Wood Selection for Puzzles: Grain, Stability, and Movement

A mechanical wood puzzle typically runs clearances measured in a few thousandths of an inch. Sugar maple moving through a normal seasonal humidity swing can change dimension by more than that across a piece only an inch wide. That single comparison is the whole problem of wood selection for puzzles, and it is why species choice is an engineering decision rather than a matter of taste.

The three tradeoffs

You are balancing three things, and they pull against each other.

• Hardness. Harder woods hold a crisp notch shoulder and survive repeated assembly. They also blunt tools faster and are less forgiving of a slip.
• Grain tightness. Fine, even textured species cut cleanly across the grain, which is what a notch corner demands. Coarse, open pored woods tear out at exactly the places a puzzle needs to be sharp.
• Machinability. Density interacts with cutting in ways that are not intuitive. A 2024 BioResources study by Pinkowski and colleagues on chip thickness, tool wear and surface roughness in up-milling found that for low density species such as balsa and obeche, a blunt knife actually produced better surface quality than a sharp one, while for denser species such as alder and beech sharpness clearly improved the finish.

What shrinkage figures actually predict

Wood does not move equally in all directions, and the ratios matter more than the absolute numbers. Purdue Extension's publication by Carl Eckelman on the shrinking and swelling of wood and its effect on furniture sets out the pattern clearly: wood shrinks and swells most in the tangential direction, about half as much radially, and only about 0.1 to 0.2 percent along its length, which is why longitudinal movement is normally ignored in design.

The species differences are large. The same publication gives sugar maple shrinking 9.5 percent tangentially and 4.9 percent radially going from green to oven dry, against mahogany at 5.0 and 3.6 percent. Mahogany is not simply a softer wood, it is a fundamentally more stable one, and it is also much closer to isotropic in its movement.

For a practical sense of scale, Eckelman notes that a 6 percent change in moisture content can produce roughly a 2.5 percent dimensional change in a sugar maple furniture part. Scale that down to a one inch puzzle stick and you get movement far larger than the clearance the puzzle was cut to.

Quartersawn versus flatsawn

Because tangential movement is roughly twice radial movement, how the board was cut from the log changes how the finished piece behaves.

In quartersawn stock the growth rings run roughly perpendicular to the face, so the face moves radially, the smaller of the two. Quartersawn pieces stay flatter, cup less, and change width less with the seasons. In flatsawn stock the face moves tangentially, so it moves more and tends to cup away from the heart side.

For puzzle sticks, quartersawn or rift sawn stock is worth the extra cost. It also tends to give straighter, more consistent grain along the length, which matters when a notch leaves only a thin bridge of material.

What extractives do

Some species resist moisture chemically rather than just structurally. A 2017 BioResources study by Jankowska and colleagues on extractives, equilibrium moisture content and shrinkage in tropical species found that higher extractive content correlated with lower equilibrium moisture content, and reported unusually low fibre saturation points for several tropical timbers, with ipe at 18.7 percent, courbaril at 20.4 percent and tatajuba at 20.5 percent.

That is part of why dense tropical hardwoods have a reputation for holding tolerance. They take up less water at a given humidity in the first place.

Matching wood to tolerance

• Tight interlocking puzzles. Prioritise stability over looks. Quartersawn, fine grained, well acclimatised stock. Avoid species with large tangential figures.
• Packing puzzles and polycube sets. More forgiving, because gaps are designed in. Almost any stable hardwood works.
• Mixed species in one puzzle. Risky. If two species move at different rates, a fit that is perfect in June can bind in January. If you mix for contrast, pick species with similar shrinkage ratios.
• Acclimatise before cutting. Let stock sit in the shop at the humidity it will live at, for weeks rather than days, then mill it.

None of this makes wood behave like metal. It only narrows the range of the inevitable.