At first it was held that the aurora was simply the overflow of the outer belt:

"... we propose that the radiation belt is the reservoir whose leakage of particles is the direct cause of visible aurora. It is further suggested that solar plasma replenishes the reservoir from time to time, working its way into the outer reaches of the earth's magnetic field when its density is sufficiently great, then being trapped in the field." [Van Allen et al., 1959].

4. The Large-Scale Structure

The picture became much clearer during 1962-5 [LeGalley, 1963], when the large-scale structure of the Earth's magnetic environment (named "magnetosphere" by Gold[1959]) was established (Figure 4). This phase followed the observational confirmation of the existence of the solar wind, predicted by Parker (see BH-1). In 1962 Explorer 12, on the sunward side of the Earth, repeatedly crossed a sharp boundary between the magnetosphere and the solar wind, named the "magnetopause" [Cahill and Amazeen, 1963; Cahill, 1995; Hines, 1963; Hess et al., 1965; Dungey, 1978] and resembling in many ways the boundary predicted (for storm times) by Chapman and Ferraro (see BH-1).

Figure 4. Regions and currents of the Earth's magnetosphere, representing the state of our knowledge around 1970 [Lopex and Baker,1994]. A small normal field component on the magnetopause allows some interplanetary field lines to link up with the Earth's field.

The tail contained two low-density "tail lobes", bundles of roughly parallel field lines directed sunward in the northern lobe and anti-sunward in the southern lobe, as required for field lines connected to the appropriate polar caps of the Earth (compare Dessler and Juday [1965]). Sandwiched between the lobes was a near-equatorial "plasma sheet." The fact that the sheet was bounded by oppositely directed magnetic fields showed that it carried a dawn-to-dusk electric current of the order of 200,000 A/RE, and this current closed around the magnetopause, giving the entire flow pattern the profile of the Greek letter theta [Axford et al., 1965; Axford, 1994]. I. Axford (personal communication, 1992) has claimed that this work predated the disclosure of IMP 1 results.

The earthward edge of the plasma sheet contained ions of about 20 keV, and these blended continuously with the ions of the ring current [Frank, 1971]; on the other hand, the energetic electron population of the sheet [Vasyliunas, 1968] displayed an asymmetrical inner edge. (Because of the presence of low-energy electrons, this does not contradict the plasma's electrical neutrality.) The inner edge was originally credited to electrons precipitated into the atmosphere by plasma instabilities [Kennel and Petschek, 1966] but in hindsight it might also be related to region 2 Birkeland currents (below), carried almost entirely by electrons.

The existence of a "ring current" circling the earth during magnetic storms was inferred long before the spaceflight era, from the magnetic signatures of such storms, and was the subject of extensive speculation [Smith, 1963]. Observations in space indicated that it was carried by trapped particles of relatively low energy, as had been suggested by Singer [1957], and that it was not a transient effect of storms but a permanent feature. It was studied (launch dates given as (mo.day.yr)) by the Orbiting Geophysical Observatories [Ludwig, 1963; Jackson and Vette, 1975] OGO-1 (9.5.64), OGO 3 (6.7.66) and OGO 5 (3.4.68), later by Explorer 45 ("Small Scientific Satellite" or S3; (11.15.71)) [Longanecker and Hoffman, 1973] and by the European GEOS 1 (4.20.77) and GEOS 2 (7.14.78). Its energy is mainly carried by ions with energies around 50 keV [Frank, 1967], but its exact composition and energy (median value about 85 keV) were only mapped by the Charge Composition Explorer (CCE) of the AMPTE mission, launched in 1984 [Williams, 1987; Lui and Hamilton, 1992].

The ring current grows stronger during magnetic storms, and it was shown theoretically that the intensity of the magnetic disturbance observed at Earth is then very nearly proportional to its total energy [Dessler and Parker, 1959; Sckopke, 1966; Carovillano and Siscoe, 1973]. Between such injections the ring current slowly decays, mostly by charge exchange collisions [Liemohn, 1961; Smith et al., 1976] with atoms of the hydrogen cloud ("geocorona") surrounding the Earth [Hunten and Donahue, 1976 ; Carruthers et al., 1976]. Each collision produces a slow proton and a fast neutral atom; such high-energy atoms are able to trigger particle counters, and since they are (like albedo neutrons) unaffected by the Earth's magnetic field, it has been proposed to use them for remote sensing of the ring current [Williams et al., 1992]. In 1971 it was also observed that a sizable portion (about 15%) of the particles added to the ring current during a magnetic storm were O+ ions of atmospheric origin, evidence of a near-earth acceleration process [Shelley et al., 1972].

A special region is the polar cusp, the region inside the high-latitude dayside magnetopause which separates field lines that close near the "nose" of the magnetosphere from those swept into the tail. The cusp was first explored in 1969-72 by the European spacecraft HEOS 1 and HEOS 2 (and later by Iowa's Hawkeye-1) which found a region of disordered weak magnetic fields [Mencke-Hansen, 1976]. Such fields are unable to exclude the solar wind, or more accurately, the magnetosheath plasma, a name given by Dessler and Fejer [1963; see Fig 1 there] to the solar wind plasma slowed and heated by passage through the Earth's bow shock. It is therefore filled with magnetosheath plasma, which spills into a funnel-shaped region reaching all the way to the atmosphere and produces a characteristic auroral glow [Shepherd, 1979].

The HEOS satellites also discovered a variety of boundary layers just inside the magnetopause, with a thickness of the order 0.3-1 RE, typical density 1-3 cm-3 and typical tailward velocity of 100-150 km/sec. A narrow "low latitude boundary layer" was found on the day side [Paschmann et al., 1976] and may be the result of reconnection processes (below) near the nose of the magnetosphere, though other origins are also possible. The much thicker "plasma mantle" found at high latitudes tailward of the cusp [Rosenbauer et al., 1975] may be formed by sheath plasma crossing an Alfvenic transition extending from the site of magnetic reconnection, discussed further below [Levy et al., 1964] , and it gradually widens with increasing distance from Earth [Hardy et al., 1975, 1979].

In what follows, coordinates in the magnetosphere will be given in the so-called geocentric solar magnetospheric coordinate system (GSM), described by Russell [1971; also Hapgood, 1992]. In these coordinates, the x-direction points to the Sun, the x-z plane contains the Earth's dipole axis and north is on the positive side of the plane z=0. The x-direction is also assumed to be the one from which the solar wind is blowing, although in very accurate work the small aberration angle (about 4 degrees) due to the Earth's orbital motion is also taken into account. The actually observed direction of the solar wind may differ from this by a few degrees (in both elevation and azimuth) due to effects occurring closer to the Sun.