The radiation impedance of six feet of cable isn't very low, so the
incremental benefit of the second is much smaller.
Imagine the wire as a piece of antenna. If it's 1/4 wave long from the
driving source, it acts as a whip antenna, when the far end is "open
circuit" (read, sufficiently high impedance).
If you have a wire grounded at both ends, it can resonate at 1/2 wave,
basically an inverse dipole (which is grounded in the middle and has
antinodes at the open ends).
The quality of the ground matters; a computer monitor would have to be very
large, physically, to make a good ground against a 6' quarter or half wave
resonance. Since it therefore does not make a great ground, it develops
some voltage on its exterior, which means it serves, in part, as the antenna
as well, and so on to anything connected to it.
The quality of the wire is most important at the nodes, where current is
highest. For any arbitrary frequency at any arbitrary position, this is
difficult to define -- the nodes could be distributed anywhere along the
wire. However, nodes are almost guaranteed at the ends (where the monitor
or whatever serves as something of a ground plane, even if a poor one), so
increasing the impedance there with a ferrite bead does an excellent job at
dampening wideband resonances, and thus propagation of noise.
Damping at the antinodes is also possible, but less likely to help, because
you need a ground to work against, and dielectric loss rather than core
loss.
Tim
