The Pretty Physics Lie About Slippery Ice

A 70-kilogram speed skater balanced on a single blade generates enough pressure to depress ice's melting point by roughly three and a half degrees Celsius. Olympic speed-skating rinks are chilled to about -7°C [S5]. Hockey rinks run colder still [S5]. The schoolbook answer — that the skater's own weight liquefies the ice beneath her — has never closed the arithmetic.

That schoolbook story was not James Thomson's. Thomson, Lord Kelvin's older brother, published the underlying thermodynamics in January 1849: pressure lowers water's freezing point, end of paper [S7]. The skating application came in 1886, from Irish engineer John Joly, who pinned the glide on that same pressure effect [S1]. Thomson himself never mentioned skates [S1]. Joly did, and textbook writers loved him for it.

Bowden and Hughes finally did the arithmetic in 1939 — 53 years after Joly — in an ice cave dug above the Jungfraujoch research station, 3,346 metres up in the Swiss Alps, where cave temperatures never rose above -3°C and the researchers used solid CO2 and liquid air to chase lower ones [S2]. The pressure under a ski could not melt ice at the temperatures where people demonstrably skied [S2]. They proposed friction heating instead: metal skis dragged worse than wooden ones, which they read as evidence that heat conduction, not pressure, ran the show [S2].

Friction melting also fails at low temperatures, and the Bowden-Hughes proposal has been quietly contested ever since [S1][S4]. Two wrong answers in a row, both published in journals of record, both repeated for decades.

The right answer is older than either. On 7 June 1850, Michael Faraday gave a Royal Institution discourse proposing that ice carries a thin liquid film on its surface even below 0°C, no pressure required [S1]. He had noticed that two ice cubes pressed together fuse, and reasoned that their surfaces must already be partly liquid [S1]. Thomson disagreed publicly, insisting the fusing was just pressure melting followed by refreezing — and produced no new experiment to support the rebuttal [S1]. The textbooks sided with Thomson. Faraday was overruled in his own lifetime and largely ignored for roughly 140 years [S1].

In the 1990s and 2010s, two techniques caught up to him. Heterodyne-detected sum-frequency generation spectroscopy — a method tuned to the vibrational signature of the topmost monolayer of water molecules — picked up something behaving like supercooled liquid water on the ice surface down to 245 K, about -28°C [S6]. Molecular-dynamics work and complementary experiments push the disordered topmost bilayer down past 170 K, about -103°C, and the layer persists in vacuum [S6]. Whatever it is, no plausible weight or friction is melting it at those temperatures.

It is not melted bulk ice. It is what physicists now call the quasi-liquid layer: a few molecular layers thick, where water molecules vibrate and rotate freely while staying loosely bonded to the crystal beneath [S1][S6]. A surface-only phase of matter. And it doesn't grow smoothly as ice warms — a 2017 PNAS paper using SFG plus molecular dynamics found the layer thickens bilayer by bilayer, with a sharp jump around -16°C as the second bilayer breaks loose [S3].

Once the quasi-liquid layer is in the picture, sports stop being mysterious. The coefficient of friction on ice bottoms out around -7°C and climbs on both sides [S5]. Speed-skating rinks chase that minimum [S5]. Hockey rinks run colder, for harder and faster ice; figure-skating ice is kept warmer for grippier edges [S5]. Curling — where the stone actually has to grip and curl — is engineered around roughly -4.5°C [S5]. To restore slipperiness at that warmer temperature, ice-makers spray the surface with droplets of warmer water that freeze into rounded bumps called pebbles; the stone rides on the tips [S5].

As of December 2025 the question is not settled. A group at the Universität des Saarlandes proposed a fourth mechanism this year: that "sticky" contact between slider and surface electrically pulls water molecules out of the lattice, disrupting the crystal independent of pressure, friction, or thermal premelting [S4]. Quanta, surveying the four live hypotheses 139 years after Joly first blamed pressure for skating, reports the community has not yet converged on which dominates [S4].

Faraday's clinching observation — two ice cubes squeezed in a warm hand fuse — is also how a snowball holds together. The quasi-liquid layer on each flake makes contact and refreezes solid the instant the pressure releases [S1]. He inferred a surface phase of matter in 1850 from a children's trick. A high-schooler with a calculator could have killed Joly's pressure story in 1887. Three generations of physics students learned it anyway, taught by people who had also learned it, in a chain of unverified pretty stories that ran for a century and a half.