More than 200 hours of ionospheric electric field measurements taken on balloons flown from three polar cap sites have been analyzed to determine average properties of the large‐scale polar cap electric field and its dependencies on the interplanetary magnetic field. The major component of this electric field was directed from dawn to dusk and produced an average polar cap potential drop of about 55 kV. The magnitude of this potential provides an upper limit of about 700 RE for the length of the magnetospheric tail and implies an energy input from the solar wind to the magnetosphere of about 5 × 1019 ergs/sec. The dawn to dusk component of the high‐latitude polar cap electric field responds to Bz, the northward component of the interplanetary magnetic field, on a time scale ≲1 hour and with an average increase of about 3 mV/m for each 1 γ decrease of Bz. The hourly averages of the electric field data at each of the three sites are well described by a two‐cell convection pattern whose location depends on the y component of the interplanetary magnetic field. When By is positive (negative), the two‐cell convection pattern shifts toward dawn (dusk) in the northern hemisphere with the following consequences: the maximum intensity of the northern polar cap dawn to dusk electric field component occurs at local morning (evening), and the auroral zone return flow reaches higher latitudes in the evening (morning). Evidence of the vector nature of the interaction between interplanetary and terrestrial magnetic fields is provided by the observation that the above By dependent signatures are most evident when Bz is most negative.

More than 100 cases of major decreases of proton and electron fluxes at the synchronous orbit have been observed and interpreted as a movement of the energetic particle trapping boundary earthward of 6.6 R r before the substorm expansion phase. These events are observed only between 1700 and 0800 LT and are consistent with the existence of a westward magnetospheric electric field of a few tenths of a millivolt per meter before substorm expansion. Most substorm particle events seen on the nightside do not exhibit this behavior, presumably because the trapping boundary moves inside 6.6 /i•B only during major events. Such events show evidence of particle acceleration, which probably was concentrated at higher L values, near the instantaneous locationof the trapping boundary. Many authors have published results showing an inward motion of many regions of the magnetosphere during substorm disturbances. These results have come from satellite measurements [Vasyliunas, 1968; Aubry et al., 1970; Mc-Pherron, 1970; Lezniak and Winckler, 1970; Meng, 1970] and from ground and balloon observations [Carpenter and Stone, 1967; Mozer, 1971]. Unfortunately, the functions and even the timings of these inward motions in the total substorm development are not yet clear. Equally unclear is the relationship of energetic (tens of kiloelectron volts) particle acceleration to other substorm phenomena. Synchronous orbit measurements of such particles have been made by Lezniak et al. [1968], Parks e't al. [1968], and Lezniak and Winckler [1970]. Magnetic field measurements on the same synchronous satellite [Cummings et al., 1968] have demonstrated that the nightside magnetic field frequently becomes more taillike before substorm particle injection events, as seen at the synchronous orbit. Some theoretical work on the motion of the tail sheet current has been pro-vided by Coroniti and Kennel [1972]. This paper will describe a large sample of observations of nightside energetic particles by ATS-5; such measurements on a synchronous Copyright (•) 1973 by the American Geophysical Union. satellite relieve some of the usual temporal and spatial distortions of elliptical orbits and provide information on. the acceleration region for auroral particles.

Pitch‐angle distributions have been obtained for protons and electrons of energy ∼30–300 kev on the geostationary satellite ATS 5. Those obtained near and in the loss cone in December 1969 and January 1970 provide evidence for the existence of a depleted loss cone except during two active intervals when a ‘filled‐in’ loss cone was approached. The slopes of the proton and electron angular distributions outside the loss cone at small pitch angles (10° to 25°) remained unchanged as the loss cone filled in or became depleted, and the flux of particles in the loss cone at any energy did not necessarily depend on the intensity of particles outside of the loss cone at that energy. Data are also examined for 8 days in August and September 1969 when the detector sampled a wide range of pitch angles (usually 25°–90°). The diurnal variations of these angular distributions are similar for protons and electrons and are in qualitative agreement with the predictions of a theory of drift‐shell splitting, although the short lifetimes of the measured particles prevent the good quantitative agreement found by other experimenters at higher energies. Several substorm events are studied, with the result that both protons and electrons are introduced preferentially at large pitch angles near midnight and are introduced more uniformly with angle at other local times.

Energetic particle observations from ATS‐5 are presented that directly show asymmetric injection of 30‐ to 300‐kev proton fluxes into the nightside magnetosphere to form the asymmetric ring current. The temporal structure of these substorm‐correlated proton bursts is consistent with a westward gradient drift from the nightside acceleration region. The drifting protons tend to be accompanied by nondiamagnetic decreases of the magnetic field at the satellite. Adiabatic decreases of the energetic electron fluxes often occur with these changes in the field strength. Nearly 200 proton‐only events have been observed, occurring uniformly between 1500 and 2200 LT and not at all before 1200 LT or after 2300 LT.

Energetic protons and electrons introduced into the vicinity of the synchronous orbit by geomagnetic substorms and observed on ATS 5 at various local times have been analyzed for velocity dispersion effects associated with their longitudinal drift. Particles with energies between about 30 and 200 keV are shown to be produced simultaneously within tens of minutes at local times between about 0000 and 0400 near 6.6 RE during substorm activity. Those with energies ≳75 keV move in longitude in the direction and with the magnitude expected from gradient B drifts. Lower‐energy protons and electrons appear at the satellite sooner than expected from their gradient B drifts, as though the observed particles of such energies were not those originally accelerated but were newly produced from or by the higher‐energy drifting component.