The Pound-Rebka Experiment Explained

In 1959, Harvard physicists Robert Pound and Glen Rebka Jr. sent gamma rays up and down a 22.6-meter (74 ft) stairwell inside Jefferson Physical Laboratory and detected a frequency shift smaller than one part in a quadrillion. That shift confirmed a prediction Albert Einstein made in 1911: clocks (and light) run at different rates depending on how deep they sit in a gravitational field. The measurement matched general relativity's prediction to within about 10 percent, and a 1964–65 refinement by Pound and his student Joseph Snider tightened that to roughly 1 percent — still one of the most cited confirmations of general relativity in a university basement, not a satellite or an eclipse expedition.

What gravitational redshift actually predicts

Gravitational redshift — the stretching of a photon's wavelength (and corresponding drop in frequency) as it climbs away from a mass — is a direct consequence of general relativity's central claim: gravity is not a force pulling on light, but a curvature of spacetime that changes how clocks tick at different heights. A clock closer to a massive object runs slower than a clock farther away, and since a photon's frequency is essentially a tiny clock ticking at an enormous rate, that photon appears to lose energy, and therefore drops in frequency, as it moves upward and out of a gravity well.

Einstein first proposed this effect in 1911, four years before he finished the full theory of general relativity, as a straightforward consequence of the equivalence principle — the idea that being in a gravitational field is locally indistinguishable from being in an accelerating rocket. If you accept that equivalence, gravitational time dilation and the associated redshift follow almost immediately from special relativity's treatment of accelerating frames.

The trouble, for nearly five decades, was that nobody could measure it. On Earth's surface, gravity is weak enough that the predicted frequency shift between the floor and the ceiling of a lab is vanishingly small — around 1 part in 10^15 for a height of roughly 20 meters (66 ft). No spectral line known before the mid-1950s was sharp enough to resolve a shift that tiny. Attempts using starlight from the Sun and from white dwarfs like Sirius B produced results but with enormous uncertainty, because stellar redshift measurements are contaminated by pressure effects, rotation, and other astrophysical noise that has nothing to do with general relativity.

The frequency shift Pound and Rebka needed to detect was about 2.5 parts in 10^15 — equivalent to noticing a single extra second added to a clock that has been running since roughly the beginning of the universe.

The Mössbauer effect: the tool that made the impossible measurement possible

The breakthrough that made Pound and Rebka's experiment feasible had nothing to do with astronomy. In 1958, German physicist Rudolf Mössbauer discovered that under the right conditions, atomic nuclei embedded in a solid crystal lattice can emit and absorb gamma rays without losing any energy to recoil. This is now called the Mössbauer effect — recoil-free emission and resonant absorption of gamma-ray photons by nuclei bound in a solid.

Normally, when a nucleus emits a gamma ray, the nucleus recoils, the way a rifle kicks back when it fires a bullet. That recoil shifts the emitted photon's energy slightly away from the exact energy needed for another identical nucleus to absorb it, which normally makes resonant absorption between two identical nuclei nearly impossible to observe cleanly. Mössbauer showed that when the emitting nucleus is locked into a rigid crystal lattice, the recoil momentum can be absorbed by the entire crystal rather than the single nucleus. Because the crystal is billions of times more massive than a single atom, the recoil-induced energy loss becomes negligible, and the emitted gamma ray keeps almost exactly the right energy for a matching nucleus to absorb it.

The practical result was a gamma-ray "clock" of extraordinary precision. Pound and Rebka used the 14.4 keV gamma ray emitted by iron-57 (⁵⁷Fe) nuclei, a transition with a natural linewidth sharp enough to resolve frequency shifts on the order of 1 part in 10^12 or better under the right conditions — several orders of magnitude more precise than anything available before. Mössbauer received the Nobel Prize in Physics in 1961 for the discovery, just two years after Pound and Rebka put it to work testing general relativity (Nobel Prize in Physics, 1961).

Why iron-57 specifically?

Iron-57 was chosen because its 14.4 keV nuclear transition has an unusually narrow natural linewidth relative to its energy, which is what makes Mössbauer spectroscopy possible at all. A narrower line means a smaller, more precisely defined target frequency, so even a tiny gravitational shift moves the emitted line detectably off resonance with the absorber below or above it.

Building the experiment inside a Harvard stairwell

Pound and Rebka's setup was almost aggressively low-tech compared to the astronomical observations that had failed to nail down gravitational redshift for decades. They placed a radioactive cobalt-57 source, which decays into excited iron-57 and emits the 14.4 keV gamma ray, at one end of the Jefferson Physical Laboratory's stairwell, and an iron-57 absorber at the other end, separated by a vertical distance of 22.6 meters (74 ft).

The logic was simple in principle: a gamma ray emitted at the bottom and traveling up to the top should be very slightly redshifted, losing just enough energy that it would no longer resonantly match the absorber's 14.4 keV transition. If the source and absorber were perfectly at rest relative to each other, the gravitational shift alone would knock the two out of resonance, and the absorber would absorb noticeably fewer gamma rays than it would without gravity's effect.

To detect that tiny mismatch, Pound and Rebka needed a way to compensate for it and measure how much compensation was required. They mounted the source on a loudspeaker cone driven by an oscillating signal, which moved the source up and down at a controlled velocity. That mechanical motion introduced a Doppler shift — the same effect that makes an ambulance siren rise and fall in pitch as it passes you — which could be tuned to exactly cancel the gravitational redshift. By finding the exact Doppler velocity needed to restore maximum resonant absorption, they could back out the size of the gravitational effect with high precision, since the required Doppler velocity was itself directly measurable.

The Doppler velocity Pound and Rebka needed to cancel the gravitational shift was only about 7.5 micrometers per second (roughly the speed of a fast-growing fingernail) — and their apparatus still measured it.

Why not just measure the shift directly?

Directly resolving a frequency shift of a few parts in 10^15 with 1959-era gamma-ray detectors was not possible. Modulating the source's velocity with a loudspeaker and hunting for the Doppler compensation point turned an immeasurably small static shift into a measurable, tunable, and repeatable dynamic signal, which is the technique that made the whole experiment work.

The 1959 result and Pound-Snider's 1964–65 refinement

Pound and Rebka published their first result in Physical Review Letters in 1959, and a fuller account followed shortly after (Pound & Rebka, Physical Review Letters, 1959). They reported the ratio of the observed frequency shift to the value predicted by general relativity as 1.05 ± 0.10 — in other words, the measured shift matched Einstein's prediction to within about 10 percent, with the theoretical prediction sitting comfortably inside the experimental error bars.

That 10 percent precision was already remarkable for a tabletop measurement of an effect that astronomical observations had struggled to pin down for nearly 50 years. But Pound and his graduate student Joseph Snider went back to the same tower and the same basic method between 1964 and 1965, refining the temperature control of the source and absorber crystals, improving the electronics used to drive and measure the Doppler modulation, and tightening the analysis of systematic errors from effects like the second-order Doppler shift, which arises from the thermal vibration of atoms in the crystal lattice and has to be carefully separated from the gravitational signal.

The Pound-Snider refinement brought the measured-to-predicted ratio down to about 1 percent agreement with general relativity, a tenfold improvement in precision achieved with essentially the same 22.6-meter tower and the same Mössbauer technique, just executed with more careful control of noise sources (Pound & Snider, Physical Review Letters, 1965). For a theory that had previously been tested mainly through Mercury's orbital precession and starlight bending during solar eclipses, a 1 percent laboratory confirmation was a significant addition to the evidence.

How Pound-Rebka fits alongside the other classic tests of general relativity

The Pound-Rebka experiment did not stand alone. It joined a small set of experiments and observations that, together, form the historical backbone of evidence for general relativity: Mercury's anomalous perihelion precession (explained by Einstein in 1915), the 1919 Eddington eclipse expedition that measured starlight bending around the Sun, and later, space-based and orbital tests that pushed precision even further. Each test probes a different consequence of curved spacetime, and each was constrained by the technology available at the time.

Test	Year	Method	Precision achieved
Pound-Rebka	1959	Mössbauer gamma-ray redshift, 22.6 m tower	~10%
Pound-Snider refinement	1964–65	Same tower, improved temperature/electronics control	~1%
GPS relativistic correction	Ongoing since 1978	Atomic clocks pre-tuned to offset combined special + general relativistic drift	Sub-microsecond/day, continuously verified
Modern optical-clock tests	2010s–2020s	Optical lattice clocks measuring redshift over millimeter to meter height differences	Parts in 10^17 to 10^20

Each row narrows the error bars on the same prediction — from a 10 percent tower measurement in 1959 to optical clocks in the 2010s and 2020s sensitive enough to detect gravitational time dilation across a height difference of about a millimeter (one thousandth of a meter).

The GPS system offers the most consequential everyday descendant of this line of research. Satellite atomic clocks orbiting roughly 20,200 kilometers (about 12,550 miles) above Earth experience less gravitational time dilation than clocks on the ground, and engineers pre-correct for this before launch so that the navigation system stays accurate; without that correction, GPS position errors would accumulate by kilometers within a single day. Read the full mechanism in our explainer on how GPS corrects for relativity.

Modern optical lattice clocks, built from techniques that trace back to the same atomic-transition precision that made the Mössbauer effect useful, have pushed gravitational redshift measurements to parts in 10^17 or better, detecting time dilation across height differences of just a few centimeters (a few inches) on a lab bench (National Institute of Standards and Technology optical clock research). The core physics being tested is identical to what Pound and Rebka measured in a stairwell in 1959; only the precision has changed by roughly 15 orders of magnitude.

Why the equivalence principle made this test so important

General relativity rests on the equivalence principle: the statement that gravitational mass and inertial mass are the same thing, and that no local experiment can distinguish being at rest in a gravitational field from being accelerated in the absence of gravity. Gravitational redshift is one of the cleanest, most direct consequences of that principle, because it follows from asking what an accelerating observer would see, then translating that answer into gravitational language.

The equivalence principle — the idea that gravitational and inertial mass are identical, making a uniform gravitational field locally indistinguishable from constant acceleration — is easier to grasp with an analogy. Imagine standing in a windowless elevator that's accelerating upward through empty space, far from any planet. A light beam entering one side of the elevator would appear to bend downward relative to the floor, because the floor is accelerating up to meet it during the light's brief transit time. Einstein's insight was that this bending should look identical to what happens in a genuine gravitational field, and by extension, that a clock (or a photon, acting as a clock) at the "bottom" of an accelerating frame should tick slower, or shift toward lower frequencies, than one at the "top." Pound-Rebka tested that exact prediction with real gravity instead of a thought-experiment elevator.

A second useful analogy: think of a photon climbing out of a gravity well the way a ball loses kinetic energy climbing a hill. The ball slows down; the photon can't slow down, since light always travels at the same speed, so instead it "pays" the energy cost by dropping in frequency, the same way a runner going uphill might keep the same stride length but slow their cadence. That's gravitational redshift in a sentence: the photon keeps its speed but loses energy, and lower energy for a photon always means lower frequency.

Common misconceptions about the Pound-Rebka experiment

A frequent misreading treats Pound-Rebka as a test of special relativity's time dilation, the effect where a fast-moving clock runs slow relative to a stationary observer. It is not. Pound-Rebka measured a purely gravitational effect tied to height in a gravitational field, not velocity through space. The two effects, gravitational and velocity-based time dilation, both appear in the GPS correction system, but they arise from different parts of Einstein's theory and were tested by different experiments historically.

Another common confusion is assuming the experiment detected light bending near a massive object, the effect made famous by the 1919 Eddington eclipse observations. Pound-Rebka measured a frequency shift, not a change in light's direction. Bending of starlight and gravitational redshift are both consequences of curved spacetime, but they are conceptually distinct predictions that required entirely different experimental setups to confirm.

A third misconception treats the 1959 result as the final word. In reality, Pound-Rebka was a beginning, not an endpoint. The Pound-Snider refinement in 1964–65 was itself superseded decades later by rocket-based tests, such as the Gravity Probe A mission in 1976, which confirmed gravitational redshift to about 70 parts per million by comparing a hydrogen maser clock on a suborbital rocket to one on the ground (NASA Gravity Probe A results), and more recently by optical-clock experiments sensitive to redshift differences over centimeter-scale height changes. Precision testing of general relativity has never really stopped; Pound-Rebka simply opened the door.

It's also worth being honest about what the experiment could not do. Pound-Rebka confirmed general relativity's prediction within the Sun-Earth gravitational regime, which is a weak-field environment. It says nothing directly about strong-field regimes, such as near a black hole's event horizon, where general relativity's predictions diverge much more sharply from Newtonian intuition and remain an active area of research using instruments like the Event Horizon Telescope and gravitational-wave observatories.

The takeaway

Pound and Rebka turned a prediction that had eluded physicists for nearly five decades into a measurement that fit inside a university stairwell, using a Nobel Prize-winning nuclear physics trick rather than a telescope. The core insight worth keeping is how the experiment's power came from combining two separate pieces of physics: the Mössbauer effect's ability to create an almost perfectly sharp gamma-ray "clock," and the equivalence principle's prediction of exactly how much that clock should shift over a known height. Neither piece alone would have produced a confirmable result; together, they turned a theoretical curiosity into hard data, first to 10 percent precision in 1959, then to about 1 percent in 1964–65, and eventually into a working correction system running continuously in every GPS satellite overhead.

---

Read next:

• How GPS Corrects for Relativity
• The Michelson-Morley Experiment Explained
• Albert Einstein: The Eccentric Genius Who Hated Socks

FAQ

Q: What did the Pound-Rebka experiment actually prove?

It confirmed Einstein's 1911 prediction of gravitational redshift: that light loses frequency as it climbs away from a massive object. The 1959 measurement matched general relativity's predicted value to within about 10 percent, and the 1964–65 Pound-Snider refinement tightened that agreement to roughly 1 percent.

Q: Why did Pound and Rebka need the Mössbauer effect?

The gravitational frequency shift over their 22.6-meter tower was only about 2.5 parts in 10^15, far too small for any spectral line available before 1958. The Mössbauer effect, recoil-free gamma-ray emission and absorption in a solid crystal, produced a gamma-ray transition sharp enough to resolve a shift that tiny.

Q: How tall was the tower used in the experiment?

The experiment used the stairwell of Jefferson Physical Laboratory at Harvard University, with a vertical separation of 22.6 meters (74 ft) between the gamma-ray source at the bottom and the absorber at the top.

Q: Is gravitational redshift the same thing as gravitational time dilation?

They're the same underlying effect described two ways. Gravitational time dilation describes clocks running slower deeper in a gravitational field; gravitational redshift describes the resulting drop in frequency of light (or any periodic signal, since light acts as a clock) climbing out of that field.

Q: Does GPS use the Pound-Rebka result directly?

GPS relies on the same physics Pound-Rebka confirmed, gravitational time dilation, but the correction applied to GPS satellite clocks is calculated from general relativity's equations directly, not from the Pound-Rebka data itself. The experiment's importance was establishing, with hard laboratory evidence, that the underlying prediction was correct.

Q: How much more precise are modern optical-clock redshift tests than Pound-Rebka?

Enormously more precise, and over a far smaller height difference. In 2010, NIST physicists Chou, Hume, Rosenband, and Wineland compared two aluminum-ion optical clocks separated by just 33 centimeters (about 1 ft) of elevation and still detected the predicted gravitational frequency shift, publishing the result in Science (Chou et al., Science, 2010). Pound and Rebka needed a 22.6-meter tower to catch a shift of about 2.5 parts in 10^15; the NIST clocks resolved a comparable effect across a height difference nearly 70 times shorter, because optical-clock frequency stability had improved by several orders of magnitude in the intervening five decades.

The gap has widened further since. A 2021–2022 JILA experiment, described in Nature and covered by NIST, split a single cloud of ultracold strontium atoms into upper and lower layers separated by roughly one millimeter (0.04 in) and still measured the gravitational redshift between the two layers, the smallest height difference over which the effect has ever been directly resolved (JILA/NIST, Nature, 2022). Where Pound and Rebka needed a Harvard stairwell and a loudspeaker rigged to cancel a shift smaller than one part in a quadrillion, the JILA team resolved essentially the same physics inside a single lattice trap smaller than a fingertip. The underlying prediction never changed; what changed was the ability to build a clock stable enough to notice.