The bias is a mechanism

Take a strip of woven cloth cut along the warp and pull it. It gives a few per cent and then goes solid. Take a strip of the same cloth cut at forty-five degrees and pull it. It stretches by a third, narrows dramatically, and springs back when released.

The cloth is the same. The threads are the same. Nothing in it is elastic in any useful sense — a cotton yarn breaks at a few per cent extension.

The wrong explanations #

Three accounts circulate and none of them survives contact with the numbers.

“The fabric is elastic in that direction.” It is not. Elasticity is a material property and the material has not changed; a strip cut at forty-five degrees is made of exactly the same yarn as one cut at nought.

“The threads are longer diagonally.” The diagonal of a square is longer than its side, which is true and irrelevant — a longer path does not stretch more, and in any case no thread runs along the diagonal.

“The crimp comes out.” Crimp does come out, and it accounts for the few per cent available along the threads. It does not account for thirty per cent, and it is available in the thread directions where the bias effect is absent.

The right explanation is not about materials at all.

The trellis #

Model a woven cloth as what it structurally is: two families of very nearly inextensible threads, crossing, free to rotate where they cross. A pin-jointed net — a trellis fence, a garden lattice, a scissor gate.

Such a structure has a degree of freedom that has nothing to do with the stiffness of its members. The bars do not change length; the angles between them do. A trellis fence extends and contracts by tens of per cent, and no part of it is elastic.

That is the bias. Pull a woven cloth at forty-five degrees to its threads and the cells shear from square to rhombic: the cell sides stay exactly as long as they were, the angle between them closes, one diagonal lengthens and the other shortens.

The mechanism is kinematic, not elastic. It is a change of configuration in a mechanism with a degree of freedom, and calling it stretch is a category error even though it looks exactly like stretch.

The arithmetic #

A trellis cell is a rhombus of side $p$ with included angle $\theta$. Its diagonals are $2p\cos(\theta/2)$ and $2p\sin(\theta/2)$.

Unsheared, $\theta = 90°$ and both diagonals are $p\sqrt2$. Shear by $\gamma$ so that $\theta = 90° - \gamma$, and the long diagonal grows while the short one shrinks — by amounts that depend only on the angle and not at all on the material.

At thirty degrees of shear the long diagonal is about twenty-two per cent longer than it was and the short one about twenty-nine per cent shorter. At forty-five degrees, thirty-one per cent longer and forty-six per cent shorter. The cell area falls throughout, as $\sin\theta$.

Those numbers are consequences of the geometry and they are checkable. Every shear figure on this site verifies segment by segment that no thread has changed length, and a drawing in which one had would throw.

What the mechanism explains #

Once bias is a mechanism rather than a material property, a number of otherwise disconnected facts line up.

A bias strip narrows enormously. The short diagonal shrinks as fast as the long one grows, so a bias-cut strip loses width dramatically as it is pulled. Anyone who has cut bias binding has watched a two-inch strip become an inch and a half in the hand.

Bias-cut garments hang differently. Cloth on the bias has almost no resistance to shear, so it conforms to a body rather than standing away from it. Madeleine Vionnet built a career on this in the 1920s, and the technique is expensive because a bias piece needs far more cloth than a straight one.

A bias-cut hem drops. The garment shears under its own weight over time, lengthening where the load is highest, which is why bias garments are hung for days before hemming.

Bias tape goes round curves. A straight-grain strip cannot follow a curve without buckling; a bias strip shears into one.

Woven tape narrows under load. Any strap cut off-grain does this, which is why webbing is woven as a narrow fabric on the straight rather than cut from a broad one.

Watching it happen #

The clearest way to see that nothing is stretching is to watch the same net at a series of angles and check the segments each time.

The area has fallen too, by about thirty per cent, which is worth noticing because it is where the cloth’s thickness goes: a sheared fabric is thicker as well as narrower, and a bias-cut garment is measurably heavier per unit area than the same cloth on the straight.

The figures check every segment because the claim is entirely about segments. An account of the bias in which the threads had quietly lengthened would be describing a different fabric — one of elastic yarn, which behaves quite differently and is genuinely a separate subject.

Where the bias is, exactly #

A detail worth being precise about, because “the bias” is used loosely and the loose usage causes real cutting errors.

The true bias is at forty-five degrees to both thread systems, and it is where the shear mechanism is most efficiently engaged — a pull along it puts equal and opposite demands on the two systems, so the cells shear symmetrically.

Pull at any other off-grain angle and the cloth still shears, but the two thread systems are loaded unequally: one takes more tension and stretches less, so the piece skews as well as extending. That is why a garment cut a few degrees off the true bias hangs crooked and cannot be corrected by pressing.

In an unbalanced cloth, with different setts in the two directions, the most compliant direction is not at forty-five degrees at all — it shifts toward the denser system. So the true bias of a denim, which is warp-dense, is not the same as the true bias of a balanced shirting, and a pattern laid at a nominal forty-five degrees is not on the bias of both.

That is a case where the weave matrix is genuinely no help. The draft is symmetric between warp and weft in a balanced weave; the finished cloth need not be, and the asymmetry is a consequence of the sett rather than the structure.

The extension is bounded, and hard #

A material has no limit short of breaking. A mechanism reaches a configuration where it can go no further, and that limit is abrupt.

Shearing a trellis closes the perpendicular distance between parallel threads. When that distance equals the yarn diameter, the threads are touching and the mechanism has run out of travel. The cloth locks.

The angle at which that happens follows from the yarn diameter and the thread spacing alone. A cloth set at half cover locks at sixty degrees of shear; one set at eighty per cent cover locks at twenty-six. The locking angle has its own essay, because it is the constraint that decides whether cloth can be persuaded round a curve.

The practical signature is one anybody who has pulled a bias strip knows: easy, easy, easy, and then abruptly immovable. That knee is not the yarn beginning to take load. It is the mechanism jamming.

Where crimp fits #

Two mechanisms are available in a woven cloth and they act in different directions, which is why the fabric behaves so differently along and across the grain.

Along the threads, the available extension is the crimp — a few per cent, taken out by straightening the threads, and coupled so that pulling one way narrows the other. Crimp interchange is that mechanism.

At forty-five degrees, the available extension is the shear — tens of per cent, taken out by changing the cell angles, bounded by locking.

Both are rearrangements rather than stretches. Both conserve thread length exactly. They differ by an order of magnitude in how much they give, and the direction decides which one is available.

That is why a woven fabric’s extension curve depends so strongly on the angle at which it is tested, and why a single number for “fabric stretch” means nothing without one.

Why cutting on the bias is expensive #

The mechanism explains a piece of trade practice that otherwise looks like snobbery.

A bias-cut garment needs far more cloth than a straight-cut one. A pattern piece laid at forty-five degrees does not nest with its neighbours the way a rectangular one does, so the waste is large — half again, commonly, and sometimes more.

It also has to be handled differently at every stage. A bias piece stretches under its own weight, so it must be cut on a single layer rather than folded, hung to settle for a day or more before hemming, and sewn with the machine’s feed carefully balanced or the seam grows. The crimp mechanism does something similar in the thread directions and far more gently; bias distortion is an order of magnitude larger and cannot be ignored.

So the expense is not the cloth alone. It is the cloth, plus the handling, plus the skill, and all three follow from the same degree of freedom that makes the fabric conform. That is the general shape of the trade in this subject: compliance is bought, and it is bought with something.

Two mechanisms, one fabric #

Setting the two side by side makes the scale of the difference obvious, and the comparison is the most useful thing in this essay.

Along the threads, the available extension is the crimp: a few per cent, taken out by straightening. At forty-five degrees, it is the shear: tens of per cent, taken out by changing the cell angles.

Both conserve thread length exactly. Both narrow the cloth as they extend it. Both are bounded — one by the crimp running out and the other by the threads jamming. They differ by roughly a factor of ten in how much they give, and the direction of the pull decides which is available.

That is why a single number for “how much a fabric stretches” is meaningless without an angle, and why fabric testing quotes extension at nought, forty-five and ninety degrees as three separate measurements. They are not three samples of one property; they are two different mechanisms measured in three directions.

What the model does not know #

Three limits, and they matter because the trellis is a strong model within a narrow scope.

It has no friction. A real crossing resists rotating, so shear takes force even where the geometry is free. The trellis says what is possible and not what it costs, and the initial resistance a real cloth shows is entirely friction.

It has no bending stiffness. A real thread resists being bent, which is where a fabric’s hand comes from and why a stiff cloth resists shear more than a limp one of the same geometry.

It has no thread thickness except at the jam. Between square and locked, the model treats the threads as lines. Real threads flatten and interfere gradually rather than suddenly, so the approach to locking is a curve rather than a corner.

So the trellis is a statement about kinematics. It predicts how far, not how hard, and reading it as a mechanical model is exactly the over-claim this site is trying to avoid.

Where it turns up outside cloth #

The pin-jointed net is not a textile idea, and recognising it elsewhere makes it easier to trust.

A trellis fence is the same mechanism at a scale where nobody is tempted to call it elastic. A scissor lift and a lazy-tongs are the same again. A Chinese finger trap is a braid rather than a weave, and it works because pulling it shears the braid angle and narrows the tube — the same coupling between extension one way and contraction the other.

Chainmail behaves this way, which is why it drapes over a shoulder better than a solid plate could. Chicken wire does not, because its cells are rigid triangles rather than deformable quadrilaterals.

That last comparison is the sharp one. A quadrilateral of pin-jointed bars has a degree of freedom and a triangle does not, and that is the whole difference between a mechanism and a structure. A woven cloth is quadrilaterals, so it moves. Add a third thread system at an angle — as a triaxial weave does — and the cells become triangles, and the fabric loses its bias almost entirely.

Triaxial fabrics exist and are used where dimensional stability matters more than drape. They are the exception that shows what the bias actually is: remove the degree of freedom and the cloth stops behaving like cloth.

Where the ladder goes next #

The limit is the locking angle, which turns out to be the number that decides whether a cloth can be draped at all.

The consequence in three dimensions is why clothes need darts — a flat sheet cannot cover a curved surface without shear, and when the shear required exceeds locking, something has to give.

What the pictures here cannot show. Every figure here is a net of lines with no thickness, no friction and no stiffness. A real cloth resists shear from the first degree, and the force required is the part of the story these drawings do not contain.