In September 1792, a French astronomer was marched out of a country house by armed militia and taken to the town hall to explain why he had spent the morning pointing a brass instrument at church steeples. Jean-Baptiste Delambre had a good answer. He was measuring the world. With the palace stormed and the king jailed a few weeks earlier, it went down about as well as you would expect.

He lost a day to it. Set against what came next, a day was nothing.

What a metre was supposed to be

Revolutionary France ran on a shambles of local units. A pound of one thing weighed differently from a pound of another, and a measure of grain in one village meant something else two valleys over. Excellent for sellers, miserable for buyers.

The Commission of Weights and Measures, Lavoisier and Laplace among its members, proposed a fix in March 1791 that nobody could accuse of local bias. One metre would be one ten-millionth of the distance from the North Pole to the equator, a length borrowed from the planet itself, as the MacTutor History of Mathematics archive records.

Somebody still had to go out and measure it.

The plan was to survey the stretch of the Paris meridian between Dunkirk and Barcelona, then scale up the full pole-to-equator distance. Delambre took the northern leg as far as Rodez. Pierre Méchain, a comet hunter with a reputation as the most careful observer in France, took the southern leg through the Pyrenees. Both worked by triangulation, chaining sightlines between hilltops and belfries with a Borda repeating circle, an instrument built to average away error.

Surveying gear in a country coming apart

Méchain barely made it out of Paris. At Essonnes, his crates of brass and glass struck the locals as the kit of a counterrevolutionary. He was held for two months before anyone in authority worked out where he was, as Owen Gingerich recounts in the Complete Dictionary of Scientific Biography.

Spain went better, briefly. He ran the triangulation up to Barcelona through September and October, then spent the winter observing stars from the castle on Mont-Jouy hill to fix the city’s latitude. In the spring a friend showed him a new hydraulic pump. The friend and an assistant got caught in the mechanism, Méchain went in after them, and a lever threw him into a wall, breaking ribs and a collarbone. He was unconscious for three days and could not move for weeks. By June his right arm still would not work, so he made the solstice observations left-handed.

Back north, the politics caught up with Delambre. In December 1793 the Committee of Public Safety purged the commission of everyone not judged worthy of confidence through republican virtue and hatred of kings, taking out Borda, Laplace, Lavoisier and Delambre himself, as documented in a review essay for the journal Isis. Lavoisier went to the guillotine the following May. Work stopped for fifteen months, because the men doing it had been declared politically unreliable.

What happened at Mont-Jouy

Which reading do you trust when you cannot go back and check?

War with Spain had handed the castle to the military, so Méchain measured Barcelona’s latitude a second time from the inn he was lodging in. By March 1794 the two sets of figures disagreed by about three seconds of arc, roughly the width of a coin seen from a kilometre away.

He told nobody.

Historian Ken Alder reconstructed what followed from sealed observatory papers for his book on the expedition. Méchain refused to come to Paris, refused to show anyone his data, achieved almost nothing for two years, and began quietly adjusting his readings so they sat tighter around their average without moving it. Colleagues eventually sent his wife to fetch him. They met at Rodez in 1798, their first meeting in six years, and she could not shift him either.

The cruel joke is that the discrepancy was never really his fault. Nobody in the 1790s had a theory of measurement error, so Méchain assumed enough care could deliver any accuracy he wanted. A systematic bias sits inside every reading, and repetition never averages it out. He spent a decade being tortured by physics.

The error that got locked in

“Hell and all the plagues it spews upon the earth,” Méchain wrote to Delambre from Spain in 1804, still chasing the latitude that had beaten him. He died of fever on the Spanish coast that September, the question still open.

The triangulations themselves had finished in November 1798. The figures went to an international conference of scientists in Paris, and the answer was cast as a platinum bar. The resulting metre was slightly short. It came out about 0.2 millimetres under the length the definition called for, a shortfall the US National Institute of Standards and Technology traces to errors in calculating the Earth’s curvature. Across a quarter of the planet that becomes a 2 kilometre miss, as a 2019 paper in the Journal of Geodesy sets out, blaming an uneven gravity field that tugged the surveyors’ plumb lines fractionally off true.

Delambre published the survey, treated the awkward numbers as findings rather than failures, and sealed the private letters in the observatory archive.

The length never got corrected. A tougher platinum-iridium bar took over as the international prototype in 1889, and krypton wavelengths replaced the bar in 1960. In 1983 the metre became the distance light travels in a vacuum in one 299,792,458th of a second, a definition the International Bureau of Weights and Measures still treats as continuous with the last. Every step was chosen to sit as close to that first platinum bar as possible, shortcomings included, as NIST notes.

Pole to equator runs roughly 2 kilometres longer than ten million metres, and always did. Instead of fixing that, the world cut the metre loose from the Earth it was copied from and kept the figure two exhausted astronomers brought back.