Not all pounds are equal: a history of mass

A 125-million-dollar probe burned up in the atmosphere of Mars because one team was counting in pounds and the other in newtons. A Boeing 767 turned into a glider over Canada because someone multiplied litres by a coefficient given in the wrong unit. Behind both accidents lies the same surprisingly fresh wound: for a thousand years, "pound" did not mean one thing. And the kilogram — supposedly the perfect answer — spent a century as a metal cylinder in a vault near Paris, slowly losing weight.

Three pounds in one kingdom

The pound begins in Rome as the libra. That is where the abbreviation lb comes from, still used today, while the word "pound" (German Pfund, Polish funt) is a Germanic adaptation of libra pondo — "a pound by weight." After the empire fell, Western Europe splintered metrologically: every town, every guild, and every commodity had its own measure.

In England alone, three pounds settled in over the last millennium, each for a different purpose. All were counted in grains — a unit that survives to this day and equals exactly 64.79891 mg.

The avoirdupois (commercial) pound — 7,000 grains, divided into 16 ounces. Introduced by London merchants around 1303 for weighing bulk goods; in the fourteenth century, under Edward III, its role in the wool trade was formalized and the stone was fixed at 14 pounds.
The troy pound — 5,760 grains, divided into 12 ounces. The name is usually traced to Troyes in France, a great medieval fair town, though the etymology is disputed. In the form of the troy ounce it still governs precious metals.
The apothecaries' pound — the same total mass as the troy pound (5,760 grains), but an entirely different internal structure: the ounce split into 8 drachms, the drachm into 3 scruples, the scruple into 20 grains.

Hence the paradox that trips up everyone buying silver for the first time. A troy ounce (31.1035 g) is nearly 10% heavier than a commercial ounce (28.3495 g). But a troy pound (373.242 g) is almost 18% lighter than a commercial pound (453.592 g) — because it holds twelve ounces, not sixteen.

Order was imposed by decree. In 1527, Henry VIII abolished the Tower pound (5,400 grains) and the mercantile pound (6,750 grains), leaving the troy and avoirdupois pounds in circulation. Two instead of four — progress, but not a solution.

The Warsaw, New Polish, and Prussian pound

Outside the Anglosphere things were no better, and in Poland the pound could mean three different things depending on which partitioning power you lived under.

The Old Polish system rested on the Warsaw pound: 405.2 g, divided into 2 grzywnas or 32 łuts (a łut ≈ 12.66 g). In Congress Poland it was replaced by the New Polish pound, introduced by an 1818 decree and in force from 1819 — the reform was led by Stanisław Staszic. It was a clever hybrid compromise: the familiar names were kept (pound, łut, hundredweight), but each was anchored directly to the young French metric system. The New Polish pound was defined as 0.405504 kg, and the hundredweight (100 pounds) as 40.5504 kg.

In Prussia the pound weighed 467.711 g, having been derived from the mass of a Prussian cubic inch of distilled water. An 1856 law replaced it with the customs pound (Zollpfund) of a round 500 g, divided into 30 Loth; the new unit took effect on 1 July 1858 across the German Customs Union. That is why, in a German shop, "ein Pfund" still means half a kilogram.

Unit	Mass	Division	Where and when
Avoirdupois pound	453.59237 g (exact)	16 ounces / 7,000 grains	International standard since 1959
Troy pound	373.2417216 g	12 ounces / 5,760 grains	Precious metals, UK and USA
Warsaw pound	405.2 g	32 łuts / 2 grzywnas	Old Polish system
New Polish pound	405.504 g	32 łuts	Congress Poland, 1819–1849
Prussian pound	467.711 g	32 Loth	Prussia until 1858
Customs pound (Zollpfund)	500 g	30 Loth	German Customs Union, from 1858

When America stopped understanding Britain

After 1776, the US Congress was given the constitutional right to fix weights and measures — and declined to use it. The Americans kept the English standards they had inherited. Thomas Jefferson's 1790 report proposed a decimal system; the proposal died.

Britain, meanwhile, carried out its own reform in 1824–1826 and established the imperial system, redefining among other things the gallon (a single imperial gallon instead of separate ones for wine, ale, and corn). Two diverging systems now shared identical vocabulary:

Hundredweight: 100 pounds in the US, 112 pounds in Britain — where the 8 × 14 lb relationship with the stone was preserved.
Ton: the American "short" ton is 2,000 pounds (907.18474 kg), the British "long" ton is 2,240 pounds (1,016.047 kg).
Gallon: the American one descends from an old wine measure and is noticeably smaller than the imperial gallon, poisoning every liquid-density calculation.

What blocked full US metrication was money, not pride: the entire machine-tool industry stood on inch and pound dimensions, and re-tooling would have cost hundreds of millions of dollars. The 1975 act signed by Gerald Ford made the switch to metric voluntary — and that was that.

In the end, both systems were anchored to the metric one. An international agreement in 1959 defined the commercial pound directly in terms of the kilogram: 1 lb = 0.45359237 kg, exactly. That is where the numbers in any correct mass converter come from: the ounce is a pound divided by 16 (28.349523125 g), the stone is 14 pounds (6.35029318 kg), and the US ton is 2,000 pounds (907.18474 kg). None of these is an approximation — all are exact by definition.

The bill for carelessness

Two systems side by side are an invitation to disaster. Here are four invoices.

Mars Climate Orbiter (1999). Lockheed Martin's ground software reported the impulse of the trajectory-correction thrusters in pound-force seconds (lbf·s). The navigation software at JPL read the same numbers as newton-seconds (N·s). Since 1 lbf·s ≈ 4.44822 N·s, for nine months of flight the navigators believed the manoeuvres were more than four times weaker than they actually were. The probe entered the Martian atmosphere at roughly 57 km instead of the planned 140–150 km — below the survival threshold of about 80 km — and was destroyed on 23 September 1999. The orbiter itself cost 125 million dollars; the programme as a whole, including the Mars Polar Lander lost three months later, came to 327.6 million.

The Gimli Glider (1983). On 23 July, the Boeing 767 flying Air Canada 143 lost both engines at 41,000 feet (about 12,500 m). It was the airline's first fully metric aircraft — its gauges showed fuel in kilograms while the rest of the fleet counted in pounds. After a gauge failure, the crew dipped the tanks manually, got 7,682 litres, and multiplied by 1.77 lb/L instead of 0.803 kg/L. The result — 13,597 — was treated as kilograms. To reach the required 22,300 kg they therefore added only 4,917 litres. The aircraft took off with about 10,100 kg of fuel, less than half of what it needed. The bitter irony: 10,100 kg is roughly 22,300 pounds — the crew had exactly the number they wanted, in the wrong unit. Captain Bob Pearson glided the jet onto a decommissioned runway at Gimli, Manitoba. Nobody died.

Space Mountain, Tokyo Disneyland (2003). In 1995 the specification for the ride cars' axles was converted from inches to millimetres, the diameter rounded to 45 mm — but the old drawings were never withdrawn from the archive. When spare parts were ordered in 2002, the older version was used and axles were manufactured at 44.14 mm. Bearing clearance grew from an allowable 0.2 mm to over 1 mm; on 5 December 2003 the fatigued metal snapped and the train derailed just short of the station. Nobody was hurt; the attraction stayed closed for many weeks.

A prescription (1999). The US Institute for Safe Medication Practices (ISMP) documented a case in which "0.5 gr" on a label was read as 0.5 grams of phenobarbital rather than 0.5 grains. A grain is about 65 mg, and apothecary practice rounded it to 60 mg — so for three days the patient received 500 mg instead of roughly 30 mg, more than fifteen times the intended dose. He survived. The ISMP drew one conclusion: in medicine, only the metric system may be used.

The cylinder near Paris

The answer to this chaos was the kilogram. In 1795 it was defined as the mass of a litre of water at the melting point of ice (soon corrected to 4 °C, the temperature of maximum density), and in 1799 its first physical embodiment was cast — the platinum Kilogramme des Archives.

In 1889 the Metre Convention anointed a new standard: the International Prototype of the Kilogram (IPK), known as Le Grand K. A cylinder about 39 mm in height and diameter, made of an alloy of 90% platinum and 10% iridium — extremely dense (about 21.5 g/cm³, so little surface area to contaminate), barely magnetic, and practically non-oxidizing. It was cast in 1879 by the London firm Johnson Matthey and placed in a triple-locked safe at the Pavillon de Breteuil in Sèvres, under three glass bell jars. Several dozen official copies went to the signatory states: the USA received K20 (primary standard) and K4 (check standard), Britain received K18.

And then the trouble began. Three periodic verifications (1899–1911, 1939–1953, 1988–1992) showed that the IPK was losing mass relative to its own copies — on the order of 50 micrograms per century. A beautiful paradox followed: since Le Grand K was the kilogram as a matter of law, it was not losing weight — the rest of the universe was gaining it, and with it the newton, the joule, and the watt.

Surface physics was to blame. Despite the bell jars, the platinum adsorbed water vapour, hydrocarbons, and mercury from the air — the BIPM estimated the gain at about 1.11 µg per month over the first three months after cleaning, settling later at roughly 1 µg per year. Cleaning itself — rubbing the metal with chamois leather soaked in equal parts ether and ethanol, then a jet of steam from doubly distilled water — removed 5 to 60 µg at a stroke, but risked microscopic abrasion. Just afterwards, unexplained fluctuations of about 30 µg over a month were still observed.

A kilogram without a cylinder

On 16 November 2018, the member states of the BIPM voted to redefine four SI units — the kilogram, the ampere, the kelvin, and the mole. Since 20 May 2019 the kilogram has been defined through the Planck constant, whose value is fixed as exactly h = 6.62607015 × 10⁻³⁴ J·s. Not through a thing, but through a measurement.

The physical realization of that definition is the Kibble balance. It balances the weight of a mass against an electromagnetic force, working in two phases. In the static phase, a current I flows through a coil of wire length l sitting in a field of flux density B, and the Lorentz force balances the weight: M · g = B · l · I. In the dynamic phase the mass is removed and the coil is moved through the same field at velocity u, while the induced voltage is measured: V = B · l · u. Multiply the two equations and the awkward product B · l vanishes, leaving a clean equality of powers: M · g · u = V · I. Voltage and current are measured against quantum standards — the Josephson junction and the quantum Hall effect — and both depend solely on the Planck constant and the elementary charge. The best Kibble balances reach uncertainties of around ten parts per billion.

The second, independent route is the Avogadro project: counting the atoms in an almost perfect sphere of single-crystal silicon-28 of better than 99.9995% purity. These 1 kg spheres, 93.6 mm across, are among the roundest objects ever made by hand — deviations from sphericity are measured in tens of nanometres. Knowing the sphere's volume (laser interferometry) and the crystal lattice constant a (X-ray diffraction), the number of atoms follows directly: N = 8V / a³, where the 8 is the number of atoms in the unit cell of a diamond-type lattice.

The prettiest reading of the new definition comes from Nobel laureate Wolfgang Ketterle. Combining Planck's formula (E = h · ν) with Einstein's (E = m · c²) gives the mass of a photon: m = h · ν / c². Take as the reference frequency the hyperfine transition of caesium-133 that defines the second (9,192,631,770 Hz), and such a photon "weighs" about 6.777 × 10⁻⁴¹ kg. The kilogram is therefore the mass of exactly 1.4755 × 10⁴⁰ such photons. Trap them in a perfect mirrored cavity and it would grow heavier by one kilogram. Technically impossible — pedagogically perfect.

	The artefact era (1889–2019)	The quantum era (from 2019)
Definition	Mass of a Pt-10Ir cylinder (Le Grand K)	Planck constant h = 6.62607015 × 10⁻³⁴ J·s
Where the standard lives	The BIPM vault at Sèvres	Everywhere — no physical anchor
Stability	Drift of about 50 µg per century	A physical constant does not age
Realization	Comparison on a mechanical balance	Kibble balance, Avogadro project
Limit on accuracy	Contamination and cleaning of the standard	The measuring apparatus alone

The twelve micrograms nobody talks about

You cannot put a Kibble balance in every factory, so everyday calibration still rests on secondary metal weights. To keep the world from drifting apart during the handover from Le Grand K to quantum physics, the BIPM publishes a consensus value of the kilogram — an average of the independent quantum realizations, compared every few years.

Since 1 March 2026, the third such value has been in force. It is 1 kg − 12 µg (standard uncertainty 20 µg) with respect to the mass of the old prototype at Sèvres. That is 5 µg below the 2023 value (1 kg − 7 µg), which was itself 5 µg below the first one from 2021. Note that the uncertainty is larger than the correction itself — this is a scale at which metrology laboratories must adjust their procedures and the rest of the world notices nothing.

A measure that depends on no object

The history of the pound is a story about how expensive convenience turns out to be. Every local measure once made sense: a merchant weighed wool differently from how a goldsmith weighed gold or an apothecary weighed powder. The trouble started when those worlds began talking to each other — through software, a refuelling slip, or a technical drawing seven years old.

The kilogram travelled the same road as the metre: from object to constant. Le Grand K, for all its three bell jars and triple-locked safe, remained a piece of metal that could be lost, scratched, or — worst of all — simply left alone for a hundred years. Today the kilogram is a definition, not a thing. It can be reproduced in Washington, Braunschweig, and Ottawa, and the results will agree to within a few tens of parts per billion. The pounds have not gone away — but at least we now know exactly what they weigh.

Not all pounds are equal — how we weighed the world