Setting and geometry

Thread count is not quality

It counts threads. It says nothing about how much of the cloth they cover, it can be inflated by counting plies, and it ranks fabrics in nearly the wrong order. The quantity it is mistaken for is cover factor, and that one is computable.

Two sheets, both labelled two hundred threads to the inch. One is dense and opaque; the other is thin enough to read through. Both labels are accurate.

The same thread count, twiceTwo cloths with identical thread counts and different yarn. The count is the same number in both; the fraction of the surface the threads actually occupy is not, and that fraction is what thread count is usually taken to mean.100 ends and 100 picks per inch in bothfine yarncover 78%coarser yarncover 97%cover from diameter and spacing, circular sections
Fig. 1 The same thread count in two yarns. The number of threads is identical; the fraction of the surface they occupy is not, and that fraction is what the count is usually taken to mean.

Thread count counts threads. It is a perfectly well-defined number and it does not measure the thing everyone uses it to measure.

What it is

Ends per inch plus picks per inch, counted in the finished cloth. A fabric with 110 ends and 90 picks is a 200-count fabric.

That is all. It is a count of how many threads cross a given distance, and it makes no reference to how thick they are, how much space they occupy, what weave holds them, or what they are made of.

Why it fails

Three separate failures, and they compound.

It ignores yarn thickness. Two hundred threads of a very fine yarn cover far less of a surface than two hundred of a coarser one. Since fineness is easy to achieve and coarseness is cheap, a high count can indicate a finer, more open, less durable cloth rather than a better one.

It can be inflated by ply. A yarn made of two finer yarns twisted together is one thread by any structural measure — it occupies one space in the reed and behaves as one thread in the weave. Counting its components doubles the number without changing anything. A “1000-count” sheet is almost always a 250-count cloth in four-ply yarn, and the physical fabric is coarser than a genuine 400-count in singles.

It ignores the weave. The weave decides how densely threads can be set at all: a satin in a given yarn takes half as many threads again as a plain weave before jamming. So the same count is a loose cloth in one structure and an impossible one in another, and the number alone does not say which.

The weave, which is doing most of the work

The third failure deserves more than a paragraph, because it is the one this site is best placed to quantify.

The maximum thread count a cloth can carry is set by its weave. Every interlacing makes a thread bend and a bend takes room, so a structure with few interlacings can be crowded much closer than one with many.

What an interlacing costsEach weave's interlacing count beside the closest it can be set in the same yarn. The two run opposite ways, because a thread that changes face often has to bend often and a bend takes room.weaveinterlacingsdensest settingplain1.0020 per inch2/2 basket0.5027 per inch2/1 twill0.6724 per inch2/2 twill0.5027 per inch3/1 twill0.5027 per inch5-end satin0.4029 per inch8-end satin0.2532 per incha geometric jamming model, circular sections, yarn diameter 1/40 inch
Fig. 2 The densest setting each weave allows, in one yarn. Plain weave jams at about twenty threads to the inch; an eight-end satin at about thirty-two. The same yarn, sixty per cent more threads, and nothing but the weave has changed.

So a two-hundred count means quite different things in different structures. In a satin it may be a comfortably loose cloth with room to drape. In a plain weave of the same yarn it is above the jamming point and cannot be woven at all. A count without a weave is not a description of a fabric; it is a number that happens to be true of one.

This is also why the crisp-versus-fluid distinction people notice between percale and sateen sheets has nothing to do with their counts. Percale is plain weave and sateen is a satin, and the difference in handle is the interlacing count doing what it always does.

What the number should be

The quantity thread count is taken for is cover: the fraction of the cloth’s area occupied by thread rather than by hole.

Warp cover is ends per inch times warp diameter — how much of each inch is filled by warp. Weft cover likewise. Cloth cover adds them and subtracts the overlap, since the places where both cross are counted twice:

K=K1+K2K1K2.K = K_1 + K_2 - K_1 K_2.

That is computable from two numbers a mill already knows, it is dimensionless, and it means what people think thread count means. A cover of one is a cloth with no visible gaps; a cover of a half is an open fabric a lamp shows through.

The one quantity it needs that a thread count does not is yarn diameter — and diameter is the thing a thread count conspicuously omits.

Diameter, and the count systems that hide it

Yarn is not usually sold by diameter. It is sold by count, which is a linear density, and the systems for expressing it are a genuine mess.

In the indirect systems a bigger number means a finer yarn, because the number is length per unit mass: cotton count is hanks of 840 yards per pound, worsted count is hanks of 560 yards per pound, and linen count is leas of 300 yards per pound. In the direct systems a bigger number means a coarser yarn, because the number is mass per unit length: denier is grams per 9,000 metres and tex is grams per 1,000 metres.

So “a 40” means a fine cotton yarn and a coarse one in denier, and the same physical yarn has half a dozen names. Diameter follows from count and fibre density with a further assumption about how tightly the fibres are packed, which varies with twist.

None of that is difficult, and all of it is a barrier to the sort of comparison a shopper is trying to make. Thread count survives partly because it is the only number in the chain that needs no explanation.

What the same count actually gives

Two cloths at two hundred threads to the inch, in yarns differing by a factor of two in diameter, differ by nearly a factor of two in cover — one nearly solid and the other visibly open.

The same thread count, twiceTwo cloths with identical thread counts and different yarn. The count is the same number in both; the fraction of the surface the threads actually occupy is not, and that fraction is what thread count is usually taken to mean.200 ends and 200 picks per inch in bothfine yarn, high countcover 75%coarser yarn, same countcover 89%cover from diameter and spacing, circular sections
Fig. 3 Two hundred ends and two hundred picks in both. The threads are drawn to scale, and the cover differs by more than a factor of two. Nothing about the count distinguishes these fabrics.

The ordering is not subtle and it is not a marginal effect at the edges of the measurement. It is the dominant term, and thread count omits it entirely.

Crimp takes some of the length too

One further complication, and it is the one that makes thread count hard to reason about even when everything else is held fixed.

A thread in cloth is longer than the cloth, because it goes over and under. So a given weight of yarn produces less cloth than its length suggests, and the shortfall depends on the weave — a plain weave crimps its threads several times as much as a satin.

A warp end in section — plainOne warp thread drawn through the cloth, with the weft threads it crosses shown end-on. The thread is longer than the cloth it spans, and the excess is the crimp — measured here from the drawn path rather than quoted beside it.the cloth this thread spansplainwarp crimp 22.3%8 interlacings per repeatso the crimp shown exceeds a real cloth'sthread thickness exaggerated for legibility3 face changes
Fig. 4 Where the length goes. A warp end in plain weave deviates at every crossing, so a good deal of its length is spent travelling up and down rather than along. A satin spends almost none, which is part of why a satin is heavy for its count.

That means two cloths of equal thread count in the same yarn can contain measurably different amounts of material, because one has put more of its yarn into crimp. Weight per unit area catches this and thread count does not, which is one more reason to prefer it.

It also produces a fact that surprises people: pull a fabric lengthways and its thread count changes. The ends per inch rise as the cloth narrows, and the picks per inch fall as it lengthens, without a single thread having been added or removed. A count measured on a fabric under tension is not the count of the relaxed cloth.

Where the claim came from

Thread count is not a marketing invention. It is a legitimate mill measurement with a real use, and the problem is where it travelled to.

In a mill, comparing two cloths in the same yarn and the same weave, thread count is exactly the right measure: it says how much yarn is in the fabric and how tightly it is set, and both quantities matter. The number was never intended to compare a percale to a sateen or a fine cotton to a coarse one, and inside its intended scope it does not fail.

What happened is that it escaped onto packaging, where none of the controlling variables are stated. A sheet’s label gives the count and not the yarn count, not the ply, and often not the weave — so the one number a shopper is given is the one that means least without the others.

The industry’s own bodies have said so. Ply-inflated counts have been the subject of complaints, guidance and litigation, and the standards position is straightforwardly that plies should not be counted as separate threads. That position is not universally observed.

Counting, which is harder than it sounds

Even taken on its own terms, the measurement is not quite as objective as it looks.

A thread count is made with a magnifier over a marked square, and several decisions have to be made before a number comes out. Is the fabric relaxed or under tension? Is it measured before or after washing, given that most cloth shrinks by a few per cent and some by ten? Is a two-ply yarn one thread or two? In a fabric where the picks are inserted in pairs, is a pair one pick or two?

Each of those is a defensible choice, and different answers give different numbers for the same physical cloth. Standards exist and specify the answers — relaxed, conditioned to a stated humidity, plies not counted separately — and the counts printed on packaging do not always follow them.

The basketThe basket on point paper, a filled square meaning the warp is on the face. Its longest float, its interlacing count and the number of separable cloths it describes were all counted from the matrix that drew it.basketlongest float 216 interlacings per repeat1 separable clothrepeat 4 × 4generated, then counted4×4
Fig. 5 A basket weave, where the ambiguity is at its worst. Threads work in pairs, so a counter has to decide whether a pair is one thread or two — and the answer changes the count by a factor of two without changing the cloth at all.

The moral is not that the measurement is worthless. It is that a number quoted without its protocol is not a measurement, and a number quoted without the variables it needs to hold fixed is not a comparison.

What to look at instead

Four things, in rough order of usefulness.

The weave. Percale is plain weave and sateen is a satin; they behave differently in every respect, and knowing which is worth more than any count. The weave decides the direction of most properties.

The yarn count and ply. A 300-count in singles is a finer, denser cloth than a 600-count in two-ply of the same fibre.

Weight per unit area. Grams per square metre is unglamorous and hard to inflate, and it is a direct measure of how much material is present.

Cover, if it is available. It rarely is, which is a pity, because it is the number the whole conversation is trying to have.

The claims that fail the same way

Thread count is the standing example and it is not alone. Three others fail by the same mechanism — a number defined for a narrow comparison, quoted where its controlling variables vary.

“Breathable”. What is usually meant is air permeability or moisture-vapour transmission, both of which are measurable and neither of which is what “breathable” is used to assert. Air permeability follows from cover and thickness, so an open weave in polyester out-breathes a dense one in cotton — and the fibre is what the claim usually names.

“Satin is strong because it is smooth”. The causation is backwards and the conclusion is half right for the wrong reason. A satin’s few interlacings make it less abrasion-resistant and more tear-resistant, because loosely held threads can bunch at a tear tip and share the load. Two opposite effects from one cause, and the claim picks neither.

“Natural fibres are cooler”. Thermal behaviour in cloth is dominated by trapped air, which is a structural property. A dense linen and an open polyester knit differ in the direction opposite to the claim.

The general remedy is the one this site applies everywhere: work out which quantity the claim is reaching for, compute it, and see whether it says what was asserted. The refutation has to be a computation rather than a contrary assertion, or nothing has been established.

Where the model stops

The cover calculation on this page uses circular yarn sections of a fixed diameter — Peirce’s 1937 assumption, and the simplest thing that works.

Real yarn is not round in cloth. It is flattened where it is gripped and rounder where it floats, so its effective width varies along its length. The racetrack model treats the section as a rectangle with semicircular ends and the elliptical model as an ellipse, and both give higher cover for the same yarn because a flattened thread covers more. The numbers differ by ten or twenty per cent between models and the ordering is unchanged, which is the general pattern in this subject.

So the cover figures here are structural rather than exact, and they are quoted with the model named. The same caution applies to setting calculations, where the models disagree by a similar margin and for the same reason.

A general habit

The pattern this essay is an instance of is worth naming because it recurs.

A quantity is defined for a narrow, well-controlled comparison. It works. It escapes into a context where the controlled variables vary freely. It stops working, but it keeps being quoted, because it is the only number anybody has.

Thread count is the textile case. Every field has one, and the remedy is always the same: find out what the number would have to hold fixed to mean what it is being used to mean, and check whether those things are fixed.

Here they are not. The weave varies, the yarn varies, the ply varies, and the count is reported without any of them — which is why two accurate labels can describe fabrics that are not comparable at all.

What a number has to hold fixed

The general lesson is worth extracting, because it applies well beyond textiles.

A measurement compares. To compare, everything except the thing being compared has to be held fixed — and a number is only as good as the list of things it assumes are constant.

Thread count assumes the yarn is the same, the ply is the same, and the weave is the same. Inside a mill those assumptions hold, and the number works. On a packet none of them holds, and the number does not.

The weaves everything else is built fromPlain, twill and satin on point paper. Reading across, the floats lengthen and the interlacings fall — and every property a weaver cares about follows from that one trade.plainfloat 1 · 8 interlacingsfirmness 1.002/2 twillfloat 2 · 16 interlacingsfirmness 0.505-end satinfloat 4 · 20 interlacingsfirmness 0.40a filled square means the warp is on the facecounted from the matrices, not from the pictures
Fig. 6 Three of the things thread count assumes away. The same count in these three structures gives three different cloths, and the number reports none of the difference.

The remedy is always the same and it is not to abandon the measurement. It is to ask what the number would have to hold fixed to mean what it is being used to mean, and then to check whether those things are stated. Where they are not, the number is a fact about the fabric that happens not to answer the question.

Cover does better because it needs fewer assumptions — yarn diameter and spacing, both of which it uses directly rather than assuming away. Weight per unit area does better still, because it assumes almost nothing at all.

Where the ladder goes next

The geometry underneath cover is how close threads can be set, where the models are compared rather than assumed.

The structural variable that dominates everything is the float, and the quantity behind setting limits is the interlacing count.

What the pictures here cannot show. The panels on this page draw threads as rectangles of a computed width, which is a model rather than a photograph. Hairiness, twist, flattening and finish all change how much of a surface a real yarn covers, and none of them is drawn.