The Justinianic Plague, the pandemic that brought waves of plague to eastern Eurasia between 541 to about 750 CE, has been a feature of narratives about the “decline and fall of the Roman Empire” ever since the eighteenth century, when Edward Gibbon featured Procopius’s vivid description of plague’s assault on Constantinople in 542 during the reign of the Emperor Justinian. For two centuries since Gibbon, the Justinianic Plague (“JP” for short) received a few passing nods in accounts of the period we now call “late antiquity.” But overall, it elicited little attention from historians. Even in my own training as a historian of medicine in the 1970s and ’80s, it was barely a blip on the radar.
Not anymore! The JP has been known for some time as the “First Plague Pandemic” because it was followed by two others: the Black Death in the mid-fourteenth century and the five-hundred-year reign of plague outbreaks it initiated in Eurasia and Africa, and then the Third Pandemic, which spread plague around the globe in the age of steamships at the turn of the twentieth century. (We have plague in Arizona, where I live, thanks to that last proliferation.)
All three pandemics, we know now, were caused by a single bacterium: Yersinia pestis. That finding is not actually surprising news, since Y. pestis’s role had been suspected ever since the bacterium was first discovered in 1894. Why, then, has the JP become a hot new topic, with more than a dozen new publications in the past five years, and more promised soon? For the same reason we often revisit old questions: because new evidence has become available. But some of that new evidence — in this case coming from genetics — is so new that most people don’t know what to make of it.
This might seem a debate of interest only to specialists in the history of medicine or demography and economics (who are concerned about assessing the pandemic’s effects on population levels), but for one urgent factor. The JP has now been linked to a pronounced change in climate that occurred in the sixth century. Indeed, such claims have been made for the Black Death, too. Does this mean that we should be looking at the events of the sixth century with new fearful eyes, scrutinizing the evidence for signs of disasters that might befall us, too? Perhaps not. The climate change that happened in the sixth century was global cooling, not warming. And viral diseases are a more likely threat to us than a bacterium like Y. pestis.
But the questions raised about the climate-disease connection are nevertheless profound. And that is in large part because the methods used to investigate such questions are themselves under scrutiny. How will these new methods help us interpret the past as a guide for future decision-making? What constitutes evidence when science is invited to the table of historical interpretation? Discussing the scope, impact, and causes of the JP is worthwhile for any number of reasons, but it has a particular value in allowing us to talk about how and why the field of genetics is playing a unique role in new approaches to the history of pandemic disease — and why it is broadening the question well beyond the boundaries of the Roman Empire as Gibbon envisioned it.
I’ve already said what the cause of the JP was: the bacterium Yersinia pestis. We actually found this out very recently. Even though everybody who ever lived, up to 1900, the turn of the twentieth century, is dead, of those billions of people, the number of those for whom we can confirm the cause of death is infinitesimally small. For people who died before the development of modern mortality record-keeping, the percentage is even smaller.
Until recently, causes of death for the pre-modern period were determined from either physical remains or verbal descriptions. Thanks to the archaeological record, the cause of death for victims of extreme physical trauma can be discerned from paleopathological analysis: a severed cranium, for example, is a pretty clear indication of how that person died. But the infectious diseases that leave distinctive traces on the bones are, ironically, mostly ones that we live with (leprosy, tuberculosis), not ones that immediately kill us. In other words, inferring cause of death from visual inspection of skeletal remains is usually very difficult. With plague, which kills within two to ten days, it is impossible.
Plague does, of course, often produce distinctive symptoms that contemporaries might have described: the painful and sometimes gangrenous buboes that give “bubonic plague” its name, for example. For the JP, Procopius and other sources gave us descriptions of such symptoms, and scenes of mass mortality that sound a lot like those we would now associate with plague. But could we be sure? In fact, we couldn’t, for the simple reason that medical understandings of the body differed in past societies from the concepts we use now to explain disease.
We explain infectious diseases as being caused by microscopic organisms that come into our bodies (germ theory). That’s why the very fact that we can positively identify anybody who died specifically of plague is an extraordinary development.
Yersinia pestis is, like all bacteria, a single-celled organism. To see a living bacterium, you have to look under a microscope. A microscope wouldn’t do much good for historical bacteria, however, because any cellular life that was alive in the sixth century has long since died and decayed. For historical work, you need molecular biologists, who have developed techniques to gather up the fragments of DNA of microbial forms and reassemble them, reconstructing their genomes and thus identifying them.
This new field of palaeogenetics has been developing for the past twenty years, at the same time that historians have been more systematically gathering evidence for the extent and impact of the JP. An invited conference in Rome in 2001 was the first instance when historical and genetics approaches were debated in the same venue. The proceedings volume from the conference (published in 2007) became the focal point around which an impressive body of scholarship now collectively paints a broad-ranging picture of the plague’s effects from Iran to Ethiopia to Spain to England. As noted above, climate history has also been playing an increasing role in defining the global events of the 530s and 540s created by two or more sequential volcanic explosions, including a dust veil in 536 CE that shrouded the world in gloom and markedly reduced global temperatures. In short, we know more about the JP — its biological cause, its environmental circumstances, its physical extent and social impacts — than we have ever known before. The genetics evidence on the pathogen clinches the deal, right?
What’s the Big Deal?
One would think that the recovery, starting in the 2000s, of fragments of Y. pestis from sixth-century graves in Bavaria, and then (in 2014) of the whole genome of Y. pestis, would have elicited great excitement. Another identical genome was retrieved from a nearby town in 2016. And just in the past year, researchers in Germany recovered eight additional complete genomes (out of 33 remains showing traces of Y. pestis) from various sites in western Europe. But “excitement” among historians has been in short order. Rather, the JP has become contentious terrain.
On the one hand, historians who had been paying attention to the trends in palaeogenetics noted that these findings had already been anticipated. The latest study wasn’t paradigm-shifting, then, but paradigm-confirming: it fit into an understanding of plague’s history that has been forming for over two decades. Other historians, however, simply shrugged, assuming that the identification of a few plague-infected bodies didn’t constitute evidence of a “pandemic.” As one study claimed — burying the comment in a footnote to stress its unimportance (p. 29, fn. 107) — “A few additional cases of plague cannot confirm a demographic, social and cultural collapse across the Mediterranean.” Another commentator opined (at 41:00 in the linked podcast), “We [historians] want bodies. We want a clear sense of what the impact was on the population. […] The geneticists, however, they don’t care how many people died.”
Is that a fair assessment? Palaeogenetics can do a phenomenal job of telling us, on a one-by-one basis, that particular people died of plague. If you have Y. pestis in your teeth when you die, that means you died of plague — the organism only proliferates that profusely in the bloodstream when you’re at the point of death. But the extremely complicated technical effort that goes into producing this data (we’re talking terabytes of computational power to do analyses on millions of DNA fragments) is not intended to give us a “head count” of the total number of people who died in a pandemic event. No, beyond confirming the presence of a particular bacterium (or, more rarely, a virus), the geneticist is concerned with studying the evolution of that organism. And that turns them into de facto epidemiologists. Who are, as it turns out, a lot like historians.
Tracking Disease: What Pathogen Palaeogenetics Actually Tells Us
The new genetics of plague may not tell us how many people died, but that’s because its potential lies not in counting bodies, but rather in helping construct richer narratives of how plague moves across landscapes and evolves. The new genetics, for the first time ever, allows us to ask some of the questions real epidemiologists do: Why is this specific disease presenting here? Why in these populations? Where did it come from in the first place, and by what pathways did it (will it) spread?
How can just a few genomes tell us so much? Because by studying whole genomes of pathogens, geneticists can construct phylogenetic trees, like the one below. These “family trees” allow us to see the larger dimensions of the “Justinianic Plague.” In evolutionary terms (from the perspective of the pathogen), it was a disease event that spanned at least half a millennium and covered many thousands of miles across Eurasia.
So, how do we “read” a tree like this? The image is a detail of the phylogenetic tree published by Keller and colleagues in June 2019, with added notes from a follow-up study the same core team published in October. These give some sense of how the genomic branch lengths (which show the gradual acquisition of genetic changes) might correlate to absolute chronology. The green and blue samples are from burials from the time of the JP, between the sixth and eighth centuries. EDI001.A is the Edix Hill sample, which comes from England, near present-day Cambridge. Note its branching-off position on the tree. It is evolutionarily prior to all the other green and blue strains. In historical terms, that means that the sample the furthest distance from the Mediterranean (England) is the one showing the “earliest” evolutionary state of the pathogen as it has thus far been documented in Europe. Until now, indications (from documentary records) that plague reached England could only be dated to the seventh century. Genetics has just given to historians (and archaeologists, too) a new research problem: the surprisingly early arrival of plague in post-Roman Britain, possibly as early as 544 CE, the same year it seems to have reached Ireland.
The close phylogenetic relationships between the different strains thus confirms that the reported outbreaks weren’t just coincidental. Rather, our historical sources were indeed reporting the spread of a single infectious disease. Moreover, just as the 530s volcanic dust veil and the climatic cooling it caused make us look at factors as far away as Central America (where the volcano Ilopango exploded probably in 539 CE), so the genetics expands our geographic horizon by allowing us to ask where the JP ultimately came from.
In 2018, several months before the latest JP genome study appeared in pre-print, a team based in Copenhagen brought out a massive study sequencing 137 human genomes from across the Central Eurasian steppe. The samples dated from the late Neolithic through the middle of the second millennium of the Common Era. Among them was an individual, found in a burial site in what is today Kyrgyzstan, who died sometime in the third century. (This is sample DA101 on the tree, in orange; dating follows Keller et al. 2019b.) The body carried in its teeth a strain of plague that is evolutionarily older than the strain of Y. pestis that caused the JP. The Kyrgyzstan sample is about three centuries earlier than the JP. And the origin of its Y. pestis lineage is even older than that: the arrow on the upper left of the tree points to a divergence event that caused the Kyrgyzstan and JP lineages to separate sometime near the beginning of the Common Era.
Equally interesting is the fact that we can now postulate an origin for these two lineages. That would likely have been in Kyrgyzstan — and the area just north of it that forms the border of western China — which is still home to strains of Y. pestis closely related to that common ancestor. In other words, the JP now needs to be conceived as a disease event originating in central Asia and extending to the Mediterranean and Europe: well beyond what even Gibbon imagined in the eighteenth century. By what route plague made its trek from Central Asia to the Mediterranean we still don’t know; right now, a path involving the western Indian Ocean and the Red Sea seems most likely.
Obviously, what would be most informative would be to have, instead of this bare phylogenetic tree, a dynamic map showing the disease’s spread across time and space. Consider this video that was developed for the then unfolding outbreak of Ebola virus disease in West Africa in 2014. It shows how the epidemic began in southeast Guinea and then was carried, person by person, across borders into Sierra Leone and Liberia, and back again. The map is built on data tracing the genetic evolution of the virus’s strains. On the right side of the screen, you can see the phylogenetic tree grow, as branches of disease spread and continue to diversify. This is phylogeny in action!
This kind of epidemiological detail was actually envisioned for the JP more than a decade ago. True, it is not yet possible to date individual disease transmissions. The lack of detailed records and the limits of radiocarbon dating will leave many specific questions unanswered. But in terms of expanding our knowledge of the pandemic into regions where we never knew it reached, palaeogenetic research is unmatched. This will, for example, aid efforts to determine the pandemic’s effects on northeast Africa; documentary records already suggest that Ethiopia and other areas around the Red Sea were also affected. Even regions right on the Mediterranean hold great intrigue, since we have new methods to document population contractions by, remarkably, noting when trash collections ceased.
When Numbers Do Count
So, how many people did die of plague? That, frankly, is not palaeogeneticists’ problem to solve. We may soon be looking at hundreds or even thousands of samples of ancient DNA (aDNA), from pathogens as well as humans. Maybe we’ll soon have aDNA from rats and other animals that served as intermediate hosts for plague or other diseases. There is in fact explicit talk now of doing palaeogenetic work on an “industrial” scale. Yet the limits of archaeology (not to mention the still not-inconsiderable costs of aDNA research) will set an upper limit to what can be expected.If there is to be an answer to the question of cumulative mortality levels from the JP, it will likely come from demographic modelers in alliance with archaeologists who document locales where human populations disappeared or substantially contracted.
Yet it should be stressed that the question “how many died?” is not the only, nor the most important, question to ask about the plague. From the perspective of the history of medicine, there is no threshold, no magic number of deaths that will automatically determine “significance.” The 1918–19 Flu Pandemic killed “only” an estimated 2% of the world’s population; the HIV pandemic (whose spiraling numbers only started to decline in the early 2000s) killed even less. Yet because many of the dead from both pandemics were young, previously healthy adults, the economic and cultural impact was profound. Plague need not have had equal effects on all parts of western Eurasia and North Africa to have been profoundly influential as an infectious disease in late antiquity.
Instead of approaching palaeoscience as a threat, or summarily dismissing it, we might engage with it as we would with any other new source of historical evidence: by asking how it yields historical knowledge. Just as we do with numismatics, for example, or tree rings, we need to know the basics. How can one measure time? With coins, we use regnal years and minting dates; with trees, we count the rings of individual trees and then create “lumber libraries” that overlap trees harvested from different eras. With genomes, we use SNPs (single nucleotide polymorphisms), the individual physical changes in the genome. If we’re counting, what are we counting? And how much is enough to be meaningful? In other words, what does one need to be an “informed consumer” of the latest work in disease history?
My prediction, in fact, is that scholars of the ancient world will prove themselves exceptionally adept at bridging the disciplinary divides between humanistic and scientific approaches. Evolutionary thinking is hardly anything more than what Classicists already do when they engage in philological stemmatics, tracking the changes that texts acquire over time, which then become defining characteristics of their descendants. There is, of course, one crucial difference: whereas it is possible that a book might sit unused, unopened on a shelf in a monastic library for centuries, only to be rediscovered and recopied later (hence the philologist’s dictum, juniores non semper deteriores — “younger copies of texts are not always more corrupt”), with life forms, there can be no inactive period. These are living organisms that must always be replicating.
Debates about infectious diseases have more than historical import. As we move into the third decade of the twenty-first century, we are facing a world without precedent. Documenting not simply climate change, but the ways altered temperature and weather patterns create permissive environments for new (or newly circulating) diseases is of the utmost urgency. Of the diseases known or believed to have struck the ancient world, all, save smallpox, still exist. If there was ever a time when cross-disciplinary understanding, dialogue, and cooperation were necessary, it is now. Plague may not be the next pathogen to threaten us, but better understanding its history will prepare us to face whatever does loom on the horizon of our Anthropocenic age.