Vincenzo stumped me with a question about how you distinguish vicariance (what I termed the dumb bell model) from dispersal (the peripheral isolate model) as mechanisms for allopatric speciation using phylogenetic data.
It turns out the answer is quite simple but that I'd been thinking too specifically. We just need to think outside the box for a moment and consider using phylogenies for groups other than the one we might be specifically studying.
A vicariance event will affect many different taxa and leave similarities in their cladograms. For example the breakup of Gondwana caused deep divisions in the cladograms of many groups representing the now separated continents. These similar cladograms are 'congruent'. Lack of congruence in cladograms is better explained by dispersal. Obviously this only works on a larger scale. I thought this page explained it all quite clearly.