The broad emission lines in NGC 3516 as observed with HUT show subtle
differences compared to other Seyfert 1's and low redshift AGN.
The optical to X-ray spectral index for NGC 3516, 
,
is typical of other Seyfert 1's ([Kriss & Canizares 1985]),
but the flux ratio of O VI 
to Ly
is only 0.18. This lies below the correlation
of O VI/Ly
with 
discussed by Zheng, Kriss, &
Davidsen (1995)ZKD95, but is comparable to the
mean value of 0.17 seen in high redshift quasars ([Laor et al. 1994]).
As in the case of Fairall 9, however, the lack of a soft X-ray excess in
NGC 3516 may be a significant factor in producing lower O VI/Ly
compared to other Seyfert 1's ([Zheng et al. 1995]).
The other noticeable difference is the relatively high strength of broad
C III 
and N III 
emission
in our spectrum.
Although these lines have been seen in other low-redshift quasars
([Laor et al. 1994]; [Laor et al. 1995]), they are not detected in other HUT spectra of
Seyfert 1's. The ratios of these lines to each other and to O VI
is also typical of that seen in the low redshift
quasars observed with HST ([Laor et al. 1995]).
Reverberation mapping experiments have shown that the broad emission-line
region (BELR) of AGN is highly stratified in both spatial distribution
and in ionization parameter (e.g., [Clavel et al. 1991]; [Krolik et al. 1991];
[Reichert et al. 1994]; [Korista et al. 1995]).
Nevertheless, it is useful to match a single-zone photoionization model to
the observed broad line emission in NGC 3516 to obtain a fiducial
for comparison to other Seyferts. In addition, it provides a comparison to
models of the warm absorbing medium seen in the X-ray spectrum and in the
UV absorption lines that we discuss in §4.
We use the photoionization code XSTAR ([Kallman & Krolik 1993]) to compute a grid of
photoionization models varying the total column density N and the ionization
parameter 
, where
is the number density of ionizing
photons between 13.6 eV and 13.6 keV illuminating the cloud and
is the density of hydrogen atoms.
As we are constraining only high ionization lines that are insensitive to the
density, we assume constant density clouds with 
and
solar abundances.
For the incident photoionizing continuum we use the extinction-corrected UV and
absorption-corrected X-ray spectrum of NGC 3516 as
described by Kriss et al. (1996)Kriss96b.
The UV power law was extrapolated to higher energies following
with a break at 51 eV to the slope of the X-ray
power law, 
.
To match the observed line ratios we varied the parameters until we achieved
the closest fit to the strongest lines --- O VI+Ly
, Ly
,
Si IV+O IV], C IV, and He II 
.
The closest match is for an ionization parameter 
.
The resulting line ratios from this model are compared to the observed values
and their error bars in Table 3.
Choosing 
is mainly a balance between
the O VI emission and the C IV line.
Producing sufficient O VI requires 
, but then
the C IV/Ly
ratio becomes too high.
fits the C IV/Ly
intensity ratio well, but the
O VI emission is then a factor of 2 less than observed.
As the lines we are matching originate in the high ionization illuminated
faces of the BELR clouds, their intensities are rather insensitive to changes
in the column density above 
in our models.
The best-fit ionization parameter 
is a typical value for
single-zone models of AGN broad-line regions, but one can see from
Table 3 that this model is insufficient to explain all the
observed line ratios.
The most noticeable differences are in the relatively high strengths observed
for C III 
, N III 
, and
He II 
.
Given the compromise we made between matching O VI and C IV,
one can envision that a higher ionization zone producing relatively more
O VI and He II 
could account for part of these
differences.
An additional population of higher density clouds may be required to
produce the enhanced C III 
, which becomes a more
important coolant as other lines become optically thick ([Netzer 1990]).
The strong N III 
is more of a puzzle.
Most photoionization models predict it to be weaker than
C III 
, yet in our spectrum and in the HST quasar spectra
([Laor et al. 1995]) it is stronger.
It is possible that fluorescent mechanisms could enhance the strength of
this transition under favorable circumstances ([Ferguson, Ferland, & Pradhan 1995]).
Although the broad lines in NGC 3516 seem to be mostly produced in a region
with an ionization parameter typical of the BELR in other AGN,
is an order of magnitude lower than the ionization parameter
required for the warm absorbing gas detected in the ASCA X-ray spectrum
([Kriss et al. 1996]).
This does not rule out an origin in the BELR for the X-ray warm absorbers,
but these absorbers must be physically distinct from the clouds producing
the bulk of the broad-line emission.
