The Anvil Test

Although the tongue added to the rabbet joint made it stronger mechanically, it left a weak point in the wood that broke on the joint line.

The last box-type joint we tested was a miter, reinforced with cross-grain splines. This held as well as the dovetail under the 26 lb. anvil, but it broke when the heavier anvil fell.

Stout Frame Joinery
We then turned our attention to joints commonly seen in tables and chairs as opposed to boxes and cases. In particular: mortise-and-tenon joints, and some of the methods that have been developed as alternatives such as dowels and pocket screws.

One of the big questions we had was how pocket screws would compare to mortise-and-tenon joints. In this test, the pocket screws did well under both anvils.
While this was impressive, I would be reluctant to use pocket screws instead of mortises and tenons in all applications. As the traditional joint fails, the tenon pulls away from the mortise gradually before letting go completely.

With pocket screws, when the joint is stressed, the screw holds on to the wood until the wood breaks, leaving something that likely can’t be repaired. The long-term performance is also a concern. In many antiques with screwed joints, annual shrinking and swelling crushes the wood surrounding the screw, defeating the screw’s ability to hold.

In testing our drawbored mortise-and-tenon joint, the 3/8″-thick tenon was almost half the thickness of the 7/8″-thick pieces being joined. The amount of wood remaining next to the mortise was thin enough to be fractured on impact.

A larger joint was then tested, with a 1/2″-thick tenon  in 1-3/4″-thick material. The tenon was pulled tightly together with a 1/4″-diameter oak dowel. The thicker components, along with a tenon less than one-third the thickness of the stock, produced a joint of great strength.

A joint will fail at the weakest point, even if it is in the surrounding wood, as seen in this biscuit joint.

A dowel joint of the same-size components, however, did not hold together. The dowels held when the grain in the dowel was parallel to the grain of the wood. In the other direction, where the grain of the dowel was 90? to the grain of the wood, the glue joint failed and the dowels popped out of their holes.

Seasonal wood movement is also a long-term issue with dowel joints. As the wood and the dowel expand and contract, the dowel and the hole change shape from round to oval. Over time, only a tiny area of contact between the dowel and the wood remains.

Designing for Time
Nearly any joint can work short-term, but if your sights are set higher than that, consider not only how you will join your components, but also the sizes and scale of the components.

In preparing these joints for testing, I used sizes and proportions that I would normally use. Most of these joints I would still make the same way. If a biscuit or rabbet location was changed in location or proportion, the failure would likely still occur, but on the other side of the joint. But I don’t think I will ever again make a mortise that is more than a third of a piece’s thickness.

Understanding what the wood will likely do, and what to expect from a glue in any given situation is a lesson well worth going to a few extremes to learn, though you can do this without an anvil.

When you try a new method, make some practice joints and see how well they hold together. Sam Maloof once tossed one of his early chairs from the roof of his garage just to see how strong it was. He learned something about his joinery methods that day. And that’s the same lesson here: Our successes emerge from the splinters of our failures. PW

Notice that the wood failed at the short grain on the tails, and that the glue held on the end grain between the pins.

Comments or questions? Contact Bob at 513-531-2690 ext. 1327 or robert.lang@fwmedia.com. Visit his website at www.craftsmanplans.com