to events that have caused wild fluctuations after Obama became the presumptive nominee, despite being stable throughout the primary season.  To what could Gallup, assuming that it's not just random statistical noise, attribute this bounce?

1) Gramm gaffe on economcy

2. McCain gaffe on Viagra/birth control
   
3. Obama/Hillary in NYC
   
4. Jackson turning Obama into the candidate of personal and parental responsibility

Rasmussen Reports daily Presidential Tracking Poll for Friday is reporting the opposite, showing Obama's lead over McCain shrinking to 1 point, 43% to 42%, the first time his support has fallen below the 45% level since May 31. With leaners, it's 47% - 45%, a low point for Obama and a high point for McCain.

The polls have been, and continue to be, all over the place, yes.

http://www.rasmussenreports.com/public_c ontent/politics/election_20082/2008_pres idential_election/daily_presidential_tra cking_poll

Yeeehhhaw!!

I'd like to see Obama break 50 before the convention.  

In my view, you should ignore individual polls and look at the poll averages at pollster.com and realclearpolitics.com.

The only reason to look at individual polls is to look at the internal data -- breakdowns of specific questions other than preference and demographics.

but after the conventions is when polls begin to have real veracity.

As of late I've heard both Peter Hart and Zogby talk about the difficult in obtaining reliable and predictive sampling. The problem is of course that people habits/patterns with telephony have changed rather dramatically over the past decade. Voicemail and cell phones ( which can not be called for polling purposes) have create an incomplete picture of the electorate.

The honest fact is that polls are all over the place because no one really knows how to reach people. The exit polls were very illuminating because they were face to face with actual live voters, who had in fact, just voted.

From now until November pollsters are really flying blind.
Perhaps a better way to judge is the flow of contributions to the candidates. It was certainly an accurate predictor of the dem primary.

And as we all know, the national numbers are useless since Barack can run up huge margins in some states and lose marginally in swing states, with the net result of a popular win and an electoral loss.

This poll combined data from July 8-10. The McCain and Gramm gaffes are too recent to have affected the results. The birth control gaffe and the 'whiners' gaffe only came out yesterday. Most people still haven't heard about them and the ones that have, other than those like us who follow closely, didn't learn about it until the evening news. It's doubtful any of the things you mentioned affected this poll. On top of that, these 'swings' in poll numbers are so small that they are meaningless. A one point increase for Obama and a one point decrease for McCain adds 2 points to the gap, but is well within the MOE. The only thing you can take from them is that Obama has led by a small margin consistently. That's encouraging, but hardly anything to get excited about.