This is  a one-time  thing due  to reloading the  archives from  the Napa
site, which at 5M per hour took  forever to FTP. Following the disk crash
that  forced us  to  reinstall the  system from  scratch  and reload  the
backup, it happened another time  due to messages having disappeared from
the archive. Finally,  as the Napa archives were reloaded  just after the
daily backup, we had to re-reload them and this messed up the indexes one
more time. From now on they should be ok.
 
  Eric