2.1. Source Classification Challenges

Classification of deep X-ray survey sources is challenging for several reasons. First, many of the X-ray detected AGN are simply too faint for straightforward optical spectroscopic identification even with 8-10 m class telescopes (note the multitude of small dots in Figure 4). Intensive optical identification programs on deep Chandra and XMM-Newton fields typically have 50-70% redshift completeness at best (e.g., Barger et al. 2003a; Szokoly et al. 2004), and some of the obtained spectra that do yield redshifts are of insufficient signal-to-noise for reliable optical classification work. Furthermore, many of the X-ray sources have modest optical luminosities, often due to obscuration. Thus starlight from their host galaxies can make a substantial diluting contribution to the flux measured in a ground-based spectroscopic aperture, plausibly "overwhelming" subtle nuclear spectral features from an obscured or low-luminosity AGN (e.g., Moran, Filippenko, & Chornock 2002). Finally, another challenge is an apparent "schism" between optical (type 1 vs. type 2) and X-ray (unobscured vs. obscured) schemes of classification; not all X-ray obscured AGN have type 2 optical spectra, and not all AGN with type 1 optical spectra are X-ray unobscured (e.g., Matt 2002).