eduKate Learning Manual · Waves × Relativity × Measurement Science · Secondary → JC · Split → Travel → Recombine → Rotate → Constrain
Wait, What? One of Physics’ Most Famous Experiments Became Famous for the Effect It Did Not Find
In the nineteenth century, light was understood as a wave. Familiar waves needed something to oscillate: sound used air, water waves used water, and stretched-string waves used the string. It therefore seemed natural to imagine that light travelled through an invisible medium filling space — the luminiferous ether.
If Earth moved through a stationary ether, light travelling along Earth’s motion should take a slightly different round-trip time from light travelling across it. Albert Michelson and Edward Morley built an interferometer sensitive enough to search for the resulting shift in interference fringes as the apparatus was rotated.
The expected ether-wind signal did not appear at the predicted size.
That null result did not, by itself, prove special relativity. But it severely constrained simple stationary-ether models and became part of the experimental landscape from which modern relativity emerged.
split one light beam → send parts along perpendicular paths → recombine them → interference converts tiny travel-time difference into fringe position → rotate apparatus → expected ether signal should change → predicted shift is absent → model must change.
The Big Question
How can finding almost no change in an interference pattern rule out a physically plausible model of how light travels?
Quick Answer
A Michelson interferometer divides coherent light into two perpendicular arms, reflects both beams and recombines them. The phase difference depends on the optical travel-time difference. In a simple stationary-ether model, Earth’s motion through the ether would make the round-trip light time depend on arm orientation. Rotating the apparatus should therefore shift the interference fringes periodically. Michelson and Morley observed a shift far smaller than the classical ether-wind expectation.
The durable lesson is not “nothing happened.” The experiment made a quantitative prediction for a hidden medium and then failed to find its required signature.
What You Will Learn
- why nineteenth-century physicists proposed an ether
- how a Michelson interferometer converts path difference into fringes
- why parallel and perpendicular light paths were expected to have different times
- why rotating the apparatus is a powerful differential test
- what a null result means scientifically
- why the 1887 result constrained simple ether models but did not uniquely prove relativity
- how FitzGerald–Lorentz contraction entered the historical response
- why vibration, temperature and mechanical strain matter in interferometry
- how modern Lorentz-invariance tests inherit the same measurement logic
Part 1 — Why an Ether Seemed Reasonable
Maxwell’s equations describe electromagnetic waves propagating through space at speed c. Before relativity, physicists were accustomed to wave speeds being defined relative to a medium.
A hypothetical ether could provide that preferred rest frame. If the ether were stationary on cosmic scales while Earth moved through it, an observer on Earth would experience an “ether wind,” analogous to apparent airflow felt by a moving cyclist.
The theory therefore made a testable prediction: measured light travel time should depend on direction relative to Earth’s motion through the ether.
Part 2 — The Interferometer Turns Time Into Phase
A beam splitter sends light along two perpendicular arms. Mirrors reflect both beams back to the splitter, where they recombine.
If one beam travels a slightly different optical path length from the other, it accumulates a phase difference. Recombined waves can interfere constructively or destructively, producing bright and dark fringes.
A path difference of one wavelength changes relative phase by 2π. Therefore a displacement far smaller than anything visible mechanically can produce a measurable fraction of a fringe.
tiny time difference → tiny phase difference → visible fringe displacement.
Part 3 — Along the Ether Wind: Outbound and Return Speeds Differ
In the simple classical ether picture, suppose Earth moves at speed v through the ether and the arm length is L.
Along the direction of motion, the outbound light effectively chases a moving mirror while the return light meets the apparatus coming toward it. The classical round-trip time is:
t∥ = L/(c − v) + L/(c + v)
For v much smaller than c:
t∥ ≈ (2L/c)(1 + v²/c²)
Part 4 — Across the Ether Wind: The Light Must Travel Diagonally
For the perpendicular arm, the apparatus moves sideways while the light is in transit. In the ether picture, the light would need a diagonal path so that it still reaches the moving mirror.
The classical round-trip time becomes:
t⊥ = 2L/√(c² − v²)
For v ≪ c:
t⊥ ≈ (2L/c)(1 + ½v²/c²)
So the simple ether model predicts a round-trip difference approximately:
Δt ≈ Lv²/c³
The effect is second order in v/c, which is why extreme interferometric sensitivity is required.
Part 5 — Rotation Should Reverse the Difference
Now rotate the entire interferometer through 90°. The arm that was parallel to the supposed ether wind becomes perpendicular, and vice versa.
If the ether model is correct, the travel-time difference should change sign. The interference pattern should shift accordingly.
This is much more powerful than comparing an absolute fringe position with a theoretical number. Rotation modulates the predicted signal while many fixed imperfections remain fixed.
The experiment therefore asks not “where are the fringes?” but “do the fringes move with the apparatus orientation exactly as the ether-wind model requires?”
A Quantitative Window — Expected Fringe Shift
If the effective optical arm length is L and the apparatus is rotated by 90°, the simple classical model gives an orientation-change path difference of order:
Δ(path) ≈ 2Lv²/c²
The corresponding number of fringes shifted is approximately:
ΔN ≈ 2Lv²/(λc²)
Michelson and Morley increased effective path length using multiple reflections so the predicted shift would be large enough to detect if Earth’s orbital-scale motion represented motion through a stationary ether.
Part 6 — What They Observed
The 1887 experiment did not show the expected orientation-dependent fringe displacement corresponding to the simple ether-wind estimate. The observed effect was much smaller than predicted and inconsistent with the straightforward stationary-ether model.
This is called a null result: the targeted effect was not detected above the experiment’s sensitivity at the expected magnitude.
A null result is not “no data.” It is a constraint:
if the proposed cause existed in the stated form and strength, the apparatus should have moved the fringes by X; it did not; therefore that version of the model is disfavoured.
Part 7 — Why “It Proved There Is No Ether” Is Too Simple
The null result strongly challenged the simplest picture of a stationary ether through which Earth moved without changing its dimensions or clocks.
But theorists could modify ether-based frameworks. George FitzGerald and Hendrik Lorentz proposed that objects moving through the ether might contract along the direction of motion by just the amount needed to cancel the expected interferometer shift.
Such ideas could reproduce the null result while retaining an ether ontology. Therefore one experiment did not logically force one unique interpretation.
Scientific model choice depends on the wider network of evidence, internal consistency, explanatory economy and new predictions — not one dramatic experiment alone.
Part 8 — Special Relativity Changes the Starting Assumptions
Einstein’s 1905 special relativity begins from principles including the equivalence of inertial frames for physical laws and invariant light speed in vacuum. Space and time then transform between inertial observers through Lorentz transformations.
Within special relativity, no ether-wind fringe shift of the simple classical type is expected because there is no privileged inertial frame in which light alone has its “true” c while other observers should obtain c ± v by Galilean addition.
The Michelson–Morley result is compatible with special relativity and historically important to the collapse of classical ether thinking. But saying “Michelson–Morley directly proved Einstein’s theory” compresses a much richer theoretical and experimental history.
The Historical Carrier — Michelson, Morley and Precision Optics
Albert A. Michelson had pursued ether-drift measurements before the 1887 collaboration with Edward W. Morley. Their improved apparatus used a massive stone support floating in mercury so that it could rotate smoothly while reducing mechanical distortion.
Michelson later received the 1907 Nobel Prize in Physics for precision optical instruments and spectroscopic/metrological investigations carried out with them.
The experiment’s enduring importance lies partly in that instrumentation culture: theoretical questions about space and motion became measurable as optical phase differences.
Part 9 — Why Temperature Is Dangerous
Interference is exquisitely sensitive to path length. Thermal expansion of a support or mirror mount can change optical distance enough to move fringes.
Temperature gradients can also change the refractive index of air along one arm. A false fringe shift could therefore arise from the laboratory itself rather than cosmic motion.
This is why slow rotation, repeated readings, thermal control and comparison across orientations matter.
Part 10 — Why Vibration and Mechanical Strain Are Dangerous
Footsteps, traffic and floor vibration can move mirrors. Rotating a heavy instrument can flex its frame. If rotation itself changes arm lengths mechanically, the experiment could manufacture the orientation-dependent effect it is trying to measure.
The experimental design therefore needs a mechanically symmetric, slowly rotatable platform and repeated angular patterns. A real ether signal should track celestial orientation, not merely a particular creak in the apparatus.
RFE Stress Test — Null Result or Insensitive Instrument?
- Sensitivity calibration: can a known path change produce the expected fringe displacement?
- Rotation test: is the signal searched for at the angular periodicity predicted by the model?
- Thermal control: can temperature-correlated fringe shifts be separated from orientation effects?
- Mechanical reversal: does the platform deform systematically as it rotates?
- Time-of-day/year test: does the supposed preferred direction track Earth’s changing velocity through space?
- Independent replication: do redesigned interferometers with different materials and geometries recover the same null result?
A null result is scientifically meaningful only when the instrument had enough demonstrated sensitivity to detect the effect it failed to find.
Observation vs Inference
Observation: rotating the interferometer did not produce the fringe shift expected from the simple ether-wind calculation.
Immediate inference: the simple stationary-ether model with classical velocity addition is inconsistent with the measured result at the experiment’s sensitivity.
Broader theoretical inference: special relativity provides a coherent framework compatible with the null result, but that framework is supported by a much wider body of evidence than this experiment alone.
Common Misconceptions and How to Repair Them
- “A null result means the experiment failed.” Repair: if sensitivity was sufficient, non-detection can strongly constrain a model.
- “Michelson–Morley proved special relativity in one experiment.” Repair: it constrained classical ether models; relativity is supported by a wider theoretical and experimental network.
- “The interferometer measured light speed directly with a stopwatch.” Repair: it measured relative optical phase/path differences through interference.
- “Rotation was decorative.” Repair: rotation modulated the predicted ether signal and rejected fixed offsets.
- “No ether wind means Earth is stationary.” Repair: Earth moves; the failed assumption was the simple preferred-medium model and Galilean light-speed addition.
- “Interference fringes are just pretty patterns.” Repair: each fringe is a sensitive record of phase difference and therefore optical path.
Checkpoint Questions
- Why did nineteenth-century physicists consider an ether plausible?
- What does a Michelson interferometer compare?
- Why did the simple ether model predict different parallel and perpendicular travel times?
- Why rotate the apparatus?
- What is a null result?
- Why does the null result not uniquely prove special relativity?
- Name three systematic effects that can move interference fringes.
Apply It — Double the Optical Arm Length
In the simple ether calculation, expected fringe shift scales approximately with effective arm length L. Doubling L doubles the predicted signal. This explains why multiple reflections were useful: they increased optical path without requiring a building-sized apparatus.
Unfamiliar Transfer — Null Experiments Can Discover Boundaries
Some of the most powerful modern experiments also search for tiny departures from a theory and find none.
Tests of Lorentz invariance, equivalence-principle violations, photon mass, electric dipole moments and rare particle decays often produce upper limits rather than discoveries.
The reusable reasoning structure is:
alternative theory predicts effect X → instrument sensitivity is demonstrated below X → X is absent → parameter space for that alternative shrinks.
Part 11 — Modern Descendants
Modern Michelson–Morley-type experiments use optical cavities, lasers, cryogenic resonators, rotating platforms and atomic clocks to test whether physical laws depend on orientation or velocity relative to a preferred frame.
The sensitivity is enormously beyond the 1887 apparatus, but the scientific job is recognisable: compare orthogonal directions, modulate orientation, and search for a periodic anisotropy that should not exist if Lorentz symmetry holds.
Answer Key
1. Known waves propagated through media, so an electromagnetic medium seemed natural. 2. Relative optical travel time/phase along two paths. 3. Classical velocity addition made light’s effective travel relative to the apparatus orientation-dependent. 4. It swaps the predicted parallel/perpendicular roles and modulates the signal. 5. A predicted effect is not detected above known sensitivity. 6. Modified ether theories could also accommodate the result; relativity depends on a wider evidence structure. 7. Temperature, refractive-index change, vibration, mirror motion and mechanical strain are examples.
Can You Explain WHY?
Explain why “we saw almost no fringe shift” can be a strong scientific result. A strong answer should connect ether model → quantitative directional prediction → interferometric sensitivity → rotation modulation → observed null → rejected parameter/model region → alternative theory evaluation.
Singapore Secondary and JC Science Bridge
Secondary Physics supplies waves, interference, speed and relative motion. JC Physics adds phase, superposition, reference frames and relativity. Michelson–Morley is especially valuable for the Practices of Science because it teaches that non-detection becomes evidence only after sensitivity and alternative explanations are quantified.
Deep Science Windows
- Lorentz transformation: space and time coordinates mix so c remains invariant between inertial frames.
- Length contraction: in relativity it follows from spacetime transformation, rather than being an ad hoc mechanical deformation through an ether.
- Resonant-cavity tests: orientation-dependent resonant frequencies provide modern ultra-sensitive Lorentz tests.
- Gravitational-wave interferometry: LIGO uses related phase-measurement principles on kilometre-scale arms to detect spacetime strain.
- Metrology: the metre is now defined through the exact speed of light, reversing the historical problem of measuring c relative to a hypothetical medium.
Evidence Boundaries
The simple equations above represent a stationary ether with Galilean velocity addition and equal geometric arms. Real optical media, refractive indices, apparatus deformation and source coherence require detailed treatment. Historically, the 1887 null result was one important constraint among many developments in electrodynamics; it should not be presented as a single-experiment proof of special relativity or as proof that Einstein derived his theory directly from this one result.
Manual Summary — KNOW → CONNECT → EXPLAIN → APPLY → CHECK
- KNOW: an interferometer turns optical path difference into phase/fringe shift.
- CONNECT: a stationary ether predicts orientation-dependent round-trip times.
- EXPLAIN: rotation should modulate the fringe pattern if the simple ether wind exists.
- APPLY: predict how path length, wavelength and assumed v change sensitivity.
- CHECK: demonstrate sensitivity, control thermal/mechanical drift and distinguish model constraint from theory proof.
Teaching Guide for Parents, Tutors and Teachers
Why this opening works: students often think experiments matter only when they detect something. A famous non-detection forces them to understand prediction, sensitivity and falsification.
- Central reasoning model: hidden medium → directional travel-time prediction → interference → rotate → expected modulation → null constraint.
- Teaching sequence: wave-medium intuition → beam splitting → phase → classical parallel/perpendicular calculation → rotation → null result → alternative theories → special relativity boundary.
- Diagnostic question: “How do you know zero fringe shift means anything if the instrument might simply be insensitive?”
- If stuck: use a swimmer crossing a moving river for the classical perpendicular-path geometry, then translate the time difference into phase.
- Ready for more: derive Lorentz transformations, cavity tests and modern limits on Lorentz violation.
Quiet Teaching Standard: do not teach “Michelson–Morley disproved ether and proved Einstein” as a slogan. Require the learner to state the precise ether model tested, the predicted fringe signature, why the null result was meaningful, and why theory choice still required a broader evidence network.