Theoretical considerations of the relation between the microstructure of polymer networks and their elastic properties are followed by a description of an experimental program in which stress-strain data for vulcanized silicone rubber have been obtained and analyzed. It is demonstrated that suppression of junction fluctuations cannot solely account for the discrepancy between experimental modulus values and the predictions of the phantom-network theory. The large-strain data obtained show the ability of Flory's strain energy function to correctly model tension-compression data over the range of crosslink densities covered by this work. Finally, the extreme importance of the careful analysis of the materials used, the reaction employed, and the resulting networks was demonstrated. The simplest available method for the verification of the network structure is by the determination of the sol fraction. The extraction of solubles in the case of highly crosslinked networks was found to be susceptible to weighing uncertainty and the presence of unreactive material. The former can be avoided by the use of larger samples, while the latter could be removed by vacuum stripping for our material.