Seventy years of thickness can't be woven
Share
I was given the chance to have denim fabric made from the warp thread up.
Send a disassembled original for analysis, then have a full bolt woven to those numbers. That's the story.
To give you the conclusion first: what I hoped for did not materialize. But the reasons it didn't taught me more than the disassembly itself. This is the record.
"Shall we build it together, from the warp?"
About a year ago, that offer arrived.
It came just as I was hoping to push the Fukuwohoru research one step deeper. A boat at the exact moment I needed to cross. I said yes on the spot.
The terms: all fabric risk on their side. I only hand over a disassembled original for analysis. No added risk to me, and the research findings are shared.
Writing it out now — there isn't a single downside on my end. I'm still amazed they brought me the offer.
What went out for analysis: a BIG E, disassembled from deadstock. My favorite era.
Why disassembled deadstock?
The conversation settled on "if we're doing this, use a disassembled piece" — and there are reasons.
Fabric analysis means measuring yarn count, twist, and density. The fabric has to lie flat for the numbers to be accurate. Areas folded into seam allowances, crushed under stitching, or tripled up in felled seams are unusable.
Disassembled, you get flat surfaces panel by panel — a full garment's worth. You can even vary the conditions slightly, measuring from the front body versus the back.
And deadstock.
That, I believe, is for shrinkage. Only a garment that has never touched water can give you the shrinkage rate. Wash it once, and all you can measure is the number after it shrank.
Disassembled, deadstock. As research material, a luxurious state of affairs.
The analysis
Here is what came back.
| Item | Warp | Weft |
|---|---|---|
| Yarn count | 6.87s | 6.81s |
| Twist | 12.3 | 10.5 |
| Density | 70 | 48 |
| Shrinkage | 8.2% | 5.0% |
The dye: pure indigo.
Unevenness overall on the weak side — judged "reproducible." Meaning today's equipment can get close to this face.
And the selvedge: ecru / pink / ecru, at 5, 1, and 6 threads. One pink line held between ecru on either side. The left and right counts differing is a nice touch.
One caveat came attached:
Mint condition or not, this is analysis from a finished garment, so twist and count will carry some error. That said, we've analyzed pieces from the same era before, and these results are close.
These are, strictly, numbers taken from this one garment. Even within the same era there will be individual variation — keep that premise in place.
Up to here, smooth sailing. The numbers were in. All that remained was to weave.
Then came two walls.
Wall 1: pure indigo can't produce a faded color
The dye is settled: pure indigo. No room for doubt there.
The problem: the target garment has been through seventy years of yellowing and fade.
Stay faithful to the original and dye pure indigo, matching only the depth — and the red cast comes out strong. The actual garment has dulled toward yellow.
Which means:
Aim for "the color it is now," and the dyer has to do something else entirely.
Aim for "the color it was then," and you get a color no one alive has seen.
Which do you aim for? That's not a question of technique but of stance. I stopped there for a while.
Wall 2: a 6.8-count yarn does not exist
The other wall was the count.
The analysis says roughly 6.8s, warp and weft. But what's actually available is a choice between 7s and 6s. There is no 6.8 on the shelf.
Worse, that 7s measures out at around 7.4s. You end up weaving with thinner yarn than you aimed for.
The result: 0.5–1oz short.
They had precedent. A previous request had also aimed at 6.8s and been woven at 7s. The comment from that time stayed with me:
The hand and the drape are better at 7s, warp and weft. But it comes out thin.
Character at 7s; weight at 6s. You can't have both. That time they solved the weight by running 6s in the weft only — which is, of course, no longer the thing you aimed at.
Move the count by 0.2 and the fabric's character changes. And 0.2 is not on the menu.
In pattern work, 0.2mm is a number I decide myself. Fabric isn't like that. You choose from what's on the shelf. I didn't understand that feeling until I actually tried.
What color for the weft?
The fabric that came up had a brown-cotton weft.
They had been looking for white. The advance communication crossed wires there — but it's not something to blame anyone for. Looking at the original, they judged "the sample fabric reads browner." And in fact, seventy years on, the original's weft looks brown, not white.
So which is correct?
Brown cotton gets you closer to how it looks now.
Ecru gets you closer to how it was then — and time will build the yellowing for you.
Their answer and mine, without conferring, were the same.
The original isn't that brown, really — the yellowing is coming out of the ecru. Ecru, with time producing the yellow, is the best answer, we think.
Exactly, I replied.
In the end, this too is about time.
What came up: 81 meters
The plan was 8–10 bolts. What came up was 81m.
Selvedge denim pants take about 2.5m per pair. 81m is roughly 30 pairs. To bring to a factory, honestly, a precarious quantity.
And picking it up — it is thin, after all.
On paper, 0.5–1oz. One ounce shouldn't feel like a large difference. And yet, side by side, the difference is unmistakable.
Here begins the part I most wanted to write about.
Cotton swells
Put the original next to the new bolt, and the original is clearly thicker.
Why? Their read:
The original's cotton has most likely swelled over time. Give this fabric a few decades and it will gradually swell too.
Cotton swells.
Perhaps obvious, once said. But I didn't have that lens. I had assumed fabric "thins" with age — abrading, losing thread, wearing down.
The interesting part: the hand tells you the opposite.
The original is far softer — decades of being handled, washed, worn. Soft, yet thick. The new fabric: stiff, yet thin.
Thickness and stiffness were different things.
I think I had conflated them all along — taking that first-touch "solidity" for thickness. It isn't.
And one more realization, arriving late:
The analysis numbers are the numbers of a seventy-year-old fabric.
Count 6.87. Density 70×48. Those are not the numbers off the loom. They are the numbers after seventy years of swelling, washing, shrinking, swelling again.
Weave to those numbers today, and what you get is not a fabric that will reach those numbers in seventy years — it's a fabric at those numbers now.
The starting line itself is different.
To be honest, I only worked this out after the fabric came up. Had I understood beforehand, I might have said "go one step heavier." I couldn't.
Then how do you derive the off-the-loom numbers of that era?
I don't know. Deadstock or not, it has breathed seventy years of air. Unused is not new. Deadstock isn't fabric where time stopped — it's fabric where time moved slowly.
I don't hold the answer here. "Which is it, I wonder" is my honest position today.
Doing it again requires a production run
The fixes for the weight were clear:
- Thicken the weft to 6.5s
- Drop the weaving tension substantially
Especially the second. Lower the tension and the yarn keeps room to swell instead of packing tight. "Tension is the main one," they said.
But running another trial had a condition:
These are samples, after all — if we adjust and weave again, a production run has to be the premise.
A production run is about 40 bolts. We cannot carry 40 bolts today.
And lead time: the next chance to weave the same fabric would be about two years out. That is how tight the world around selvedge denim has become.
Ending at the trial bolt isn't unusual — it's normal, they told me. Even so, asking for a second trial with no intention of a production run felt wrong.
So I stopped it here, for now.
What remains
A fabric number.
Once a fabric is woven, it gets a number. As long as the number remains, the density, the yarn, the dye can all be adjusted from that base. In other words, it stays re-strikable at any time. When we can carry 40 bolts, we can come back holding the answer.
81 meters.
I took these. About 30 pairs' worth.
And the numbers, and the reasons it didn't work.
This may be the biggest thing.
"Seventy years of thickness can't be woven" — written out, obvious. I needed a whole bolt woven to understand it. Expensive tuition... except I paid no tuition. They carried all of it. I can only bow.
When you disassemble, the garment in front of you starts to feel like "that era itself." In reality, you're looking at a figure that has walked seventy years from that era.
The pattern can still be taken out as it was. Lines don't age.
But fabric ages.
From one and the same garment, out come a thing that doesn't age and a thing that does. That was the most interesting part of all.
The products made from these 81m are written up on the FUKUBORI side: RE DENIM Development Log #02
Fabric research: concluded. ...Except I can't quite say that. The fabric number remains.
