All the tracking polls yesterday showed Dean, at a minimum, stopping his decline and in most cases making up some lost ground. Now the movement seems clear. Tonight's ARG poll has Kerry 38%, Dean 20%, Edwards 16%, Clark 15%.

(ARG now also has tracking polls out for South Carolina, Arizona and Oklahoma.)

I heard Dean on the radio tonight as I was driving back to my hotel. And he sounded very much at ease and commanding -- frank, smart, quick on his feet. Earlier today, I was wondering just where he'd have to place on Tuesday to win in the media's expectations game. And it seemed to me that if Dean could manage a convincing second -- that is, with real distance between him and the third place finisher -- that he could play the comeback kid angle. He could argue, with some merit, that he took a huge hit, fell dramatically in the polls but was then able to fight back into contention by corralling a lot of doubting supporters back into the fold.

The problem for Dean is that none of the February 3rd contests strike me as natural Dean states -- with the possible exception of New Mexico. And a victory for Kerry in New Hamsphire would still mean that Dean had failed to win in two states which seem to play heavily to his strengths.

One more point: I haven't had a chance to write about it yet. But I was struck by the iffy advance work for the Clark event that I went to on Friday, and what I've heard about a few others. (The rally on Sunday was much, much better.) For all the money Clark's raised and the polished Internet presence, this is still a campaign that was cobbled together quickly and then had significant internal shake-ups in its first couple months. I don't want to judge on limited evidence -- which mine very much is. But it's just left me wondering whether it might be a sign of a broader problem.