That Ceiling Stain Is From an Old Roof Leak: Find the Cause First, Then Seal It
Somebody calls us about a ceiling roughly every week with the same opening line. There is a brown mark in the corner of the living room, it has been there for years, the roof was fixed a long time ago, and can we just paint over it.
The honest answer is that we probably can, and that the word “just” is doing an enormous amount of work in that sentence. A water stain on a ceiling is not a cosmetic blemish. It is the physical residue of an event, and it holds information about that event that you cannot get any other way once the water has dried. Reading it takes twenty minutes. Painting over it without reading it takes about the same time and costs you the job twice, because roughly a third of the ceilings we are called to repaint have already been painted over once by somebody else.
Here is how we work through one, and what you can check yourself before you call anyone.
What the Stain Is Actually Made Of
Start with what you are looking at, because it explains everything that follows.
Water gets in above the ceiling. It runs along framing or along the top of the plaster, then finds a low point and soaks through. On the way it dissolves whatever it touches: tannins out of old Douglas fir framing and wood lath, iron oxide off any nail or screw it passed, sometimes soot from a chimney chase, and mineral salts out of the lime plaster itself. When it eventually evaporates through the ceiling surface, everything it was carrying stays behind, concentrated at the point of evaporation.
That is the brown ring. The dark edge is where the wet area stopped advancing and evaporated hardest, so the deposit is thickest right at the boundary. The paler middle is where the water sat but kept moving. And that deposit is not on the ceiling, it is in it, sometimes the better part of a millimeter down into the plaster.
Now the crucial property. All of that residue is water soluble. It got there dissolved in water and it will redissolve in water. When you roll a coat of latex paint over it, you are applying a wet, water borne film directly onto a dried solution, and it goes straight back into suspension and wicks upward through your paint as it dries. Two days later the mark is back, usually looking softer and slightly larger than it was, with a fuzzy edge instead of a sharp one. That fuzzy edge is the diagnostic. It means the stain came through the paint rather than showing under it.
Three more coats will not fix it. We have seen ceilings with five coats of good quality ceiling white on them and the ring still showing, because every coat was another delivery of water to the same soluble deposit.
Old Leak or Live Leak
This is the part that actually matters, and the part almost everyone skips.
A repaired leak from eight years ago and a slow flashing failure that only runs in a driven rain leave marks that look identical in July. Same color, same shape, same hard edge. You cannot tell them apart by looking, and neither can we. What separates them is measurement, access, and your own account of the history.
The meter. We use a pin type moisture meter and take a reading in the stained plaster, then another in clean plaster a few feet away in the same ceiling. The absolute number is close to meaningless on old lime plaster, because the material and its age skew it. The comparison is what tells you something. If the stain reads meaningfully wetter than its own neighbors, something is still feeding it. If the two readings sit on top of each other, the ceiling is uniformly dry today, which is good evidence but is not proof, because a leak that only runs in a hard rain will read completely dry after a Pasadena summer.
The attic. Where there is access, this is worth more than the meter. Twenty minutes with a headlamp above the ceiling usually settles it. You are looking at the underside of the sheathing for dark tide lines and for the pale grey fuzz of old fungal growth, at the framing for staining that runs in a line, at the top of the plaster or the back of the drywall directly above the mark, and at anything that penetrates the roof anywhere uphill of it. Vent stacks, swamp cooler curbs, old antenna mounts, the low side of a valley. Deep overhangs and exposed rafter tails, which nearly every Craftsman bungalow in this city has, mean the leak often starts further up the slope than the stain suggests. Water travels along a rafter and drops off wherever the rafter is interrupted.
Two more things you can read up there. Old staining goes matte, flat and slightly chalky, and any fungal growth on it is dry and dusty. Recent wetting looks darker, sometimes still slightly glossy at the edge, and growth on it is soft. And if the sheathing is patched with new plywood above the stain, that is your repair, and it tells you the roofer found the same spot.
Your history. When did you first see it? Has it grown? Has it changed after a storm? A stain that appeared once and has been the same size for six years is almost certainly dead. A stain that gets a new darker halo after a wet winter is alive, no matter what the meter says on a hot afternoon.
If those three sources agree, you have your answer. If they disagree, the right move is to wait for a rain and look again. That is not us stalling. It is cheaper than sealing, painting, and doing it all a second time in February.
Then, and Only Then, Seal It
Once the source is finished and the material is dry, the fix is not more paint. It is a stain blocking primer, which is a specific product category and not the same thing as the primer built into a self priming paint.
Pigmented shellac is the workhorse. It is carried in alcohol rather than water, so it never redissolves the deposit and never offers it a path upward. It grips old chalky plaster that other primers struggle to bond to. It dries hard in about half an hour and it blocks tannin, rust, smoke and water marks equally well. Its one real drawback is a sharp solvent smell for the first hour, which is fine with a window open and genuinely difficult in a child’s bedroom.
Water based stain blockers are much better than they were, and on lighter marks in occupied rooms they are the sensible choice. They want a properly dry, sound surface and they are less happy on chalky lime plaster, which is why on the older housing stock here we still lean shellac.
One coat, properly applied, does more than four coats of finish paint. Feather the primer edge and sand it flush, because a primer ridge under a flat ceiling paint will catch the raking light from a window and be just as visible as the stain was. Then two finish coats in a dead flat ceiling white. Flat, not eggshell, because any sheen at all turns a hand floated plaster ceiling into a relief map.
When the Plaster Itself Has Failed
Sometimes the sealing conversation never gets started, because the plaster is gone.
Lime plaster on wood lath is held up by keys: the squeezed out ridges of the first coat that pushed through the gaps between the lath strips and hooked over the back. Water that sat for long enough softens those keys and they snap. The plaster then hangs on friction and on the paint film, which is exactly as reliable as it sounds.
You test it with a flat hand, pressing gently over a wide area. Sound plaster feels solid. Failed plaster gives, and often makes a faint gritty crunch as broken key fragments shift in the cavity. A ceiling that does that does not need primer. It needs the loose area cut back to a sound edge and rebuilt in coats, or in a bad case, screwed back to the lath with plaster washers before anything else happens. Skimming over it buys a year at most.
What This Means for the Money
Sealing and blending a single dead stain in a decent ceiling is a small job. It is a couple of hours, one room, and the result should be invisible. That is by far the most common outcome and it is worth asking for by name.
Full ceiling repaints happen when there are several marks, or when the ceiling has yellowed unevenly with age, because a bright fresh patch on an aged ceiling reads as a repair from the doorway even when it is technically perfect. We hold a work light flat along the surface and show people which of those two situations they are in, then let them choose.
The one thing we will not do is seal over something still wet, and the one thing we would ask you not to do is roll a coat of white over the mark before someone has looked above it. That coat costs you nothing much in materials and it removes the evidence, because a painted over stain is harder to read and its edge no longer tells you where the water stopped.
A photograph is a good start. Take one straight on in daylight, and one from a low angle so the surface texture shows. Call (626) 263-0391 and we can usually tell you from those two pictures whether you are looking at a twenty minute answer or a rainy day wait.