A SPECTROSCOPICALLY COMPLETE SAMPLE OF QUASARS WITH BJ-LESS-THAN-OR-EQUAL-TO-22.0

We present and discuss the spectroscopic data for a complete sample of faint optically selected quasars extracted from the quasar candidates published by Marano, Zamorani & Zitelli. These data, obtained at the ESO 3.6-m and the Anglo-Australian telescopes, have provided a sample of 54 confirmed quasars with J less-than-or-equal-to 22.0. Fifty-two of them constitute a complete sample and provide the highest surface density of spectroscopically confirmed quasars measured so far (115.0 +/- 16.5 per square degree). Splitting the sample into two redshift ranges (z < 2.2 and greater-than-or-equal-to 2.2), the corresponding surface densities at J less-than-or-equal-to 22.0 are 86.3 +/- 14.5 and 28.7 +/- 8.4. We briefly discuss the log N-m relation for various subsamples of objects extracted from our data. In particular, we find that, while these data confirm the relatively flat slope at faint magnitudes of the total counts of all the quasars, with no cut in absolute magnitude, the slope of the log N-m relation for low-luminosity objects (M(B) > -23.0) is steeper than the Euclidean value, thus suggesting also for them a substantial cosmological evolution. For each object we give, in addition to J magnitude and redshift, the equivalent widths of the main lines. The distributions of the equivalent widths of these lines are, on average, consistent with other existing data at brighter magnitudes. The main difference is seen in the Mg II line, for which our average equivalent width is significantly higher than in other quasar samples and is, instead, consistent with the data of a `local' sample of Seyfert galaxies. While our data for the Mg II line are consistent with the existence of a correlation between the equivalent width and the absolute magnitude, such a correlation (`Baldwin effect') is not seen for the equivalent width of the C IV line. Although this absence of correlation for the C IV line could, in principle, be due to the more limited range in absolute magnitude for this line, we point out that at the typical absolute magnitudes of the objects in our sample the latest version of the Baldwin relation, as derived from a complete sample of radio-selected, flat-spectrum radio quasars, would predict average widths about two or three times larger than our average equivalent widths. We can therefore conclude that our objects do not follow the standard Baldwin relation. A more detailed discussion of the Baldwin effect (or absence of it) in optically selected samples has been given elsewhere.