Because a gasket seals by deforming to fill both surfaces - once compressed it has already done that and cannot do it again.

The material yields into the microscopic peaks and valleys of the mating faces under clamping load.

Removing it releases the load but not the deformation, so the material no longer matches either surface.

Composite and fibre gaskets also lose thickness permanently, which changes the clamping geometry.

Multi-layer steel gaskets rely on precisely formed embossed beads that flatten in service and do not spring back.

The cost of a gasket set against the labour to reach it makes reuse indefensible even where a gasket looks intact.

Smoother than the older composite types - multi-layer steel gaskets are much less tolerant of surface roughness.

Composite gaskets were relatively thick and conformed to a rougher machined finish.

Multi-layer steel gaskets seal through formed beads on thin steel layers and need a fine, consistent finish to seal against.

That means a head or block machined to the older standard may leak with a modern gasket, which surprises people rebuilding older engines with current parts.

Flatness must also be checked with a straightedge and feeler gauges against the manufacturer's limit, across the length and both diagonals.

Remove old gasket material with a plastic scraper and a suitable solvent rather than an abrasive disc - abrasives gouge the surface and drop grit into the engine.

Usually not. Most modern engines use torque-to-yield bolts that are stretched permanently on tightening.

A torque-to-yield bolt is tightened to a modest torque and then through a specified angle, taking it past its elastic limit so it maintains a very consistent clamping load.

Having yielded once, it will yield again at a lower load, so reusing one risks insufficient clamp or outright failure.

Where bolts are reusable, they carry a maximum length specification and must be measured; a stretched bolt is scrap.

Thread condition in the block matters equally - dirty or damaged threads give a false torque reading, so they should be chased and cleaned.

Follow the manufacturer's sequence, torque stages and any specified waiting period exactly; these are engineering requirements, not suggestions.

Where the manufacturer designed the joint for formed-in-place sealant - and then only the specified product, bead size and assembly time will do.

Many modern joints, particularly sumps and timing covers, are designed with no gasket at all and rely on an anaerobic or silicone sealant.

The bead size and path are specified because too little leaks and too much squeezes into the engine, where cured sealant can block an oilway or a pick-up screen.

Assembly time matters: many sealants must be assembled within a stated window and then left to cure before the engine is filled or run.

Adding sealant to a joint designed for a gasket is a common error that changes the joint's thickness and can prevent it sealing at all.

Where a gasket is specified, fit it dry unless the manufacturer says otherwise - most modern gaskets have their own coating.

Usually because the shaft has a groove worn by the old seal, and the new seal's lip sits in the same place.

A seal lip runs continuously on one narrow band of the shaft or pulley and slowly wears a groove into it.

A new seal fitted to the same depth runs in that groove and cannot maintain contact pressure.

The remedies are a thin repair sleeve pressed over the worn area, or fitting the new seal at a slightly different depth where the design allows it.

Seals also leak because they were fitted at an angle, with a damaged lip, or dry - most need the lip lightly lubricated and the shaft chamfer protected during fitting.

And a seal will leak regardless if the crankcase is pressurising, so check the breather and PCV system before blaming the seal.

Coolant loss with no visible leak, a pressurising expansion tank, emulsion under the oil cap, or a misfire that clears as the engine warms.

A head gasket can fail between a cylinder and a coolant passage, between two cylinders, or externally, and each gives different symptoms.

Combustion gas entering the coolant pressurises the system and pushes coolant out, frequently with no external leak at all.

Coolant entering a cylinder gives white vapour that persists once hot, and a cold misfire on that cylinder.

Emulsion under the oil cap is suggestive but not conclusive - many failures never mix oil and coolant, and short-journey condensation produces it without any fault.

A combustion-gas test on the coolant is quick, cheap and conclusive, and belongs before any dismantling.

Levels and leaks after a full heat cycle, correct coolant bleeding, and any specified re-torque.

A marginal seal shows itself after the engine has been to temperature and cooled again, not on the first start.

Cooling systems must be bled properly; an air lock causes overheating that gets blamed on the repair and can damage the new gasket immediately.

Some engines specify a head re-torque after a running-in period, and where they do it is not optional.

An early oil and filter change after major engine work removes assembly debris and sealant residue.

Record the gasket set, the fasteners and the torque procedure used - on a fleet that record is what makes a repeat failure interpretable.