Time and tide wait for those granted sufficient observing hours on a big telescope

More on those galaxies which appear to lack dark matter. Hot on the heels of the recent simulations showing that dark matter can be tidally removed with little disruption to the stars, here's an observational paper claiming that's exactly what's happening.

This paper looks at one particular galaxy, NGC1052-DF4, which is not quite as famous as its neighbour DF2. However the distance controversy surrounding DF2 doesn't affect DF4 nearly as much, so even if it's quite a lot closer than the supposed 20 Mpc, it would still appear to be a weird object.

The "it's due to tides" theory has a lot going for it. There's no particular reason to expect any significant numbers of galaxies to form without dark matter all by themselves, so to find two in close proximity is suspicious. That they're in a group means tidal encounters are all but inevitable, and indeed, clear signatures of such interactions have already been detected elsewhere in the group. It's an elegant, simple way to explain what's otherwise a thoroughly perplexing observation.

But that doesn't mean the idea is automatically correct. Indeed, I find the results of this latest paper far from convincing.

They proceed in two ways. Using a combination of very deep ground-based and HST imaging, they search for both globular clusters and the diffuse light of the DF4 target galaxy. Globular clusters are handy because they're bright and compact - they should survive a tidal encounter but their overall distribution should give clues that something happened. The diffuse light, directly from the stars of the galaxy itself, is harder to detect but potentially a better diagnostic of tidal encounters.

First, the globular clusters. A few of these were already known and confirmed as being associated with the host galaxy by the gold standard of spectroscopic (velocity) measurements. This is hard to do, so they do the next best thing and do a lot of careful work to define sensible colour ranges to identify new candidates. After all this they find four new possible clusters. And faster than you can say, "small number statistics", here's their figure claiming to identify a signature of a tidal encounter :

The one nearest the bottom of the image is (they say) likely associated with the nearby blue smudge, not DF4. Okay, so that's really three new candidates. Two of them are close to the centre of the DF4 and don't change the distribution at all, so in fact we're down to a single new useful candidate. And the purple swathe ? That identifies the elongation of the "structure" they might have identified.

To be blunt, I often say to people, "I've seen worse claims", and this would be one of those semi-proverbial worse claims. It's highly dependent on two data points and has no 3D information. I'm glad they reported their results, but I don't believe for one second that this supports the claim of a tidal encounter.

On to the direct detection of the stellar light, which I find slightly better but still nowhere near as strong as the authors claim. They reach an impressive sensitivity level, again after a lot of careful work (this time to remove any contaminating light from the nearby brighter galaxies). The characteristic signature of a tidal encounter is a tail/counter-tail structure which they describe as S-shaped. Do they find anything ? Not really. Here's their figure :

Which they seem to think has a very clear S-shape. I looked at it and went, "huh ?", because it doesn't. It does show some hints of weak, broad, double-sided extensions, but it looks nothing like the classic tidal tails. It could be indicative of tidal disruption, but it's bloody faint.

Their surface brightness profile shows a distinct break, which I'll totally accept as real. They also identify a point in the isophotes which shows a "twisting", but this looks extremely weak to me and could be due to anything. The key thing they're missing is any kind of comparison sample : what would these same procedures reveal for isolated galaxies ? This is something we went to great lengths to explore when looking at gas tails, but there's no kind of equivalent analysis here. When an extension is this weak, I'd want to be damn sure it could only be the result of an interaction.

It looks even worse when they plot the isophotal contours :

Look, there might be something there. That it can be seen in different bands is somewhat more convincing, but the structure itself is so weak I'm not sure I'd even bother to report it. There is - maybe - some sort of extension. Is it compatible with a tidal encounter ? Sure. Is it evidence that this is what happened and not some other process ? Absolutely not. A tidal encounter is a great way to explain some very weird observations - I just don't believe that this data is good enough to support that conclusion yet. And there's no way to show that the small amount of stellar disruption (7% of the total stellar mass is in the diffuse extended light) really corresponds to significant levels of dark matter loss.

Finally, there are a lot of mutually-exclusive explanations floating around for these objects. As well as tidal debris and distance uncertainty, another idea is that the mass has been measured incorrectly. In one press release for the latter explanation, the authors were quoted with the Johnsonian phrase of "it's time to move on". But it can't be both of normal mass and the result of a tidal disruption. As long as there are conflicting explanations, "moving on" is exactly what we shouldn't do

The galaxy "missing dark matter" NGC1052-DF4 is undergoing tidal disruption

The existence of long-lived galaxies lacking dark matter represents a challenge to our understanding of how galaxies form. Here, we present evidence that explains the lack of dark matter in one of such galaxies: NGC1052-DF4. Deep optical imaging of the system has detected tidal tails in this object caused by its interaction with its neighbouring galaxy NGC1035.