The Scratch Bench · mineralogical working copy
Plate 066 Kit MH‑10 Bench clock 00:00:00 Lamp north light

After Friedrich Mohs, 1812 · ten minerals · one rule · drag to scratch

Ten minerals, in order.
Scratch them yourself.

Hardness is a rumour until you have held one stone inside another’s furrow. Here the scale is a kit: draw a specimen from its drawer, seat a block in the clamp, and pull the harder across the softer. It will cut. Pull the other way and the bench simply skids — and powders.

§ 01–03

The kit, the test, the record

one tile · two canvases · one red pen

Drawers 1–10 · Mohs kit

coll. 1812

Common objects · cross-checks

tab.
In hand
—
In clamp
—

Test tile · slate slab

0 passes
lamp · upper left grid ≈ 5 mm
drag a specimen across the block
  • harder cuts
  • softer skids
  • repeat passes deepen the furrow
  • lens follows the cursor · S on a drawer seats it
awaiting Seat a block in the clamp, take a specimen in hand, then draw the hand across the block.

Streak plate · porcelain

bare
unglazed · powder line

Specimen dossier

no. —
—

Field notes · red marker

0 stamped

nothing stamped yet — the pen rests

Ledger · this session

0 rows
PenBlockVerdictPassesDepth
— empty —
§ 04

The scale is ordinal, whatever the posters say

bars are relative scratch resistance · log scale

3 · soft as soapstone1 000 · quartz = datum16 000 · diamond

Red tick on each bar marks where a linear reading of the scale would put it. The bars themselves are the numbers mineralogists actually measure — the gap between 9 and 10 is a cliff; the gap between 1 and 2 is a step.

Friedrich Mohs did not measure hardness. In 1812 he ranked ten stones by what scratched what, and the ranking has been copied onto a million classroom walls ever since. It is a perfectly good ordering device and a terribly misleading ruler.

Set the numbers on a log axis, as above, and the shape of the scale becomes obvious: the steps are shallow from talc to calcite, then apatite multiplies calcite sevenfold, feldspar more than doubles apatite, quartz quadruples feldspar, and diamond leaves corundum four times behind. Anyone who tells you that topaz is “four-tenths as hard as diamond” has read the classroom wall, not the stone.

The job of a kit is not to be accurate. It is to be repeatable: the same furrow, in the same direction, under the same lamp, for whoever comes after you.Bench rule 4 · this collection

The test on this page is deliberately unforgiving about that. A groove is only a groove once the pen is genuinely harder than the block; anything else is dust, and dust is recorded as dust.

§ 05

Provenance & honest caveats

what the kit is, and what it is not

The 1822 table

Mohs first published the ten in 1812; the list settled into its familiar form in his Grundriß der Mineralogie of 1822, where every verdict on this page is checked against it. Talc, gypsum, calcite, fluorite, apatite, feldspar, quartz, topaz, corundum, diamond — the same ten, the same order, two centuries of copying later.

About the powder

Every one of Mohs’ ten streaks white on unglazed porcelain; they are silicates, carbonates, sulphates and, at the end, carbon. So the streak plate here is a bookkeeping device, not a diagnostic: the powder line is tinted from the specimen’s body colour purely so you can tell whose dust is whose on a crowded plate. Real streak diagnosis belongs to hematite, limonite, chalcopyrite and their darker relations.

Two caveats

A weathered crust fakes hardness: the dusty rind of talc resists like something far above its station, and every scratch on this bench is drawn on a freshly broken face. And a crystal is not the same stone in every direction — kyanite famously measures about 4½ across the blade and 7 down it. One number per mineral is a courtesy, not a law.

Not a laboratory

Nobody at this bench owns a corundum plate or a Vickers indenter, and the red pen is a felt-tip, which is why the notes are stamped instead of signed. Diamond still cuts everything in the drawer, including the glass. It does not cut the porcelain: nothing in the kit will mark it, which is exactly the point of the plate.