Here’s another sample in the current series, based on 2 metres of pearl cotton 5 for each colour and five-colour diamonds.

The colour scheme is, again, not my own. I’m still far from choosing my own colours. It originates in a karakumi/diamond pattern worked on maru-dai or round stand by renowned American braider, Michael Hattori. Though the finished product looks similar, the way of working the diamond braid on the two stands is different. I’ve yet to work karakumi on a marudai; my own marudai is too small at 9.5″ diameter – I need a marudai which is 12″, or even 17″, wide.

This time, I’ve increased the number of diamonds to three and tensioning is improving. There are no gaping holes anymore in my work and no ‘slubbing’ of colours where the tension is too weak. I’m now much more keenly aware of which threads to tension when. It’s vitally important to be somewhat “mechanistic” or “industrial” in the kata or hand-movements. The hand movements have to be the same throughout; once something works, it has to be repeated. I’m constantly reminded of an experience I had in Kyoto, watching a young apprentice doll-maker at work, demonstrating at the Kyoto Traditional Arts & Crafts Museum. Every movement was carefully considered and replicated as if he was a machine. Everything he picked up went back into the same place; his movements looked automated. It was the only way to get the dolls looking exactly like each other. We in the West find this odd – we too often like to break the rules and too often feel uncomfortable about submitting ourselves to the discipline of repetition.

I know now there are three vital areas to monitor regarding tensioning: one is the top threads of the half-diamond which must be tightened before working the centre-point thread; another is going back and tensioning the previous threads on the top half of each full diamond and lastly there’s the very centre of each full diamond: I always drape these particular ones over the top of the stand which helps in the tricky reversing of direction involved for the bottom half of each full diamond. And there’s just one more: where the top of each full diamond interacts with the surrounding diamonds – I know it sounds counter-intuitive, but it just needs an extra tug to nudge the point of the diamond into position. This last one is the scariest of all, but just keep tugging!

Apart from tensioning, the other major issue at this stage of development is the ‘boundary threads’, here in yellow, which “outline” all the diamonds and half-diamonds. It’s vital to keep these ‘passive’ threads out of the way of the ‘worker’ threads involved in each diamond, so I’ve used plastic freezer clips/barrettes from IKEA to ‘keep me honest’. The clips are not critical in working so few diamonds, but they help.

Obviously I’m now on the way to working a Four Diamond sample. But first, I want to backtrack and work a Double Diamond again, this time with six colours rather than five. The six colours will allow me to work a combination of small and large diamonds; the varying sizes of diamonds is aesthetically enticing. But only achievable with six colours! More on that later.

I’m deliriously happy that I now no longer need to refer to diagrams. On reflection, it’s taken me about 100 hours all up to memorise the hand movements.


Working the mitsubishi (Triple Diamond)


Start of the five-colour mitsubishi (Triple Diamond)

Reasonably comfortable with my competence regarding the required hand-movements, I upgraded from 8-ply acrylic yarn to pearl cotton 5. The dimensions of this sample are the same as the previous ones: for each bobbin, 2 metres of thread, folded in half and attached via a cow-hitch or strangle knot or half-hitch to a plastic curtain ring.

This is a “5-element diamond” design; by that, I mean that the colours at the curtain ring, from left to right, are two lots of five pairs of threads, or set up (from left to right) Grey-Yellow- Pink-Purple-White-Black-Black-White-Purple-Pink-Yellow-Grey. This is repeated twice for a Double Diamond braid: Diamond 1 ([GYPiPuWB][BWPuPiYG])+Diamond 2 ([GYPiPuWB][BWPuPiYG]).

Why do I mention 5 in the sequence of colours? I mention it because if there are six colours, it’s possible to work small and large diamonds, mixing up the regularity of the full diamonds so their colours intermingle. Later down the track, I’ll work up a sample of “6 element diamonds” to show what I mean.

For the moment, though, this sample was devoted to trying to memorise the hand-movements and to acclimatise myself to the tensioning required in using pearl cotton 5.

What became increasingly obvious during the braid was how to deal efficiently with what I call “boundary threads”, that particular colour which runs around all the diamonds but is never part of the ‘worker’ pairs producing the diamonds and half-diamonds. In this sample, the ‘boundary threads’ are grey.

The colour scheme is not my own – it comes from watching the Youtube video featuring UK braider Marjie.

My next sample progresses from Double Diamonds to Triple Diamonds, requiring me to source additional EzyBob plastic bobbins from the current 48.

This sample is 3cm or 7/8″ wide and 35.5cm/14″ long (braided), finishing in 4cm/1.5″ of twelve four-element braids, acting as a fringe, done while the bobbins are still on the threads and strung over the four sides of the braiding stand.


Double Diamonds Karakumidai

karakumidai s pretty collage

Rather than revising traditional karakumi braiding on the karakumi-dai,
jumping in at the deep end by using very fine (and slippery and
expensive) Karakumi Silk, I opted for self-directed exercises, albeit
slower and more carefully-considered than a formal braiding workshop,
lasting eight hours a day for a week.
My starting point was Steve Pretty's useful instructions as published on
behalf of the UK Braid Society, a class handout from Makiko Tada's
class in 2014, a pdf outlining the experience of another braider and
the very fluid braiding style on show in a Youtube videoclip.
Following Steve Pretty's exercises in a very general way, the points of
difference in my braid samples were:
* threads were attached to a small plastic curtain ring,  instead of
something flat, like a double-eyed or long single-eyed needle;
* threads were secured with a larks head knot (with consequential
narrowing once the braiding started); 
* for the chevrons and single diamonds, ten lengths (two metres/two
yards long and folded in half) of five 8-ply acrylic yarn, graded
from white through yellow, orange, red and dark red; they yielded
braid samples half a metre (18”) long; 
* small 4cm/1.75” plastic EZ Bob bobbins (without cotton thread
leaders), rather than traditional hiradama bobbins.
The finished braid samples were about 0.5m/18” long (50% takeup) and
2.5-3cms/0.5-1” wide.
Steve Pretty Exx.1&2.
The process involved becoming familiar with the thickness of acrylic
yarn, appropriate finger-tensioning and familiarity with the
kata/bobbin movements. My distribution of colours didn't match Steve's which
proved problematic: my advice is to adopt his colourway and then
change the colours in future samples. The sample took eight hours.
Steve Pretty Ex.3.
The main challenge here was making the necessary adjustments in
moving between braiding half-diamonds and full-diamonds. Each full
diamond (and equivalent two half-diamonds) took about an hour each,
but I had the advantage of prior experience and visual schematic
instructions from Makiko Tada's workshop. The sample, twelve single
diamonds in total (not counting the two half-diamonds in between),
took about 10 hours.
Steve Pretty Ex.4.
By the time I'd finished this sample, I was working independently of
Makiko's visual schematic showing bobbin movements.  The key learning
was what I call the path of “structural threads”, that is, the
thread pair which crisscrosses the braid forming the boundary of more
than one half- or full-diamond. The secret is to consciously
“exclude” the 'structural thread' from the karakumi interweaving
threads in each half- and full-diamond: consider instead the
structural thread as a 'boundary' or 'framing' device. In my sample,
the darkest red is quite separate from the white-yellow-orange-red
which are the busy 'interweaving' threads. The darkest red threads
both move diagonally across the width of the braid (literally from
one edge of the braid to the other, framing three full diamonds and
two half-diamonds). Critically, they lock together the full- and
half-diamonds along the way. The sample, eleven double-diamond sets
(not counting the single full-diamond and two half-diamonds in
between), took about 14 hours.
Acrylic yarn is hardly very beautiful, though threads naturally bind
themselves to each as part of the braiding process, unlike more
slippery threads with a tighter twist like Japanese karakumi silk. In
terms of progressing to karakumi silk, it's useful to consider pearl
cotton as suggested both by Steve; pearl cotton #5 and #8 are thicker
than one type of traditional karakumi silk I have; pearl cotton #10
is closest, but a wide range of colours may be difficult to obtain.
Working with graded colour is particularly useful in learning the bobbin
movements; braiding in a single colour is not recommended until the
kata become automatic. 
Finger-tensioning problems are obvious in the beginning, but improvement 
is relatively fast. Similarly, gaps caused by incomplete understanding of
“structural thread” paths are obvious in the beginning but
quickly disappear.
During this revision stage, a mirror showing the backside of the braid isn't
required, but I imagine with formal braiding it may be useful, with
checks required at the end of braiding each diamond.
The yotsu-gumi/yotsu-me knots as a finishing device, indicated by Steve,
definitely help secure final diamonds at the end of the braid from
The finished braids will show up any errors in the middle of the
half-diamonds: a twist is required in the threads as they hit the
edge and then reverse back into the braid.
What's next? A sample of double diamonds in pearl cotton 5, mainly to
consolidate the 'structural thread' paths and finger-tension
associated with this yarn. Other options include applying oblique
interlacing to the finger-woven organza ribbon scarf and the
kumi-muffler, as taught by Makiko Tada, as well as revising karakumi
on other braiding stands - sasanami on maru-dai and anda-gumi on
taka-dai stands.

Fujinami no Kaede, How to do karakumi: early period Japanese flat braids:
Howes, Yuko (Japan Outpost), Youtube video:
Triple Diamonds. Particularly useful are the kata for the centrepoint of
each full diamond.
Pretty, Steve. “An Introduction to Karakumi: how to make braids on the
karakumidai”, in Strands. Braid Society (UK), 2009: 24-31. 
See also;



In her second book on takadai braids, Makiko Tada details four different flat braids of the ryuko (dragon/tiger) design, all done with 50 bobbins on the takadai braiding stand. I worked my first one about seven years ago, but am doing them now as ‘miniatures’ with just 4 and 6 strands (20/2 weaving yarn) per bobbin, producing a much narrow braid than I’ve done in the past.

I got the idea of “miniature” braids from ‘apprentice’ braids on public display in the museum section of the Adachi braiding workshop in Kyoto. In addition, I got specimens of sageo from eBay and saw examples in my local art gallery, The Art Gallery of New South Wales, Sydney, with their fine Japanese swords on display, with sageo attached. I’ve seen similar at the Osaka City Museum and I’ve got my own insignificant Korean repro sword, with a double-layer takadai sageo braid (a pickup braid with less than the usual 66 or 60 bobbins).

My warps are four metres so I thought I’d get around 60″ or so of braid (I scored 2.72 metres). There are four ryuko braids: Ryuko-gumi, Ryuko-shima-gumi (“striped”), Uroko-ryuko-gumi (“dragon scales”) and Hirosuji-ryuko-gumi (“wide striped”). All of these are done with equal numbers of bobbins in two colours. Working at the rate of several centimetres a day, this first one took the best part of six weeks.

The function of these braids was sageo or sash for Japanese sword, attached to the metal loop on the saya or scabbard of the sword and attached to another braid or running entirely around the waist proper, under or over the heko-obi or men’s obi waist sash. We know this design was used for sageo in the Edo Period because it’s mentioned in the kumihimo classic, Shika Suuyou, a braiding manual of 1826. So the aperture of the saya is the governing factor for the width and depth of the braid. I’ve not designed these for any Japanese sword in particular.

First up is the Ryuko-shima-gumi (shima meaning “stripe”). There are just eight hand-movements through one complete cycle; it’s nice to know that there is consistency in the “big” jumps between bobbins which are always over six. I did this in Praslon (synthetic) 2/20 weaving yarn, so it as a very matt appearance compared to shiny silk. I’m okay with this pattern and its braiding to upgrade to rayon or silk sometime.


I successfully submitted this as part of the Complex Weavers kumihimo Study Group’s six-monthly braid exchange.


Makiko Tada, Comprehensive Treatise of Braids IV: Taka-dai braids 2. Tokyo: Texte, 1998. Pattern 48.

I called by the Art Gallery of New South Wales, Sydney, this weekend to check on what Thai art is on display. I also checked the Japanese exhibits and there has been some re-location of works: there is a spectacular green kimono in ro, gauze weave with gold paulownia flowers. On show currently are netsuke and inro too, plus sword blades and sword hilts and accoutrements.

I made some quick notes on a flat braid in a dark chocolate brown and cream. The colours are not dissimilar to illustrations of similar flat takadai braids in the book devoted to the takadai braiding loom published last century by the Domyo School, a braiding studio in Tokyo. No reference was made to the kumihimo braid in the exhibition labels, so I can only conjecture about the age and origin and maker of the sageo sword hilt braids.


The notes won’t mean much to a non-braider or someone not familiar with takadai double-layer pickup braids, but basically it’s cream on one side and brown on the other; every 6 inches or so the colours reverse appearing in a long lozenge, perhaps around 4-5″ long, in the centre of which is a white paulownia flower, the traditional stylized design ubiquitous in Japan of three descending flowers from a bunch of leaves.

My next step is to convert the surface design into a graphed pattern I can use to braid it. I’ll post it here when I’m done.  If it’s a “routine” double-layer pickup braid, it will be braided with 66 bobbins, but I need to work out whether it’s braided with fewer bobbins. This is all seredipitous because I was wondering what sort of braid to tackle next, given the sageo braid currently on my takadai is about to finish. I’ve allowed myself 12 braids to be braided in 2013 and already the first one was not finished by 1 February! I am encouraged by the fact that even at 15mm wide – quite a narrow braid by my standards – , it’s working out well.


There were other sageo attached to other swords in this exhibition, but I’ll concern myself with them later.

ito sword braid

I’m reviewing what I know (and don’t know) about Japanese sword braids, because alongside braids for making Japanese decorative knots, for making obijime waist sashes to wear with kimono, there are at least two more kinds of kumihimo, being made today, which are functional and associated with Japanese swords: the thin tsuka-ito used for wrapping the hilt of the sword and sageo used for attaching the sword to the waist.

To begin, tsuka in Japanese means “sword handle” and ito means “thread”. There are various types of ito associated with Japanese samurai armour and accessories, so tsuka-ito contains the idea of “handle wrapping”.

Today my length of dark red tsuka-ito arrived from China and I thought I’d compare its width with a bunch of sundry other Japanese braids I’ve made, test or learning braids with no anticipated functional purpose.

From the top, then:

1. photo of a sageo attached to a Japanese sword in the Osaka City Museum. They have a special section on permanent display devoted to sword blades as well as examples of sageo such as this, as well as a range of tsuka-ito braids of many many colours in drawers (I imagine they must have someone come in and demonstrate to the public how the tsuka-ito is tied to the sword hilt.) Like others of its type, you can see its not too thin and nice and plump. I don’t know for sure how wide sageo are when used by iaido sword play practitioners, but you can see that my braids approximating this type are around 1.5cm to 2cm wide. The thing about the sageo on this braid is that it’s tied in a standard decorative, non-functional knot – for display purposes only when not in use. These days, there are videoclips on Youtube and elsewhere showing how this elaborate display knot is tied.

So if I’ve posted a photo of a sageo here, what has that to do with tsuka-ito? The white-outlined kikkoh braid is obviously a sageo, but have another look and take in the thinner dark green braid underneath. I’ve no idea what this green braid is doing here but it looks awfully like a “stray” piece of tsuka-ito, since tsuka-ito is supposed to be on the sword hilt, not on the outside of the sword as here.

2. The second is the dark red Chinese tsuka-ito, 1cm wide, made from rayon thread and commonly available for Japanese sword enthusiasts. It’s sold by the metre or, in this case, in 10m lots. It’s a double-layer Takadai braid so if I was attempting to copy it, I’d be going for 56 bobbins using a standard, single-colour Nimai-Kourai-gumi pattern, which would give me six ridges (whereas 60 would give me seven). Most Japanese sword experts would stick to Japanese tsuka-ito and perhaps go for traditional silk. The everyday variety is Chinese, of rayon. It’s possible these days to see it available in cotton (in 6mm and 8mm), leather, high quality doeskin and suede. It is normally sold as 10mm wide, i.e. when “relaxed” which becomes 8mm wide when pulled tight on the sword handle.

3. This is the two-colour une-gumi sample I’ve braided this week on Takadai. I prepared the bobbins with only six strands of 20/2 weaving yarn so the whole thing felt like I was braiding a miniature braid. When I finished, I had a flashback to the Kyoto kumihimo studio/workshop/retail outlet, Adachi Kumihimo-kan, where I saw what looked like miniature braids, as I imagine apprentices would braid – what seemed like 50% the size of ordinary obijime, arranged vertically on cardboard supports.

4. What follows are braids I’ve done in years past. I guess I had a need to make ‘substantial’ braids with some heft and handle. This fourth is the kikkoh from Rod Owen’s book on Takadai and it’s 15mm wide.

5. My test Saidaiji-gumi has nothing to do with Japanese sword braids. On the contrary, the original was 3mm wide, a square braid, and was used for wrapping up Buddhist sutras. Mine is nearly 9mm wide. Its religious function is a far cry from braids deployed in armour for the battlefield.

6. Next is a 2cm-wide braid, another plain kikkoh, red hexagons on a yellow ground, done this time on Ayatakedai.

7.  The last one is another Ayatakedai braid (this time done in 20/2 tencel), 22mm wide.

Here’s how the tsuka-ito is tied with its distinctive cross-over point. The photo below left shows how it’s tied on pieces of wood; the second below right shows it wound on the hilt of the sword with the stingray skin (same in Japanese) underneath showing through. The textured rayskin was excellent for keeping a strong hold on the sword when in the battlefield. The skin is normally an off-white colour; if it’s any dyed colour (e.g. red, black), it’s as a  result of lacquer being applied after the skin is fitted to the sword and before the tsuka-ito braid is wound on.

STAA2  800px-Tsuka_

Actually braiding by hand a length of tsuka-ito is more an academic exercise than anything else these days. A 12-inch sword handle would require 16 feet of tsuka-ito and commercially machine-made braid in silk is available for as little as $US3.75 a foot. That said, it is possible to buy, for 24 pounds sterling per metre, silk tsuka-ito braided in Sasanami (in one or two colours), so a kumihimo braider could create some bespoke tsuka-ito in Sasanami-gumi (see Makiko Tada, Book III-Takadai I, design 14 on p.37 with 49 bobbins) but it would have to be just 10mm wide.

Will I have a go at braiding some tsuka-ito in an “exotic” pattern such as sasanami? Sure, but probably only in the context of moving towards examples of thinner “miniature” Takadai braids than I have in the past, rather than metres and metres of tsuka-ito. Far better to have a go at making 2-metre-long sageo braids instead. More on sageo in an upcoming post!



10/2 dyed tencel, 6 strands per bobbin; 24 bobbins, square braid using the standard Saidaiji-gumi bobbin movements.

Yesterday’s revision of Une-gumi quickly came unstuck when my red yarn started to give way under the pressure of being worked on a takadai loom.  As a weaving yarn, it was a tad weak and I’m scouting around for some substitute yarns.

I quickly hopped back in the saddle though with some leftover 20/2 tencel (in Persian Red, Green and White) still on bobbins from a previous project. Knowing how slowly I braid Saidaiji-gumi, I was thankful the lengths weren’t too long! It feels strange braiding with a yarn that is now endangered if not extinct, as a so-called environmental wonder, it’s no longer being sold anywhere as a dyed yarn, as beautiful as it is.

Because of is (unexpected) pattern, I’ve dubbed this 24-bobbin braid “Harlequin”. So far, I’ve braided eight full repeats of the lozenge in about five hours or so (a bit less than six inches).

The eight bobbin movements are standard Saidaiji-gumi, as per Makiko Tada’s Book 2 of Takadai Braids, braid 71 on pp.152-154: basically you create a square braid with an “inner tube” (I think).

The visual problem associated with Saidaiji-gumi is that braider only sees the side of the braid (right photo) whereas the “proper” identifiable pattern occurs on the sides (left photo). I’m in the habit now of checking my work visually every eight movements but frankly it’s nearly impossible to reverse the braiding if I’ve made a mistake.

My takadai set up was Left arms: Green,Green,Red,Red,Green,Green (top to bottom) and Right arms: Red, White, Red, White, Red, White (also top to bottom). So where are the 56 bobbins you need to braid the standard, authentic Saidaiji-gumi? Well, here’s the interesting bit: I’ve managed to show a mathematical progression in this braid such that the smallest number of bobbins you can create the bar-and-lozenge pattern is 24. The next size up is 40 bobbins; the next size up is 56 (the standard Saidaijigumi) and the ones after that are 72 bobbins (‘half” the standard Chuzon-ji-gumi, which reaches the outer limit of most takadai braiding stands in the West) and 144 bobbins, the full version of Chuzon-ji-gumi  (and so on to infinity). Culturally what’s important is that, as historic temple braids, Saidaiji-gumi was done on 56 bobbins (or its equivalent if done in the hand or on marudai) near Kyoto while Chuzon-ji-gumi, found hundreds of kilometres to the north of Saidaiji up in Northern Honshu, was made with 144 bobbins, using the same hand-movements

See my earlier post on Saidaiji-gumi for schematics showing the progression from 24 to 144 bobbins. Now I’m sure Kinoshita showed this in her book in Japanese, Archaic Braids, but it’s been personally fascinating to see this relationship for myself. While I’ve braided short lengths using 56 bobbins and 72 bobbins (Han-Chuson-ji-gumi or “Half” Chuzonjigumi, as opposed to the “Full” version using 144 bobbins), I’ve not braided before the smaller 26 and 40 bobbin braids. So it’s nice to have some actual braids showing the mathematical relationship.

In a similar (but different) way, braiding colleagues elsewhere have recreated Saidaiji-gumi on marudai and outsized O-marudai, so they are real breakthroughs too.