Ok, some weirdness.
I backup up my 512Mb USB flash drive, wiped it and deleted all partitions. I then created a single primary partition and formatted it with ReiserFS. I copied over about 200Mb of data, and shrunk the filesystem by 164Mb with "resize_reiserfs".
Next, I opened up fdisk, deleted the primary partition, created an extended partition and two logical partitions within that, one the same size as my shrunken ReiserFS. The other filling up the rest of the space.
I then tried to mount the logical partition, but it didn't work

Complained about being unable to find the superblock.
So, I went back to fdisk, deleted all the partitions again and recreated the single primary partition. After that I was able to mount and use the filesystem ok, and my files were fine.
Very strange. Creating the extended partition didn't seem to damage the ReiserFS (or its superblock), so that would seem to verify that ReiserFS leaves a blank space at the beginning. But then why couldn't I mount the ReiserFS if it was perfectly fine and residing within the logical partition?
Makes me wonder about a couple of things. Are sector sizes on USB drives guaranteed to be 512b?
It seems the problem was this: ReiserFS *always* looks for the super block beginning at byte 65536 of the partition. Now if that's where it was on the original partition, and the new extended partition takes the first sector as its partition table, then the superblock is actually going to be at byte 65024 of the new logical partition. Damnit.
Leaves me at a bit of a loss about what to do with my HD though...
I'm going to try to use "reiserfsck" on the USB drive to rebuild the super block in my logical partition and then rebuild the file tree and see what happens...