Thank you Bernt (and Sebastian):

I now understand that it is necessary to pull the changes to the alternate tree before pushing.  This will make a big difference in the future, I believe.  It worked as advertized on the /tmp/junk experiment. 

I think the other point I had missed, that would be necessary in an FAQ or tutorial, is that it is necessary to initially clone each of the working trees, on each machine, from the bare repo on the USB flash drive.  Then it now makes sense to me to name that as org.git, to distinguish it as a bare repo.  And once the working trees are cloned from there, then the org.git on the USB drive is recognized automatically as "origin".  That point eluded me.   And that git pull will perhaps automatically find origin.

Meld is starting to make sense too.  Thought it would be nicer to have a tool that would be able to make simple updates from two edits on the same file if they don't conflict.  Usually, this will not be the case.

Now, is it useful to use the switches on an ext2 formatted USB drive, as follows?

    $ git clone -l --no-hardlinks --bare . /path/to/usb/stick

Thank you again,

Alan