Evidence that high-temperature superconductivity originates in the charge-reservoir layers rather than in the cuprate-planes of the host materials is discussed. Hence PrBa2Cu3O7 superconducts when grown under conditions that minimize Ba-site Pr. The supercurrent in high-temperature superconductors is located primarily in the layers where the holes are and the magnetic moments are not. Gd2-ZCeZCuO4 fails to superconduct because Gd+3 has L=0 and does not have its Cooper-pair-breaking limited by crystal-field splitting. The observed charges-transferred in these materials are compatible with charge-reservoir superconductivity, and not with cuprate-plane superconductivity. The observed ionic charge of Ce in Nd2-ZCeZCuO4 and its homologues is consistent with p-type superconductivity, doping by the p-type defect (Ce,interstitial-O) pair, and a size effect on the superconductivity. The chemical trends in T-c and the amount of Ni or Zn required to destroy superconductivity by pair-breaking clearly indicate that the primary superconducting condensate is in the charge-reservoir layers.