As the serials technician at the Fox Chase Cancer Center library in
Philadelphia, I can report that our subscription decisions also depend
almost exclusively upon the requests of the scientists at the center.  The
library seldom initiates subscriptions based upon its own determination of
a journal's usefulness to the staff.  We maintain approx. 450 open titles
in oncology and biochemistry, so our size allows our decision making process
to remain fairly informal.

We also put great emphasis upon the availability of a sample issue for evalu-
ation.  The advice I could give to a journal publisher is to _never_ ignore
the request for a sample issue to evaluate!  As the library staff member whose
task it is to actually try to get these sample issues from the publishers,
it has been my experience that publishers not responding to requests for
sample issues do not end up with a subscription to their journal.  Journals
that don't get evaluated usually don't end up being subscribed.

Scott Denlinger
denlinger@hslc.org