|Annu. Rev. Astron. Astrophys. 1997. 35:
Copyright © 1997 by . All rights reserved
2.3. Images from Hubble Space Telescope
Since the review by Koo & Kron (1992), multicolor HST images have become available for galaxies to the limits achievable with ground-based spectroscopy and beyond. The images are spectacular and provide morphological data for large numbers of distant galaxies, a growing number of which have redshifts. The potential for directly tracking the evolution of the Hubble sequence of types is clearly an exciting opportunity. However, interpretation of the HST data in this way has not been straightforward (cf discussion by Abraham et al 1996a, b). Although many workers have concluded that the faint field population is dominated by "irregular" and "merging" galaxies (Glazebrook et al 1995b, Driver et al 1995b), both the reliability of this result and a physical understanding of what it signifies require careful study. When comparing distant images with possible local counterparts, the reduced signal-to-noise ratio, relative increase in background, and cosmological surface brightness dimming tend to accentuate the visibility of high contrast features. Furthermore, bandpass shifting effects arising from the k-correction mean that low redshift optical images are being compared with high redshift UV images (O'Connell & Marcum 1996). An illustration of these effects is given in Figure 4, following the work described by Abraham et al (1996b). Such studies indicate that, although the broad classification system (E/Spiral/Irr) is reliable to z 1 for most large regular galaxies, any unfamiliarity in the images of higher z systems may be spurious.
Figure 4. Simulated appearance of NGC 4450 (Sab), 3953 (Sbc), 5669 (Scd), and 4242 (Sdm) at various redshifts. With the exception of NGC 4242, which is subluminous, all galaxies have L L*. From top to bottom the panels show (a) the rest-frame B band images, (b) the appearance as viewed at z = 0.75 in a three-orbit F814W WFPC-2 image, and (c) the appearance at z = 1.5 in an exposure equivalent to the HDF. (Courtesy of R Abraham.)
The concept of comparing HST images of galaxies grouped by class at different redshifts raises a very important problem. Most of the above discussion has centered on statistical results from surveys. By examining the entire field population accessible at each redshift, robust statements are made about the global changes in, for example, the galaxy LF and volume-averaged star-formation rates (SFRs). Yet, to physically understand these changes in terms of the elaborate models now available, we are encouraged to divide the samples into "subclasses" in the hope of determining the evolutionary behavior as a function of type. This can only be done if there is some fundamental slowly changing property of a galaxy that can act as a label. Clearly, both color and spectral class (cf Lilly et al 1995, Heyl et al 1997) could be transient properties affected by short-term changes in star formation induced, for example, by merging or other processes. Likewise galaxy morphology, even if correctly assigned in HST images that sample a range of rest-frame wavelengths, may change with time. White (1996) discussed physical situations where gas-rich disk galaxies may merge to form gas-poor spheroidals, only to accrete more gas and become spirals again. The possible migration of galaxies in and out of a faint blue category by some unidentified time-dependent process may be an important stumbling block to progress.