>>     I don't think it *is* normal, since this was not the observed 
>> behavior before I updated my binaries to 950702.  Prior to that, the 
>> address provided to me by my ISP simply replaced the <Link> entry and 
>> there remained a single instance of both sl0 and lo0...

>I think you are mis-remembering things.  This behavior is normal and has
>been around a long time.  It is certainly present on one of my HP systems
>that has not been upgraded since October 94.  I don't remember what NetBSD
>0.9 did.

    Hmmm...  Well, anything's possible I suppose, but I can't imagine how 
I could have missed all those console messages from routed ever since 
I've been running NetBSD...  I haven't changed *anything* in my 
configuration.

>> there is no change in the output of 'netstat -r').  Further, why does it 
>> *not* delete the route to the inactive lo0?

>lo0 is NOT inactive.  Try telneting to "localhost" or to "0".  You'll note
>that inactive interfaces have an asterix after their name.

    I am aware that lo0 (at least *one* of the two listed) is not
inactive; I am operating on the assumption that at least one of the two
*is*, since, clearly, one of the two sl0s is.  Running tcpdump shows
plenty of routing information being exchanged via sl0, but routed insists
on telling me that the route to sl0 has timed out after a few minutes. 
Tcpdump continues to show plenty of routing information being exchanged
via sl0.  I take this to mean that one of the two instances of sl0 that 
are listed is inactive, but the other is not.  Wouldn't this same logic 
apply to the two instances of lo0 as well?