The clustering of the first galaxy haloes

We explore the clustering properties of high-redshift dark matter haloes, focusing on haloes massive enough to host early generations of stars or galaxies at redshift 10 and greater. Haloes are extracted from an array of dark matter simulations able to resolve down to the 'mini-halo' mass scale at redshifts as high as 30, thus encompassing the expected full mass range of haloes capable of hosting luminous objects and sources of reionization. Halo clustering on large-scales agrees with the Sheth, Mo & Tormen halo bias relation within all our simulations, greatly extending the regime where large-scale clustering is confirmed to be 'universal' at the 10-20 per cent level (which means, e. g., that 3 sigma haloes of cluster mass at z = 0 have the same large-scale bias with respect to the mass distribution as 3s haloes of galaxy mass at z = 10). However, on small-scales, the clustering of our massive haloes (greater than or similar to 10(9) h(-1) M-circle dot) at these high redshifts is stronger than expected from comparisons with small-scale halo clustering extrapolated from lower redshifts. This implies 'non-universality' in the scale-dependence of halo clustering, at least for the commonly used parametrizations of the scale-dependence of bias that we consider. We provide a fit for the scale-dependence of bias in our results. This study provides a basis for using extraordinarily high-redshift galaxies (redshift similar to 10) as a probe of cosmology and galaxy formation at its earliest stages. We show also that mass and halo kinematics are strongly affected by finite simulation volumes. This suggests the potential for adverse affects on gas dynamics in hydrodynamic simulations of limited volumes, such as is typical in simulations of the formation of the 'first stars', though further study is warranted.