image Appendix image


The Nature of the Pestilence: Current Debates

THIS BOOK does not set out to prove whether the great mortality or pestilence, as it was known to those who suffered it, was bubonic plague, anthrax, typhus, haemorrhagic fever or some other disease. However, the London evidence can contribute to the debate in this contested arena, since it would appear to point away from the argument for bubonic plague and towards something else. It is impossible for this author in this volume to attempt a full critical survey of all published theories on the Black Death and its character, but in order to provide the basis for any such argument, a summary of the different approaches and theories is necessary; it is hoped that the key elements of the various arguments have been acceptably precised here.520

The pestilence was, for the vast majority of schoolchildren and researchers alike, bubonic plague. This has been the accepted identification of the Black Death for most of the twentieth century, since the outbreak of a plague in Hong Kong in 1894 by a team including Alexandre Yersin: the bacterium was eventually called Yersinia pestis after him. In the early 1900s, the link between the bacillus, fleas and rats was made, and in 1914 the mechanism which caused rat fleas to attack humans.521 Yersin himself made the first connection between the nineteenth-and early twentieth-century outbreak of bubonic plague and the medieval scourge, linking it also to the even earlier outbreak in c. ad 540 known as the Justinian plague.522

The main arguments in favour of this being the medieval pestilence are based around its observed characteristics – the common presence of bubos in most of the chronicles and many medical observations, and the similarities of other symptoms and of the progress of the disease. Variants in description (such as the spitting of blood, or of victims dying within hours almost unmarked) are explained through the presence of three varieties of plague caused by Yersinia: bubonic, where the infection passes from the bite site into the lymphatic system causing bubos; pneumonic, where the bacteria are transported from the bite site to the lungs causing secondary pneumonia; and septicaemic, where the infection is directly into a vein effectively allowing rampant bacterial reproduction and assault before the body can mount any effective defence. Interpersonal infection risk in all cases is low. The disease can be moved effectively around urban and rural areas by the presence of rats and the capacity for people to transport infected fleas in grain or clothing.523 Death rates for nineteenth and twentieth-century outbreaks can be as high as 35 per cent in particular localised settlements, but at a regional level seldom exceed 3 per cent.524

The requirements for a bubonic plague outbreak are a suitable rodent vector (the black rat is the usual suspect) and a flea population that is active (most of the year except winter). 525 A forceful supporter of the bubonic plague argument is Professor Ole Benedictow, who has attempted a detailed reconstruction of the spread of the first plague outbreak across Europe in terms of a sequence of epizootics (where first the rat population in a given area contracts and perishes from the disease, and then, as the infected fleas seek new food, creates epidemics within the associated human population), spread and established by human transport of rat and flea vectors. His assertion that modern bubonic plague and experiences of its outbreaks are directly comparable with those of medieval pestilence has, however, been questioned, and his particular use of chronicle sources to support an epidemiological analysis of spread is considered by some to be rather too risky.526

Support for the bubonic model has been offered on the basis of the reported extraction of ancient Yersinia DNA (aDNA) from the dental pulp of skeletons recovered from cemeteries argued to be from plague outbreaks, with the interpretation that this evidence concludes the debate. However, questions have been raised about the basic assumptions underlying the samples used for the aDNA research, and some scientists have questioned the possibility of contamination of some samples. A test to obtain aDNA from skeletons from the London East Smithfield plague cemetery (among others) proved negative, although many other kinds of aDNA were recovered; but a more recent study of the mid-fourteenth-century mass burials excavated at Hereford Cathedral (and other European sites) has again identified the bacillus.

Correspondence among researchers demonstrates the difficulties in using such approaches to obtain a definitive answer.527 There are two principal issues with claims, for example by Haensch et al., that aDNA discovery ‘ends the debate about the etiology of the Black Death’. The first is the assumption during sample selection for testing that mass burial equates to plague (‘mass’ has been as few as two burials in a single grave in one study). This is unsupportable, since outbreaks of infectious disease other than plague must have been extremely common in the medieval period, and we know that severe famine caused mass deaths on at least two occasions (in the midthirteenth century and in the early fourteenth century). The second is that the discovery of Y. pestis aDNA simply tells us that the plague bacillus was present; to offer categorical assertion that it was the agent of the 1348 disaster requires us to prove both its prevalence in skeletal material firmly tied to the outbreak, and in addition its absence in skeletons dating before this date. One thing that all researchers seem to agree on is the need for suitable independent and blind testing in order to solve the issue.

The fit of modern bubonic plague to the documented evidence for the pestilence has always been problematic to a degree. As early as 1893 (and thus a year before Yersin started his work in Hong Kong), Gasquet noted of the 1348 outbreak that ‘together … with the usual characteristics of the common [meaning bubonic] plague, there were certain peculiar and very marked symptoms, which although not universal, are recorded very generally in European countries’. These were, he said, ‘gangrenous inflammation of the throat and lungs, violent pains in the region of the chest, the vomiting and spitting of blood, and the pestilential odour coming from the bodies and breath of the sick’. He concurred with Charles Anglada, a Parisian physician, who in his study of extinct and new diseases published in 1869 expressed a ‘profound conviction that the Black Death stands apart from all those which preceded or followed it’. 528 This recognition of the highly unusual nature of the 1348 pestilence was set aside as the post-Yersin bubonic model became firmly established. Nonetheless, subsequent researchers into the impact and effects of the pestilence, while content to accept bubonic plague as the culprit, continued to have some problems with making the evidence fit the diagnosis. In 1963 historian J.M.W. Bean noted that the Black Death must have:

contained a very considerable pneumonic element, a fact which is suggested by three of its features: first, the exceptionally high death rate; second, the descriptions by contemporary observers of the symptons which are nowadays met in pneumonic plague; and third, the fact that the Black Death raged in winter as well as summer, so that many bubonic victims are likely to have contracted pneumonia and ‘triggered off’ pneumonic outbreaks.529

The inherent problem with this analysis was recognised just a few years later, in 1970, by the bacteriologist J.F.D. Shrewsbury, who noted that ‘the pneumonic plague cannot occur in the absence of the bubonic form and it cannot persist as an independent form of plague’. He went further:

Our modern knowledge of the bionomics [of the rat flea] explains why in medieval England epidemics usually erupted towards the end of spring, rose rapidly to a peak of intensity in late summer or early autumn, collapsed or declined sharply with the arrival of autumn frosts, and were extinguished in or remained dormant throughout the winter months.530

Indeed, his whole thesis comprised a sustained attack on the documentary evidence for high mortality and winter spread precisely because it did not fit the bionomics of rat and flea populations and the aetiology of bubonic plague. He argued for a range of coterminous diseases being responsible for the mortality seen in the fourteenth-century outbreaks.

Despite the inherent contradictions that this kind of situation gives rise to, it was to be another decade before a cautionary note was sounded. In 1979 Stephen Ell, undertaking an examination of more than 300 medieval plague accounts, concluded that the human flea, and thus human-to-human transmission, more closely fitted the bill.531 Three years later, the risks associated with extrapolating modern plague experience backwards to the 1348–9 outbreak were raised by medical historian N.G. Siraisi,532 and in 1984 the first co-ordinated confrontation of the bubonic model was mounted by zoologist Graham Twigg. In it he argued that the specific problems of the reproductive capacity of fleas in cold winters and cold climates, and the presence of appropriate rodents (i.e. the black rat) in appropriate numbers for them to prey on, were too great to be surmounted. He expressed doubts about whether the speed of transmission was consonant with the requirements of bubonic plague. The bubos and other symptoms described in the contemporary literature he thought were potentially associated with other diseases as readily as with bubonic plague. He also brought into play references to chroniclers who mentioned the death of larger animals as well as men, and concluded that the answer was more likely some form of disease of animals that also affected humans, and made a tentative case for anthrax, spread person-to-person.

Detractors from this view raised the issue of our ability to compare like with like, the dangers of using sources selectively, and the lack of any real evidence for widespread larger animal deaths in the plagues of the later sixteenth and seventeenth centuries. It was in one case also noted that the physicians of the day knew the difference between pestilence buboes and anthrax carbuncles.533

Bean rejoined the debate in 1982. Accepting bubonic plague as the agent, he reasserted that since there was undoubtedly a rapid spread of the disease during winter months, ‘it was reasonable to assume … there was pneumonic plague’. However, he recognised that if the plague had indeed assumed different seasonal characteristics (bubonic in summer, pneumonic in winter), we should expect to see different mortality rates – much higher in the areas affected in winter. The fact that this was not borne out in studies of summer mortality in Orvieto, Siena and San Gimignano (with mortality rates of 50 per cent or more) he sought to explain away by arguing that cities were at far greater risk.534

In 1986 the issue of the black rat itself came under scrutiny by David E. Davis, a zoologist at North Carolina State University. Following a survey of archaeological and documentary indications of rats and a consideration of the ecological requirements of Rattus rattus, he concluded that ‘the accumulated evidence, interpreted by ecological and epidemiological methods, does not support the traditional view that black rats were responsible for the Black Death’. Subsequent studies concluded that the archaeological evidence for Rattus in the medieval period was more considerable than Davis had allowed for, but one French study raised the issue of the viability of the flea carrier itself:

the rat population had gradually grown from a fairly restricted one in the early Middle Ages to a significant one in the 11th and 13th centuries. The rodents spread along the major highways explaining the very different geographical impact of the various plague epidemics of the early and late medieval periods. However, the mystery of the exact mechanisms by which plague spread has still not been entirely elucidated, since the Asian rat flea, Xenopsylla cheopis, whose role as vector was demonstrated by P.L. Simond [an early twentieth-century bubonic plague researcher], could not have survived in the temperate European climate.535

The role of rats came under increased pressure, this time from Iceland in 1996. This related to two plague outbreaks argued to be of the Black Death type, in 1402–3 and 1494–5, which may have resulted in more than 50 per cent mortality rates, and which were able to attack the human population even in winter. The researcher, Gunnar Karlsson, argues that there were no rats in Iceland at the time, and that in winter the plague spread rapidly into areas of the country which enjoy a mean temperature of 0°C at that time of year. He supported the view of an earlier scholar, Jon Steffensen, who had proposed that pneumonic plague, not bubonic plague, was the cause.536

A further alternative was proposed by Susan Scott and Christopher Duncan in 2001.537 They had taken the very detailed documentary evidence from the outbreak of plague in Penrith, Cumbria, in 1597 and used analysis of this, in conjunction with wider epidemiological and environmental concerns over the identification of bubonic plague as the killer, to argue that whatever it was, the pestilence was transmitted person to person, had a long incubation period (around thirty-two days, allowing it to be spread widely before a plague event was obvious), and was haemorrhagic in nature. They adopted the term ‘haemorrhagic plague’. They were also able to develop a simulated model of what should happen in terms of death rates and compare this with real late sixteenth-and seventeenth-century examples of plague outbreaks (assuming that these were the same as the Black Death itself, of course). The argument against bubonic plague as the culprit was based around the low speed of transmission of this disease, the impact of climate on rat and flea vectors, and the low mortality rate of modern bubonic plague in comparison with reported and inferred death rates during the medieval pestilence. As with Twigg, critics of this approach cited selective use of documentation.538

The most sustained published attack on the bubonic model, drawing on many of the specific concerns raised previously, has come from historian Sam Cohn.539 His approach was essentially to deconstruct the century-old assumptions regarding the relationship between modern bubonic plague and the documentary evidence for the spread not only of the 1348–9 outbreak of pestilence, but the subsequent fourteenth-century outbreaks across Europe. In it he uses important new evidence for contemporary description and understanding of the plague from plague tractates, as well as assembling a formidable array of primary documentary evidence from wills and other quantifiable sources. He does not pretend to know what the disease was, but, like the pre-Yersin historians, he is certain that it was not the bubonic plague. His arguments against that diagnosis are well worth reading, but can be summarised as:

– The symptoms do not satisfactorily tie the disease down as bubonic plague.

– There is no contemporary report of elevated rat mortality (although there are chronicles reporting birds and sheep and other kinds of animals dying during the plague period).

– The mortality rates far exceed anything like modern bubonic plague.

– The seasonality of the outbreaks defy standard bubonic plague progression.

– The evidence for reducing mortality rates in successive outbreaks, combined with increasing proportions of child mortality, suggest increasing acquired immunity, something not possible for modern bubonic plague.

– In some well-documented Italian cities, epicentres of plague outbreaks were not associated with granaries as modern bubonic plague has been, but were associated with poor, artisan parishes.

– The pestilence struck regions where the rat/flea vectors could not have been present.

Also using documentary evidence, but bringing statistical analysis to bear on them, historical anthropologist James Woods, and colleagues from Pennsylvania State Universities and Georgetown University, re-examined the use of bishops’ registers (recording the replacement dates of parish priests who had died) in the diocese of Coventry and Lichfield.540 They discovered that the standard one-month estimate for the replacement used by most researchers actually varied considerably between a few days and several months, hiding the true dates of the plague’s arrival at any given parish, and thus potentially underestimating the speed of transmission. They also found that hidden within the overall blanket of nine months (March to November 1349), in which the pestilence raged in the diocese, individual archdeaconries suffered in shorter bursts of four to six months, and noted that if this were true at archdeaconry level, the speed of transmission from parish to parish would almost certainly be faster again. This they saw as profoundly different from anything presented by the nineteenth/twentieth-century evidence, and as an argument against bubonic plague.

An unusual explanation for the plague has been presented by Mike Baillie, a palaeoecologist specialising in events dated in palaeoenvironmental sequences such as tree rings and ice-cores.541 In 2006 he presented his idea that the Justinian and Black Death plague outbreaks were associated with very significant environmental and climatic changes, possibly as a result of comet strikes; what the plague was actually caused by he remains cautious about, but he believes it was air-borne (and not bubonic) and raises the possibility of biological pathogens being introduced into earth’s atmosphere by the comets. Such a proposal, he recognises, will be seen as far-fetched by many, but the environmental markers he establishes for the fourteenth century are convincing and important for the study of this period.

One final twist to this swirl of debate and research has been the report that bubonic plague may actually be transmissible faster than previously thought. Researchers at the Division of Vector-Borne Infectious Diseases at the National Center for Infectious Diseases in Colorado conducted a study of an American carrier of bubonic plague (O. montana). In their words:

in contrast to the classical blocked flea model, O. montana is immediately infectious, transmits efficiently for at least 4 days post-infection … and may remain infectious for a long time because the fleas do not suffer block-induced mortality. These factors match the criteria required to drive plague epizootics as defined by recently published mathematical models.542

While this still requires a rodent or similar vector to produce the plague reservoir, the long incubation period and short infection window described in the existing bubonic model could be dismissed. If the disease was bubonic plague, the impact this research might have on, for example, Benedictow’s carefully plotted chronology of spread and infection could be considerable. In any event, their acceptance that the characteristics of modern bubonic plague do not match at all well with the documentary evidence of past epidemics is important.

This short summary demonstrates the difficulties in getting a definitive identification. The capacity for pathogens to evolve, the variability of the primary documentary sources, and the complexities surrounding the science of the epidemiology and DNA character of the plague agent all present formidable challenges. However, further research into the evidence from specific cities, towns, villages and manors remains of the greatest importance in helping to solve this mystery.

A Consideration of the Evidence – The Plague’s Arrival

The other great unknown about the plague, contingent of course on what caused it, is the definite timing of its arrival and the speed and manner of its spread across the country. Again, as a prelude to a consideration of London’s contribution, it is necessary to look at current evidence. The most frequently deployed evidence comes from a range of medieval chroniclers. The advantage of this kind of source is that many had lived through, or were near contemporaries of, one or more of the outbreaks; some were eye-witnesses. The disadvantage is that there were differing concepts of truth within chronicles, used to impart separate messages to the reader,543 and that the sources for such chronicles might rely on second-or third-hand information. The evidence is, therefore, not straightforward.Table 5 shows the range of dates in a selection of chronicles broadly contemporary with the time of the epidemic. Considerable variation exists bracketing a three-month period from late June to the end of September. There is also a disagreement over a south or west coast entry point. On this uncertain basis, scholars have ascribed different levels of dependability to the chronicles, from overall caution to confident assertion that a date in late June or early July must be accepted.544 Most agree that the south coast, specifically Weymouth (Melcome), was the first port affected.

There are other more objective records which can be brought to bear. These fall into three categories. The first source is the evidence in the bishops’ registers. These documents, kept by each bishop for his diocese, recorded among other matters the filling of clerical vacancies within the diocese, whether caused by resignation, cessation, exchange or death. The latter group, where they can be identified, are clearly of great significance in helping to track the plague, since abnormal clusters of vacancies may be taken to illustrate heightened mortality in a particular area. However, the dates recorded are of replacement of the rector or vicar, not of the death itself. To be able to use the evidence, we need to understand the lag between the two. This lag was the time it took to report a death to the episcopal authorities, identify a suitable replacement, institute him to the benefice and record the fact in the register. It has been generally estimated at between twenty-one and forty-two days, suggesting a standard period of thirty-five days to some.545However, closer examination of the register for the diocese of Coventry and Lichfield, unique in recording the actual dates of death for the clergy, shows that the median period was twenty-two days and that there are reasonable grounds for arguing that the earlier in the outbreak the death occurred, the shorter the lag.546

Dorset lay within the diocese of Salisbury and the bishop at the time was Robert Wyville (fl. 1330–75).547 The earliest realistic contender for a plague-related institution was at West Chickerell (near Weymouth) on 30 September 1348. Prior to this, institutions were relatively infrequent and clustering was absent; afterwards, the frequency of institutions markedly increases in parishes and towns such as Warmwell, Dorchester and Wool from 9 to 24 October. If the vacancies were due to deaths, this would point (on the basis of the Coventry and Lichfield analysis) to deaths mounting from early September into October in the Weymouth/Poole area. Further institutions occurred in the Salisbury area (some 50 miles north-east) between 20 and 28 October.548 There is good coherence in this distribution, providing the basis for estimating a general speed of spread of 1 mile to 1.5 miles per day.


When Written

Suggested Plague Date

Outbreak Location


Geoffrey le Baker


Before 15 August 1348

A Dorset seaport

Thompson 1889a, 98–9

Ralph Higden


24 June 1348


Lumby 1882, 344–6

Robert of Avesbury

1340s and ’50s

1 August 1348


Thompson 1889b, 406–7



7 July 1348

Melcombe, Dorset

Haydon 1863, 213–4

John of Reading


After Midsummer 1348

Not given

Tait 1914, 106–10

Greyfriars of King’s Lynn

1360s and ’70s

21 June 1348

Melcombe, Dorset

Gransden 1957, 274

Thomas Stubbs


30 September 1348

Not given

Raine 1886, 418

Anonimalle, St Mary York


1 August 1348


Galbraith 1927, 30

Henry Knighton




Martin 1995, 99

Table 5. Chroniclers’ estimations of the date and location of the arrival of the plague in England.

Nevertheless, the uncertainty over the lag between vacancy and institution requires corroborative evidence. This exists to a degree in some surviving royal documents, particularly the Patent Rolls. Among many other items, these recorded the occasions of royal presentations of candidates for vacant churches in the king’s hands. A church might enter royal control for one of two principal reasons at this time: the lord (secular or religious) may have died; or the religious house, having rights of presentation, may have been foreign (normally French), and thus its properties were sequestered by the Crown during the war. Either way, the king was entitled to propose his candidate for the vacancy, which the bishop either confirmed or not. The key point here is that the date of royal presentation lay somewhere after the vacancy/death but before the actual institution, so it acts to narrow down the date of death. Taking the Weymouth/Salisbury area of Dorset, we find a cluster of presentations between late September and mid-November 1348, none further than 20 miles from Weymouth: Dorchester on 20 September; Bincombe on 8 October; Bradford Peverell on the 16th; Piddlehinton on the 27th; Tolpuddle on the 30th; Portesham and Abbotsbury on 6 November; two churches in Wareham on the 8th and 12th, and on 13 November Owermoigne and Blandford.549 These may well represent deaths from mid-September through to the beginning of November.

A further source of evidence is sometimes available from manorial documents. The estates of lords, bishops and monastic houses all required complex administrative and legal apparatus to account for production and to regulate the labour force working the land. Such apparatus gave rise to periodic documentation in the form of account rolls and court records. These survive in England better than in any other European country and, while their composition and survival are highly variable, they can provide exceptionally clear and objective information about the arrival and duration of the pestilence.

Where clerical vacancies occur in the same locale for which such manorial evidence exists, these documents can be used to ‘anchor’ the evidence from the registers. An example of how this type of evidence can assist comes from Gillingham in Dorset, 40 miles north of Weymouth. A recent discovery of a series of court rolls for three manors centred on Gillingham shows quite clearly that exceptional mortality was recorded during the first two weeks of October 1348.550 Deaths peaked between 5 and 27 November (the dates of successive courts) but had effectively subsided by early February 1349. Two clerical vacancies in nearby Shaftesbury (just 4 miles southeast) were filled on 29 November and 10 December. Allowing for around twenty-two days’ lag, these vacancies would have occurred between 7 and 18 November, agreeing remarkably well with the manorial documents. This timing appears consonant with further manorial evidence, this time from the manor of Curry Rivel, Somerset (also about 40 miles north-east of Weymouth), where numerous deaths were recorded in court rolls between 15 October and early December.551

Taken as a whole, there is a very strong impression that the plague made itself apparent in September at the earliest and developed rapidly through October and November. The implied later date for the plague’s appearance makes much more understandable the letter of Ralph of Shrewsbury on 17 August (see Chapter 1), requiring preparation against an expected threat, and that of the Bishop of Winchester dated 24 October, establishing that the ‘coastal’ regions were, finally, under attack.

Other evidence previously offered in support of the chroniclers’ summer date for the outbreak of plague can be readily refuted. Two detailed references have been published (and repeated) by way of proof and confirmation of the chroniclers that pestilence was abroad in July and August 1348 in Somerset and Dorset. The first stated: ‘By the beginning of August [implying 1348], most of the tenants of Frome Braunch in Somerset were dead and there were other deaths in North and South Cadebury.’ The reference, found in the Calendar of Inquisitions Post Mortem, is in fact to 1349.552 The second recorded the ‘statement that the king had noted in July 1348 that the mortality of men in the prebend of Bere Regis and Charminster in the present pestilence is so great that the lands thereof lie untilled’. This reference, found in the Calendar of the Fine Rolls, is again to July 1349.553

While more synthesis is clearly needed in this area, there is a strong possibility that the deadly character of the pestilence became apparent in September and not June or July 1348554 and spread outward in October just as the season began to worsen and temperatures to drop. This implies that its speed of transmission into southern England was greater than has been previously entertained, and that this speed was not apparently unduly influenced by cooling temperatures.

The Experience of London & its Contribution to the Debate

Currently, the jury is still out on two key and linked debates: what the disease actually was, and the manner of its arrival and spread into England. London can contribute to these debates in a number of ways.

The first way in which the London study can inform this debate is over the issue of the rats themselves. Leaving aside for the moment the problem of Iceland, the bubonic model requires that the initial epizootic phase occurs for each rat colony before the fleas depart their dying hosts to begin the infection of humans. Archaeologists in London (and indeed across the UK) now routinely sample key stratigraphic groups (pit fills, ditch fills, cesspit fills, etc.) for environmental remains, using established sampling procedures which involve both the hand-collection and wet-sieving of bulk soil samples. Recovered through such procedures are the bones of fish, birds, amphibians and small mammals. Such tiny bones do not routinely survive, being readily susceptible to decay and destruction by natural processes, but the general incidence of rat bones (Rattus rattus) can be considered.

London has produced unequivocal evidence of the presence of R. rattus in the period both immediately before the outbreak of plague and during the later fourteenth century. Excavations adjacent to the medieval Guildhall have produced fifteen positively identified R. rattus bones from among a collection of discarded cat skeletons (used by furriers and skinners), scavengers no doubt, along with the remnants of red kites and crows. The yard deposits were dated to 1280–1350, and probably towards the latter date, since the yard also contains features associated with construction works on the Guildhall which were completed in the 1350s. 555 Four bones were located in a flood-prone area of New Palace Yard, Westminster, dated to c. 1230–1350.556 In refuse dumps within the city ditch, eight R. rattus bones were recovered at Aldersgate dated to 1350–1400; and four ‘rat’ bones (not specifically identified as R. rattus) at Newgate dated 1340–1400.557 Rat bones were recovered from a hearth in an undercroft of the substantial mansion house of Sir John de Pulteney (himself a victim of the 1349 pestilence), dated to the later fourteenth or early fifteenth centuries; and from fourteenth-century domestic rubbish dumps at the quay adjacent to Baynard’s Castle. Two rat bones have also been recovered from the Augustinian priory of St Mary Merton, Surrey (about 7 miles south-west of the city), one from within the stone infirmary hall (dating to 1360–90) and one from the adjacent garden (1300–90).558

Further afield, a partial skeleton of a black rat was recovered from an ashy occupation layer around a hearth in the stone hall of a moated manor at Chalgrove, Oxon, dated to the late fourteenth or fifteenth centuries; rat bones were found at the Bishop of Winchester’s manor in Witney, Oxfordshire, dating to the fourteenth century, and a similar partial skeleton was found in the guardroom of the inner ward of Barnard Castle in County Durham dated to 1330–1479. Finally, the presence of rats in a mid-fourteenth-century monastic setting is confirmed by references of payments made to rat-catchers at Durham Cathedral priory in 1347 and 1356. 559 One researcher claimed: ‘There is nothing surprising about the almost complete exemption of the English nobility and landed class from “The Great Pestilence”. It just happened that the house-rat could not make itself at home in their castles.’560 The evidence both from London itself and other sites across England refutes this.

There are, however, significant blanks in just the places we might expect to find the evidence – for example, many of the thirteenth- to fourteenth-century waterfront reclamation dumps which have been sampled but from which no rat bones have been identified. By definition, such deposits are normally waterlogged (encouraging excellent organic preservation), they are composed of dumped urban rubbish behind wooden or stone revetments (and so should be expected to contain disposal evidence of any mid-fourteenth-century mass rat deaths) and they are precisely the locations where many plague historians have posited that the impact would be the worst – along waterways. These references do not represent a comprehensive national survey but, as far as the bubonic model is concerned, it would appear that the black rat was present and reasonably widely distributed in and around the London region. But we cannot yet speculate on the size and number of colonies that may have existed, and thus the degree to which they could have supported a widespread epizootic. The evidence, patchy as it is, proves that rats were present, but not that they carried the plague.

The second contribution concerns the seasonality of the outbreak. We can be certain that deaths had not reached plague proportions in London by 25 October 1348 (when Edington wrote his letter in his Southwark palace, praying that his own city and diocese of Winchester be spared the impending onslaught from the coast). We can be equally certain that that scale had been achieved by 14 November (when the blanket indulgence to all London citizens was issued by the Pope). Depending upon its incubation rate, the infection probably established itself in the city no earlier than late September or early October. This is much later than the date provided, for example, by Benedictow (4 August), which he calculated from the earliest chronicled date of 29 September, including an eight-week period for the bubonic plague to become epizootic and then epidemic. This delay in development may be crucial in terms of the environmental argument.

If the plague was bubonic, and the biology of the black rat and the flea is as has been stated, we would have to accept that this winter was a particularly mild one, as below 10°C bubonic plague spread is significantly slowed. London’s current average monthly temperature range does not exceed this figure from November through to March, and evidence from broader climate studies suggests that the fourteenth century saw deteriorating climatic conditions for much of northern Europe (and indeed globally), with oak growth rates at a consistently low level from the early 1340s for a decade, and one study of eastern England indicating at least three severe winters in the decade up to 1348.561 So a mild winter does not appear to have been the case.

Fig. 6 (see p. 85), showing both the dates of writing and the dates of enrolment of wills, and the mortality rates for Stepney manor, establishes with as much certainty as can be hoped for that in the largest city in the land, the plague’s transmission speed increased as London slipped into the grip of winter. The only previous attempt to integrate the progression of the pestilence through London with the environmental and seasonal requirements of the bubonic model proposed that the first epidemic manifestation emerged in later September, leading into a phase of infection among the poor, drawn out by the cool autumn months. This triggered a scare among the wealthy leading to a rise in will-making in November and December, and was followed by a full explosion of plague with the arrival of warm, spring weather.562 The evidence from the Stepney manorial deaths refutes this ‘considerable’ time-lag between the onset of plague among the poor and the subsequent panic of the wealthy: the poor began to die exactly at the time when the richer began to draw up wills in exceptional numbers. The will enrolment trajectory matches very closely that of the will-making, and, taking the available evidence for lag between probate and enrolment, the likely death curve sits just three weeks later than the will-making curve. Therefore, the plague quite clearly escalated throughout the winter and had already peaked by Easter and the onset of warmer weather. This combination of circumstances does not appear to match the expected bubonic model. So for the traditional bubonic model to stand up we would have to accept that there were more rats than archaeological evidence is as yet offering, and that the winter was particularly mild. Even then, the chronology of the outbreak has been significantly revised.

This speed of transmission ought also to be considered in the light of the date at which the plague struck England as a whole (see above). Since official documentation suggests that the plague broke out in earnest possibly in late September and more certainly in October 1348, in a range of places from Dorset and north Devon right across to London, there would be insufficient time for the process of epizootic development in rat colonies, followed by spread into the human population, to allow for land-based transfer. We would need to accept that either London was infected directly from the Continent (quite possible) or that the agent was not reliant on rats and fleas.

A fourth contribution of the London evidence is that, as was first identified by Sam Cohn, the impacts of the four successive plagues under review reduces steadily and markedly. He suggests that some kind of immunity to the pathogen was acquired over a relatively short time. Such immunity would not be characteristic of modern bubonic plague.563 Cohn argues that the pestilence in Florence broke out not in the zone where the grain store was situated, a recurrent feature of twentieth-century outbreaks in India, but particularly in the impoverished, highly populated quarters of the city.564 London can advance no definitive evidence on this aspect of the plague since rich and poor lived very much side by side in the city, but there are hints that the pestilence killed the poor more readily.

A comparison of the rates of prevalence of bone lesions (picked up during life prior to contraction of the plague), from which those who were buried at the East Smithfield cemetery suffered, with those from pre-1348 medieval cemetery assemblage, has concluded that once infected, the weaker and less healthy individuals were more likely to succumb to the disease.565 It seems likely that, in general, poverty would contribute to an increased incidence of such environmental or occupational stresses. Coupled with the fact that the physical stature of those buried at East Smithfield was generally lower than other late medieval cemetery assemblages, this supports a general notion that the poor were more affected than the rich. In her consideration of the squeeze on day-patient care at Westminster Abbey from 1350 onwards, Barbara Harvey speculates that one possible reason may have been that ‘those who survived [the plague] were perhaps the fittest who did not have these kinds of needs’.566

Therefore, while rats were present in London to act as a potential bubonic go-between, the season and speed of the outbreak would require that the epidemic was either very significantly a pneumonic variant bubonic plague, or was not bubonic plague at all. The apparent and rapid decrease in mortality rates of successive outbreaks after 1349 would suggest the latter. The need for more detailed scientific studies on pre-plague skeletal collections to establish the presence of Y. pestis ancient DNA would certainly help inform this debate.

If you find an error or have any questions, please email us at Thank you!