How did George Nakashima do it?

[smorgasbord]

I'm lucky enough to have inherited my parents' dining room table, built by George Nakashima in 1972. It's a 2 piece bookmatched top, about 40" overall, with the "Frenchman's Cove" base. Here's the center underside:
[attachimg=1]

In the cross-member there's a screw near the center with no elongation. Note just a regular slotted screw, btw.

Here's the furthest from center support view:
[attachimg=2]

Here's kind of an overview of the underside:
[attachimg=3]

BTW, note that he used screws to pull the butterflies in:
[attachimg=4]

So, how is Nakashima accounting for wood movement of the wide top pieces with plain screws and no hole elongation?

 

[tsmi243]

Wood doesn't move as much as everybody says on Youtube. 

 

[Crazyraceguy]

The table width is already halved by the two pieces not being glued together. There is an obvious gap between the pieces. So any movement is isolated to each section. The battens are pretty thick, but the screw holes seem fairly deep too. This allows for a shorter screw, which has very little of its length being held, in a tight hole. It doesn't necessarily take much slack in the hole, for it to move back and forth, even tilting the screw.
Next is also the amount of care the object has seen in its time. The shrinking/swelling is purely humidity based. So, an item that is in an environmentally controlled area may not see those variations.
If you built that in the swampy air of a costal city, then moved to the Arizona desert, you would see a pretty big swing.
Some species (or even ways of cutting timber from the tree) have more movement than others. It just depends on a lot of factors. Someone could own that table successfully for 50 years, then one day....

 

[smorgasbord]

The table width is already halved by the two pieces not being glued together.

Yes, but there is still over 20" of width to accommodate. The top is only about ¾" thick, so yeah those screws don't go deep, but seems they would have loosened over the years. Maybe because the top pieces are close to quartersawn? The base is walnut but I think the top is maple, not the most stable of woods.

Piece was built in New Hope, PA, ended up on Long Island, and is now in Northern California. Of course, inside heated homes, but there was a year or so where it was in an unheated storage place. So I'm sure movement has occurred, or successfully prevented.

 

[ChuckM]

Is it possible that the other side of the holes is larger/elongated, which can be verified if one of the screws is unscrewed?

I haven't seen the YT on wood movements, but it's a real thing. This oak table was made in the 50s/60s as a junior high (middle?) school project in the UK and moved with the family (neighbor) in the 80s to Canada. The coffee table top is 16" wide (the small footprint is line with the size of typical British houses). I glued the top back four years ago and elongated one side (the underside) of all the mounting holes. The owner (maker's grandson) has since reported no more issues with the table.

[attachimg=1]

[attachimg=2]

 

[Crazyraceguy]

So I'm sure movement has occurred, or successfully prevented.

You can't prevent it, just mitigate as much as possible. You can quite literally split rocks, by driving a dry wooden plug into a drilled hole, and soaking it with water.

When it swells, it will attempt to split whatever is restraining it. If there isn't enough mass to the wood, it just crushes the fibers.
When it shrinks, it will split open, since it can't stretch. Where it splits is anybody's guess, likely in the biggest eyesore location though  [unsure]

This is the whole point of running the grain from side to side, with a solid wood drawer bottom. If it runs front to back, when expended, the drawer might not open. Fortunately, sheet goods have mostly eliminated this issue.

 

[Michael Kellough]

Here is a simple chart of the coefficient of expansion/contraction of domestic woods.

As you surmise quarter sawn wood is much more stable. Silver maple is one of the more stable hardwoods. A surprise to me is that the softwoods are generally more stable than the hardwoods.

I estimate that the screws farthest apart have had to give at most 3/16” each. They might have bent if the pilot holes in the stretcher/batten are tight but if generous they might have simply leaned.

Please don’t take them out to find out…

 

[Svar]

Wood doesn't move as much as everybody says on Youtube.

That depends on so many things: species, growth conditions, drying mode, cut direction, finish, climate outside and inside the house.
Maple/padauk desk I'm sitting at moves 1/8" over 24" width seasonally. Is this more or less than everybody says on Youtube?

 

[Svar]

Yes, but there is still over 20" of width to accommodate.

The width to accommodate is between two screws. Looks a lot less than 20". Other half has a crack, which takes care of some slack. Screw holes not being tight will give few mm of wiggle room.
[attachimg=1]

 

[fritter63]

Wood doesn't move as much as everybody says on Youtube.

Wood doesn't move as much as everybody says on Youtube.

Here is a picture of small stickley style desk I made 25 years ago. Quartersawn white oak. The tongue was flush with the end cap when I made it (we have also moved from Denver to California since).

I also have a dining table (36" wide) in red oak (non QS) that has a dovetail joint in the end cap. I've seen that top retract by at least 1/8"  on each side, maybe more, it was shocking. But my son has it so I can't get a picture of right now.

Guess I should make a video about this on my YouTube channel. 8)

 

[Michael Kellough]

A good while back, before Sam Maloof passed, I toured his shop and house with some of the Talk Festool guys.

In the house were several tables made of walnut with breadboard ends like the photo above. I was very surprised to see that on every single table the sides were absolutely flush with the end piece.

I asked Sam about that and he seemed surprised at the question. He said he’d never had a problem. I still don’t understand how that could be.

 

[ChuckM]

Wood doesn't move as much as everybody says on Youtube.

I also have a dining table (36" wide) in red oak (non QS) that has a dovetail joint in the end cap. I've seen that top retract by at least 1/8"  on each side, maybe more, it was shocking. But my son has it so I can't get a picture of right now.

Guess I should make a video about this on my YouTube channel. 8)

To cater for the wood movements of the top, some follow the guidelines of deciding on the length of the breadboard as described in this article:
https://www.woodmagazine.com/woodworking-tips/techniques/joinery/breadboard

 

[jeffinsgf]

Wood doesn't move as much as everybody says on Youtube.

As others have said, that depends on a lot of things. One of my very first woodworking projects, some 45 years ago, I glued up 3/4" red oak boards about 18" wide and 4'-6" long. I bolted together an aluminum angle frame, rack mounted my stereo gear in the frame and bolted the oak panels to the sides. This was done in Southwest Missouri in the middle of the summer. That fall I moved to Des Moines and rented a small old house with inadequate insulation. It was bitter cold that first winter and the heater ran all the time. Being young and stupid (and it being the 80s), I didn't know about or consider a humidifier. I still remember the sound of one of the sides splitting at 3:00 in the morning. The wood moved and the aluminum frame didn't. Sounded like a gun shot inside the house. That was my lesson in wood movement, and I've never ignored it since.

 

[smorgasbord]

The width to accommodate is between two screws. Looks a lot less than 20". Other half has a crack, which takes care of some slack. Screw holes not being tight will give few mm of wiggle room.
[attachimg=1]

1) There are 4 screws in each top piece, not 2. The distance between the outer ones on each piece is 14.5"
2) The top two pieces are joined with butterflies that are not just inlaid, but screwed in from the bottom. In essence, the top behaves as one 40"+ wide board.
3) The distance between the outermost screws is 32".
4) That crack you pointed out is not visible in the top, but there is a hole. The other side attachment doesn't have either.

[attachimg=1]

All of the scews appear to be plain old flat head screws, slotted head, with the counterbore just the diameter of the screw head. The top is tight with the base. I do not see how the screws can hold AND allow for wood movement in a 32" wide screw to screw distance, but it's lasted decades with no adverse signs.

 

[ChuckM]

Perhaps, you can contact the maker (now under the management of his daughter), and ask about it.

 

[jeffinsgf]

2) The top two pieces are joined with butterflies that are not just inlaid, but screwed in from the bottom. In essence, the top behaves as one 40"+ wide board.
ost screws is 32".

I'm going to throw the challenge flag on that statement. Butterflies and all, I think the segments are acting...at least somewhat...independently.

 

[smorgasbord]

I'm going to throw the challenge flag on that statement. Butterflies and all, I think the segments are acting...at least somewhat...independently.

I can't imagine physically how that could be unless the 1/8" gap in the middle gets wider or narrower, which, if it does, is imperceptible, which would seem too small a movement for a 40+" wide top.

[member=57948]ChuckS[/member] , sometime in the past I did reach out to get a replacement quote for insurance purposes. I don't seem to have that letter anymore, unfortunately, but she did mention something about the challenges of keeping the wide tops flat. I think she gets a lot of inqueries these days, so I'd have probably (and rightfully) have to pay for another insurance quote (probably worth doing again, and maybe dig up data on the older bedroom set we have), and ask in that request.

 

[Michael Kellough]

The flathead screws sit in conical countersinks so an oversized countersink wouldn’t make any difference. The head won’t slide in response to the tabletop moving.

But if the pilot hole the shank of the screw passes through is oversized then the screw can lean as the wood moves. If the pilot hole was/is tight then the (brass?) screws will likely now be bent.

 

[fritter63]

I managed to find a couple of (inadvertent) pictures of the aforementioned dining room table.

First picture is from late 90s right after I build it (in Denver). You can (almost) see how the dovetail key is flush with the end cap.

Second picture is from last year during repairs (ahem, long story). In California, 15 miles from the coast, it's a bit more humid here, so I'm surprised that the top has SHRUNK so much, rather than expanded. As you can see the, the dovetail key is now recessed quite a bit into the end cap. (look just below the end collar on the pipe clamp)

 

[smorgasbord]

OK, well it's dry Jan now, so I'll take some careful measurements.
Someone come back here in wet July/Aug and remind me to take measurements again.

I'm thinking of doing the sliding dovetail thing - accommodate wood movement AND help keep the top flat.