In my previous column, in WB No. 304, I focused on bending wood and suggested that for ring-porous hardwoods such as oak and ash, tension surface failures could be avoided by bending against the radial (quarter-sawn) rather the tangential (flat-sawn) face. I provided a photo to demonstrate failure in the weak earlywood zone near the outside (tensional face).
In that same issue, the boatbuilder Harry Bryan wrote lovingly of the superb craftsmanship he found in a Lawley tender, built in 1925, that had been stored for more than 50 years in a barn. In a caption for one of his photographs (No. 10), he suggested that the ½″ × ⅝″ frames were oriented non-traditionally—that is, bent against the radial face. This made me wonder whether bending against the tangential face was the norm for boatbuilders.
Although Harry and I have never met in person, we are connected through a magical fishing location in Downeast Maine. So, I emailed him to find out more about bending traditions among boatbuilders. This led to several emails between us. In his first reply, Harry provided citations from his boatbuilding library. He wrote: “In Howard Chapelle’s Boatbuilding there is a picture on page 358 titled ‘End-Grain Position in Bent Frames.’ He says, ‘The frames should always be bent so that the annual rings run more or less parallel to the ribbands or planking.’ Bud McIntosh in How to Build a Wooden Boat, page 68, implies that there might be a very slight improvement with rings oriented fore-and-aft. Sam Manning’s illustration on the following page shows rings oriented close to fore-and-aft with respect to the hull. Robert Steward’s Boatbuilding Manual, page 118, says, ‘It is best to bend the stock on the flat of the grain.’ The illustration on the following page shows rings parallel to the planking.”
Harry added an experiment that he had conducted at WoodenBoat School in which he “clamped two steamed ½″× ½″ pieces of oak so they cantilevered an equal distance out from the bench. One had annual rings running vertically (floor to ceiling) and the other had rings parallel to the floor. My conclusion (admittedly with only these two pieces) is that the one with rings parallel to the floor bent down a bit more when an equal weight was hung from the end.” From this experiment (probably repeated by other boatbuilders in the past) one could conclude that ease of bending has been a key reason for choosing tangential-face bending.
Then why did the skilled builders at the George F. Lawley & Son shop choose to break with tradition and bend the frames for the Lawley tender against the radial face?
We, of course, will never know. But I will make a stab at an answer. I don’t think the builders just randomly chose the radial orientation. I think they tried the traditional orientation, and after producing a number of frames that failed on the tensional surface, they rotated the frames 90 degrees when building Lawley tenders.
I suspect that in other hulls when frames were bent over forms and the tension face was strapped and end-clamped, the traditional orientation likely prevailed—and was successful due to the tension strap. But in the Lawley tenders, the frames were steamed and then bent into the planked hull where no tension strap could be used.
If my conjecture is correct, bending orientation is optional if a clamped tension strap is used. But if a tension strap can’t be used, radial-face bends should produce fewer failures in ring-porous woods. And, of course, straight grain is critical regardless of orientation.
Kiln-dried Wood
In a subsequent email, Harry wrote, “As a teacher, you must have recollections of a few embarrassing moments in front of a class. I certainly have a few of these. One was at WoodenBoat School: Our class was about to bend some knees, and the director asked if I would mind demonstrating in front of all three classes in session. Also in attendance was a visiting boatbuilder whom I had met in Tasmania. I stationed myself at the jig and called for the hot piece of oak, which came in a dramatic sprint from the steambox. I clamped it in place and gave it all I had. The piece did not bend even ¹⁄₁₆″. The piece brought to me was kiln-dried. It could have been solid steel, for all that it bent. The oak cooled as my ears heated up. My feeling is that when a piece of wood is put in the steambox, particularly if it has any size, there is little penetration of moisture into the wood, steaming being a way to heat the wood with no loss of the moisture already within the piece being bent.”
I have had my own classroom moments of chagrin, but none as dramatic as Harry’s. He is correct in surmising that steam in a steambox does not add moisture to the wood—as I noted in my column in WB No. 304. What I didn’t emphasize in that column was the general futility in trying to bend kiln-dried lumber. At best, if the wood has been very carefully kiln-dried on a slow schedule, you would still need to soak the wood for an extended period of time to condition it to a moisture content of 15 to 20 percent. At worst, no amount of soaking will soften kiln-dried lumber that has been hurried through a fast-drying schedule at higher-than-recommended temperatures—a common practice with construction-grade lumber. Under those conditions, the wood will be partially torrefied—a process I described in WB No. 304—making it extremely resistant to bending.
Heat Without Steam
In a final email, Harry wrote: “I cannot remember whether you have addressed dry-heat bending in your writing. I have not explored its limits but have had good success with oak in moderate bends such as slats for a chair, kayak backrests, and small-boat tillers. I do not own a heat gun, which I suspect would be better than the propane torch I use. It is a bit like toasting marshmallows, seeing how hot I can get the wood without scorching it. I hold one end of the piece in a vise and pull on the other end as heat is applied to what will be the inside of the bend. When the desired shape is reached, I hold it for perhaps 30 seconds, after which it will retain its shape. I am not sure why the bend holds its shape almost immediately, unlike steaming, which seems to take 24 hours or so before it can be trusted to retain its shape.”
The lignin component of wood is thermoplastic, with a “glass transition” temperature of 170˚C (338˚F). Above that temperature, lignin will flow, permitting wood to be bent. Unfortunately, as we have already learned, a temperature that high degrades the cellulose component of wood. Adding moisture to the wood reduces the glass transition temperature of lignin and allows for wood-bending without cellulose degradation.
In dry-heat bending, where heat is applied quickly to the compression surface, cellulose degradation will be confined to that surface and not affect the tension side. Luthiers use a hot, curved platen to rub thin guitar sides to facilitate bending. The heating must be done quickly, and as Harry notes, only works for small bends in relatively thin stock. The reason the wood sets quickly is that no water-induced cellulose loosening has occurred—just lignin flow—so as soon as it cools, the lignin solidifies. With steam-bending of wet wood, the set doesn’t occur until the wood loses some moisture.
I will end by mentioning one more way of bending wood: microwave heating. For most boatbuilding situations this method is not practical, but for very small, thin items, it may have application. But a warning is necessary. Like dry-wood heating with torch, heat gun, or platen, timing is critical. Also, while those methods heat only the compression surface, microwaves heat the entire board, beginning in the center. This is less desirable than heating only the compression surface. Precise control is difficult with microwave heating, but the short heating times limit moisture loss. 
Dr. Richard Jagels is an emeritus professor of forest biology at the University of Maine, Orono. Please send correspondence to Dr. Jagels by mail to the care of WoodenBoat, or via e-mail to Senior Editor Tom Jackson, tom@woodenboat.com.