Martin,
that's not the style in which you can expect any serious communication from me. First you make
such unqualified, "generic" statements:
A):
I have the impression there's a bug in Celestia's database, since it doesn't match any of the databases I tried,
despite the fact that your cited "databases" either were out of date with 1950 coordinates only or they were part of my own sources. My cross-checks with your quoted list gave a perfect agreement with my respective distance values in
deepsky.dsc.
After challenging you to quote specific disagreements, this statement has suddenly gone and I get ONLY this instead:
B)
the problem with your database is it's so full of LARGE STRAIGHT RADIAL lines, it's looking extremely unconvincing. I can't believe the uncertainties are so strong.
+++++++++++++++++++++++++++++++
Either you give me your list with disagreements found (outside reasonable uncertainties) along with a citation of your sources, or we better stop communicating about this issue.
+++++++++++++++++++++++++++++++
It is a familiar effect that the
3d display can enhance inherent SYSTEMATIC uncertainties as compared to the usual 2d projections to the skyplane. Usually in the various distance determination methods used, there are unavoidable underlying assumptions, which can certainly give rise to such biases as you noted. We discussed and explained similar phenomena already much earlier, as I remember. At that time these "artifact patterns" were still much stronger, since a much higher fraction of galaxies did not yet enjoy reliable distances.
But there is NOTHING that can be done about it. It's just an expression of SYSTEMATIC uncertainties inherent e.g. in some distance determination methods.
Let me give you a typical example that is presumably at the root of the problem: notably for
galaxies from within a big cluster, the velocity vector does not only have a radial component (<=> redshift) but also "peculiar", i.e. NON-radial ones due to substantial gravitation effects from the surrounding cluster and neighbors thereof. This effect can be so strong as to leave us with galaxies that are effectively BLUE-shifted, rather than redshifted. For such extreme cases I never used the Hubble law, of course.
Nevertheless, all one can do to extract the distance
via the Hubble law is to assume that the
modulus of such velocity vectors approximates the
radial component as associated with the expansion of the Universe. This can easily account for "artificial patterns". That's life, and astronomers know very well about this problem. Still that's often the basis of distance determinations that are given in the published sources. Of course in such cases I have always tried to use an alternative method that does not need this kind of assumption. But there is NO "free lunch" here, one has to accept other assumptions instead. In my database, always the best method is selected. But unfortunately, alternative methods are not always available!
++++++++++++++++++++++++
For making any progress, I
urgently need at least ALL names of the galaxies along the "string" that you marked in red here
If I know the names, I can trace the distance method used and the name of the data base source. E.g. from the local group? (NED1d database??) .
++++++++++++++++++++++++
The question is whether the galaxies along that string actually ALL belong to a cluster that should be located in the blue "box" of yours? The distance of that "box" is, of course, also uncertain and may actually be significantly smaller, like e.g halfway along that string! If so, one might consider replacing the above standard asumption about the "peculiar" velocity components, by attributing the same distance to all members of that cluster. That would be another assumption which is NOT exact and may have other "cons", of course. But it may look nicer as to a display of galaxy clusters.
Clearly, the radial streaks towards the outer boundary of the database volume, have largest uncertainties! If you don't believe it, pick some of them, find out the methods available for distance determination...and you will be convinced!
There, all we got usually are measurements of the
redshift z. One way of crosschecking is to display the same galaxies NOT versus "distance" but versus redshift z as in my previous celx script z-dist.celx. See whether the "radial" streaks have gone in such a display,
where distance has been substituted by z!? If some streaks are gone, you will know that for them the problem was as I indicated. You can then tell me how you get a more accurate distance JUST from measuring z
According to this WEB list of yours (where is it
published in peer reviewed form!!??)
http://www.atlasoftheuniverse.com/galaclus.htmle.g
the very big Virgo I cluster is associated with 1 single distance of 52 Mly, around which you plotted your blue box. That is a STRONG and presumably wrong assumption, since the members of the virgo cluster have a widely differing redshift!! Accordingly, one would expect a much wider extension along the 3rd dimension (distance) than what your blue boxes indicate. Don't forget, the
basic measurement information that is available is about the 2d transverse extension of the virgo cluster in the sky plane supplemented by measurements of the redshift z!
Fridger
I simply don't care about them and whatever you like it or not.
