After the loom

An unbalanced cloth wrinkles one way and not the other

A fold has to find its outer face a length, and a woven cloth finds it by moving crimp — but a fold across the warp spends the warp's budget and a fold across the weft spends the weft's, and in an unbalanced cloth the two are different numbers. A poplin runs out of crimp at 4.22 millimetres one way and 1.85 the other, so a two-millimetre wrinkle stretches its threads folded across the warp and costs nothing folded across the weft.

Worth reading first: A crease is a fold the crimp cannot supply · A wrinkle cannot settle what a crease settles · A cloth has one budget for two directions.

Every fold in the account of how a crease forms was a fold across the warp. Its axis lay along the weft, the warp ends bent round it, and the length the outer face needed came from the warp’s own crimp until that ran out — at a radius of millimetres, after which the fibres took the rest. One number came out per cloth, 4.22 millimetres for a poplin, and the account moved on to the fibres.

A cloth can be folded the other way. Turn the fold through a right angle and the weft bends round it, and the weft has a budget of its own. In a balanced cloth it is the same budget and the question is idle. In an unbalanced one it is not, and the two numbers can differ by more than a factor of two.

What follows from that is a statement about which wrinkles a cloth keeps, and it has a direction in it that a garment makes visible.

Where eight cloths run out of crimp, folded across the warp and across the weftFor each of eight cotton cloths, the fold radius below which the crimp can no longer supply the length the fold's outer face needs — once for a fold across the warp, which bends the warp and spends the warp's budget, and once for a fold across the weft, which spends the other end of the same locus. cheesecloth 11.00 and 11.00 mm; voile 2.88 and 2.88 mm; batiste 1.83 and 1.71 mm; muslin 2.60 and 2.26 mm; poplin 4.22 and 1.85 mm; sheeting 4.74 and 3.04 mm; duck 4.79 and 3.85 mm; filter 4.00 and 3.30 mm. Balanced open cloths have one radius; the poplin's two differ by a factor of 2.3, and a fold of 2.0 mm falls between them for poplin. Between its two dots a cloth folds free one way and strains its threads the other. The picture cannot show the fibres, which take the rest of the length below each dot.A cloth runs out of crimp at one radius folded one way and at another folded the otherthe poplin's two are 4.22 and 1.85 mm; a balanced cloth's are one number024681012fold radius (mm)a 2.0 mm foldcheeseclothvoilebatistemuslinpoplinsheetingduckfilterfold across the warp (the warp bends)fold across the weft (the weft bends)dots: the radius below which the fold has spent that direction's crimpcotton cloths · computed thickness
Fig. 1 Eight cotton cloths and the radius at which each runs out of crimp, folded across the warp (large dots) and across the weft (small dots). The open cheesecloth and voile have one radius each, 11.0 and 2.88 millimetres. The six close cloths all run out sooner across the warp, the poplin most of all: 4.22 millimetres against 1.85. A fold of two millimetres, ruled across, falls between the poplin’s two radii and no other cloth’s; the slider moves the fold from one millimetre to five and names the cloths it falls between.

The fold the other way spends the other end of the locus

A woven cloth has one budget for two directions. It can lengthen along the warp only by moving crimp into the weft — the interchange that every extension of a woven cloth starts with — and along the weft only by moving it back, and both moves are the same slide along one curve. Where the cloth sits on that curve when it is quoted, its reference state, is somewhere in the middle; the distance to one end is the warp’s budget, and to the other end is the weft’s.

A fold across the warp needs the warp on its outer face to lengthen. That spends the first distance. A fold across the weft needs the weft to lengthen, and that spends the second. Nothing else about the fold changes: the cloth is as thick, the outer face needs the same half-thickness over the radius, and once a direction’s budget is spent its threads must stretch for the rest.

So each direction has its own handover radius, the cloth’s thickness over twice that direction’s budget. For a fold across the warp, this is the number the crease account already computed; for the fold across the weft, it is new, and it comes out of the same locus with nothing fitted.

A poplin folded, and the length its outer face has to find. A cloth of thickness 0.332 mm bent to a radius R. Its outer face must be longer than its middle by t/2R, and a woven cloth's way of supplying a length is to move crimp — the warp outside the fold straightens while the warp inside takes on more. That runs out when the interchange budget does, at a radius of 4.22 mm for this cloth on the computed thickness and 2.80 on the measured one. Both are millimetres and a pressed crease is a fraction of one, so the coarse route gives out first and hands the problem to the fine one, which is the fibres. What the drawing cannot show is that the two mechanisms are not alternatives: every fold uses the cloth-level route first and the fibre-level one for whatever is left.
Fig. 2 The poplin folded across the warp, which is the only fold the crease account drew. Its outer face needs to be longer than its middle by the half-thickness over the radius, and the warp supplies that by straightening while the warp inside the fold takes on more crimp. The warp’s budget runs out at 4.22 millimetres on the computed thickness and 2.80 on the measured one. Turned through a right angle, the same drawing is a fold across the weft, and the weft’s budget is more than twice as large.

Two budgets that are equal in an open cloth and not in a close one

The eight cloths split cleanly into two kinds, and the split is about what stops the locus at each end.

The two open cloths are symmetric. The cheesecloth and the voile extend along the warp until the warp is straight, and along the weft until the weft is straight; their quoted states share the crimp nearly evenly, so both ends are the same distance away. Their budgets are 1.91 and 4.59 per cent both ways, and each folds the same either way — at 11.0 millimetres for the cheesecloth, which is so thin in crimp that almost any fold strains its threads, and at 2.88 for the voile.

The six close cloths are not. Each of them stops extending along the warp when the weft jams — the weft can take on no more crimp because there is no room left round the warp — and that happens before the warp is anywhere near straight. Along the weft the limit is reached further out in every one of the six. Their warp budgets run from 3.93 to 6.59 per cent; their weft budgets from 6.27 to 8.97. Every close cloth in the table has the larger budget across the weft, and every one of them is set closer in the warp than in the weft.

The poplin is the extreme, and it is extreme in the expected way: the most unbalanced construction in the table, thirty-two ends of 15 tex against twenty-two picks of 20 tex, with a warp budget of 3.93 per cent and a weft budget of 8.97. Its two radii are 4.22 and 1.85 millimetres, a ratio of 2.3. The sheeting comes second, 4.74 against 3.04, and not because of its setts, which are nearly balanced at twenty-eight and twenty-six. It is the closest set of the eight, and its weft jams when the warp has given up a quarter of the crimp it has.

That the unbalanced cloth has the widest window is not a surprise. That a balanced-looking sheeting has the second widest is, and it says the asymmetry belongs to how the locus ends rather than to the two counts on a specification.

A two-millimetre wrinkle, both ways

A wrinkle is the same fold at twenty times the radius of a pressed crease, and the wrinkle account took two millimetres as its example. For the poplin that radius sits between the two handover radii, and the consequence can be read straight off the arithmetic.

The poplin folded to 2.0 mm both ways, and where the length comes from. A poplin 0.332 mm thick folded to a radius of 2.0 mm needs its outer face 8.29% longer than its middle. Folded across the warp the crimp supplies 3.93% of that and the threads must stretch 4.36%; folded across the weft the crimp supplies 8.29% and the threads 0.00%. Of the thread strain a cotton fibre keeps 2.13% across the warp and nothing across the weft, from its measured recovery. What the drawing cannot show is friction, which holds a fold where the crimp has moved even when no thread has stretched.
Fig. 3 The poplin folded to two millimetres each way. Its outer face needs 8.29 per cent more length than its middle. Folded across the warp the crimp supplies 3.93 points of that and the threads must stretch 4.36, of which a cotton fibre keeps 2.13 according to its measured recovery. Folded across the weft the crimp supplies all 8.29 and no thread stretches at all.

The outer face of a two-millimetre fold in a cloth a third of a millimetre thick needs 8.29 per cent more length than the middle. Folded across the weft, the weft’s budget of 8.97 covers it with room to spare: the fold is made entirely by moving crimp, no thread changes length, and when the fold is released the crimp moves back. Folded across the warp, the budget covers 3.93 of the 8.29, and the threads must stretch the remaining 4.36 per cent.

What a cotton fibre does with a 4.36 per cent strain is measured rather than modelled: it gives back a little over half, and at this strain the measured recovery puts what it keeps at 2.13 per cent. That is a permanent set in the threads on the fold’s outer face, which is what a wrinkle that stays looks like at the scale of a thread.

So the same cloth, folded to the same radius, keeps a wrinkle one way and not the other. The direction is not a matter of degree. Across the weft the cloth does not reach the fibres at all, and the fibre’s recovery — which is measured and which nothing predicts — never enters.

On the measured rather than the computed thickness the poplin is thinner, 0.22 millimetres — the gap between the two is the same cloth at two moments, round sections against flattened ones — and its two radii fall to 2.80 and 1.23. A two-millimetre fold still lies between them. The threads across the warp then stretch 1.57 per cent, below the smallest strain at which cotton’s recovery has been measured, so what they keep is bracketed between nothing and 0.41 per cent; across the weft they still do not stretch at all. The direction survives the choice of thickness; only the size of what is kept depends on it.

The window, and how wide it is

The two-millimetre case is one radius. Sweeping the radius shows the whole shape.

The thread strain a poplin is left with when folded each way, against the radius. For a poplin folded to radii from 1 to 9 mm, the strain its threads must take once the crimp in that direction is spent, for a fold across the warp and a fold across the weft. Across the warp the crimp runs out at 4.22 mm and across the weft at 1.85. At 1 mm the threads take 12.66% and 7.62%; at 2 mm, 4.36% and 0.00%. Inside the shaded band one curve is zero and the other is not, so the cloth folds free one way and strains its threads the other. The picture holds the thickness fixed as the fold sharpens, which a real fold does not.
Fig. 4 The thread strain left over once the crimp is spent, for the poplin folded each way, from one millimetre to nine. Across the weft the threads stretch only below 1.85 millimetres; across the warp, below 4.22. In the shaded band between, the cloth folds on crimp alone one way and stretches its threads the other. At one millimetre both directions stretch, 12.7 and 7.6 per cent.

Above 4.22 millimetres the poplin folds freely both ways: a gentle drape, a soft fold in a sleeve, the curve of a cloth over a shoulder. Below 1.85 it stretches its threads both ways, though always less across the weft, because the weft’s crimp has paid for more of the length first. Between, there is a band 2.4 millimetres wide in which only one direction reaches the fibres, and that band is exactly the range in which ordinary wear makes its wrinkles — creases from sitting, from a bent arm, from being packed.

The balanced voile has no band at all.

The thread strain a voile is left with when folded each way, against the radius. For a voile folded to radii from 1 to 9 mm, the strain its threads must take once the crimp in that direction is spent, for a fold across the warp and a fold across the weft. Across the warp the crimp runs out at 2.88 mm and across the weft at 2.88. At 1 mm the threads take 8.64% and 8.64%; at 2 mm, 2.02% and 2.02%. The two curves lie on one another, because the cloth's two budgets are nearly equal. The picture holds the thickness fixed as the fold sharpens, which a real fold does not.
Fig. 5 The same sweep for the voile, whose two budgets are equal. The two curves lie on one another, crossing zero together at 2.88 millimetres. At two millimetres both directions stretch their threads by 2.02 per cent, and nothing about the cloth prefers one way of folding to the other.

A voile folded to two millimetres stretches its threads 2.02 per cent whichever way it is folded. It wrinkles a little, evenly, in every direction. That is the control the poplin is measured against: same thread-level mechanism, same fibre, and the directional behaviour gone because the locus is symmetric.

A garment shows which way

Shirting is cut with the warp running down the body, along the lengthwise grain. A fold across the warp is then a horizontal line: the crease at the waist where a shirt is tucked in, the rings inside a bent elbow, the bands across the back from sitting against a chair. A fold across the weft is vertical: the soft pleats down the front of a shirt hanging loose, the folds along a sleeve.

For a poplin the arithmetic says the first set is the one that keeps. At wrinkle radii between about two and four millimetres the horizontal folds stretch the warp’s threads and leave some of that stretch behind; the vertical folds are made of crimp and go back when the cloth is shaken out. A poplin shirt worn for a day should come off with its horizontal wrinkles set and its vertical ones gone — and that is the ordinary experience of shirting, which the trade has usually put down to where the body bends rather than to which way the cloth was folded.

The two explanations make different predictions, and that is what makes the second worth stating. Where the body bends decides where folds are made. Which way the cloth was woven decides which of them stay. A poplin shirt cut on the cross grain, with the warp running round the body, would keep its vertical folds and lose its horizontal ones; a voile shirt would keep neither more than the other.

The sheeting makes the same point with a narrower window and at a larger radius.

The sheeting folded to 3.0 mm both ways, and where the length comes from. A sheeting 0.382 mm thick folded to a radius of 3.0 mm needs its outer face 6.36% longer than its middle. Folded across the warp the crimp supplies 4.03% of that and the threads must stretch 2.33%; folded across the weft the crimp supplies 6.27% and the threads 0.09%. Of the thread strain a cotton fibre keeps 0.68% across the warp and 0.00% to 0.02% across the weft, from its measured recovery. What the drawing cannot show is friction, which holds a fold where the crimp has moved even when no thread has stretched.
Fig. 6 The sheeting folded to three millimetres each way. Its outer face needs 6.36 per cent. Across the warp the crimp supplies 4.03 and the threads stretch 2.33, of which cotton keeps 0.68. Across the weft the crimp supplies 6.27 and the threads stretch 0.09, which is below any measured recovery and is kept somewhere between nothing and two hundredths of a per cent.

A sheet’s creases from a night’s sleep are folds of a few millimetres in every direction at once, and at three millimetres the sheeting keeps 0.68 per cent across the warp and next to nothing across the weft. So a bed sheet’s wrinkles, too, should be mostly the ones running across its length — the ones lying along the weft — and the effect should be strongest in the closest-set sheeting, which is where the locus ends soonest.

Holding a cloth makes it worse

A sleeve at an elbow is not folded loose. The outside of the elbow is stretched along the arm, which is along the warp, while the inside folds. That is exactly the situation the one budget for two directions was about: a cloth held along the warp has spent part of its warp budget already and gained weft budget by the same slide.

The poplin's two fold radii as the cloth is held at a rising warp strain. A poplin held stretched along the warp has spent part of its warp budget and gained weft budget, because the two are the two ends of one locus. Its radius for a fold across the warp rises from 4.22 mm with no hold to 5.72 at 1% and 8.77 at 2%, and runs off the chart as the hold approaches the whole 3.93% budget; its radius across the weft falls from 1.85 to 1.52 and 1.23. A 2 mm wrinkle across the warp strains threads at every hold; across the weft it is free throughout. The picture assumes the hold is reached along the locus, with no friction keeping the cloth where it was.
Fig. 7 The poplin’s two handover radii as the cloth is held at a rising strain along the warp. Across the warp the radius rises from 4.22 millimetres with no hold to 5.72 at one per cent and 8.77 at two, and runs off the chart as the hold approaches the warp’s whole budget. Across the weft it falls from 1.85 to 1.52 and 1.23. A two-millimetre wrinkle is ruled across.

Held at two per cent along the warp, the poplin’s radius across the warp more than doubles, to 8.77 millimetres, so almost any fold across a stretched sleeve reaches the fibres. Its radius across the weft falls to 1.23. The window widens from a ratio of 2.3 to 7.1, and the direction effect is never stronger than at the elbow, where the folds are horizontal rings round a sleeve held taut along its length.

That gives the familiar pattern of a worn shirt sleeve — a set of rings inside the elbow that do not shake out — a reason in the cloth. It also says something the usual explanation does not: a looser sleeve, which stretches less at the elbow, should keep fewer of them, and by an amount the locus prices.

What the standard tests measure, and where

Crease recovery is tested in both directions as a matter of course. A crease-recovery angle test folds a small specimen, holds it under a weight for a few minutes and measures how far it opens, and it is done on specimens cut along the warp and along the weft and reported separately. A wrinkle-recovery test crumples a larger specimen and grades it against reference replicas.

The arithmetic here says the two tests sit in different regimes. The angle test folds sharply, to a radius of the order of the cloth’s own thickness, which is far below both handover radii: there the outer face needs half as much length again as the middle or more, the crimp supplies a few per cent of that, and the fibres take nearly all of it in either direction. The direction should matter little to it. The wrinkle test folds gently, at radii of millimetres, which is exactly where the window lies. The direction should matter most to it, and most for the most unbalanced and closest-set cloths.

So the prediction is specific and has a sign: for a poplin, the wrinkle grade of a specimen whose folds lie along the weft should be worse than one whose folds lie along the warp, by more than the angle test’s warp and weft readings differ. That comparison needs no new instrument, only both tests reported on one cloth in both directions.

How the radii were computed

The cloths are the eight in this collection’s table: their counts, setts and a stated packing factor give a thickness and a crimp in each direction, and the cloth relaxes along one locus whose two ends are where it can extend no further along the warp and along the weft. The warp’s budget is the distance from the reference state to the first end, which the budget account computed; the weft’s is the distance to the other end, read from the same locus.

A direction’s handover radius is the cloth’s thickness over twice that direction’s budget. The thread strain at a radius below it is the fold’s demand, the half-thickness over the radius, less the budget. What is kept is the thread strain times one minus cotton’s measured recovery at that strain, interpolated between the measured points; below the smallest measured strain it is a bracket from nothing to the lowest measured value, and above the largest it is refused rather than extrapolated. A held cloth has its two budgets measured from the state it is held at rather than from the reference.

Required of it, and not shown: that the fold across the warp reproduces the crease account’s radius to the last digit for every cloth; that the cheesecloth’s two radii agree within five per cent; that the poplin’s differ by more than a factor of two, with a two-millimetre fold keeping strain across the warp and nothing across the weft; that holding the poplin along the warp raises one radius and lowers the other at every step; and that crimp and thread strain account for the whole demand in both directions. A fold of zero radius, a hold beyond the budget, a cloth not in the table, and a fold so sharp that cotton’s recovery is unmeasured at its strain are all refused.

What the drawing cannot show, and what the model leaves out

Friction. Every figure here is the geometry: a fold made of crimp is a fold no thread has stretched, and the arithmetic says it comes back. A crimp that has moved is held where it moved to by friction at every crossing, and a tensioned cloth loses its load by the same barriers that would hold a crimp-made fold in place for a while. So “free” here means free of permanent set in the threads, not free of a fold that lingers until the cloth is shaken.

The fibres inside the yarn. Past the handover, the threads’ outer fibres bend about their own middles as well as stretching, and which of two bending models applies is the question the crease account settled for a sharp fold. At wrinkle radii both models survive, which is why only the cloth-level stretch is priced here.

The stack. Every fold here is of one layer. Where a crease crosses a seam it is a fold of several, and a crease cannot cross a seam because the layers’ outer lengths are pinned together; which way each layer was cut then decides which budget each spends, and nothing here sums them.

The weave. Every cloth here is a plain weave. A fold that falls on a float has no crimp to spend there at all, and a twill or satin’s floats lie along one direction more than the other, so the weave adds a second asymmetry on top of the budgets’. Which way it points depends on the weave, and nothing here combines the two.

The thickness while folded. The cloth is taken to keep its thickness through the fold. A real fold compresses the inside, and the thinner cloth needs less length — which moves both radii down by the same factor and leaves their ratio, and so the direction, unchanged.

Who found what

That fabrics recover differently from creases in the warp and weft directions is old textile-testing knowledge, and the standard test methods report both. The crimp-interchange locus is Peirce’s, from 1937. What is added here is the second handover radius, taken from the other end of the same locus; the window between the two, in which a cloth folds on crimp one way and stretches its threads the other; the observation that every close cloth in the table has the wider budget across the weft, and that a nearly balanced sheeting has the second-widest window because of how close it is set; the widening of the window under a hold along the warp; and the prediction that a wrinkle test should see the direction where an angle test hardly does.

Still open: whether a weave’s floats add to the budgets’ asymmetry or cancel it

Every cloth here is a plain weave, so the only thing that differs between the two directions is the locus. A twill or a satin adds a second difference. A warp-faced twill’s ends float over several picks at a time, and a fold that lands on a float finds no crimp there to spend; its picks interlace more often and carry their crimp in more places. So a fold across the warp of a warp-faced cloth has less crimp locally available than the warp’s budget says, and a fold across the weft has about what the weft’s budget says.

In a warp-dense, warp-faced cloth — a gabardine, a drill — the two asymmetries point the same way, and the prediction is a window wider than the budgets alone give, with the cloth keeping its horizontal wrinkles even more decidedly than a poplin. In a weft-faced sateen they point opposite ways and could cancel. Which one wins needs the weave-scale count of turns per end, already taken across the warp, to be taken across the weft as well and set against the two budgets for the same cloth; that is a calculation with every piece in hand and not yet made.

What links here

Computed from the collection rather than written here: the essays that point at this one.

Reads more easily once this is understood

Essays that name this one as worth reading first.

Shares its objects with

Essays naming at least two of the same things, that neither author linked.

Named objects

A flat tag is an object no other essay names yet.

BalanceCreaseFold radiusInterchange budgetPre-tensionWrinkle