Cut registry › Jewellery cuts › Marquise
Jewellery cuts · Elongated brilliant
The navette, and the longest shape here.
The marquise is the most elongated outline in regular production. It is also, on our measurements, the best-performing of the three elongated brilliants, which is not what its reputation would suggest.
The short version
A marquise cut diamond is an elongated brilliant with two pointed ends and a curved outline between them, typically 58 facets.
It is also called a navette, meaning little boat, and it carries the longest length-to-width ratio of any shape in common use.
Definition
The marquise is a brilliant cut, so its facet families are the round brilliant’s: table, stars, bezels, upper and lower girdle facets, and pavilion mains. The outline is a pointed ellipse, and the two tips are where all of the shape’s particular problems live.
The name comes from eighteenth-century France. The cut is said to have been commissioned by Louis XV and named for the Marquise de Pompadour, its outline echoing her lips. The story is repeated everywhere and rarely sourced, so treat it as the trade’s founding anecdote rather than documented history. The alternative name, navette, is plainly descriptive and probably older in practical use.
Symmetry is the making of a marquise. Both points must sit on the long axis, and the two curved sides must mirror each other. Wings that do not match, or shoulders of different heights, show immediately because the eye has two identical halves to compare.
Proportions
Commonly cited ranges against our measured reference stone. The marquise is one of the few shapes where the traditional ratio really is close to the optical optimum, and we can show that rather than assert it.
| Measure | Commonly cited | Our reference stone | What it does |
|---|---|---|---|
| Length-to-width | 1.85 to 2.10; GIA submissions cluster 1.6 to 2.2 | 2.06 | Two to one is the classic proportion. |
| Depth | 58% to 62% | 62.0% | Shallow enough to spread, deep enough to hold the points. |
| Table | 53% to 63% | 58.7% | The flat top, as a share of the width. |
| Pavilion angle | not commonly published | 31.4° | Decides whether light returns or escapes. |
| Facets | 58 | 87 | Counted as facets that actually reach the surface. |
GIA issues no cut-quality grade for marquise diamonds. The report will carry symmetry and polish. On a shape with two points that must align on one axis, the symmetry grade is doing more work than usual.
The bow-tie
A bow-tie is a dark band lying across the waist of an elongated brilliant. It is a shadow rather than a fault: the facets beneath it draw their light from the part of the sky that your own head and shoulders are blocking as you look down at the stone.
That distinction decides what you can do about it. Light leakage is a failure of the pavilion angles, it is present under any light, and it shows on a light-return map. A bow-tie only appears when something is standing in the way of the light, which is why it darkens and lifts as you move your hand. A good deal of published advice calls the bow-tie a leak. It is not, and the two want different remedies.
Some central contrast is inherent to the shape, and we will not pretend otherwise. GIA describes the mechanism plainly: the band darkens as the difference between length and width grows, and as variation in the pavilion angle becomes more extreme. A well-cut stone reduces it. Nothing eliminates it.
What we can and cannot measure. Our engine illuminates from a narrow overhead cone and contains no observer, so by construction it cannot cast a head shadow, and our reference geometry for this shape carries no twist in the pavilion mains. We tested for a bow-tie twice, once under the standard face-up cone and once under full-hemisphere light with a modelled observer blocking the central 25°, and found none in either run while the round-brilliant control behaved exactly as it should. So we do not publish a bow-tie severity figure for this cut. We would rather leave the column empty than fill it with a number we cannot stand behind.
What we measured instead. The ratio question has a real answer for the marquise, and it is not the answer we expected. Stretching the marquise’s own facet plan across the range the trade uses, useful light rises from about 52% at a ratio of 1.6 to a maximum around 2.1, then falls away again. The best Global score in the sweep lands on 2.06, which is the traditional proportion. Unlike the oval, where ratio is nearly free, a marquise cut short of about 1.9 measurably gives up light. The old two-to-one rule has something behind it.
Light
Same census as every cut on this site: 3,200 rays face-up, 2,200 at 20° tilt, on exact facet geometry, measured against T57.
| Material | Useful light | Leak | Fire | Scintillation | Global | vs the standard |
|---|---|---|---|---|---|---|
| Diamond | 64.06% | 17.23% | 88.52 | 85.53 | 79.73 | −8.88 |
| Moissanite | 58.75% | 19.46% | 93.50 | 85.53 | 79.28 | −10.54 |
| Cubic zirconia | 66.20% | 20.11% | 92.97 | 85.53 | 79.82 | −11.95 |
| T57 standard, diamond | 92.28% | 1.88% | 78.85 | 71.82 | 88.61 | 0.00 |
Our reference marquise returns 64.06% of entering light usefully against the round brilliant’s 92.28%, leaks 17.23%, and scores 79.73 Global against 88.61. The gap of 8.88 points is the smallest of the three elongated brilliants, ahead of both the oval and the pear.
That is worth stating plainly, because the marquise is usually written about as a difficult, compromised shape. On one consistent protocol it out-returns the oval by a wide margin.
It carries 87 active facets for 85.53 on scintillation against the standard’s 71.82, and 88.52 on fire. The long outline gives it room for more facets than a round of the same weight, and it uses them.
Where the light is: useful return along the length
Five equal bands from point to point, each returning ray binned by where it leaves the crown. The belly returns 86.8% against 57.5% at the tips, a spread of 29.3 points, where the round-brilliant control reads 1.0. That is the flattest profile of the three elongated brilliants: the marquise loses less at its ends than the oval or the pear does, which is part of why it scores as well as it does.
Against the standard
Light return. 92.28% for the round against 64.06% for the marquise. The round still wins comfortably, as it does against every fancy shape here.
Sparkle character. 87 facets against 57, for 85.53 on scintillation against 71.82. A marquise throws noticeably more separate flashes than a round brilliant of the same weight.
Face-up size. The marquise spreads further per carat than almost anything, which is the main reason people choose it. As with the other elongated shapes we do not quote an Index figure for this, because our spread axis reads depth rather than outline.
Durability. Two points instead of one, and both need protecting. French tips, which replace the large bezel facet at each point with smaller facets to spread the stress, are the cutter’s answer. V-prongs or a bezel are the setter’s.
Price. Marquise diamonds are usually among the least expensive shapes per carat, and have been unfashionable long enough for that to be a real saving.
The baseline for all of this is the round brilliant itself. Its facet plan is set out in the anatomy of a round brilliant, and the measured standard we compare against is T57, the stone Tolkowsky published in 1919.
Materials
In moissanite the marquise returns 58.75% for 79.28 Global, with fire at 93.50 against 88.52 in diamond. Moissanite is doubly refractive, and on a long outline with many small facets that doubling is easier to see than on a compact shape.
In cubic zirconia it returns 66.20% for 79.82 Global. Zirconia is the softest of the three at Mohs 8.25, which matters more than usual on a shape with two exposed points.
Compare all three for this outline in the Index.
Trade-offs
Questions
The questions buyers actually ask, answered in full. Where the answer is ours to prove, the figure is measured on the Index; where it is an established gemological fact, it is stated as one.
A marquise cut diamond is an elongated brilliant with two pointed ends and a curved outline between them, typically carrying 58 facets. It is also known as a navette, French for little boat. Its facet families are the same as a round brilliant’s; only the outline differs.
The classic range is 1.85 to 2.10, with two to one the most preferred, and GIA lab submissions cluster between 1.6 and 2.2. Unusually among fancy shapes, this is not purely a matter of taste: sweeping the marquise’s own facet plan across that range on our engine, useful light peaks near 2.1 and falls away below about 1.9. The traditional proportion is close to the optical one.
Most show some. It is a shadow rather than a leak: the facets across the waist draw their light from the region of sky blocked by the viewer’s own head and shoulders, so the band lifts and darkens as the stone moves. GIA describes it as worsening with greater length-to-width difference and more extreme pavilion angle variation.
The trade commonly cites 58% to 62% depth and 53% to 63% table. Our reference marquise measures 62.0% depth and 58.7% table. Outside those ranges a marquise either passes light through the pavilion or buries weight below the table where it adds nothing face-up.
They are, once the points are protected. Three approaches are standard and can be combined: French tips, where the cutter replaces the large bezel facet at each point with smaller facets that spread the stress; V-prongs, which cup each tip; and a bezel, which encloses the outline entirely. An unprotected marquise tip in a plain claw setting is the arrangement that chips.
Yes, typically more so than any other common shape, because a two-to-one outline spreads the same weight across far more face-up area. We do not attach our Index spread figure to this claim, since that axis reads depth rather than outline. Compare millimetres.
Fifty-eight is the trade standard. Our reference marquise measures 87 facets that actually reach the surface, and that count is what our scintillation figure of 85.53 is calculated from.
Yes, generally noticeably so per carat at comparable quality. The shape retains more of the original rough than a round does, and it has been out of fashion for long enough that demand has not closed the gap.
No. GIA issues cut grades for round brilliants only. A marquise report carries symmetry and polish grades. On a shape whose two points must align on a single axis, the symmetry grade is the more informative of the two.
From the French court. The cut is said to have been commissioned by Louis XV and named for the Marquise de Pompadour, its outline echoing her lips. The story is universally repeated and rarely sourced, so it is best treated as the trade’s founding anecdote. The alternative name, navette, simply means little boat.
Next
See the whole set on the jewellery cuts index, or the measured house and historic cuts in the cut registry.
Method
Monte-Carlo on exact facet geometry: 3,200 rays face-up plus 2,200 at 20° tilt, cosine-weighted from a 7° near-vertical cone. Useful light is what exits through the crown within 64° of vertical. The same protocol runs on every cut on this site, so the figures are comparable to each other.
They measure our reference geometry for the shape at the proportions listed above, not an average of stones on the market. A marquise cut to different proportions will read differently. Treat them as a like-for-like comparison between shapes, not as a grade for any individual stone.
Engine estimates, ±2 points, not lab-certified grades. We do not publish the Index symmetry axis on these pages: it scores facet regularity within each pavilion family, which is meaningful for a round brilliant and misleading for a shape whose pavilion is deliberately irregular.
85.53 here against the standard’s 71.82. The term is 100(1−e−n/45) on the count of facets that actually reach the surface, so it is a pure function of facet count and nothing else.