3.1. Deep F410M surveys for faint Ly emitting candidates at z 2.4
Pascarelle et al. (1996b; P96b) report evidence in support of the 'bottom-up' hypothesis, in the form of the detection of 18 significant compact emitters in deep HST images in the medium band filter F410M - likely Ly at z 2.4 (see Section 3.2) - with spectroscopic confirmation thus far for eight of these subgalactic-sized clumps at z 2.391 ± 0.004 with the MMT and the KPNO 4 m. Deeper Keck spectroscopic follow-up of the faintest z 2.4 candidates is presented by Armus et al. (1997; A97). These may be time building blocks from which present-day galaxies were made. In this section, we provide further evidence that galaxies, or time small star-forming subsystems from which they formed, tended to exist in small groups or proto-structures in the early Universe.
The initial discovery of a group of z 2.4 candidates was made from ground-based photometry with a medium-band (150 Å-wide) filter centered at 4130 Å (Ly at z 2.39) in the field surrounding the weak radio galaxy 53W002 at z = 2.390 (W91), yielding two other candidates at the same redshift (P96a). Fortuitously, the existence of a nearly identical medium-band filter on HST (F410M, centered at 4100 Å or Ly at z 2.4) allowed the same observations to be conducted with WFPC2 in Cycle 5 at much higher sensitivity and spatial resolution than could be achieved from the ground. The HST is able to achieve nearly the full point-source sensitivity gain, which is lacking in ground-based data, because of the characteristically small sizes of the z 2.4 candidates (Figs. 1 & 5a). These z 2.4 candidates are labeled in the color Plate of P96b, which covers about 2.5 x 2.5 arcmin of the sky or ~ 0.7 x 0.7 Mpc at z 2.39.
To confirm these results, two long parallel observations were made with WFPC2 using the F410M, F450W, & F814W filters in Cycle 6 (out of three fields scheduled thus far). The first parallel field was at 21h-5°, and the second at 16h+82°, each consisting of seven orbits (about 4 x 2400s in F410M, 2 x 1500s in F450W, and ~ 800s in F814W for color information). Because of the more limited number of orbits, proper image averaging and CR-rejection is crucial, and was done with a custom-written IDL routine to deal specifically with a small number (n 4) of low-signal images (Co97). Photometry was done as in W91, utilizing user-input apertures and a sky box that is interpolated underneath the object aperture by fitting a sloped plane to the pixels unaffected by faint neighbors (P96b). Approximately 40 objects were detected simultaneously in the three filters in each parallel field, and 115 in the deeper 53W002 field. Each photometric "curve of growth" was individually examined, and apertures were adjusted to give the best possible total magnitudes, while maintaining the same size aperture in the three filters for each object. Compared to the interactive package of O96, sky-estimates were produced consistent within 0.07% and aperture magnitudes grown-to-total consistent within 0.05 mag (with an rms 0.22 mag for each algorithm).
We believe that the discovery of this z 2.39 group is not likely strongly biased by targeting the weak radio galaxy 53W002, because the field galaxy counts in this region are consistent with those in several other randomly selected fields (Ca95, D95b, O96), and the surface density of radio sources at this flux level (~ 5 mJy at 8.4 GHz) is high enough that one such source would be found in every few WFPC2 fields (W93). Despite its steep spectral index, 53W002 is ~ 30 x weaker than most of the luminous 3CR sources, which are known to cluster strongly, and therefore its presence is likely to have less of an effect on the surroundings.