Plain miter joints are strong enough by themsleves

[ChuckM]

Miters need no dowels, splines or dominoes for added strength if you believe in these (the effect of wood movement over time has not been factored in). (But dominoes are still a great aid during clamping of miters. [tongue])

FW's findings (2009, based on pushing):

[attachimg=1]  [attachimg=2]

More recent findings (2023, based on pulling):

[attachimg=3]    [attachimg=4]

Video:

 

[Packard]

As a former picture framer, I can attest to the fact that miters will fail over time from just the seasonal movement.

In most cases, when I was asked to re-frame a failed miter frame, the joints were glue starved.  But nonetheless, there are no stresses placed on the joints, and still the miters failed.

The wider the molding, the more likely that seasonal movement will make the joint fail.

I use miters on my kitchen cabinet doors, with two dowels per corner.  I prefer it because the painted joints look nicer.

But, I suspect that the reason miter joints will not ever be very popular is because you need to use twice as many bar clamps to clamp it.

Also, I question the data.  I tested the picture framers’ standby glue, Corner Weld, against Woodworkers III.  I made miter joints on 12” pieces of 3/4” x 1-1/2” poplar.

I only made three samples of each, with generous squeeze-out, but the results were highly consistent.  Hanging barbell weights from the end of the miters all failed when I went from 50 pounds and jumped to 60 pounds. 

The big difference between the glues was that Corner Weld started to setup in about 30 seconds, and you could handle the piece in about 5 minutes.  If all you are doing is joining miters on an underpinner, then long open times are not necessary and are not your friend.

My results do not sound anything like their results.  Keep in mind that you probably need a minimum of 10 to 20 samples and average out the results for the results to be meaningful.

I did only three samples.  I was not looking for absolute results, only relative results.

When Corner Weld came out, I tested it against the industry standard back then, white glue.  The Corner Weld was vastly stronger than white glue—nothing ambiguous about that.  So three samples was fine to find a winner. 

I really question the results.

 

[ChuckM]

This box-maker has this to say about different miter constructions, and why he doesn't reinforce his mitered boxes:
https://www.finewoodworking.com/2015/02/27/not-all-miters-need-reinforcement

Box maker Doug Stowe does use some form of reinforcement in most (all?) of his boxes

 

[Svar]

More recent findings (2023, based on pulling):
[attachimg=3]

Box joint 754 and Butt Joint 1303? That can't be right. It's poor technique, flawed methodology, or lack of replications.

 

[ChuckM]

Box joint and dovetail joint should be similar or close in strength.

 

[Packard]

Mitered edges absorb large amounts of glue.  So you can apply an even coat of glue and still end up with a glue starved joint.

I always apply glue to both surfaces allowing about a minute for the glue to be absorbed.  If I applied enough glue, then there will still be a minute amount of squeeze-out. 

On my Shaker doors with mitered corners, I place two 3/8” x 1” dowels per corner.  It not only strengthens the joint, it also makes alignment nearly automatic.  I still need four bar clamps per door though to pull all the corners completely closed.

I have sliding table saw which cuts 45 degree miters only.  The miters are always perfect, and the length scale and stops make for a high degree of repeatability.  But the dowels will help in alignment.

Splines, biscuits (and I believe) dominoes have too much lateral play to ensure alignment.

 

[jeffinsgf]

What these instant tests fail to address is that the two pieces of wood are not going to expand and contract at the same rate over time. The glue line is going to be eroded away as the two pieces move in different directions at different rates.

I find it amusing that techniques that have been foundational for millennia are routinely questioned by "experts" that bought their first table saw 18 months ago.

I haven't (and probably won't) watch the video. I wasted a half hour on a similar one 4 or 5 weeks ago. I could be completely off base on the "expert", but I'll keep putting splines, dowels or Dominos in my miter joints, thank you very much.

 

[mino]

What these instant tests fail to address is that the two pieces of wood are not going to expand and contract at the same rate over time. The glue line is going to be eroded away as the two pieces move in different directions at different rates.
...

This.

I will add that below effects are fighting against a mitter joint:
- humidity cycling
- wood hardness - with softwood the glue will "bend" the fibers a bit, preventing a break up to a point, less so with hardwood
- glue age, glue degrades over time and becomes more brittle, so testing a freshly-glued piece is pretty much useless

Basically, the tests are ideal scenario for a mittered joint. Reality is way harsher.

There is one scenario where a mitred joint is fine in my view - short-term use item, toy and in general small items with careful choice of same-grain direction like art pieces etc.

Nice case are the small drawer carcasses Peter Parfit (over)built in one of his videos. At 10x10", even if half of the joint breaks off, the pieces will still hold.

 

[ChuckM]

Snip
(and I believe) dominoes have too much lateral play to ensure alignment.

You can change your belief on that. Dominoes (when mortises are milled in the narrow setting) have no play. Sometimes, they're too tight in hardwood even in dry fitting that they need to be removed with a pair of pliers or something.

 

[ChuckM]

What these instant tests fail to address is that the two pieces of wood are not going to expand and contract at the same rate over time. Snip.

Since they were instant in nature, the tests were limited in some aspects. But, in both tests (FW and YT), they did caution that wood movement or the passage of time could affect the observations.

By the way, the two pictorials I pulled from the YT video are pretty much all you need. I didn't watch most of it; I only viewed the part about the mechanism used and the end about the summary.

 

[Packard]

Snip
(and I believe) dominoes have too much lateral play to ensure alignment.

You can change your belief on that. Dominoes (when mortises are milled in the narrow setting) have no play. Sometimes, they're too tight in hardwood even in dry fitting that they need to be removed with a pair of pliers or something.

I did not know that.  Every video I have ever seen shows fairly loose fits.  You can tell by how easily they slide the dominoes in.  A tight fit domino would work like the two dowels.  It would prevent alignment issues while clamping. 

For mitered rails and stiles, (2-1/2” wide stock, I made a small measuring fixture for marking the placement of the dowels.  I use a Rockler clamp on fixture, so only one line is required for the two dowels.  All my corners are interchangeable, but out of habit I mark them A/A, B/B, etc. 

If anyone is making similar miters, then a marking fixture is faster and more accurate than a ruler.

 

[Crazyraceguy]

Snip
(and I believe) dominoes have too much lateral play to ensure alignment.

You can change your belief on that. Dominoes (when mortises are milled in the narrow setting) have no play. Sometimes, they're too tight in hardwood even in dry fitting that they need to be removed with a pair of pliers or something.

That is absolutely true. I have left dry-fitted assemblies clamped up overnight before and had to struggle to get them back apart again.....just to put it back together again  [huh]
That little line down the edges of the Dominos makes them fit very close.

 

[luvmytoolz]

I did not know that.  Every video I have ever seen shows fairly loose fits.  You can tell by how easily they slide the dominoes in.  A tight fit domino would work like the two dowels.  It would prevent alignment issues while clamping. 

I have both Domino's and haven't come across any size Domino between the models that comes even close to a loose fit. The 4's are far easier for sure as they're tiny, but even they fit snugly. I'd say the loose fitting ones you've seen have either been sanded or their technique leaves something to be desired?