Optical Galaxy Selection

Our view of the properties of galaxies is strongly affected by the way in which we survey for them. I discuss some aspects of selection effects and methods to compensate for them. One result is an estimate of the surface brightness distribution. I believe this is progress, but considerable uncertainty remains. A basic goal of galaxy research is to identify and characterize the galaxy populations which inhabit the universe. Astronomers undertake large surveys to discover galaxies, but this is only part of the battle. In order to transform from the observed, apparent distributions of galaxy properties and recover the desired intrinsic distributions, one must understand and correct for the selection characteristics of galaxy surveys. An essential fact about galaxies, as they appear on survey plates, is that they are not point sources. Some of the consequences of this simple fact were first quantified by Disney (1976), who pointed out that the intrinsic distribution could be very different from the apparent one. At the least, it is necessary to generalize the formal procedures developed for point sources (Disney & Phillipps 1983; McGaugh et al. 1995). The appearance of galaxies requires many parameters to approximate: they have multiple components (bulge and disk); they have spiral arms and dust lanes; their morphology can be a strong function of wavelength depending on their recent star formation history; etc. So, we simplify. Here I will pretend that a tolerable approximation can be had with axially averaged radial surface brightness profiles. This reduces the dimensionality of the problem immensely. The surface brightness profiles of galaxies obtained in this way are tolerably described as a combination of bulge and disk (4 parameters) or with a generalized (Sérsic 1969) profile (3 parameters). For brevity, I will reduce this further to 2 parameters by limiting the discussion to galaxies dominated by exponential disks: µ(r) = µ 0 + 1.086 r h. (1) This reduces complex galaxies to a characteristic central surface brightness µ 0 and scale size h. In some cases this is a fair approximation. Galaxies inhabit a large range in surface brightness and scale length (Figure 1). Real limits appear at high surface brightness and luminosity. Here, objects are prominent and are easily selected by surveys.