Ahead of my fourth and final lecture, a crude but useful rendering of the so-called "Great Plane of Andromeda". With all these lectures I'm picking topics I already know about (obviously), but I also go back to the source literature to make sure I've got the details right. In this case this has resulted in a "the hell ?" moment.
Ibata et al. 2013 claim they've discovered a bunch of dwarf galaxies around Andromeda which are rotating in a narrow plane. Such planes would be a challenge to standard models, which predict that satellite galaxies should be in roughly spherical clouds. Such a plane definitely exists around the Milky Way, with certainty. Here they claim there's a similar plane around Andromeda with 99.998% confidence (more on this in a future post). Doubts had already been sown for me by a seminar last year by one of the leading experts on satellite planes, which had the unintended consequence of persuading me that the Andromeda plane is not as astonishing as is sometimes claimed (the speaker was trying to do the opposite !), but when I finally got around to looking at this in detail.... urrgh.
The problem is the way they select the galaxies in the plane is bloody weird. For the Milky Way, all the galaxies are in a plane, so there's no need for any selection effects. Here, they select galaxies which look like they'd be in a narrow band across the sky as seen from Andromeda, without initial reference to the true 3D position. So they select things based on projected position, not true position. Worse, they do so explicitly by searching for the narrowest structure the statistical algorithm can find. It's hardly surprising that they found something !!!
EDIT : It would be far more convincing if they selected by some other, physical parameter (luminosity, for example) and then demonstrated that, say, all the brightest galaxies are in a plane. As it is, what they've done is pretty much equivalent to saying, "all the galaxies in a plane are in a plane".
The gif shows first the entire satellite galaxy population (grey) around Andromeda (white), then the galaxies in the supposed plane (red) and out of the plane (blue). I defy anyone to claim they can spot a plane without this colour-coding. Yet true 3D position is what's physically meaningful, not sky position.
They also claim that this structure shows evidence of rotation from the velocities of the galaxies (not shown here). That's a bit better - I'd buy the claim that the narrowest structure around Andromeda is rotating. But if you select by rotation, there are other galaxies that should be included as well, which would make the plane an awful lot thicker. As far as I can tell, there's absolutely no physical justification for selecting these particular galaxies at all - they do so solely because they form a thin plane. The fact that you can select a thin structure from an isotropic cloud doesn't really mean that such a structure has physical significance, because you could equally select many other such structures which have very similar physical parameters.
I have yet to read the more detailed follow-up paper but if I was refereeing the Ibata paper I'd have rejected it. I'm not wholly convinced that there isn't a plane of satellites around Andromeda, but I think the evidence is far too marginal to justify these very strong anti-standard model claims that these structures are often used for.
Right, that's my re-emergence for the day, now I submerge again into the world of PowerPoint....
confused
ReplyDeletegoo.gl/CBsdZf