GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2006-11 > 1162503133


From: "Ken Nordtvedt" <>
Subject: Re: [DNA] Neolithic J2 and E3b in Britain? Maybe not.
Date: Thu, 2 Nov 2006 14:32:13 -0700
References: <3b2a446a0611010918j45594e12r28817e9f8821f252@mail.gmail.com><042401c6fe97$5ce8b830$6401a8c0@Precision360><3b2a446a0611020852u457ad5e8n68b92b2596ec2566@mail.gmail.com><454A57CB.3070208@sbcglobal.net><3b2a446a0611021318r31c7979bq4b4ccffcda6081a9@mail.gmail.com>


Suppose the Roma population was cut in half, but still numbered in the
thousands or tens of thousands after. How much variance do you think is
lost? Epsilon! To really lose variance you need to collapse to close to 1;
if you are left with two, for example, the variance could equally well have
increased drastically. Population collapse increases the variance of the
variance; it does not necessarily diminish the variance. In any case, such
events don't produce a systematic "effective" rate without knowing details
of what actually happened.

After you get away from small population numbers ---- 1,2,3.... 10,
population variance does not care about population size. The defined
variance is a normalized attribute which quotes variance in something akin
to "per person" units. Geographical expansion can promote development of
subclades within a population, but how does it promote health for the
variance of the overall population?

Needless to say, I consider the fudge factor a terrible chapter in the
field. When something is not understood yet, that's how it should be
presented.

If a different "effective rate" can be invoked for each applied case, it is
not an "effective rate"; it is a fudge factor rather than a phenomenological
law.

Ken



----- Original Message -----
From: "Sasson Margaliot" <>

> Zhivotovsky et all (2004) are using two very unusual populations, one in
> Europe that was nearly wiped out in forties (reducing variance), and
> another
> in Pacific, that was in total isolation (mthematitians proved that
> geographical expansion is healthy for preservation of variance).
>
> Therefore, it is not surprising that these two populations indeed DO have
> the "fudge factor" of 3.
>
> But when you look at lineage like J2 that in few thousand years expanded
> from Iberia to India, why would their "fudge factor" be anything but 1.0?


This thread: