Just want to draw attention to a new article on estimating the age of R1b haplotypes using *just STR values* - the synopsis by Dienekes is exactly the same criticism that I leveled at attempts on this list to do the same for R-L21 M222+, without reference to SNPs, just using STR values: *It is also probably consistent with the idea that Y-STR based estimates are > suspect because of deviations from the linear model.* Like I previously mentioned, STR values increase over time, but importantly decrease over time too, and not in a linear fashion. So an STR that has been steadily increasing from it's upstream value, can suddenly decrease. When I look at my own STR values, and my closest matching Dunn(s) families, who share a number of Off-Modal values with me, such as DYS449=32 and CDYb=41, I'm struck by the fact that my DYS391=10, and theirs is DYS391=11. DYS391 is a slow moving marker, let's say for arguments sake its mutation rate is once every 1,000 years, well it would seem that such a mutation occurred in my line recently, after the other Off-Modals occurred. So essentially lineages go on a statistical "random walk", and one lineage can end up looking similar, and another different from a common founder lineage, in a non-linear fashion. So extracting the correct aging is fraught with difficulty. As the 1,000 Genomes project and others have shown recently, there are many SNPs out there, and many more to be discovered, which are invaluable, as they do follow a linear pattern, and exact branching of lineages can be determined by their use. http://dienekes.blogspot.com/2011/08/back-to-drawing-board-for-r-m269-busby.html Cheers, Paul