measure for measure As fast as it gets Bart Verberck refects on measuring the speed of light, its role in metrology, and special relativity.

or many, the speed of light in vacuum 22 minutes. This insight led to a value of (c) probably brings to mind physics’ 220,000 km s–1 for c, based on a reasonable Fmost famous equation. For me, the estimate for Earth’s orbit2. power of c is that, when postulated to be the Proper terrestrial light-speed same for different observers regardless of the measurements were carried out around motion of the light source, the special theory the middle of the nineteenth century by of relativity follows — which, also requiring Hippolyte Fizeau. He let light reflect back the invariance of physics’ laws in non- to its source, with a rotating cogwheel put accelerating frames of reference, of course in the path of the double beam. When gives you E =​ mc2. gradually increasing the rotational speed For metrologists, c’s quintessential of the wheel, the situation becomes such connotation is probably that of defining that the incoming beam passes between the constant for the International System of notch between two cogs and the reflected Units (SI). Since 1983, the value of c has beam passes through the next notch. (In been fixed at 299,792,458 m s–1, which today’s beam-scientist lingo: Fizeau used an enables a definition of the metre relating optical chopper.) In 1849, he published a to it. The main advantage of defining units value of 315,000 km s–1 (ref. 3). through physical constants with assigned Léon Foucault (of pendulum fame) Credit: Bettmann/Getty Images values is that such definitions can forever developed a similar set-up. Instead of a remain unchanged — it’s the practical cogwheel, he put a rapidly rotating mirror realizations that can evolve and improve between the light source and the (now similar results, led to a recommendation of over time. Yet any such constant fixing needs spherically curved) reflecting mirror. the 15th General Conference on Weights to be preceded by high-precision, state- For large rotational speeds, a slit source and Measures in 1975 of the value that of-the-art measurements of the constant is imaged at a small angle away from the would become the definition in 1983. in question in the existing metrological source. The formula relating distance, angle Back to Albert Michelson, c measurer par framework. Only then can continuous and and rotational speed allows a calculation of excellence. In 1887, together with Edward backward consistency be guaranteed. the speed of light. In 1862, he published the Morley, Michelson produced what is arguably Until very recently, c was one of only value 299,796 km s–1 (ref. 4). the most famous null result of all time: a two defining constants for the SI. But last Both Fizeau’s and Foucault’s time-of- comparison of measurements of the speed month, during the 26th General Conference flight experimental set-ups were refined in of light in perpendicular directions revealed on Weights and Measures, fixed values for the following decades, leading to ever more no difference, doing away with the then- the Planck constant, the elementary charge, accurate measurements of c. Albert Michelson prevalent concept of the ‘aether’, a medium the Boltzmann constant and the Avogadro made several measurements of c between pervading the cosmos believed necessary for constant, and together with it the ‘new SI’, 1877 and his death in 1931, using adaptations transmitting forces. The Michelson–Morley were ratified1, to become effective on of the Foucault apparatus allowing for longer result proved a stepping stone for Einstein to 20 May 2019 — World Metrology Day. distances and larger source–image angles. In develop the theory of special relativity — and One of the first attempts to measure 1930, he started performing Foucault-type so physics’ most famous equation. ❐ light’s speed goes back to Galileo Galilei; experiments in a one-mile-long vacuum his experiment trying to establish whether chamber (pictured). His work was Bart Verberck there’s a delay between uncovering a lit continued posthumously by Fred Pease and Regional executive editor for Nature Research. lantern and observing the emitted light at Francis Pearson, who obtained the value e-mail: [email protected] a distance was inconclusive, however. 299,774 ±​ 11 km s–1 (ref. 5). Ole Rømer made an important step Alternative methods that provided Published online: 4 December 2018 forward. Around 1676, he compared the better accuracy, developed in the following https://doi.org/10.1038/s41567-018-0374-7 timings of eclipses of Io, one of the moons years, were based on cavity resonance References of Jupiter, at different times throughout the or interferometry. By 1972, using laser 1. NPL https://go.nature.com/2Bbf4EJ (2018). year. The observed variation of the time interferometers, a fractional uncertainty of 2. Phil. Trans. R. Soc. 12, 893–95 (1677). intervals between eclipses was due to the the order of 10–9 was achievable. Scientists 3. Fizeau, M. Comptes Rendus Acad. Sci. 29, 90–92 (1849). changing Earth–Jupiter distance, he argued. at the National Bureau of Standards in the 4. Foucault, M. L. Comptes Rendus Acad. Sci. 55, 501–503 (1862). 5. Michelson, A. A., Pease, F. G. & Pearson, F. Contrib. Mt Wilson Rømer estimated that to pass the Earth’s US achieved the result c =​ 299,792.4562 ±​ Observ. 522, 1–36 (1935). orbital diameter, light travels for about 0.0011 km s–1 (ref. 6). This, combined with 6. Evenson, K. M. et al. Phys. Rev. Lett. 29, 1346–1349 (1972).

1232 Nature Physics | VOL 14 | DECEMBER 2018 | 1232 | www.nature.com/naturephysics