Some additional information:
From my calculations, it looked like, for the entire job, both USB and
Firewire 400 clocked about 120 Mbps.
If I copy the files from one internal drive to another, it takes 9
minutes. Therefore copying to externals, which took almost precisely
21 minutes, is slightly less than half the speed. Ignoring overlapped
operations, etc., it would appear that the 12 minutes difference is the
overhead caused by the USB or Firewire connection. That should seem to
be the speed at which USB and Firewire are rated. In either case, my
calculations seem to show that the USB and Firewire are tranmitting at
about 211 Mbps, with the remainder of the time attributable to hard
drive seeks, searches, reads and writes. It would appear that both are
running about half the published rated speed.
I just happened to pick up the 9/21/04 issue of PC Mag where they
tested the Wiebe TECH Firewire800 external HD. They said "Testing
showed that FireWire 400's performance is essentially identical to that
of USB 2.0 for large directory files. Firewire 800 didn't provide as
big a throughput boost as we anticipated. We measure USB 2.0
throughput at 129 Mbps, with Firewire 400 clocking in at 133 Mbps.
Firewire 800 raised the rate to 148 Mbps. For multimedia files,
however, the difference was notable: from 190 to 235 Mbps, with USB 2.0
lagging a little at 167 Mbps.
Now their results of 129 and 133 Mbps are very close to the 120 Mbps
that I got for the overall test.
At any rate, I am somewhat placated in finding that my externals are
running not much worse than half the speed of the internal EIDE drives.
I had run all these tests to determine if going from USB 2.0 external
drives to Firewire 400 or even 800 would significantly improve external
drive performance. It would appear that FW 400 would provide only
slight gains; and going to FW 800 would probably not be worth the
outlay in funds for the slight gain in speed.