MOND, Original Phenomenological Law
Newton's law breaks down at very low accelerations. Below a universal scale a_0 ≈ 10^-10 m/s², gravity is effectively stronger, and galaxies rotate as observed without dark matter.
Placeholder for a 3D visualisation of Modified Gravity / MOND. The interactive scene will land in Phase 3. Modified-gravity programs share a single editorial commitment: instead of postulating invisible particles (dark matter) or a cosmological term (dark energy), modify general relativity at the scales where ΛCDM invokes them. MOND's original observation is that galaxy rotation curves obey a tight phenomenological law tied to a universal acceleration scale a_0 ≈ 10^-10 m/s². This pattern is too clean to be coincidence and is not derived from first principles by ΛCDM. The hard part is extending the same idea to clusters, the CMB, and large-scale structure. Whether the same principle scales to clusters and cosmology is the open empirical question. Cluster and cosmological data favor ΛCDM cleanly; galaxy-scale data favor MOND-style patterns just as cleanly. Both empirical signals are real.
In one sentence
Milgrom proposed in 1983 that Newton's law of gravity breaks down at extremely low accelerations, below a universal scale a_0 ≈ 10^-10 m/s². Below this scale, gravity effectively gets stronger than Newton predicts, and galaxies rotate as observed without invoking dark matter.
The claim
MOND is not a relativistic theory, it's a modification of Newton's force law that kicks in below an acceleration scale a_0 ≈ 10^-10 m/s². Above a_0, gravity behaves exactly as Newton predicted. Below a_0, the force becomes effectively stronger, so an object orbiting at low velocity feels more gravity than a Newtonian calculation would predict. In the deep-MOND regime where acceleration a is much less than a_0, the Newtonian acceleration g_N maps to actual acceleration a via a ≈ sqrt(g_N · a_0). This single rule fits galaxy rotation curves across enormous dynamical range, with one free parameter (a_0 fixed empirically once) and no free halo parameters per galaxy.
The galaxy-scale success is striking. McGaugh, Lelli, and Schombert (2016, 2017) showed that around 150 galaxies in the SPARC database satisfy a tight relation between centripetal acceleration (from rotation velocity) and the Newtonian acceleration computed purely from baryonic matter, the radial acceleration relation. Scatter is consistent with observational error; the intrinsic scatter is consistent with zero. Whether MOND is the right underlying theory or not, this empirical regularity is real and is not yet derived from first principles within ΛCDM, where it must emerge from baryonic feedback in N-body simulations.
The honest two-level picture: MOND captures a real galaxy-scale empirical pattern that any successful theory of the dark sector must explain. But at cluster scales, the Bullet Cluster (Clowe et al 2006) shows lensing mass cleanly separated from baryonic gas in a way MOND alone cannot explain without invoking some additional dark component. And the CMB acoustic-peak structure is reproduced precisely by ΛCDM with cold dark matter; relativistic MOND extensions struggle here. Wide-binary tests using Gaia DR3 are an active empirical question right now: the joint Banik-Pittordis-Sutherland 2024 analysis sets strong constraints, with different sub-analyses pointing in different directions.
The family stance
Galaxies don't need invisible matter; they need modified gravity at low accelerations. The radial acceleration relation is real and tight, and any successful theory of the dark sector must reproduce it. Whether the same principle scales to clusters and cosmology is the open empirical question. Cluster and cosmological data favor ΛCDM cleanly; galaxy-scale data favor MOND-style patterns just as cleanly. Both empirical signals are real.
Predictions
- Radial acceleration relation: galaxies follow a specific tight relation between observed centripetal acceleration and the Newtonian acceleration computed from baryonic matter, with universal a_0 and no per-galaxy halo parameters
- Baryonic Tully-Fisher relation: V^4 ∝ M_baryon · a_0 for rotationally supported galaxies; no scatter from dark halo properties because there is no halo
- Wide binaries at separations above ~10^4 AU have orbital velocities probing the deep-MOND regime; specific Newton-Kepler deviations are predicted
Evidence
- McGaugh, Lelli & Schombert 2016 (PRL 117) measured the radial acceleration relation in ~150 SPARC galaxies; scatter is consistent with observational error, no per-galaxy free parameters needed
- Lelli, McGaugh, Schombert & Pawlowski 2017 (ApJ 836), 'One Law to Rule Them All,' extended the analysis and found the relation holds across diverse galaxy types
- Baryonic Tully-Fisher relation matches MOND prediction across roughly four decades in baryonic mass
- Banik, Pittordis, Sutherland and collaborators 2024 (joint Gaia DR3 wide-binary analysis) report strong constraints on the gravitational law in the deep-MOND acceleration regime; the precise interpretation of the bounds vs MOND vs Newton remains actively debated
Counterpoints
- Clowe et al 2006 Bullet Cluster: weak-lensing mass is spatially separated from baryonic gas after a cluster collision, requiring some dark component beyond what MOND alone provides
- CMB acoustic-peak structure is fit precisely by ΛCDM with cold dark matter; no MOND-derived cosmological framework reproduces the CMB without invoking some particle-like dark component
- Pittordis & Sutherland 2022 wide-binary analysis using Gaia EDR3 found a preference for GR over MOND, contradicting earlier claims of MOND-like tension; subsequent analyses are ongoing
- Galaxy cluster dynamics (X-ray temperature profiles, lensing mass) require more mass than baryons alone, even in MOND, undercutting the 'no dark matter' motivation at cluster scales
- MOND has no fundamental motivation: a_0 is fit empirically, not derived from first principles, and the interpolating function between Newtonian and deep-MOND regimes is also free-form
Variants in this family
▸Go deeperTechnical detail with proper terminology
MOND interpolating function: a · μ(a / a_0) = g_N, where g_N is the Newtonian acceleration computed from visible matter and μ is a smooth function with μ(x) -> 1 for x >> 1 and μ(x) -> x for x << 1. The deep-MOND limit gives a = sqrt(g_N · a_0). The value a_0 ≈ 1.2 × 10^-10 m/s² is fit to galaxy data; it is approximately c · H_0 (the speed of light times the Hubble parameter), which is suggestive but not derived.
Radial acceleration relation form: g_obs = g_bar / (1 - exp(-sqrt(g_bar / a_0))), as proposed by McGaugh-Lelli-Schombert 2016. This is empirical, derived from data, not from a fundamental theory.
Cluster scales: in MOND, cluster dynamics require additional unseen mass even with the modified law (often invoked as hot neutrinos or hot dark matter at cluster scales, which would still leave galaxy-scale MOND intact). This combination is not widely accepted.
Wide-binary parameter space: Gaia DR3 wide binaries with separations 10^4 to 10^5 AU and orbital periods 10^5 to 10^6 yr probe the deep-MOND regime. The expected MOND deviation from Newton-Kepler is detectable if a_0 is universal, but the analysis depends critically on excluding hierarchical (triple) systems where a third star contaminates the dynamics. The Banik-Pittordis-Sutherland 2024 joint paper is the current authoritative analysis; Pittordis-Sutherland-Shepherd 2025 added triple-modelling refinements.
References
- EstablishedMilgrom (1983). A modification of the Newtonian dynamics as a possible alternative to the hidden mass hypothesis. Astrophys. J. 270, 365
- EstablishedMcGaugh, Lelli & Schombert (2016). Radial Acceleration Relation in Rotationally Supported Galaxies. Phys. Rev. Lett. 117, 201101
- EstablishedLelli, McGaugh, Schombert & Pawlowski (2017). One Law to Rule Them All: The Radial Acceleration Relation of Galaxies. Astrophys. J. 836, 152
- DebatedBanik et al. (2024). Strong constraints on the gravitational law from Gaia DR3 wide binaries. Mon. Not. Roy. Astron. Soc. 527, 4573
- DebatedPittordis & Sutherland (2022). Wide Binaries from GAIA EDR3: preference for GR over MOND?
- EstablishedDesmond (2025). Modified Newtonian Dynamics: Observational Successes and Failures
Last reviewed May 17, 2026
Spotted an error? Have a source to add?
Prefer email?
You can also send a prefilled email with the variant URL already filled in.