2 Observations, data reduction and the production of the HII region catalogue

The observations of NGC 7479 were made during the night of May 20$^{\rm th}$, 1992 with the 4.2 m William Herschel Telescope (WHT) on La Palma. We used the TAURUS instrument in imaging mode as a re-imaging camera, mounted at the Cassegrain focus of the telescope. The detector used was an EEV CCD 7 with projected pixel size 0$.\!\!^{\prime\prime}$279 $\times$ 0$.\!\!^{\prime\prime}$279. Observing conditions were very good with 0$.\!\!^{\prime\prime}$8 seeing (FWHM as measured in the final images) and photometric sky. We obtained two exposures of 1200 seconds: one through a 15 Å ($\sim$685 km s^-1) wide filter whose central wavelength coincided with that of the redshifted H$\alpha$ emission from the galaxy, and another through a nonredshifted H$\alpha$ filter ($\lambda$6565 Å with 15 Å width) for continuum subtraction.

Standard reduction routines were used; bias level was first subtracted, and the images were then corrected using appropriate dawn sky flatfields. Next, the images were aligned and cleaned of cosmic ray effects and then the continuum image was subtracted from the line (H$\alpha$ + continuum) image, giving the net H$\alpha$ flux. The procedure is described with more detail in Rozas et al. (1996a).

Absolute flux calibration was carried out using observations of standard stars from the lists of Oke (1974), Stone (1977), and Filippenko & Greenstein (1984). The luminosity in H$\alpha$, corresponding to a single instrumental count is $9.27 \ 10^{34}$ erg s^-1 count^-1.

Before producing the HII region catalogues, we flagged the foreground stars in the image. These are distinguishable from HII regions by their regular, circular shapes in the original, unsubtracted image, and because they show much more intensity in the continuum than in the corresponding H$\alpha$ continuum-subtracted image. (Ideally, foreground stars should not show up at all in the HH$\alpha$ continuum-subtracted image, but in most cases some residual is seen, due to e.g. differences in point spread function, alignment, or stellar emission between the line and continuum image; or because the star is saturated in one or both images.) Emission in the H$\alpha$ image coincident with a foreground star on the continuum image was considered residual starlight, and not entered in the catalogue as an HII region. Any HII region with a superposed foreground star is readily detected both morphologically and via its anomalous line to continuum ratios, and rejected. As a selection criterion for HII regions we specified that a feature must contain at least nine contiguous pixels, each with an intensity of at least three times the rms noise level of the local background. Any object not meeting this criterion was indistinguishable from noise and therefore treated as noise. The rms noise of the background-subtracted H$\alpha$ image is  15 instrumental counts, the lower limits to the luminosity of the detected HII regions, and to the radius of the smallest catalogued regions (the last two quantities are derived directly from the adopted selection criterion) are, respectively, log $L_{\rm H\alpha}$ = 37.65 erg s^-1 and $\approx 75$ pc.

In identifying and quantifying the parameters of the HII regions we had to overcome three complicating effects. Firstly, many HII regions appear to overlap on the image. Without attempting to analyze what fraction of these overlaps implies real contact and what fraction is merely a projection effect, we adopted the solution proposed in Rand (1992) and followed in Knapen et al. (1993) and Rozas et al. (1996a) of counting each peak in H$\alpha$ as representing a single HII region. The flux of each HII region was then estimated by integrating over the pixels which could be reasonably attributed to a given region. One will undoubtedly miss a number of HII regions that are too weak to be detected in the vicinity of stronger emitters close by. This will influence the lower end of the LF but is not a significant factor in the determination of the shape of the true LF at the higher luminosity end (Rand 1992). Secondly, an HII region is not necessarily circular. Thirdly, the presence of diffuse H$\alpha$ may lead to ill-defined edges of HII regions, introducing some systematic errors, above all for the weakest regions (see Sect. 9, and Zurita et al., in preparation). The detection and cataloguing of the HII regions were performed using a new program developed by one of the authors (Heller et al., in preparation).

The program, identifies each HII region, measures the position of its centre, derives the area in pixels and the flux of each region, integrating all the pixels belonging to the region and subtracting the local background value. The background values were fixed before running the code. As the background was not constant across the frame, we selected 96 areas in the image with circular aperture of radius 7 pixels, so that an appropriate value determined from the nearest background area was applied to each HII region. As a result of test measurements, we found that the uncertainties in the HII region fluxes caused by the variation in the background lie between $\sim$10% for the faintest regions and < 1% for the brightest.

The code allows us to edit the catalogue by deleting, adding, separating or rounding (with a circular aperture) the regions by hand where the automatic process produces regions with faint wings of very irregular shape.

The number of catalogued HII regions in NGC 7479 is 1009, and for all the HII regions we determined equatorial coordinate offsets from the nucleus and deprojected distances to the centre (in arcsec), using the inclination angles and position angles given by Laine & Gottesman (1998) ($i=51\hbox{$^\circ$}$, PA = 22$^\circ$). We also determined the diameter and the HH$\alpha$ luminosity (in erg s^-1) of each HII region. The catalogues are available via the CDS, or from the authors. In Fig. 2 we show schematically the positions of the HII regions in the disc of NGC 7479, on a deprojected RA-dec grid centred on the nucleus of the galaxy.

  

Figure 2: Representation of the positions of the measured HII regions. Symbols show ranges of log L. Coordinates of the centre of the image are RA = 23$^{\rm h}$ 4$^{\rm m}$ 56.64$^{\rm s}$ Dec = 12$^\circ$ 19' 22.9'' (J2000)