The Cavendish experiment was a laboratory measurement of the gravitational attraction between lead spheres, carried out by Henry Cavendish and published in 1798. Using a sensitive torsion balance, Cavendish compared the attraction produced by nearby masses with the gravitational attraction of Earth, thereby determining Earth’s mean density. The experiment is now commonly described as the first laboratory determination of the gravitational constant , although Cavendish reported his result as Earth’s density rather than as a value of .
Historical background
Isaac Newton’s law of universal gravitation described the attraction between two bodies as proportional to the product of their masses and inversely proportional to the square of their separation. It successfully explained planetary and lunar motion, but astronomical observations alone did not independently establish the absolute scale of masses and the gravitational constant. For example, an orbit can determine the product of a central body’s mass and much more directly than either quantity separately.
Before Cavendish’s work, attempts to estimate Earth’s density included measurements of the deflection of a plumb line by a mountain. The Schiehallion experiment of the 1770s used this approach, but its interpretation depended on estimates of the mountain’s shape and internal density.
The apparatus used by Cavendish was originally devised and constructed by John Michell, who died in 1793 before completing the intended measurements. Cavendish subsequently acquired the apparatus, modified it, and carried out the experiment. His paper, Experiments to Determine the Density of the Earth, appeared in the Philosophical Transactions of the Royal Society in 1798.
The phrase “weighing the Earth” became associated with the experiment. This is shorthand: Cavendish did not balance Earth against a known weight, but inferred its mean density by comparing gravitational effects.
Apparatus and procedure
The central instrument consisted of a light horizontal wooden rod, approximately 1.8 metres long, suspended at its centre by a thin wire. A small lead sphere, about 5 centimetres in diameter, was attached to each end. Two much larger lead spheres, each approximately 30 centimetres in diameter and about 158 kilograms in mass, could be moved into positions near the smaller spheres.
The attraction of the large spheres exerted a torque on the suspended assembly. As the rod rotated, the suspension wire twisted and produced an opposing restoring torque. The rod’s equilibrium position therefore depended on the strength of the gravitational attraction.
The apparatus was enclosed in a wooden case to reduce air currents and temperature disturbances. Cavendish observed scales at the rod’s ends through telescopes from outside the experimental room, reducing interference from his own movement and body heat. The large spheres could also be repositioned without entering the room.
The experiment involved two complementary measurements:
- Angular displacement: moving the large spheres between opposite configurations changed the gravitational torque and shifted the rod’s equilibrium position.
- Oscillation period: the time taken for the suspended assembly to oscillate provided a measure of the suspension wire’s torsional stiffness.
Reversing the positions of the attracting masses helped distinguish their gravitational effect from a fixed offset in the suspension. Cavendish also analysed oscillations and changes in equilibrium rather than relying solely on a single stationary reading.
Physical principle
Gravitational attraction
For two spherically symmetric masses and , separated by a centre-to-centre distance , Newtonian gravitation gives
Here, is the attractive force, and sets the strength of the interaction. The force between laboratory-sized objects is extremely small. A torsion balance makes it measurable because a fine suspension wire can provide a very weak restoring torque, allowing a small force to produce an observable rotation.
For a small twist angle , the restoring torque is approximately
where is the wire’s torsional stiffness.
In a simplified symmetric arrangement, each small sphere experiences an approximately tangential force at a distance from the suspension axis. At equilibrium,
This expression is illustrative rather than a complete model of the historical apparatus. A precise calculation must account for the directions and distances of all relevant attractions, including each large sphere’s attraction on the more distant small sphere and on other parts of the balance.
Determining the wire’s stiffness
The balance’s oscillation period , for small motions and negligible damping, is
where is the assembly’s moment of inertia about the suspension axis. Thus,
This supplies a mechanical calibration of the wire without requiring a separately applied force of comparable smallness.
Combining these relations gives the simplified modern expression
The angular displacement in this expression is measured from the position corresponding to zero gravitational torque in the idealized model. If the observation is instead the difference between two symmetric, opposite source-mass configurations, that difference is approximately .
In actual measurements, the attracting masses can also change the effective restoring torque as the balance rotates. Their influence on the oscillation period must therefore be included when significant.
From laboratory attraction to Earth’s density
Cavendish’s original result can be understood as a comparison between the attraction of the large lead spheres and the weight of the small spheres in Earth’s gravitational field.
Approximating Earth as a sphere with radius , mass , and surface gravitational acceleration ,
Its mean density is
A measurement of laboratory attraction can therefore be expressed either as a value of or, with suitable terrestrial measurements, as Earth’s mean density. Conversely, a measured density determines through the same relationship.
Cavendish reported that Earth’s mean density was 5.48 times that of water. The modern value is approximately 5.51 times that of water, so his result was within about one percent of the modern value. Determining Earth’s mean density also allowed its total mass to be inferred from its known size.
The often-used description “measurement of ” is physically appropriate in modern notation, but it should not obscure the experiment’s original purpose or the form in which Cavendish presented his findings.
Experimental difficulties and uncertainty
The principal difficulty is not the complexity of the gravitational law but the weakness of the signal. Several effects can produce apparent rotations comparable to, or larger than, those caused by the source masses.
Thermal disturbances and air motion. Uneven heating can generate air currents and slowly change the apparatus’s dimensions. These effects can shift the balance even when the source masses remain stationary.
Suspension behaviour. A real wire is not a perfectly elastic spring. Its restoring torque can depend on its previous loading and on the timescale of deformation. Such behaviour is particularly important when stiffness inferred from oscillations is used to interpret a nearly static displacement.
Geometry and mass distribution. The inferred attraction depends on accurate knowledge of sphere positions, dimensions, and masses. Departures from spherical symmetry, density inhomogeneity, and attraction on the support structure require attention.
External disturbances. Ground vibration, movement of nearby objects, and unwanted electrical or magnetic interactions can disturb the balance. Reversal procedures, shielding, environmental control, and repeated measurements help separate these effects from gravity.
These issues distinguish the sensitivity of an instrument from its overall measurement uncertainty. A balance may detect very small changes while still suffering from a systematic error in the inferred value of .
Significance and later developments
The experiment established a practical method for measuring the gravitational attraction between controlled laboratory masses. It connected terrestrial measurement with the determination of planetary properties and demonstrated the effectiveness of torsion balances as precision instruments.
Later experiments retained the central idea while changing the measurement method. Some determine from equilibrium deflections; others measure changes in oscillation period, use feedback to hold a balance at a fixed orientation, or measure angular acceleration. Additional approaches include precision weighing and measurements of the motion of freely falling atoms using interferometry.
The gravitational constant remains unusually difficult to measure accurately compared with many other fundamental constants. Its commonly quoted value is approximately
Unlike constants assigned exact numerical values in the modern International System of Units, must be determined experimentally.
Cavendish-type apparatus is also used in education to demonstrate gravitational attraction between ordinary objects. Precision tests of departures from the inverse-square law employ related techniques, although those tests require different experimental geometries and analyses from a straightforward determination of .