That was a big help.
I went back through the wiki documents left by me predessor, and an email from him while he was configuring this switch, last year.
What I did was
- delete VME
add ip 192.168.1.1 to me0, unit 0, no static route.
This enables it to be reached from our Cisco switch, which has a 192.168.0.0 subnet.
Now, this has some drawbacks: namely if unit 0 does down, we won't be able to login to the switch and manage it. And that one could login to the port of the switch when it is the slave.
The former problem (why doesn't VME work with our setup) is one we set aside to troubleshoot at a later day.
The latter problem is fixed by executing 'request session member 1' when one is logged into member 0.