I’ve noticed that sometimes a tenon, on a wide board, is split into two separate tenons. See attached picture. And it is shown going into two separate mortises. Why? Seems like it only is making it harder to make. If the idea of a tenon is to make face to face contact for gluing, this lowers the area available. (What is the proper term for that? I know endgrain is weak in gluing, what is the other face of the board? Facegrain??)
Thanks, Edward
Discussion Forum
Get It All!
UNLIMITED Membership is like taking a master class in woodworking for less than $10 a month.
Start Your Free TrialCategories
Discussion Forum
Digital Plans Library
Member exclusive! – Plans for everyone – from beginners to experts – right at your fingertips.
Highlights
-
Shape Your Skills
when you sign up for our emails
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply. -
Shop Talk Live Podcast
-
Our favorite articles and videos
-
E-Learning Courses from Fine Woodworking
-
-
Replies
A full depth mortise 4" wide (an extreme) is very easy to split apart with just a little side force on the tenon member. The mortise shown in your picture is much stronger because of the intermediate bridge of wood holding the two sides of the mortise pocket together.
With this type of mortise, many people would only apply glue to one of the tenons to allow for seasonal expansion and contraction of the tenon member.
Just a brief overview I know, but someone else may be able to provide more detail.
I was taught to split tenons into sizes of 2.5" or less because each tenon would expand and contract. When they are small they will expand and contract less causing less of a chance for failure.
ef,
the tenon is split into two or more sections for three main reasons:
1) to mitigate the effects of shrinkage on a wide beam. The amount of movement that takes place from moisture loss is a certain percentage of board width. Wood in most species moves a greater amount tangentially ("face" grain) to radially ("edge" grain). Since the tangential movement is greater typically, and given that in most cases where wide boards are used the widest face is the tangential one, then the use of a single wide tenon, whether glued or pinned, may result in the tenon splitting when the material shrinks. Dividing the tenon allows some of the shrinkage stress to be taken up in the divided portion, and the shrinkage in the remaining smaller tenons is going to be less, so there is less likelihood of splitting. There is a traditional joiners rule for tenon thickness to width, I seem to recall it being about 1:5; that is, the maximum width for a 1" thick tenon is 5". If the material is wider, the tenon is divided. Some classic 19th century carpentry texts disagree on this point however. In species that exhibit very little movement, like Honduran Mahogany, or Black Locust - especially when using kiln-dried material - this rule is perhaps not so applicable any more. However, in timbers that are pegged, as shown in your illustration kiln drying is not so effective - only radio-frequency drying has been shown to thoroughly dry timbers without excessive degrade, and there are only 4 or 5 such kilns in North America, so it is not a practical or cost-effective choice for many. With timbers that are likely to move significantly from moisture loss (or movement in service), the tenon division is a wise move I think.
2) Tenons are divided (or multiplied, if you like) to increase the available glue area and strengthen the joint. Do the mathematics and you will see. Take a tenon that is 1" thick, 5" wide, and 5" long: the total area for gluing is 2(1"x5")+2(5"x5"), or 60 square inches. Now, divide the tenons, as your picture illustrates, into two, giving two tenons each 1"thick x 2.25"wide x 5"long. That gives a surface area of 4(1"x5")+ 4(2.25"x5"), or 65 square inches. It's not a huge difference when dividing the tenons that way, however if you divide the tenons along their width instead, you get much more surface area: say twin tenons that are each 0.25" thick and 5" wide and 5" long. That adds up to 4(0.25x5")+4(5"x5")=105 square inches. Dividing into 3 or 4 or even 5 tenons continues to increase surface area. The trade off to all this tenon division is threefold: the greater amount of labor required, which affects the bottom line in a commercial setting; poor cut out of the tenons can result in a weaker connection; it is much more difficult to cut out such connections accurately.
3) tenons may be divided, in the case of through-tenons, for purely aesthetic reasons.
The three grain orientations to be concerned about are radial (edge), tangential (face) and end grain. You are correct that end grain gives poor glueing result, since it is like trying to glue the ends of straws together. Radial glue surfaces are optimal since moveemnt in service is the least, in most species, and second best is tangential. In the case of a 90˚ tenoned connection, as you illustrated, the movement of the tangential faces of the tenoned piece is going to be large, and the faces to which they meet in the mortised piece are also going to move a fair bit, but in a direction that is 90˚ different. The radial sufaces in the tenoned piece, on the other hand, press against end grain in the mortised piece, so the connection there is not so good. If gluing, there isn't much point putting glue on those abutments, though a case can be made for 'sizing' the end grain in the mortised piece to minimize moisture loss over time.
I've never seen this joint before, and some of the points made have some flaws, as I see it.
Dividing the tenon allows some of the shrinkage stress to be taken up in the divided portion, and the shrinkage in the remaining smaller tenons is going to be less, so there is less likelihood of splitting.
No, it really doesn't. The shrinkage number should be applied from peg to peg regardless of how many tenons there are in between. If anything, the divided tenon would be worse since the little gap between them is uncovered (butted) end grain. The tenons aren't seperate pieces of wood so they offer no advantage in seasonal movement.
There is a traditional joiners rule for tenon thickness to width, I seem to recall it being about 1:5; that is, the maximum width for a 1" thick tenon is 5". If the material is wider, the tenon is divided.
I think you've got this backwards. The tenon is made wider! This is just basic physics to maintain a strong tenon that can carry the load it is sized for. A tenon in a 10" deep beam (which can take a lot of load) mustn't be 1/2" thick. It should be 2" thick. Two 1x5" tenons aren't the same. Dividing a thin tenon would do nothing structurally.
Tenons are divided (or multiplied, if you like) to increase the available glue area and strengthen the joint. 3) tenons may be divided, in the case of through-tenons, for purely aesthetic reasons.
Gluing tenons is very tricky business. Some glues extrude out of tight joints, others develop no properties in loose joints, That's why structural joints are still pegged. I don't think the increase in surface area warrants this design. A better way to increase surface area would be a twin tenon (where they are parallel, not inline).
I think the picture offered is just bogus i.e. it's somebody's misconception of a twin tenon. The only place I've ever seen a joint remotely close is where you have a really wide side panel of a Queen Anne carcass or a bread board end of a table. In both instances, having three 2" long tenons across a 20" or 30" wide board represents an effort to save on mortise cutting alone. There's no structural advantage, and no seasonal movement advantage in creating a mortise of the sort shown. (Which is probably why we don't see more of these.) I appreciate Sparrow's thoughtful analysis. I just think he/she is trying too hard to find the good in this pointless joint.
Adam
Adam,
Agreed, the drawing is proportionally off, the tenons should have more space between them.
Pointless joint? Not at all. As you mentioned, it is used in the leg-to-end joints of highboys and lowboys; but also sideboards, and writing tables, and others, with deep aprons. Also as you mentioned, bread-board ends, not only on tabletops, but also slant-front desk lids, and blanket chest lids, many sideboard doors (under the veneer) and candleslides. The bottom rails on traditional house doors, being 10" or more wide, and some cabinet doors, (for example, rails above tombstone panels) also will have divided tenons. Face frames for cupboards with round- topped doors will have divided tenons on their top rails. The list goes on.
The structural advantages to the joint occur in the mortised member, where a 20" long mortise in a highboy's leg, for instance, would invite a split from the end of the mortise to the top of the leg the first time the joint is torqued (sliding it across a floor for example), or upon warpage of the case end. Multiple mortises are a way to retain some of the integrity of the solid square, by leaving webs of wood between the mortises.
Often, these joints are mistaken for a simple tongue-and-groove joint; frequently they are housed in a 3/8" or so deep groove ploughed in the mortised member, and the tenons are only partially haunched, leaving a 3/8" long stub between the tenons and at the ends, to fit into the groove, and aid in alignment of the parts, between the tenons.
You are correct, there is typically no allowance for seasonal movement with this joint. But, in the days before central heating, seasonal movement was much less of an issue for woodworkers. The drop in humidity we get in the winter occurs when frigid outside air (which cannot hold much moisture) is warmed in our homes, causing an increase in the amount of water it can hold (lower humidity). When the air remains cold, as in a home "heated" by a fireplace alone, the relative humidity remains much higher in the winter, and results in less shrinkage.
I recently had in the shop, a blanket chest with a heavy molding around the top. The lid was tenoned (three tenons per end) and pinned to the end moldings, in the German fashion, and the top had split between two of the tenons. I repaired the break, and before reassembling the molding onto the ends of the top, I lengthened the mortises and opened the pins' holes in the tenons into slots, to allow for some mevement. Glued the miter at the front and the first tenon, and put the old pins back. That, I think, will be an improvement on the original technique, for our times.
Regards,
Ray Pine
I appreciate your response Adam, and I think we see things differently. I was surprised that you had never seen this type of joint before, so I'll attach a couple or three (four actually) pictures of similar joints from my Japanese texts - hopefully you will not think they are 'bogus'. I am away from home, and thus the bulk of my books at the moment, so I can't post further references, though I have many more, including some in my 19th century texts of western carpentry. In any case, multiple tenon joints are not that uncommon in my experience and research.To address your first point, that dividing the tenon does not improve the performance/integrity of the joint in regards to shrinkage...well, the shrinkage is a given percentage. Let's say it will be 3-4% of the width (tangential orientation of movemement) - on a 10" wide tenon, then, the total shrinkage would be around 0.375". Now as the surface tries to shrink, stresses build up that would normally find outlet at the first point of weakness/stress riser and make for a split or check. Also, with pegs, the tenon is restricted in where it may shrink somewhat. If the tenon is divided, in this example, into two tenons that are 4.5" high, then the 'outlet' for shrinkage is already given by the division, and each remaining tenon will experience the same 3-4% shrinkage. Even with a single wide tenon, running a kerf up the middle would give the shrinkage stresses a place to go.Over a 4.5" individual tenon, that 3-4% amounts to 0.135", or a little more than 1/8". It is more likely that two tenons each accomodating a little more than 1/8" shrinkage each will hold up better than one large tenon trying to accomodate 1/2" of shrinkage, no? The same princple of localizing shrinkage at a division is seen in the German and Japanese practice of runnning a kerf along a boxed-heart timber so that drying stresses are taken up in the kerf - i.e., the kerf widens into a wedge-shape as the material loses moisture, and the remaining surface of the material remains largely check free. Also, of course, such a kerf allows the inside of the boxed-heart piece to lose moisture more evenly with the outer surfaces, thus reducing drying stresses. Perhaps these ideas are bogus to you as well, I don't know, but I can assure you that such are standard practices with proven results.Your point that, "if anything, the divided tenon would be worse since the little gap between them is uncovered (butted) end grain" doesn't make sense, since that portion of end grain you mention is present whether the tenon is divided or not, or whether the tenons have a web/stub between them (a better practice in this sort of joint I would suggest) or not.On your next point, that I have the traditional joiners rule "backwards", well, perhaps I did not explain it clearly enough. Tenons are not simply "made wider", as you assert, since maximum single tenon width is going to be determined from the outset largely by the capacity of the mortised piece to receive the tenon. Typically, if the mortised and tenoned pieces are the same size of stock, the tenon ends up being, by tradition and stuctural logic, approximately 1/3 the thickness of the piece. This does not change when going to a twin tenon as depicted in the original post. I was not suggesting that a tenon in a 10" deep beam be 1/2" thick, or any particular thickness - I was giving a hypothetical division, given a starting point of a 1" thick tenon. Your comment that the tenon "should be 2" thick" - what is this based on? What if the timber is itself 2" thick, or 4" thick, or 6-8-10" thick? Would you suggest the same thickness of tenon in each case? I don't imagine so, though some people tend to make tenons and mortises by the limitations of the narrow range of chisels/mortisers, etc at their disposal. While it makes sense to a point, it makes better sense to me to size tenons based upon the capacity of the mortised pice to receive them. A 2" thick timber that is to be tenoned into a 6" wide timber, need only be shouldered slightly to conceal shrinkage - the tenon can be nearly full thickness in other words. If the pieces are more nearly the same size each, then, the 1/3 rule generally applies.If you were tenoning a 10" x 10" timber into another of the same size, then by the 1/3 rule you would get a tenon of about 3.5" thick x 9" or so wide. I would suggest in such a case that making a pair of tenons 2.5" thick x 9" wide, shouldered 0.5" from the faces and with a 4" space inbetween would make a much stronger joint. Simply increasing the tenon size, as you suggest, would result in the mortised piece becoming unduly weakened.And the rule I mentioned about the 1:5 ratio then...well, if you had a 2" thick tenon, the maximum width of that tenon would be 10". If the timber was the same thickness, but, say, 14" wide, then instead of making a 2"x14" tenon, you would divide them into a pair of tenons each about 2"x6" (as in the picture in the original posting). The same rule holds for really thick timbers, where the tenon thickness becomes a liability due to shrinkage concerns. If you look at attachment 2 you will see a tenon divided both in width and thickness into twin tenons each side.Your comment about 'basic physics' is fine - however the original post was not making any qualifications about the load-bearing capacity of pegged joints, rather he was asking about glue areas and grain orientations. Dividing a tenon along it's width will dramaticaly increase available glue surface. The strength of a joint is given by the available surface areas that abut within whether gluing or pegging- the more surface area the better, though of course, as I menioned in my previous post, this ideal needs to be counterbalanced against time/money factors and difficulty of executing the cutout cleanly. Another important factor governing the shape and configuration of the tenon, and it's pegging, is to design it to resist the forces that it might be subject to - compression vs.tension, bending or twisting. That's a whole other long topic.In the case of pegging it, as shown in the original post, I would perhaps try to stick to a single tenon. However, if I were concerned, say, about the joint opening as result of the receiving (mortised) post twisting, then dividing the tenons as in attachment 2, and running them through the receiving piece and then double wedging each tenon would be one option to deal with that. It's a fairly standard practice in Japanese joinery, though I know to some people that such ideas about joinery, despite their long history of development, are 'excessive', 'purely decorative' or 'impractical'. Not to me - these ideas are entirely rational, though it is also true that aesthetic concerns play a significant role, usually in the direction of concealing the joinery as much as possible.Gluing is tricky business as you say. I'm glad you agree with me that a better way to increase the surface area is to form parallel tenons, as I stated by detailed example in my original post.I don't agree with you take on the picture as bogus, and your comment that, "having three 2" long tenons across a 20" or 30" wide board represents an effort to save on mortise cutting alone" makes not much sense to me either. Attachement 4 above, I believe depicts the situation you mention. It seems that in terms of cutout, making multiple tenons does not represent any effort to save on mortise cutting - it is more cutting. Laying out and then plowing and running a long mortise dado across, either with a powered or hand router is easier than laying out and cutting a series of smaller mortises.Again, I don't feel such joints as multiple tenons are 'pointless', nor was I "trying to hard to find the good" - the good is apparent enough to me.
Thank you all for your comments.
I am basically trying to figure out why/when to use one tenon and when to use two or more tenons.
In my original picture, won't the outer holes not move apart the same amount with one or ten tenons? Are they not forced to move with the whole board they are attached to? Unless they bend or break.
Glue vs. pegged: Seems like smaller tenons are mostly glued and post and beam house are pegged and never glued. And the middle sizes some times have both? So where in size is the cut off?
Sparrow: For the "localizing shrinkage at a division is seen in the German and Japanese practice of running a kerf along a boxed-heart timber so that drying stresses are taken up in the kerf" how deep is the kerf vs the timber size?
To give credit for the picture it comes from:
http://www.taunton.com/finewoodworking/FWNPDF/011147046.pdf
Not that it really matters, just that seeing a picture I could lift to show what I was talking about made asking my question easier.
Thanks, Edward
I happen to be building a lowboy. The design shows 3 tennons in the side and back panels.From my point of view:The top tennon should be sufficient in size to provide any strength the joint needs. This tennon should be glued in place.The other tennons should be sufficient in size and number to keep the panels flat. These tennons should not be glued.---In general, to argue engineering on details like this is worthless since the pieces are not engineered. But ...The engineering of a tennon on a lowboy should account for 2 issues. One - shear at the base of the tennon from loading. Two - shear in the glue (creep issues) due to racking of the joint.
In answer to your question, the kerf goes to the center - the pith, or close to it. The greatest amount of tangential shrinkage is towards the outside of the log, so that is where the need for the kerf is greatest.Also, while it might be generally true that smaller tenons are glued and that timber framed structural connections are not, it is also common enough that small tenons were/are drawbored and pegged in green timber pieces like chairs and other frame & panel assemblies. The Japanese favored glue made from mashed rice paste for furniture and interior fittings, that, along with the use of pegs and other locking mechanisms allowed the piece to be easily disassembled for repair.I have been using yellow glue and pegs/wedges in my furniture, and, wanting to build to last as a primary consideration, I am moving more and more towards making stuff that can be repaired in the future. That leaves the use of pure joinery, or gles that can have their bonds broken, like rice paste glue, hide glue, etc.
Edited 5/28/2006 7:00 pm ET by sparrow
This forum post is now archived. Commenting has been disabled