A seam slips before it breaks

The seams ladder has two rungs and they are about the two things a seam is made of. What holds a thread in a seam is about the cloth: a sewn edge is a set of threads gripped by their neighbours, and how well they are gripped decides whether the seam opens. The stitch that weakens the seam is about the needle: every stitch is a hole, and a hole cuts threads.

Both are careful about what they cannot say. The first ends without a force, because there was none on this site. This rung supplies it, and the useful form turns out to be a length rather than a force.

The claim #

A seam allowance is a length that has to clear a crossover, and it is not a sewing convention.

A thread is held one crossing at a time: its resistance to being withdrawn rises with the length it is gripped over, while its own breaking load does not. The length at which the two are equal is the crossover, and a seam allowance is precisely a gripped length. So:

• allowance below the crossover — the threads slide out and the seam opens with the cloth undamaged, which is slippage;
• allowance above it — the thread or the fabric breaks first, which is rupture.

Two failure modes the trade names separately turn out to be one inequality, and which side of it a cloth is on is computable from its construction.

The table, and the boundary in it #

At an allowance of ten millimetres — the ordinary garment allowance — and a friction of 0.3:

Four break and four slip, and the boundary falls between a voile and a muslin — two ordinary cotton cloths that nobody would describe as differing in kind. A voile is fine and openly set at 24 ends per centimetre in 12 tex; a muslin is set the same way in 20 tex. The finer, weaker cloth breaks and the coarser, stronger one slips, which is the opposite of the intuition and follows directly from the arithmetic: a coarse yarn’s breaking load rises faster than the grip on it does.

It is also worth reading against what thread count is sold as buying. A densely set cloth does grip its threads harder, and that is one of the few things a high count genuinely delivers — though it delivers it as a shorter crossover rather than as anything a specification records.

That inversion is the practically useful part. A maker choosing between two cloths for a garment that will be stressed at its seams cannot read the answer off the cloth’s strength, its weight or its handle. The relevant comparison is between the allowance and the crossover, and the crossover is a ratio of two things that both rise with the yarn’s coarseness at different rates.

Why the allowance is where it is #

An allowance of about a centimetre is standard across the garment trade and has been for as long as anybody has written patterns down. This rung says what it is standard for.

It is not a comfortable margin. Half the cloths in the table are on the wrong side of it at that allowance, and the allowance that would put every one of them on the right side is 73.7 millimetres — set entirely by the openest cloth in the table, which frays for seven centimetres and could not be seamed conventionally at any sensible allowance at all.

So the trade’s rules make sense as a set rather than as a rule. A centimetre works for the cloths it works for, and everything else gets a different construction: a French seam, which encloses the raw edges and doubles the gripped length; a bound edge, which replaces the grip with a second fabric and is what a selvedge does at the loom; an overlocked edge, which adds a thread that holds the cut ends; or a resin finish, which raises μ and moves the crossover directly.

Each of those is a way of dealing with a cloth whose crossover is longer than its allowance, and the trade reaches for them on exactly the cloths this table puts on the slipping side.

Why the coarse cloth slips, in the terms of the expression #

The inversion between the voile and the muslin is the table’s most useful reading and it is worth taking apart, because the expression says which way each term pushes.

The crossover is a breaking load divided by a grip per unit length. The breaking load is a tenacity times a cross-section, so it goes as the yarn’s count. The grip per unit length is the number of crossings in a millimetre times the force at each, so it goes as the other system’s sett times the sine of the weave angle. So

crossover ∝ count ÷ (sett × sin θ)

and the coarse yarn wins the numerator outright. A muslin’s 20 tex against a voile’s 12 is a factor of 1.67 in breaking load with the setts identical, and nothing in the denominator moves by anything like that.

The denominator does move, and it moves the right way, which is why the effect is smaller than the count ratio. A coarser yarn at the same sett covers more of the cloth, so its threads have to climb further over one another and its weave angle is larger — 24.6° for the muslin against 18.1° for the voile, which is a factor of 1.34 in the sine. So the grip does rise with the yarn’s coarseness; it simply does not rise as fast as the load the yarn can carry.

That is the whole of the inversion and it generalises past these two rows. Anything that makes a yarn stronger without making the cloth grip it proportionately more pushes a construction towards slipping — a higher tenacity fibre, a folded yarn, a better spinning system. A cloth made of a better yarn is more likely to slip at its seams, which is not what anybody expects of an improvement and follows from one term appearing in only one of the two quantities being compared.

The crossover moves with the load, and the direction is unhelpful #

One quantity in the crossover is not a property of the cloth at all, and it is worth separating from the rest.

The grip at a crossing is μ times twice the thread’s tension times the sine of the weave angle, so the whole grip term is proportional to a tension the calculation states rather than derives — half a newton, here, as everywhere in this group. The breaking load does not contain it. So the crossover is inversely proportional to whatever tension is holding the cloth’s other system, and the comparison the table makes is a comparison at one assumed state of loading.

A seam under load is not in that state. Pulling a seam apart pulls the cloth in the direction across the stitch line, and a cloth pulled one way grips harder in the other — the crimp moves, the weave angle in the unpulled system deepens, and the tension in it rises with the seam’s own load. So the crossover shortens as the seam is loaded, and a cloth’s failure mode is not fixed before the test begins.

The direction is the awkward one for a specification. A lightly loaded seam is nearer the slipping side and a heavily loaded one nearer the breaking side, so a test run at a low force measures a different failure mode from the same seam under a real load, and the mode reported depends on the force chosen. That is not an argument against the test; it is an argument for reading its stated force as part of the answer, in the same way the allowance already is.

What a slippage specification is measuring #

Seam slippage is a specified property with its own test: a seam is loaded and the opening beside the stitching is measured at a stated force, or the force to produce a stated opening is reported.

Read through this rung, that test is measuring how far below the crossover the allowance is. It is not a property of the cloth alone, and it is not a property of the seam alone; it is a comparison between a length the cloth sets and a length the maker chooses.

Two consequences follow that a specification stated as a cloth property obscures.

The same cloth passes or fails depending on the allowance. Doubling the allowance doubles the grip and can move a cloth from one side to the other — a muslin at ten millimetres slips and at fifteen breaks. So a slippage figure without an allowance beside it is a dimension without a state in a different vocabulary.

And a finish moves it without touching anything a specification records. μ is the only parameter in the crossover, and a softening finish can raise the crossover by a factor of nearly three across the reported range. A cloth that passed a slippage test in the greige can fail it after finishing, with the same construction, the same yarn and the same allowance.

The three consequences of one coefficient, met for the third time #

This is the third rung in this group to arrive at the same trade-off from a different direction, which is worth naming rather than repeating.

Friction appears once in the crossover expression and nowhere else. So the same coefficient decides how far a cut edge frays, how firmly a tuft is held, and whether a seam slips — and there is no way to move it for one and not the others.

A finisher softening a cloth is making its edges fray further, its seams slip sooner and its pile less firmly bound, all by the same factor, simultaneously. The trade treats those as three faults with three remedies and they are one parameter with three consequences.

What can be traded is the geometry: a wider allowance, a denser sett, a shorter float, a different seam construction. None of those touches μ and every one of them moves the comparison. So the practical statement is that a cloth’s slippage behaviour is fixed at the loom and at the finish, and everything a garment maker can do about it is done by changing the length rather than the grip.

What was counted, and how #

Each cloth’s crossover is the contact ladder’s: a Peirce state at the quoted construction, a weave angle from it, a contact force of twice the thread tension times its sine, a per-crossing grip of μ times that, and a crossing count of one per end spacing. The breaking load is the yarn’s tenacity times its translation efficiency times the twist obliquity.

The grip at a stated allowance is that per-millimetre rate times the allowance, and the verdict is a comparison.

One assertion guards the table and it is the one worth having: an ordinary seam allowance must divide the table rather than settling it either way. A computation that put every cloth on one side would be consistent with a sign error, a units error or a friction coefficient off by an order of magnitude, and it would look entirely reasonable. Requiring both outcomes to appear is a weak check and it is the check that a plausible-looking wrong answer would fail.

The thread tension is stated at half a newton, as everywhere in this group, and every grip is proportional to it. The allowance that would save the whole table is reported as what it is — a number set by one cloth — rather than as a recommendation.

Where the model stops #

The stitching is not in it. A seam’s threads are also cut and displaced by the needle, which is the other rung of this ladder and is a strength reduction the grip calculation knows nothing about. A real seam fails at whichever of three things gives first: the grip, the fabric, or the sewing thread.

The load is taken as a pull-out. A seam under load is not simply withdrawing threads; it is opening, which rotates the threads at the stitch line and changes the geometry as it goes. The static comparison here is the first term of that and not the whole of it.

The float correction is missing, exactly as it is on the rung below: this counts every intersection as a grip, which is right for a plain weave and overstates a satin’s hold by something like the ratio of their interlacing rates. That is why satin-faced cloths slip at seams far worse than this table would suggest, and it is recorded as owed rather than approximated.

And the allowance is treated as fully engaged. A real seam’s grip is not uniform over the allowance — it is largest at the stitch line and falls away — so the effective length is shorter than the geometric one and every cloth is nearer the slipping side than the table says.

The generalisation #

The shape is the same characteristic-length argument the rung below sets out, and what this rung adds is what happens when a standard meets it.

A conventional dimension that has survived a long time is usually a value that clears the characteristic length for the common cases and not for the rest. A centimetre of seam allowance, a lap length in a joint, a splice length in a rope, an anchorage length for reinforcement: each is a number the trade converged on, each works for the middle of its range, and each has a family of special constructions attached to it for the cases it does not cover. The special constructions are the evidence that the standard is a threshold rather than a margin.

The second point is about specifications. A property that is a comparison between two lengths cannot be specified as a property of one of them, and doing it anyway produces exactly the confusion around seam slippage: a cloth is said to slip, when what is true is that a cloth of that construction, finished that way, seamed at that allowance, slips. Every one of those four is load-bearing and only the first is on the label.

Who found it, and when #

Seam slippage as a distinct failure mode is old garment-trade knowledge and is codified in the seam-strength standards; the distinction between a seam that opens and a seam that breaks is the whole reason those tests exist.

Yarn pull-out as a measured quantity belongs to the fabric-mechanics literature of the 1960s, and the observation that pull-out force rises with engaged length is in every account of it. The connection to seam slippage is also made there — a slipping seam is a pull-out at the stitch line.

What this site adds is the crossover as a number computed from the construction rather than measured, and the observation that it puts an ordinary allowance in the middle of the range of ordinary cloths rather than comfortably above it. That reading makes the trade’s family of special seam constructions look like a rational response to a threshold rather than a set of separate traditions, which is a more flattering account of the trade than the trade usually gives itself.

Where the ladder goes next #

The float correction is the next thing this ladder owes, and it is the same one the rung below owes: a satin’s grip is not a count of its intersections, and the cloths that slip worst in practice are exactly the ones the model handles worst.

Sideways, the same comparison at a much shorter grip is what holds a tuft in, and at a raw edge it is how far a cloth frays — three rungs, one inequality, read at three lengths.

Further out is the seam as a mechanism rather than as a pull-out: a seam under load opens, and the opening rotates the threads and changes their wrap angles as it goes. That is a large-displacement problem and it needs the elastica this whole group keeps arriving at the edge of.