1500-Year Tree


Herriott Surname DNA Project

A 1,500-Year Tree for All Our Family Lines

By Scott Herriott

December 2016

Our Herriott Surname DNA Project began in 2007 when we subscribed to the Family Tree DNA service. We had done a fantastic job of filling out the tree of the descendants of David Herriot, who landed in New Jersey in 1685. Our historical research had met a dead end, as we were unable to find any reliable information about David Herriot's family in Scotland. We hoped that we might turn up some related DNA in the UK. But as we got DNA from more Herriotts, Harriotts, Herriots, Harriots and Heriots in the US and UK, we encountered two disappointments. First, we have not yet found any "New Jersey" DNA in the UK which might give us clues about the ancestry of our immigrant. Second, we have hoped to find out how our New Jersey branch is related to George Heriot, the jeweler to King James VI, who founded the famous Heriot School in Edinburgh. We are continuing these efforts.

As more people had DNA tests done, we began to identify many different genetic "lines" that carried the surname with one or another spelling. Even from the simple tests we use, geneticists can tell which broad group of humans a male belongs to. (Our DNA tests give information about the Y-chromosome, which is passed down from father to son.) To date, in our Surname Project, we have four groups of people carrying a Northern European (Scandinavian) DNA and 8 or 9 that have a Western Atlantic (Scots) DNA. The New Jersey group is Scots, as is the group of Heriots who trace their origin back to Longniddry, Scotland, whom we believe to be related to George Heriot. The puzzle all along has been whether these Scots lines descended from a common ancestor who lived some time after 1100AD, when surnames began to come into use, or whether their common ancestor predates the use of surnames, so their current family names are only coincidentally "Herriott".

For the first 10 years of our project, we were stumped by the differences among the DNA tests from our various "lines." The DNA test reports used for this 1,500-Year Tree consists of 25 numbers, each of which is a count of the repetitions of a small segment of DNA at specific sites on the Y-chromosome. See the table. Geneticists use abbreviated names for these sites. For simplicity, I call them Y01 to Y25 in the order they are shown on a DNA report.

Mutations of the DNA in the passage from father to son are rare, but if they happen, the result is to add or subtract one repetition at a particular location, so the count goes up by 1 or down by 1 at that location in the son's DNA report. On average, among all 25 sites, such a change occurs roughly once every 10 generations down a line of descent from an ancestor. When a change happens to one brother but not to another, their two lines start to look different, so the lines diverge and begin to change independently. As a result, the descending "cousins" will accumulate one difference (a change in the one or the other) about once every 5 generations. Such a change of DNA over time accounts for the differences between the New Jersey DNA report and the Heriot report shown in the table at right. There are 7 sites with a difference of 1 and one site (Y21) having a difference of 2, so the total "distance" between these two reports is 9. As a very rough guess, one would estimate that these two lines have been evolving separately for about 45 generations, or roughly 1,200 years, back to a time before the advent of surnames.

In the past, we calculated distance measures like that for each pair of "lines," the major groups in our Surname Project, but distance alone does not show which changes to the DNA report took place early on and which are more recent. So we could not know how each line is related to the others, as in a family tree.

We now have a method to answer to solve that problem. Using a large database of DNA reports available through the Family Tree DNA service, we were able to compare a New Jersey Herriott (me) with many Scots having similar DNA reports (differing by up to 6 changes). All of them had surnames other than Herriott. The technical analysis is described in a much longer report (email me: srherriott@gmail.com), but the procedure allowed us to see the sequence in which the DNA changes had taken place in the descent from the most recent common ancestor (MRCA) of all of the Scots who were within a distance of 6 changes from my DNA report (the Scots in my data "file").

Doing the same type of analysis with all the other groups in our Herriott Surname DNA project, including the Longniddry Heriots, I was able to sequence their changes and thereby construct a genetic family tree that reveals how each of our lines is related to each other and to a well-known, ancient DNA pattern called the Scots Modal, which geneticists believe to represent Colla Uais, progenitor of the Dalriadic royal house, who died circa 330 AD. See the chart.

The Surname DNA Journal is interested to publish the results of our work, because no one else has used Y-DNA tests to create such a family tree. Genetic genealogists use different type of test, called a SNP test ("snip" for single nucleotide polymorphism), to identify the branches in humanity's genetic family tree as it evolved do some SNP tests to identify exactly how our ancestral lines are positioned in relation to the Scots Modal. The article I would like to write for them will describe our use of the Y-DNA test to get a more accurate picture of changes over just the past 1,500 years, which enabled us to create a tree that unifies the 9 lines of our Surname Project participants (the "leaf" boxes in the tree below, counting the two New York (NY) boxes as one line).

The main conclusion is that our various surname lines have a common ancestor who predates the use of surnames. Only the two "New York" lines are likely to be from a common "Herriott". Whenever we downloaded Scots DNA reports that were even a short distance away from one of our "Herriott" reports, the file consisted entirely of non-Herriott surnames, except for a few whom we knew from genealogical research to be related. If we can do the SNP tests suggested by the Surname DNA Journal, we will have more results to report in the June 2017 issue of the Herald.