We have presented examples of waveform data obtained in the high altitude cusp and at crossings of the plasma sheet boundary, where large amplitude solitary waves are commonly observed. The high altitude solitary waves have high propagation speeds of ~1-2 x 103 km/s, and are positive potential structures (i.e. electron holes or ion enhancements). This is very different from the type of solitary waves first observed in the auroral zone at low altitudes [Temerin et al., 1982, Bostrom et al., 1988] which are negative potential structures (identified as ion holes) travelling at 100s of km/s [Mozer et al., 1997; Bounds et al., 1998]. The structures described herein are more similar to the recently identified fast solitary waves in the low altitude auroral zone [Mozer et al., 1997; Ergun et al., 1998; Bounds et al., 1998] which have been classified as electron holes traveling at velocities up to ~5000 km/s. The lack of an AC magnetic signature in association with the solitary waves described herein is not inconsistent with this possibility since the expected amplitudes are generally below the instrument noise level.
Ion acoustic solitary waves, which have been extensively studied theoretically for low altitude auroral zone parameters [Lotko and Kennel, 1983; Hudson et al., 1983; Marchenko and Hudson, 1995; and references therein], are associated with ion beams. Simulations with both a cold background and an ion beam have shown that the solitary waves develop as ion holes in the background plasma when the ion beam is weak. The expected propagation speed, Vsw, is given by Vsw=Vb +/- cs in the frame of the background plasma, where Vb is the beam speed, consistent with analytic studies [Lotko and Kennel, 1983]. For the case of a strong beam, the resulting solitary waves are ion enhancements and propagate at Vsw=2cs in the background frame. Electron mode solitary waves, electron acoustic solitons [Dubouloz et al., 1991; Mace et al., 1991] and electron holes [Ergun et al, 1998], move at higher velocities, greater than the electron acoustic speed, and comparable to the electron beam speed, respectively. They require two electron components. The two possible explanations for the observed high altitude solitary waves, ion and electron modes, will be examined in more detail below to determine which is most consistent with the observations.
Although simulations of ion solitary waves assumed the existence of a cold background plasma, Hudson  suggested that the results can be applied to other regimes. For example, in the low altitude auroral zone ion beam regions where FAST data are consistent with no cold plasma [Strangeway et al., 1998], the oxygen beam could act as the cold background plasma, and the hydrogen beam as the beam. Hot plasma sheet electrons are the third component and determine the sound speed. Using these particle populations and an ~ 1 keV beam, the solitary waves would propagate in the satellite frame at ~few hundred km/s upward. This is consistent with the low altitude auroral zone observations of negative potential (ion hole) structures. To ascertain whether the observed characteristics of the high altitude solitary waves are consistent with an ion mode, we extend this argument to the high altitude regions to determine if a plausible explanation for the observed propagation speeds can be made. In cusp injections, H+ and He++ downflowing ions and cool magnetosheath electrons are the main particle populations. The He++ ions could act as the background population for the growth of ion holes. The sound speed is low, ~40 km/s, while the injected ion speed is ~500 km/s. The observed solitary waves are positive potential structures which are propagating downward. This is consistent with the conclusion of Lotko and Kennel  that compressive (i.e. positive potential) ion acoustic solitons can grow when Vb > ~10 cs, and with the simulations results [Marchenko and Hudson, 1995] which showed the compressive mode growing for strong beams (beam density comparable to background ion density). The structures propagate in the direction of the ion beam at approximately twice the sound speed in the background frame. For the observed parameters, the predicted speeds are ~600 km/s which is comparable to, but somewhat less than, the observed speeds.
At the high altitude plasma sheet crossings, the hot, plasma sheet ion population may act as the background plasma. The sound speed is ~500 km/s. If the positive potential mode could grow, its propagation speed would be ~1 x 103 km/s, comparable to the observed velocities of ~ 1-2 x 103 km/s. However, the Lotko and Kennel  study predicts that positive potential modes could not grow for the observed plasma parameters. The Marchenko and Hudson simulations suggest a very dense ion beam would be needed, and such beams are not always observed. In the 3/28 event, there was no ion beam. Although a dense beam was observed in the 3/10 event, it was an upward beam which should result is upward propagation for ion mode solitary waves. The observed waves, however, were travelling downward.
Although the above comparisons suggest that the structure and speed of the cusp solitary waves might be explainable by an ion mode, the plasma sheet boundary solitary wave characteristics are not consistent with an ion mode. An additional problem is that simulations [Barnes et al., 1985] have suggested that the growth of ion solitary waves requires that the plasma be strongly magnetized (fce/ fpe >>1) which is not the case for the observed high altitude events. For the plasma sheet boundary crossings, fce/ fpe=~2; for the cusp case, fce/ fpe<1. It is likely, therefore, that the high altitude solitary waves are an electron mode, either an electron hole or an electron acoustic solitary wave, rather than an ion mode. The observed velocities are consistent with an electron mode, as is the existence of 2 electron components. Electron acoustic solitary waves can be positive potential structures and electron holes always are; neither require that the plasma be strongly magnetized. Further studies of the solitary waves and associated plasma distributions, as well as more detailed theoretical studies, are needed to determine if the structures are electron acoustic or electron hole mode, to determine whether ion mode solitary waves ever occur at high altitudes, and to determine whether the mechanisms are the same as for the low altitude electron holes or for those observed in the distant plasma sheet [Matsumoto et al., 1994].