Limits of Newtonian Gravity
For two centuries, Newton's law of universal gravitation predicted planetary motion correctly. Then it didn't: Mercury's orbit shifted by an amount the theory couldn't account for. This chapter follows that discrepancy, and two others, through the late nineteenth century.
Table of Contents
Introduction
The previous chapter ended with a bargain. Newton could not say what gravity is, or how a body reaches across empty space to pull on another; he declined to invent an answer, and the demand for one eventually faded. The objection was never met—see Chapter 3, §6. It was outvoted by results.
For a century and a half the results kept coming. Comets returned on schedule. Tides rose and fell as predicted. The Earth turned out to be the shape Newton said it should be. And when Uranus strayed from its calculated path, a few—Airy among them—wondered aloud whether the inverse-square law simply gave out at that distance. Most did not. They used the law to predict a new planet, and Neptune appeared where the equations said it would, which settled the argument for the next half-century.
This was the zenith of Newtonian physics: a universe governed by a single, elegant mathematical law. If you knew the positions and velocities of all bodies at one instant, you could—in principle—calculate the future (and past) of the cosmos with absolute certainty. Laplace said as much, and meant it literally: the universe was a mechanism, and its equations were now known.
By the late 1800s, three discrepancies had accumulated. Mercury's orbit precessed faster than Newtonian gravity predicted. Light traveled at a fixed, finite speed, which sat awkwardly with a theory of instantaneous gravitational action. And a widely repeated experiment failed to detect the medium light was assumed to travel through.
This chapter tells the story of how success bred crisis. Newton's theory was so powerful that its failures, when they finally appeared, couldn't be dismissed as measurement errors. They demanded a complete rethinking of space, time, and gravity itself—a revolution that would culminate in Einstein's General Relativity.
The Crumbling of the Clockwork
The Man Who Found a Planet with a Pen
Urbain Le Verrier
(1811 - 1877)Master of celestial mechanics, and from 1854 Director of the Paris Observatory. In 1846 he was not yet running the place—he was the man who found Neptune, as the phrase went, at the tip of his pen.
His method was to trust Newton's laws absolutely. If a planet moved oddly, it wasn't the law that was wrong—it was the map of the solar system.
The year was 1846. The French mathematician Urbain Le Verrier had spent months on the irregularities in the orbit of Uranus, and concluded that an unseen planet was pulling it off course. He worked out where that planet had to be, and wrote to an observatory that would go and look.
That night, they looked. And there it was: Neptune, within one degree of the predicted position. A planet nobody had seen had been located with pen and paper, and then found where the paper said it would be.
The Letter
The popular version of the story compresses Le Verrier's instruction to a single line—look there, and you will see it. What he actually sent to Berlin on 18 September 1846 was a working observer's brief: a position on the ecliptic, a tolerance, and a physical detail to identify the object by.
"Direct your telescope to a point on the ecliptic in the constellation of Aquarius, in longitude 326°, and you will find within a degree of that place a new planet, looking like a star of about the ninth magnitude and having a perceptible disc."
The letter reached the Berlin Observatory on 23 September. Johann Gottfried Galle took it to the director, Johann Franz Encke, who was reluctant but let him use the telescope that night—it was Encke's birthday. Galle worked with a student assistant, Heinrich Louis d'Arrest, and d'Arrest made the suggestion that decided the matter: rather than sweep the sky, compare it against Carl Bremiker's brand-new star chart of that exact region, which the observatory happened to possess and almost nobody else did.
Galle called out positions; d'Arrest checked the chart. Within an hour they had a star of the eighth magnitude that was not on it. The following night it had moved. It stood 55 arcminutes from Le Verrier's predicted longitude—less than one degree, exactly as promised.
The Englishman Who Nearly Got There First
The credit was disputed within weeks, and the dispute was national. In Cambridge, a young mathematician named John Couch Adams (1819–1892) had attacked the same problem from 1843, and by September 1845 had a solution of his own—arrived at independently, and earlier than Le Verrier's published one.
It went nowhere. Adams took his figures to George Biddell Airy, the Astronomer Royal, who replied with a technical question about the radius vector of Uranus that Adams, for reasons never satisfactorily explained, did not answer. When Le Verrier's results appeared in 1846 and Airy saw how closely the two predictions agreed, he asked James Challis at the Cambridge Observatory to begin a search. Challis had no up-to-date chart of the region, so he was reduced to mapping every star in a wide field and re-mapping it later to see what had moved. He recorded Neptune on 4 August and again on 12 August 1846 and did not notice, because he had not yet reduced his own observations.
The Anglo-French row that followed was bitter, and it settled into a compromise—joint credit to Adams and Le Verrier—that suited the diplomats of both countries. Historians have been less generous since. The Royal Greenwich Observatory papers on the affair went missing for decades and resurfaced only in 1998, after the death of the astronomer who had removed them; the reassessment that followed showed Adams's predicted position drifting substantially between versions, and none of it published. He had a calculation. Le Verrier had a calculation, a publication, and the decisive instinct to write to an observatory that owned the right chart.
Four names, one planet
Le Verrier computed the position and sent it to someone who would act on it. Adams computed a position and told almost no one. Airy and Challis held the prediction and the telescope and were too slow. Galle and d'Arrest pointed the instrument and looked.
Neptune had in fact been seen before. Galileo recorded it twice, in December 1612 and January 1613, close to Jupiter, and noted that one "fixed star" appeared to have shifted. He did not follow it up.
The Observers Behind the Calculation
A prediction of this kind is only as good as the data it is fitted to. Le Verrier's calculation rests on two centuries of patient positional astronomy—generations of observers recording where things were, night after night, and checking that Newtonian gravity held not only for planets but for comets, moons, and asteroids. Caroline Herschel was one of the most remarkable of them.
Caroline Herschel
(1750–1848)Caroline Lucretia Herschel was born in Hanover in 1750 and followed her brother William to England in 1772, intending to train as a singer. She was drawn instead into grinding mirrors, building reflecting telescopes, and observing; after William discovered Uranus in 1781, astronomy became the whole of both their lives.
In 1787 George III granted her £50 a year as William's astronomical assistant. It made her the first woman known to have been paid for scientific work, and the first woman in England to hold a salaried government post.
Comets, and the data behind them
Between 1786 and 1797 she found eight comets—the first woman to discover one. The best known is 35P/Herschel-Rigollet (1788), a periodic comet still carrying her name. What mattered for gravitation was less the discoveries than the positions: night-by-night measurements of where each comet stood, which is exactly the input Newtonian celestial mechanics needs to fit an orbit.
This was the programme that had begun with the predicted return of Halley's comet in 1758 (see Chapter 3, §7): every comet whose path could be computed in advance was another test the inverse-square law passed. She was among the first women to publish under her own name in the Royal Society's Philosophical Transactions, beginning in 1787, with five of her comets reported there by 1796. She also found fourteen nebulae independently, and helped catalogue some 2,500 with William.
Catalogues and recognition
In 1798 she presented the Royal Society with a Catalogue of stars, taken from Mr. Flamsteed's observations: an index to Flamsteed's work, 560 stars omitted from the British Catalogue, and a list of that publication's errors. Her reduction of the Herschel nebulae became one of the foundations of the New General Catalogue, whose NGC numbers are still in use.
The Royal Astronomical Society gave her its Gold Medal in 1828, the first awarded to a woman; the next went to Vera Rubin in 1996, one hundred and sixty-eight years later. In 1835 she and Mary Somerville became the RAS's first women members. The King of Prussia sent her a Gold Medal for Science in 1846, when she was ninety-six. She died in Hanover in 1848, aged ninety-seven.
Her working life spans the high-water mark of Newtonian astronomy almost exactly: Uranus in 1781, precision observation pressed to its limits through the 1780s and 1790s, and—two years before her death—the perturbations of Uranus yielding Neptune. She helped build the observational base on which that prediction stood.
Main sources: Herschel, C. (1787), Phil. Trans. Roy. Soc. London 77, 1–3; Herschel, Mrs. J. (ed.), Memoir and Correspondence of Caroline Herschel (1876); MacTutor History of Mathematics; Royal Museums Greenwich; Herschel Archive, Royal Astronomical Society.
The Pebble in the Shoe
Le Verrier's Discovery (1859)
Understanding the Anomaly
Forty-three arcseconds per century is an almost invisible discrepancy—less than one-hundredth of a degree per century. To put it in perspective, it would take roughly 3 million years for this extra precession to complete a full circle.
And yet, it was real. It was measurable. And it was completely unexplained by Newton's inverse-square law. In a theory as precise and successful as Newtonian gravity, there is no room for almost.
The Hunt for Planet Vulcan
Newtonian Fixes That Didn't Work
Over the next half-century (1859–1915), physicists and astronomers proposed various modifications to Newtonian gravity to explain Mercury's precession:
- An undiscovered inner planet ("Vulcan"): Searched for intensively throughout the late 19th century, but never found. Total solar eclipses were ideal times to look, but systematic searches came up empty.
- A modified inverse-square law: Perhaps gravity falls off as F ∝ 1/r2+ε for a tiny ε? Asaph Hall proposed exactly this in 1894, with an exponent of 2.00000016, and it did fit Mercury. That was all it did: the value was chosen to fit, there was no independent reason for it, and Newcomb showed the same exponent misbehaved for the Moon.
- An oblate (non-spherical) Sun: If the Sun bulged at the equator, its non-spherical mass distribution might produce extra precession. But the bulge needed was far larger than anything observation would allow—by orders of magnitude, on modern figures—and a Sun that oblate would have shown itself in other ways.
- Interplanetary dust or a zodiacal cloud: Perhaps unseen matter near the Sun was tugging on Mercury. But the amount of mass required was implausibly large and would have produced other observable effects on sunlight and planetary motions.
Every proposed fix either violated other successful predictions or lacked independent evidence. By the early 20th century, the anomaly had become a persistent thorn in the side of Newtonian gravity—small enough to ignore for practical purposes, but troubling enough that it could not be entirely dismissed.
The Numbers: Where the 43 Arcseconds Come From
| Source of Precession | Arcseconds/Century |
|---|---|
| Gravitational pull from Venus | 277 |
| Gravitational pull from Earth | 90 |
| Gravitational pull from Jupiter | 153 |
| Gravitational pull from other planets | 11 |
| Total Newtonian prediction | 531 |
| Observed precession | 574 |
| Unexplained anomaly | 43 ✗ |
| Einstein's GR correction (1915) | 43 ✓ |
A note on what 574 means. Measured the way an observatory actually measures it—relative to the equinox, against a reference frame defined by the Earth's own axis, which is itself turning—Mercury's perihelion advances by roughly 5,600 arcseconds per century. Some 5,025 of that has nothing to do with Mercury at all: it is the precession of the equinoxes, the slow wobble of the Earth. Subtract the Earth's wobble and 574 is what is left, and that is the figure in the table: the advance against the fixed stars. Both numbers are correct; they answer different questions, and quoting one for the other is the commonest error in accounts of this problem.
Why This Tiny Number Mattered So Much
Forty-three arcseconds per century amounts to about 0.012 degrees per century—an incredibly tiny angular shift. Why did this matter so much to physicists?
Because it was precisely measured, reproducibly observed, and completely unexplained within Newtonian physics. Small discrepancies are only troubling when they resist every reasonable attempt at explanation. By 1915, Mercury's stubborn precession had withstood half a century of theoretical attacks.
This made it the perfect litmus test for any new theory of gravity. When Einstein's General Relativity predicted exactly 43 arcseconds per century of extra precession—with no adjustable parameters, no hidden planets, and no ad hoc modifications—the match was exact, with no free parameter left to tune.
Where This Goes
Newton's gravity accounted for the planets, the moons, the comets and the tides. What it could not account for was 43 arcseconds per century of Mercury's motion—a shift that would need three million years to work its way once round the orbit. That was not, at the time, enough to unseat anything. It sat in the tables for half a century as an untidy residual, and it was still sitting there in 1915, unexplained and unrepaired, when a theory built for other reasons entirely turned out to predict it.
Interactive Visualization: Mercury's Precession
Three orbits, drawn on top of each other. A single planet round a single Sun traces the same closed ellipse forever; add the other planets and the ellipse starts to turn, by 531 arcseconds a century; the observed figure is 574. Toggle the layers to separate the 43 arcseconds nobody could account for. The rates are exaggerated enormously, and so is the eccentricity—at true scale the whole anomaly is 0.16 degrees per century, which is nothing you could watch.
Mercury's Perihelion Precession
In 1859, Urbain Le Verrier discovered Mercury's perihelion advances 43 arcseconds per century faster than Newton's theory predicted. For 56 years this remained a mystery. In 1915, Einstein's General Relativity solved it perfectly—the first major triumph of GR.
Precession Budget: Where Does the 574"/century Come From?
The Key Insight: Newton's gravity (with planetary perturbations) explains 92.5% of Mercury's precession. But it predicted a closed Keplerian ellipse for an isolated system. The extra 7.5% from General Relativity was the missing piece that solved a 56-year mystery (1859-1915).
Compare directly: The blue Kepler orbit stays perfectly closed (no precession), while the red observed orbit creates a rosette pattern. The divergence grows with each orbit, making the 7.5% GR contribution visible after just a few orbits.
| Source | Contribution |
|---|---|
| Venus | 277" |
| Jupiter | 153" |
| Other planets | 101" |
| Total (Newtonian) | 531" |
| GR correction | +43" |
| Total with GR | 574" |
| Observed | ~574" |
Classical mechanics predicts 531" per century from planetary perturbations. However, observations show ~574" per century—43" more than expected. General Relativity adds exactly this extra shift, turning an anomaly into a precision test of Einstein's theory.
The Speed Limit of Reality
Ole Rømer
(1644 - 1710)Working at the Paris Observatory, he was the first to argue that light has a finite speed. He used the moons of Jupiter as a cosmic clock, noticing they were "late" when Earth was far away. His own figure was for the delay, not the speed: about 22 minutes for light to cross the Earth's orbit (the true value is 16.7). Huygens turned it into a velocity in 1678. Cassini, his own patron, argued against the conclusion, and the Paris Academy stayed unconvinced for decades.
While Le Verrier was hunting for Vulcan, a deeper crack in the foundation had already been exposed—not by looking at planets, but by looking at light.
In Newton's system gravity acts instantaneously: if the Sun vanished, the Earth would fly off into the dark in the same instant. In 1676 a Danish astronomer named Ole Rømer found the first evidence that something else in physics does not work that way.
Rømer noticed that eclipses of Jupiter's moon Io arrived "late" when Earth was far from Jupiter. His conclusion? Light has a finite speed.
Aha Moment: In Newton's world, eclipses would happen like clockwork. But Rømer saw they arrived late when Earth was far from Jupiter. The light had to travel an extra distance! Watch the delay accumulate in the graph below.
Schematic, not to scale. Orbit radii and angular speeds are pixel values chosen for visibility, not physical ratios: Jupiter's period here is 12.5× Earth's, but Kepler's third law would put the orbit-radius ratio at 12.52/3 ≈ 5.4, not the 2:1 shown. The light speed (100 px/s) is likewise an arbitrary visualization unit. The “eclipse” event fires when Io crosses a fixed angle in its orbit, not the actual Sun-Jupiter-Io shadow geometry.
Does Gravity Travel?
Bradley Closes the Case
Rømer's argument was indirect: it inferred a speed from a discrepancy in a table of eclipse timings, and it depended on an orbital diameter nobody knew accurately. Cassini, who had encouraged the work, turned against the conclusion, and the Paris Academy was not persuaded for a generation.
The proof arrived from an unexpected direction in 1728. James Bradley, hunting for stellar parallax—the annual wobble that would finally prove the Earth moves—found a wobble in γ Draconis of the wrong size and, decisively, the wrong phase. It peaked three months away from where parallax should peak. Bradley eventually saw why: the effect was not parallax at all but aberration. Rain falling vertically appears to slant when you run through it, and starlight appears displaced because the Earth is moving across the incoming beam.
The tilt is a right triangle: one side is the speed of light along the incoming ray, the other is the Earth's orbital speed across it, and the aberration angle α is the angle between them.
Light comes straight down at speed c; the telescope is carried sideways at the Earth's orbital speed v⊕, so it has to tip forward by α to keep starlight running down the tube — the same geometry as rain appearing to slant when you run through it. The angle is so small that tan α ≈ α in radians, so c ≈ v⊕ / α — the orbit's size in miles never enters.
Diagram angle massively exaggerated for visibility (drawn at about 24°); the real α is 20.5 arcseconds, roughly 0.006°.
Bradley had also, in passing, produced the first direct physical evidence that the Earth really orbits the Sun, ninety-five years after Galileo's trial.
Then What About Gravity?
Once light is known to take time, the obvious question is whether gravity does. Chapter 3 laid out why Newton's contemporaries hated instantaneous action at a distance as philosophy—see §6, Action at a Distance. What changed after Bradley is that the question became answerable by measurement.
Pierre-Simon Laplace took it up in the Mécanique céleste (1805). His reasoning was the aberration argument again, transposed to gravity. If the gravitational pull of the Sun propagates at finite speed, then the Earth is attracted not toward where the Sun is but toward where it was. That misalignment puts a small component of the force along the direction of motion rather than across it, and a force with a component along the motion does work: the orbit would gain energy and spiral outward, secularly and cumulatively.
The Moon's motion has been recorded for two millennia and shows no such secular acceleration. Laplace turned the absence into a bound, and the bound was staggering: gravity, if it propagates at all, must do so at least some seven million times the speed of light. For practical purposes, instantaneously.
Newton was not being lazy
It is tempting to read instantaneous gravity as an eighteenth-century failure of nerve—a placeholder nobody bothered to replace. Laplace shows the opposite. Finite propagation was tested, and the observations came back against it by seven orders of magnitude.
Which is what makes the eventual resolution so strange. General relativity does propagate gravity at exactly c, and it survives Laplace's test anyway: in the full theory the velocity-dependent terms very nearly cancel the aberration, so a body in a steady orbit is pulled almost exactly toward the instantaneous position of its companion even though nothing travelled faster than light.
The residue that does not cancel is not zero. It is the radiated part—gravitational waves—and measuring it took until 1974.
The Most Famous Failed Experiment
Michelson & Morley
(1887)Albert Michelson (Nobel 1907) and Edward Morley. Their 1887 interferometer was the most sensitive instrument of its kind then built, and it was built to find the Aether. Instead, they found nothing — no shift attributable to Earth's motion through the aether. The null result had no explanation within Newtonian physics; Einstein's 1905 paper gave it one.
If light has a finite speed, it must travel through something, right? Sound needs air; waves need water. Physicists called this hypothetical medium the Luminiferous Aether.
In 1887, Albert Michelson and Edward Morley built a device to detect the "wind" of this Aether as Earth moved through space. They expected the speed of light to change depending on the direction.
The Michelson-Morley Experiment (1887)
The interferometer floats on mercury and can rotate freely. If the Ether exists, rotating should shift the interference pattern. It never did.
Playback Controls
Parameters
Orientation of the interferometer. Rotate to test for ether wind effects
Display Options
Current Data
Hypothesis Selector
💡 The Key Insight
If the Ether exists: As you rotate the apparatus, the two arms change their orientation relative to the "Ether wind." This should cause the interference pattern to shift left and right as the light takes different times to travel each arm.
What actually happened: No matter how they rotated the table, the fringes never moved. The speed of light was the same in all directions. The Ether doesn't exist.
This "null result" was one of the most important negative results in physics history. It paved the way for Einstein's Special Relativity (1905).
What the Null Result Meant
Two Attempts, Not One
Michelson tried first in Potsdam in 1881, on money from Alexander Graham Bell. The result was null, but the apparatus was barely sensitive enough to say so, and the analysis was wrong: Michelson had neglected the effect of the Earth's motion on the transverse arm, which halved the fringe shift he should have been looking for. Alfred Potier pointed this out to him in Paris the same year, and Hendrik Lorentz published the corrected treatment in 1886. The expected signal was smaller than Michelson had thought, and his instrument had not really been able to see it.
The 1887 experiment at Cleveland, with Edward Morley, was built to remove that excuse. Multiple reflections folded the light path out to about eleven metres. The whole optical bench sat on a sandstone slab floating in a trough of mercury, so it could be rotated smoothly through every orientation without flexing. The runs that made the experiment famous were taken over four days in July 1887, at noon and in the evening, so that the Earth's rotation would swing the apparatus through a range of orientations relative to its orbital motion. Michelson and Morley meant to repeat the whole thing at three-month intervals, and never did.
The fringes did not move—or rather, they moved by less than a fortieth of what the aether theory predicted, which was consistent with nothing at all.
Saving the Aether
The reaction was not to abandon the aether. It was to ask what could hide it. In 1889 George FitzGerald, and independently in 1892 Lorentz, proposed the same escape: matter moving through the aether is physically compressed along the direction of motion, by the factor
√(1 − v²/c²)
The arm pointing into the aether wind gets shorter by exactly the amount needed to compensate for the light's slower round trip. The interferometer cannot detect its own motion because the contraction and the delay cancel, precisely, at every speed.
It works. That was the problem. Henri Poincaré objected that a hypothesis invented for the sole purpose of cancelling one experiment, with no independent evidence and no other consequence, explains nothing—and that nature seemed to be running a conspiracy to keep the aether hidden. A theory requiring a fresh conspiracy for each new experiment is a theory in trouble.
The same formula, twice
In 1905 Einstein derives length contraction—the identical factor, √(1 − v²/c²)—from two postulates about how measurement works, with no aether anywhere in the argument.
Lorentz had the equations before Einstein had them. What he did not have was a reason for them. In his hands the contraction was a mechanical accident befalling rods that plough through a medium; in Einstein's it is a statement about space and time themselves, which is why it applies to everything and needs no medium to apply to.
The equations were right; the mechanism was wrong. The next chapter replaces Lorentz's aether with Einstein's postulates about space and time.
Kelvin's Two Clouds
27 April 1900
Lord Kelvin—William Thomson, then seventy-six and the most decorated physicist in Britain—rose at the Royal Institution to deliver a lecture with a deliberately unglamorous title: Nineteenth-Century Clouds over the Dynamical Theory of Heat and Light.
The dynamical theory he meant was the great synthesis of the century: matter made of moving particles, light made of waves in an elastic medium, both governed by mechanics. Kelvin's point was that this edifice, magnificent as it was, had exactly two things wrong with it. He named them.
The first cloud was the motion of the Earth through the luminiferous aether. Light is a wave; waves need a medium; the medium must be there; nobody could detect the slightest trace of the Earth's motion relative to it. Kelvin discussed Michelson-Morley explicitly, and discussed the FitzGerald-Lorentz contraction as the proposed way out, and was not comfortable with it.
The second cloud was the equipartition of energy—the Maxwell-Boltzmann doctrine that energy distributes itself equally among all the available degrees of freedom of a system. It is a theorem of classical statistical mechanics, and it gives systematically wrong answers for the specific heats of gases and solids, and catastrophically wrong ones for radiation.
Where the clouds went
Cloud 1 → special relativity. Five years later Einstein removed the aether rather than the discrepancy, and the first cloud dissolved along with the medium it was about.
Cloud 2 → quantum theory. Equipartition fails because energy is not infinitely divisible among modes. Planck had already, in December 1900, been forced into quantised oscillators to fit the blackbody curve; it took two more decades to see what that meant.
Two clouds, two revolutions, and between them essentially the whole of twentieth-century physics. Kelvin's aim was poor in one respect only: he thought they were clouds, not weather systems.
What Kelvin Did Not Say
Mercury was not one of the clouds. The 43 arcseconds were well known in 1900 and Kelvin does not mention them, which is itself informative: to a nineteenth-century physicist an unexplained residual in a planetary orbit was a problem for celestial mechanics—a missing planet, an oblate Sun, a tweak to the exponent—and not evidence against the framework. It became a crisis retrospectively, once something else explained it.
Nor did Kelvin ever announce that physics was finished. The sentence attributed to him—that there is nothing new to be discovered, and all that remains is more and more precise measurement—has never been traced to anything he wrote or said, and it directly contradicts the lecture he actually gave.
Myth: "nothing left to discover"
The quotation is unsourced. No text of Kelvin's contains it, and the attribution appears only in the twentieth century, long after his death.
The nearest genuine statement belongs to someone else. At the dedication of the Ryerson Laboratory in Chicago in 1894, Albert Michelson said:
"It seems probable that most of the grand underlying principles have been firmly established and that further advances are to be sought chiefly in the rigorous application of these principles to all the phenomena which come under our notice."
Even that is not quite Michelson's own claim—he presents it as the view of "an eminent physicist" and appends the warning that future discoveries may well lie in the sixth decimal place. He was right about the decimal place. His own null result was already sitting in it.
The Road Out
Two hundred years after the Principia, the ledger stood like this. Newtonian gravity had predicted a planet nobody had seen and been proved right within a degree. It had failed by 43 arcseconds a century at Mercury, and the fix that had worked for Uranus—another planet, further in—had been searched for and not found. Light had a finite speed, and gravity, so far as anyone could measure, did not. And the medium that light was supposed to be a wave in refused to show itself, no matter how the interferometer was turned.
None of these was fatal on its own. Working physicists in 1900 had reasonable-looking repairs for each: an unseen mass near the Sun, a small correction to the exponent in the inverse-square law, a contraction of moving bodies. What they did not have was a single repair that fixed more than one.
The way out did not come from patching gravity. It came from taking the strangest of the results— that the speed of light is the same however you move—and treating it not as an anomaly to be explained but as a law to be obeyed. That is the subject of the next chapter, and it is a chapter about space and time, not about gravity. Gravity has to wait ten more years, because the theory Einstein publishes in 1905 cannot accommodate it at all.
What Chapter 3 asked, and when it gets answered
Chapter 3 ended on an unpaid debt: Newton had no mechanism for gravity, said so, and physics agreed to stop asking. Leibniz's objection and Mach's were never refuted—they were set aside because the mathematics kept working.
This chapter is what happens when the mathematics stops working. The debt does not come due in Chapter 5: special relativity makes the problem worse, by ruling out the instantaneous action Newton's gravity depends on. It comes due in Chapter 6, and the payment is not a mechanism in the sense Leibniz wanted. It is the discovery that the question had the wrong shape.
Why Newton Still Matters
Why Newton Still Matters
Einstein's General Relativity replaced Newtonian gravity. So why do we still teach Newton? Why do engineers still use F = Gm₁m₂/r² to design satellites, predict tides, and navigate spacecraft?
Because Newton's theory is not wrong—it's a limiting case.
The Newtonian Regime
General Relativity reduces to Newtonian gravity when:
- Gravitational fields are weak: Far from black holes and neutron stars, where spacetime curvature is gentle.
- Velocities are slow: Much slower than the speed of light (v ≪ c).
- Timescales are short: We don't need nanosecond precision over centuries.
For the solar system—planets, moons, asteroids—these conditions hold, and Newton's equations give an answer good to more decimal places than the measurement usually is.
Even for GPS satellites orbiting Earth, Newtonian corrections account for most of the trajectory. Relativistic corrections matter for timing precision: gravitational time dilation (GR) makes satellite clocks tick faster by ~45 µs/day, while orbital velocity (SR) slows them by ~7 µs/day—satellite clocks run fast by a net ~38 µs/day and must be pre-corrected. But the bulk of the trajectory calculation is pure Newton.
Practical Accuracy
Consider some examples where Newton's gravity is "good enough":
- Apollo missions to the Moon: Navigated with Newtonian mechanics throughout. Relativistic corrections were below the error budget and were not applied.
- Mars rovers: Landing trajectories calculated with Newton's laws. GR corrections unnecessary.
- Tidal predictions: Centuries of accurate tide tables based on Newtonian gravity and celestial mechanics.
- Satellite orbits: For most satellites, Newton gives accuracy within meters. Only atomic-clock precision (GPS, Galileo) requires GR corrections.
Newton's theory is an approximation, and for the regime we live in it is a very good one.
Conceptual Foundations
Beyond practical calculations, Newton's framework supplied the scaffolding later physics was built on—though not always in the form he left it:
- Conservation of momentum: implicit in the third law, and still exactly where he put it. The two laws we now group alongside it arrived later and from elsewhere—angular momentum was given its general form by Euler, and energy conservation is a nineteenth-century result (Mayer, Joule, Helmholtz) growing out of the vis viva tradition that ran through Leibniz and du Châtelet rather than through the Principia.
- Inertial frames: The concept of reference frames where F = ma holds unchanged—still fundamental in special and general relativity.
- A field, written in afterwards: Newton gave a force acting between bodies and pointedly declined to say what carried it. Recasting that force as a field filling space—force per unit mass, obeying a differential equation—is the work of Laplace and Poisson mathematically, and of Faraday and Maxwell conceptually. It is Newtonian gravity rewritten in a language Newton did not have and would not have accepted.
- Differential equations: Newton developed the calculus to describe motion—as, separately, did Leibniz, whose notation is the one we actually use. The method (differential equations plus initial conditions) remains the language of physics.
Pedagogical Value
Learning Newton first is not just historical courtesy—it's pedagogically essential:
- Newton's theory is intuitive. Forces, vectors, F = ma—these are concepts we can visualize and test in everyday experience.
- General Relativity is abstract. Curved spacetime, covariant derivatives, tensor equations—you need the Newtonian foundation to appreciate what Einstein changed.
- Understanding why Newton works and where it fails teaches you more about GR than jumping straight to Einstein's equations.
The Relationship Between Newton and Einstein
Think of it this way:
Newton built the first floor, and it carries everything most of us ever need it to carry. Einstein built on top of it. You still live downstairs, and the stairs up are Newton's.
General Relativity doesn't invalidate Newton—it contains Newton. In the limit of weak fields and slow speeds, Einstein's equations become Newton's equations. Newton is not wrong; he's special-case correct.
The Limits Reveal the Depths
The places where Newton fails—Mercury's perihelion, gravitational lensing, GPS timing, gravitational waves—are the useful part. A theory that never breaks tells you nothing about what lies underneath it.
Each of those failures marks a boundary, and the boundaries turn out to have something in common: they appear wherever speeds approach that of light, or fields become strong. What replaces Newton there is not a better force law. Working out what it is takes the next two chapters, and it starts by leaving gravity alone entirely.
Newton gave us the tools to ask the questions. Einstein gave us the tools to answer them.
"All models are wrong. Some models are useful."
Newton's model of gravity is wrong, and useful, and has been both for a hundred years.
The Enduring Legacy of Newton
What Newton established: one law for falling bodies and for planets, Kepler's ellipses derived rather than fitted, and a way of writing physics down—as differential equations— still in use.
Newton's limitation: Could not explain the mechanism of gravity, the equivalence of masses, or the behavior near extreme conditions.
Newton's legacy: Still the theory actually used for almost every gravitational calculation anyone performs, and still the thing you have to know before what Einstein changed means anything.
Newton found the rules. Einstein changed the board they are played on, and left the pieces where they were.
Sources and Further Reading
This chapter covered three discrepancies in Newtonian physics that emerged during the late 19th century: Mercury's anomalous perihelion precession, Rømer's measurement of light speed, and the Michelson-Morley experiment's null result. The following resources provide deeper exploration of this pivotal period in the history of physics.
1. Primary Sources and Historical Papers
Le Verrier, Urbain. "Lettre de M. Le Verrier à M. Faye sur la théorie de Mercure et sur le mouvement du périhélie de cette planète." Comptes rendus hebdomadaires des séances de l'Académie des sciences, vol. 49 (1859): 379–383.
Le Verrier's original announcement of Mercury's anomalous perihelion advance. The paper that started the crisis in Newtonian gravity.
Rømer, Ole. "Démonstration touchant le mouvement de la lumière trouvé par M. Römer de l'Académie Royale des Sciences." Journal des sçavans (1676).
Rømer's original paper demonstrating that light has a finite speed, based on observations of Jupiter's moon Io. English translation available in Philosophical Transactions of the Royal Society (1677).
Michelson, Albert A. and Edward W. Morley. "On the Relative Motion of the Earth and the Luminiferous Ether." American Journal of Science, vol. 34 (1887): 333–345.
The famous "null result" paper. Freely available through Wikisource.
Newcomb, Simon. "The Elements of the Four Inner Planets and the Fundamental Constants of Astronomy." Supplement to the American Ephemeris and Nautical Almanac for 1897 (1895).
Newcomb's refined calculation of Mercury's perihelion precession (43 arcseconds per century), which became the standard value Einstein would later explain.
2. Historical Studies
Mee, Nicholas. Gravity: From Falling Apples to Supermassive Black Holes (2nd edition). Oxford University Press, 2022.
A continuous narrative from Babylonian astronomy to gravitational-wave detection. Its account of the Neptune prediction and of the hunt for Vulcan—including Lescarbault and Le Verrier's refusal to abandon the planet—underlies §1 and §2 of this chapter.
Topper, David R. How Einstein Created Relativity out of Physics and Astronomy (Astrophysics and Space Science Library, vol. 394). Springer, 2013.
Chapter 6, "The Michelson-Morley Muddle," is the source for the treatment of the 1881 and 1887 experiments here: the Potsdam attempt, Lorentz's correction to Michelson's analysis, and the FitzGerald-Lorentz contraction as an aether rescue rather than a discovery. Chapters 1 and 4 supply the Galileo-to-Maxwell background against which this chapter's closing crisis is set.
Roseveare, N.T. Mercury's Perihelion from Le Verrier to Einstein. Oxford University Press, 1982.
The definitive historical account of the Mercury problem. Traces all attempts to explain the anomaly from 1859 to Einstein's solution in 1915, including the search for Vulcan.
Swenson, Loyd S. The Ethereal Aether: A History of the Michelson-Morley-Miller Aether-Drift Experiments, 1880-1930. University of Texas Press, 1972.
Comprehensive history of the aether experiments and their interpretation. Shows how the "null result" was initially controversial and not immediately accepted.
Staley, Richard. Einstein's Generation: The Origins of the Relativity Revolution. University of Chicago Press, 2008.
Examines the broader context of late 19th-century physics and how the crisis in classical physics emerged from precision measurement and electromagnetic theory.
Galison, Peter. Einstein's Clocks, Poincaré's Maps: Empires of Time. W.W. Norton, 2003.
Explores how the practical problem of synchronizing clocks and the theoretical problem of simultaneity converged in Einstein's special relativity.
3. Technical and Scientific Treatments
Will, Clifford M. Was Einstein Right? Putting General Relativity to the Test (2nd edition). Basic Books, 1993.
Accessible account of how general relativity solved the Mercury problem and passed subsequent experimental tests. Excellent for readers with basic physics background.
Pais, Abraham. 'Subtle is the Lord...': The Science and the Life of Albert Einstein. Oxford University Press, 1982.
The definitive scientific biography of Einstein. Chapter 9 discusses Mercury's perihelion and how Einstein knew his theory was correct when it predicted the right value.
Shankland, R.S., et al. "New Analysis of the Interferometer Observations of Dayton C. Miller." Reviews of Modern Physics, vol. 27 (1955): 167–178.
Modern reanalysis of the Michelson-Morley experiments, showing how systematic errors were eliminated and the null result confirmed.
4. Online Resources
AIP History Center: Michelson-Morley Experiment (history.aip.org): Historical exhibit with photographs, original apparatus descriptions, and context.
Stanford Encyclopedia of Philosophy: Early Philosophical Interpretations of General Relativity (plato.stanford.edu): Scholarly article on how Einstein's theory resolved the Mercury problem and other anomalies.
NASA: Tests of General Relativity (science.nasa.gov): Modern experimental tests of Einstein's predictions, including gravitational waves.
5. Recommended Reading by Level
For General Readers
Bodanis, David. E=mc²: A Biography of the World's Most Famous Equation (2000). Engaging narrative covering the historical context of Einstein's breakthrough.
Levenson, Thomas. Einstein in Berlin (2003). Accessible account of Einstein's life during the development of general relativity.
For Students
French, A.P. Special Relativity (1968). Classic undergraduate text connecting the Michelson-Morley experiment to special relativity.
Rindler, Wolfgang. Relativity: Special, General, and Cosmological (2nd edition, 2006). Comprehensive treatment from special to general relativity, including Mercury's perihelion.
For Advanced Study
Misner, Thorne, and Wheeler. Gravitation (1973). The comprehensive treatise on general relativity. Chapter 40 discusses Mercury's perihelion in detail.
Pais, Abraham. 'Subtle is the Lord...' (1982). Essential for understanding Einstein's path to general relativity and how it resolved Newtonian anomalies.
A Note on the Next Chapter: The anomalies described in this chapter—Mercury's perihelion, the constancy of light speed, and the failure to detect the aether—were not isolated puzzles. They were symptoms of a deeper problem with the Newtonian framework. Einstein's general relativity would resolve all three by reconceiving gravity not as a force, but as the curvature of spacetime itself. The next chapter explores how this revolutionary idea emerged and how it was confirmed through observation.
Reader Notes
No notes yet. Be the first to leave one.