To be cored or not to be cored, that is the question....

Dark matter density profiles in simulations tend to be "cusped" - they increase very rapidly towards their centre. Unfortunately reality tends to disagree, with at least some galaxies having cored profiles : below some radius, the dark matter density seems to be roughly flat. Measurements of the density profile are difficult, relying on precision measurements of the motions of stars and gas, but they can be done, and the results are pretty clear. Assuming that dark matter exists at all, that is, and that the standard model of cold dark matter (CDM) is correct. Modifications to CDM are possible but they tend to make all dwarves have cored profiles, which is rather extreme.

In this very interesting paper submitted to MNRAS (not yet accepted), the authors show that at a given total mass, the central density of the dark matter is sufficient to determine if there's a core or a cusp. This still needs good data, but it's much easier to get than a whole density profile. They demonstrate that it gives results in good agreement with previous findings, though their modelling is surely the weakest section of the paper. I'm not qualified to say if their model is sensible, but they seem quite careful to point out potential weaknesses, address concerns where possible, and highlight where further research is needed. They're limited to a very small sample (16) of dwarves in the Local Group, but even so their findings strongly suggest this is one to watch.

According to this research, a galaxy's inner dark matter is either cored or cusped depending on its star formation history. How does this work if dark matter can only interact with normal matter through gravity ? Well, star formation is known to be able to disturb gas, even removing it entirely from dwarf galaxies, e.g. through supernovae explosions and stellar winds of hot, short-lived stars. And the mass of the gas can be significant, so having all that gas moving around effectively drags the dark matter around too. Each incidence of star formation doesn't do much, but if it goes on for long enough, it can slosh enough gas around to destroy the dark matter cusp.

The data here seems remarkably clear. Yes, it's a small sample, and yes there are uncertainties. But the difference in core or cusp based on star formation history is awfully good, and still with enough scatter that I wouldn't be suspicious. It's compelling stuff.

Even more interesting, it's tough to explain this in theories of modified gravity. Most claims of refuting modified gravity tend to run into the problem of the External Field Effect of MOND, where proximity to a large nearby mass (i.e. another galaxy) sends everything back into the Newtonian regime. For example, recent claims that dwarf satellites lie off the standard radial acceleration relation have intriguingly noted that their acceleration might depend on that of their parent galaxy. And that galaxy without dark matter has been criticised since the main study didn't account for the EFE correctly.

So what you want, really, are two identical galaxies in an identical environment. If MOND (or other non-CDM theories) are correct, then such objects ought to have very similar dynamical properties. If CDM is correct, then it's at least possible that they won't - star formation is one way to change the dark matter distribution, but there are others. It's not guaranteed, but possible that the two galaxies would have different dynamics, because the dark matter is essentially independent of the visible matter. If, however, MOND or some such is correct, then it shouldn't be possible at all : gravity is gravity, and if the mass and mass distribution are the same, then the dynamics should be the same.

Remarkably, such an ideal test system may have been found. The authors note that for the Draco and Carina dwarfs :

These two galaxies require different dynamical mass profiles for almost the same radial light profile. This is a challenge not only for MOND, but for any weak-field gravity theory that seeks to fully explain DM.

There is of course a caveat. They could have different dynamics if at least one of them wasn't in equilibrium. There is some observational evidence for this for Carina, but it's weak, and models, they say, suggest that MOND couldn't explain it anyway. It's just about possible that these two galaxies actually have very different orbits, but to get a result that would be compatible with MOND "is likely to require significant fine-tuning". And the latest data from Gaia suggest that if anything, it's Draco which should be experiencing the greatest disruption, yet none is evident. More models are definitely needed, but it's undoubtedly one to watch.

http://adsabs.harvard.edu/abs/2018arXiv180806634R