It would probably be best for John [McCain] if there were still three potentially viable opponents splitting up the Florida pie. ~John Weaver 
But even if that vote were split just two ways, Thompson wasn’t drawing that much support in Florida anyway, so the gain for either candidate would be minimal. If Thompson supporters in Florida are anything like his supporters in South Carolina , more will break for McCain and Huckabee than for Romney. Working even more to McCain’s advantage, Huckabee is reducing his presence in Florida, which may not bode ill for his campaign if he can hang on for two more weeks to those strong leads in Georgia, Alabama, Oklahoma and the like. But whatever happens to Huckabee later, some of his supporters in Florida will probably drift to McCain, while only a few will go to Romney. Romney seems to have a marginal advantage among Thompson supporters, and no advantage among Huckabee supporters. McCain stands to expand his lead over the field during the next week, and there is every reason to assume that weak Giuliani supporters will decide to back a similar candidate who already has won a couple of primaries. Romney will gain strength, but he won’t be able to gain as much as quickly as McCain. The remarkable thing about all of this is that reporters and pundits have assumed, as have I, that Thompson’s supporters were obvious Romney voters, but nearly two-thirds (at least in South Carolina) were apparently more interested in other candidates. That doesn’t just reflect Romney’s last-minute retreat from the state, but hints at a deeper resistance to Romney’s candidacy.
Of course, based on my track record of making predictions about this race, you can almost certainly ignore all of that.
P.S. Nationally, if I’m reading this right, Thompson supporters  seem most likely to favour McCain and Giuliani, but they have the profile of a Huckabee voter. (Don’t ask me to explain it!)