i'm using FB version 2.1.1.17910; classic version

i have two simple tables defined, "master" and "detail"
each has a primary key on column "recno", and table "detail" has an
index on a column named "master"

i've loaded the "master" table with about 5000 rows, and the "detail"
table with 100 rows per each master -- so a total of 500k rows

this query executes within milliseconds:

select * from detail where detail.master in
(select recno from master where master.recno > 1 and master.recno <
500 order by recno)

whereas this query executes in over 10 seconds

select * from detail where detail.master in
(select first 10 recno from master where master.recno > 1 and
master.recno < 500 order by recno)

i need the second query to execute much faster.
i know i can accomplish something similar using join, but i really
need this to work for sub-select as i have a lot of source code
already written for sub-selects.

help is greatly appreciated!

here's the ddl to create the tables:

CREATE TABLE MASTER (
RECNO INTEGER NOT NULL);

ALTER TABLE MASTER ADD PRIMARY KEY (RECNO);

CREATE TABLE DETAIL (
RECNO INTEGER NOT NULL,
MASTER INTEGER);

ALTER TABLE DETAIL ADD PRIMARY KEY (RECNO);

CREATE INDEX IDX_DETAIL ON DETAIL(MASTER);