Fellowship-level gemology tests your ability to order observations, not just collect them. The same red colour, the same single refractive character, or the same UV reaction can belong to several different stones, and each instrument has recognised failure modes such as the refractometer's upper limit or strain effects in the polariscope. The study approach here is to build a written decision sequence for each colour family, log every reading with the method used to obtain it, and practise justifying each step. Start today by writing the ruby-versus-spinel-versus-garnet sequence from Section 3 and testing it on any red stone you can borrow.
Why a memorised RI table fails you at the refractometer
The refractometer rewards technique, not memory. A spot reading on a domed stone, a shadow edge near the instrument's cut-off, or a misaligned facet can each produce a value that misleads unless you record exactly how it was obtained.
The instrument works by critical angle, which imposes an upper measuring limit around 1.81 on the standard scale. Stones above that range, such as diamond or demantoid garnet, return no shadow edge at all, so the refractometer can exclude rather than identify. Precision also differs by method: a flat-facet reading can resolve birefringence, while the spot method on a cabochon is approximate and cannot reliably separate two close indices.
Build one habit from the first day: every RI entry in your log states the method, the facet or spot position, and how many shadow edges appeared. On a doubly refractive stone, rotating the stone through a full turn should show the two edges shifting with a constant separation. A single reading with no method note is not data; it is a guess dressed as a number, and your later decisions will inherit that weakness.
Telling true birefringence from anomalous double refraction
Birefringence is a genuine doubling measured as two shadow edges on the refractometer. Anomalous double refraction (ADR) is strain-induced behaviour in a structurally single-refractive stone, seen in the polariscope, and it does not change the measured refractive index.
Between crossed polars, an isotropic stone should stay dark through a full rotation. Strained glass, flame-fusion synthetic spinel, diamond, and some garnets instead show anomalous colours or periodic brightening. This is a real and useful observation, but it is a property of internal strain, not of crystal optics, and it must never be entered in your log as birefringence. The two live on different instruments and mean different things.
Apply a fixed reconciliation rule. If the polariscope shows a reaction, return to the refractometer and take a proper rotated flat-facet reading. One shadow edge across all rotations means a single-refractive stone displaying ADR; two edges with constant separation means true double refraction, whatever the polariscope did. Writing this rule down and following it mechanically is what keeps the two concepts apart under time pressure.
Red stone in the tweezers: ruby, spinel, or garnet?
Red stones span three overlapping families: corundum, spinel, and garnet, with tourmaline as a lower-index outsider. The decision order should be refractometer first, then optic character, then specific gravity, then fluorescence and inclusions.
Scenario 1: a red cabochon gives a spot reading near 1.76, and the tester stops there, calling it almandine garnet. That is the plausible mistake. Near 1.75 to 1.77, a single approximate spot value cannot separate a garnet from a ruby, because the spot method is coarse and the ranges meet. The better decision is to attempt a flat-facet reading on any polished facet, rotating the stone and watching for a second edge; a pair near 1.762 and 1.770 with small separation points to corundum instead.
Why it matters: ruby and red garnet sit in completely different value and identity classes, so a conclusive call from one coarse reading is indefensible. The full sequence rescues the case: garnet is single-refractive and typically inert under longwave UV, while ruby is doubly refractive, around 4.0 in specific gravity, and often fluoresces strong red. The table below gives typical textbook values for drill use; verify exact figures against your own course references.
Treat the table as a drill aid, not a lookup answer key. In practice you will often have a mounted stone, an abraded facet, or a spot-only surface, so the skill being trained is choosing the next discriminating test when the first one is ambiguous.
| Stone | Typical RI range | Optic character | Typical SG | Typical longwave UV |
|---|---|---|---|---|
| Ruby (corundum) | 1.762-1.770 | Doubly refractive, uniaxial | about 4.0 | often strong red |
| Red spinel | about 1.718 | Single refractive | about 3.6 | often strong red |
| Pyrope-almandine garnet | about 1.74-1.80 | Single refractive | about 3.8-4.1 | usually inert |
| Rubellite tourmaline | about 1.62-1.64 | Doubly refractive, uniaxial | about 3.0 | weak to inert |
When the polariscope and the refractometer disagree
When two tests point to different answers, the less discriminating one yields. The polariscope detects reactions; the refractometer with a rotated flat facet measures optic character directly, so the refractometer settles optic character and the polariscope result is reinterpreted.
Scenario 2: a blue stone shows strong anomalous reaction in the polariscope, and the tester concludes it is glass, since glass is isotropic and strained glass shows ADR. The mistake is skipping the reconciliation rule. A flat-facet reading later shows two shadow edges with a constant separation of roughly 0.009: genuine uniaxial double refraction, which glass cannot produce. The stone is consistent with corundum, and the polariscope behaviour is reinterpreted as strain or twinning-related response in a doubly refractive stone.
Why it matters: misreading ADR as proof of glass turns a natural stone into an artificial one, the most serious category error in identification. The corrected reasoning also teaches the reverse discipline: natural corundum can show alternating extinction from twinning, so a lively polariscope is a prompt to measure, never a verdict on its own. Log both results and the reconciliation, because a reasoned conflict resolved correctly demonstrates more competence than a conflict-free run.
Naming treatment evidence without overcalling it
Treatment identification is a conclusion from converging evidence: colour concentrations, flash effects, glassy fillings, or altered inclusion features. A single visible feature justifies a cautious wording, not a definitive treatment name.
Learn the classic evidence by family. Surface-reaching fissures in ruby that flash blue or orange under the microscope suggest lead-glass-type filling, ideally supported by flattened gas bubbles or a glassy texture within the fissure. Dyed material shows colour concentrated along fractures and in cavities. Heated corundum is associated with altered or ruptured silk and discoid inclusion features. Each of these is a magnification skill, so reserve dedicated sessions where you sketch what you see rather than read about it.
Then impose a two-tier conclusion format: first the raw observation, then the inference. For example, 'blue flash in a surface-reaching fissure, consistent with glass filling' is a defensible entry; 'glass-filled ruby' without corroborating bubbles is an overcall. In your logs and in any assessment answer, separating observation from inference shows the reasoning chain, protects you when evidence is partial, and is exactly the discipline a multi-level evidence question is built to reward.
A useful weekly check is to reread three old log entries and rewrite any bare conclusions in the two-tier format. If you cannot supply the observation tier, you did not look carefully enough the first time.
- Fissure-filling evidence: flash effect, gas bubbles, glassy lustre difference within fissures
- Dye evidence: colour concentration along fractures, around cavities, under swabbing where appropriate
- Heat-related corundum evidence: altered silk, discoid halos around included crystals
- Coating evidence: uneven colour in reflected light, colour concentrated on specific facets
- Rule of thumb: one feature suggests, several independent features converge
A daily unknown-stone drill with a scoring rubric
Run a twenty-minute daily drill on one unknown stone: a full observation log in a fixed order, then a written justification for the identification. Score yourself on the sequence and the reasoning, not only on the final name.
The fixed order matters because it simulates a real testing sequence: unaided observation, then 10x magnification with an inclusion sketch, refractometer (spot and flat facet where possible), polariscope through four rotations, longwave and shortwave UV, and a specific gravity estimate. As a benchmark for expected observations, a rock-crystal quartz unknown should give two shadow edges roughly 0.009 apart, four extinction positions per polariscope rotation, no UV reaction, and clean inclusions or none. If your quartz log lacks the double refraction, the drill has found a technique gap, not a knowledge gap.
Score each completed drill against the rubric below. Five out of five is a learning milestone indicating your method is stable, not a prediction of any exam outcome. When a drill scores three or less, the fix is almost always in the order of tests or in an unrecorded method note, so reread the log before choosing a new stone.
Keep every log, dated. The collection becomes your personal casebook of lookalike pairs and your own failure patterns, which is far more valuable in the final weeks than any generic summary sheet.
- 1 point: every instrument step completed in the fixed order, none skipped
- 1 point: each RI value recorded with method (spot or facet) and edge count
- 1 point: optic character supported by two independent tests, with any conflict resolved and noted
- 1 point: inclusions or features sketched with approximate size and position
- 1 point: conclusion written in two tiers, observation first, inference second
A study sequence by discrimination problem, plus readiness checks
Sequence your study by discrimination problem rather than by species list: instrument technique first, then one colour family at a time, then treatments and synthetics, then timed unknowns. Pace the blocks against your rubric scores.
Block one is pure instrument mechanics: you are done when spot and facet readings are reproducible across repeated attempts. Block two works through colour families, red, blue, green, yellow, colourless, and for each family writes a one-page decision tree naming which test discriminates which pair and why. Block three covers treatments and synthetics through magnification practice. Block four is timed unknowns, with every mistake traced back to the block that produced it. Adjust the duration of each block to your own rubric results rather than to a fixed calendar; a candidate strong in microscopy may spend days where a refractometer novice needs weeks.
Use these readiness checks in the final phase. You are ready to sit a practice paper when you can: reproduce a spot RI within about 0.01 across three tries on the same stone; recite, for five colour families, the reasoning behind your test order; write two-tier treatment conclusions unaided; and complete a full unknown log within your own target time. Administrative matters such as seminar dates and enrolment are published by the issuer, the Deutsche Gemmologische Gesellschaft, at dgemg.com; that site is also the right place to confirm anything about the credential's structure that this study method deliberately leaves open.
- Reproducible spot reading within about 0.01 across three attempts on one stone
- One-page decision tree written from memory for five colour families
- Two-tier treatment conclusions produced without prompting
- Full unknown log completed within your self-set time target
- Every past drill mistake mapped to the block that caused it
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.