Why wood cracks, cups and twists when it dries
Every drying defect starts with shrinkage. Wood does not shrink while it loses free water from its cell cavities; it starts to shrink when the cell walls dry below the fibre saturation point, at about 30% moisture content (MC). It then shrinks about twice as much along the growth rings (tangentially) as across them (radially), and hardly at all along the grain.4 Those differences, between directions and between the outside and inside of a board, are what crack and distort it.
Shell and core: the stress reversal behind most defects
When drying starts, the outer layer of a board, the shell, dries below the fibre saturation point within hours and tries to shrink. The wet core does not shrink yet and holds it back. The shell is therefore in tension and the core in compression. In 50 mm oak on a standard schedule, the tension in the shell peaks after about five days.2
If the shell is pulled beyond its elastic limit, it dries in a permanently stretched state, called tension set. If the pull exceeds the strength of the wood, it tears: a check. Later, when the core dries and tries to shrink, the oversized shell stops it. The stresses reverse: the shell goes into compression and the core into tension. That late core tension can open internal cracks.1,2 The FPL compares tension set to a rubber band that has been held stretched too long and no longer returns to its original size.2
The four stages of drying and their defects
The FPL's hardwood drying manual divides drying into four stages, each with its own typical defects. The stages are a teaching device, not sharp boundaries, but they explain when each problem is created and when it merely becomes visible.2
| Stage | Moisture content | Main defect risk |
|---|---|---|
| I | Green to two-thirds of green | Surface and end checks, stain, warp |
| II | Two-thirds of green to 30% | Existing surface and end checks get worse |
| III | 30% to final | Checks turn into honeycomb; cupping; overdrying |
| IV | Final (equalising and conditioning) | Uneven final moisture content; case-hardening |
Source: Denig, Wengert & Simpson (2000), Drying Hardwood Lumber, USDA FPL-GTR-118, Table 2.4 (after Bois 1998).
The practical lesson is that the most important hours of a kiln charge are the first ones. Checks are created only in Stage I; later stages can only make them worse. For red oak, that critical stage runs from green to about 50% MC, for white oak from green to about 45%.2 How kiln schedules handle this is covered in our article on kiln drying versus air drying.
Checks, splits and honeycomb
Checks are cracks that open during drying. Where they appear tells you when they were made and why.
Surface checks
Surface checks open on the faces of boards, nearly always on flat-sawn faces, because the junction between the wood rays and the long fibres is weak and tangential shrinkage forces are greatest there. Oak, with its large rays, is especially prone.2 Many surface checks close again later in drying and are invisible on rough dry lumber; they reappear only after planing. If a check contains dirt or stain, it was open during air drying.2
The cause is always drying too fast at the start: air that is too dry, moves too fast, or both. Only a few hours of overly harsh conditions are enough.2 The FPL's limits for the first stage of kiln drying are conservative:
- Initial dry-bulb temperature no higher than 43 °C, and 38–41 °C for critical species or thicknesses.
- Air speed no higher than 1.9 m/s in the early stages.
- Relative humidity not below 87% during Stage I, unless the air speed is well below 1.9 m/s.
- At least ten properly prepared kiln samples per load, with the daily drying rate checked against the species limit.2
The yard and the saw matter too. Direct sun and strong wind in hot or dry weather can check 25 mm heartwood of oak and beech, so check-prone lumber should be under a roof if it is exposed for an hour or more.2 A dull saw tears the surface and makes checking more likely: in one FPL experiment, red oak planed before drying had 18 times less checking than rough-sawn boards.2
End checks and splits, and why end coating pays
Moisture moves along the grain 10–15 times faster than across it, so the ends of a board or log dry first and crack first.1 Severe end checks grow into end splits, and the honeycomb they cause can reach 0.6 m into a board.2 The most effective defence is a water-resistant end coating, applied to the fresh ends of logs or freshly trimmed boards.1,2
Show the data
| Storage time | Uncoated end | Coated end |
|---|---|---|
| Week 2 | 11 mm | 3 mm |
| Week 3 | 16 mm | 0 mm |
| Week 4 | 22 mm | 4 mm |
| Week 5 | 35 mm | 0 mm |
| Week 7 | 73 mm | 0 mm |
| Week 9 | 101 mm | 0 mm |
| Week 12 | 156 mm | 0 mm |
Coating cost was below US$3 per thousand board feet; savings reached US$37 by week 12 (1990s dollars). Source: Denig, Wengert & Simpson (2000), FPL-GTR-118, Table 3.1 (Linares-Hernandez & Wengert 1997). American red oak.
The FPL data show both how well coating works and how quickly its benefit is lost. In a Wisconsin summer, split length on the uncoated ends of red oak logs grew from about 11 mm after two weeks to about 156 mm after twelve, while the coated ends stayed practically free of splits.2 On air-dried 25–32 mm (5/4) red oak, coating fresh ends left 62% of the ends with checks no longer than 25 mm; waiting three days before coating cut that to 22%, and waiting ten days to 16%.2
Other measures help: trim pile ends even, place a row of stickers at or near the board ends, and in the kiln butt the piles end to end so that air does not stream past the ends.2
Honeycomb: internal checks found at the saw
Honeycomb is a crack inside the board, running across the growth rings. It forms late in drying, when the core is in tension, typically if the core is still fairly wet while temperatures are raised too high for too long.1 Spontaneous internal checking is uncommon: most honeycomb is a surface or end check that has deepened.2 Honeycomb rarely shows on the surface. It is usually found only when the wood is cross-cut, ripped, heavily planed or routed, which makes it one of the most expensive defects for a parts maker.1,2
The FPL's rules against it are to prevent surface and end checks in the first place, to dry already-checked wood gently at high humidity and moderate temperature, and to be cautious about raising the temperature on thick stock while the core is above about 28% MC.2
Collapse, shake and ring failure
Collapse is the flattening of wood cells. Severe collapse shows as corrugations, a washboard surface; mild collapse just looks like excessive thickness shrinkage. It occurs when very wet wood is dried fast, and high temperatures encourage it.1,2 It is rare in most hardwoods, but bacterially infected wood, called wetwood, is prone to it: for several American oaks the FPL lists collapse, ring failure and honeycomb as typical wetwood defects.2
Shake or ring failure is a separation along the growth rings. It is almost always linked to bacterial weakening in the standing tree, or sometimes to strong wind, and drying stresses are seldom large enough to cause it. Little can be done about it in the kiln, apart from sorting out affected logs.2
Case-hardening and the prong test
Case-hardening is the tension set described above, still locked in the wood at the end of drying. The board looks perfect, but the shell and core are fighting each other. As soon as the balance is disturbed by resawing or ripping, the pieces bow, or the kerf closes and pinches the saw. The FPL also links residual stress to planer splits and poor gluing.2
It is relieved by conditioning at the end of the kiln run: briefly raising the equilibrium moisture content of the kiln air to about the target MC plus 4 points, at around 82 °C, so the shell takes back moisture and relaxes.2 Our kiln drying article describes equalising and conditioning step by step.
The check is the prong test. A thin cross-section is cut from a board and sawn into prongs: three prongs, with the middle one removed, for boards under about 4 cm thick; six prongs, with the second and fifth removed, for thicker stock. If the outer prongs bend inwards, tension set remains; with moderate to severe stress they touch.2 Timing matters. A moisture gradient in the prongs can fake or hide stress, so the FPL advises letting the lumber cool for a day or two before testing, or heating the cut prongs in a microwave oven for 15 seconds to even out their moisture.2
Warp: bow, crook, twist and cup
Warp is any deviation of a face or edge from flat or square. The Wood Handbook lists six types: bow, crook, twist, cup, oval and diamond, the last two being distortions of square and round sections.1 Warp comes from two sources, the differences between radial, tangential and longitudinal shrinkage, and growth stresses in the tree, and is made worse by irregular grain, juvenile wood and reaction wood.1
What causes each type of warp
- Bow, a lengthwise curve like a ski, often comes from sawing that is not parallel to the bark, and from poor stacking: stickers out of line or too far apart, uneven stickers or uneven foundations.2
- Crook, a sideways bend along the edge, follows off-centre growth rings or crooked logs: one edge shrinks more lengthwise than the other.2
- Twist, the corners of a face no longer in one plane, appears in wood with spiral, wavy, diagonal or interlocked grain.2
- Cup, a curve across the width, is the natural tendency of flat-sawn boards. The bark side always shrinks more than the heart side, and the effect grows the closer the board is cut to the pith.2
Most warp, the FPL notes, comes from characteristics of the lumber rather than errors in drying. Anything that rewets the surface during drying makes it worse: steam spray on cold lumber at start-up, fan reversal cycles longer than two hours, or schedules that are too humid for partly dried wood.2
Why flat-sawn oak and beech cup: the tangential/radial ratio
The ratio between tangential and radial movement is a good guide to how strongly a species cups. For European species, the Bavarian State Institute of Forestry gives about 2.2 for oak and 2.1 for beech, against 1.8 for ash and about 1.6 for lime.3,5
This is why quarter-sawn boards, whose wide faces lie across the rings, cup, surface-check and split less in drying and in use.2 The choice is made at the saw, at a cost in yield, as explained in our article on plain, quarter and rift sawn timber.
Stacking that keeps boards flat
Stickers and bolsters work together to hold lumber flat while it dries.2 The FPL's guidance:
- Sticker thickness. Usually 19 mm, and above all uniform: uneven stickers cause warp. Each extra 3 mm of sticker thickness reduces kiln capacity by about 7%.
- Sticker spacing. Every 30–46 cm for upper grades of 25–38 mm (4/4–6/4) hardwood, about 30 cm for lower grades and warp-prone species, and 60 cm is usually acceptable for stock thicker than 38 mm. Add stickers within 30 cm of the pile ends when lengths are uneven.
- Alignment. Stickers vertically in line, with a bolster under every sticker column and level foundations.
- Weight. About 7 kPa on top of the load, equal to some 25 cm of concrete, controls some warp, although weights are seldom used in hardwood drying.2
Discolouration: stains that appear after planing
Colour defects matter most on light woods and on oak under clear finishes. Some are chemical, some fungal, and several stay hidden until the surface is planed.1,2
Iron tannate stain in oak
Oak is rich in tannins. Tannin, water and iron react to form iron tannate, which is blue-black. The worst source is condensate dripping from rusty kiln fittings or from steam lines that do not drain; with a constant supply of iron and water, the stain can penetrate more than 13 mm into the lumber. Forklift forks, metal rollers, steel strapping and saw filings also stain, but seldom deeper than 0.4 mm.2 Prevention means removing the iron and sloping steam lines so that water drains away. Oxalic acid removes the stain from green or dry wood.2
Sticker stain and grey chemical stain
Sticker stain, also called sticker shadow, is a band of darker or lighter wood across the board where a sticker lay. It usually shows only after planing.2 Despite the name, fungi are seldom involved and the stickers do not cause it: the FPL attributes it to oxidation of natural chemicals in the wood, started by slow drying, probably below 5% MC a day, at 10–54 °C while the wood is wetter than 40% MC.2 Closely related chemical stains include interior greying, noted especially in maple, hackberry, southern oaks and ash, and brown "coffee" stain, mainly in white oak.2 A deep greyish-brown discolouration can occur in many hardwoods if initial drying is too slow or the first kiln temperature too high.1
The cure is speed at the start. The FPL's checklist includes using logs stored less than two weeks in warm weather, stacking lumber within 12 hours of sawing, using dry stickers at 8–10% MC, grooved stickers, keeping stacked lumber out of the rain, high air speed for light-coloured woods, and kiln temperatures no higher than 71 °C except during equalising and conditioning.2
Blue stain and mould
Blue stain is a fungus that grows in wet sapwood when drying is too slow, especially in summer; its colour is the colour of the fungus. It does not weaken the wood, but it ruins clear-finished surfaces. Moulds live on dirt and dust rather than on the wood. Both are controlled by drying fast enough. Temperatures above 54 °C also stop the fungi, but that much heat is too severe for many species while they are still wet.1,2
What drying defects cost
The FPL calls drying defects often the most costly item in drying hardwood lumber.2 The most often quoted figures are old but telling: in a Pennsylvania study of 25–32 mm No. 1 Common lumber air dried in typical open yards, value losses were 6% for yellow birch, 8% for basswood, 13% for red oak, 14% for hard maple and 15% for white oak, not counting shrinkage.2 We have not found a comparable modern figure for European species.
Losses do not end at the kiln. The FPL estimates that perhaps 90% of machining and gluing defects, from fuzzy grain to open glue joints, can be eliminated by proper drying, and it sets the target for final moisture content at the equilibrium moisture content of the place of use, plus or minus 2 points.2 Our article on wood moisture content and EMC explains how to set that target.
Drying defects at a glance: cause and prevention
| Defect | Main cause | Prevention |
|---|---|---|
| Surface checks | Shell dries too fast early on: RH too low, air too fast | Mild start (≤43 °C, ≥87% RH in Stage I); sharp saw; shade green stock |
| End checks and splits | Ends dry 10–15× faster along the grain | End coating at once; stickers at board ends; piles end to end |
| Honeycomb | Surface or end checks deepening; core wet while temperature is raised | Prevent surface checks; caution with thick stock while core >28% MC |
| Collapse | Very wet wood dried fast or hot; wetwood | Low initial temperature; sort out wetwood |
| Shake, ring failure | Bacterial weakening or wind in the standing tree | Little control in drying; sort logs |
| Case-hardening | Tension set left at the end of drying | Equalise, then condition at target MC + 4; prong test |
| Bow, crook, twist | Grain, growth stresses, reaction wood, poor stacking | Saw parallel to the bark; uniform 19 mm stickers in line |
| Cup | Tangential shrinkage about twice radial | Quarter-sawing; avoid rewetting during drying |
| Iron tannate stain (oak) | Iron + tannin + water | No rusty drips or steel contact; drained steam lines |
| Sticker stain, grey stain | Slow initial drying of wet wood | Fresh logs; stack within 12 h; dry stickers; good airflow |
| Blue stain, mould | Fungi in wet sapwood when drying is slow | Fast initial drying; clean equipment |
Source: Compiled by SARAL from Bergman (2021), Wood Handbook FPL-GTR-282, chapter 13, and Denig, Wengert & Simpson (2000), FPL-GTR-118, chapters 3, 4, 8 and 9 and Table 9.1.
How we prevent drying defects at SARAL
How to inspect a delivery of kiln-dried hardwood
- Look at the ends. End checks and splits are visible at once. Long splits that start at the ends often hide honeycomb further in.
- Plane a few boards. Surface checks, sticker stain and grey chemical stain often show only after planing. Planing a sample from the middle of the packs takes minutes.
- Cross-cut a sample board. A cut well away from the end reveals honeycomb and gives you a section for a prong test.
- Do a prong test if you will resaw or rip the wood. Let the section stabilise before you read it.
- Measure moisture content on at least 10% of the pieces and compare it with the agreed target and tolerance.
- Check flatness of a few boards for bow, crook, twist and cup, and look for dark marks where steel strapping or forks touched oak.
- Ask for the drying records: schedule, final moisture content of the kiln samples, and whether the charge was equalised and conditioned.
Frequently asked questions
Why does wood crack when it dries?
Because the surface dries and shrinks before the inside. The wet core restrains the shrinking surface, which goes into tension; if the tension exceeds the strength of the wood across the grain, it cracks. Ends crack first because moisture escapes along the grain 10–15 times faster than across it.1
What is the difference between checking and cracking or splitting?
A check is a separation along the grain that forms during drying, on the surface, at the end or inside the board. A split runs through the piece from one face to the other, typically from an end. Severe end checks become end splits; deep surface checks can become internal honeycomb.2
How do you dry oak without cracks?
Protect it from sun and wind from the moment it is sawn, coat the ends immediately, stack it on uniform stickers, and dry it slowly at first: on the FPL's guidance, a kiln start no hotter than 43 °C at 87% relative humidity or more, with modest air speed, until the wood has lost a third of its moisture.2 Quarter-sawn oak checks less than flat-sawn.
Which wood cracks the least when drying?
Species with modest shrinkage, a low ratio between tangential and radial shrinkage and no large rays are the most forgiving. Among European hardwoods, lime (about 1.6) and ash (1.8) are gentler than oak (2.2) and beech (2.1).3,5 Within any species, quarter-sawn boards check and cup less than flat-sawn ones.2
