4 Epidemiology of the Black Death and Successive Waves of Plague

Abstract

Open any textbook on infectious diseases and its chapter on plague will describe three pandemics of bubonic plague. The first, the plague of Justinian, erupted in the Egyptian port city of Pelusium in the summer of ad 541 and quickly spread, devastating cities and countryside in and around Constantinople, Syria, Anatolia, Greece, Italy, Gaul, Iberia, and North Africa: “none of the lands bordering the Mediterranean escaped it”, and it reached as far east as Persia and as far north as Ireland in less than two years and spread through their hinterlands.1 Historians have counted eighteen waves of this plague through Europe and the Near East that endured until ad 750, if not longer.2 The second pandemic originated in India, China, or the steppes of Russia, touched the shores of western Europe (Messina) in the autumn of 1347, circumnavigated most of continental Europe in less than three years and eventually struck places as remote as Greenland. While the first lasted just over two centuries and the third a mere twenty-five years in pandemic form, this second wave returned periodically for nearly five hundred years in western Europe. Its last attack in Italy was at Noja (Noicattaro), near Bari, in 1815,3 but it persisted longer in eastern Europe and Russia. Its cycles, however, lengthened from a hit about every ten years for any locale during the latter half of the fourteenth century to absences of 120 years or more for major cities at least in Italy by the seventeenth century. Despite repeated claims in textbooks, the plague of Marseilles in 1720–1 was not this pandemic's European finale.4 In 1743, 48,000 perished from plague in Messina; in 1770–1 over 100,000 in Moscow; and in the Balkans, Egypt, Asia Minor and Russia this Black-Death-type of contagious plague may have persisted as late as 1879.5 The “third pandemic” began in the mid-nineteenth century and crept slowly through the Yunnan peninsula until it reached Hong Kong in 1894. From there, steamship commerce carried it across much of the world. However, except for China and India and a few other subtropical regions, its spread (unlike that of the other two pandemics) was limited in epidemic force to coastal cities and even there hardly penetrated beyond docklands. Instead of millions killed, as happened with the previous two pandemics and as Europe feared at the beginning of the twentieth century, death counts of this third pandemic in temperate zones rarely exceeded one hundred. Few quantitative records such as burials or last wills and testaments or narrative sources that describe the signs or symptoms of plague survive for the first pandemic. But several—Procopius of Caesarea, John of Ephesus, Gregory of Tours, the Antiochene lawyer Evagrius “Scholasticus”, the Chronicle of Zuqnīn, and Paul the Deacon—report swellings in the groin, armpits, or on the neck just below the ear. Like later chroniclers of the Black Death, Procopius also observed that black pustules covered victims’ bodies.6 These narratives, as well as the archaeological evidence, provide certain clues about the epidemiology of the “first pandemic”. It extended rapidly, not only with its first appearance in 541 but also with subsequent waves. In 664, the plague took only ninety-one days to travel 385 kilometres (as the crow flies) from Dover to Lastingham7 (4.23 km a day)—a far faster overland spread of any plague ever seen for Yersinia pestis since its discovery in 1894. At various places during the early twentieth century, such as New Orleans and South Africa, Yersinia pestis was shown to be a slow mover, travelling overland at only 12 to 15 kilometres a year. This is because Yersinia pestis is a disease of rats in which humans participate, to use Robert Koch's succinct definition of 1900 (now we should update it by substituting rodents for rats),8 and rats are homebound creatures. Thus modern plague, even with the advantage of railways, automobiles, and much higher population densities, would have needed twenty-five years to cover the distance this early medieval plague travelled in a mere three months.9 Connected with this great difference in the speed of transmission, a second discrepancy arises: the bubonic plague of the “third pandemic” was a rodent disease. In “typical” epidemics—those which have shown the highest mortalities in the twentieth century—ailing rats staggered drunkenly out of their hiding places and then collapsed and died, or the stench of their bodies seeped through walls and floors; a clear indication of plague to villagers and western doctors alike.10 By contrast, no scholar has found any evidence, archaeological or narrative, of a mass death of rodents that preceded or accompanied any wave of plague for the first or second pandemic.11 A third discrepancy arises from contemporary descriptions of the “first pandemic”: in contrast to the “third pandemic”, which never killed more than 3 per cent of the populations of major cities, the first wiped out the inhabitants of entire communities and regions, leading to mass burials or even the absence of burial altogether because of the paucity of survivors. According to the archaeological evidence, coinage, and inscriptions, its demographic effects were pivotal and long lasting.12 This overland speed and high mortality in cities and the countryside suggest a fourth discrepancy: while plague doctors of “the third pandemic” discovered to their surprise that the bubonic plague of the late nineteenth and twentieth centuries was rarely contagious, contemporaries of the first suggest a highly contagious person-to-person disease. Procopius, Evagrius, John of Ephesus, and Gregory of Tours characterized the disease as contagious and, in keeping with this trait, described it as clustering tightly within households and families; the evidence from burial sites supports their claims.13 A fifth discrepancy arises from a close correlation between plague and famine with the first, with famine often preceding and sparking a wave of plague, as seen in Alexandria and Constantinople in 618 and 619.14 By contrast, dearth has tended to dampen epidemics of Yersinia pestis of the “third pandemic”, while bumper crops have intensified them: increase grain yields feed higher populations of infected rats and their fleas, giving rise to higher numbers of human cases and fatalities. Finally, the “first pandemic”, like the second, but unlike “typical” bubonic plague of the Yersinia variety, could strike at any time of year before usually settling into a summer pattern for the southern Mediterranean and the Near East, that is during the hottest and driest points of the year, the least hospitable season for the most efficient flea vector of Yersinia pestis—Xenopsylla cheopis. Despite these epidemiological differences between the first and third pandemics of plague, scholars of the Justinianic plague have nonetheless been absolutely certain that the causative agent of this first pandemic must have been the bacterium, Yersinia pestis. Even more extraordinary, according to some scholars of the first, little if anything differentiated the epidemiology of the three pandemics, although the first two spread more rapidly and were more contagious than any disease ever witnessed by contemporaries, while the third, whose agent was discovered in 1894, was a slow mover, surprising health carers around the world by its lack of contagiousness.15 The narrative sources for the “second pandemic” set off by the Black Death of 1347–52 explode in number and variety. In addition to hundreds of chronicles from abbeys, city-states, and principalities across Europe, the plague tract, written principally by university-trained doctors, takes off; by the fifteenth century, it had become one of the earliest forms of “popular literature”.16 Further, the survival of tens of thousands of last wills and testaments, monastic and confraternity necrologies, burial records, manorial rolls, lists of ecclesiastical vacancies, and doctors’ examinations of causes of death allow quantitative analyses of the late medieval and early modern plagues: their cycles of mortality, seasonality, topographical details of transmission, and some basic characteristics of the victims—age, sex, occupation, class, and locality. Among these riches, neighbourhood burial records exist for Siena in 1348 and a city-wide one for Givry in Burgundy during the plague's first onslaught in 1348–9. In Arezzo, they begin in the 1370s and in Florence in the 1420s. By 1452, Milanese doctors and health workers examined all city deaths and supplied detailed clinical reports for each body. Similar records (although lacking the rigorous medical observation of victims’ symptoms and signs as seen in the Milanese death books) appear for cities north of the Alps during the second half of the fifteenth century.17 Yet, despite this surfeit of information, no evidence links any late medieval or early modern European plague from 1347 to the end of the eighteenth century to a description or even suspicion of an epizootic of rodents that preceded or accompanied human plague. The closest claim for such a connection has been the doctors’ repetition of Avicenna's tenth-century cliché about the remote cause of epidemics in general: the inversion of nature—birds dropping from their nests in mountaintops to lower valleys and underground beings such as topi rising to the surface as the harbingers of epidemics. Moreover, in this flip-flop of nature, doctors did not note mice and rats taking centre stage; rather they pointed to snails, frogs, scorpions, snakes, bugs, and other subterranean creatures surfacing from their holes, and, unlike the drunken prance of rats that soon dropped dead on doorsteps, as poignantly described in Camus's La peste, these underground creatures climbed out in the bloom of health.18 Occasionally, in the early modern period, contemporaries even described the increased number and vigour of live mice and rats during plague time, the direct result, as they probably rightly surmised, of governments’ systematic culling of the domestic cat populations.19 The epidemiological evidence that can be assembled from the late medieval and early modern sources raises further doubts that the so-called second pandemic was the same disease as the bubonic plague that spread to ports globally at the end of the nineteenth century. First, like the early medieval plague, the “second pandemic” was a fast mover, travelling in places almost as quickly per diem as modern plague spreads per annum. George Christakos and his co-researchers have recently employed sophisticated stochastic and mapping tools to calculate the varying speeds of dissemination and areas afflicted by the Black Death, 1347–51, through different parts of Europe at different seasons. They have compared these results to the overland transmission speeds of the twentieth-century bubonic plague and have found that the Black Death travelled at 1.5 to 6 kilometres per day—much faster than any spread of Yersinia pestis in the twentieth century.20 The area of Europe covered over time by the Black Death in the five years 1347 to 1351 was even more impressive. Christakos and his colleagues maintain that no human epidemic has ever shown such a propensity to cover space so swiftly (even including the 1918 influenza epidemic). By contrast to the spread of plague in the late nineteenth and twentieth centuries the difference is colossal: while the area of Europe covered by the Black Death was to the 4th power of time between 1347 and 1351, that of the bubonic plague in India between 1897 and 1907 was to the 2nd power of time, a difference of two orders of magnitude.21 Furthermore, it should be noted that the spread of bubonic plague in India at the turn of the twentieth century was exceptional. Since 1894, 95 per cent of cases of this plague have occurred in India. If these scientists had compared the time–space propagation of the Black Death with other twentieth-century regions of plague, even subtropical ones such as China, the differences in orders of magnitude would be greater still. Doctors and chroniclers marvelled at the Black Death's lightning transmission, reporting that mere speech was enough to pass it directly and immediately from one person to the next. Tales of sailors leaving their ships and infecting local populations instantly cannot be attributed merely to fear and exaggeration by stunned observers in 1347 and 1348 as some have asserted.22 Nor is it reasonable to assume that somehow the plague mysteriously raced ahead of the ships docking in Messina's harbour in October 1347 and it was mere coincidence that the resident population began dying at the very moment when the sailors from the east happened to disembark, or that these Black Death chroniclers simply pointed to the sailors as convenient scapegoats for the disease, as Susan Scott and Christopher Duncan have asked us to believe.23 Towards the end of the European plague experience, the Enlightenment doctor and polymath Richard Mead, using contemporary reports from the plagues of Rome in 1656, Marseilles in 1719, and Messina in 1743, recorded the same rapid infection of dock workers, sailors, and magistrates who inspected ships and cargo at the beginning of these later European plagues. In almost every case, these plague officials could identify the spread of the disease as the immediate consequence of an infected ship's dropping anchor in a port. Mead was astonished how rapidly victims could die from the time of first exposure to plague; for example, “the Porters who opened the infected Bales of Goods in the Lazaretto's [sic] of Marseilles, died upon the first Appearance of Infection, as it were by a sudden stroke … death insued sometimes in a few hours”.24 Before 1348 the word contagium (contagion) was rarely used, especially outside the medical profession. When chroniclers and theologians used the term, it applied almost exclusively to heresy or revolt and not to disease. By contrast, after 1348 its usage became widespread and not only among doctors. Chroniclers and merchants in their diaries and letters used it to describe the spread and destruction wrought by disease and principally by plague.25 By contagion, they clearly meant person-to-person transmission by breath, touch, or occasionally by sight, and, although the term was applied to other maladies, it was mainly reserved for describing plague, in order to distinguish it from other diseases.26 Early on, contagion and the swiftness of death became as important “pestilential accidents” as the bubo to identify their new disease and to differentiate it from others such as pondi, dondi, or smallpox, which contemporaries often saw as producing similar skin afflictions. Thus in England a chronicler known as the Brut described a 1369 epidemic that combined smallpox and plague. He distinguished between the two, not by the diseases’ signs but on elementary epidemiological grounds, pointing to the levels of contagion and the speed of the death: with “pestilens” people went to bed healthy, “hool & in good poynt”, and “suddenly they died”; whereas smallpox (“an illness that men callen ‘the pokkes’”) was “much slower to infect both men and women”.27 In 1390 the merchant chronicler of Florence, known as the pseudo-Minerbetti, also observed two diseases that he saw as concurrent, spreading through parts of the territory of Florence. In terms of seasonality—the summer months—and skin disorders that spread over the victims’ bodies, he claimed the two were indistinguishable, but he nonetheless made a distinction between them, as did the English chronicler, on epidemiological grounds: one, which Minerbetti called pondi, moved slowly, while the other was the fast mover and quick killer, the plague.28 In addition, Florentines such as the merchant Giovanni Morelli had grown only too well accustomed to the new “big killer” after it had appeared in the city five times since 1348. Morelli also turned to epidemiological patterns to recognize this disease by noting its exceptional contagiousness and high mortalities as well as its seasonality: regularly, it made a faint appearance in Florence at the end of winter with only a few hidden cases and smouldered until the first heat waves in May, which set the plague ablaze with a mounting death toll that would peak in June or early July.29 In plague tracts of the fifteenth and sixteenth century, plague was often called simply “the contagion”—morbo contagioso, contagioso male, voracissimo contaggio 30—and for many the first distinguishing factor of “true plague” was precisely this, that it was the most contagious of diseases.31 Similarly, by the end of the so-called second pandemic, commentators such as Mead continued to emphasize the plague's contagion and its capacity to spread rapidly. For this reason he advised against “Crowding the Sick together in Hospitals”, claiming that it “can serve to no good purpose; but instead thereof will promote and spread the Contagion”.32 He too distinguished plague from other illnesses that gave rise to boils and pustules by turning to the disease's extraordinary contagion,33 and, because of this, used his plague pamphlet to preach the benefits of quarantine.34 There is indeed usually some difference between these Swellings in the Plague, and in other Fevers, especially in the time of their coming out: for in the Plague they discover themselves sooner than in most other Cases. But the principal difference between these Diseases, is, that the Plague is infectious, the other not; as least not to any considerable Degree. And this leads me to another Character of this Disease, whereby it is distinguished from ordinary Fevers, which is the Contagion accompanying it.35 In contrast to this defining characteristic of the Black Death and its subsequent waves through the eighteenth century, doctors and health workers at the turn of the twentieth century, from Bombay to Sydney, in one hospital after another, remarked to their great surprise that the safest place to be in time of plague was the plague ward.36 Despite the habits of Indian relatives crowding round plague victims in plague hospitals and passing the patient's sputum from one relative to the next, Yersinia pestis rarely had pneumonic complications (primary or secondary) and was even more rarely spread by contaminated clothes or goods (other than rat-infested grain). Instead, it depended largely on the complex and inefficient mechanisms of the rat carrier and rat-flea vector.37 Moreover, despite two serious epidemics of pneumonic plague in Manchuria in 1911 and 1922, several outbreaks of plague in which Pulex irritans may have been the ectoparasite, and plague deaths resulting from handling skins of marmots, squirrels, cats, and other carriers, or from the consumption of diseased camel meat, the vast majority of plague cases from 1894 to today (certainly over 95 per cent of them) have resulted from rodent fleas feeding off rodents (principally rats) and then biting and regurgitating the bacillus into humans. Thirdly, the seasonality of the late medieval plagues does not resemble that of modern plague. Because of the flea vector, Yersinia pestis in bubonic form is sustained at epidemic levels only within a narrow temperature band (10°C to 25°C), accompanied by high levels of relative humidity. On the other hand, bouts of late medieval plague could occur at any time of year, including January in places as inhospitable to the subtropical rat flea (Xenopsylla cheopis or the European one, Nosopsyllus fasciatus) as Norway, Sweden, and Scotland.38 Further, in the warmer Mediterranean, the Black Death and its recurrent strikes consistently peaked at the warmest and driest points of the year (June and July), the least likely months for modern plague to spread, given the fertility cycles of fleas and especially that region's most prevalent rat flea.39 Cases of Yersinia pestis in the Mediterranean zones during the twentieth century (although few in number) have followed closely the fertility cycle of Nosopsyllus fasciatus, whose numbers and fertility peak in late September to November40 with corresponding human cases in Mediterranean towns such as Taranto in September to November.41 Thus, these Italian twentieth-century plague cases have arisen two to three months after plague had declined or disappeared entirely in Florence, Bologna, Rome, and other cities during the fourteenth and fifteenth centuries. In fact, the time when late medieval plague most often peaked in Bologna, Tuscany, Umbria, and Rome was June, the very month when the appearance of Nosopsyllus fasciatus reaches its nadir in Mediterranean Europe.42 It is, therefore, even less likely that an eruption of Yersinia pestis will occur in June in these zones than in January.43 Fourthly, modern plague has never attained the mortalities seen with major strikes of the Black Death or even with most of its subsequent assaults. In the summer months of 1348 alone, tax records show that Florence lost three-quarters of its population or more.44 It is clear from manorial records that villages in Cambridgeshire,45 the bishopric of Worcester,46 and the area around St-Flour (Auvergne)47 lost up to 80 per cent of their populations. According to chroniclers, places such as Trapani on the west coast of Sicily became totally abandoned after 1348.48 Further, although later strikes of plague in the seventeenth century were not as widespread as waves of the Black Death in the latter half of the fourteenth century, they could be as devastating for cities such as Genoa and Naples in 1656–7, which had not experienced plague for 120 years. Both lost two-thirds of their populations.49 By contrast, modern plague has never approached these levels of human destruction—not even in India, where over 95 per cent of twentieth-century plague casualties have occurred. The highest mortality for any large city in any plague year was at Bombay City in 1903, when less than 3 per cent of its population perished from plague. True, in the countryside and in small villages, the proportion killed by bubonic plague rose higher. In several villages in the Punjab (the region most devastated by Yersinia pestis in any form since the bacillus was discovered), mortality rates may have reached 30 per cent. But, as E H Hankin speculated in 190550 and Major Greenwood demonstrated with statistical rigour seven years later,51 the case and mortality rates of bubonic plague were inversely correlated with human population because of Yersinia pestis’ reliance on a pool of diseased rodents and not on the density or absolute figures of human population. Recently, Ole J Benedictow has vigorously tried to demonstrate that Black Death demographics followed a similar pattern. But, despite his efforts to scale down the mortalities in large cities and magnify them for smaller places, his own figures still fail to show any such inverse correlation.52 Christakos and his co-authors have argued cautiously that the mortality figures for 1347–51 show no correlations, positive or negative, between the size of a place and the epidemic's mortality.53 However, according to population estimates supplied in tax records in late fourteenth- and fifteenth-century Tuscany, mortality levels were roughly higher the larger the population of a city or town.54 Furthermore, by the sixteenth century in England, France, and Italy, the contrasts between cities and countryside became more marked, leading contemporaries and historians to conclude that plague was essentially a disease of towns and of overcrowding.55 Fifthly, the cycles and trends of the “second pandemic” differed strikingly from those of the “third”. Because humans possess little or no natural immunity to Yersinia pestis and cannot acquire it over the long term,56 mortality rates from this bacillus do not fall steadily as a population becomes more accustomed to the pathogen. Thus in India plague mortalities first increased for a decade or more, then jumped randomly from year to year before declining in the 1920s, as a result of rats (not humans) acquiring immunity to the bacillus.57 Similar patterns can be seen in Brazil, Thailand, Vietnam, and other subtropical regions later in the century, even after the introduction of DDT, the use of effective antibiotics, and modern sanitary measures.58 Moreover, in subtropical zones, where plague at the end of the nineteenth and early twentieth centuries scored by far its highest mortalities, the disease returned annually to the same localities five to ten years running, and in villages of the Punjab and the Presidency of Bombay for twenty years or more. Finally, the age range of victims of modern plague has not changed significantly over the past 112 years of the third pandemic's history. As with the first strike on virgin-soil populations, those in the prime of life, between the ages of twenty and forty, have remained most often this plague's principal victims.59 By contrast, the plague cycles of the Black Death and its subsequent strikes over the first hundred years of its history were radically different. Firstly, Scott and Duncan maintain that during the initial sweep of plague, 1347 to 1353, not a single place was re-infected two or more years running,60 which is all the more remarkable given Pope Clement VI's decision to call a jubilee year for Rome in 1350. Pilgrims from northern and eastern Europe, where plague was just beginning to rage, travelled southward; yet I know of no reports of any locality being re-infected by this movement of persons and, no doubt, pathogens. The absence of re-infection, I would speculate, resulted at least in part from the immunity these individuals had acquired to their bacillus or virus, one that had outlasted the six months or less that humans today can acquire to resist Yersinia pestis. Secondly, the mortality trends across Europe illustrate a remarkable tendency for the pathogen of the Black Death and its human hosts to adapt, again unlike the trends reflected in cases and mortality from Yersinia pestis in India and elsewhere during the late nineteenth and twentieth centuries.61 Contemporaries were aware of the Black Death trend and attributed the sharp decline in morbidity and mortality of successive strikes of plague to their own scientific and political intervention. After four strikes of plague in Avignon, the pope's personal doctor Raymundus Chalmelli de Vinario gave a retrospective in 1382 on their plague experience since the Black Death: In 1348, two thirds of the population were afflicted, and almost all died; in 1361, half the population contracted the disease, and very few survived; in 1371, only one tenth were sick, and many survived; while in 1382, only one twentieth of the population became sick, and almost all of these survived.62 While Chalmelli's assessment may have been overly sanguine, last wills and testaments63 and monastic and confraternal necrologies point in this same downward direction, charting a rapid adaptation between the Black Death pathogen and human hosts.64 All of these late medieval records pose problems for quantitative analysis. Although testaments did not pertain only to the “affluent” as some have charged but included peasants, artisans, and even disenfranchized workers (especially in Italian cities), these records concentrate among propertied groups and, perhaps more problematic for demographic trends, do not include infants or many of the young. Nor are burial records without their problems as a demographic record, especially in 1348, when the clerics, who recorded the burials, either died themselves or ran for the hills, as happened in Siena in June of that year. Just as the plague was heating up to a crescendo, these records stop and resume only after the disease had disappeared in late August. For Florence, the Dominican necrologies may be a more precise source for 1348. They give a complete record of deaths and specify whether brothers died of plague or not. These records, however, pertain to a specific group of men, who were cloistered, and as the Black Death was a contagious disease with a high household clustering of cases, the documentation may not provide an equivalent mortality among the general population. Yet these various death records show a remarkable consistency from city to city and amongst themselves within cities where multiple types of records exist for the same period as with Florence. Along with this steady decline in mortality through the fourteenth century, chronicles across Europe described plagues from 1361 onwards as “plagues of children”. The contemporary chronicler of Pisa, Ranieri di Sardo, was more specific than most. By the third plague in 1374, he recorded that 80 per cent of the plague deaths in his city were among those aged twelve or younger65 (precisely the number of years since the last plague had hit Pisa). Statistics drawn from the burial records of Siena confirm Ranieri's report. From the second plague in 1363 to the third in 1374, the proportion of deaths attributed to children rose from a third to over half (136 of 233), and with the fourth in 1383, children had become a staggering 88 per cent of the plague's victims (230 of 260).66 To repeat, the bubonic plague of the twentieth century has never shown this tendency, even in places such as the Punjab, where plague returned annually for twenty or more years and infected larger proportions of the population than anywhere else.67 Further differences in the epidemiology and ecology of the late medieval–early modern plague and Yersinia pestis are apparent. As with the first pandemic, chronicle descriptions tallied by Henri Dubois show a close correlation between famine and outbreaks of plague during the late medieval and early modern periods.68 By contrast, the opposite has been seen for late nineteenth- and twentieth-century India and elsewhere:69 bumper crops have led to higher plague mortalities. From the early twentieth century, plague commissioners in India, Sydney, and other international ports, as well as later in Vietnam in the 1960s and 1970s, discovered that grain deposits were the powder kegs and epicentres of bubonic plague with grain workers usually the first victims. Statistics from the burial records of Florence in 1400 allow a detailed reckoning of the neighbourhood distribution of plague mortality, where the disease first struck, how it spread, and in what wards it proved most lethal. The relationship between grain storage and the dissemination of the late medieval plague can probably be studied better here than elsewhere because of the strict enforcement of grain distribution from a single point in the city, first at Orsanmichele and, after 1367, a hundred or so metres away at the Mercato Nuovo del Grano. Instead of being the first area of affliction or the neighbourhood most severely visited by pestilence, the tiny parishes of and surrounding the grain market were the last to be hit and the least heavily blighted by the plague of 1400. Instead, like typhus or other infectious diseases that haunt primarily the poor, the plague began in the poor parishes on the periphery and then worked its way into the wealthier neighbourhoods of the centre, where in fact Florence's central grain market was located, but never scored here dramatic rises in rates of mortality.70 * Given, therefore, such wide epidemiological differences on so many fronts, why have historians and scientists been so certain that the Black Death and its successive strikes through the early modern period had the same agent—Yersinia pestis?71 Usually scientists point to Boccaccio or occasionally to a handful of chroniclers to insist that contemporary descriptions of swellings were the “unmistakable signs” of the same bubonic plague that was cultured in 1894.72 For instance, the immunologist and Nobel Prize winner Sir Macfarlane Burnet judiciously cautioned historians that in order to understand diseases in past times, the scholar must go beyond the signs and symptoms of a disease and study its epidemiological patterns.73 When he came to the Black Death and bubonic plague, however, he concluded: “The symptoms are characteristic enough to make it easy to recognize the disease from classical or medieval descriptions.” He then went further: “We can be sure that the two greatest European pestilences, the plague of Justinian's reign (A.D. 542) and the Black Death of 1348, were both the result of the spread of the plague bacillus.”74 Yet, as health workers in Asia are taught today and as medieval contemporaries were aware, swellings in the lymph nodes are not unique to Yersinia pestis, and cultures of the infected regions must be taken.75 Plague doctors such as J W H Chun remarked in the 1930s that diseases such as malignant typhus may be indistinguishable in their signs and rapid fatal ending from bubonic plague. He concluded: “Summarizing, we may say that the danger of mistaking plague for other diseases is great.”76 Boccaccio and many other chroniclers and physicians from Michele da Piazza in Messina (1347) to doctors of the plagues of 1665 in London, 1720 in Marseilles, and 1743 in Messina did not end their descriptions with the bubo forming in the groin, the armpits, or behind the ears. They added that other “tokens” joined the buboes covering victims’ bodies. These so-called “tokens” or lenticulae were carbuncles and pustules of various sizes and colours. In addition, occasionally the buboes themselves are described as migrating and multiplying from the lymph nodes to places not regularly seen in cases of “typical” Yersinia pestis—the penis, vulva, elbow, face, chest, legs and even up the nose (as in the case of an unfortunate Bavarian nobleman in the early fifteenth century, who also had a big one in his groin).77 Contemporaries such as the fourteenth-century English chronicler Geoffrey le Baker and the early fifteenth-century Florentine diarist Giovanni Morelli pointed to these pustules as being more deadly than the buboes: the stricken could survive the larger glandular boils but not the lentil-like bumps, “shower of peas giving rise to affliction, messenger of swift death”, “brittle coal fragments”, to use the words of the Welsh poet Llywelyn Fychan, who lamented the death of his four daughters felled by plague in 1363.78 Moreover, through the fifteenth to the eighteenth century, plague doctors and chroniclers continued to point to these smaller pustules and carbuncles accompanying the larger tumours in periods of plague.79 By contrast, various editions of Manson's tropical diseases maintain that “carbuncles” are extremely rare in cases of bubonic plague: the only case cited comes from the London plague of 1665.80 Over 3,000 clinical reports of plague from hospitals around Bombay City in 1896–97 assembled by Brigadier-General W F Gatacre in 1897,81 show much the same: less than 5 per cent of plague victims developed more than one plague boil, and in not a single case did these boils migrate from their lymph nodes, nor did smaller spots and pustules spread over victims’ bodies. Gatacre concluded that “true carbuncles were never met with any of the Parel patients”.82 Similarly, in the Glasgow plague of 1900, the clinical reports do not describe a single case of pustules or carbuncles forming or any other skin disorder apart from a single bubo or in two cases, two boils (both of which were in the groin).83 In addition, with modern bubonic plague 60 to 75 per cent of boils form in the groin,84 because fleas generally bite on or below the shins. Although contemporaries occasionally called their plague inguinal, not a single medieval source pointed explicitly to the groin as the bubo's principal site. Instead, from miracle cures found in saints’ lives to doctors’ reports, the late medieval boils’ pride of place was the neck, behind the ears, or on the throat.85 By the late fifteenth century, at least in Milan, buboes in the groin and on the hip became the principal node. Yet many more swellings formed outside the three principal lymph nodes than has been thus far described for any twentieth-century plague, “typical” or not: 20 per cent as opposed to less than 2 per cent.86 However, plague signs and symptoms, even with bubonic plague in various parts of the world since 1894, have been neither so regular nor so consistent as many textbooks on infectious diseases sometimes describe them or as historians often presume. “Pyogenic, necrotic, infarctive, inflammatory, hemorrhagic, and edematous lesions” forming over modern plague victims are not impossible.87 In 1909 Choksy found pustules (which he labelled “cellulocutaneous plague”) developing in 3.7 per cent of 13,600 cases of bubonic plague collected at two hospitals in Bombay over the course of twelve epidemics, 1897 to 1908.88 From cases in Brazil, Peru, Ecuador, and Chile, Atilio Macchiavello observed in the 1930s and 1940s carbuncles on the wrists and ankles of plague victims forming a day or two before the bubo or simultaneously with it. Further, cutaneous lesions—“petechiae, ecchymoses, pustules, and gangrene, [generally called carbuncles] … of variable size may appear over the buboes or independent of them, especially on the abdomen or extremities”. He also observed “pustular plague eruptions” that at times resembled smallpox, which could occur “on any part of the body” but most often in the gluteal and scapular regions.89 Unfortunately, Macchiavello did not quantify the proportion of such cases. But this form of bubonic plague (rarely noticed or at least commented upon by others working in the US, the Soviet Union, India, and China where plague has been endemic in the twentieth century90), nonetheless differs in two crucial respects from medieval and early modern plague with their own variety of pustules and boils. Chroniclers and doctors from 1348 to the very end of the so-called second pandemic in the eighteenth century saw multiple pustules and carbuncles as an integral part of their plague and as its most deadly sign, more so than the bubo.91 By contrast, Macchiavello, Choksy, Chun, and others in the twentieth century concur that these “dark pustules” (sometimes called “plague smallpox”) were never very common even in incidents of “atypical plague”, and, more critically, they saw their appearance predicting the very opposite outcome from that of Black Death ones: in the twentieth century the pustules have been of low toxicity and their prognosis generally benign.92 * From at least the beginning of the twentieth century, scientists have noted discrepancies dividing the characteristics of the so-called second and third pandemics.93 For the absence of any reports of an epizootic of rats or other rodents in 1348 or thereafter, some such as Philip Ziegler have argued that medieval men and women were too overwhelmed by the terrible loss of human life to notice all those rats that surely covered plague-wreaked streets and houses.94 Others such as Jean-Noël Biraben, after a thorough combing of texts, medieval and early modern, have harvested very few references to rats during plagues and none are descriptions of epizootics. Biraben has instead tried to convince us of the supposed tremendous mortality of rats during plague by pointing to less than a handful of examples such as that at Uelzen in 1576, when one man alone was paid 6 schillings to poison rats in the town.95 If so many rats were dying, why did contemporaries throughout Europe and for five centuries not comment on it? In the instance above, what was the explanation for killing still more if so many already littered streets and houses? By contrast, peasants in subtropical regions even going back to the Middle Ages have left traces in records and folklore of rat epizootics when their rat-based disease gave rise to boils and quick death. The appearance of dead and dying rats was the sure sign that they should abandon their dwellings and camp out nearby until the end of the plague season. Why should Europeans alone have been so blind, not only during the horrific events of 1347–51 but with successive and less deadly attacks of this plague to the end of the eighteenth century? As for differences in seasonality and in particular the spread of plague through winter months in northern Europe, historians have argued that the winter of 1348 was extraordinarily wet and mild, but they have failed to look beyond 1348 and account for the continuation of plagues in places as far north as Lübeck, where last wills and testaments show plague peaking consistently in late October or November through the fourteenth and fifteenth centuries, or to sixteenth- and seventeenth-century reports showing plague peaks in places as far north as Trondheim, Norway, in February.96 On the other hand, as already noted, plague in Mediterranean zones (places such as Genoa, Florence, and Naples) continued through the seventeenth century to peak often during the hottest and driest months of the year—the least likely months of plague if a rat-flea was the vector.97 Why the north and south of Europe should have had such seasonal differences and variability might prove a valuable key for further understanding the late medieval–early modern plagues. At this point, I have no explanations, but such differences certainly do not accord with the narrow temperature and humidity bands found with the fertility cycles of rat fleas or incidents of Yersinia pestis. For the mixture of skin disorders—buboes in the lymph glands, migrating buboes, and a host of other pustules that covered plague bodies—microbiologists and medievalists have claimed that the bubonic plague activated a penumbra of disease, possibly typhus and smallpox—and that these diseases account for those other skin disorders not generally seen with twentieth-century plague. They fail to explain, however, why this penumbra of disease suddenly appeared with such ferocity and became even more deadly than the new virgin-soil bubonic disease of medieval plague. Nor do they explain why Yersinia pestis in modern times has never displayed this synergy, suddenly causing an explosion of a wide host of other diseases, much less ones with skin disorders similar to typhus or smallpox but which were more contagious and deadly than these diseases are today. Perhaps most vexing for those who wish to label the Black Death Yersinia pestis has been the drastic difference between the transmission of the two diseases: one travelling with astonishing speed and efficiency; the other discovered early on by plague commissioners to have been hardly contagious at all, especially in bubonic form. To resolve this riddle, historians and doctors first turned to pneumonic plague to explain the seeming absence of rats with the Black Death along with its contagion and rapid dissemination. Yet the late medieval plagues especially in the Mediterranean recurred consistently during the hottest months of the year, not the usual time for respiratory diseases and for pneumonic plague in particular to reach their peaks. By the mid-fifteenth century, if not earlier, moreover, descriptions of coughing and spitting blood disappear from records such as the detailed clinical reports of the Milanese necrologies and in doctors’ plague tracts; yet the plague continued to be as contagious as earlier, with deaths clustering tightly in families and households.98 More damaging still, assumptions that Yersinia pestis in pneumonic form needs no rodents, is highly contagious, and can result in much higher death tolls than in bubonic form are mistaken. The head of the Manchurian plague commissions in 1911 and 1922, Wu Lien-Teh, observed that pneumonic plague, as in its other forms, remains a rodent disease and must begin with an epizootic of rodents. In Manchuria it was the tarabagan, whose furs soared in price in the early twentieth century, leading to an influx of inexperienced trappers into the region. Pneumonic plague can spread person-to-person, especially in the extraordinary conditions in which the tarabagan trappers found themselves, forty or more crammed into unventilated underground huts measuring 15 by 12 feet in sub-freezing temperatures.99 Wu, however, observed that the pneumonic form of Yersinia pestis was not, even in these extraordinary circumstances, terribly contagious. He noted that in tightly packed train cars travelling across Manchuria those infected with pneumonic plague rarely passed it on to fellow passengers. As a consequence, the worst-known epidemic of pneumonic plague, that of Manchuria in 1911, infected and killed less than 0.3 per cent of the population affected by the disease. As Wu concluded, it did not spread with anything like the speed or efficiency of influenza or the Black Death (although later in life and further removed from the events of 1911 and 1922, he thought that the Black Death may have been Yersinia pestis).100 More recently, Steve Leach of the Centre of Applied Microbiology and Research has shown that pneumonic plague has had low rates of transmission (0.25–.55), much lower than influenza, measles, and many other infectious diseases.101 The pneumonic plague that flared in Gujarat, Surat, and other districts and cities of India further substantiate these conclusions: despite this plague's centre in the highly crowded slums of Surat, “almost no confirmed secondary cases were reported in any of the investigations” and in “hardly any family was more than a single member affected”.102 Historians and now many from the scientific community speculate that the human flea, Pulex irritans, was the vector—a proposition first raised cautiously by C J Martin in 1913. If present in great abundance, they argue, this flea might explain the absence of rodents as carriers (at least after an initial outbreak) and the wide differences in epidemiology between the second and third pandemics, that is, the lightning speeds of late medieval plagues, their high contagion, household clustering of cases, and high mortality. According to Martin, however, “A variation of the plague bacillus in the direction of greater infectivity, with perhaps diminished toxicity leading to a higher degree of septicaemia in man” would have needed to exist to explain the direct transmission by human fleas on the scale of the late medieval plagues.103 As far as I am aware, the most detailed study of a plague transmitted person-to-person and perhaps by the human flea was that of the 1967 outbreak in the mountain village of Nawra, Nepal. It resulted in six cases of tonsillar plague,104 one of primary pneumonic plague and seventeen of the bubonic variety. Unlike bubonic plague in India, Sydney, China, and most other places, plague here clustered tightly in households and could be shown to have been transmitted person-to-person. Further, no epizootic of rodents was reported immediately preceding or accompanying the human plague, “despite persistent inquiry and careful searching”. Finally, in the bubonic cases, again unlike classic plague in India and most other places, the buboes did not concentrate in the femoral regions and groin but formed on shoulders, calves, facial and cervical regions. In at least two cases, “multiple Black Spots scattered over the skin prior to death”.105 So, even if Pulex irritans can explain small-scale epidemics in the twentieth century without evidence of rodent epizootics and with skin disorders that look much more similar to late medieval and early modern plagues than the classic “Indian” variety, should we conclude that the epidemiology of the second pandemic is solved by Pulex irritans? To date, no one has shown that this vector has ever caused an epidemic on any noticeable scale or explained how a flea that is far less efficient in transmitting Yersinia pestis than Xenopsylla cheopis can account for the medieval mortalities that were at least an order of magnitude higher than anything ever scored by Yersinia pestis since its discovery in 1894 and that spread more widely over space in a given period of time by two orders of magnitude and without twentieth-century modes of transportation.106 This remarkable difference in transmission is to be explained by a flea that Albert Burroughs and others showed rarely, if ever, became blocked and in the laboratory never transmitted the bacillus to its hosts even in ideal conditions of temperature and humidity.107 The plague commissioner Fabian Hirst in the 1950s showed the difficulty in grasping this straw to solve the riddle. First, he argued vigorously that “all the available evidence” points to Pulex irritans as “a feeble transmitter of plague” anywhere in the world and questioned the conclusions of Georges Blanc and Marcel Baltazard that plague in Morocco in the 1930s and 1940s had been transmitted by this flea. He then went further, maintaining that the reverse transmission of plague from humans to rats, other mammals, or other humans was highly unlikely: the concentration of the bacillus in humans is far too low to transmit the plague effectively to other animals or humans.108 But when Hirst came to explain the Black Death, for some reason he ignored his earlier conclusions and speculated that its person-to-person spread might be explained by the human flea. Other entomologists, rat specialists, and plague scientists such as Atilio Macchiavello, and even C J Martin, have repeated Hirst's first doubts about this flea transmitting plague in the twentieth century and have connected the severity of plague since 1894 with the prevalence of the most efficient vector, Xenopsyllus cheopis. Along with Graham Twigg, who has put it bluntly, they have concurred: “man is a biological dead end” as far as plague goes.109 Even if we reject this axiom and accept the studies of Blanc and Baltazard for Morocco and Iran that implicated the human flea, the “third pandemic” has yet to provide a single such case of an important epidemic scoring over a thousand deaths, and most outbreaks in Iran, northern and central Africa, and Nepal, where Pulex irritans has been suspected as the vector, have counted far fewer casualties. Secondly, while plague transmitted by Pulex irritans might spread rapidly through a single village (as it may have done in Nepal in 1967), it can extend beyond the village only with great difficulty. This failure, moreover, cannot be attributed to efficient medical intervention. In the cases of plague in the mountains of Iran and Nepal, health workers from the Institut Pasteur at Tehran arrived on the scene only after these epidemics had almost completed their course. Even the most vigorous supporter of human ectoparasites as the vectors of the late medieval, early modern, and twentieth-century plagues, Georges Blanc, recognized these difficulties: Inter-human contagion is strictly limited to those who enter the mortuaries of plague victims, or who change or wear their clothes. For the most part, they are family members, or of the same farm or hamlet. The transport of infected ectoparasites at any distance is thus nil, the opposite of what happens with typhus.110 He then went on, however, to speculate that in urban areas such restrictions on the mobility of the disease would not pertain. And it was there, he suggested, that the late medieval plagues achieved their “réputation catastrophique”. But, in addition to failing to explain the rapid and devastating spread of the Black Death or other plagues such as those of 1629–30 through isolated rural districts, such an urban transmission has yet to happen even on a minuscule scale in urban areas since 1894, even in cities such as Dakar and St Louis (Senegal) where both plague and human fleas have been plentiful. Instead, the suspected examples of outbreaks of inter-human plague transmitted by human ectoparasites have come solely from nomadic tribesmen in cold climates or those in desert circumstances such as on the steppes of Russia,111 where herdsmen have been accustomed to wearing thick layers of clothing for protection.112 These characteristics of plague transmitted by Pulex irritans do not, moreover, tally well with the epidemiology of the Black Death and its subsequent strikes through the early modern period or with the habits and conditions of its people. First, these late medieval and early modern plagues were capable of spreading rapidly from village to village and from one rural region to another even in isolated and sparsely populated mountainous territories such as Snowdonia,113 the Apennines of Italy,114 and the cantons of Switzerland.115 Secondly, the seasonality of the plague especially in the warm Mediterranean does not correspond with conditions of plagues spread by human ectoparasites, which, if they arise at all, occur when people are wearing extra clothing and using the maximum bedding for warmth or protection against wind and sand. Such clothing habits in summer time are not seen among Florentines or Romans in the Middle Ages or today. As early as the late thirteenth century, Italians had developed spinning technology to weave light summer cloth for their Mediterranean climates. David Herlihy and John Munro have shown that these new fabrics spread to the peasantry and may well have constituted the first mass- market goods in the west.116 Elsewhere, after the Black Death, contemporaries hardly give us the impression of people—even the common folk—constantly bundled up to excess in layers of clothing. Indeed, Geoffrey Chaucer, John of Reading, the legislators of new sumptuary laws, and many others complained of luxury clothing penetrating the ranks of commoners, and of the “inordinat scantnesse” of their garments; instead of an excess of clothing, moralists of the late Middle Ages, especially in plague time, complained of new fashions that were “extremely short” and which exposed the arses and “private parts” of men and women alike.117 Thirdly, the spread of plague by Pulex irritans makes certain assumptions about medieval life across social classes—that they were all equally dirty, insanitary, and rarely washed themselves, their bedding or clothing.118 But late medieval and Renaissance paintings show obsessions with clothing, beauty, and cleanliness now corroborated by economic historians and those studying fashion for wider strata of the population before and especially after the Black Death.119 The first several strikes of plague through the fourteenth century did not victimize solely the poor; rather they more or less hit all equally, regardless of social class. In various places chroniclers even commented that the rich fell to the disease in greater numbers than the poor.120 Certainly by the sixteenth century (and probably well before) doctors recommended changing clothing frequently,121 and at least the rich and middling classes could afford to do so; yet plague continued to strike communities with massive epidemic force as in 1575–77, 1629–30, 1656–7, 1709–12, 1719–20, 1743, and 1770–1, and these plagues (after a generation or longer absence from a locale) killed the rich and well attired along with the poor and dirty.122 Fourthly, from the plague legislation at Pistoia in 1348123 to the eighteenth century, communities made strenuous efforts to restrict the movement of goods as well as persons by imposing quarantines; governors and health magistrates were particularly suspicious of clothing, bedding, and textiles as prime culprits in the transmission of plague. More than any other articles, these were subjected to careful scrutiny and, by the eighteenth century, even experimentation in the transmission of plague.124 Yet, I know of no account that describes fleas popping out of bedding and clothing like popcorn as has been observed among the articles of herders today, who live in tents with carpets and thick layers of bedding and clothing.125 Unlike the microscopic pathogen, whether a bacterium or virus, the human flea, especially in the great numbers necessary for the spread of plague to even a few people in a single village, is visible.126 In ancient times as well as the Middle Ages, contemporaries knew about insects and classified them. Albertus Magnus, for instance, described various characteristics of fleas (pulices)—their eggs, biting mechanisms, how they drew blood, their leap, and seasonality. He saw them, however, exclusively as an insect that drew blood from animals, not man. In marked contrast to the flea, he described lice (pediculi) as an insect of man as well as of animals, “generated from the filth that is found in the opening of human skin pores” and “found especially in the seams of clothing”.127 Finally, with early industrialization to the late nineteenth century or even later, overcrowding, public hygiene and sanitation worsened, infectious diseases rose in intensity, and life expectancies often dipped dramatically below levels of the late Middle Ages and early modern period.128 Yet these new levels of filth did not suddenly spark renewed outbreaks of the Black Death in Manchester or Calcutta, despite a probable increase in the density of human ectoparasites. For instance, Pulex irritans became the principal rat flea in Dakar and St Louis (Senegal) by the 1950s.129 But despite these being overcrowded urban districts with appalling poverty and sanitary conditions on a par with or worse than nineteenth-century cities in the West, no major epidemics of Yersinia pestis have broken out there in the past fifty years, much less ones that have approximated the horrific mortalities and lightning speeds of the Black Death or any which have depended principally on the human flea as its vector.130 Instead, with the spread of Pulex irritans as a major flea in Dakar and St Louis, outbreaks and cases of Yersinia pestis declined. The two trends, moreover, may be related. More recently, researchers have claimed to have discovered ancient DNA in tooth marrow from early medieval, late medieval, and early modern grave pits thus “putting to an end” the controversy over what the Black Death was. Geneticists and archaeologists such as Mike Prentice, Alan Cooper, Thomas Gilbert, Carsten Pusch, and others have cast doubts on these claims on several fronts—laboratory contamination, misused methods, and the absence of such results in well-preserved plague pits in Britain, Denmark, France, and Italy.131 Let us suppose, however, that medical opinion will accept the claims of Michel Drancourt, Didier Raoult, and their team: that Yersinia pestis was the causal agent of these earlier European pandemics.132 Would the riddle of the Black Death then be suddenly solved? Or would such a result instead open new problems and questions for microbiologists, geneticists, others in the scientific community, and historians? How in the space of a century (but perhaps even less given the evidence from nineteenth-century Russia133) would a disease that once spread quickly and efficiently, person-to-person without the complexities of a rodent carrier or an insect vector, have suddenly become transformed into the inefficient, barely contagious bubonic plague, whose agent was first cultured in 1894? Why would this pathogen have suddenly withdrawn from its previous pathways toward symbiosis with its human host? Why would humans, who once possessed natural immunity to this pathogen and could acquire it over the long term, suddenly have lost these characteristics? What biological event in the nineteenth century led to such a quick and radical genetic mutation in both humans and bacteria at precisely the same moment and not just in one or two locales but across the globe? These are questions the historian cannot answer, but which multidisciplinary research and cooperation (as seen at the Oslo conference in November 2005) might help to unravel. Both camps—those who believe the Black Death was Yersinia pestis and those who doubt it—must now be open to new possibilities. As Robert Pollitzer commented in 1960: “Plague is a disease of so protean a character that it would be misleading to generalize the results of observations in one or a few areas, however suggestive they appear to be.”134 For the moment can we agree: the signs and symptoms of both pandemics—the bubonic plagues of late medieval and early modern Europe on the one hand, and those touched off in the late nineteenth century on the other—were hardly so “unmistakable” as to end all debate on the agency and character of these two diseases? Even if new advances in palaeopathology should one day settle the question that Yersinia pestis was the agent of all three pandemics, historians and scientists would then be confronted with new questions: how and why did the epidemiology of the “third” differ so radically from the first two?