By simply plotting the critical temperatures Tc for the major superconductors versus the distance d between a cuprate-plane Cu-site and the nearest oxygen site in a charge-reservoir, it becomes apparent that the superconductivity originates in the charge-reservoirs, not in the cuprate planes. By plotting uc versus d, where uc is the amount of impurity Ni or Zn on a Cu-site (presumably in a cuprate-plane) required to suppress Tc to zero, it is evident that pair-breaking requires exponentially more impurities as the cuprate-planes become further from the charge-reservoirs, indicating directly that the superconducting condensate is primarily in the charge- reservoirs, not in the cuprate-planes. The fact that uc(Ni) < uc(Zn) for Nd2-zCezCuO4 indicates that this material is a BCS-like polarization- pairing superconductor, not a spin-fluctuation paired superconductor. The fact that uc(Ni) approximately equals uc(Zn) for most other high-temperature superconductors indicates that the primary superconducting condensate is more distant than the range of the exchange interaction from the cuprate- plane Cu-sites where the impurities reside. The data for YBa2Cu4O8 and for La0.6Ca0.4Ba1.35La0.65Cu3Ox indicate that the behavior of Ni and Zn in YBa2Cu3Ox, uc(Ni) > uc(Zn), is an artifact of different solubilities on different sites, and not evidence of spin-fluctuation- pairing. The trends in the data suggest that there is only one mechanism of high-temperature superconductivity, that the mechanism involves some form of polarization-pairing, and that the superconductivity originates primarily in the charge-reservoirs, not in the cuprate-planes.