1.5. How much clustering is allowed at high redshifts and on what scales?

It is becoming clear that galaxies over a wide range in redshift are more inclined to be in groups or clusters than isolated, giving the term "field galaxy" less meaning than it once had. As such, it is perhaps not surprising to find that in many "pencil-beam" galaxy redshift surveys, often several galaxies are found in a single narrow redshift bin (e.g., the HDF redshift distribution reported by C96 and Cohen et al. this volume). On a slightly larger scale, "spikes" have been found in the redshift distribution of several more extensive galaxy surveys out to z 1 (B90; LF94), indicative of possibly some sort of larger structure at earlier epochs. What is questionable, according to CDM models, is whether or not any such structure could have existed at much higher redshifts than z ~ 1. RHS97 are able to reproduce the observational characteristics of QSO absorption systems at z ~ 3 over a wide range of column densities with a hydrodynamical model consisting of subgalactic clumps embedded in "sheet"-like structures and often lying along "ribbons" or "filaments" (see also Ostriker, this volume). Their simulations indicate that their ~ 20 clumps will merge to produce three L* galaxies by z = 0, which agrees remarkably well with the independent observations of P96b, who conclude that their 18 z 2.4 candidates could produce a few L* galaxies by z = 0 (Section 3). Is it possible that groups, clusters, or even some larger-scale structure existed to some extent at earlier epochs, although at much lower amplitude than that seen at lower redshifts? The random deep Cycle 6 WFPC2 parallel fields in the F450W and F410M filters presented by P97 are a first attempt to address this question (Section 3.4).