Wood Movement: Why Boards Swell and Shrink With the Seasons

Wood Movement at a Glance

Wood is hygroscopic. It absorbs moisture from humid summer air and releases it to dry winter air, and as that moisture moves in and out of the cell walls, the board swells and shrinks across the grain. A flatsawn 12-inch white oak board can move close to 1/4 inch between July and January in most of North America. That's why a breadboard end on your dining table feels flush in spring and proud in August, and why drawer fronts that fit in February jam in July.

You can't stop it; you can plan for it. The two facts to keep in your head: wood moves across the grain, not along it, and tangential movement (the way a flatsawn board cups) is roughly twice radial movement (the way a quartersawn board widens). Once those numbers are loaded, the rest of this guide is engineering.

In this guide:

• Why wood moves
• Tangential, radial, longitudinal
• Plain-sawn vs quartersawn
• How much a board actually moves
• Designing for movement
• Common mistakes
• FAQ

Why Wood Moves

A living tree is full of water. Mill the log, dry the lumber, and most of that water leaves. Two kinds were in there: free water in the cell cavities, and bound water locked inside the cell walls. Free water leaves first, with no change in size. Below about 30% moisture content (the fiber saturation point), the cell walls start releasing their bound water, and that's when the board starts shrinking.

Wood never stops trading water with the air. The board settles at whatever moisture content matches the surrounding humidity. That number is the equilibrium moisture content, or EMC. At 30% relative humidity and 70°F, EMC is about 6%. At 80% RH, it climbs to about 16%. A heated home in January typically holds wood at 6 to 8% MC; the same home in July, before the AC catches up, pushes wood to 10 to 12% MC. The USDA Forest Products Laboratory's Wood Handbook is the canonical engineering reference, free as a PDF; chapter 4 has the full EMC table.

A board you milled flat at 8% MC in February will be sitting at 11% MC by August. Three percentage points sounds like nothing. On a 12-inch flatsawn red oak panel it's the difference between a flush joint and a 1/8-inch gap.

The Two Directions That Matter

Wood expands and contracts in three directions, but only two of them matter in the shop.

Longitudinal (along the grain, end to end). About 0.1% across the full moisture range. A 96-inch board changes length by less than 1/16 inch over a full seasonal cycle. Build as if it doesn't move at all.

Tangential (across a flatsawn board, parallel to the growth rings). This is the big one. About 4 to 12% across the full moisture range depending on species, or roughly 0.3% per 1% MC change in the working range you actually see (6 to 12% MC).

Radial (across a quartersawn board, perpendicular to the rings). About half of tangential. Roughly 2 to 6% across the full range.

The tangential-to-radial ratio (T/R) is the reason quartersawn lumber costs more. For white oak, T/R is about 1.9; for hard maple, 1.8; for cherry, 1.9. Beech is the outlier at 2.4. For most domestic furniture species, tangential is roughly twice radial. Cut a board so its rings run vertical (quartersawn) and you cut its movement in half.

Plain-Sawn vs Quartersawn

How a board was sawn out of the log determines which kind of movement you get. There are three cut patterns, distinguished by the angle of the growth rings to the board face:

• Plainsawn (flatsawn). Rings run mostly parallel to the face. Cathedral grain pattern. Cheapest to mill (least waste); moves the most.
• Quartersawn. Rings run between 60° and 90° to the face. Straight ribbon grain. Ray fleck in oak. Costs 30 to 50% more than plainsawn.
• Riftsawn. Rings hit the face at 30° to 60°. Straight grain, no ray fleck. The least common, often a byproduct of quartersawing.

RELATED: Wood & Materials — Moisture & Movement Hub Browse every Woodwiki guide on EMC, kiln drying, dimensional stability, and seasonal movement in one place.

Why does the cut matter? Look at it in cross-section. A flatsawn board is a stack of growth rings laid sideways across the face. As humidity changes, the wood expands tangent to those rings, which is across the width. The face cups, the edges curl, and the panel grows visibly. A quartersawn board has rings running through the thickness instead. Now expansion is radial relative to the rings, still across the width, but radial movement is half of tangential. Same direction, half the swing.

Riftsawn falls in the middle. So does any flatsawn board cut near the center of a wide tree, where the rings are flatter relative to the face. In a real lumber stack you usually find a mix of cuts on the same board.

How Much It Actually Moves

The Wood Handbook publishes shrinkage coefficients for every common species. The number you want is the dimensional change coefficient (Cr for radial, Ct for tangential), which gives the fractional change per 1% MC change in the working range. To use it, multiply: coefficient × board width in inches × expected MC swing in percentage points.

Numbers for the species you're most likely to build with:

Walnut and mahogany move noticeably less than the oaks and maple. Mahogany is famously stable; its T/R ratio near 1.0 means cup is barely a thing, which is why eighteenth-century cabinetmakers chose it for veneered carcase work and museum cases.

Worked example. You're building a dining table with a 36-inch white oak flatsawn top. Annual MC swing in your shop is 6% (8% in February, 14% on a humid August day before the AC kicks in). Tangential coefficient is 0.00365.

36 in × 6 × 0.00365 = 0.79 inches

Your tabletop's width swings by 13/16 of an inch over the year. That has to go somewhere. Bolt it rigidly to the apron and the top will crack, usually at the screws. Use elongated slots or figure-8 fasteners and the top floats while the apron stays put.

For the same 36-inch top in black walnut, 0.59 inches. In quartersawn white oak, 0.39 inches. Same principle, different gap to plan for.

Designing With Movement

Every joint that crosses the grain either lets the wood move or fails. There is no third option. Three hundred years of shop tradition is built around a small set of answers.

Tabletops and wide panels. Never glue the top to the apron. Use figure-8 fasteners (about $0.50 each at Rockler), wooden buttons in slotted apron mortises, or metal Z-clips in a router-cut groove. Plan slots for 3/8 inch of travel on a 36-inch top, split front and back with the screw centered.

Breadboard ends. A breadboard caps the end grain of a wide top with a narrow cross-grain board. The center tenon is glued; the outer tenons sit in elongated mortises with their pins through slotted holes. Center stays put, ends slide. The breadboard sits slightly proud or shy depending on the month, and that's fine.

Frame and panel. The panel sits in a groove with no glue and no fasteners. It floats. A bead of silicone in the corners stops rattles, but the wood itself stays unrestrained.

Drawers and solid-wood doors. Size the drawer front to the opening at worst-case (mid-summer) humidity, and use joinery that allows movement: sliding dovetails or pinned tenons. Solid-wood doors are the same problem. Build frame-and-panel, or hang a single-board door with 1/8 inch top-and-bottom clearance.

The Wood Handbook heuristic: 1/4 inch of movement per 12 inches of width on a flatsawn domestic hardwood across a typical year. Build accordingly.

Common Mistakes

Five errors show up in every beginner's first failed top, panel, or door.

Gluing across the grain to a rigid substrate. A solid-wood top glued flat to a plywood case cracks within a year. Plywood doesn't move; solid wood does. Something has to give, and it's always the solid wood.

Treating wood like steel. The "clamp it tight" instinct treats movement as a problem to muscle. A 36-inch oak panel pushing 0.8 inches of expansion will split the wood, strip the screws, or pull the apron apart. The wood always wins.

Gluing up in the wrong season. Tight breadboard end glued in February with the wide board at 7% MC, then August arrives at 12% MC and the center board pushes the outer pins into the mortise walls. Crack. The reverse (assembling in summer, watching the joint open in winter) is just as common.

Mixing species without checking the math. Glue a maple top to a walnut frame: maple moves about 30% more than walnut for the same humidity swing. Pick species with similar coefficients, or design joinery that lets them move independently.

Skipping acclimation. Your supplier keeps cherry at 7% MC; your shop runs at 11% MC. Mill on Saturday, glue up Sunday, and the panel moves overnight. Sticker the lumber in your shop for at least a week before milling.

FAQ

Does the species really matter that much?

Yes. Hard maple has more than double the tangential coefficient of Honduran mahogany. On a 36-inch top across a 6% MC swing, maple moves 0.76 inches; mahogany moves 0.37 inches. Twice the gap to design around. If you're building something where movement is hard to engineer around (a bookmatched panel, a wide unbreadboarded top, an inset door), switch to a calmer species before you reach for fancier joinery.

Can I just dry the wood more before I work it?

No, and this is the most common beginner mistake. You can dry a board to 4% MC, but the moment it sits in your shop it starts trading water with the air. Within a few weeks it will be back at your shop's EMC. The right number isn't "dry," it's "stable at your shop's average humidity." Sticker the lumber, take a moisture-meter reading, repeat a week later, and mill once two readings a week apart match within 0.5%.

Does finish stop wood movement?

Not really. A film finish like polyurethane slows moisture exchange but doesn't stop it. Studies cited in the Wood Handbook show that even a heavy finish on all six surfaces reduces seasonal movement by about 30 to 60%, never to zero. Oil finishes barely slow it at all. Finish evens out short-term humidity spikes; it does not lock dimensions.

How long should I let lumber acclimate?

At least one week for kiln-dried hardwoods stickered in your shop, two is better. Rough lumber from an air-dried source can need a month. The right test is the moisture meter, not the calendar. Take a reading on day one and day seven. If the number moved more than half a percent, the wood is still adjusting. Wait for two readings a week apart to match within 0.5% MC, then mill.

Sources

These sources informed the wood movement coefficients, EMC tables, and design rules in this guide.

• USDA Forest Products Laboratory — Wood Handbook: Wood as an Engineering Material (2010 edition) — the canonical engineering reference. Chapters 4 (moisture relations) and 13 (drying and dimensional stability) provide every shrinkage coefficient and EMC table cited above. Free PDF.
• Forest Products Laboratory — Equilibrium Moisture Content of Wood for U.S. Cities — regional EMC data. Use it to estimate the seasonal MC swing in your shop.
• Hoadley, R. Bruce. Understanding Wood: A Craftsman's Guide to Wood Technology (Taunton Press, 2nd ed.) — the classic shop reference. Translates the Wood Handbook's engineering tables into language for woodworkers.
• Bruce Hoadley's species shrinkage tables, reproduced in the Fine Woodworking archive — practical worked examples for tabletop, panel, and door construction.