4.3. Evolution of the Luminosity Function
One of the most striking observed features of quasars is the evolution of their luminosity function. The space density of luminous quasars increases by a factor of ~ 1000 between the present epoch and redshift 2 - 3 and then falls steeply toward higher redshifts (Warren, Hewett, & Osmer 1994, WHO; Schmidt, Schneider, & Gunn 1995, SSG; Kennefick, Djorgovski, & de Carvalho 1995). A straightforward explanation of this behavior is that we are seeing back to the epoch of peak quasar activity, an epoch that presumably has to do with the formation of black holes at the centers of galaxies and the time of significant fueling of the quasar activity via the infall of material to the center.
However, a persistent question about the nature of the peak is whether it is affected significantly by dust absorption along the line of sight. If so, there could be an important population of quasars at high redshift that are hidden at optical/UV wavelengths, indicating that the epoch of peak activity was even earlier. There is no doubt that some quasars are highly reddened; the basic question is how many.
One way to answer this question is to use samples of radio-selected quasars with complete optical identifications. Dust is transparent to radio radiation, and so samples with complete optical identifications provide an excellent test, as long as the ratio of radio quasars to the total number of quasars does not change significantly with epoch.
Hook, Shaver, and McMahon (1998) have carried out just such a program and find that the evolution of quasars in their sample is remarkably similar to that found by WHO and SSG. This suggests that dust is not the cause of the apparent decline in activity at z > 3. Similarly, Benn et al. 1998 used IR observations in the K band of radio-selected quasars and found no evidence for a large population of reddened and dust-absorbed quasars. These results are in contrast to those of Masci (1998), who does claim evidence for a population of reddened objects.
The ultra-deep ROSAT survey of Hasinger et al. (Hasinger 1998) has yielded important new X-ray results. The good positional accuracies of the survey show that most of the sources are quasars/AGNs and narrow emission-line galaxies are only a small fraction, in contrast with some previous work. Their new determination of the X-ray luminosity function is not consistent with pure luminosity evolution but can be fit by pure density evolution from z = 0 to z 2. Their results suggest that black holes should be common in massive galaxies at the present epoch, as discussed in more detail below.