Every joint answers a question. The dovetail answers tension at a box corner. The dado answers shelf load without hardware. The bridle answers two members meeting at a visible corner.
The mortise and tenon answers the big one: how do you attach a horizontal member to a vertical one, under load, in solid wood, for good?
That's it. That question drives everything about this joint. The proportions, the depth, the fit tolerance. None of it is arbitrary. It's physics dressed up as convention.
What the joint is actually doing
Picture a table apron meeting a leg. Three loads hit that joint simultaneously.
Gravity pulls the apron straight down. That's shear. Someone bumps the table sideways and the rectangle tries to become a parallelogram. That's racking. When it racks, the mortise walls get pulled outward. That's tension.
And then there's the one most books skip: seasonal movement. The apron and the leg are solid wood, oriented in different grain directions. A joint that's tight in your dry January shop will be under swelling stress by July. The mortise construction is a movement management decision as much as a strength decision.
One joint, one set of proportions, handles all four loads. That's why it's been the default for about six centuries.
The one-third rule is not a suggestion
Standard rule: tenon thickness equals one-third of the rail stock. A 1/4" tenon in 3/4" stock.
This is not tradition. It's math.
When you cut a 1/4" mortise in 3/4" stock, you leave 1/4" of wall on each side. Those walls resist racking. Go thinner and they crush, check, or split under dynamic load, especially in softer species. Go wider than one-third and you get more glue surface but the walls become the weak point.
At one-third, both failure modes balance. Maximum glue surface (long-grain to long-grain, the strongest glue joint in wood) without compromising wall integrity. It's an optimization, not a rule of thumb.
The corner exception matters here. Two aprons meeting at the same leg means two mortises competing for interior space. Cut both to full depth and they either intersect or leave a paper-thin wall between them that splits under clamping pressure. Solution: miter the tenon ends at 45 degrees so both can reach full depth without conflict. Mitered tenon ends exist because of geometry, not decoration.
The depth rule and the 1/16" that matters most
Mortise depth: two-thirds to three-quarters of the receiving stock thickness.
The lower bound sets minimum glue surface for shear resistance. The upper bound keeps the bottom of the mortise far enough from the far face to prevent checking at the joint location.
But the most important number in the whole joint is smaller than either of those: cut every tenon 1/16" shorter than the mortise is deep.
A tenon that bottoms out before the shoulders seat will rock under load. The shoulder gap is invisible from most angles. The joint looks assembled. It isn't. It's pivoting on the tenon end instead of compressing against the shoulder face.
That 1/16" gap gives displaced glue somewhere to go and guarantees the shoulders bear fully. It's the difference between a joint that tightens under load and one that works itself apart.
Three ways a mortise fails
The bottomed tenon. Shoulders don't seat, joint rocks. Test during dry-fit: pull the joint apart and check the tenon end. Chalk or pencil marks transferred from the mortise floor mean it bottomed before the shoulders landed. Cut 1/16" off the tenon. Done.
The short mortise. Depth is 1/3 or 1/2 of stock because someone went deep enough by feel. In a table this might hold for a decade. In a chair, where dynamic load works the joint at every shift of weight, a short mortise pulls out in three to five years. Measure. Always measure.
The forced fit. The tenon needs a mallet to go home. Looks tight. What it actually is: mortise walls under compression stress before any external load gets applied. When seasonal movement adds to that stress, the walls check or the joint shears at the shoulder. Correct fit is firm hand pressure. If it needs the mallet, it needs to come back to the bench.
What changes now
Before this: you cut mortises to the dimensions on the plan.
After this: you know why those dimensions exist. The depth is a structural minimum, not a design choice. The one-third ratio is a balanced optimization. The 1/16" gap is non-negotiable. A forced fit is a failed joint, even when it looks like the tightest one in the shop.