A polariscope answers one question: does light travel through this stone at one speed, or two? Singly refractive material (garnet, spinel, glass, diamond) has one; doubly refractive material (corundum, beryl, tourmaline, quartz) splits light into two rays travelling at different speeds. A third answer is possible — the stone is not a single crystal at all, but a mass of microscopic ones, like chalcedony or jadeite.

That answer is cheap, fast, and it is a fork rather than a filter. In Loupewise's own engine, a polariscope reading alone takes 132 gem species to a median of 75 candidates — 23 in the best case, 75 in the worst — and every one of the 132 species has the data for it. In plain terms: it roughly halves the field. The three outcomes are not evenly sized, which is why the median sits where it does — 75 species are doubly refractive, 34 singly refractive, 23 aggregates.

It never finishes the job on its own, and it is not the strongest thing you can buy. Across all 132 species, a polariscope reading identifies exactly zero of them outright, and a refractometer narrows far harder (median 18). What the polariscope gives you is a decision no other cheap instrument makes, in seconds, with nothing to calibrate — and it keeps working on stones a refractometer cannot read at all: over-the-limit material, cabochons, rough. Run it early because it costs you almost nothing; do not expect it to carry the identification.

These figures simulate a perfect reading on every species. A real polariscope call on a strained garnet is not this clean, so read them as an upper bound.

What a Polariscope Actually Is

Two polarising filters and a light source. The lower filter (the polarizer) sits over the lamp; the upper one (the analyzer) is the one you look through and the one that rotates. Turn the analyzer until the field goes as dark as it will go — the two filters are now crossed — and put the stone in between.

There is no optics to align, no liquid, nothing to calibrate, and nothing to spill. Gemological models add a rotating stage, a swing-in condensing sphere for interference figures, and a stand that holds the working distance steady, but the instrument in its essential form is those two filters. This is the reason people build their own, and the reason it is worth owning before a refractometer.

A polariscope, or two polarising filters. Gemological units are purpose-built and hold the geometry for you. A pair of filters over any steady, diffuse light source does the same job with more fiddling.

A diffuse, reasonably bright light. Even illumination matters more than power. A hot spot in the field is easy to mistake for a reaction in the stone.

Tweezers or a stone holder. You will be turning the stone into several different orientations, not just spinning it on one face.

A small dish of water. Optional, and the fix for one specific failure described below — a well-cut, high-RI stone that refuses to transmit light at all.

Taking the Reading

Cross the filters

With nothing on the stage, rotate the analyzer until the field is at its darkest. That is the crossed position, and it is your baseline. If the field never goes properly dark, the filters are not truly crossed or the light is leaking in around them — fix that before you read a stone, because every outcome below is defined against a dark field.

While the field is empty, look at it. The cover glass over the polarizer is supposed to be unstrained; if it is not, it shows its own cross-hatch under crossed filters, and you will read the instrument's window as anomalous double refraction in every stone you put on it. A diffuser under the polarizer is doing the other half of this job — an uneven field with a bright patch in it gets read as a reaction in the stone.

Check that light is getting through the stone

Before you interpret anything, confirm the stone transmits. A stone that passes no light reads dark no matter what it is made of, and that darkness tells you nothing about its optics. Opaque material is outside this instrument's reach entirely, and a reading taken through the thin edge of a semi-translucent stone is not worth much either.

Rotate the stone through a full turn

Slowly, and watch the whole stone rather than one spot. You are looking for whether the stone goes dark, goes light, or alternates — and if it alternates, whether the entire stone extinguishes at once or only parts of it do. Speed hides the difference; a full turn should take a few seconds, not one.

Repeat in at least three directions

Turn the stone onto a different face and rotate again. Then a third. A doubly refractive stone viewed straight down an optic axis stays dark through a complete rotation and is, in that one direction, indistinguishable from a singly refractive one. No amount of spinning recovers it — only a new viewing direction does. Three is a sensible working minimum drawn from teaching practice, not a figure any examining body mandates.

Run the confirmation

Whatever you think you saw, confirm it with the analyzer flip described below. This is the step that separates a reading from a guess, and it takes about two seconds.

The Four Outcomes

There are four, not three. Most short explanations give three and omit the fourth, which is the one that causes almost every wrong polariscope call.

Dark all the way round

The stone stays dark through a full 360° rotation. Reading: singly refractive. The textbook version of this is cleaner than the bench version, though — the teaching sources call a perfectly dark singly refractive reaction “seldom encountered in practice”, so a stone that is nearly but not quite dark has not necessarily done anything unusual.

Typical: garnet, spinel, glass, diamond, fluorite — but see “when dark means nothing” below, because three quite different things produce this.

Four blinks per turn

The whole stone goes light and dark four times in a complete rotation, extinguishing every 90°. Reading: doubly refractive.

Typical: corundum, beryl, tourmaline, quartz, topaz, peridot.

Light all the way round

The stone never goes dark at any point in the rotation. Reading: an aggregate — microcrystalline or cryptocrystalline material.

Typical: chalcedony, agate, jadeite, nephrite. This is not the doubly refractive reaction; a doubly refractive stone blinks.

Anomalous double refraction

The stone appears to blink, but the extinction is patchy or cross-hatched and the field never goes fully bright. Reading: still singly refractive, under internal strain.

Typical: synthetic spinel, almandine garnet, diamond, strained glass. This is the outcome that gets misread as double refraction.

The first three are straightforward. The fourth is not, and the whole of the rest of this guide is about telling it apart from the second.

The Confirmation Step: Flip the Analyzer

Light and dark is not the test. The confirmation is.

Turn the stone to its lightest position. Then rotate the analyzer 90°, from crossed to parallel, and watch what the stone does.

It jumps much brighter

Singly refractive, showing strain. What you were watching was anomalous double refraction.

It stays about the same

Genuinely doubly refractive — or an aggregate, which also does not change.

The reason it works is worth understanding, because it tells you where the test stops. Strain birefringence is tiny — small enough that the light emerging from a strained singly refractive stone is still, essentially, polarised the way it went in. Crossed filters block it and the stone looks dark; uncross them and it all comes through. The stone jumps.

A genuinely doubly refractive stone is different. Over a few millimetres of corundum the two rays fall tens of wavelengths out of step, and by the time you average that across the visible spectrum the emerging light is scrambled. Roughly half of it gets through whatever the analyzer is doing. Crossed or parallel, the stone looks much the same — and that lack of change is the signature.

If the difference is hard to judge, put an opaque card with a hole cut in it over or under the stone. Comparing the stone against a black surround rather than against a bright field makes a small brightness change much easier to see.

What the flip does not do. It separates strained singly refractive material from everything else. It does not separate a true aggregate from a single crystal that merely scatters light — both emerge depolarised, and both give “no change.” That case is dealt with below, and the fix is to look at the stone rather than at the instrument.

Where It Goes Wrong (1): Strain Read as Double Refraction

This is the most common wrong polariscope call, and it is expensive: an isotropic stone carried forward as anisotropic eliminates the correct answer at the very first test, and everything after it is spent in the wrong half of the database.

The signature to recognise is that the extinction is not clean. Anderson described anomalous double refraction as “mottled or grid-like” rather than clear-cut across the field; a fuller description is patchy or cross-hatched, resembling a coarse woven texture or irregular fragmented brushes, with some part of the stone in extinction during most of the rotation. A stone showing anomalous double refraction is never completely light.

That follows from what strain is. Strain varies across the stone in both strength and direction, so at any given rotation some patches happen to be aligned with the polarizer and go dark while their neighbours do not. Turning the stone moves which patches are dark; it never aligns them all. True double refraction has one continuous orientation through the whole stone, so the whole stone extinguishes at once, four times a turn.

Do not key your answer on how neat the extinction looked. A real doubly refractive stone does not always give a clean whole-field blink either — gemologists testing genuine rubies report a dark brush sweeping across the stone rather than a clean light-to-dark switch, and that is a strong indication of true double refraction, not of strain. Patchy extinction is your cue to run the flip. It is not your answer.

The cross-hatched pattern sometimes called “tabby extinction” is often quoted as diagnostic of synthetic spinel. Treat that as a strong hint and not as an identification: strained glass and plastic show cross-hatch too, and diamond's plastic-deformation strain has its own cross-hatched signature. Narrow it with refractive index, not with the pattern alone.

Two strained materials have signatures specific enough to name. Glass often shows a heavy black cross in the centre of the stone and no interference colours at all — a reaction the sources treat as typical of glass, or of another strained singly refractive transparent material. Amber, other natural resins and plastics go the other way: patchy interference colours mixed with greys and blacks, visible from any direction inside the stone. Neither is an identification on its own, and both are worth recognising rather than filing under “patchy”.

And if the stone is simply too dark to resolve any pattern — a deeply coloured synthetic spinel, for instance — the honest reading is “too dark to resolve,” followed by magnification. That is a better answer than a guess.

Where It Goes Wrong (2): When Dark Means Nothing

“It stayed dark, so it is singly refractive” is right often enough to be dangerous. Dark is one observation with three completely different causes, and only one of them is about the stone's optics.

You are looking down an optic axis

Along an optic axis the two rays travel at the same speed, so there is nothing to split and nothing to extinguish. The stone is dark through a full turn and is optically indistinguishable from singly refractive material in that direction. Uniaxial stones have one such direction and biaxial stones have two, and no amount of lateral rotation escapes either — only a new viewing direction does. This is the reason for the three-directions rule above.

The stone is returning the light before it ever gets in

This is the trap the sources call the most serious, and it is invisible if you do not know to look for it. A well-cut brilliant with a high refractive index — diamond, zircon — placed table-down is designed to return light entering the crown straight back out through the crown. Placed in a polariscope that way, it may transmit nothing except through the culet. It reads dark, and the darkness is about the cut, not the crystal.

The fix is to turn it so light actually passes through: roughly at right angles to the bezel facets is the easiest direction. If reflections still get in the way, immerse the stone in a small dish of water and read it there.

The stone is opaque

A floor tile in a polariscope stays dark, and that does not make the tile singly refractive. It means light cannot get through it. The polariscope has nothing to say about opaque material, and its verdict on semi-translucent material that transmits only at the edges should be treated as questionable.

Which way up? It depends on the species, and the rule reverses. Ruby, sapphire and emerald are cut to show their best colour, which puts the optic axis at right angles to the table — so looking straight down through the table of a ruby is looking down its optic axis, the worst polariscope direction available. Most other uniaxial stones are cut to save weight, with the optic axis lying along the length of the crystal and therefore in the plane of the girdle — for those, the table-to-culet direction is the direction of maximum double refraction and the best one to read. Learn the pair, not half of it.

A stone examined in three or more directions, rotated fully in each, transparent enough to pass light cleanly, and dark throughout, is singly refractive — and that call can be accepted without much hesitation. A doubly refractive reaction deserves more caution than a singly refractive one, not less.

Where It Goes Wrong (3): False Aggregates

Two mirror-image errors live on the same observation.

The first is calling a light-throughout stone doubly refractive. It is not: doubly refractive stones blink four times a turn. Light all the way round is the aggregate reaction, and reading it as double refraction skips the entire chalcedony and jadeite branch.

The second is the opposite — calling an aggregate on a single crystal that is merely scattering. A heavily included, badly fractured, twinned, or frosted-backed stone diffuses light on its way through and appears light regardless of position. A natural ruby has been misread this way. Repeated twinning in corundum can do it too, with alternating sets of internal plates never simultaneously in extinction.

In an aggregate, light passes through many randomly oriented microscopic crystals and emerges depolarised, so some of it always reaches the analyzer and the stone can never go dark. That is why the reaction is diagnostic. But scattering depolarises light by an entirely different route and produces exactly the same appearance. Same reading, different physics — and as noted above, the analyzer flip cannot separate them, because both give “no change.”

The discriminator is not an instrument at all. Look at the stone. Is the back polished or frosted? Is the light path clean, or is it crowded with fractures and inclusions? Get clean transparent material and a polished back in the light path, and then the reading means something.

One singly refractive exception worth knowing: the green jade-like grossular garnet legitimately reads light-throughout despite being singly refractive. Reporting the observation faithfully and then running the flip is the right procedure even when the raw reading looks like an aggregate.

What the Polariscope Cannot Settle

Dr Bill Hanneman puts the governing rule about as plainly as it can be put: do not conclude that a stone is singly refractive, doubly refractive, or an aggregate on the basis of a polariscope observation alone.

The instrument is a splitter. It divides the field cheaply and early, and it hands you a smaller problem — a median of 75 candidates from 132, in our own engine, with a perfect reading. It does not hand you a name, and it does not get you close to one. Confirm with something independent:

A dichroscope. If you can see genuine dichroism, the stone is doubly refracting and therefore not glass and not a cubic mineral. That is a direct, independent confirmation of the polariscope's second outcome. Be careful of the near-misses: a dichroscope always shows two windows — the calcite in it makes them — so two windows alone means nothing, and an elongated singly refractive stone can look pleochroic simply because the light path through its ends is longer than through its sides.
A refractometer. Two shadow edges that move as you rotate the stone confirm double refraction and measure it. One steady edge points the other way. This is also how you separate the cross-hatch identities that the pattern alone cannot.
Specific gravity. Independent of optics entirely, which is exactly what you want when an optical reading is ambiguous.

Read together, those tests either agree or they tell you that one of them was mistaken. That is the whole method.

Loupewise takes your optic character reading — singly refractive, doubly refractive, aggregate, or uncertain — and immediately shows which of 132 gem species remain consistent with it, plus which test will narrow the field fastest from there.

Open Loupewise

Studying for the FGA or GIA practical? Loupewise drills instrument-based identification — optic character, refractometer, SG and UV readings — and runs timed mock exams built around the real exam formats.

Practice exams →

Where This Comes From

The four-outcome table, the analyzer flip, and the failure modes above are drawn from published gemological literature rather than assembled from general knowledge, because the failure modes are exactly the material that gets garbled in retelling.

B. W. Anderson, Gem Testing, 6th ed., Heywood, 1958 — the extinction mechanics, anomalous double refraction in garnet, diamond and synthetic spinel, the dichroscope cross-check, and the cutting orientation of tourmaline.
The Gemology Project wiki, Polariscope — the four outcomes and the analyzer-flip confirmation.
Gem-identification course material circulating as “Identification by Polariscope” — the total-internal-reflection case, the water immersion remedy, the contrast shield, the three-directions practice, and the cutting-orientation pair. Widely reproduced teaching material of unattributed origin; treated here as consensus practice rather than as any institution's official position.
Practising gemologists on the record at gemologyonline.com — the moving-brush appearance of genuinely doubly refractive rubies, the disputed diagnostic value of tabby extinction, and Dr Bill Hanneman's rule against concluding from the polariscope alone.

Where those sources disagree — most sharply on how far cross-hatched extinction narrows an identification — this guide says so rather than picking a side.

Frequently Asked Questions

What does a polariscope measure?

Optic character: whether a stone is singly refractive, doubly refractive, or an aggregate of microscopic crystals. It does not measure a value the way a refractometer or a scale does — it sorts material into three categories. In Loupewise's engine that single sort takes 132 species down to a median of 75, in seconds, for the price of two pieces of film.

What it does not give you is optic sign, or whether a doubly refractive stone is uniaxial or biaxial. Those come from an interference figure, which needs a conoscope, and a retardation plate for the sign — separate glass. Worth knowing what that upgrade buys: adding the sign takes a doubly refractive stone from 75 candidates to a median of 19. Almost all of the narrowing people attribute to "the polariscope" is really coming from those extra instruments.

Can I make my own polariscope?

The working parts are two polarising filters and a light source, so yes, and people do. What a purpose-built gemological instrument adds is convenience and repeatability: a stage that holds the stone at a fixed working distance, an analyzer that rotates smoothly through a known angle, and a condensing sphere for interference figures. If you build your own, the two things worth getting right are even diffuse illumination and a genuinely dark crossed position — every outcome above is defined against that baseline.

My stone blinks, but the extinction looks patchy. Is it doubly refractive?

Unknown until you run the analyzer flip. Patchy or cross-hatched extinction that never lets the whole stone go bright is the classic appearance of strain in singly refractive material — but genuinely doubly refractive stones can also show a brush sweeping across rather than a clean whole-field switch. Turn the stone to its lightest position and rotate the analyzer 90°: a marked jump in brightness means singly refractive with strain; little change means genuinely doubly refractive.

Why does my diamond or zircon stay dark in every position?

Most likely because it is table-down. A well-cut, high-refractive-index brilliant returns light entering the crown back out through the crown, so almost nothing reaches your eye except through the culet. Turn it so the light path runs roughly at right angles to the bezel facets, or immerse it in water. Diamond also commonly shows strain birefringence when you do get light through it, which is a separate thing to be ready for.

The stone never goes dark. Is it chalcedony?

Maybe, and check the stone before you commit. Light-throughout is the genuine aggregate reaction, but a single crystal with a frosted or unpolished back, or one crowded with fractures and inclusions, scatters light and produces exactly the same appearance. The analyzer flip will not tell them apart — both give “no change.” Inspect the light path instead: clean, transparent, polished material first, then trust the reading.

How many directions do I really need to test?

Three is the working minimum, and it comes from teaching practice rather than from any examining body's published requirement. The reason for it is the optic axis: a doubly refractive stone viewed straight down one stays dark through a complete rotation, so a single orientation can never rule double refraction out. If your first two directions disagree, keep going until you understand why.

Is a polariscope worth buying before a refractometer?

Not on narrowing power. In our engine a refractive index reading leaves a median of 22 candidates from 132; a polariscope reading leaves 75. The refractometer wins that comparison decisively, and it is worth saying plainly because the opposite is easy to assume from the price difference.

Buy it for what it does that the refractometer cannot: it costs almost nothing, it takes seconds, there is nothing to calibrate and no contact liquid to run dry, and it still works on stones that defeat a refractometer entirely — material over the instrument's limit, cabochons, and rough. It is the one cheap instrument that gives a categorical answer rather than a number in a range.

And if you are buying exactly one thing on the numbers, it is neither of these. A 10× loupe and the patience to read inclusions leaves a median of 7 candidates and settles 35 species outright — the strongest single observation in the whole set, on the cheapest instrument in it.

How to Use a Polariscope for Gemstone Identification