I think this is a pretty confused discussion of Hispanic voting patterns from Lamar Smith:
Exit polls reported by CNN and updated this week reveal that a historically robust 38 percent of Hispanic voters cast ballots for House Republican candidates in 2010 – more than in 2006 (30 percent) and 2008 (29 percent). In fact, since 1984, Republican House candidates have only won a higher percentage of the Hispanic vote in one election: 2004. This level of Hispanic support for Republican candidates came despite widespread pre-election claims by advocates for illegal immigration that the Arizona law and a pro-rule-of-law stand would undercut Hispanic support for Republicans.
All we’re seeing here is that Latino support for the GOP moves up and down, just as white support does. 2010 was a better year for Republicans among all demographic sub-groups, and Hispanics are no exception. But it’s possible to look at the evolution of the white/Latino voting gap over time. In 2004, when I don’t think there was a clear partisan difference on immigration, Republican House candidates did 13 percentage points better with white voters than with Latino voters. By 2006, it was a 21 point gap. In 2008 that expanded to a 24 point gap. And in 2010 it was a 22 point gap. The hypothesis that the creation of this gap was driven by the opening up of a partisan difference on immigration strikes me as plausible and at any rate isn’t debunked by merely citing the fact that the overall level of Latino support for the GOP fluctuates with general political conditions.