I think it's well-nigh impossible for a chess program not to have an opening book. Maybe the first attempts on analog computers had none, but since the days of core memory, I'm going to say not a single chess program has lacked one. There's simply no point to doing without one, and even the dumbest freeware has always had them -- chess openings themselves are not copyrightable, so they can just copy them from earlier programs or transcribe them from MCO or whatever they want.
So it appears to me that the programmers were simply too disinterested in playing the game and testing their own code not to disable search when on the book early in the game. On difficulty 8 out of 10, it takes 2 seconds of hourglass to respond 1 c5 (Indian defense, certainly from the book), and on setting 9 it takes 10 seconds to make the same move; but on setting 10, it's 20 seconds, which is just absurd.
Incidentally the computer is a dual-core AMD from last year, though reading from the book should be just as fast on a 68000 machine from 1984.
As an aside, its book seems really committed to the c pawn, every game I've started with e4 (6 or so now, just to test it) it's either played an Indian (c5) or more unexpectedly the Caro-Kann (c4).