On 06/07/2006, at 10:37 PM, David Hobby wrote:
Sorry, there is too reason to doubt their numbers. The above model sounds too simple and homogeneous. Even if such a model incorporates geography, it still doesn't do better than guesswork when it comes to the cross-contamination probabilities.
Doesn't it? I think the assumptions seem quite reasonable, even a little conservative.
All it takes is one region, somewhere in the world, with negligible cross-contamination probabilities. If this exists, people in the middle of it will not be descendants of Genghis Khan, Charlemagne, or whoever.
One small region that's managed to stay TOTALLY isolated for 10,000 years? That's extremely improbable, because firstly, nowhere is THAT remote (everywhere is contactable from *somewhere*), and second, the maths of evolutionary genetics is against you - while direct chromosomal inheritance goes down exponentially by generation, family tree goes up exponentially by generation (to within population limits). Or do you really think you had 2,147,483.648 *individual* ancestors 30 generations ago? No, of course not - family trees converge as well as diverge. Third, totally isolated populations tend to die out.
Charlie _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
