The impact of HST on the field of galaxy evolution cannot be overstated. The revolution that we have witnessed in the last five years in the field of distant galaxies is due primarily to the high resolution imaging capability of the WFPC2 camera (though Keck spectroscopy has also played a key role). The morphological and structural information that can now be obtained for high redshift galaxies has placed the studies of distant galaxies onto a much more quantitative footing. So much so, that it is hard to conceive of this field without the low background, high resolution, wide-field imaging capability that we get from HST. The impact that the Hubble Deep Field (Williams et al. 1996) had on this field is well known. The number of papers generated from this tiny, five (arcmin)2 region is remarkable. Even now, four years after the data were taken, it is still a joy to look at the color images of the HDF. The range in scales and morphologies of the galaxies in this image is astonishing, as can be seen in Figure 2, where an expanded section of the HDF is shown.
The HDF-S added a comparable region to the southern sky opening up more opportunities for ground-based imaging (particularly JHK IR imaging) and spectroscopy (e.g., with the VLT telescopes). A number of additional fields have been imaged with HST, one of which is proving to be valuable in a quite different way. Ground-based multiobject spectrographs typically have much larger fields than the WFPC2 field, and so single HST pointings are not ideal for efficient spectroscopic followup on the ground. For example, the LRIS field on Keck covers about three contiguous WFPC2 areas, and the next generation of even wider field, multiobject spectrographs will cover more than twice the field of LRIS (e.g., DEIMOS on Keck has over 15' of useful field in one direction) and so 6-7 contiguous pointings in a row are really needed from HST. Thus, HST imaging surveys like the WFPC2 GTO team survey strip that has 28 contiguous WFPC2 images (2.5' x 42') in F606W (2800 s) and F814W (4200 s) are optimal for the new redshift surveys, like the Keck DEEP project (Davis and Faber 1999; Koo 1998). While not as deep as the HDF, the ``Groth strip'', as the GTO team survey area is known, is proving to be an invaluable survey region. The magnitude limit for obtaining photometric and structural parameters for distant galaxies is comparable (~ 23 mag) to that for which redshifts can easily be determined using the Keck telescope and a multiobject spectrograph.
The future for HST imaging, particularly with the new wider-field, more efficient ACS, will likely be a combination of very deep, multiband images (like the HDF) for photometric redshift studies of the very faintest objects, and larger field, multiple-pointing, multiband imaging surveys (like the ``Groth strip'') that can be combined with the redshifts and velocity widths determined from spectroscopic surveys with the many 6.5-10 m ground-based telescopes.