A similar dataset to that used for the bivariate luminosity functions was studied to examine the relationship between the u-r colour, galaxy morphology and galaxy density environment, including residual relations to ascertain which is more fundamental. The paper detailing the results, Ball et al. (2008), is at MNRAS 383 907 and astro-ph/0610171.
Here the colour is the restframe model u-r from the Fourth Data Release of the Sloan Digital Sky Survey. The morphologies are those assigned from artificial neural networks, the Petrosian inverse concentration index and the Sersic index of the light profile. The environmental density is that 5th nearest neighbor within a redshift slice of +/- 1000 km/s, the surface density sigma5. The magnitude limits are r < 17.77, -22.5 < Mr - 5logh < -19.5 for nearest-neighbour galaxies and -22.5 < Mr - 5logh < -17.5 for galaxies for which sigma5 is measured. The brighter Mr corresponds to an apparent r < 17.77 at the limiting redshift of z < 0.0889. This prevents densities biased low by Malmquist bias at higher redshifts in the sample.
In agreement with previous results, we find that the colour is a more predictive indicator of environment, and that there is no evidence for a residual relation in morphology once the relation due to colour is removed. We also find that the robust bimodality displayed in colour is less clear in morphology, indicating either insensitivity to the process that causes the bimodality in colour (an numerous other quantities), a larger overlap of the populations, or a third (or more) population in transition between the other two. Our results are consistent with other studies in which the bimodality is hypothesized to be caused by bulges and disks, formed by merger and accretion. The third population may then be S0 galaxies.
Future work could include adding the star formation rate to the analysis and the use of marked correlation functions as a more general and less arbitrary description of environment, e.g. sigma5 can correspond to intra- or inter-halo galaxy separations. One could also analyze the galaxy population in terms of bulges and disks instead of galaxy 'types'.
The paper includes several large figures which I do not reproduce here, instead I just show the main result, which is that when a single measure of environment and colour/type is used, there is a residual in colour-density with morphology-density removed, but not vice-versa. The importance of adding more detail, for example the morphology of a nearest neighbor (e.g. Park et al. 2008) is the subject of ongoing work in the community.
Figure 1: Residual colour-density with morphology-density removed. From Ball et al. 2008 (MNRAS 383 907).
Figure 2: Residual morphology-density with colour-density removed. From Ball et al. (2008).