Collecting · Condition

How to tell if a mineral specimen is repaired

Repair is normal in the mineral trade and undisclosed repair is not. A six-step inspection order, what raking light and UV each reveal, and the effect on value.

First move Turn it overBest cheap tool A raking lamp at 30 cmSecond tool Longwave UV, in the darkDisclosed repair Trade-normalUndisclosed repair Not

Hold the specimen 30 cm from a bare lamp and rake the light across it: adhesive shows as a change in reflectance before it shows as a change in colour. Then check it under longwave UV in the dark, because most epoxies and cyanoacrylates fluoresce differently from the mineral and the matrix around them.

In short

  • Repair is not fraud. Undisclosed repair is. A clean break in a specimen's own material, rejoined and declared, is trade-normal on old collection pieces. A crystal glued on from somewhere else is not a repair at all.
  • Work in a fixed order. Most people start by squinting at the crystal; most repairs are on the back, the base, or a matrix contact, because that is where specimens break and where a repairer expects nobody to look.
  • Raking light is the highest-yield tool and it costs nothing. A glue line differs in gloss from the mineral around it even when it matches in colour.
  • Longwave UV is the confirmation, not the search. It will not find a repair you have not already suspected, but it will settle one you have.
  • The value question is not whether a piece is repaired but what was repaired and whether you were told. Rejoined matrix costs little; a rejoined crystal termination costs a great deal.
Where the trade draws its lines
What was doneWhat it meansTrade viewEffect on value
Cleaning — acid, ultrasonic, mechanicalRemoval of coating, not addition of materialNormal, usually not itemisedNone, if the species survived it undamaged
Matrix trimmingExcess rock removed to improve balanceNormal, usually not itemisedUsually positive
Repair: the specimen's own broken piece, rejoined, disclosedA clean break put backAccepted, and universal on 19th-century materialReduces, mildly on matrix, sharply on a crystal
Repair, undisclosedThe same act, without the sentenceNot acceptedMaterial, and it damages the seller more than the piece
Restoration: added filler, tinted resin, rebuilt terminationNew material that was never part of the specimenMust be disclosed explicitlySubstantial
Composite: parts of two specimens married into onePresented as a single natural pieceNot acceptedDestroys collector value
Fabrication: crystals set into a base with adhesive and dusted matrixManufactured objectFraudNo collector value at all

The six-step check, in order

Order matters, because each step is cheaper than the one after it and rules out more.

  1. Turn it over. Look at the back and the base before the display face. Specimens break where they were struck loose or where they were dropped, and both are usually behind or underneath.
  2. Rake the light. Hold the piece about 30 cm from a bare bulb or a phone torch, at a low angle so the light skims the surface, and rotate it slowly. You are watching for a line where the gloss changes — adhesive is almost always either shinier or duller than the mineral, even when a repairer has matched the colour perfectly.
  3. Follow every crystal-to-matrix junction. Nature grows crystals out of matrix, so the contact is intergrown and irregular. Adhesive sits on matrix, so the contact is a seam. A crystal whose base meets the rock along a clean flat line did not grow there.
  4. Look for filler. Restoration commonly uses adhesive loaded with crushed matrix or mica to bridge a gap. It matches in colour and fails in texture: it is matte where the rock around it has micro-glitter, and it has no grain.
  5. Longwave UV, in the dark, after your eyes adjust. Most epoxies and cyanoacrylates fluoresce — typically a dull bluish white — against a mineral and matrix that do something else or nothing at all. Interpret with care: plenty of minerals fluoresce in their own right, so a glowing area is a question, not an answer.
  6. Ask, and ask for it in writing. The reply is the most informative test on this list. A seller who says yes, there is a repair through the base, here is a photograph of it has told you both about the specimen and about themselves.

What raking light shows that a photograph never will

Almost every online listing is lit flat and frontally, which is exactly the lighting that hides a repair. That is not usually deliberate — flat light is also the lighting that makes colour read correctly — but the effect is the same. A glue line is a specular feature: it announces itself in the reflection, and a photograph optimised for colour has removed the reflections.

So when buying at distance, ask for one extra photograph taken with a single light source low and to one side, showing the back and the base. It is a small request, it is instantly recognisable to any experienced seller as a condition question, and a refusal is itself informative. On old British material this is close to routine: much of it was collected in the 1800s, has changed hands many times, and has been repaired at some point by somebody. We say the same thing about condition on our page about buying old collection material, and it applies with particular force to Cornish copper species, which are both fragile and old.

Repair, restoration and fabrication are three different things

The vocabulary matters, because the words are used loosely and the difference in value is enormous.

Repair puts a specimen's own broken piece back where it came from. Nothing is added but adhesive. On a 150-year-old Cornish specimen this is close to expected, and a well-executed, disclosed repair through a matrix is a minor matter.

Restoration adds material that was never part of the specimen: filler in a void, tinted resin rebuilding a chipped termination, a sculpted replacement for a missing crystal tip. This must be disclosed explicitly and described, because the buyer is being asked to pay for a surface that is not mineral.

Fabrication is manufacture. Loose crystals stood upright in a bed of adhesive with crushed mica dusted over it to hide the join is the pattern documented most often, and the giveaway is usually statistical rather than physical: real vugs do not produce a dozen equally sized crystals all standing at the same angle. This is not a condition issue, it is a fake, and we treat it as one on how to tell if a crystal is fake or treated.

A fourth category sits awkwardly between them: the composite, where genuine pieces of two different specimens are married into a better-looking one. Everything on it is real. The object is not.

What a repair does to what a specimen is worth

We do not publish price guidance and there is no market rate to quote, but the direction and rough proportion of the effect are consistent enough to state:

  • A repair in the matrix, away from the crystals, is the mildest case. The specimen still shows what it is meant to show.
  • A repair through a crystal is much more serious, and it is worse the closer it is to the termination, because the termination is the part that carries the aesthetic and the part that is hardest to fake convincingly.
  • A repair on a species that is common matters more than the same repair on a species that is rare, because the buyer has alternatives. A repaired specimen of something that comes to market twice a decade is still the only one available.
  • Disclosure changes the transaction, not the object. A repaired specimen bought knowingly at a fair price is a good purchase; the identical specimen bought unknowingly is a bad one. That is why the sixth step of the check is asking.

Doing this yourself is entirely realistic, and worth saying plainly: a torch, a dark room, a cheap longwave lamp and the discipline to always turn the piece over will catch most of what is out there, and it costs an evening to learn. What it will not do is scale. Checking one specimen properly takes ten minutes; checking a whole cabinet coming out of a deceased estate takes days, and the pieces you most want are the ones you will see photographs of rather than hold. Where that becomes the constraint, tell us what you are after on the wanted list — the value is in somebody having had it in their hands under a raking lamp before you commit, not in us holding anything, because we hold no stock.

Questions

Does a UV lamp reliably show glue?
It shows most of it, and it is not infallible. Common epoxies and cyanoacrylates fluoresce under longwave UV against a mineral that usually does not, which is why it is the standard screening tool — dealers describe it as a fast screen rather than a proof, as in McDougall Minerals' overview Beware the Hand of Man. Some adhesives are formulated not to fluoresce, and many minerals fluoresce on their own account, so use it to confirm rather than to search.
Is it normal for old specimens to be repaired?
Yes. Most classic British material was collected in the 1800s and has moved between cabinets ever since, and pieces get dropped. On Cornish copper species and Caldbeck Fells material in particular, a disclosed repair is unremarkable. What is not normal is finding out afterwards.
Should I avoid buying a repaired specimen?
Not necessarily. Buy it knowingly, at a price that reflects it, and record the repair on your own label so the next owner is told too. For a rare species from a locality that no longer produces, a repaired specimen is often the only realistic option, and refusing all of them narrows a collection more than it improves it.
Can I repair a specimen myself?
You can, and a clean break rejoined carefully is better than a specimen kept in two pieces. Use a reversible adhesive rather than a permanent epoxy, apply it sparingly so nothing squeezes out onto a crystal face, and write what you did on the label. The irreversible, over-applied repair is the one that destroys value.
How do I tell a natural crystal contact from a glued one?
Look at how the two materials meet. A crystal that grew in place is interlocked with the matrix — the contact is irregular, and small crystals or crusts of other species usually run across it. A glued contact is a seam: flat, continuous, and interrupting whatever was on the matrix surface rather than growing with it.