Drawer Bottoms: Glued, Floating, or Captured? Racking Tests Settle It

Yo, let me tell you a story. There's a slab of plywood leaning against a brownstone stoop on my block, been there three weeks, and every time I walk past it I think about drawer bottoms. Why? Because last winter a buddy of mine handed me a kitchen drawer that had gone rhombus. Pulled it out for the hundredth time, and instead of a rectangle it was a parallelogram. Contents slid to the low corner like change in a couch. And his question was the same one that starts a fight in every woodworking forum every six months: do I glue the bottom in, float it, or what?

Turns out the answer's been tested. With actual machines. Let's break it down.

Racking Is the Enemy, Not Pull-Apart

Here's the number that changed how I think about drawer boxes. WOOD Magazine ran what they call their Wood Joint Torture Tests, samples of ½" poplar and yellow glue, lab-tested at Iowa State University. The average drawer joint survived about 1,489 pounds of pull-apart force. Shearing force? The force that pushes the box out of square, the racking force? 80 pounds.

Read that again. Nearly nineteen to one. Every argument about which joint is strongest is mostly an argument about a failure mode that almost never happens. Your drawer doesn't get pulled apart. It gets knocked out of square, over and over, a few pounds at a time. That's the fight you're actually in.

Which means the question "glued, floating, or captured?" isn't a wood-movement question first. It's a racking question first. And that flips the whole traditional answer on its head.

Why Plywood Changes the Glue Debate

The old-school floating bottom exists for one reason: solid wood moves. Popular Woodworking has the math, based on the USDA Forest Products Lab's Wood Handbook. Movement equals width times annual moisture-content change times a species coefficient.

Worked example: a 24"-wide quartersawn yellow-pine panel with a 3-point swing moves 0.127", about 1/8". Flatsawn of the same species? 0.189", closer to 3/16". That's real movement, every year, forever. Which is why the traditional drawer bottom leaves the back edge uncaptured, per Fine Woodworking's forum tradition: grain runs parallel with the front, the front edge often gets glued into its groove, and the back edge just gets nailed to a back that's shorter than the sides. The panel breathes without pushing the drawer apart.

But plywood? Plywood doesn't play that game. Cross-laminated, dimensionally stable. Fine Woodworking's Andy Rae says it straight up: since your bottoms are plywood, glue each one into the drawer grooves. It significantly strengthens the corner joints. WOOD Magazine's lock-rabbet article says the same, because you can glue the bottom in without worrying about expansion and contraction.

And here's the payoff: a glued-in plywood bottom is a shear panel. It's a giant brace locking the box square, which is exactly the 80-pound problem. WOOD even tells you to measure corner-to-corner diagonals at glue-up and clamp across the long diagonal. The glued bottom locks that squareness in permanently. One sheet of 1/4" ply is the cheapest anti-racking device in woodworking. In my second-bedroom shop, cheapest-and-actually-works is the whole religion.

The honest downside, per Rae: gluing it in makes future repairs harder. If a bottom fails, you might be rebuilding the drawer. I'll take that trade.

How Thick? The Practitioner Ladder

Now, nobody's lab-tested sag deflection for bottom thicknesses at various spans. What exists is the consensus of cabinetmakers who've built thousands of drawers, and it's a clear enough ladder:

• 1/8": the home-center standard, and WoodWeb pros call it about as cheesy as it gets. If you're stuck with it, keep the drawer under 12" wide.
• 1/4" plywood: the norm for kitchen drawers, good up to roughly 24". One WoodWeb pro ran a stand-in-the-drawer demo where the thing collapsed, and the 1/4" bottom didn't fail. The slide suspension did.
• 3/8" to 1/2": for the wide, heavy stuff. Fine Woodworking forum builders doing a 45-drawer kitchen with 26" to 40" wide pot-and-pan drawers stepped up to 1/2" Baltic birch bottoms in 5/8" sides, glued all around, with a 31/64" × 3/16" dado.
• Andy Rae, for a 22¼" × 26" drawer, says 1/4" multi-ply should suffice but sounds "tinny," and recommends 3/8" or 1/2" for anything but the smallest drawers.

There's a real debate about whether 1/2" is overkill. One Fine Woodworking guy calls it "awfully thick" and admits sheer cowardice keeps him from capturing bottoms on big drawers. Other pros standardized on 1/2" for everything just to avoid stocking two thicknesses. I'm with the stock-one-thickness crowd. Apartment shop, no room for a plywood library.

One constraint people miss: groove depth. In 1/2"-thick sides, a 1/4"-deep dado leaves 1/4" of wall, which caps how thick your bottom can be. Another pro running dovetailed maple drawers says 3/8" is the widest groove that won't cut into the pins. Standard practice locates the bottom groove 1/2" up from the bottom edge of the parts. If your sides are 1/2" stock, your bottom thickness ceiling is basically set.

Hardboard: Only If You Hate Money Slightly Less

Tempered hardboard shows up as drawer bottoms in cheap furniture, and Fine Woodworking forum folks report 1/4" hardboard working great for templates and drawer bottoms. WoodWeb pros reject 1/8" Masonite as way too cheap.

Match the Bottom to the Slide

Here's a spec that ties it together. Accuride's 3832E, the classic light-duty full-extension slide, is rated for 100 lbs max load, 24" max drawer width, side space 0.50", running $16.15 a pair on their site. General guidance from the drawer-slide crowd: side-mount ball-bearing slides commonly run around 75 lb capacity, undermounts around 50 lb because of the clip mechanism.

Now do the math on the mismatch. A 100-lb slide under a 24"-plus span with a 1/4" bottom, fully loaded with plates? That's exactly where the sag complaints come from. The slide can carry the load; the question is what the bottom does between the slides. Match bottom thickness to the load class you're actually building for. Wide drawer, heavy contents, slides rated for it? 1/2" bottom, done.

The Joints, Ranked by What the Test Actually Found

WOOD's torture test gives us a real ranking, and it's got some surprises:

• Box joint: number one for pull-apart, thanks to 3¾ square inches of glue surface in ½" stock, where the glue tore out wood fibers. But its shear strength was only 4% of its pull-apart strength, so WOOD says reinforce with #18 × 1" brads if racking matters. Great for heavy utility drawers.
• Lock rabbet: the tablesaw workhorse. Glue bonds outlasted the surrounding wood; failures were edge-to-edge cracks in the mortised piece. Strengthen it by positioning the mortise away from the drawer side's end and keeping glue surfaces from burning.
• Through-dovetail: the only joint where glue failure didn't mean total failure. After the glue cracked, the pieces stayed mechanically locked. That's the repairability argument, and it's real.
• Sliding dovetail: weaker than the lock rabbet in both pull-apart and shear, it fails when the 1/8" strip beside the groove breaks off, and it complicates inserting a floating bottom. Skip it for drawers.

One wrinkle worth being straight about: WOOD's own prose says the lock rabbet is "not as strong as a dovetail joint," while their test ranking put the lock rabbet above the through-dovetail, and their video claims it proved stronger than even dovetails. I read the test as the data and the prose line as editorial caution. Either way, both joints live or die by racking, and the glued plywood bottom does more for either one than agonizing over the joint choice does.

What I'd Build Tomorrow

Small drawer, under 12" wide, light duty? 1/4" ply bottom, glued in, lock rabbets or even dados. Standard kitchen drawer to 24"? 1/4" bottom glued in, lock rabbet joints, squaring the box at glue-up. Wide pot-and-pan drawer over 24"? 3/8" or 1/2" Baltic birch, sized to your groove depth, glued all around, and don't cheap out on the slides.

The one-sentence version: pull-apart strength is nearly free, racking is what kills drawers, and a glued-in plywood bottom is the cheapest shear brace you'll ever install. The slab on the stoop is still there, by the way. Someone's going to beat me to it.