At a glance
| Species | Side hardness (lbf, 12% moisture) | Tangential shrinkage (%) | Radial shrinkage (%) | Plain-sawn movement, 150 mm board, 4 point swing (mm) |
|---|---|---|---|---|
| Iroko | 1,260 | 3.8 | 2.8 | 0.8 |
| African mahogany | 830 | 4.5 | 2.5 | 0.9 |
| Teak | 1,000 | 5.8 | 2.5 | 1.2 |
| Afrormosia | 1,560 | 6.4 | 3.0 | 1.3 |
| Sapele | 1,510 | 7.4 | 4.6 | 1.5 |
| Black cherry | 950 | 7.1 | 3.7 | 1.4 |
| Black walnut | 1,010 | 7.8 | 5.5 | 1.6 |
| White ash | 1,320 | 7.8 | 4.9 | 1.6 |
| Northern red oak | 1,290 | 8.6 | 4.0 | 1.7 |
| Sugar maple | 1,450 | 9.9 | 4.8 | 2.0 |
| White oak | 1,360 | 10.5 | 5.6 | 2.1 |
| American beech | 1,300 | 11.9 | 5.5 | 2.4 |
Hardness from Wood Handbook (FPL-GTR-190, 2010) Tables 5-3b and 5-5b; shrinkage, green to oven-dry, from Tables 4-3 and 4-4. The movement column is our arithmetic (section 2), a ranking aid and not a design figure. 1,000 lbf is about 4,450 N. US-grown species and imports; European oak, ash and beech are not listed.
What hardness tells you, and what it does not
The Wood Handbook defines hardness as resistance to indentation in a modified Janka test: the load needed to push an 11.28 mm (0.444 in.) ball to half its diameter into the wood, averaged over the radial and tangential faces. The tables call it side hardness and give it in pounds-force (lbf). It is the number behind most “Janka charts”.
On the chart the twelve timbers run from African mahogany at 830 lbf to afrormosia at 1,560 lbf. Ipe sits far off the scale at 3,680 lbf. Read the number for what it is:
- It is dent resistance. It does not measure abrasion, scratching or how much a scratch shows. Those depend on the finish, the grit on the floor and the colour of the timber.
- It is a clear-wood average. The handbook tests small clear specimens at 12% moisture content. A real board has knots, sapwood and variation. For imported species the handbook says its values come from a more limited number of specimens.
- Botanical names. Iroko, Milicia (Chlorophora) spp.; African mahogany, Khaya spp.; teak, Tectona grandis; afrormosia, Pericopsis elata; sapele, Entandrophragma cylindricum; black cherry, Prunus serotina; black walnut, Juglans nigra; white ash, Fraxinus americana; northern red oak, Quercus rubra; sugar maple, Acer saccharum; white oak, Quercus alba; American beech, Fagus grandifolia.
- A European species is not listed. European oak is the commonest specification in the UK and is not in the tables. White oak is its nearest relative in them, not the same timber. Ask the maker for hardness on the species supplied, and for the test method.
- The European test is different. BS EN 1534 sets a Brinell-type indentation test for wood flooring. Its figures are not the same as Janka lbf and should not be converted by a rule of thumb. We did not read the standard.
Movement: turn the shrinkage figure into millimetres
Shrinkage in the tables runs from green wood to oven-dry, which no floor ever reaches. The handbook gives a way to scale it to the small moisture swings of a room: movement is the shrinkage figure times the change in moisture content divided by the fibre saturation point, which averages about 30%. Use the tangential figure for plain-sawn boards and the radial figure for quarter-sawn boards. The handbook says to use tangential values where the cut is mixed or unknown, and that individual pieces vary from the prediction (a coefficient of variation of about 15%).
Width × shrinkage % ÷ 100 × swing in points ÷ 30. For a 150 mm board and a 4 point swing in moisture content:
| Species | Plain-sawn: 150 mm board, 4 point swing | Quarter-sawn: same board |
|---|---|---|
| Iroko | 0.8 mm | 0.6 mm |
| Teak | 1.2 mm | 0.5 mm |
| Black walnut | 1.6 mm | 1.1 mm |
| White oak | 2.1 mm | 1.1 mm |
| American beech | 2.4 mm | 1.1 mm |
Two things stand out. Species matters: American beech moves more than three times as much as iroko in the plain-sawn case. Cut matters as well: quarter-sawn faces move between about a quarter and a half less than plain-sawn for these species, which is why they are chosen for wide boards and for floors over underfloor heating. The swing itself, the gap budget and the fixing method are covered in seasonal movement in timber floors and underfloor heating under wood and stone.
Matching the species to the job
The table sets out which number to look at first for four common briefs, and which species in this source sit at the favourable end. It is a reading of the figures, not a recommendation: the maker's own data for the board supplied overrides it.
| The job | Number to look at first | On the figures read, the species at the favourable end | What the figures cannot tell you |
|---|---|---|---|
| Busy hotel or restaurant floor | Side hardness, then finish and board width | Afrormosia 1,560, sapele 1,510, sugar maple 1,450, white oak 1,360 lbf | Wear and scratching depend on the finish and the grit carried in; hardness is dent resistance only. |
| Floor over underfloor heating | Tangential shrinkage, then construction | Iroko 3.8%, African mahogany 4.5%, teak 5.8% | Heat dries the board; board width, cut and construction matter as much as species. See the UFH sheet. |
| Wide plank, large room | Tangential shrinkage; cut | Iroko, teak, then the middle group; quarter-sawn cuts the figure by a quarter to a half | A wide board moves in proportion to its width; the gap budget matters more than the species. |
| Pale or dark scheme | A finished sample, aged | None: no colour data was read | Colour at delivery, and how it moves in light, must be seen on a sample. |
For grade, see wood flooring grades; for the wear layer on an engineered board, solid versus engineered wood flooring; for a hospitality checklist, specifying timber floors for high-traffic interiors; for joinery timbers by the same method, timber species for joinery.
Colour change: test it, do not guess it
Timber changes colour in light, and finishes change it too: oil deepens the tone, and the finish and timber should be approved together (see timber floor finishes compared). The direction and the speed differ with species, finish and the light on the floor. We found no authoritative table of colour change by species, so this sheet gives none. What it gives is a way to see the answer on the real sample.
- Get a finished sample from the batch, finished the way the floor will be.
- Cover a strip with opaque card and leave the board where the floor will sit, in daylight, for [4 to 8] weeks.
- Lift the card and compare. The step at the edge is the change you can expect.
- Keep a sealed offcut as the reference, and ask for spare boards from the same batch for repairs.
Smoked, fumed and stained colours are made by different routes (a reaction in the timber, or a pigment on it), and they may not age alike. Ask the maker how the colour is made and how it behaves in light; we did not read anything on this. Window film, curtains and rugs change the light on the floor and so the outcome.
Legality and the tropical timbers
Several of the twelve are tropical (iroko, African mahogany, teak, afrormosia, sapele). Whoever first places timber on the UK market has legal duties to check that it was harvested legally, and some species are controlled under CITES. See UK Timber Regulation due diligence and specifying FSC and PEFC, and ask for the species, country of harvest and any certificate in writing before you fix the specification.
Specification wording
Paste into your schedule or specification and replace anything in [square brackets] with your own figures. Wording is a starting point: your specification governs.
Species [European oak, Quercus robur or Q. petraea], [quarter-sawn / rift / plain-sawn], grade [ ] to [ ]; board [ ] mm wide, [ ] mm overall, [ ] mm wear layer above the tongue and groove. Common and botanical name stated on the order and the delivery note.
The Supplier states the hardness of the species supplied, with the test method [BS EN 1534 / Janka], the moisture content at test and the source of the figure, and confirms in writing that the floor is suitable for [hotel restaurant] use at [location].
Board width not more than [ ] mm. The Supplier states the expected movement for the boards supplied across a moisture content swing of [ ] points and the expansion gap this requires at [perimeter / fixed points]. Installed to the maker's written method for [bonded / floating / nailed] fixing over [underfloor heating].
Colour to match the approved sample from the delivery batch, finished by the proposed method. A covered-strip light test sample, [ ] mm, to be supplied for approval and kept sealed for reference; spare boards of [ ] m² from the same batch.
For each species the Supplier states the common and scientific names and the country of harvest, and provides [FSC / PEFC / other] certificate details or due diligence records on request. No species listed under CITES without a permit.
Terms used on this page
- Janka hardness
- A measure of the force needed to press a steel ball into a timber surface; used to compare species for denting, although Brinell (BS EN 1534) is the European method.
- Flat-sawn
- Timber cut tangentially to the growth rings, giving broad cathedral-shaped figure and, for a given width, more movement across the board than quarter-sawn timber.
- Quarter-sawn
- Timber cut so the growth rings meet the face at roughly 60-90 degrees, producing straighter grain, ray fleck in oak, and generally better dimensional stability.
- Fumed oak
- Oak darkened by exposure to ammonia vapour or by a reactive treatment that reacts with the tannins, so the colour is within the wood rather than a surface stain.
- Tangential shrinkage
- Shrinkage across the growth rings, along the face of a plain-sawn board, from green to oven-dry. It is larger than radial shrinkage, which runs through the thickness of a plain-sawn board.
- Fibre saturation point
- The moisture content, about 30% on average, below which wood starts to shrink as it dries.
Standards and sources
- Wood Handbook, Wood as an Engineering Material (FPL-GTR-190, 2010), chapters 4 and 5: Tables 4-3, 4-4, 5-3b and 5-5b. US Forest Products Laboratory.
- BS EN 1534: Wood flooring and parquet: indentation (Brinell-type) test method. A search summary listed the 2020 edition as current; not read. Confirm against the current edition.
Questions we are asked
What is the hardest timber for a floor?
Among the twelve in our table, afrormosia (1,560 lbf), sapele (1,510) and sugar maple (1,450). Ipe is far harder at 3,680 lbf in the same source. Hardness is dent resistance only; it does not decide wear, scratching or stability.
Is oak or walnut better for a floor?
On side hardness, white oak (1,360 lbf) and northern red oak (1,290) are harder than black walnut (1,010), and walnut moves less than oak (7.8% against 8.6 to 10.5% tangential). Neither is a European species: ask for data on the timber supplied. Choose on hardness, movement, colour and finish together.
Which species moves least?
Of the twelve, iroko (3.8% tangential shrinkage), African mahogany (4.5%) and teak (5.8%). Quarter-sawn boards move roughly a quarter to a half less than plain-sawn for these species.
Does hardness matter for underfloor heating?
Less than movement. Heat dries the board, so low shrinkage, narrower boards, quarter-sawn faces and a construction approved for heating matter most. See the underfloor heating sheet.
Will a wood floor change colour?
Yes, in light and with the finish, but the direction and speed depend on the species and finish and we found no authoritative table. Run a covered-strip light test on a finished sample and keep a sealed reference.
Can I convert Janka lbf to a Brinell figure?
Not safely. They are different tests. Ask the maker for the figure by the method the specification names.
Download the factsheet
A two-page A4 PDF with the species table, the hardness and movement chart, the light test and specification wording, to keep with your project file.