Annu. Rev. Astron. Astrophys. 2005. 43: 861-918 Copyright © 2005 by Annual Reviews. All rights reserved

11. ARE DAMPED Ly SAMPLES BIASED BY GRAVITATIONAL LENSING ?

On the other hand, Prochaska et al. (2005) find that biasing due to gravitational lensing could be important. They compared the full SDSS sample of damped Ly systems with subsamples comprising the brightest 33% of background quasars and the faintest 33% of background quasars. While the incidence of damped Ly systems, d / dX, was found to be insensitive to quasar magnitude, the mass density, _g(z), was found to vary significantly. Specifically, the bright subsample showed sysematically higher values of _g(z) than the faint subsample. To explain the independence of d / dX on quasar magnitude, the difference must lie in the incidence of systems with large N(H I), which is observed to be larger in the bright subsample.

Prochaska et al. (2005) argue that gravitational magnification of the background quasars by massive halos or disks associated with the foreground damped Ly systems could account for this systematic effect, which was first detected at the 2 level by Murphy & Liske (2004). Whereas obscuration by dust would cause _g(z) to be lower in the bright subsample, magnification by lensing due to exponential disks is greatest for damped Ly systems with large values of N(H I) (Bartelmann & Loeb 1996; Maller et al. 1997).

Consequently, the values of _g(z) in Figure 5 may be 10-20% too high, but the evolution of _g(z) with z is still likely to be correct. However, the effect could be more important for damped Ly system samples with z < 2 (Rao & Turnshek 2000).