The Wilberforce Pendulum: Trading Bounce for Twist

A mass that bounces and twists at once

Hang a weight from a long, soft helical spring and it has two independent ways to oscillate. It can bounce — the spring stretches and recoils, moving the mass vertically at frequency ω_z = √(k/m), where k is the spring's stiffness and m the mass. It can also twist — the mass rotates back and forth about the vertical axis as the coils wind and unwind, an angular oscillation at frequency ω_θ = √(δ/I), where δ is the spring's torsional stiffness and I the mass's moment of inertia. In an ordinary demonstration these two motions have very different rates and barely notice each other.

The trick, discovered by Lionel Robert Wilberforce (1861–1944), a demonstrator at the Cavendish Laboratory in Cambridge and later professor of physics at Liverpool, is to deliberately make the two frequencies equal. His 1894 paper in the Philosophical Magazine, “On the vibrations of a loaded spiral spring,” described a bob fitted with small adjustable radial masses on threaded arms. Sliding those masses in or out changes the moment of inertia I — and therefore the twist frequency ω_θ — without changing the bounce frequency ω_z at all. This gives the experimenter a fine, independent knob to bring the two into exact resonance.

Why stretching a helix also twists it

The reason bounce and twist can talk to each other lies in the geometry of the coiled wire. A helical spring is stiff not because its wire stretches, but because the wire is twisted: pull on the ends of a helix and each little element of wire is loaded chiefly in torsion. That is why a spring's stiffness depends on the material's shear modulus G rather than its Young's modulus, k = Gd⁴/(8D³n) for wire diameter d, coil diameter D and n turns.

Because axial stretching is really wire-torsion in disguise, a helical spring does not extend and rotate independently — the two are elastically coupled. Stretch the spring and the free end rotates slightly; twist the free end and the spring changes length slightly. Physically, pulling on the coils changes the helix pitch angle, and conserving the wire's elastic geometry forces a small rotation. This translation–rotation coupling is built into every real helical spring; the Wilberforce pendulum simply arranges for it to matter. We can encode all three elastic effects in a single potential energy,

U = ½ k z² + ½ δ θ² + ½ ε z θ,

where z is the vertical displacement, θ the twist angle, and ε is the coupling constant that ties them together. The cross term ½εzθ is small, but it is the entire show.

The coupled equations and their normal modes

Differentiating the potential gives the force on the mass and the torque about the axis, and Newton's laws become a pair of coupled linear differential equations:

m z̈ = −k z − ½ε θ,     I θ̈ = −δ θ − ½ε z.

Each equation is a simple harmonic oscillator with an extra term feeding in the other coordinate. Because they are linear, the system has two normal modes — special patterns in which the bounce and the twist oscillate together at a single, pure frequency, in a fixed amplitude ratio. Substituting z, θ ∝ e^iωt yields the characteristic condition (k − mω²)(δ − Iω²) = ¼ε².

Now impose Wilberforce's tuning, k/m = δ/I ≡ ω₀². The two solutions split symmetrically about ω₀:

ω_±² = ω₀² ± ε / (2√(mI)).

In the lower mode the mass bounces and twists in phase; in the upper mode they run out of phase. Crucially, the coupling prevents the two frequencies from ever crossing even as you tune through resonance — they veer apart and repel, an avoided crossing (level repulsion) that is a universal fingerprint of coupled oscillators, from molecular vibrations to quantum energy levels.

Beats: energy sloshing from bounce to twist

Here is where the magic appears. Suppose you start the pendulum in pure bounce, with no twist. That initial condition is not a normal mode — it is an equal superposition of both normal modes, one at ω₊ and one at ω₋. Because the two modes tick at slightly different rates, they gradually drift out of step. When they are exactly out of phase, their vertical contributions cancel and their torsional contributions add: the bounce has vanished and the motion is pure twist. Keep waiting and they realign, returning the energy to bounce. This is a textbook beat phenomenon, identical in spirit to two tuning forks producing a throbbing tone.

The bounce energy follows an envelope ∝ cos²[(ω₊ − ω₋)t / 2], so a complete bounce → twist → bounce cycle takes the beat period T_beat = 2π / (ω₊ − ω₋). The number of individual bounces the mass makes before the energy has fully migrated into twisting is roughly N ≈ ω₀²√(mI)/ε — typically 10 to 30 oscillations per handoff for a classroom pendulum with a bounce period near one second and a beat period of tens of seconds. Because the coupling ε is small, the transfer is slow and stately, which is exactly what makes it so mesmerizing to watch. No energy is lost in the exchange; it is merely reshuffled between the two degrees of freedom, a perfectly reversible sharing that continues until damping finally drains the whole system.

Tuning the resonance — and why detuning kills it

Complete energy transfer — the bounce dying all the way to zero — happens only when the two uncoupled frequencies are exactly matched, ω_z = ω_θ. This is why the adjustable inertia is essential: the resonance window is narrow. If a detuning Δ = ω_z − ω_θ creeps in, the maximum fraction of energy that can flow into the twist mode falls to roughly 1 / [1 + (2Δω₀√(mI)/ε)²]. Once the detuning grows comparable to the coupling strength, the two modes essentially ignore each other and the mass just bounces the way any ordinary spring would. The same mathematics describes a driven two-level atom, where off-resonant driving gives only shallow, incomplete Rabi flopping — the Wilberforce pendulum is a purely mechanical cousin of that quantum effect.

Two other subtleties bound the demonstration. First, damping: air drag and internal friction give the pendulum a quality factor Q of order 100–several hundred, so you see a handful of clean bounce↔twist exchanges before the amplitudes decay away — the higher the Q, the more beats survive. Second, amplitude: at large swings the spring stops being perfectly linear (anharmonicity), the coupling constant drifts, and the beautiful clean beating degrades into a messier, aperiodic exchange. The ideal demonstration lives in the sweet spot of exact tuning, modest amplitude, and low damping.

How we measure it, and where else it hides

Modern treatments make the two-mode picture quantitative. The definitive analysis by Richard E. Berg and Todd S. Marshall, “Wilberforce pendulum oscillations and normal modes” (American Journal of Physics 59, 32, 1991), derived the coupled equations above, measured the normal-mode splitting directly, and confirmed that the bounce and twist envelopes are exactly out of phase — energy leaving one appears in the other. Today the motion is routinely captured with video tracking or a laser-and-photogate: the vertical position and the rotation angle are logged simultaneously, and their amplitude envelopes trace out the complementary sloshing curves that make the coupling undeniable.

• Molecular vibrations. The same avoided crossing appears when two vibrational modes of a molecule are accidentally degenerate. In carbon dioxide, the symmetric stretch (~1337 cm⁻¹) is nearly resonant with the first overtone of the bend (2 × 667 cm⁻¹); anharmonic coupling splits them into a Fermi doublet at ~1388 and ~1285 cm⁻¹, exactly as Enrico Fermi explained in 1931.
• Structural and MEMS engineering. Unwanted translation–rotation coupling in springs, shafts and micro-resonators can steal energy from a desired mode into a parasitic one at resonance — the Wilberforce mechanism is a cautionary tale for designers.
• Teaching normal modes. As a slow, visible, single-object demonstration of mode mixing, the Wilberforce pendulum remains a staple of physics lecture halls, a mechanical analog of everything from coupled pendulums to two-level quantum systems.

Its enduring appeal is that a single humble spring makes an abstract idea — that energy in a resonant system belongs to modes, not to individual motions — something you can watch unfold in real time.

The Wilberforce pendulum among two-mode coupled oscillators

System	The two modes	What couples them	Visible signature
Wilberforce pendulum	Vertical bounce ↔ torsional twist	Stretch–twist elasticity of the helix	Bounce fades as twist grows, then reverses
Two pendulums + spring	Left swing ↔ right swing	Weak connecting spring	One pendulum stalls while the other builds
Coupled LC loops	Loop-1 current ↔ loop-2 current	Mutual inductance	Charge sloshes back and forth between loops
CO₂ molecule (Fermi resonance)	Symmetric stretch ↔ bend overtone	Anharmonic coupling	Fermi doublet split in the IR/Raman spectrum
Two-level atom (Rabi)	Ground state ↔ excited state	Resonant driving field	Rabi flopping of the population