Is the Eiffel Tower Taller in Summer?
Every August, some version of this fact does another lap of the internet: the Eiffel Tower is several inches taller in summer than in winter, because its iron frame expands in the heat. It is a satisfying claim because it sounds like it should be checkable — thermal expansion is ordinary physics, not folklore. So we checked it, against the tower's own explanation of itself and against the arithmetic.
The claim
Heat makes the Eiffel Tower's iron structure expand, so the tower is measurably taller in the height of summer than in the depth of winter — by an amount commonly given as 12 to 15 centimetres (roughly 5 to 6 inches).
Source trail
The physics is not in dispute: puddled iron, like any metal, expands when it warms. The dispute, once you look, is entirely about the size of the number. Two lineages of that number circulate:
- The tower's own operator, SETE, publishes a short explainer on toureiffel.paris, dated 3 February 2021, that describes the seasonal change in structural height as "a few millimetres" — and separately describes an unrelated daily effect: the sun heating one face of the tower more than the others causes the tip to trace a circular path "measuring approximately 15 centimetres in diameter" over the course of a clear day. That is a sideways wobble from uneven heating, not a height gain.
- A widely shared claim of "12 to 15 centimetres" of actual height gain circulates through science-explainer outlets. The most recent well-sourced version we could trace is a 6 August 2025 article on The Conversation, written by Federico de Isidro Gordejuela, an architecture professor at Universidad CEU San Pablo, syndicated onward to Phys.org, Yahoo News, and others. The article states the figure as something "specialists have estimated" — it does not name the specialists, cite a study, or show a measured before/after reading. It is a calculation, not a measurement, and the article does not claim otherwise.
Both lineages trace back cleanly to a named, dated source. Neither traces back to an actual measurement of the tower before and after a heatwave.
What's confirmed
- The tower's iron structure does expand with heat — this is not in question. Puddled iron's coefficient of linear thermal expansion is given as 12 × 10⁻⁶ per °C by The Conversation and independently as 12 × 10⁻⁶ per °C for iron (13 × 10⁻⁶ for steel) by Georgia State University's HyperPhysics reference tables (checked 2026-08-17). Two independent sources within 10% of each other on the physical constant is about as settled as a material property gets.
- The tower's own operator describes the effect as "a few millimetres," not centimetres — quoted directly above, checked 2026-08-17.
- A separate, genuinely 15 cm official figure exists — for a different phenomenon. The same official page states the tower's tip moves in a roughly 15 cm-diameter circle over a clear day, from one face heating and expanding more than the others. That number is real, official, and not about seasonal height.
- Running the arithmetic ourselves lands closer to the popular figure than to the official one. Using a 300 m iron structure (the tower's original name was "La tour de 300 mètres," per The Conversation) and the 12 × 10⁻⁶/°C coefficient: a 15 °C swing gives 300 × 12 × 10⁻⁶ × 15 = 0.054 m (5.4 cm). A 40 °C swing — plausible for sun-heated metal, which runs well above air temperature, versus a cold winter night — gives 300 × 12 × 10⁻⁶ × 40 = 0.144 m (14.4 cm), which is inside the commonly quoted 12–15 cm range. The fact-check site factmyth.com runs the same formula with a 324 m height and an 11 × 10⁻⁶ coefficient over a 15 °C swing and gets 5.34 cm — in the same neighbourhood as our low-end figure, for the same reason: a smaller assumed temperature swing produces a smaller answer, from the same physics.
What isn't confirmed
- No source we found — official or independent — cites an actual measured height before and after a heatwave. Every centimetre figure in circulation, ours included, is a calculation from an assumed temperature swing, not a reading. "A few millimetres" is likewise a description, not a cited measurement with instrumentation named.
- The size of the answer depends entirely on which temperature swing you assume, and no source states which one it used. Our own range above — 5.4 cm to 14.4 cm — came from picking 15 °C versus 40 °C. Neither The Conversation's "12–15 cm" nor the tower's "a few millimetres" states the temperature swing behind it, so neither can be checked against ours.
- A frequently repeated detail — that the largest recorded gain was 18 cm in 1976 — could not be traced to a primary source. It appears only in aggregator and syndicated articles with no citation. We are not publishing it as fact.
- We could not reconcile the official "few millimetres" with the physics. A temperature swing small enough to produce only "a few millimetres" — under a centimetre — would need to be under about 3 °C by the same formula, which is smaller than the difference between a cold Paris morning and a hot Paris afternoon, let alone winter to summer. We do not know whether the operator's phrase is a rounding-down for public communication, a description of something other than the full seasonal swing, or based on a different figure than either of our sources used. We have not found anything that resolves it, and we are not guessing at an explanation.
Verdict
Yes, the tower is measurably taller in summer than in winter — the physics is real and not contested by any source we checked. But no single centimetre figure is settled: what circulates as "12 to 15 centimetres" is a back-of-envelope calculation attributed to unnamed specialists, what the tower's own operator says is "a few millimetres," and the two do not agree with each other or, in the operator's case, cleanly with the underlying physics. The likeliest source of the inflated popular figure is a mix-up with a real but unrelated official number: the roughly 15 cm daily sideways wobble of the tower's tip as the sun heats one face more than the others. That is a genuinely confirmed 15 cm — just not the 15 cm the claim is about.