Technical note 5

TN-5: Finishes and Moisture

Whether to finish the chair, and if so with what. First draft 19 August 2026, from KU's questions and his 2010 Baubilt post "Polywhat? Interior Wood Finishes." Status: DRAFT; the wood-science claims are VERIFIED against the Forest Products Laboratory and the cited literature, the chemistry is STANDARD, and the recommendation is DERIVED from the project's own repairability doctrine.

The claim

For a chair kept dry and sheltered, a finish is not needed for survival. The longest-lived wood we have is unfinished. A finish on such a chair is chosen for other reasons, cleanliness, wear at the touch surfaces, a little moisture buffering, and appearance, and if it is chosen it should obey the same doctrine as the joints and the parts and the mark: THIN and RENEWABLE, so the next custodian can refresh it in place, rather than a thick built film they must strip to repair. On a hundred-year object, renewability is longevity. Note the correction, caught by KU: the property that matters is renewable (re-coatable in place with no witness line), not necessarily reversible (dissolvable off). Thin wipe-on poly is renewable, so KU's 2010 recommendation largely survives the longevity test, which the first draft of this note wrongly denied.

Does dry, sheltered wood need a finish? (longevity)

No. Wood degrades by three mechanisms, and all three are exposure phenomena [VERIFIED against FPL GTR-113 ch. 15, Finishing of Wood, 19 Aug 2026]: ultraviolet light, which photo-oxidizes the surface and preferentially degrades lignin so the cellulose fibers loosen and gray; liquid water cycling, whose repeated wetting and drying roughens and checks the surface; and biological attack by fungi, which needs free moisture to work. Remove sun, rain, and damp and you have removed the agents. The FPL's own number for how slowly even fully exposed, unfinished wood is lost is the punchline: about 6 mm, a quarter inch, of surface erosion in a century. Sheltered indoors, the loss is essentially nil.

The historical record agrees, and it is the same record the thousand-year chapter draws on. The oldest wooden buildings in the world, the seventh- and eighth-century hinoki timbers of Horyu-ji, are unfinished. European cathedral oak roofs carry unfinished timber for centuries. And in the dry American Southwest, the roof beams of the Chaco Canyon great houses, felled between roughly 850 and 1150 CE, survive as original wood, unfinished, so intact that they are routinely tree-ring dated to the year, which is only possible because the high desert kept them dry and sheltered [VERIFIED in general terms against the Chaco timber literature, Smithsonian and SAPIENS, 19 Aug 2026]. Unfinished structural wood, kept dry, is the longest-lived wood there is. A finish is not what makes wood last.

So why finish a chair at all? For the honest reasons a chair, unlike a roof beam, actually presents: it is touched by hands that carry skin oil and dirt, sat on, and spilled on, so it wants to be cleanable and abrasion-resistant at the contact surfaces; a finish slows the rate at which the wood takes up and gives off humidity, which reduces the cyclic movement that checks a surface over decades; and it looks and signals in ways discussed below. None of these is survival. All of them are service and appearance.

Does wood need to "breathe"?

The folk belief is half right, and the FPL states the exact truth: a finish cannot change wood's equilibrium moisture content, it can only change the RATE at which the wood reaches equilibrium [VERIFIED, FPL ch. 15]. Wood always comes to balance with the ambient humidity; no finish stops that. What a finish can do is slow the exchange, and there the danger appears: a film finish with high moisture-excluding effectiveness is MORE prone to fail by peeling, because it traps moisture that then lifts it off. This is why exterior paint blisters and why a sealed tabletop in a wet kitchen fails at the film. Indoors and dry the trapping risk is modest, but the lesson still points one way: a penetrating finish that buffers without sealing is the forgiving choice, and it fails gracefully, wearing thin, where a film fails catastrophically, peeling. "Breathing" is the wrong word for the right instinct. The instinct is: do not build a skin that can trap and lift.

Is there a molecular distinction between natural and synthetic finishes?

The natural-versus-synthetic axis is nearly meaningless at the molecular level, and KU argued this correctly in 2010: nearly every finish is a polymer, wood itself is a polymer, and the "plastic look" people dislike is high gloss, not synthetic chemistry. That still holds. But the first draft of this note then drew the NEXT line in the wrong place, and KU caught it: it treated "crosslinked" and "film-forming" as the marks of a bad finish, which would wrongly condemn tung oil and wipe-on poly, both of which crosslink and both of which in fact repair beautifully. Here is the corrected picture. Three axes matter, and only the third decides a hundred-year finish.

So beeswax and wipe-on poly do differ, but not as "good reversible penetrating" versus "bad irreversible film." Beeswax is a soft, non-crosslinked, reversible, renewable, near-zero-film wax. Wipe-on poly is a hard, crosslinked, non-reversible but RENEWABLE, thin film. Both renew invisibly in place; the wax adds full reversibility and the easiest repair, at the cost of hardness and protection. Neither is the peeling villain. The peeling villain is the thick built film, and it is defined by thickness, not by chemistry.

What is desirable in a 100-year or 1000-year finish?

Four qualities, and they are the project's doctrine applied to the surface. RENEWABLE: it can be cleaned, locally scuffed, and re-coated over itself without stripping, invisibly, in place, once a generation; oil, oil-wax, and shellac all re-coat over themselves and blend, while polyurethane leaves witness lines where old meets new. REVERSIBLE: a future conservator can remove it without harming the wood beneath, the museum principle, which is why conservators keep wax and shellac and dread a century-old polyurethane that can only be taken off by stripping, and stripping removes patina and a little wood every time. NON-EMBRITTLING: it does not, over decades, yellow, craze, and peel; film finishes fail this at century scale, penetrating finishes fail gracefully by thinning. FAILS LOCALLY: wear shows where hands and bodies actually touch and can be fixed just there; a film shows a hard boundary between worn and unworn, oil and wax feather into it.

By these tests the property that matters most is RENEWABILITY, and it is a property of THIN finishes, shared by true oils, oil/varnish blends, and thin wipe-on poly alike. The finish to AVOID is the THICK BUILT FILM, brush-on varnish, floor polyurethane, thick lacquer, conversion coatings, which fails at century scale by yellowing, crazing, and peeling, and which cannot be spot-repaired invisibly; you strip and recoat the whole part, and stripping costs patina and wood. The failure mode the first draft pinned on "polyurethane" belongs to THICK FILM poly, not to thin wipe-on poly, which behaves like an oil because it is applied like one: it wears, it takes another coat, it blends. KU's account of both tung oil and wipe-on poly, that they renew without a line between old and new, is correct and is the corrected note's central evidence.

A second, rarer property is REVERSIBILITY: full removal by a future conservator without stripping. Only wax, shellac, and lacquer offer it. It is the conservation ideal, but it costs durability (wax soft, shellac water- and alcohol-sensitive) and it is not required for renewability. So the corrected recommendation is: use a THIN, renewable finish, a true oil (linseed, tung), an oil/varnish blend (Danish oil), or a thin wipe-on poly, all of which renew in place without a witness line, and refuse only the thick built film. KU's 2010 recommendation of wipe-on poly therefore largely stands even at the hundred-year scale; its one shortfall against oil, wax, and shellac is that it is renewable but not REVERSIBLE, so a distant conservator cannot dissolve it off cleanly. What all of these share, and what actually decides the century, is that they stay thin and renew in place.

Repairability, which is the whole point

KU's instinct is correct: sand, clean, reapply. A penetrating reversible finish is designed for exactly this. Wipe on grime remover or a little solvent, scuff the worn patch with fine abrasive, wipe on more oil or wax or alcohol-cut shellac, and the repair blends invisibly because the finish is in the wood, not a layer over it. It is done in place, by anyone, in an hour, per the maintenance covenant, and it sits at the bottom of the failure order: finish before wood, wood before hardware, nothing before the geometry. The finish is the first consumable, and it should be the easiest thing on the chair to renew. A polyurethane film resists this precisely: you cannot spot-repair it invisibly, so you strip and recoat the whole part, and each strip costs patina and wood. A finish that can only be repaired by removal is, on a hundred-year timescale, a slow way of sanding the chair away.

Why we like wood "finished": the aesthetics

The question is not stupid. Two mechanisms from the aesthetics chapter explain it, and a third qualifies it. GLOSS: humans carry a vestigial attraction to reflective surfaces, plausibly because the only reflective thing on the Pleistocene savanna was water [STANDARD, Coss and Moore, cited in the aesthetics chapter]; a finished surface reads as faintly wet, and we are drawn to it. HONEST SIGNALING: a well-finished surface is costly to fake and therefore signals care and quality, and over a long life it does more, a maintained finish signals that the object is loved, which is the very mechanism, custody recruited by legible care, that the longevity chapters say keeps objects alive. The BIOPHILIA qualification: one might expect a biophilic pull toward raw, bark-like texture, and there is something to the love of visible grain, but the evidence and the market both say we prefer wood smooth with its figure revealed under a low sheen, not rough. The appeal of "finished" wood resolves into three things at once: the grain made legible (fluency, and the honesty of showing the real material), a soft sheen (the water vestige), and the signal of care. A finish that reveals figure and adds a low sheen hits all three. High gloss over-plays the water card and tips into the "plastic" read, which, as KU noted in 2010, is the gloss talking and not the chemistry.

There is a longevity corollary in the aesthetics. A reversible penetrating finish AGES into patina, a wanted signal of age and care; a film finish DEGRADES into damage, yellowing and crazing, an unwanted signal. Patina recruits custody; damage invites discard. The finish choice is therefore also a choice about how the chair will look when it is old, and the reversible penetrating finishes are the ones that grow more wanted with age rather than less.

For the Standard

A finish clause: the finish is optional and non-structural. If applied, use a penetrating, renewable, reversible finish, an oil, an oil-wax, or shellac, and avoid permanent film finishes on any part meant to be maintained across generations. The finish is a consumable surface, renewed on the maintenance covenant, and it sits below the wood in the failure order.

Open questions

  1. Which finish for the Keel Chair specifically: raw, oil, oil-wax, or shellac, tested on both the plywood panels and the hardwood keel, which will take and wear finish differently.
  2. THE PLYWOOD PROBLEM. This doctrine assumes solid wood you can sand and re-oil many times. Structural plywood has a thin face veneer, so it can be sanded very few times before breaking through, and film-faced ply (as on the built chair) already carries a phenolic film that is itself a permanent, non-reversible layer. The reversible-finish argument is strongest for the solid hardwood parts and weakest for the panels; the panels may be better treated as the consumable layer they are, refinished rarely and replaced when spent. Worth its own short note.
  3. Does a penetrating finish measurably reduce cyclic-movement checking on the hardwood keel? Testable: two offcuts, one oiled and one raw, cycled through a humidity swing, checked for surface checking.
  4. The wet-storage case: the whole doctrine assumes dry and sheltered. A chair that gets damp or sits in sun is a different problem and may want a different, more protective finish, at the cost of the reversibility argued for here.

Revisions

19 Aug 2026: first draft, from KU's questions and his 2010 Baubilt post.

19 Aug 2026, corrected the same day: KU caught a real error. The first draft treated "crosslinked" and "film-forming" as the marks of a bad finish and condemned polyurethane wholesale, which wrongly implicated tung oil (crosslinked, yet repairs invisibly) and wipe-on poly (thin, renewable, behaves like an oil). The correction: the axis that decides a hundred-year finish is FILM THICKNESS, not chemistry; nearly all durable finishes crosslink; and RENEWABLE (re-coatable in place) is distinct from and more important than REVERSIBLE (dissolvable off). Thin oils, oil/varnish blends, and thin wipe-on poly are all renewable; the thick built film is the villain. KU's 2010 wipe-on poly recommendation largely survives. Wiping varnish as thinned varnish verified against Flexner.

Renumbered 19 Aug 2026: first written the same day as TN-12 (Finishing), then folded into its reserved slot, TN-5 (Finishes and Moisture), to retire the duplicate. The content is unchanged.

Technical Notes of the Hundred-Year Chair: public from first draft, refined iteratively, forever. A claim is VERIFIED, STANDARD, DERIVED, or UNVERIFIED, and says which. Sources: FPL GTR-113 ch.15, the drying-oil polymerization literature, the Chaco timber accounts (SAPIENS), and KU's Polywhat? (Baubilt, 2010). The register · the verification ledger · hundredyearchair.com