More than 2 years of magnetospheric O+ data from the Plasma Composition Experiment on the ISEE 1 spacecraft are compared, in a statistical manner, with concurrent data on the interplanetary magnetic field (IMF) and the solar wind plasma in an attempt to clarify the geophysical responses to varying solar wind inputs, in particular to variations in the GSM Bz component of the IMF. It is found that the energy density of 0.1- to 16-keV O+ ions in the plasma sheet, inside of 23RE, is well correlated with the average solar wind energy flux during the preceding few hours, whether that flux is measured by its electromagnetic (Poynting) component P or by its far greater kinetic component K. Although P and K are well correlated with each other, the correlation of the O+ energy density is slightly better with K than with P during times of positive Bz, while the opposite holds for negative Bz. In either case the O+ energy density is more nearly proportional to K, and, given a typical value of K, there is at most a marginal (less than a factor of 2) increase in this density associated with a negative Bz. Except for this latter effect, which can perhaps be taken as evidence that a change in polarity of Bz from positive to negative may induce ''unloading'' of internal tail energy, there is no evidence in these O+ data, including data from the inner magnetosphere (L<10), that a negative, or southward, Bz is inherently favorable to the transfer of solar wind power across the magnetopause. These findings, together with the results of correlating the hourly AE index with P and K, suggest that the strong dependence of the AE on the Bz polarity is substantially exaggerated by the lack of AE magnetometers above 71¿ magnetic latitude. ¿ American Geophysical Union 1995