A larger skirt can cost nothing for its nap
Worth reading first: A nap is paid for by the taper of the pattern · A grade charges by the length and a cutter pays by the panel · Raising moves the surface onto the hairs.
A pattern for a velveteen skirt prints two fabric requirements for every size: one without nap and a larger one with nap. The larger figure exists because raising moves the surface onto the hairs, and hairs that lean one way show a different shade from each end, so every piece has to be laid with its top pointing the same way along the cloth.
A nap is paid for by the taper of the pattern found what that rule costs. A rectangle costs nothing laid one way. A tapered panel costs extra cloth that depends only on its taper: in lanes and exactly half that in the densest one-way lay there is, where is the narrow width over the wide. And on a real cloth width, it found, the price comes in whole panel lengths, not in percentages.
It ended on a prediction it could not test from where it stood: across the sizes of a graded pattern, the with-nap surcharge should jump in whole lengths rather than grow. This essay grades a skirt and lays every size. The surcharge does jump, and it jumps the wrong way: on the commonest wide cloth it is a full panel length for the small sizes and nothing for the large ones.
A skirt graded by the gore
A six-gore skirt is six tapered panels, narrow at the waist and wide at the hem. Grading it means making every gore wider at both ends as the size goes up, and the grading used here is the plainest one that is still realistic. Each gore’s top is the waist plus two centimetres of ease, divided by six. Its hem is eighteen centimetres wider than its top. Its length is 75 centimetres at every size.
That keeps the flare per gore fixed and lets the hem grow with the waist, from 170 centimetres round at a 60-centimetre waist to 220 at 110. It is the only assumption in this essay that is a choice rather than a consequence, and a real publisher’s grading will differ in detail, often adding length to the larger sizes too. What follows is about the shape of the result, and the shape survives any grading that widens the gores steadily.
The taper of each gore eases as the size goes up, since a fixed eighteen centimetres of flare is a smaller share of a wider panel. The ratio of top to hem goes from 0.36 at the smallest waist to 0.51 at the largest.
The endless price falls smoothly
On an unbounded cloth, the earlier essay’s price is a smooth function of that ratio. A staggered one-way lay, columns of gores with every other column shifted half a length, needs more cloth than a two-way lay in which alternate gores are turned end for end.
For this skirt that is 23.3 per cent at a 60-centimetre waist, falling steadily to 16.3 per cent at 110. A gentler taper wastes less when it cannot be turned, so the larger sizes should pay a little less, in proportion, for the nap. If a pattern envelope printed its with-nap surcharge as a percentage of its without-nap figure, that is roughly the column it would print: about a fifth, easing slowly as the sizes rise.
A home sewer does not cut an endless marker, though. A home sewer cuts one skirt, from a cloth with two edges, and the surcharge on one skirt is the difference between two staircases.
One skirt on a cloth with edges
A 150-centimetre cloth is the commonest wide width for a napped cotton. On it, the two-way lay sets gores side by side, alternately turned, each pair making a parallelogram with nothing between them. A band 75 centimetres long holds as many gores as fit across the width that way: of them need centimetres, where and are the gore’s top and hem.
The one-way lay has two arrangements. In lanes, every gore needs its full hem width , so fit across. Staggered, columns sit apart and alternate columns start half a length down, so one more column often fits, at the cost of a half-length’s offset. The cutter takes whichever is shorter.
For the 65-centimetre waist the gores are 11.2 centimetres at the top and 29.2 at the hem. Six of them fit across the width when alternately turned, and only five fit when they all point one way, in lanes or staggered. Six gores is one band two ways and two bands one way. The nap doubles the cloth, from 75 centimetres to 150.
That is a hundred per cent surcharge, not twenty. It is the same mechanism the earlier essay found at six gores on a fixed panel, and the reason is that one band either holds the whole skirt or it does not. The fraction of a band left empty is no use to anybody cutting one garment.
The large sizes have nothing left to lose
Go up to a 90-centimetre waist. The gores are 15.3 centimetres at the top and 33.3 at the hem, and now only five fit across the width even when they are turned. The two-way lay needs a second band for its sixth gore, so it takes 150 centimetres. The one-way lay needed two bands already, and still does, so it also takes 150.
The surcharge is nothing. Every size from 90 to 110 centimetres pays nothing for its nap on this cloth.
The large skirt is not cheaper to cut. Its gores cover 1.17 square metres of a cloth 2.25 square metres in area, so a little under half the length it pays for is cut away as waste. The nap costs nothing because the two-way lay has already lost the advantage the nap would have taken away. That advantage was always one thing: fitting the last gore into the first band by turning its neighbours. Once the gores are too wide for that, turning them buys nothing and forbidding the turn costs nothing.
So the surcharge column for this skirt on this cloth reads 37.5 centimetres at a 60-centimetre waist, 75 at every waist from 65 to 85, and nothing from 90 up. A rule that charges a fixed share of the plain requirement would print the opposite pattern, a surcharge that grows as the plain figure grows. That rule is the one a grade charges by the length and a cutter pays by the panel found running through fault grading as well: a charge proportional to a length, applied to a cost that comes in whole pieces.
The smallest size pays half
The 60-centimetre waist is an edge case worth a sentence. Its gores are narrow enough, 10.3 centimetres at the top and 28.3 at the hem, that six staggered columns fit across, alternately offset by half a length. The sixth gore then sits in the second half of a column that started half a length down, and the skirt takes 112.5 centimetres one way instead of 150. The surcharge is half a panel length.
That is the stagger doing exactly what the earlier essay said it would: recovering half of what lanes lose. On a finite width it recovers it in half-lengths, and only at a size where one more column happens to fit.
The width decides which sizes pay
None of this is a property of the skirt alone. It is a property of the skirt and the cloth width together, since both lays count gores across the width.
The grid has a clear structure. The paid cells form a band running diagonally from the small sizes on a 130-centimetre cloth to the middle sizes on 160. Inside the band, the two-way lay fits all six gores across and the one-way lay does not. Above it the cloth is too narrow for either to fit six; below it, wide enough for both. A wider cloth lets bigger gores fit six across two ways, so the band moves to larger sizes as the cloth widens.
Between 115 and 125 centimetres no size pays anything, because no size fits six gores across in either lay. A 115-centimetre cloth, the old 45-inch width, is the one most napped cottons and corduroys were once woven at. On it a velveteen six-gore skirt costs exactly the same cloth whether the velveteen has a nap or not. It also costs 150 centimetres of a narrower cloth at every size, which for the 65-centimetre waist is 53 per cent of the cloth used against 81 per cent on a 150-centimetre width laid two ways. The nap is free there because the narrow width was already wasteful, not because the narrow width is a bargain.
At the narrowest widths, 100 to 110 centimetres, the band reappears at the far corner. There the contest is over three gores a band rather than six. On a 100-centimetre cloth the large sizes still fit three gores across when they are turned, so six take two bands, but only two fit in a one-way lane. The stagger fits three columns and pays for it with half a length of offset, 187.5 centimetres against 150; at a 110-centimetre waist even the third column is lost and the nap costs a whole length. The band has wrapped round: a different count of gores per band, the same mechanism.
The smooth price is never sampled
Put the one-skirt surcharge beside the endless-marker price, both as shares of the two-way length, and the relation between them is almost none.
The smooth curve is what a cutter’s marker converges to over many pieces. The staircase is what one garment pays. For a single skirt, the smooth price is never the price at any size: every size pays either nothing, half or all of a two-way lay’s length again. The earlier essay found that the staircase settles toward the smooth price as pieces are added to one marker, which raises the question of how many skirts that takes.
A batch pays about a fifth
A cutting room does not cut one skirt. It lays a batch of one size, or several sizes together, in a single long marker, and the gaps one skirt leaves at the end of a band are filled by the next skirt’s gores.
The two sizes trade places as the batch grows. One small skirt pays a hundred per cent, two pay fifty, and by five the share is twenty, moving about the endless price thereafter. One large skirt pays nothing and two pay nothing, then three pay a quarter, since three skirts are eighteen gores and the one-way lay needs a fifth band where the two-way lay manages with four. By twelve skirts, the small size pays 25 per cent and the large one 20.
So the same pattern has two honest with-nap figures, and they disagree. A cutting room buying for a production run should add about a fifth whatever the size, and a little less for the larger ones. A home sewer cutting one skirt should add a whole panel length at the small sizes and nothing at the large ones. An envelope printing one surcharge per size can be right for one of those customers, and not for both.
What the two columns on an envelope could be checked against
The arithmetic here makes a definite prediction about published patterns. For a gored skirt, the with-nap figure on an envelope should equal the without-nap one at some sizes and exceed it by about a panel length at others, with the switch where the gores stop fitting six across turned. The switch should move to larger sizes for a wider cloth.
A straight skirt or a shift should show no difference at any size, since a rectangle costs nothing to lay one way. An envelope whose surcharge grows by a steady share as the sizes rise is printing an average over batches. That is defensible, but it is not what one garment costs.
Nothing about this is special to a nap. A repeat has to fit the panel, a one-way print restricts translation exactly as a nap does, and corduroy and velvet are laid one way for the same reason as a brushed cotton. Every one-way cloth pays this staircase and not the smooth curve.
The staircase also has relatives outside the cutting of garments, and they share its character. A paperhanger pays for a pattern’s repeat in whole drops, and a bay window wastes less than a wall because short drops can share a repeat that a full wall’s drops cannot. A half-drop wastes less, cut in the right order, for the same reason the stagger here recovers half a length. And a bolt’s usable width, which a cutter values by what its faults leave, feeds straight into the grid above: a bolt that loses five centimetres of its width to a selvedge fault moves a 150-centimetre cloth to the 145 row of the grid, where the 85-centimetre waist stops paying for its nap and the 60-centimetre waist pays a whole length instead of half. In every one of these the waste is counted in whole units of something the job is made of, and an average over many jobs is a poor guide to any one of them.
How the yardages were reached
Every skirt is six trapezoids, each with its top at (waist + 2)/6, its hem eighteen centimetres wider and its length 75, at waists from 60 to 110 centimetres in steps of five. Each lay is counted across the width exactly. Two ways, gores alternately turned fit to a band. One way in lanes, . Staggered, columns, with the length grown in half-lengths until the columns hold every gore. The one-way length is the shorter of lanes and stagger. The endless price is the difference-body bound for a trapezoid, the one the earlier essay proved the stagger attains.
Required of the arithmetic, and not shown: that every surcharge is a whole number of half lengths and never negative; that on 150-centimetre cloth a 70-centimetre waist pays a whole length and a 100-centimetre waist nothing; that rectangular panels pay nothing at any size; and that a batch of twelve comes within ten points of the endless price at both a small waist and a large one.
What a marker of gores alone leaves out
A skirt is not only its gores. A waistband, facings and pocket bags go into the same marker, and a cutter puts them in the gaps a one-way lay leaves. In the 100-centimetre two-way lay, the second band holds one gore and is mostly empty. A waistband 112 centimetres long and a few wide would fit in it, and then the one-way lay, also two bands, would have to fit it too. Small pieces close the gap in some sizes and not others, and a full pattern would have to be laid piece by piece to say where.
Gores can sometimes be laid across the grain. A gore rotated a quarter turn keeps its nap consistent with the others if they all turn together. That changes which dimension the width limits, and so moves the band on the grid. It does not change the mechanism. And the grading is one grading. A pattern that also lengthens its larger sizes adds length to every band, which scales both lays alike and leaves each surcharge a whole number of the new length.
What none of this can say is how much the shade difference between two panels actually matters to a wearer, which is the reason for the rule in the first place. Only a float can be raised, and only a raised cloth has a lean to show; how much its two directions differ to an eye is a question about light that the lean’s own geometry prices only as a covered length.
Who worked out which part
With-nap yardages on pattern envelopes and one-way markers in cutting rooms are long-standing practice. Grading a gored pattern by widening its gores and counting pieces across a cloth width are ordinary pattern-room arithmetic. The packing bound for a convex shape under translation is Minkowski’s, through the difference body. What is added here is the size run itself: that a six-gore skirt on 150-centimetre cloth pays a whole length at small sizes and nothing at large ones; that the paid sizes form a diagonal band across widths and sizes, absent entirely between 115 and 125 centimetres; that the endless price is never the price of one garment at any size; and that a batch of one size settles between a fifth and a quarter by a dozen skirts while one garment pays all or nothing.
Still open: whether the small pieces close the gap or open it
Everything above lays the gores alone. A real pattern’s waistband, facings and pockets are the pieces a cutter uses to fill gaps, and they are often allowed to lie either way because no one sees them beside a main panel. So a one-way lay that loses a band to its sixth gore may win part of it back by packing the small pieces into the second band’s spare width. The two-way lay, which had the first band full, has to start a second band for them anyway.
The prediction is that the small pieces shrink the surcharge most in exactly the sizes that pay it, since those are the sizes where the one-way lay’s second band is mostly empty, and leave the free sizes free. Testing it needs a whole pattern laid piece by piece at each size on two widths, which is a small exhaustive search rather than a formula. It would turn the grid above from a statement about gores into a statement about garments, and it would say whether an envelope’s with-nap column can be predicted from its pieces alone.
What links here
Computed from the collection rather than written here: the essays that point at this one.
Shares its objects with
Essays naming at least two of the same things, that neither author linked.
- Two lots of one cloth drift apart down a drop — both name cutting, specification, waste
- A fault map is worth most where the grade is worst — both name marker efficiency, specification
- A jersey has no float for a wire to catch — both name nap, specification
- A missing end is a fault the length of the piece — both name marker efficiency, specification
- A skew steps at every seam and a bow at none — both name cutting, specification
- The bias cut and the selvedge — both name marker efficiency, specification
Named objects
A flat tag is an object no other essay names yet.