>it would seem to me that the only way you could gain from spreading the
>data base across different volumes, would be if there were several
>(parallel) queries to it, able to run concurrently to the different
>physical data sets.

There is a gain because the access is spread over multiple disk arms (and
hopefully I/O paths), as accessed by a server which is either multi-processed
or multi-threaded.
   Richard Sims, BU