Plaster in thin coats

Why is a wall built up in three thin coats of plaster rather than one thick pass that would fill it in a morning?

A plasterer facing a rough wall could, in principle, mix a large batch and lay it on thick in one pass. Instead the wall gets a scratch coat, then a float coat, then a thin finish — three operations, three waits, three chances for something to go wrong, and a job that takes days rather than a morning.

The obvious explanation is that thick material would sag off the wall, and that is true as far as it goes. But it does not explain why the practice persists on ceilings held up by suction, on horizontal work where sag is irrelevant, or with modern materials that stay put perfectly well in a thick pass. Something else is being managed, and the layering is the way of managing it.

Start with what plaster does as it sets. Whether it is a gypsum or a lime-and-cement material, the wet mix contains far more water than the chemistry needs — the excess is there to make it workable. That excess has to leave, and as it does the material shrinks.

Now ask what happens to a layer that shrinks while stuck to a wall that does not. The layer wants to become smaller in area; the background holds its edges. So the drying layer goes into tension, and it is at its weakest precisely while this is happening, because it has not yet developed full strength. If the tension exceeds what the young material can carry, it relieves itself the only way available: it cracks.

So the question becomes what governs how much tension develops. Two things, and both point the same way.

Drying is a gradient. Water leaves from the exposed face, so the surface dries and starts shrinking while the depth is still wet and plastic. In a thin layer the whole thickness dries at nearly the same rate and the differential is small. In a thick one the surface can be shrinking hard against a soft wet interior — and against a background it cannot move relative to — so the stress is much larger. The skin cracks, or it debonds.

Total shrinkage scales with thickness. A layer twice as thick contains twice the water per unit area, so its total movement is roughly twice as large. Splitting the same total thickness into two layers halves the movement each one has to accommodate — and, crucially, the first layer has already done its shrinking and hardened before the second goes on. It is now part of the background rather than part of the moving material.

That is the core of it. Layering converts one large shrinkage event into several small ones separated in time, each occurring against a background that has already finished moving.

Now every feature of the practice follows. Coats get progressively weaker and finer outward, which is not aesthetics: a strong finish would shrink and pull a weaker base apart, so each coat must be no stronger than the one beneath. The scratch coat is keyed with a comb for grip. Each coat must take up before the next, or the second coat's shrinkage acts on a base still moving. And suction is controlled by wetting the background, since a dry one pulls water from the mix too fast and cracks it for the same reason in miniature.

The thickness of the finish coat is the tell. It is very thin — a skim — because at that thickness shrinkage stresses are negligible and the surface can be worked to a polish without cracking. You could not get that surface on a thick coat at all.

Sag is a real constraint but not the main one. A thick wet mix slumps, limiting vertical work. But the practice survives where slump is not the issue and with materials formulated to resist it, so the shrinkage and gradient arguments do the deeper work. Where thickness can go on in one pass — some one-coat products on a prepared background — it is because the material shrinks less, not because the old reasoning was wrong.

Different materials shrink for different reasons. Gypsum sets by hydration and moves little afterwards; lime putty carbonates slowly and keeps moving for a long time; cement renders shrink on drying and on continued hydration. The layering logic applies to all three, but the waiting times are governed by different chemistry, which is why lime rules look absurdly patient beside gypsum practice.

"Weak over strong" is the rule most often broken, and the failure is delayed. Applying a hard cement render over soft lime, or a strong finish over a weak base, produces a surface that survives the job and fails later — because the stronger layer wins the argument and takes the weaker one off the wall with it. This is the standard damage mechanism in older buildings repaired with modern materials, and the delay between the work and the failure is exactly why the practice keeps recurring.

Suction is the variable that ruins otherwise correct work. A very absorbent background robs the mix of the water it needs to hydrate, so the plaster dries rather than sets and never reaches strength; a non-absorbent one leaves it wet with nothing to grip. Both are managed before any plaster is mixed, by wetting or by a bonding treatment — which is why an experienced plasterer spends time on a wall that looks ready.

The falsification test. If the mechanism is restrained shrinkage plus a drying gradient, then a single thick coat should crack where an equal total thickness in layers does not, the cracking should appear during drying rather than after, and reducing the water content of the mix should reduce it. If a thick coat and a layered one of the same total thickness cracked equally, the account would be wrong and something about adhesion or the background would have to explain the practice instead.