Well, Microsoft solved this problem for me: this morning I found it had shut
down and restarted the computer I was running HtTrack on :-(
(A disadvantage of running programs from other people's computers..)

So besides escaping spiders (see elsewhere) and pages taking hours to scan
(not yet reported as I don't have a good small example), here is another
reason why one might want to restart an interrupted download.

But I'm not going to restart this one this year, as I won't have an
uninterrupted session that lasts long enough.  Trying again next summer..