Lead in short lengths
A Socratic walk-through of lead in short lengths — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a roofer cut flashing into short pieces when one continuous strip would obviously leak less?
A run of flashing along a roof abutment is weatherproofing, and the instinct with weatherproofing is to avoid joints. Every joint is a place water could get through. A single continuous strip along the whole run has no joints at all, and lead comes in rolls long enough to do it.
The roofer cuts it into pieces of a metre or so, laps them, and leaves each piece free to move. The job now has a joint every metre. Done the other way, the continuous strip fails — not by leaking at first, but by tearing itself apart — and it does so in a way that looks nothing like a weatherproofing failure.
Reasoning it through
REASONING #Start with the material. Lead expands and contracts with temperature like everything else, but two features make it unusual here. Its coefficient of thermal expansion is high compared with the masonry and timber it is fixed to. And a roof surface is a punishing thermal environment: dark, exposed, in direct sun, swinging through a large range between a summer afternoon and a clear night, every day of the year.
So a long strip of lead is going to change length appreciably, twice a day, indefinitely. Ask what happens to that movement if the strip is fixed at both ends.
The lead cannot lengthen freely, so it goes into compression on heating, and being soft it does not resist — it buckles or creeps sideways, taking up the extra length as a small permanent deformation. On cooling it contracts from that deformed state, so it is now in tension, and it stretches slightly to make up the difference. Next cycle, the same again.
Notice the shape of that: each cycle leaves a small permanent change, and the changes accumulate in one direction rather than cancelling. The lead is being worked back and forth, a little more each day, at whichever point is most restrained. Lead creeps readily at ordinary temperatures, so it does not spring back.
Follow it far enough and the result is fatigue — ripples and thickening where material has been pushed, thinning where it has been drawn, and eventually a split. The split appears at the point of greatest restraint, which is usually near a fixing or at a change of direction, and it appears after years rather than months, which is part of why the cause is so often misdiagnosed as poor material or a manufacturing fault.
Now the remedy is obvious once the mechanism is clear. If the trouble is accumulated movement over a length, reduce the length. A piece of a metre expands by a fraction of what a piece of ten metres does, and if each piece is fixed along one edge only and free elsewhere, its movement is small enough to be absorbed without permanent deformation. The laps between pieces overlap generously in the direction water runs, so they shed water by geometry rather than by being sealed.
That last point is the conceptual move worth naming. The joints are not holes that have been tolerated; they are the mechanism. A lap where the upper piece overlaps the lower by a sufficient distance is weatherproof without being watertight — water runs across the joint and away, gravity does the work, and the two pieces are free to slide relative to each other. The detail is designed to leak-proof and move, which a sealed continuous strip cannot do.
Two practices follow. Each piece is fixed at its head and left loose at its foot, so movement runs one way rather than being pinched. And the lap is sized by roof pitch: shallower pitches need longer laps, since water travels more slowly and wind can drive it further up the joint.
The analogy
THE ANALOGY #Think of a long steel railway rail in the days before continuous welding, laid in lengths with a visible gap at each joint.
The gap is a defect from the passenger's point of view — it is what makes the noise, and it is a weak point that needs maintaining. It is there because a rail fixed rigidly along miles of track would buckle in summer heat, and the resulting derailment is a considerably worse outcome than a noisy joint. When continuous welded rail did arrive, it did not remove the expansion problem; it managed it differently, by stressing the rail deliberately and holding it with far heavier restraint than any joint bar could provide.
A rail's expansion is resisted by a very strong material that would buckle elastically, whereas lead is soft enough to creep and take a permanent set — so lead does not fail suddenly by buckling, it fails gradually by being worked, which is why its symptom is a split appearing years later rather than a distortion appearing on a hot afternoon.
Clarifying the model
THE MODEL #The failure is fatigue, not corrosion, and this is the most common misdiagnosis. Lead's durability against weather is exceptional — it forms a stable surface layer and lasts a very long time. So when a lead detail fails, the material is rarely the problem, and replacing like with like in the same over-long lengths reproduces the failure on the same schedule. The give-away is where the split appears: thermal fatigue splits at restraint points and shows the characteristic thickening and thinning, whereas a corrosion failure would not be so precisely located.
Thickness is not a substitute for shorter lengths. It is tempting to answer a fatigue failure with heavier lead. Thicker material is more durable in other respects, but the mechanism here is driven by length and restraint, and a thicker piece that is still too long still accumulates the same movement. Heavier lead in over-long runs is a common and expensive way to get the same failure a few years later.
Sealants make it worse, and the reasoning explains why. The instinct on seeing a lap is to run a bead of mastic into it. That converts a free-moving lap into a restrained one — reintroducing exactly the fixity the detail exists to avoid — and it fails anyway, because the sealant is now the thing taking the cyclic movement. A lap that is correctly sized does not need sealing, and one that needs sealing is the wrong size.
The falsification test. If the mechanism is accumulated thermal cycling restrained over length, then splits should appear preferentially in the longest continuous pieces, at the most restrained point, on the most sun-exposed elevations — and pieces of the same lead in short free-moving lengths on the same building should be sound. If failures were distributed evenly regardless of piece length, restraint and exposure, the account would be wrong and something about the material or the fixing method would have to explain it.
A picture of it
THE PICTURE #How to readRead the boxes as conditions of one piece of lead through a daily temperature cycle. The two exits from Tensioned are the whole design decision: a short piece free to move returns to where it began, closing the loop harmlessly, while a long restrained one carries a small permanent change into the next cycle. The loop through Worked is what accumulates — note that it never returns to Flat, which is why the failure arrives after years of cycles rather than after any single hot day.
What became clearer
WHAT CLEARED #The joints are not a compromise on weatherproofing. Lead in a long fixed run cannot expand, so it creeps, and because it creeps rather than springing back, every daily cycle leaves a small permanent change that adds to the last — until the material splits at whichever point was most restrained. Cutting the run into short pieces makes each day's movement small enough to be reversible, and lapping them sheds water by geometry rather than by a seal. The roofer is not accepting more joints for the sake of easier handling; they are choosing a detail that can move, because the alternative is one that must eventually tear.