The observation that would not go away
In the 1970s, Vera Rubin and others measured how fast galaxies rotate. The outer parts move as if there were far more mass than the stars and gas can account for. The standard answer is dark matter — an invisible halo holding the disk in.
In 1983, Mordehai Milgrom proposed a different reading: the rotation curves are not evidence of missing mass but of a modified force law. Below a characteristic acceleration a₀, the pull of gravity falls off as 1/r rather than 1/r². Flat rotation curves follow immediately, with one parameter and no per-galaxy tuning.
Where it works remarkably well
The radial acceleration relation is the strongest evidence. Across hundreds of galaxies, the observed acceleration tracks the predicted one with very little scatter, over five orders of magnitude. That is a regularity dark-matter simulations had to be tuned to reproduce.
MOND also predicted, before observation, that low-surface-brightness galaxies would rotate faster than they appear to warrant, and it anticipated features of the baryonic Tully–Fisher relation. Predictions that come true before the data are the currency of a good theory.
Clusters of galaxies. MOND still needs roughly twice the mass it accounts for in cluster cores. Dark matter handles clusters without extra parts.
The Bullet Cluster. The mass, mapped by gravitational lensing, sits apart from the hot gas that collided. That separation is natural if the mass is dark matter and awkward if gravity is the only thing being modified.
The cosmic microwave background. The acoustic peaks fit a dark-matter model closely; MOND requires additional components to match them.
Relativistic completion. MOND is a non-relativistic law. TeVeS and similar extensions exist, and are heavily constrained by gravitational-wave speed and lensing data.
Why it is still worth taking seriously
Because the alternative has its own embarrassments, and because MOND's successes are not decorative. A theory that predicts galaxy-scale regularities with one constant is telling you something about the data, even if it is not the final law.
The most likely resolution is a hybrid: dark matter on cluster and cosmic scales, with a modified or emergent gravitational behaviour appearing in the low-acceleration regime. That is not elegant, but neither is the current standoff.
Sources for this page
- Wikipedia, “Modified Newtonian dynamics” — history, predictions, criticismsen.wikipedia.org
- McGaugh, Lelli & Schombert (2016), the radial acceleration relationjournals.aps.org
- Clowe et al. (2006), the Bullet Cluster mass mapiopscience.iop.org
- NASA/ESA, on dark matter and cluster lensingscience.nasa.gov