Page Actions

Non-paternity event

From ISOGG Wiki

This page contains changes which are not marked for translation.

Other languages:
English

Non-paternity event is a term used in genetic genealogy to describe any event which has caused a break in the link between an hereditary surname and the Y-chromosome resulting in a son using a different surname from that of his biological father.[1] The definition excludes minor changes in the spelling of the surname, and is implicitly limited to events after the relevant branch of the surname became hereditary.

The term non-paternity event was first used in the context of genetic genealogy by Bryan Sykes and Catherine Irven in their 2000 paper "Surnames and the Y-chromosome".[2]

The term non-paternity event is also used in law, sociology and clinical genetics but with a narrower definition to denote cases of false paternity where the biological father of a child is someone other than who it is presumed to be.

Sections 1 and 2 of this page introduce the use of the term non-paternity events (NPEs) in genetic genealogy. Sections 3 and 4 offer some statistics on NPEs that show these events are likely to occur in every surname DNA project. Sections 5, 6 and 7, intended primarily for administrators of surname DNA projects, suggest ways in which NPEs, one of the more controversial issues they are likely to encounter, can be recognized and their associated features handled sensitively and constructively.

Synonyms

The following is a list of apparent synonyms though strictly speaking some of these terms are not synonyms but describe specific NPE scenarios:

  • Ancestral introgression
  • Break in transmission (BIT)
  • Cuckoldry
  • Exogenous ancestry
  • Extra-pair paternity
  • Extra paternity event
  • False paternal event
  • False paternity event
  • False paternity
  • Incorrectly assigned parentage
  • Male introgression
  • Misattributed parentage
  • Misattributed parentage event (MPE)
  • Misattributed paternity
  • New Surname Event
  • Non-parental event
  • Non-paternal event
  • Non-paternity event (NPE)
  • Non-patrilineal transmission
  • Not the Parent Expected [3]
  • Out-parenting event
  • Paternal discrepancy
  • Paternity discrepancy
  • Paternity fraud
  • Surname discontinuity event (SDE)
  • Surname from mother (SFM)
  • Surname switch
  • Undocumented adoption

The term non-paternity event has a more specific meaning to clinical geneticists, as does the term paternity fraud to lawyers. It would have been better if genetic genealogists had adopted some less pejorative term such as Surname Discontinuity Event to meet their need for a “catch-all” interpretation, but pending some standardization NPE will continue to have different meanings to different people.

NPE scenarios in the context of genetic genealogy

These include:

  • Illegitimacy outside marriage: boy taking maiden name of mother
  • Infidelity within marriage: boy taking surname of mother’s husband
  • Re-marriage: boy taking surname of step-father
  • Rape: boy taking surname of mother or partner
  • Changeling, surrogacy, sperm donation, unintentional embryo/baby swap: boy taking surname of mother or partner
  • Adoption, incl. ‘hidden’, orphan & foster: boy taking surname of guardian
  • Apprentice or slave: youth taking surname of master
  • Tenant or vassal: man taking surname of landlord or chief
  • Anglicisation of gaelic or foreign name: man taking translated/phonetically similar name
  • Formal name-change, e.g. to inherit land: man taking maiden name of wife or mother
  • Name-change to hide criminal past, embarrassing surname, or a stage name: man taking unrelated surname
  • Informal name-change, alias, by-name: man taking name of farm, trade or origin
  • Mistake in genealogy, or in DNA analysis

Informal name changes, the use of aliases and by-names, and name changes by tenants, vassals, apprentices and slaves were prevalent in the 13th-18th centuries, in some cases before surnames became hereditary, and in this latter context, strictly speaking, they are not NPEs. Similarly, a genealogical mistake is not strictly an “event”, but this too can be manifest as an NPE.

Anglicisation of surnames occurred in Ireland in the 16th century, in the Scottish Highlands in the 18th century, and in America in the 18th through early 20th centuries. Formal adoption and unintentional baby swapping in hospitals have only arisen in the last two centuries, while surrogacy, sperm donation and unintentional embryo swapping are obviously recent developments. The other scenarios above have been on-going for centuries. NPEs may have been more common in Scotland than elsewhere because of the right and custom of mothers to retain the use of their maiden name after marriage.

The important point emerging from the above is that in genetic genealogy the potentially embarrassing possibilities of previously unsuspected illegitimacy and infidelity are only two of many scenarios. Probably the most likely scenario is the death of a young father, perhaps due to accident, combat or disease, resulting in the mother remarrying and the young boy taking the name of his step-father.

Incidence rates of NPEs

The rate of non-paternity events has often been quoted to be around 10%, and while this figure is now deprecated,[4] there is no widespread consensus on a better substitute.

Contemporary NPE statistics

A large number of studies of have been conducted. Primary studies of current (as opposed to historical) non-paternity rates include:

  • France: 2.8%, in a sample of 362 children. [5]
  • Mexico: 9.8% to 13.8%, in a sample of 396 children. [6]
  • Switzerland: 0.3% to 1.3%.[7]
  • United Kingdom: “few %”;[8] 1.3%;[9] 1 to 2%, in a sample of 1,678 men. [10]
  • USA, Hawaii: 2 to 3%, in a study of 1748 families with 2,839 children. [11]
  • USA, Michigan, Detroit: 1.4% and 10.1%, in studies of 1,417 white and 523 black children. [12]

Several attempts have been made to analyse such observed studies:

  • In 1996 Allison found that estimates of non-paternity in the general population vary, ranging from 1% to as high as 20%.[13]
  • In 2005 Bellis et al. published international studies of paternal discrepancy. They found a range in incidence from 0.8% to 30% (median 3.7%). In situations where disputed parentage was the reason for the paternity testing, there were higher levels, an incidence of 17% to 33% (median of 26.9%). Most at risk of parental discrepancy were those born to younger parents, to unmarried couples and those of lower socio-economic status, or from certain cultural groups. [14]
  • In 2006 Anderson examined non-paternity rates from 67 published studies. Non-paternity rates for men who were judged to have high paternity confidence ranged from 1.9% in the U.S. and Canada, 1.6% in Europe, and 2.9% elsewhere. Men with “high and unknown” levels of paternity confidence exhibited a 3.9% non-paternity rate. In contrast, for men in studies of disputed paternity, who were considered to have low paternity confidence, the rates of non-paternity were higher – 29% in the U.S. and Canada, 29% in Europe, and 30% elsewhere. [15]

The studies of Bellis et al. and Anderson, and the remarks attributed to Jobling, suggest an emerging consensus that for families where paternity has not been an issue, paternity tests are showing average illegitimacy rates in the order of 2%. However for families where paternity is disputed the average rates may approach 30%. Presumably disputed paternity is relatively uncommon in such tests, as Bellis suggests a median rate of 3.7%.

But it is unclear whether such test data represents non-paternity rates amongst the general population, and today approaching half of all births in UK and USA are out of wedlock.

Historical NPE statistics

Evidence of historical rates of non-paternity, more relevant to genetic genealogy, is more sparse:

  • Laslett et al. assessed illegitimacy rates in English parishes from 1581 to 1810, falling from 4% to 0.5% c.1650, and then rising to 6%;[16]a later study found bastardy rates of 1-7%, averaging 3-4%, in 98 English parishes from 1540 to 1900, and 4-11% in Scotland in the mid 19th century. [17]
  • Smith estimated that in three Suffolk manors in the second half of the 13th century 5-12% of births were illegitimate.[18]
  • Razi thought illegitimacy rates in Halesowen in the 14th century were probably even higher. [19]
  • Illegitimacies recorded in the parish/civil registers held by the Scottish General Register Office have averaged 5.5% (1845-1899), 4.6% (1900-1945) and 16.3% (1946-2005). [20]
  • McEvoy and Bradley calculated one Irish family had 1.6%.[21]
  • King and Jobling (2009) demonstrated that, historically, false paternity rates probably lie between 1% and 4.5%, and adopted 2% for their simulation modelling.[22]
  • Larmuseau et al claimed that the rates of 1-2% they found in Flanders over the last four centuries were unlikely to differ over time or by locality.[23]
  • Greeff and Erasmus (2015) studied 1273 conceptions in 23 Afrikaner families in South Africa over 200 years and found that the cuckoldry rate was just 0.9%.[24]
  • FTDNA believe that the rate of non-paternal event, unannounced adoption or false paternity is about 1-2% per generation.[25]

Commentary on observed rates of NPEs

As Dickason (2005) noted, observed rates of non-paternity are highly idiosyncratic, depending on location, time, religion and other factors. These other factors include:

  • the scatter inherent in small sample sizes and survey error;
  • underlying differences in populations being sampled, e.g. surveys of illegitimate births should always yield smaller rates than surveys of non-paternity, as the latter also include infidelities within marriage; as Dickason said, “remember that a 4% non-paternality rate could be made up of about 2% illegitimacy and 2% other causes”.
  • differences between per-generation studies and cumulative studies over several generations: DNA projects will typically reflect the latter.

Cumulative NPE rates

All the data in section 3 above is explicitly or implicitly “per generation”. But all these “per generation” rates are cumulative over the genealogical timeframe. Plant suggested that “one can theoretically expect that around half of randomly selected, modern bearers of a populous, single family surname will remain free of ancestral introgressions.” [26] He has developed the arguments further, suggesting the theoretical probability P of a participant not being a biological descendant of his surname’s progenitor can be approximated from the formula P% = 100(1-(1-p) n)<, where p is the fractional probability of a false paternity per generation and n is the number of intervening generations: [27] [28]

Table: Cumulative effects of NPE

Generations p = 1% p = 2% p = 5% p = 10% p = 30%
n = 5 5% 10% 23% 41% 83%
n = 15 14% 26% 54% 79% 100%
n = 25 22% 40% 72% 93% 100%
n = 35 30% 51% 83% 97% 100%

Furthermore, the number of intervening generations may be calculated from n = (T/t + 1), where t is the average generation interval in years, and T is the number of years since the earliest hereditary holder of the name.

Simplistically it can be assumed that hereditary surnames were first found in Ireland in the 10th century, in England in the mid 11th century, and in the Lowlands of Scotland in the late 13th century. Combining these elements, and adopting FTDNA’s upper guide for NPE rates of 1%-2% per generation, gives the following cumulative probabilities that may be expected for the proportion of participants in a surname DNA project who may, in theory, be expected to have NPE ancestry:

Table: Expected proportion of participants to be NPE

Date of earliest ancestor
with hereditary surname
T 25 yrs/
generation
33 yrs/
generation
40 yrs/
generation
  AD1700 300 years 12-22% 10-18%   8-17%
  AD1500 500 years 18-35% 15-28% 12-24%
Scotland AD1300 700 years 25-46% 19-36% 17-32%
England AD1100 900 years 33-55% 25-44% 20-37%
Ireland AD 900 1100 years 46-65% 31-52% 25-43%

And as been shown, per-generation NPE rates higher than 2% are quite possible.

The difficulty in identifying NPEs within surname DNA projects, especially in plural- and multi-source surname projects, means that in practice fewer observed NPEs may be found than the theoretical proportions suggested by this table. Conversely, it also means that most surname projects probably include more NPEs than is readily apparent or currently realized, although other factors such as genetic drift and founder effect will also be relevant.

Types of NPE in genetic genealogy

In surname DNA projects two distinct types of NPE arise:[29]

  1. Egressions, or e-NPEs, that is in surname project A, men with a haplotype characteristic of surname A but who are now using surname B, e.g. men with Adams Y-DNA but who use the Bell surname.
    1. An egression arises when a Y-DNA testee finds his haplotype is a near match[30] to none or only a few of those of other participants in the project of his current surname, but is a near match with those of many in another surname project, and asks the administrator of this other project if he may join that project.[31]
  2. Ingressions, or i-NPEs, that is in surname project A, men now using surname A whose haplotypes are more characteristic of men who still use surname B, e.g. men with Bell Y-DNA but who use the Adams surname.
    1. An ingression arises when a participant in a surname project is found to have a haplotype that is a near match with none or only a few of other men using his surname, but is found to be a near match with many in another surname project, AND there is evidence, specific or circumstantial, of a NPE involving the two surnames at some place and date which could explain the earlier use of the “other” surname.

Three corollaries arise from this distinction:

  1. It follows that typically one surname project’s e-NPE is another surname project’s i-NPE.[32] And so care has to be taken to avoid double counting when making statistical comparisons involving NPEs. For example, the statistics in sections 3 and 4 above can apply to e-NPEs or i-NPEs, but not to the sum of e- and i-NPEs identified within a single surname project.
  2. While the identification of previously unsuspected i-NPEs may cause disappointment and embarrassment, this can be avoided by sensitive handling of the matter, and will not arise when the event is already suspected, or, by definition, with e-NPEs. This issue is addressed in section 7 below.
  3. Deciding whether a participant with a mismatching haplotype is an i-NPE or is a representative of an original branch of a surname can also be a sensitive issue – see section 7 below.

Communicating NPE status to a project participant

Some project administrators are reluctant to consciously include NPEs in their project, while others are apprehensive about how best to advise a participant that his ancestry apparently includes an NPE. Obviously it would be most insensitive and tactless for a project administrator to bluntly inform a participant that his Y-DNA implies an illegitimacy or infidelity in his ancestry. But this is quite unnecessary: one of the objectives of this page is to show why no participant need be confronted with such a conclusion, and that there are many reasons why NPEs can and should be included in surname projects. Two points arise.

Participant awareness

The first point is that most participants who have a NPE in their ancestry will have already been aware of this possibility or, if unaware, are soon likely to suspect this possibility. All those who are e-NPEs have by definition already become aware, as it is they who have taken the initiative of asking to join the project of their “new” surname. Many i-NPEs are already aware they have a name change in their paternal ancestries, and some of these suspect or even be aware of the particular circumstances. And those i-NPEs who were unaware of a name change in their paternal ancestries will soon recognise that the high proportion of their “Matches” who bear some other surname is at least suggestive of the likelihood of there having been an i-NPE in their ancestry. In other words, if a participant with a NPE in his paternal ancestry is not already aware of this fact, he is likely to find out sooner or later. [33]

NPE not synonymous with illegitimacy

The second point is that a discontinuity between surname and inherited Y-DNA, the broad definition of a NPE, is far from being synonymous with illegitimacy. Today nearly 50% of births in UK and USA are outside wedlock,[34] and leaving aside the philosophical question of why this is condoned by today’s society while we frown at illegitimacy in our own ancestry, births out of wedlock were much less common in the past ( see section 3 above).

And leaving aside the more esoteric possible causes of NPEs, there were probably no more boys born out of wedlock or from infidelities than there were who took the name of their step-father after their own father had died from disease or been killed in combat or in an accident at work.

The impact of NPEs on surname projects

NPE’s raise a number of issues for administrators of surname DNA projects.

For e-NPEs the issues are relatively straightforward. As the testee has already taken the initiative of asking to join the project he is unlikely to feel particularly sensitive about the matter, and it will be readily apparent to which genetic family[35] within the project that the testee “belongs”. [36]

Two issues remain. The first is a matter of principle: should testees not using the project surname but having a haplotype that matches the Y-DNA signature of one of the genetic families in the project be included in the project? At one extreme the simple answer is “No”, i.e. to restrict the project to only those participants who use the surname. But in whose interest is an administrator acting if they choose to exclude, or remove, [37] an adopted testee with a different surname who knows his biological father bore the project surname? Or whose earliest confirmed paternal ancestor used the surname? Or who has a 67/67 match with a modal participant of one of the project’s genetic families? The choice is down to the administrator. But as the project grows, consistent treatment will be important, and the choice made should be compatible with the project’s terms of reference, goals etc. which in turn may need to be modified.[38] The wider the interpretation that is chosen the more the project is likely to learn of the context of the surname, and to satisfy the aspirations of individual members.[39]

The second is a matter of degree: if testees with matching Y-DNA but different surnames are accepted in the project, how close does the “match” need to be? This again is a matter of judgement for the project administrator. It is quite justifiable for the administrator to stipulate a more stringent definition of “near match” when the surnames differ than when they are the same, and to err on the side of caution, explaining why, in the absence of additional evidence, he believes membership of the project would be inappropriate. [40] However in practice the determining consideration will usually not be the “nearness” of the match, however measured, or the need for a SNP test, but the simple fact that the common ancestry will be very clear from the first page of the testee’s Y-DNA matches, where all or nearly all the Matches will be with participants in the project.

For i-NPEs the issues facing the project administrator are very different. It would be difficult for an administrator to justify exclusion from a project of participants who share the same surname, no matter what their Y-DNA signature may be, even when, for example, it is already known or suspected that a paternal ancestor had been adopted. And if exclusion was desired it would be difficult to develop some objective definition of i-NPE to justify exclusion on a consistent basis.

A compromise adopted by some project administrators is to include all participants using the project surname, but not to draw attention to the knowledge, or suspicion, that some have, or may have, an i-NPE in their ancestry. For example, administrators may divide participants into genetic families, each family comprised of participants who are all “near matches” to the modal DNA signature of that family, but even if it is known that one or more members of the family are adoptees, neither the participant nor his genetic family is described as being a NPE. This “comfortable” approach avoids risk of embarrassing interpretations, and the need for subjective decisions.

But as a project grows in size, so too will the limitations of such an approach become more and more apparent.

If one member of a genetic family is known to be an adoptee, or when one or more of its members notice that most of their “Matches” across FTDNA’s entire database are with testees of some single other surname (apart from “Matches” with other members of their own genetic family), then the project administrator will come under increasing pressure to make other members of the genetic family concerned aware of the possibility/likelihood of a NPE in their ancestries. This pressure will increase if evidence comes to light that links the two surnames, for example if the surnames originated in the same district.

The alternative to these issues being “brushed under the carpet” is to accept that it is almost inevitable that every surname have some i-NPE lineages (see sections 3, statistical rate, and 4, cumulative rate, below), and seek to handle the implications arising sensitively but objectively. Two issues arise.

  • The first is the importance of avoiding unnecessary embarrassment. As we have seen, there need be no embarrassment when an i-NPE participant already knows of, or suspects, an adoption or illegitimacy in his paternal ancestry. Nor need there be embarrassment if a new participant asks his project administrator to comment after noticing that he has no or few “Matches” with his own surname but many “Matches” outside the project, all or nearly all of which are with some other, single, surname. It is thus a small step, if the same participant has not noticed such matches, for his administrator to draw such matches to his attention, and to explain the possible implications, stressing that there are many possible explanations and there is no need to assume that the more embarrassing explanations are more likely than the non-embarrassing explanations. [41] And of course the “event” may have taken place many generations ago. It can be argued that giving advice along these lines is one of the responsibilities as a project administrator, and the initiatives of at least one administrator[42] have upset any participant in his project. However other project administrators make take a very different view.

While a participant remains a “singleton”, i.e. is not a member of a genetic family within the project, there is no need to draw attention to the possibility that he may be a NPE. But a more difficult scenario arises when a project has a “new” group of two or more participants who use the project surname and have genetic signatures that “match” each other but do not match those of participants in other clusters/groups in the project, especially if the administrator identifies such groups as genetic families and seeks to identify their likely origins. Under the broad definition of NPE, whenever a “new” genetic family is identified within a surname project, logic dictates that it must fall into one of three categories:

  • (a) it includes an i-NPE in the ancestry of its participants, or
  • (b) it represents an original branch of the surname[43], or
  • (c) its status is unknown, i.e. it is unclear whether (a) or (b) applies.

Project administrators who respect the interests of their more zealous participants will seek to minimise the latter (c). How this may best be done a complicated matter. Two extremes can be identified:

  • in a mature[44], single-source surname[45] project, it has to be assumed, by definition, that all participants and genetic families using the project surname but having genetic signatures that do not match the surname signature must be i-NPEs;
  • in a young, multi-source surname project, it should be assumed, pending further evidence, that all participants or genetic families using the project surname but having genetic signatures that do not match any of the other genetic families in the project, represent a new, original branch of the surname. Further evidence that they might instead be i-NPEs include a known adoption of a shared ancestor, or the “Matches” across the entire FTDNA database showing clearly that there has been an i-NPE in the paternal ancestry of the participant(s), i.e. the “Matches” at 37 markers or more are exclusively, or almost exclusively, with some single “other” surname.

These two examples suggest that identification of i-NPE ancestry is dependent on the type and maturity of the surname project. Three types of evidence may help clarify the conundrum:

  1. one or more of the participants concerned already suspected an i-NPE in their ancestry;
  2. “Matches” on the personal web page of each participant in the genetic family with FTDNA’s entire database reveal many “Matches” to some surname other than that of the project;[46]
  3. further research discovers evidence of some link between the two surnames, for example a previously unknown first marriage, or the two families having been neighbours, SNP tests etc.

The interrelationship of these factors for determining whether or not the ancestry of a newly identified genetic family is categorised as an i-NPE or an original branch of the name may be summarised thus:

Decision table for i-NPE or original branch

  specific
i-NPE previously known or suspected
“other” surname previously known or suspected unsuspected i-NPE suggested from FTDNA’s “Matches” research suggests surnames were neighbors All surname projects Single-source surname projects Plural-source surname projects Multi-source surname projects
young mature young mature young mature
A yes yes yes or no yes or no DNA test effectively confirms or partly refutes previous knowledge
B yes no yes yes or no As A above, plus DNA test identifies “other” surname
C no no yes yes   strongly suggestive of being NPE further evidence desirable specific evidence necessary
D no no yes no Further evidence needed, especially if “other” surname is multi-source
E no no no   in these circumstances two possible exceptions may apply: either (a) the non-matching genetic family is an i-NPE
  likely certain
{in theory)
possible, further evidence desirable unlikely, dangerous to assume possible, difficult to identify
or (b) the non-matching genetic family is not an i-NPE

but is a new, previously unsuspected branch of surname

  possible, but surname becomes plural source impossible (in theory) possible, further evidence desirable very likely possible, difficult to identify

How much evidence is needed to label a genetic family as an i-NPE or as an original branch of the surname is a subjective matter. If in doubt it will generally be safer, unless the administrator is very confident the surname is “single-source”, to err on the side of caution and assume that the genetic family is an original branch of the surname until there is sufficient evidence to show it is an i-NPE.

Conclusions

Two fundamental points arise from this discussion:

  • Even if administrators of surname projects wish to avoid the subject, it is very likely that every surname has some NPE’s in the paternal ancestries of some branches of the name; [47] and
  • How best to handle the subject is a matter of choice for each individual administrator, and despite the guidance offered above it is not unreasonable to avoid making proactive judgments.

Further reading

Blog posts

See also

Acknowledgements

ISOGG would like to thank James Irvine who provided the content for this updated page in July 2014 assisted by Debbie Kennett and Ralph Taylor.

Notes

  1. This definition of NPE follows the thinking of Sykes and Irven 2000; Smolenyak and Turner (Trace Your Roots with DNA 2004, p40); Pomery (DNA and Family History, 2004 p76); Dickason, DG 2005: “Familial and Genetic Descendancy; Conflict or Complement?”: http://www.isogg.org/fgd.htm); Plant (“Surname Studies with Genetics” 2009, p12: http://cogprints.org/6595/); Irvine 2010, p9: http://www.jogg.info/62/files/Irvine.pdf); and Kennett (DNA and Social Networking 2011, p212; The Surnames Handbook 2012, p152). Such a “wide” definition is necessary to enable the handling of the many contingencies that may arise within a DNA surname project, where the cause of the event is usually unknown, and its existence is typically derived from DNA evidence of a second surname. However some genealogists prefer a narrower definition, e.g. Matthiesen: “NPE's and Their Resolution”.
  2. Sykes B, Irven C (2000). Surnames and the Y Chromosome. American Journal of Human Genetics 66 (4): 1417-1419.
  3. Term coined by Emily Aulicino.
  4. Attributions have included
    • http://freepages.genealogy.rootsweb.ancestry.com/~gkbopp/DNA/falsepaternal.htm (which in fact says 5-10%, based on folklore);
    • Macintyre S and Sooman A (1991) “Non-paternity and prenatal genetic screening” Lancet volume 338 (8771) pp869-71; doi:10.1016/0140-6736(91)91513-T. PMID 1681226;
    • Neale MC, Neale BM, Sullivan PF (2002) “Nonpaternity in Linkage Studies of Extremely Discordant Sib Pairs” in American Journal of Human Genetics volume 70 (2) pp526-9; doi:10.1086/338687. PMC 384925. PMID 11745068. (which in fact gives a rate of 5-10% for sib pairs, not the general public);
    • Rincon P (“Study debunks illegitimacy 'myth'” BBC News, retrieved 11 February 2009) quotes Jobling as saying the oft quoted figure of 10% for illegitimacies is an urban myth.
  5. Le Roux, Marie-Gaelle; Pascal, Olivier; Andre, Marie-Therese; Herbert, Odile; David, Albert; Moisan, Jean-Paul (1992). “Non-paternity and genetic counselling”. The Lancet volume 340 (8819) p607. doi:10.1016/0140-6736(92)92139-7
  6. Cerda-Flores RM, Barton SA, Marty-Gonzalez LF, Rivas F, Chakraborty R. (1999) “Estimation of nonpaternity in the Mexican population of Nuevo Leon: A validation study with blood group markers” in American Journal of Physical Anthropology volume 109 (3) pp281-93.
  7. Sasse G, Müller H, Chakraborty R, Ott J (1994) “Estimating the frequency of nonpaternity in Switzerland” in Human Heredity volume 44 (6) pp337-43
  8. Brock DJK and Shrimpton AE (1991) “Non-paternity and prenatal genetic screening” in Lancet volume 338 (8771) p1151
  9. Sykes, B and Irven, C (2000).
  10. King TE and Jobling MA (2009) “Founders, drift and infidelity: the relationship between Y chromosome diversity and patrilineal surnames” in Molecular Biology and Evolution volume 26 (5) pp1093-1102. This is in fact a misquote: the published article does not mention 1-2%, simply saying “Historical rates of nonpaternity are difficult to estimate, though modern rates, where these have been measured, are of the order of a few percent per generation (Brock and Shrimpton 1991; Sasse et al. 1994). Elsewhere (Rincon 2009) Jobling is quoted as saying: “If you look directly at [single-source] families without any prior suspicion of non-paternity, then you find a value [for illegitimacy] of about 1% or 2%."
  11. Ashton GC (1980) “Mismatches in genetic markers in a large family study” in American Journal of Human Genetics volume 32 pp601-13
  12. Schacht LE, Gershowitz H “Frequency of extra-marital children as determined by blood groups” in Gedda L (ed). Proceedings of the Second International Congress on Human Genetics Rome, Sept 6-12, 1961
  13. Allison DB (1996) “The use of discordant sibling pairs for finding genetic loci linked to obesity: practical considerations” in International Journal of Obesity volume 20 pp553-60.
  14. Bellis MA, Hughes K, Hughes S, Ashton JR. “Measuring paternal discrepancy and its public health consequences” in Journal of Epidemiology and Community Health September 2005 volume 59 (9) pp749-54.
  15. Anderson, KG (2006) “Evidence from Worldwide Nonpaternity Rates” in Current Anthropology volume 47 (3) pp513-20. This article updates a 2004 version quoted by Dickason 2005.
  16. Laslett P and Oosterveen (1973) “Long-Term Trends in Bastardy in England: A Study of the Illegitimacy Figures in the Parish Registers and in the Reports of the Registrar General, 1561-1960” in Population Studies vol.27 pp.255-286.
  17. Laslett P et al (editors) (1980) Bastardy and its Comparative History.
  18. Smith. RM (1974) "English Peasant Life-Cycles and Socio-Economic Networks" unpublished Univ. of Cambridge Ph.D. thesis, pp1454-6.
  19. Razi, Z (1980) Life, Marriage and Death in a Medieval Parish: Economy, Society and Demography in Halesowen 1270-1400, p 70. Further information may be found in: Bennett, JM (2003), “Writing fornication: medieval leyrwite and its historians” in Transactions of the Royal Historical Society, volume 13 pp131-62. Hanawalt, B (2001) “Childrearing among the lower classes of late medieval England” in Robert I. Rotberg (ed.), Population History and the Family: a Journal of Interdisciplinary History Reader (Cambridge, Mass.).
    • Razi, Z and Smith RM (eds.) (1996), Medieval Society and the Manor Court.
  20. http://www.one-name.org/members/wiki/index.php?page=Illegitimacy%20rates
  21. Plant 2009, p 8.
  22. King and Jobling 2009, p1095.
  23. Larmuseau MHD et al (Dec. 2013) “Low historical rates of cuckoldry in a Western European human population traced by Y-chromosome and genealogical data” in Proceedings of the Royal Society vol. 280 no. 1772.
  24. Three hundred years of low non-paternity in a human population. Heredity advance online publication, 6 May 2015.
  25. What is the likelihood of a non-paternal event or false paternity? FAQ, Family Tree DNA Learning Center.
  26. Plant 2009, 3. Plant’s comment was no doubt based on Sykes’s finding that only 44% of his participants matched the modal haplotype, and Sykes’s study was in turn based on only four markers.
  27. Although Plant was discussing single-source surnames, this argument is broadened here on the assumption that multi-source surname holders behave similarly to single-source surname holders.
  28. In Dr JS and Prof RE Plant “Getting the Most from a Surname Study” (2012) this approach has been reversed: their study has observed 35.5% of their participants are non-matching and hence, as it is a single-source surname, these are all assumed to be i-NPEs, and if 23 generations have elapsed since the origins of the surname, the formula gives a theoretical false paternity rate of 1.9%. This is seen to be compatible with the 2% used by King and Jobling. Similar data gave theoretical NPE rates of 1.3% and 2.2%.
  29. This terminology is attributable to Plant (2009, p9).
  30. The term “near match” may be considered similar to the term “Match” (formerly “Close match”) used by Family Tree DNA (i.e. 12/12, 23/25, 33/37 or 60/67), but caters for individual project administrators who may prefer some other definition, for example a TiP score (a 24-generation, no paper-trail comparison with a modal participant) of 60%.
  31. Some Surname DNA project administrators may not recognise NPEs without clear evidence of an actual event (see below); others may choose to exclude e-NPEs from their project. As with other such issues, these choices remain the prerogative of the individual project administrator.
  32. A tedious logic can identify some i-NPEs which need not be another project’s e-NPEs
  33. How a project administrator might convey this to a participant is discussed in section 7 below.
  34. http://www.ons.gov.uk > Population > Birth Summary Tables; http://www.cdc.gov/nchs/data/nvsr/nvsr61/nvsr61_05.pdf.
  35. By genetic family is meant a group of participants in a surname project whose genetic signatures are near matches with the “modal participant” (the participant with the most common values for each marker). The modal participant is usually the best available indication of the genetic signature of the common ancestor of the genetic family.
  36. For example the Winchesters in www.familytreedna.com/public/taylorfamilygenes/default.aspx?section=yresults\, Group I1-01.
  37. Exclusion could be exercised from a closed project, or from an open project when a request to join is made, removal after a testee has joined unilaterally.
  38. For example “the Project is open to all bearing the surname XXXX or similar spellings, and others who share the Y-DNA of one of the genetic families in the project”.
  39. For an example of how an e-NPE identified by DNA can be followed up with genealogical research, see Supplementary Paper No. 8 at [www.dnastudy.clanirwin.org www.dnastudy.clanirwin.org].
  40. For example, if 11/12, 33/37 or 60% TiP scores are used for “acceptable” comparisons of participants sharing the project surname, then a minimum of 37 markers with at least a 36/37 match and a 95% TiP score might be required for e-NPEs, perhaps together with some additional evidence such as a comparison of terminal SNPs. Such stringency is necessary to avoid accepting “false positives”, that is the convergence of two genetically-unrelated haplotypes with different surnames during the genealogical era. If the terminal SNP of a suspected e-NPE (i.e. with a different surname) differs then it is definitely a “false positive”, i.e. does not share a common paternal ancestor during the genealogical era; if it is the same as that of the surname genetic family with whom it “matches”, then it may be a e-NPE, but this is not conclusive as the match may be coincidental, at least until terminal SNPs are not exclusive to a surname.
  41. Publication of a pro-forma explanation saves the administrator from having to “invent” suitable wording each times these circumstances arise. This also helps the participant recognise that he is not the only NPE in the world. And it may be appropriate for the administrator to allow some delay to enable the participant to digest the implications of this background and his particular circumstances before discussing embarking on a detailed discussion
  42. See, for example, groups NB through NR in http://dnastudy.clanirwin.org/latest-results and associated discussion.
  43. An original branch of the surname is a branch whose common ancestor adopted the surname independently of another branch, with no NPE involvement, for example two branches of a surname each represented by descendants of two unrelated blacksmiths, or two unrelated tailors, or two unrelated men from the the same town, or two unrelated men adopting the same personal surname.
  44. The maturity of a surname project may be characterised thus:
      • “young” projects, typically small and recently established, which have a penetration (see Irvine 2010, p4.) of less than, say, 0.05 participants per million surname users.
      • “mature” projects, typically large and well-established, which have a penetration of more than, say, 0.10 participants per million surname users.
  45. For the purpose of this article three types of surname are considered:
      • Single-source surname projects, defined as those known by genealogy and DNA studies to have a single common ancestor, the only exceptions being NPEs. Paradoxically this definition is circular. Examples include Creer, Owston.
      • Plural-source surname projects, defined as those with a few unrelated paternal ancestors. They may include surnames that were once thought to be single-source, and/or locative surnames, such as Irwin, Plant.
      • Multi-source surname projects, defined as those with many, possibly thousands, of unrelated paternal ancestors, such as Black, Jones, Smith, Taylor, Williams
  46. It must be remembered that while “Matches” with some other single surname is indicative of an i-NPE, the absence of such matches does not necessarily indicate an original branch of the surname: the absence may be due to no/few individuals with the “other” surname appearing on the FTDNA database, especially when the “other” surname is not a common, multi-source surname.
  47. Some projects have even identified double NPEs in a lineage.