Results and discussion

In the figure, versus radius, with the associated errors, is plotted for all the regions catalogued and measured in NGC 4449. For a classical H II region (i.e. in the absence of extinction and gas inhomogeneities) the log of the luminosity is expected to be proportional to the log of the radius according to the relationship

 
Figure 1: Overplotted on the observed data-points is the function .

We have performed a linear fit to the and values of the lower luminosity data points (). The best fit to the physical model of classical H II regions is given by:

From the figure two ranges, both in R and in luminosity, can be distinguished. For radii larger than arcsec or larger than erg s (the giant H II regions, hereafter GHIIRs) most H II regions are less luminous than expected. For the less luminous and smaller regions the increase of luminosity with size follows, within the errors, the trend expected for ionization bound classical Strömgren spheres. This segregation is not related with any particular spatial distribution of the regions. Both classes, the smaller ones and the GHIIRs, are uniformly distributed in the galaxy. This also allows us to rule out any possible differential extinction as the cause of the leakage of photons coming from regions located in especially dusty areas. Also the possibility that the most luminous regions are also more heavily shrouded in dust is unlikely in dwarf irregulars known to be relatively metal-poor systems (see Kunth 1995, and references therein).

The change for the more luminous regions might be due to the fact that these regions have exceeded the thickness of the galactic disc and are therefore losing photons. These photons, leaking out from the large H II regions of the galaxy, might be an important source in ionizing the DIG. The total luminosity, , of the subsample is erg s. If, on the other hand, one assumes that they should follow the same behaviour as the smaller H II regions, provided they are not losing photons, their luminosity, using equation (2), would be, erg s. The difference is erg s or 54% of the luminosity escaping from the GHIIRs and contributing to the DIG. The contribution to the total H luminosity coming from the other H II regions, those less luminous than erg s, is erg s. Adding all these contributions, a total luminosity in H erg s is obtained. This is lower that the erg s reported by Hunter & Gallagher (1986). The discrepancy comes from the fact that our observations cover only the central arcsec, whereas the galaxy extends more than twice as far. However, most GHIIRs of NGC 4449 are included in our data, and so if photons leaking from giant star formation complexes are the only source of the DIG luminosity the relative contribution of the DIG ( erg s) to the total measured H luminosity ( erg s) would be about 30%.

Previous studies, using CCD images, to determine the DIG luminosity of NGC 4449 (Kennicutt et al. 1989; Hunter & Gallagher 1990) from the total H luminosity of the galaxy give values of the relative contribution in the DIG of 50% and 20%, respectively. The difference is probably due to the different lower isophotal surface-brightness cut used to define the DIG; the method is therefore also dependent on the exposure time and on the observational set-up.

Among the 44 H II regions isolated and classified in this galaxy, the biggest, or brightest, larger than about 2 arcsec and with L brighter than about erg s, follow a linear relationship between and with a slope smaller than that for less luminous regions, departing from the expected behaviour of ionization bound classical HII regions. This is considered as direct evidence of photons being lost from GEHRs, perhaps limited by the disc size and undergoing champagne flows. The boundaries of GHIIRs are ionized and champagne flows must be taking place. These can be an important source in providing the material detected as DIG in the galaxy NGC 4449. The evaporation of the GEHRs may well be the mechanism to supply the diffuse medium which bathes other irregulars with important star formation activity.

© Copyright Astronomical Society of Australia 1997