On the universality of the two-point galaxy correlation function

We examine the behavior of the two-point galaxy correlation function ξ(r) in volume-limited subsamples of three complete redshift surveys: the CfA survey of northern galaxies with m_B_ < 14.5, the diameter-limited Southern Sky survey, and a full-sky 60 micron flux-limited survey of galaxies detected by the IRAS satellite. We use the subsamples to investigate the possibility that the correlation amplitude depends on the absolute luminosity of the galaxies. The correlation length r_0_ is found to scale approximately as the square root of the distance limit in both the CfA and Southern Sky catalogs, but is independent of the distance limit in the IRAS sample. We show that this effect is caused by a combination of factors: large positive density fluctuations in the foreground of the optically selected catalogs bias the estimate of r_0_ downward, while the brightest galaxies (M < -20 + 5 log h) do appear to be more strongly clustered than the mean. This result demonstrates that weighting the pair counts by the reciprocal of the selection function, the usual practice with redshift catalogs, can lead to substantial overestimation of ξ(r) for L_*_ galaxies. We suggest r_0_~5h^-1^ Mpc as a best estimate for the correlation length of optically selected L_*_ galaxies.