37-marker DNA

DNA analysis offers a new approach to genealogy. Men get their Y chromosome from their fathers, and  – were it not for mutations  – all male descendants of a given male ancestor would have the same Y "markers." Although the mutations are unpredictable, they occur at a reasonably predictable rate. So, statistically, the degree of agreement between the Y chromosomes of two men says how far back their common ancestor was.

As of October 2008, more than 110 male Montgomerys have had their DNA analyzed, and most have agreed to share the results in a surname project. The results of my DNA 37-marker analysis show near matches that I would like to use in my work in genealogy. Though no one matches my DNA exactly on 37 markers, there are three other Montgomery males with "genetic distance" (g.d.) of 1 and 16 others with g.d. of 2 measured with respect to me. [G.d. is essentially the number of mutations. See the footnote.] Two of the DNA participants unfortunately have not shared further information about their lines. The following summary, then, uses the 17 for whom I have genealogical information. I identify participants only by codes in order to protect their privacy. After some lengthy description, I will get to conclusions and applications. In addition, other DNA analysis leads to findings on "haplogroups." Although they do not help the immediate paper trail, I will report on these findings as a matter of some interest.

The table shows the 17 testers in two major groups, those with g.d. of 1 and of 2.

G.d. M code Most distant known ancestor
selfM 0038John PA>NC 1773 d 1795 son Samuel b ? 1761 >TN
1M 88/94Rev William b Cumberland PA 1768 >NC>GA d MS1848
1M 0028Elias b ? 1775 >NY d ? 1846
2M 0059William b Antrim Ulster?>ME 1718>MA d MA1731
2M 0066Samuel b Antrim Ulster 1745 d Ulster 1827
2M 0080James b TN 1805 d Pontotoc Co MS 1856
2 CBMJohn b ca 1703 Armagh Ireland
2M 0031Robert b Milton PA 1811 d Keokuk IA 1880
2M 0110Robert b Milton PA 1811 d Keokuk IA 1880
2M 0039John b Ulster 1718-1720 >PA d VA 1795
2M 0042James Hamilton b TN 1835 >AR>NV>CA>OR d TX 1915
2M 0067Luther High b ? 1855 d Sewanee TN 1926
2M 0086Hugh b VA1786 d TN 1860 m Rebecca Overton KY
2M 0113John b VA 1775 >TN>AL d Lawrence Co AL 185?
2M 0085Joseph b Waxhaws SC 1766 >MS 1802 d Concordia LA
2JMJohn b TN 1836
2M 0029John d SC? 1784 son William >GA 1798 d GA1818

In the g.d.-of-1 group, M-94 is the father of M-88. So we get no additional information on ancestry from that duplication; their most distant known is Rev. William Montgomery. For online elaboration about Rev. William, see Foote's book, Sketches of North Carolina, Historical and Biographical, Illustrative of the Principles of a Portion of Her Early Settlers.) The other with g.d. of 1 is M-28, whose most distant known is Elias Montgomery, associated with New York. Neither of these lines goes back earlier than the time of my John; so it is hard to apply their paper trails to my work. I tried to use the Cumberland connection in searching for my John in Pennsylvania without success.

For the 14 with g.d. of 2, 13 have identical 37 markers; their likelihood of a common ancestor within the last 7 generations is 95%. The 14th is M-29, whose most-distant John died in SC in 1784 (a contemporary of my John). Four of the project participants have paper documentation going back earlier than do I, and all four believe their ancestor was from Ulster. Three of these have identified Antrim as the source.

The probability of my having an ancestor in common with any of these participants is affected by the assumption that he is not my John nor is he a descendant of John; that is, he does not lie within 6 generations of me. The following table (generated with the Family Tree DNA calculator) shows the probability of having the common ancestor within N generations, using the observation that he is not within 6 generations.
N g.d.=1
(13 identical on 37)
7 27.1% 19.7 19.5
8 47.5% 36.8 36.5
9 67.6% 51.2 50.8
10 73.6% 62.9 62.5

For example, since my John is 6 generations back, there is a 47.5% probability that his grandfather (generation 8) is also in the line of M-88/94/28. To get above the "50-50" level for g.d. of 2, we have to go back 9 generations, my John's great-grandfather. There seems to be a fair chance that my John was no more than 3 generations away from Ulster, probably County Antrim.

Note that all g.d.-2 testers are not the same. I and my g.d.-1 cousins differ from the "13 identical on 37" on markers 10, 11, and 12; whereas M-29 matches us on those three markers. His line may have "peeled off" from the common line of the other 13 at the same generation as mine. (His, of course, mutated further.) This scenario would seem to require at least two generations between my John and the common ancestor, one to split off the 10-11-12 markers (then to beget the lines of M-88/94/28) and another to account for the other mutations. That is, the common ancestor with respect to the g.d.=2 group can be no closer than John's grandfather.

More detail of the project and updated results are shown on the website, www.familytreedna.com/public/Montgomery/.

After a long period of quiet, email in November 2010 raised some dust. I have now been in contact with a man having the surname of Ball who believed himself to descend from Edward Ball (c1642-1724) and Abigail Blatchley of Newark NJ. His DNA test showed that we differ by only 1 on the 37 markers. That is, he is as close to me genetically as M-88/94/28, with all the implications as above. Descendants of that Edward Ball (of Newark NJ) migrated through Pennsylvania and then down into the southern states, but a few went north to Canada. The social ancestry of the Ball whose DNA is so close lies in Canada. (The main file on the Ball line is at www.newenglandballproject.com/index.htm.) Analysis of this new connection is complicated beyond the problems of M-88/94/28 by the lack of knowledge of when the Ball/Montgomery line changed from social to biological. A Pennsylvania connection seems possible, but I have not been able to make use of this additional information.

A couple of results are enough to make the DNA work worth it to me: The analysis implies that, in my line, the father of record in all the generations back (surnamed Montgomery) was indeed the biological father. As a second significant note, there is a Montgomery line that ended up in Blount County, close to the Knox-County home of my Samuel (son of John), that I have been running across for years (identified by the couple married in Pennsylvania, John Montgomery and Esther Houston). It has been tempting to say, as a number of well-meaning observers have said, there "must" be a close connection between Samuel and this line. Two of the participants in the DNA analysis, M-0113 and 0039, however, are in that line, and the chromosomes say that Samuel was not all that close: my g.d. from that line is 2.

Every male can be placed in a haplogroup. A haplogroup is a group of identical Y-chromosomes that share a very distant common ancestor in whom a particular mutation occurred. The mutation that defines group J2, in particular, is thought to have occurred in Mesopotamia about 18,500 years ago. The defining characteristics of each haplogroup mutate very rarely, and analysis over time and location of the characteristics in males allows the tracking of ancient migrations. The J2 group spread in all directions, including westward, from the "fertile crescent." In Europe today, J2 reaches its highest frequency in Greece. The frequency of J2 in the European population decreases as the sample moves north.

To bring us back to the Montgomery ancestry, the Montgomery name goes back to Roger de Montgomery (called "The Great") of Normandy. He was father to another Roger, born about 1030 who assisted William the Conqueror in the invasion of England in 1066. Analysis now concludes that these Montgomerys were in haplogroup J2.

The Montgomery Surname DNA Project has identified five different major haplogroups to date, with a handful of different haplogroups outside the five main groupings. Our surname has had plenty of opportunity to spread over the centuries since it first came into use. Beyond biological heritage, some Montgomerys got that surname through adoption of the surname (In England the adoption of a surname was not compulsory until the 13th century.), adoption of children, and illegitimacy.

The remarkable point is that FamilyTreeDNA confirms me as being in haplogroup J2. So despite historic opportunities to the contrary, I am a biological descendent of the "original" Montgomery family of Normandy.

The lack of matches limits the use of these results for me. If I were in the group of 13 with identical markers at the 37 level, I would be eager to extend the analysis to 67 (as most of them have). With no identical matches on 37 markers and the tenuous connections described above for g.d. of 1 and 2, the value of the DNA analysis to me is suggesting where to look. Basically, I am hoping that first cousins of my John are in the lines identified. If the connection is closer, then great.

I tried to use the implications of the DNA work accordingly in the search for John. See that search particularly regarding other John Montgomerys.

Footnote: To first approximation, genetic distance is the number of makers that are different, but if certain pairs of markers - known to change at the same time - differ, the pair counts as only one in calculating g.d. In comparing my markers with the latter 11 participants, there are three differences. The mismatch on two markers, however, counts as only one since they are not independent. That is, these 11 participants have a "genetic distance" of 2, not 3.

Revised January 2011

[To the top][ To the genealogy page][ To my home page>]