Friday 13 July 2018

Goodbye NPE, Hello SDS - some causes of Surname or DNA Switches

The term NPE stands for Non-Paternity Event, or Non-Parental Event, or alternatively Not the Parent Expected - a much more descriptive and accurate term coined by Emily Aulicino. It refers to the break in transmission either of the surname or of the DNA from parent to child. The break in transmission can occur from father to child, or from mother to child, or both. For fathers, the break in transmission can involve either the surname or the DNA; for mothers, the break in transmission involves only the DNA.

In relation to surname research, an NPE refers to a break in transmission of either Y-DNA or the family surname from the father to the son, along the direct male line (the father's father's father's line). Either the surname is switched for another one, or the father's Y-DNA is switched for some other man's Y-DNA. For this reason, I prefer to use the term Surname or DNA Switch (SDS), because it is much more descriptive and easier to understand than the term NPE which is somewhat obscure.

Surname or DNA Switches are important because they can mislead your family tree research, especially if you are using Y-DNA to research your surname, which should have passed down your father's father's father's line hand in hand with your Y-DNA ... but that is not always the case. 

In Ireland (as in many European cultures) where a system of hereditary surnames has been in existence for almost 1000 years, the chances of having an SDS on any one of your ancestral lines is about 50%. Here's the reason why:
  • According to several different studies, the average incidence of SDSs seems to be about 1-2% per generation.
  • So, if we allow 25 years per generation and hence 4 generations in 100 yrs, that will give us 40 generations in 1000 years.
  • multiply that by 1-2% per generation and that gives us a 33-55% chance of an SDS on any ancestral line going back 40 generations [for these corrected calculations see the first comment in the Comments section below]
  • If we go back along the father's father's father's line, then there is a 33-55% chance that your Y-DNA does not go back to the person who originated that surname.
  • If we allow 33 years per generation (and hence 3 generations per 100 years), then the chances of an SDS in the past 1000 years is 28-49% on each ancestral line [for these corrected calculations see the first comment in the Comments section below]
  • Whether you choose the 33-55% option or the 28-49% option, about half of us will not share the same DNA as the man who first held our surname … but half of us will.

There are many causes for SDSs and in relation to Irish family history, there are some unusual Irish-specific reasons for SDSs.

Allegiance to the Lord or Chieftain of the Clan (Sept, or Tuath)
It was not unusual for the servants, soldiers, vassals, tenants, or slaves of a clan to take (or be given) the name of their chief. This was a sign of respect for the chief but may also have conferred protection on the bearers of the surname. As a result, one would expect surnames that were associated with powerful clans to have many different Y-DNA signatures. And conversely, surnames that were associated with non-powerful clans and septs would have relatively few Y-DNA signatures. Surnames that might demonstrate this kind of pattern of multiple Y-DNA signatures associated with a single surname might include: O'Brien, Kennedy, O'Neill, O'Donnell, etc

This is an example of a surname-switch SDS. It means that the father who is doing the child-rearing will raise a son with the same Y-DNA as himself but with a different surname.

The Past was a different country - different customs, different reasons for Surname  or DNA Switches (SDS)

A similar situation arose with the Trans-Atlantic Slave Trade, where many enslaved people were given the surname of the Plantation Owner. Thus many African Americans today will carry the name of the slave owner of their ancestors (i.e. SDS present), and some will carry the Y-DNA of that slave owner (i.e. no SDS present), or his family (possible SDS), or the overseers he employed on his plantation (SDS present).

Adoption / Fostering / Guardianship
Modern adoptions are an obvious cause of surname switches. The adopted child takes on the surname of the adoptive father but carries the Y-DNA of his birth father. In times past, orphans or waifs would take on the surname of their guardians. These examples result in the child-rearing father having a son with the same surname as himself but with different Y-DNA.

Fostering on the other hand, did not frequently result in a surname switch. Under the Brehon Law of fosterage, children were brought up by relatives who were usually within the 5th degree of kinship (e.g. great uncle, 1st cousin once removed). However, the child did not adopt the family name of his foster family. Instead he retained his own family name and thus there would have been no SDS as a result.

Similarly, with modern-day fostering, the child in question does not usually take on the name of the foster family i.e. there is no SDS.

Young widow remarries
It would not be unusual for the young children of a woman recently remarried, to adopt the surname of her new husband. This results in the child-rearing father (i.e. the second husband) having a son (i.e. stepson) with the same surname as himself but with different Y-DNA (i.e. from the first husband).

"Widowed and fatherless"

Legal Condition of Marriage / Inheritance
Sometimes it was stipulated that a man would have to change his name in order to marry the daughter of a wealthy man to ensure that the family name was carried on. A good example is the Pennington family of Muncaster. There are two recent generations where the husband had to change his surname to Pennington as a condition of marriage. Thus, there have been two successive switches in the DNA associated with the Pennington surname along the same direct male line of descent.

A similar condition arises where the prospective heir has to change his name as a condition of inheritance - "You won't inherit a penny, my boy, unless you change your surname to Sidebottom".

"Sidebottom, grandfather? ... Seriously?!"

In his 2009 paper, Dennis Wright gives another perfect example of a surname switch to secure inheritance of land: Morgan O'Brien of Ballyphillip, Co. Limerick got married  in 1690 to Eleanor, daughter of Capt. Edward Butler of Bansha, descendant of Thomas Butler, 10th Earl of Ormonde. But Morgan O'Brien took Butler as his surname, probably as part of his and his wife’s succession to the lands of Bansha, Co Tipperary. Thus all their children would carry the Butler surname but O'Brien Y-DNA.

Taking wife’s name upon marriage (because of her higher social status)
In times past, it was not unusual for a man, upon marriage, to take the name of his wife if she was considered to be of higher social status. A famous example is that of Oliver Cromwell. His real name (perhaps we could say his maiden name) was Oliver Williams. His great-great-grandfather, Morgan Williams, married Katherine Cromwell in 1497. She was sister of Thomas Cromwell, Henry VIII’s chief minister, who was famously promoted to the earldom of Essex but later executed in 1540 when he fell from the King’s favour.

Morgan and Katherine’s three sons took the surname Cromwell in honour of their famous maternal uncle. This practice was repeated by many of their descendants, who also occasionally used the surname Williams-alias-Cromwell. Or at least they did until the Restoration, when some members of the family reverted to the surname Williams temporarily to distance themselves from any links to Oliver.

Morgan Williams and Katherine Cromwell’s eldest son Richard had two sons, Henry and Francis, both of whom also used the surname Cromwell. Henry was knighted and one of the youngest of his eleven children (Robert) was Oliver's father.

Thus the Cromwell surname became associated with Williams DNA.

Oliver Williams

Customary coupling with Dignatories
In ancient Celtic culture, apparently it was customary to offer one's wife to visiting dignitaries for "entertainment". This custom of guest-friendship-prostitution is still practiced in certain cultures today and is somewhat similar to the droit de seigneur or jus prime noctis that was apparently practiced in some places. Thus the dignitary's Y-DNA would become associated with any offspring of the union, but the child could be raised with the surname of the woman's husband.

Why was such a custom practiced? A child by a powerful man would give protection to a family and also had rights to succession under Brehon Law. A good example is that of Mathew (Fear Dorcha) O'Neill (1520-1558), the illegitimate son of Conn O'Neill. He was accepted by Conn as his natural son and was made his heir. However, Shane (Mathew's half-brother) contested his right of inheritance and Shane's men eventually killed Mathew, but not before Mathew had a son, Hugh O'Neill, who later went on to become The Great Earl and Clan Chief of the O'Neill. On this occasion the Y-DNA and surname remained intact (no SDS) but in other cases, the surname of the father who reared the dignitary's child may have been retained (i.e. SDS present).

There was also an associated custom of “Naming” children on a wife’s deathbed (i.e. she would reveal which of her children were the offspring of famous people). Apparently, as she lay dying, the wife would call her husband into the bedroom and something like the following conversation would ensue:
WIFE:           Husband dear, I need to have a word with you.
HUSBAND:   My dearest wife, what do you want?
WIFE:            Well ... you know our eldest child?
HUSBAND:     Of course I do ... sure isn't he my son?!
WIFE:              Well ... that's what I want to talk to you about ...
This sort of conversation does not happen very often today (or does it??), but it was relatively commonplace 500+ years ago.

Infidelity - a different thing under Brehon Law
In many ways, the old Brehon Law system practised by the mediaeval Irish, as well as their social attitudes, were more advanced and less restrictive than the laws we have today, especially in relation to the rights of women. There was a completely different concept of infidelity under Brehon Law. And such thinking is quite foreign to our modern ears.

How common infidelity was in medieval society is open to question. But if it was commonplace, and in addition married couples practiced regular sexual relations, then it must have been very difficult for a man to know which children were his and which were someone else's. Maybe they played guessing games at medieaval cocktail parties.

Attitudes to infidelity have varied with time and place. In Shakespearean times, "cuckolding" was the subject of derision and the cuckolded husband would be mocked. In Victorian and more modern times, infidelity carried a stigma and was taboo. As a result, it would probably be kept secret and remain undetected. Husbands would be blissfully unaware that their child carried another man's DNA.

Infidelity usually goes undetected, but sometimes there may be clues in the genealogical records. I came across a note written by the parish priest in the margin of an 1890 baptism record which read: this is not the husband's child - he has been working on the docks in Liverpool for 18 months.

"Not tonight, Joseph"

Illegitimacy - another different thing under Brehon Law
There was a completely different concept of illegitimacy under Brehon Law and it was not unusual for men to have concubines. In addition, sex with servants was commonplace.

There was also apparently a custom whereby a woman could decide which man she wanted to sire her child and which she wanted to raise him. So Dim Dave may have been in demand for his awesome good looks, whilst Four-Eyed Stan was lumped with child-rearing because he was an excellent Dad (despite the limp and halitosis).

In more recent times, and certainly since the introduction of Victorian morality, the illegitimate offspring of any "out of wedlock" encounter would usually be given the maiden name of the mother. However, this was not always the case and I have seen illegitimate children christened with their biological father's surname (in which case, no SDS is present).

Girl Power under Brehon Law

Anglicisation of surname
It was common for surnames to be anglicised in Ireland. And as a result there are some unlikely genetic associations between different surnames. The surname Green was at times an anglicised version of Hooney or Fahy, both of which are derived for an Irish word for the colour green.

Surname change to match the prevailing social circumstances 
Sometimes people would change their name to suit their circumstances. A lot of Jewish immigrants to the US changed their name to make it sound more "English" and thus avoid discrimination. Similarly Catholics living in a predominantly Protestant community might change their name to "fit in".

There is a popular belief that a lot of surnames were misspelt or assigned as people entered the US via Ellis Island but this has been contested for a variety of reasons. Firstly, passenger lists were not created in Ellis Island but at the port of departure of the immigrant i.e. in his or her native country, probably by a native speaker of the immigrant's language. The inspector at Ellis Island did not write down the immigrant's name - he worked from the passenger list.

Abandonment by husband
In cases where the husband deserted his family, the children sometimes took name of their mother out of respect for her and perhaps to distance themselves from their wayward father.

Legal Name Change after an important ancestor
Some people changed the family's name to reflect the family's ambition to achieve higher social status.

Babies switched at Birth
Occasionally we hear stories of babies being swapped at the hospital and the mother going home with the wrong baby. Several recent cases have come to light with the help of DNA testing.

More modern causes
In more modern times, the nature of SDSs is subtly changing with the introduction of new causes such as surrogacy, sperm donation, mitochondrial donation, and three parent families (one parent donates the egg, another the sperm, and a third the mitochondria).

Other causes
This is not an exhaustive list of the causes of Surname or DNA Switches, so if you can think of other examples (and I am sure there will be many), please leave a comment below.

But hopefully it will be abundantly clear from the above, why you only have a roughly 50:50 chance that your Y-DNA goes all the way back to the person who originated your surname some 1000 years ago (or thereabouts). And why, as you go back along each ancestral line in your family tree, there are likely to be 3-parent and 4-parent families. In other words, one particular family may consist of a genetic father as well as a genealogical one. And some families (e.g. where an adoption has taken place) will consist of two genetic parents and two genealogical parents.

As a result, we all have two types of family tree superimposed on each other - one purely genetic, one purely genealogical. But both equally valid in terms of our family history. Both types of ancestor (genetic & genealogical) have contributed to who their offspring became ... and who we are today.

Maurice Gleeson
July 2018


  1. Nic3e post, bout your math to get to "40-80% chance of an NPE" in 40 generations with a 1-2% change of a NPE per generation is not correct.

    The probability of having no NPE's in 40 generations given a 1% NPE probability is 0.99^40 = 0.669 (66.9%). So the probability of having at least one NPE is 1 - 0.669, that is 33.1%.
    For a 2% NPE probability in one generation the probability of no NPEs is 0.98^40 = 0.446 (44.6%). And the probability of at least one NPE is 1 - 0.446, that is 55.4%.
    Instead of 40-80%, it should be 33.1-55.4%

    For the case of 33 generations, instead of 30-60%, it should be 28.2-48.7%


    1. How did I comment on correction with two typos on the first sentence?!

    2. Thanks for pointing this out, Gonçalo - statistics are clearly not my strongest point! :-)
      I've incorporated your corrections into the post.
      Cheers, Maurice

  2. Sharing your blog on our surname group project... thank you Maurice!

  3. Great article! Permission requested to re-post on the Genetic Genealogy SIG Website as well as the Irish Genealogy SIG website of The Villages Genealogical Society.
    Jim Lannin

  4. Very interesting article, great information and guidelines. Sharing on Genealogy a la Carte blog. Thank you GWSFHS and Maurice Gleason.

  5. Thank you so much for this wonderful post! So many applies to my family branches and other extended branches (for my relatives). Couple that with the fact that surnames are new for both my paternal (Filipino) and maternal (Hawaiian) side, it adds more complication with Y-DNA testing.

    1. Nice to hear from you, Kalani. Hope you are doing well.


  6. I worked with a male who hated his father and took his wife's surname when they married. He is now in his mid-30s. I also have a friend who adopted a son when they were 70y/o and 23y/o and changed the surname of the son and the son's wife and children (this will be a genealogical nightmare). One surname is Spanish and the other Irish. I will be sharing this article on the Bucks County Genealogical Society FB page. Thank you for a great article.

    1. Thanks Sara. And thanks for the interesting examples!

  7. Great article. My Florance Driscoll great-grandfather changed the family name to Paul in Poplar, East London c 1870. Paul being his mother's surname. I think this was as a result his wife's (Mary Anne) lobbying. She came from the same Irish background and you can see her parent's family switching from RC to CoE through the parish registers and also changing the mother's surname from Gosnell to Gosling as the later children are registered. Anyway Mary Anne was baptised Mary and added Anne along the way and registered her marriage to Florance in 1867 using the surname Paul. He clearly did not like it as the two UK born children were registered as Driscoll and they are Driscoll in the 1871 census. She won out and when they migrated to NZ in 1873 it was as the Paul family and that has been in use from then on. Florance registered his first NZ born son with Driscoll as a middle name providing me the clue to the change 125 years later. Why the change? I'll never know for sure but suspect anti-Irish sentiment in UK at the time and a desire to assimilate by Mary Anne's family.

    Victor Paul

  8. Thank you for this article. I will share in Lancashire Genealogy group. I have two recent SDS in my family. One being my Grandfather was given his mothers maiden name, his mother then married (twice) and he still retained his mother's maiden name. But then her 4th child was given the name of her husband - 4 years after he died!

  9. Great article, may we reuse on site Worldwide Ancestry Finders? Many thanks

  10. Hi Maurice. Analysis in the Manx Y-DNA study see shows an incidence of NPE/SDS of around 0.5% per generation or ca 12% in the tested male population. This is based on a sample size of over 470 men tested so far.

    John Creer

    1. Interesting analysis, John. Most of the published data suggests the rate is 1-2% per generation so it is interesting to see this data from your surname project. I wonder if other people are doing this type of analysis? It would certainly be useful to have some additional data.

  11. Comments from a Dim Dave,

    Very nice post Maurice! FYI The ISOGG Wiki entry at has a very long list of possible SDS causes as well as pointers to various incidence rates and studies. Not surprisingly the estimated rate varies considerably across the ages and cultural and social norms.

    There are many anecdotal cases of SDS in traditional genealogy. One of our Vance DNA project subgroups has a story handed down from medieval times of two knight best friends where the Vance ancestor died in the Scottish-English wars and the second adopted his children "out of respect" (although we've never located this line matching current Vance descendants). But adoptions do seem to have been much less strictly regulated in former ages.

    I've encountered many researchers who firmly believe that the practice in Ireland of tenants taking their landlord's surname was wide-spread, although I'm sure it may have happened I have not yet seen any data that backs this up as a wide-spread cause of SDS.

    While I agree with you that the Ellis Island "wrote it down like it sounds" enduring story is more myth than reality, the same "anglicization" of a surname was more common once an immigrant to the US settled in a local community. We have several Vance subgroups who descend from German or Russian immigrants whose last names were originally "Wentz" and especially with the "W" pronounced as "V" in German, they eventually just went with the flow when all the English speakers around them wrote their name as "Vance".

    It would be interesting to see a follow-up article about how to recognize SDS in DNA Projects. For example, we have another Vance subgroup with a very distinct Y-DNA haplogroup descending from one immigrant to the US. That immigrant is known to have emigrated with and settled next to a man named Craig. That Vance line is an exact match to a subgroup of the Craig project, and the Craig subgroup is both older and more diverse than the Vance subgroup. This seems like an obvious case of a SDS although the exact cause is still unknown. Obviously it is can be a controversial and emotional subject to broach with testing participants so it would be great to have some common guidelines for how to recognize and handle it!

    Dave Vance

    1. Hi Dave, excellent comments. And I thoroughly agree about developing hints & tips for recognising an SDS ... and most importantly, how do you determine which of the alternative surnames came first and which came afterwards!

  12. Wonderful article, Maurice! As an adoptive parent, I especially appreciated the closing paragraph. I have four adopted children and my wife and I have been hard at work on their genetic and genealogical trees as well as our own. Our hands (And hearts) are full!

  13. Thanks for the informative article!
    The Cromwells are in my family tree, and it's somewhat of a mess.
    On another note, "[P]assenger lists were not created in Ellis Island but at the port of departure" is not completely the case. One example concerns German immigrants to the U.S. before very recent times. Many of the records of departure were destroyed in the wars, and the extant records were made upon the ships' arrival in the U.S. The only records of the various members of my father's family's coming here in the 19th century were written upon their arrival in Castle Garden, NYC. The expected mistakes were made in the transcriptions of their names and magnified by careless transcription, making tracing that genetic line a longer task.
    My Native American husband's surname was invented by the invading California missions. The list of names is short and the names are Spanish surnames, so the impression is that the entire ethnic group consists of very few families that are often assumed to be of Mexican or other Latin-American origin . Naturally, for this reason, and because of the dearth of written records and political/historical reasons, creating accurate family trees is a complicated task.
    And, lastly, current refugees' records tend to be written, often not accurately, upon their arrival, for obvious reasons.

    1. Thanks for this helpful information. Good points all. :-)

  14. Hi Maurice - Great admirer of all your presentations and style, even if not necessarily relevant to my researcher. For this blogpost, an additional possible name source for you, which I believe applies to my surname of GRACE (not the Irish one), and that it is church foundling name. Orphans who may have been left on church steps and have no name can be given foundling names such as GRACE, PILGRIM, CHURCH etc. I am on a Y-DNA hunt to connect my surname family (name root abt 1811) to any other. Not even close yet.

    1. Very true! The same thing happened in Ireland - foundlings being named after the street on which they were found, etc.

  15. Maurice,

    Quite interested in your comment about the Penningtons. Our Holt line (Nicholas) which is the MA/NE line with many descendants is an exact DNA match to one of the Pennington lines linked to Lancashire. If you could share the details and or documentation of the incidents you mention, it could provide extremely helpful clues! Thanks!

    1. Hi Christine, drop me an email and we can chat ... mauricegleeson AT

  16. Many thanks for your post. This is a topic that consumed several years of my own family research. From an emotional POV, it was not near and dear to my heard! lol Finally, through diligent paper research and DNA testing, I found the genetic Y-DNA break in my line. I suppose I was lucky since the NPE was of fairly recent happening. I have no idea if my paternal great-grandfather knew if someone else was his biological father. We’re dealing with very rural Southern families, some illiteracy, and ca. 1858 (the year of my gr-grandfather’s birth) Looks like great granny spent too much time at the spring house! lol Although I have not a clue if the matter was consensual (I shudder to think otherwise) I am certain the culprit was a male member of the farm NEXT DOOR. But why, at least for all my effort, could I not be rewarded with a family who had a really rare or unique surname. What was their surname? boring. lol

    1. Hi Phil, glad you liked it. My colleague Lisa Little was in a similar situation to yours, and again, through diligent research and some targeted DNA testing, was able to find out at which generational level the DNA Switch occurred. You can read her blog post about it here ...

  17. Maurice thanks for clarifying surnames. In our family we have several examples where the mother gave her illegitimate child her surname with the name of the "reputed father" as a middle name. An example : Martha Carnie & reputed father William Thomson had an illegitimate child & Martha gave her child the name Elizabeth (after her mother) Thomson (after reputed father) Carnie (family surname). A huge help when it comes to tracing family trees.

  18. Maurice,
    How do we constrain the timing of a SDS?
    Our family links with the Menzies clan go back some 300 ybp to the early 1700s, a hard paper trail. YDNA reveals that our "Celtic/Early Scots" line (DF13-L1335-S746) differs from "Menzies" brought by the Normans & linked to the Clan Chief. So sometime pre-1700 we adopted the Menzies surname. Given that surnames were probably not in general use in the remoter parts of Highland Scotland in the 1600s how can we constrain when the SDS happened?
    Much thanks for this article?

  19. This is a wonderful article that highlights everything I have been saying for years, despite my own extensive family history research, especially to people who say they have gone back a thousand years. The one thing you don't emphasise is simple clerical errors. One of my family names is Loughborough, and I have over 40 possible 'wrong' spellings, ranging from Loughbrough and Loufbro via Luffbro and Lickburrow, to Lovebarrow and more. This name, with its two different sounds for 'gh' is readily open to corruption.

    Martin Watson

  20. Interesting and useful article. For the DNA Project of my One-Name Study of my surname Pedlow I have found two main haplogroups: (1) Pedlow branches with origins in County Armagh, (Northern) Ireland, are from the E-M35 haplogroup (five of us have done Big-Y; three are E-FT153754 and two are E-FT257703; one older Y-DNA test on of the last surviving male of another Co. Armagh branch was simply E1b1b; unfortunately no further testing of him is possible); and (2) four Pedlows from different branches with origins in County Down, (Northern) Ireland; are all I-FT242053. There are also two Pedlows with completely different haplogroups who are suspected NPEs (SDS). I believe that I can explain the difference in haplogroups for these two counties, because most records for Pedlows in County Down spelled their surname as Pedlar prior to the 20th century, whereas the Armagh Pedlows were generally spelled correctly or with minor variations such as Pedlo or Pedloe in records going back to the early 18th century. In some County Down families the surname was still spelled Pedlar in the 1901 census of Ireland but had changed to Pedlow by the 1911 census, at which point there were no longer any individuals with the surname Pedlar in Northern Ireland. The Pedlow surname itself is originally Norman, and the earliest written record is in an Irish land grant from 1188, witnessed by Richard and Roger Pedelow. The surname evidently originated in France, where it was "Pied de Loup" (wolf's paw), a surname which still exists today in France as Piedeloup or Piedelou. I would love to be able to test some of them, but DNA testing still seems to be illegal in France. I would be curious to see if any of them are in the same haplogroup as the Armagh Pedlows; this would require around 1000 years of unbroken DNA transmission from father to son, but according to the article there is a 33-55% chance of finding this (more than I had expected after so many generations). So I guess I need to find some Piedeloups willing to break the law in France!

  21. I would to add the possibly common story of my surname to this epistle. The name is Spinner -assigned in the Austrian empire in the 18th century? Yet much more complicated. Sometime in the 1900s, a Galician bureaucrat decided that the birth of a male Jewish child 'deserved' a new surname for the family. He combined Spinner and Better - the parents surnames - into Betterspinner! For a few decades this was the surname used near Lviv (now in Ukraine), Yet when some young males emigrated to North America in the 1900s, they reverted to using Spinner as their surname. And to even colour this naming story further, the Y DNA evidence shows that back to mediaeval times we are actually LURIA descendants!