The first evidence of high-energy particles from the Sun was obtained 50 years ago when Forbush [1946] used sea-level ion chambers to study the large solar events of February and March of 1942. Over the next 20 years, observation of these solar energetic particle (SEP) events using neutron monitors and riometers (that measure radio opacity of the ionosphere) and, later, with detectors on balloons and satellites, led to an extensive body of knowledge on the time profiles, spectra and particle abundance in the large events. Meanwhile, there was already a rich history of the study of solar flares spanning 100 years since the first observations reported by Carrington [1860]. With no knowledge of the existence of coronal mass ejections (CMEs) [ Kahler 1992], it was tempting to assume that the particle acceleration somehow occurred in spatial and temporal conjunction with the solar flare itself. Thus the ``solar flare myth'' [ Gosling 1993] of particle acceleration began nearly 30 years ago.
However, it was not so easy to explain the SEP observations in terms of a flare source. Flare activity at the Sun lasts, at most, for hours while SEP events can persist for many days. Particles could be seen from events anywhere on, and sometimes behind, the visible disk, yet it was well known that particles could not cross the interplanetary magnetic field lines over such large distances. The proposed explanation, which dominated the study of energetic particles, was the Reid-Axford model [ Reid 1964; Axford 1965]. The key to this model was a solar corona that stored the particles, somehow allowing them to diffuse easily in longitude, then slowly diffuse outward once the particles leaked from the corona into interplanetary space. No mechanism for the ``coronal diffusion'' was ever identified, however, the model offered enough adjustable parameters to fit some of the more well-behaved time profiles of the particle intensities in large SEP events.
At about the same time, Wild, Smerd and
Weiss [1963] proposed a much different picture to explain their
radio observations. They observed type III bursts, in which fast
frequency drifts were produced as electrons streamed outward from
the Sun through plasma of decreasing density, and type II bursts,
where drift rates corresponded to the lower speed (
1000
km/s) of an outbound shock. This led to the idea of two-phase
acceleration, with electron acceleration early in the flare
followed by proton acceleration at the expanding shock. While we
now know that protons and electrons are accelerated in both
``phases,'' these authors emphasized the presence of two
acceleration mechanisms and prophetically suggested that particle
abundances might distinguish them. Of course, interplanetary
(IP) shocks had also been observed directly near Earth by this
time, but the idea that shocks could dominate the acceleration of
particles over a long time periods at longitudes far from the
flare was not appreciated. We now know from Voyager and Pioneer
observations that multiple mass ejections merge to drive shocks
that continue to accelerate particles far beyond the orbit of
Earth, perhaps all the way to the heliospheric boundary.
An essential ingredient of our current picture of both SEP
events and large non-recurrent geomagnetic storms [e.g.,
Webb 1994] is the coronal mass ejection (CME). Observations on
the Skylab mission [ Gosling et al., 1974] (a decade
after the ``coronal diffusion'' models) showed that CMEs can
eject up to 10
g of material at speeds above 1000 km/s (an
energy of 
10
ergs). These sudden, violent ejections
probably drive all of the traveling interplanetary shocks and are
associated with gradual or long-duration (>1 hr) soft X-ray
production associated with field reconnection as the large
filament and surrounding magnetic structure of the CME tears away
from the restraining fields. CMEs are poorly associated with
flares (see review by Kahler [1992]) but, in very large
events, CMEs and flares do occur together.