Quark stars are what you might get if a collapsing star were to proceed
beyond the neutron-star condition in which the star's constituents are
chiefly neutrons (each neutron is itself a sovereign parcel consisting,
in the main, of two down quarks and one up quark) all the way to a condition
in which the very neutron membranes would be dissolved, allowing the
quarks to run together. Brookhaven's RHIC collider attempts to do something
like this on a much smaller scale when it smashes together two gold
atoms.
In the case of a quark star it is immutable self-gravity rather than
manmade accelerator gradients that provide the needed crushing power.
Under these conditions it might be energetically feasible for many quarks
to exist as strange quarks rather than the lighter up and down quarks.
Hence the name "strange stars."
Evidence for quark stars comes now in the form of observations of two
neutron stars, viewed at x-ray wavelengths by the Chandra x-ray telescope
and in the visible by the Hubble Space Telescope.
One of the objects, RXJ1856, is too small (to judge by its wealth of
x ray and dearth of visible emissions) to be a conventional neutron
star made primarily of neutrons. The other object, 3C58, seems to have
cooled too quickly (to judge by its present measured warmth and its
known lifetime, drawing upon medieval Chinese records of the object's
birth as a supernova in 1181 CE) to be an ordinary neutron star.
In both cases the observations tally better with a star of quarks (one
big nucleus), or a mix of quark and neutron layers. (Two articles to
appear in Astrophysical
Journal; 3C48 preprint, Slane et al., astro-ph/0204151;
RXJ1856 preprint, Drake et al., astro-ph/0204159).
A day after the press conference at which these results were announced,
a new preprint (astro-ph/0204199)
appeared which suggests that the distance to JXJ1856 is actually further
away than the earlier estimate and that the object need not be a quark
star after all.