2 Sample selection

The alternative expression for X-ray-to-optical flux ratio $\log (f_X/f_{{\rm opt}})$comes from the statistical analysis of known RASS-BSC X-ray sources. The sources which have an entry in SIMBAD, NED, the catalogs of AGN (Véron-Cetty & Véron 1996), or other available databases are considered as known.

Figure 1: Count rates as a function of R magnitudes for known sources

The X-ray sources with the highest X-ray-to-optical flux ratio are: BL Lac objects, emission line AGN, clusters of galaxies etc. However, X-ray flux is difficult to evaluate before optical identification. Although for a given spectral energy distribution the conversion of X-ray count rate to flux depends on the column density $N_{\rm H}$, the X-ray flux ($f_{\rm X}$) is nevertheless roughly proportional to the count rate (C), so X-ray flux in the X-ray-to-optical flux ratio criterion can be replaced by X-ray count rate, i.e., $\log(f_{\rm X}/f_{{\rm opt}})\propto \log C + 0.4\,R +$ constant, where we use the R magnitude flux to represent the optical flux. For fixed $\log (f_{\rm X}/f_{{\rm opt}})$, $\log C + 0.4\,R =$ constant. According to our statistical analysis of known RASS-BSC X-ray sources, we found that there is an apparent gap between Galactic stars and extragalactic objects. As can be seen in Fig. 1, 92% AGN concentrate in the region: $\log C \geq - 0.4\,R$ + 4.9. $\rm O {-} M$ spectral type stars in this region are only 3% of its total amount included in the statistics. Thus, choosing $\log C \geq - 0.4\,R + 4.9$ is expected to be efficient for the preselecting of AGN and can avoid too many $\rm O {-} M$ stars being included in the sample.

Considering the observatory site, instruments and possible observation times, we selected our sample from RASS-BSC according to the following criteria:

1.

Unknown sources which do not have an entry in SIMBAD, NED, or other available databases.

2.

Declination $\delta \geq 3\hbox{$^\circ$ }$.

3.

Galactic latitude $\vert b\vert \geq 20\hbox{$^\circ$ }$.

4.

Optical counterparts within a circle with radius $r=r_{1}+5\hbox{$^{\prime\prime}$ }$, where r₁ is the RASS position error given by Voges et al. (1996).

5.

Optical counterparts with R magnitudes between 13.5 and 16.5.

6.       $\log C \geq - 0.4\,R + 4.9$.

For the search of optical counterparts we adopt the error circle with radius r, 87% of known AGN are found within r from X-ray position. For each object inside the error circle, we obtain its R magnitude from USNO-A1.0 and check whether it satisfies the flux ratio criterion. Only X-ray sources which have at least one object inside the error circle which meets the criterion are included in the sample. A sample with 165 unidentified X-ray sources with Right Ascension between 19$^{\rm h}$ and 8$^{\rm h}$ was selected for optical spectroscopy from RASS-BSC. Table 2 summarizes the general information for the observed objects as following:
Column (1): ROSAT (1RXS J) source designation.
Columns (2) & (3): Optical right ascension $\alpha$ and declination $\delta$ of the proposed optical counterpart (epoch 2000.0) derived from the Digitized Sky Survey (DSS) images which should be better than $2\hbox{$^{\prime\prime}$ }$ (Irwin et al. 1994).
Column (4): ROSAT X-ray position error as given in RASS-BSC.
Column (5): Angular Separation between the X-ray source and the proposed optical counterpart in arcsecond.
Column (6): ROSAT total count rate (in counts s^-1) as given in RASS-BSC.
Column (7): Hardness ratio HR1 = (B - A)/(B + A), where A means the count rate in the energy range 0.1-0.4 keV and B in the energy range 0.5-2.0 keV.
Column (8): Hardness ratio HR2 = (D - C)/(D + C), where C means the count rate in the energy range 0.5-0.9 keV and D in the energy range 0.9-2.0 keV.
Column (9): Extent in arcsecond as given in RASS-BSC.
Columns (10) & (11): The R and B magnitudes obtained from USNO-A1.0 with an accuracy between 0.25 and 0.40 magnitudes depending on the declination (Monet et al. 1996).