Gaia’s Milky Way Rotation Curve Favours Dark Matter Over MOND

Using millions of highly accurate stellar-motion measurements from Gaia, a new study contends that the Milky Way’s rotation favours unseen mass over altered gravity.

Gaia’s new map shifts the balance

Rotation curves show the orbital speeds of stars at varying distances from a galaxy’s centre. For decades, observations of many galaxies have revealed “flat” curves, with speeds remaining almost unchanged at large distances. Such flatness implies that there is more mass than can be accounted for by starlight and gas alone. This additional material is usually represented as an enormous, spherical dark-matter halo.

Interpreting our own galaxy is more difficult because we observe it from inside. Earlier maps suggested an almost flat rotation profile, but recent Gaia data releases have made the view clearer. Several research groups, including Jiao and colleagues in a 2023 study, now find that orbital velocity declines steadily beyond about 15,000 light-years from the centre. The reduction is around 3.5 kilometres per second for every extra 3,200 light-years. This behaviour continues over more than 30,000 light-years, and independent teams using different tracers and techniques have confirmed it.

Gaia reveals a clean, sustained fall in rotation speed in the outer Milky Way, replacing the old idea of a strictly flat curve.

The result is important because it provides a clear test between two competing explanations: a dark-matter halo under the standard framework, or modified-gravity theories designed to explain galactic motion without invisible mass.

Why Milky Way rotation curves matter

With classical gravity and visible matter alone, orbital speeds should decrease at large radii. The flat curves seen in many galaxies therefore prompted a major reassessment. Two broad solutions followed. One introduces extra mass as a halo of non-luminous particles. The other, MOND (Modified Newtonian Dynamics), changes gravity at extremely low accelerations, using a constant usually denoted by a0. Flat rotation curves can often be described by MOND with one nearly universal a0. A persistent decline, however, is more challenging for that model.

The standard model with hidden mass reproduces the decline

Even Coquery and Alain Blanchard constructed a detailed Milky Way mass model incorporating a central bulge, a stellar disc and a gas disc, all with measured masses and shapes. On their own, these visible components cannot account for the outer regions. The researchers therefore included a dark-matter halo with realistic properties taken from the standard cosmological framework.

When the halo density profile is varied within accepted limits, the model matches Gaia’s measured fall in rotation speed, particularly beyond 50,000 light-years. Achieving this agreement does not depend on unusual assumptions about the galaxy’s visible matter.

Their best-fitting model gives a total mass close to 4.28×10^11 solar masses. This is well within estimates obtained from satellite orbits, stellar streams and the kinematics of halo stars. Neither the halo’s concentration nor its size is extreme. Instead, it provides a credible reservoir of additional mass that progressively influences the velocity field at larger distances.

A conventional halo fit delivers a Milky Way mass of about 428 billion Suns and naturally produces the observed decline in speed.

• The observed gradient in the outer disc is approximately −3.5 km/s per kiloparsec.
• The Sun lies near 8.2 kiloparsecs from the centre, within the region where the decline begins.
• Contributions from the bulge, disc and gas remain close to independent estimates based on star counts and emission maps.
• The halo properties remain compatible with simulations and limits derived from satellite galaxies.

Modified gravity encounters difficulties

MOND has remained attractive because it can describe flat rotation curves using a single low-acceleration scale. Gaia’s Milky Way measurements impose a more demanding test. With standard bulge, gas and disc values, and using commonly adopted MOND versions, predicted velocities do not decline in the way Gaia observes. The closest fit under these conditions requires an a0 far above the values that successfully fit other galaxies. Even allowing broad uncertainties does not remove the discrepancy.

Even flexibility brings severe trade-offs

The authors next carried out a fully flexible Markov chain Monte Carlo search. They allowed the masses of stars and gas to vary substantially, relaxed the disc’s thickness and scale size, and permitted a0 to vary freely. Their aim was straightforward: to determine whether any realistic combination could reproduce Gaia’s declining curve.

The analysis does identify a mathematical fit, but only with substantial costs. The stellar disc would have to be about three times heavier than conventional estimates, exceeding 100 billion solar masses. That would clash with star counts, stellar-population models and independent dynamical measurements. Meanwhile, the a0 value that brings MOND closest to the observations becomes exceptionally low, approaching zero in some chains. In effect, this removes the modification that defines the theory.

In simple terms, MOND can only come close to Gaia’s trend by distorting fundamental Milky Way properties well beyond believable limits, or by forcing its central parameter into a range that defeats its purpose.

Aspect	Dark-matter halo	Modified gravity (MOND)
Match to declining speeds	Achieved with realistic halo profile	Poor with standard parameters
Stellar disc mass needed	Near literature values	~3× higher than observations
Key parameter behaviour	No special tuning	a0 shifts to unrealistic ranges
Consistency with other data	Aligned with streams and satellites	Conflicts with independent constraints

What could still mislead researchers

Rotation curves can be affected by movements that are not perfectly circular. The central bar creates streaming motions, while spiral arms disturb both gas and stars. At large radii, the disc is warped and flared. Asymmetric drift also changes the behaviour of stellar tracers relative to gas. Calibration further depends on the Sun’s exact distance from, and speed relative to, the galactic centre.

Recent studies have considered these systematic effects. Different teams have used a range of tracers and corrections, yet the declining pattern remains across methods, increasing confidence in the finding. Nevertheless, improved treatment of non-circular motions and selection effects should narrow the uncertainties in forthcoming data releases.

Why the result matters beyond theory labels

Dark-matter halos are more than a way of balancing the mass budget. Their form determines how satellite galaxies fall towards the Milky Way and are disrupted. They influence expectations for substructure capable of lensing background stars and galaxies. They also help set local targets for direct-detection experiments on Earth, which depend on the density and velocity distribution of particles near the Sun.

A falling rotation curve offers clues about how halo density varies with radius. This can refine estimates of local dark-matter density, an essential input for detectors. It also changes predictions for the trajectories of long, narrow stellar streams including GD-1 and Palomar 5. These streams can in turn probe the halo’s clumpiness and its history of growth.

What to watch next

• Gaia’s next data release will provide longer observational baselines, increasing velocity accuracy for faint stars at great distances.
• New 21-cm surveys will produce clearer maps of outer gas, helping to distinguish circular motion from streaming.
• Very-long-baseline maser observations will improve distance and velocity anchors towards the inner galaxy.
• Forthcoming wide-area surveys will follow more stellar streams, placing tighter limits on halo mass and shape.

Helpful context and definitions

a0 is the acceleration scale at which MOND departs from classical gravity. Its commonly used value is based on fits to numerous spiral galaxies. If one a0 applies everywhere, that would support a universal modification to gravity. If separate systems need different a0 values, the proposal loses predictive strength.

Halo mass estimates differ because they depend on which tracers are measured and how far from the galaxy observations extend. The 4.28×10^11 solar-mass value in this case describes the mass inside the region sampled by Gaia’s rotation data and the adopted model assumptions. Estimates including very distant satellites can produce higher totals because they examine a larger fraction of the halo.

A useful mental model

Imagine the Milky Way’s visible disc as the exposed tip of an iceberg. Close to the centre, luminous matter supplies much of the gravitational weight. Further out, that contribution becomes weaker. If the rotation curve remains flat, extra mass must take over. If it drops gradually, extra mass is still present but distributed in a manner that allows speeds to fall. Gaia indicates that the second picture better describes our galaxy.

Students and enthusiasts can try a straightforward exercise: use a published rotation curve, subtract the calculated contribution of stars and gas, then investigate the halo density profile required to explain what remains. Varying the disc mass within observational limits shows how strongly the outer curve continues to require unseen mass.