We discuss here several aspects of the 
emission
in nearby galaxies which are most relevant to their application
in surveys of faint objects.
Kennicutt (1983) calibrated the relationship between the integrated
luminosity and total SFR (equations 5 and 6) in
disk galaxies, and the spectrophotometry presented in Kennicutt (1992)
can be used to derive a corresponding relation between
flux and SFR.
Practically, the calibration between SFR and 
luminosity can be
trasformed to 
luminosity once a mean values
for the 
and
ratios are known.
From the study of a homogeneous sample of normal galaxies
Kennicutt (1992) found mean value 
= 0.3
and 
= 0.5
resulting in an extinction-corrected calibration
where L is the observed emission line luminosity in units of
.
One drawback of this relation for applications to distant
galaxies is that they require a photometric measurement of the
. But absolute line fluxes are difficult to measure
in faint galaxies. However, the strong correlation between
and 
EWs (Fig. 8)
suggests that 
can be used as a quantitative
indicator of the SFR in a distant galaxy, in much the same way that
is used in nearby galaxies.
Among galaxies dominated by stellar photoionization (solid points)
the data follow a mean relation EW[OII
= 0.4
EW(
), with a rms dispersion of 
%.
Kennicutt (1992) showed that an approximate SFR can be derived if the EW and integrated broadband magnitudes are known.
where the B luminosity is expressed in units of the Sun.
However, Fig. 8 clearly shows that
the 
versus 
correlation breaks down for
galaxies with luminous active galactic nuclei. These galaxies fall
indeed into two classes, those with abnormally strong 
emission,
independent of Balmer line strength, and those with strong Balmer,
and 
emission but weak emission in 
.
The former most often are Seyfert 2.
Figure: Relation between the equivalent widths of the 
and 
emission lines. Notation the same
as in Fig. 5. The line shows a simple model with
.
