Assuming that curvature perturbations and gravitational waves originally arise from vacuum fluctuations in a matter-dominated phase of contraction, we study the dynamics of the cosmological perturbations evolving through a nonsingular bouncing phase described by a generic single scalar field Lagrangian minimally coupled to Einstein gravity. In order for such a model to be consistent with the current upper limits on the tensor-to-scalar ratio, there must be an enhancement of the curvature fluctuations during the bounce phase. We show that, while it remains possible to enlarge the amplitude of curvature perturbations due to the nontrivial background evolution, this growth is very limited because of the conservation of curvature perturbations on super-Hubble scales. We further perform a general analysis of the evolution of primordial non-Gaussianities through the bounce phase. By studying the general form of the bispectrum we show that the non-Gaussianity parameter $f_{\mathrm{NL}}$ (which is of order unity before the bounce phase) is enhanced during the bounce phase if the curvature fluctuations grow. Hence, in such nonsingular bounce models with matter given by a single scalar field, there appears to be a tension between obtaining a small enough tensor-to-scalar ratio and not obtaining a value of $f_{\mathrm{NL}}$ in excess of the current upper bounds. This conclusion may be considered as a "no-go" theorem that rules out any single field matter bounce cosmology starting with vacuum initial conditions for the fluctuations.

The origin of the seeds which develop into the observed super-massive black holes at high redshifts may be hard to interpret in the context of the standard $\Lambda CDM$ of early universe cosmology based on Gaussian primordial perturbations. Here we consider the modification of the halo mass function obtained by introducing skewness and kurtosis of the primordial fluctuations. We show that such primordial non-Gaussianities constrained by the current observational bounds on the nonlinearity parameters of $f_{NL}$ and $g_{NL}$ are not effective at greatly increasing the number density of seeds which could develop into super-massive black holes at high redshifts. This is to be contrasted with the role which cosmic string loops could play in seeding super-massive black holes.