Assignment 2/ World Civilization

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to Reviews of Infectious Diseases.

http://www.jstor.org

Three Days in October of 1630: Detailed Examination of Mortality during an Early Modern Plague Epidemic in Venice Author(s): Stephen R. Ell Source: Reviews of Infectious Diseases, Vol. 11, No. 1 (Jan. - Feb., 1989), pp. 128-139 Published by: Oxford University Press Stable URL: http://www.jstor.org/stable/4454756 Accessed: 20-08-2015 17:45 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

OXFORD JOURNALS O ·~OJ i~!iD• UNiV£AS:l'i"V PRE S S

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

REVIEWS OF INFECTIOUS DISEASES • VOL. 11, NUMBER I • JANUARY-FEBRUARY 1989 © 1989 by The University of Chicago. All rights reserved. 0162-0886/89/IIOl-0014$02.00

HISTORICAL ARTICLES

Three Days in October of 1630: Detailed Examination of Mortality During an Early Modern Plague Epidemic in Venice

Stephen R. Ell From the Department of Radiology, University of Chicago Hospitals, Chicago, Illinois

The epidemiology of medieval and early modern European plague remains highly con- troversial. It now seems likely that the epidemiology was not uniform throughout either the geographic or temporal boundaries of the plague in Western Europe. The Venetian plague of 1630 was extensively documented; day-by-day records were kept, and each mor- tality in the city was recorded in a set format. The days 23-25 October 1630, representing a period when mortality was beginning to increase sharply, are examined. In all, 1,163 deaths were recorded. They show a large preponderance of women; a mean age of 28, but a majority of cases clumped between ages 0 and 25 years; and an unequal sex ratio among children. Further, there was an identifiable smallpox epidemic raging simultane- ously with plague, and more than one-quarter of all the deaths in this period of high mortality were clearly due to nonplague causes. Deaths due to wounds and associated with violence were prominent in one parish, which suggests that in times of plague the breakdown in the normal machinery of government, in everyday patterns of life, and pos- sibly of mental well being resulted in an even more exaggerated death toll. These factors - violence, accidents, and other epidemics- have never been so definitively tied to a Euro- pean plague epidemic. In addition, there are hints that plague has a marked proclivity to kill pregnant women-their deaths far outnumber those anticipated-and that plague was very localized at a given moment within Venice itself, even during times of peak mor- tality.

Plague in medieval and early modern Europe has fostered a huge number of scholarly works. When Biraben wrote his two-volume work on European plague, published in 1975 [l], he consulted several thousand references, and the flow has not ceased. At least partly, this fascination lies in the extent of the epidemics. It is generally believed that Western Europe lost one-third of its population from 1348 to 1350 [2-5]. Subsequent epidemics did not match this level, but in cities they were especially devastat- ing. The Venetian plague of 1630-1631, a fraction of which this paper examines, left one-third of a city of 150,000 dead.

A further fascination (as well as a problem) comes from the scanty records available. Because disease was not conceptualized as it is today, because statis- tics were not the very lifeblood of government, be- cause descriptions were impressionistic and incom- plete, historians and historical epidemiologists are left with tantalizing pieces of a vast puzzle, the solu-

Received for publication 31 December 1987 and in revised form 2 May 1988.

Please address requests for reprints to Dr. Stephen R. Ell, Department of Radiology, Box 429, University of Chicago Hospi- tals, 5841 South Maryland Avenue, Chicago, Illinois 60637.

128

tion of which is nowhere in sight. Huge extrapola- tions are made from minute amounts of data, while reliable generalizations are few. It is not even univer- sally agreed that the disease called plague in the period in question was the same disease as the bac- terial infection with Yersinia pestis. A recent book has argued, albeit unconvincingly, that the disease in question was anthrax [6]. For all that has been written, there is little that can be said with certainty.

Nowhere is this more obvious than in the case of plague epidemiology. Scholars have placed the greatest mortality variously among children, young adult males, adult females, and elderly males [7-11]. For each view there is fragmentary evidence exten- sively extrapolated. It seems most reasonable sim- ply to admit that such an overall question lacks an answer at this point and may have none in an abso- lute sense. Considering the number of known fac- tors influencing epidemiology in plagues (let alone those not yet discovered) coupled with the meager knowledge we have of climatic conditions, rodent infestation, flea and other ectoparasite levels, rela- tive exposure, means of spread, forms of the disease, possible concurrent epidemics, and a myriad of other matters, there is little wonder that such a grand ques- tion (and others like it) cannot be answered.

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630

By contrast, this paper will try to address some smaller questions but, I hope, on surer footing. I shall try to determine, whenever possible, what people ac- tually died of and what factors can be discerned to have affected overall mortality. Particular groups within a population seem often to have been at par- ticular risk. Efforts will be made to disentangle the unique from the general factors involved. The spa- tial characteristics of a major plague epidemic in Western Europe have never been adequately de- scribed, although Carmichael has done work on smaller epidemics in Florence [7]. At least for the three days in question it is possible to see where the cases clustered within the city. Further, the possibil- ity that the magnitude of the epidemic increased mortality from other causes is examined in detail. These are small steps in a very long and uncertain journey, but I hope they will be secure ones.

The Venetian plague of 1630-1631 stands out among European plague epidemics as perhaps the most carefully documented of those from which documentation survived. Each death was recorded in a set format. The name, age, surname, occupa- tion, cause of death, days of illness, and place of residence of victims were routinely recorded . From the point of view of anyone trying to reconstruct the epidemiology of the epidemic, it is important that there was no bias in reporting. Infants who died dur- ing birth were reported. Women, children, non- Venetians, Jews - all found their way to the pages of the Venetian public health board. This removes the recording biases that have damaged, often in un- known ways, many efforts at reconstructing plague epidemics.

A number of analyses of this epidemic have re- lied on the curious existence of roughly contem- porary summary statistics of the epidemic (12-15]. These give totals of 46,490 or 51,903 fatalities for Venice and either 93,661 or 82,175 for Venice, Chiog- gia, Malamocco, and Murano (the adjacent portions of the Dogado or Duchy of Venice). The statistics for the Dogado break down the deaths, albeit in a confusing way, by age and sex. Although of immense interest at first glance, these documents raise more questions than they answer. If we accept that Venice suffered f\.J50,000 deaths, a quite reasonable figure given plague in other European cities, we must postu- late /"\J35,000-45,000 deaths in the towns mentioned above. In 1630, these towns did not possess 40,000 inhabitants among them [16]. In fact, Chioggia, one of the largest of these towns, lost 6,000 of a total

129

population of 13,000 [16]. This figure leaves the over- all statistic even more unlikely. The figures for Venice alone are broken down by month, and numerous graphs based on these documents have been pub- lished. Again, however, problems arise. In October, 2,100 deaths are reported, yet nearly 1,200 are recorded for the three days examined here alone, and a brief examination of the daily death logs shows the true figure to exceed 7,000. Despite their popular- ity as source documents, these primitive statistical breakdowns of the epidemic are both clearly impos- sible and of unknown origin. They are found in manuscripts now in the library of the Correr Museum in Venice, one set in a volume that also con- tains a narrative history of the epidemic. Yet no one knows who compiled them or when or how. Despite their historiographic prominence, these documents must be treated with extreme caution. As will be seen in the discussion of plague and pregnancy, one of these compilations posits an impossibility.

This author himself has undergone a full revolu- tion regarding the value of these documents. From a whole-hearted acceptance [17], he passed to grave doubts [18] and now flatly denies their value. To pass them by without comment, however, is to invite the suggestion of neglect of well-known and widely used documents.

If these documents leave us little better off than we were in ignorance of them, our ignorance of the demography of Venice in 1630 is more painful and ultimately more damaging. While it is generally be- lieved that the population of the city was 150,000 before the epidemic began [19, 20], we have no break- down of that number. We cannot reliably estimate what proportions of the population lay in what age/ sex groupings. It is reasonable, and widely ac- cepted on the basis of fragmentary evidence from other European cities and on generalized patterns in underdeveloped societies, to postulate that one- third to 40% of all inhabitants were children under age 15 years. Old age was rarely achieved; persons who reached adulthood generally lived into late mid- dle age [21]. As for the ratios of the sexes, we find severe problems. It is generally thought that women outnumbered men on the order of 1.1 :1 (10, 21]. The reasons for this pattern (if indeed it was the pattern) are unclear.

Venice presents special problems. Aspects of Venice's social situation (the city moved from being a major European power and the greatest port of Europe to being the center of a more regional econ-

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

130

omy and power structure) make it extremely hard to guess what the age and sex composition of the adult population was.

In the spring of 1630, Venice went to war with the Haps burgs over the rule of Mantua. Hapsburg arms prevailed, but plague proved the ultimate victor, claiming 40,000 lives in the region. The Venetian army was so decimated it could not be reorganized [22]; Venetian commitments at sea demanded sev- eral thousand oarsmen, and Venice garrisoned some of its remaining colonies [23]. These facts suggest that thousands of young men were absent or already dead when the epidemic took hold. Again appear- ances are somewhat deceiving. Venetian armies were almost exclusively mercenary, and a reasonable es- timate of the Venetian component at Mantua is "-' 1,000 men -with the number of accompanying Ve- netian prostitutes being totally conjectural. Patterns in the fleet were similar, with many of the galley crew- men being foreigners [22-24]. While it may further be argued that Venetian merchants resided in for- eign lands, many foreign merchants lived in Venice. In short, while it is reasonable to postulate an ex- cess of women in the Venetian population, there is little ground for making that excess any larger than the general trend of the period.

This study, the first step in a complete analysis of the epidemic, focuses on three days chosen not at random but to represent a time when the average number of deaths per day was rising sharply. I thought that days showing an inflection of the mor- tality curve would be among the most revealing, and so the results indicate. Complete examination of all recorded deaths will be necessary before this epidemic can demonstrate all it has to show, but the conclu- sions available from this limited work bring to light and prove phenomena previously hypothetical.

Materials and Methods

The records for this paper are from the Archivio di Sta to di Venezia - specifically the records of the Provveditori alla Sanita, series Necrologio, registers 858 and 859. Neither register is occupied solely by records of these three days, and all registers cover- ing the epidemic period (856-863) were examined to be sure no deaths from these dates were recorded else- where. Pagination is a problem in this series because the pages are rarely numbered at all. The following convention was used: The first page recording deaths was the point of reference for establishing page 1.

Ell

If the writing began on a left-hand page, that page was number l; if on the right, that page became num- ber 2. This preserves the normal pagination pattern of books. Anyone wishing to check a particular ref- erence would have to keep this in mind, especially since in rare cases the registers have a few apparent page numbers written in, although whether they were intended as such is unclear.

Each case was recorded on a computer-generated form by use of Database III Plus software (Ashton- Tate, Inc., Torrance, Calif.). This same software was used for simple sums, such as number of persons from the same parish. Statistical analysis was per- formed with Statgraphics (STSC, Rockville, Md.), Javelin (Javelin Software Corp., Cambridge, Mass.), and PFS: Professional Plan (Software Publishing Co., Mountain View, Calif.), all three of which soft- ware packages generate graphs. Other database func- tions, such as cross-tabulations and string searches, were performed with RBase System V (Microrim, Inc., Redmond, Wash.), because of its simpler com- mand language. Particularly because Javelin and Statgraphics use slightly different methods to arrive, in many cases, at the same statistical end, most im- portant calculations were performed on both.

Another important consideration in methodology was how to decide whether someone died of plague. Most records mention fever or fever and petechiae. Obviously, plague is far from the only disease that can produce these symptoms. The epidemic, how- ever, was part of one that affected almost the whole of Italy, and contemporary descriptions exist show- ing that plague, as we know it, was the major dis- ease involved [25]. Records were kept by a large num- ber of persons, not all physicians, so that the symptoms mentioned vary for that reason, also. For- tunately, no single physician accounted for all the reporting for a single parish even on one day. This eliminates much of the chance that a given locality would seem to vary due to the words a particular person employed to record cases.

In general, if there was nothing to exclude plague and the patient died after an illness of 10 days or fewer, the case was taken to represent plague. Un- less there was some compelling reason to accept the case as plague (such as specific mention of plague or of one or more buboes), longer duration excluded the case from the plague grouping because it is rare for a plague victim to linger long and not recover [26]. Mention of smallpox (variola), of a wound, of death from "old age" (vechiezza, and always with

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630

a long duration), of a fall down a well, and of many other specific causes of death also ruled out the vic- tim as a plague casualty. Inevitably, there were many difficult cases. In the fall, what of a 75-year-old woman who dies with fever in 6 days? Pneumonia would not be unreasonable, as would many other causes, but such cases were accepted uniformly as plague because it was assumed that plague was the major cause of mortality and was only excluded when the cause of death was clearly different. An- other exceptionally difficult area is that of children. Mention of vermi (worms) was taken to indicate a predominantly gastrointestinal presentation. Most such cases had long durations, but some did not. These were not taken as plague cases, however, as worms in seventeenth century medicine, despite its literal meaning, represented a gastrointestinal illness, which simply does not fit plague. Even more diffi- cult is the child reported dead, almost always in fewer than IO days, due to spasemo (seizures). In infants, this was somewhat arbitrarily attributed to tetanus. In older children, unless duration was long or the child was noted to have had seizures and therefore presumably epilepsy for some time, the death was attributed to plague. While seizures are not a prom- inent feature of plague, the high fever of plague would be expected to produce a significant number of seizures in children (it is reckoned now that some 20% of all children experience a febrile seizure at some time in their lives). It is extremely important overall to recognize that the methodology employed here gives figures for other causes of death that are almost certainly minimums, with the possible excep- tion of deaths in infants.

Plague as a Specific Disease

It is important at the outset to offer a few remarks about plague as a specific disease for two reasons. During the historical period in question, the terms for plague became generic; that is, they were used to describe many epidemics, a reasonable number of which were not the disease, plague, understood as a clinical entity then and as a specific infectious disease with a known etiology now. The Venetian plague of 1630-1631 was both, but it was clearly pre- dominantly an outbreak of the infection produced by Y. pestis. Second, plague is a complex disease in ecologic terms, with nonhuman carriers and vectors involved.

Human infection by Y. pestis is an acute bacterial

131

infection with an incubation of anywhere from 1 or 2 to "-' 7 days. Onset is acute with high fever, pros- tration, and a characteristic lesion at the site of drain- age of a regional lymph node, usually in the groin, axilla, or subauricular area. This is the bubo, hence the name bubonic plague, an infected mass of lymph nodes, grossly enlarged, from which the organism can readily be recovered. Patients are said to ex- perience a characteristic sense of restless anxiety and dread. Untreated, the disease has a mortality of 70%. Septicemia occurs in many fatal cases and may also occur without a bubonic stage. Plague pneumonia is almost always a hematogenous metastatic infec- tion. The conditions under which plague may spread directly from person to person via a respiratory route are extremely specific. The temperature must be near freezing and the humidity in a high but rather nar- row range. Basically, primary pneumonic plague is a winter disease, and documented epidemics have been very few in number [21-29].

Plague, however, is immensely more complex as a phenomenon in human society because humans are usually incidental victims of what is first and foremost an enzootic disease. In the classic model of plague transmission worked out at the beginning of this century, human epidemics require certain preconditions. First, an enzootic reservoir is required. That is, some wild animal species or, as is more and more recognized, groups of species are chronically infected by the plague bacillus. Many such animals are not killed by the disease but harbor it over in- definite periods. When, in some manner rarely documented or understood, these wild animals in- fect the human-commensal rodent, Rattus rattus (the black rat), the rats show an exquisite sensitivity to the organism and die in great numbers. The rats' fleas, of the species Xenopsyl/a cheopis, are heat- dependent. As the rat corpse cools, the ectoparasites abandon it for other hosts. They prefer other rats but, in their absence, will attack humans. These par- ticular fleas will have fed on the blood of an animal with a high-grade Y. pestis septicemia, and their digestive tracts may eventually become blocked by the multiplying bacteria. When such a flea bites a new host, it regurgitates bacteria into rather than feeding from that host [28-31].

In this chain of transmission, humans are inciden- tal victims; they are the rat flea's last resort for a host. Further, this type of transfer and, in fact, plague epi- demics among humans, can only occur in particu- lar circumstances. The fleas are only active enough

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

132

within certain temperature ranges, and humidity lev- els are crucial. The rat itself, has an immunity to plague that fluctuates with the seasons. It has been shown, for example, that the immunity of black rats to plague is lowest in the spring. This is probably one reason that, while an initial outbreak of plague among humans can occur in any season, recurrences in the same locale tend to take place in the summer. Yet, even this is not absolute, for in places like Viet- nam, plague follows rainfall regardless of season. (Of course, temperature variation in Vietnam is much less extreme than in most of Europe.) In Vietnam, the transmission pattern of plague is not well un- derstood, however, and may not depend on rats [32-37].

Indeed, considerable controversy exists over the extent to which the classic model of plague transmission -which was elaborated in India and China and seems, without much question, to explain the expression of plague in those regions - can be employed without modification to explain plague in medieval and early modern Europe. While we need not tarry excessively on these questions, some sense of the problems involved is valuable in evaluating the results of any paper attempting to reconstruct the epidemiology of European plague between 1347 and 1800.

For example, J. F. D. Shrewsbury attempted to ap- ply the classic model of plague transmission to me- dieval England [38]. As a result, he rejected as plague any epidemic that began in the fall, and the net ef- fect of his analysis was to assign plague a minimal role in the demography of England. At least three detailed studies, in which original accounts of the disease in question left no doubt that it was plague, have since appeared, and all three concluded that fall was the commonest season for plague in medieval Europe [39-41].

In the 1940s, the French claimed to have shown that plague in North Africa (and later Iran) was passed from person to person by human-commensal fleas [42-44]. Opinion on the ability of human- commensal fleas to support a major epidemic has tended to follow linguistic lines. Its value in explain- ing European plague, however, has not been over- looked [1, 8].

If the reader will excuse the length of this digres- sion, its point will be elucidated. The classic descrip- tions of plague transmission simply do not fit plague in medieval and early modern Europe. A great many ingenious theories have been advanced to explain

Ell

these discrepancies (for example see [4, 6, 17]), but none has proven satisfactory. The extremely high death rates, patterns of multiple cases per household, the extremely long epidemics (that under discussion lasted almost 2 years in one city), and a host of other factors render the rat-flea transmission model irrele- vant, but we have no accepted substitute. Simplistic claims (such as a high incidence of primary plague pneumonia [4, 51) are totally unsupported by con- temporary descriptions of the disease, which clearly identify metastatic (secondary) pneumonic plague but ascribe no role to the primary form, nor do they explain how definite climatic conditions not com- mon in an Italian December could obtain repeatedly in August. In short, despite our extensive knowledge of plague transmission in the modern world and in spite of innumerable learned explanations of the Western European plague experience, we do not know much about the mechanisms of plague in West- ern Europe at all.

To the question of who was most at risk to die of plague and why, the answers have been wildly at odds. As noted, basically every possible age and sex grouping has been proposed as that most highly at risk. One proponent of high risk for women and chil- dren has claimed that risk was due to the consump- tion of grain containing aflatoxins, which suppressed immunity [9]. The hypothesis that young men were most at risk has been attributed to variations in iron metabolism with age and sex [17, 45]. In short, the historiographic arena of medieval plague is one with a great many more answers than questions and many more hypotheses than primary source studies.

Results

Sex ratios. Records of the Provveditori alla San- ita for the dates 23-25 October 1630 reveal 1,163 fa- talities. The most striking initial finding is the dis- parity between the sexes: 59% of the dead were females and this ratio extends upward from age 5 years (table 1). In general, plague tends to show no sexual proclivity in children before puberty [46]. While some argument might be made that in Venice there tended to be more women resident at any given moment because of the demands of trade and the fleet, this could hardly apply to children. It is clear from the laws passed by the Senate relative to this epidemic and from the edicts of the Provveditori alla Sanita that women and children were recognized as being at special risk. In March 1631, the government

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630 133

Table 1. Mortality by age, sex, and cause of death: Venice, 25-28 October 1630.

No. of deaths due to plague/ nonplague (total)

Age group (y) Males Females

0-4 25/ 41 (66) 28/ 39 (67) 5-9 34/20 (54) 43/28 (71)

10-14 45/6 (51) 59/9 (68) 15-29 109/13 (122) 134/28 (162) 30-59 120/32 (152) 181/42 (223) 60+ 18/ 14 (32) 52/43 (95)

Total deaths 477 686

restricted the movement within the city of both groups (47].

Why girls and women were at such exaggerated risk is unknown. Exposure factors often do account for relative risk between the sexes but not ordinarily among small children (46]. There is no known Ve- netian custom whereby boys of 5-10 years of age were treated markedly different from girls (48-49]. The sex distribution over these 3 days (mirrored in pre- liminary examination of the rest of the records) can- not be explained with our current level of knowl- edge. Speculation over possible metabolic differences between the sexes in adults crumble when such a pat- tern is repeated in children before puberty. It seems most reasonable to invoke an environmental cause. Perhaps women and girls were most often at home and therefore had greater exposure to the fleas of dying rats. On the other hand, it is very clear that women in this period were not really stay-at-homes and often provided second incomes, as did some chil- dren [22-23). Did they go more often to markets and other places where infection was spread?

One possible reason for the disparity in mortality between the sexes may lie in the preplague popula- tion of Venice itself. Figures from 1642 show a marked predominance of females in the population (roughly 1.2:1), and these figures follow an epidemic that killed many more women than men [20). Un- fortunately, these figures come from a period of in- tense immigration (it is said that 50,000 persons moved to Venice between 1631 and 1660) [20] and are, therefore, of dubious value in determining pre- plague proportions. Because this immigration was encouraged openly by the government and favored males, it is not unreasonable to postulate a signifi- cant excess of females before this period. But the gap between what is reasonable and what occurred

Percentage of nonplague deaths, by gender

Male Female Total deaths

62 58 133 37 39 125 12 13 119 II 17 284 21 19 375 44 45 127

1,163

is often startling. It does seem that most cities had an excess of females in the population, but again the reasons for this are unknown, and it is possible that the marked predilection of the disease for females of all ages merely reflected the composition of the population and is not a true disparity. At present, however, this is merely conjecture.

Age distribution of mortality. The mean age of a person dying during the Venetian plague of 1630-1631 for the days in question was 28 years. The median age was 24 years, however, and a frequency histogram shows that cases tended to cluster from age Oto 25 years (figure 1). As age increased, it was often estimated. The commonest age listed is 40, but it is usually followed by the abbreviation ca., indi- cating that the age is not exact. There were several persons listed as being in their nineties, but one sus- pects that this was an exaggeration. Overall, the mor- tality histogram is very much what would be expected of a population profile by age. In preindustrial soci- eties, it is reckoned that at least one-third of the population is less than 15 years of age and, before modern times, old age was not commonly attained [4, 19-21). Thus, the profile of plague mortality by age is simply that of the population itself, which renders the sex distribution of cases even more dif- ficult to understand. If environmental or cultural fac- tors so strongly affected sex distribution even into childhood, why did those (admittedly unidentified) factors not favor one age over another?

Deaths due to causes other than plague: other dis- eases. Of the 1,163 deaths, more than 300 (27%) were clearly not due to plague. In other words, at the height of a plague epidemic, nonplague mortal- ity was so high that it would produce alone almost the same number of deaths as the entire epidemic, if it continued at the same level. Let us consider this

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

134

IJ) Q)

~

150 -

~ 100- lll u. 0

Age (years)

l£l Males m Females

m ~ f

/

Figure 1. Histogram of total mortality, by age and sex, in Venice, 23-25 October 1630.

assertion for a moment. At the rate of about 100 deaths per day (due to nonplague causes) during an epidemic that lasted for 16 months, one would pre- dict >50,000 nonplague deaths. This number is thought to equal all the mortality from the epidemic. In short, plague increased nonplague mortality as- tronomically.

The exceptional detail of the Venetian records in this regard is of tremendous interest. Among other things, it is clear that there was a simultaneous small- pox epidemic. While many cases are totally unclear as to cause of death or in some ways suggestive of smallpox, 12 cases are specifically listed as being due to smallpox. Of these, only one was an adult, nine were 3 years old or younger, one was 11, and one 12 years of age. Seven took more than 12 days to die; the duration of illness is not given for three. The age distribution and the length of illness reinforce the impression that these were indeed smallpox cases. Smallpox in the seventeenth century was primarily a childhood killer [4, 50]. At an average of four fa- talities per day, smallpox could not have been a ma- jor factor among the deaths in this epidemic unless its numbers of fatalities increased relative to the period in question. During the three days in ques- tion here, smallpox accounted for almost exactly 1 OJo of mortality. Preliminary examination of records for the whole epidemic indicates that it did accelerate and was a numerically more significant cause of death. The fascination lies in the fact that a second epidemic can be clearly discerned in the records of

Ell

this plague epidemic, something postulated but never well documented before [7].

Much commoner among children and account- ing for the vast majority of nonplague deaths in the age group were gastrointestinal diseases, almost in- variably described as vermi (worms) or flusso (diar- rhea). Symptoms of this type are not a significant part of plague and the duration of illness before death averaged 18 days, much longer than would be expected of plague. So common were nonplague causes of death among children less than 5 years of age, that plague and other causes claimed nearly equal numbers of victims (figure 2 and table 1). Since the city could not have sustained a background child- hood mortality of the magnitude seen during the period under study, it seems likely that plague was indirectly increasing death from other causes.

Some reasonable speculation can be offered for this. Under normal circumstances the Provveditori alla Sanita enforced strict provisions regarding what drinking water and food could be sold in the city [51]. During the plague, such rules could not be en- forced. Thus, although the mechanism of disease was not understood, it is possible that empiric knowl- edge of what was safe to eat and drink existed. In default of such regulations, water- and food-borne pathogens may have killed many. Typhoid fever, for example, might easily have caused many deaths in this way. Beyond the question of food, there is the whole matter of the care of children. Even in the 3- day period intensively examined here, at least five sets of parents perished, three sets along with at least one child. In the midst of so much carnage, it is dif- ficult to imagine that orphaned children could have been cared for satisfactorily. Thus, the breakdown of the normal structure for provisioning the city as well as neglect of children whose parents had per-

(JJ 80 Oo-4

! years Q) :;:; [TI5-9 ~ 60 8a10-14 Ill u..

0 40 ,._ Q)

20 .0 E :, z 0

PLAGUE OTHER PLAGUE OTHER

MALES FEMALES

Figure 2. Childhood mortality by age, sex, and cause of death, in Venice, 23-25 October 1630.

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630

ished probably account for the exceptionally high nonplague childhood mortality. While the govern- ment of Venice never failed to continue functioning or providing its basic services, it is inconceivable that all public health provisions were enforced when pub- lic health officials were examining hundreds of corpses a day.

Nonp/ague deaths: violence. Among adults, non- plague causes of death varied more than among chil- dren and actually accounted for a smaller percent- age of the total, except among persons more than 60 years of age. Old age, lingering illnesses of sev- eral years, cancer, and many other causes are cited, including the gastrointestinal illnesses noted among children. There is one intriguing exception to this commentary, however, and that is the curious case of the parish of San Pietro. The church of San Pietro stands not far from the Arsenal, the great Venetian ship-building and general-purpose government fac- tory. The Arsenal, in fact, was one of the largest em- ployers in Europe at the time, with a staff, in its days of glory, of several thousand [22, 23, 52]. The Ar- senal, at the orders of the government, manufactured beds for the ill and boats for transporting the dead, despite heavy casualties among its employees [53]. On 23-25 October 1630, 31 people are listed as hav- ing died of wounds. Of these, 18 came from the sin- gle parish of San Pietro. Again, these fatalities were recorded by several different persons, excluding any recording bias. Thus, more than 58% of deaths sec- ondary to wounds originated in a single parish near the Arsenal. Unfortunately, narrative histories are silent on this point; we do not know whether there was a major accident in the shipyard (certainly a pos- sibility) or an outbreak of violence, perhaps a riot. The latter seems possible, for accidental deaths were usually referred to as deaths due to accident specifi- cally, sometimes with elaboration such as "fell in well." The language used even suggests that violence was involved, for the termferito (wounded) was used by the Council of the Ten in its regulations regard- ing what sorts of occurrences physicians were obliged to report to the neighborhood authorities [54]. It seems possible that there was a civil disturbance near the Arsenal at the end of October 1630. This impression is reinforced by the fact that San Pietro was the patriarchal church of Venice until after the fall of the Republic. Not only was it in a neighbor- hood full of potential strife, but it was the seat of ecclesiastic government in the city as well. The com- bination may well have proven too explosive.

135

Yet such an argument must be made from silence. How skillfully could Venice cover up such an inci- dent? The author of the only contemporary narra- tive of the epidemic was openly critical of the govern- ment, yet he makes no mention of such an occurrence. An alternative may lie in the Arsenal's situation it- self. With about 1,500 [12, 22] employees at the be- ginning of the epidemic, the Arsenal was repeatedly called on to produce beds, boats, and other necessi- ties of the battle against plague. Meanwhile many of the most-skilled artisans themselves died of plague. It seems more likely that accidents in a be- leaguered and unusually inexperienced work force explain this localized phenomenon. It was, nonethe- less, another way in which plague increased back- ground mortality.

Plague and pregnancy. Among the dead were 29 women listed as pregnant (da parto ). This represents 2.5% of all mortality and 8.5% of women 15-50 years of age. Thus, mortality among gravid women was extremely high. In a summary of fatalities dur- ing this epidemic, which is unfortunately of unknown provenance and reliability, pregnant women account for an even greater number, about 12% of the total mortality [13]. The latter figure is impossible, since a population with that many pregnant women would have a crude birthrate beyond any ever known to have existed.

The figures derived from the records themselves, however, are quite plausible. Depending on what proportion of pregnant women died in the epidemic, the anticipated birthrate would be in excess of 25/1,000 population annually. Such birthrates are be- lieved to have been the rule in early modern Euro- pean society [21]. It was widely noted among com- mentators on European plague, from the age of Justinian on, that pregnant women were at particu- lar risk of plague death [1, 55]. The records of this epidemic are probably the first to show this relation clearly.

Again, it is important to note that recorded deaths among pregnant women are to be taken as a mini- mum. It is unclear at what stage pregnancy was recognizable by physicians or skilled laymen in this period. The extreme pressure exerted upon those recording the deaths probably limited the recogni- tion of pregnancy to the beginning of the second tri- mester or thereabouts. Such a consideration obliges us to postulate that the minimum number of women dying who were actually pregnant should be consid- ered to be at least 50% higher, suggesting that 13%

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

136

of deaths among women of reproductive age were of gravid females (5% of all mortality). Obviously all pregnancies would not end in birth, but the num- ber of victims of the epidemic who were gravid was exceptionally high.

It is generally thought that reproduction in cities in the period was not adequate to maintain popula- tion [4, 21, 56]. A city of 150,000 would need 4,000 births per year to support its population, assuming an average life expectancy of 35-40 years. At least 20% of pregnancies abort spontaneously, potentially raising the true portion of gravid dead to 5%. As- suming that Venice did not maintain its population without immigration, it is not possible to assess the birthrate, although 2,000-3,000 per year is reason- able. Yet with one of 20 dead being a pregnant woman, we would postulate about 7,500 (6,000, al- lowing for a 20% abortion rate) births per year, a figure suggesting that pregnant women are dis- proportionately represented among the dead. These figures assume that every pregnant woman in Venice died. If the proportion were equal to that of the whole population (that is, one-third mortality), births would rise to rv20,000 per year, clearly an im- possibility. If no correction is made ( every pregnant woman died) and the 2.5% of overall mortality is used, about 2,500 births would be expected - a fairly appropriate number but clearly showing a tremen- dously increased risk for pregnant women.

Another problem lies in the question of when in pregnancy women became particularly susceptible to plague. In many other diseases, this is a late (usu- ally third trimester) problem. If that is so, it explains why the diagnosis of pregnancy was made so confi- dently but forces us then to say that the dead who were recognized to be pregnant may have come from only one-third of those pregnant. No matter how one computes it, the risk of death from plague was ex- aggerated for the gravid woman.

To render matters more difficult yet, very little is said about the pregnant women who died other than that they were pregnant. These cases were taken to represent plague mortality for the following reasons: two other clearly pregnant women died, but the cir- cumstances are elaborated- they died in childbirth specifically and were counted as nonplague mortal- ity. If we consider that childbirth in this period resulted in the death of the mother in 2%-10% of cases, as is commonly assumed, then we would postulate a crude birthrate of 24/1,000 population annually at 10% perinatal mortality and up to

Ell

C/l 200 -~ - 015-29 ( <IS -i50 <IS : :0

6 -059 years LL

0 100 ~

Q) 50 .c

E ::, z 0

PLAGUE OTHER PLAGUE OTHER

MALES FEMALES

Figure 3. Adult mortality by age, sex, and cause of death, in Venice, 23-25 October 1630.

120/1,000 at a 2% rate. Thus, the separation of those women actually dying in childbirth or in the post- partum period (seen on some dates not discussed here) argues that the other pregnant women were be- lieved to have died of plague. Given the association of pregnancy and plague by those recording the deaths, pregnancy was probably taken as an adequate indication of plague as the cause of death, unless another cause was specified.

If we consider the immunology of plague, a plau- sible reason for the exaggerated mortality among gravid women is readily discerned. It has recently been recognized that pregnancy represents an im- munodeficient state, mainly due to a suppression of T cell immunity [57]. Until recently [58], it was thought that plague immunity was almost solely an- tibody mediated. Recent studies have demonstrated, however, that there are three types of immune re- sponses to plague - antibody, T cell, and nonspecific, with T cell immunity most important among them [59]. Plague is neither always fatal nor always clini- cally apparent, so that we are probably seeing the result of an increased susceptibility to clinical dis- ease and an almost complete mortality among in- fected pregnant women. Mortality due to untreated plague is typically 60%-80% [25], with an unknown number of subclinical cases. The results of this study suggest both a higher proportion of fatal cases and a lower number of subclinical ones within the gravid portion of the population.

Spatial characteristics of the epidemic. Fatalities are recorded from 77 parishes in Venice. This represents the majority of parishes and initially sug- gests that the spatial distribution of the epidemic was relatively uniform, given the vagaries of reporting fatalities during a limited period. A more detailed examination, however, suggests the opposite. A

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630

limited number of parishes actually accounted for 700/o of fatalities. There appear to have been four main foci of high mortality during the three days in question. One was centered near the Arsenal and the parish of San Pietro; more than 160/o of all deaths fell in the contiguous parishes of Sant Apostoli, San Canciano, San Bartolomeo, San Salvatore, Santa Maria Nuova, and Santa Maria Formosa. These par- ishes lie in a region just north and east of the Rialto Bridge, roughly in the direction of San Marco. Other centers of mortality were San Marcuola and San Geremia (near the modern railway station) and San Giacomo dall'Orio. Twenty parishes accounted for 773 of these deaths (700/o), with the main cluster- ings as noted. Thus, the spatial distribution was far from uniform.

This attribute of the epidemic must be considered to be very tentative at this time because it may rep- resent a reporting artifact. That is, those persons recording mortality and examining the dead may not have visited all parishes in a 3-day period. Nonethe- less, there is some reason to exrapolate from these data. For parishes like San Pietro where mortality was very high, deaths were recorded by several differ- ent persons and were also recorded on all three days. This suggests that reporting was relatively consistent. Even those parishes with only a handful of cases show those cases spread over all three days, and each patient was visited by more than one official of the public health board. Thus, there is some reason to consider this early pattern of spatial distribution as at least possibly characteristic. It certainly agrees with many medieval and early modern commentators on plague, who described epidemics in terms of the se- quential destruction of neighborhoods. A charac- teristic fourteenth-century tract states, for example, that plague first killed everyone in a house, then ravaged a parish, then spread to one section of a town after another [60]. This sequential pattern also helps to account for the very long temporal span of the epidemic, some 16 months in this case. The fuller implications of the pattern, should it be borne out, remain to be discovered.

The reader should not be mislead by these re- marks. The epidemic of 1630-1631 involved essen- tially all of Italy. Plague was not localized to a few neighborhoods in Venice; it was a widespread phe- nomenon. The emphasis here is on how, at a time of high death rates, the disease was spatially situ- ated in Venice. The spotty localization seen here has been described elsewhere [7, 40]. In Venice, the sharp-

137

ly limited land available made population density relatively uniform and high (some major cities in this period still had grazing land within them). Thus, high numbers of deaths near San Pietro di Castello com- pared with negligible numbers in the Rialto district on the opposite side of the Grand Canal reflect areas of, as far as we know, comparable population density.

Conclusion

This paper represents the early fruits of a larger proj- ect, the complete examination and analysis of the day-by-day records of the Venetial plague of 1630- 1631. Three days at the point of inflection of the curve of death rates were chosen for a detailed anal- ysis. The records reveal no discernible reporting bias but suggest a number of features not well docu- mented in previous studies of plague in this period.

The age and sex distributions of mortality are striking. While the mortality curve follows the an- ticipated age profile of the population, there is an excess of females in every age group. Overall, 590/o of the victims were women. While this may to some extent reflect an excess of females in the population, the disparity is still striking.

Nonplague causes of death were extremely high. Notable among such deaths were a small but signifi- cant number due to smallpox, an epidemic of which occurred simultaneously with plague. Also of con- siderable interest are deaths attributable to neglect of children and probably of the old, along with those due to accident. The pattern of nonplague mortal- ity suggests a breakdown of some aspects of normal social order, with increased risk of accidental death and a loss of the usual means of providing for the disadvantaged. This is not a surprising conclusion, but the records examined here are almost unique in that they provide sufficient detail about causes of death to substantiate such an argument.

For reasons that have only recently been eluci- dated, the long-held belief that pregnant women were particularly at risk for plague is also given credence. The proportion of pregnant women among the dead is much too high to be accepted as a chance occur- rence. Lastly, the spatial distribution of deaths within the city suggests that, even during a major epidemic, neighborhoods were ravaged sequentially (another widely held belief at the time). Lack of spatial unifor- mity helps to explain how such an epidemic could last so long.

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

138

The reasons for the phenomena observed in this study are not yet clear. It is extremely difficult to supply an explanation• for the sex distribution of cases, for example. I hope that the continuation of this study and the reconstruction of other primary sources will take the historiography of plague in me- dieval and early modern Europe from the very broad hypotheses that have so often characterized it to firm and reliable observation and explanation. Then in- deed will we cease to see as through a mirror darkly and step into the light of a more confident knowledge.

References

I. Biraben JN . Les hommes et la peste en France et dans les pays europeens et mediterraneens. 2 vols. Paris: Mouton, 1975;1:13, 111-86;2:51-88

2. Gottfried RS. The Black Death. New York: Free Press 1983:11-1, 106-48

3. Ziegler P. The Black Death. New York: Harper 1971:16, 23, 43, 78, 163

4. McNeill WH . Plagues and peoples. Garden City, NY: An- chor Press/Doubleday 1976:79-97, 149-98

5. Carpentier E. Autour de la peste noire: famines et epidemies dans l'histoire du XIVe siecle. Annales: Economies, So- cietes, Civilisations 1962;17:1062-92

6. Twigg, G. The Black Death: a biological reappraisal. New York: Schocken, 1985

7. Carmichael AG. Plague and the poor in Renaissance Flor- ence. Cambridge: Cambridge University Press, 1986:48-71

8. Swellengrebel NH. Interhumane transmissie bij builenpest. Ned Tijdschr Geneesk 1967;111:647-9

9. Matossian MK . Did mycotoxins play a role in bubonic plague epidemics? Perspect Biol Med 1986:29:244-6

10. Bullough VL. Nutrition, women and sex ratios. Perspect Biol Med 1987;30:450-60

11. Russell JC . British medieval population. Albuquerque: Univer- sity of New Mexico Press, 1948:147, 216-27, 260-2

12. Peruzza P. Venezia e la peste . Venice: Marsilio Editore, 1979:97-9, 99-113, 123-55, 294-342

13. Archivio di Stato di Venezia, Provveditori alla Sanitil. Decreti de! Maggior Consiglio e del Senato communicati ai Prov- veditori alla Sanita, Filza 1, folios 407 R-V

14. Bibliotheca Museo Correr. Codex Cicgogna 1509, folios 98r-99v

15. Bibliotheca Museo Correr. MS 3055/ 8 fasc . 8 16. Guy RJ. Chioggia and the villages of the Venetian lagoon .

Cambridge: Cambridge University Press, 1985:86 17. Ell SR. Immunity as a factor in the epidemiology of medie-

val plague. Rev Infect Dis 1984;6:866-79 18. Ell SR. The Venetian plague of 1630-1631: a preliminary

epidemiologic analysis. Janus, 1989 (in pr ess) 19. Pollitzer R. A review of recent literature on plague. Bull WHO

1960;23:313-400 20. Beltrami D. Storia della popolazione di Venezia dalla fine

de! secola XVI alla caduta della Repubblica. Padua : Cedam, 1954:11-78

Ell

21. Flinn MW. The European demographic system, 1500-1800. Baltimore: Johns Hopkins, 1985:22-68, 85-110

22. Pullan B. Rich and poor in Renaissance Venice. Cambridge, Mass: Harvard University Press, 1971:253, 315, 373 1-9

23. Pullan B, ed. Crisis and change in the Venetian economy in the sixteenth and seventeenth centuries. London: Methuen, 1968

24. Mallett ME . Hale JR. The military organization of a Renais- sance state : Venice c. 1400 to 1617. Cambridge: Cambridge University Press, 1984:313-460

25. Cipolla CM. Fighting the plague in seventeenth century Italy. Madison, Wis: University of Wisconsin Press, 1981:7-18, 51-88, 89-109

26. Poland JD . Plague . In: Hoeprich PD, ed. Infectious diseases: a modern treatise of infectious processes. 3rd ed. Philadel- phia: Harper and Row, 1983:228-9

27. Hunter GW Ill, Frye WW, Schwartzweldler JC . A manual of tropical medicine. 4th ed . Philadelphia : W. B. Saun- ders, 1966:209-12

28. Teh WL. A treatise on pneumonic plague. Geneva: The League of Nations Health Organization, 1926:1-77

29. Pollitzer R, Mayer KF. The ecology of plague. In: Mays J, ed. Studies in the ecology of disease. New York: Hafner, 1961:433-501, 476

30. Hirst LF. The conquest of plague : a study in the evolution of epidemiology . Oxford: Clarendon Press, 1953

31. Heisch RB. Zoonoses as a study in ecology with special ref- erence to plague, relapsing fever, and leishmaniasis. Br Med J 1956;2:669- 73

32. Beloch J. Beviilkeriingsgeschichte Italiens. 3 vols. Berlin: De Gruyter , 1953;2:410-4

33. Bahmanyar M, Cavanaugh DC. Plague manual. Geneva: WHO, 1976:4

34. Delprat J. La peste bubonique, ii propos de dix cas observes ii l'Hopital Principale Fianarantsoa (Madagascar). Medna Trop 1974;34:567-72

35. Velimirovic B. Untersuchiingen iiber die Epidemiologie und Bekampfiing der Pest in Sudvietnam . Zentbl Bakteriol I 972;228:482-508

36. Davis DE . The characteristics of rat populations. Q Rev Biol 1953;28:373-401

37. Benenson AS. Plague. In: Top FH , Wehrle PF, eds . Com- municable and infectious diseases. 7th ed. St. Louis: C. V. Mosby, 1972:450-6

38. Shrewsbury JFD. A history of bubonic plague in the British Isles. Cambridge: Cambridge University Press, 1970:49

39. Eckert EA . Spatial and temporal distribution of plague in a region of Switzerland in the years 1628 and 1629. Bull Hist Med 1982;56:175-94

40. Ell SR. Interhuman transmission of medieval plague. Bull Hist Med 1980;54:497-510

41. Morison AS, Kirshner J, Molho A. Epidemics in Renaissance Florence. Am J Public Health 1985;75:528- 35

42. Blanc G, Baltazard M. Recherches experimentales sur la pesti: !'infection de la puce de l'homme. Acad Sci [D] (Paris) 1941;213:813-816

43. Baltazard M, Seydian B. Enqiiete sur les conditions de la peste au Moyen-Orient. Bull WHO 1960;23:157-67

44. Baltazard M, Bahmanyar M, Mostachfi P, Eftekhari M, Mofidi C. Recherches sur la peste en Iran . Bull WHO 1960;23:141-55

This content downloaded from 131.94.186.23 on Thu, 20 Aug 2015 17:45:21 UTC All use subject to JSTOR Terms and Conditions

Plague Epidemic in Venice, 1630

45. Ell SR. Iron in two seventeenth-century plague epidemics. Journal of Interdisciplinary History 1985;15:445-57

46. Butler T. Plague and tularemia. Pediatr Clin North Am 1979;26:355-66

47. Archivio di Stato di Venezia. Provveditori alla Sanita. Decreti Registro 17, folio 293

48. Arnaleti G, Pastore J, eds. Storia della cultura veneziana: ii seicento. 2 vols. Vicenza: Neri Pozza, 1984;2:186-289

49. David JC. A Venetian family and its fortune, 1500-1900. Philadelphia: American Philosophical Foundation, 1977

50. Hopkins DR. Princes and peasants: smallpox in history. Chicago: University of Chicago Press, 1983:52, 74-5, 87

51. Archivio di Stato di Venezia. Provveditori alla sanita. Rub- brica delle leggi 1:2R, llR

52. Lane FC. Venice: a maritime republic. Baltimore: Johns Hop- kins, 1973:362-4

53. Archivio di Stato di Venezia. Provveditori alla sanita. Decreti Registro 17, folios 133r-v, I 59r, I 83r-v

139

54. Zago F, ed. Consiglio dei dieci: deliberazioni misti. 2 vols. Venice: Comitato Editore, 1962:42, 48, 61, 67

55. Procopius. History of the wars II, xxxii:30-36 56. Lee WR. European demography and economic growth. Lon-

don: Methuen, 1979:lll-68 57. Weinberg ED. Pregnancy-associated depression of cell-

mediated immunity. Rev Infect Dis 1984;6:814-31 58. Burrows TW, Bacon GA. The basis of virulence in Pasteurella

pestis: an antigen determining virulence. Br J Exp Pathol 1956;37:481-93

59. Wake A, Morita H, Wake M. Mechanisms of long and short term immunity to plague. Immunology 1978;34:1045-52

60. Jakob de Stockstals. Ex regimine magistri Jacobis de Stock- stals doctoris in medicina. In: K. Sudhoff, ed. Pestschriften aus den ersten 150 Jahren nach der Epidemie des "Schwar- zen Todes" 1348. Archiv fiir Geschichte der Medizin 1912-1913;6:373