Everything here is a number the trade already uses. Where it is a published standard I say so. Where it is a rule of thumb I say that too, because the difference matters and almost nobody marks it.
It gets invoked constantly in fashion, nearly always as decoration on an argument that did not need it. There is one spot where it genuinely holds up, and it is worth knowing which, if only so you can be sceptical about the rest.
| landmark | fraction of height | nearest φ value | apart by |
|---|---|---|---|
| natural waist | 0.620 | 1/φ = 0.618 | 0.3% |
| mid thigh | 0.390 | 1/φ² = 0.382 | 2.1% |
| knee | 0.270 | — | — |
| shoulder | 0.818 | — | — |
On a body of 160 cm the golden point sits 98.9 cm from the floor. The natural waist sits at 99.5 cm. Six millimetres apart.
That is why a high waistband reads as proportioned and why the myth has such a long life. It is also the full extent of the claim. The waist is in the right place; your shoulders, your knees and your hemline are not governed by anything mystical.
Stylists divide a look one third to two thirds rather than down the middle, because an even split shortens the figure.
| division | at 160 cm | what sits there |
|---|---|---|
| a third from the floor | 53.3 cm | upper calf, which is a midi hem |
| two thirds | 106.7 cm | just above the natural waist |
| golden, 1/φ | 98.9 cm | the natural waist itself |
| golden, 1/φ² | 61.1 cm | mid thigh, which is a short hem |
Both survive because they land within a few centimetres of each other on most bodies, so whichever you invoke afterwards will appear to explain why something looked right. Thirds is the useful one, because you can do it in your head in a changing room.
European sizes step by four centimetres in bust, waist and hip. That part is a standard and it is reliable.
| size | bust | waist | hip | step |
|---|---|---|---|---|
| EU 34 | 82 | 64 | 88 | — |
| EU 36 | 86 | 68 | 92 | +4 |
| EU 38 | 90 | 72 | 96 | +4 |
| EU 40 | 94 | 76 | 100 | +4 |
| EU 42 | 98 | 80 | 104 | +4 |
| EU 44 | 102 | 84 | 108 | +4 |
Length is not part of that standard, and this is the whole reason this website exists. A range graded 34 to 44 covers twenty centimetres of body circumference and, very often, zero centimetres of length. Everyone in the range gets the same hem.
Which is less stupid than it sounds. A size 44 woman is not taller than a size 34 woman, so grading length with size would hand the longest skirt to the widest customer, who might be 158 cm. There is a longer version of this argument on the Why it is not done page, and it is more sympathetic than you would expect.
The most misunderstood number in online shopping. A chart of body measurements and a chart of garment measurements can differ by fifteen centimetres and both be perfectly honest.
| cut | ease at the bust | ease at the hip | what it does |
|---|---|---|---|
| negative | −2 to 0 cm | −4 to 0 cm | only possible with stretch |
| fitted | 1 to 4 cm | 0 to 3 cm | skims the body, almost no spare |
| close | 5 to 7 cm | 4 to 6 cm | follows the body, restricts a little |
| standard | 8 to 11 cm | 7 to 9 cm | the default for a woven |
| relaxed | 12 to 17 cm | 10 to 16 cm | hangs away from the body |
| oversized | 18 cm and up | 17 cm and up | the volume is the design |
These are trade rules of thumb, not a standard. They shift by house and by category. They are the same bands the Fitting page uses when it recommends you a size, so if you disagree with a recommendation, this table is where the disagreement lives.
Elastane does not make a garment run large. It lets a smaller flat measurement fit a larger body, because the cloth supplies movement that ease would otherwise have to provide. Those are different things and the difference is the reason I am a 38 in a woven and a 36 in a knit without having changed shape.
| elastane | usable give in wear | can you size down | what it will not fix |
|---|---|---|---|
| 0 to 1% | almost none | no, size for your widest measurement | anything |
| 2 to 3% | 3 to 5 cm at the hip | if you are between sizes, yes | length, shoulders, rise |
| 4 to 6% | 6 to 10 cm | usually, and you should if you want it fitted | length, shoulders, rise |
| 7% and up | 10 cm and more | yes, and it will relax further with wear | length, shoulders, rise |
Look at the last column. Stretch is a circumference solution and nothing else. You can take a 36 in a stretch trouser and still receive a 36 inseam and a 36 rise on a body that needs a 38 in both, and no amount of elastane will lengthen a leg.
The figures in that table are trade rules of thumb, not measurements. The same percentage of elastane behaves differently in different yarns: in laboratory tests on denim, recovery after stretching ran from about 57 per cent for plain cotton to about 90 per cent for cotton spun round elastane and polyester.
A woven cloth barely extends along either set of threads. Cut it at 45° to both and it extends five to ten times as much and falls in a spiral.
This is geometry rather than chemistry. Forty-five degrees is the only angle at which both thread systems are equally oblique, so the cloth can deform as a lattice instead of pulling on either set of threads directly. Madeleine Vionnet built an entire house on it in the 1920s and it still looks like a trick.
| direction of cut | extension in a plain weave | behaviour |
|---|---|---|
| warp, along the grain | 2 to 5% | holds its length, hangs straight |
| weft, across the grain | 3 to 6% | a little softer |
| bias, at 45° | 15 to 30% | clings, elongates, drapes in a spiral |
The Cusick test drapes a circle of cloth over a smaller disc and measures the shadow it casts. The drape coefficient is the ratio of that shadow to the flat circle. Lower means more fluid.
| cloth | drape coefficient | reads as |
|---|---|---|
| silk chiffon, fine viscose | 20 to 35% | fluid, follows the body |
| crepe, soft twill | 35 to 55% | falls close, some volume |
| cotton poplin, gabardine | 55 to 75% | holds its own shape |
| denim, canvas, heavy linen | 75 to 95% | stands away from the body |
It is a real laboratory number with a real test method behind it, and it would answer the question every product photograph fails to answer, which is whether a dress will follow your shape or replace it.
Body-type geometry. Pear, hourglass, apple. Useful shorthand for the relationship between three measurements and nothing more. There is no equation underneath, and nothing on this site uses one.
Universal proportion systems. Le Corbusier’s Modulor turns up in fashion writing with some regularity. It is architecture, and it was built on a 183 cm man.
The eight-head figure. A real adult is about seven and a half heads tall. Fashion illustration stretches the figure to eight or nine because it flatters the line of a garment, and it has been doing so for long enough that a drawing convention is now widely mistaken for anatomy. It is also, if you think about it, the same problem as the 178 cm model, a century older.
The fractions in the first table — the waist at 0.620 of height, the knee at 0.270, the shoulder at 0.818 — are the foundation of every reading this site produces, so they deserve a note of their own.
Body landmarks are defined in ISO 7250 and ISO 8559, and European size labelling in EN 13402. The proportional figures used here come from national anthropometric survey data; the large European sources are SizeUK, SizeGERMANY and CAESAR.
These are population averages, and your own proportions will differ from them, which is exactly why the site asks for your inseam separately instead of estimating it from your height. The landmarks above the crotch scale with your torso and the ones below scale with your leg, and getting those two apart from each other is most of what the engine does.