In spite of the null results from BALs, there is a growing
consensus from the BELs and intrinsic NALs for typically
solar or higher metallicities in high-redshift QSOs. These
results support models of galaxy evolution wherein vigorous
star formation in galactic nuclei (or dense proto-galactic
condensations) produces Z 
Z
in the gas at redshifts z 4. It is interesting
to note that the high metal abundances
in dense quasar environments sharply contrast with
the metallicities measured elsewhere at high redshifts.
For example, the damped-Ly
absorbers in QSO spectra, which
apparently probe lines of sight through intervening disk galaxies
(Prochaska & Wolfe 1998
and refs. therein),
have mean (gas-phase) metallicities of order 0.05 Z
at z ~ 2-3
(Lu et al. 1996,
Pettini et al. 1997,
Lu et al. 1998).
The Ly forest systems,
which presumably probe more extended and tenuous inter-galactic structures
(Rauch 1998),
typically have metalicities < 0.01 Z
at high redshifts
(Rauch et al. 1997,
Songalia & Cowie
1996).
The much higher metal abundances near QSOs are consistent
with the rapid and more extensive evolution
expected in dense environments
(Gnedin & Ostriker
1997).
These are exciting times for quasar abundance work.
The results so far have only scratched the surface of what
is possible. The new generations of large ground-based and
space-based telescopes are or will soon make it possible to
greatly extend the results discussed above.
In particular, we will be able to 1)
measure a wider variety of both emission and absorption
diagnostics and 2) compare the derived abundances in large QSO
samples that span a wide range of redshifts and luminosities.
The new data will thereby
test further the reliability of each diagnostic, search
more definitively for trends with redshift or luminosity,
examine the range of QSO abundances at any given z or L,
and make better measurements of specific evolution probes like
the Fe / clock. It will be
particularly interesting to compare emission and absorption diagnostics
in the same objects. For example, one goal should be to test the BEL result
for high Fe/Mg abundances by observing FeII/MgII
or perhaps FeII/SiII in NAL systems.
Inevitably, these studies will also improve our general understanding of observational trends between QSO luminosities and the BEL spectra. Abundance variations might be an important ingredient in calibrating the Baldwin Effect for tests of cosmology (Section 6, also Korista et al. this volume).
Acknowledgments
I am greatful to my close collaborators, T. Barlow, F. Chaffee, G. Ferland, C. Foltz, V. Junkkarinen, K. Korista and J. Shields, for their contributions to this work. I also acknowledge support from the Space Telescope Guest Observer program and from NASA through grant NAG 5-3234.