The Bimodal Galaxy Color Distribution: Dependence on Luminosity and Environment

Abstract

We analyse the u-r color distribution of 24346 galaxies with Mr<=-18 and z<0.08, drawn from the Sloan Digital Sky Survey first data release, as a function of luminosity and environment. The color distribution is well fit with two Gaussian distributions, which we use to divide the sample into a blue and red population. At fixed luminosity, the mean color of the blue (red) distribution is nearly independent of environment, with a weakly significant (~3sigma) detection of a trend for colors to become redder by 0.1-0.14 (0.03-0.06) mag with a factor ~100 increase in local density, as characterised by the surface density of galaxies within a +/-1000 km/s redshift slice. In contrast, at fixed luminosity the fraction of galaxies in the red distribution is a strong function of local density, increasing from ~10-30 per cent of the population in the lowest density environments, to ~70 per cent at the highest densities. The strength of this trend is similar for both the brightest (-23<Mr<-22) and faintest (-19<Mr<-18) galaxies in our sample. The fraction of red galaxies within the virialised regions of clusters shows no significant dependence on velocity dispersion. Even at the lowest densities explored, a substantial population of red galaxies exists, which might be fossil groups. We propose that most star-forming galaxies today evolve at a rate that is determined primarily by their intrinsic properties, and independent of their environment. Any environmentally triggered transformations from blue to red colors must either occur on a short timescale, or preferentially at high redshift, to preserve the simple Gaussian nature of the color distribution. The mechanism must be effective for both bright and faint galaxies.Comment: ApJ Letters, in press. 5 pages and 3 inline figures,using emulateapj style. Fig 1 larger than in published version. Fit parameters available at http://star-www.dur.ac.uk/~balogh/SDSS/colourfits.htm