Keeping furniture from getting worse: humidity, light and moving it
Most of what happens to furniture happens slowly, in the room it sits in. What the year does to a tabletop is arithmetic, and the rest is where you put the piece and how you carry it.
- Start here: the essentials
- 1. The slow things are the ones that matter
- 2. Measure before you decide
- 3. The arithmetic
- 4. Telling seasonal movement from damage
- 5. Vents, radiators, windows, basements
- 6. Light
- 7. Moving it without breaking it
- 8. If it gets wet
- 9. See it: why Priya's doors stick in August
- 10. Practice: the room survey
- 11. Your piece
- 12. Check your work
- 13. Knowledge check and scenario
Start here Essentials
At minimum this chapter asks you to put a hygrometer in the room, measure the piece, move it off whatever is harming it, and measure and date any crack. The equation, the light figures and the water emergency are Go deeper on a first pass.
- Put a hygrometer in the room and read it twice, six months apart. Nobody can tell you your own numbers (section 2).
- Measure the piece and find out what the year does to it. Width across the grain, and the species if you can narrow it (section 3, and Humidity and movement).
- Move it off the thing that is hurting it. A vent, a radiator, a sunny window, a damp basement (section 5).
- Measure and date every crack, so the next reading means something. One reading cannot tell you anything (your condition report).
Minimum time: about 30 minutes now, and five minutes six months from now. Done when: your report says where the piece lives, what the humidity was when you read it, and the width and date of any crack.
1. The slow things are the ones that matter Understand
Everything else in this manual is about something that has already happened. This chapter is about what is happening now, at a rate too slow to notice, and it is where the most damage is done to furniture over a lifetime.
Three slow agents account for most of it: moisture, which makes wood move and can open joints and split panels; light, which fades and degrades; and handling, which breaks things in a moment but is included here because the damage is cumulative across many moves.
Physical principle The moisture one is worth understanding rather than just obeying, because it explains behavior people otherwise read as damage. Wood exchanges moisture with the air around it and keeps doing so for as long as it exists. It gives up moisture in a dry indoor winter and takes it up in a humid summer, and as it does it shrinks and swells across the grain. Along the grain it barely moves; the federal Wood Handbook puts longitudinal shrinkage from green to ovendry at between 0.1 and 0.2 percent for most species, against several percent across the grain. That asymmetry is why furniture is built the way it is, and why a panel held rigidly in a frame splits while a panel free to move does not.
The conservation guidance's version of this is short: wood and other organic materials respond to changes in relative humidity by expanding or contracting as they try to maintain equilibrium with the moisture in the environment. Ideally, it says, relative humidity should be maintained within a 40 to 60 percent range. If it is too high, above 70 percent, wood and other materials expand, and if they are constrained in any way they may split on shrinking when the humidity drops. It adds that changes can cause a coating to detach, and that prolonged high humidity promotes mold growth and insect infestation.
Those are museum figures, measured and recommended for collection objects in managed environments, and this manual does not tell anyone to hold 40 to 60 percent relative humidity in a house. Most houses do not, most furniture survives, and a reader who takes a museum number as a household requirement will spend money on humidifiers rather than on moving a chest off a radiator.
What the figures are good for is a scale. They tell you what the people who study this consider comfortable for wood, which gives your own two readings some meaning. And the arithmetic in the next sections tells you what your own numbers do to your own tabletop, which is a far more useful thing than a target you are not going to hit.
2. Measure before you decide
The first step is to buy a hygrometer and put it in the room, which is the honest answer to "what humidity is my furniture in." Nobody can tell you from a map or a season, because indoor humidity depends on your heating, your ventilation, your cooking and how many people breathe in the room.
The conservation guidance says the same thing: monitoring temperature and relative humidity can be done with thermometers and hygrometers bought at electronic or hardware stores, and the humidity can be modified to stay within acceptable ranges with humidifiers and dehumidifiers.
Two readings matter, and they are six months apart: one in the driest part of your year, which in a heated house is usually deep winter, and one in the most humid, usually late summer. A single reading tells you where the room is today. The pair tells you the swing, and the swing is what opens joints and splits panels.
Write both down with their dates on your condition report, and put the second reading in a calendar now, because that is the step that otherwise does not happen.
3. The arithmetic
This is the one place in this manual where a reader needs a number rather than a judgment, and there is a federal reference for it.
The USDA Forest Products Laboratory's Wood Handbook publishes two things that together answer the question. First, an equation relating temperature and relative humidity to the equilibrium moisture content of wood, which is the moisture content at which wood is neither gaining nor losing moisture in those conditions. Second, tables of shrinkage by species: how much a given wood changes dimension, radially and tangentially, as it dries.
Put together, those let you work out roughly how much the top of your table changes in width between your winter and your summer. Humidity and movement does the arithmetic, with the equation, the species figures and the handbook's own cautions printed under it.
Three things about the calculation matter more than the answer.
It depends on the cut, not just the species. Wood moves roughly twice as much tangentially as radially in many species, so a plainsawn board moves considerably more across its width than a quartersawn one of the same wood. The handbook's instruction is to use tangential values for estimating the width shrinkage of plainsawn material, radial values for quartersawn, and tangential values for mixed or unknown ring orientation, which is why the calculator defaults to the larger figure where you are not sure.
The species matters, and you probably cannot name it. White oak and black cherry differ by about a third in tangential shrinkage. What your piece of furniture is actually made of explains why naming a species by eye is unreliable, and the practical consequence here is that the calculator's answer carries that uncertainty with it. Run it for two candidate species and see whether the difference changes any decision you would make.
The handbook contradicts its own precision, deliberately. Its moisture content table is printed to 0.1 percent, and the chapter says that level of precision is provided only to illustrate trends and that in reality it is not meaningful. On shrinkage it says that accurately predicting the shrinkage of an individual piece of wood is impossible, while the relationship between average shrinkage and moisture content over a sufficient quantity of pieces can be predicted accurately. It puts transverse and volumetric shrinkage variability at about a 15 percent coefficient of variation.
Physical principle So the right way to read the calculator's output is as an order of magnitude rather than a measurement: whether the year moves your tabletop by a thirty-second of an inch or by half an inch. That distinction is enough to settle every question in this chapter, and the second decimal place is not.
4. Telling seasonal movement from damage
This is the practical payoff, and it saves people from gluing things that did not need gluing.
Seasonal movement is reversible and repeats. A gap that opens in February and closes in August is the piece breathing. Damage does not reverse: a split that was 1/32 inch last winter and is 1/16 inch this winter is progressing.
The way to tell them apart is the only way there is: measure, date, and measure again. One reading cannot distinguish them, and most reports contain exactly one.
| What you see | Measure | If it is the same next season | If it is wider in the same season a year on |
|---|---|---|---|
| A gap between two boards in a top | Width at a marked point, to the nearest 1/32 inch | Stable, and probably a glue line that opened once | Progressing: worth a professional opinion |
| A crack in a panel | Length and width at a marked point | Stable | Progressing |
| A drawer that binds | Which season, and where it binds | Movement if it frees in the other season | Wear or racking if it binds in both |
| A door that will not close | Which season, and which edge rubs | Movement | A structural change: see Joints, racking and breaks |
| A joint showing a line at the shoulder | Gap at a marked point | Seasonal, if it closes | A failing glue line |
Mark the measuring point. A pencil tick on an unseen surface, or a photograph with a ruler in frame, so the second reading is taken in the same place. A crack measured at its widest point in February and at a random point in August produces a comparison that means nothing.
5. Vents, radiators, windows, basements
Most furniture that is suffering from its environment is suffering from one specific thing it is sitting next to, and moving it is free.
The conservation guidance names the places to avoid directly: place furniture in areas of minimum temperature and humidity extremes, and avoid storing furniture in attics and basements, or placing pieces near fireplaces and heating vents.
| Location | What it does | What to do |
|---|---|---|
| Over or beside a heating vent | Repeated local drying, often on one side of a piece only, which is the recipe for a split | Move the piece, or deflect the vent. This is the single most common fixable problem |
| Against a radiator | The same, plus heat | Move it, or put something insulating between |
| In front of a fireplace in use | Heat and drying in bursts | Move it |
| In direct sun | Fading and finish degradation, cumulative and irreversible | Move it, or manage the light: section 6 |
| In an unheated attic | Extremes of temperature and humidity, both directions | Named by the guidance as a place to avoid |
| In a basement | Prolonged high humidity, which promotes mold and insect infestation | Named by the guidance as a place to avoid |
| Against an exterior wall in a cold climate | A cold surface where moisture can condense | Pull it a few inches off the wall for air circulation |
Note what the first three rows have in common: they dry one part of a piece faster than the rest. Uneven moisture change is harder on furniture than a uniform one, because the parts that change are restrained by the parts that do not. That is why a tabletop with one end over a vent is at more risk than a whole room that is simply dry.
6. Light
Light damage is the quietest entry in this chapter and the one with the least recourse.
The conservation guidance states it plainly: light, particularly visible and ultraviolet light, is very damaging to organic materials such as wood, and damage from light is cumulative and irreversible. Clear finishes often turn yellow or opaque in response to light, and the color or cellular structure of the wood itself can also change. And, as Surfaces, veneer and hardware sets out at length, it says the resulting damaged finishes and bleached wood cannot be restored to their original color without stripping and refinishing, which it calls a practice not recommended.
Its remedies are ordinary and cost nothing much: move all furniture out of direct sunlight, use blinds or curtains to block the light's intensity, and apply UV-filtering film to windows, which it says screens out ultraviolet light.
For a figure, the Park Service's leaflet on upholstered furniture gives one: attempt to keep light levels below 50 lux, which it gives as 5 footcandles, and do not display upholstered furniture in direct sunlight. It adds that furniture should be totally protected from light when not being viewed, by switching the light off at source or covering the object with clean cotton sheets.
Convention Fifty lux is a museum gallery level and it is dim: a lived-in room is usually several hundred. This manual gives the figure with its setting named rather than as a target, and the useful household version is the direction rather than the number: less light is better, direct sun is the thing to eliminate, and the damage adds up over years rather than arriving in an afternoon.
7. Moving it without breaking it
The conservation guidance gives a specific sequence, and it is worth following because every instruction in it corresponds to a way furniture is commonly broken.
- Before moving a piece, examine it for loose or damaged joinery. If what you find is a break in something load-bearing, Joints, racking and breaks applies and the piece may not be safe to lift.
- Remove elements such as shelves, doors and drawers. Where doors will not come off, secure them by locking or by wrapping the case with soft cotton straps.
- Lift tables by the apron or legs rather than by the top, which the guidance warns could detach.
- Lift chairs by the seat rails, not by the arms or crest rail.
- Lift, do not drag. The guidance's reason: excessive lateral pressure on legs and feet can cause them to shear off.
- Carry a marble top separately and vertically, as you would a sheet of glass.
- In a vehicle, place the piece on its back or its top, not on its legs.
The Park Service's water-emergency leaflet gives the same principle in a different form, which is a useful generalization: move furniture by its lowest structural member rather than by grasping the top board or the original handles. Lift tables by their aprons, non-upholstered chairs by their seat rails rather than arms or crest rails, upholstered chairs by their legs with two people, and case pieces by their bottom rails, removing drawers to lighten the load.
On how many people: the State Historical Society of Iowa's collections guidance says to limit movement to only when necessary, and to use at least two people for anything heavier or bulkier than a typical dining room chair. That is a usefully concrete threshold.
8. If it gets wet
A burst pipe, a leaking roof, a flooded basement. The published emergency guidance has one counterintuitive instruction at its center, and it is the one worth knowing in advance.
The Park Service's leaflet on water-soaked furniture lists the damage modes: surface mold growth, splitting and deformation of wood members, lifting of veneers, staining of finishes and failure of glue joints. Its procedure, in brief: get the piece off the water, or where moving it risks more damage, block it up off the floor instead. Blot rather than scrub. Promote even air circulation: open doors, lift lids, remove drawers, block short-legged cases up, and use fans but do not aim them at the object.
The leaflet's central instruction: furniture and other wooden objects should not be dried out too rapidly, because rapid drying and the resulting stresses can cause splitting and deformation of the wood members. It suggests monitoring relative humidity and letting it fall slowly over the course of days to approximately 55 percent, and says that if the humidity is falling too rapidly, furniture can be temporarily covered with polyethylene sheeting to retard evaporation.
This runs against every instinct, which is to get a wet thing dry as fast as possible. On furniture, fast drying is how a soaked piece becomes a split one.
The leaflet also notes that period furniture was often made with protein-based glues such as hide and fish glue, which are soluble in water to varying extents, so joints, veneer and appendages may come loose on a water-soaked object. Expect it, and keep any detached parts bagged and labeled, as Writing a condition report sets out.
On mold, which follows water: it says lowering the ambient temperature slows mold growth, that growth slows below 70 °F and ceases entirely below 40 °F, so in an emergency it advises turning the heat down in winter and using air conditioning in summer, with a dehumidifier if condensation appears on walls. The health side of mold is in What can hurt you in old furniture.
How the equation gets from a room to a moisture content
The Wood Handbook's Equation 4-5 is a sorption model with four temperature-dependent parameters, and the calculator uses the chapter's published Fahrenheit parameter set. You put in a temperature and a relative humidity, and it returns the moisture content the wood in that room is heading for.
Two sanity checks worth knowing, because they are in the handbook's own table and this manual's calculator reproduces them. At 70 °F, the equilibrium moisture content is about 6.2 percent at 30 percent relative humidity, about 9.2 percent at 50 percent, and about 16.0 percent at 80 percent. So the same room at the same temperature puts wood at nearly three times the moisture content across the humidity range people actually live in.
The handbook notes that these values were derived primarily for Sitka spruce, under a condition described as midway between absorption and desorption, and that actual values vary considerably with species, humidity, temperature and how the measurement was made. Species with high extractive content tend to sit lower at high humidity. So the moisture content figure is a good general model rather than a measurement of your board.
From moisture content to inches
The second half uses Equation 4-9, which gives shrinkage from the green condition to a given moisture content as the full green-to-ovendry shrinkage figure scaled by how far below fiber saturation the wood is. Fiber saturation is the moisture content below which wood begins to shrink, and the handbook says the point below which most wood properties start to change averages about 30 percent moisture content while varying by several percentage points between species and individual pieces. Where the real value is unknown, it offers 30 percent as an approximation, which is what the calculator uses.
Differencing that between two moisture contents gives the change in width between winter and summer. The arithmetic is the handbook's; the differencing is the ordinary application of it.
What the result assumes, and the calculator says so under itself: that the whole board has reached equilibrium with the room. A thick top takes months to get there and a finish slows the exchange further, so a real piece lags its room and falls short of the extremes the calculator shows. Treat the figure as the outer bound of the movement rather than as this week's dimension.
Why the same room is two different environments
Worth stating because it is the insight the whole calculation exists to deliver. A heated house in January and the same house in August are not the same environment for wood, even at the same thermostat setting, because heating dry outdoor air drives indoor relative humidity down in winter while summer air brings it up. The furniture does not experience a stable home with occasional events; it experiences an annual cycle, every year, for as long as it exists. Every joint, every panel and every drawer in your house is working through that cycle right now.
That is why the two readings six months apart are the measurement that matters, and why a single reading, however precise, answers a less interesting question.
UV film, and what it does not do
The conservation guidance recommends UV-filtering film on windows, and the Park Service publishes a leaflet on choosing one. What such a film does is screen out ultraviolet radiation, which is the most damaging part of the spectrum for organic materials. What it does not do is stop visible light, which the same guidance names as damaging in its own right, saying light, particularly visible and ultraviolet, is very damaging and that the damage is cumulative and irreversible.
So film is worth having and it is not a solution on its own. A piece in front of a filmed window in direct sun is still in direct sun. The order of effectiveness is: move the piece out of the sun, manage the light with blinds or curtains, and use film as well.
9. See it: why Priya's doors stick in August See it
Priya's sideboard. Sliding doors that stick in August and run freely in February. The piece lives in a living room with a forced-air vent about four feet away.
What she measured. A hygrometer in that room read 30 percent in February and 60 percent in late August, both at about 70 °F. The door panel is about 18 inches wide across the grain.
What the calculator says, under her assumptions. At 70 °F, 30 percent relative humidity puts the wood at about 6.2 percent moisture content and 60 percent puts it at about 11.0 percent, a swing of about 4.8 points. Assuming the wood is teak and the panel is plainsawn or of unknown cut, which takes the tangential figure of 5.8 percent, an 18 inch width works out at roughly 0.17 inch of seasonal change, or about three sixteenths of an inch.
Known: the two humidity readings with their dates, the panel width, and the doors' behavior across the year.
Assumed, and both assumptions are Priya's: that the wood is teak, which What your piece of furniture is actually made of says a reader cannot settle by eye, and that the panel is plainsawn or mixed, which is the handbook's own default where the cut is unknown. The calculator uses the larger of the two shrinkage figures in that case, so the estimate leans toward more movement rather than less.
What that settles, under those assumptions. Three sixteenths of an inch across an 18 inch panel is a plausible amount of seasonal movement, and it is the right order of magnitude to explain a door that binds in one season and runs in the other. So the sticking is potentially the piece responding to its room rather than a fault, and planing the door in August would leave a loose door in February.
What that does not settle. It does not establish that the wood is teak, that the piece has no other problem, or that the movement is exactly three sixteenths. The handbook says accurately predicting the shrinkage of an individual piece is impossible, and this estimate carries that. What it does is tell Priya which kind of explanation to prefer and what to measure next.
What she actually does. Moves the sideboard away from the vent, which is free and which addresses an uneven drying problem the calculator says nothing about. Measures the gap at the sticking edge in August and writes the date. Puts a reminder in her calendar for February to measure it again. Does nothing to the door.
The open question that the second reading answers: whether the gap really does change by about that much, which is the test of whether the whole explanation holds.
The sources behind this chapter:
- Moisture relations and physical properties of wood (Glass and Zelinka, chapter 4 of the Wood Handbook, FPL-GTR-282, USDA Forest Products Laboratory, 2021): the equation, the shrinkage tables, the fiber saturation approximation, and the chapter's own cautions about its precision. The whole of Humidity and movement rests on it.
- Caring for Your Treasures (American Institute for Conservation): the furniture leaflet, for the humidity range, the places to avoid, the light statements, UV film, and the moving instructions in section 7.
- Emergency treatment for water-soaked furniture and wooden objects (National Park Service, Conserve O Gram 7/7): the water procedure, the lifting points, and the instruction not to dry too fast.
- Upholstered furniture: agents of deterioration (National Park Service, Conserve O Gram 7/4): the 50 lux figure, the humidity thresholds for textiles, and covering with cotton sheets.
- Collections care: wooden objects and furniture (State Historical Society of Iowa): limiting movement, and at least two people for anything bigger than a dining chair.
10. Practice: the room survey Practice
Minimum: one hygrometer reading, one measured and dated crack or gap, and one thing moved. Time: about 30 minutes now, five minutes in six months. Materials: a hygrometer, a ruler or a tape, a pencil, a phone camera, and your condition report open.
- Put the hygrometer in the room with the piece, not in the hall, and leave it an hour before reading it.
- Walk the room and name what the piece is next to: vents, radiators, fireplaces, exterior walls, windows and where the sun actually falls at different times of day.
- Measure the piece across the grain on the widest panel or top, and record the width.
- Run Humidity and movement with your reading, and again with a second candidate species, to see whether the uncertainty changes anything.
- Mark and measure every crack and gap, with a pencil tick at the measuring point and a photograph with a ruler in frame. Write the date.
- Move the piece off whatever it is next to, if anything, using the lifting points in section 7.
- Put the second reading in your calendar, six months out, and note what you will measure.
The two-species run. Run the movement calculator twice for the same piece, once with your best guess at the species and once with the next most likely candidate, keeping the width and humidity readings the same. Then ask whether the two answers would lead you to different actions. Usually they do not, which is the reassuring result: the uncertainty about the species, which the materials chapter says you cannot resolve by eye, turns out not to matter for the decision in front of you. Where the two answers do differ enough to matter, you have found the one case where getting the wood identified would actually change something.
11. Your piece Your piece
12. Check your work Check your work
Criteria
- Does the report say where the piece lives, specifically enough to include what it is next to?
- Is there a humidity reading with a date, and a calendar entry for the second one?
- Does every crack and gap have a width, a marked measuring point and a date?
- If you ran the calculator, is the species assumption recorded alongside the answer?
- Have you moved the piece off any vent, radiator or direct sun, or recorded why not?
- If you moved it, did you check for loose joinery first and lift rather than drag?
Worked example
Priya's entry records both humidity readings with dates and temperature, the panel width, the calculator's answer with both assumptions named, the action taken, which was moving the piece off the vent, and a dated measurement of the sticking gap with a February recheck in the calendar. Checked against the criteria it passes on all six. What makes it genuinely useful is the last part: the February reading is the test of the whole explanation, and without it this is a plausible story rather than a finding.
Common mistakes
- Guessing the humidity. Indoor humidity depends on your house, and a hygrometer is inexpensive.
- Taking one reading. The swing is what matters, and one reading does not contain it.
- Treating the museum range as a requirement. It is a scale for interpreting your numbers, not a target for a house.
- Reading the calculator to two decimal places. The handbook says that precision is not meaningful; use it to tell a thirty-second from a half inch.
- Measuring a crack at a different point each time. Mark the point.
- Planing a sticking door in August. Find out whether it is seasonal first.
- Drying a soaked piece fast. The emergency guidance says the stresses from rapid drying split and deform members.
- Lifting a table by its top. The guidance says it could detach; use the apron or the legs.
- Dragging a piece across the floor. Lateral pressure on legs and feet can shear them off.
What is still unresolved
The calculator gives an estimate carrying the species uncertainty from What your piece of furniture is actually made of and the handbook's own statement that an individual piece is not predictable accurately. It does not tell you whether a particular crack is seasonal, which only two dated readings can; it does not tell you what humidity your house should hold, which this manual does not claim to know; and it does not address faded or bleached wood, which Surfaces, veneer and hardware explains is largely final.