2. Quality of the data

We classified our redshifts from 1 (best) to 3, according to their quality. For galaxies with absorption lines: spectra of quality 1 have at least three lines clearly visible; spectra of quality 2 have two; and spectra of quality 3 have only one. The signal to noise parameter R given by the cross-correlation measure is also given in Table 3. A histogram of this quantity is shown in Fig. 4 (click here).

  
Figure 4: Distribution of the Tonry & Davis R parameter given by the cross-correlation measure on absorption lines

For galaxies with emission lines: spectra of quality 1 have all [OII]3727, H, and [OIII]4959-5007 lines clearly visible; spectra of quality 2 have at least two emission lines, and spectra of quality 3 have only one (usually [OII]3727); in that case, the identification of the emission line had to be confirmed by the shape of the continuum. The identification of a single line was made possible by the fact that all exposures were doubled, so we could remove cosmic rays and check that the emission line was indeed present in both spectra. Notice the high number of redshifts (100) obtained from emission lines.

  
Figure 5: Typical absorption line spectrum of quality 1 (best)

  
Figure 6: Typical absorption line spectrum of quality 2

  
Figure 7: Typical absorption line spectrum of quality 3

  
Figure 8: Typical emission line spectrum of quality 1 (best)

  
Figure 9: Typical emission line spectrum of quality 2

  
Figure 10: Typical emission line spectrum of quality 3

Typical spectra of various qualities are displayed in Figs. 5 (click here) to 10 (click here).

In order to check the intrinsic quality of our velocity measurements, two velocity standard stars from the Maurice et al. (1984) list were observed each night. The errors, derived by cross-correlating the star spectra to the spectrum of M 31, range (from night to night) from 16 to  km s^-1 for HD 24331, and from 17 to  km s^-1 HD 48381. The mean internal error on velocities derived from the accuracy in the wavelength calibration is 66 km s^-1. The distribution of errors on velocities is displayed in Figs. 11 (click here) and 13 (click here) for absorption and emission line measurements respectively (these histograms only include the galaxies that we observed). For absorption lines, the correlation between the Tonry & Davis R parameter and the error on the velocity is displayed in Fig. 12 (click here) for the 262 galaxies for which we had both quantities. For emission line measurements, the errors on velocities were estimated from the dispersion on the velocities derived from the various emission lines present. When only one emission line was present we averaged the emission and absorption line redshifts whenever possible; if no reliable absorption line redshift was available, we estimated the internal error on a single emission line to be the intrinsic value of 66 km s^-1 mentioned above.

  
Figure 11: Distribution of internal errors on velocities derived from absorption lines

  
Figure 12: Relation between the Tonry & Davis R parameter and the errors on velocities derived from absorption lines

  
Figure 13: Distribution of internal errors on velocities derived from emission lines

In order to test the agreement of our redshifts with those of previous surveys, we reobserved 7 and 6 galaxies from the velocity catalogues by Beers et al. (1991) and Malumuth et al. (1992) respectively. The comparison of these various measurements are given in Table 1. Except for a few totally discrepant values, which are probably due to galaxy misidentifications, our values agree with those of the literature, with BWT mean differences of +8 and +35 kms^-1 between our values and the Beers and Malumuth samples respectively, and corresponding BWT dispersions of 132 and 156 km s^-1 (for the comparison with the Beers data, we kept the two first of our objects in the above Table which gave the smallest velocity difference with the Beers data). These values are comparable to the BWT mean (-71 kms^-1) and BWT dispersion (125 kms^-1) that we estimate between the Beers and Malumuth samples. However, the number of objects in common is of course small and may pervert statistics.

 

DFLS         DFLS DFLS Beers         Beers         Beers Difference

position        velocity error position        velocity error

0 41 28.112, -9 13 51.49 14102 92 0 41 27.202, -9 13 44.11 14233 46 -131

0 41 28.112, -9 13 51.49 14102 92 0 41 29.000, -9 14 00.14 14022 33 80

0 41 50.396, -9 18 09.48 16719 81 0 41 49.767, -9 18 33.40 16536 44 183

0 41 50.396, -9 18 09.48 16719 81 0 41 50.070, -9 18 10.41 16734 48 -15

0 41 50.157, -9 25 47.54 17224 78 0 41 50.240, -9 25 47.41 17360 69 -136

0 41 53.400, -9 29 39.10 15331 48 0 41 53.625, -9 29 45.45 15392 34 -61

0 42 12.944, -9 17 50.97 19950 75 0 42 12.750, -9 17 50.70 19758 59 192

DFLS         DFLS DFLS Malumuth       Malumuth Malumuth Difference

position        velocity error position        velocity error

0 40 31.650, -9 13 20.02 14102 42 0 40 31.550, -9 13 20.41 15410 91 -1308

0 40 43.550, -8 57 27.33 16530 48 0 40 43.550, -8 57 27.33 16341 32 189

0 41 21.950, -9 03 31.18 16646 43 0 41 21.950, -9 03 31.18 16462 77 184

0 42 12.940, -9 17 50.85 19950 67 0 42 12.940, -9 17 50.85 19930 45 -20

0 42 35.640, -9 03 48.28 16287 43 0 42 35.640, -9 03 48.28 16301 74 -14

0 43 10.730, -9 26 18.43 28724 30 0 43 10.730, -9 26 18.43 28890 93 -166

Malumuth       Malumuth Malumuth Beers         Beers Beers Difference

position        velocity error position        velocity error

0 38 30.1, -9 29 01 16256 96 0 38 30.0, -9 29 02 16213 48 43

0 38 33.7, -9 27 60 16617 83 0 38 33.7, -9 27 59 16725 64 -108

0 38 55.2, -9 30 09 14178 109 0 38 55.4, -9 30 11 14233 46 -55

0 38 56.9, -9 30 25 16530 61 0 38 57.2, -9 30 27 14022 33 2508

0 38 58.4, -9 32 13 13393 26 0 38 58.6, -9 32 15 13429 28 -36

0 38 58.6, -9 30 33 16241 96 0 38 58.7, -9 30 35 16328 46 -87

0 39 00.1, -9 36 29 13811 56 0 39 00.2, -9 36 30 13781 41 30

0 39 01.7, -9 25 51 18141 59 0 39 01.6, -9 25 52 18154 39 -13

0 39 02.9, -9 38 17 14125 46 0 39 03.3, -9 38 19 14234 42 -109

0 39 11.2, -9 42 50 16792 107 0 39 11.3, -9 42 49 16886 35 -94

0 39 16.4, -9 33 30 15851 22 0 39 16.6, -9 33 31 15912 69 -61

0 39 18.6, -9 34 38 16447 126 0 39 18.3, -9 34 37 16734 48 -287

0 39 18.4, -9 42 15 17174 118 0 39 18.5, -9 42 14 17360 69 -186

0 39 20.4, -9 46 42 17054 51 0 39 20.5, -9 46 43 17164 33 -110

0 39 21.7, -9 46 13 15181 46 0 39 21.9, -9 46 12 15392 34 -241

0 39 41.1, -9 34 17 19930 45 0 39 41.0, -9 34 17 19758 59 172

0 40 01.9, -9 27 05 16891 99 0 40 02.0, -9 27 07 16761 58 130

0 40 22.9, -9 30 20 22910 51 0 40 22.8, -9 30 16 5230 59 17680

 

To check the consistency of our calibrations between the two observing runs, we reobserved in 1995 four galaxies already observed in 1994; the agreement between the various values is better than 100 kms^-1, confirming that the mean error on our velocities is smaller than 100 kms^-1.

The final redshifts given in the catalogue are those derived from the cross-correlation with M 31, since this template gave the best results. A correction was applied to obtain heliocentric velocities.

  
Figure 14: Velocity histogram of all observed galaxies

The histogram of all the velocities in the catalogue is displayed in Fig. 14 (click here). It will be discussed in detail in a companion paper (Durret et al. in preparation).

 

Limiting 700 1000 1500 2000 2500

Completeness 88.5% (85/96) 91.9% (137/149) 87.6% (219/250) 84.7% (316/373) 77.4% (384/496)

 

We have estimated the completeness of the spectroscopic catalogue presented here by comparing the numbers of galaxies with redshifts to the total number of galaxies from our photographic plate catalogue. Results are shown in Table 2 (click here).