CMEs carry large amounts of plasma and magnetic fields into the
heliosphere. This material is detected by remote sensing and
in-situ spacecraft observations. Remote sensing of CME plasma has
been primarily with the Helios 1 and 2 white light photometers and
by measurements of IP scintillation of galactic radio sources from
the ground [e.g., Hewish and Bravo, 1986]. The Helios
photometers were analogous to heliospheric coronagraphs and viewed
200 CMEs over a period of 9 years [ Jackson et al.,
1994]. The occurrence rates, spans and speeds of these CMEs are
consistent with measurements made by coronagraphs near the sun [
Webb and Howard, 1994]. The appearance of the expanding CMEs is
also similar, often with two brighter ``legs'' behind a tenuous
leading front. However, derived masses tend to be higher than those
determined by coronagraphs [ Jackson and Webb, 1995]. The IP
scintillation technique detects transient disturbances as enhanced
scintillation along the line of sight to a distant steady source.
However, there are problems with the accuracy and reliability of
this technique for detecting CME plasma.
A variety of in-situ signatures have been proposed as proxies for the passage of CME ejecta past spacecraft. These include shocks, density changes, decreased temperatures, flows with enhanced helium abundances [He(A)], and magnetic field structures consistent with looplike topologies [cf., Gosling, 1993; Webb, 1993]. Recent efforts have focussed on signatures considered to be indicative of the topology of the ejected magnetic fields [see McComas, 1995]. Most such studies involve observations and modeling of magnetic clouds and bidirectionally streaming particle flows. Magnetic clouds are long-lived solar wind flows having enhanced field strengths which exhibit smooth, coherent rotations Burlaga [1991].
Wilson and Hildner [1986], Bothmer and Schwenn
[1994], Rust [1994] and others have associated some
magnetic clouds with solar filament disappearances. Solar
filaments consist of dense plasma embedded in helical, horizontal
magnetic fields. The close association of CMEs with filament
eruptions and shearing fields near the surface has led to the
modeling of CMEs as flux ropes. Gosling, [1993] describes a
process wherein the interior fields of a rising, sheared CME
reconnect, resulting in an ejected flux rope and new, closed
coronal loops. He suggests that 
of all
bidirectional electron events and magnetic clouds, i.e., CMEs, in
the heliosphere have the characteristics of flux ropes.
Bothmer and Schwenn [1994] found in 4 of 5 cases that Helios
magnetic flux ropes (clouds) had the same orientation and
polarity as associated erupting filaments at the sun. Furthermore
S. Martin [1994, private communication] finds that high latitude
filaments always have twist in the same sense in a given
hemisphere, regardless of the cycle number. These results suggest
that the sign of the helicity of the erupted fields can be
predicted assuming one can associate a given flux rope with a
high-latitude filament eruption. Rust [1994] did so for a
list of magnetic clouds modeled as force-free structures at 1 AU
and claimed good agreement. Thus, filament eruptions and CMEs may
be important ways that the sun sheds helicity or twist.
Bidirectional flows of electrons have been interpreted as evidence that the associated transient field lines are entirely or partially closed and, thus, are a good proxy for CMEs [e.g., Gosling, 1993]. Bidirectional proton flows have also been detected in the solar wind, often but not always in association with shocks and ejecta flows [e.g., Richardson and Reames, 1993]. Although magnetic clouds and bidirectional flows are usually considered proxies for closed structures and CMEs, the observational evidence is not sufficient to differentiate between the partially closed (flux rope), entirely closed (bottle or plasmoid), or even entirely open field configurations [e.g., Kahler and Reames, 1991; Suess, 1993].
Recently, a number of bidirectional events were observed as
Ulysses moved to higher southern heliographic latitudes [
Gosling et al., 1994]. The speeds of these events (740 km/s on
average) are much higher than those observed in the ecliptic
plane and comparable to the high speed of the surrounding solar
wind flow. This suggests that CMEs in general are affected by the
same forces as the surrounding solar wind. Since the typical speeds
of CMEs observed in the field of view of coronagraphs (
10
solar radii) is 
400 km/s, much of this acceleration must
occur beyond this point.
Solar energetic particle (SEP) events are commonly associated with IP shocks and can arrive at Earth within minutes following onset of a flare [see Reames, 1995]. They are of concern because of their deleterious effects on communications systems, satellites, astronauts and high flying aircraft. In the new paradigm, CMEs are important to the production of the SEP particles because the bulk of the particles are not accelerated at the low flare site or during its impulsive phase, but relatively high in the corona and for prolonged periods of time by the shock driven ahead of the CME. The key supporting observational results are: large, prompt SEP events are nearly always associated with fast CMEs, they are not well correlated with the amplitude of a flare's impulsive phase, their abundances and ionization states are typical of ambient coronal rather than flare-heated plasma, and isolated erupting filaments can also produce SEPs [e.g., Kahler, 1992].
As mentioned earlier, CMEs tend to arise in coronal streamers which form a belt that encircles the sun and is the base of the heliospheric current sheet (HCS). The HCS is the separatrix between the oppositely-directed fields on either side of the streamer belt and, in projection, is the heliomagnetic equator of the sun. A new model by Crooker et al. [1993] suggests that the base of the HCS may often be broad, encompassing multiple helmet streamers tens of degrees across with associated multiple current sheets. Most CMEs might then be spatially associated with the HCS. In this view the HCS is more dynamic than previously thought, and acts as a conduit for a range of activity from slowly evolving streamers to large CMEs. This activity also acts to compress, amplify and align preexisting magnetic discontinuities in the HCS and at the leading edge of high speed streams [ Neugebauer et al., 1993]. These factors enhance the geoeffectiveness of CMEs as discussed next.