In this paper I examine cosmological models that contain a stochastic background of nonlinear electromagnetic radiation. I show that for Born-Infeld electrodynamics the equation of state parameter, $w=P/\rho$, remains close to 1/3 throughout the evolution of the universe if $E^2=B^2$ in the late universe to a high degree of accuracy. Theories with electromagnetic Lagrangians of the form $L=-{1/4}F^2+\alpha F^4$ have recently been studied in magnetic universes, where the electric field vanishes. It was shown that the $F^4$ term can produce a bounce in the early universe, avoiding an initial singularity. Here I show that the inclusion of an electric field, with $E^2\simeq B^2$ in the late universe, eliminates the bounce and the universe "begins" in an initial singularity. I also examine theories with Lagrangians of the form $L=-{1/4}F^2-\mu^8/F^2$, which have been shown to produce a period of late time accelerated expansion in magnetic universes. I show that, if an electric field is introduced, the accelerated phase will only occur if $E^2<3B^2$.

It is well known that general relativity does not admit gravitational geons that are stationary, asymptotically flat, singularity free and topologically trivial. However, it is likely that general relativity will receive corrections at large curvatures and the modified field equations may admit solutions corresponding to this type of geon. If geons are produced in the early universe and survive until today they could account for some of the dark matter that has been "observed" in galaxies and galactic clusters. In this paper I consider gravitational geons in 1+1 dimensional theories of gravity. I show that the Jackiw-Teitelboim theory with corrections proportional to $R^2$ and $\Box R$ admits gravitational geons. I also show that gravitational geons exist in a class of theories that includes Lagrangians proportional to $R^{2/3}$.

The field equations associated with the Born-Infeld-Einstein action including matter are derived using a Palatini variational principle. Scalar, electromagnetic, and Dirac fields are considered. It is shown that an action can be chosen for the scalar field that produces field equations identical to the usual Einstein field equations minimally coupled to a scalar field. In the electromagnetic and Dirac cases the field equations reproduce the standard equations only to lowest order. The spherically symmetric electrovac equations are studied in detail. It is shown that the resulting Einstein equations correspond to gravity coupled to a modified Born-Infeld theory. It is also shown that point charges are not allowed. All particles must have a finite size. Mass terms for the fields are also considered.

In recent papers (astro-ph/0306630, gr-qc/0312041) I have argued that the observed cosmological acceleration can be accounted for by the inclusion of a 1/R term in the gravitational action in the Palatini formalism. Subsequently, Flanagan (astro-ph/0308111, gr-qc/0403063) argued that this theory is equivalent to a scalar-tensor theory which produces corrections to the standard model that are ruled out experimentally. In this article I examine the Dirac field coupled to 1/R gravity. The Dirac action contains the connection which was taken to be the Christoffel symbol, not an independent quantity, in the papers by Flanagan. Since the metric and connection are taken to be independent in the Palatini approach it is natural to allow the connection that appears in the Dirac action to be an independent quantity. This is the approach that is taken in this paper. The resulting theory is very different and much more complicated than the one discussed in Flanagan's papers.

Recently Flanagan [astro-ph/0308111] has argued that the Palatini form of 1/R gravity is ruled out by experiments such as electron-electron scattering. His argument involves adding minimally coupled fermions in the Jordan frame and transforming to the Einstein frame. This produces additional terms that are ruled out experimentally. Here I argue that this conclusion is false. It is well known that conformally related theories are mathematically equivalent but not physically equivalent. As discussed by Magnano and Sokolowski [2] one must decide, in the vacuum theory, which frame is the physical frame and add the minimally coupled Lagrangian in this frame. If this procedure is followed the resulting theory is not ruled out experimentally. The discussions in this paper also show that the equivalence between the generalized gravitational theories and scalar tensor theories discussed by Flanagan [gr-qc/0309015] is only mathematical, not physical.

Corrections to Einstein's equations that become important at small curvatures are considered. The field equations are derived using a Palatini variation in which the connection and metric are varied independently. In contrast to the Einstein-Hilbert variation, which yields fourth order equations, the Palatini approach produces second order equations in the metric. The Lagrangian $L(R)=R-\alpha^2/R$ is examined and it is shown that it leads to equations whose solutions approach a de Sitter universe at late times. Thus, the inclusion of 1/R curvature terms in the gravitational action offers an alternative explanation for the cosmological acceleration.

One of the most intriguing unsolved problems in modern cosmology involves the nature of the observed dark matter in the universe. Various hypothetical particles have been postulated as the source of the dark matter, but as of yet none of these particles have been observed. In this paper I present an alternative view. The observations indicate the presence of a new force that is much weaker than gravity on small scales and stronger than gravity on large scales. This new force is described by a Liouville field that consists of a scalar field with an exponential potential. The solutions have a Newtonian behavior on small scales and produce forces that are weak compared to gravity. On large scales the field has a logarithmic behavior and will therefore produce flat rotation curves. The Liouville field may therefore be able to account for the observations that are usually taken to imply the existence of dark matter.

Following renewed interest, the problem of whether the cosmological expansion affects the dynamics of local systems is reconsidered. The cosmological correction to the equations of motion in the locally inertial Fermi normal frame (the relevant frame for astronomical observations) is computed. The evolution equations for the cosmological perturbation of the two--body problem are solved in this frame. The effect on the orbit is insignificant as are the effects on the galactic and galactic--cluster scales.