Janka hardness, correctly

What the test is

A steel ball 11.28 mm across is pressed into a sample until it is embedded to exactly half its diameter, and the load required is recorded. The ball is that size because the resulting indentation has a projected area of precisely one square centimetre. Side hardness — the figure normally quoted — is measured with the load perpendicular to the grain, and it is the average of several indentations.

That is the entire test. It is a static indentation measurement, and every extension of it beyond that is somebody's inference.

Four things it does not measure

Scratch resistance. Nothing in the procedure involves dragging anything across the surface. A search of the Wood Handbook chapter that publishes these values returns no instance of "scratch", "abrasion" or "wear resistance". Flooring retailers describe the scale as measuring scratch resistance constantly.

Wear. Abrasion under traffic is a different property with different test methods. A widely distributed retailer document calls the scale the worldwide standard for gauging the ability of wood to endure normal wear and tear, which the test does not attempt.

Density. Hardness correlates with density, which is not the same as measuring it. Two woods of equal density can have measurably different side hardness.

Rot resistance. Decay resistance is an entirely separate property covered in its own chapter, driven by extractives in the heartwood rather than by hardness. Very hard woods can rot readily and some soft ones barely rot at all.

The wrong standard is cited constantly

The governing standard for solid wood is ASTM D143, Standard Test Methods for Small Clear Specimens of Timber, in which hardness is section 13. A large number of flooring pages instead cite ASTM D1037 — which is the standard for wood-base fiber and particle panel materials: particleboard, MDF, hardboard and OSB. It has nothing to do with solid hardwood flooring.

The error appears to originate on one heavily-referenced encyclopedia page and has been copied outward from there, including onto the sites of flooring manufacturers and at least one testing-machine vendor. The flooring trade's own magazine has published a correction article pointing this out, which has not slowed the spread.

Two conditions that are almost never stated

Side or end. Side hardness takes the load perpendicular to the grain; end hardness takes it on the end grain and gives a substantially different figure. Published tables almost universally mean side hardness and almost never say so.

Moisture content. The tabulated values are at 12% moisture content. Side hardness changes by roughly 2.5% to 2.75% for every one per cent change in moisture content, so a bare number carrying no moisture condition is ambiguous over a surprisingly large range. A board at 8% is meaningfully harder than the same board at 14%.

Two figures to check before quoting them

Eastern redcedar has no 12% Janka value in the Wood Handbook at all — the table gives 650 pounds-force green and leaves the 12% column blank. The 900 that circulates everywhere originates somewhere else entirely, and I could not establish where. This site publishes what the Handbook has and marks the gap.

Teak is given as 1,000 pounds-force by the Handbook, while 1,070 is the number usually quoted. And the values for the true hickories come from a different study than the rest of the table, which its own footnote records. None of these is a large error. They are a good illustration that a number repeated a thousand times still has exactly one origin, and it repays knowing which one.

What it is actually good for

Comparing woods for resistance to denting, which is a real and useful thing to know when choosing a floor or a benchtop. A 3,680 pound-force ipe floor will take a dropped pan better than a 540 pound-force poplar one, and that comparison is exactly what the test supports. What it will not tell you is which finish survives a dog's claws, how the floor looks after ten years in a hallway, or whether the boards will rot on a deck.

Asked at the bench

Is a wood twice the Janka number twice as hard?

Twice the load to make the same indentation, yes — that is what the ratio means. Whether that translates into twice the service life is a different claim entirely and one the test cannot support, because service life depends on finish, traffic, grit and maintenance far more than on the substrate.

What units is it in?

Pounds-force in North American tables, newtons elsewhere. It is a force, not a pressure, so quoting it in psi — which some retailer pages do — is a unit error. Converting is straightforward at 4.448 newtons per pound-force, though the Handbook’s own two columns disagree by about 1% from rounding, so pick one and convert consistently rather than mixing them.

Does hardness affect how the wood moves?

Not directly. Hardness and shrinkage are separate properties and a hard wood is not necessarily a stable one. Ipe is very hard and also has an excellent T/R ratio; African ebony is very hard and shrinks 10.8% tangentially. If movement is your concern, the shrinkage table is the one to read.

Where does the species data here come from?

The same USDA tables as everything else on this site, at 12% moisture content, side hardness, in pounds-force, with blanks where the Handbook has blanks. It is in the last column of the species table.