Here's a cool article for those of you who ever wondered how researchers make such narrow estimates of the dates and duration of human migration patterns. A new methodological development made at the University of Leeds makes it possible to do more accurate dating of ancient human migration patterns, which has been a constant problem for researchers.
The most common method for dating such patterns is to analyze mutations in samples of mitochondrial DNA (mtDNA) in order to find the last common ancestor of any set of lineages. This method has problems, however, as natural selection affects the number of mutations observed in a family tree in phasing out deleterious ones, thus skewing any dating results from the mtDNA.
The new method, developed by Pedro Soares, utilizes a formula developed with the other researchers at Leeds to correct for this effect of natural selection in dating. Implementation of this formula has already yielded some surprising results -- humanity's out-of-Africa migration has been dated with this method to 60-70,000 years ago, 10-20,000 years earlier than the 50,000 years held by popular belief.
This method has a lot of potential for future research on ancient human migration; the improved accuracy means that dating of migration could be done without supporting archaeological evidence. Furthermore, there are many ongoing debates regarding when and how different populations settled, like my previous post on whether or not Native American ancestors migrated to the Americas as a single population or as multiple. Such debates now have a chance of being put to rest.