The most important nail was hammered in by this article — a symbolic slap across the face to anyone who uses mtDNA, and mtDNA only, to study demography. An upcoming article in BMC Biology lays down another nail:
A study published today in the open access journal BMC Biology reveals that fewer than 10% of African American mitochondrial DNA sequences analysed can be matched to mitochondrial DNA from one single African ethnic group. There has been a growing interest in the use of mitochondrial DNA to trace maternal ancestries, and several companies now offer to analyse individuals’ mitochondrial DNA sequences to obtain information about their origins. The current study suggests that only one in nine African Americans may be able to find clues about where their ancestors came from, in their mitochondrial DNA.
I can’t find the actual article, so I’m only able to go by the press release.
Note: I am not opposed to people using mtDNA to make inferences regarding demographic history. I am opposed to them using only mtDNA to make those inferences. I am opposed to this practice for the same reason as I am opposed to people using any single marker to study demography. Demography affects the entire genome, so we should use as many markers as possible when investigating demographic history. One marker is not as many as possible.