The Caltech Faint Galaxy Redshift Survey XII: Clustering of Galaxies

A clustering analysis is performed on two samples of $\sim 600$ faint galaxies each, in two widely separated regions of the sky, including the Hubble Deep Field. One of the survey regions is configured so that some galaxy pairs span angular separations of up to 1 deg. The median redshift is $z_{med}\approx 0.55$. Strong clustering is obvious, with every pencil-beam field containing a handful of narrow redshift-space features, corresponding to galaxy structures with sizes of 5 to 20 Mpc. The structures are not obviously organized on planes, though one prominent, colinear triplet of structures is observed, spanning $\sim 20$ Mpc. This may be evidence of a filament. A galaxy--galaxy correlation function calculation is performed. No significant evolution of clustering (relative to stable clustering) is found in the redshift range 0.3<z<1.0. This is not surprising, since uncertainties in the correlation amplitude estimated from surveys like these are large; field-to-field variations and covariances between data points are both shown to be significant. Consistent with other studies in this redshift range, the galaxy--galaxy correlation length is found to be somewhat smaller than that predicted from local measurements and an assumption of no evolution. Galaxies with absorption-line-dominated spectra show much stronger clustering at distances of <2 Mpc than typical field galaxies. There is some evidence for weaker clustering at intermediate redshift than at low redshift, when the results presented here are compared with surveys of the local Universe. In subsets of the data, the measured pairwise velocity dispersion of galaxies ranges from 200 to $600 km s^{-1}$, depending on the properties of the dominant redshift structures in each subset.