4 Image analysis

The data reduction of the CCD images follows a procedure identical to the one described in previous papers of the series (Gavazzi et al. 1998 Paper I), based on IRAF/ STSDAS data reduction packages. To remove the detector response each image is bias subtracted and divided by the mean of several flat field exposures obtained on the twilight sky. Calar Alto flat fielding is sometimes poor because of filter vignetting. In some cases images were fitted with a 2-D polynomial function to remove second order structures in the image. Vignetting problems, combined with generally shorter exposures, make the Calar Alto images of poorer quality with respect to the OHP ones. When three images in the same filter are available, a median combination of the realigned images allows removal of cosmic rays. For galaxies with only one available image, direct inspection of the frames allows manual cosmic rays removal. Subtraction of contaminating objects, such as nearby stars and galaxies, is done by direct editing of the frames. The sky background is determined in each frame in concentric object-free annuli around the object. The typical uncertainty on the mean background is estimated 10% of the rms in the individual pixels. This represents the dominant source of error in low S/N regions.

Total counts in the two frames have been obtained by integrating the pixel counts over the area covered by each galaxy, as derived by the optical major and minor diameters. H$\alpha$+[NII] fluxes and equivalent widths have been determined using Eqs. (1) and (2) given in Paper I.

 

 

Table 4: The characteristics of the narrow band filters.

$\lambda$	$\Delta \lambda$	$R(\lambda)$	$\int R(\lambda)$d$\lambda$	Telescope
6450	45	81	37.80	1.20 m OHP
6561	48	78	38.45	1.20 m OHP
6612	55	68	40.60	1.20 m OHP
6569	113	83	101.00	1.23 m CA
6744	97	80	82.15	1.23 m CA

Equation (2) shows that once the normalization constant $n_{\rm\frac{ON}{OFF}}$ between ON and OFF band frames is known, the H$\alpha$+[NII] equivalent width can be estimated also in non photometric conditions. The zero point of each galaxy is determined assuming an extinction law of slope 0.1 for the OHP and 0.12 for the Calar Alto observations respectively.

At the redshift of Virgo the Calar Alto OFF band filter ($\lambda$ 6744  Å) is partly contaminated by the emission of the [SII] dublet at $\lambda$ 6717  Å and 6731  Å. We corrected for this effect using Eqs. (3), (4) and (5) given in Paper I.

Errors on the H$\alpha$+[NII] flux and EW are estimated from Eqs. (6) and (7) of Paper I.