Subscribe | Give a Gift | Subscription Customer Service
How We Are Different Order Back Issues Interactive Glossary Weblog Contact the Staff Subscribe to Woodworking Magazine

Woodworking Magazine Weblog

Posted 3/31/2006 in All Weblog Posts | Joinery


In China, 2005 was the year of the rooster. In our shop, 2005 was the year of the anvil. We built a guillotine out of framing material and dropped anvils of three weights on joints to see how they fail.

We learned a few things. First: You can get paid for doing juvenile stuff with anvils. Second: Modern PVA glues (yellow glue) are a lot stronger in end grain applications than woodworking wisdom suggests. And third: How you size the parts of your joint (thickness, width and length) has a lot to do with how sturdy it ultimately is.

Nowhere was this more evident than with the venerable mortise-and-tenon joint. Changing the thickness of a part of the joint, such as the mortise wall, could greatly weaken or strengthen the joint under the crush of the anvil.

There are some well-worn rules about how to scale a mortise-and-tenon joint, and they are worth thinking about the next time you lay out a tenon. Let's take a look:

Tenon thickness: This one gets debated a lot, and with good reason. Traditional texts say the tenon's thickness should be one-third the thickness of the stock being mortised (an important distinction). So if you are joining two pieces of 3/4"material for a door, the tenon should be ¼" thick. If you are joining a 7/8"-thick apron to a 1-1/2"-thick table leg, the tenon should be 1/2" thick.

Some modern texts say the tenon should be one-half the thickness being mortised – not one-third. My opinion is that this difference relates to the tools being used. If you mortise by hand, with chisels, the one-third rules makes more sense in my experience. Using a 3/8"-wide mortise chisel on 3/4"-thick material invites destruction in many cabinet woods.

But if you've ever used a hollow-chisel mortiser, then you've probably been amazed at the difference in performance between the 1/4" chisels and the 3/8" chisels. The 1/4" chisel gets clogged up much more easily because its escapement is much small. Plus, the hollow-chisel mortiser doesn't put the kind of lateral strain on your work that hand-mortising does. So a 3/8"-wide mortise works with machines.

Tenon length: The general rule is that the minimum tenon length is five times its thickness. So a 1/4"-thick tenon should be 1-1/4" long. Of course, if you look at antique furniture, you see this "rule" violated – or maybe the furniture was made before they made the rule. Longer through-tenons are the rule of the day in much 19th and 18th century work. These are wedged tenons, generally. Check out George Ellis's "Modern Practical Joinery" for a trip through the land of the through-tenon. Personally, I try to follow the "five times the thickness" rule for most cabinetwork. But when I'm building something that will encounter more wracking forces (such as a dining table), I go long.

Tenon width: This one is more complex. The rule in Ellis's book is two-fold. First, make the tenon one-half the width of the rail you're cutting it on (a 2"-wide rail would get a 1"-wide tenon). Second: If that tenon's width would be greater than six times its thickness, then you should split it into two (or more tenons). Example: You want to cut a 1/4"-thick tenon on a 6"-wide rail. Ellis's rule says that your tenon should be 3" wide. But a 3"-wide tenon is greater than 1-1/2", which is six times the tenon thickness. So you have to break that tenon into two 1-1/2"-wide tenons.

Is your head swimming yet?

This rule seemed odd to me at first. The tenons it made seemed too narrow in width, which would allow the corners of your to frame warp (or cast) over time. But when you look at Ellis's illustrations, it makes sense. He shows all his tenons with a short haunch that runs the entire width of the work. Ah!

And what about double mortises, such as when you join a narrow drawer rail to a leg in a chest of drawers? This drawer rail is usually somewhat squarish and stout, and it doesn't follow the rules laid out above – you don't need a double tenon.

Some sources seem to suggest that the double tenon can be made for convenience. You might not have a 1/2" mortising chisel for that 1-1/2" drawer rail. But you have a 1/4" mortising chisel (of course you do!). So making two ¼" mortises that are set by the tool are easier to make than a 1/2" mortise that you would have to make with an odd-size chisel. (It's a theory – not much more than that.)

All this math and theory and contradiction is enough to make you want to smash something, with an anvil.

Christopher Schwarz

4/6/2006 2:06:51 AM (Eastern Daylight Time, UTC-04:00)
On Tenon Thicknesses:
This tenon thickness thing here is wrong. See below.

First, the reason in a door to have 1/4" thick mortises in a 3/4" thick door is that your panel grooves are 1/4" square and can't get thicker without weakening the door. The tenon and haunch fill the groove cut for the panels. This was true in the days of grooving planes and is still true today with dado heads and shapers.

On Tenon Lengths:
The long tenon is a good idea. You make the tenon as long as possible without weakening the mortised part. If you look at Victorian house doors and pre-industrial furniture, asian or western, the tenons are long, often through mortised and wedged where appropriate and still are together today. They are not the cope and stick shaper cut doors, which if you are lucky, are then doweled together. Modern construction is often dictated by the bean counters and the generally low skill of low cost industrial labor helped by better glue. The total width of the tenon is the width of the rail less the panel groove(s). The interior reduced length area that this author calls a haunch is good, but Ellis calls it franking. The Ellis book is excellent and Ellis is right.

On double tenons at drawer dividers like traditional tables such as illustrated in my Beginning Class Table on the PAAS Website.
This is wrong. The drawer rail has width at a right angle to the post it is tenoned into and a double tenon is better, because it gives more glue & mechanical surface, keeps the post stronger and avoids the warp problem that franking is doing on wide flat rails planar to the posts (stiles), as in doors.

On using a smaller hollow chisel mortiser chisel for making larger thickness mortises:
Making a 1/2" mortise with two cuts of a 1/4" hollow chisel will tend to result in a mortise with cheek walls out of parallel, since the chisel will deflect into the area already cut, now occupied by air.

On mortise and tenon joint design generally:
What is missed here is that you aren't just designing tenons, you are designing mortises and furniture that are a combination of both joint parts. It is important that the post not be weakened by the tenons coming into it. So to get a long tenon, and not weaken the post/ table or chair leg with tenons coming into it at right angles, you try to leave mass in the post between the mortises. When you have a 1-1/2" thick, square post met by a 7/8" thick rail, a 1/4 or 5/16" thick tenon will give you more length, still leaving more wood in the post between the mortises, than if the tenon were thicker. You design joints for strength in the plan view particularly, drawing everything you need and maximizing the strength of each part. Your joint is a compromise, but a planned one.

My teacher, Lance Patterson at the North Bennet Street School, always said to leave some meat between the mortises. Deepening the mortises in a post into one L shaped mortise and mitering your tenon ends looks pretty in a drawing, but it leaves stalagmites of dried glue in the second mortise of a two stage glue up and it weakens the post a lot.

I hope that this helps.

John McCormack
7/19/2008 3:46:40 PM (Eastern Daylight Time, UTC-04:00)
I had no idea that there are certain rules. Very interesting article I must say.
Name
E-mail
Home page

Comment (HTML not allowed)  

Enter the code shown (prevents robots):

 Copyright © 2005 F+W Publications Inc. All rights reserved.