How to identify fake gold: UK guide

Why fake gold exists

A piece mistaken for solid gold may be plated, filled or rolled jewellery whose original description has been lost. That is different from a counterfeit sold dishonestly. The usual suspects are a rolled-gold watch from the 1960s or 70s, a chain bought abroad and remembered as better than it was, and costume pieces that ended up in the same inherited jewellery box as the real thing. An at-home check is there to flag those items before anyone mistakes them for solid gold. On arrival, inspection considers construction and marks alongside readings at suitable test points. Exposed wear edges can help, but an intact thick gold layer may conceal the core from XRF.

Test 1: The magnet test

Pure gold is not magnetic. If a piece is attracted to a strong neodymium magnet, it contains a magnetic base metal (typically iron, nickel or steel) and is not solid gold. The reverse is not proof: many plated, filled or low-carat alloys are also non-magnetic. One caveat before you condemn a chain: the spring inside a genuine gold clasp is usually steel, so the clasp will often snap to a magnet even when every link is solid gold. Test the links, not the fastening. A magnet is a fast fail-test, not a pass-test.

Test 2: Density (the bench-jeweller’s test)

Gold is very dense. Pure 24ct weighs about 19.3 g/cm³. 18ct weighs about 15.5 g/cm³, 9ct about 11.3 g/cm³. If you can measure the item’s weight and the volume of water it displaces, the density tells you a lot. A worked illustration: a plain band weighing 6.20 g that displaces 0.40 ml of water has a density of 15.5 g/cm³, right where 18ct sits. If the same ring displaced 0.70 ml, the density would be 8.9 g/cm³, which is brass territory. You need scales that read to 0.01 g and a narrow measuring vessel to get anywhere near this at home. The catch is hollow items, stone settings and solder, all of which throw the figure off. A bench jeweller will isolate the metal first; at home, this is a rough sanity check, not a verdict.

Test 3: The ceramic plate scratch

Do not use a ceramic scratch as a home verdict on an unknown piece. It can mark the object, and the colour of a streak does not establish the construction or purity of the whole item. Keep the surface intact, photograph the marks and any existing wear, and ask what examination is appropriate. A test that damages the piece without resolving the question leaves you with less evidence and potentially less value.

Test 4: Read the hallmark properly

A UK hallmark is the easiest test most people miss. 375 = 9ct, 585 = 14ct, 750 = 18ct, 916 = 22ct, 999 = 24ct. A standing lion is a sterling-silver mark, not a gold mark, so a "lion" on a yellow piece usually means the yellow is gold-coloured plate over silver. Marks like GP, GF, RG, HGE and 1/20 identify plated, filled or rolled-gold items rather than solid gold. Check the wear points too: plating goes first where the piece rubs, the edges of a ring shank, the crowns of chain links, the back of a pendant. A different colour, grey, pink or brassy, showing through at those spots is plate wearing off a base metal.

Counterfeit marks exist. A convincing-looking punch is not a substitute for examining the complete mark set and the item. See the gold hallmark guide for the symbol set.

Test 5: The acid test (and why we do not use it)

The traditional jeweller’s acid test scratches the metal onto a touchstone and applies an acid graded for a carat (e.g. 9ct, 18ct, 22ct). If the streak dissolves under a higher-carat acid, the item is below that carat. It is informative, and it permanently marks your piece. We replaced acid testing with XRF for this reason: XRF measures the same composition without leaving a trace, so if you decline our offer your items go back to you exactly as they came in.

When in doubt, ask before posting

If a clear photo of the mark, the back of the clasp and the inside of a ring is sent on WhatsApp, an experienced eye can usually tell whether a piece is worth posting in. There is no charge for that conversation and no obligation to send anything afterwards. If you do send it, the piece is checked by XRF rather than acid, so nothing is scratched.

Record a clue without turning it into a verdict

Use a short observation list for each item: visible marks, where the colour differs, whether a fastening reacts differently from the body, and which areas you could not inspect. Write “clasp attracted” rather than “chain is fake”. Write “750 visible, other punches unclear” rather than “confirmed 18ct”. These notes give an examiner something to work with while leaving room for an explanation that a home check cannot settle.

Take one photograph of the entire item before close-ups. A tightly cropped image of a number can hide the fact that it is on a detachable clasp rather than the main piece. For a pendant on a chain, photograph and describe them separately. The same applies to a pair of earrings with replacement backs: the backs, posts and decorative fronts should not automatically be treated as one material.

If a supposed test gives an odd result, repeat the observation without escalating to a damaging test. Check that the scales were zeroed, that the item was not touching another object and that the mark photograph is in focus. Do not file a fresh patch, cut a link or remove a stone to make an informal experiment easier. At that point you would be changing the evidence and potentially reducing the value of something you have not yet identified.

Ask what an examination can actually establish

Before agreeing to any procedure, ask whether it is non-destructive, which part of the item will be examined and what the result can and cannot establish. “Tested” is too broad a description on its own. An examination of an exposed surface, an assessment of construction and an identification of a hallmark are different pieces of evidence. A responsible conclusion should explain how those pieces fit together, including any remaining uncertainty.

If you already have a report, send its relevant findings with the photographs and say which item it covers. Keep the report date and item reference visible. Do not assume that a result for one ring applies to a visually matching ring, or that a reading from a clasp describes every link in a chain. Ask for clarification where the report does not identify the part examined.

For jewellery bought recently and suspected to have been misdescribed, preserve the listing, receipt, messages and original photographs before making changes to the item. Getting a material identified and resolving a purchase dispute are separate tasks. A sale for metal could remove the very object someone needs to examine, so establish what evidence is required before deciding what to do with it.

A useful first message for an uncertain piece

An enquiry can be simple: “I have a yellow-coloured chain with 750 on the clasp. The other punches are too worn for me to read. I have attached the whole chain, both sides of the clasp and the worn area on one link. I have not scratched or cut it. Can you tell me what else you need to assess whether it is suitable to send?” This is an illustrative message, not a promise that those photographs will identify every construction.

That message gives more useful information than a confident label such as “real gold” or “fake gold”. If the answer is that a photograph cannot resolve it, ask what examination is proposed and what happens if the piece is outside the buyer’s service. You can decide whether to proceed once those practical questions are clear.

Common questions