Our windows and doors have the paper pasted on the inside face. That leaves the slats and stile-and-rail frame exposed outdoors, taking direct sun and rain. That alone puts them on different footing from a door like a shoji that stands entirely indoors — starting with how you even choose a finish.

How Wood Left Outdoors Breaks Down

UV in sunlight breaks down the lignin in wood. Lignin is what binds cell to cell, so once it breaks down the surface turns chalky. Rain washes the broken-down material away, mold settles into its place, and within about half a year the wood has turned grey.

Fortunately, this kind of breakdown happens only at the surface. UV is said to penetrate only about 75µm into wood, visible light about 200µm. Leave checking aside, and the damage stays within the outer 1mm. Even bare softwood left to the weather with nothing applied erodes at only around 6mm per century. That's why a door gone grey still turns out sound underneath once you run a plane over it.

Four Branches of Finishing

The first is applying nothing at all. Plane it and leave it. Interior joinery in Japan is often left this way, and when it yellows with age the fix is to wash it rather than repaint it — a method called aku-arai, "lye washing." Lye is an alkaline water made by steeping and straining ash from burned straw or wood; it used to be used for laundering and cutting grease. These days, it seems, the name often survives while the actual work is done with a dedicated commercial cleaner. Left outdoors, the wood turns grey and slowly erodes — but at least there's no film to peel.

The second is a penetrating oil. Raw perilla oil, or the egoma (perilla) oil used in Japan, is rubbed in by hand. It wets the wood, deepening the color a shade and shedding a little more water. But it blocks almost no UV. With no film, there's nothing to peel, either — once it fades, you just rub more in. The trade-off is that you have to apply it often. The USDA Forest Products Laboratory (FPL) puts this kind of finish's outdoor service life at 1–2 years initially, 2–4 years between recoats. It's also vulnerable to mold: the same source notes that drying oils serve as a food source for mold, and domestic testing found raw perilla oil did almost nothing to prevent it (a 4-week mold-resistance rate of 4.10%).

The third is a clear, film-forming finish. This covers boiled oils like myeongyu, tung oil (dongyu), varnish, polyurethane, and lacquer. These shed water well, but on a surface that gets direct sun, the film will eventually lift.

The fourth is a pigmented finish — tinted oils, semi-transparent stains, paint, anything that lays down color. The pigment blocks light, so these last far longer than clear finishes, but at the cost of obscuring the grain, partly or entirely.

Why Clear Films Lift

On the jobsite you often hear "an oiled film lifts in the sun." That's true — but it isn't the film that fails first. It's the wood underneath it.

A clear film doesn't block all UV and visible light. Light passes through it, reaches the wood surface underneath, and breaks down the lignin there, degrading the wood. That very layer the film was bonded to weakens, and the film loses its grip — less the film itself shattering than the bond beneath it giving way. Meanwhile the film itself grows brittle under sunlight and more prone to cracking. It can't keep pace with the wood as it swells and shrinks, so it cracks, and water gets in through the cracks. In the end a soft film lifts and peels; a hard one flakes off in pieces.

USDA literature notes that varnish tends to fail first over earlywood. Hokkaido's Forest Products Research Institute recorded clear coatings turning grey starting from cracks, with mold growing between the film and the wood. In test panels left outdoors for three years, the film still looked sound on the surface while the latewood tissue underneath had already begun to crack.

Once a film has lifted, it's also awkward to repair. Peeled and still-bonded patches sit side by side in a mottled pattern, which makes spot repair difficult; usually you have to strip the whole thing and start over. This is the sharpest contrast with a penetrating oil.

How Long Does It Last?

I've put together figures from the USDA Forest Products Laboratory (Williams 1999) and a 2022 presentation by Atsushi Kataoka of Japan's Forestry and Forest Products Research Institute.

÷ The Japanese figures are transcribed from a table in a lecture handout, so the original source still needs to be verified.

Both the American and Japanese data agree that clear films are the shortest-lived option. FPL goes so far as to say it doesn't recommend varnish for exterior wood at all, and Japan's building-standard specification for construction work (JASS 18) apparently has no provision at all for clear exterior wood coatings. Hokkaido's Forest Products Research Institute likewise advises reserving clear finishes for spots that get neither direct sun nor rain. A water-based polyurethane sign kept out of the rain, for instance, reportedly developed only minor surface crazing even after 5 years — so where you put a given finish matters nearly as much as which finish it is.

Lacquer Is a Different Story

Lacquer also forms a film, but it fails differently. Rather than peeling, it erodes from the surface inward and takes on a whitish, powdery look. Kyoto City's Industrial Technology Research Institute tracked lacquered metal fittings left outdoors for ten years; where exposure to sunlight was strong, the black-lacquer film had lost 243µm of thickness.

Because of this, the original application was built around a film over 400µm thick from the start, with a maintenance routine of wiping it down with water every ten days. In testing of tsugaru-nuri lacquerware in Aomori Prefecture, a surface exposed to a year of direct sunlight had powdered away entirely, leading researchers to conclude that lacquer film simply isn't suited to outdoor use. Domestic testing found the same: raw-lacquer film exposed to UV developed pitting and cracks on the surface, and after 600 hours its strength had dropped sharply.

As a finish for interior faces or indoor sliding doors, lacquer is excellent and long-lasting — but it isn't suited to slats that take direct sun outdoors.

What the Korean Record Shows

Myeongyu comes up often as a finishing oil in Joseon-era architecture. It's made by boiling raw perilla oil together with pyrolusite (manganese dioxide), litharge (lead oxide), and alum, and it appears in nearly every uigwe (royal protocol record); one entry records 400 mal of myeongyu used for palace finishing work in the 7th year of King Taejo's reign. Raw perilla oil takes a long time to cure, but myeongyu cures in roughly 12 hours, and in mold-resistance testing it blocked 87.67% of mold growth when applied by soaking. There's also research suggesting that the dongyu (tung oil, literally "paulownia oil") mentioned in the uigwe actually refers to boiled perilla oil or myeongyu itself — the scholar Seo Yu-gu reportedly wrote that "since we don't know how to cultivate the tung tree, we call boiled perilla oil 'dongyu' instead."

Unfortunately, I haven't been able to find published test data on how well myeongyu holds up in sunlight, or whether it lifts. As a film-forming oil, it seems hard to avoid the same fate as other clear films eventually — but for now that's only a guess. In current heritage restoration work, raw perilla oil is what's used as the finishing material for dancheong (traditional painted ornamentation).

Surveys of Japanese folk houses tell a similar story. In interior lattice windows and the boxes storing storm shutters, researchers found oil that had been boiled with a lead drier (presumed to be tung oil); but on exterior woodwork, weathering was so severe that it was impossible to tell what, if anything, had been applied. Exterior finishing, in other words, has left behind neither record nor trace.

Maintenance

The first priority is not missing the window for recoating. Hokkaido's Forest Products Research Institute holds that the best time to recoat is right as degradation is just beginning. Once a film starts lifting, you have to strip it all; catch it before that, and a thin recoat is enough. In testing by Japan's Forestry and Forest Products Research Institute, the point at which water-repellency dropped below 80% of its original value was used as the trigger for recoating.

Rounding over the edges also helps a finish last. When Hokkaido's Forest Products Research Institute left 11 film-forming coatings outdoors for four years, square edges and 45° chamfers cracked and began peeling within a year, while rounded-over edges degraded far less. Even with the same coating, film thickness came out to 257µm on a flat face versus only 97µm on a square edge — the film simply runs thinner at corners. The rounding on a slat or stile-and-rail edge, in other words, isn't just about shape; it's tied directly to how long the finish survives.

Color faded by sunlight comes back with a plane, because the degradation stays at the surface. By contrast, black staining from contact with iron, or discoloration from mold, runs deeper and doesn't come out as easily. It's best to simply keep metal from touching wet wood in the first place.

Even within the same house, different windows and doors weather at different rates, because how deep the eaves run, which direction a window faces, and whether it has storm shutters all change how much sun and rain it actually takes. One Japanese source notes that a south-facing vertical surface erodes at only about half the rate of a surface tilted 45° toward the south.