There is a particular disappointment that comes with opening the kiln, lifting out a bowl you are pleased with, and then noticing a week later a fine web of cracks across the inside. Nothing was mixed badly. Nothing was fired to the wrong cone. The glaze and the clay body simply disagree about how much to shrink as they cool, and that disagreement has to go somewhere. Potters call it glaze fit, and it is one of the few problems in ceramics you can reason about before you ever light the kiln.

What follows is the practical version: what fit means, why it fails in two opposite ways, and how to test a combination so you know where you stand.

What glaze fit actually means in the kiln

A glaze is a thin sheet of glass welded to a clay body. While the kiln is hot, both materials are soft enough to move. At some point on the way down — often somewhere between 500 °C and 600 °C, though it depends entirely on the recipe — the glaze stiffens into a rigid solid. From that moment the two materials are glued together and contracting at their own rates. The clay body keeps shrinking as it cools. The glaze, now solid, either goes along with it or gets pulled apart trying.

The comfortable situation is a glaze held in very slight compression by the body beneath it. Glass is strong under compression and weak under tension, so a glaze that is being gently squeezed is tougher than one that is being stretched — harder, more chip-resistant, more durable in the dishwasher. Every good clay-and-glaze pairing you have admired is, at least in part, a well-managed squeeze.

Too little compression or too much tension, and the glaze cracks. Too much compression, and it buckles off the pot. There is no neutral setting, no glaze that works equally well on every body, which is why a recipe that performs perfectly in one studio can fail in the next.

Thermal expansion, in plain language

Every material has a coefficient of thermal expansion: a number describing how much it grows when heated and shrinks when cooled. You will see it written as something like 6.5 × 10⁻⁶ per °C. The exact figure is less important than the comparison between two materials. A glaze that contracts more than its body ends up in tension. A glaze that contracts less ends up in compression.

In glazes, the chemistry gives you levers. Silica and boron are the great shrinkers — more of them means less expansion. The alkalis, sodium and potassium, are the opposite; they raise expansion sharply, and they are present in every soda feldspar and nepheline syenite you scoop. Calcium, magnesium, barium, zinc and strontium sit between the extremes. Alumina does not shift expansion dramatically, but it makes a harder, more durable glass that tolerates stress better, which is why small additions of clay or alumina hydrate show up in so many crazing fixes.

Clay bodies play by related rules. Bodies with a lot of free quartz tend to expand more and are the usual suspects when a glaze crazes on one clay and behaves on another. Talc is the traditional body-side remedy: adding a little raises the body's expansion, which lets a slightly tensioned glaze relax.

Two failures, two opposite fixes

The two fit failures look nothing alike, and confusing them is the most common way to make things worse.

• Crazing is a network of fine cracks across the glaze surface, like the crackle on an antique vase. The glaze is in tension because it contracts more than the body. Crazes can be decorative when deliberate, but on functional ware they are a flaw.
• Shivering is the opposite fault: small flakes lifting and detaching, most often at rims, edges and sharp corners. The glaze is in too much compression and has nowhere to go. It is less common than crazing and considerably more dangerous — a chip of glass in food or on a fingertip is a real hazard.

The directions of repair are therefore mirrored. Fighting crazing means lowering the glaze's expansion or raising the body's: add silica, trim the alkaline flux, add a little alumina, or introduce talc into the body. Fighting shivering means raising the glaze's expansion: more feldspar or nepheline syenite, less silica, or a body with lower expansion. Apply a crazing fix to a shivering glaze and you will make the flaking worse.

Why the cracks often appear later

Fit problems do not always announce themselves at the kiln door. Three things delay the verdict.

The first is cooling rate. Cool a kiln quickly and you freeze stress into the glaze that slower cooling would have allowed it to shed. If your work is crazing only on the top shelf, or only in the pieces you unload early, question the cooling before you reformulate.

The second is glaze thickness. A heavy layer of glaze carries more accumulated strain than a thin one, in either direction. A recipe that survives at a normal dip can craze or flake where glaze has pooled inside a foot ring.

The third is moisture expansion. Fired clay is not inert; over months it slowly takes up moisture from the air and from use, and it grows very slightly in the process. That growth keeps tightening the tension on a glaze that was already marginal. This is why a mug can leave the studio flawless and craze after a summer of daily coffee.

How to test a combination properly

1. Make tiles from the actual body. Not a convenient porcelain from the shelf — the clay you throw with, wedged and rolled to an even thickness, with a hole punched for hanging.
2. Apply the glaze at the thickness you really use. Dip half the tile at your normal timing and leave the other half thinner. Note which is which.
3. Fire exactly as you fire your work. Same cone, same schedule, same cooling. Then fire a second tile through the schedule you would use for a glaze test, because the two may not agree.
4. Wait, then look. Give the tiles a week before judging, and check again at a month.
5. Run an ink test. Wipe a felt-tip pen or strong tea over the surface, leave it a minute, then wipe it off. Crazes hold the colour as fine dark lines; an intact glaze comes clean.
6. Abuse one tile. Pour boiling water over it, then plunge it into iced water. Repeat a few times. Repeat dishwasher cycles will do similar work over a longer period. Crazing that needs stress to appear is crazing you would have met later on a customer's shelf.
7. Check rims for shivering with a fingernail along the edge and a close look under good light. Any lifting flake means the pairing is wrong, not that it needs more glaze.

Glaze a test tile on the body you actually use, fire it the way you actually fire, and then wait a month before you commit to a batch.

Which side should you change?

Working on the glaze

For crazing, the gentlest first move is usually a small addition of silica — a few percent — plus a touch of alumina, then retest. If the glaze is high in sodium or potassium, reducing the feldspar and replacing some of it with a calcium or magnesium source will bring expansion down further. Boron is a useful lever at low-fire and mid-fire, since it lowers expansion while keeping the melt fluid.

For shivering, do the reverse: increase the alkaline flux a little, or cut back the silica. Change one thing at a time and keep notes, because two small adjustments in the same direction can overshoot from one fault straight into the other.

Working on the body

If a glaze you love is marginal on your clay, the body is often the easier place to move. Adding talc raises expansion and can rescue a crazing glaze without touching the recipe. Removing some of the coarsest quartz, or choosing a body with less free silica, reduces the strain. These changes alter working properties — talc makes a body soapier and can soften it at temperature — so test the clay on the wheel before you commit to a hundred kilos.

A studio routine that saves years of guessing

Keep a fit file. When you settle on a body and a glaze that work together, write down the tile date, the firing, the thickness and the result, and keep the fired tile in the drawer with the notes. When you buy a new clay, or a supplier changes a recipe, you have a benchmark to compare against rather than a hunch.

Never introduce a new glaze into production on the strength of one successful firing. Fire a tile, abuse it, wait, and check the rims. And remember the useful asymmetry at the heart of all of this: a little compression is your friend, and a crazed glaze on a food surface is not a small cosmetic issue — it weakens the pot and gives moisture a way in. If a piece is going to hold someone's dinner, give it the month.

Photo: Mathias Reding / Pexels