Per> Yes and no. It was meant to measure "post-duplicate-elimination". Per> The 'print $1' selects out the *addresses*, not names. Yeah, I completely misread that. Oops. Tom