Imagine if you look you'll see this is not the case. For instance, the
lower shelf of Toller's sofa table, made with a single board of butternut,
if tightly held between the legs has a 50/50 chance of popping or loosening
the joinery as it expands beyond its ability to compress what holds it and
itself. Simple remedy is good woodworking, either pinning center on the
cross-grained rail and floating both ends with a sixteenth gap by making
oversize holes for the pocket screws, or realizing that there will be a
front - where the drawers open - and a rear, pinning flush to the leg at the
front, floating the center and rear, allowing the full eighth expansion
where it's not noticeable.
Using your example of frame and panel construction, which, I assume, you
know is designed to maintain virtually constant exterior dimension through
changes in MC, you always want to finish the panels prior to framing them,
because they can shrink and reveal bare wood at their edges. If made too
large when dry, they can find the weakest glue joint in the frame and
destroy it, or finding themselves the weaker, compression set their fiber
and develop a rattle. A competent woodworker anticipates and compensates
for all these.
I'm sure you've seen your share of tables with split tops (and chairs with
split seats) which were restrained without regard for wood movement, as well
as those with ears from having well-varnished tops and virtually finish-free
bottoms. Then there are the open corners on mitered wood ... the list is
endless. It's never been my objective in woodworking to emulate the shoddy
manufacturing processes of commercial work. Therefore I use techniques
designed to accommodate future movement.
This is where I see his construction differently, as viewed on the web site.
He has a support between the front and rear legs upon which the shelf sits.
I would agree that if it were simply a piece of butternut, he'd have
problems without a doubt. What he has though, with the support screwed in
place, is very much like the standard construction of any dresser or other
piece of furniture. Secured to the support, the butternut is not going to
move as it would if it were simply sitting exposed to the changes in
climate. Such is the nature of furniture construction. All furniture is
constructed with cross grain construction and it is constructed in a rigid
manner. Yet - it does not expand and contract so as to create a wobbly
piece after a few cycles.
I fully agree that this would work, but I have to keep going back to ask the
same question - and I'm not trying to be difficult with you George. I'd be
reluctant to float the board as I'd fear cupping or bowing. Those two
possiblities are something I'd consider to be more likely. Look at a
dresser. It has to contend with cross grain expansion as much as any other
piece of furniture does but the top is not floated. It is secured to the
frame rails. Since however, he's fitting a piece of wood in between the
legs, I could see where a very slight shave might be OK, to allow for some
movement, but even that I'm not sure is necessary, since the shelf will be
united with the support and the support being long grain, will limit the
movement of the shelf.
Precisely the point that I made. The panel, while floating is only doing so
in that it's not glued in. It's not loose though, when it's fabricated. In
other words it won't rattle about. Yet - it does not push the joints apart
as it absorbs moisture. It is constrained by the rest of the wood around
As well, I think it could be said of everyone here that they pursue proper
woodworking techniques. Yes, I've seen bad fabrication, but I've also seen
excellent fabrication. I tend to base my observations, my practices and my
thoughts on those examples of excellent fabrication. Basically, a wordy way
of saying "ditto".
Sorry to say it, but your design, as shown in the photo is probably not the
best for longevity. :(
Then again, it may last forever.
Next time consider using M&T joints for the "rail" (your "support") between
the legs, and attaching a "cleat" to the inside of the rail as the shelf
support. Then attach the shelf to the cleat with screws in oversize or
Here is poor picture, but a much better way, but not the only one, to attach
a lower shelf between table legs:
Be particularly wary of the advice you've received in this thread to
"ignore" cross grain situations ... it is simply wrong and misguided.
Hell, just ask to see some samples of their work before making up your mind
to rely on such advice.
Yes, a M&T is clearly the best way to do the support, but by the time I
realized I had a problem the side and legs were all glued up.
But why not screw to the two legs, and put a support under the shelf,
attached to the legs in mortises; attaching the the shelf to the support
with the slotted holes?
That is what I thought about doing before I decided it wasn't practical at
that point of construction. (another problem was that I had never done a
mortise and tenon before; but I've never done a dovetailed drawer before,
and there are 6 of them in it...)
Sorry, I left a word out.
Why not screw the shelf to the two legs, and put a support under the shelf,
attached to the legs in mortises; attaching the the shelf to the support
with the slotted holes?
The shelf, with grain the same as the sides and top can be screwed between
The support, with grain perpendicular to everything else, mortised into the
legs. And screw the shelf to the support with slotted holes' though they
could probably be left out.
We're talking solid wood here, not laminate or veneer, right?
Why? You already know that solid wood of that width needs room to expand and
contract with seasonal changes.
A design which calls for "screwing the shelf to the legs", unless I am
missing something in your description, does not appear to be taking the
dimensional instability of the solid wood into account.
As a furniture maker, you do that at your eventual peril.
A better way, as in the posted picture link, is to attach the shelf support
to the rail between the legs, leaving room for the shelf between the legs to
expand and contract. This also has the subjective benefit of hiding the end
grain of the shelf and giving a seamless look to the joinery.
Alternately, if the design calls for the shelf to be supported by the rails,
and the shelf is solid wood, you would be better off not "screwing the shelf
to the legs" in any manner, instead leaving space for expansion and
contraction of the solid wood shelf that you already know is a distinct
possbility. Just incorporate that space into your design, and be sure to
"join" the shelf to the rail with joinery that takes wood movement, and the
cross-grain, into account.
One way to pick up on time tested methods is to study the construction of
antique furniture. You will notice that the many antiques still in use have
similiar joinery methods ... with good reason.
Forget today's "commercially" made furniture, regardless of its "quality"
... a good bet is that much of it won't be around in 100 years, and it is
doubtful that most of it is actually "solid wood", despite its appearance.
You will be far better off learning what you can from those pieces which
have stood the test of time.
In today's vernacular, "solid wood" often means plywood plus a few pieces
of hardwood face-frames. Kind of a different meaning to "solid" wood than
most of us would use.
If you're gonna be dumb, you better be tough
Maybe I am not understanding the problem here. If I am, I would be grateful
for an explanation.
The two legs are held together by:
1) the top, which is parallel to the leg grain (pocket screwed, glue?)
2) the side panels, parallel to the leg grain (glued)
3) the shelf, parallel to the leg grain (pocket screwed)
Presumably the top, side panels, shelf, and legs all expand pretty much the
same rate; so there is no relative movement. The legs are quarter sawn, but
they are small and on the outside, so that shouldn't matter.
The problem that prompted the whole thread is 4) the cross grain support
(loosely pocket screwed).
I understand the support should have been done better; presumably mortised
to the legs. But what difference does it make whether the shelf is screwed
to the legs or not? The danger is that the shelf (and the legs and top)
will expand and stress the screws holding the support. The shelf will not
expand any more,or stress the support anymore, just because it is screwed.
That would not be true if this were August, since then I would have exactly
the opposite problem; but is it January.
(Thinking about it now, I could probably have avoided the problem by simply
making the shelf 1/16" narrower everything else. Does that make sense?)
So, where is the wood movement that I am not considering?
(I haven't put the top on yet, and am undecided about gluing it.)
Remember that link to the fpl site, wood handbook chapter three? Look at
the orientation of the annual rings and the characteristic direction of
movement. Your top and or shelf are going to do some cupping and making
ugly unless you change you design. Not to mention - again - the instability
of a pocket screw leg into a plank.
I can't disagree that the shelf is at risk, but why should the top cup? It
is firmly attached to the body on all four sides. If it cups, then anything
I am familiar with alternating ring direction to avoid cupping.
Interestingly, a book I just read (returned it to the library today) says
that you should have all the rings oriented the same so that the grain
across the panel matchs better than it would if you alternated. He didn't
think alternating rings gained any stability.
You don't avoid cupping by alternating ring direction, what you supposedly
avoid is cumulative errors resulting in one big bow. Not so, of course.
Geometry doesn't work like that. Match heart to heart, sap to sapwood for
best display, and if you're just screwing the two sides of the top, which is
what the message implied, you have an unsecured middle which will choose its
Out of curiosity, have you studied any basic woodworking texts? Are you
working from any sort of plan?
"why should the top cup? It is firmly attached to the body on all four
That implies it just screwing two sides of the top with an unsecured middle?
"Interestingly, a book I just read"
Have I studied any basic woodworking texts?
Am I working from any sort of plan?
If you were trying to be helpful before, I appreciate that; but now you are
just randomly argumentative.
Primarily the sides are holding the legs together, though the top could
probably adequate also. Was my 1,2, 3 ambiguous?
The shelf adds a little rigidity probably, but the legs couldn't possibly
come apart if the shelf were removed.
The top and the sides have paralllel grain. The top and the front and rear
aprons have parallel grain. Why would I want to float it? With all the
grain running the same, why would anything move?
You said your were going to glue the top on? Regardless of the underlying
grain direction, that is generally not a good solution.
Have you heard of "figure eight fasteners" for attaching wood table tops?
Yeh, I know about figure eight fasteners; but I am perfectly serious about
this question. If the grain runs the same as the top on all 4 sides (and
all the wood is butternut, bought at the same time, and stored together for
several weeks); what is going to move? It is no different than a glued up
board, or any other matched grain joint, is it?
If I am missing something, I would sincerely appreciate having it pointed
out. I probably won't glue because it will be messy and I don't think it is
necessary, but I can't see why it would hurt.
HomeOwnersHub.com is a website for homeowners and building and maintenance pros. It is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.