One of the outstanding unanswered questions in extragalactic
astronomy is the nature of the IR-luminous galaxies discovered by IRAS.
These galaxies emit nearly all their power in the far-IR but their
nuclei are so thoroughly enshrouded by dust that previous studies have
failed to reveal the source of their quasar-level luminosities.  It is
not known whether they contain heavily obscured non-stellar central
engines or if their extraordinary luminosities are produced by stellar
processes such as enormous bursts of star formation. This is of
particular interest because they may dominate the space density of
high-luminosity objects in the Universe, and therefore are being
discovered at cosmologically significant distances.
Only spectroscopy at long wavelengths can penetrate the enormous
extinctions present (Av=20--40). Whereas the Core Programs are observing
several such galaxies in depth, here we take a complementary approach: a
larger survey for the 2 strongest high-ionization forbidden lines
emitted by Seyfert nuclei, from our complete IR-flux-limited sample of
galaxies. The [O IV]26um and [Ne V]14.3um lines are the most
definitive indicators of photoionization by an obscured Seyfert nucleus,
and are not produced by O and B stars. Neither line can be detected by
any instrument other than ISO. Our sample was constructed using every
available indicator of nonstellar nuclear activity that still works in
the presence of heavy reddening: hot IRAS colors, extremely high IR
luminosity, or strong radio or x-ray emission. We will search for hidden
AGNs in each galaxy which meets these criteria, yet is not optically
classified as a Seyfert. Our sample will be sufficiently
large for the required statistical accuracy to determine the
true space density of all AGN types--free from incompleteness due to
dust obscuration--for the first time.