I am surprised you find this surprising. There are many studies based upon the Sloan Digital Survey that characterized redshift-dependant effects – this one http://arxiv.org/abs/astro-ph/0408578 is a good example: To make any sense at all of the eigenspectra, they had to first bin the quasars by redshift:

They go on to state:

So it is easy to write off the redshift dependence of the quasar survey as evolutionary effects. It is also easy for a geocentrist to insist this is evidence the earth is the center of everything. I think both conclusions are completely wrong. Why? Next sentence:

The Baldwin effect! The propensity for spectral line width to get smaller with increasing redshift distance. This effect is present is every reshift survey that is not comparing apples and oranges. Why should spectral line widths get smaller with increasing distance? Line widths should not evolve, even if luminosity does. They don’t.

Part of the data reduction package for analyzing distant spectra are the k corrections. These correct for chromatic aberration caused by lenses and shift the spectrum to the local reference frame. They also correct for chromatic dispersion inherent in Doppler redshifting. Remove the frequency dispersion term from the k corrections, and the Baldwin effect disappears (in fact, the line widths broaden, which is what should happen to light traveling for eons – Zeeman and related effects).

Removing the Doppler frequency dispersion term is valid if and only if most of the frequency shift is non-Doppler. This would mean the universe is not expanding.

[quote"Astronomer"]Imagine if I said that green m&ms in a jar tended to hang out near red m&ms…[/quote]
Good analogy, but to understand my point, we need a different one:
If I take my bag of M&Ms and look at a handful of them under a microscope, I might discover the tiny bugs like the green M&Ms better than the red ones, and conclude that maybe we should not eat the red M&Ms either. (Come to think of it, a better conclusion might be that we should not eat the green ones covered with bugs.)

Anyway, You decide to repeat my experiment by scattering a bag of M&Ms across an anthill. You might reasonably conclude the bugs do not give a damn about the color. We are both right in our conclusions, but wrong to draw inferences that stretch beyond our observational boundaries. The large surveys do not include the observational detail Arp is evaluating.

Why?
[/quote]
Because every time we have opened our eyes a little further, the bandwidth of ‘evolutionary’ trends is extended. One obvious example is red/blue galaxy ratios. A few decades ago, we could only see blue galaxies in extended space. It was quite reasonably assumed blue galaxies evolve into red ones.

With improved optics, it has become clear red galaxies extend into the past just as blue ones do, and another ‘geocentric’ effect is found, because eventually the red/blue count inverts again. This can also be explained as an artifact of a small intrinsic redshift component in blue galaxies:

Assume blue galaxies have an average an intrinsic redshift of ~0.16 with an SD ~ 0.06. Also assume the red/blue galaxy ratio is about 5/4. Finally assume blue galaxies are more likely to be centrally located in clusters. Now look what happens:

Almost all of the blue galaxies in our own galaxy cluster have intrinsic redshifts, so we assume they are more distant “field” galaxies not associated with any clusters. As we count galaxies just beyond our local cluster, we included the ‘field galaxies’ that are really redshift displaced members of our own cluster, so our count of blue galaxies is too high. Also, in the nearby galaxy clusters, the random intrinsic redshifts in the blue galaxy causes us to assume they are not located near the red shift centers of these clusters, so the centers appear to be dominated by red galaxies.

At increasing distances, the intrinsic redshift of the blue galaxies becomes less and less significant compared to the cosmic redshift factor. Now we observe the blue galaxies dominate the centers of clusters, but since we are no longer erroneously including a high percentage of misplaced ‘field’ galaxies, the ratio of red to blue galaxies increases.

I have just described the observation known as the Butcher-Oemler effect as an artifact, an optical illusion caused by the type of intrinsic redshifts that Russell has so carefully quantified. Can you tell me why blue galaxies would have dominated galaxy cores in the early universe, and then suddenly decided to leap out of cluster centers and become ‘field’ galaxies?

__________________
jwj

It's a big universe out there...is it really unwinding, really burning out?