The Election Day tracking polls still show a statistical dead heat and Romney's white share at 56%-59%, at the high end of the mediocre post-Reagan range. Remember, George W. Bush won with 58% of white votes in 2004; John McCain lost with 55% in 2008. (More comparisons here). But as immigration (which both supported) is slowly tipping the electorate, this may no longer be enough.

• Politico/ GWU Bettleground Tracking Poll (Nov. 6) again shows Romney and Obama tied, 47-47.

Romney's white share (on Page 86!!): 56%, for an eighteen-point lead over Obama.

• Rasmussen Reports (Nov. 6) has Romney still a point ahead Obama, 49-48. This poll has mostly shown narrow within-the-margin-of-error Romney leads for the last month

Platinum Subscribers to Rasmussen can view the crosstabs. Romney's white share: still 58%, so his lead over Obama among whites remains 18 points.

• IBD/TIPP, returning after a Hurricane Sandy break, shows (Nov. 6) Romney trailing Obama 50.-48.7. This poll was long an outlier showing Obama with a statistically significant lead; its latest result is a statiscal dead heat.

Very unusually, the IBD/TIPP commentary mentions the white share—because Obama has reached 40% for the first time. IBD/TIPP describes this as "a critical threshold he had to hit in order to make a win possible."

This is an odd remark, given Obama's vanishing overall lead—and the fact that Romney's white share has also increased to 59%, for a nineteen-point lead, both among his better performances.