Conventional theories of nuclear interactions predict a net increase in the distribution of virtual pions in nuclei relative to free nucleons. Analysis of data from several nuclear experiments has led to claims of evidence against such a pion excess. These conclusions are usually based on a collective theory (RPA) of the pions, which may be inadequate. The issue is the energy dependence of the nuclear response, which differs for theories with strong NN correlations from the RPA predictions. In the present paper, information about the energy dependence is extracted from sum rules, which are calculated for such a correlated, noncollective nuclear theory. The results lead to much reduced sensitivity of nuclear reactions to the correlations that are responsible for the pion excess. The primary example is ( p, n) spin transfer, for which the expected effects are found to be smaller than the experimental uncertainties. The analysis has consequences for Deep Inelastic Scattering (DIS) experiments as well.