|Annu. Rev. Astron. Astrophys. 1992. 30:
Copyright © 1992 by Annual Reviews. All rights reserved
Another potentially powerful and independent probe of evolution and cosmology is the angular clustering of faint galaxies, since such measures are sensitive to the redshift distribution of the sample and the cosmological model (Peebles 1980). Koo and Szalay (1984), Stevenson et al. (1985), Pritchet and Infante (1986), and Jones et al. (1988) have measured the two-point angular correlation function to the limit of photographic counts (B = 23 to 24), where the scaled correlation amplitude appears to be the same as found locally. More recently, CCD data have extended the results to B > 25 (Neuschaefer et al. 1991, Efstathiou et al. 1991). Despite the nominal simplicity of the experiment, the various measures of the correlation function at low amplitudes are consistent only within a factor of 2 to 3, and may reflect not only variations in photometric zero points, but also the exact method for calculating the clustering and for correcting the background.
Efstathiou et al. (1991) found a low clustering amplitude at 30-arcsec separations, a result that was especially apparent for the bluer galaxies. If there had been extensive galaxy-galaxy interactions at visible redshifts, then the clustering amplitude at short scales might be expected to show an enhancement, rather than the decline seen in the data. Efstathiou et al. (1991) prefer a picture in which weakly clustered blue galaxies are apparent at high redshift, but which are not observed locally. These and other interpretations of the correlation amplitude in faint data assume that galaxies of differing types and luminosities share the same intrinsic clustering properties. In fact late-type spirals are found typically to have about a factor of 4 lower clustering amplitude than early-type galaxies (Davis and Geller 1976; Giovanelli, Haynes, and Chincarini 1986). This dependence should be taken into account, since fainter samples of galaxies are naturally expected to be weighted more towards bluer and lower-luminosity galaxies.
It is preferable to measure spatial correlation with complete redshift samples. A pioneering attempt with faint galaxies was made by Loh (1988), who estimated redshifts from multicolor photometry for a sample of 1000 galaxies brighter than I = 22 and explored the two-point spatial correlation function to z = 0.75. He found no evidence for a change in power-law slope nor convincing changes in the clustering amplitude. We anticipate that the study of clustering evolution will gain prominence as larger samples of spectroscopic redshifts for faint field galaxies are acquired.