A brief review is presented to illustrate how the knowledge and experience gained from investigations of ionospheric topside-sounder ionograms has benefited scientific research beyond the ionosphere. In particular, to the interpretation of (1) natural radio emissions from space plasmas, (2) sounder-stimulated echoes and plasma emissions, and plasma diagnostics in planetary magnetospheres and (3) X-ray spectra of disks around neutron stars and black holes. The relevant data are from 60 satellite years of ionospheric topside-sounder operations by the four satellites of the International Satellites for Ionospheric Studies (ISIS) program from 1963 through 1989. Not all of these data were reduced to ionograms on 35-mm film. An ongoing effort at the NASA Goddard Space Flight Center is producing digital topside-sounder ionograms directly from a selected subset of the original telemetry tapes. More than 300,000 digital ionograms are now in the National Space Science Data Center at the NASA GSFC (http://nssdc.gsfc.nasa.gov/space/isis/isis-status.html).
 The Alouette 1 and 2 and ISIS-1 and -2 satellites were launched into polar orbits in 1962, 1965, 1969, and 1971 at 1000, 500–3000, 550–3500, and 1400 km altitudes, respectively. Alouette 1 and 2 each produced topside ionograms for 10 years; ISIS-1 and -2 for 21 and 19 years, respectively. The ISIS program, NASA's first international space program, was a spectacular success. Nearly 700 scientific papers based on the ISIS data, mainly dealing with ionospheric physics, were published through the end of 1985 [Jackson, 1986] and many more have been published since that date. Not all of the data were reduced to the ionogram format (virtual range of the intensity-modulated echo amplitude vs. sounder frequency) on 35-mm film. In 1996, a subset of the original seven-track analog telemetry tapes were shipped from Canada to the NASA Goddard Space Flight Center (GSFC) for analog-to-digital (A/D) conversion in order to produce digital topside-sounder ionograms. The subset consisted of 13,800 tapes selected from time intervals centered on the solstice and equinox dates from 24 telemetry stations chosen (with the assistance of H. G. James and D. Bilitza) so as to obtain global coverage over a solar cycle [Benson, 1996]. The main coverage was for the interval from 1972 to 1983. Another 4500 tapes were deposited in the National Archives in Canada; the bulk of the tapes (81,500) went to a Canadian landfill. The goal of the Goddard A/D effort was to preserve as many of the long-term ionospheric records as possible, within projected cost constraints, in a format suitable for digital retrieval and analysis techniques. Most of the selected ISIS-2 analog topside-sounder data have been reduced to digital topside-sounder ionograms (approximately 300,000) and have been archived in the National Space Science Data Center at the NASA GSFC and ISIS-1 digital ionograms are currently being produced. The data, and search and analysis software, are available from http://nssdc.gsfc.nasa.gov/space/isis/isis-status.html. The search page enables efficient retrieval of desired data. The analysis program allows the user to (1) view the ionograms in detail in either the ionogram format or the amplitude vs. time delay (at a selected frequency) format, (2) invert ionospheric reflection traces into vertical topside electron-density (Ne) profiles, (3) produce Ne contours along the satellite orbit from the satellite altitude down to the altitude of the ionospheric Ne peak and (4) scale and identify sounder-stimulated plasma resonances. These capabilities were used by Benson and Grebowsky  in a study involving ISIS-2 digital ionograms recorded over the polar cap, which suggested that the absence of an F layer Ne peak may be common at high latitudes. The ISIS-2 ionograms are currently being used as input to a program that automatically inverts the ionospheric reflection traces into vertical Ne profiles [Bilitza et al., 2004]. The resulting topside profiles are being used to improve the International Reference Ionosphere model, which is in dire need of this input to supplement its ground-based Ne profile database in order to improve total electron content model results. Such information is necessary for optimum performance of communication and navigational satellite systems.
 The importance of the ionosphere to such systems, and the magnetic storm effects on power grids and long-distance cable operations, is well known [Lanzerotti et al., 1998]. Our goal here is to emphasize the importance of also considering the ionosphere as the nearest space plasma that is available for cost-effective research. Thus the goal of this brief review is to illustrate that ionospheric research provides knowledge and experience pertaining to fundamental space-plasma processes that are relevant to our understanding of the terrestrial magnetosphere and planetary and astrophysical plasmas in addition to the terrestrial ionosphere. Our discussion will emphasize the plasma resonances stimulated by ionospheric topside sounders.
2. ISIS-2 Example and Resonance Interpretation
Figure 1 shows an ISIS-2 topside ionogram that displays the intensity-modulated amplitude of the various signal returns that are received during the 22-ms listening period following the transmission of a short (0.1 ms) sounder pulse as a function of the sounder frequency. Two scales are presented on each axis, i.e., the time delay of the echo (right scale) can be expressed in terms of apparent range (left scale) out to 3000 km (assuming propagation at the free-space velocity of light) and the swept frequency (from 0.1 to 10 MHz in this example) is a function of time after the start of the ionogram (from 3.3 to 14 s in this example) (bottom two scales). The first 3.3 s is dedicated to fixed-frequency sounding (at 1.0 MHz in this example). Most of the signal returns observed in Figure 1 are due to either short-range electrostatic (es) echoes or long-range electromagnetic (em) echoes. The former, called plasma resonances, are observed at
where fH is the electron cyclotron frequency, fN is the electron plasma frequency, fT is the upper hybrid frequency, e is the electron charge, me is the electron mass, εo is the permittivity of free space, ∣B∣ is the ambient magnetic field strength, and Ne is the ambient electron density (see, e.g., reviews by Muldrew [1972a] and Benson ). One member of a sequence of resonances known as Qn resonances is present in Figure 1. This sequence is observed at frequencies between the nfH harmonics and above fT given by the approximate expression
This equation was derived by Dougherty and Monaghan ; the Qn notation was introduced by Warren and Hagg , who showed that fQn corresponded to resonances stimulated by the Alouette 2 topside sounder. Equation (4) corresponds to zero group-velocity vg solutions of the es approximation to the hot-plasma dispersion relation with a Maxwellian electron velocity distribution for finite values of the wave number k = 2π/λ, where λ is the wavelength, when k is perpendicular to B, i.e., for the Bernstein modes. Muldrew [1972b] showed that the Qn resonances are due to sounder-stimulated es Bernstein-mode waves with group velocity vg nearly matched to the satellite velocity vs. A resonance is also observed at fH (i.e., n = 1 in (1) above) in Figure 1. The associated propagation conditions have not been determined for this fundamental resonance [Muldrew 1972a]. Prominent resonances at nfH with n up to 5 are observed in Figure 1; they are designated by the numerals 1–5 at the top of the ionogram.
 The two free-space long-range electromagnetic (em) echoes, one corresponding to the ordinary mode (O) and one to the extraordinary mode (X), extend out to the 10-MHz limit in Figure 1. Since these reflection traces extend beyond the observational high-frequency limit in this case (an extended mode allowed sounding to 20 MHz), it would not be possible to invert them into Ne profiles all the way down to the vicinity of the F layer peak. The cutoff frequencies at the satellite for these O- and X-mode traces are given by equation (2) and by
respectively. There is also a slow branch of the X mode, called the Z mode (Z), which is confined to the ionospheric plasma with a maximum possible frequency given by equation (3) and a cutoff frequency of
 There are also plasma resonances that correspond to incoherent plasma emissions stimulated by the sounder pulses (as opposed to coherent es echoes). An important class of these are observed at frequencies between the nfH harmonics and below fT and are know as Dn resonances [Nelms and Lockwood, 1967; Oya, 1970]. Osherovich [1987, 1989] and Osherovich and Benson  showed that they can be described by the following equations:
The D1 and D1+ resonances present in Figure 1 satisfy the hybrid relationship given by equation (8) just as fN and fT, also clearly observed in Figure 1, satisfy the hybrid relationship given by equation (3).
 Since the parameters defined in equations (1)–(9) depend only on fH and fN it is useful to present them in a normalized (by fH) form as a function fN/fH (Figure 2). In this representation, a self-consistent interpretation of the resonance and cutoff phenomena of equations (1)–(9) correspond to the vertical line appropriate for the fN/fH value for the corresponding ionogram. Thus accurate determinations of ambient plasma parameters can be made. In Figure 1, most of these phenomena are identified and there is an additional feature just prior to the 0.75-MHz frequency marker that approximately corresponds to fT/2. It is observed when fT/2 ≈ fN − fH [Hagg and Muldrew, 1970]. The normalized frequencies of the features observed in Figure 1 correspond to a value of fN/fH = 2.4 in Figure 2.
3. Scientific Applications Beyond the Ionosphere
 The knowledge and experience gained from ionospheric topside sounding has been applied to the interpretation of natural wave emissions, sounder-stimulated echoes and plasma emissions, and plasma diagnostics in space plasmas other than the ionosphere. It has also stimulated an entirely new interpretation of the X-ray spectra of exotic objects.
 The ISIS observations have been applied to at least three types of natural emissions. First, high-altitude ISIS-1 observations made the first determination that fN/fH ≪ 1 in the Auroral Kilometric Radition (AKR) source region [Benson and Calvert, 1979]. These observations were of fundamental support to the cyclotron maser generation mechanism of Wu and Lee , which has been applied to the interpretation of solar, planetary, and stellar radio emissions [see, e.g., Wu, 1985]. The astrophysical application of this mechanism has been modified and extended based on the comprehensive source region measurements using the FAST satellite [Ergun et al., 2000]. Second, lower-altitude ISIS-2 observations demonstrated that wave emissions often identified with fT in planetary magnetospheres are seldom observed at that frequency in the terrestrial auroral region; rather, the emission peak is more often at fN or at a frequency between fN and fT [Benson, 1993]. Third, Fredricks  suggested a relationship between magnetospheric-banded emissions observed between the harmonics of fH and sounder-stimulated plasma emissions with similar frequency patterns. Evidence for a close relationship between the sounder-stimulated emissions and the natural emissions, which have been observed in all planetary magnetospheres, has been presented by Oya , Benson and Osherovich , and Benson et al. .
 The experience gained from interpreting the wave cutoff and resonant phenomena observed by ionospheric topside sounders, as described by equations (1)–(9), has been applied to the interpretation of sounder records in planetary magnetospheres. These interpretations have been used to determine fN, and hence Ne, to within a few percent. In the terrestrial magnetosphere they have been used to provide the starting points for the inversion process leading to Ne polar cap profiles [Reinisch et al., 2001a; Nsumei et al., 2003] and hemisphere-to-hemisphere magnetic field-aligned profiles [Reinisch et al., 2001b] based on active soundings by the Radio Plasma Imager (RPI) on the Imager for Magnetopause-to-Aurora Global Exploration (IMAGE) satellite. In Jupiter's Io plasma torus they have been used to interpret Ulysses relaxation-sounder observations and determine Ne along the trajectory of the spacecraft [Osherovich et al., 1993]; an alternate interpretation of these observations was presented by LeSager et al. .
 The upper hybrid relationship, which is well known in space plasma physics, has recently been applied to the field of astrophysics to explain X-ray spectra of the accretion disks around neutron stars and black holes. Two examples of the upper-hybrid relationship have been discussed in connection with Figure 1 and equations (3) and (8). (A similar relationship is expressed in equation (9), after rearranging terms, and the corresponding ionogram examples have been presented by Osherovich and Benson .) There are astrophysical analogs to the fN and fT of plasma physics as expressed by equations (2) and (3), respectively. In plasma physics, fN describes the oscillation frequency of a plane body of electrons displaced in one dimension from their equilibrium position in a neutral cold plasma where ion motions are neglected because of the large ion/electron mass ratio. In astrophysics, a neutral test body of small mass displaced from the center of a sphere of constant density ρ (a zero-order model for a star) will oscillate at the frequency f ∝ ρ1/2 [Allen, 1973] in analogy with the expression for fN in equation (2). Mathematically, these problems are equivalent. Outside this sphere, a test body located at a distance r from the center will rotate with the Keplerian frequency fK ∝ r−3/2. The astrophysical analogy with the upper hybrid frequency in a magnetized plasma is with the quasiperiodic oscillations in the accretion disk surrounding a neutron star. (Twenty soft X-ray sources are now known to have such quasiperiodic oscillations [van der Klis et al., 1997].) This analogy is based on the well-known analogy between the Lorentz force/unit mass/unit charge v × B in a plasma with magnetic field B, and the Coriolis force/unit mass v × (2Ω) in a frame rotating with angular velocity Ω; where for the Lorentz force v is the velocity of a charged particle in a magnetized plasma while for the Coriolis force v is the velocity of a test mass in a rotating frame. Again, consider a body of electrons displaced in one dimension from their equilibrium position in a neutral cold plasma with ion motions neglected. This time, however, assume that an external B, oriented perpendicular to the direction of electron displacement, is present. The electrons will now experience the simultaneous influence of Langmuir plasma oscillations given by equation (2), parallel to the original displacement direction, and the transverse Lorentz force. The result will be to introduce a characteristic frequency for transverse oscillations equal to the upper hybrid frequency given by equation (3) [see, e.g., Goertz and Strangeway, 1995]. In a magnetosphere rotating with angular velocity Ω about a neutron star, a blob of mass near the accretion disk will simultaneously be subject to radial oscillations at fK and oscillations due to the Coriolis force resulting in hybrid oscillations with a frequency given by [fK2 + (∣Ω∣/π)2]1/2 [Osherovich and Titarchuk, 1999]. (Note that in an accretion disk that rotates with a frequency fK, there are always oscillations with the same frequency fK due to radial perturbations.) It is important to note that the magnetospheric rotation frequency ∣Ω∣ is different from fK because, according to Ferraro's theorem [Alfven and Falthammar, 1963], ∣Ω∣ depends only on the magnetic structure surrounding the star (if the electrical conductivity is high). Thus the rotating magnetosphere introduces a frequency that is different from fK that, when combined with fK, leads to hybrid oscillations. These oscillations formed the basis of a two-oscillator model of the twin-peak phenomena observed in the X-ray spectra of neutron stars [Osherovich and Titarchuk, 1999]. This model, which has recently been applied to the spectra of black holes (V. A. Osherovich and J. Fainberg, manuscript in preparation, 2004), provided a radical new approach to explain a fundamental observational problem in astrophysics, namely, that observations do not support the predicted constant frequency difference between the twin peaks. The two-oscillator model attributes this frequency variation to the differential rotations of the magnetospheres surrounding neutron stars and black holes. Its genesis was based on recognizing the importance of the upper hybrid relationship to the interpretation of ionospheric topside ionograms and applying this experience to another field of science.
 A large database of digital topside ISIS ionograms are now available for ionospheric research using state-of-the-art analysis techniques. In addition to the well-known applications of ionospheric research to scientific problems associated with the ionosphere and applications to technological systems, the significance of this research to other space plasmas has been highlighted. Particular emphasis has been placed on the scientific spin-offs from obtaining a proper understanding of the plasma resonances stimulated by ionospheric topside sounders. These scientific spin-offs include (1) the determinations of electron densities from space-borne sounders in planetary magnetospheres based largely on the knowledge and experience gained from ionospheric topside sounding, (2) the interpretation of natural emissions from magnetized space plasmas based on spectral similarities with both natural and sounder-stimulated ionospheric emissions where precise Ne and ∣B∣ measurements in and near the radiation source regions have provided critical theoretical guidance, and (3) the interpretation of X-ray spectra associated with neutron stars and black holes based on the mathematical analogies between ionospheric collective plasma oscillations in the presence of an external magnetic field and astrophysical mass oscillations in a rotating system.
 The digital ISIS ionograms (as used in Figure 1) were produced with the support of NASA/OSS Applied Information Systems Research Program (AISRP) RTOP grant 370- 03-00-04. We are grateful to the late W. Schar and also to G. Burgess and P. Rozmarynowski for assistance in the data presentations.