This seemed to be one of those things that "just shouldn't happen", as the new Panther installation was confined to my second drive, and as far as I could determine had never laid a finger on my 10.2.8 drive. It hadn't, for example, replaced the USB kexts -- the cause of much exercising of the new KP interface in early versions of 10.2, as I recall. As an aside, the revised KP interface introduced around this time is VERY BAD piece of human interface design. It tells you simply that "Your computer needs to be restarted...", which is deeply misleading. It should say, in virtually all the cases of KP that I've ever come across on any UNIX: "There is a very serious instability in your hardware/software which will recur unless dealt with." But I digress...
The only cause I could think of is that somehow Panther has introduced some kind of change in my second drive that Jaguar doesn't know how to deal with. This pointed very obviously to the new journalling version of HFS+. Sure enough, once I had rebooted into Panther, used the Disk Utility to switch off journalling on the secondary drive, and rebooted back into Jaguar, my production system became wholly stable again. So I can now continue my gentle and controlled transition to the new version of the operating system.
[robg adds: For another data point, I ran a Panther/Jaguar dual boot machine for a month or so, and never had a single KP ... and I did not disable journaling. I'm not sure what's different between our systems, but if you're having KPs with dual boot systems, give this fix a try.]

