A kilogram of feathers and a kilogram of lead — why the scale disagrees

The riddle is as old as the school desk: which is heavier, a kilogram of feathers or a kilogram of lead? Everybody knows the answer. The same, because both are a kilogram. The teacher scores a point, the class nods, the matter seems closed.

It isn't. The riddle has three floors, and the schoolyard answer stops on the first. On the second waits the history of two weight systems, in which a pound of feathers really is heavier than a pound of gold. On the third waits the air, which lifts feathers far more than it lifts lead — so that a kilogram of feathers placed on a scale reads about sixty grams light.

To reach the bottom you need a quantity that is neither mass nor weight. You need density.

The second floor: a pound of feathers beats a pound of gold

Swap the kilograms for pounds and ask again: which weighs more, a pound of feathers or a pound of gold? Intuition says the same as before. This time intuition is wrong.

The Anglosphere never settled on a single pound. Everyday goods — feathers, flour, wool — are weighed in the avoirdupois system (from Old French aveir de pois, "goods of weight"), where the pound splits into 16 ounces and equals exactly 453.59237 g. Precious metals and gemstones fall under the troy system, whose pound has only 12 ounces and weighs 373.2417216 g.

A pound of feathers is therefore heavier than a pound of gold by more than 80 grams. The confusion is complete once you notice that single ounces reverse the relation: a troy ounce of gold (31.10 g) outweighs an avoirdupois ounce of feathers (28.35 g). Where these parallel systems came from, and why both survived, we tell separately in Not all pounds are equal. The moral is enough here: until units are coherent, arithmetic will not save you.

Three words we conflate

Reaching the third floor requires separating three words that everyday speech uses interchangeably.

Mass (m, in kilograms) is the amount of matter — an intrinsic property, a measure of inertia.

Weight (F = m · g, in newtons) is the force with which gravity pulls that mass down. It depends on where you are: the same weight-piece pulls differently at the pole and at the equator, and six times less on the Moon than on Earth. The mass, throughout, does not change.

Density (ρ = m / V) adds the spatial dimension: how much mass fits into a unit of volume. It is an intensive quantity — it does not depend on sample size. A drop of water has the same density as the whole ocean, provided temperature and pressure match.

And density is what the whole riddle turns on. A kilogram of lead (11,340 kg/m³) occupies 88 cm³ — it fits in your palm. A kilogram of loosely piled feathers has a density on the order of 20 kg/m³, so it takes up roughly 50 litres, six hundred times more. That difference in volume is innocent only in a vacuum.

The third floor: air displaces too

Archimedes' principle is not about water alone. Air is a fluid, and it pushes up on every body immersed in it with a force equal to the weight of the displaced air. The density of dry air at 20 °C and 1 atm is about 1.204 kg/m³ — a conventional figure, sensitive to temperature, pressure and humidity, but good enough for the arithmetic.

Let us weigh the air each of our kilograms shoves aside.

Lead: 88.2 cm³ × 1.204 kg/m³ = 0.106 grams of displaced air. Negligible.

Feathers: 50 litres × 1.204 kg/m³ = 60.2 grams of displaced air. Not negligible.

So if both bodies truly carry one kilogram of mass, a scale in an ordinary room will disagree with itself: lead reads almost exactly 1000 g, feathers barely 940 g. Sixty grams of difference, manufactured by the air alone.

Now invert the experiment. If the scale is to read one kilogram for both, you must heap more matter onto the feather pan — its true mass has to reach about 1.06 kg. The same physics runs in both directions; only the starting point differs. It is in a vacuum, and only there, that the schoolyard answer becomes unconditionally true. Metrologists know this well, and weighing mass standards to the microgram involves a routine air buoyancy correction — otherwise a steel weight and a platinum standard of identical mass would give different readings.

Why 1 g/cm³ is exactly 1000 kg/m³

Anyone who has ever computed a density has noticed the suspicious elegance of its conversions. 1 g/cm³ is a clean 1000 kg/m³. 1 g/mL is exactly 1 g/cm³. 1 g/L is exactly 1 kg/m³. No fractions, no constants pulled from tables. That is not luck; it is what the decimal architecture of the metric system buys you.

Write out the first identity. A gram is 10⁻³ kg. A metre is 100 cm, so a cubic metre is (100 cm)³ = 10⁶ cm³, which makes 1 cm³ = 10⁻⁶ m³. Substituting:

1 g/cm³ = 10⁻³ kg / 10⁻⁶ m³ = 10³ kg/m³ = 1000 kg/m³

Going from centimetres to metres in the denominator costs a factor of a million; going from grams to kilograms in the numerator costs only a thousand. A million divided by a thousand is a thousand. That is the entire mystery.

The second identity is simpler still, because it follows straight from the definition of the litre: a litre is a cubic decimetre, that is 1000 cm³. A millilitre is a thousandth of a litre, hence exactly 1 cm³. Identical volumes mean identical densities: 1 g/mL = 1 g/cm³. By the same token, 1 g/L = 10⁻³ kg / 10⁻³ m³ = 1 kg/m³.

You will find all of these in our density converter — along with the one unit that breaks the harmony, and to which we devote a section below.

There is one catch. The sentence "a millilitre is exactly a cubic centimetre" has been true since 1964. Before that, it wasn't.

The litre that spent 63 years not being a cubic decimetre

In 1901 the 3rd General Conference on Weights and Measures (CGPM) set out to make chemists' lives easier. Since laboratories measure everything against water anyway, let the litre be defined not geometrically but by mass: a litre is the volume occupied by 1 kg of pure, air-free water at the temperature of its maximum density, under standard pressure.

The decision rested on a silent assumption — that the platinum-iridium kilogram standard outside Paris matched the mass of exactly one cubic decimetre of such water. The assumption was false. As measurement improved, it emerged that the standard was slightly too heavy, which meant a kilogram of water occupied marginally more than 1 dm³:

1 L ≈ 1.000028 dm³

Twenty-eight parts per million. To someone buying milk, nothing. To an analytical chemist, a nightmare: the millilitre had stopped being a cubic centimetre, and a density quoted in g/mL was no longer the same number as the same density in g/cm³.

This metrological double life lasted 63 years. Only in 1964 did the 12th CGPM cut the knot with Resolution 6: the 1901 definition was annulled and the word "litre" demoted to a special name for the cubic decimetre.

The water that expands when it freezes

Why tie units to water in the first place? Revolutionary France. A decree of 7 April 1795 defined the gram as the mass of a cubic centimetre of pure water at the melting point of ice. The reference proved poor: water at 0 °C is unbearably sensitive to pressure fluctuations, and it behaves perversely — warmed from zero, it grows denser rather than thinner. So the temperature at which water reaches its natural maximum density was chosen instead, and on that basis the Kilogramme des Archives was cast in 1799.

Today, thanks to measurements on isotopically standardised VSMOW water, we know that point precisely. The maximum falls at 3.983 °C, where the density is:

ρ_max = 999.974 95 kg/m³ ≈ 0.999 975 g/cm³

About 25 parts per million below a round thousand. This is the very discrepancy that broke the litre — the 28 ppm adopted then and the 25 ppm known now differ only by how much better we measure today. The familiar "water has a density of 1 g/cm³" is a rounding, convenient and entirely sufficient outside a laboratory. And a double rounding at that, since at room temperature (20 °C) water is already down to 998.21 kg/m³, and just short of boiling, at 95 °C, to a mere 961.89 kg/m³.

That maximum at just under four degrees is one of nature's most consequential anomalies. Almost every substance grows denser as it solidifies: molecules lose energy and pack tighter. Water does the opposite. Below 3.98 °C thermal energy no longer suffices to override the geometry of hydrogen bonds, and the molecules settle into hexagonal ice — a lattice in which each molecule holds four neighbours at arm's length, leaving void between them. On freezing, volume jumps by roughly 9% and density falls from 999.84 kg/m³ (water at 0 °C) to 916.7 kg/m³ (ice at 0 °C).

That is why ice floats, submerged to 91.7% of its volume in fresh water. (An iceberg floats in seawater, denser for its salt, so it hides "only" about 90% below the surface — a proverbial proportion, rarely quoted exactly.)

The consequence is enormous. In autumn, chilled surface water sinks and pushes warmer water up, and this convection continues until the whole lake reaches the temperature of maximum density, around 4 °C. Cooling the surface further makes it lighter, so it stays on top and freezes there. Ice, a poor conductor of heat, then acts as a blanket. Beneath it, at the bottom, a layer of liquid water always remains at a survivable four degrees. Had water densified on freezing like everything else, lakes would freeze from the bottom up and freeze solid.

16.018463… — an imperial relic in the numerator

American engineering, for all that, still counts density in pounds per cubic foot. And its conversion factor looks like a typo:

1 lb/ft³ = 16.018 463 373 96 kg/m³

That ugly number is not arbitrary — it is the collision of two philosophies of measurement, worked out to the last digit. The International Yard and Pound Agreement of 1959 fixed both conversions exactly: 1 lb = 0.453 592 37 kg and 1 ft = 0.3048 m. A cubic foot is therefore (0.3048 m)³ = 0.028 316 846 592 m³, and the rest is division:

0.453 592 37 kg ÷ 0.028 316 846 592 m³ = 16.018 463 373 96 kg/m³

No magic, no rounding — simply a pound over a foot cubed. The unit holds on tenaciously: geotechnical engineers quote soil bulk density in it, structural engineers weigh concrete with it, and ventilation engineers call 0.075 lb/ft³ "standard air", which is roughly the familiar 1.20 kg/m³.

Specific gravity: a number without a unit

Trades and industry, though, prefer a quantity with no unit at all. Specific gravity is the ratio of a substance's density to that of water. Since water is approximately 1 g/cm³, the number comes out nearly identical to the density in g/cm³ — except that you drag no units behind it and need not remember which system you are in. Three professions use it daily, and none of them thinks about physics while doing so.

The brewer. Before the yeast goes in, wort is thick with malt sugars — its specific gravity might read 1.050. The yeast turns sugar into ethanol, which is far thinner than water (789 kg/m³), so the solution's density falls. A final reading of 1.010 tells the brewer two things at once: that fermentation has finished, and that the glass holds about 5.2% alcohol.

The mechanic. In a lead-acid battery the electrolyte is sulfuric acid diluted with water. On discharge, sulfate ions bind to the plates and water returns to the solution — the dense acid (ρ ≈ 1.84 g/cm³) grows dilute. A simple hydrometer reads off the state of charge: about 1.265–1.280 means a full battery, 1.190 is roughly half, and 1.120 says plug in the charger now.

The geologist. In the field there is no laboratory, only a hand. Quartz and calcite have specific gravities of 2.6–2.7 and "weigh what they ought to". But a sample of barite (4.5) feels unnaturally heavy for a pale, non-metallic mineral, and galena (7.4–7.6) simply feels like a lump of metal. This test, the heft test, narrows the search in a second.

A ladder of densities: from air to a neutron star

The range of densities in nature is hard to hold in the mind. Below, values at 20 °C and 1 atm unless noted otherwise.

Two questions fall out of this table, and both deserve an answer.

Why is gold denser than lead? A lead atom is the heavier of the two (207.2 u against gold's 196.97 u), and both metals crystallise in the same face-centred cubic structure. Yet gold (19.30 g/cm³) is 70% denser than lead (11.34 g/cm³). The answer lies in spacing: in gold's lattice, neighbouring atoms sit 288.4 pm apart; in lead's, a full 350.0 pm. With packing identical, density scales as atomic mass divided by the cube of that spacing. Run it: (196.97 / 207.2) × (350.0 / 288.4)³ = 0.951 × 1.789 = 1.70. Exactly the observed density ratio. The heavier atom loses because it sits loose — gold's atoms, squeezed by the lanthanide contraction and by relativistic effects, simply pack tighter.

Osmium or iridium? The contest for densest stable element dragged on for decades, because the two values differ by about 0.1% — less than the error of classical hydrostatic weighing, which is hostage to microbubbles and impurities in the sample. X-ray crystallography settled it, by measuring the unit cell rather than weighing the metal and computing density from atomic mass. Osmium won: 22.587 against 22.562 g/cm³.

And at the foot of the table, physics stops pretending to be intuitive. When a massive star collapses after a supernova, gravity overcomes electron degeneracy pressure and drives electrons into protons, turning matter into a sea of neutrons. A single cubic centimetre of that material — a cube one centimetre on a side — would weigh between 370 and 590 million tonnes on Earth.

The bathtub that never was

Since this whole story began with buoyancy, we owe a visit to the most famous anecdote in the history of science. Archimedes lowers himself into a bath, sees the water spill over, grasps the principle of buoyancy, and runs naked through the streets of Syracuse shouting Eureka! — then exposes the goldsmith who cut King Hiero II's crown with silver.

The first warning sign: Archimedes did not write it down. His surviving works are strictly mathematical. The story was recorded by the Roman architect Vitruvius in the preface to Book IX of On Architecture, late in the 1st century BC — nearly two hundred years after the scholar died during the siege of Syracuse.

The second warning sign is physical. Vitruvius' version rests on measuring the water spilled over the vessel's rim. Let us see how much water that is. A one-kilogram crown of pure gold occupies 51.8 cm³. The same mass of pure silver occupies 95.2 cm³. Had the goldsmith swapped 20% of the gold for silver, the crown's volume would grow to 60.5 cm³ — that is 8.7 cm³ more than pure gold. Under two teaspoons.

The crown has to fit in the vessel, so assume one 20 cm across; its water surface is then 314 cm². The rise in level caused by the fraud comes to:

Δh = 8.7 cm³ ÷ 314 cm² ≈ 0.28 mm

A quarter of a millimetre. Surface tension, the meniscus, water clinging to the walls, and bubbles trapped in the crown's ornament introduce errors many times larger than the effect being measured. The method Vitruvius describes is simply unworkable.

A 22-year-old Galileo noticed as much. In 1586 he wrote a short treatise, La Bilancetta ("The Little Balance"), arguing that crediting Archimedes with so crude an idea insults the author of rigorous treatises on floating bodies. And he proposed a reconstruction worthy of the original: the hydrostatic balance.

The crown hangs from one arm and is counterpoised in air. Then it is lowered into water. The crown apparently loses weight equal to the water it displaces — exactly as our feathers shed sixty grams in air — and the balance tips. To restore it, the counterweight slides toward the fulcrum, and Galileo read that displacement off by winding fine wire around the arm and counting the turns. The method collects no spilled water, fears no meniscus, and catches a silver adulteration without difficulty. It measures precisely what was at stake: density.

The invisible architect

Density is a derived quantity, wedged between mass and volume, and easy to overlook. Yet it is density — not mass, not weight — that decides what sinks and what floats: a warship of many thousand tonnes stays up because its average density, hull air included, is lower than the sea's, while a fistful of steel nails massing one kilogram goes straight to the bottom.

On a larger scale, density moves the planet. Cold, denser air sinks and builds weather fronts. Density differences in ocean water, arising from temperature and salinity, drive the thermohaline circulation that redistributes heat around the globe. Convection in Earth's mantle shifts the tectonic plates.

The schoolyard riddle about feathers and lead is thus shrewder than its authors intended. "The same" is the right answer — in a vacuum. Everywhere else you need to know how much room a kilogram takes up. And that is a question about density.

Further reading

• Vitruvius, On Architecture, Book IX (preface) — the original record of the legend of Hiero's crown.
• Galileo Galilei, La Bilancetta (1586) — a reconstruction of Archimedes' method built on the hydrostatic balance.
• BIPM, Resolution 6 of the 12th CGPM (1964) — the annulment of the "water" definition of the litre; bipm.org.
• BIPM, The International System of Units (SI Brochure), 9th ed. (2019) — the definitions in force.
• M. Tanaka et al., Recommended table for the density of water between 0 °C and 40 °C, Metrologia 38 (2001) — the VSMOW data.
• IAPWS, Formulation 1995 (IAPWS-95) — the reference equation of state for water; iapws.org.
• J. W. Arblaster, Densities of Osmium and Iridium, Platinum Metals Review (1989) — the X-ray verdict on the densest element.
• CRC Handbook of Chemistry and Physics — density tables for elements and compounds.
• Główny Urząd Miar, National density standard (gum.gov.pl) — Poland's density measurement infrastructure.