stored procedure problem

I have a stored procedure on two machines (same hardware, one with Solaris 2.5.1 and the other Solaris 2.6), server 1 w/ online 7.22 UC1 and the server 2 with online 7.23uc6. In this stored procedure there is a select statement which is puzzling me. Here is why: on server 1: the select statement uses an index within the SPL but if I take the same select statement and run it thru dbaccess it does a sequential scan. on server 2; the select statement uses sequential scan in and outside the spl. The query takes 8 seconds when using the index and 1 minute when it doesn't. My question is why is the engine not being consistent all around. I've done update statistics on the tables and procedure involved and the behavior does not change. Any help would be greatly appreciated since it is holding up my upgrade.