In the absence of some mitigating processes, the outward
expansion of CMEs would cause the number of open field lines,
and hence IMF magnitude, to grow without bound. This ``magnetic
field magnitude catastrophe'' was first pointed out by
Gosling [1975]. The amount of open magnetic field in
interplanetary space has been examined most recently using an
improved proxy, the ``total flux integral'' which combines the
scalar inward- and outward-pointing fluxes and removes the
effects of variations in the solar wind speed in determining the
amount of magnetic field crossing 1 AU [ McComas et al.,
1992a]. If we assume that all counterstreaming electron events
represent simply connected magnetic tongues (B in Figure 2), then
for CME rates characteristic of solar maximum, the field
magnitude at 1 AU would double over 
9 months in the absence of
any mitigating processes. If CMEs remain magnetically attached
but are flux ropes, where field lines can loop around many times
in interplanetary space before connecting back to the Sun, then
the rate of build-up would be smaller. However, it must be
stressed that if CMEs retain any attachment to the Sun
whatsoever, the field magnitude catastrophe will ultimately occur
in the absence of some other process to close off previously open
fields.
The average magnitude of the IMF displays 
50% variation from
the mid-1970s to the mid-1980s [ Slavin et al., 1986] while
the total flux integral displays 
60% variation [ McComas
et al., 1992a,b]; both show basically the same modulation over
the solar cycle with maxima shortly after solar maximum and
minima shortly after solar minimum [ McComas et al., 1992a,b;
McComas, 1994]. Both this variation over the solar cycle
and the fact that the field magnitude does not display unbridled
growth, indicate that in addition to CMEs opening new, previously
closed field regions from the corona into interplanetary space,
there must be some process for closing off previously open field
regions and returning magnetic field lines to the Sun.