Evidence for a magnitude-dependent bias in the Hamburg/ESO survey for damped Lyman–$\alpha$ systems

We present preliminary results from the Hamburg/ESO survey for damped Ly-$\alpha$ (hereafter, DLA) systems. This survey is characterised by (i) good knowledge of the biases affecting the parent QSO survey, and (ii) the brightness and (iii) relatively wide magnitude distribution of the background QSOs. Therefore, it is well–suited to study possible magnitude–dependent biases in DLA surveys, such as the one expected from dust obscuration. We have systematically searched for damped Lyman-$\alpha$ line candidates in 5 Å, resolution spectra of the 188 QSOs that constitute our statistical sample. These candidates have later been re-observed with UVES at the ESO–Very Large Telescope (VLT) for confirmation and accurate $N(HI)$ measurements. In the redshift range covered by the survey, 19 DLA systems have been discovered. Over the whole survey, we find that the number density $n(z)$ and cosmological density of gas $\Omega_\mathrm{gas}$ have comparable values to the ones obtained by CORALS (Ellison et al. 2001). However, the number densities of DLA systems $n(z)$ in two sub–samples of equal absorption distance path defined by the magnitude of the background QSOs differ by a factor of $\approx 5$. We estimate that the probability that $n(z)$ is equal in the two sub–samples is $<0.003$. A similar, only slightly less significant difference is found for $\Omega_\mathrm{gas}$.