The effect of the clumping of matter on the form that Einstein's equations should take on various scales (e.g. those including stars, galaxies, and clusters of galaxies) will be explored. It is already clear from work done on the first two of these scales, that such parameters as the rate of expansion of the universe, as estimated from observations, will be affected. Before quantitative estimates of such effects can be made, it will be necessary to refine the procedures used for averaging quantities in the Einstein's equations over a discrete matter distribution. A Swiss-cheese model will be used for the galaxies, rather than the almost-point model used so far, and this can be extended to clusters of galaxies. Such clustering must obviously be taken into account. Filament-void structure may also be investigated eventually, using computer simulation to find an appropriate probability distribution. The other refinement needed is finding a more realistic stress-energy tensor for a galaxy, and afterwards for a cluster. The zero-pressure dust model used so far was sufficient to establish the general nature of the results, but not to allow evaluation of constants in the expression for the rate of expansion of the universe.
