Small-strain moduli reported in the literature for polydimethylsiloxane (PDMS) model networks have been critically re-examined, with account being taken of recent studies which have demonstrated the occurrence of significant side reactions in hydrosilylations. When the effects of such side reactions on network structure parameters are properly taken into account, the elasticity results can be well explained using the recent molecular theory of Flory and Erman. The revised calculations show a transition in the modulus from the affine to the phantom limit of deformation as the degree of chemical crosslinking increases. This is to be expected when the constraints on the fluctuation of junctions vanish (because of the decreased interspersion of chains), and such constraints are expected to vanish even in the small-strain region when the chains are sufficiently short. In lieu of carefully controlled reactions with well-defined stoichiometries, it appears that the procedure best suited for testing the various theories of rubber-like elasticity is a plot of the modulus G (as approximated by the sum 2C1 + 2C2 of the Mooney-Rivlin constants) against the phantom modulus [f*]ph (as approximated by 2C1). Also of importance is the difference between the 'effective' number of network chains nu (relevant to theory) and the commonly used 'active' number of chains nu(a).