Counts of Low-redshift SDSS Quasar Candidates

We analyze the counts of low-redshift quasar candidates selected using nine-epoch SDSS imaging data. The co-added catalogs are more than 1 mag deeper than single-epoch SDSS data, and allow the selection of low-redshift quasar candidates using UV-excess and also variability techniques. The counts of selected candidates are robustly determined down to g=21.5. This is about 2 magnitudes deeper than the position of a change in the slope of the counts reported by Boyle et al. (1990, 2000) for a sample selected by UV-excess, and questioned by Hawkins & Veron (1995), who utilized a variability-selected sample. Using SDSS data, we confirm a change in the slope of the counts for both UV-excess and variability selected samples, providing strong support for the Boyle et al. results. 1. A Controversy About Quasar Counts The quasar luminosity function, Φ(L), provides fundamental information about their nature. Boyle et al. (1990, 2000) found that Φ(L) for quasars at redshifts z < 2.3 resembles a “broken” power law which becomes flatter at the faint end. They also demonstrated that the “break” luminosity at which the slope changes increases with redshift, and between z = 0.8 and z = 2.2 becomes more luminous by 1.9 mag. A peculiar aspect of this result is that the apparent magnitude corresponding to the “break” luminosity changes by only 0.8 mag. over this redshift range (B=18.6–19.4; that is, the differential distribution of quasar apparent magnitudes has a shape very similar to the shape of their luminosity function). Further, this apparent magnitude range is only ∼ 1-2 mag. above the photographic plate limit used for selection. Therefore, it is possible that the flattening of the luminosity function is due to a systematic underestimate of the increasing incompleteness at the faint end of the UV-excess selection technique employed