The bridle joint, the box joint, and the sliding dovetail don't belong cleanly to one structural family. They borrow geometry or function from two. That's not a defect — it's the point. Each one exists because the problem it solves is itself a hybrid.
The bridle joint is M&T with the wall removed
Take a standard mortise-and-tenon. Open the mortise wall at the end. Now you have a bridle joint: tenon with two shoulders, U-shaped slot in the receiving piece, open at the top.
The racking resistance is the same. Two long-grain cheeks, two shoulders — identical geometry to a closed mortise in the joint plane. What you lose is tension resistance. A closed mortise captures the tenon on all four walls. A bridle slot doesn't. The tenon could theoretically pull straight out. Glue handles this fine for most furniture applications, but the bridle is not where you put joints under sustained tension.
What you gain is speed. The slot cuts on a table saw: cheeks as rip cuts, walls as crosscuts or a dado stack. No drilling, no chopping, no mortising machine. For face frames, trestle structures, and rail-to-post connections where you want M&T-class racking resistance without M&T-class setup time, the bridle is correct.
Proportions: tenon thickness = 1/3 of receiving stock. Keep slot walls at least 3/8" thick on each side. The end grain at the open top is fragile and blows out if you get greedy.
The box joint is not a dovetail
People compare them. They're not the same thing.
A dovetail resists tension mechanically. The angle of the tails creates interlock — pull the joint apart and the geometry fights you. Unglued dovetails held for centuries before modern adhesives because they didn't need glue to stay together.
A box joint's fingers are parallel. Pull the pieces apart and they slide straight out. No mechanical interlock. The box joint's strength comes entirely from glue.
What it offers instead is surface area. The interleaving fingers present long-grain faces to each other across the full depth of both pieces. Long-grain-to-long-grain PVA is strong enough that a well-fit box joint fails in the wood before the adhesive lets go. It's a strong joint. It's just not a mechanical one.
The reason to choose it over the dovetail is speed. A table saw jig with a dado stack produces identical box joints quickly. Once dialed in, variation is zero. Hand-cut dovetails require skilled layout, careful fitting, and time. For drawer boxes, tool cases, shop storage — anywhere speed matters more than the dovetail's credential — the box joint is the right answer.
Fit is everything. The fingers need to slide together under firm hand pressure. Too tight and humidity will lock them during glue-up. Too loose and you're gluing a gap, which PVA handles poorly. Cut test pieces in actual stock. Sneak up on the fit.
Finger width: match your stock thickness for a balanced proportion that also produces practical glue surface. Minimum 3/16" in hardwood, 1/4" in softwood. Thinner fingers increase glue area and visual interest but demand more precision at every step.
The sliding dovetail is a housing joint that won't pull out
A standard dado resists racking in one axis. It doesn't resist tension. A dadoed shelf can pull straight out of its housing if the fit is loose or the load is direct. The sliding dovetail is the fix.
The geometry: one member has a trapezoidal male tenon — wider at the base than the top. The receiving piece has a matching groove. The male slides in from the open end. Once seated, the angled walls resist withdrawal the same way a standard dovetail does. Tension without glue.
This matters for construction. A sliding dovetail assembled dry will stay. Useful for adjustable shelving, breadboard ends, and anything meant to come apart. Glue it permanently and you get both mechanical lock and adhesive strength — the strongest version of a housing joint available.
The movement rule applies here, critically. If the sliding dovetail runs cross-grain, as it does in a breadboard end or a case divider across a solid panel, the wood moves seasonally. A fully glued cross-grain sliding dovetail fights that movement and cracks the surrounding wood. Glue cross-grain sliding dovetails at the center only, or not at all. The mechanical geometry holds the joint. Movement handles itself within the groove.
To cut: router with a dovetail bit. Cut the male tenon first with a two-fence setup, then cut the groove to match. The tenon should slide by hand in dry conditions. Test in actual stock — not a separate offcut of the same species. A tenon cut at 12% MC in your shop will lock when installed at 6%.
Stopped vs. through: a through dovetail shows the tenon end at the panel edge. Fine for shop furniture. For visible-face work, stop the groove before the edge and chisel the corner square. More work. Also not optional.
Why the hybrids exist
The four structural families from Post 224 are conceptually clean. Real problems aren't always.
Bridle: M&T racking resistance at lower setup cost and lower tension resistance. Table saw joinery. Right for face frames, trestles, corner joints where the open top is hidden or irrelevant.
Box joint: Dovetail-class corner strength through glue surface, not mechanical interlock. Jig-dependent. Right for drawer boxes, cases, shop storage, anywhere you're optimizing for speed.
Sliding dovetail: Housing joint placement with tension lock added. Right for breadboard ends, fixed dividers, shelving without fasteners.
None of them replace the parent families. Each one exists because a specific problem needed a specific answer. Know the problem before you pick the joint.