1.2. Clues to galaxy formation with HST

Among the many theories of galaxy formation, the idea that galaxies may form by the accumulation of smaller star-forming subsystems has recently received much attention (NW94; P96b, Low97). Two major lines of evidence lend support to this 'bottom-up' scheme. First, deep HST images and the angular size-redshift (or -z) relation of luminous early-type galaxies (E/S0) and mid-type spiral galaxies (Sabc) indicate that these objects were assembled largely before z 1, and that they have been passively evolving since z 1 (D95a, Mu94). In addition, there is evidence from both HST and ground-based work that the galaxy-merger rate was higher in the past, roughly increasing with redshift in proportion to (1 + z)^m, with m 2-3 (B94; C94a; YE95). Taken together, these two points suggest that the luminous galaxies seen today could have been assembled from the merging of smaller systems sometime before z 1. The second piece of evidence comes from the excess of blue objects at faint magnitudes which has been found in many deep imaging studies (e.g., T88, NW95a, M95). Although recent spectroscopic studies (E96; G95b; L95) have found many of these FBG's to be at modest redshifts (z ~ 0.5), the existence of a significant high-redshift 'tail' to the distribution (CHS95a) raises the possibility that a non-negligible fraction lies at z 2. 096 noted that for B 25 mag, the FBGs are extremely compact ( 0".2-0".3). This (likely) higher-redshift FBG population, which is dominated by late-type or irregular galaxies (D95a), that have undergone substantial evolution since z 1, could provide a reservoir of building blocks from which the luminous, present-day galaxies were made at z 1. Indeed, to produce the number of giant galaxies seen today, many more objects at z 2 are required by most merger-dependent models of galaxy evolution (CHS95a; B92) than we see for z 1. Recent findings, such as a population of faint, compact (B 24 mag) blue galaxies at 1 z 3.5 with rather unusual morphologies which suggest dynamical formation processes or mergers (P96b, S96a, CHS95a), seem to support this hypothesis.