> I'd like to use HTTrack to download a site (html, 
gif 
> etc), WITHOUT saving it to the local hard disk.  The 
> purpose would be to fill our proxy server overnight 
> with pages we're interested in.
> Can it do this?  Or is there anything else I should 
> use?
Well, you can only scan HTML structure without saving 
it (--spider) but not other files.
The best thing to do would be a batch file with 
temporary file.

If you don't want to waste (even temporarily) space, 
use the batch -V option, to specify a command to 
execute after each file being downloaded: (example for 
Linux/unix)

httrack www.foobar.com -V "rm \$0"