How to Tell If Gold Is Real: Seven Tests, Ranked
A stamp is a claim, not a fact. Here are the tests in order of how much they prove and how much damage they do, starting with the ones that cost nothing and ending with the one that settles it.
Every test below answers a slightly different question, and the order matters. Do the free, non-destructive ones first, because two of them settle most cases in under a minute.
1. Look for the wear, not the stamp
Free, instant, and more reliable than people expect. Plated jewelry wears through at the points of friction: the underside of a ring shank, the edges of a bangle, the high points of any raised design, the back of a clasp. Under a loupe you are looking for a different colour showing through, usually a duller yellow, a pink from copper or a grey from nickel.
Solid gold wears down. Plated gold wears off. The difference is visible at the edges long before it is obvious to the naked eye.
2. The magnet test
Free, thirty seconds, and it eliminates a specific class of fake. Gold, silver and platinum are not magnetic. If a piece jumps to a strong magnet, the body is steel or another ferrous alloy and it is plated.
What it does not prove: many base metals used in plating, brass and copper in particular, are also non-magnetic. Passing the magnet test means only that the piece is not made of steel. Use a strong neodymium magnet rather than a fridge magnet, and test the body of the piece rather than a clasp, since spring mechanisms are often steel in perfectly genuine gold jewelry.
3. Weigh it in your hand
Free, and surprisingly discriminating once you have handled a few pieces. Gold is heavy. 18 carat gold is roughly twice as dense as silver and around two and a half times as dense as most base metals. A gold ring feels wrong in the hand if it is not gold: too light for its size.
Platinum is heavier again, about ten per cent denser than 18 carat gold and roughly twice the density of sterling silver. A white metal piece that feels much heavier than it looks is usually platinum.
4. Read the marks properly
Free, and the step most often done badly. A fineness figure struck alongside an assay office mark means the metal was tested by a licensed body. A fineness figure standing alone means somebody owned a punch.
The marks that mean the piece is definitely not solid are worth knowing by heart: GF, RGP and fractions such as 1/20 12K GF mean gold filled; GP, GE, HGE and 18KGP mean plated; EPNS and EP mean electroplated. Full detail in reading hallmarks and every figure in the fineness numbers.
5. The acid test
Cheap, quick, and it damages the piece slightly. A test kit contains bottles of acid mixed for specific carats and a testing stone. You rub the piece on the stone to leave a streak of metal, apply the acid for that carat, and read what happens: the streak stays for gold at or above that standard, fades for lower, and disappears for base metal.
What it proves: the fineness of the surface. What it does not prove: what is underneath. A thick gold-filled layer or a heavily plated piece can pass an acid test on the surface, which is why a good tester files a small notch into an inconspicuous edge first and tests the metal exposed underneath.
That notch is real damage. On an antique piece it is a decision, not a routine step. Filing into a Georgian shank to satisfy curiosity is a poor trade.
6. Specific gravity
Non-destructive, accurate for solid pieces, useless for anything hollow or stone-set. Weigh the piece in air, then weigh it suspended in water. Divide the dry weight by the loss of weight in water to get the density.
| Metal | Density, grams per cubic centimetre |
|---|---|
| 9 carat gold | about 11.3 |
| 14 carat gold | about 13.1 |
| 18 carat gold | about 15.5 |
| 22 carat gold | about 17.7 |
| Sterling silver | about 10.4 |
| Platinum, 950 | about 21.0 |
| Brass | about 8.5 |
Any stone, hollow section, enamel or solder of a different alloy throws the result off, which rules the method out for most jewelry and makes it excellent for plain bands and solid chains.
7. XRF, the one that settles it
An X-ray fluorescence analyser reads the elemental composition of the surface in a few seconds without marking the piece. Most serious dealers, pawnbrokers, assay offices and many jewelers have one, and many will test a piece for a small fee or for nothing.
What it proves: exactly which elements are present and in what proportion, at the surface and a little way beneath. What it still does not prove: what is at the centre of a thick piece. Tungsten-cored fakes exist in bullion bars and are essentially unknown in jewelry, because the shapes are wrong for it.
If the answer matters financially, this is the test to ask for, and it is the reason not to bother with the home methods when a jeweler is ten minutes away.
Tests to ignore
The bite test. It tells you the metal is soft. Lead is soft. It also damages the piece and your teeth.
The ceramic scratch test. Dragging the piece across unglazed tile is supposed to leave a gold streak for gold and a black streak for base metal. It scratches the piece and is unreliable on plated items.
The vinegar test. Widely repeated online and close to worthless. Vinegar is too weak to distinguish reliably between gold alloys and several base metals.
The float test. Real gold sinks. So does almost everything else jewelry is made of.
What to do if it turns out to be plated
Rather less disappointing than it sounds. A great deal of good nineteenth and early twentieth century jewelry is gold filled or rolled gold, and it was sold honestly as such at the time. Edwardian gold filled work in particular can be beautifully made, has real collectable value, and will outlast a thin modern gold chain.
Value in that case rests entirely on design, maker and condition rather than on metal, which is a different question but not an empty one. See what is my antique jewelry worth.
Frequently asked questions
How can I tell if gold is real at home?
Look at the wear points with a loupe for a different colour showing through, test with a strong magnet, and weigh the piece in your hand against its size. Those three cost nothing and settle most cases. Anything more definite needs an acid test or an XRF reading.
Does real gold stick to a magnet?
No. Gold is not magnetic. A piece that jumps to a strong magnet has a ferrous core and is plated. Passing the test does not prove the piece is gold, since brass and copper are also non-magnetic.
Is a 585 or 750 stamp proof that jewelry is gold?
No. It is a claim. Where an assay office mark accompanies it the metal was tested by a licensed body; where the figure stands alone, nobody has verified it. Plated jewelry is routinely stamped with fineness figures.
Can gold plated jewelry be stamped 18K?
Yes, and legitimately, when it reads 18KGP or 18KGE. That means plated with 18 carat gold rather than made of it. The carat figure describes the coating.
Where can I get gold tested for free?
Most pawnbrokers and many jewelers will XRF-test a piece while you wait, often at no charge, because they hope to buy it. Assay offices in hallmarking countries offer testing for a fee and give an authoritative answer.
Does real gold tarnish?
High carat gold does not. Lower standards can, because most of the alloy is copper and silver. 9 carat and 333 gold in particular will darken over years, especially against skin, and that is normal rather than a sign of a fake.