Warning: this contains a long, boring explanation of theories, with
only minimal real-world experience to back it up.  Skip it if you're
not into that sort of thing.

"Bill Clouser"  wrote:

> > The stock for these cannot be too dry.
>
> Reading this makes me wonder if the combination of very dry
> pegs and wetter surrounding timber will be susceptible to the
> same effects as the mortise and tenon joints in Windsor chairs.

For those interested, there was a great article a number of years ago
in FWW describing why dowel joints (round mortise and tenon, like
chair rung joints) are doomed to failure just about no matter what you
do.  I'll find the pointer if someone wants it.  I'm not sure if these
descriptions apply to pegged lap or mortice and tenon joints or not,
though.

> In his early stages, Mike Dunbar used to recommend that the
> tenon material be very dry so that it would swell to lock the
> joint together.  But later, he said that was a mistake, because
> he realized that, as it swelled, the super-dry tenon would
> compress the wood fibers around the mortise which would lead
> to a looser joint during subsequent wet/dry cycles where the
> tenon and mortise might shrink away from each other.  I
> believe this is his current thinking anyway.

Your description of what happens in the joint over time is true, but
according to the FWW article, it doesn't matter if the parts are wet
or dry to start out with.  Original moisture content at best postpones
the inevitable.

With the grain of the leg, the mortise will not grow and shrink much
with the weather.  This is against the grain of the rung, so the tenon
will try to grow and shrink in this direction.  Since it is
constrained by the unchanging mortise, it will be compressed in that
direction, and will shrink smaller during the next dry spell.  You can
line up the grain in the plane perpendicular to the rung and leg by
choosing where you drill the tenon, and the rotation of the tenon in
the hole.  In this case, the tenon won't be compressed sideways by the
mortice because both parts will swell and shrink the same amount as
the weather changes.  But, parallel to the tenon, the mortice will
shrink and grow (in depth), but the tenon won't, so the glue bonds on
the sides of the mortice will break eventually anyway.

As for real-world experience... does anyone have a wooden post and
rung chair more than a few years old that doesn't have loose tenons?

In the FWW article they were testing flexible gap filling glues
(silicone based, I believe) to solve this problem: make the tenon
undersized, and use a compressible gap filing glue.  They found that
it worked, but probably wasn't generally feasible.  The chairs were
springier feeling, and the joints were weaker when the chairs were
first constructed, but after a few wet/dry cycles, the flexibly glued
tenons hadn't changed much in strength, but the normal ones were much
weaker.

I don't really want to speak for Mike Dunbar, but as I understand it,
the main reason he stopped doing wet/dry locking tenon construction of
his chairs was because he realized it was unnecessary.  The stretchers
in a Windsor chair are compression members, not tension members (in
his book he makes the stretchers 1/4" longer than they should be, to
make sure they're pushing the legs out).  Since they work to push the
legs apart, it doesn't help much to construct them to resist a
pulling-out force.  If the mortices and tenons are well constructed,
the bottom half of a Windsor will stay together with no glue at all,
and sitting on the chair only strengthens the joints.

As you can tell, I definitely need to spend more time in the shop, and
less time reading...

Alan