Introduction

The centrifugal acceleration of a friction-free ring on a rotating pole is equal to the product of the velocity component of the ring perpendicular to the pole and the angular rotation rate. Similarly, the centrifugal acceleration along the magnetic field of the guiding center of an ion spiraling on a ''rotating'' field line is equal to the dot product of the $\vec{E}\times\vec{B}$ drift of the ion (motion of field line) and the rotation rate of the field-line, ${\rm {d}}\hat e/{\rm {d}}t$, along the ion motion ( $\hat e = \vec{B}/B$). The importance of this effect on the transport of ions from the dayside cusp region to the magnetotail was first demonstrated by Cladis [1986]. Field lines in the cusp region are populated by ions that flow upward from the ionosphere and solar-wind ions that are injected from the magnetosheath. When these field lines are convected generally tailward, the up-ward flowing ions trace a region in the magnetosphere called the plasma mantle. Since both the $\vec{E}\times\vec{B}$ drift and the centrifugal acceleration are independent of the ion charge and mass, all ions with the same magnetic moment and parallel velocity component, initially, will follow the same path in the magnetosphere and will be accelerated by the same amount. Typically, the parallel velocity components of ionospheric ions that reach the center plane of the magnetotail, at distances greater than about 12 ${\rm {R}}_{\rm {E}}$ from the earth, become increased by factors of 10 to 100 by this process, depending on the magnitude of the convection electric field and the configuration of the magnetosphere. Cladis and Francis [1992] have shown by computer simulation that, during times of enhanced magnetospheric convection, the pressure of these cusp-ionospheric O⁺ ions injected in the magnetotail ( $-10 > X_{\rm {GSM}} > -15 {\rm {R}}_{\rm {E}}$) is comparable to the highest O⁺ pressures measured in that region during disturbed times [e.g. Kistler et al., 1992]. Such plasma injections in the near-earth tail, coupled with the fact that the transport time (1 to 2 hours) of these ions is comparable to the delay time of substorms relative to the onset of enhanced magnetospheric convection, suggests they may trigger substorms [Cladis and Francis, 1992]. This important acceleration term has since been incorporated in various types of transport codes and used for numerous studies of ion motion in the magnetosphere [e.g., Horwitz, 1987; Cladis, 1988; Delcourt et al., 1990; Swift, 1990; Delcourt and Sauvaud, 1992].

Although ion measurements in the magnetosphere have not been in disagreement with the effects of this acceleration process, pertinent data were insufficient to fully verify the process. During the event described here, (i) measurements were made of the time-varying interplanetary magnetic field (IMF) and solar-wind pressure that compressed the magnetosphere; (ii) the Polar satellite was fortuitously located in the region of the magnetosphere where the rotational motion of the magnetic field was high, and (iii) the satellite instrumentation provided excellent data on the local magnetic field components, the perpendicular drift of the ions, and the parallel velocity of the ions that were flung along the magnetic field. In this report we describe these data and a computer simulation that relates the data to the centrifugal acceleration process.