A rib pulls back on a force the loop supplies

A cuff has an awkward specification. It must open far enough to admit a hand, close far enough to grip a wrist, do the first at almost no load, do the second every time, and go on doing both after a hundred washes. Nothing that behaves like a material does all of that: a stiff material fails the first requirement and a soft one fails the second.

A rib does it by separating the two. The extension comes from folding, which is geometry and costs nothing. The return comes from the loops reconfiguring, which is a force this collection could not compute until now and which turns out to be under four newtons a metre across the whole working range.

The folding is the extension #

A one-by-one rib draws alternate wales to opposite faces of the fabric, so the fabric corrugates. In its relaxed state each wale hides the one beside it, and the fabric’s width per two wales is about the width of one.

Open it out and the wales rotate into the plane, and the width roughly doubles before any loop has changed shape at all. That is the first hundred per cent of a rib’s extension and it is free, in the sense that it costs only whatever it takes to unfold a corrugation.

This collection has said that much before and it has never had a number for what comes back.

Where the return comes from #

Not from the fibre. A cotton fibre recovers a few per cent of extension elastically and a wool rather more, and neither is what a cuff is relying on — a knit recovers for a reason that is not its fibre’s: a cuff made of cotton grips a wrist perfectly well, and cotton has almost no rubber-like recovery in it.

The return is the loop wanting to be the shape it was set into. Once the yarn’s natural shape is the relaxed loop, every departure from it stores bending energy, and the fabric pulls back with the gradient of that energy — which is exactly the load–extension curve, read in reverse.

What the force is #

Small, and rising monotonically. At a fifth of extension it is under a newton per metre of fabric; at a half, one and a half; at a doubling, three and a half.

Over a cuff of ten centimetres’ circumference that is a few tenths of a newton — tens of grams-force — which is about right for something that should grip without marking. And it is a rising force, so a cuff opened further grips harder, which is the behaviour a wearer expects and which a constant-force mechanism would not give. It is also, at these loads, three orders below what a woven cloth would need to reach the same extension.

Why the two mechanisms are worth separating #

Because they fail differently, and a cuff that has gone slack has failed in one of two ways that need different fixes.

The folding can be lost. A fabric stretched far enough that the wales stop returning to their folded position has lost its geometric extension, and no amount of loop recovery brings it back. That is what has happened to a cuff that is permanently splayed.

The recovery can be lost. A yarn that has been re-set in the open position — by heat, by a hot wash, by prolonged wear — has moved its own natural shape, and the loops no longer want to return — which is the setting mechanism working against the garment instead of for it. That is a set problem rather than a geometry problem and it happens without the fabric looking different.

The first is a mechanical failure and the second is a thermal one, and telling them apart matters because the second is preventable by specifying a fibre and a finish and the first is not.

What the curve looks like over a wearer’s range #

The useful reading is not the whole curve but the piece a garment uses, and it is worth putting the two on the same axes.

That is why the rib figure at the head of this essay is drawn on its own axes. A curve whose interesting range is a twentieth of its extent is a curve that needs two figures, and drawing only the full one is how a real working region gets described as “approximately zero”.

The general point is worth carrying: a load–extension curve for a knit is nearly always plotted to failure, and nearly always used over the first tenth of it.

Why a rib and not a slacker jersey #

The obvious alternative to a rib is a plain fabric knitted loosely, and it does not work. The reason is now stateable.

A slacker jersey does extend further, but its recovery force falls with the cube of the loop length, so the fabric that gives most is the fabric that returns least. The two requirements move in opposite directions along the only lever a plain knit has.

A rib decouples them. Its extension comes from folding, which is set by the structure rather than by the loop length, so the loop can be knitted tight — with a high recovery force — while the fabric still opens by a factor of two. That is the whole reason ribs exist, and it is a structural trick rather than a material one.

What this ladder cannot compute about it #

The folding. The wales rotate out of the fabric’s plane, and the model is planar: it has no way to represent a wale leaning to the back while its neighbour leans to the front.

So the fifty per cent pull-in quoted here is geometric and classical rather than computed, and the force to unfold a corrugation is not available — the same absence that leaves the curl radius uncomputed. What is available is the loop’s own contribution, which is what the fabric returns with once it is open, and that is the half a cuff’s designer needs.

The honest statement is therefore: the extension is asserted from geometry, the return is computed, and the crossover between them — the extension at which folding is exhausted and the loops start to work — is not known.

The same asymmetry that curls a jersey #

The folding is not an accident of the structure; it is the same loop asymmetry that makes a plain fabric curl, acting sideways instead of along.

The two directions are worth separating carefully, because the connection is easy to overstate. A jersey’s curl is the loop’s asymmetry resolved out of the plane: the fabric has no partner face to pull against, so the asymmetry appears as a moment and the edge rolls. A rib has the same asymmetry on both beds and the two faces are back to back, so the moments oppose one another out of the plane and there is nothing left to roll — which is why a rib lies flat. What is left instead acts along the course, drawing the two faces together and pulling the fabric in. One asymmetry, two structures, two entirely different visible consequences.

So a rib’s pull-in and a jersey’s curl are one mechanism seen twice. That is a satisfying connection and it is also a limitation: neither is computed here, because both live in the dimension the model does not carry.

The two-bed structures, side by side #

Reading the three structures as a sequence makes the trade-off explicit. A plain fabric has no folding, curls at every edge, and is the most extensible per unit of yarn. A rib folds, does not curl at its selvedges, and buys its extension structurally. An interlock cannot fold, does not curl at all, and is the most stable and the least giving.

Each is the right answer to a different question, and the questions are about which of curl, extension and stability matters most. There is no structure that gives all three, and the reason is that curl and folding are the same asymmetry — removing one removes the other.

What an interlock does instead #

Interlock is two ribs knitted together and is symmetric by construction, so its wales cannot fold and its pull-in is small.

That makes it the stable fabric — it does not curl, it does not corrugate, and its dimensions move less than either a jersey’s or a rib’s. It is also the least extensible of the three for the same reason, and it needs about twice the yarn for a given area, which is what a fabric’s weight is made of. The three fabrics are three positions on one trade-off, and the trade-off is between having a geometric extension mechanism and having dimensional stability.

What decides how long a cuff lasts #

The failure mode is worth pricing, because a cuff that has gone slack is the commonest complaint about knitted garments and it has two candidate causes with different arithmetic.

A fabric with no energy minimum of its own does not return to a preferred state; it returns to wherever friction stops it, and each cycle can stop it somewhere slightly further out. That is a ratchet, and it predicts that a cuff’s slackening should be cumulative and roughly logarithmic in the number of cycles rather than sudden.

Which is what happens, and it is the reason a rib is knitted tighter than the fabric it borders. Tightness raises the contact force, which raises the friction, which narrows the range each cycle can move it through.

What a cuff’s specification should say #

Four things, of which the trade usually states two.

The loop length, which sets the recovery force and everything else about the loops. The structure, which sets the geometric extension. The fibre and its setting treatment, which decides whether the recovery survives washing. And the extension the cuff is expected to work across, because the force is a rising function and quoting it at one extension says nothing about the others.

The last is the one most often missing, and this ladder makes the omission visible: a recovery force quoted as a number is a point on a curve that rises by a factor of four across an ordinary working range.

What the picture cannot show #

A rib. Every drawing on this ladder is a plain jersey, because the model is a plain jersey’s, and the rib figure here is a structural diagram rather than a solved shape.

That is worth being plain about. The forces quoted come from a plain knit’s loop at the rib’s loop length, on the argument that a rib’s loops are the same object doing the same thing; the fabric-scale geometry around them is different and is not solved. A reader should take the force as an estimate for a rib and a computation for a jersey.

Where the number would go wrong #

Two places, and both push the same way.

A rib’s loops are not quite a jersey’s: at the fold, the loop is bent about an extra axis and stores more energy than the planar model gives it. And the wales’ contact with each other where they fold adds friction the model does not carry.

Both make a real rib pull back harder than the number here, so this is a lower bound and should be used as one. The direction is at least known, which is more than can be said for the folding.

Why elastane changes the argument rather than the fabric #

Most modern cuffs carry a few per cent of elastane, and it is worth saying what that does to everything above, because it does not replace the mechanism — it adds a second one in parallel.

An elastane filament is a rubber: it stores energy in a stretched polymer network and returns nearly all of it, over hundreds of per cent, with a force that rises steeply and does not depend on any geometry. Put alongside a knitted structure it dominates the recovery force by an order of magnitude and adds nothing to the extension available, because the geometry was already supplying that.

So an elastane cuff is a structure whose folding still supplies the extension and whose return is now a material property rather than a configurational one. That makes it stronger, more repeatable and immune to the ratchet above — and it makes it fail in a new way, because a rubber degrades with heat, chlorine and time while a loop’s geometry does not.

The point worth carrying is that the two mechanisms are separable, and the arithmetic in this rung is the one that survives when the elastane is gone. A hundred-year-old cotton rib still grips.

What was known before #

That a rib pulls in to about half its width, since ribs existed. That the mechanism is the loops folding to opposite faces, from the structural literature. That the recovery is not the fibre’s elasticity — this is understood in the trade, because everybody knows a cotton rib works.

What has not been available is the force. A cuff has been specified by its structure and its loop length and tested by being pulled, and the quantity connecting the two is the one this rung supplies — with the caveat that it supplies the loops’ half and not the folding’s.

The number, and what to do with it #

Collecting the arithmetic in one place: at a fifth of extension a rib of ordinary construction pulls back with about eight tenths of a newton per metre of fabric, rising to three and a half at a doubling, with the loops’ contribution computed and the folding’s not.

For a cuff of a hundred millimetres’ relaxed circumference opened to a hundred and fifty, that is around a tenth of a newton around the wrist — ten grams-force — which is the right order for something that grips without marking and is easily an order below the load that would leave a mark.

The design lever is the loop length, through the inverse cube, and it is the only one that moves the force without changing the structure. Doubling the recovery force means shortening the loop by a fifth, which also raises the contact force, the run resistance and the fabric’s weight, and lowers the extension available. Everything moves together, because everything is a function of one ratio.

Grip per gram, which is what a cuff is really buying #

The design lever is named above as the loop length through an inverse cube, with the caution that everything moves together. Writing down what else moves shows that the movements do not cancel, and the direction they leave is a useful one.

The recovery force per metre of fabric goes as the yarn’s bending rigidity over the cube of the loop length, and at the free bound a spun yarn’s rigidity is proportional to its count. A knit’s areal weight is its stitch density times its loop length times its count, and the stitch density goes as the inverse square of the loop, so the weight goes as the count over the loop length. Divide:

force ∝ tex ÷ ℓ³, weight ∝ tex ÷ ℓ, so force per unit weight ∝ 1 ÷ ℓ².

A tighter cuff grips harder per gram, as the inverse square of its loop length, and the only stop is what the machine can knit. That is not obvious from the cube alone, because the cube is paid for partly in weight and the question a garment maker asks is what the grip costs.

Written in the knitter’s own index it is sharper still. With the tightness factor being the root of the count over the loop length, substituting gives

force per metre ∝ K³ ÷ √tex, weight ∝ K√tex.

So at a fixed tightness factor, a finer yarn gives more grip and less weight at once — the force rises as the reciprocal root of the count and the weight falls as its root. There is no trade at all in that direction: halving the count multiplies the grip by 1.41 and divides the weight by the same.

That is a strong statement and it lands exactly on practice. Cuffs and welts are knitted in finer yarn and at a higher tightness factor than the body they border, which every knitter does and which is usually explained as making them firmer. The arithmetic says it is buying grip per gram by two independent routes, and that both of them run the same way.

Two limits stop it, and they are the reasons a cuff is not knitted in the finest yarn at the tightest setting available.

The machine. A tightness factor has a ceiling set by what the needles will take, and a finer yarn on a coarse gauge simply falls through — so the two levers are not independent of the machine’s own gauge.

And the yarn. A finer yarn is more irregular by an exponent and weaker, and a cuff is the part of a garment that is pulled hardest. The grip per gram rises without limit in the arithmetic and the yarn runs out first.

So the honest form of the design rule is: go as fine and as tight as the machine and the yarn allow, because within those limits nothing is being traded away. That is an unusual position in this collection, where almost every lever costs something, and it is worth flagging as the exception rather than letting it read as the norm.

Where the ladder goes next #

Into the consequences of the contact force rather than the recovery force. A pill needs fibre to be drawn out of a surface, and how readily depends on the grip — which is the same normal load multiplied by a coefficient, and which has been argued for two ladders without a number.