Statistical analysis of galaxy surveys - III: The non-linear clustering of red and blue galaxies in the 2dFGRS

We present measurements of the higher-order clustering of red and blue galaxies as a function of scale and luminosity made from the two-degree field galaxy redshift survey (2dFGRS). We use a counts-in-cells analysis to estimate the volume averaged correlation functions, xi_p, as a function of scale up to order p=5, and also the reduced void probability function. Hierarchical amplitudes are constructed using the estimates of the correlation functions: S_p=(xi_p/xi_2)^(p-1). We find that: 1) Red galaxies display stronger clustering than blue galaxies at all orders measured. 2) Red galaxies show values of S_p that are strongly dependent on luminosity, whereas blue galaxies show no segregation in S_p within the errors; this is remarkable given the segregation in the variance. 3) The linear relative bias shows the opposite trend to the hierarchical amplitudes, with little segregation for the red sequence and some segregation for the blue. 4) Faint red galaxies deviate significantly from the "universal" negative binomial reduced void probabilities followed by all other galaxy populations. Our results show that the characteristic colour of a galaxy population reveals a unique signature in its spatial distribution. Such signatures will hopefully further elucidate the physics responsible for shaping the cosmological evolution of galaxies.