Module 8: Dark matter · Lesson 8.7 · about 16 min
Could gravity be wrong instead? MOND and friends
MOND changes Newton's law at very weak pulls instead of adding dark matter. It predicts galaxy rotation well but struggles with clusters, the Bullet Cluster and the CMB.
Module 8 · about 16 minutes. Read first: 8.2 Rotation curves and 8.3 The Bullet Cluster. Every piece of evidence for dark matter is gravitational: things move as if there were more mass. There is another possibility: the mass is all there, and our law of gravity is wrong where gravity is very weak. This lesson takes that idea seriously. By the end you will be able to explain what MOND changes, where it works brilliantly and where it struggles.
1. Milgrom’s idea
In 1983 the Israeli physicist Mordehai Milgrom noticed something about rotation curves. They do not stop following Newton at a particular distance; they stop at a particular pull. Wherever the acceleration produced by the visible matter drops below about
the stars move faster than Newton allows. That is a tiny acceleration: it would take a car about 7,000 years to go from 0 to 100 km/h. Milgrom proposed MOND, Modified Newtonian Dynamics: below a0, the real pull is not Newton’s gN but
- gN
- the pull Newton's law gives from the visible mass: G M / r²
- a0
- MOND's one new constant of nature, about 1.2 × 10⁻¹⁰ m/s², found from galaxies
- g
- the pull that actually acts. Between the two limits a smooth "interpolating function" joins them
- Physical meaning
- Where gravity is strong (the Sun's pull on the Earth is about 50 million times a0), nothing changes. Where it is extremely weak, gravity fades with distance as 1/r instead of 1/r², so it stays strong enough to keep outer stars fast.
Far out in a galaxy, put g = v²/r = √(G M a0) / r. The r cancels:
The speed no longer depends on distance: a flat rotation curve, from the visible mass alone.
2. Where MOND shines
That last equation says a galaxy’s mass in stars and gas grows as its rotation speed to the fourth power. Galaxies follow this baryonic Tully-Fisher relation remarkably tightly, over a factor of more than 10,000 in mass, the cousin of the distance tool in Lesson 2.5. In 2016 Stacy McGaugh, Federico Lelli and James Schombert plotted, for 153 galaxies, the pull observed at each point against the pull expected from the stars and gas. All 2,693 points fell on one tight curve, with a0 as the only number.
MOND made real predictions in advance. It said that faint, spread-out “low surface brightness” galaxies, where the pull is below a0 almost everywhere, should show the biggest gap between visible and dynamical mass, before such curves were measured; that turned out right. In the dark matter picture this tight link between visible matter and rotation has to come out of how galaxies form, and simulations are still working to reproduce it in detail.
3. Where MOND struggles
- Clusters. Apply MOND to galaxy clusters and the visible mass (galaxies plus hot gas) still falls short by about a factor of 2. MOND needs extra unseen matter there, for example massive neutrinos.
- The Bullet Cluster (Lesson 8.3). The lensing mass sits with the galaxies, away from the gas that holds most of the atoms. A modified law of gravity alone cannot move the pull away from the mass.
- The CMB and the growth of structure (Lesson 8.4). The high third peak and the growth of lumps from such small ripples need something that clumps like dark matter before recombination.
- Gravitational waves. In 2017 gravitational waves and light from the same neutron star merger, 130 million light-years away, arrived within 2 seconds of each other. That ruled out many relativistic versions of MOND in which the two travel differently.
MOND by itself is only a rule for slow motions; to deal with light bending and the universe as a whole it needs a relativistic theory, as Newton’s gravity needed Einstein’s. Jacob Bekenstein built the first, TeVeS, in 2004; it is now ruled out by the 2017 gravitational waves. In 2021 Constantinos Skordis and Tom Złośnik built AeST, which fits the CMB peaks, but does so with an extra field that behaves like dark matter in the early universe.
A recent test uses wide binary stars, pairs so far apart that their pull on each other is below a0. In 2023 one study using the Gaia satellite claimed a MOND-like boost; others, including a 2024 analysis by Indranil Banik and colleagues, found Newton’s law works and MOND does not. The debate continues.
- 1983 Milgrom publishes MOND in three papers.
- 1984 Bekenstein and Milgrom turn it into a proper theory of Newtonian-style gravity (AQUAL).
- 2004 Bekenstein's TeVeS, the first relativistic version.
- 2006 The Bullet Cluster lensing map.
- 2016 McGaugh, Lelli and Schombert's radial acceleration relation.
- 2017 GW170817: gravitational waves and light arrive together, ruling out TeVeS-like theories.
- 2021 AeST fits the CMB.
An odd coincidence: a0 is close to c H0 / 2π ≈ 1 × 10⁻¹⁰ m/s². Nobody knows whether that means anything.
“Scientists just ignore the alternatives to dark matter.” MOND has been tested for over forty years in hundreds of papers, and its successes in galaxies are taken seriously; any full theory of dark matter has to explain them. But the same evidence that MOND is tested on, clusters, lensing, the CMB and structure, is explained by cold dark matter with one set of numbers. That is why cold dark matter is the mainstream picture, and why MOND remains a minority view that keeps everyone honest.
- MOND (Milgrom 1983) changes Newton’s law below a0 ≈ 1.2 × 10⁻¹⁰ m/s²: there the pull is √(gN a0).
- It predicts flat rotation curves and the baryonic Tully-Fisher relation, v⁴ = G M a0, from visible matter alone.
- It needs unseen mass in clusters (about 2 times), cannot explain the Bullet Cluster’s offset, and needs a dark-matter-like ingredient for the CMB.
- Relativistic versions: TeVeS (ruled out by GW170817) and AeST (fits the CMB with a dark-matter-like field). Cold dark matter remains the mainstream picture.
Check yourself
Sources
- M. Milgrom (1983), A modification of the Newtonian dynamics as a possible alternative to the hidden mass hypothesis, ApJ 270, 365 https://doi.org/10.1086/161130
- S. S. McGaugh, F. Lelli and J. M. Schombert (2016), Radial acceleration relation in rotationally supported galaxies, Physical Review Letters 117, 201101 https://doi.org/10.1103/PhysRevLett.117.201101
- B. Famaey and S. S. McGaugh (2012), Modified Newtonian dynamics (MOND): observational phenomenology and relativistic extensions, Living Reviews in Relativity 15, 10 https://doi.org/10.12942/lrr-2012-10
- J. D. Bekenstein (2004), Relativistic gravitation theory for the modified Newtonian dynamics paradigm, Physical Review D 70, 083509 https://doi.org/10.1103/PhysRevD.70.083509
- C. Skordis and T. Złośnik (2021), New relativistic theory for modified Newtonian dynamics, Physical Review Letters 127, 161302 https://doi.org/10.1103/PhysRevLett.127.161302
- I. Banik et al. (2024), Strong constraints on the gravitational law from Gaia DR3 wide binaries, MNRAS 527, 4573 https://doi.org/10.1093/mnras/stad3393
Your ticks stay in this browser. Save them as a file on the course home.
Updated 2026-11-22. Numbers are rounded for teaching; measured values come from the sources listed in each lesson. Simulations run in your browser; nothing you type is sent anywhere.