3 Results of follow-up spectroscopy

In total 108 new candidates and 5 known ELGs have been observed. Among them, 81 are new or confirmed emission-line galaxies, 4 are quasars (all with redshifts in the range 3.07 to 3.20), and 8 are galaxies without emission lines. Only 2 of the latter have good enough S/N ratio to identify absorption features enabling measurements of their redshifts. The remaining 20 objects appeared to be either stars with characteristic absorption lines or stellar objects with featureless spectra where the signal-to-noise ratio was insufficient to identify lines.

3.1 Emission-line galaxies

The new emission line galaxies are listed in Table 2 containing the following information:
Column 1: The object's IAU-type name with the prefix HS. We note by asterisk objects observed at Calar Alto.
Column 2: Right ascension for equinox B1950.
Column 3: Declination for equinox B1950. The coordinates were measured on direct plates of the HQS and are accurate to $\sim$ 2 $\hbox{$^{\prime\prime}$ }$ (Hagen et al. [1995]).
Column 4: Heliocentric velocity and its rms uncertainty in km s^-1.
Column 5: Apparent B-magnitude obtained by calibration of the digitized photoplates with photometric standard stars (Engels et al. [1994]), having an rms accuracy of $\sim$ $0\hbox{$.\!\!^{\rm m}$ }5$ for objects fainter than $m_{B} = 16\hbox{$.\!\!^{\rm m}$ }0$ (Popescu et al. [1996]). Since the algorithm to calibrate the objective prism spectra is optimized for point sources the brightnesses of extended galaxies are underestimated. The resulting systematic uncertainties are expected to be as large as 2 mag (Popescu et al. [1996]). For about 1/3 of our objects, B-magnitudes are unavailable at the moment. We present for them blue magnitudes obtained from the APM database. They are marked by a "plus" before the value in the corresponding column. According to our estimate they are systematically brighter by $0\hbox{$.\!\!^{\rm m}$ }92$ than the B-magnitudes obtained by calibration of the digitized photoplates (rms $1\hbox{$.\!\!^{\rm m}$ }02$).
Column 6: Absolute B-magnitude, calculated from the apparent B-magnitude and the heliocentric velocity. No correction for galactic extinction is made as all objects are located at high galactic latitudes and because the corrections are significantly smaller than the uncertainties of the magnitudes.
Column 7: Preliminary spectral classification type according to the spectral data presented in this article. BCG means that the galaxy posesses a characteristic H II-region spectrum and that the luminosity is low enough. SBN and DANS are galaxies of lower excitation with a corresponding position in line ratio diagrams, as discussed in Paper I. SBN are the brighter fraction of this type. We here follow the notation of Salzer et al. (1989). Seyfert galaxies are separated mainly on diagnostic diagrams as AGN. But if their emission lines are quite narrow, they probably should be classified as Sy2. SA is a probable super-association at the rim of an edge-on nearby disc galaxy. Six objects are difficult to classify. They are coded as NON.
Column 8: One or more alternative names, according to the information from NED.

The spectra of all emission-line galaxies are shown in Appendix A, which is available only in the electronic version of the journal.

The results of line flux measurements are given in Table 4. It contains the following information:
Column 1: The object's IAU-type name with the prefix HS. By asterisk we note the objects observed during non-photometric conditions.
Column 2: Observed flux (in 10^-16ergs^-1cm^-2) of the H$\beta$ line. For the few objects without an H$\beta$ emission line the fluxes are given for H$\alpha$ marked by a "plus''. For the objects observed on Calar Alto during non-photometric conditions this parameter is unreliable and marked by (:).
Columns 3, 4, 5: The observed flux ratios [O II]/H$\beta$, [O III]/H$\beta$ and H$\alpha$/H$\beta$.
Columns 6, 7: The observed flux ratios [N II]$\lambda$6583 Å/H$\alpha$, and ([S II]$\lambda$6716 Å + $\lambda$6731 Å)/H$\alpha$.
Columns 8, 9, 10: Equivalent widths of the lines [O II]$\lambda$3727 Å, H$\beta$ and [O III]$\lambda$5007 Å. For the few objects without a detected H$\beta$ emission line the equivalent widths are given for H$\alpha$ marked by a "plus''.

Below we give notes on several individual objects:

HS 1015+3717: In the spectrum of this object a cosmic ray hit is exactly on the line [O III]$\lambda$4959 Å. This was not corrected in the figure shown in Appendix A.
HS 1214+3801: This is seemingly a supergiant H II-region at the very rim of the nearby edge-on disc galaxy (SA(s)cd) NGC 4244 ( $V_{\rm hel}$ = 224 kms^-1 and $B_{\rm T}$=10.88). At the accepted distance of NGC 4244 (D = 4.5 Mpc) M_B of HS 1214+3801 is about - $11\hbox{$.\!\!^{\rm m}$ }8$. The difference between the systemic radial velocity of the host galaxy and H II-region is small (32 kms^-1) and does not contradict that HS 1214+3801 belongs to NGC 4244. However, the velocity field of NGC 4244 near the position of the H II-region is unkown. Both, the single-dish H I-measurements as summarized in Huchtmeier & Richter ([1989]), and an estimate of the maximum rotational velocity $V_{\rm rot} \leq$ 130 kms^-1 (which we obtained through the Tully-Fisher relation from the absolute B-band magnitude of NGC 4244 of $\sim$ - $17\hbox{$.\!\!^{\rm m}$ }8$), yield a range of expected velocity differences between the galaxian material and HS 1214+3801 of up to +160 or -100 kms^-1. But since the 2-D spectrum of HS 1214+3801 with a total spatial extent of about 20 $^{\prime\prime}$ ($\approx$0.5 kpc) shows evidence of internal motions with an amplitude of about 50 kms^-1 we need to consider an alternative interpretation for this object as a companion BCG. Its SF burst may be triggered due to the tidal effect from the more massive galaxy, similar to the case of HS 1717+4955 described in Kniazev et al. ([2000]). To check this option one needs a detailed map of the NGC 4244 velocity field including HS 1214+3801.
HS 1214+3922: This BCG was reobserved with higher S/N ratio in order to measure the flux of the [O III]$\lambda$4363 line, necessary to determine unambiguously the electron temperature $T_{\rm e}$([O III]) of the H II-region and the oxygen abundance. A preliminary determination according to the procedure described by Izotov et al. ([1997]) shows that it has the low oxygen abundance of log(O/H) + 12 = 7.76.

3.2 Quasars

In the course of our follow-up spectroscopy, four QSOs were discovered with a strong emission line in the wavelength region between 5000 Å and the sensitivity break of the Kodak IIIa-J photoemulsion near 5400 Å. In all of them, we identified Ly$\alpha$$\lambda$1216 redshifted to z $\sim$ 3 as the responsible line. This strong line produces an easily visible emission peak in the digitized prism spectra even for very faint objects (B $\sim 19\hbox{$.\!\!^{\rm m}$ }0 - 20\hbox{$.\!\!^{\rm m}$ }0$) which is hard to distinguish from low-redshift [O III] features. Else, QSOs were not selected as candidates for follow-up spectroscopy.

The data for these four new high-redshift quasars are presented in Table 3. Finding charts and plots of their spectra can be found on the www-site of the Hamburg Quasar Survey (http://www.hs.uni-hamburg.de/hqs.html).

3.3 Non-emission-line objects

In total, for 28 candidates no (trustworthy) emission lines are detected. We divided them into three categories.

3.3.1 Absorption-line galaxies

For two bright non-ELG galaxies the signal-to-noise ratio of spectra was sufficient to detect absorption lines, allowing the determination of redshifts. The data are presented in Table 5.

3.3.2 Stellar objects

To separate the stars among the objects missing detectable emission lines we cross-correlated a list of the most common stellar features with the observed spectra. In total, 13 objects with definite stellar spectra and redshifts close to zero were identified. Four of them are obvious K or M-stars. The rest were classified roughly in categories from definite A-stars to F or G-stars, with most of them intermediate between F and G. The data for these stars are presented in Table 6.

3.3.3 Non-classified objects

Thirteen non emission-line objects are hard to classify at all. Their continua have too low signal-to-noise ratio to detect trustworthy absorption features, or the equivalent width of the emission lines is too small. Six of them are certainly non-stellar on DSS images, and classified as well as non-stellar in the APM database. From our spectra in the range $\approx$ 4000 to $\approx$ 7300-8000 Å, we can exclude the presence of strong H$\alpha$. The remaining 7 objects are indistinguishable from stellar ones, and we suggest that most of them are galactic stars. One of the galaxies - namely HS 1232+3609, was presented after our observations in the paper by Popescu et al. ([1998]) as an ELG with z= 0.2529. Our spectrum is too noisy, and we could not identify any significant emission with this redshift.