I think John Hawks' two posts (part I and part II) express a lot of the reservations I have with the "ancestry" testing industry. To me, they resemble nothing more than online IQ tests where you pay $20 to get confirmation about what you already know! The problem, as I have said before, is that people want to know in a gestalt manner the crux of their ancestral narrative, when in reality they are a phenotype generated from discrete bits of genetic material with varied phylogenetic histories, and most of what they want to know they know (take a look at yourself in the mirror). Of course, the autosomal tests are better than the Y or mtDNA tests, and I can understand why African Americans would get some psychological satisfaction about knowing the regions where their male or female lineage are common. But for everyone else, I say stick to old school genealogy. Or else, these results should come with a warning label: for entertainment purposes only!