How wood moves

Wood is hygroscopic, permanently

Wood exchanges moisture with the air around it and never stops. Put a board somewhere and it drifts toward the moisture content that air will support; move it and it drifts somewhere else. This is not a defect of poorly dried timber. It is what wood is, and kiln drying does not cure it — drying sets the starting point, not the behaviour.

Only water in the cell walls changes the size

Water sits in wood in two places: inside the hollow cell cavities, and bound within the cell walls themselves. Water in the cavities is just cargo — it adds weight and nothing else. Water bound in the cell walls physically holds them apart, and losing it is what makes the wood shrink.

Once the cavities have drained but the walls remain saturated you are at the fibre saturation point, and dimensional change starts there. Above it, drying a board changes its weight and not its size. Below it, every point of moisture lost takes some width with it.

That threshold is conventionally 30%, and the Wood Handbook uses 30% for most species while assigning 22% to eleven of them. It also notes the value varies by several points between individual pieces and depends on how you measure it. A calculator built on 30% as a universal constant is wrong by more than a third for teak, redwood and western redcedar.

Moisture content is measured against the dry weight

The convention is water weight divided by ovendry wood weight. Not divided by total weight, which is the intuitive reading and gives a different number. A board described as 20% has water equal to a fifth of its dry substance, meaning water is one sixth of what you are lifting, not one fifth.

Because the denominator excludes the water, the figure has no ceiling. Green western redcedar sapwood averages 249%. The theoretical maximum runs from 267% for light woods down to 44% for the densest, since dense wood has less cavity space to fill.

It shrinks about twice as much one way as the other

Wood shrinks very little along its length — a tenth to a fifth of one per cent, ignorable in practice. Across the grain it shrinks a great deal, and unequally: tangentially, along the growth rings, roughly twice as much as radially, across them.

The ratio between those two is what makes wood interesting rather than merely shrinky. If shrinkage were equal in both directions, boards would get smaller and stay flat. Because it is not, a board's shape depends on how the rings run through it.

Which is why boards cup away from the heart

Take a flatsawn board, where the rings arc across the wide face. The face that lay nearer the bark has rings running more nearly parallel to it, so it shrinks by something close to the tangential figure. The face nearer the pith has rings meeting it at a steeper angle, so it shrinks by something closer to the radial figure. One face loses more width than the other, the board bends, and it bends away from where the pith was.

Turn that log ninety degrees before sawing and the rings run through the thickness. Both faces now shrink identically, nothing drives a bend, and you have quartersawn stock. Its width is governed by the smaller radial figure, and its thickness by the larger tangential one — the movement has been moved, not removed.

The wood is always behind the weather

Equilibrium is where a board is heading, not where it is. Moisture travels through wood by diffusion, which is slow and gets slower with thickness — squaring the distance roughly squares the time. So a thick board never reaches either seasonal extreme; it lags, and averages. A thin one tracks the room closely. This is why panels split in the first hard winter of central heating rather than the first cold week, and why a board brought in from a damp store feels stable for a month before it starts to argue.

Finish slows it and cannot stop it

Coatings retard moisture exchange but do not prevent it. The Forest Products Laboratory tested 91 finishes and found none that entirely prevented uptake; only eleven of eighty-seven commercial finishes showed much success even over a fortnight. Paraffin wax and two-part epoxy performed best; paste wax and linseed oil scored essentially zero.

The practical reading is that a film finish damps the seasonal peaks, which is worth having, and that it must go on both faces equally. Seal one side and the other exchanges faster, which produces exactly the imbalance that makes a board cup.

Asked at the bench

Does old wood stop moving?

No. Antique furniture moves with the seasons exactly as new furniture does, which is why old panels are set in grooves rather than glued and why long-case clock doors stick in August. What ages out is the initial settling of stock that was never dry to begin with, not the seasonal cycle.

Does kiln drying make it more stable afterwards?

It sets the moisture content and kills anything living in the wood, both worth having. It does not change the shrinkage coefficients. The first drying below fibre saturation does permanently shift the wood’s sorption behaviour somewhat, but the amount it moves per point of moisture change afterwards is a property of the species.

Why do some boards twist instead of cupping?

Usually reaction wood or juvenile wood near the pith, where longitudinal shrinkage can reach 2% instead of the normal 0.1% to 0.2%. When one edge of a board shortens ten times as much as the other along its length, it winds. This is why boards cut close to the centre of a log are the troublesome ones.

Is any of this avoidable?

Not avoidable, only accommodated — which is what every traditional joint is doing. Frame and panel, breadboard ends, tabletop buttons, drawer bottoms in grooves: all of them are ways of holding wood firmly while letting it change size. The tradition long predates anyone measuring shrinkage coefficients.