4 Data reduction

4.1 Bias and flat-field corrections

The science frames have been first bias subtracted. For each observing run, we have obtained the proper bias by combining several bias frames with a median filter. Then the images have been corrected for pixel to pixel response variations. For each night, its own flat field has been constructed by medianing several flat field frames carried out during the night.

Two different types of flat fields have been used during the three observing runs: the first one has been obtained by medianing twilight sky images and it has been used to reduce the data of 1996. The other one, used in 1995 and 1997 runs, has been constructed by medianing flat field frames taken on the dome illuminated with twilight sky. No significant differences have been measured by using the two flat fields.

These two steps of data reduction are based on the NOAO IRAF package, developed at the Center for Astrophysics.

Finally, cosmic rays and bad pixels have been removed from each frame using Munich Image Data Analysis System (MIDAS).

4.2 H$_\alpha$ emission-line map

The map of the H$_\alpha$ emission-line flux for each HCG (H$_\alpha$image) has been obtained by removing the contribution of the underlying continuum, that it is by subtracting the H$_\alpha^{\rm off}$ from the H$_\alpha^{\rm on}$ (Pogge 1992). There are several reasons why the number of continuum photons per integration time unit passing through the on filter can be different from the one through the off filter. For example differences between the transmission curves of the two narrow-band filters, such as different width and/or transmission peak; or variations of the sky transparency during the night. This implies that in order to obtain the H$_\alpha$ emission-line flux image a careful estimation of the underlying continuum to subtract from the H$_\alpha^{\rm on}$ is required. In practice, the H$_\alpha^{\rm off}$ have to be rescaled to the continuum of the H$_\alpha^{\rm on}$ wavelength. Since stars do not show H$_\alpha$ emission, the number of continuum photons coming from the stars in each HCG field and passing through the on and off filters have to be the same. In-fact, although the on and off filters are sometimes separated by more than 150 Å, implying that the number of photons coming from the stars is not rigorously the same, such a difference is negligible. Thus for each HCG field (and for each couple of filters) we have selected at least three stars and we have calculated the mean scaling factor K

$$K=\frac{1}{N}\sum_{i=1}^{N} \left( \frac{C_{\rm on}}{C_{\rm... ..._i=\left<\left(\frac{C_{\rm on}}{C_{\rm off}} \right) \right\gt$$ (1)

where $C_{\rm on}$ and $C_{\rm off}$ are the counts from stars in the on and off image respectively, and N is the number of stars. Once rescaled, the H$_\alpha^{\rm off}$ have been spatially aligned to the H$_\alpha^{\rm on}$.The alignment has been performed by applying the IRAF tasks geomap/geotran using the position of at least five stars in the field as reference coordinates. Finally, after having additively rescaled the on and off images to the same median value, we have subtracted the H$_\alpha^{\rm off}$from the H$_\alpha^{\rm on}$ thus obtaining the image of the H$\alpha$emission-line flux.