I know that wood movement is an issue, but it's a bit vague for me where I should be concerned with it.

This page Suggests that there is almost no movement in the longitudinal direction, but that there is significant movement in radial and tangential directions respective to the grain.

Are there particular orientations which I should avoid, e.g. radial in one piece in a plane with tangential of another?

With what orientations is it safe to use screws without hardware that would allow movement?

  • Suppose it is worth mentioning that it depends on the grain direction of your wood. Longitudinal and tangential movements differ depending on that. Then there is temperature, humidity, assembly location and permanent location....my god so much to consider. Look forward to an answer
    – Matt
    Commented Apr 13, 2015 at 0:56
  • Longitudinal should be "with" the grain, radial is inward from the curve, and tangential is tangent to the curve ... I'm not sure what you mean by "it depends on the grain direction of the wood?" Do I need to clarify the question?
    – Daniel B.
    Commented Apr 13, 2015 at 1:10
  • I was more trying to suggest that there could be too many conditions to account for. Not all boards react the same way to their environment. I was basing that comment on what I was reading about grain direction on wood. Your question is fine.
    – Matt
    Commented Apr 13, 2015 at 2:04

3 Answers 3


That page you link to is correct, wood movement is almost completely across the grain. Longitudinal or long-grain movement is so slight it can nearly always be ignored.

Just to note, we must remember movement is both expansion and contraction, which typically happens seasonally:

Seasonal wood movement

Obviously there are specific instances where movement cannot be accurately predicted. Where the grain shifts direction, as much as 90° within the same board (e.g. wood taken from various crotches within the tree) and particularly where there is wildly changing grain (e.g. in a burr or burl).

Are there particular orientations which I should avoid, e.g. radial in one piece in a plane with tangential of another?

This is the standard cross-grain situation you want to try to avoid where possible. So the answer is basically yes but it's complex because you can't avoid it in many cases.

Where you have no choice but to make a cross-grain join, you must build in some means to allow for movement, for example by fixing loosely as with various tabletop fasteners, fixing at one end of the joint only or allowing something to float within a groove or rebate (US: rabbet), as in the floating panel in frame-and-panel construction.

Unless you take these sort of steps to avoid an issue you can take it that something will happen, not that it might.

With what orientations is it safe to use screws without hardware that would allow movement?

That depends for a start on how you define hardware. The standard 'buttons' used to attach tabletops and allow movement are wood, and are made by the woodworker, but are technically hardware.

Rather than think about the orientations it is safe to use screws visualise the expansion and contraction of the wood and how you allow this to happen after fixing. So essentially screws could be safe in any joint they can be used in, as long as they're utilised appropriately.

A good example is the top to any table, which typically will be cross-grain to some portion of the underframe or apron assembly. One standard way is to fix firmly in the centre and allow the front and back edges to move. Where you want an edge to remain fixed relative to the piece (e.g. a table intended to go flush against the wall) you fix firmly there and allow movement to occur out from that edge:

Allowance for tabletop wood movement

Screws are used here both for the points where firm fixing is desired as well as where movement is being allowed for, by screwing directly into the tabletop through the apron and through some form of movable fixing/fastener respectively.

  • So these all have to do with large surfaces, but there are plenty of instances where wood is joined cross-grain on smaller surfaces. The simplest example would be a lap joint on a small frame, at which point I have movement up and down, as well as outward/inward from the center of the frame, do these not put stresses on the joints, or is it that they are all expanding/contracting in roughly the same shape and so the stress on the joint is minimized?
    – Daniel B.
    Commented Apr 13, 2015 at 13:40
  • 1
    @Daniel B, movement is proportional to the size of the piece of wood. Small pieces = small movements, which is why we can get away with things like corner lap joints on things like picture frames (species and cut of the wood, as well as the type of glue used are also factors). But even at that scale joints can be pushed apart, or pull away from a glue line due to shrinkage. I have a lap tray (you can visualise the approximate size) where all 4 corners are now broken in some way. The gaps were large during the winter but are closing up as the moisture level in the air goes up as it gets warmer.
    – Graphus
    Commented Apr 13, 2015 at 14:42
  • So what's really important is the size of the joint. If I tried to screw every corner of the table top down, it'd snap screws or split wood because there's a potential for(for a 4 foot width) as much(or more) as an inch of movement, but if I make an open frame of the same size, the only movement is the radial expansion of the boards, which wouldn't be more than 1/64th, a negligible amount. Is that about right?
    – Daniel B.
    Commented Apr 13, 2015 at 15:29
  • @Daniel B, yes that's about right. I'm not sure if you'd ever see a full 25mm or 1" of expansion on a tabletop of that width though!
    – Graphus
    Commented Apr 14, 2015 at 7:15
  • 1
    @Daniel B, I think the takeaway from this is to err on the side of caution. Particularly since it often doesn't require much effort to make allowance for movement so it's really lazy not to :-)
    – Graphus
    Commented Apr 14, 2015 at 14:22

Lets use a worse case scenario. A table top 30" wide made from flat-sawn wood connected to a cross brace (so the grains of the two pieces run perpendicular). If you look at how table tops are connected, they are usually just attached in the middle, allowing the top to expand to each side. In this case, the table could expand or contract up to 3/8 inch in total width.

In a best case, you have a 4 sided box. Even though the sides will expand and contract, they should all change the same amount at the same rate, so the joinery can be tight.

Wood moves more in areas that are cold and dry in the winter then hot and humid in the summer, than in areas that are relatively consistent year around. Items stored in a climate controlled house won't see nearly as much movement as others.

You can minimize the effects of wood movement by choosing quartersawn and rift-sawn boards.


Graphus gave a great answer covering many of the design considerations and pretty much covered the strategies for dealing with movement.

Shrinkage and expansion also vary by material (wood, MDF, OSB, plywood...), species, climate, and application, as well as other factors. If the humidity level is relatively consistent, you won't need to allow for as much movement as if you're building a piece of outdoor furniture which will be subjected to the elements, for example.

If you aren't sure how much movement to account for in your project, you can use the Shrinkulator to get an estimate based on the wood species you're using.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.