Modelling galaxy clustering at high redshift

We discuss the theoretical interpretation of observational data concerning the clustering of galaxies at high redshifts. Building on the theoretical machinery developed by Matarrese et al. (1997), we make detailed quantitative predictions of galaxy clustering statistics for a variety of cosmological models, taking into account differences in spatial geometry and initial fluctuation spectra and exploring the role of bias as a complicating factor in these calculations. We demonstrate that the usual description of evolution (in terms of the parameters $\epsilon$ and $r_0$) is not useful for realistic galaxy clustering models. We compare the detailed predictions of the variation of correlation functions with redshift against current observational data to constrain available models of structure formation. Theories that fit the present-day abundance of rich clusters are generally compatible with the observed redshift evolution of galaxy clustering if galaxies are no more than slightly biased at $z\sim 1$. We also discuss the interpretation of a concentration of Lyman-break galaxies found by Steidel et al. (1998), coming to the conclusion that such concentrations are not unexpected in `standard' models of structure formation.