Reconstruction of the local velocity field from the overdensity field and a gravitational acceleration that falls off from a point mass as r^-2 yields velocities in broad agreement with peculiar velocities measured with galaxy distance indicators. MO
NDian gravity does not. To quantify this, we introduce the velocity angular correlation function as a diagnostic of peculiar velocity field alignment and coherence as a function of scale. It is independent of the bias parameter of structure formation in the standard model of cosmology and the acceleration parameter of MOND. A modified gravity acceleration consistent with observed large scale structure would need to asymptote to zero at large distances more like r^-2, than r^-1.
We present a simple heuristic model to demonstrate how feedback related to the galaxy formation process can result in a scale-dependent bias of mass versus light, even on very large scales. The model invokes the idea that galaxies form initially in l
ocations determined by the local density field, but the subsequent formation of galaxies is also influenced by the presence of nearby galaxies that have already formed. The form of bias that results possesses some features that are usually described in terms of stochastic effects, but our model is entirely deterministic once the density field is specified. Features in the large-scale galaxy power spectrum (such as wiggles that might in an extreme case mimic the effect of baryons on the primordial transfer function) could, at least in principle, arise from spatial modulations of the galaxy formation process that arise naturally in our model. We also show how this fully deterministic model gives rise to apparently stochasticity in the galaxy distribution.
