6.2 Luminosity Function
How do active star forming galaxies, in particular dwarf systems contribute to the luminosity function (hereafter LF) as compared to normal galaxies? To some extent, although somewhat paradoxical, the LF of local objects is more difficult to derive. Completeness problems are severe and combined with the need of large spatial coverage. Even when samples become large enough, the problem remains to establish the nature of selected objects. Low dispersion spectra must be used and require a diagnostic diagram to disentangle H II region-like spectra from AGN. Most LF determinations use a Schmidt estimator (Schmidt 1968) combined with a V/Vm test for completeness. One must bear in mind that such a test may not be adequate if large structures of our local Universe are not sufficiently averaged out. The very faint end of the LF is naturally the most difficult to establish. Deep CCD imaging using narrow-band filters complemented by follow up spectroscopy (Boroson et al. 1993), dedicated to this problem points towards a moderate slope of the LF at low luminosities in agreement with larger samples from the Case surveys or the UM. These studies are not in agreement with others, showing that this problem will be settled only by a better understanding of selection biases. Certainly future surveys involving CCD techniques will allow to build larger and deeper samples.
The space density of dwarf emission line galaxies has been addressed by Salzer (1989), who finds ~ 0.03 galaxies per Mpc3, corresponding to ~ 7% of the local field galaxy density based on the UM survey. This number is dominated by intrinsically faint systems. A rough calculation of the space density of SBS galaxies from Pustilnik et al. (1995) immediately shows these to be an order of magnitude less abundant, probably an effect of different selection criteria. The Case survey is even richer than the UM survey and the emission line galaxies (ELGs), which it contains, may contribute up to one third of the field galaxy population (Salzer et al. 1995). The reason is that the Case survey also is sensitive to galaxies with a low level of H II activity. A comparison of derived space densities of emission line/UV-excess dwarfs is given by Comte et al. (1994) and Gallego et al. (1997). These differences demonstrate once more that different surveys effectively target different types of galaxies.