The history of radio astronomy has been one of ever increasing sensitivity averaging about an order of magnitude per decade. The most sensitive observations with the current VLA reach levels of the order of 10 microJy in selected regions and about 1 mJy over the whole sky, or one to two orders of magnitude weaker than any other radio telescope. A wide variety of sources are observed at all flux density levels including radio galaxies and quasars, BL Lac Objects, Seyfert and other star forming galaxies, low luminosity AGN and normal galaxies. Each population has its own luminosity function which may evolve independently with cosmic epoch.
The enhanced VLA promises to continue this process reaching the 1-microJy level for the first time. As we have probed deep we have found new populations of sources, most recently the apparent domination of starburst galaxies below 1 mJy. It seems quite possible the universe will change again when we probe the next step deeper.
However, what we predict most clearly is that we will be able to study the changes in the properties of the currently known populations with epoch. It seems very likely that these changes are related to the other properties of the universe which are different at earlier epoch such as the clustering environment. In particular, it is important to look for changes in the environments of Fanaroff-Riley I and II radio galaxies with epoch. Does the change in morphology occur at the same luminosity at earlier epochs? Does the cluster environment of the different types of sources change? Is there an epoch at which the morphological classes no longer divide in this way at all, like the changes HST sees optically?
Thus we want to study samples of radio galaxies which in areas
of sky that will be studied intensely at other wavelengths. We
want to make high quality images, not merely detections, of these
sources. For sources below the Fanaroff-Riley I/II break, radio and
optical work are now practical to z 
. With the enhanced VLA,
planned space missions in the IR and ground-based millimeter
and sub-millimeter instruments like the MMA, it should be possible to
reach z 
for these radio luminosities.
Another example of such a study is the properties of the radio quiet
quasars. Although the first quasars were discovered as a result of
their strong radio emission, subsequent observations have shown that
quasar radio emission is bi-modal with only a small fraction, of the
order of 10-15% of optically selected quasars found as strong radio
sources with 5 GHz radio luminosity greater than 
or a ratio of radio to optical luminosity greater than 10:1.
Understanding the relation, if
any, between the radio loud and radio quiet quasars has been a
challenging problem of contemporary quasar astronomy.
Only the VLA has been able to even detect the weak radio emission from the radio quiet quasars, and only from the nearest ones. The greater sensitivity of the enhanced VLA will let us explore the lower end of the radio quiet luminosity function out to greater redshifts where the quasar population density rapidly increases and to determine whether indeed there is a population of truly radio silent quasars.
We want to image these objects and understand their relation to nearby objects like Seyfert galaxies and the nature of their changes with epoch. How do these objects fit into the changes in clustering properties with redshift? Does their radio structure indicate they smoothly join with the properties and Seyferts or do they have radio structures which indicate they are related to the Fanaroff-Riley I's or II's as seen at high z? The key here is the increased sensitivity that will allow high quality imaging of these weak sources at reasonably high redshift.
The tremendous increase in sensitivity at the short centimeter
wavelengths also offers the possibility of surveying large parts of
the sky at these frequencies. At 30 GHz a survey as deep as the
current 
20cm D-configuration (NVSS) survey could be made to the same
limiting flux density in the same length of time ! The telescopes
would have to be continuously scanned rather than tracking one spot
and correlator dumps would be necessary every 0.05 seconds.
Alternatively, much deeper surveys of much smaller regions could be
made. We know we would find large numbers of steep spectrum sources,
as have been cataloged at 6cm and longer, and probably
populations of faint flat spectrum QSOs; however, based on surveys at
other wavelengths it is likely that other sunrises would await such
studies.