A dark galaxy candidate that's not messing about

Of all the candidate dark galaxies that have come and gone over the years, this new one quite likely deserves to jump straight to the top of the tree.

Apparently there's a dedicated filler project at FAST to search for dark galaxies, though the authors don't say very much about it. It's found "some" candidates, but this one is so good that I guess the others just aren't worth mentioning yet.

What makes it so compelling ? It's isolated. Like, very, very isolated. They don't actually emphasise this much, but a cursory NED search (and examination of WWT images) shows there's bugger all nearby. The object is about 29 Mpc away and there's no other galaxies at all within a 500 kpc search radius, at least with known optical redshift measurements. And that's about as isolated as it's ever going get. And the optical images, as a sanity check, don't show any obvious galaxies anywhere nearby. That immediately makes the popular "debris" explanation incredibly hard to sustain in this case.

The object itself is a hydrogen cloud of about 80 million times the mass of the Sun, a few times larger than the ones I've spent an inordinate amount of time investigating in Virgo. Its spectrum appears to have a hint of a double-horn profile that's characteristic of rotating discs, which is backed up by the position-velocity diagram which looks quite a lot like the standard flat rotation curves which are associated with normal, dark matter-dominated galaxies. Optically, it appears to be completely dark : not just dim, but nothing there at all, whereas typical galaxies with this much hydrogen ought to be readily detectable at optical wavelengths.

This means it's ticking all the boxes : it's detected in radio but not other wavelengths, it's isolated, it's got a flat rotation curve. Numerically its estimated dynamical mass is a few tens of times greater than its hydrogen mass so it would have to be strongly dominated by dark matter. And estimates of the gas density put it well below the threshold above which star formation tends to occur.

In terms of caveats I'm struggling to find any. It can't just be hidden behind a nearby dust cloud because the reddening in this region is negligible. It fits the baryonic Tully-Fisher relation between mass and velocity width for much brighter galaxies. In all honesty it's pretty darn close to the Platonic ideal of a dark galaxy candidate.

Is there anything at all to spoil the excitement ? Not really. I think it's virtually certain that whatever this turns out to be, it'll be something interesting. My only slight concern is that the double-horn shape at detection of a flat rotation curve are both marginal, but that it shows both makes this more convincing. And this marginal nature isn't unexpected for a small galaxy like this one anyway : for low mass galaxies, the gas tends to only just reach the flat parts of the curve.

To me the isolation is the crucial feature here. Higher resolution gas observations ought to be able to pretty decisively determine if it's really rotating or not, and it should be easy enough to get some really deep optical data to make sure there's no low surface brightness component hiding nearby. Since the object is so small and lonely, there's no need for massive levels of surrounding data for this : just point the telescopes and go.

The only thing I can think of that would change the picture is if there were more hydrogen clouds nearby, which would point towards some much larger scale feature. But that is hugely unlikely, because if the gas came from galaxies then at least few ought to be known about, and if it's a large-scale gas feature then there's no way anyone wouldn't report this. I'm scraping the bottom of the barrel pretty hard in an effort to find cautionary notes to end on here.

In terms of excitement levels I... have to rate this one 10/10. I have to. To be blunt, in terms of research directly relevant to me, it's potentially the most exciting thing I've heard about in my whole career. There have been dark galaxy candidates aplenty over the years but none of them come close to matching this one : every single one has particular extenuating circumstances whereas this one has none of that.

The only thing that's keeping me together is the pressing need to debug code and a very very strong habituation to avoid getting excited until everyone else does, in case there's some rudimentary error in the whole thing that I've missed, or the whole thing turns out to be bunk for some reason. That's been known to happen to similar discoveries before, so fingers crosses that this is one that's genuine.

This is exactly what I expected

Today's post is not a paper but a poster (doesn't format well on mobiles so please use an actual PC). I wouldn't normally do anything this preliminary, but we've been waiting long enough for this that I'll make an exception. Keep in mind that content that appears in a poster can change drastically by the time of formal publication, so everything I'm about to say is subject to revision.

To give some necessary context : Robert Minchin and I had this running joke in Arecibo that when set he'd automated a task, he really meant he'd delegated it to the postdoc, i.e., me. Well, it took several years and Robert moved institutions twice, but I finally have my revenge. Mwhahahah !

What's this all about then ? Well, way back in 2012/2013, we published a couple of papers which, among other things, noted the discovery of eight optically dark hydrogen (atomic HI) clouds in the Virgo cluster. These were particularly strange in that they have velocity widths normally associated with rotation-dominated galaxies, but no optical counterparts. They were also pretty isolated, with no obvious nearby parent galaxies to explain them.

What followed for the next decade has been a series of papers looking at the various possible explanations. In the "Flying Snakes" paper, we catalogued other, similar objects, did some simple analytical calculations, and also ran a bunch of simulations to see if we they were more likely to be optically dark galaxies or tidal debris. This was followed up with more, similar but more realistic simulations a little later. And then we tried even more advanced simulations looking at more novel suggestion that the objects were held together by the pressure of the intracluster medium.

Let's cut to the chase. Tidal debris doesn't work for these objects - they are much too far from their parent galaxies and should have dispersed already (clouds with smaller line widths, however, are no problem at all for the tidal debris scenario). And the confining pressure of the ICM just isn't significant.

This leaves the most dramatic explanation : that these objects are indeed rotation dominated, and thus require a substantial amount of dark matter to prevent them from exploding. They are in effect dark galaxies.

... except, this explanation has problems too. We don't really find objects like these except in clusters, and the ram pressure they experience there ought to destroy them almost instantly. So the dark galaxy hypothesis remains at most the best of a bad bunch.

What we've yearned for is data with better spatial resolution. With Arecibo they're just point sources. If we could see their morphology, we could tell whether they had the all-important ordered motions that would truly indicate rotation, while more haphazard motions would mean some other, probably less interesting explanation.

At last, we have it. We put in the proposal to the VLA and got the data back in 2017 (!), but then muggins here left it lying on a shelf ever since. Why ? Basically, stuff kept coming up. Reducing the data is damned hard if you don't know what you're doing and there's always lower-hanging fruit to pick. I mean, sure, I could spend months doing this incredibly difficult thing to get this probably very interesting result, or I could do this easier thing with a guaranteed moderately interesting result instead.

Anyway, Robert has reduced the data and made a shiny poster for the AAS meeting. What did we find ?

Of our eight targets, six were observed. Not sure what happened to the other two but never mind. Of those, three were detected. The non-detection of three is not at all surprising, as the VLA can be hundreds of times less sensitive to low-density gas than Arecibo. So it goes.

Of those three which were detected... (drumroll please)...

Two are galaxies. One is optically dark.

Ho-hum. Exactly as I expected, we found results which have nothing much to do with what were expecting but were interesting all the same - which is exactly why detailed observing proposals are a waste of everyone's time.

One of these, AGESVC1 231, appears to be associated with a very small, compact blue blob. Actually we spotted this blob from the word go, but we can only now say that the HI is associated with it thanks to the higher resolution of the VLA data. Without this, the HI map doesn't really coincide well with the object at all, it's so compact it could be anything. Also the high velocity width would be completely atypical for an object this small, so originally, there was no good reason to think the HI and the blue blob were related. But the match with the VLA data is so good it can't possibly be coincidence.

The other thing the VLA data shows is that the blob appears to be associated with the brightest, narrowest peak of the HI emission, with the high velocity width component being considerably displaced - i.e. a tail. Because of its lower sensitivity, the VLA only detects the narrow, densest component, with the wider, more diffuse component only seen with Arecibo. 

Tentatively, this looks like a dwarf galaxy experiencing ram pressure stripping. The blob is extremely blue (the bluest object in our original sample), which might be because ram pressure can initially compress the gas and enhance star formation before it removes and disperses it. So the object is indeed a galaxy, but whether it has any bearing on the notion of "dark galaxies" is harder to say. It's gas rich but not enormously so. More interesting from this one would be to ascertain the age of the stellar population : given that it's so blue, perhaps it only started forming stars very recently. 

What we're probably seeing here is the narrow velocity width component giving a decent representation of the object's true dynamics. The high width component, which is what marked it out as interesting and made us think the blue bob was unrelated, is likely deceptive : this might exist only because ram pressure is disturbing the gas, not because of rotation. So most likely the dark matter content is nowhere near what we initially guessed, though that needs to be more properly quantified.

The second object, AGESVC1 274, also has an optical counterpart. If you look at Robert's poster I think you'll see why we missed this one : it's truly pathetically faint, and you'd never claim it was associated with the HI based on the lower-resolution Arecibo data. However, we also had more sensitive optical data where this is more clearly visible, so I'm a little surprised I missed this. I'll have to go back and check to see why this one didn't get mentioned.

This object is one of two of our sample which have very low velocity widths. In this case, unlike 231, the VLA and Arecibo spectra are in excellent agreement. In some ways it appears to be a typical though extremely gas rich dwarf galaxy, which puts me in mind of other strange stellar systems. I also wonder if it might classify as an Ultra Diffuse Galaxy. But it really is very gas rich indeed, and certainly has a very low surface brightness. It's a galaxy for sure, but whether this is just an extreme example of the general population or is actually an outlier from the general trends is not yet clear.

The third object, AGESVC1 258, is the most interesting. It's only marginally detected in the VLA, but what we see is a clearly elongated structure. It has no apparent optical emission at all (though maybe I do see a pathetic little blob in the densest part of the gas ?). Now this elongated structure could indicate a disc, but it doesn't show any obvious velocity gradient that would be the smoking-gun signature of rotation. On the other hand, as with 231 unfortunately the VLA is missing a lot of flux compared to the Arecibo data, so we might just not be able to detect the rotation because the thing's so faint. In addition, the density of the detected gas is well below the typical threshold for star formation.

Again tentatively, this object might still be consistent with the dark galaxy hypothesis, but it could also be some really weird debris. If it's debris then it's still a mystery how it survived.

So as usual, this data answers some questions but raises whole new ones. Why are some optically bright but some are dark ? Do they have a common formation mechanism or we looking at completely different objects ? If they are debris then how do they survive and reach such great distances from their parent without dispersing despite their high velocity widths ? How much - if any - dark matter do they have ?

For me this is pretty exciting stuff. It's not a huge amount of new data, but it's more than enough to provoke a whole new inquiry. Watch this space.

When is a galaxy not a galaxy ?

Today's paper is a long one, so I'm going to heavily condense it down to the main result. I should say in advance that, although I think there are a couple of points (literally, just two) which are just gut-wrenchingly wrong, it's an excellent, detailed yet very readable work which is interesting from start to finish.

First, you might remember a couple of oddballs found in the Virgo cluster. The first is SECCO 1, a collection of very young, blue stars without any older component that's apparently just wandering around the cluster nowhere near any plausible parent galaxy. It's got gas, and simulations show it can survive moving slowly over such long distances, but why it only has a young stellar population is a mystery. What started this recent episode of star formation, given that there's bugger all around it ?

The second object is more controversial. It appears to be a case of gas stripping from a ultra diffuse (faint, spread out) galaxy with star formation occurring in the stripped gas. That would be the first time we've seen such a galaxy caught in the act of gas loss, and relatively few of them appear to even have any gas, so this would be a pretty neat result. And it would make sense : lots of UDGs known in clusters but virtually all are gasless, whereas those with gas are a bit more common in other, less dense environments.

But then, another team (actually the same as today's) came along and said that the UDG was actually not in the cluster at all, and that the star forming stuff was embedded in a stream of hydrogen coming from another galaxy.

I'm actually a bit skeptical about this second result. We've seen galaxy distance estimates be revised and revised again, and revised some more... so I wouldn't automatically trust any one claim. And the "stream" of hydrogen looks... well, maybe. It's not definitely wrong, but it's not definitely not wrong either. It looks quite a lot like an artifact of smoothing the data, to me.

Still, there are weird bits of star formation happening in the Virgo cluster. Today's paper is an attempt to tackle such objects a bit more systematically. The authors decided to do a search of the entire Next Generation Virgo Survey data and get a more complete catalogue of such objects, rather than stumbling on them at random as in all previous cases.

They found... a total of six candidates, including those two previous cases. And one of those they rule out, as it's most likely just a bunch of distant background galaxies. So three genuine new objects, bring the total to a staggering five.

Yes, that was slightly sarcastic, but never mind.

As usual with such small number studies, it's unclear if these objects really are all that similar to each other. They're not found in any particular location in the cluster. It's not really known if they have a distinct gas content - two are detected and have similar amounts of neutral hydrogen (a few tens of millions of solar masses), but of the other three, their redshifts are so low that any gas would be hard to distinguish from the much brighter gas in our own Galaxy. Only one is at all a significant non-detection, but if its gas was even just slightly less than the others, it would still be extremely gas rich compared to typical galaxies.

But in other ways they are actually remarkably similar. They're all very blue patches of starlight with no obvious underlying old stellar component - everything seems to have happened in the last hundred million years or even less. Their morphologies also look similar, and their total stellar masses are all a few tens of thousands of solar masses. And they have similar chemical compositions, indicating they weren't formed from primordial gas, but likely from material that was pre-enriched inside another galaxy. Star formation rates are also similar, and if they're not found in the same particular region of the cluster, they do seem to have similar levels of isolation - which is pretty unusual by itself for objects in a cluster.

So, how do you form such an object ? There are various possibilities. None are especially compelling, although one does seem more promising than the rest.

The mechanism I would be naturally drawn towards is that the objects might be very faint galaxies, with their stars bound in a dark matter halo just like much brighter ones. This, however, is the first point where I think the authors commit and aaaaargh noo this is wrong level of mistake. It doesn't help that they cite me in order to point out that there are "few convincing" results to date : this is technically correct, but very much missing the point. Worse is the claim that because the gas is not primordial, the objects themselves cannot be primordial and therefore they cannot be galaxies.

I don't think this is fair. If you want to decide if something is a galaxy, you need to look at its dynamics to infer its dark matter content. Whether or not it's composed of primordial material is secondary, especially given that nobody know much about how chemical enrichment works in objects like these. Two points in particular are that they can't rule out an old stellar population, and they say the stars detected are so blue they're hard to fit with a model even assuming very young ages. So there might be something unknown at work here. As it stands, this is confusing two quite separate issues.

Much more convincing is the data on dynamics which does exist, little though it is, doesn't suggest any hint of the high rotation speeds that would indicate the presence of dark matter. If anything it seems that their velocities are extremely low. Ironically, given the size of the objects, this implies they're not gravitationally self-bound and so are likely transient : they exist for just a short time and then disperse.

But how can this be ? How they reach a high distance from their parent galaxy, suddenly form stars, and then bugger off into void ?

There are two possibilities. One is that they could be tidal dwarf galaxies, formed from material stripped off in gravitational encounters between two or more galaxies. But this has a lot of problems. Tidal dwarfs need to be quite massive to survive the gravitational field from their parent galaxies (otherwise they just disintegrate immediately), and these guys are just too little. They could be from smaller parents, but then it's hard to eject material with the necessary velocity to get them to such isolation. And there's a relationship between mass and metallicity, which these objects would violate : their chemistry suggests bigger parents, not smaller.

A more promising alternative is that these are a new class of stellar system : ram pressure dwarfs. Note the lack of the word "galaxy" here, commendably emphatic in its absence*. We know ram pressure in clusters can push out gas at basically any velocity – this is dependent on the mass of the cluster itself, not that of the galaxy. We also know that star formation can happen in this stripped gas, with star formation initially boosted by the compression, though declining afterwards in the parent as its gas supply is rapidly exhausted.

*Claiming a new type of galaxy is tantamount to claiming you've discovered a new planet in the Solar System, so kudos for them for not doing this.

So this sounds great. This mechanism would allow the big, metal–rich parents to eject gas at high enough velocities to produce these slightly star-forming blobs that appear to be very isolated. Hurrah !

But even this is not without issue. One is the second oh god no why did you say that moment. They note that the objects are only found in clusters as evidence for their ram pressure origin. Well, yes, ram pressure is a distinctly cluster–based process... but they only looked for these objects in clusters ! Who knows if they're found in groups or not - they're not at all easy to spot, and no-one's been looking.

Certainly it's fair to say that they must be produced by some mechanism that does operate in clusters, however. The other issues here might be more difficult : the objects don't seem to be found near to objects with known gas tails where we know stripping is still ongoing, and they estimate that such objects could remain detectable for about half a billion years – a not insignificant amount of time.

There's a couple of possible future options for really nailing down what's going on. First, their search was "visual", meaning they took the data and looked at it. That's fine (although I would have like a lot more details), but doesn't tell us anything about their expected completeness, which could surely be quantified, and search techniques refined. Second, getting molecular gas measurements with ALMA or other instruments would help understand if star formation really is occurring the way they think it is.

So in the end this feels like an incremental result. And it is. But it's important to realise the amount of work that goes into some of these incremental results, which is... not small. A better way to think about it would be that this one interesting result has spawned the potential for yet more interesting results, in a never–ending cataclysmic chain reaction of interest that will eventually consume us all. Or something.

Wibbly-wobbly curvey-wurvey

Fresh from having mastered the art of fitting the baryonic Tully-Fisher relation the hard way, I thought I'd resume a long hiatus from reading papers with a look at this one.

The BTFR is an empirical relation between the rotation speed of a galaxy and its baryonic mass (stars and gas). In logarithmic units, this is a nice straight line with (arguably) no intrinsic scatter. This is rather surprising since even very simple calculations show that star formation and environmental effects ought to disturb the gas and stars - galaxies with low surface brightness (brightness per unit area) ought not to follow the same relation as high surface brightness galaxies, but they do anyway.

It's been argued that this is evidence for alternative theories of gravity. If it's just mass that dictates rotation speed, then it doesn't matter if the mass consists of stars or gas : everything should follow the same relation. Under the standard paradigm of galaxies being dark matter-dominated, this would involve a conspiracy between different components that's comparatively hard to justify.

But while the BTFR has been shown to hold true over a wide range of rotation speeds and masses, for the very largest galaxies this is known not to be the case. Establishing whether this is also the case for the smallest galaxies is harder. Not only are they rare, but to get accurate measurements of the rotation speed requires high resolution observations - and resolving the smallest galaxies of all is naturally more difficult. And since you want the systems to be stable, they need to be isolated. Ideally, you want them to be gas-dominated as well, as it's easier to measure the mass of gas than it is for stars.

So getting a statistically significant sample to probe this end of the BTFR is difficult. That's where this paper comes in. At long last, the authors have obtained just such a sample, of 25 objects. Enough, at least, to say whether galaxies of the lowest mass typically follow the same BTFR as for normal galaxies.

The answer ? They don't. They rotate significantly more slowly than the relation found for more massive galaxies.

The really tricky part comes in demonstrating whether this is consistent with standard model expectations or not. The problem is you can only measure how fast a galaxy is rotating out as far as you can observe its material, whereas the dark matter halo ought to extend much further. To correct for this, they compare with analytic predictions for different dark matter halo density profiles. With enough clever tweaking, it's possible to get a result that agrees really well with numerical simulations.

You may well read a note of skepticism here. Thing is, since we don't know the true distribution of dark matter, there's quite a lot of freedom to choose which one to use to get a result which agrees with simulations. And the corrections are so extreme that these galaxies then have a faster rotation than the BTFR for more massive galaxies predicts. But to be fair, they state their results quite cautiously, emphasising the need for more data. So I have no problem with saying this is an intriguing result that needs to be reported.

That said, I'm more persuaded of the claim that the galaxies are more likely to have a constant density inner "core" than an extremely high-density spike (a.k.a. a "cusp"). Trying to get a self-consistent cusped profile that agrees with observations just doesn't seem to work very well, a well-known problem that might be solved because of all the gas and stars sloshing around in the middle.

The provisional, take-home message, then, is that the BTFR is really more the result of the complicated physics of star formation and galaxy formation than it is something as fundamental as gravity. Naively, while you can reproduce the BTFR for ordinary galaxies using nothing more than dynamics, it looks likely that this is really just an elaborate selection effect : once we go outside the standard range of galaxies we've been accustomed to observing, we don't see the nice neat linear relation any more.

Where does this leave Ultra Diffuse Galaxies, ones of extremely low surface brightness that look to be rotating so slowly they may have no dark matter at all ? The authors say, "dunno". My bet remains that we've simply estimated their velocities incorrectly, but if there is something funky going on with how star formation works at these mass scales, that'd be much more interesting.

The wrong sort of gassy

Back in June last year, a nice little paper showed that it was possible to use stacking to measure the atomic gas content of very distant galaxies indeed. Normally, directly measuring the atomic hydrogen content of galaxies becomes excessively difficult beyond the relatively local Universe. Beyond a couple of hundred megaparasecs it becomes really quite challenging; to reach a gigaparsec is just about possible, but at the extreme limits of technical capabilities. The 21cm emission from hydrogen is just too bloody faint.

So the paper used a standard technique of "stacking" the detections : observing lots and lots of galaxies and averaging all their signals together. In this way, you get the sort-of-but-not really equivalent to spending an insane amount of observing time on an individual galaxy (see link for subtleties). The previous paper didn't do much but establish the main result, which is fair enough, but in this new paper they've gone quite a bit further.

Using new, better data for over 11,000 galaxies, they can now recover a signal from a redshift z=1.4, equivalent to a lookback time of 9 billion years. This is well in the range where we'd expect to see evolution in the typical gas content. Last time they established that, as expected, earlier galaxies have a higher gas fraction than their nearby contemporaries, but now they do two additional things. First, they can split their sample to check the gas content at z=0.74-1.25 and for z=1.25-1.45. Adding a separate sample of very nearby objects means they can plot the evolution of the gas content with no less than three, count 'em, three data points - enough, some would say, to see a trend. Second, they also estimate the molecular gas content, so they can compare how the atomic and molecular fractions evolve.

I'm going to gloss over some pretty hefty caveats. First, not only is stacking itself a subtle and oft-misunderstood procedure, but it's also subject to a host of biases and selection effects. However, their detections are convincing, so I note that only for caution, not as criticism. Potentially more problematically, their estimate of the molecular gas comes not from direct measurement but relationships between star formation rate and molecular gas content established elsewhere. As they note, this too may be biased.

However, let's give them the benefit of the doubt and take their results at face value. It seems that the atomic fraction has been monotonically decreasing over time, dropping rapidly from z=1.4 to 1.0, but then more slowly to the present-day z=0.0 (with the reminder that this is based on three data points). The molecular gas fraction is more complicated : it's lowest at z=0.0, peaks at z=1.0, then decreases a bit at z=1.4.

What's going on here ? Well, the prevailing view is that atomic gas the reservoir of fuel for star formation, but it's molecular gas which is actually where star formation happens. If atomic gas is what's in the tank, then molecular gas is what's in the engine. So we might naively expect the gas to be predominantly molecular during the peak of star formation activity. The problem is that's not what we see here : star formation peaks when the gas is dominated by atomic gas instead.

That's a bit strange, but I wouldn't be overly-concerned that everything is broken just yet. Conversion of molecular gas to stars involves a great deal of highly complex feedback, where the hot young stars and supernovae inject energy back into the gas and helping to slow further conversion into molecular gas and thus stars. It looks like the early rapid drop of atomic gas is balanced by the corresponding, combined increase in molecular gas and stars, so that fits the basic narrative. The hard bit is explaining why star formation activity was higher at earlier times, when there apparently wasn't all that much molecular gas around at all.

There's been some hints that the basic scenario is too simple, that actually the atomic gas itself can be directly involved in star formation. That could well be the case. No galaxy, so far as I'm aware, in the nearby Universe, shows atomic gas fractions as high as the ones at high redshift described here - so things might just work differently. Alternatively, it could be that the molecular content has been underestimated, or that the statistical effects of all this stacking are washing out a plethora of important details. Unfortunately, we're probably not going to be able to remedy this without direct detections of atomic hydrogen as z=1, which isn't going to happen anytime soon... though it's likely to be years (rather than decades) away. So, we'll see.

Nothing to see here

The last few years have seen a lot of hoo-hah about galaxies apparently lacking dark matter, but this may at last be drawing to a rather boring conclusion.

First, we had a whole kerfuffle about two dwarf galaxies, NGC 1052-DF2 and DF4. These seemed to have such a low velocity dispersion that they wouldn't require any additional matter apart from their stars to hold them together. There was some dispute about the magnitude of the velocity dispersion (in my opinion not a very credible one) which went away quite quickly because yes, their velocities really were very low. A much more protracted and complex dispute arose about their distance, which was only settled quite recently and found that yes, they are far away (and lacking dark matter) after all.

Those first discoveries were pure starballs. Then, we've also had the saga of galaxies with gas, for which we can much more easily examine the rotation velocity. They were also so far away that there was no prospect of a distance controversy. But just the other day, despite many arguments in favour of these being a population of weird objects, simulations showed that such objects can't be stable. It's far more likely that there's been a slight measurement error in their inclination angle. I wouldn't personally call that one settled just yet, but it's definitely going in that direction.

There was also bad news for those of us hoping the original objects would turn out to be something truly weird : simulations found a plausible way of producing objects like this through unusual but entirely plausible encounters, specifically requiring direct collisions between galaxies, which are relatively rare. Which is great news for the boring old standard model of cosmology.

Today's paper feels like another nail in the coffin for anyone hoping these objects would turn out to be truly strange. Not the final nail by any means, and it's still just about possible they could burst back out like an enraged vampire, but it's definitely making it harder for any would-be children of the night to go on a killing spree.

It's a nice little paper but I do have to wonder why on earth it made it into Nature. Noting the simulations showing that collisions are a viable explanation, they go out and have a look at the context of DF2 and DF4 on a larger scale. They show they're in a neat line of maybe 11 objects in total, stretching over 2 Mpc. This is nicely consistent with the collisional origin hypothesis.

And honestly... that's it. Yes, it's a nice discovery, but they don't go on - as I assumed they would - to do numerical simulations to back this up. They don't have any new observations of these additional galaxies to see if any of them have anomalously low velocity dispersions. They just report the basic discovery along with their (undeniably) very interesting hypothesis. I mean, I'm glad they reported this, but it feels much more worthy of a letter to MNRAS than to Nature.

As I said though, this is not the final nail. More observations of those new galaxies are essential to determine their distances and velocities, otherwise this is just a chance alignment. More simulations to examine exactly what we expect in a collision similar in quantitative detail (major kudos to the authors for proposing this, which largely offsets publishing it in the wrong journal) would also help, a lot. But it's awfully tempting to declare the mystery... well solved is too strong a word, but shall we say, addressed then ? Yeah, that'll do. Mystery addressed !

There is also perhaps some prospect of a weird afterlife for these oddballs. If they do form through good old-fashioned collisions, they're like teeny-weeny versions of the famous Bullet Cluster. This has caused a lot of controversy because Standard Model enthusiasts have pointed and laughed at MOND believers by saying, "Hahah ! Look, we found a case where dark matter can be separated from normal matter, and you can't do that just by modifying gravity !". To which MOND devotees responded by saying, "Yes it can, you don't understand our theory at all, and look the collision velocity you need is just too dang high." But then the Standard Modellers retorted with, "Yeah, but you don't have a unique relativistic solution, and anyway, the collision velocity is way lower than we though, so pfffffp !".

Or something like that anyway. I tend to heavily favour the Standard Model, but if we had more cases to compare, especially on such dramatically smaller scales like these galaxies, this might give us a much more rigorous comparison. The thing about MOND though it's that it gives such similar results to the dark matter paradigm that one wonders what the point of it is. So, we'll see.

Curiously cloudy

Today's paper is one that fell through the cracks of time from March 2021 all the way to the present day. Meaning, I should have read it a year ago, but didn't.

Returning to my favourite topic of optically dark gas clouds, this is an early science project of one of the SKA pathfinder telescopes. Naturally they chose a nearby galaxy group as a target. It's not an especially interesting group as far as I can tell, but looking at any galaxy group with the higher sensitivity and resolution of a new instrument is always a good idea.

What they found is... not much. None of the galaxies seem to be doing much of anything. Except, there are these two quite hefty HI clouds with no optical counterparts floating around and no obvious reason how they got there. Hooray !

Unfortunately though, God appears to have been very cruel by aligning both of these with optically bright sources. One is right behind a bright foreground star, which must be within our own galaxy and cannot possibly have anything to do with the cloud. The other is in front of a more distant galaxy, likely too distant to have any relation either. But of course this doesn't mean that the clouds don't have their own optical emission and it isn't simply being wiped out by the brighter intervening sources. They try to model and remove these, and don't find anything obvious, but I'd be very cautious about drawing any strong conclusions about this. There's just too much other optical crap in the way to be sure they've removed it all.

There's still a lot the data can say though. One of the clouds looks quite a lot like a standard rotating disc. The other is an elongated blob, but here the velocity gradient is across the shortest axis which is frankly just confusing - a spinning cylinder ? Doesn't make a lot of sense to me.

Combined with their size of a few kpc, the velocity gradients aren't very high. In fact they're low enough that the objects could be gravitationally self-bound with their HI alone. And their HI mass is pretty substantial, comparable to dwarf galaxies. So are these the long-sought "dark galaxies" of popular radio astronomy myth and legend ?

Maybe ! But no explanation is very satisfying, which is why these are really neat objects.

First, the standard go-to explanation for all HI clouds is that they must be tidal debris. This is a perfectly sensible default option. The problem in these cases is that there aren't many galaxies around, and the only others with HI are many hundreds of kpc away and don't show any signs of disturbances. Where are the long HI tails ? You'd expect to see these even if much smaller amounts of gas were removed, but that we see substantial amounts of gas in just these two clouds implies a factor of several more must have been removed - tidal encounters invariably produce streams, not just isolated clouds.

But given that the clouds appear to be self-bound, this may not be such a problem. Perhaps the encounter just took place a long time ago, so that most of the rest of the gas has dispersed. Without some dedicated numerical simulations, and even more sensitive observations to look for streams, we can't really rule out this scenario yet.

On the other hand, if they were dark galaxies, their self-bound nature is very surprising. The clouds we found in Virgo had such high line widths that they directly imply the presence of a dark matter halo to keep them stable, just like for normal galaxies. These objects don't fit that. The authors search existing numerical simulations for analogues, and they do find a few, but those tend to have higher line widths.

I would also note (though the authors don't plot this themselves) that these clouds have such narrow line widths and such high amounts of gas that they deviate from the standard Tully-Fisher relation for normal galaxies. Actually this looks very similar to those controversial galaxies without dark matter, which have been quite a challenge to explain. So it does seem a bit strange to invoke a dark galaxy for objects where there doesn't seem to be much indication of a dark halo. On the other hand, they do fit the standard mass-size relation for normal dark matter-dominated galaxies. 

What's going on ? I don't know, but this opens up quite a bit of parameter space for HI clouds. The Virgo clouds I've spent so much time on really can't be self-bound by their HI because of their high line widths : they'd have to be so compact that they'd have densities so high we'd expect explosive levels of star formation. These clouds, though they have ten times as much gas, are in a regime of size and velocity where they could indeed sit around quietly doing nothing. That they're so massive is really interesting : why haven't we found more objects like this already ? They were in fact already detected with much less sensitive instruments. That kindof implies a special explanation, rather than them being part of a larger population.

My preference would be old tidal debris. But I can't see any good reason why such clouds are found only in this group and not elsewhere, so this is more instinctive than rational. Those bloody intervening stars and galaxies make things especially tricky... yet even if they do have optical counterparts, that wouldn't necessarily make the situation any clearer : objects which are self-bound by ordinary visible matter are just strange, regardless of whether they're purely gas or purely stars. This is definitely a case where the rallying cry of "we need more data !" is absolutely legitimate.

Is it a plane ? No, it's a... uhh... orbital anisotropy, or something

We haven't heard from satellite planes from a while, so let's re-open that old wound and fill it with hornets. 

My stance on satellite planes calcified quite some time ago. Claims that satellite galaxies orbit their parents in narrow planes are, in my view, mainly due to an over-reliance on numerical estimates of statistical significance and an apparent disdain for actually looking at the bloody data. Most such structures don't look like anything very much at all to me, let alone the insurmountable challenge to the standard models of cosmology they're supposed to be. See the previous link for a much more in-depth and angry examination of this.

The major exception is the plane around our own Milky Way galaxy. This one is visually unmistakable, and regardless of whether such planes are indeed common or not, this surely deserves an explanation. Of course, it's possible our own galaxy just happens to be unusual, but this isn't very satisfying.

This paper takes another look and says there's nothing remarkable about this at all.

Frankly, I am not sure I either agree with or fully understand their arguments here. They examine the plane only using the 11 brightest "classical" satellites, noting that the anisotropy is dominated by just one or two objects. Move them around a bit and actually the "plane" becomes nothing very remarkable.

Hmm... maybe. It would have been nice to see an example, but more fundamentally, they don't seem to consider that basically all the more recent satellite galaxies also follow the plane. I think it's unfair for people to add other features like stellar streams as independent evidence of the plane (because if the galaxies are in a plane, then all features resulting from their interactions must be there too), but ignoring the other galaxies seems strange. To be fair, this seems motivated by previous studies which also limited themselves to this subset.

Their other claims feel on firmer (or at least clearer) ground. Earlier studies couldn't reproduce the planes, they say, because of numerical limitations in their simulations. Apparently some satellite galaxies are disrupted through purely numerical effects (though I am not sure what exactly), but they are able to account for this and reproduce systems more closely resembling the Milky Way.

What seems to me their strongest argument is new data from Gaia for proper motions of the galaxies. If the plane was indeed long-lived and stable, as claimed, we'd expect there to be little velocity dispersion perpendicular to its axis of rotation : in fact, they say, the dispersion both parallel and perpendicular is about the same. That's not at all consistent with rotation, and indicates the plane is just a chance alignment. So while people have generally concentrated on what fraction of galaxies in simulations show planes at any one moment, it might be interesting to consider what fraction ever show a plane at any point in their history.

All this is nice, but I can't help looking at the structure and still thinking, "really ? that's normal, is it ?" Just as it would take something dramatic to convince me of the counter-claim that "planes are everywhere !", so it's going to take me a bit more than this present work to convince me that "the Milky Way's plane is nothing special."

Un-missing matter

When you find that the missing matter of a galaxy is itself missing, and then you realise that actually it isn't missing missing matter, it's just the regular missing matter, you're in a right proper linguistic pickle.

Readers are advised to consult this previous post for a full summary. In brief, there's this galaxy which doesn't seem to have the usual dark matter (unseen and hence missing by direct observations) needed to hold it together. Based on the mass of its visible stars and gas, its rotation speed is so slow that it's apparently gravitationally bound without needing any of this additional matter at all. And this just doesn't fit the general picture of galaxy evolution : yes, some systems like this can be produced in interactions, but this object is nowhere near anything else. And a generally overlooked detail is that the slow rotation speed means the galaxy wouldn't have had much time to settle into its apparent stability.

My major worry last time was that the rotation speed of the galaxy might have been underestimated. This is dependent on estimating how inclined the object is along our line of sight : if it's directly edge on, we'd be able to measure its full rotation directly, whereas if it's face on, we wouldn't be able to measure rotation at all; at intermediate inclinations we can apply a correction factor. The observations indicate a low inclination angle, which makes this correction already uncertain, but the very ragged edge of the gas disc used for estimating the angle makes this process extra difficult.

To be fair, the authors of the observational studies have had many good arguments in favour of the higher inclination angle they say is preferred by the data. Indeed, it wasn't until the last paper that I would even have questioned this. And they do admit that the inclination angle is the biggest uncertainty.

Today's paper is by a different group coming at the problem with numerical simulations. They cite a lot of private communication with the original team, so it's nice to see that this dispute is being addressed cooperatively.

You might remember my own extensive efforts to model a galaxy without dark matter that met with assorted dismal failures. Only when I added a nice big dark matter halo, to ensure stable circular orbits, did everything magically click into place. This is nothing very new, as the author's of the latest paper explain - in fact, disc stability was even one of the motivations for dark matter in the first place.

What they do is much as you might expect. They model the galaxy as according to the original observations and see what happens. Which, as I found myself, is that the object tears itself apart in short order. Even if they use a very peculiar dark matter halo the original team say the observations do permit, it just doesn't work - it's a bit better, but not much. And then, lo and behold, if they allow a lower inclination angle and hence higher velocities and a big dark halo, everything works out nicely.

This is an awfully tempting explanation. It's just possible that there is some combination of parameters that would be compatible with the observations where a dark halo is not required, but this needs a lot of fine tuning and feels contrived. In contrast, accepting the inclination angle is wrong and there actually is a dark halo after all is a hell of a lot simpler and more satisfying. There is one other option that the object is not in fact stable, and is in the process of disintegrating, but this would seem to be incredibly unlikely in the case of an isolated object.

Not that we should discount these alternatives altogether, mind you. Remember one of those other claimed galaxies without dark matter, which had a prolonged controversy over its distance. A lower value would have made it a very satisfactorily normal galaxy after all, but in fact it was the more puzzling higher value that won out. So galaxies without dark matter are still, for the time being, very much a thing. But in this particular case, my money would now be quite firmly on the "inclination angle measurement" problem. And given that those other cases may well be tidal objects, this particular controversy just might be wrapping itself up.

Not the One Ring but the wrong ring

I have to confess that I read today's paper by mistake. The title is about "HI debris in the NGC 7232 group", which I misread as NGC 7332. And we have our own observations of that group, so I wanted to see if the new data found anything that we didn't. Well, I suppose technically it does, but only because it's of another group entirely...

This meant I was at first disappointed, and then intrigued, then annoyed, then intrigued, and then mildly annoyed again. But overall I'm interested. In fact I think the authors potentially under-sell their main result, which is always better than the more common opposite case.

Their target is a relatively nearby little group of galaxies, using radio data to study gas loss. Why look at groups ? Massive clusters tend to get all the glory because there are so many galaxies crammed in that there's always tonnes of stuff going on, plus you can observe lots of targets in a single observation - but it's small groups where by far most galaxies tend to hang out. Clusters may be sexy, but that's like saying Instagram influencers are sexy. Perhaps they are, but if you wanted to study a cross-section of humanity, looking on Instagram would be about as poorly representative as selecting people randomly from the public library. Galaxy groups, on the other hand, they are much more like the great thriving mass of Joe Public - it's them you want to study if you want to know about more typical behaviours.

The authors use the fancy new MeerKAT array of 64 antennas to get higher resolution and sensitivity HI observations than has been possibly in this group before. And this does help, substantially, though I should add that the column density sensitivity is still about two orders of magnitude worse than Arecibo was capable of. But such is life.

Anyway, at the heart of the group is a triplet of galaxies - two spirals and one lenticular. They find... a great big HI ring (see their figure 3). This is already really strange, but for some reason they don't comment on the morphology at all. I find that to be a very strange omission, especially considering that the ring isn't small - it's about 100 kpc across. Polar rings, which orbit directly around the plane of a galaxy, are at least sort-of understood, but this isn't one of those : the three galaxies are found all on one side of the ring, embedded within its HI gas - there's no galaxy in the centre of the ring. Not mentioning that the gas is a giant ring is a bit like finding a giraffe and neglecting to mention that it was purple.

Actually, they point out that only two of those galaxies may be associated with the ring material. At the velocity of the third spiral, which is quite a bit different from the other galaxies, they don't detect any obvious HI extensions. So probably there are just two galaxies involved here. But the ring isn't anything remotely like the tail and counter-tail structure seen elsewhere and in umpteen numerical simulations of interacting galaxies.

Connected to the ring is a long HI tail terminating in a big blob. A bit further away in the same direction is another, smaller tail, orthogonal to the first and pointing directly towards (almost intersecting) a lenticular galaxy. So this also, they say, might have been involved in the interaction. Okaaay... yes, it might be, you're not wrong, but this is glossing over a lot of interesting stuff ! How do you get a ring-shaped structure with a neat linear tail and then another tail that's orthogonal to it ? That's just weird ! And lenticular galaxies don't often have gas, so that's of note all by itself.

And the mass of the gas present doesn't really help. They estimate the deficiency of the galaxies and find that overall this system has more gas than expected. While the galaxies themselves have lost significant amounts of gas (though I would dispute the use of the words "vast majority" here), overall, the total mass of HI present is actually somewhat more than the whole system would be expected to contain if it was a bunch of normal, isolated galaxies.

What's going on ? To my mind this is one case where accretion ought to be considered a serious possibility. We know that galaxies in general have star formation rates that require external replenishment of gas, but claims for detecting the accretion material are always fraught with the difficult question as to why we don't see such structures everywhere. Most such circumgalactic material (as it's called) can indeed probably be explained as gas removal through tidal interactions. This is the interpretation the authors exclusively employ here, but I don't think it fits well at all. The morphology is all wrong, the kinematics don't fit (though their velocity resolution is low), and the total mass of gas doesn't fit the picture for gas loss. Surely, this deserves some more exotic considerations.

So yeah, this is a really interesting little system. I'm slightly annoyed that they don't cite any of my papers, more annoyed that they restrict themselves to a single interpretation, but it's a very nice result indeed and clearly we need more MeerKAT.