Testing a stone at home

Search for how to tell if your gua sha is real jade and you’ll be given a short list of tests presented with far more confidence than they deserve. Scratch it with a knife. Put it in water. Hold it against your cheek and see if it stays cold.

Each of those has some physical basis. None of them does what it’s usually claimed to do. The useful version of this post isn’t a better list of tests — it’s an accurate account of what each one rules out, because ruling things out is the only thing home testing does.

What the tests are actually measuring

Every home test is a crude proxy for one of four measurable properties: hardness, density, thermal conductivity, or optical appearance under magnification. Gemmological identification uses those same properties plus refractive index and spectroscopy, measured precisely with instruments. The gap between “measured precisely” and “estimated in a kitchen” is where almost all the false confidence lives.

The scratch test

What it does. Establishes an approximate lower or upper bound on Mohs hardness by finding out what scratches what. A steel knife blade sits somewhere around 5 to 6.5 depending on the alloy; window glass is about 5.5; a quartz point is 7.

What that rules out. If a steel point scratches the stone easily, it’s softer than the steel, so it isn’t nephrite (6–6.5), jadeite (6.5–7) or quartz (7). That’s a genuine elimination and it’s the single most informative thing you can do with household objects. Soft green stone is very often serpentine.

What it does not do. It cannot distinguish nephrite from jadeite — they’re a fraction of a Mohs step apart. It cannot distinguish either from a hard imitation. And a stone that resists steel is merely “harder than about 6,” a bracket containing quartz, both jades, amazonite, and a large number of other rocks.

The trap. Scratch tests damage the object, and people routinely misread the result. Steel dragged across a harder stone leaves a grey metal streak that looks exactly like a scratch and rubs off. More than one stone has been declared fake on the strength of a metal smear.

Test on the back, in a small area, or don’t test.

The water test

What it does. Depending on which version you mean, either nothing or a rough density check.

The common internet version — put it in water and see if it sinks, or whether water beads — has no diagnostic value. Everything solid here sinks. Beading tells you about surface finish, not material.

The version with content is specific gravity by displacement: weigh the stone dry, weigh it submerged (or measure the water it displaces), and compute density. Jadeite is around 3.30–3.38, nephrite 2.90–3.03, serpentine 2.4–2.8, quartz about 2.65, glass typically 2.4–2.6. Those are genuinely separated numbers.

Why it usually fails anyway. Doing this well needs a scale accurate to about 0.01 g and careful technique; the objects are light, so small errors dominate. A kitchen scale reading to 1 g on a 40 g tool gives you an error bar wide enough to span two of the categories above. It’s real physics executed at the wrong precision.

The temperature test

What it does. Assesses thermal conductivity by feel. Stone conducts heat away from your skin faster than resin or plastic, so it feels colder and stays cold longer.

What it rules out. A tool that warms almost immediately and feels like a warm object is likely resin or plastic. That’s worth something.

What it does not do. Glass is intermediate and often ambiguous. Everything genuinely stony feels much the same. And the test is completely confounded by the room, by your hands, and by whether the thing has been sitting on a cold surface.

The loupe

The best value per pound of anything in this list. A 10× loupe costs very little and is the only tool here that produces positive identifications rather than eliminations.

Under magnification, in daylight, three findings are strong:

Spherical bubbles — glass or resin. Natural stone doesn’t contain them. Genuinely diagnostic, with the standing exception that obsidian is itself natural glass.

Colour concentrated in a web along fine cracks rather than distributed through the body — dyed. Very strong indicator.

Visible particles in a uniform matrix — reconstituted powder in a binder.

Weaker but useful: mould seams at the edges, flowing swirl patterns rather than mineral banding, and surface crazing.

What you can tell

WHAT YOU CAN TELL — the full honest ledger

  · Softer than steel
                    → not nephrite, jadeite or quartz.
                      Real elimination.

  · Bubbles under 10× magnification
                    → glass or resin (obsidian excepted).
                      Real positive.

  · Colour pooled along cracks
                    → dyed. Strong.

  · Warms instantly in the hand
                    → probably resin. Weak but useful.

  · Very light for its size
                    → lower density than stone; suspect
                      resin or reconstituted material

  · Nephrite or jadeite?
                    → NOT DETERMINABLE at home.

  · Type A or Type B jadeite?
                    → NOT DETERMINABLE at home. Needs
                      infrared spectroscopy.

  · Natural or synthetic quartz?
                    → NOT DETERMINABLE at home, and
                      often difficult in a lab.

  · Whether an undyed-looking stone is untreated
                    → NOT DETERMINABLE at home.

  · Whether it came from a claimed locality
                    → NOT DETERMINABLE, at home or
                      generally, from the object alone.

Five eliminations, four hard walls, and one question that no equipment answers. That’s the true shape of home testing, and it’s a much smaller box than the genre implies.

The tests to skip

The hair-burning test. Wrap a hair around the stone, apply a flame, see whether the hair survives. Circulating for jade; heats the object unevenly, risks thermal shock, and establishes nothing about species.

The sound test. Tapping two pieces together and judging the ring. Experienced dealers do use sound, and there is something in it — dense interlocking material rings differently from porous or resin-filled. But it requires a trained ear and a known reference piece, and it’s read wrong far more often than right by people without both. It also chips edges.

The tongue test. Not diagnostic, and given what you don’t know about the surface treatment, inadvisable on any grounds.

What to do with an inconclusive result

Usually: nothing. This is the part the genre never says.

For a tool at typical gua sha prices, professional testing costs more than the object. A gemmologist will identify the species properly and can detect polymer impregnation, and that service is worth buying for a stone with real value. For most tools, the rational endpoint is knowing roughly what bracket the object is in — soft green stone, hard green stone, glass, resin — and stopping there.

Knowing you own a nice piece of serpentine is a complete answer. It’s just not the answer the listing gave you.

The honest limit

Everything on this site converges here. Home testing eliminates; it does not confirm. You can prove something is glass. You can prove something is softer than jade. You cannot prove something is nephrite, and any page that tells you a knife and a glass of water will do it is selling either a product or a certainty.

The stone in your hand has a determinate composition. Finding out what it is requires instruments you don’t have. What you can do — and it’s not nothing — is narrow the field honestly and know which questions you’ve actually answered.