sequential scan when using bigint value

I have a table of 2mil records. The table consists of two columns, id
(BYTEA/GUID, PK) and i (INT, UNIQUE INDEX). Could someone explain why,
when using a bigint value like this:

  select * from partition where i=3000000000;

or

  select * from partition where i in (1,2,3,3000000000);

Postgres immediately switches from index scan to seq scan?

--
dave