An Indian counter runs three tests, and each answers a different question. Kasauti, the black touchstone, tells you whether a piece is gold at all: the metal leaves a streak on the stone, and acid clears whatever in that streak is not gold. A karat meter tells you what the surface of one spot is made of. Fire assay tells you what a sample of the piece truly is, and destroys that sample doing it.
The mistake that costs money is treating a meter reading as if it were an assay. A karat meter prints a figure to a decimal place, and that figure is true of the one spot it read. What lies underneath that spot never reaches the figure.
This guide belongs to the metal and the bench. Where a piece already carries a BIS stamp, and the 6 characters of a HUID etched beside it, the answer is a lookup, and the decoder below does it.
Sona asli hai kaise pata kare?
Start with the cheapest test that can end the conversation. If a streak fails on acid, nothing further is needed. If it passes, a meter narrows the answer. If the money at stake is larger than a laboratory fee, only an assay closes it.
| The test | What it actually reads | What it costs the piece | The question it can settle |
|---|---|---|---|
| Kasauti and acid | A streak rubbed off the surface, beside needles of known purity | A little metal, and a dull spot where acid touches the piece | Is this gold at all, and does it sit near the stamp |
| Karat meter (XRF) | The elements in a thin layer, at the one spot the window sits over | Nothing. No scratch, no filing, no metal | What that surface is made of, where you pointed it |
| Fire assay (cupellation) | A sample taken from the piece, read through its whole thickness | The sample, which never comes back | What fraction of that sample is gold |
Two of the three read a surface. Most arguments at a counter about purity start there.
What can a kasauti streak actually tell you?
Rub the piece on the stone and it takes a streak of the metal. Beside that streak the tester rubs touch needles of known purity, puts acid across all of them together, and watches which streaks fade and which hold.
Base metal reacts and clears. Gold stays. The acid is chosen for the purity being tested, so a strong acid on a low-karat streak and a weak one on a high streak both give an answer worth nothing.
A skilled hand reads a band from that comparison, and the judgement is made by eye, so the tester, the light over the counter and the freshness of the acid all sit inside the result.
It has the same blind spot the meter has. A streak comes off the outside, so a brass core under a good gold skin gives a streak of the skin. The stone is strong at “this is not gold” and weak at the gap between one high karat and the next.
Karat meter kitna sahi hota hai?
Exact about what it measures, and that exactness is the trap. A karat meter is an XRF machine: X-ray fluorescence. It sends X-rays into the piece, the atoms near the surface send back a signature that differs for every element, and the machine sorts those returns into a composition.
What it reads is a thin layer. How thin depends on the metal and the machine, and the maker states it in the unit’s manual. Ask to see that line.
It reads one spot at a time, so moving the window to a clasp, a joint or the inside of a shank makes the same bangle return different figures, which is why a careful reading is taken at more than one place.
And what it reports is a proportion. The meter cannot see a core, a filler or a hollow sitting behind the surface it read, so it has nothing to say about how much gold is in the piece.
So ask two things. Where on the piece was the reading taken, and when was the machine last calibrated. A meter drifts between one servicing and the next, and a shop that looks after its own machine knows that date.
Why does the same piece read high in one spot and low in another?
A bangle straight off the polishing wheel reads high. Pickling, the acid bath that cleans a piece after casting, strips copper from the skin and leaves a surface richer in gold than the alloy beneath it, and polishing works the same way. A thin flash of gold or rhodium plated on last does it deliberately. Plating over brass does it completely.
The one that moves a reading the other way is solder, and it is the one that starts arguments. Solder has to melt below the metal it joins, so it is a lower alloy than the piece. A chain is a row of joints. A jhumka is a stack of parts soldered together. Read the meter on a joint and the figure sinks, because the meter is reading solder and reading it correctly.
So a heavily soldered piece stamped 22K can melt down to a tunch below its stamp, because a refiner tests the whole melted lot — solder, clasps, springs and all.
| What the piece carries | Which way the reading moves | What to do at the counter |
|---|---|---|
| A freshly pickled or polished surface | High | Read a filed window instead of the finished face |
| A gold or rhodium flash over the finished piece | High | Read where wear has taken the flash off, then open a window |
| Plating over brass or a base alloy | High, and by a long way | Open a window. A plated piece reads gold at the skin and nowhere else |
| Solder at joints, links, clasps and hooks | Low | Read away from the joints; a melted lot tests near the average |
A window is a small flat filed where nobody sees it: inside a shank, under a clasp, behind a bangle. It costs seconds and a trace of metal.
Why does an assaying centre destroy a sample to test purity?
Fire assay weighs a sample exactly, wraps it with silver and lead, and heats it in a cupel. The base metals oxidise into the cupel wall, leaving a bead of gold and silver. Acid dissolves the silver away, and what survives is annealed and weighed. Taking the alloy apart is the only way to read all of it.
Whatever share of the starting weight is left at the end is the fineness: a weight before and a weight after.
The centre keeps the drilling or clipping it takes and hands the article back. On a refining lot the whole lot is melted first and the sample dipped from the melt, which is why a refiner quotes on a melted figure and a shop on a surface one.
Assaying and hallmarking centres and refiners run it. It is worth paying for when two surface readings disagree by more than the test costs. Ask the centre what it charges, and which standard the report is issued under. BIS publishes that standard, and a report with no method named on it settles no argument.
Should I ask for the purity reading and the deduction separately?
Two numbers, never one. A single figure said across the counter has somebody’s deductions already folded into it, and once they are folded, neither half can be checked.
So ask twice. What did the test read, and where on the piece was it taken. Then, as its own line: what is being taken off, and for what. Melting and testing loss, refining and the buyer’s margin are three separate deductions, and that second half is the half you can argue, because it is arithmetic.
Where your own shop sits inside the usual bands depends on your refiner, because one refiner does not charge what the next one charges. Read the split written in front of you against the deductions a buyer normally takes, and get purity, weight and every deduction onto paper before the piece leaves your hand.
What do none of these tests tell you about the piece?
Purity is not quantity. All three tests answer how pure the metal is. How much metal is there is a separate question, and it is where old-gold buying actually goes wrong.
Lac, poured into hollow jhumkas and bangles to hold their shape, weighs on the scale. So does the thread through a necklace, the glue under a stone, the stone itself, and the spring in a clasp. A meter reading 22K on a lac-filled jhumka is telling the truth about a shell.
The order that protects you is fixed. Stones and lac out first, then the weight, then the test. Gross weight to net weight to fine metal, in that direction, because a purity figure on a gross weight prices somebody else’s resin as gold.
The water test has the same hole in it. Weighing a piece in air and again in water reads a plain bar of one metal honestly, and a jewellery piece is hollows, solder, stones and filler.
Set one rule for the next piece that crosses your counter. Nothing is weighed for settlement until the surface has been opened somewhere it will not show, and the reading has been taken at two places. That costs one stroke of a file. Skipping it costs the difference between the skin and the metal.