We examine the effects of $f(R)$ gravity on Jeans analysis of collapsing dust clouds. We provide a method for testing modified gravity models by their effects on star formation as the presence of $f(R)$ gravity is found to modify the limit for collapse. In this analysis we add perturbations to a de Sitter background. Depending on the characteristics of a chosen $f(R)$ model, the appearance of new limits is possible. The physicality of these limits is further examined. We find the asymptotic Jeans masses for $f(R)$ theories compared to standard Jeans mass. Through this ratio, the effects of the $f(R)$ modified Jeans mass for viable theories are examined in molecular clouds. Bok globules have a mass range comparable to Jeans masses in question and are therefore used for comparing different $f(R)$ models. Viable theories are found to assist in star formation.

The recent LIGO observation sparked interest in the field of gravity wave signals. Besides the gravity wave observation the LIGO collaboration used the inspiraling black hole pair to constrain the graviton mass. Unlike general relativity, $f(R)$ theories have a characteristic non-zero mass graviton. We apply the constraint on the graviton mass to viable $f(R)$ models to find the effects on model parameters. We find it possible to constrain the parameter space with the gravity wave based observations. We make a case study for the popular Hu-Sawicki model and find a parameter bracket. The result generalizes to other $f(R)$ theories and can be used to contain the parameter space.

The evolution of the scale parameter in the Hu-Sawicki model is examined. We search the parameter area for instabilities. It turns out the parameter area of physically meaningful evolution is non-existent. For greater stability a stabilising term such as R^2 should be added. With the addition, the area mainly coincides with phantom models. We also find that the evolution of the scale parameter can be examined through a linear expansion and not only in the Hu-Sawicki case. The oscillatory behaviour found in the evolution raises questions of the physicality of the model. The behaviour can also cause current observations of the Hubble parameter to be misleading. This paper has been withdrawn by the authords due to a critical error in equation (10)

Hamiltonian perturbation theory is used to analyse the stability of f(R) models. The Hamiltonian equations for the metric and its momentum conjugate are written for f(R) Lagrangian in the presence of perfect fluid matter. The perturbations examined are perpendicular to R. As perturbations are added to the metric and momentum conjugate to the induced metric instabilities are found, depending on the form of f(R). Thus the examination of these instabilities is a way to rule out certain f(R) models.

The affine connection in a space-time with a maximally symmetric spatial subspace is derived using the properties of maximally symmetric tensors. The number of degrees of freedom in metric-affine gravity is thereby considerably reduced while the theory allows spatio-temporal torsion and remains non-metric. The Ricci tensor and scalar are calculated in terms of the connection and the field equations derived for the Einstein-Hilbert as wells as for f(R) Lagrangians. By considering specific forms of f(R), we demonstrate that the resulting Friedmann equations in Palatini formalism without torsion and metric-affine formalism with maximal symmetry are in general different in the presence of matter.