psql nested queries with 2000+ records

I have noted with 2000+ records in a table (using a btree unique index on
the primary key, btw) performance of a query takes 15+ minutes to respond.

The query is in the structure of this

select foo from bar where foo in (select foo from bar where bazzfoo like
'Word%');

Just reporting a possible bug..

With smaller amounts of records, responce time is within a few seconds.

Other information you might want:

postgres version: 6.3
os: BSDI 3.1

Thanks
-coronach@hill-b-073.resnet.purdue.edu