Treebeard might have something to say – eventually – but when it comes to wood movement, we should not pay much attention to Mr. Tolkien.

And when I say “wood movement” I do not mean the Northern Light Treehugger and Fondler’s Ass-ociation (sic.) NLTFA. I’m talking about the fact that wood do change size in conjunction with changes in the humidity. As the book readin’ folks says it: wood is hygroscopic; it takes up moisture readily. Sort of like the silica gel packs from mother nature. Do not eat*.

* silica gel is not harmful and can safely be eaten in small amounts. The packs are marked with a warning label because they are a chocking hazard. Still – eating silica gel? Get back to the kindergarten!

In some cases, we do want the wood to shrink on us. If you insert bone dry cross members into green wood legs, the result will be a near permanent joint that’ll last a hundred years. When you know how things work, you can avoid problems and exploit the effects.

“I’ll just screw the sucker down! STEEL, DUDE (or dudette)!” you might think, slamming huge hex nut bolts through some 10mm steel into the Moria of your creation. Think again! There is nothing you can do to prevent wood from moving, except slicing it in very thin sheets and glue them back together with the grain going perpendicular per layer. That’s plywood, baby! Or you can shred or otherwise maul the forest into powder, mix with glue and press to shape. That is particle board, LDF and MDF (to name a few). But wooden planks? They’ll move. Do not overunderestimate (yes, I know that’s not a word) the effect of small forces acting on big and/or hard things.

If you screw steel channels to a table top without accounting for movement, one of three things will happen. Maybe not the first year or in three years, but it WILL happen:

1. The metal bends, leaving a top with a cup or a twist, depending on where the misery starts out on the moisture content scale. That’s hard to plane away – the 30° secondary bevel will get a workout on that steel girder…
2. The wood splits because it shrinks too much, for instance during a particular dry season.
3. The wood rips itself loose from what ever is holding it. It might be 20mm steel bolts, but the threads ain’t. You can rip that thing right out of the wood with a crowbar. Motha nature has more girl power than Spice Girls, TLC, Dolly Parton and Dixie Chicks combined, so she won’t be stopped with some lousy steel…

Wood. Will. Move.

The good news is that this fact is not the heeuge issue one might think it is. In fact, it is really easy to live with – once you understand what it is and how to deal with it properly.

“BUT! BUT! BUT! WHAT IF I GLUE IT TO PLYWOOD? GLUE IS STRONGER THAN WOOD!”

Yes. But wood is known to be able to rip itself apart if it has to. And it will bend plywood much easier than steel. Something’s gotta give. Even Jack’s character had to swallow the blue pill. Hard facts.

Example

The fact that wood will do it’s thing regardless of our feelings is why we can’t just lop down a tree, cut it to planks and start making stuff. Green wood shrink a lot (and crack, but that is another topic). The moisture content in fresh cut wood will generally be above 60%, and in some cases up to 120%. For oak, which I do have a lot of, the MC is appr. 75% when you fell it (will vary a bit depending on time of year). The MC will drop pretty rapidly at first – the trunk “drains” any free water. Cut a tree in the spring and place it upright against the wall, and there will be a puddle forming in no time. Then there is water bound within the cells of the wood, and that takes a long time to dry out. The harder the wood, the longer it takes. When the “free” water is gone, the wood is at its “fibre saturation point” which is around 30%. For shrinkage calculations, this is the starting point as the dimensions does not change above the fibre saturation level.

To avoid issues, exterior projects need 12-15% or below, while interior projects should be in the 6-8% (kiln dried) or 8-10% (air dried) range* to avoid issues. Let’s look at the slab I cut from this log:

* When we air dry wood, we cannot get much lower than about 10-12% depending on where you live. In kilns, temperature and dehumidificators is used to get the wood down to very low levels. You would need about 30% relative humidity for a long time to get down to 6%. Unless you drive camels to work, the humidity outside is most likely a lot higher.

It was about 95cm wide freshly cut. A good guesstimate indicates it will shrink about 5.5% in width by the time it reaches 10% MC. Most of the slab is quarter sawn, which shrinks a lot less than flat sawn. I explain the different cuts in my article “It’s slaburday!”, in addition to a LOT more info on making your own materials from trees.

5.5% – that should leave the once 95cm slab at 90 cm. 5 cm, or about 2”, less. That’s quite a significant amount of shrinkage. And we can’t exactly blame cold water in a pool…

So what do we do to avoid issues? First of all, we need to understand what to expect. Let’s look at some elm slabs:

These are flat sawn and contains both flat sawn, rift sawn and quarter sawn sections (more on that later on). We also need to toss in a few other words in the mix. Tangential, radial and longitudinal shrinkage.

• Radial: across the annual rings. The annual rings will be thinner.
• Tangential: Think diagonally across the annual rings.
• Longitudinal: The length of the board. The wood will shrink 0.1-0.2% in length, so we disregard that completely. It won’t affect our work at all, but now that you know this – keep it in that lizard brain of yours. A dead tight fit – and I mean cold, dry bone dead – is nothing you need to strive for. It is only good for the evening you make the fit.

The radial shrinkage coefficient is a lot smaller than the tangential, often around 50%. Since the annual rings on quarter sawn material is perpendicular to the surface, such boards will experience radial shrinkage. A flat sawn will experience tangential shrinkage – and thus proportionately more shrinkage than QS material.

This shows that quarter sawn material is a lot more stable than flat sawn, since the wood shrink a lot less across the annual rings than along them, so to speak. Add to the mix that QS material won’t cup or twist as much, the ray flecking is very pronounced on certain species and the low shrinkage factor will play nice with our projects, and you know why quartersawn material is so sought after and highly valued. This is the reason why I aimed for as much quarter sawn materials over wide slabs as I could, when I cut my oak to planks.

Speaking of quartersawn – in the above image I have a small dovetail box with a 100mm wide quartersawn oak panel. I created a reveal around the perimeter to experiment with that look. In addition I planned for movement by making the panel a good millimeter narrower than the width between the dado bottoms on both sides.

There’s quite a few tables online for wood expansion with values for radial and tangential difference. If you plan for an average of the two values, you should be on solid ground for most applications. For oak, the values are 0.0018 and 0.0036. If you use 0.0028, you are roughly in the middle. Then we need to look at the change in moisture content (MC). Here in Norway, the relative humidity (RH) in a normal house varies between 20 and 50 percent. Outdoors, the values are normally 40 to 90 percent RH. Here is a table showing which MC you can expect at a given RH. The values are indicative.

Relative Humidity	Moisture Content
10%	2.5
20%	4.5
30%	5.2
40%	7.7
50%	9.2
60%	11.0
70%	13.1
80%	16.0
90%	20.5

For a normal home, the RH usually is rather stable, but we should factor in very dry weather. In our example, we are going to assume RH values between 30 to 50 percent. This gives a difference in MC of (9.2 – 5.2) 4 percent. We put that into the equation and get:

100 (mm) x 0.0028 (average factor) x 4 (% change) = 1.12mm

This means that the panel could be expected to change width by 1.12mm. The 1mm allowance I put in should then be sufficient to prevent any problems. Especially since my guesstimates in this case are taken well on the safe side for the application.

 dimension x factor x change in moisture content = change in dimension

Note: I did not plan for any room longitudinally since we can ignore those 0.006mm the panel will change in length…

Doing all that math is of course a bit cumbersome, so shoot for a good average. For oak, I shoot for a “generous” 1% change in width (meaning that I add a bit of wiggle room). Of course this adds up if the pieces are very wide. For a one metre wide table top, the width can change with up to one centimeter over the seasons. In a stable environment where the humidity and temperature values are pretty much the same year round, there will be little change – but what happens if there is a particularly dry winter? Or something changes down the line, such as installing a heat pump? If your project is built to account for wood movement, these things won’t matter.

Just remember one important thing though: the calculated movement is the total. It does not tell you what will happen to the part from here on and out. If a part is expected to move 1mm and is right smack in the middle of its moisture content range, the actual movement will be 0.5mm in either direction. For small parts, this is really not important – make that 1mm allowance. But for a 2 metre (6’7”) wide panel, we are talking about a total movement of up to 20mm (¹³/₁₆”) in total.

If you want to geek out over this (and find more values for other species of wood), I recommend these links:

• The Wood Handbook, a 500+ page monster from US department of agriculture. A comprehensive study and research paper on this very subject.
• Moisture-related wood movement by Swedish Wood, a part of the Swedish Forest Industries Federation.

The last page in this article contains a table over some common woods woodworkers use. I sourced it from USDA, so it is naturally focused on American woods – but it is a good reference.

Now that we have all that knowledge, what to do about it?? Turn to the next page!