qemu-doc.html, section 3.9.1, in the qemu directory, left me doubting whether I could access a USB flash drive from qemu, but I thought, "Hmm, windows just thinks its another disk drive", so...
-hda \\.\PhysicalDrive1
(my system has only one actual hard disk, (PhysicalDrive0, in this context), so perhaps the USB stick is PhysicalDrive1. Sure enough, that works -BUT-...
1) TC bitched about a problem on hda5 (my USB stick is 4GB partitioned as hda1=Primary/fat32/2GB and hda2=extended/2GB w/hda5=logical/ext2/2GB)
I didn't mess with hda5 at all.
2) Changes written to hda1 were not visible in windows until I physically removed and reinserted the USB stick, which I did after
a: unmounting the filesystem in TC (changes still not visible to Windows)
b: shutting down TC and qemu (changes still not visible to Windows)
c: using View / Refresh in Windows explorer (changes still not visible to Windows)
After reinserting the USB stick, the files looked fine (as far as I can tell) in Windows.
The moral of the story - this looks like it works, but you want to test -thoroughly- on a scrap filesystem before trusting this setup.