Posted for "Ted Noel" <tednoel@cfl.rr.com>:

 Angier,
 
 The discussion on this in the AeroElectric Connection reveals ground loops.
The basic issue is that there are very small potential differences between
"ground" at various points. In an aluminum airplane, this is less because of
the immense size of the conductor (the airframe). In our composite birds, we
have to depend on the relatively small (wire) connections, so that even a
couple of feet difference between "ground" locations can result in a small
potential, and therefore current flow. Without a lot of testing, the results
of these ground loops can be either nothing or nasty. Ground at both ends. It
removes one potential (pun intended) problem.
 
 Ted Noel