Statistical surveys have always been of great value in ULF wave
studies, but only during the past five years have systematic surveys
of the large ULF wave data bases provided by the AMPTE and ISEE
satellites appeared. The studies of Pc 1-2 waves by Anderson
et al. [1992a,b] are mentioned above. AMPTE CCE observations of Pc
3-5 (1--50 mHz) waves from L 
5 to 9 were surveyed by
Anderson et al. [1990] and updated by Anderson [1994a].
Takahashi and Anderson [1992] surveyed the same frequencies in the
inner magnetosphere [
], and noted a number of as yet
unexplained phenomena, including local time structuring of
compressional and toroidal Pc 3 waves and a minimum in toroidal
intensity just outside the nominal plasmapause. Zhu and
Kivelson [1991, 1994] surveyed compressional pulsations using
magnetic field and particle data from ISEE 1 and 2, and concluded
that they were most likely to be generated by internal instabilities
in regions where plasma beta and field line curvature are large.
Cao et al. [1994] presented a comprehensive survey of waves
with all polarizations in the Pc 3, 4, and upper Pc 5 bands using
ISEE data. The wider radial and latitudinal coverage provided by
the ISEE orbits largely confirmed but also greatly extended the
AMPTE CCE results.
Although attention has been increasingly focused on external sources and the continuing question of their transmission into the magnetosphere, questions remain about the means by which field line resonant pulsations are established, both at high and at low latitudes. There is still little or no evidence of global cavity modes in the outer magnetosphere, and it remains difficult to reconcile satellite observations of a continuum of locally oscillating field lines with ground-based observations of fixed-frequency resonances [ Hughes, 1994]. An important realization is that a time-varying field-aligned sheet current of moderately large azimuthal extent will have the same magnetic field signature as a toroidal pulsation with low azimuthal wave number. Recent studies of nonlinear effects of Alfven (hydromagnetic) waves by Rankin et al. [1993] and of ULF modes associated with field-aligned currents by Elias et al. [1994] may possibly lead to new approaches to the establishment of fixed-frequency resonances by such mechanisms as Kelvin-Helmholtz instabilities at the site of large-amplitude waves and ionospheric coupling to the field-aligned currents associated with the waves. Zhu et al. [1994] modeled the development of multiple polar cap arcs, incorporating an ionosphere that dynamically responds to incoming Alfven waves and associated field-aligned currents. They found that the magnetosphere-ionosphere coupling effects coming from this dynamic response resulted in the formation of multiple regions of localized small-scale magnetospheric shear flow. To our knowledge such a dynamic model has not yet been applied to the field line resonance problem.
Other innovative studies include:
(1) Walthour et al. [1993, 1994] demonstrated a novel technique for analyzing remote measurements of two-dimensional perturbations on the magnetopause, such as those produced by flux transfer events or pressure perturbations.
(2) Morrison et al. [1994], in a study of high latitude quasi-periodic radio emissions with frequencies from 500 Hz to 4000 Hz (in the ELF and VLF frequency bands), pointed out the potential significance of the high-latitude field line minima, occurring near local noon on field lines in the outermost magnetosphere, as sites for generation and/or modulation of electron cyclotron waves. This strongly nondipolar region of the dayside magnetosphere, which has rarely been explored by spacecraft, should be similarly important for a variety of wave-particle interactions and specifically should be considered as a source of very high latitude dayside ion cyclotron waves.
(3) Song et al. [1994c] investigated the effects of magnetic helicity and the resulting dynamo process on the response of the magnetosphere to external pressure pulses and/or flux transfer events, and on the generation of field-aligned currents and transverse ULF waves. The Alfvenon concept introduced by Song and Lysak [1994] may aid in connecting the microphysics of reconnection to macroscopic theory and observations.
As 1994 ends we await new results from the SuperDARN (dual auroral radar network) radars and from the satellites in the International Solar-Terrestrial Program (ISTP), as well as from the several ground magnetometer networks newly installed or under development. It is hoped that the successful synergism recently shown between observation and theory in the study of magnetosheath waves and ground signatures of magnetic impulse events can be extended in the coming years to several of the other outstanding problems in the study of ULF waves.
The author is grateful to the many colleagues who aided in identifying papers to be included in this review, and to L. J. Cahill, Jr., W. J. Hughes, and the two referees for helpful comments on the manuscript. This work was supported by the National Science Foundation under grants ATM-9112108, DPP-8918689, and DPP-9217024.