Modified Newtonian Dynamics, universally known as MOND, is an alternative to dark matter that proposes gravity itself behaves differently when gravitational accelerations drop below a very small threshold. First introduced by physicist Mordehai Milgrom in 1983, the idea was born from a simple observation: galaxies spin in ways that Newton’s laws cannot explain unless we either add invisible mass or change the rules of gravity. MOND takes the second path, and four decades later it remains the most developed rival to the dark matter hypothesis, with a track record of predictions that is impressive in some domains and troubled in others.
The Problem MOND Was Built to Solve
When astronomers measure how fast stars orbit the centers of spiral galaxies, they find something strange. Stars far from the galactic center move just as quickly as stars closer in, producing flat “rotation curves” instead of the declining speeds that Newtonian gravity predicts from the visible matter alone. This mismatch between what you see and what gravity seems to require is the foundation of the entire dark matter debate.1The Astrophysical Journal. The Published Extended Rotation Curves of Spiral Galaxies: Confrontation with Modified Dynamics The standard explanation is that galaxies sit inside enormous halos of invisible dark matter, whose gravity holds the fast-moving outer stars in orbit. MOND offers a radically different explanation: there is no dark matter, and gravity simply works differently at the extremely low accelerations found in galactic outskirts.
The Acceleration Threshold
MOND’s central claim is that Newton’s second law, or equivalently Newton’s law of gravity, breaks down when gravitational acceleration drops below a specific value called a₀, roughly 1.2 × 10⁻¹⁰ meters per second squared. Above that threshold, gravity behaves exactly as Newton described. Below it, gravitational attraction declines more slowly with distance than the familiar inverse-square law predicts, effectively becoming stronger than you would expect from visible matter alone.2AIP Conference Proceedings. Modified Newtonian Dynamics, an Introductory Review
To put the scale in perspective, a₀ is absurdly small by everyday standards. The gravitational acceleration you feel standing on Earth’s surface is roughly a trillion times larger. Even the Sun’s pull on the outermost planets is well above this threshold. You only cross into MOND territory in the sprawling outer reaches of galaxies, in the dynamics of galaxy clusters, or in systems where objects orbit each other at separations of thousands of astronomical units. This is why MOND has no effect on anything you experience in daily life or even in most of the solar system.
Galaxy Rotation Curves and the Tully-Fisher Relation
Where MOND truly earns its reputation is in predicting how galaxies rotate. Given only the distribution of visible matter in a galaxy (stars, gas, and dust), MOND can calculate the expected rotation curve with essentially one free parameter per galaxy: the mass-to-light ratio of its stellar population. This works across a striking range of galaxy types. Low surface brightness galaxies, which are dominated by gas and have very little stellar mass, are an especially strong test because dark matter models need to be tuned individually for each one. MOND fits to 15 such galaxies produced good matches to the observed rotation curves with reasonable stellar mass-to-light ratios.3IOPscience / The American Astronomical Society. Testing Modified Newtonian Dynamics with Low Surface Brightness Galaxies: Rotation Curve Fits
Perhaps MOND’s single most celebrated success involves the baryonic Tully-Fisher relation, a tight correlation between a galaxy’s total ordinary mass and the fourth power of its flat rotation speed. In the dark matter framework, this relationship emerges somewhat indirectly from the interplay between dark halos and baryonic matter, and its tightness is not guaranteed. In MOND, the relation falls out directly from the theory itself. Independent data from gas-rich galaxies confirm the relationship with a form fully consistent with MOND’s prediction, yielding an acceleration constant of about 1.3 × 10⁻¹⁰ meters per second squared, right where MOND says it should be.4Astronomical Journal. The Baryonic Tully-Fisher Relation of Gas Rich Galaxies as a Test of LCDM and MOND The fact that MOND predicted this tight, specific relationship before it was confirmed observationally is one of the strongest arguments in its favor.5The Astrophysical Journal Letters. The Challenge to MOND from Ultra-faint Dwarf Galaxies
A closely related success involves the radial acceleration relation, which connects the total gravitational acceleration at any point in a galaxy to the acceleration expected from visible matter alone. Observations of rotationally supported galaxies show this relationship to be remarkably tight, with characteristics that line up with what MOND predicts. Interestingly, the standard dark matter model can also reproduce this tightness, which means the radial acceleration relation does not by itself settle the debate, but the fact that MOND predicted the relationship’s shape decades before it was measured with precision remains a point in its column.6Monthly Notices of the Royal Astronomical Society. Hooks & Bends in the radial acceleration relation: discriminatory tests for dark matter and MOND
The External Field Effect
MOND makes a prediction that dark matter does not: the internal gravity of a system should depend on the gravitational environment it sits in. In standard physics, if you are in a freely falling elevator, you feel weightless regardless of how strong the external gravity is. This is the strong equivalence principle, and general relativity demands it hold for self-gravitating systems. MOND violates it. A galaxy falling through a region of strong external gravitational field should have its own internal dynamics altered, even if the external field is uniform. This is called the external field effect.
A 2020 study tested this by looking at the rotation curves of galaxies in different gravitational environments. Galaxies embedded in stronger external fields should, according to MOND, show less of the modified-gravity boost to their rotation curves. The results supported exactly this pattern, pointing toward a breakdown of the strong equivalence principle consistent with MOND’s prediction.7The Astrophysical Journal. Testing the Strong Equivalence Principle: Detection of the External Field Effect in Rotationally Supported Galaxies A follow-up study connected the external field effect signal to the large-scale structure of the universe, finding that the direction and magnitude of the effect tracked with the expected gravitational field from surrounding cosmic structures.8The Astrophysical Journal. Testing the Strong Equivalence Principle. II. Relating the External Field Effect in Galaxy Rotation Curves to the Large-scale Structure of the Universe
If the external field effect is real, it would be extremely difficult to explain within the standard dark matter framework. Dark matter halos do not care about external fields in this way. This makes it one of the few truly discriminatory tests between the two paradigms, though the statistical significance and potential systematic effects in these measurements remain subjects of active discussion.
Where MOND Struggles
For all its galactic success, MOND runs into serious trouble at larger scales. Galaxy clusters are the most prominent example. When you apply MOND to clusters, the visible mass still falls short of what is needed to explain the observed dynamics by a factor of about two to three. In other words, even with modified gravity, clusters seem to need some form of unseen mass. An updated analysis of the Bullet Cluster using gravitational lensing data from the James Webb Space Telescope confirmed that this residual missing mass problem persists, and moreover that the missing mass appears to be collisionless, centered on the galaxies rather than the gas, which is precisely what you would expect from dark matter.9Research Notes of the AAS. On the Residual Missing Mass of the Bullet Cluster
MOND proponents have suggested that this cluster-scale missing mass could be ordinary matter in a form we have not yet detected, such as cold dense gas clouds, or possibly massive neutrinos. Neither explanation has gained broad acceptance. The Bullet Cluster in particular, where two clusters collided and the mass distribution separated from the hot gas, has long been cited as a smoking gun for dark matter. Under MOND, the collisionless nature of the residual missing mass is an awkward fit.
Dwarf satellite galaxies also present challenges. Some Milky Way dwarf spheroidal galaxies appear to have mass-to-light ratios that are difficult for MOND to explain with stellar populations alone. An analysis of eight such galaxies found that six could be fit with reasonable stellar mass-to-light ratios, but two, Sextans and Draco, required implausibly high values, suggesting either hidden mass or a problem with the MOND framework in these systems.10Monthly Notices of the Royal Astronomical Society. Dwarf spheroidals in MOND Ultra-faint dwarf galaxies pose an even sharper challenge, since their extremely low baryonic masses and high velocity dispersions push MOND into regimes where its predictions become harder to reconcile with observations.11The Astrophysical Journal Letters. The Challenge to MOND from Ultra-faint Dwarf Galaxies
The Wide Binary Controversy
One of the most dramatic recent episodes in the MOND story involves wide binary stars. These are pairs of stars orbiting each other at separations of thousands of astronomical units, where their mutual gravitational acceleration drops into the MOND regime. If MOND is correct, the relative velocities of widely separated binaries should be systematically higher than Newtonian gravity predicts. Data from the Gaia satellite, which has mapped the positions and motions of over a billion stars with extraordinary precision, made this test possible for the first time.12Monthly Notices of the Royal Astronomical Society. A critical review of recent Gaia wide binary gravity tests
The results have not been kind to MOND. A detailed analysis comparing Newtonian and MOND models against Gaia data found Newtonian gravity preferred with very high statistical confidence, excluding the standard MOND prediction at the level of 16 standard deviations.13Monthly Notices of the Royal Astronomical Society. Strong constraints on the gravitational law from Gaia DR3 wide binaries A subsequent study using a refined quality framework for selecting pure binary samples and spanning both Newtonian and MOND regimes independently found no evidence for MOND-like behavior.14Monthly Notices of the Royal Astronomical Society. A quality framework for testing gravity with wide binaries: no evidence for MOND
MOND advocates have pushed back, arguing that the external field effect could suppress the MOND signal in wide binaries. Because these systems are embedded in the Milky Way’s gravitational field, which is above the MOND threshold, the external field effect might cause them to behave more Newtonianly than a naive MOND calculation predicts. The debate hinges on exactly how much the external field suppresses the effect and whether the data can distinguish between these scenarios. A critical review of the various wide binary studies noted that sample contamination, unresolved triple systems, and methodological choices all affect the conclusions, though the overall trend of the data remains more consistent with Newtonian gravity.15Monthly Notices of the Royal Astronomical Society. A critical review of recent Gaia wide binary gravity tests
Making MOND Work With Einstein
Milgrom’s original MOND was a modification of Newtonian mechanics, which is a non-relativistic framework. For MOND to be taken seriously as a fundamental theory, it needed to be embedded in a relativistic theory consistent with general relativity’s well-tested predictions. This turned out to be a formidable challenge. Early extensions like AQUAL and QMOND succeeded in preserving conservation laws and the weak equivalence principle, but remained non-relativistic.16The Astrophysical Journal. Exact Semianalytical Calculation of Rotation Curves with Bekenstein–Milgrom Nonrelativistic MOND
The first full relativistic MOND theory was Bekenstein’s Tensor-Vector-Scalar gravity (TeVeS), published in 2004. TeVeS showed that it was possible to construct a relativistic framework that reduced to MOND in the appropriate limit. Calculations using TeVeS demonstrated that it could potentially reproduce the cosmic microwave background power spectrum and galaxy distributions.17PubMed. Large scale structure in Bekenstein’s theory of relativistic modified Newtonian dynamics However, TeVeS was dealt a blow when the 2017 detection of gravitational waves from a neutron star merger showed that gravitational waves travel at the speed of light, which ruled out certain versions of TeVeS.
A newer relativistic framework was proposed in 2021, designed specifically to evade the gravitational wave speed constraint. This theory demonstrated agreement with observed cosmic microwave background and matter power spectra on large cosmological scales, bringing relativistic MOND closer to viability on the very scales where the original framework struggled most.18PubMed. New Relativistic Theory for Modified Newtonian Dynamics Whether this newer theory can survive the full battery of cosmological tests that the standard dark matter model passes remains an open question.
Gravitational Lensing
Light bending around massive objects, gravitational lensing, provides another arena where MOND must prove itself. Since MOND changes how gravity works, it also changes how much light bends around galaxies. An analysis of 36 strong gravitational lenses found that MOND’s predicted “fundamental plane,” a relationship between galaxy mass, size, and velocity dispersion, coincided with the observed lensing fundamental plane to within about 30 percent. The small discrepancy could be explained by how modified gravity contributes to photon deflection. The study found no evidence for significant missing mass within these high-surface-density systems, exactly as MOND would predict.19Monthly Notices of the Royal Astronomical Society. MOND and the lensing fundamental plane: no need for dark matter on galaxy scales
Future surveys may sharpen this test considerably. A recent theoretical analysis showed that MOND predicts a substantially higher number of disc galaxies acting as gravitational lenses compared to dark matter models, a difference that upcoming wide-field surveys should be able to detect.20arXiv. A Novel Test for MOND: Gravitational Lensing by Disc Galaxies If MOND-type gravity enhances lensing by disc galaxies more than dark matter halos do, a statistical comparison of lens counts could become one of the cleanest discriminating tests between the two frameworks.
Solar System Constraints
Because the Sun’s gravitational field keeps accelerations well above the MOND threshold throughout most of the solar system, you might expect MOND to be invisible here. But the external field effect introduces a subtlety. The solar system sits within the Milky Way’s gravitational field, and the transition between the galaxy’s external field and the Sun’s internal field should produce a small but potentially measurable anomalous force. Specifically, MOND predicts an extra quadrupolar correction to planetary orbits, aligned with the direction of the galactic gravitational field, which would show up as an anomalous precession of planetary orbits. Numerical calculations found that this effect is largest for the outer gas giant planets. For Saturn, the predicted precession is comparable to the limits set by the best planetary orbit models, meaning it is either marginally consistent with observations or marginally excluded depending on which version of MOND is used.21arXiv. Testing MOND in the Solar System
Improvements in planetary ephemerides, driven partly by missions like Cassini’s long stay at Saturn, continue to tighten these constraints. If the anomalous precession were convincingly detected, it would be a dramatic confirmation. If it is definitively ruled out, certain formulations of MOND would need to be abandoned or modified further.
Entropic Gravity and the Deeper “Why”
One persistent criticism of MOND is that it lacks a compelling theoretical motivation. The acceleration scale a₀ is simply posited, and there is no widely accepted explanation for why gravity should change behavior at that particular value. Several theorists have explored a possible connection to entropic gravity, an idea that treats gravity not as a fundamental force but as an emergent phenomenon arising from thermodynamics and information theory.
In one approach, generalizing the entropic gravity framework by introducing a minimum temperature for microscopic degrees of freedom on a holographic screen naturally produces a gravitational law that takes the MOND form. The resulting acceleration constant turns out to be proportional to this minimum temperature, which can be interpreted as the Unruh temperature associated with the expansion of the universe.22Modern Physics Letters A. Entropic Gravity, Minimum Temperature, and Modified Newtonian Dynamics Related work using nonextensive thermodynamics and the Debye model of statistical mechanics has independently arrived at MOND-like gravitational laws from entropic principles.23Europhysics Letters. MOND and cosmological issues from entropic gravity and nonextensive thermostatistics correspondence24Journal of Cosmology and Astroparticle Physics. Einstein equations and MOND theory from Debye entropic gravity
These connections are suggestive rather than definitive. They hint that the MOND acceleration scale might not be an arbitrary parameter but could be tied to the cosmological constant or the expansion rate of the universe, both of which yield a similar numerical value when converted to an acceleration. If that connection turns out to be real, it would transform MOND from an empirical recipe into a window onto deep physics. For now, these remain theoretical explorations without direct observational confirmation.
The Philosophy of the Debate
The MOND-versus-dark-matter argument is not just about data. It raises genuinely interesting methodological questions about how science evaluates competing paradigms. A 2023 paper in the philosophy of science examined the debate through the lens of major 20th-century philosophers, finding that widespread misconceptions and misapplications of ideas from Popper, Kuhn, and Lakatos crop up regularly in the technical literature.25PubMed. Methodological reflections on the MOND/dark matter debate
One common framing treats MOND as “unfalsifiable” because it can sometimes be adjusted to accommodate new data. But the same criticism applies to dark matter models, where the properties of the dark halo can be tuned galaxy by galaxy. A fairer assessment is that MOND makes riskier, more specific predictions at the galactic scale (and sometimes gets them right in advance), while the standard dark matter model makes stronger predictions at cosmological scales (and sometimes gets galaxies right only after the fact). Neither paradigm neatly fits the textbook ideal of a clean falsification framework. The sociology of the debate also matters: because dark matter is embedded in the standard cosmological model, which has passed an enormous array of independent tests, MOND faces a much higher evidential bar to be taken seriously, fairly or not.
Whether MOND ultimately survives, is incorporated into a broader theory, or is replaced by something that captures its galactic successes while handling clusters and cosmology, the framework has already served a valuable role. It identified regularities in galaxy dynamics that the dark matter paradigm did not predict and still struggles to explain naturally. Even researchers who are skeptical of MOND as a theory of gravity acknowledge that the empirical patterns it highlights are real and demand explanation.
Cosmological Structure Formation
One of the least developed aspects of the MOND research program is structure formation: how the large-scale cosmic web of galaxies, filaments, and voids emerged from the nearly uniform conditions of the early universe. In the standard model, dark matter plays a starring role in structure formation. Because dark matter does not interact with light, it could begin clumping under gravity before ordinary matter could, creating the gravitational scaffolding into which gas later fell to form galaxies. Without dark matter, MOND needs an alternative mechanism to explain how structure grew fast enough to match what we observe.
Hydrodynamical simulations of structure formation in a Milgromian cosmology have been attempted, exploring whether the enhanced gravitational pull at low accelerations can drive the formation of structures comparable to those in the observed universe.26Monthly Notices of the Royal Astronomical Society. Hydrodynamical structure formation in Milgromian cosmology These simulations are still in relatively early stages compared to the vast library of dark matter N-body simulations that have been refined over decades. Getting the cosmic microwave background power spectrum, the distribution of galaxy clusters, and the abundance of large-scale voids all right simultaneously is a tall order for any MOND-based cosmology, and it remains one of the framework’s most significant open challenges.
Some researchers argue that this is not a fair comparison, since dark matter cosmology has had thousands of person-years of development while MOND cosmology has had a fraction of that investment. Others counter that the difficulty is not a matter of effort but of principle: without a cold, collisionless, gravitationally dominant component seeding early structure, the physics simply does not work the same way, and no amount of refinement will close the gap without introducing something functionally equivalent to dark matter. The coming decade of surveys, including data from the Vera Rubin Observatory, Euclid, and next-generation cosmic microwave background experiments, will likely tighten constraints on both sides enough to shift the balance of evidence one way or the other.