Please clarify some areas of confusion for me.
1) Is a 2 1/2" "laptop" hard drive preferred for an external USB
enclosure?
Not necessarily, a 3.5" is faster and far higher capacity
per dollar. On the other hand, a 2.5" is more portable
(assuming it's in the smaller enclosure that only accepts
2.5" drives), including that the drive may be able to run
from the ~500mA (minus what the enclosure itself consumes)
that USB supplies, or supplies additionally through a second
cable plugged into another USB port for more power, so the
whole thing is only one smal piece of gear plus a USB cord
or two. With the 3.5", you have it even larger if it has an
internal PSU, or more commonly a power brick similar to
those used on notebooks.
I'd go with the 3.5" unless you plan on transporting it off
premises.
2) Is it true you do not need separate power for a 2 1/2" external USB
enclosure because the drive can be powered by the USB port?
Usually, though as mentioned above, some have a 2nd cord or
a special cord with a USB plug at one end that turns into a
small barrel connector at the other end which plugs into the
enclosure's power jack.
3) What kinds of sustained transfer rates (in megabytes per second -
MB/s) to another hard drive can be expected with external USB
enclosures? What is the fastest kind?
Sustained large file transfers are typically under 28 MB/s,
sometimes much lower (depends on the enclosure and even more
the system's host controller chipset). If you want some
kind of random average I'd call it 18MB/s on large files,
10MB/s on smaller but without any contention for bandwidth
from other devices on same USB hub (controller).
4) Is SATA preferred over PATA for speed and low power use?
eSATA is preferred over USB for external enclosures.
SATA is only marginally faster when comparing same drive
family/model in PATA vs. SATA interface. NCQ on SATA300 can
help in certain types of uses but moreso fileservers than
typical PC uses, where you have multiple I/O requests
simultaneously.
Power differences are too trivial to consider.