PDA

View Full Version : Paper on M51-type interactions accepted.



turbo-1
2008-May-16, 03:02 PM
A long time ago in a thread far, far away, some of us were discussing apparent interactions between astronomical bodies with different redshifts. The "More from Arp et al" thread was initiated by antoniseb, and during the discussion he asked (post 1320):


That being said, I am also interested in the actual point of turbo-1's post. Do we have a catelog of interacting galaxies in which we record the redshifts of the larger vs. smaller galaxy? Is it in fact the case that the smaller galaxy always has a higher red shift as turbo-1 implies? I don't know the answer, but if the assertion is correct, I agree it is pretty meaningful.

I won't go into the details here but my first crack at the study yielded some interesting results, summarized here:
http://www.bautforum.com/723310-post1339.html

Very soon, I was joined by Ari and Dave and we decided to study galactic associations that fit the M51 archetype. It took us a couple of years to assemble the catalog and gather the data for the tables (thanks Ari!) and get the paper ready for publication. The paper has been accepted for publication by Astrophysics and Space Science - a peer-reviewed journal published by Springer. We have made a pre-print available on arXiv:
http://arxiv.org/abs/0805.1492

The data tables (too large to publish in print) and annotated images of every galaxy association in the M51 sample and the possible M51 sample are available on our web site:
www.jorcat.com/

We would like to acknowledge the generosity of Mike Petersen, who provided us with a domain name, server space, and his web-design skills in building and maintaining the site. It is much appreciated. Thanks to antoniseb for asking the innocent-sounding question that launched this project.

To anybody with access to decent-sized scopes with spectrometers - look at the redshift numbers of the companion galaxies. Fewer than 20% of them have any published redshifts. Given the potential value of this catalog for the study of galaxy morphology, dynamics, triggered star-formation, etc, it would be wonderful to see these gaps filled. Anybody up for it?

ExpErdMann
2008-May-16, 05:15 PM
Congrats to you, Ari and Dave. That's no small feat! I can't imagine the hours of work that went into this.

A giant step for BAUTkind!

turbo-1
2008-May-16, 05:42 PM
Congrats to you, Ari and Dave. That's no small feat! I can't imagine the hours of work that went into this.

A giant step for BAUTkind!Thank you for the kind words. A prominent observational astronomer told us that we were crazy for taking on such a large project in a non-academic setting. He said that since we had no grad students to enslave, we would have to enslave ourselves. He was right!

There's much more to be done, even with the limited spectroscopy available to us on the companion galaxies, and our second paper is already in the works.

parejkoj
2008-May-16, 05:54 PM
Congratulations to all of you!

If you weren't already aware of it, you might be interested in what Galaxyzoo (http://galaxyzoo.org/) is currently working on. A lot of potential mergers were found during the first stage of galaxyzoo, and they are now doing more detailed classification of these systems. They have a PDF tutorial available (http://www.galaxyzooblog.org/wp-content/uploads/2008/04/tutorial.pdf) if you want to take a look. Eventually there will be a catalog published with these results, which you might want to compare to.

turbo-1
2008-May-16, 06:01 PM
Congratulations to all of you!

If you weren't already aware of it, you might be interested in what Galaxyzoo (http://galaxyzoo.org/) is currently working on. A lot of potential mergers were found during the first stage of galaxyzoo, and they are now doing more detailed classification of these systems. They have a PDF tutorial available (http://www.galaxyzooblog.org/wp-content/uploads/2008/04/tutorial.pdf) if you want to take a look. Eventually there will be a catalog published with these results, which you might want to compare to.I'm quite interested in the Galaxyzoo project. I have attempted to participate through the classification process, but always run into an error. I emailed them with no resolution. You wouldn't think that after staring at multi-band IRSA imagery for hours every day, I'd want to classify galaxies for them, but it seemed like a worth-while endeavor. On behalf of Dave, Ari, and myself, thank you for the kind words.

parejkoj
2008-May-16, 06:24 PM
What error were you getting, and during what stage of the process? Also, when did you try it? There were some bugs early on that have since been smoothed out.

Staring at galaxies is always fun!

Their first two (http://arxiv.org/abs/0803.3247) papers (http://arxiv.org/abs/0804.4483) have been submitted, if you want to see how it is going.

A thought for the three of you: do any of the public telescopes (like SLOOH (http://www.slooh.com/)) have access to spectrographs of any kind? If all you care about are redshifts, you could make do with quite low resolution spectra, so long as their wavelength calibration is decent.

turbo-1
2008-May-16, 06:55 PM
Thanks for the suggestions, parejkoj. I'll have to alert my co-conspirators to the possibility of using publicly-available instruments. Magnitude may be a limiting factor, since many of our companion galaxies are tiny and faint.

The Galaxy zoo error is:
"An Error has occurred

Galaxy Zoo has experienced a technical difficulty. The problem has been logged and we are looking into it. Please make sure you have cookies and javascript enabled. You can keep browsing and classifying but if the problem persists check back soon as we're sure to have the issue resolved."
The error has never resolved, and it appears every time I try to classify a galaxy.

antoniseb
2008-May-16, 07:07 PM
A long time ago in a thread far, far away, some of us were discussing apparent interactions between astronomical bodies with different redshifts. The "More from Arp et al" thread was initiated by antoniseb...

You and I conversed a bit through PM in the last day or two, and I congratulated you there, but let me do it publicly too. Well Done.

turbo-1
2008-May-16, 07:16 PM
Thanks antoniseb. When you posed your question, I was unsure if I could pull together the information that you asked about. Thankfully, Ari and Dave jumped in. They are both far more organized and detail-oriented than I am and they both added organization, structure, and discipline to the project that greatly enhanced the quality of the work. We've got a lot more ahead of us, and the next paper will contain another author whose name you're sure to recognize... Thanks again.

parejkoj
2008-May-16, 08:54 PM
The error has never resolved, and it appears every time I try to classify a galaxy.

Huh... I looked around in the forums (http://www.galaxyzooforum.org), and there were a few other people who had similar problems, some of which got resolved after updates, some apparently not. Have you tried creating a new account, or posting in the forum? I'm sorry I can't help you more than that.

StupendousMan
2008-May-17, 12:10 AM
Very soon, I was joined by Ari and Dave and we decided to study galactic associations that fit the M51 archetype. It took us a couple of years to assemble the catalog and gather the data for the tables (thanks Ari!) and get the paper ready for publication. The paper has been accepted for publication by Astrophysics and Space Science - a peer-reviewed journal published by Springer. We have made a pre-print available on arXiv:
http://arxiv.org/abs/0805.1492


Congratulations!



To anybody with access to decent-sized scopes with spectrometers - look at the redshift numbers of the companion galaxies. Fewer than 20% of them have any published redshifts. Given the potential value of this catalog for the study of galaxy morphology, dynamics, triggered star-formation, etc, it would be wonderful to see these gaps filled. Anybody up for it?

I've very briefly browsed one of the datafiles, and I think there is a way for you do quite a bit of work without any spectroscopy.

Many of the companion galaxies are faint enough -- fainter than 14'th mag -- that they might be measured by the SDSS. Now, the SDSS only covers one quarter of the sky, so it will have measurements for only some of your candidates .... but it's a sizeable fraction.

If the SDSS gives you magnitudes of the companion galaxies in multiple passbands (it observed in five passbands across the visible), then you can use the colors of the companion galaxies to estimate their redshifts. There are a number of slightly different formulas for converting color to redshift, but any one of them should be fine. Go to ADS and look for papers with keywords "SDSS photometric redshift", and you should find plenty.

This really shouldn't be too hard. Make a few queries into the SDSS database to gather information on all the companion galaxies which fall into its region, run the resulting colors through a photometric redshift formula, and *voila* you can compare the companion redshifts to the main galaxy redshifts.

Go for it!

parejkoj
2008-May-17, 02:36 AM
Good point StupendousMan! There are actually two different photometric redshifts already calculated in Skyserver (http://cas.sdss.org/dr6/en/help/docs/algorithm.asp?search=photoz&submit1=Search). Though since most of the galaxies in your sample are likely z<0.05, their SDSS photozs may not be very reliable (low redshift sources don't have much change in spectral shape across the SDSS filters at low-z).

Ari Jokimaki
2008-May-17, 06:02 AM
Thank you all for the kind words and suggestions! This was an ultimate learning experience.


I can't imagine the hours of work that went into this.

Having been in a similar situation yourself, I'm sure you can. :)

I did have a session or two in galaxyzoo, but didn't classify much, somewhere around 100-200 I think. But for this work I looked at thousands of galaxies, and many of them I looked again and again and again... The galaxyzoo newsletter mentioned about the merger project. Might be interesting for us.

I don't know if we have any use for photometric redshifts because I think they're too inaccurate. I have understood that they can be used statistically for very large samples, but when looking at individual objects in a sample of 232 cases, they seem to be too inaccurate for that in my opinion. As an example, I extracted photometric redshifts for 10 objects from SkyServer and compared them to their spectroscopic redshifts. All 10 objects are nearby NGC 450, which is the object that appears when you open SkyServer's navigate-page. Table below gives the object name, spectroscopic redshift, and the two types of photometric redshifts with their errors. One can see that photometric redshifts for this sample of 10 are quite inaccurate, not all of them match the spectrum even if error bars are considered. When we are talking about comparing main galaxy and companion redshifts, they generally have very small differentials, 100 km/s differential in radial velocity (0.0003 in z) might be typical.

We might have a use for photometric redshifts in detecting the grossly discordant redshift "companions" in our sample, though.



Object Spectrum photo-z (D1) err photo-z (CC2) err
NGC 450 0,006 0,005693 0,005382 0,12339 0,064016
SDSS J011534.40-005146.0 0,0056 0,051301 0,025991 0,081658 0,047387
SDSS J011536.91-005710.9 0,0186 0,031732 0,012347 0,037934 0,019779
SDSS J011503.26-005322.4 0,1225 0,089677 0,013508 0,083894 0,015551
SDSS J011444.76-004951.4 0,3499 0,375376 0,024968 0,375938 0,029532
SDSS J011445.43-004737.3 0,0905 0,096752 0,02015 0,082571 0,020185
SDSS J011437.64-005044.6 0,1821 0,144031 0,013259 0,145813 0,015427
SDSS J011439.85-004741.4 0,1843 0,157905 0,009753 0,16186 0,010565
SDSS J011637.36-004903.8 0,116 0,080585 0,01662 0,083938 0,01824
SDSS J011511.50-010039.3 0,1202 0,130957 0,021202 0,130886 0,022357

StupendousMan
2008-May-17, 11:29 AM
We might have a use for photometric redshifts in detecting the grossly discordant redshift "companions" in our sample, though.


Yes, that sounds like the best use for them. If you can reduce your sample of 232 cases to one of 100 cases, that saves you a lot of time doing spectroscopy ....

turbo-1
2008-May-19, 01:41 AM
Dave and Ari turned a light-weight project into a real research program with paper(s) to ensue. Data collection was a grind but that's OK , since it's done. Analysis and expression may not be so time-consuming. We have a new team member and we're going to flog him to the finish line.:)

dgruss23
2008-May-20, 02:54 AM
Dave and Ari turned a light-weight project into a real research program with paper(s) to ensue. Data collection was a grind but that's OK , since it's done. Analysis and expression may not be so time-consuming. We have a new team member and we're going to flog him to the finish line.:)

Ari and Skip deserve a tremendous amount of credit for keeping the project moving forward as my contributions were often slowed by the ridiculously busy schedule I keep.

Tensor
2008-May-21, 12:57 PM
How did I miss this thread. Dave, Skip, Ari, great job. Congratulations.

trinitree88
2008-May-21, 02:32 PM
How did I miss this thread. Dave, Skip, Ari, great job. Congratulations.

I'll second that. Nice work guys. pete:clap:

turbo-1
2008-May-21, 03:20 PM
Thanks, Tensor and trinitree88!

Cougar
2008-May-22, 07:48 PM
A long time ago in a thread far, far away, some of us were discussing apparent interactions between astronomical bodies with different redshifts....

Unsurprisingly, I disagree with the attempted implications of your paper, and especially with the after-the-fact statistical treatment of Section 4.6, BUT I must add my congratulations to all of your efforts, perseverance, and the ultimate publication of your findings.

dgruss23
2008-May-22, 09:54 PM
Unsurprisingly, I disagree with the attempted implications of your paper,

What attempted implications?


and especially with the after-the-fact statistical treatment of Section 4.6,

What do you disagree with in that section?


BUT I must add my congratulations to all of your efforts, perseverance, and the ultimate publication of your findings.

Thanks! :)

turbo-1
2008-May-22, 10:14 PM
Hmm, "attempted implications" sounds very underhanded and suspicious. Perhaps you'd like to share your concerns with us, Cougar. This paper was accepted in a peer-reviewed journal after being reviewed by a referee who made some very helpful and astute suggestions. Perhaps you'd like to critique the paper here in this thread. We would be happy to answer any questions that you might like to pose, and I would be very interested to see just how and why you feel that we have fallen short of your standards.

And thanks for the congratulations, no matter how qualified.:lol:

Cougar
2008-May-23, 02:19 AM
What attempted implications?


Hmm, "attempted implications" sounds very underhanded and suspicious.

Well, your section 4.8 says....



...only 41 (18%) of the companion galaxies have a published redshift.... It should be noted that 17 of the 41.... are “discordant redshift” pairs in which the companion has a redshift at least 1000 km s-1 larger than the main spiral galaxy.

Didn't the compilation of your Catalog start with your prior knowledge of many of those 17 "discordant redshift" pairs? I'm not sure how you end up with a random sample on which to do statistics.

And while 1000 km s-1 might be a good cut-off, how many were greatly more discordant? Huge discrepancies might require much more of an "excuse" for retaining dubious associations in a catalog of associations, at least more than the one you provided:

While chance projections of background galaxies may occur in a sample such as this, we have kept the discordant redshift examples in the catalog because (1) they meet the morphological criteria used to compile the catalog and (2) the extreme incompleteness of the companion redshifts in the catalog and high percentage of discordant redshift pairs suggests that a significant number of associations in the catalog for which there is currently no spectroscopy on the companions may also be discordant redshift systems. If it is assumed that the companions without measured redshifts have the same fraction of discordant redshifts, then an additional 79 discordant redshift systems would be expected when the redshift sampling of the catalog is complete. Therefore keeping the 17 discordant redshift examples in the catalog will allow for better constraints on the statistics of possible chance projections.

To your credit I note that your title is "A Catalogue of M51 type Galaxy Associations" and that there is no mention of discordant redshifts in the Abstract or the Conclusion. However, I do wish you would help stamp out the "early" or "late" type galaxy nomenclature, each of which unfortunately means the exact opposite of what it says. :doh:

turbo-1
2008-May-23, 03:20 AM
Since our sample was selected on the basis of visual appearance of interaction (as was the Arp and Arp-Madore interactions used in the first iteration of this study), the onus is upon you to prove bias if you wish to assert it. Arp and later Arp-Madore looked for apparently-interacting galaxies based on plates in visual bands. If these interactions were later found to be "disproved" by redshift differentials measured later, they were dismissed as chance alignments. We scoured the major catalogs for years to see if we could discover meaningful trends in apparently-interacting associations of the M51-type kind, and we presented our findings in a paper. If you would like to refute the findings, we have made the entire data tables available to you so you can flail away. Have a good time.

Carl_Smith
2008-May-23, 03:37 AM
As one who followed the Arp thread regularly while you were thrashing it all out, but with nothing of value I could contribute, I congratulate you on a job well done, and wish you all the best on filling in the remaining blanks.

I'm looking forward to seeing what your future analysis shows.

Ari Jokimaki
2008-May-23, 05:29 AM
Thank you Carl! :)

-----------------------------------------


Didn't the compilation of your Catalog start with your prior knowledge of many of those 17 "discordant redshift" pairs? I'm not sure how you end up with a random sample on which to do statistics.
We have explained how the sample came about in the paper. Random sample follows from our procedures. Either you haven't read the paper, or you don't understand it, or then you are suggesting that we have been dishonest in our classification phase. Which one is the case here?


And while 1000 km s-1 might be a good cut-off, how many were greatly more discordant? Huge discrepancies might require much more of an "excuse" for retaining dubious associations in a catalog of associations, at least more than the one you provided:
Amount of discordancy is not a factor here. The reason why discordant redshift (DR) cases are in our sample follows from this logical trail of thought:

- most companion galaxies in our sample don't have measured redshift available
- of those cases where companion has redshift, quite large portion are DR cases
- it is to be expected that there are lot of DR cases in the cases that don't have companion redshift
- if we want to treat the sample statistically, we can't reject the known DR cases from our sample because we can't reject the unknown DR cases from the part of the sample where companion redshifts are not available (because without redshifts we don't know which ones are DR cases). If we would reject the known DR cases, part of our sample would be biased.

Either we do statistical treatment only for 41-17 = 24 non-DR cases that have redshifts for both objects, or we keep the DR cases and do statistical treatment for the whole sample.

Furthermore, if we clean the DR cases out, other researchers, who might want to use our sample for some study, might be mislead to think that our whole sample is clean of DR cases, even if it is almost certain that there are DR cases among the cases that don't have companion redshift. If someone wants to remove the known DR cases afterwards, and do something for the sample without the DR cases, it is very simple thing to do.

To sum things up, I think we have to keep the DR cases in the sample for now.


However, I do wish you would help stamp out the "early" or "late" type galaxy nomenclature, each of which unfortunately means the exact opposite of what it says.
We used the common terminology that our references use, we can't help it.


Unsurprisingly, I disagree with the attempted implications of your paper, and especially with the after-the-fact statistical treatment of Section 4.6,...
How do you "know" that we attempted to put some implications in this paper? Do you think you are a mindreader? Just to let you know, we didn't have the chance projection probability calculations there initially, the referee asked us to add them, and then later agreed to our calculations.

What we have attempted here is to publish a catalog of M51 type galaxy associations, nothing more.

RussT
2008-May-23, 05:39 AM
I do wish you would help stamp out the "early" or "late" type galaxy nomenclature, each of which unfortunately means the exact opposite of what it says.

Please, more detail! On "Your" take on 'Early' or "late' type galaxies meaning exactly the opposite....opposite of what?

parejkoj
2008-May-23, 06:30 AM
The history of the Hubble Diagram is where the "early" (elliptical) and "late" (spiral) type galaxies come from, and it is exactly opposite, because Hubble thought that ellipticals turned into spirals (it's sort-of, but not really, the other way around). The names have stuck, but have begun to fall somewhat out of fashion among younger astronomers. But many astronomers still use them in papers and even talks, much to the confusion of students and the public.

dgruss23
2008-May-23, 10:48 AM
Well, your section 4.8 says....



...only 41 (18%) of the companion galaxies have a published redshift.... It should be noted that 17 of the 41.... are “discordant redshift” pairs in which the companion has a redshift at least 1000 km s-1 larger than the main spiral galaxy.
Didn't the compilation of your Catalog start with your prior knowledge of many of those 17 "discordant redshift" pairs? I'm not sure how you end up with a random sample on which to do statistics.

That is discussed in section 2.1 of the paper. We had multiple source catalogs for which we visually inspected images of candidate galaxies. Actually, - and Ari would be able to verify whether or not I'm right - but I believe through the process some of the original discordant redshift M51 systems we discussed here on BAUT were actually rejected or moved to "possible M51".


And while 1000 km s-1 might be a good cut-off, how many were greatly more discordant? Huge discrepancies might require much more of an "excuse" for retaining dubious associations in a catalog of associations, at least more than the one you provided:

Ari did a nice job explaining this. Frankly, I don't think we can satisfy you on this one. You're objecting to us keeping the discordant redshifts in the catalog, when we had no choice for the reasons Ari explained. But we felt obligated to mention that many of the companions are discordant redshift cases because any researchers that wish to utilize the catalog should be aware of all of the charcteristics of the sample. It would have been irresponsible not to mention the high fraction of discordant redshift cases, and had we not done so then we'd get people griping that we were trying to sneak something through. Our criteria are spelled out in section 2. It is better to try and understand what we did.

Cougar
2008-May-23, 03:02 PM
We scoured the major catalogs for years to see if we could discover meaningful trends in apparently-interacting associations of the M51-type kind, and we presented our findings in a paper.

And for that you are to be commended, even though the trends you uncovered may not have been what you hoped. I expect many scientific investigations work out that way.

TomT
2008-May-23, 03:32 PM
The history of the Hubble Diagram is where the "early" (elliptical) and "late" (spiral) type galaxies come from, and it is exactly opposite, because Hubble thought that ellipticals turned into spirals (it's sort-of, but not really, the other way around). The names have stuck, but have begun to fall somewhat out of fashion among younger astronomers. But many astronomers still use them in papers and even talks, much to the confusion of students and the public.

If I understand what you are saying, the current belief is "sort of" that spiral galaxes turn into ellipticals. If spirals are "born" at different times over the ages, wouldn't it be correct to say that spirals are at an "early" phase of the process cycle, and ellipticals at a "late" phase?

TomT
2008-May-23, 03:41 PM
Unsurprisingly, I disagree with the attempted implications of your paper, and especially with the after-the-fact statistical treatment of Section 4.6, BUT I must add my congratulations to all of your efforts, perseverance, and the ultimate publication of your findings.

From what I understand, thousands of galaxies were examined by the present authors, and the authors of the references cited, to get to the M51 type catalog in ths paper. So I would assume that the intial sample size was large enough to be representative of the whole. So I would be interested in yours, or others, opinions on what would be the correct statistical treatment of this catalog regarding redshifts of the parents and companion galaxies.

TomT

turbo-1
2008-May-23, 03:47 PM
And for that you are to be commended, even though the trends you uncovered may not have been what you hoped. I expect many scientific investigations work out that way.

I am not sure what you mean by that. Our purpose was to assemble the most complete catalog possible of apparently-interacting galaxies of the M51 type. You have made reference to the fact that we knew the redshifts of some of the first members of the catalog - you seem to forget that when I undertook that preliminary study in response to antoniseb's question, I had NO idea what I would find, nor (except for a few highly-publicized galaxy associations) did I have a clue what the redshifts of the members were. I selected the galaxy associations by appearance in optical bands ONLY, and when I had mined my initial sources, I looked up their redshifts in NED. It was then that I discovered that most of the companion galaxies had no published spectroscopy. Though there appeared statistically-significant trends in the sample, the sample was too small, so we had to expand it.

With Ari and Dave, the situation was similar. We refined the guidelines by which we would visually select M51 type associations, threw out some of the associations in the initial sample that did not meet the more stringent guidelines, and went to work. I stayed out of NED and HyperLeda and worked with DSS survey images. I would make an initial examination in IRSA using the "reproject" option and study each case in multiple bands (and often in multiple surveys - DSS1 and DSS2). This was slow going, as I could only process about 10 of these associations per hour (perhaps 12 if there were several unambiguous rejects in the list). I would categorize the associations and pass them along to Ari. He would review them and if he disagreed with my assessment or wanted another opinion on some of the associations, we sent them along to Dave (who was often busy, as he has said) and we traded emails until we had come to some sort of agreement about the classification of the objects. When the associations were added to the tables, Ari would then look up the relevant data and populate the data set. All this was designed to be as comprehensive and unbiased as possible. Ari kept a log of our emailed comments about the associations, and when our tables were complete, we went back and reviewed each case again in light of these comments, and looking at the DSS images and others.

About the trends - the trends that we found, including those regarding main galaxy morphology, number of arms, etc were not clear when we were building our tables. It's only after examining the tables that the trends became evident. That's the value of a large catalog compiled under well-defined guidelines. It is our hope that someone in both the Northern and Southern hemispheres will take spectra of the objects in this catalog.

ngc3314
2008-May-23, 04:55 PM
The history of the Hubble Diagram is where the "early" (elliptical) and "late" (spiral) type galaxies come from, and it is exactly opposite, because Hubble thought that ellipticals turned into spirals (it's sort-of, but not really, the other way around). The names have stuck, but have begun to fall somewhat out of fashion among younger astronomers. But many astronomers still use them in papers and even talks, much to the confusion of students and the public.

(Nit: Hubble tuning-fork diagram - "Hubble diagram" started out as a magnitude-redshift relation and is often co-opted into a distance-redshift relation...)

So whatever terms do you young whippersnappers use when finding it necessary to talk about (elliptical plus lenticular/S0 galaxies) versus (spirals plus irregulars) or compare galaxies at various stages along the tuning fork or various T values in de Vaucoulers' system? (I freely admit the inconvenience of having early-type galaxies comprised mostly of late-type stars and neither actually having anything to do with an evolutionary sequence).

Cougar
2008-May-23, 09:30 PM
You're objecting to us keeping the discordant redshifts in the catalog, when we had no choice for the reasons Ari explained.

Well, it's an odd situation. The paper is A Catalogue of M51 type Galaxy Associations. M51 and its companion are close to each other and are gravitationally interacting. In the past, researchers have -- and I would guess most researchers would -- throw out candidate pairs from such a catalog whose redshifts are vastly different, concluding that one galaxy is far in the background of the other and therefore they could not be interacting, so they're not really "M51-type". They only appear to be "associated" because they are both close to the same line of sight.

And I recognize that you authors of this paper are not exactly like "most researchers." :) Further, I don't criticize the inclusion of these disparate redshift pairs in your paper. Having more information is usually better than having less. But to include them as standard entries in your M51-type catalog is rather strange. Putting them in an appendix might have been more appropriate.

Having been a participant in the very lengthy Arp thread, I expect that each of you would have liked your results to have supported the contention that the disparate redshift pairs are actually close and interacting, and the grossly different redshifts are the result of some "intrinsic" effect that we don't know anything about. Unfortunately, as you found, there isn't enough data on these apparent companions, and like many papers, you point to this area where more data is needed. No complaints there!

turbo-1
2008-May-23, 10:07 PM
Well, it's an odd situation. The paper is A Catalogue of M51 type Galaxy Associations. M51 and its companion are close to each other and are gravitationally interacting. In the past, researchers have -- and I would guess most researchers would -- throw out candidate pairs from such a catalog whose redshifts are vastly different, concluding that one galaxy is far in the background of the other and therefore they could not be interacting, so they're not really "M51-type". They only appear to be "associated" because they are both close to the same line of sight.You have missed the point entirely. We have compiled a catalog of optically-selected galaxy pairs that appear to be interacting, and we designed the process to produce the least amount of bias in the compilation. Afterward, we studied the tables, and discovered that our optically-selected samples had some pretty interesting morphological trends in Table 1, and the trends did not follow consistently to the associations in Table 2.

You may object to the inclusion of associations with discordant redshifts, but as Ari and Dave have explained, we have compelling reasons for doing so, not the least of which is the stunning incompleteness of the spectroscopy. Rejecting members based on some threshold redshift differential would result in an arbitrary bias in the sample, which we need to avoid. The morphological trends in our sample seem to support our decision not to split out candidates based on redshift, though you would know that if you have read the paper carefully.

Cougar
2008-May-24, 02:09 AM
You have made reference to the fact that we knew the redshifts of some of the first members of the catalog....

That argument of mine was not very well thought out. Of course your catalog is not a random collection. It has very specific rules for membership. So you compare all the members and see what you can figure out. This was obviously a major effort. And you got a publication out of it. Congratulations.

dgruss23
2008-May-24, 02:10 AM
Well, it's an odd situation. The paper is A Catalogue of M51 type Galaxy Associations. M51 and its companion are close to each other and are gravitationally interacting. In the past, researchers have -- and I would guess most researchers would -- throw out candidate pairs from such a catalog whose redshifts are vastly different, concluding that one galaxy is far in the background of the other and therefore they could not be interacting, so they're not really "M51-type". They only appear to be "associated" because they are both close to the same line of sight.

When you construct a sample, you need to apply a consistent set of selection criteria to the larger candidate sample from which you compile your final sample. For example, part of our selection criteria was that the companion would be 50% or less the size of the main galaxy. When we submitted the paper the referee suggested that we modify our selection criteria by calculating flux ratios from magnitude data. This was a great suggestion. However, we found that only a small fraction of the sample had the necessary magnitude data. Ari would probably know the exact percentage but it was certainly less than 30%. Therefore we could not apply the suggested criteria - so we stuck with the visual criteria described in the paper (see section 2).

As with the magnitude criteria, it would be inappropriate to eliminate the discordant redshift cases from the catalog when 82% of the companions don't have measured redshifts. That would be introducing a selection criterion that cannot be applied to the entire sample. We didn't apply the magnitude criteria for the same reason.


And I recognize that you authors of this paper are not exactly like "most researchers." :) Further, I don't criticize the inclusion of these disparate redshift pairs in your paper. Having more information is usually better than having less. But to include them as standard entries in your M51-type catalog is rather strange. Putting them in an appendix might have been more appropriate.

No Cougar, applying a selection criteria to only 18% of the sample would be strange. You don't apply selection criteria in such a careless fashion. The catalog is a classification based upon visual inspection of images. The criteria selected was consistent and appropriate for the available data.

If redshift data for the companions was complete (or even nearly complete) we could have created a table with M51's that have redshift differentials smaller than 1000 km s-1 (or whatever cutoff is most appropriate) and a second M51 sample with discordant redshifts. But you cannot start separating the sample that way when only 18% of the sample that made the visual inspection cut has the necessary redshift data. That would be ridiculous.


Having been a participant in the very lengthy Arp thread, I expect that each of you would have liked your results to have supported the contention that the disparate redshift pairs are actually close and interacting, and the grossly different redshifts are the result of some "intrinsic" effect that we don't know anything about. Unfortunately, as you found, there isn't enough data on these apparent companions, and like many papers, you point to this area where more data is needed. No complaints there!

Your statements in this post do not show defensibly consistent logic. You say above that there isn't enough data. Yet you claim earlier in the post that the discordant redshift M51 galaxies are not interacting without providing a single shred of independent data to back that up. The redshifts themselves are not proof that the galaxies are not interacting. You need more evidence - evidence that establishes with reasonable certainty the M51 companions are not at the same distance.

Given the sparse nature of the data on most of these M51 companions, you lack that evidence needed to make your assertion that the discordant cases are chance alignments anything more than a re-statement of the standard view of redshifts that everyone understands is the standard view of redshifts.

And if comprehensive data is ever collected on these systems, actual tests that could be applied to the interaction scenario would be possible. For example, if the discordant cases are actually in the background, then the incident of enhanced star formation could be smaller in the discordant cases than in the non-discordant cases. What about direct distance estimates for both the main galaxy and the companion. Many of the companions are ellipticals which could at some point have their distances determined by Fundamental plane or Surface brightness fluctuation methods - when the resolution of telescopes allows the data to be collected. And what about the spectra - are there any peculiarities that contrast the discordant systems from the non-discordant systems?

We need more data to answer these questions. Prematurely ruling out the possibility that the discordant cases are interacting based upon zero data that actually allows one to rule out that possibility only serves to inhibit progress in our understanding of the universe.

Ari Jokimaki
2008-May-24, 03:45 AM
Well, it's an odd situation. The paper is A Catalogue of M51 type Galaxy Associations. M51 and its companion are close to each other and are gravitationally interacting. In the past, researchers have -- and I would guess most researchers would -- throw out candidate pairs from such a catalog whose redshifts are vastly different, concluding that one galaxy is far in the background of the other and therefore they could not be interacting, so they're not really "M51-type". They only appear to be "associated" because they are both close to the same line of sight.
None of our reference catalogs did that, so in that sense there's nothing odd about it. What works on M51 systems are you referring to?


Having been a participant in the very lengthy Arp thread, I expect that each of you would have liked your results to have supported the contention that the disparate redshift pairs are actually close and interacting, and the grossly different redshifts are the result of some "intrinsic" effect that we don't know anything about. Unfortunately, as you found, there isn't enough data on these apparent companions, and like many papers, you point to this area where more data is needed. No complaints there!
What you fail to notice is that this is not the paper we set out to do during the Arp thread. This is only a description of our sample, and has next to nothing to do with discordant redshift issues. We have only now begun to look at this sample from that angle for subsequent work.

And kindly stop guessing our motivations and what we would like or hope, it's quite insulting.

Ari Jokimaki
2008-May-24, 03:55 AM
Actually, - and Ari would be able to verify whether or not I'm right - but I believe through the process some of the original discordant redshift M51 systems we discussed here on BAUT were actually rejected or moved to "possible M51".
Yes, that's true. There are quite a lot of DR cases both in rejected and possible piles.


Ari would probably know the exact percentage but it was certainly less than 30%.
Of our 232 Table 1 cases 40 had K-band data for both objects (which referee suggested to be used) = 17 %.

VanderL
2008-May-24, 05:14 PM
Oy,

Missed this thread by a week! Sorry Skip, Ari and Dave, but congrats for this huge effort! Hopefully this is the beginning of a fruitful line of investigation in an interesting area of research.

Cheers.

turbo-1
2008-May-24, 07:19 PM
Thanks for the good wishes. We hope that others can find uses for the data, too, and that they will help spur the demand for the missing spectroscopy.

Cougar
2008-May-26, 12:51 AM
...you claim earlier in the post that the discordant redshift M51 galaxies are not interacting without providing a single shred of independent data to back that up. The redshifts themselves are not proof that the galaxies are not interacting. You need more evidence - evidence that establishes with reasonable certainty the M51 companions are not at the same distance.

That's why I asked about the cases with extremely discordant redshifts, where the redshift difference "cannot be ignored."

But perhaps we'll get lucky - when the additional spectra trickle in - and find intermediate redshift absorption lines in the "companion" spectrum that are absent from the host's, confirming it as a background object.

turbo-1
2008-May-26, 12:59 AM
Hopefully, the spectra will not "trickle in", but someone will take it upon themselves to fill these huge gaps. I hope that you will not be too offended if some of the companion galaxies that have apparently pulled tidal streams from (or triggered accelerated star formation in) their host galaxies have redshifts that place them impossibly distant from the hosts to be in interaction. You'll have a blast explaining how any secondary evidence for interaction can be safely ignored. It'll be fun!

"When you believe in something that you don't understand, then you suffer. Superstition ain't the way." My Man, Stevie.

dgruss23
2008-May-26, 03:03 AM
That's why I asked about the cases with extremely discordant redshifts, where the redshift difference "cannot be ignored."

But perhaps we'll get lucky - when the additional spectra trickle in - and find intermediate redshift absorption lines in the "companion" spectrum that are absent from the host's, confirming it as a background object.

Certainly multiple lines of evidence will be needed to find a resolution to the matter. Absorption lines will be one of them. We're hoping to put together a list of promising candidates for follow-up study.

parejkoj
2008-May-27, 05:17 AM
To the three paper authors: one thing you might want to try if you are going to look at other objects is using Sky (http://earth.google.com/gallery/kml_listing.html#csky#s1#e10) in Google Earth (http://earth.google.com/): it uses data from SDSS and DSS, and you can do a search by object name and it will take you straight to a color image of that object. Might be faster than using NED. Some of the KML plugins from that page include data overlays from other surveys. There are others floating around the web (and STScI is working on a Hubble archival data overlay, which will be great!).



So whatever terms do you young whippersnappers use when finding it necessary to talk about (elliptical plus lenticular/S0 galaxies) versus (spirals plus irregulars) or compare galaxies at various stages along the tuning fork or various T values in de Vaucoulers' system? (I freely admit the inconvenience of having early-type galaxies comprised mostly of late-type stars and neither actually having anything to do with an evolutionary sequence).

First of all, how do you always manage to find my nits to pick? Yes, I realize you have a license, but still, it's creepy sometimes...

Second of all, I've heard things referred to as elliptical-type, spiral-type and irregular/other (I swear someone used "deformed" or the like for the last grouping) by us whippersnappers. Unfortunately, the best citation I can come up with at the moment is from the first galaxy zoo paper (http://arxiv.org/abs/0804.4483), which you're already familiar with, and which still uses the terms occasionally.

I fully admit that the terms will probably be around until after Hubble's great-great-great-great-etc. grandchildren (does he have any?) are long dead.

But, besides tradition, is there a compelling reason to keep the early/late terms? They are both confusing and misleading and mask a number of subtleties regarding galaxy evolution. I'm not much of a fan of outdated traditions. Guess I shouldn't get in an argument with Tevye...