The spectroscopy was carried out at Lick Observatory, Kitt Peak National Observatory, and La Palma. The observations at Lick Observatory were made on the Shane 3-m telescope with the Kast spectrograph spanning the wavelength range 3500-8000Å. The Kitt Peak spectra are from the 4-m telescope with somewhat redder wavelength coverage (4500-9000Å) and have some overlap of second order at the red end. The La Palma 2.5-m telescope spectra span 5000 - 8000Å. All the spectra have a resolution of Å. Integration times were typically 10-15 minutes. Observing conditions varied markedly both in transparency and seeing. Sample QSO spectra are shown in Figure 1, including the brightest new quasar, BQ 0751+2919 with E magnitude = 14.6, and the highest redshift new quasar, BQ 0933+2845 with z = 3.42.
Figure: Sample optical spectra from the Lick Observatory Kast Double Spectrograph. BQ 0751+2919 is the brightest of the new quasars, with E = 14.6; BQ 0933+2845 is the highest redshift new quasar, with z=3.42.
To date we have obtained optical spectra for 151 of the 194 previously unclassified QSO candidates. Because of the large range in brightness, the spectra vary in quality from signal-to-noise of 50 to . We have classified objects into five categories. Spectra with broad emission lines with , much greater than typical galaxy velocity dispersions, are classified as QSOs. This classification draws no distinction between bona fide QSOs and Seyfert 1 galaxies; better optical images are required for this refinement, although the optical luminosities can be used to make a rough separation between the two (see below). Featureless spectra are designated as BL Lacs. Spectra with narrow emission lines, , are classed as ELG, making no distinction among true AGN, starburst, and ordinary star forming galaxies. The two remaining categories are galaxies with absorption lines only (ALG), and Galactic stars. All but 4 of the spectra provided positive classifications; these remaining 4 are among the lowest S/N spectra but are good enough to confidently rule out the presence of strong emission lines. On the basis of the radio association and a relatively red optical spectral energy distribution, we tentatively classify these 4 as ALGs, though the possibility exists that they are red Bl Lacs. The classifications of the 151 spectra and 25 NED-identified objects break down as 69 QSOs, 3 BL Lacs, 32 ELG, 41 ALG, and 31 stars. Despite the bright magnitude limit, only 15 of the QSOs were previously cataloged.
For each of the 69 QSOs found in the pilot survey region of 306 , Table 1 lists RA and Dec (J2000), O and E apparent magnitudes from the APM catalog, 20 cm flux density (S ) for the nuclear radio source, total 20 cm flux density including radio lobes (for objects with multiple components), emission line redshift, the logarithm of the inferred radio luminosity for the nuclear radio source, and the absolute E magnitude. For consistency with other studies, we adopt and to compute the distance-dependent quantities using the relations from Weedman (1986). We assume that the spectral energy distributions are described by a simple power law in frequency with an exponent of . Notes to Table 1 describe the radio morphology where it differs from a single compact component. Eleven of the QSOs have extended emission on a scale while another five objects have measured single component sizes , the resolution limit of the FIRST Survey (White et al. 1996).
In Tables 2 and 3, we list similar data for the narrow emission line galaxies and the absorption line galaxies respectively. Data for the Galactic stars are contained in Table 4. Many, but not all, of the stars are chance coincidences (Becker et al. 1995).
Figure 2 shows the distribution of observed nuclear radio flux for the QSOs; every QSO in the sample with mJy has been previously identified, an indication of the completeness threshold of previous radio-selected samples.
Figure: Distribution of observed nuclear 21 cm flux for the QSO sample. Previously known QSOs are shaded.
About 50% of the QSOs in the sample are radio quiet with , using a simple luminosity cut as the criterion for radio loudness (Schneider et al. 1992). With a flux density limit of 1 mJy, some radio quiet QSOs will be detectable in the FIRST survey out to a redshift of . A histogram of the radio luminosities (Figure 3) has a suggestion of a deficit at , hinting at a bimodal distribution, roughly consistent with radio loud and radio quiet objects. As the survey progresses and the sample grows, the reality of the bimodality will be resolved, though it is already apparent that there is considerable overlap in the distributions. The narrow emission line galaxy (ELG) and absorption line galaxy (ALG) luminosities are shown for comparison (dashed line). The latter two classes have indistinguishable distributions.
Figure: Histogram of the QSO/Seyfert 1 nuclear radio luminosities; there is a suggestion of a bimodal distribution between radio loud and radio quiet. The distribution of narrow emission line and absorption line galaxy radio luminosities (dashed histogram) is plotted for comparison.
Figure: Redshift distribution of the QSO sample. Radio loud sources are shaded.
Figure 4 displays a histogram of the redshift distribution for the 69 QSOs. The radio loud QSOs have been shaded to illustrate clearly the bias against detecting high redshift radio quiet QSOs in a radio-selected sample. The fraction of radio loud objects increases from for z< 0.5 to for to for z > 1.5. The overall distribution is roughly consistent with that of the LBQS, though detailed comparisons must await better statistics when we have a larger sample.
Figure: Histogram of the QSO/Seyfert 1 optical (red) absolute magnitudes (solid line); the excess in the faint tail can be explained by the lower luminosity Seyfert 1 objects present in the sample. The distributions of narrow emission line and absorption line galaxies are indistinguishable and are plotted as a single histogram (dashed line) for comparison.
Figure 5 shows the histogram of absolute E magnitudes derived from the APM magnitudes, adopting and the same cosmological assumptions as above. These have not been corrected for Galactic extinction or contributions from emission lines; the errors in the APM magnitudes dominate these corrections. The QSOs again suggest a bimodal distribution. The usually accepted luminosity cutoff between Seyfert 1 galaxies and QSOs is M , which is equivalent to M for typical O-E of 1. The low luminosity component, containing objects, is probably the Seyfert 1 contribution to the sample and perhaps accounts for the bimodality. All but five of the 18 were classified as stellar on both POSS plates, so deeper optical images are necessary to confirm them as Seyferts. The ELG and ALG distributions are again very similar and are plotted for comparison as a single dashed histogram. Eight of the nearest absorption line objects have either unusually bright APM magnitudes, or the APM could not assign a reliable magnitude because of complicated image structure; these are indicated by a 0. in Table 3. All of these are previously cataloged objects and a comparison with data from NED shows that the APM numbers are in error by many magnitudes. These have been excluded from the histogram and no entry for M is present in Table 3.