Thanks to those who responded.  I set the timeouts
and packet sizes for the mount thusly:

# mount_nfs -t 1000 -r 1024 -w 1024 euclid:/u13/netbsd /mnt

and the result seems to be success.  I have written (so far)
about 5 meg of a 7 meg file, and it is slowly continuing.

There are still silo overflows, but the transfer continues.

This is at 19200 baud, using PPP.  I'd like to see something in
the FAQ about this...