Start here: UK gold hallmarks
Before reaching for any physical test, always look for hallmarks first. In the UK, all gold items over a certain weight that are sold commercially must by law be assayed and marked with a hallmark confirming their gold content. This requirement has been in force in various forms since the 14th century and is administered by the four UK assay offices — London, Birmingham, Sheffield, and Edinburgh.
A full UK hallmark on a gold item typically includes four components: the sponsor's mark (the maker or importer), the millesimal fineness number (indicating gold purity), the assay office mark, and sometimes a date letter. For practical purposes, the fineness number is what you are looking for.
| Fineness mark | Carat | Gold content | Common use |
|---|---|---|---|
| 375 | 9ct | 37.5% gold | Most common in the UK; everyday jewellery, chains, rings |
| 585 | 14ct | 58.5% gold | Common in continental Europe and the US; less common in UK-made pieces |
| 750 | 18ct | 75.0% gold | Fine jewellery, luxury watches, engagement rings |
| 916 | 22ct | 91.6% gold | High-carat jewellery; traditional Asian and Middle Eastern pieces |
| 999 | 24ct | 99.9% gold | Bullion coins, investment bars; too soft for most wearable jewellery |
Hallmarks are small — typically 1–2mm — so a 10x loupe or magnifying glass is useful. You can also use the zoom on a smartphone camera: place the item on a white surface, zoom in on the area inside the clasp, inside a ring band, or on the back of a pendant bail, and photograph it. The image can then be enlarged on screen.
If you find a fineness number from the table above, you can be confident you are looking at genuine gold. Items that carry no hallmarks — or marks that simply read "GF" (gold-filled), "GE" (gold electroplate), or "GP" (gold plated) — are not solid gold, whatever they look like.
Antique items made before 1999 may carry older carat marks (a crown with a number) rather than the current millesimal fineness. Items made abroad — particularly in continental Europe, the US, and Asia — may use the fineness system but without UK assay office marks. These can still be genuine gold. When no hallmark is present or legible, the physical tests below become your next step.
Seven home tests to try
Gold is not magnetic. Hold a strong rare-earth magnet (available cheaply online) close to the piece. If the item is strongly attracted to the magnet, it is either iron, steel, or contains a magnetic metal core. Real gold will show no attraction at all.
Use a rare-earth (neodymium) magnet, not a fridge magnet — the latter is too weak to detect the small magnetic responses that cheaper alloys can produce. This test takes about two seconds and costs nothing if you already own a suitable magnet.
Gold is extremely dense — approximately 19.3 g/cm³ for 24ct, and between 13 and 18 g/cm³ for lower carat alloys. Fill a glass or jug with water and drop the item in. A genuine gold piece will sink immediately and fall straight to the bottom.
If the piece floats, hovers in the water column, or sinks very slowly, it is likely gold-plated over a much lighter base metal such as aluminium or plastic. This test works best for chains and loose pieces; it is less decisive for hollow items, which can trap an air pocket that slows the sink rate.
Hold the piece firmly between two fingers and press it against the palm of your hand for thirty seconds. Real gold does not react with skin and leaves no mark. If you notice a green, black, or blue-black mark on your skin, the piece contains copper, nickel, or other base metals in significant quantity.
This test is less reliable for 9ct gold, which contains enough copper to occasionally leave a faint mark on some people's skin — particularly if skin pH is acidic. An 18ct or 22ct piece should leave absolutely no mark. A vivid green mark is almost certainly not gold at all.
Find an unglazed ceramic surface — the underside of a ceramic plate or tile, or an unglazed terracotta piece. Gently drag the item across it. A real gold piece leaves a gold or yellow streak. A gold-plated item with a non-gold core will either leave a black streak (if the base metal is exposed) or scrape through to reveal a different colour beneath the gold surface layer.
Use only a light touch. This test does leave a small scratch on the item's surface, so use it on an inconspicuous area — or avoid it entirely on pieces you wish to preserve in perfect condition. It is most useful for confirming a positive result from other tests.
Nitric acid reacts with base metals and low-grade alloys but does not react with gold. A small drop on a scratch mark will: turn green if the piece is not gold or gold-plated copper; turn milky if it is silver or silver-plated; leave no colour change if gold is present. More concentrated acid formulas test for different carats.
This is the most chemically reliable home-level test, but it requires acid that must be handled safely, used in a ventilated space, and disposed of correctly. Many high-street jewellers will perform this test for a small fee or for free if you are considering selling. We recommend having this done professionally rather than sourcing the chemicals yourself.
Examine the piece carefully under good light with a magnifying glass. Look at wear points — clasps, edges, prongs, the inner surface of ring shanks. On a gold-plated piece, the base metal often shows through at points of friction: a different, duller colour beneath a yellow surface layer.
Real gold is consistent in colour all the way through. You may also notice that very cheap pieces have an inconsistent surface colour or visible brush marks in the plating. Genuine gold pieces tend to be heavier than equivalent fake pieces of the same size.
Electronic gold testers measure the electrical conductivity of the metal and use this to estimate purity. They are available for around £15–£50 and can test to within one or two carats of accuracy without any acid or physical marking. They work by touching a probe to a conductive surface of the item.
For a collection of mixed pieces — common in an estate or loft clearance — an electronic tester pays for itself quickly and gives consistent results without the ambiguity of physical tests. The limitation is that they cannot distinguish between a solid gold piece and a gold-plated piece if the plating is thick enough to conduct in the same range.
Tests that do not reliably work
Several widely-shared methods are unreliable and should be treated with scepticism.
- The bite test. Biting gold does not tell you much. Gold is soft and will dent slightly, but so will lead and many plated alloys. This test risks damaging dental work and is not diagnostically useful.
- The bleach test. Real gold is unaffected by household bleach. But this test damages the finish of the piece whether or not it is genuine, and the result is not conclusive.
- Colour alone. Gold plating on high-quality fakes can be visually indistinguishable from solid gold. Never rely on appearance without at least one of the tests above.
- Weight by feel. Useful as a very rough indication but easily fooled by tungsten-filled pieces, which are designed to mimic gold's weight.
When you know it's real — what next?
Once you have confirmed a piece is genuine gold, the next question is what it is worth and what to do with it. Gold value has two components: the intrinsic metal value (based on the weight and purity) and any additional collector or antique value (based on the maker, style, age, and condition).
For unwanted pieces with primarily metal value — broken chains, single earrings, mixed jewellery — the gold price is the main reference. The London gold fix (published daily) sets the baseline. You can calculate the approximate melt value yourself: weigh the piece in grams, multiply by the fineness fraction (0.375 for 9ct, 0.585 for 14ct, 0.75 for 18ct), and multiply by the current gold price per gram in sterling.
For pieces that appear to have maker's marks, hallmarks suggesting significant age, unusual craftsmanship, or brand identifiers — Cartier, Van Cleef, Boucheron, a notable Scottish silversmith — the collector value may substantially exceed the metal value. In these cases, a specialist assessment is important before selling through any channel that prices only on metal weight.
When you submit gold jewellery to Fair Vintage, our specialists assess both the metal value and any maker, period, or collector value above the melt price. You receive a written valuation per item explaining the offer and the basis for it. Payment within 72 hours of acceptance, or we add 3% — no obligation to accept.
Know what your gold is worth
Send a photograph of your gold pieces for a free specialist estimate. No obligation, no commitment. Our team will identify the pieces, confirm the hallmarks from your image, and give you a valuation before you post anything.
Get a free estimate →Frequently asked questions
The most reliable method at home is to look for UK hallmarks using a magnifying glass. Genuine gold jewellery sold in the UK carries an assay office hallmark confirming its carat. For older pieces or items bought abroad, additional tests help: the magnet test (real gold is not magnetic), the float test (real gold sinks immediately in water), and the skin test (real gold leaves no mark on skin) are all quick, safe checks that give useful information without specialist equipment.
In the UK, gold is hallmarked with a millesimal fineness number: 375 (9 carat — 37.5% gold), 585 (14 carat — 58.5% gold), 750 (18 carat — 75% gold), 916 (22 carat — 91.6% gold), and 999 (24 carat — 99.9% pure). Items are also stamped with one of the four UK assay office marks: the London anchor, the Birmingham anchor, the Sheffield rose, or the Edinburgh castle. You may need a 10x loupe to read hallmarks clearly — they are typically 1–2mm in size.
Real gold sinks immediately. It is one of the densest metals — denser than almost any metal used to fake it. If a piece floats, hovers, or sinks very slowly, it is either not gold or is gold-plated over a light base metal such as aluminium. Simply fill a glass with water and drop the piece in. A genuine gold item will fall straight to the bottom. Hollow items can occasionally give a false result if they trap air.
No. Real gold is not magnetic and will show no attraction to a magnet. Use a strong rare-earth (neodymium) magnet rather than a fridge magnet, which is too weak to detect small magnetic responses in alloys. A strong pull toward the magnet means the piece is not solid gold. No magnetic response does not confirm gold on its own — some base metals are also non-magnetic — but a magnetic response is a clear negative sign.
These are millesimal fineness marks showing the gold content per 1,000 parts. 750 = 18 carat (75% gold). 585 = 14 carat (58.5% gold). 375 = 9 carat (37.5% gold) — the minimum standard legally described as gold in the UK. These marks are the international equivalent of the UK carat system and appear on pieces made in continental Europe and many other countries. If you see these numbers clearly stamped on a piece, you are looking at genuine gold.