Assuming the stellar spectrum to be spatially uniform, we can subtract the
spectrum of the host galaxy from the total nuclear spectrum in order to
isolate the spectrum of the active nucleus. In practice, we progressively
subtract more of the stellar spectrum from the total spectrum until the
CaII doublet feature at 
Å vanishes. The resulting
difference spectrum is taken to be the spectrum of the active nucleus.
This is shown in Figure 2b. A similar procedure was not performed for the
red (FORS) data because the poorer spatial resolution of our FORS data
makes isolation of the host galaxy spectrum more difficult. This would
make subtraction of the stellar component rather subjective and
consequently diminish its value.
Table 1: UV/Optical/NIR line spectrum for MCG-6-30-15. Column 4 shows
the velocity redshift of the line centre with respect to the reference
frame defined by the [OIII] 
emission (z=0.00779). Column 6
shows the total flux in the line, corrected for Galactic extinction. Those
equivalent widths marked with an asterix have been measured from FORS data
which have not been galaxy-subtracted. All errors and limits are
quoted at the 1- 
level.
As is typical for Seyfert 1 nuclei, the spectrum consists of a strong
non-stellar continuum, broad Balmer lines and narrow permitted and
forbidden lines. Both the red and blue spectra were visually examined for
known prominent lines. All such identified lines were characterized by
fitting a single Gaussian profile whilst modeling the local continuum as a
power-law. This procedure was performed on the galaxy-subtracted spectrum,
with the exception of those few lines that were identified in the FORS data
(for the reasons given above). Since our stellar spectrum is contaminated
with forbidden oxygen line emission (presumably from an extended NLR), we
note that these oxygen lines will be suppressed by 
per cent in
the galaxy-subtracted AGN spectrum. The single Gaussian parameterization
is a (visually) good fit to all of the emission lines except H 
.
Three Gaussian components are required to properly describe this line:
a) A narrow line component at the systemic velocity of the galaxy (defined as
the velocity of the [OIII] emission line region) with FWHM 
and flux 
.
b) A broad line component blueshifted by 
relative to the
systemic velocity with FWHM 
and flux 
.
c) A very broad line component redshifted by 
relative to the
systemic velocity with FWHM 
and flux 
.
The resulting line identifications, wavelengths, line widths and total line
fluxes for all of the identified lines are given in Table 1. The errors
quoted in this table (and those for the H components above) include
both statistical errors and an estimate of any systematic errors resulting
from the wavelength/flux calibration. The statistical errors are derived
from 
fitting of the Gaussian models to the data, assuming that the
pixel-to-pixel variation is due to random Gaussian noise.
The optical spectrum clearly shows the effect of dust extinction: the continuum flux declines towards the blue end of the spectrum and the Balmer decrements are large. The line-of-sight extinction, and a comparison of this extinction to the line-of-sight X-ray absorption will be addressed in Section 3.