03. The Metre — The Hidden Error in the Standard

A unit taken from the planet

In 1792 the new French Republic sent two astronomers to measure the meridian arc from Dunkirk to Barcelona, so that a unit of length could be defined from the Earth itself. The metre was to be one ten-millionth of the distance from the north pole to the equator, a definition owned by nobody and therefore acceptable to everybody. Jean-Baptiste Delambre took the northern half and Pierre Méchain the southern, and the survey was expected to take about a year. It took seven, through a revolution, a war, and Méchain's imprisonment, and both men triangulated across the country one sighting at a time. The point of the exercise was political as much as scientific: a standard derived from nature could not be the property of a king. The metre that resulted is about 0.2 millimetres short of the quarter meridian it was meant to represent.

The error Méchain could not report

Méchain measured the latitude of Barcelona from two different stations and the two results did not agree. The discrepancy was small, well under the width of a fingernail at that distance, and it was larger than his method was supposed to permit. Ken Alder's archival account describes Méchain concealing the inconsistency, reworking his data, and being destroyed by it over the following years; he died in 1804 having never resolved it. The irony is that the discrepancy was not his fault at all — the Earth is not a perfect ellipsoid, and local variations in gravity tilt a plumb line by small amounts no eighteenth-century observer could have modelled. The survey also assumed a value for the Earth's flattening, near 1/334, that is some way from the modern figure of roughly 1/298. The standard the world adopted was therefore built on a real error and an honest concealment, and it worked anyway.

Why it did not matter, and then did

A unit does not have to be true to be useful; it has to be agreed and stable, and the metre was both. The definition was cut loose from the planet in stages, until in 1983 the metre was defined by fixing the speed of light, which is where it still sits. That move ended the possibility of this failure mode entirely: a unit defined by a constant of nature cannot contain a surveyor's mistake, because there is no survey.

Example: A shared clock that runs four minutes slow keeps a whole town punctual, right up to the moment someone checks it against another town's.

Verdict. The metre was founded on a measurable error its own author hid, and remained fit for purpose because a standard needs agreement and stability rather than truth.

Overturned if: the archival record shows Méchain reported the Barcelona discrepancy openly to the commission, which would remove the concealment on which the standard account rests.

Sources. 2002 · Ken Alder, "The Measure of All Things: The Seven-Year Odyssey and Hidden Error That Transformed the World" — American Scientist review · undated · NIST, "Meter" — NIST