves do. With each succeeding knot and join, I got better. I need to refer to bobbin lace making for better joins. In Miao textiles, joins are disguised as part of the applique/couched embroidery process and, frankly, with such thin thread, joins are required at very short intervals. I’ve been used to warping bobbins “by hand” so far, but will upgrade to warping between two doorknobs several metres apart: the thinner the braid, the longer the thread needs to be. I wonder how long the silk threads are which Miao buy at their markets.

My aim, before too long, is to copy in all respects the tiny panel at Fig.99 in the book. I doubt it will be anything like the 8×4.5cm of the original, but I like the idea of (re-)creating a piece of Miao textile which shows the braid in context, among cross-stitch and folded silk-ribbon applique.

Photo 1. The 2ply synthetic, showing a near-empty bobbin ready for another piece of thread to be joined.

Photo 2. First attempts at braids JC1-4.






Today, braids JC3 and 4 from Jacqui Carey’s book, Chinese Braid Embroidery.

All are 2 metres in length in pearl cotton 5, one strand per bobbin, yielding a braid 5mm wide.

Note erratum page 67: The arrangement of colours for JC3 should be the starting bobbins 1,2, 7 and 8 are one colour and bobbins 3,4,5 and 6 of another colour. The surface design resembles the takuboku woodpecker braid in Japan, as featured in samurai armour.

JC4 resembles the kata-sasanami braid in Japan.

Left to right: JC1, JC2, JC3 and JC4.




Today, I worked braid JC2 from Jacqui Carey’s book, Chinese Braid Embroidery, creating a two-colour chevron braid; again in pearl cotton 5, one strand per bobbin, 2 metres in length, creating about 125cm/50″, about 5mm wide.

I decided to try attaching the thread to the bobbin using a Western bobbin lace-style slip knot. This worked extremely well, enabling me to quick-release thread almost with one hand. I noticed an improvement if I kept the bobbin thread about 3cm/1″ from the cuphook on the top; any lower down and you have to exaggerate the hand movements to keep out of the way of the cuphook; any higher and the thread and slip knot catch on the cuphook. I’m not intending to think about making my own bobbins until I’ve done all the braids in Jacqui Carey’s book.

Borrowing from the Japanese bobbin-and-stand tradition (the takadai in particular), I decided also to attach a cotton leader to the roller and secure the bobbin threads to that leader. This stopped the braid from slipping.

As I go, I’m imitating the photos of original Miao textiles she has published with each braid diagram, by keeping the colour schemes identical to the photos. This lends an air of authenticity to what I’m doing. Today’s was white and light purple.