During the days of Roman occupation in Britain (43AD to 450AD), did the Roman occupants of Britain interbreed and intermarry with the native British population substantially (is not only a couple of cases but repeatedly)? If so, to what extent was the British population also derived from Roman stock?

I have already done some preliminary research and the results seem to be conflicting. One article from the Telegraph says that 1 million Britons have Roman DNA, and then another article states that:

"there is little Roman DNA in the British genetic makeup".

How can both of these articles be correct? I'm particularly interested in contemporary sources if possible.

  • 7
    @Charlie - are you looking from a historical perspective (narrative) or proof (as in genetic results, your reference to the articles)? I ask because they may not agree (consilience). Good question, by the way.
    – J Asia
    Commented Sep 11, 2017 at 17:27
  • @PauloEbermann I'm not purely interested in a genetic perspective; sources from the time would also be handy
    – Charlie
    Commented Sep 11, 2017 at 22:44
  • 12
    Whats your definition of a Roman? Much of the Roman army (and population of the empire) in this period we're from all over Europe; Spaniards, Gauls, Germanic tribes. If by a Roman you mean someone from Rome, then hardly any of the Romans we're actually Roman.
    – user7002
    Commented Sep 12, 2017 at 15:16
  • 3
    @Liam: And conversely, during the several centuries when parts of Britain were part of the Roman Empire, many Britons were Roman citizens. A good many people from other parts of the Empire settled in Britain, or spent time there (and vice versa). To suggest that there could possibly have been no intermarriage (or less formal relationships) between native Britons and these others displays a complete lack of knowledge of human behavior :-)
    – jamesqf
    Commented Sep 12, 2017 at 17:28
  • 4
    @Liam Two of the Legios that invaded Britannia were previously settled in Hispania for a century.
    – roetnig
    Commented Sep 13, 2017 at 6:51

6 Answers 6


The answer to your question is actually to be found in the two articles you have mentioned.

Official figures show that the UK population was 65.6 million in June 2016. A little under 50% of the population is male, although the exact ratio varies by age. This gives a male population of about 32 million.

Your first article is about research into genes carried on the Y-chromosome. Males have one X-chromosome and one Y-chromosome, while females have two X-chromosomes. So, assuming the population used for the study was representative, that would mean that at least 1 million men (out of 32 million) are direct descendants of the "Italian Romans", i.e. the population of the Roman population born in Italy that lived in the UK during the Romano-British period (43-410AD).

That is just over 3% of the population. More would be descended from "Romans" who arrived from other parts of the Roman empire than Italy, and, of course, even more would be descended from the female children of immigrants to the UK during the Roman period.

It is worth noting that this particular study focused only on the incidence of the R1b-S28 haplotype. Caution should be exercised when drawing conclusions from the data (see my observation below about taking your research further)).

The study was carried out by the company BritainsDNA.

The second article is talking about data from the People of the British Isles DNA project. This study looked at both DNA from Y-chromosomes (passed only along the male line) and also mitochondrial-DNA which is only passed along the female line. This project does have some selection bias since it only analysed the DNA of people from rural areas of the UK, whose four grandparents were all born within 80km of each other.

If we assume that a similar percentage of the female UK population is descended from "Italian Romans" as with the male (i.e. about 3.1%), then, by the Law of Total Probability, almost 97% of the population are not descended from "Italian Romans".

Now, obviously this is based on supposition, rather than data (you'll have noted I used the word "assume" in regard to the percentage of the female population above!), but this accounts for the quote in the article:

"There is also little Roman DNA in the British genetic make-up."

In fairness, journalists - even on the "broadsheets" - are not generally expected to be experts in genetics (or statistics!). More information about the People of the British Isles DNA project can be found on their website.

In terms of contemporary sources, obviously we have the funerary inscriptions, which often mention wives and children. Where names of wives are preserved they are often "Celtic", rather than Latin, and the assumption is that they were Britons. Occasionally, even the wife's tribal association may be recorded, and in these cases we are able to state with certainty that the wife was British.

Many of these inscriptions are found close to a vicus associated with a fort (particularly on the province of Britannia's northern frontier). We certainly have inscriptions recording veterans from all over the Roman empire living at these vici. Some of these veterans may well have been born in Italy.

If you are interested in further research into the subject of DNA "haplogroups", including some of the limitations in drawing conclusions from the data, there are a number of good websites available. As is so often the case, Google is your friend here, but two good starting points are Wikipedia and the International Society of Genetic Genealogy Wiki.

To date, much of the research* that has been published in regard to genetic ancestry has come from commercial companies whose business models are predicated on selling DNA testing kits. That may, or may not, influence the way in which the claims have been made. However, I find that it is worth bearing that fact in mind when reading their press-releases.

In this specific case, there is, perhaps, some indication of the degree of scientific rigour to which this research has been subjected in the number of hits you get when you search for Alpine R1b-S28 in Google Scholar.

* Very little has been published in peer-reviewed journals as yet.

  • 20
    "almost 97% of the population are not descended from Italian Romans" -- that's true only if by "descended" you mean "descended in unbroken male line". If you allow descent through a mixed-gender line, it becomes almost certain that a random British person (who does not descend solely from recent immigrants) has some ancestor from your Italian-Roman population. Commented Sep 11, 2017 at 20:46
  • 2
    @HenningMakholm Perhaps (although there are other complicating factors), although I was just talking about the measurable Y-DNA & mtDNA components. In the absence of evidence though, I'd be careful about using terms like "almost certain". However, even if that random person did have an ancestor from the Italian-Roman population, they might not exhibit the relevant DNA haplotypes. Commented Sep 11, 2017 at 21:00
  • 8
    and that's assuming that no migration to the British Isles took place after the Romans left. Something we know is wrong, there was in fact a very high degree of migration into the British Isles, seriously diluting any Roman blood present in the population during the Roman era.
    – jwenting
    Commented Sep 12, 2017 at 10:15
  • 1
    @jk. In theory, those should have been filtered out (along with Italian Scots) by the testing company, BritainsDNA on the basis of the genealogical information supplied with the DNA samples. Commented Sep 12, 2017 at 14:00
  • 2
    @Liam Why would you think that? Do you think that the Vikings or Normans carried a significant proportion of the R1b-S28 haplotype? The Viking and Norman invasions brought different haplotypes that now contribute to the British genetic makeup. Commented Sep 12, 2017 at 15:22

From both a narrative (general history) and scientific (genetics), the answer is No (there was not a lot of interbreeding).

(We get more precise as we go from narrative history to genetics, as shown below -- but science requires certainty which creates some confusion in narrative history).

The narrative history of Roman Britain, was fairly straightforward until the pesky thing called science (genetics, archaeology, etc) started the confusion (with males vs females, slaves vs nobles, true romans vs barbarian romans ... i.e. what do we mean by romans?, which part of Britain?).

Before the science, in terms of (classical) narrative, I believe the short answer is No, there should not be much intermarriage (legal interbreeding). This was the general belief because of several factors:

  • Roman Identity - this was a prized possession and not everyone could become a citizen (via marriage)
  • Rights of Women - was less much than the modern world (i.e. right to vote, etc.) but one crucial aspect of having a mother who was a (barbarian) citizen was her children would also become citizens
  • Rights of Inheritance - property (land and chattel) were not taken lightly and even wills of Roman soldiers had to get approval.

In short, the social structure of Roman society was not to inter-mix with the barbrians of the provinces. Here is a speech by Claudius that did not go down well at the senate because he wanted to admit noble citizens of Gaul to the Senate:

Once upon a time Kings ruled this City, but they were not fated to have home-grown successors. Outsiders took over their rule, foreigners in fact, for when Numa succeeded Romulus he came from the Sabine lands—not far away to be sure, but it made him a foreigner in those days. When Tarquin the Elder succeeded Ancus Marcius, well he was of mixed race, for his father was Demaratus the Corinthian, while his mother was born in Etruscan Tarquinii. She was not a wealthy women, as you might imagine given she had agreed to such an inferior marriage, and for that reason he was unable to hold offi ce at home. But he migrated to Rome, and here was made king.

and (the factors as explained above)(emphasis mine)

During their (peregrini) twenty to twenty-five years of service, soldiers of all kinds naturally formed relationships with local women: but these were not formally marriages and children born to them would not be citizens. Special dispensations allowed soldiers to make wills, but their wives and children had no automatic rights in respect of them. An auxiliary veteran could make his slave a citizen by freeing him, but any children he had fathered before he was discharged would have to join up themselves if they wanted the same status.

Source: Rome: An Empire's Story (Oxford, 2013), p.218 and p. 222, respectively.

(On sources, as per your request in comments) The main sources of (classical) history of Roman Britain should include R. G. Collingwood from 1923, and less old but still excellent, both 1981, Roman Britain by Peter Salway (Oxford) and Roman Britain by Malcolm Tood (Fontana); both books excellently reviewed together by the LRB (worth a read). The latter was also editor of A Companion to Roman Britain (Blackwell, 2004).

The answer from a scientific perspective is a somewhat more involved, and difficult to make a simple and clear-cut statement (as indicated by your articles) because we have to differentiate between male and female lines. But in general, the conclusion is No, as there is very little evidence of interbreeding for both maternal and paternal lines.

According to the world's first genetic archaeologist, Bryan Sykes, a Professor of Human Genetics at the University of Oxford, who conducted extensive DNA research within the British Isles, covering a much longer (earlier) history from pre-history to modern-day (emphasis mine):

The first conclusion, blindingly obvious now I can see it, is that we have in front of us two completely different histories. The maternal and paternal origins of the Isles are different. And that should be no surprise, given the opposing characters of the chroniclers. The matrilineal history of the Isles is both ancient and continuous. I see no reason at all from the results why many of our maternal lineages should not go right back through the millennia to the very first Palaeolithic and Mesolithic settlers who reached our islands around 10,000 years ago. ...

and (for females)

Lastly, I have found a tiny number of very unusual clans in the southern part of England. Two of these are from sub-Saharan Africa, three from Syria or Jordan. These exotic sequences are found only in England, with one exception, and among people with no knowledge of, or family connections with, those distant parts of the world. I think they might be the descendants of Roman slaves, whose lines have kept going through unbroken generations of women. If this was the genetic legacy of the Romans, they have left only the slightest traces on the female side...

and (for males)

I have tried to find Roman Y-chromosomes, but they left very few traces that I can be sure were theirs. Only one very rare patrilineal clan, without even a name, may be the faint echo of the first legions. It is found in southern Europe, including Italy. What makes me think, as well as this link to Italy, that it might be linked to the Romans is that it is entirely restricted to England. There are no traces beyond the borders with Wales or Scotland. There may be others, but as was pointed out, the tradition of recruiting legionaries and auxiliaries from Gaul and other parts of the Empire, as well as from Britannia itself, makes them very difficult to spot among the descendants of later arrivals from the same areas. But true Roman genes are very rare in the Isles.

Source: Blood of the Isles (Penguin, 2006), Chapter 18 (final chapter).

  • "This was the general belief..." is my opinion too but not universal because I have read that some historians, though I cannot recall whom/which group, does not agree with this.
    – J Asia
    Commented Sep 11, 2017 at 19:40
  • 1
    You can't use genetics, as they are invalidated by time. The current population of the British Isles is NOT just descendants of people who lived there during the Roman era. In fact those probably make up only a small percentage of the total.
    – jwenting
    Commented Sep 12, 2017 at 10:17
  • @jwenting - Yes, I have read that too, in Blood of Isles and others (but they go on to use it as guidance). Can I ask, how does it work if it's to be of use to history? None at all or are we saying that it can be used, but only in a limited way?
    – J Asia
    Commented Sep 12, 2017 at 12:08
  • genetics can't really tell you much on its own, it needs to be accompanied by other things like family trees as deep as possible (and we're talking 2000 years here, that's longer than anyone I know of can reliably trace their ancestry). If you have a small isolated community (like 2 amazonian tribes that have next to no contact with the outside world) it becomes a lot more reliable.
    – jwenting
    Commented Sep 12, 2017 at 12:46
  • Agreed. And in which case, I did not do justice by merely providing summaries of Prof Sykes' work (Blood of the Isles) because he did go into this point quite a bit (esp Ch - The Nature of the Evidence). As a rule, I think textual and material sources are more valuable, as compared with genetics and linguistics.
    – J Asia
    Commented Sep 12, 2017 at 13:13

Your question seems to answer itself.

As in yes, per your first source they have interbred to the tune of the UK having 1M or so people with Roman ancestors. Romans intermixed of course - they did so everywhere they settled, just like people do nowadays when they change countries. Populations mix, then like now.

Per your second source, only 1M Britons out of 65M or so have Roman ancestors. That's quite low: it's a few percent of the population. But how low is that really?

Even if you ignore a generous percentage of the total population owing to them not being nationals, descending from recent waves of UK immigrants, descending from earlier settlers or invaders like Saxons, Scots, Picts, Vikings, French, and I'm sure I'm missing many others, not to mention internal migrations from places that didn't intermix with Romans like Scotland, Ireland, and islands up north, and other reasons one might come up with, it'll still be a fraction of the population.

But then flip the problem around and it actually looks quite high, actually. Romans didn't migrate wholesale to Anglia. Consider how awful the place is for a Roman. It's cold and wet by any reasonable Mediterranean standard, and it barely - if at all - grows olive trees or wine grapes. Romans that went there were administrative and military staff first and foremost, and the population of London estimates when Romans were around would suggest they weren't that many to boot. Yet they interbred to the tune of having a whopping 1M descendants. Not bad.

  • Is it also possible that the 1m has spread from a smaller number? EG if one Roman had two children with a Briton woman that would be 2 descendants. If they each had 2 kids that would be 4 descendants, etc etc. If you extrapolate that you get 2^n = 1'000'000 . n would be about 20 which means from 1 Roman in 20 generations that would create 1 million descendants. This logic may be incorrect by genealogical standards though as I don't have knowledge of the subject.
    – Charlie
    Commented Sep 11, 2017 at 18:06
  • 1
    @Charlie: Yeah, sort of. Until the industrial revolution families of 6-10 children were commonplace, and it was fairly common to marry distant (or not so distant) cousins, if only because most people wouldn't travel much further than walking distance from where they'd live. In light of that it's reasonable to assume one Roman ancestor spread its gene in some village and those around it, and maybe a few villages further. But not that further until the industrial revolution, which is more like 8-10 generations back - with lower fertility but the continued trend of breeding with cousins. Commented Sep 11, 2017 at 18:33
  • 5
    @Charlie It is possible, but that 1m is just those that descended down the male line. If your hypothetical Roman had two daughters, then no matter how many descended they have, they won't appear in that 1m estimate. That is because the daughters got one X-chromosome from each of their parents and no Y-chromosome. Also, in the UK it was more usual for daughters to travel to live in their husband's parish ... Commented Sep 11, 2017 at 18:45
  • 1
    @Charlie yes, but some of them also would have migrated out, messing up the numbers even more. IOW genetics is an extremely bad instrument to get any idea about the amount of amorous relations between Romans and native Britons during the Roman era.
    – jwenting
    Commented Sep 12, 2017 at 10:18
  • 1
    This is still failing to answer an important question: what exactly is a Roman? Someone whose ancestors dwelt in the city of Rome since the days of Romulus and Remus? Anyone who had Roman citizenship and lived in the city, including people descended from freedmen whose ancestors were brought from all the lands the Empire conquered? Or anyone in the Empire who attained Roman citizenship, which includes a number of native Britons?
    – jamesqf
    Commented Sep 13, 2017 at 3:41

1 million isn't "a lot", it's only a few percent of the male population of the British Isles. But that's not the worst that's wrong with your assertions.

Not only do you assume implicitly that no migration into or out of the British Isles took place after the Roman era, but also you assume that the number of children of mixed relations between Romans and natives was identical to that of relations between 2 natives, and that that was consistent throughout history.

Thus using genetics to make sweeping statements about social customs of 2000 years ago is not going to work. It can only be used to determine migration patterns in very broad terms. It could for example be used to conclude that yes, Romans visited the British Isles and interbred to some degree with the locals. But even that'd be sketchy as that mixed blood could itself have been imported. For example someone from mixed Roman/Germanic blood immigrates to Britain in the 18th century and has children there. Your genetics do not take that into account.


Your first link is easily answered by asking another question: "Is there even such a thing as Roman DNA, distinct from other European DNA?"

The evidence says almost certainly not. This is a good summary of the actual state of the science. To quote one professor about this kind of DNA testing, "the business is genetic astrology". Haplogroup comparisons may make sense for widely-separated populations, but Europeans emphatically were not widely separated. In the case of Britain, it only became an island roughly 8000 years ago, which is not nearly enough time for significant genetic divergence.

So DNA really isn't a good place to start looking.

  • Cheers. I however want to stress I'm more interested in the history (hence why this SE) than genealogy and hence DNA isn't much of a bother for me.
    – Charlie
    Commented Sep 12, 2017 at 15:15
  • 1
    @Charlie: DNA should be "much of a bother for you" because it could be the single most important thing that proves the history you are trying to establish.
    – Tom Au
    Commented Sep 12, 2017 at 17:27

Admittedly, I don't have any specific anthropological or genetic evidence to prove that Romans and early Britons were interbreeding 1600-2000 years ago.

The majority of English peoples are of Anglo-Saxon ethnic descent; this is to say that when the Anglo-Saxons arrived in Roman "Britannia"-(at the very end of the Roman Empire), both Romans and Celtic Britons were the primary inhabitants.

The Anglo-Saxons were a Germanic tribe originally from the Saxony region of Germany and like many other Germanic tribes at the end of the centuries old Pax Romana, they conquered and settled into a distant land. When the Angles and Saxons initially arrived in Roman colonial Britain 1500 plus years ago, they would have encountered a small-(but powerful) percentage of Mediterranean looking Roman Politicians, Generals and soldiers, as well as a small civilian population working for the larger Roman imperial bureaucracy. However, the indigenous Celtic Briton population residing within Ancient England would have been far larger than the Roman population and the indigenous Celts would have had a distinct anthropological appearance when compared with both the Romans, as well as the Anglo-Saxons. In other words, 1500 years ago, Roman Britain would have had red haired, blue or green eyed, fair skinned Celtic Britons, blonde haired, blue eyed, fair skinned Angles and Saxons, as well as dark haired, brown eyed, olive complexioned Romans.

As for the present population of England, I would say that from a distant view, the majority of contemporary English peoples are primarily of Anglo-Saxon ethnic descent with a small percentage of Celtic ancestry and a smaller percentage of distant Roman ancestry. Occasionally, one will find English persons with a heavy Roman appearance and at times, one will see English persons with a heavy Celtic appearance. However, it was the Anglo-Saxon invasions 1500 plus years ago that transformed the ethno-racial demography, culture and actual name of the country; from Roman Britannia, to the Germanic, "Land of the Angles".....better known as, "England".

Not the answer you're looking for? Browse other questions tagged or ask your own question.