Contents

A Review of Airy's 'Failure'

An Aether Cosmology Research Group Presentation
Dedication
Malcolm Bowden

Malcolm Bowden

Civil engineer · Creationist author

Malcolm produced the earliest and longest-standing video presentation of Airy's experiment as a proof of geocentricity, sharing his work from 2010. This review stands on the trail he cut.

Michelson & Morley, 1887 — the crucial experiment

Michelson's Historical Recap

A 155-year-old unsolved problem in heliocentric cosmology.
  • Michelson recounts stellar aberration and Airy's water-telescope null, the very experiment we are examining.
  • Michelson and Morley then built an interferometer to catch the Earth's motion through the ether directly, by the shift of its interference fringes.
  • A fringe is one full light-and-dark band. The Earth's 30 km/s orbit should slide the pattern by 0.4 fringe, nearly half a band.
  • The instrument could resolve about 0.004 fringe, a hundredth of the expected shift.
Michelson & Morley 1887 opening: aberration and the Airy water-telescope null
MMX derivation: fringe displacement 2D v²/V² = 0.04 fringe
Michelson, A. A. & Morley, E. W. (1887). On the relative motion of the Earth and the luminiferous ether. American Journal of Science, 34, 333–345.
Michelson & Morley, 1887 — the measurement

A small signal, not a clean zero

  • The observed shift was tiny, under 0.01 fringe, less than a fortieth of the 0.4 predicted.
  • Because the shift grows as v², that puts the velocity at roughly a sixth to a quarter of the orbital speed, about 6 to 8 km/s, not the full 30, and not a clean zero.
  • The textbook "null result" overstates it: the data carried a small residual, right at the instrument's floor.
  • Their own conclusion: the result is small enough "entirely to refute Fresnel's explanation of aberration", and if the ether is at rest at the Earth's surface, Lorentz's own theory "also fails."
  • Consoli and Pluchino (2023) reanalyze the ±0.004-fringe records and read the residual as a real signal.
MMX 1887 noon and evening observations and Figure 6, displacement about 0.01 wavelength
MMX conclusion: relative velocity less than a sixth, certainly less than a quarter, of orbital
Consoli & Pluchino 2023: MMX precision about ±0.004 of a fringe
Consoli, M. & Pluchino, A. (2023). Michelson-Morley experiments: at the crossroads of relativity, cosmology and quantum physics. Int. J. Mod. Phys. A, 38(35n36), 2330017.
The option left off the table

One reading they would not take

Every ether theory on offer had failed in their own words: Fresnel refuted, Stokes incompatible, Lorentz's own fix collapsed. Yet one reading was never entertained.
  • The near-null sat with the simplest reading of all: no relative motion — the Earth at rest.
  • The Royal Society and the Royal Astronomical Society hold the Copernican principle as a starting premise, not a result to be tested.
  • So a stationary Earth was excluded before the experiment began. The search turned instead to ever more elaborate ways to hide a motion that would not appear.
"It is evident that in the post-Copernican era of human history, no well-informed and rational person can imagine that the Earth occupies a unique position in the universe."
Rowan-Robinson, M. (1996). Cosmology, 3rd ed. Oxford University Press, pp. 62–63. ISBN 978-0-19-851884-6. Quoted in Wikipedia, "Copernican principle."
Wikipedia, Copernican principle: the Rowan-Robinson quote in context
Section One

Stellar Aberration

The history and significance of stellar aberration in heliocentric cosmology.
Bradley, 1728

Starlight Drift

  • Bradley took zenith measurements of γ Draconis as it transited his telescope.
  • Across the year the star drifted about 40″ north to south, farthest north in September and farthest south in March.
  • Half of that swing, 20″.2, is the aberration constant.
  • Aberration is a first-order effect: it scales with v/c, not v²/c².
Bradley's zenith sector, the sector as mounted on its spindle, engraved plate
Bradley's zenith sector installed, an observer seated beneath the vertical tube looking up toward the zenith
Bradley's 12.5-ft zenith sector, built by George Graham (1727). The observer reclines beneath a near-vertical tube to read stars at the zenith, where refraction vanishes.
The stakes

The only first-order measurement

  • Kepler's kinematics and Newton's dynamics derive from one another. The mathematics alone cannot decide the model.
  • Aberration is the one measurement that returns a first-order v/c relationship giving 30 km/s.
  • From 1812 (Arago) to 1887 (Michelson), every further attempt at a first-order measurement returned null.
  • The nulls were explained by ad-hoc aether constructions built to cancel exactly what was being looked for.
  • The absence of a measurable velocity became the argument for the velocity.
Klinkerfues, 1867

Dynamic Consequences of Earth's Motion

If aberration is the ratio of the observer's velocity to the velocity of light, then slowing the light inside the instrument must change the angle.
"Whatever direction the motion of the earth may have, and in what proportion the Aether may take part in it, such an appearance as the direction of the ray will not be affected by it."
tan(θ) = v / .75c water: c / n = .75c → 27.33″ internal 36.35″ read

The moving frame worked out frame by frame in Rosser's Figure 4.3 (right). See it live in our simulator — toggle Simulation above and pick a theory.

Rosser Figure 4.3: telescope inclination in the heliocentric frame Σ and the geocentric frame Σ′
Klinkerfues, W. (1867). Die Aberration der Fixsterne nach der Wellentheorie, p. 40. · Rosser, W. G. V. (1964). Introduction to the Theory of Relativity, §4.4, Figure 4.3.
Greenwich, 1871

The instrument

  • A zenith sector: fixed vertical, double spirit-levelled, not pointable.
  • Rotated 180° between readings: the star is measured from both sides and averaged, cancelling instrument error to about 0.8″.
  • The micrometer reads the star's shift on the plate: about 3.5 µm per arcsecond of sky. He measures from both sides and averages.
  • Light path 37.0 in (0.9 glass + 35.3 water + 0.8 air); the micrometer is scaled to its 27.8 in air-equivalent (reduced) length by dividing each component by its refractive index.
  • Observed at 6-month intervals, sampling opposite points of Earth's assumed motion.
Airy's water telescope, engineering plate, figures 1 to 9 from Greenwich Observations 1871
The surviving Airy water telescope instrument
Figure 7, the micrometer wire-plate detail
Airy, G. B. (1871). On a supposed alteration in the amount of astronomical aberration of light, produced by the passage of the light through a considerable thickness of refracting medium. Proceedings of the Royal Society of London, 20, 35–39.
Instrumentation

Micrometer & Scale

The water plate is calibrated n times finer. That scaling is why a smaller internal angle reads the same arc.
θ_int = 15.45″ inside the water n = phys_L / eff_L = 37.0 / 27.8 = 1.33 reading = θ_int × n = 15.45″ × 1.33 = 20.55″
Reading

The celestial arc the telescope reports: the star's sideways drift on the wire plate divided by that telescope's own scale. Airy's 20.55″ is a reading, not a direct angle.

Internal angle θ_int

The physical tilt the ray takes inside the tube. Never measured directly. It is the angle required to produce the observed reading. In water, a 20.55″ reading needs 15.45″ inside.

Plate scale

1″ of sky = 3.45 µm on the plate. The micrometer reads the star's drift in µm; Airy's scale turns that displacement into celestial arc.

Later, the analysis gives a 15″ internal angle. That is correct: the micrometer recovers the measured 20.55″ by converting its plate displacement into sky arc.

Airy, G. B. (1871). On a supposed alteration in the amount of astronomical aberration of light. Proceedings of the Royal Society of London, 20, 35–39.
The result

Nothing changed

Every theory's internal angle and the celestial arc it would read, against the measured 20.55″.
Modelθ_int (internal) reads (celestial arc)speed in watervs 20.55″
Air telescope — measured20.55″20.55″cbaseline
Water telescope — measured—20.55″—no change
Emission (corpuscular)15.45″20.55″1.33 cangle ✓, speed refuted (Foucault 0.75 c)
Undulatory / Snell / Klinkerfues 186727.33″36.35″0.75 c✗ fails by n²
SR / Pauli / Rosser / Ligabue — Earth moving11.88″15.80″0.75 c✗ predicted
Jones 1972 — transverse drag, MEASURED11.88″15.80″0.75 cdrag confirmed to 0.02% → 15.80″, not 20.55″
SR / Pauli / Rosser / Ligabue — Earth at rest15.45″20.55″0.75 c✓
Geocentric — Earth at rest15.45″20.55″0.75 c✓

Readings six months apart differed by 0.8″; the instrument resolves 2 to 3″ easily, and the predicted change was 15″. The wave prediction failed by a factor of n², far outside its resolution.

Airy, G. B. (1871). On a supposed alteration in the amount of astronomical aberration of light. Proceedings of the Royal Society of London, 20, 35–39. · Every value derived in the ACRG spreadsheet ↗
Section Two

The Predictions

Each theory's own math, worked to the single angle it lands on.
Emission theory — the prediction

Right angle, impossible speed

|u| = n·c = 1.33c corpuscle speeds up in water u‖ = v u⊥ = √( (n·c)² − v² ) tan θ_int = u‖ / u⊥ = β/n → θ_int = 15.45″ reads = θ_int · n = 20.55″ ✓ matches Airy
  • Tangential velocity is conserved, so the angle lands right, 20.55″.
  • But it demands light travel at n·c = 1.33c in the water. Foucault measured 0.75c.
  • Michelson: to rescue it you must assume "the motion of the water carries the ray in the opposite direction" — absurd, in print.
Emission theory prediction, full step-by-step derivation in the ACRG spreadsheet
Michelson & Morley (1887), footnote to p. 333. · Full derivation in the ACRG spreadsheet ↗
Wave theory — the prediction

Four routes, one wrong answer

Slow the light to c/n while the tube keeps moving. Every wave-theory route lands on the same magnified angle.
Treatmentθ_int formulaθ_int readsvs 20.55″
Undulatory (wave)tan θ_int = n·β27.33″36.35″✗
Snell (moving tube)tan θ_int = n·β27.33″36.35″✗
Snell × γ (velocity comp.)tan θ_int = γ·n·β27.33″36.35″✗ (γ−1 ≈ 5×10⁻⁹)
Klinkerfues 1867θ_int = nα27.33″36.35″✗ (n²α)
Four different routes, one answer: 36.35″, off from the measured 20.55″ by a full factor of n². Michelson: it "should be four-thirds of its true value."
Relativity — the prediction

The Lorentz transform, both frames

Pauli (311b) / Rosser (4.30): tan α = sin α′·√(1−β²) / (cos α′ + β·w′/c) water at rest: θ_int = β/n = 15.45″ → reads 20.55″ ✓ water moving: θ_int = β(n − 1/n) = 11.88″ → reads 15.80″ ✗
  • Transform to the frame where the water is at rest: Snell gives 15.45″, reads 20.55″. Matches.
  • That rest frame is dynamically the stationary Earth.
  • Carry the same transform into the moving frame and it gives 11.88″ → 15.80″. Never matches.
water at rest
15.45″
reads 20.55″ ✓
water moving
11.88″
reads 15.80″ ✗
Pauli, W. (1921). Theory of Relativity, §36γ, eq. (311b). · Rosser, W. G. V. (1964). Introduction to the Theory of Relativity, §4.4, eq. (4.30).
Jones, 1972 — the experimental answer

We know exactly how the ray behaves

The transverse drag every explanation leans on is not an assumption. R. V. Jones measured it directly, in Airy's geometry, to 0.02%.
θ = β(n − 1/n) = 11.88″ δ = (v/c)·L·(n − 1/n) measured 6.174 nm vs Fresnel 6.175 nm
Jones's transverse-drag equation, δ = rωt(n − 1/n)/c, and the Airy inference
Jones's apparatus: light through a rotating glass disk, source, detector, mirror
Jones Figure 6: the measured deflexion trace, reversing with disk direction
Jones results, measured 6.174 nm vs expected 6.175 nm

In a moving medium the ray drifts by exactly β(n − 1/n) = 11.88″, reading 15.80″, not Airy's 20.55″. Read as a speed, that moving frame demands 1.73c; Jones's measured drag needs only the real 0.75c.

Jones, R. V. (1972). 'Fresnel aether drag' in a transversely moving medium. Proc. R. Soc. Lond. A, 328, 337–352. · Apparatus schematic: skullsinthestars
Jones, 1972 — the last escape hatch

It matched the phase velocity

  • The remaining objection is that the drag speed is ambiguous, phase velocity or group velocity — the distinction Ligabue leans on (right). Jones's numbers resolve it.
  • Phase velocity predicts 6.175 nm; group velocity would be ~1.5% larger, 6.27 nm.
  • He measured 6.174 nm — squarely the phase-velocity value.
  • So the drag is fixed by the ordinary refractive index at the phase speed. The ambiguity the "timing" argument depends on is closed by direct measurement.
Ligabue: the timing objection, that the speed of light in moving water is not knowable
Jones p.351: the measured drag agrees with the phase velocity, not the group velocity
Jones, R. V. (1972). 'Fresnel aether drag' in a transversely moving medium. Proceedings of the Royal Society of London A, 328, 337–352, at p. 351.
Section Three

The Transform

The relativistic account, the frame it requires, and the frame it never shows.
The resolution offered

A frame where the water is at rest

  • The Lorentz transformation is applied in the frame where the water is stationary.
  • In that frame the starlight arrives already aberrated, refracts by Snell, and reads 20.55″.
  • That frame is the one in which the Earth carrying the water is at rest.
  • The transform's lineage runs Voigt (1887) → Lorentz → Einstein, built inside aether theory to make every experiment come out as if the Earth were at rest.
A device engineered to produce nulls, later presented as the discovery that explains them.
The omission

What the moving frame predicts

For 155 years, published treatments stop at the water-at-rest frame.
"The law of refraction at moving boundary surfaces can also be obtained from the stationary case, by means of a Lorentz transformation, but it leads to complicated formulae … if it is observed from the rest system, Airy's result is self-evident."
Pauli, W. (1921). Theory of Relativity, §36γ, p. 114.
Water at rest
15.45″
reads 20.55″ ✓
Water in motion
11.88″
reads 15.80″ ✗
Pauli p.114: the moving-boundary law leads to complicated formulae; only the rest system is worked
We are supposed to be the moving frame. Pauli shows only the rest frame — the one that matches.
The trap

Frames must agree

  • Reciprocity between frames is the backbone of the theory. It is how invariance is validated.
  • Every frame must agree on what the instrument reads.
  • Two different answers means you can distinguish between frames.
The escape, and its cost

"We do not live in that frame, so it does not matter" abandons the framework entirely. If the analysis is allowed to disagree with the lab, invariance is gone.

Ligabue's derivation: the medium's frame gives −15″, the Sun's frame gives β(n−1/n); the two frames disagree
Either the moving frame reproduces 20.55″, or the theory distinguishes frames. Pick one.
Closing the escape hatch

How fast is light in water?

Faster · n·c
1.33 c
superluminal, refuted by Foucault 1850
Slower · c/n
0.75 c
measured, still and moving water
"Cannot be known"
—
measured for 150 years
  • Feed the real 0.75c into the naive aberration formula and it returns 27.33″. To get 15″ you need 1.33c — not a speed, the refractive index in disguise.
  • Ligabue's own moving-frame derivation (right) admits the speed "will not be c/1.33" — so not the real 0.75c either. The moving frame manufactures an unphysical speed.
  • The 15.45″ comes from refraction, not from a light speed. That is why the speed does not move it.
Ligabue's Airy's-case derivation: the moving-frame speed will not be c/1.33
Section Four

The Verdict

Every model of light propagation, measured against one instrument reading.
The verdict

Heliocentric fails, Geocentric holds

Modelθ_intreads read as a speedverdict
Heliocentric — the Earth is moving
Emission (corpuscular)15.45″20.55″needs 1.33c (real)speed refuted, Foucault 0.75c
Undulatory / Snell / Klinkerfues27.33″36.35″0.75cwrong angle, fails by n²
SR / Pauli / Rosser / Ligabue (moving)11.88″15.80″implies 1.73cnever reaches 20.55″
Jones 1972 — the drag, MEASURED11.88″15.80″0.75c measuredconfirms 15.80″, not 20.55″
Geocentric — the Earth is at rest
SR / Pauli / Rosser — rest frame15.45″20.55″0.75c (real)matches — but this frame is the stationary Earth
Snell refraction, single frame15.45″20.55″0.75c (real)matches, real speed, one frame

The Jones row is not a measurement of aberration in water — Airy made that, and got 20.55″. Jones measured the transverse drag itself, which forces the moving frame to 11.88″ / 15.80″. Those are the aberration values a moving Earth ought to produce, now backed by experiment, and they are not what Airy read.

Every moving-Earth account needs a superluminal speed or the wrong angle. Only the Earth at rest, at the real 0.75c, reproduces 20.55″.
Conclusion

There is no model that gives 20.55″ twice for a moving Earth

However you treat light — corpuscles, waves, wave packets, by phase velocity or group velocity — no propagation model predicts 20.55″ in air and 20.55″ in water for an Earth in motion.
  • Relativity reaches the measured value only in the water-at-rest frame.
  • That frame is dynamically the stationary Earth.
  • Emission theory fails on speed. The wave theory fails on angle. The moving frame fails on both.
If experiment is to match theory, the Earth must be dynamically at rest.
Rebuttal — Franco's timing objection

We do know how light behaves in a moving medium

"'Timing' reasonings are incorrect, since they unjustifiedly assume the direction and the value of speed of light in water in the sun's frame (where water is moving)." — F. Ligabue
  • Not assumed — measured. Snell (refraction), Fizeau and Fresnel (longitudinal drag), and Jones (transverse drag) all pin down light in the moving medium.
  • A time-of-flight measurement, a single-photon stopwatch, reads light taking longer through water. That is c/n = 0.75c by the clock, not a phase-velocity artifact.
  • Which kills the "c = c by emission / zig-zag" red herring: slowed wave or longer zig-zag path, either picture gives the same real timing delay. Light is slower in the medium, full stop.
Ligabue: the claim that the speed of light in moving water is unknown in the Sun's frame
Ligabue, F. (2025). Airy's experiment and relativity.
Franco's own derivation

He derived it, and so did the critics

  • Ligabue's own moving-frame derivation gives θ′₂ ≈ θ₂ + βn = β(n − 1/n) = 11.88″, reading 15.80″ — and he notes the speed "will not be c/1.33."
  • On stream, Toon and Cody ran the same Lorentz transform and landed on the same θ₂ ≈ 12″ / 15.80″.
  • Everyone who does the moving-frame math, defender or critic, arrives at the same wrong answer for a moving Earth. The disagreement is only about whether to look.

On stream, Toon & Cody work the same transform and arrive at ~12″ — watch at 56:48 ↗

Ligabue's derivation: the moving-frame drag θ′₂ = β(n − 1/n), speed not c/1.33
Ligabue, F. (2025). Airy's experiment and relativity — the moving-frame derivation.
Rebuttal — "the Sun frame doesn't count"

You cannot discard the Sun frame

  • The objection: "we don't live in the Sun frame, so its 15.80″ prediction doesn't count."
  • But the same figures apply to the Earth frame — Pauli states it outright and Rosser draws both frames in Figure 4.3. By covariance, all frames are indistinguishable and must agree on what the instrument reads.
  • The Sun frame giving 15.80″ while the Earth frame reads 20.55″ is not a frame you can wave away — it is a reciprocity violation, the theory contradicting itself.
  • And the double standard is fatal: if Sun-frame analysis is void because no one was there to measure it, then the entire heliocentric ephemeris — computed in that same unmeasured Sun frame — goes out with it.
Throw out the Sun frame and you throw out heliocentrism's own bookkeeping.
Sources

References

Primary papers
  • Bradley, J. (1728). A new apparent motion discovered in the fixed stars. Phil. Trans., 35, 637–661.
  • Foucault, L. (1850). Méthode générale pour mesurer la vitesse de la lumière. C. R. Acad. Sci., 30, 551–560.
  • Klinkerfues, W. (1867). Die Aberration der Fixsterne nach der Wellentheorie.
  • Airy, G. B. (1871). On a supposed alteration in the amount of astronomical aberration of light. Proc. R. Soc. Lond., 20, 35–39.
  • Michelson, A. A. & Morley, E. W. (1887). On the relative motion of the Earth and the luminiferous ether. Am. J. Sci., 34, 333–345.
  • Jones, R. V. (1972). 'Fresnel aether drag' in a transversely moving medium. Proc. R. Soc. Lond. A, 328, 337–352.
Analysis, video & data
  • Pauli, W. (1921). Theory of Relativity, §36γ. · Rosser, W. G. V. (1964). Introduction to the Theory of Relativity, §4.4.
  • Rowan-Robinson, M. (1996). Cosmology, 3rd ed. Oxford University Press.
  • Consoli, M. & Pluchino, A. (2023). Michelson-Morley experiments. Int. J. Mod. Phys. A, 38(35n36), 2330017.
  • Ligabue, F. (2025). Airy's experiment and relativity — the moving-frame derivation.
  • Bowden, M. — Airy's Failure · Science Mafia · channel
  • ACRG — spreadsheet ↗ · interactive simulator ↗
Aether Cosmology Research Group

Thank you

Interactive simulation, full spreadsheet, and every derivation are open. Check the numbers yourself.
alanspaceaudits.github.io/airy_sim · discord.gg/aethercosmology
Arrows or swipe to navigate · L for contents · S for sim · F fullscreen