Notice: Due to the conversion of some ACS WFC or HRC observations into
WFPC2, or NICMOS observations after the loss of ACS CCD science
capability in January, there may be an occasional discrepancy between a
proposal's listed (and correct) instrument usage and the abstract that
follows it.

Nearby compact galaxy groups are uniquely suited to exploring the
mechanisms of star formation amid repeated and ongoing gravitational
encounters, conditions similar to those of the high redshift universe.
These dense groups host a variety of modes of star formation, and they
enable fresh insights into the role of gas in galaxy evolution. With
Spitzer mid-IR observations in hand, we have begun to obtain high
quality, multi-wavelength data for a well- defined sample of 12 nearby
{<4500km/s} compact groups covering the full range of evolutionary
stages. Here we propose to obtain sensitive BVI images with the ACS/WFC,
deep enough to reach the turnover of the globular cluster luminosity
function, and WFPC2 U-band and ACS H-alpha images of Spitzer-identified
regions hosting the most recent star formation. In total, we expect to
detect over 1000 young star clusters forming inside and outside
galaxies, more than 4000 old globular clusters in >40 giant galaxies
{including 16 early-type galaxies}, over 20 tidal features,
approximately 15 AGNs, and intragroup gas in most of the 12 groups.
Combining the proposed ACS images with Chandra observations, UV GALEX
observations, ground-based H-alpha imaging, and HI data, we will conduct
a detailed study of stellar nurseries, dust, gas kinematics, and AGN.

Ultraviolet observations have revealed remarkable diversity among old
stellar populations in globular clusters and E/S0 galaxies. We recently
discovered with HST/STIS that globular clusters in the giant elliptical
galaxy M87 have the most heavily populated hot horizontal branches of
any stellar systems yet studied. Their far-UV/optical colors are up to 1
mag bluer than any Milky Way globular cluster and approach the
theoretical limits for production of hot-HB stars in old stellar
populations. The differences among the metal-poor clusters are
particularly interesting, because it is thought that these objects
reflect the earliest stages of galaxy formation at high redshifts. Here
we propose deep ACS far-UV imaging of a second gE galaxy, NGC 1399, with
a cluster system that is well-studied at longer wavelengths, to
determine whether it shares characteristics with M87. These observations
bear on aspects of advanced stellar evolution, on the histories of
globular clusters in different environments, and on the interpretation
of the "ultraviolet upturn'' phenomenon in elliptical galaxies and its
value as a population probe in distant galaxies.

Eta Carinae is now known to be undergoing some unusually rapid changes
on a timescale of several years. They are probably essential for
modeling the star's long-term recovery from its Giant Eruption 160 years
ago -- the prototype "supernova impostor" event. Since high spatial
resolution is needed to isolate the central star, and the present state
will probably not recur in the future, it is important to obtain HST
data during the next two years. We propose a cost-effective set of ACS
observations with three goals: {1} to obtain a continuing record of the
star's rapid UV and visual brightening; {2} to lengthen the temporal
baseline of ACS images enough to settle an important question concerning
ejecta ages; and {3} to extend the record of morphological changes in
the inner ejecta past the midpoint of eta Car's 5.5-year cycle.