Common Roofing Slate Defects: A Field Guide to Cracks, Pyrite and Delamination

Part of the Roofing Slate Guide's Quality, Testing & Standards section — the "what's on the stone" companion to our inspection checklist, and the surface-level side of the carbonate story. If you've ever held a tile, squinted at a spot, and wondered is that a problem or just a rock being a rock — this is the page.

I've spent the better part of twenty years with a hand lens, a bottle of water, and an unhealthy amount of curiosity about what goes wrong with roofing slate. I've examined thousands of tiles — from a quarry in Jiujiang that ships to thirty countries, and from plenty of places that don't. And I've learned that the word "defect" gets thrown around a lot more than the thing itself deserves.

Most of what people call a roofing slate defect is one of four things: a natural mark that's perfectly harmless, a flaw that actually changes the stone's future, damage from the journey, or something that only shows up after a few winters. The whole skill is telling the four apart before you buy. That's what this article is for.

The three questions that decide everything

Before we get to the list, forget the list for a moment. Every defect on it — every crack, spot, line and band — answers three questions, and the answers decide what you do:

1. Is it in the stone or on the stone? A mark on the surface usually means transport, handling, or something that'll stay cosmetic. A mark in the stone means the slate was born with it, and born-defects travel deeper.
2. Does it follow the grain or cut across it? This is the most repeated rule in the trade: slate splits along its grain, so a line that runs with the grain is often just where the next split would go anyway — harmless. A line that runs across it is a line of weakness the roof will find eventually.
3. What does the ring test say? Hold a tile by a corner and tap it lightly with a steel trowel or hammer handle. A clean ring means the stone is sound. A dull thud means it's delaminated or hiding a crack. A row of dull thuds is a bad batch, not bad luck.

Keep those three in your head and the rest of this guide is just names for what you're seeing.

Defect #1: Cracks — the one that needs a direction

Here's the thing people don't expect: a crack is not always a defect. Natural roofing slate is split by hand, along its grain, and a hairline that runs in the same direction as that grain — along the length of the tile, parallel to the cleaving line — is often just the stone following its natural parting. It doesn't change the strength of the tile, and it doesn't make the roof leak. I've put thousands of slates like that on roofs and watched them outlast the house that followed.

The crack that is a defect goes the other way: across the grain, usually visible on the surface as a fine line you can feel with a fingernail, or running in from an edge. That's a crack that will grow — under the weight of the roof, the pull of the wind, and the freeze of a cold night. It's the kind of crack the ring test catches: the tile thuds instead of rings.

So the practical rule is simple: crack along the grain = watch it. Crack across the grain = don't buy it. If you're ever unsure, you've got a thumbnail and a tile — run the thumbnail along the line. If it catches, it's a real crack. If it doesn't, it's a grain mark, and a grain mark is part of what makes slate slate.

Defect #2: Delamination — the one that robs a century

This is the defect I want you to be most afraid of, and the reason is worth stating once: delamination is the difference between a slate that lasts a century and a slate that lasts a winter.

Roofing slate is a rock that was once mud, squeezed and baked under huge pressure until it gained a plane — the cleavage — along which it splits into sheets. Delamination is when the layers of a single tile start to separate from each other. You'll see it as a flaky, crumbly edge, or a corner where the stone lifts like the pages of a book. Tap it and it thuds, because the stone has stopped being one piece and become several thin ones pressed together.

Why does it happen? It's usually a sign the stone wasn't quite right in the first place — a patch of clay or a microscopic gap between the layers, often at the edges of a tile where the cleavage is less regular. It can also appear after the roof is up, when water gets into that edge and freeze-thaw cycles gradually push the layers apart over decades. Either way, the outcome is the same: that tile is done, and it's the kind of done that sneaks up on you.

Here's the field rule: delamination is a reject. Not a "watch it" — a reject. You can't fix it, you can't dry it, and it's the single most common reason a slate roof that looked perfect at year five needs a re-felt at year twelve. On inspection day, if a sample tile thuds and its edge flakes under your thumb, that's a crate you want to look at twice. And the ones that flake under your thumb — those are the ones that will flake under the sky.

Defect #3: Pyrite spots — the sparkle that's a catch

Now the one everyone notices: those tiny golden or brassy specks on a slate's surface, catching the light like glitter you didn't ask for. That's pyrite — iron sulphide — and it's so common in slate that a quarry that has none of it is either very lucky or very well sorted.

Here's what pyrite does: it oxidises when it's exposed to air and water, turning into a rust-coloured mineral that swells as it forms. On a roof, that swelling slowly pushes tiny pits out of the slate face — pinhead craters you notice after a few winters. A couple of dots per tile is normal, and a responsible slate will carry them without drama. A face covered in them is a slate that's telling you it's a poor short. And when the pyrite sits on the cleavage plane rather than the face, it can force the layers apart from the inside — a delamination that starts without a single visible crack.

The honest numbers from our own stone, because I'd rather give you a real one than a safe one: our Lushan slate runs about 0.2% water absorption, and the pyrite in it is fine-grained and widely scattered — the sort the roof never feels. That's the chemistry you want: not "no pyrite", but "so little and so stable it never becomes anything." If you're looking at a slate with a face full of sparkle, ask for the pyrite content on the test report — a serious supplier will have the number.

Defect #4: White marks and ribbons — the one that looks pretty

If there's one defect that's trickiest to call, it's this one — a pale, often lighter streak or a soft band running through the slate, sometimes looking almost decorative. And decorative it is not. What you're looking at is a carbonate ribbon — a vein of calcium carbonate that got folded into the rock. The carbonate dissolves in slightly acidic rainwater, so the band erodes faster than the slate around it. On a roof that's been up a decade or two, you'll see it as a groove, and eventually a line where the tile has thinned enough to lose its edge. The detail is in the carbonate explainer if you want the chemistry.

The important distinction — and the one I'm asked about every year — is carbonate ribbons vs. ordinary light marks. Real roofing slate carries all sorts of pale streaks that are harmless: a light grain of clay, a patch of pale iron, a bit of white where the surface has a natural cleft. The rule of thumb: a band that runs clearly through the whole tile, with soft edges, is the one to worry about. A patch or a speck is usually fine. The carbonate band also feels different under the thumbnail — slightly soft, chalky, easy to scratch.

And the quick test: wet the band with a drop of water and run your thumb over it. If it's carbonate, you'll feel a slipperiness — like soap. If it's just a pale grain, it feels like the rest of the stone. It's not a lab test, but it's a pretty good one.

Defect #5: The edge chips, and the one that's really the installer

You've read the checklist story about edge chips: a few chips on a delivery are normal, the container does its. A corner knocked off in transit is a logistics accident, not a stone defect. But when you find chipped edges on a lot of tiles, or a chip in the same spot on every tile, the story changes — it's either the handling at the dock or the packaging, and both are your problem before the roof is.

As for surface chipping, the kind that shows up on the roof itself — it's most often a loading-and-unloading wound, and it's not the same as delamination. A chipped edge is a broken edge; a delaminated edge is a split. The first is annoying and usually stays put. The second is a slow leak of the tile's whole future.

The defect that shows up after a winter: weathering, not factory

There's a breed of "defect" that isn't in the stone at all — and I want you to recognise it, because it's the most expensive one of the lot. It's the mark that shows up after a winter or two, and it's not pyrite, not delamination, and not a crack. It's the frost finding what the eye missed.

Here's the honest sequence: a slate that's naturally fine at the quarry gets wet, gets shipped in a container across an ocean, gets stacked in a yard, and gets wet again. If a tile has any hidden weakness — a hairline, a soft band, a tiny delamination at the edge — that water finds it. Then the first hard frost comes and the water inside the hairline freezes, expands, and turns the hairline into a crack. You've seen the result on a few roofs: a line of tiles on the north side that look a little more cracked than the ones on the south, a few years in. That's not the "crack defect" from the quarry. That's weathering doing its slow, patient job on a stone that wasn't quite up to it.

This is why the tests matter: water absorption (a low one, like our 0.2%) and freeze-thaw cycles (a lot of them, with no weight loss) are exactly the numbers that tell you a slate won't be the one doing this. A slate with a low absorption and a clean freeze-thaw report is a slate that's quietly boring — and boring is exactly what you want your roof to be.

The defect family you should learn to live with (the harmless)

Before you panic at every spot, let me list the ones that are not defects:

• The natural cleft surface. The split face, with its slight irregularity and faint ridging, is the slate's normal face — not a flaw. It's what gives natural slate its depth.
• Light grain and pale specks. Iron oxide, fine pyrite dust, a whisper of white bedding — all normal in a natural product, all harmless in the quantities that matter.
• A straight hairline along the grain. As we said: the stone's own parting line, and usually a nothing.
• Variation in tone from tile to tile. That's the charm. A "patchy" roof is a roof that's alive. If you want uniformity, you're buying a factory tile, not a slate.

The word to remember is "proportion". A slate with a single pale mark or a single pyrite dot is not defective — it's natural. A slate with a face full of sparkle, a ribbon, and a delaminating corner has told you its whole story at a glance. The difference is what the benchman sees when he sorts, and it's what separates the honest from the floor.

How to write "OK" and "REJECT" on a stack of slate

Now the wall-chart, the part of this page I'd like you to print and put where the tiles get sorted:

You see…	You call it…	What you do…
A hairline running along the grain, no nail catch.	Grain, not defect.	Use it. It doesn't change the roof.
A hairline running across the grain, or a crack you can feel.	Real crack.	Reject it. It'll grow.
An edge that flakes, a corner that splits into layers.	Delamination.	Reject it. The tile is done. Don't buy it in.
One or two tiny gold sparkles on a face.	Pyrite, and it's fine.	Use it. Normal in natural slate.
A face full of sparkle, or rust streaks forming.	Pyrite overload.	Question it. Ask for the pyrite/iron content on the report.
A pale band across the whole tile, soft and chalky.	Carbonate ribbon.	Reject it. It'll dissolve ahead of the roof.
Light speckles and pale grain, scattered.	Natural mark.	Use it. It's the character.
The tile thuds when tapped.	Hidden crack or delamination.	Look at it twice. The sound knows what your eyes missed.

And the two habits that make the difference: ring every tile — it takes two seconds and catches what the eye can't — and wet the top side — a wet tile shows its pale bands, its soft edges and its pyrite far better than a dry one. The roof is going to be wet more often than it's dry. So inspect it the way it's going to live.

The defects that are actually a supply-chain conversation

One last thing, and it's the one buyers forget: how many defects you see is a factory conversation, not a stone conversation. A good supplier doesn't just have a quarry — it has a sorting bench, a delamination cage, a pyrite magnet, a standard of what it calls a tile and what it calls a reject. When a crate of roof slate shows up at your door with too many flakes and a couple of ribbons, the question isn't "is this slate bad" — it's "was this batch packed by a supplier who sorts, or one who ships everything?"

The sorting is the difference between a slate roof that lasts a century and one that leaks at fifteen. It's the difference between the test results we publish and the ones we don't. And it's the reason you ask, every time, "what's your reject rate at the bench?" A supplier that knows that number, and is happy to say it, is a supplier that understands that the defects aren't in the stone — they're in the sorting.

Frequently asked questions

Related Reading

• Roofing Slate Inspection Checklist — the field drill that uses this field guide, minute by minute, when a delivery arrives.
• How to Evaluate Roofing Slate Quality — the six-station factory walk-through that sorts the defective before it ever gets a crate.
• Carbonate Content and Weathering Risk — the chemistry behind the white bands, and why the test numbers matter.
• Roofing Slate Water Absorption Explained — the number that decides how fast any hidden defect turns into a leak.
• Roofing Slate Terminology — cleavage, grain, cleft, ribbon — the words in this guide, defined for the first read.