The window that seems to sag
Look closely at some old stained-glass windows and a few panes are thicker at the bottom. It is an irresistible piece of evidence for a familiar story: glass is actually a liquid, and over the centuries it has flowed under gravity. The story is memorable, but the conclusion is wrong. A pane of ordinary glass behaves as a solid, and the observed unevenness is not a record of it running downhill.¹
The confusion begins with a real and fascinating difference between glass and many other solids. In table salt or quartz, atoms repeat in an ordered three-dimensional pattern. In glass, the arrangement is disordered. It resembles the jumbled structure of a liquid more than the regularity of a crystal, which is why scientists call it amorphous. Yet the material holds its shape and resists shear, the practical hallmarks of a solid.²
Calling it an amorphous solid does not hide a secret liquid inside it. Structure and behaviour answer different questions. A crystal and a pane can both be solid while their atoms are arranged differently. The lack of a neat lattice is a property of glass, not proof that it is currently flowing.
From a melt to a rigid pane
Glass does begin as a flowing melt. Heat the ingredients sufficiently and their atoms move past one another. As the melt cools, that movement becomes slower and slower. Its viscosity, a measure of resistance to flow, rises steeply. Rather than freezing into a regular crystal at one sharp temperature, the liquid enters a glass-transition range in which its structure can no longer rearrange readily on the timescale of cooling.²,³
This is useful to glassmakers. At one temperature the material can be poured or formed; at another it can be stretched, blown or shaped; after further cooling it is rigid enough to become a bottle or window. The transition is not a magical instant at which every atomic motion stops. Atoms can still move or rearrange very slowly, and a piece of glass retains some memory of how it was cooled. But a cold window is not behaving like a pourable liquid.³
The word viscosity helps explain the misleading slogan. Honey flows more slowly than water, and a glass-forming melt becomes dramatically more viscous as it cools. It is tempting to extend that thought indefinitely and imagine room-temperature glass as honey in extreme slow motion. For practical purposes, however, its resistance to deformation is so enormous that gravity cannot visibly thicken the bottom of a window within the life of a building. Corning, the glass manufacturer, notes that calculations make such a change fantastically slower than the age of Earth.¹
Why old panes really are uneven
Before modern flat-glass production, makers could not simply order sheets of perfectly uniform thickness. Historic methods formed glass into cylinders or discs and then flattened or cut them. The resulting pieces often varied across their surface. A glazier choosing how to set an irregular pane might place its heavier edge at the bottom, where it seemed more stable. The thickness is therefore a clue to manufacture and installation, not a time-lapse record of flow.⁴
There is another reason the myth survives: it pairs a visible fact with a striking explanation. The thick edge is real. The explanation sounds scientific because hot glass undeniably flows and room-temperature glass is structurally unlike a crystal. But those facts do not establish the rate of movement needed to explain a cathedral window. Once that rate is considered, the story falls apart.
Does glass change at all as it ages?
Yes, but the changes are subtler. Glass is not necessarily in its lowest-energy arrangement after cooling. Over time, its structure can relax towards a different state, a process sometimes called physical ageing. Researchers study this because it can affect properties such as density and mechanical response, and because the rate depends on the glass composition and its thermal history.³
That is a more interesting account than either slogan, “glass is a liquid” or “glass is completely frozen”. In physics, describing a material often means specifying both the timescale and the kind of change being measured. A scientist can detect slow structural relaxation without seeing a window slump. Likewise, glass can soften and be reshaped when heated, while remaining a solid at ordinary temperatures.
The next time an old pane looks thicker at one edge, the best question is not “how far has it flowed?” but “how was it made?” Its unevenness preserves a trace of craft. Its continued place in the window shows just how effectively an untidy arrangement of atoms can make a solid.