Selection method

The basic ideas of the HSS and its selection methods of ELG candidates are described in Paper I. The final selection was slightly modified to improve significantly the detection rate of ELGs in the follow-up spectroscopy as described in Paper II. As it was outlined in Paper I, the selection procedure provided us finally with two candidate lists (first and second priorities): 1st - objects showing a clear density peak near $\lambda$ 5000 Å and blue continuum in the Hamburg Quasar Survey objective-prism spectra scanned with high resolution; 2nd - candidates with a blue continuum but without prominent emission features or candidates with indications of emission peaks but with an unusual continuum shape. In short, the ELG candidate selection criteria applied are a blue or flat continuum (near $\lambda$ 4000 Å) and the presence of strong or moderate [O III] $\lambda\lambda$ 4959,5007 Å emission lines recognized on digitized prism spectra. Candidates accepted had B-magnitudes in the range $16^m - 19\fm5$ .

Based on the experience with a training sample of BCGs drawn from the Second Byurakan Survey (see Paper I for details) the first priority candidates were considered as highly probable H II/BCG type emission galaxies. The follow-up snap-shot spectroscopy confirmed that among all detected ELGs this type of galaxies constitutes up to 70-80%. Thus, our main goal to create a large new sample of BCG/H II-galaxies in the HSS region is achieved by follow-up spectroscopy of the full sample of first priority candidates.

In this paper we will present the results for follow-up spectroscopy of three samples listed in Table 1. The first sample is made up by 139 first priority candidates in the strip $\delta=$ 35-40 of which 26 are known ELGs. For the latter galaxies additional spectra were required to improve their classification.

The other two samples were drawn from the second priority candidates of the same strip. We found in Paper I that the detection efficiency of H II/BCG galaxies is rather low among them, prohibiting follow-up spectroscopy of all second priority candidates. Moreover, the second priority objects are about twice as numerous than the first priority ones. At the faint end, this sample is also dominated by candidates selected because of noise peaks in their objective prism spectra. We created therefore a random selected sample of 43 second priority objects from this strip matching in its magnitude distribution the sample of first priority candidates. These objects make up 10% of the second priority sample in the magnitude range of the first priority candidates. We obtained follow-up spectroscopy of all objects to study the general content and to determine the fraction of BCGs in this sample.

The third sample (referred to as APM Selected in Table 1) was created by applying additional selection criteria to the second priority candidates to increase the detection efficiency for BCG/H II-galaxies among them. These criteria and the results of follow-up spectroscopy are described in Section 4.2.2.

**Table:** Summary of the samples observed and breakdown of the classifications after follow-up spectroscopy
Candidate	Sample	N	BCGs	Other	QSO	Galaxies	Stars	Not
Classification				ELGs		without ELs		Classified
First priority	new	113	47	24	7	3	7	25
	already known	26	19	7	-	-	-	-
	Total	139	66	31	7	3	7	25
Second priority	6m observations	26	1	10	1	1	9	4
	2.2m observations	17	3	2	-	1	3	8
	Random (total)	43	4	12	1	2	12	12
	APM selected (total)	11	7	4	-	-	-	-
Objects presented in this paper		176	74	45	8	4	16	29

**Table 2:** Log of observations at the SAO 6m telescope
Run	Date			Instrument	Grating	Wavelength	Dispersion
No					[grooves/mm]	Range [Å]	[Å/pixel]
(1)	(2)			(3)	(4)	(5)	(6)

1	17-19	Dec	1998	CCD, SP-124	300	3600-7800	4.6
2	08-13	Feb	1999	CCD, LSS	325	3600-7800	4.6
3	02	Sep	1999	CCD, LSS	650	3700-6100	2.4
	04	Sep	1999	CCD, LSS	325	3600-7800	4.6
4	02	Feb	2000	CCD, LSS	650	3700-6100	2.4