The Milky Way's Outer Rotation Curve Is Falling. Dark Matter and MOND May Both Have Predicted Wrong
Gaia's measurement of the Milky Way's outer rotation curve looks like a Keplerian decline, while both dark matter halos and MOND expect it to stay flat — both sides need tuned parameters or external field effects to survive.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
A flat curve is the starting point for dark matter
What Vera Rubin and others observed is this: velocities in the outskirts of galaxies do not fall with radius, they stay roughly constant. By Kepler's laws, at large radii nearly all the mass is enclosed, so velocity should decay like 1/√r; but a constant velocity means the enclosed gravitational mass must keep increasing with radius. Visible luminosity is falling off while gravitational influence is not, and that mismatch is one of the clearest reasons dark matter became central to modern astrophysics. The standard explanation is that a larger dark matter halo wraps around the visible disk, emitting almost no light yet continuously contributing enclosed mass, which keeps the rotation curve flat.
— Brian KeatingWe are measuring the pitch from inside the stands
Measuring the Milky Way has one built-in disadvantage: we are inside it. Keating's analogy is inferring the shape of an entire football pitch from inside the stands. Gaia provides precise measurements of stellar positions and motions, but it does not have a button labeled "true Milky Way rotation curve." To get the curve you must first choose a model, correct for asymmetric drift, assume equilibrium and symmetry, and then invert to circular velocity. The instrument provides the data; the pipeline tells us what those data points mean — revolutionary conclusions require looking at both chapters together.
— Brian KeatingAsymmetric drift matters most in the outskirts
Stars do not travel on perfect circular orbits; they oscillate radially and vertically, which makes the mean azimuthal velocity of a stellar population lower than the circular velocity of the gravitational field. That difference is called asymmetric drift. Recovering the circular velocity requires modeling the random motions. The problem is that the correction matters most precisely in the sparse outer regions, which is exactly where the enclosed mass is largest and the controversy is greatest. The analysis infers a velocity decline of about 30 km/s between 19.5 and 26.5 kiloparsecs from the center, with a fitted outer slope of minus 0.47 plus or minus 0.15.
— Brian KeatingTurning a curve into mass rests entirely on geometric assumptions
If the circular velocity really declines as 1/√r, then velocity squared declines as 1/r, and plugging that into the circular velocity equation gives an approximately constant enclosed spherical-equivalent mass — walking outward yet inferring almost no additional gravitational mass, which is awkward for a large, extended dark matter halo that should keep adding enclosed mass. But note the phrase "spherical equivalent": the Milky Way has a disk, gas, a bulge, a warp, and a three-dimensional halo, and turning one curve into a mass requires a great deal of geometric insight.
— Brian Keating260 billion solar masses is below most independent estimates
The analysis gives a mean total mass estimate of about 2.6×10^11 solar masses, i.e. 260 billion solar masses, including stars, gas, and dust — below many independent estimates approaching one trillion solar masses. That number comes from fitting the measured rotation curve and extrapolating the mass model beyond the region directly constrained by stars. Gaia's rotation curve only constrains roughly 9 to 27 kiloparsecs, while the Milky Way's halo may extend to about 200 kiloparsecs, so most of the mass comes from the fitted model rather than direct tracers. Observation and extrapolation are not synonyms.
— Brian KeatingThere is no radius in MOND's formula
MOND holds that below a certain characteristic acceleration, the effective dynamics deviate from Newtonian relations, and in the MOND regime the acceleration is approximately the Newtonian acceleration times the square root of the MOND scale A0. In the low-acceleration limit, velocity to the fourth power equals Newton's constant times the baryonic mass times A0 — there is no radius in the equation. That is exactly why MOND naturally yields asymptotically flat rotation curves, and also its fragility. If the Milky Way's outer curve really is a Keplerian decline, MOND likewise expects something flatter and did not find it. Dark matter and MOND reach a flat curve through completely different physics, and a robust Keplerian decline would challenge both at once.
— Brian KeatingDefects lead to new laws, not to who won
Keating says defects lead to new laws — that is what he teaches his students. A Keplerian decline does not immediately falsify all versions of MOND: MOND is nonlinear, external gravitational fields can affect a galaxy's internal dynamics and change the Milky Way's outer behavior, and the external field effect can produce a decline. So the defensible conclusion is not that Gaia killed MOND, but that a robust Keplerian curve would create tension with the isolated prediction and require external gravitational fields or other modifications to do a lot of quantitative work. The standard for a scientific test is whether a theory can fit the measured curve with independent, justified parameters, not whether it can tell a nice story after the fact.
— Brian KeatingDistance errors, selection effects, and the warp all interfere
Distance errors change both a star's inferred position and its tangential velocity; the asymmetric drift correction depends on tracer density and velocity dispersion; selection effects change which stars enter the sample; the Milky Way's warp violates simple disk geometry; and the Sagittarius dwarf galaxy and the Large Magellanic Cloud also drive non-circular motions, acting like external gravitational mass. These tracers become sparse at large distances, and one must also question whether the stars are truly in equilibrium. In the Gaia analysis, neglected dynamical terms and the systematic error budget grow at the outermost radii. None of this proves the decline is fake, but it means that smooth curve may be masking a messy galaxy.
— Brian KeatingAll three explanations are still alive
Keating offers three still-viable possibilities. First, the decline is real but mild, and the Milky Way has a lighter or more concentrated halo than older models — both dark matter and MOND can survive by tuning parameters. Second, the decline is real but exaggerated, affected by systematics and non-equilibrium — the least glamorous answer, which is why scientists must take it seriously. Third, the outer disk is perturbed enough that the reconstructed curve simply cannot be assumed to be equilibrium circular motion. One pattern, three possible physical stories, and the existing evidence cannot uniquely constrain or choose among them.
— Brian KeatingWhat would actually settle it
Different objects can be used to test it: dynamically colder Cepheids, stellar streams that probe farther out, and globular cluster satellites. Future Gaia data releases will improve astrometry, radio surveys will provide different tracers such as gas, and better theoretical models can explicitly incorporate the warp, the Sagittarius dwarf galaxy effect, and the Large Magellanic Cloud effect. If the same decline is still recovered after combining different methods and different systematics, then you have to listen — it could go from 3 sigma to many sigma. But the Milky Way may be asking a trickier question: not whether dark matter lost or MOND won, but whether we are expecting the wrong curve.
— Brian KeatingIn their own words · checked verbatim
The luminosity of the galaxy fades away, but the gravitational influence does not.
Brian Keating2:02
So it's like trying to infer the shape of a football stadium from inside the bleachers.
Brian Keating3:05
Revolutionary claims need us to keep those chapters together.
Brian Keating4:05
So what's missing from this equation? Radius! There's no radius in there.
Brian Keating8:12
I usually say that flaws lead to new laws.
Brian Keating9:15
The scientific test is whether the theory fits the measured curve with independent, justified parameters, not whether or not we can tell a cool story after seeing it.
Brian Keating9:15
Sometimes the galaxy is telling us about dark matter, sometimes it's telling us that it recently had a close encounter of the third kind.
Brian Keating12:26
Not did Dark Matter lose, not did Mond win, but rather, are we expecting the wrong curve?
Brian Keating13:30
Figures
| Fitted outer slope | minus 0.47 plus or minus 0.15 | 4:05 |
| Independent estimate of the Milky Way's mass | approaching one trillion solar masses | 6:10 |
Glossary
- MOND / Modified Newtonian Dynamics
- The idea that gravity deviates from Newton's laws at low accelerations, removing the need for dark matter particles.
- asymmetric drift
- The difference by which random stellar motions make a population's mean azimuthal velocity lower than the circular velocity.
- external field effect
- A nonlinear MOND effect in which an external gravitational field can alter a galaxy's internal dynamics.
- baryonic mass
- Mass contributed by ordinary matter (stars, gas, dust).
- kiloparsec
- A unit of distance equal to about 3260 light-years.
How to listen
Physicists following dark matter, MOND, or Milky Way dynamics; astronomy enthusiasts; and readers who want a fast take on the field's latest controversy.
The methodological outlook after 13:30 can be skipped; the core conclusions are in the first 13 minutes.