The Scottish Origenes McMillan Case Study
"Be particularly careful to take
into account anything that conflicts with your beliefs".
The belief that it is possible to pinpoint the location where one's ancestors first took their surnames more than 1000 years ago is the basis of case study reports produced and marketed by Scottish Origenes. The reports rely on YDNA test results which are used to identify surnames which match a test taker's surname. The 1841 census is used to identify farmers with these surnames. The location of the farmers is plotted and the area where the surnames cluster together is used to identify the surnames “genetic homeland”.
It should be noted that the McMillan case study relies entirely on the interpretation of 25 marker tests. Inferences regarding the time frame associated with tests with this resolution are notoriously unreliable and matches will often not appear in higher resolution tests. These issues are not addressed in this review.
.The "genetic homeland" case:
In the preamble of the McMillan study the following proposition is advanced with respect to the McMillan surname.
“By plotting the location of farmers called Chesney/McChesney, Glendining, Paxton and Elliott it reveals that these surnames are all associated with Southern Scotland and that the majority cluster close to a McMillan farming cluster centred upon Dumfriesshire...” This claim will be the basis of the following review.
The case study considered surnames from the YDNA test result, and a list of matching surnames derived from Y-search.org. Each name in the list was categorized as either Scottish, English, Irish or unknown. Turning our attention first to the YDNA list, and beginning with the closest match first:
Badger (36/37): There are no Badgers in the 1841 Scottish census. The Origenes case study does not consider this surname. We will return to the Badgers later.
Smith (35/37): The surname is not used in the study. Smith is a ubiquitous surname and would add nothing to the analysis.
Fedderly (23/25): The Fedderly surname is not found in in the 1841 Census of Scotland.
Chestnut (23/25): There are Chestnuts in the 1841 Scottish census, one in Paisley Renfrewshire, and five in Urray, Ross and Cromarty. They are well removed from the "genetic homeland" and are not considered in the Origenes analysis. The study claims Chestnut is a corruption or variant of McChesney/Chesney. No evidence is provided that would support a link between the two names. The Chesney's are found in Dumfriesshire but more heavily concentrated in Kirkudbrightshire and Wigtownshire. The surname is linked to a notable Norman family with srttlement in England as well as Scotland. The presumed "corruption" of the Chesney name into Chestnut is key to the identification of the Origenes "genetic homeland". The Chestnuts will also feature prominently in an alternative interruption of the results in the YDNA report.
Paxton(23/25): The Paxton surname is located in the borders of Scotland focused on the parish of Bunkle and Paxton, and near the village of Paxton in Berwickshire. The distribution is well removed from the McMillan "genetic homeland". There are no farmers listed among the 401 Paxtons found in the Scotttish census.
Turning our attention to surnames found in the Y-search.org list:
Elliot: The Elliot's are found in the borders of Scotland and strongly associated with it's clan territory in Roxburghshire. The Elliot's were notorious "Border Reivers" and their homes and farms were located close to the border between England and Scotland.
Glendining: The Glendining surname is also located in the borders of Scotland, principally in the county of Dumfriesshire. However the distribution does not overlap with the genetic homeland and their clan territory is found well to the east of the suggested "genetic homeland".
Jordan: The final name from the matched list used in the Origenes analysis is Jordan. The surname distribution is found in several locations within the borders and lowlands of Scotland. The Jordan surname is not used in the study but assumes rather that Jardine is a corruption of Jordan. Both surnames are Norman in origin and Jordan's can be found in England Ireland and Scotland. Jordan is derived from a Norman personal name while Jardine can be either an occupational or locative surname. The Jardine's are concentrated in Dumfriesshire south east of the "genetic homeland". The Jardine's fought with the English in the Battles of Stirling Bridge and Falkirk. The Clan would later support King Robert the Bruce at the Battle of Bannockburn in 1311. Their presence in Dumfriesshire is thought to date from the 14th century.
The preamble to the case study states: “one can plot where farmers with the surnames that appear in one’s Y-DNA results cluster and identify an area common to all...”
In this instance there is certainly no "area common to all". The genetic homeland identified in the McMillan study is predicated on the location of 2 Chesney farmers found among the 108 "related" lowland farmers which are distributed from Wigtownshire in the west, to Berwickshire in the east. None of the actual matching surnames provide any evidence in support of the genetic homeland's location in the borders of Scotland. Elliot clearly contradicts the study's conclusion as does the Paxton surname. In summary, none of the names from the test results provide any support for the location of a McMillan "genetic homeland".
Finally the claim by the study that: "the majority cluster close to a McMillan farming cluster centred upon Dumfriesshire" is very puzzling. To put this statement into perspective for a majority of related lowland farmers (53%) to cluster close to the McMillan farming cluster, a circle with a radius of 40 miles and an area of 5,000 sq miles would have to be drawn.
An alternatative interpretation:
In order to develop an alternative explaination we would need to relax two of the Origenes assumptions which are a built into its methodology.
1. That the names in the matching list were genetically linked prior to the adoption of surnames.
2. That farmers are the key to finding a "genetic homeland".
If these conditions are removed data from the YDNA table can be used to provide a reasonable alternative explanation of the findings in the test takers YDNA report.
The Ulster Analysis:
In 1609, following the ascension of James IV, I to the throne of England and Scotland, the newly united country began the process of plantation and settlement of Ulster with Lowland Scots and English. The consequence of this economic experiment was a mass migration of thousands of lowlanders, among them the McMillan's. If we were to assume that the matching surnames were the result of a series of non paternity events involving the surnames on the list, an alternative interpretation emerges. The following analysis is based on data extracted from the 1901 census of Ireland.
The Chestnut surname is found in the 1901 Census of Ireland. In light of Irelands population of 4,429,886 the Chestnut surname is extremely rare with a total frequency of 138 individuals. Despite their scarcity they are found in two distinct clusters, one in north west Antrim, the second just west of the shores of Lake Neage. In the Antrim cluster we find 69 Chestnuts, 33 McMillans and 3 Elliots within a five mile radius. Extending the search radius to 10 miles we can identify 100 Chestnuts and 47 McMillans.
In the second cluster adjacent to lake Neagh, we find 8 Chestnuts, 20 Jordans, 17 Mcmillans and 1 Paxton. However the most significant finding was the presence of 58 Badgers. Badger was the closest match(36/37) in the matching surname list. The search area had a nine mile radius.
Findings and Conclusion:
The likelihood of two extremely rare surnames appearing In precisely the same local in association with the McMillan surname, and then independently, 160 years later in the McMillan DNA test results, strongly suggests it was not simply a coincidence. It would have to defy astronomical odds if it were truly caused by chance. The Origenes case study assumed there was no spatial association with the Chestnut surname because it could find none in Scotland, therefore Chestnut was conveniently dismissed as a corruption of McChesney/Chesney. However the fact that the McMillan, Chestnut, Badger, Jordan, Elliot and Paxton surnames can be found in two clusters associated with the McMillan surname in Ireland calls the Origenes homeland conclusion completely into question. An Irish "way point" has dramatic implications in the search for a "genetic homeland" in Scotland. The McMillan ancestor could have arrived in Ireland from scores of possible locations in the Scottish lowlands. This information would have been helpful to the MacMillan client in the search for the truth behind their ancestral journey. Having exhausted the matched list with no satisfactory results, the study should have pursued other avenues of investigation. This would been a normal research protocol.
The discovery of the surnames in Ireland would not be beyond the capability of the case study's author. However to do so would have involved abandoning a set of pre conceived beliefs. One can only wonder, was the goal of the case study to validate its methodology or to serve its clients needs? If the former, it was at the expense of advancing the research needs of it's client. If the latter, they were very poorly served.