Consider me part of the choir; Gary's plan sounds great.

One other thing you gain by sending lots of probes is that you're more willing to risk some on targets that are either small or in which hazzardous terrain fills muchof the landing ellipse.

There are some areas where the MER system simply won't work -- if you know that the terrain *everywhere* in the ellipse is very rough/steep/whatever, then it doesn't help you. Nor would MER work at extremely high lattitudes due to sun angles.

I suspect there are targets though, where the "sweet spot" is fairly small, either because the target of interest is small, or there are hazzards that could kill the rover on landing, say, a huge, sharp boulders spread every few hundred meters. Those obstacles would have taken a site out of the candidate list for 2004, but if you are sending 10, why not devote a few to targets where you have a only 50% chance of safe landing, but are of great scientific importance. Send two or three to the Holden North delta and hope one lands someplace trafficable, for example.

Balloons also sound like a great idea for surveying some serious area. Right now, that's probably the biggest need.