Large scale clustering from Non-Gaussian texture models

A basic ingredient to understand large scale fluctuations in our present day Universe is the initial conditions. Using N-body simulations, we compare the clustering that arises from Gaussian and non-Gaussian initial conditions. The latter is motivated by a global texture model which has initial $J$-order correlations $\xibar_J$ close to the strongly non-Gaussian regime $\xibar_J \simeq \xibar_2^{J/2}$. The final amplitudes $S_J \equiv \xibar_J/\xibar_2^{J-1}$ in the non-Gaussian (texture) model evolves slowly towards the (Gaussian) gravitational predictions but, even at $\sigma_8=1$, are still significantly larger, showing a characteristic minimum with a sharp increase in $S_J$ with increasing scales. This minimum, which is between $10$ and $15 \Mpc$, depending on the normalization, separates the regime where gravity starts dominating the evolution from the one in which the initial conditions are the dominant effect. In comparing this results with galaxy clustering observations, one has to take into account {\it biasing}, i.e. how galaxy fluctuations trace matter fluctuations. Although biasing could change the amplitudes, we show that the possible distortions to the shape of $S_J$ are typically small. In contrast to the non-Gaussian (texture) predictions, we find no significant minimum or rise in $S_J$ obtained from the APM Galaxy Survey.