[a0-prw]

I had never heard of "cumulative Bayesian skyline plots". A search on Google gave one hit, which was a dissertation quoting this study.

I assume they mean "coalescent Bayesian skyline plots".

Why invent a new name for something which already has a more informative name?

I have great difficulty imagining a culture which could socially sustain an N_e of 1:17 (male to female). Just do a few sample calculations. The priviledged males would be massacred.

A few things could explain it: (1) large scale male infanticide or male child castration (2) A disease which kills mainly male children.

Both of these alternatives suffer from one of the same problems as their proposal: How would the cause (whatever it was) spread so relatively quickly among distantly seperated populations ?

I think it far more likely they have made mistakes somewhere.

[Reubend]

I agree. 10,000 years ago it would have been difficult for majors trends to spread across different continents.

And I have a guess (though unsupported) that every civilization's mythology would tell similar stories of this event if it really were occurring globally.