Texas

Harris County (Houston and many of its suburbs): 2004: Kerry 44.64%, Bush 54.87%, 2000: Gore 42.9%, Bush 54.3%, Nader 2.4% (so, factoring in Nader, this county bascially stayed the same from 2000. If anything, Bush made a tiny gain here)

Dallas County (Dallas and many of its suburbs (? - Texas folks, correct me if I'm wrong): 2004: Kerry 49.03%, Bush 50.44%, 2000: Gore 44.9%, Bush 52.6%, Nader 1.9% (this represent a significant and (to me) surprising gain for Kerry on Gore's total)

Travis County (Austin and many of its suburbs (again, correct me if I'm wrong): 2004: Kerry 56.31%, Bush 42.24%, 2000: Gore 41.7%, Bush 46.90%, Nader 10.4% (now clearly, Kerry 15% gain from Gore's total came in large measure from Nader voters. But he also gained substantially from 2000 Bush voters as well. This county represents a fascinating and substantial swing, and mirrors the gains Kerry made in Ohio's Franklin County (home of Columbus))

Bexar County (San Antonio and some of its suburbs): 2004: Kerry 44.49%, Bush 54.49%, 2000: Gore 44.9%, Bush 52.2%, Nader 2.4% (now here, we see a small - 2 to 3% - Bush gain from 2000. The Latino factor at work? I would venture that this is the case)

Colorado

Denver County (the city of Denver and some suburbs (?)): 2004: Kerry 69.51%, Bush 29.41%, 2000: Gore 61.9%, Bush 30.9%, Nader 5.9% (now Nader voters were clearly a factor, but Kerry also made some decent gains on Gore's 2000 totals at Bush's expense - interesting)

Adams County (Denver suburbs and exurbs(?)): 2004: Kerry 48.39%, Bush 50.4%, 2000: Gore 44.1%, Bush 50.2%, Nader 3.9% (Basically the same results, factoring in Nader)

Arapahoe County (Denver suburbs and exurbs(?)): 2004: Kerry 47.29%, Bush 51.66%, 2000: Gore 43.5%, Bush 51.5%, Nader 3.9% (another unchanged county, factoring in Nader)

Jefferson County (Denver suburbs and exurbs(?) - contains Littleton, CO): 2004: Kerry 46.59%, Bush 51.9%, 2000: Gore 42.9%, Bush 51.0%, Nader 4.4% (at best, this represents a very slight gain for Bush - assuming everyone of Nader's 2000 voters voted Kerry in 2004. But basically unchanged)

El Paso County (Colorado Springs and environs - home of Focus on the Family (Dobson's outfit) and the US Air Force - known as a bastion of conservatism): 2004: Kerry 32.04%, Bush 66.82%, 2000: Gore 30.8%, Bush 63.9%, Nader 3.5%, Buchanan and the Constitution Party candidate 1.8% (so Bush made a small gain here - probably not surprising - but if Bush isn't making large gains in places like El Paso county, where in "urban" America would he be making such gains?)

Larimer County (Fort Collins and environs): 2004: Kerry 46.58%, Bush 51.87%, 2000: Gore 38.9%, Bush 52.7%, Nader 6.9% (Fort Collins is not exactly what one would think of when one pictures cosmopolitan blue state stereotypes. But even taking into account Nader, Kerry made a small, but not insignificant gain on Gore's 2000 performance here. Again, if Bush isn't making his "urban" gains in Fort Collins, where is he making them?)

North Carolina (note: Nader was not on the North Carolina ballot in 2000)

Mecklenburg County (Charlotte city and its exurban and suburban environs): 2004: Kerry 51.85%, Bush 47.85, 2000: Gore 48.2%, Bush 51.0% (again, this is a very solid gain for Kerry at Bush's expense. What went right for the Dems in Charlotte? They should find out. The "ideopolis" thesis at work? Perhaps.)

Wake County (Raleigh city and some suburbs, typically the most conservative part of the "research triangle area): 2004: Kerry 48.35%, Bush 51.26%, 2000: Gore 46.0%, Bush 53.1% (another decent gain for Kerry on Gore's 2000 total, if not as big as in Charlotte - but definitely worth noting nonetheless)

Durham County (Durham city and some suburbs (? - North Carolina folks help me out): 2004: Kerry 67.90%, Bush 31.75%, 2000: Gore 63.7%, Bush 35.6% (yet another substantial gain for Kerry on Gore's 2000 totals in a "red state" city)

Orange County (Chapel Hill - home of UNC's flag ship campus - and environs): 2004: Kerry  67.03%, Bush 32.46%, 2000: Gore 62.7%, Bush 36.3% (I'm starting to see a pattern in North Carolina . . . )

Georgia (now, here is a state where Bush made a significant overall improvement from his 2000 win - but he didn't do it in Atlanta)

DeKalb County (eastern Atlanta suburbs): 2004: Kerry 72.71%, Bush 26.65%, 2000: Gore 70.2%, Bush 26.7% (with Nader considered, this represents a very slight Kerry gain on Gore's total in 2000)

Fulton County (Atlanta city and some suburbs): 2004: Kerry 60.16%, Bush 39.18%, 2000: Gore 57.8%, Bush 39.8% (even with Nader considered, this represents a small Kerry improvement on 2000)

Cobb County (conservative western Atlanta suburbs and exurbs - Newt Gingrich's former district): 2004: Kerry 37.18%, Bush 62.09%, 2000: Gore 36.9%, Bush 59.8% (so a small - 2% or so - Bush gain on his 2000 total)

Gwinnett County (conservative northern Atlanta exurbs): Kerry 33.46%, Bush 65.82%, 2000: Gore 32.1%, Bush 63.7% (so another small Bush gain - 1 to 2%. Not surprising, but one would perhaps that Kerry would have done worse in a county like this. I'm somewhat surprised he didn't)

Missouri

St.Louis City (the city of St. Louis, surprisingly enough): 2004: Kerry 80.39%, Bush 19.3%, 2000: Gore 77.4%, Bush 19.9% (even factoring in Nader, this represents a small Kerry gain on Gore's 2000 totals)

St. Louis County (much of St. Louis suburbia): 2004: Kerry 54.38%, Bush 45.15%, 2000: Gore 51.5%, Bush 46.1% (this one is surprising to me. Kerry clearly made gains here at Bush's expense. At least 1% of his additional support came from Bush's 2000 totals. And this is a very suburban county)

Jackson County (the Missouri part of Kansas City and some suburbs): 2004: Kerry 62.48%, Bush 37.13%, 2000: Gore 59%, Bush 38.4% (again, this is another Kerry gain that came in part at Bush's expense. I would not have expected this before looking at the data)

SO: What can be concluded from my investigation so far?

1.    That Bush's increased totals came from three sources: 9/11 "belt voters" in suburban New Jersey, Connecticut and New York, some gains amongst Latinos (but not everywhere - and not in Florida), but most importantly, amongst small town and rural folks
2.    The most important divide in American politics today is between Democratic cities and "inner" suburbs and Republican exurbs and rural areas. It is not really the red state/blue state divide that is most important.
3.    That Ruy Teixeiria's ideopolis thesis is real, but will not - in the short term at least - be enough for the Democrats to be a majority party.

I will present further analysis of metropolitan voting patterns in later diaries. To be continued.