1 Introduction

During the ROSAT All-Sky Survey (RASS; Voges 1992, 1997) about 80000 X-ray sources with a detection likelihood $\geq 10$ were detected, from which 18 811 sources having a PSPC count rate larger than 0.05 cts s^-1 and detection likelihood $\geq 15$ were compiled in the RASS-BSC (Voges et al. 1996a). More than 65% RASS sources remain unidentified (Voges et al. 1996b). The optical identification of this large numbers of X-ray sources is a challenge. In order to obtain the statistical classification of the RASS sources, several optical identification programs have been carried out (Appenzeller et al. 1998; Fischer et al. 1998; Hasinger et al. 1998; Schmidt et al. 1998; Thomas et al. 1998).

The identification of X-ray sources has proved to be an effective means of creating samples of Active galactic nuclei (AGN). The present program aims at creating a new bright sample of AGN which have the highest X-ray-to-optical flux ratios for statistical investigation. The results of the EMSS (Stocke et al. 1991) have shown that different classes of X-ray sources represent different narrow ranges in the X-ray-to-optical flux ratios. These bounds mean the X-ray sources with Galactic and extragalactic counterparts can be separated at high confidence level prior to any optical spectroscopy (Maccacaro et al. 1988).

Sections 2 and 3 review the creation and follow-up optical spectroscopic observations of the sample. Section 4 gives details of the optical identification procedure. The final section presents the results. A Hubble constant of H₀=50 km s^-1 Mpc^-1 and cosmological deceleration parameter of q₀=0 are used throughout.