Archiver > GENEALOGY-DNA > 2003-10 > 1065235553

From: "William Hurst" <>
Subject: [DNA] Practical mtDNA test, or Kelly sisters project
Date: Fri, 3 Oct 2003 22:46:11 -0400

Hi all,

Back on May 20, I proposed a practical test of mitochondrial DNA to try to
prove that two women named Kelly were sisters. Martha Kelly was born in 1803
and Catherine Kelly (my ancestor) was born in 1807, probably both in Wythe
County, Virginia. The original message is in the Archives at:

Having come up with the idea, I had to find direct maternal descendants of
the two women. For Catherine, any of the children of my father's sisters,
and their daughters' children, would qualify. I finally got a volunteer. For
Martha, I had an obvious candidate, since one of her descendants had
suggested the relationship of the two women back in 1983. It helped in
devising this test that daughters of the two women were the first and second
wives of my great grandfather. So my two test candidates were already
half-second cousins, once removed - but have never met.

Finding candidates was one step, getting them to agree to take the test was
another step, getting them to actually take the test was yet another step.
Then it was a matter of waiting for the results.

A few weeks ago, the first result came in from FamilyTreeDNA. The first
cousin had the following mtDNA:

Haplogroup H
HVR1 mutations: 16183C, 16189C, 16519C
HVR2 mutations: 263G, 315.1C

That result produced one low-resolution, or HVR1, match, who is apparently
is not related in any reasonable time frame. Ann Turner checked the Oxford
Ancestors database for me and found three matches on just the first two
numbers - OA not testing for 16519C or HVR2.

Today, the second cousin's results came in. His results were EXACTLY THE
SAME!! They are the only two with that exact haplotype in the FTDNA

Based on these results and on my study of the old census reports and other
data, I think it's safe to say that Catherine Kelly and Martha Kelly were
indeed sisters. Their father was John Kelly. Their mother was Elizabeth
Cummins - the first name from estate and census records and the last name
only from my grandmother's notebook.

Yes, I know the results would be the same if they were cousins instead of
sisters. For first cousins, that would have required that two men named
Kelly - perhaps brothers - married two Cummins sisters. However, John Kelly
was born in Ireland and Elizabeth Cummins was born in Virginia. Not exactly
two brothers and two sisters from the same little village having a double

Still not definite, but if I can find two women from a century ago, with no
documented relationship, then determine that their descendants have the same
rare DNA, then I can call them sisters if I wish.

Me, I now have a new set of 3rd great grandparents.

Bill Hurst

This thread: