Whistlers have, for many years, been used as probes of the ionosphere and magnetosphere. Whistlers received on the ground have been shown (Smith [1961], Helliwell [1965]) to have propagated, in almost all cases, through ducts of enhanced ionisation aligned along the magnetic field direction. Analysis of these whistlers, using for example the Ho and Bernard [1973] method, allows determination of the L-value of the field line along which the signal has propagated, the equatorial electron density and the time of the initiating lightning strike. Satellite received whistlers, known as fractional-hop whistlers, are not restricted to propagating through ducts and, in this case, ducted whistlers are probably rarer than unducted whistlers. Analysis of these whistlers is consequently much more difficult as the propagation path is often not known. This study is an attempt to understand some of the characteristics of whistlers received on the 18182 satellite at low latitudes during October 1976. Haselgrove's [1954] ray tracing equations, together with realistic density and magnetic field models, have been used to determine the ray paths and travel times. The whistler dispersions, calculated from the travel times, are compared with the results obtained from analysis of the 18182 data. Values given by the density models used were also compared with density values obtained from other models and values recorded by ionosondes during the same period and at locations close to the latitude and longitude of the 18182 satellite. Another part of this study considers the cyclotron resonance interaction between ducted whistler mode waves and energetic electrons. During this interaction, electrons can diffuse into the loss cone and will then precipitate into the upper atmosphere causing secondary ionisation. This ionisation patch modifies the earthionosphere wave guide and can be observed as phase and/or amplitude perturbations on VLF transmitter signals, known as Trimpi events (Helliwell et al [1973], Dowden and Adams [1988], 1nan and Carpenter [1987]) . Trimpi events and associated whistlers were observed at Marion Island (46°53" 5, 37°52" E, L = 2.63) during May 1996. Analysis of the associated whistler groups confirms that the Trimpi events can be explained by the above mentioned cyclotron resonance interaction and subsequent electron precipitation. During this process the whistlers were propagating towards Marion Island while the electrons were propagating away. The electrons must therefore have mirrored in the northern hemisphere before precipitating near Marion Island causing the observed Trimpi. The calculated time delays are shown to confirm this process. During the unusual 2-hour period of observation, the Trimpi associated whistler groups were, in all cases, followed by a second, fainter whistler group which has been called a whistler 'ghost' . The dispersion of whistlers within this second whistler group are shown to be the same as those within the initial whistler group indicating that these whistlers must have propagated through common ducts at different times and hence must have been caused by different atmospheric discharges. It is thought that during the wave-particle interaction, which caused the observed Trimpi, some of the energetic electrons may have precipitated into the northern hemipshere triggering this second discharge. The timing between the two whistler groups is such that, if the above triggering is correct, the interaction must have taken place about 10° from the equatorial plane .

Thesis (Ph.D.)-University of Natal, 1997.