Death in Venice: A Digital Reconstruction of a Large 1 Plague Outbreak During 1630-1631

The plague, an infectious disease caused by the bacterium Yersinia pestis , is widely con- sidered to be responsible for the most devastating and deadly pandemics in human history. 14 Starting with the infamous Black Death , plague outbreaks are estimated to have killed around 100 million people over multiple centuries, with local mortality rates as high as 60%. However, detailed pictures of the disease dynamics of these outbreaks centuries ago remain scarce, mainly due to the lack of high-quality historical data in digital form. 18 Here, we present an analysis of the 1630-31 plague outbreak in the city of Venice, using newly collected daily death records. We identify the presence of a two-peak pattern, for which we present two possible explanations based on computational models of disease dynamics. Systematically digitized historical records like the ones presented here promise 22 to enrich our understanding of historical phenomena of enduring importance. This work 23 contributes to the recently renewed interdisciplinary foray into the epidemiological and 24 societal impact of pre-modern epidemics.


Main text 26
Disease outbreaks of the plague in the past centuries have been so devastating throughout 27 Eurasia that the term plague has become synonymous with a terrible disease. By killing 28 a substantial proportion of the human population, which took multiple generations to 29 recover, plague pandemics have had enormous impacts on the development of Eurasia. 30 Correspondingly, historical questions, such as the role of institutions and the socioeco-31 nomic impact of plague outbreaks [1], as well as epidemiological questions, such as the 32 causes, nature and interactions of vectors [2,3,4,5], seasonality and climatic patterns 33 [6,7] and even the distinction between plague and the Black Death [8], are still being 34 investigated. While previous studies have highlighted some common traits to plague epi-35 demics [9], such as the high impact on densely-inhabited cities acting as hotspots [10,11], 36 the importance of human-to-human transmission [12] and the effect of the plague on dif-37 ferent sexes [13], little is known about local outbreaks, due to the lack of detailed historical 38 data. 39 We analyze high-quality data from death records created during the 1630-31 plague 40 epidemic in Venice, whose initial investigation is limited and by now dated [14]. This 41 epidemic was part of the so-called "Second Pandemic", which started with the Black 42 Death and lasted until the early 19 th century. Originated in northern Europe (modern 43 France and the Rhineland) in 1623, this epidemic crossed the Alps approximately in 1629, 44 in the case of the territories of the Republic of Venice likely carried by imperial armies 45 on their way to Mantua. The cause of this specific outbreak in Venice has been linked to 46 the bacterial species Yersinia pestis [15], and with a set of surprising results, including an 47 uneven and unexpected impact on different cohorts by sex and age, a high parallel increase 48 of mortality due to a synchronous smallpox epidemic and a raise in public violence [16]. 49 Venetian death records from this period, also referred to as necrologies, are organized 50 by parish and contain the systematic registration of every death among the resident pop-51 ulation. These necrologies, edited by the parson, were established by decree since 1504 52 and kept in the archives of the responsible magistracy [17]. While death records were 53 commonplace in all Christendom since the late Middle ages, and are commonly used for 54 demography studies including on the plague [18,1], Venetian records were particularly 55 detailed. In the Patriarchal Archives of Venice, 54 out of more than 70 existing parishes at year are few and scattered. Based on our assessments, these record series are overlapping 59 and one (the former) constitutes the source for the other (the latter). We thus focus our 60 efforts on the Patriarchal records. An example page from a necrology record is shown in 61 Figure 1. Necrology records were kept in tiny and oblong books, with entries grouped 62 chronologically by day. Typically, the most recurring details given for every entry were:   Tables SI1, SI2 and SI3 for details on which causes of death were considered to be 101 plague.

102
The statistics of the causes of death give us a first insight. In Figure SI2a we show the 103 distribution of deaths grouped by cause and (conservatively) classified as related to the 104 plague or not. One can see how the two distributions are skewed, meaning that a small 105 4 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020  fraction of causes (5%) contributes to a large fraction of deaths (63%). However, while 106 the number of deaths clearly due to plague (N plague = 1007) and possibly non-plague 107 are similar (N not plague = 778), only 56 out of 156 causes could be clearly attributed to 108 plague, leaving more vagueness around the non-plague causes (in Figure SI2b the causes 109 with more than 50 deaths are listed). This seems to suggest that our plague-death counts 110 likely constitute a lower bound of the total number of deaths directly linked to plague, 111 which we cannot further refine from the records. In Figure 2b we show the time-series 112 of deaths belonging to the Sant'Eufemia parish, distinguishing between those caused by 113 the plague and the ones possibly due to other causes. Surprisingly, the first peak of the 114 epidemic begins with few references to the common symptoms of the plague (October to 115 November), when the records point instead to more generic and common illnesses, such as 116 fever or spasms [17]. Only afterwards the records start to extensively mention the plague 117 as the cause of death, well into the Fall of 1631. This might indicate an initial reticence to 118 acknowledge the epidemic outbreak, as well as a subsequent possible overemphasis of it.

119
This reticence might be caused by the public authorities' practice to quarantine the whole 120 household in their house when someone from it died of plague. It might also be due to 121 a surveillance issue generating a bias in the records: while many deaths were occurring, 122 medical examination was no longer taking place and the registrations of the causes of 123 death were not happening regularly, but instead in batches, leading to approximations.

124
Furthermore, several people were moved to quarantine areas (lazzaretti ) and died there, 125 while their registration happened subsequently, possibly by reporting generic causes of 126 death. It is thus likely that these deaths are also in large part attributable to plague. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020 of 1631. 131 We further verify that deaths by plague were not significantly affected by sex, under the 132 reasonable assumption that sexes were equally distributed in the population of Venice at 133 the time [20]. Indeed, the male to female deaths ratio was close to one (N male /N f emale = 134 865/917 ∼ 0.94), a result confirmed by the majority of the literature [1,12,22,23,135 24, 25], with few exceptions [16,13]. Furthermore, the distribution of illness duration 136 and of age at death did not significantly change with sex (see Figure SI3a and SI3b 137 respectively). Assessing the effect of the plague on age is challenging, as assumptions on 138 the age distribution of population at that time are quite difficult to make and historical 139 statistics are hard to find. Furthermore, the literature on the effect of the plague on 140 different age cohorts is still ambiguous. Nevertheless, our data are in line with previous 141 studies [26,27,18,28,1] indicating that the plague had higher relative impact among age 142 cohorts of typically low mortality, in particular adolescents and adults between 14 and 44 143 years of age, as shown in Figure SI3c and Figure SI3d. Even though these clusters seems to be well separated in time, there is no clear evidence 166 of a specific process or event in the history of the city that might have driven this spatial 167 distribution of localized epidemics in different parishes during 1631. We therefore assess 168 epidemiological models on data aggregated over all parishes. The plague is generally 169 modeled as a zoonosis, in which the transition from an epizootic (typically, in rodents) to 170 a human epidemic is mediated by animal fleas, the vector carrying Yersinia Pestis [30,29].

171
From here on, we refer to this model as the Rats-Fleas-Humans (RFH) model. At the 172 same time, other studies suggest that these models are not always preferable to explain the 173 outbreaks dynamics, especially due to the 'efficacy and speed' of some historical plague 174 6 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020.  Figure SI5d. (c) Example realization of a stochastic delayed behavioral SIR; the evolution of transmission rate β(I) is shown in Figure SI5e.
outbreaks [1], if compared to the typical dynamics of RFH models. We first confirm 175 that neither a deterministic RFH nor a deterministic Susceptible-Infected-Removed (SIR) 176 model can explain the presence of the 1631 secondary outbreaks (see Figure SI5b). We  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020.03.11.20034116 doi: medRxiv preprint a network model which is likely to resemble the modular structure of social contacts [33] 203 (further details on the simulations are given in the Methods). We find that few simulated 204 epidemics do resemble the data, as shown in Figure SI6a. However, as this happens in only 205 about 0.1% of the simulations, such alternative interpretation of the 1631 tail based on 206 pure stochastic effects and network structure, although reasonable, remains very unlikely.

207
In summary, we find a novel epidemic pattern of two peaks in the 1630-31 plague 208 outbreak in Venice. The first peak in 1630 was very high, and the outbreak highly 209 synchronized among all parishes; the second peak in 1631 shows temporal variability, and 210 was much less pronounced in strength. Most previous recorded cases show a single main 211 peak [6, 29, 5] of varying duration [18,12], with possible cyclical recurrence [6]. Relying 212 on fine-grained daily death records [1], we are able to confirm that the plague spanned 213 both the main peak and the long tail, over a period of more than a year and caused the 214 death of approximately 30% of the city's population.

215
Providing an interpretation of the two-stage process remains challenging with the 216 evidence at our disposal. Firstly, not all deaths could be clearly attributed to the plague 217 during the early weeks of the main peak. Generic causes of death such as fever and spasms 218 might indicate plague deaths as well as deaths due to other causes. A first hypothesis 219 is therefore that the same plague epidemic went on for more than a year, while being 220 aggravated by other concomitant causes during the main peak. An alternative hypothesis 221 is that two distinct plague epidemics took place instead, one during the main peak and 222 another during the long tail. Previous studies suggest the possibility of a transition from a 223 mainly bubonic to a mainly pneumonic plague, for example. Furthermore, we show that it 224 is also possible that such temporal pattern could be generated by the adaptation of hosts'  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10. 1101/2020 We would like to thank the support of the Patriarchal Archive of Venice and the State 244 Archive of Venice during data collection. We thank Paolo de los Rios and Giulio Rossetti 245 for useful discussions on diffusion processes on networks, and gratefully acknowledge the  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020.  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020.

11
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020 Secondly, two different co-authors have counted again all deaths from a sample of 20 352 parishes out of 70 (8 and 12 each), to further assess our main dataset, with the following    The delayed behavioral SIR model (Figure 4c) was defined using the following expression 390 for the transmission rate β(t) = β 0 e −I(t−τ )/I * , where β 0 , τ and I * were fitted parame-391 ters, together with the usual (constant) death rate γ and initial number of infected I 0 392 1 For more details, find here the description of possible metrics: scipy.spatial.distance.pdist.

12
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

396
Data availability 397 All code and data needed to reproduce plots and analysis presented in the manuscript 398 will be made available in a dedicated GitHub repository before publication.

399
13 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. Total_deaths 0 to 500 500 to 1,000 1,000 to 1,500 1,500 to 2,000 2,000 to 2,500 2,500 to 3,000 Missing 14 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. Supplementary Figure SI2: Distribution of number of deaths by cause, for the parish of Sant'Eufemia. In the records, 156 unique causes of death are found, of which 56 were attributed to plague. One can clearly see the skewed distribution, with the top 5% of causes accounting for about 63% of the overall deaths (a). For the sake of clarity, only causes with more than 50 deaths are listed; in bold the ones attributed to plague (b).

15
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020   16 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020  Supplementary Figure SI4: Pairwise Pearson correlation between cases time-series of each couple of parishes, as function of distance between the two parishes. The same scatter plot for all parishes (N parishes = 54) for the entire time-windows (a) and for the largest parishes, for the 1631 outbreaks only (b). The largest parishes are defined as those reporting more than 500 deaths (N parishes = 28).

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020.  Figure SI6: (a) Selected stochastic simulations of a simple SIR model on top of a small-word network. Two particular epidemics are highlighted, in order to show the possibility of having a large peak followed by a long tail, as present in the data. In shaded orange we only show, for sake of clarity, the simulated epidemics with lowest deviation from the data (RM SE < 50 − N = 93) (b) Best fit of deterministic behavioral delayed SIR. Although the model can fit very well the first part of the epidemic, it does not show a secondary outbreak, in 1631.

19
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted March 13, 2020. 20 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 13, 2020.

21
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 13, 2020.

22
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 13, 2020. . https://doi.org/10.1101/2020