Have you tested your favorite wood glue to see if it can be reglued successfully should either your work be damaged, or a cross-grain glue joint fail with age and seasonal movement?

In 4 decades as a woodworker, I’ve done a good bit of conservation, repair and restoration work, including pieces in a few federal museums both here and overseas. As I pass what I know down to my boys, included will be what I know about glues. I know that some glue types can’t be glued over, often requiring new wood to be let in during repairs, and the joint recut. I discovered that the hard way some decades ago restoring furniture, and simply switched to other glues for all my work. Since then, those glues I rejected may have been reformulated; plus there are a number of new glues worth checking out, so to make sure I’m not providing bad or outdated advice, it’s time to check out the current crop of wood glues for repairability.

I make no pretenses toward science, here…this is all anecdotal based on experience, not chemistry…all I want to show is whether marine epoxy will adhere to the glue lines or residue of the various wood glues during repairs. You can look up strength and other test data in your USDA Wood Handbook; I care about repairability because I’ve never seen any test or even anecdotal data on anything but hide glue in that regard, and it’s important if your work is to survive beyond typical damage and wear and tear over time. I chose epoxy as the regluing agent because it’s the usual choice in professional structural repair work and it adheres to a greater number of diverse substances than any other wood glue I know of. In fact, it usually rebonds a failed but fully cured glue joint much better than the original glue would, and as it also bonds to itself very well, epoxy is a good, repairable choice for many applications.



On identical tiles of freshly planed, vertical grain, second-growth Doug Fir, I saturated the faying surfaces with glue and let them cure to full strength by the manufacturer’s instructions for time and temperature….



…then I keyed each faying surface with 100-grit abrasive paper, reglued them with marine epoxy, and “clamped” the assemblies to the degree favored by epoxy. For glues that left a rough surface like polyurethane, the epoxy was applied twice…an unthickened coat followed by a second coat thickened with West 404 High-Adhesive Thickener, per the manufacturer’s instructions. I let the epoxy cure for 6 days to reach full strength.

I purposely chose small blocks of wood with easily broken short grain because strength here isn’t the issue, adherence is, and I can check adherence using a sharp chisel without trying to break long glue joints in a press. Of greater concern was that the glues to be tested were applied without any clamping pressure, but as it turned out, several glues that require high clamping pressure fared very well, so I believe the results are reasonably valid.


The results offered no surprises.



The epoxy thoroughly adhered to the strongest of the off-the-shelf glues, the 2-part resorcinol, breaking completely at the wood rather than the glue line. Attempts to slip the chisel between the glue lines revealed a thorough and unified bond between all three layers of glue.



Epoxy on epoxy showed similar results……and so did liquid polyurethane (Elmer’s Ultimate)…



…and powdered urea formaldehyde plastic resin glue.



Titebond, a Poly Vinyl Acetate glue, however, broke some wood but failed the chisel test…. the chisel easily separated the two layers of Titebond, indicating poor adherence of the epoxy in between.



Titebond II broke even less wood, with poor adherence…



…and Titebond III, while a much stronger glue, still did not adhere to the epoxy.

The implications of all this can be minor if we are talking about a first-effort coffee table….but they can be serious if we are talking about a strip-planked boat hull made of 1 X 1 strips glued together using an unrepairable glue. Picture the requirement to feather in a large patch to repair hull damage, and you can see that patch will be pinstriped with unsound repair at every glue line, leading to early failure of the repair.

You can draw your own conclusions. Mine are that the work most easily restored is often the work that survives the longest, that you may not care about longevity, but that may break you granddaughter’s heart some day, and I’d check out my glue choices thoroughly before committing them to any 20-hour high-end project, let alone a 700-hour project.