Main Findings

The main findings that have emerged from our Study so far are listed below:

1.   The success of this Study is manifest by its steady growth over thirteen years to about 500 participants, so that it is now one of the 50 largest of over 7,000 surname DNA projects worldwide.   

2.   About 5% of these participants, including senior representatives of all the long-established Scottish and Irish branches of the surname, have been asked to participate in the Study to enable traditional genealogical relationships to be proved or disproved.  The remaining 95% of participants have joined the study at their own volition, in the expectation that their tests will identify hitherto unsuspected relationships.

4.   96% of the participants have tested to 37 STR markers or more.  Testing to other resolutions has generally been found to be less cost effective, and at present upgrading to 67 or 111 markers is not recommended without prior consultation with the Study Administrator.  561 STR markers are now available to BigY testees, but these additional markers are of much less value than the more reliable SNP data derived from these newer tests (see below). 

5.   To qualify for membership of one of the Study’s genetic families (aka surname branches), comparison of the participant’s DNA signature with the modal haplotype of the Study using FTDNA’s 24-generation, no paper-trail TiP must exceed a probability of 90%.

6.   On this basis, the geographic origins in the Old World of the paternal ancestry of about 90% of all participants have been ascertained.  Scottish origins predominate.

7.   Some local spelling variants of the surname remain dominant within Great Britain (Irving in Central and western Dumfriesshire, Irvine elsewhere in Scotland, Urwin in Northumberland and Durham, Irwin elsewhere in England). However the spelling of the surname today, especially in Ireland and in the New world, has been shown to be a most unreliable indicator of the geographic origin of paternal ancestry.

8.   Amongst those now tested are descendants, direct or collateral, of the Lairds of Drum, Bonshaw and Castle Irvine, of the Dumfries, Eskdale and Limerick branches of our surname, and of individuals including the American poet Washington Irving (1783-1859), the orator Rev. Edward Irving (1792-1834), and the mountaineer Sandy Irvine (1902-1924).    
9.  Over 40 distinct genetic families (aka surname branches) using the surname and its spelling variants have been identified, each unrelated to one another during the surname era.   The number of distinct genetic families in our Study is now increasing very slowly, suggesting there are few more large branches of our surname waiting to be "discovered". 

10.   Two thirds of all the Study's participants are members of a single genetic family, sharing a common ancestor who probably lived in the Scottish Borders during the 14th century, although his name is unknown.  This "Borders" family includes participants representing the Irvines of Eskdale, the Irvings of Bonshaw and Dumfries (all in Dumfriesshire), the Urwins of Durham and Northumberland, and the Irvines of Castle Irvine (Co. Fermanagh).  Some of these participants still live in Dumfriesshire, some are descended from ancestors who migrated direct from there to USA, but the majority now living in USA are descended from ancestors who probably migrated from the Borders to Ulster in the 17th century, and from Ireland to America (typically to PA, VA, NC, SC or GA) in the 18th or 19th centuries, often for religious or economic reasons.  This proportion of participants sharing a single common ancestor is higher than found in most other Scottish surname DNA projects, and our Borders genetic family is probably the largest such family in any Surname DNA project.  

11.  In 2011 this Borders genetic family was tentatively divided into 14 sub-groups, each with a distinctive STR signature.  Some of these sub-groups were assigned a local geographic origin, and some an approximate age.  The sub-groups were identified with the help of cladograms (both phylogenetic trees and network diagrams).  However recent SNP tests are now showing most of this sub-grouping to be unreliable (see below).   

13. A curious feature of the largest of these sub-groups was that 30 participants have identical DNA signatures to 37 markers (and 12 of which have identical signatures to 67 markers), and yet no genealogical relationship between any of them can be identified.  However this uniformity does not continue in the 68-111 marker panel.  SNP tests now suggest that the apparent homogeneity at 37 and 67 markers is probably due to convergence and is misleading. 

14. Conversely another sub-group has two brothers with a genetic distance of 2 in 25 markers, illustrating the dangers of relying on a single participant to represent the DNA signature of a genealogically related family.  

16. All of the remaining genetic families are represented by only a few participants:  the largest has only 10 participants.

17. Nearly half of these small genetic families, all of whose members now use the surname Irwin (or spelling variants), share the DNA signatures of other Border surnames, suggesting non-paternity events ("NPEs") such as a young boy taking the name of his Irving stepfather, probably before the 17th century when many Irvings migrated to Ireland.

18. A further nine of the small genetic families have origins elsewhere in Scotland, including one from Drum in Aberdeenshire, one from Forfarshire, one from Perthshire, two from Orkney, and three from Shetland.  None of these families are genetically related to one another or to the Borders family within the surname era, and the lack of close matches with any other surname suggests most are not NPEs but are descended from unrelated individuals who adopted the surname independently at an early date, possibly because they came from the town of Irvine in Ayrshire, which was an important port in medieval times. 

19. The specific finding that representatives of the families of Bonshaw and Drum do not share a common paternal ancestor during the genealogical era contradicts - though does not unequivocally disprove - the tradition that the two lines were related in the fourteenth century. 

20. None of the participants share the DNA signatures of the Carlisle, Dunbar, Hume or Robertson families whose traditions, like that of the Irvings, claim descent from Crinan, the grandfather of King Duncan I.

21. It is accepted that the probabilistic basis of genetic genealogy prevents dogmatic conclusions, and that genetics alone cannot disprove long-established traditions. 
22. The origins of a further eight small genetic families have not yet been determined; all are probably Scottish; some may be NPEs.

23. These findings are suggesting that even within Scotland the surname has plural origins. This challenges the tradition that all Scots bearing the surname had a single ancestor.

24.  Eight other small genetic families appear to have non-Scottish origins:  four apparently use anglicised versions of Gaelic surnames that originated in southern Ireland, two appear to be descended from the Dutch or German surname Arnwin, and one has African origins.

25. 10% of the participants in the Study have clearly inherited the DNA signature of the relevant Irwin genetic family but are no longer using the surname Irwin (or variant spelling).   The ancestries of these participants are clearly include NPEs.  Probable circumstances, locations and approximate dates for some of these "events" have been identified.  

26. Only 6% of all participants cannot (yet) be unassigned to a genetic family because they are singletons apparently unrelated to any other participant.

27. So far there has been only limited success in identifying genealogical relationships within our genetic families, but with only about 0.3% of the world's 100,000 adult males who today use the surname, however spelt, having undergone a Y-DNA test, this is hardly surprising.  As new participants continue to join, and analysis methods improve, more relationships will be found.

28. These developments, together with FTDNA's predicted haplogroups of participants who have not undergone SNP tests, has enabled all of the Study's genetic families to be placed on a Clan Irwin haplotree (aka phylogenetic tree, or genetic family tree) which shows the descent of each genetic family, and hence its members, from the genetic "Adam".
29. First identified in 2011 thanks to the initiative of one of our participants, the L555 SNP is almost unique to our Borders genetic family.
30. A small number of participants in our Study have opted to take the Geno2 test, but the results have contributed little to our Study. 
31. As of April 2019, 40 participants have taken the BigY test, of which we have results for 40, of which 30 are L555+.  20 BigY700 tests are on order.  Six of our early BigY tests were funded in part or in whole by donations by a few most generous participants.  Although assessment of the BigY test results initially proved difficult, a methodology has been developed independently that is giving interpretations of BigY data that is compatible with those of our STR and conventional SNP data and with leading L21 "citizen scientists" including Dennis Wright, Mike Walsh and Alex Williamson.  Initial analyses of these L555 BigY tests suggests that the Border surname probably dates from about 14th century, roughly as expected. 
32. Building on the results of our BigY L555 results, FTDNA launched their L555 SNP Pack test in April 2016.  This tests for 86 SNPs contemporary with or downstream of L555.  This test enables members of our large Borders genetic family who have not taken the BigY test to determine which of SNPs they share with those who have taken the BigY test.  Over 70 participants have taken this test.
33.  These developments have confirmed the reliability of using 37 STR markers to establish membership of the relevant genetic family, but conversely the unreliability of most higher resolution STR markers for determining membership of sub groups within our Borders family. 
34.  Results of the L555 Pack tests have revealed the existence of further branches downstream of L555, but not their identities.  They have also shown that most of our tentative sub-groupings based on STRs to be misleading.  Those have taken BigY or L555 Pack tests now have a more reliable indication of which other participants they are most closely related to, and a more detailed haplotree is steadily evolving as more participants take this Pack test. 
35.  More specifically these tests have enabled distinctive SNPs to be identified with the Bonshaw and Castle Irvine lines, and this triangulation means that the human haplotree can now be extended down to within pedigrees derived by conventional genealogy.  The Study is one of a very few that have so far been able to do this.
36.  Building on the lessons learnt from this Study as the Administrator thereof I have been privileged to be asked to lecture at a number of genetic genealogy conferences in England, Ireland and USA.  These occasions enable me to extend my understanding of genetic genealogy and at the same time enable others to learn from our Study.  

37.  While much has already been learnt, including some surprises, about the origins of our surname and its genetic families and their sub-groups, it is confidently believed that considerable further understanding will result from the efforts of individual participants to improve their paternal pedigrees, from selective upgrading of some SNP tests, and from the rapidly evolving wider field of genetic genealogy.

38.  The Study respects the diverse contributions of all participants to the Study, the Clan Irwin Association and the clan and its heritage. It continues to welcome all new participants around the world who share our surname, however they may spell it, whatever their origin, and immaterial of the length of their paternal pedigree, as well as individuals who share our DNA but not our surname.  Through the generosity of some existing participants, financial assistance may be available for those with long paternal pedigrees.

39.  The Study's Privacy Statement and other administrative arrangements are thought to comply fully with the European Union's new General Data Protection Regulation, ISOGG's Interim Guidance on GDPR, and FTDNA's Terms and Conditions.   The new GDPR regime has not resulted in the loss of any participants from our Study, and 96% of our participants have opted to share their results in public. 

40.  The on-going success of this Study is attributed to the technical and financial support given by a number of its dedicated participants, to the support (non-financial) and publicity given by the Clan Irwin Association, by several members of the International Society of Genetic Genealogists, and by the staff of FTDNA, to the regular publication of detailed analyses of test results, and to our good fortune of inheriting a surname with such interesting genealogical and genetic features.

James Irvine, Study Administrator, May 2019