Which clue in a photograph is the most useful?

Internal features. Needles, veils, clouds and colour banding are the clues that separate families, and whether they are angular or rounded is often visible at ordinary phone magnification. Colour is the clue most people lead with and the one most easily corrupted by lighting.

Why does the report name families rather than individual stones?

Because a family is what visible evidence supports. Stones from different families can look identical face up in a photograph, and what separates them are readings taken with the stone in hand. "Quartz family, amethyst-type purple" is a claim a picture can carry; a species stated as a certain fact is not.

Can a stone be read while it is still in its setting?

Yes, with less certainty. A setting removes the girdle, the pavilion, the depth of the stone and often the true body colour from evidence, and metal reflects into the stone. A reading of a set stone should say which clues were lost and lower its confidence accordingly.

How do I see the inside of a stone in a photograph?

Put the light behind it. Stand the stone at the edge of a window with the window behind it, or rest it on paper lit from below. Needles, veils, colour banding and rounded voids appear in a backlit frame and stay invisible in a frame lit from the front.

What can a photograph never tell me?

Anything measured: refractive index, specific gravity, weight, dimensions, and how the stone behaves under polarised light or a spectroscope. A photograph also gives no price and no certification. Those need the stone in hand and a qualified person.

Gemstone identification guide: reading a stone from what you can see

This is a reference for the clues a stone shows to the eye and to a camera: what each one is, what it narrows down, and what it cannot settle. It is the same order of reasoning that every reading from the tool follows, written out so you can check the reasoning yourself.

One rule runs through all of it. Visible evidence supports families and look-alike groups, not names. Everything below is written to that limit.

Body colour, and the way the colour is distributed

Body colour is the colour of the stone itself rather than the colour of its highlights, and there are two separate observations in it: the hue and tone, and how evenly they sit.

Even colour from edge to edge is one piece of evidence. Colour that changes across straight-edged zones is another, and it is the stronger of the two: zoning follows the directions in which a crystalline material grew, so its presence narrows the field more than a hue does. Parallel colour banding is a third pattern, and a curved smooth swirl is a fourth that points away from a crystalline host altogether.

What colour cannot do is separate families on its own. Purple, green, blue and red each occur across several families, and a tone read from a photograph is only as trustworthy as the light it was shot in. A warm lamp, a coloured tabletop, a metal setting reflecting into the stone, or a phone deciding the white balance for you all shift it. Photograph in daylight against a plain neutral card, and treat colour as a filter that removes candidates rather than one that chooses them. The post on photographing a stone for identification covers the lighting in detail.

Transparency

Three states are worth distinguishing: the background reads clearly through the stone; light passes but nothing is legible through it; no light passes at all. Judge this separately from colour, because a dark stone can be fully transparent and a pale one opaque.

Transparency only reads honestly with the light behind the stone. In a front-lit frame a transparent stone sitting on paper often reads as opaque, because everything you see is surface reflection. Through a setting it frequently cannot be judged at all: the metal closes the back of the stone and there is no path for light to take.

Internal features, and how they differ from bubbles and swirls

Internal features are the most useful thing a photograph can carry, and they come in two families of their own.

Features grown into a crystalline host are angular and directional. Straight needles with flat reflective faces, often lying along two or three distinct directions. Feathery veils. Clouds of fine particles. Bands of colour following growth directions. When these appear, they narrow the field usefully, and they argue strongly that the host grew rather than set.

Features left in a poured material are rounded and smooth. Spherical or teardrop voids where gas was trapped. Smooth curved streaks running through the body where liquid moved while it set. Angular versus rounded, straight versus curved, is the distinction to hold onto; it is often visible at phone magnification and it carries more weight than anything else you can see without instruments. The post on the indicators separating a stone from moulded glass weighs each of them.

What internal features cannot tell you is which mineral they are, and therefore they narrow the host to a family rather than naming it. Their absence tells you very little either: at ordinary magnification anything under about a millimetre simply does not resolve, so "no inclusions visible" is a fact about the photograph, not about the stone.

Cut style and the state of the facet junctions

Cut style is a human decision rather than a property of the material, so it does not identify anything by itself. It still carries information. A faceted crown with a table, a cabochon dome, a polished rounded form and a rough uncut lump are each read differently, and the proportions of a faceted stone are readable in profile.

The junctions matter more than the style. Where two faces meet, a crisp line reads one way and a narrow rounded band reads another: moulded pieces tend to have soft junctions because the mould left them soft. The trap is wear. A cut stone worn daily for decades also has softened junctions, and no photograph separates a soft junction from a worn one. Take the state of the junctions as one indicator among several, never as a conclusion.

Surface luster

Luster is the character of a reflection rather than its brightness: bright and glassy, soft and slightly waxy, oily, silky, or dull. It is a genuine clue and a weak one. It reads only when the light is even, the surface has been wiped, and the reflection is not blown out by a flash or by direct sunlight, and even then it is a comparative judgement with nothing in the frame to compare against. Use luster as a tie-breaker between two readings that otherwise fit equally, and read a soft luster as a question rather than an answer.

Fracture and cleavage, where they are already visible

Where a stone is already chipped, the shape of the damage is evidence, and it is worth describing carefully.

A break that curves like the inside of a shell, with fine concentric ridges, reads as one kind of surface. A break that steps along flat planes, leaving straight edges at consistent angles, reads as another: flat planes suggest the material parts preferentially along directions, which is a property some families have and others do not. Chips at the girdle of a set stone and abrasions along an exposed edge fall in the same category of visible damage.

This is a clue to observe, never one to produce. Nothing a break surface could add justifies damaging the stone, and no reading from a photograph will ever ask you to make one.

Scale, and why a frame without it loses information

A photograph has no inherent size. A five millimetre stone and a fifty millimetre one fill a frame identically, and that costs more than it seems: the plausibility of a family, the meaning of a visible inclusion and the magnification the picture actually achieved all depend on it. Put a ruler edge, a coin or a known ring band in the frame, or state the size if you know it. A jeweller's scale for a loose stone adds another dimension the picture cannot.

The families a careful reading will name

A visual reading works with a deliberately short list of families and look-alike groups. Each entry below is a group of materials that share an appearance, not an identification of any stone in it.

Quartz family, including rock-crystal type colourless, amethyst-type purple, and types carrying needle inclusions.
Corundum family, including sapphire-type blue and other fancy-colour types.
Beryl family, including aquamarine-type pale blue-green and emerald-type green.
Topaz family, which overlaps the quartz family closely in a photograph.
Opal family, recognisable by play of colour across the surface rather than by a body hue.
Garnet family, a wide group spanning red through orange to green.
Tourmaline family, often strongly zoned along the length of the stone.
Diamond-type colourless, a look-alike group rather than a single material, because several colourless materials sit in it and a photograph does not separate them.
Glass or moulded imitation, indicated by rounded voids, curved swirls and soft junctions.
Plastic or resin imitation, usually giving away a softer luster and a lighter feel in the hand.
Undetermined, which is the correct answer whenever the frame does not carry enough to choose.

"Undetermined" at the top of a list is a result, not a failure. A poorly lit frame of a set stone honestly supports nothing else, and inventing a family to fill the slot would make the whole report worthless. Run your own photograph and the preview will tell you which of these situations you are in before anything is unlocked.

Loose or set, and what a setting hides

A loose stone photographed face up gives the outline, the crown, the table, the girdle and the colour. Add a profile frame and you also get the depth, the pavilion facets and the girdle line along its whole length.

A setting takes most of that back. A closed back removes transparency entirely. Bezels and cups cover the girdle, which is exactly where a mould seam or a chip would show. The pavilion and the depth of the stone go with them, so the proportions cannot be judged. Metal reflects into the stone from below and from the sides, lightening the body colour on one flank and adding highlights that are not the stone's. Prongs sit over the outline and can hide a damaged edge.

None of that makes a set stone unreadable. It makes a set stone a weaker case, and the honest response is a shorter list of clues and a lower confidence, plus an explicit note of what was lost. A side-on frame recovers some of it; a jeweller unsetting the stone recovers all of it, and that is only worth doing for its own reasons.

Where a photograph stops being evidence

A picture carries appearance. The readings that actually separate the remaining candidates are measurements, and they need the stone in hand and a person qualified to take them:

Refractive index, read on a refractometer, and birefringence with it.
Specific gravity, from weight in air against weight in water.
Behaviour under polarised light, and the optic figure that follows from it.
Absorption seen through a spectroscope.
Magnified examination of the interior at powers a phone cannot reach.
Weight and dimensions, taken on a balance and with a gauge.

Two further things sit outside a photograph entirely. No price comes from a picture, and no certificate does either. This tool does not price stones and does not appraise them, and it is not a laboratory report; those are three different documents and the difference is set out in the post on identification, laboratory reports and appraisals.

The safe steps between a photograph and an appointment are short: retake the frame in daylight, add a second angle, look through a 10x loupe, try a different light source, weigh a loose stone on a jeweller's scale. Then take the short list and your one open question to a qualified gemmologist. Nothing that scrapes, heats or soaks the stone belongs in that sequence.

Using this guide on your own stone

Work down the clues in order, write what you see rather than what you hope, and note which candidates survive. Then upload the photograph and compare your reading against the report's, clue by clue. The order in which the clues are read is set out in the post on identifying a gemstone from a picture, and the preview is free on every run, so a disagreement costs nothing to investigate.

Questions

Which clue in a photograph is the most useful?: Internal features. Needles, veils, clouds and colour banding are the clues that separate families, and whether they are angular or rounded is often visible at ordinary phone magnification. Colour is the clue most people lead with and the one most easily corrupted by lighting.
Why does the report name families rather than individual stones?: Because a family is what visible evidence supports. Stones from different families can look identical face up in a photograph, and what separates them are readings taken with the stone in hand. "Quartz family, amethyst-type purple" is a claim a picture can carry; a species stated as a certain fact is not.
Can a stone be read while it is still in its setting?: Yes, with less certainty. A setting removes the girdle, the pavilion, the depth of the stone and often the true body colour from evidence, and metal reflects into the stone. A reading of a set stone should say which clues were lost and lower its confidence accordingly.
How do I see the inside of a stone in a photograph?: Put the light behind it. Stand the stone at the edge of a window with the window behind it, or rest it on paper lit from below. Needles, veils, colour banding and rounded voids appear in a backlit frame and stay invisible in a frame lit from the front.
What can a photograph never tell me?: Anything measured: refractive index, specific gravity, weight, dimensions, and how the stone behaves under polarised light or a spectroscope. A photograph also gives no price and no certification. Those need the stone in hand and a qualified person.