Joining Wood

 

How many things are single pieces of wood?
I can quickly visualize salad bowls and spoons, baseball bats,
slicks. . . it's not a very long list. As you compose such a list, you soon realize how many producis are made of more than one piece of wood, somehow joined, held or stuck together, or of pieces of wood attached by male-rials other than wood. The list of products made of joined pieces is endless, and the more functionally elaborate and important the things—furniture, boats, houses, pianos—ihe more complex and varied are the means used to join their components.
More often than not, performance depends not so much on the physical or mechanical properties of the separate parts, but upon how successfully ihey are fastened together. Wooden items are rather more prone to fall apart than to break. It is the rule rather than the exception that the weakesl point in any wooden construction is at a joint. The successful woodworker focuses not just on each piece of wood, but on the interrelationship of pieces of wood and especially on how they are joined logelher.
In our modern world, it is important to sense the historical evolution of wood use. Some functional forms, like the walking slick, need no improvement, nor are they liable to become obsolete. But new and more complex forms of wood are constantly being invented. These new materials and products owe their success mostly to some kind of marriage between wood and other materials. The evolution of products is paralleled by the evolution of systems for fastening the parts involved. New methods of fastening evolve in response to our changing resources and as a reflection of changing technology, and each advance in turn paves the way for a new array of products.

The elements of joints
I shall not attempt to catalog or discuss all the joints in woodworking. Rather, 1 shall discuss the four critical considerations lhat determine the success ofl any given joint. One of the four may be of overriding importance in some particular situation, or any one may be interrelated with one or more of the others. These four basic considerations are the stress system involved, the grain direction of the joined parts, wood movement in re-ing parts.
The stress system is what the joint is being asked to do mechanically, as a consequence of its being part of a
sion, shear or racking (bending), and usually the great difference in the strength of various joints depends on the stress situation. Most compression joints give little trouble, and shear stress is not too difficult to overcome. Joints subject to tension and racking are usually the most troublesome. Figure 1 shows joints under representative stress systems. Although it is not usually necessary to figure the loads precisely, one must have a general grasp of the direction and relative magnitude of stress in order lo design and construct good joints. This analysis can come only from realistic examination of the structure tn which the joint will be used, and the loads it is likely to encounter in service.
The second element is the grain direction in each mating surface of the joint, as related to the stresses involved. For example, the most difficult surface combination to fasten is end grain to end grain (1A). It would not matter if the load were exclusively compression, but this is most uncommon without some racking stress also. As a result, limber framers who must often lengthen stock to span a space have evolved an elaborate system of scarf joints, many with mechanical interlocks, whose principal purpose is to convert mating end-grain surfaces to long-grain surfaces. End grain to side grain (1B) is a very common situation, which can be accomplished quite satisfactorily if all factors are considered. When stressed in compression, such a joint is usually limited by the perpendicular-to-grain compression strength of the side-grain piece. When stressed in tension, the fastening to the end grain may be difficult, and when under racking slresses, either part may be the limiting factor. The solution usually involves a mechanical interlock formed on the end-grain piece or made by adding a third piece of wood to cross the joint. Side-grain to side-grain joints (1C.D) can be as strong as the wood itself when they are adhesive-bonded, even when the grain directions of adjacent members are not parallel. But here the third element comes strongly into play, the dimensional properties of the wood in response to
changing moisture conditions.
Dimensional change in response to moisture is usually no problem in the case of end-grain (1A) or parallel side-grain to side-grain joints (1Ñ) because the orientation of the growth rings can be the same in boih pieces. In these same joints, if the growth rings are not similarly oriented, the difference between radial and tangential movement might cause visual difficulties, if not structural problems.
In perpendicular side-grain to side-grain joints (1D) and in end-grain to side-grain joints (1B), the conflict between dimensional change along the grain and across the grain (especially where tangential direction opposes longitudinal direction) may become more important than the stress/strength of ihe original joint. The potential self-destructiveness of such joints should always be anticipated, A lap joint (1D), for example, might be very strong when glued, but it could self-destruct as a result of dimensional change.
The last element is the surface condition of the mating parts, including the precision of fit and evenness of bearing, the trueness of the surfaces, and the severity and extent of damage to cell structure resulting from the surfacing process. Uneven surfaces may concentrate enough stress to overcome the strength of wood or glue, while the same joint would survive very well if it had fit properly. Poorly fitted parts may also allow unintended motion, ending in destruction. It is quite common for joints to fail not along a glueline, but in adjacent wood tissue that had been mangled by poorly sharpened tools or bad woodworking technique while the joint was being cut. In joinery, the combinations of stress, strength, dimensional change and surface quality are endless. But careful analysis of the factors involved in each joint will develop your judgment and minimize your mistakes. I'll discuss a few of the more common joints, to suggest how you might approach your particular joinery problems.

 

Basic types of joints
The term joint has various and sometimes overlapping meanings. In its broadest sense, it refers to any junction between two components or materials. Without reference to any accompanying means of fastening, joints can be characterized on the basis of the grain orientation of the mating surfaces as end to end, and end to side or side to side. Flat mating surfaces are loosely termed butt joints, although this term usually suggests either end-to-end butt joints (1A) or end-to-side butt joints (1B)- Side-grain to side-grain joints are more clearly designated as edge joints when the mating grain directions are parallel (1Ñ) or as lap joints when the grain directions are perpendicular (1D). (The special case for miters and scarf joints would be termed cross grain to cross grain.)
Obviously, such joints have no structural integrity without some means of holding or fastening the pieces together. I prefer to consider three basic types of joints or a combination thereoi:
Worked joints, where the wood is physically interlocked or fitted together;
Fastened joints, where a "third party," the fastener, is attached mechanically to both components;
Glued joints, where an adhesive forms a continuous bond between two pieces by surface attachment.
Each type has its ancestor far back in history The first worked joints might have been tree stems with forks or splits, interlocked with others, or perhaps circular branches inserted into knotholes. The first fasteners were probably thongs of hide or vines, used to lash wooden parts together. Who can say when the first crude metallic nail replaced a wooden pin to hold two pieces of wood together? Adhesives were probably discovered when residues from cooking meat accidentally stuck two pieces of wood together. Modern fastening systems are refinements of each of these, in complex and igenious combinations.

Worked joints
Creating interfitting or interlocking shapes to provide strength and integrity in a joint has been a hallmark of the woodworking tradition. Alihough modern machines can simplify the making of fitting parts, the pride of accomplishment in hand-execuiion of difficult and beautiful joints will always be among the challenges, pleasures and rewards of woodworking.
The mortise and tenon—Fastening of end-grain to side-grain joints can be accomplished with a high level
of success using the mortise and tenon. The basic join is fashioned by forming the end-grain component, the tenon, into around or reclangular cross section and inserting it into a hole, or mortise, of the same size and shape in the side-grain component. By closeness of fit alone, this joint can have positive resistance in compression, shear and racking, in which cases the strength of the wood in compression perpendicular to the grain limits movement in ihe joint (2). The mortise and tenon is commonly associated with frame construction. In chairs, round shapes are usually used. In window frames, paneled doors and other squared frames, the rectangular form is common.
The mortise-and-tenon joint has mechanical restraint in every direction except direct withdrawal of the tenon from the hole. Although this is the way a joint usually comes apart, it most often does so only after damage due to racking. A racking load on a rectangular frame acts to deform it diagonally. Under racking loads, the tenon pivots in the mortise.
The basic "dry" joint can be improved in several ways. The most obvious is to glue it, thus adding side-grain to side-grain shear resistance along the mating mortise-and- tenon cheek surfaces to oppose the rotational effect, A second approach (3) adds a shoulder to the tenon, giving additional bearing surface to share the compressive resistance on the outside of the mortise.
In our earlier discussion of beam strength (Chapter 6), it is pointed out that the greater the depth of a beam, the lower the axial stress developed when the beam is loaded. The mortise and tenon (end-grain to side-grain joint) can be thought of as a cantilever-beam attachment, so increasing the height of the "beam" will reduce stress (4). This also increases the surface area of the cheeks, and thus the gluing area. Lengthening the insertion depth will further increase the resisting glueshear areas.


 
  2—A mortise-and-tenon joint has posilive resistance to compression, shear and racking, even without glue, but no resistance to tension.

 

3—In a shouldered tenon subjected to racking, rotation is restricted by shear strength in the glueline bonding the cheeks of the tenon. The bearing surface of the shoulder carries corn-may open so (ha! shear in the upper area of the tenon (S) will be greater than below (S'). 4—A mortise-and-tenon joint func-lions like a cantilever beam, so increasing the heighl of the tenon will reduce the stress on the joint, if the force (M) remains the same.

At the same time, the improvement in mechanical advantage obtained by increasing height is offset by increased dimensional conflict between longitudinal and transverse grain orientation. Some careful compromises must therefore be worked out. For example, since tangential shrinkage (and swelling) is about twice the radial movement, it is best to have the radial (rather than tangential) grain direction of the tenon matched to the long dimension of the mortise. It is also better to have the radial direction of the mortise matched to the longitudinal direction of the tenon. In Figure 1, joint A would be best, since radial/longitudinal grain direction is matched along the mortise cheeks both vertically and horizontally, and tangential grain is matched perpendicularly to the plane of the mortise. Joint D has the worst dimensional conflict. As the height of the mortise (along the grain) is increased, joint survival increasingly depends upon moisture control. The usual solution to dimensional conflict is to divide
the joint into multiple sections (2). By keeping the dimensions
of each tenon within limits, dimensional conflict can be reconciled by mechanical restraint. Thus there is considerable advantage, especially in wide or thick stock, in multiple tenons or multiple splines. These considerations also emphasize the importance of well-made, well-glued joints designed for mechanical restraint of the dimensional conflict as the wood moves, as well as for initial strength.
In summary, the rectangular mortise and tenon seeks to join side-grain to side-grain gluing surfaces and to offer optimum mechanical resistance by maximizing the depth of the joint while still surviving dimensional conflict. For example, in Figure 3, joint  offers three distinct advantages over joint A. First, the depth of the individual tenons more than offsets the loss of width. Second, the height of the glued side-grain surfaces is increased. Third, the number of side-grain to side-grain interfaces is multiplied.
The round mortise-and-tenon joint has advantages and
disadvantages. By turning the tenon on a lathe and by drilling the mortise, it can be produced with a high level of precision. However, poor tool geometry or poor
sharpening commonly leaves drilled hole surfaces and
tenon surfaces in poor condition. The joint may therefore have weak mating surfaces. Also, the proportion of side-grain to side-grain gluing surface is somewhat limited and cannot be improved by increasing the dowel diameter. The best side-grain to side-grain gluing area is located at the mid-depth of the dowel, where it can do the least for racking resistance.

 
3—The multiple tenon (B| is preferable to a single wide tenon (A). Tenon depth more than offsets the loss of width, and the increase in tenon height and the number of side-grain to side-grain interfaces greatly improves the strength of the joint.

 

1—The best possible orientation of growth rings in a mortise and tenon is with radial/longitudinal grain direction matched along the mortise cheeks both vertically anil horizontally, as in A. Joinl D, the worst orientation, is apt to split.
 
2—Although joints A and B have the same amount of  wood in each component, joint B has triple the bonding surface and more balanced dimensional restraint.
Moreover, it seemsapparent that gluing can do little to improve lateral shear strength. Much of the integrity of a round mortise-and-tenon joint depends on racking strength (that is, resistance to pivoting) from surface bearing. It is therefore critical to maintain the depth-to-diameter ratio above a certain level (about 3:2) (o distribute the stress as much as possible. Most joint failures seem to be associated with shallow insertions.
A further complication of the round joint, the result of bolh poor gluing characteristics and dimensional conflict, is the development of compression-set loosening at the top and bottom edges. This emphasizes the importance of well-machined surfaces and of matching tangential-grain to tangential-grain directions. Experiments with the moisture-cycling performance of joints indicate the high degree of success achieved by using a good glue to restrain mechanically at least some of the dimensional conflict. Some additional reduction of compression-shrinkage loosening of tenons may be afforded by pre-splitting the tenon (4). As compression shrinkage develops in moisture cycling, the split can open to fur-mechanism contributes to the success of wedged tenons. Although the wedge is primarily intended to produce
lateral pressure to the glue surfaces and perhaps also to
splay the tenon for a dovetail-style mechanical lock, it may play the more important role ot providing a stress-relief slot that helps the glueline survive.

Dovetail joints
Nothing is more symbolic of the woodworking tradition than the dovetail joint (5). It is a strong and beautiful way to execute the corner side-grain to end-grain joint and is commonly used in carcase construction. The joint consists of interlocking tails and pins, giving it strength in tension along the tail member but not along the pin. It
should therefore be oriented to resist tension against
the tails. In a drawer, for example, the tails should be in the drawer sides, the pins in the drawer front. In case construction, the pins should be in the sides and the tails in the top to prevent the sides from spreading. Although the joint strength results from the wedging action of the tails against the pin faces, the joint is held in place principally by gluing the side-grain to side-grain mating faces of the tails and pins. In designing ihe joint, the slope of the tails must be a compromise. If the angle is not great enough, the wedging-locking action will be lost. If the angle is too great, the splayed tips of the tail will be too fragile, and a component of end grain will be introduced. This impairs the side-grain integrity of the gluing surfaces. An angle of 11° to 12° has proven satisfactory. The joint strength depends largely on the success of the glue bond between the side-grain faces (the end-grain areas behind the pins and tails cannot be depended upon for any substantial contribution to the strength) and the shear strength (parallel to the grain) of the wood of the tails. Joint strength therefore increases
as tre number of tails increases, as long as anough wood
remains across the narrow part of the tails. However, if the joints are cut by hand, the added labor should also be considered in determining the number of pins and mils per joint.
When a large number of pins and tails can be cut with precision, as is possible with machines, the glued surfaces alone can develop adequate strength. In fact, the joint can relinquish the wedging taper and have straight tails. Thus evolved the finger joint or box joint, which gains its strength from the many closely fitting side-grain to side-grain gluelines.

 


 

4—In this yellow birch joint, the tenon was split radially and tangentially before assembly. After moisture cycling? compression shrinkage has developed entirely in one direction, opening the radial split, while the tangentioal split remains tight, and the glueline has remained intact.
5 - The dovetail, a standard carcase joint, has strength in tension along the piece with the tails.

Miter joints
Parallel side-grain to side-grain miters (1Ñ) make a very efficient corner joint. Miter joints where the grain direction meets at a 90° angle are attractive but present serious technical problems. Because of the difference between dimensional change along and across the grain, the joint may open if the moisture-content change is great or if the members are wide. If nailed perpendicular to the outside face of either member, nailing into the end grain must penetrate deeply to ensure adequate holding. Gluing miter joints is of marginal effectiveness because of ihe large component of exposed end grain. Doweled miters, splined miters or combination lap-and-miter joints can improve strength by providing side-grain to side-grain gluing surface.


 
1—End-grain mitered joints canbe strengthened by dowels (A) or by splines (B). A properly glued side-grain miter (C) will have adequate strength by itself.

Doweled joints
Dowels are cylindrical wooden rods used in a number of ways to fasten and strengthen joints. I think of dowels as Tailing into three categories: as tenons, as pins and as gluing accessories.
In cases where dowels are used to modify end-grain òî side-grain joints (2) a dowel is inserted into a hole in the end of the perpendicular member. Since the piece has side-grain to side-grain contact, a high degree of integrity can be expected. The dowel becomes a tenon extension of the piece (multiple dowels, of course, can also be used). The mating hole in the side-grain surface of the joint becomes a mortise into which the tenon is fitted. In double-dowel joints that are subjected to racking, one of the dowels carries a critical share of the load in tension. The remaining load is transferred as surface compression. In designing such a joint, increasing the height of the member is advantageous, because it enables the dowels to be spaced as far apart as possible (3). The design should incorporate dowels that are large enough in diameter to carry tensile load and deep enough to resist pullout. The dowels should also be able to carry and transfer the shear load parallel to the side-grain member of the joint.
This design, when slightly modified, becomes a doweled miter joint, as in Figure 1A.
Dowels as pins provide physical constraint to a joint without glue (although glue may still be applied). Examples might include a pinned slip joint, a wooden hinge pin, the guide pins in a table leaf, flooring pins, etc. Historically, large wooden dowels called trunnels (tree nails] pinned framing members together in buildings. The pin usually has ils grain direction perpendicular to the grain direction of both parts being joined. These parts may have parallel grain directions but they are more often perpendicular to one another. Pins are sometimes added as a fail-safe measure or as a way of providing clamping pressure to draw a joint home. An example of the former is the pin usually concealed in the neck of a duck decoy, which will hold the head on in case the neck breaks because of weak grain direction. An example of the latter is the draw-bored mortise and tenon, where the hole drilled through the tenon is slightly offset from the one through the mortise, so pounding a pin home pulls the joint tightly together.

 
2 —Dowels used in end-grain to sidegrain joints in effect  become tenons; the mating holes become mortises.

Despite the obvious success of the draw-bored mortise and tenon, dowels are most often misused as gluing accessories to hold parts in alignment. For example, in making a tabletop, boards might be edge-glued and held with a series of bar clamps. To ensure alignment of board surfaces at the joints, dowels are sometimes used. However, if gluing is correctly done, full wood strength can be developed by a plain side-grain to side-grain joint—no reinforcement is necessary. Because they do not provide strength, the pins therefore need only be long enough and numerousenough to ensure alignment. For edge-gluing 1-in. lumber, 3/8rin. or 1/4-in. dowels that are 1 in. long are plenty. Needless to say, the holes should be bored a little deeper than the length of the dowels. Dowels should fit snugly into accurately positioned holes. When the joints are clamped, no attempt need be made to glue the dowels into the holes in the mating edges. The loss of glueline due to the dowels is negligible. For example, in edge-gluing 3/4-in. lumber, a 3/8-in. dowel placed every 8 in. along the joint reduces the glueline area less than 2%. Although it might seem advantageous to make the dowels "good and long" and glue them in "good and light," a negative effect can actually result. The restraint to normal shrinkage and swelling may cause the wood to fail at or near the glue joint (4), If gluelines fail at edge joints, the problem should be rectified by trouble-shooting the gluing procedure rather than by pinning a bad joint with dowels in an attempt to bring it up to standard. If the gluelines are properly made, there is little to gain in trying to reinforce the joints, since the strength of the wood on either side of the joint is still the limiting factor.
Edge joints also are modified by various tongue-and-groove configurations to assist in alignment. The logic that such joints are stronger because of greater surface area is questionable. If the quality of gluing is up to standard, the glueline is as strong as the adjoining wood. In joints of end-grain to side-grain combination, the spline may become a tenon or a simple locking device. In edge-gluing, the idea of "strengthening" the joint with a longitudinal spline may be tempting, but is a serious misconception. Since the spline is continuous, the reduction in surface area of the board would be substantial (5). For example, if a 1/4-in. spline ran the length of a 3/4-in. thick joint, the strength of the joint could be reduced by one-third. If the spline is very thin and the joint will be subject to bending, only slight weakening will occur, providing the spline is centrally located along the
neutral axis.


 
3-In doweled joints subject to racking, increasing the spacing between the dowels reduces the tensile load that each musl carry.

 

4—If long dowels are used across an edge joint and glued into (he holes, shrinkage will be restrained across A-B. Tensile stress, will develop across A'-B' as the boards attempt to shrink. 5—Reinforcing an edge-glued joint with a cross-grain spline actually weakens the join at the margins of the spline.

 

Fastened joints
The term fastener refers to an item that holds together two members being joined. When used in the sense of a pin, wooden dowels are examples. Other wooden components, such as cross-battens, corner blocks and plywood gusset piates, might also be thought of as fasteners. Usually, however, the term fastener suggests nails and screws. It is likely that when civilization learned to extract and shape metals, nails and spikes for wood were among the first items produced. Until the last century, handmade nails had hardly changed and screws were relatively expensive. Today, however, with automated production, improved fastener design, and power installation equipment, mechanical fasteners have become inexpensive and efficient alternatives For assembling wood components. They are likely to remain indispensable to many forms of woodworking.
It has been estimated that some 75,000 fasteners—mostly nails—are used in the average house. Most woodworkers readily appreciate the importance of nails in general carpentry and softwood construction, but also assume a traditional notion which holds that fasteners should be avoided in cabinetmaking. But it has been perceptively observed that wood joints "can be poorly made with considerable ease," and this would certainly apply to many fastened joints. On the other hand, fastened joints can also be well made. Woodworkers should carefully study mechanical fasteners as an alternative means of joining wood.
Since most fasteners are metal and thus have superior strength, failure of the fastener itself need not be a concern. The primary requirement is holding power, which is the ability of fasteners to transfer stress from one member to another without detaching, dislodging or causing failure in either member. Holding power is closely related to the structural strength properties and condition of the wood.
Because of the endless array of styles of modern nails and screws, I will review general considerations making no attempt to summarize the technical data on individual fasteners.
 

Nails
TIil' common \÷ ire nail, with iis bi uilii Iniish, dumiond-cut point and flat head, is the most familiar of modern nails. In typical use, it is driven forcefully and rapidly through one or more materials, embedding its point into the side grain of seasoned wood. A general empirical formula for direct withdrawal (1) of a bright-finish, common wire nail immediately after driving, is:
p = 7,850G 5/2 DL, where
p — maximum withdrawal load, in pounds
G = the specific gravity of the wood {Table 3 p. 8)
D = the nail diameter in inches
L = the depth of penetration, in inches, of the nail into the member holding the point.
Under these standard conditions, direct withdrawal varies directly with the diameter and length of the nail; the greater the length or the diameter, the greater the holding power. The formula also reveals that denser woods develop greater holding power.
When a nail is driven into a side-grain surface, the longitudinal cell structure is separated or split apart and also compressed ahead of the point, depending on its taper or blunlness. As the nail progresses, many cells in the path of the nail are broken, and their severed ends are bent and compressed in the direction of driving. The tendency of the cell ends to recover causes them to press against the nail surface, resulting in resistance to withdrawal.
Withdrawing the nail restraightens the cells,

1—The holding power of a nail is a function of its diameter (D), its driven length (L) and the grain direction and density of the wood into which it is driven.

?icreasing the bearing against the fastener. Only when
?ippage occurs does the nail finally pull out.
Experiments have shown that in many woods, a spear ?oint with a slim taper results in the greatest holding ?ower. However, the wood fibers also separate, and in ?ome species this type of point causes splitting. A blunt ?oint has less tendency to cause splitting, because a ?lug of compressed wood structure is torn loose and ?ushed ahead of the point, rather than being pushed ?side to start a split. However, the greater cell damage ?educes holding power. The common diamond point, ?hen, is a compromise that seems to afford the greatest ?olding power with the least splitting in common softwood structural lumber.
The fact that nails can be driven without preboring ?ilot holes has apparently led to the assumption that ?hey should be driven without preboring. An unfortu-?;ate corollary seems to be that nails are therefore lim-?ted to use where they can be driven without splitting ?he wood or bending over. In reality, the best holding ?lower develops when nail holes are prebored. While ?nost woodworkers accept the idea of installing wood-?crews in prebored holes to prevent splitting and to ?naximize holding power, they seldom consider prebor-?ng nail holes. Pilot holes ranging from 60% of nail ?hank diameter for low-density woods to 85% of shank diameter for high-density woods give maximum with-drawal resistance. In routine construction and carpen-?ry it is obvious that preboring is not feasible. For cab-?netmaking and other woodworking, however, nails in-?talled in prebored holes are extremely effective fasten-?rs and deserve greaterIn consideration.
In nailing two pieces together, driving the nail through the first piece builds up a compression zone that may cause splitting or other disruptions as the nail ?merges and enters the second piece. The resulting rupture can keep the pieces from maintaining close contact.
Appropriate preboring eliminates this problem.
The holding power of nails diminishes over time following installation. The long-term holding power, especially where extreme moisture variation causes dimensional change in the wood, can be reduced to as little as one-sixth the loads indicated by the formula given.
Holding power can be improved considerably by surface modification of the nail shanks (2). Resin-coated nails (called cement-coated nails) have about double normal holding power but this advantage disappears over lime. Withdrawal resistance can be substantially improved by placing annular grooves on the nail shank (3). Annularly threaded nails are understandably harder to drive, but when installed into side-grain prebored holes they provide positive resisting surfaces for bent-over fibers. Spiral-threaded nails have improved holding power that appears least reduced over time, perhaps because of the minimum damage to the holes as the nails "screw" into place.
All nails have less holding power when driven into end grain than into side grain. Test results indicate that the difference is smallest with high-density woods, but with low-density woods end-grain holding power may be as little as halt of side-grain holding power.

3—When a nail is driven perpendicular to the grain, annular grooves provide added bearing surface for the bent wood fibers, which resist withdrawal of the nail.

 

1
12—From top to bottom, a common nail with smooth, bright finish, a spiral-grooved nail and an annularly grooved nail.

Unfortunately, one hears the flat statement that nailing into end grain should be avoided on the grounds that holding power is reduced. Actually, the lower holding power can be compensated for by preboring and increasing the diameter, length or number of nails. Nailing into end grain, in fact, may be beneficial in certain situations. For example, when a nail is driven into side grain, later shrinkage shifts the wood along the nail shank, causing the nail head to protrude rather than pushing the point deeper into the wood. The longer the nail, the greater the protrusion. This is why, in dry-wall work, the shortest nail that will hold the gypsum board in place is recommended. When the wood swells, it moves equally away from each side of the shank center, thereby backing the point still farther from the bottom of the hole. Repeated cycles cause further emergence of the nail. In end grain, however, since wood does not change dimension along  its grain direction? using longer nails for greater holding power does not increase nail popping.
In lateral loading (1), joint slip rather than maximum load is critical. Stouter nails offer increased bearing against the wood so lateral load resistance increases exponentially with nail diameter. (If the diameter is doubled, the lateral resistance is increased by a factor of 2 3/2 -2 3 = 2.8It is also important that the fastener be stout enough to transmit load without bending. A long, slender nail crushes the wood near the surface and bends, and then becomes loaded in withdrawal and "snakes" out of the hole.
The head design of a nail is also critical. The broad heads of common nails are usually large enough to carry full withdrawal load without pulling through the top member. However, the small heads of Finishing nails limit the effectiveness of deep penetration since they readily pull through the top board. Thus, while a joint can be made stronger by using longer common nails, beyond a critical point a joint with finishing nails might best be improved by increasing the diameter or the number of nails, if space permits.
Appearance is a strong influence in the bias against using nails in woodworking joints. However, there are many places where nails could be the most efficient and effective fastener, for example, half-blind dovetail joints commonly used for drawer sides. It would be interesting to compare a well-nailed half-blind tongue-and-rabbet joint for overall strength. Dare I even suggest nailing a dovetail joint together?
 

1—When a nail subject to laleral loading tails, it typically crushes the wood near the surface, then bends and pulls out of the hole.

Woodscrews
Most woodworkers recognize the superiority of correctly installed woodscrews (2) over other fasteners. Their great holding power is understandable in terms of the positive engagement of the threads into relatively undamaged wood structure. The key to maximum holding power is preboring pilol holes tor the threaded portions of the screw and for the shank. Slightly undersized shank holes should be prebored to provide a snug fit and firm bearing without developing enough stress to cause splitting. Pilot holes for the threaded portion should be from 70% of root diameter in low-density woods to 90% in high-density woods. (In the densest woods, it may be best to bore the pilot hole the same diameter as the root, especially for brass screws.) Lubricating the threads with wax facilitates driving and minimizes screw breakage without loss of holding power.
When woodscrews are correctly installed into the side grain of seasoned wood, maximum withdrawal loads can he estimated by the empirical formula:
p= 15,700 G2DL,
p= maximum withdrawal load, in pounds
G = the specific gravity of the wood (Table 3, p. 8)
D = the screw shank diameter, in inches
L=the depth of penetration, in inches, of the
threaded portion of the screw into the member
receiving the point.
As with nails, the holding power of screws increases directly with diameter and length, and exponentially with the density of the wood. Similarly, holding power
loading conditions of long duration might be as little as 20% of the values estimated by the formula given above. Screws driven into end-grain surfaces average oniy about 75% as much holding power as those driven into side grain, and holding power will be more erratic. As with nails, it is preferable to design joints to load screws laterally rather than in direct withdrawal.
Besides nails and screws, a host of various other fasteners are available for use with wood, such as clamp nails, corrugated fasteners and staples. In addition, various hardware items serve as fasteners in the role of a "third party," which is fastened to the two or more wood components by nails and screws. These include mending plates, flat corner irons, angle braces, T-plates and hinges. In evaluating the use and effecctiveness of each, the integrity usually depends on the holding power of the attachment fasteners in terms of stress application relative to grain direction. In construction joints,
for example, special framing anchors have been developed which transfer loads from one member to the other by loading fasteners laterally rather than in withdrawal. These anchor plates also provide for the proper number and placement of fasteners (usually nails) for maximum strength.


 
2—Common woodscrews (from left to right): flathead, round-head, ovalhead.

Adhesive joints
Laminated items assembled with glue have been discovered in the tombs of eariy Egyptian pharaohs, and it is probable that the use of adhesive substances for holding wood parts together predates recorded history. Through the ages, most glues were made from fish, animals and vegetable starch and showed little change or improvement. In this century, however, development of the plywood industry initiated drastic changes in the types and properties of adhesives. Further stimulated fay the demands of World War II and the scientific plunge into the space age, a dynamic adhesive technology has given us numerous multipurpose and specialized adhesives, with the promise of a continued parade of new ones well into the future.
Today's woodworkers use adhesives in a number of ways—to make large pieces out of smaller ones (such as carving blocks and laminated beams), to create combinations for strength or aesthetic improvement (such as plywood, veneers and marquetry) and to join parts to create a final product, as in furniture, sporting goods and structures. A complete discussion of gluing technology is impossible here. However, certain basic considerations that may be overlooked or misunderstood often cause serious gluing problems and are worth a systematic review.

  1—The five phases of a glue joint.
 
The general term adhesive includes any substance having the ability to hold two materials together by surface attachment. Those most commonly used for wood are called glues although materials described as resins, cements and mastics are equally important in the assembly of wood products.
No truly all-purpose adhesive has yet been manufactured and probably never will be. A general-purpose adhesive cannot hope to attain all the individual capabilities and attributes of closely designed ones. Although any of the standard commercial glues will do a satisfactory job if the moisture content of .the wood is controlled and the temperature remains within the human-comfort range, there is an increasing trend toward development of special adhesives. Adhesive selection must therefore take into account factors such as species, type of joint, working properties as required by anticipated gluing conditions, performance and strength, and, of course, cost.
One interesting adhesive is water. It is easily spread, wets wood well and solidifies to form a remarkably strong joint. It is delightfully inexpensive. However, it is thermoplastic and its critical maximum working temperature is 32°F. At temperatures at which it will set it has a very short assembly time. But due to its temperature limits, water will never capture a very important position among woodworking adhesives.
A wide and confusing array of adhesive products confronts the woodworker. A common pitfall is the belief that some glues are belter than others; the notion that simply acquiring "the best" will ensure success is careless and may give disastrous results. With certain qualifications, all commercially available adhesives will perform satisfactorily if chosen and used within their specified limitations. An important corollary is that no adhesive will perform satisfactorily if not used properly. Within the specified limitations, mosf woodworking adhesives will develop joints equal in strength to the woods being joined. Thus, the wood, rather than the glue or its bond, is the weak link in a well-made joint.
Glues made from natural materials have been used from earliest times and even today, hide glue (made from the hides, tendons and hooves of horses, cattle and sheep) and casein (primarily a milk derivative) are still in use. However, the bulk of modern wood glues are synthetic compounds. Perhaps the most versatile are the polyvinyl acetate emulsions (PVA), commonly called white glues. More recently the yellow glues (modified PVA) have emerged, which have greater rigidity, improved heat resistance and better "grabbing" ability. These yellow glues are satisfactory for the bonding jobs of most craftsmen. They are easy to use and are more tolerant to unfavorable conditions than are white glues. Yellow glues also give less trouble in clogging abrasive paper.
Urea-formaldehyde, or plastic-resin glues, are water-resistant but not heat-resistant, as are the resorcinol-formaldehyde adhesives. A number of other adhesives with a variety of special uses and properties are also on the market, including epoxies, contact cements, mastics and acrylic adhesives.
A logical starling point is to wonder why glue sticks at all. It is sometimes assumed that adhesion results from the interlocking of minute tentacles of hardened adhesive into the fine porous cell structure of the wood surface. However, scientific research has shown that such mechanical adhesion is insignificant compared to the chemical attachment due to molecular forces between the adhesive and the wood surface, or specific adhesion. The assembled joint, or bond, is often discussed in terms of five intergrading phases (1), each of which can be thought of as a link in a chain. The weakest phase determines the success of the joint. Phases 1 and 5 are the pieces of wood, or adherends, being joined. Phases 2 and 4 are the interpenetrating areas of wood and adhesive, where the glue must "wet" the wood to establish molecular closeness for specific adhesion. Phase 3 is the adhesive itself, which holds together by cohesion.
Fundamentally, then, gluing involves machining the two mating surfaces, applying an adhesive in a form that can flow onto and into the wood surface and wel the cell structure, and then applying pressure to spread the adhesive uniformly thin and hold the assembly undisturbed while the adhesive solidities. The "typical adhesive is obtained or mixed as a liquid but sets to form a strong glue layer, either by loss of solvent, which brings the adhesive molecules together and allows them to attach to one another, or by a chemical reaction that develops a rigid structure of more complex molecules.
Different woods have different gluing properties. In general, less dense, more permeable woods are easier to glue, for example, chestnut, poplar, alder, bass wood, butternut, sweetgum and elm. Moderately dense woods such as ash, cherry, soft maple, oak, pecan and walnut glue well under good conditions. Hard and dense woods including beech, birch, hickory, maple, Osage-orange and persimmon require close control of glue and gluing conditions
to obtain a satisfactory bond. Most softwoods glue well, although in uneven-grained species, earlówood bonds more easily than denser latewood. Extractives, resins or oils may introduce gluing problems by inhibiting bonding, as with teak and rosewood, or by causing stain with certain glues, as with oaks and mahogany.
Since most adhesives will not form satisfactory bonds with wood that is green or of high moisture content, wood should at least be well air-dried. Ideally wood should be conditioned to a moisture content slightly below that desired for the finished product, to allow for the adsorption of whatever moisture might come from the adhesive. For furniture, a moisture content of 5% to 7% is about right. However, when using urea-formaldehyde glues, the moisture content should not be below 7%. For thin veneers, which take up a proportionately greater amount of moisture, an initial moisture content below 5% might be appropriate.
Machining is especially critical. In some cases, especially for multiple laminations, uniform thickness is necessary for uniform pressure. Flatness is required to allow surfaces to be brought into close proximity. The surfaces to be glued should have cleanly severed cells, free of loose fibers. Accurate hand-planing is excellent if the entire surface, such as board edges, can be surfaced in one pass. On wide surfaces, peripheral milling (planing, jointing) produces adequate surfaces. Twelve to twenty-five knife marks per inch produce an optimum surface. Fewer may give an irregular or chipped surface; too many may glaze the surface excessively.

Dull knives that pound, heat and glaze the surfaces can render the wood physically and chemically unsuited for proper adhesion even though it is smooth and flat. Planing saws are capable of producing surfaces acceptable for gluing, but in general sawn surfaces are not as good as planed or jointed ones.
Surface cleanliness must not be overlooked. Oil, grease, dirt, dust and even polluted air can contaminate a wood surface and prevent proper adhesion. Industry production standards usually call for "same-day" machining and gluing. Freshly machining surfaces just before gluing is especially important for species high in resinous or oily extractives. Where this is not possible, washing surfaces with acetone is sometimes recommended. One should not expect a board machined months or years ago to have surfaces of suitable chemical purity. If lumber is flat and smooth but obviously dirty, a careful light sanding with 240-grit or finer abrasive backed with a flat block, followed by thorough dusting, can restore a chemically reactive surface without seriously changing flatness. Coarse sanding, sometimes thought to be helpful by "roughening" the surface, is actually harmful because it leaves loose bits. Tests have shown that intentionally roughening a surface, as in "toothed planing," does not improve adhesive-bond quality. In summary, wood should be surfaced immediately prior to gluing, for cleanliness and to minimize warp, and should be kept free of contamination to ensure an acceptable gluing surface.
Shelf life is the period of time an adhesive remains usable after distribution by the manufacturer. Unlike photographic films, adhesives are not expiration-dated. Beware the container that has been on the dealer's shelf too long. Outdated package styles are an obvious tip-off. It is wise to mark a bottle or can with your date of purchase. It is amazing how fast time can pass while glue sits idle in your workshop. If possible, refrigerate glues in tight containers to prolong shelf life. In general, if the glue is spreadablc when mixed according to instructions, it is suitable for use. Adding water to restore spreadability is not a good practice.
The adage "when all else fails, read the instructions," all too often applies to glue. It is unfortunate that instructions arc so incomplete on retail glue containers. Manufacturers usually have fairly elaborate technical specification sheets but supply them only to quantity consumers- Too often many critical factors are left to the user's guesswork or judgment. Mixing proportions and sequence usually are given clearly; obviously they should be followed carefully.
Glues with a pH above 7 (alkaline), notably caseins, will absorb iron from a container and react with certain woods such as oak, walnut, cherry and mahogany to form a dark stain. Coffee cans or other ferrous con-metallic mixing containers such as plastic cups or the bottoms of clean plastic bleach jugs work nicely.
Once glue is mixed, the pot life, or working life, must be considered. Most adhesives have ample working life to handle routine jobs. The period between the beginning of spreading the glue and placing the surfaces together is called open assembly time; closed assembly time indicates the interval between joint closure and the development of full clamping pressure. Allowable closed assembly lime is usually two or three times open assembly time. With many ready-to-use adhesives, there is no minimum open assembly time; spreading and closure as soon as possible is recommended, especially in single spreading, to ensure transfer and wetting of the other surface. If the joint is open too long, the glue may precure before adequate pressure is applied. The result is called a dried joint, f n general, assembly time must be shorter if the wood is porous, the mixture viscous, the wood at a low moisture content, or the temperature above normal. (As a rule, it is a good policy to avoid gluing where temperature of the room or of the wood is below 70 °F.) With some adhesives, such as resorcinol, a minimum assembly time and double spreading (that is, applying adhesive to each of the mating surfaces) may be specified for dense woods and surfaces of low porosity, to allow welting of the wood and to permit thickening of the adhesive to prevent excessive squeeze-out. Proper spread is difficult to control. Too little glue results in a starved joint and a poor bond. A little overage can be tolerated, but too much results in wasteful and messy squeeze-out. Though some squeeze-out is assurance that sufficient adhesive has been applied, squeeze-out may cause problems in machining or finishing. With experience the spread can be eyeballed. It is useful to obtain some commercial specifications and conduct an
experiment to see just what they mean. Spreads are usually given in terras of pounds of glue per thousand square feet of single glueline, or MSGL. A cabinetmaker
will find it more convenient to convert to grams per square foot, by dividing Ib./MSGL by 2,2. Thus a recommended spread of 50 Ib./MSGL, typical of a resorcinol glue, is about 23 grams per square foot. Spread it evenly onto a square foot of veneer for a fair visual estimate of the minimum that should be used. Usually, the recommended spread appears rather meager.
 

 

Table 15—Average clamping pressure of typical woodworking clamps. To find out just how mach load typical clamps could apply, Hoadley attached open steel frames to the crossheads of a universal timber-testing machine. With a clamp positioned to draw the frames together, the load applied was indicated directly. The clamps are described in the table, with the last column giving the average of three trials by average-sized Hoadley. tightening as if he were trying to gel maximum pressure in a gluing job. The quick-set clamp listed first in the lable was used to calibrale the setup: A secretary squeezed 330 Ib., a hockey player squeezed 640 lb., and Hoadley squeezed 550 Ib- Repealed trials by each person yielded readings that agreed to within 10%. An asterisk indicates that the clamp began lo bend, and the test was stopped at the value listed.

 

Double spreading is recommended where feasible. This ensures full wetting of both surfaces, without relying on pressure and flatness to transfer the glue and wet
the opposite surface. With double spreading, however, a greater amount of glue per glueline is necessary, perhaps a third more.
Glue should be spread as evenly as possible, even though some degree of self-distribution will of course result when pressure is applied. Toothed scrapers, rollers or stiff brushes are best for this purpose. Some bled, it pays to have them in the order of assembly.
The object of clamping a joint is to press the glueline into a continuous, uniformly thin film, and to bring the wood surfaces into intimate contact with the glue and hold them undisturbed until setting or cure is complete. Since loss of solvent causes some glue shrinkage, an internal stress often develops in the glueline during set-ling. This stress becomes intolerably high if gluelines are too thick, Gluelines should be not more than a few thousandths of an inch thick.
If mating surfaces were perfect in terms of machining and spread, pressure wouldn't be necessary. The "rubbed joint," skillfully done, attests to this. But un-evenness of spread and irregularity of surface usually require considerable external forces to press properly.
The novice commonly blunders on pressure, both in magnitude and uniformity.
Clamping pressure should be adjusted according to the density of the wood. For domestic species with a specific gravity of 0.3 to 0.7, pressures should range from 100 psi to 250 psi. Dense tropical species may require up to 300 psi. In bonding composites, the required pressure should be determined by the lowest-density layer. In gluing woods with a specific gravity of about 0.6, such as maple or birch, 200 psi is appropriate. Thus gluing up one square foot of maple requires pressure of (12 in. x 12 in. x 200 psi) 28,800 pounds. Over 14 tons! This would require, for an optimal glueline, 15 or 20 C-clamps, or about 50 quick-set clamps. Conversely, the most powerful C-clamps can press only 10 to È square inches of giueline in maple. Jackscrews and hydraulic presses can apply loads measured in tons. But since clamping pressure in the small shop is commonly on the low side, one can see the importance of good machining and uniform spread. table 15 gives an indication of how much gluing pressure can be delivered from various clamps.
Another troublesome aspect of clamping is uniformity, usually a version of what I call "the sponge effect." Lay a sponge on a table and press it down in the center; note how the edges lift up. Similarly, the force of one clamp located in the middle of a flat board will not be evenly transmitted to its edges. It is therefore essential to use heavy wooden cover boards or rigid metal cauls to ensure proper distribution of pressure (1).
Clamp time must be long enough to allow the glue to set well enough so that the joint will not be disturbed by clamp removal. Full cure time, that is, for development of full bond strength, is considerably longer. If the joint will be under immediate stress, the clamp time should be extended. Consider a dry run to check for tightness of joints and to rehearse the process.
Finally, cured joints need conditioning periods to allow moisture added at the glueline to be distributed evenly through the wood. Ignoring this can result in sunken joints (2). When edge-gluing pieces to make panels, moisture is added to the gluelines, especially at the panel surfaces where squeeze-out contributes extra moisture (A). If the panel is surfaced while the glueline is still swollen (B, C), when the moisture is finally distributed the glueline will shrink (D), leaving a joint that is sunken.
 

 

1—Cover boards (cauls) distribule clamping pressure evenly 2 — When an edge-glued panel (A) is .surfaced while the glue-line is still swollen with moisture (Â, Ñ), a sunken joint (D) is the result.