LARGE-SCALE CLUSTERING FROM NON-GAUSSIAN TEXTURE MODELS

A basic ingredient for understanding large-scale fluctuations in our p resent-day universe is the initial conditions. Using N-body simulation s, we compare the clustering that arises from Gaussian and non-Gaussia n initial conditions. The latter is motivated by a global texture mode l which has initial J-order correlations <(xi)over bar>(J) close to th e strongly non-Gaussian regime <(xi)over bar>(J) similar or equal to<( xi)over bar>(J-1)(2). The final amplitudes S-J equivalent to<(xi)over bar>(J)/<(xi)over bar>(J-1)(2) in the non-Gaussian (texture) model evo lve slowly toward the (Gaussian) gravitational predictions but, even a t sigma(g)=1, are still significantly larger, showing a characteristic minimum with a sharp increase in S-J with increasing scales. This min imum, which is between 10 and 15 h(-1) Mpc, depending on the normaliza tion, separates the regime where gravity starts dominating the evoluti on from the one in which the initial conditions are the dominant effec t. In comparing this result with galaxy clustering observations, one h as to take into account biasing, i.e., how galaxy fluctuations trace m atter fluctuations. Although biasing could change the amplitudes, we s how that the possible distortions to the shape of S-J are typically sm all. In contrast to the non-Gaussian (texture) predictions, we find no significant minimum or rise in S-J obtained from the APM Galaxy Surve y.