Is glass really a liquid that flows very slowly?

Walk into an old cathedral, squint at the stained glass, and someone in your group will almost certainly lean over and whisper that glass is secretly a liquid. The proof, they say, is right there: the old panes are thicker at the bottom, because the glass has been oozing down the frame for hundreds of years. It is a lovely story. It is also wrong by a margin so large it is almost funny.

The short answer

Glass does not flow at room temperature on any timescale that matters to humans, cathedrals or, frankly, civilisations. In a study published in January 2018 in the Journal of the American Ceramic Society, a team led by John Mauro of Penn State, working with researchers from Corning, modelled how medieval glass, including glass of the type found at Westminster Abbey, would creep under its own weight. Their verdict: it would take billions of years to produce changes measured in nanometres. That is not a visible sag. That is barely a rumour of one.

The real story

Part of the confusion is that glass really is strange stuff. Most solids, like metals or salt, have atoms locked into a tidy, repeating crystal pattern. Glass does not. Its atoms are jumbled in a disordered arrangement that looks, at the atomic level, a lot like a frozen snapshot of a liquid. Mauro has described glass as having a hybrid liquid-solid nature: the structure resembles a liquid, but mechanically it behaves as a solid. Scientists usually call it an amorphous solid.

So technically, glass does have some ability to flow. The question is how much. Mauro’s team built a new mathematical model of how glass flows, now known as the MAP equation, which accounts for how a glass’s chemical recipe affects its stickiness, or viscosity. Interestingly, their calculation showed glass flows faster than older estimates had suggested, by a staggering 16 orders of magnitude. Even so, faster still means hopelessly slow. No medieval window has been around anywhere near long enough to droop.

Other researchers have come at the question from a different angle. A team studying 20-million-year-old amber, which is a natural glassy material, used it as an ultra-stable test case for how glassy substances behave over geological time. The conclusion was the same: no meaningful flow at everyday temperatures.

Why people get it wrong

The myth survives because the evidence people point to is real. Many old panes genuinely are uneven. The reason is how they were made. Medieval glassmakers took a blob of molten glass, then spun and flattened it into a large disc before cutting it into panes. That process left panes thicker in some places than others, and glaziers tended to install the heavier, thicker edge at the bottom, where it sat more securely in the frame.

So the lumpy windows are not a slow-motion waterfall. They are a fingerprint of medieval craftsmanship. Add the fact that glass really does look like a liquid at the atomic scale, and you have the perfect recipe for a myth that sounds scientific, feels clever, and gets repeated on every cathedral tour.

“I spent years under desks believing that one. Turns out the windows weren’t melting, mate. It was just the glazier being sensible.”