I saw this on a recent show from my not-so-favorite TV woodworker. I didn't believe it would work so I tried it out, to my chagrin it actually does work. This was the Scott Philips show (American Woodshop) and if you've seen any of the recent episodes you know he does some pretty dumb things....but they always come out "perfect". Enough bashing, on to the tenon. What he did was take some blind tenons, these were the legs of a stool, cut a partial slit in them, and drilled a hole. He then put some polyurethane glue into the hole (only) and stuck the tenon into it's mortise. His claim was the glue would expand and wedge the leg into the stool seat, with no exposed tenon on the seat surface. Thinking that he was full of, well, something brown I tried it. It actually worked, the leg could not be budged from the mortise! Knowing this wasn't possible, I decided to dissect the joint. So I cut the leg off my test "seat", and then cut the joint in 1/2 with the bandsaw. I suspected the glue has simply leaked into the joint and glued everything together. That wasn't the case, the 2 halves of the tenon popped right off. There was some small amount of glue seepage, but not much and it didn't hold the pieces in place, they were quite easy to remove. So I have pics, and if you want a blind tenon this may have some application for you. (I still think Scott's show sucks.)
Anyway, the first 2 photos show how he cut the slot, notice it's on a angle front to back...I've shown both sides. You also see the 5/16" hole that's drilled roughly 2/3 of the way through the tenon (tenon is 1"). I filled the hole about 3/4 full with the urethane glue. Disclaimer: he used the Gorilla brand, I used Titebond. I don't like this glue and didn't have any on hand, so I took a $5 hit in the name of science. My seat was a 2" thick piece of scrap where I drilled the 1" mortise. After putting glue into the hole, I clamped it in a way I hoped would minimize any glue seepage into the joint. After it sit overnight, I took it out of the vise to find the joint was incredibly tight...no possible movement. So then I cut the leg off and dissected the joint. After I popped the 2 halves of the tenon out, you can see there was some glue leakage into the joint...that was where the hole was, and the expanding glue got to the mortise wall. Otherwise the glue pretty much didn't have anything to do with the tight joint. I didn't have a way to measure the expansion, but the slot is noticeably wider. So if you need such a joint, this one might work.....I may stay with the through hole and exposed wedge on the top of the seat.

That's pretty interesting! Now, I wonder, that if the tenon expanded because of the moisture would it also shrink at a later date when the moisture dried out? Barrels and other objects rely on the constant moisture to keep the wood swelled. That would not be the case here.

 

I've heard of fox wedged tenons being used when not using thru holes but that's a different process.

No idea, but maybe the glue holds it in the expanded positioned. I didn't add any moisture, the glue only reacted with whatever was in the wood.

Clever use of two materials.  Example:  Chair spindle fix:  inject 50/50 mix yellow glue and water into the joint (needle helps get deep).  Water then swells wood; the yellow glue structure hardens and "freezes" wood in expanded state.  I note that wood hardener used on rotted wood seems to do the same thing:  "freeze" even rotted wood into a matrix.  I also remember that caution about dry-fitting:  when you finally add glue, the joints will be a bit tighter than the test fit.  Clamping a joint controls the expansion, causing wood fibers to mesh at the joined surface, and then lock in place as the glue dries.  

 

I share the distaste for the show mentioned by Fred.  But I watch anyway.  I guess I figure the star of the show is really the wood, which wins the fair maid in spite of the villain?

Now the question is does this only work with one kind of glue or with any?

Well, I can only guess...but you have to have the tenon expand; that's why the urethane glue was used. It might be that using water/PVA as Pete described might do it, but it seems like enough water might dilute the PVA to the point of not working that well. But like I said, I'm guessing.

Only thing about Polyurethane Glue I don't like is the shelf life sucks.  Get a small bottle, use today, next month go to use again, it's a solid bottle.  Titebond II gallon jug 3 years later still ready to go and still holds great.  

 

Reminds me, I need another gallon saw mine was low when I refilled the glue bottle last night.  Oh goodie a trip to the borg. 

I know the poly will expand and move loose parts . But to think that the swell will move wood would take a lot of pressure and if it is locked in with no air it will not cure. So I think the moisture it absorbs swells the wood. But not being a chemical engineer we may have to do more tests

So, which way is the grain and hole/slit oriented relative to the grain in the mortise.   When I took the "Make a Chair From A Tree" class, every piece had an optimal orientation to maximize the wet-dry joint.

While Polyurethane glue fills gaps, it fills them with non-structural foam.   In a FWW test years ago, it was the only glue to fail in the "slightly loose fitting" mortise and tenon joint.

 

And thank you for your $5 + time and material contribution to science.

When Phillips did it (as I recall) it was unclear if the paid attention to the grain orientation. It was a 3 legged stool, with splayed legs. I don't believe he did anything special regarding the grain orientation. In my test I didn't, but that may have been because I really wanted this to not work . My slot was cut sort of across the grain at an angle to what would be optimal. I haven't done any testing of the relative strength of urethane glue versus anything, but past use convinced me it wasn't of any use to me. Maybe if I glued concrete blocks together or such, but not in my woodshop.

I think water is an important part of the (pardon) mix:  remember that water will raise the grain of a surface so you have to sand and extra time if using wb finishes, whereas that's not true of oil based?  That seems part of the trick:  where you're gluing two surfaces, the grain raising causes a meshing of fibers from opposite surfaces, which then get "frozen" in place.  Doesn't work very well on end grain because end grain doesn't rise much.  My wood guru showed me the 50/50 water/glue trick, and the chair hasn't come apart in the decade following, so I surmise the ultimate strength of the glue isn't too badly hurt by dilution.  In the end, the water always departs.  Maybe over-dilution results in insufficient glue residue, but I don't know where the limit is.