Isotopes and aDNA: teasing apart ancestry versus migration in early medieval England

We are amid an exciting new Scientific Revolution within archaeology, thanks to biomolecular techniques and ‘big data’. The aDNA study featured in this special issue is the first to tackle the post-Roman to pre-Viking period at scale using such data, and large-scale isotopic datasets are also coming to the fore. As John Hines points out earlier in this issue, we are now able to give definitive evidence of individual- and population-level migration and ancestry dynamics, which raises more questions than answers but is nevertheless exciting.

Despite the wealth of information that aDNA can give us, though, to understand this picture fully we need other (bio-)archaeological data to put it in context. Genetic analysis alone cannot confirm if an individual actually migrated during their lifetime, because ancestry does not equate to place of birth, which in turn cannot be equated to identity. Isotopic evidence, however, can fill in the gaps by shedding light on a person’s movements across their lifetime. It is our best proxy for place of origin when studying the ancient dead, and, when combined with aDNA and funerary archaeology, it starts to build a more-complex picture of who migrated to where, from where, when, and how those movements tie into familial relationships, social status, gender, and other aspects of early medieval culture(s). So, what is isotopic analysis?

Introducing isotopes

Isotopes are forms of the same element that differ in their atomic mass, which can be quantified in the lab. Different rocks, plants, animals, and water sources all vary in their isotopic compositions, and for certain elements these variations are predictable and measurable, making them ideal for determining where people and animals were living and what they were eating in the past thanks to the chemical signatures preserved in their bones and teeth.

Here, I will largely be referring to isotopic analysis of human tooth enamel as a proxy for mobility, using strontium and oxygen isotopes (87Sr/86Sr and δ18O). When the enamel of our teeth forms, it locks in oxygen and strontium isotopic signatures that reflect drinking water and underlying bedrock isotopic variability respectively. Once it has been formed, tooth enamel does not remodel, and our teeth themselves develop at regularly defined intervals (I bet most readers can remember when they lost their two front baby teeth with good accuracy). This means that archaeologists can be relatively confident that we have an isotopic snapshot from a few specific years of the drinking water and bedrock where someone was living while the chosen teeth were forming (for example, a second molar gives us insights into when the individual was around 2.5-8 years old). Therefore, it is also possible to track mobility over individual lifetimes by analysing multiple teeth from a single person to see how their signatures change – and with large enough sample sizes we can even look at migration patterns at site, regional, and continental scales.

A key caveat about this kind of data, though, is that it cannot pinpoint a person’s exact origins, only rule in/out certain areas based on chemical matches/mismatches and suggest several possible regions of origin. Furthermore, what we call the ‘brewing and stewing’ phenomenon can have an impact on results, where someone can look like they’re from an isotopically ‘warmer’ region when they have instead been consuming large amounts of ‘brewed and stewed’ liquids such as stews or alcohol.

A great case study to illustrate this is the analysis of Richard III (see CA 277). Incomers from cooler climes and distinctive geologies (like the Fennoscandian shield, which lies beneath much of Scandinavia and western Russia), though, are more easily identifiable. Excitingly, there are some large isotopic studies on early medieval Europe on the horizon, and I will briefly summarise some of this research here to compare with the aDNA work.

What does the isotopic data tell us?

The picture from isotopic data is clear – people have been on the move around Europe for millennia, which isn’t a surprise. These movements fluctuate through time, and there are highly nuanced patterns that we are yet fully to tease apart. For southern Britain, there is a pattern of continuous migration from at least the Roman period to well past the Norman Conquest, especially in eastern regions. There is clear evidence for people settling in southern Britain from the west (of Britain) in cemeteries further east, and for people coming from all regions of continental Europe and further afield, from Scandinavia to North Africa.

The map shown on p.45 demonstrates proportions of ‘cold’ non-locals, ‘warm’ non-locals, and locals in southern Britain between c.AD 450-1100, based on oxygen isotopic data from 659 individuals. The map alongside it shows the proportions of non-locals versus locals based on oxygen and strontium data for just the 5th to 8th centuries AD, based on 525 individuals. We can see that there are more non-locals in coastal regions, especially non-locals with ‘cooler’ origins. The gradation of ‘warmer’ signals is particularly interesting; with the aforementioned caveats in mind, this suggests movements from Ireland and western Britain. The overall proportions of locals versus non-locals in the second map is based on a more nuanced but smaller dataset than the first. It suggests fewer first-generation migrants than the aDNA study alludes to, and to me this shows the long-term settlement of people with CNE ancestry in these regions, reflected in smaller numbers of isotopic ‘non-locals’, meaning that many of these people were descendants of earlier incomers.

Another key point of difference with the story from aDNA is the gendered patterning in the isotopic data. I cannot do justice to the intricacies of the regional and gendered trends here, but two headlines are interesting to ponder in light of the aDNA patterning. First, 39% of women and 29% of men in southern Britain are definitively ‘non-local’ to their place of burial across the 5th to 8th centuries AD. Second, men are more likely to have come from overseas compared to women, whereas women show signs of some long-distance travel and more localised intraregional movement.

Can we be more specific about where some of these incomers may be from? Terminology in aDNA and isotopic papers is often deliberately vague given the caveats of our techniques. However, as the maps for four cemetery sites – Empingham 94, Wasperton 1, Adwick-le-Street, and West Heslerton – on the opposite page show, recent advances in computer modelling of isotopic data now allow us to speculate on origins in a more refined way. The four individuals used as examples here showcase the wide variety of childhood origins possible in early medieval cemeteries, which highlights the connectedness and long-distance mobility of the period.

In the table on p.48, I have highlighted a handful of individuals from Buckland, near Dover, whom I analysed (oxygen data only) and who are also part of the aDNA study. First, let’s look at a family group – Graves 346, 347, and 375. Graves 346 and 375 are second-degree relatives (e.g. aunt/uncle/niece/nephew, grandparent/grandchild, or half-sibling) and 346 is 347’s father. Analysis of their second molars tells us that all three were living outside Kent from c.2.5 to 8 years of age, and all have 100% CNE ancestry (the image shown on the left gives suggestions for where they might have been living during that time). The teenager in Grave 347 moved to Kent after the age of eight, and lived there for a maximum of eight years before she died (if she was a maximum of 16 years old when she died). The similarity of her origin map with her relative in Grave 375 suggests that they might have moved together as a family unit, although her father in Grave 346 might have grown up somewhere slightly different to his two younger relatives.

Next, we have two people who were probably born locally: Graves 323 and 426. Both have close relatives in the cemetery (although these individuals haven’t been isotopically studied yet), but they have different ancestry profiles: Grave 323 is 100% CNE ancestry; 426 has predominantly WBI ancestry. You’ll see from the maps above that they have an equally likely chance of growing up in Kent as they do for large parts of Europe – so are these people ‘locals’, are they migrants from other green areas on the maps, or are they the descendants of incomers? I’m hopeful that including more isotopic data in future and syncing it up with their family trees will help us narrow down the options, as we have done for the other three individuals above.

Biomolecular futures

Taken together, the two lines of biomolecular evidence discussed here match up well in terms of the geographical clines within Britain – particularly with the dominance of Continental incomers into eastern regions, and a greater proportion of ‘locals’ and WBI ancestry in the more westerly cemeteries. Both datasets also agree that there was continuous interaction between Britain and the Continent, which resulted in integrated and multi-origin communities.

For regions and sites that have distinctions in burial practice based on ancestry, though, claims that this is based on immigrant versus local backgrounds must be nuanced further when viewed in the light of the isotopic data. These funerary delineations seem to be based on kinship (with the implication of memories of migration, with most people being locally born) rather than recent incomer status. It is a subtle distinction but an important one to consider, and the regionality of this funerary separation on ancestry needs further research.

The need for more Roman-period data, both isotopic and genetic, is clear, as others have already pointed out, and the necessity of synchronising these biomolecular datasets all the more pressing. Where larger sample sizes were achieved for aDNA (e.g. in East Anglia), there is a paucity of enamel isotopic data, and vice versa. Both methodologies have a hurdle to overcome with the large cremation cemeteries in the east of the country, which (until recently) have posed serious methodological issues for biomolecular research. However, advances in techniques mean that strontium work is now possible on these sites, and we most likely have big surprises in store at cemeteries like Spong Hill. So this is just the beginning. With growing biomolecular datasets, we can now move beyond the binary arguments John Hines described in his introduction and theorise in a more-nuanced manner about early medieval population movements and community dynamics.

ALL IMAGES: Sam Leggett.