Contextualising Edix Hill: First-Pandemic Plague and Britain^*

Abstract:

In 2019, a team of scientists led by Marcel Keller published incontrovertible evidence that Yersinia pestis, the bacterium responsible for plague, was present in sixth-century Britain.¹ Multiple samples from individuals buried in the early medieval cemetery at Edix Hill, just outside Cambridge in eastern England, had tested positive for Y. pestis (Fig. 1). This discovery was definitive proof of the bacterium’s presence and lethality. While false negatives are still common in state-of-the-art palaeogenetics, false positives are not, so this palaeogenomic detection unambiguously attests to an active infection at the time of these individuals’ deaths.² The presence of multiple plague genomes at Edix Hill is thus significant and must reflect that plague killed there.

Figure 1:

Map of Britain and Ireland with locations mentioned in the text labelled. References to ‘western Britain’ refer loosely to the regions which are today western lowland Scotland, Cumbria, Wales, Devonshire and Cornwall, and ‘eastern’ refers to the lowland regions which became England. I intentionally leave the border between east and west vague so as not to project modern borders into the post-Roman period and to reflect scholars’ fuzzy knowledge of the extent of English cultural zones in the fifth to sixth centuries. The geographical distinction most relevant to plague ecology is the topographic one between the lowland east of Britain and the upland west and north, for which see the hillshade on the map. Figure by the author.

Open in new tabDownload slide

The most startling and unexpected factor of Keller et al.’s discovery was the genetic relationship between the Y. pestis recovered from Edix Hill and that recovered from other early medieval sites on the Continent in modern-day Spain, France and Germany. If plague had spread from the eastern Mediterranean, where surviving texts first report it at Pelusium in 541 CE, overland across Europe to Britain, we would expect the plague at Edix Hill to be derivative of that recovered from the Continent, and therefore farther down the First Pandemic branch of plague’s phylogenetic, ‘family’, tree.³ Yet the reverse is true. More early medieval plague palaeogenomes will be reconstructed, potentially including earlier genomes from the Continent, but based on current data it is striking that the Edix Hill Y. pestis branches off before multiple early Continental first-pandemic genomes. This relationship leaves room to suggest that plague did not spread to Britain overland from Constantinople, as generations of scholars expected.⁴

Indeed, this genetic relationship, coupled with recent radiocarbon dating which reveals that one Edix Hill victim most likely died between 416 and 541 CE, has led scholars to propose that the plague found in eastern Britain may pre-date the onset of the famous ‘Justinianic’ outbreak of the 540s in the eastern Mediterranean.⁵ Efforts to date the remains more precisely are underway, but the data available insinuate that Y. pestis reached Britain either before or concurrently with its appearance in Constantinople. The discovery by Keller and his colleagues thus does much more than prove plague came to, and killed at, Edix Hill—it has the potential to challenge scholars’ knowledge of the origins, spread and regional impact of the first plague pandemic, their understanding of Britain’s place in post-imperial Europe and the historiographical role of extra-Mediterranean locations in scholarship on the post-Roman period.

Existing efforts to incorporate the Edix Hill discovery into our understanding of the First Pandemic seek to determine its significance for our understanding of sixth-century Continental, and particularly Mediterranean, plague.⁶ These responses have originated from both sides of the major historiographical divide in sixth-century plague scholarship, which centres on the relative historical and demographic significance of the pandemic. One argument employs the Edix Hill data to underscore the wide scope and destructiveness of the ‘Justinianic Plague’, arguing that the pandemic’s presence in a place assumed to be as peripheral as eastern Britain underscores its magnitude and rapid, ubiquitous spread.⁷ Kyle Harper best articulates this assumption, arguing that ‘if the plague made it here, in the very far west, in the boondocks of Anglo-Saxon England, then it makes a strong case that the First Pandemic was able to spread far and wide’.⁸ While some scholars espousing this view contend that the British plague may be pre-Justinianic, they do not posit how plague reached eastern Britain, nor do they confront the implications of its interpretation for our understanding of Britain’s place in post-imperial Europe or the Mediterranean centrism of first-pandemic scholarship. Another argument has the British plague travelling directly from the Mediterranean and then southwards into Gaul, yet omits insular annalistic evidence pertinent to its theory and bases its argument exclusively on Mediterranean sources.⁹ A third view dismisses the find’s importance, arguing that the long-accepted presence of plague in seventh-century England minimises the significance of its discovery there a century earlier and probably via a different transmission route.¹⁰

Examining the Edix Hill find in the context of British textual sources, historiography and archaeology suggests a different interpretation of how plague reached Britain and how it behaved once it did, revealing that Y. pestis may have reached the island at least two distinct times early in the First Pandemic via two separate transmission routes and that its earliest outbreaks there were probably less widespread and devastating than those in the eastern Mediterranean. The first of these introductions could have been quite early—before or concurrent with the well-documented eastern Mediterranean outbreak of the 540s—and may have occurred independently of the transmission of plague to Pelusium and Constantinople.

Recentring the British context also reveals that the historiographical implications of the Edix Hill discovery are profound and wide-reaching. Scholars’ treatment of the Edix Hill find reveals fundamental issues within the subfields of early medieval British history and first-pandemic plague studies. British specialists’ neglect of the Edix Hill discovery demonstrates how the insularity of post-imperial British history and archaeology has isolated the subfield from broader scholarly debates and that this isolation has damaged scholars’ understanding of larger historical processes which involved Britain. Likewise, plague historians’ attempts to incorporate the find underscore issues with the field’s interdisciplinarity and failure to consider the best uses of textual and palaeoscientific evidence. They also reveal the subfield’s detrimental geographical generalism, Mediterranean centrism and neglect of the extra-Mediterranean ‘periphery’ and its scholars—geographical hierarchies which grow out of an understanding of the pandemic as a Mediterranean ‘Justinianic Plague’. Prior treatments of the Edix Hill evidence also underscore how the vitriolic and polarising historiographical divide among ‘maximalist’ and ‘minimalist’ interpretations of the First Pandemic has led to unproductive interpretations of new data. The local context of the Edix Hill genomes instead reveals the need for more local and regional studies to ameliorate the disconnect in many regions between plague studies and their local histories and to create a new, global picture of the First Pandemic from the ground up, which may look quite different from our current understanding of it as a monolithic, Constantinople-centred ‘Justinianic Plague’.¹¹

More broadly, the discovery should urge scholars of post-imperial European history to reconsider the ‘core–periphery’ model which grows out of the paradigm of Late Antiquity. This model encourages scholars to see the textually rich eastern Mediterranean as central and normative in this period, leading many to smear data and experiences from it to geographically ‘peripheral’ regions such as Britain, as prior interpretations of the Edix Hill findings evidence. Conversely, the paradigm leads scholars to resist using evidence and frameworks from ‘peripheral’ locations to inform their understanding of the ‘core’ or the whole. But the Edix Hill genomes intimate that these ‘peripheral’ areas were perhaps as keyed in to Eurasian trade networks as the ‘core’ eastern Mediterranean and may have as much to tell us about pan-European phenomena, such as first-pandemic plague, as the textually rich Mediterranean ‘core’.

This article confronts these issues by contextualising and interrogating the Edix Hill find’s local and broader significance. In order to resist the eastern Mediterranean centrality and normativity which the term ‘Justinianic Plague’ implies, it refers instead to the ‘initial outbreak of the First Plague Pandemic’ to describe the Y. pestis outbreaks which occurred in Europe and the eastern Mediterranean in the 540s CE and more specifically to the ‘Edix Hill outbreak’ to refer to the roughly dated Y. pestis recovered from eastern Britain. It begins by presenting a historiography of early medieval plague in Britain and a summary of current scientific knowledge of the topic to situate the discovery, contend that the discovery marks a significant shift in our understanding of early British plague, and clarify what palaeoscientific data can and cannot tell us about plague’s presence in and its impact on Britain. It then posits a theory as to how plague reached Britain so early which is predicated on the developing Edix Hill evidence and on recent historical and archaeological scholarship.¹² It argues that this evidence suggests that first-pandemic Y. pestis was transmitted to Britain at least two distinct times. This interpretation relies on eastern Britain participating in broadly Eurasian trade networks without Mediterranean mediation and underscores how differently plague is likely to have spread and killed in the rural, sparsely populated Latin West and the urbanised, densely populated Byzantine Empire. The article then reflects on why British sources have been omitted from or misused in plague scholarship, arguing that this trend speaks to British history’s damaging insularity and to methodological issues within plague studies. It concludes by connecting this issue to wider geographical hierarchies within late antique scholarship and suggests ways in which future plague scholarship can respond to these challenges.

I

Most historians for at least a century have argued that first-pandemic plague reached western Britain in the sixth century, identifying it with the ‘mortalitas magna’ recorded in the Annales Cambriae and dated to 547 or 549 CE.¹³ It is, however, unclear exactly when this ‘great mortality’ occurred; all CE dates for insular annal entries are editorial, as the manuscripts do not employ an externally verifiable dating system.¹⁴ The majority of twentieth- and twenty-first-century publications on sixth-century British disease have nevertheless followed the annals in assuming the First Pandemic struck what is now Wales in the 540s, but the few publications that have challenged a diagnosis of Y. pestis for sixth-century British epidemics have had an influence disproportionate to their number. Although attempts to argue that plague was unknown in sixth-century Britain have proven incorrect, those studies have shaped the geographical, linguistic and epidemiological assumptions of all subsequent work on the topic. It is therefore worth interrogating the source of those assumptions so that scholars can consciously decide whether to accept them going forwards.

Most notably, J.F.D. Shrewsbury argued in 1949 that the disappearance of rats from post-Roman Britain, already recognised by his time, prevented Y. pestis from reaching the island until the fourteenth century and that smallpox caused the earlier outbreaks more commonly identified as plague.¹⁵ Shrewsbury’s retrospective diagnosis rested on the name Y Fad Felen, ‘The Yellow Plague’, given to these outbreaks in later Welsh sources, which date to around the twelfth century.¹⁶ He argued that since jaundice is not one of plague’s stereotypical symptoms, the pathogen responsible for any epidemic called ‘yellow’ could not have been Y. pestis.¹⁷

In focusing on jaundice as a symptom of early medieval British epidemics, Shrewsbury followed William MacArthur’s influential examination of outbreaks recorded in the Irish Annals, which appeared only a few months before Shrewsbury’s study. MacArthur argued that the first epidemic the Annals record, mysteriously called blefed—a word which nobody has been able to translate or explain conclusively—and dated to between 540 and 545 CE, was indeed plague which also struck Britain.¹⁸ However, based on the assumption that Y. pestis does not cause jaundice, MacArthur identified later Irish outbreaks in c.556 and 665, which were termed buide chonaill (buide meaning ‘yellow’, though scholars are less sure about chonaill), as spirochetal relapsing fever.¹⁹ MacArthur believed that the fever was triggered, as in nineteenth-century Ireland, by famine, which he proposed must have followed a devastating outbreak of Y. pestis in the 540s.²⁰ He thus argued that Ireland experienced plague, and the resulting mortality caused famine, which in turn led to outbreaks of buide chonaill. MacArthur also applied his argument to Britain, drawing, as did Shrewsbury, on the twelfth-century Welsh name ‘Yellow Plague’ to justify his non-yersinial diagnosis of all British outbreaks preceding and following the mortalitas of the late 540s.²¹

These scholars’ reliance on the Irish Annals to illuminate events in Britain is sound because the Annals’ earliest identifiable stratum was probably recorded at the abbey of Iona in northern Britain from around 563 CE (Fig. 1).²² They were thus kept for a time in Britain in a monastery which housed British as well as Irish monks, and they include information pertaining to both islands.²³ Furthermore, while the annals are probably not quite contemporary to the outbreaks of the 540s and 550s, the monks who wrote the accounts of blefed and buide chonaill may themselves have experienced the epidemics or relied on contemporary written reports.²⁴ The Iona stratum of the Irish Annals survives piecemeal in later continuations, including the Annals of Ulster and the various Clonmacnoise-group annals. It also serves as the source for some of the early Annales Cambriae entries, including the 547/9 mortalitas entry, which was incorporated into the Welsh annals in the tenth century.²⁵

The Irish Annals are thus the closest we have to a contemporary account of sixth-century plague in Britain, but they have limitations. It can often be difficult to tell where their events occurred. The early annals often adopt Iona’s geographical perspective, but it is unclear whether this focus also applies to probably retrospective annals of events which occurred before the monastery’s foundation.²⁶ The Annals record the deaths of several Irish monastics, including St Columba’s mentor, Mobhí Clárainech, from blefed and buide chonaill, which suggests the outbreaks reached Ireland, particularly Leinster (Fig. 1).²⁷ It is impossible to tell from the Annals’ evidence whether these epidemics struck Britain as well. The incorporation of the entry on the 547 or 549 mortalitas magna into the Welsh Annals reveals that their tenth-century compiler believed the outbreak had reached Wales, but his evidence may not be reliable. The compiler amended the mortalitas annal to record the death of the sixth-century Welsh king Maelgwn Gwynedd who, according to a tradition most strongly attested from the twelfth century, died of ‘Yellow Plague’ in around 549 CE.²⁸ The amended annal’s inclusion is thus only evidence that this tradition of Maelgwn’s death existed by the tenth century; it tells us nothing definite of sixth-century reality. There is thus no firm evidence of the outbreak’s geographical distribution beyond its presence in eastern Ireland.

The Irish Annals, furthermore, do not employ the word ‘yellow’ uniformly. The c.556 outbreak is labelled buide in some of the surviving continuations of the Iona Chronicle, but is called cromm or cron (which could mean ‘crooked’, but Ann Dooley took it to be crón, meaning ‘dark brown’ or ‘red’), in others.²⁹ The c.665 buide chonaill is uniformly buide across the surviving annals, though it occurred nearly a century later, so it is impossible to identify it confidently as the same disease recorded in the 550s, particularly since jaundice is such a non-specific symptom.³⁰ It is thus unwise to argue from its name that the earlier outbreak was predominantly buide rather than crón. We do not know either way. The name of the c.665 buide chonaill is also widely considered to be a later interpolation, and some scholars go so far as to contend that all vernacular phrases in the early, Latin-language entries of the Irish Annals are later additions, so it is possible that the ‘yellowness’ of these epidemics is a later tradition in the Irish as well as the Welsh evidence.³¹

The assumptions which underpinned MacArthur and Shrewsbury’s influential contentions are thus problematic. It is troubling that their argument against a diagnosis of plague for sixth-century epidemics hinged on a name which is not attested in Britain until six centuries after the outbreaks, and which is uncertain in the more contemporary Irish evidence.³² It is furthermore uncertain that the name ‘Yellow Plague’ referred to the disease’s symptomatology. Few hold that the name ‘Black Death’, for instance, is based on any tendency of Y. pestis to affect its victims’ complexion, and the ‘Yellow Plague’ may also not have turned its sufferers yellow.³³ Yet, even if twelfth-century sources did record an authentic, British, sixth-century name for the disease which described jaundice as one of its symptoms, Y. pestis could still be the pathogen responsible for it. Sepsis, which can result from bubonic or septicaemic plague, can cause yellowing.³⁴ Indeed, it was first remarked that the name Yellow Plague is ‘not incompatible with the hypothesis of bubo-plague’ in 1891.³⁵ The remainder of this article thus proceeds on the premise that both the blefed and buide/cromm chonaill outbreaks, as well as the unnamed mortalitates recorded in the Irish and Welsh annals c.545/9, could be first-pandemic plague. It is, however, important to underscore that we lack palaeogenomic confirmation that Y. pestis reached the west of Britain or Ireland at all in this period, and it is equally possible that another pathogen (or pathogens) caused any or all of these outbreaks.³⁶

But that pathogen was probably not Variola, the virus responsible for smallpox. Shrewsbury’s identification of smallpox with ‘Yellow Plague’ carries its own host of issues. MacArthur challenged Shrewsbury’s contention that smallpox victims appear yellow at all, but, more crucially, scholars do not know when the disease we recognise today as smallpox arose or arrived in Britain.³⁷ Current Variola palaeogenomics suggest that smallpox could have existed as early as the fifth or sixth century, but scholars are divided on the date of the disease’s earliest literary description, and it is likely that a Variola virus capable of causing smallpox, as we know it, evolved later.³⁸ The earliest Variola palaeogenome from Britain dates to between 880 and 1100 CE, and thus to centuries later than the ‘Yellow Plague’.³⁹ It is moreover genetically distinct from the Variola associated with smallpox in ways that may have affected its symptoms, epidemiology and virulence.

Shrewsbury’s argument has nevertheless been profoundly influential, particularly because it convinced John Maddicott, author of the only major study of plague in early medieval Britain between 1997 and 2018.⁴⁰ Maddicott argued that first-pandemic plague reached Britain, but not until 664, contending that the earlier British epidemics recorded by the post-Roman writer Gildas and in the Welsh and Irish Annals probably described smallpox and may not have spread east of modern Wales.⁴¹ He did believe that the 664 and later outbreaks mentioned in seventh-century sources were yersinial based on Bede’s description of the 664 outbreak’s symptoms, seasonality and apparent spread from the coast of Britain to its interior.⁴² Maddicott’s study has remained virtually unchallenged in the decades since its publication and has been paradigmatic for British plague studies. Yet its argument did not convince everybody. Although Maddicott’s article has been widely cited, several later publications argue or imply that plague reached Britain in the sixth century.⁴³ These studies, however, often adopt Shrewsbury’s sixth-century geographical assumptions and combine them with Maddicott’s seventh-century diagnosis and epidemiology—an unsubstantiated combination, but one which has underpinned efforts to understand the plague at Edix Hill. The pre-seventh-century date of the Edix Hill ancient DNA [hereafter aDNA] is thus jarring but not utterly unforeseen, in a British historiographical context. What is wholly unexpected is the samples’ location within Britain at this date.

All surviving, near-contemporary accounts of epidemics in sixth-century Britain have a western perspective, describing outbreaks in modern Ireland and Wales, so twentieth-century scholars assumed that the outbreaks they described did not reach, or at least did not kill widely in, the eastern territories which are now England.⁴⁴ This assumption allowed historians of the early to mid-twentieth century to assign to plague a causal role in the expansion of ‘Germanic’ material culture and political control in eastern Britain.⁴⁵ This argument tied into twentieth-century scholars’ belief that the adventus Saxonum, the process by which eastern Britain became culturally and politically ‘English’, was a sudden, violent affair, involving early English warbands’ extermination of the native population and conquest and colonisation of large swathes of territory.⁴⁶

In the context of this paradigm, scholars once argued that plague weakened the native British and made them more susceptible or less resilient to ‘Germanic’ attacks.⁴⁷ They explained the epidemic’s apparent lack of impact on the early English by arguing that more densely populated ‘Anglo-Saxon’ settlements would somehow have been less susceptible to the spread of the disease than less dense, native British settlements.⁴⁸ Attention to plague ecology and epidemiology shows this line of reasoning to be erroneous, as higher population density instead seems to be correlated with higher plague mortality.⁴⁹ Alternatively, some scholars portrayed sixth-century Britain as so ethnically segregated that the two populations had virtually no contact with one another, which, they argued, would have prevented transmission of disease from the west to the east.⁵⁰ Scholars of this period also named plague as a cause of the decline of towns across Britain and of the post-Roman migrations from Britain to Brittany.⁵¹ There were challenges to the consensus of plague’s destructiveness and historical agency in Britain, but these were the exception rather than the rule.⁵²

Historians and archaeologists have since largely abandoned the ‘conquest’ paradigm in favour of one that views the rise of early English culture and political dominance in Britain as a slower process which incorporated the native population rather than replacing it.⁵³ This is a welcome development which has prompted more nuanced explanations for the adventus, but it has had an unintended side effect: most modern scholars altogether ignore the presence and possible impact of epidemic disease in Britain in this period, and what scholarship exists on post-Roman British disease now focuses on endemic, local disease burdens, such as the possible presence of malaria in the east of England.⁵⁴ Once there was no longer empirical evidence for catastrophe, discussion of epidemic disease ceased, though epidemics can be non-catastrophic or cause archaeologically invisible catastrophes. Recent scholarship on Britain which does mention plague refers to it only briefly, though the long-standing belief in the disease’s demographic destructiveness has persisted, even as confidence in its historical agency has wavered.⁵⁵

Reappraisal of the adventus Saxonum also led to significant reconsideration of the texts which give witness to post-imperial British epidemics. Gildas’s De excidio Britonum is the only surviving narrative source from Britain in this period and may include the earliest record of a British epidemic.⁵⁶ Transformative scholarship on the work, beginning in the 1980s, revealed the text’s inability to serve as an objective historical source and advanced our understanding of how to interpret and employ Gildas’s testimony effectively.⁵⁷ These studies have underscored that historians must employ extreme caution when working with Gildas. The De excidio is not securely datable or locatable—scholars of Gildas concur that we cannot definitively date the text’s composition more narrowly than to the late fifth or early sixth century, nor can we determine its author’s geographical perspective.⁵⁸ The work is moreover notoriously difficult to interpret. It lacks a firm internal chronology which would allow us to date its epidemic with any precision.⁵⁹ The work’s homiletic purpose also led Gildas to exaggerate the impact of the disasters he portrayed and to employ myriad biblical references. His ‘pestifera lues’ could thus have been a biblical metaphor, a medical reality or some combination of the two.⁶⁰ Plague scholars must therefore tread lightly and consider the De excidio’s vast, informative historiography when examining Gildas’s epidemic as a possible early textual account of plague in Britain. Gildas does not provide unambiguous evidence for any event he describes.

I. i

The Y. pestis palaeogenomes which have been recovered from Edix Hill also have source-critical issues which are crucial to consider when employing them to draw historical conclusions. There are some historical questions which plague aDNA can answer clearly on its own and others which it can only answer when combined with other data, but there are topics to which it cannot speak. The discovery of Y. pestis at Edix Hill confirms that first-pandemic plague was present in East Anglia. Scholars must, however, turn to other data and theory to tell us when plague struck Edix Hill and what local effects it had.

To determine when people infected with plague died, scholars have mostly relied on archaeological and radiocarbon dating of their graves and skeletons, though some scholars also seek to date genomes themselves by their molecular composition, on the basis of where they fit in relation to other recovered palaeogenomes in plague’s phylogeny. None of these dating techniques is precise, and each has its own methodological issues. Archaeological dating has relied on grave goods, if present, which cannot usually be narrowly dated. Radiocarbon dates, on the other hand, must be correctly calibrated to a regionally specific ‘calibration curve’, which is produced by carbon dating tree rings whose age can be independently verified.⁶¹ Unfortunately, there is a ‘plateau’ between c.420 and 530 CE—a flattening of the curve which makes carbon date ranges for this period broad and uncertain, as it causes mathematical models to return multiple possible dates.⁶² This plateau explains why the most precise carbon dates currently available for Edix Hill plague victims encompass century-and-a-half-long spans of 416–541 and 474–637 and why these differing date spans do not firmly establish which victim died first.⁶³

Scientists can use Bayesian modelling to estimate the chronology of samples within a plateau. This process takes external archaeological or historical information, such as the samples’ stratigraphy and grave goods or the date of the First Pandemic’s Mediterranean emergence, into account.⁶⁴ But this methodology has a downside: by relying on external information, Bayesian modelling produces a non-independent date, and thus cannot be used to corroborate that external information. A small sample size, unreliable external data or a mathematical error could also skew the results.⁶⁵ The dating of plague genomes is thus usually much less precise and certain than textual dating, and it cannot always corroborate the dates texts provide.

Well dated or not, Y. pestis palaeogenomes themselves can speak to the pathogen’s spread. Comparing sequenced samples allows scholars to guess at the pathogen’s path of dissemination. However, the differences between palaeogenomes are often minute (some first-pandemic sequences are seemingly identical), and specialists cannot reliably reconstruct the time, let alone the number of victims, that elapsed between mutations.⁶⁶ Analysis of recovered plague aDNA thus cannot reveal with precision where, when or why plague genomes evolved as they did or how much time passed between divergences.

The positions of published first-pandemic plague genomes relative to one another on the Y. pestis family tree suggest that first-pandemic plague was not transmitted overland, from Constantinople to Gaul to eastern Britain (Fig. 2), as often previously supposed. Had that happened, the Edix Hill genomes would presumably sit farther away from the base of the first-pandemic’s phylogenetic tree than the many early Continental strains so far uncovered, and they do not.⁶⁷ While present genomic data intimates that the accepted theory could be wrong, it is less indicative of which theory of plague’s transmission to Britain is correct. Critical gaps in our knowledge make many interpretations of the data possible. There is, for instance, currently no published early first-pandemic plague aDNA from the eastern Mediterranean, which obscures the roles of Constantinople or Alexandria in plague’s route to Edix Hill.⁶⁸ The genetic relationships between recovered samples could reflect plague’s spread to Britain by ship directly from Constantinople, as Michael McCormick has argued, but they could also indicate that plague bypassed Constantinople altogether on its way to Britain, as Peter Sarris has suggested.⁶⁹ Long-held beliefs about how first-pandemic plague spread are thus not grounded in current palaeogenomics but are instead based on assumptions of the Mediterranean’s centrality in all post-imperial Eurasian trade networks and on faith in Procopius’s claim that the first outbreak of the pandemic was at Pelusium in 541.⁷⁰ Yet even had Procopius been correct to the best of his knowledge, he may have had no way of knowing if plague had already or concurrently appeared in rural northern Europe.⁷¹

Figure 2:

Maximum-likelihood phylogenetic tree for the first-pandemic branch of Y. pestis (0.ANT4) with genetic difference between strains on the x-axis (rather than time elapsed). Note how the British plague strains branch off before the Continental ones. The dotted lines indicate two possible placements for EDI064 (the second Edix Hill plague genome published by Guellil and colleagues in 2022 and radiocarbon-dated to 416–541 CE); it may be identical to EDI001.A (the genome published by Keller and colleagues in 2019 and dated to 474–637 CE) or directly derived from it. The genome was not high-resolution enough to know for sure, and the figure reflects this uncertainty. This tree will grow as more first-pandemic plague genomes are recovered. Figure adapted from Guellil et al., ‘Haemophilus influenzae’, fig. 3C (p. 8), by Marcel Keller, under a Creative Commons BY 4.0 licence (https://creativecommons.org/licenses/by/4.0/) and reproduced with his permission.

Open in new tabDownload slide

It is also possible that plague outbreaks prior to 541 went unnoticed by Mediterranean writers because they were not severe enough to draw the attention of distant observers, or because they occurred where Mediterranean writers were not expecting or would not have noticed them. On the basis of present data, it remains possible that first-pandemic plague was introduced not only or at all via Egypt or Ethiopia, as long suspected, but from elsewhere.⁷² Palaeogenetics also cannot speak to the population mortality of individual plague epidemics, as plague genomics do not reveal how many people died in an outbreak in addition to the individual from whom the genome was recovered. The genetic stability of Y. pestis also applies to the factors which influence its virulence—those have remained essentially unchanged throughout the Common Era.⁷³ Yet, plague has behaved differently in different times and places over the last several millennia.⁷⁴ Since factors other than plague’s genes explain its varying demographic toll, the presence of Y. pestis aDNA in a region does not automatically evidence a catastrophic, Black Death-scale epidemic tragedy there or in adjacent regions, though it can reflect one.⁷⁵

Identifying plague aDNA in human remains also does not indicate how plague manifested in the individuals it killed. Y. pestis symptoms vary depending on a victim’s initial exposure to the bacterium.⁷⁶ People bitten by plague-carrying ectoparasites develop bubonic plague, while people who inhale plague-ridden respiratory droplets catch pneumonic plague. It is also possible to contract septicaemic or gastrointestinal plague via contact with or consumption of infected blood or meat.⁷⁷ These manifestations have strikingly different symptoms, and it may have been impossible for contemporaries to apprehend that they had a single cause. They are also spread in different ways and thus not equally contagious. Yet, only texts can speak to symptoms, which in turn speak to transmissibility.⁷⁸ In areas such as Britain, where there are no narrative descriptions of symptoms, scholars can only guess at by what means and how widely plague spread.

Scholars tend to assume that first-pandemic plague was transmitted by rodents, particularly black rats (Rattus rattus), but there is no hard evidence that it was.⁷⁹Y. pestis often reaches human populations via ‘commensal’ rodents such as rats.⁸⁰ But it is impossible to reconstruct how, when, where or why spillover into commensal rodent populations occurred or whether plague continued to be spread by rodents once it infiltrated human populations. It may then have shifted to human-to-human pneumonic transmission or transmission via human fleas or lice.⁸¹ Transmission methods could have differed in various regions plague touched. Zooarchaeologists have long argued, for instance, that there were no rats in pre- or post-Roman Britain, yet plague aDNA has been recovered from the island in both periods, revealing that Y. pestis could spread and kill there without them.⁸² Plague transmission is also influenced by environmental factors, including climate, agriculture, soil salinity and landscape ruggedness, as well as by differences in population density.⁸³ These environmental and social differences are what cause plague’s demographic toll to vary in different contexts. This myriad of uncertainties demonstrates why a localised approach to plague is essential and how broad, Europe-wide narratives which are not built on solid understandings of the various regions plague touched are bound to be misleading. Nevertheless, the scope of most first-pandemic plague scholarship extends to the entire pandemic.⁸⁴

Finally, palaeogenomes cannot speak to central, social and cultural historical concerns. The presence of plague in a region cannot tell us how that outbreak affected its survivors; aDNA cannot tell us how people understood the epidemic it identifies, whether they thought of plague victims as somehow different from people who died from other causes, or how they acted on those feelings. Only texts and burial archaeology can reveal these crucial facets of history.⁸⁵

With all of this in mind, what does the Edix Hill discovery tell us, and what does it not? The presence of Y. pestis and that plague’s relationship to other published Y. pestis indicate that first-pandemic plague reached eastern Britain early in the pandemic and that it was probably not carried there overland from Constantinople. The discovery proves that plague killed multiple individuals in the Cambridgeshire community that buried its dead at Edix Hill. It certainly killed more, but we do not know how many unidentified plague victims exist or where they are located. There may be more at Edix Hill itself or in nearby cemeteries, some of which are being tested for Y. pestis aDNA.⁸⁶ There may also be some elsewhere in Britain or in Ireland, though, to my knowledge, no genomes have been recovered from early medieval British cemeteries outside of East Anglia. Currently available radiocarbon dates for one Edix Hill plague victim indicate that there is a 95 per cent chance that he died between 416 and 541 CE, and thus before or concurrently with plague’s appearance at Pelusium.⁸⁷ This range overlaps with that of the victim who supplied the most basal British genome, 474–637 CE, but it is not yet clear whether the two died at or around the same time.⁸⁸ There is a possibility that the earlier dated genome is derived from the latter (Fig. 2), which could help narrow the chronology of the victims’ deaths, but this genetic relationship is not certain. More precise radiocarbon dates from several plague-positive individuals are being prepared, but they will not be absolute.⁸⁹ They also will not illuminate plague’s transmission to and within the island; only more Y. pestis genomes from Britain and Europe could speak to that, and they will not be able to if new genomes are identical to those already recovered from Edix Hill.

II

While we can learn much from a single genome, genomes are most illuminating in combination with data about the demographic patterns, ecology and trade routes of the region in which they were found. Placing the genetic evidence in the context of regional archaeological and textual data can illuminate plague’s possible transmission routes, introduction, prevalence and impact in an area. Britain is particularly well suited as a regional case-study to contextualise aDNA evidence. While it has few texts from the post-imperial period, those which survive have been intensively studied, and there exists an exceptional amount of published archaeological data for the east of Britain in the post-Roman period. Britain’s demography and ecology are, moreover, in many ways Constantinople’s inverse. It is thus an ideal example of how differently plague may have behaved in the urbanised east and the rural west of Europe.

There have been recent, significant advances in our knowledge of fifth- and sixth-century Britain which underlie our understanding of plague’s impact and dispersal. Recent publications have transformed our understanding of post-Roman demography, ecology and trade; yet these significant realisations have yet to influence plague scholarship, in part because the twenty-year dearth in publications on early medieval British epidemics coincided with the appearance and development of many of these novel perspectives. These developments suggest that plague was introduced to Britain at least two distinct times early in the First Pandemic and that it may not have had profound, long-term demographic impacts in either case.

The most dramatic recent historiographical shift in our understanding of post-Roman Britain concerns population estimates. Historians long held that the end of Roman Britain triggered a demographic catastrophe in the fifth and sixth centuries which perhaps halved the island’s population.⁹⁰ Recent plague scholarship still operates within this paradigm, but historians and archaeologists largely abandoned this consensus in the 2010s.⁹¹ Reconsideration began when pollen evidence revealed that agriculture in Britain did not collapse following Roman withdrawal.⁹² Instead, fields across Britain remained in use, though some transitioned to less intensive, pastoral forms of agriculture. Archaeologists accordingly began to re-conceptualise the end of Roman Britain, arguing that the abrupt and pronounced material and economic changes visible in the archaeological record only affected a small portion of the population (those few who had been able to afford Roman material culture in the first place), and that the long-assumed agricultural ‘collapse’ was an illusion caused by decades of grain overproduction to supply the Empire.⁹³ These theories were bolstered by a re-evaluation of the archaeological evidence indicative of the demographics of Roman Britain. This re-examination revealed that population statistics for the fourth century were inflated, and that demographic decline actually began as early as the second century.⁹⁴ Today, most scholars believe that the withdrawal of Rome made much of Britain’s population archaeologically invisible, as most of the population of Roman Britain had been all along. Virtually nobody would argue for catastrophic mass mortality across post-Roman Britain, though some still reasonably believe that there could have been minor, circumscribed and short-term episodes of population decline.⁹⁵

Contrary to the long-held consensus of plague’s destructiveness in Britain, these new population estimates seem to exclude a scenario in which first-pandemic plague decimated the entire island’s population for generations.⁹⁶ They probably indicate that plague had a more limited demographic impact in Britain than it did in the vastly different eastern Mediterranean, as we would expect given Britain’s near total deurbanisation and resulting low population density in this period.⁹⁷ It is now widely agreed that this deurbanisation was a gradual process unrelated to catastrophe or mortality, but it would have affected plague’s impact nonetheless, probably by slowing its transmission and limiting its ability to kill widely in comparison with highly urbanised areas.

These data nevertheless do not necessarily indicate that plague did not have significant short-term effects.⁹⁸ They are compatible with an argument that population levels stalled during the pandemic, as scholars believe occurred in sixth-century Ireland based on archaeological proxy evidence and changing nitrogen isotope values in the bones of early medieval animals, which speak to agricultural intensity by reflecting anthropogenic changes to the nitrogen cycle.⁹⁹ It is also possible that plague mortality caused serious short-term population decline, but levels recovered so quickly that little arable land was long abandoned. Plague mortality may also have been uneven across Britain.¹⁰⁰ The likely prevalence of malaria in the eastern fenlands near Edix Hill, for instance, may have increased the outbreak’s severity in the region compared with probably less malarial areas, such as the upland west and north.¹⁰¹ This local demographic context underscores how uncertain plague’s mortality is and how varied its impact could have been even within Britain, much less across the whole of Europe.

Equally crucial to an examination of plague in Britain is an understanding of recent developments in our appreciation of the island’s ecology, as Y. pestis is primarily hosted in and spread through an infrastructure of rodent hosts and insect vectors. Fortunately, investigation of plague ecology in Roman and post-Roman Britain is unusually well developed. Scholars have long debated the presence of rats in post-Roman Britain; considered how trade contacts (or lack thereof) may have shielded or infected regional populations; and posited transmission by human ectoparasites when they realised that the early medieval British climate probably could not support plague’s infamous and efficient vector, the flea Xenopsylla cheopis.¹⁰²

Most important for plague studies is the seeming extinction of rats from Britain between roughly the fifth and eighth centuries.¹⁰³ Rat bones are difficult to recover archaeologically, but they disappear from archaeological finds dating to this period. They vanish even from sites where archaeologists have employed sieving screens capable of catching material as tiny and delicate as 1,500-year-old rodent bones and which boast numerous rat finds in both earlier and later strata.¹⁰⁴ A lack of archaeological evidence is by no means definitive evidence of absence, but a rat extinction in the fifth century would fit into the well-attested pattern of the temporary disappearance from Britain of Roman floral and faunal imports following Roman withdrawal.¹⁰⁵ In the absence of regular reimportation and large Roman-era grain stores as habitats, rats seemingly disappeared from post-Roman Britain along with formerly prevalent Coleoptera beetle grain pests, whose populations were similarly maintained.¹⁰⁶

The apparent absence of rats from the island is significant to our understanding of how Y. pestis spread, what forms of plague people contracted and how contagious the illness was. Scholars of later plague outbreaks have long known that plague can spread and kill widely in areas which seem to lack contemporary rat populations.¹⁰⁷ Yet the transmission mechanisms and impact of plague without rats are poorly understood and may have differed from those in rat-infested areas. It is possible that early medieval British plague was spread by other rodents and insects, such as the house mice and human lice archaeologists have uncovered at post-Roman sites.¹⁰⁸ Plague in sixth-century Britain may otherwise have been partially pneumonic rather than bubonic.¹⁰⁹ A dearth of rats could also have confined outbreaks in Britain to a smaller geographical area. An absence of rats to travel alongside and within trade goods could have slowed or stalled long-distance, overland transmission, though Y. pestis may have spread well in more densely occupied settlements and along often-travelled roads were it partially transmitted via human ectoparasites. Moreover, it is unusual for pneumonic plague to spread widely outside of densely populated settlements, as its symptoms become so severe so rapidly that victims are generally too sick to interact with others by the time they are highly contagious.¹¹⁰ Plague transmission could also have been limited by terrain ruggedness.¹¹¹ Even had plague spread widely across lowland Britain, the mountainous terrain of the west and north may have slowed or prevented dissemination to those areas. Together, these factors could have prevented widespread dispersal in a rat-free Britain and confined plague to the area around Edix Hill.

The date of the rats’ disappearance may also tell us something about the timeframe of plague’s introduction to the island. Whether or not British rats became extinct in the mid-fifth century, their numbers seemingly plummeted suddenly and did not recover for centuries.¹¹² This phenomenon mirrors a recently discovered and still roughly dated genetic turnover of rat populations across the whole of Europe in this period.¹¹³ Studying this phenomenon in Britain, where rodent zooarchaeology is relatively well developed, may both speak to this poorly understood, pan-European turnover and provide a way forward for using rodent zooarchaeology to speak to first-pandemic plague.¹¹⁴ The causes of rats’ temporary extinction from either Britain or Europe have not been established and could be unrelated, both to each other and to plague. But it is possible that the rats’ disappearance could reflect epizootics (animal epidemics) corresponding with plague’s introduction to Britain and spread within Europe.

Epizootics could have occurred because black rats are usually not asymptomatic carriers of Y. pestis. The bacterium sickens and kills them as well, and scholars believe that plague spills over into human populations when it kills so many rodents that their fleas must reluctantly seek non-rodent hosts to survive.¹¹⁵ Many outbreaks of plague in the Third Pandemic were thus associated with pronounced ‘ratfalls’.¹¹⁶ It is therefore possible that the introduction of Y. pestis to Britain, and to Europe more broadly, would have triggered its own ratfalls, killing many local rats. Without Roman grain trade and stores to replenish and maintain their population, they could thereby have gone extinct in Britain. Tellingly, the rats’ seeming disappearance from Britain in the mid-fifth century falls within the radiocarbon date range of 416–541 CE reported for one of the Edix Hill plague victims, and may just pre-date the range of 474–637 reported for another.¹¹⁷ If British rats carried and died of plague, these pieces of evidence taken together could support Sarris’s speculation that Y. pestis reached Britain long before it appeared in Constantinople, or, indeed, in any written source, which underscores texts’ limited ability to speak to the pandemic’s origins.¹¹⁸ But how could the First Pandemic have reached eastern Britain so early? And how can this theory be reconciled with the textual evidence from western Britain and Ireland which suggests that the pandemic did not reach that region until the mid sixth century?

Developments in our understanding of Britain’s trade contacts with the Continent and the Mediterranean may speak to these questions. Scholars have long recognised that Britain maintained contact with mainland Europe following Roman withdrawal, both with peoples just across the English Channel and, in western Britain, with the eastern Mediterranean.¹¹⁹ Plague historians have engaged productively with scholarship on these contacts to argue the theory, originated by McCormick in 2021, that this trade enabled the spread of first-pandemic plague directly from the Mediterranean to western Britain, from where it travelled east across the island and then across the Channel to northern Gaul.¹²⁰ The first element of this theory—that trade between the Mediterranean and western Britain brought plague to that area from the Mediterranean in the mid-sixth century—could be correct. Transmission from the Mediterranean would suggest a date in the mid to late 540s for the initial Y. pestis outbreak in the region, which is precisely the timeframe indicated by the hitherto-neglected insular annalistic evidence. Because current evidence indicates that the terminus of this Mediterranean trade was in modern Cornwall and Wales, transmission via this route would suggest that plague reached Britain and then travelled west across the Irish Sea, where the Annals seem to record its impact on Leinster monastics.¹²¹

The second part of the theory, that plague spread from western Britain to the east and from there to Gaul, remains possible, but it seems insufficient to explain the Edix Hill data, and recent developments in our understanding of British trade contacts render it unlikely. The Edix Hill genomes’ basal positioning with regard to other recovered first-pandemic genomes and the radiocarbon dates currently available for them together suggest that plague may have arrived there before McCormick’s proposed date of 542/3, particularly because recent reanalysis of Mediterranean pottery assemblages in western Britain and Ireland suggests that trade between the two regions was not direct, but instead followed coastal routes, which deposited similar pottery in Spain and on the Atlantic coast of France en route to Britain.¹²² The likelihood of indirect, coastal trade rather than direct connections between Constantinople and western Britain makes the dissimilarity between the Edix Hill samples and the early first-pandemic genomes recovered from Valencia potentially more significant (Fig. 2).¹²³ This route thus has similar issues as an explanatory framework for the Edix Hill genomes as does an overland route from Constantinople. McCormick’s date, moreover, does not consider the time it would have taken for plague to spread across the island once it reached the west of Britain, which could have been considerable, if it were even possible, given the ruggedness of that area’s terrain and the absence of rats.¹²⁴ Finally, the Annals’ evidence suggests that plague reached western Britain or Ireland in the mid-540s, around 545 by most estimates, which seems late to explain the early phylogenetic placement and current radiocarbon dates of the East Anglian samples.¹²⁵ So, although it remains possible that McCormick correctly identifies a transmission route of sixth-century British plague, it seems unlikely that he has identified the route which brought plague to Edix Hill.

Recent research on cross-Channel contacts bolsters an alternative theory of plague’s path to Edix Hill. Robin Fleming has argued from archaeological evidence that eastern British trade contacts with the Continent reoriented farther east and north following the Roman withdrawal, resulting in a disruption of trade with northern Gaul between the early fifth and mid-sixth centuries.¹²⁶ Whether or not these findings represent a true break, they seemingly reflect a significant reduction in intensity of trade with Gaul, which may decrease the likelihood that the Gaulish outbreak came across the Channel from Britain. They also reveal an opening of new trade routes to the north and east, which may have reached even farther east than hitherto supposed. For it is not impossible that these eastward-oriented contacts brought plague to eastern Britain early from the pandemic’s putative central-Asian origin point.¹²⁷ This theory better explains the earlier, 416–541 CE radiocarbon date range for an Edix Hill victim and the Edix Hill plague’s position on the First Pandemic’s branch of the Y. pestis phylogenetic tree.¹²⁸ It also better fits the timeframe for the disappearance of British rats and for Gildas’s shadowy pestifera lues, and does not contradict the later, 474–637 CE date range for the other victim.¹²⁹

The evidence thus suggests that there could have been at least two distinct early introductions of first-pandemic plague to the British Isles, only one of which originated in the Mediterranean. The first, which struck Edix Hill, occurred before or concurrently with plague’s appearance in the eastern Mediterranean in 541, and may even have taken place as early as the mid-fifth century, when rats disappeared from the island.¹³⁰ The second, in the mid-540s, introduced plague to western Britain and Ireland, as recorded in the Irish annalistic tradition. Either introduction could have resulted in a widespread epidemic, as eastern Britain maintained contact with its western neighbours throughout this period.¹³¹ Indeed, there is a mortalitas recorded in the Annales Cambriae around the year 537 CE, which scholars have generally interpreted as a famine, but which could represent an earlier outbreak of plague in what is now Wales.¹³² That possible outbreak does not square with the standard narrative of early medieval plague and has yet to factor in discussions of the pandemic. Alternatively, the absence of rats could have confined plague outbreaks to the regions of Britain where they were first imported and slowed their spread. A multi-pronged transmission of plague thus best fits the currently available evidence for the First Pandemic in Britain.

This theory is only one possible reading of the small amount of evidence for first-pandemic plague in Britain, and it may turn out to be hopelessly off the mark as more palaeogenetic and archaeological information emerges. But it has advantages over current explanations of the Edix Hill genomes. It is grounded in relevant, local evidence and pushes against dominant scholarly paradigms of the First Pandemic to urge scholars to contextualise genomes and to keep an open mind when encountering new evidence that may not fit their pre-existing assumptions. It also provides a way forward for plague studies by combining a characteristically ‘minimalist’ cautious approach to the limitations and paucity of our evidence with a ‘maximalist’ effort to use extant evidence to craft narratives of the pandemic. It moreover helps to bring our understanding of the First Pandemic more in line with the messy and chaotic probable reality of plague’s spread. It seems unlikely that first-pandemic plague was introduced anywhere in Europe only once and probable that trade connections instead transmitted Y. pestis back and forth across the English Channel repeatedly. Moving from one to two identifiable introductions to Britain is a step towards better representing this reality. It is, moreover, a testable theory—with additional Y. pestis genomes from western Britain or Ireland to compare with those from Edix Hill, we could potentially confirm or disprove it. The hope here is therefore not to be conclusive, or even to be correct, but rather to provoke more productive directions in first-pandemic scholarship.

III

Examining the Edix Hill discovery in the context of its post-imperial British setting thus suggests a new interpretation of how plague reached and behaved in the region. Perhaps most importantly, it suggests that Britain may have participated in the same Eurasian trade networks as Constantinople and underscores how regional case-studies can nuance our understanding of the First Pandemic. It is thus disappointing that so many plague studies have sought to characterise the Edix Hill discovery without placing it in its local context. It is similarly problematic that such a major and well-publicised discovery has been largely neglected by British area specialists, who are elsewhere eager to address evidence of Britain’s connections with Europe and the Mediterranean.¹³³ But how did the Edix Hill genomes come to be separated from their British context?

Foremost, the Edix Hill discovery’s de-contextualisation reveals that early medieval British history has become a highly specialised field whose findings seldom influence, and are not commonly influenced by, scholarship on the rest of Europe. The insularity of post-imperial British history and archaeology has been critiqued in recent years, most vocally by Guy Halsall, who has argued that the unwillingness of British specialists to consider Continental evidence has led them to misunderstand parallel historical developments in Britain.¹³⁴ The Edix Hill evidence demonstrates the reverse—that the disengagement of scholars of Britain from broader historiographical concerns can lead others to misunderstand Eurasian processes which involved Britain.

Historians of Britain thus bear some responsibility for plague scholars’ generally outdated or erroneous characterisations of post-imperial Britain in their treatments of the Edix Hill evidence. Their failure to engage with the discovery or more broadly with the field of disease history compelled non-experts to do it for them. And, when they did, they had no topical scholarship that incorporated recent, area-specific findings on which to draw. The last major study on first-pandemic plague in Britain is nearly thirty years old and focuses almost exclusively on the seventh century.¹³⁵ The next-most-recent major articles about sixth-century British plague date from the 1970s, and those with the most historiographical influence appeared in the late 1940s.¹³⁶ This neglect of infectious disease contrasts with the explosion of interest on the topic in early medieval Continental and Irish contexts between 1997 and 2018, and demonstrates British historiography’s disengagement with broader historiographical trends.¹³⁷ Non-specialist scholars’ unfamiliarity with recent, key developments in British history and archaeology reveals how damaging this disengagement has been. The case of the Edix Hill evidence therefore underscores the necessity that scholars of early Britain communicate with scholars of the Continent and consider broader historiographical concerns.

Nevertheless, plague historians are predominantly responsible for de-contextualising the Edix Hill genomes. This issue reflects wider methodological problems within first-pandemic plague studies, which boil down to three overarching themes: insufficient consideration of the best uses of different types of evidence; geographical hierarchies which privilege the Mediterranean; and the bitter divide between those who espouse a ‘minimalist’ and those who argue a ‘maximalist’ interpretation of the pandemic.

Plague studies encompasses many disciplines, including history, archaeology, genomics, palynology, epidemiology and ecology. The evidence that each of those fields produces or interprets answers different questions with varying degrees of certainty. Yet, as Kristina Sessa has argued, there is a tendency for palaeoscientific evidence to take precedence over textual evidence in plague scholarship, even in articles authored by historians.¹³⁸ The omission of the Irish Annals from recent scholarship probably reveals an assumption that texts are of lesser importance than palaeogenomic evidence. The sudden interest of Continental scholars, and even Byzantinists, in sixth-century British plague indicates that plague scholars could not uncontroversially overlook new insular plague genomes such as those found at Edix Hill. But there has been little concern that the blefed annal, which, despite its uncertainties, is more precisely dated than any aDNA sample, has been omitted from theories of the spread of plague to the region. That incorporating it significantly nuances our understanding of that spread underscores how damaging this ‘scientism’ can be to interdisciplinary scholarship.¹³⁹ Texts are not only more precisely dated than palaeoscientific evidence; there are some topics, such as plague’s social impact, to which they speak better than any other kind of evidence, and others, such as the symptomatology of Y. pestis, which only texts can illuminate. Dismissing their testimony in favour of scientific evidence is detrimental to scholars’ understanding of that evidence.

In some cases, however, the reverse is true. As Tim Newfield has contended, plague historians have overutilised texts from the Mediterranean, marshalling them to answer questions they cannot.¹⁴⁰ Plague histories have long followed Procopius’s testimony that the First Pandemic began in Pelusium in 541, though Procopius could not have known whether the first outbreak he noticed was indeed the first of the pandemic.¹⁴¹ In some contexts, palaeogenomes undoubtedly speak better to plague’s emergence and spread than do texts, as the Edix Hill genomes, evidence of plague deaths unrecorded in any written source, indicate. If scholars do not rely on the testimony of texts to reconstruct spread, there is no reason to assume that plague came to Edix Hill from the eastern Mediterranean. The Edix Hill discovery thus underscores that plague studies possesses a bizarre relationship with texts; textual evidence is sometimes ignored and sometimes overemphasised. The root of the problem seems to be insufficient consideration of what different kinds of evidence can and cannot tell us, and a resulting overinterpretation of both textual and scientific data when it is convenient.¹⁴² Cultural histories of plague which consider plague’s epidemiology and ecology may provide a way forward by giving scholars better ground for believing the texts they do. Overinterpreted texts moreover tend to come from the eastern Mediterranean, which reflects the second major issue at play in first-pandemic scholarship.

Scholars’ treatment of the Edix Hill evidence also reflects the ‘Justinianic Plague paradigm’—geographical hierarchies which assume the centrality of the eastern Mediterranean and generalise our understanding of first-pandemic plague there to the rest of Europe and beyond.¹⁴³ This geographic emphasis dates to a time when the testimonies of Procopius, John of Ephesus and others of epidemic catastrophe in the eastern Mediterranean were the only evidence for early medieval plague. Before palaeogenomes, there was only textual evidence for epidemics, and there was no way to be sure what pathogen caused them or whether concurrent outbreaks in different locations were related.¹⁴⁴ Archaeological evidence was of little help, as the First Pandemic is not unambiguously archaeologically identifiable without aDNA even at sites we now know it struck.¹⁴⁵ The uncommon textual richness of the eastern Mediterranean thus made its centrality to scholars’ understanding of early medieval plague unavoidable and their interpretation of the pandemic as a Mediterranean ‘Justinianic Plague’ understandable.

With the advent of palaeogenomes, this Mediterranean-centric paradigm is no longer unavoidable, but it has persisted. The 2005 discovery of Y. pestis aDNA in Bavaria, an extra-Mediterranean location no text claims the ‘Justinianic Plague’ reached, was the first blow to the paradigm.¹⁴⁶ Scholars incorporated this find into the existing, Mediterranean-centric understanding of the pandemic, assuming that plague spread to Bavaria from Pelusium and Constantinople, and many assumed that the pathogen behaved similarly and killed widely in both locations.¹⁴⁷ Many similarly incorporated the 2019 discovery of plague palaeogenomes in northern Europe and Britain into the existing interpretation.¹⁴⁸ But the Edix Hill samples, no matter where they sit phylogenetically, cannot easily be made to fit this paradigm if we take their implications seriously.

There is no compelling evidence that eastern British plague passed through the eastern Mediterranean or that it spread and killed as widely in Britain as it did in the east. Contextualising the genome suggests that it did not. Yet, plague historians have still tried to force a ‘Justinianic’ interpretation of the sample, assuming its Mediterranean origin, ignoring the implications of its independent transmission route even when they accept that such a route is a serious possibility, and failing to consider that plague’s behaviour in Britain, and perhaps in many sparsely populated areas of the Latin West, probably differed significantly from its behaviour in Constantinople.¹⁴⁹ The persistence of this paradigm is clear in scholars’ continued use of the terminology ‘Justinianic Plague’, even when arguing that the eastern, Justinianic outbreak is not necessarily representative of or central to the First Pandemic.¹⁵⁰ First-pandemic plague was not exclusively a phenomenon of the ‘Mediterranean world’, and scholars’ continued interpretation of it as such in the face of evidence which disputes that view is sure to stunt our understanding of it.

It thus seems that plague scholars have omitted or misused relevant British evidence because the power of the ‘Justinianic Plague’ paradigm obscured the potential of studying in detail a geographically peripheral region which they assumed was isolated from the rest of Afro-Eurasia.¹⁵¹ Indeed, assumptions of Britain’s relative unimportance are all that can explain its omission from plague studies, as the subfield is geographically generalist and its scholarship is now often co-authored by specialists of the various fields it encompasses, a practice which sometimes even extends to co-authorship among historians specialising in different regions.¹⁵² There is thus no procedural reason why no plague study interrogating the Edix Hill evidence features a specialist of Britain as a co-author or why only one names a post-Roman British specialist in its overall acknowledgements.¹⁵³ Examining the Edix Hill find in its local context reveals that the interpretations which have resulted from these assumptions are unsupportable. It thus suggests that it is time for the paradigm to shift and for scholars to take seriously the implications of both the limitations and potential of texts and palaeogenomes.

To do so would require abandoning unfounded assumptions of the Mediterranean’s relative importance, representativeness or centrality in plague’s spread unless they can be proven and considering evidence from the ‘periphery’ on an equal footing with that of the Mediterranean ‘centre’. A good first step would be to abandon the misleading name ‘Justinianic Plague’ for the First Pandemic’s initial outbreak, or at least to restrict its usage to refer only to the eastern Mediterranean in the outbreak of the 540s. Future studies which seek to incorporate the Edix Hill find must also feature communication or co-authorship with British area specialists, a practice which should be extended to all relevant regional specialties. Regional experts possess unique skills and command of specialised sources, and experts on thematic or interregional topics should take advantage of their expertise, either by incorporating them into interdisciplinary co-authorial teams or by seeking their advice and feedback while drafting articles. Plague historians would especially benefit from their input as scholars’ growing knowledge of the presence of plague in formerly unrecognised areas stretches plague specialists’ geographical expertise. The input of regional experts will be particularly crucial in palaeoscientific studies of the pandemic, which generally feature a single historian as co-author, responsible for the entirety of the relevant history.¹⁵⁴ This practice is likely in part responsible for the long-standing approach to the First Pandemic as a single, monolithic phenomenon.

The geographically generalist approach to the First Pandemic has fed into a third major methodological issue: plague studies’ polarising divide between ‘minimalists’ and ‘maximalists’. The vitriol with which scholars have pursued the question of the pandemic’s demographic and cultural significance has created, to echo Gildas, a lugubre divortium among scholars on opposing sides of the debate.¹⁵⁵ This divide has had methodologically detrimental effects, as it is predicated upon a question that cannot currently be answered.¹⁵⁶ To characterise fairly the demographic significance or impact of the First Pandemic as a whole, scholars would have to understand its significance in each of the regions it touched, for it behaved differently in each region.¹⁵⁷ Yet, in contrast to the trend in post-imperial European history towards tight, regional focuses carefully stitched together to form a broader argument, plague studies has remained geographically generalist.¹⁵⁸ This generalism and focus on minimalist/maximalist disagreement is responsible for the interpretations which have problematically sought to present the Edix Hill sample’s pandemic-wide significance divorced from its regional context. Both the argument which casts the find as evidence of the pandemic’s ravages and that which dismisses its significance grow out of this issue.¹⁵⁹ To ameliorate this problem, plague scholarship should focus on regional expressions of the pandemic which can later be combined into an evidence-based picture of the pandemic’s broader effects.¹⁶⁰ This is both a more productive path towards understanding early medieval plague’s overall impact and a necessary check to the field’s unproductive vitriol.

Some of the issues illuminated by the Edix Hill samples also have implications beyond plague studies. Plague scholars’ adoption of a textually influenced ‘core–periphery’ model is inherited from their broader historical outlooks—most are Byzantinists, and those who are not study ‘Late Antiquity’, a paradigm which encourages focus on the Mediterranean ‘core’ and relative dismissal of the non-Roman or anomalously post-Roman ‘periphery’, both within and beyond Europe. The Edix Hill discovery suggests that one region of this geographical ‘periphery’ may not have been peripheral at all, and that eastern Britain may have participated in the same Eurasian trade routes which terminated in Constantinople. Indeed, the perception of the Mediterranean as ‘core’ may be as much an accident of textual survival in European history as it seems to have been in plague studies. Historians’ growing employment of non-textual evidence may reveal that the ‘periphery’ has more to tell us about Europe-wide developments than scholars have long supposed. The Edix Hill discovery should thus encourage scholars of the immediately post-imperial period, both within plague studies and without, to question fundamental assumptions of the Mediterranean’s exceptionalism, to consider northern and southern Europe on a more equal footing in this period, and to keep an open mind when encountering new evidence from regions assumed to be marginal. It is possible that the ‘periphery’ has something vital to tell us about the ‘core’ or the whole, should we choose to listen.

Footnotes