The ''planetary wind'' model, which holds that corotation must break down outside some Alfv¿n critical radius and that a centrifugally driven wind outflow must then develop, does not agree with Voyager observations at Jupiter. The discrepancies are most evident in the distant compositional signatures observed by Voyager that showed the plasma to be mostly of indigenous (i.e., Io) composition to ~60 RJ, solar wind-like to ~150 RJ, and again mostly indigenous at >150 RJ. Throughout most of the Jovian magnetodisk plasma sheet, the corotational energy density exceeds the magnetic energy density, but corotation-dominated flow is maintained. Furthermore, the composition of the energetic particles observed by Voyager changed significantly when the spacecraft left the rotating plasma sheet and entered the outflowing magnetospheric wind, contrary to the prediction of the planetary wind model. A new global convection model for the Jovian magnetosphere is proposed to explain these observations, drawing upon models of quasi-stationary plasma convection in Earth's magnetosphere. The model predicts a substantial dawn--dusk asymmetry in the structure, dynamics, and plasma composition of the magnetopause and magnetosheath, as well as a region of cross-tail flow (dusk-to-dawn) in the nightside plasma sheet extending from ~60--150 RJ and containing a substantial admixture of solar wind plasma. ¿ American Geophysical Union 1989