Journal of Medieval History Fo rP ee ! " # $ %# &' #' ( # ) ' # ( ' ' rR ev ie w On ly URL: http://mc.manuscriptcentral.com/rmed Page 1 of 32 Journal of Medieval History 1 2 3 The mystery of plague in medieval Iceland 4 5 6 7 Abstract 8 9 10 11 According to written sources, two memorable outbreaks of ‘plague’ occurred in Iceland in1402-4 and 12 1494-95. Here we argue that these were episodes of pneumonic plague, caused by Yersinia pestis, and 13 that the likely mortality was no more than 25% of the population in both cases. This contrasts with the 14 higher rates (50-60% and 30-50%) postulated elsewhere. Although it is recognised there are alternative Fo 15 explanations for plague in Iceland, greater caution needs to be taken in interpreting the direct and 16 indirect evidence for its demographic effects. A lower mortality rate fits better with a less widespread 17 and more fragmented epidemic. The numbers and types of Icelandic farms which might have been 18 vacant during the fifteenth and early sixteenth centuries are given more detailed consideration than in rP 19 previous accounts. ‘Farm abandonment’ in the fifteenth century was continually driven by of a series of 20 environmental and economic factors and need not be interpreted as a demographic collapse caused 21 solely by the plague. Greater attention is also given to undersanding how plague could have reached Iceland and the biological, ecological and sociological factors which might then have sustained it. 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 1 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 2 of 32 1 2 3 The mystery of plague in medieval Iceland 4 5 6 7 Introduction 8 9 10 Icelandic annals record two severe plague epidemics for 1402-4 and 1494-95. This 11 mortality, which was sufficient to merit special attention from the annalists, appears to 12 have been of the same order as that reported elsewhere for the ‘second plague 13 pandemic’ which spread through most of Europe from 1346 onwards. Some of the 14 evidence for Iceland has been taken to suggest mortality rates of over 50%.1 Fo 15 16 Both Ole Benedictow and Gunnar Karlsson have summarised the information that can 17 be gleaned from the documentary sources and work by Icelandic scholars writing in 18 Icelandic.2 It seems probable that annals were written shortly after the actual events rP 19 20 took place but they are preserved in manuscripts which date from the fifteenth century 21 onwards. The first epidemic appears to have started in the autumn of 1402 and that the 22 disease had arrived by ship, possibly in the harbour of Hvalfjörður on the south west coast. The epidemic is generally considered to have spread to the north of Iceland ee 23 24 before Christmas and continued until Easter 1404. The annals refer to a ‘plague 25 winter’, ‘plague year’ and ‘later autumn plague’.3 For the second epidemic there is no 26 contemporaneous annal but the four surviving accounts from the sixteenth and rR 27 seventeenth centuries give the same general picture of it. A significant plague started 28 29 in the south-west and spread everywhere except for Vestfirðir, the north-western 30 peninsula. The epidemic seems to have been active in the north and north-west of the country in the winter of 1494-95.4 ev 31 32 33 The nature of these epidemics is puzzling. Benedictow, in line with many scholars, 34 has argued that they were primarily rat-based bubonic plagues, the aetiological agent ie 35 being the bacterium Yersinia pestis. He rejected pneumonic plague as an explanation 36 because it is not easily transmitted and those that are infected die on average within 37 w two to three days of the onset of symptoms, having little opportunity to pass the 38 39 disease on.5 Gunnar Karlsson, however, has maintained that there was no evidence 40 that the Black Rat (Rattus rattus) had ever established itself in Iceland. Thus, because the epidemics persisted throughout the winter months when fleas are unlikely to On 41 42 43 1 Gunnar Karlsson, ‘Plague without rats: the case of fifteenth-century Iceland’, Journal of Medieval 44 History 22, no. 3 (1996), 263–84; Gunnar Karlsson and Helgi Skúli Kjartansson, ‘'Plágurnar miklu á 45 ly Íslandi', Saga 32 (1994), 11–74. Reference is made to the earlier Icelandic-language article when 46 necessary to cite its more detailed discussion. Some of the criticisms of these articles were also made 47 by Jón Ólafur Ísberg, ‘Sóttir og samfélag’, Saga 34 (1996), 177–216. His points, particularly on the 48 weaknesses of the written evidence in relation to the early fifteenth century, were elaborated on in 49 idem, ‘Annálar og heimildir um Svarta dauða’, Ritmennt 2 (1997), 55-75. The scientific evidence for 50 the case made by Jón Ólafur Ísberg has since grown and this article adds to his arguments about the 51 nature of the textual evidence. The following abbreviation is used in this paper: DI: Diplomatarium 52 Islandicum. Íslenzkt Fornbréfasafn I–XVI (Copenhagen, 1857–1972). 2 53 Gunnar Karlsson, ‘Plague without rats’, 263–5; O. J. Benedictow, Plague in the late medieval Nordic countries. Epidemiological studies (Oslo: Middelalderforlaget, 1992). 54 3 Ibid., 267. Karlsson appears to translate ‘mandeyda haustit seinnara’, DI, vol. 3, 739, as ‘posterior 55 plague autumn’ but the translation here is preferred. 56 4 Gunnar Karlsson, ‘Plague without rats’, 267. 57 5 O. Benedictow, Plague in the Late Medieval Nordic Countries: Epidemiological Studies (Oslo: 58 Middelalderforlaget, 1996), 221. 59 60 2 URL: http://mc.manuscriptcentral.com/rmed Page 3 of 32 Journal of Medieval History 1 2 3 operate as vectors, the outbreaks were most likely the pneumonic rather than the 4 bubonic form of this disease. Although Karlsson1 has pointed out that the mortality 5 rates implied by the annals might have been exaggerated, and that other diseases such 6 as typhus fever might also have been prevalent,6 he sees evidence of unoccupied 7 farms in the fifteenth century as supporting the idea of high mortality rates. We argue, 8 9 however, that empty farms are continually caused by a number of economic and 10 environmental factors as well as intermittent outbreaks of disease. This phenomenon 11 can be explained, therefore, by a series of processes rather than simply the result of 12 outbreaks of plague. 13 14 Difficulties with the reconciliation of historical and modern plagues have led some Fo 15 authors to argue that the first and second plague pandemics were not caused by Y. 16 pestis but by a completely different organism, as yet unidentified, whilst others have 17 suggested anthrax and typhus as possible causes.7 Nevertheless, this paper argues that, 18 despite Iceland’s apparent isolation and the delay in the ‘Black Death’ reaching it, a rP 19 20 strong case can still be made for identifying Y. pestis as the disease organism which 21 ravaged Iceland on the two occasions mentioned above. However, the way in which 22 the disease epidemic progressed is likely to have been more complex than has been previously understood and it needs to be recognised that aspects of human social ee 23 24 behaviour in a rural community – medieval Iceland was entirely rural – served to 25 facilitate the spread of the disease, particularly in its pneumonic form. Such a view is 26 based on an understanding of how modern plague epidemics progress and a rR 27 reassessment of the significance of the written evidence for the plague in Iceland. 28 29 30 Descriptions of the plague in Iceland ev 31 32 Before the details of mortality are discussed it is necessary to explain just why debates 33 about the Black Death in Iceland have been so heated: more so than for other regions, 34 there is little detail on the actual symptoms of the disease which is assumed to have ie 35 been the bubonic or pneumonic plague. Harsh words have been exchanged between 36 Benedictow and other scholars on the way a single account of the symptoms has been 37 w described by the single Icelandic annal entry which describes it. That entry, an 38 39 account in Lögmannsannáll for 1349, describes the symptoms of the disease that was 40 ravaging the rest of the world. The disease killed people within two or three days and was accompanied by ‘hörðum stinga’ (hard/sharp pains) and the vomiting of blood.8 On 41 42 The ambiguity here concerns the meaning of the Old Norse-Icelandic word stingi, 43 variously interpreted as being an internal pain such as might be associated with 44 pleurisy, or else the kind of glandular swellings (bubos) which are often associated 45 with bubonic plague. It is extremely hard to tell which is intended and is possibly not ly 46 relevant to the Icelandic situation of 1402 or later. Most likely, the phrase shows 47 awareness on the part of the author that victims suffered severe, focussed pain in 48 49 particular parts of their body. It seems unsafe to speculate as to the author’s meaning 50 6 51 J.F.D. Shrewsbury, A History of Bubonic Plague in the British Isles (Cambridge: Cambridge 52 University Press, 1971) suggested something similar for the fourteenth-century epidemics in the British 53 Isles. 7 S. K. Cohn, ‘The Black Death: end of a paradigm’, American Historical Review 107 (2002), 703–38 54 and C.J. Duncan and S. Scott, ‘What caused the Black Death?’, Postgraduate Medical Journal 81 55 (2005), 315–20 for scepticism; G. Twigg, The Black Death: A Biological Reappraisal (London: 56 Batsford, 1984) suggested anthrax; Shrewsbury, A History of Bubonic Plague in the British isles, 142 57 suggested that typhus could account for some episodes routinely attributed to Y. pestis. 58 8 Gunnar Karlsson, ‘Plague without rats’, 281-2; Islandske Annaler, 275-6. 59 60 3 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 4 of 32 1 2 3 beyond this rather than accept Gunnar Karlsson’s suggested analogy with the Modern 4 Icelandic usage of the word noun stingur for a sharp chest pain. 5 6 Some accounts of the 1494-5 outbreak of what is called ‘the great disease’ (mikla sótt) 7 say that it might have spread from some blue cloth or clothes (hafi komið upp úr bláu 8 9 klæði). This speculation comes only for this outbreak in what is generally a rather 10 tentative account and is followed by a sentence saying that the disease had a physical 11 form, resembling a bird and, in one account, giving off smoke.9 Whether the story 12 only reflects folk beliefs about disease in general, or derives from real knowledge of 13 the source of the disease is unclear but it may be significant that there is a connection 14 with shipments of clothing as we shall see. Fo 15 16 With this rather limited sense of what the two plagues’ symptoms and causes were, 17 the scale of deaths and the more extended narration of the diseases’ course make them 18 distinct from other recorded epidemics, as discussed below.10 rP 19 20 21 The course of the epidemics 22 The picture we get of the arrival and spread of the 1402-4 plague is vague at times but ee 23 24 generally coherent, as Gunnar Karlsson made clear. However, the annal accounts of 25 the arrival of the disease have a sparseness and simplicity to them reminiscent of 26 myths. Some historians have chosen to read them as providing a straightforward sense rR 27 of the spread of the disease from the south-west of the country out to other regions but 28 29 really each annal has its own interests. What Gunnar Karlsson identifies as ‘the 30 fragment’ (Vestfjarðarannáll) has an interest in the geography of the disease’s spread, while the other account, New Annal, has almost no detail on the movement of the ev 31 32 disease. The former identifies Hvalfjörður in the south-west as the point of entry of 33 the disease but the latter provides no precise origin.11 The former also records that the 34 cleric Áli Svarthöfði and his servants died en route back to the south from ie 35 Hvalfjörður, while the latter merely says that he was the first cleric to die in the 36 autumn of 1402.12 New Annal is also short on detail for the north of Iceland and far 37 w more concerned with the southern of Iceland’s two sees, Skálholt, hence its recording 38 39 of the deaths of priests at Skálholt and then of clergy in the south, west and east of the 40 country but not in the north. It is also noteworthy that it says nothing about widespread depopulation or its consequences, such as a reduction of rents that On 41 42 landowners might have had to suffer as their tenants died or disappeared.13 43 44 The evidence for the north of Iceland is in a different form and unambiguously 45 contemporary. For the north-east of Iceland two documents record vows of atonement ly 46 which were taken at two of the most significant religious centres, Grenjaðarstaður in 47 Þingeyjarsýsla and Munkaþverá in Eyjafjörður, written in December 1402 and 48 49 January 1403 respectively. The former document, from the more north-easterly 50 location, records that the plague had travelled eastwards (for vestan epter landinv) and 51 affected the south of Iceland, and the northern regions of Húnavatnssýsla and 52 53 9 Benedictow, ‘Plague’, 212; Annaler 1400-1800, vol. 1, 27-8; Safn til sögu Íslands, vol. 1 54 (Copenhagen, 1856), 43-4. 55 10 See below, p.10. 56 11 Jón Ólafur Ísberg, ‘Sóttir’, 194, accepts this for all his scepticism about the value of the annals. 57 12 Cf. Gunnar Karlsson, ‘Plague without rats’, 265–67. 58 13 Jón Ólafur Ísberg, ‘Sóttir’, 187. 59 60 4 URL: http://mc.manuscriptcentral.com/rmed Page 5 of 32 Journal of Medieval History 1 2 3 Skagafjörður, before it arrived in the north-east. Thus it appears not to have affected 4 either the large, relatively densely populated valley of Eyjafjörður or the sparsely- 5 populated north-eastern corner of the country before this point. The disease seems to 6 have persisted for over a year, only being remembered as disappearing after Easter 7 1404.14 As will become clear, it is frustrating that annals do not give a clearer picture 8 9 of the first epidemic’s spread in the north when other evidence exists for possible 10 mortality rates there. 11 12 For the 1494-95 outbreak, despite its being more recent in time, there are almost no 13 contemporary or near-contemporary sources for its perceived cause or course. Almost 14 all that can be safely said about this outbreak is that it did happen in the 1490s. Fo 15 Authors of annals written in the seventeenth century give various dates for its arrival 16 ranging from 1492 to 1495 and all give the impression that it lasted for no more than a 17 year but at least one will fixes it to no later than 1495.15 It is said to have reached 18 Iceland via a port in the south-west, much as for the first outbreak. Four different rP 19 20 points of entry for the disease are given, however, all of them equally credible 21 (Vestmannaeyjar, Hafnarfjörður, Seltjarnarnes and Hvalfjörður), in much the same 22 way as medieval English writers had disagreed over the plague’s point of entry to England in 1348.16 This time a source says that one region of the country was spared, ee 23 24 the West Fjords peninsular, home to perhaps 12% of the population.17 25 26 rR 27 Mortality 28 29 30 Mortality rates are notoriously difficult to calculate on the basis of the kinds of anecdotes and random samples of general economic data which are typically used for ev 31 32 other regions affected by the Black Death. For Iceland it is much the same but it has 33 been tempting for some historians to place too much confidence on the written 34 sources’ veracity without considering their form and origins. For example, one might ie 35 choose to stress that mortality rates were low on the basis that several key figures in 36 the stories about the plague’s arrival survive, figures like Einarr Herjólfsson himself, 37 for instance, and other members of the elite.18 However, much more has been made of w 38 39 other statements or apparent evidence of high mortality rates. These latter anecdotes 40 were heavily shaped by oral transmission, a point that will be made for all of the stories commonly used to indicate mortality rates. The mixture of brief statements and On 41 42 then outline narratives about particular mortalities should not fill us with any great 43 confidence beyond accepting the general impression that mortality was remembered 44 as shockingly high. 45 ly 46 Good illustrations with which to begin are the stories of Jón Egilsson (1548-1636?), 47 compiler of the Bishops’ Annals. Jón Egilsson clearly had a genuine interest in the 48 49 past but he was keen to retell the few stories which came his way without doing much 50 in the way of his own research. Other scholars have noted this propensity in his 51 52 14 53 Gunnar Karlsson, ‘Plague without rats’, 266-7; DI, vol. 3, 680, 682. 15 Gunnar Karlsson, ‘Plague without rats’, 267; Gunnar Karlsson and Helgi Skúli Kjartansson, 54 ‘Plágurnar miklu’, 18–19. 55 16 The Black Death, ed. R. Horrox (Manchester, 1994), 10. 56 17 Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 28. Based on earlier and later data 57 which, respectively, suggest 9% and 15% of the population lived here. 58 18 Jón Ólafur Ísberg, ‘Sóttir’, 199. 59 60 5 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 6 of 32 1 2 3 account of the history of the Icelandic church in the 1520s, for example, again a time 4 before he was born and where he does not seem to have had access to anything other 5 than hearsay.19 6 7 The first of Jón’s plague stories Gunnar Karlsson cites as ‘a moving picture of the 8 9 mortality in the second plague in the central South, where he [Jón] was brought up 10 and later served as a priest’. Jón was writing his account in 1605 which needs to be 11 given in full to appreciate its anecdotal nature: 12 13 In this plague the mortality was so great, that no-one remembered or 14 had heard of anything like it, because so many farms were devastated, Fo 15 and on most farms only three or two survived, sometimes children, 16 usually two or mostly three, and some of them yearlings, and some 17 suckling their dead mothers. Of these I saw one, who was called 18 Tungufell’s-Manga, she was three winters here in Galtafell, she died rP 19 20 when I was thirty years old. Another I saw was called Halldóra, she was 21 the mother of síra Grímr, who lived at Hruni, she had been one winter 22 old in that plague. She died when I was 34. Where there had been nine children, two or three were left alive.20 ee 23 24 25 In this first story we are given what appears to have been the received wisdom about 26 what had happened but Jón does not even go so far as to claim that the elderly women rR 27 identified here were his informants. In the second story a man who was ‘fourteen 28 29 years in that plague’, so seemingly born in 1480, supposedly told Jón, born in 1548 30 and writing in 1605, that there were four farms, two on each side of the river Hvítá, where the plague ‘might not have reached’. The account uses the subjunctive mood ev 31 32 three times in setting out what Jón was told, indicating that even Jón seems uncertain 33 about what he was recording including even the location of his informant’s childhood 34 home. The geographical distribution, size and status of these farms also follow a ie 35 pattern which might have been real or else shaped by the memory of the informant or 36 interviewer.21 We have no idea when Jón spoke to his witness, who would have been 37 w 68 years old when Jón was born. It is difficult to imagine at what time this 38 39 conversation might have happened such that the informant had a clear enough 40 memory of what had happened such a long time before, Jón was old enough to appreciate what he was being told, and then that Jón could remember what he was On 41 42 told clearly to write it down in 1605. We might postulate that, at best, perhaps, Jón 43 had interviewed his informant in 1568, when they were 20 and 88 respectively, and 44 45 ly 46 19 e.g. M. Kalinke, The Book of Reykjahólar. The Last of the Great Medieval Legendaries (Toronto: 47 University of Toronto Press, 1996), 25–26. 20 48 Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 25–26; Safn til sögu Íslands, vol. 1 49 (Copenhagen, 1856), 43. 21 50 Árni Magnússon and Páll Vídalín, Jarðabók, vol. 2 (Copenhagen 1925–7), 258, 272, 326–27, 357. 51 Each farm is from a different district on either side of the river Hvítá. If later data is anything to go by, 52 each farm was slightly below the average size for this region and the three other than the former home 53 of the informant, were arguably on the real geographical map and the informant’s mental map of the region. Kaldárhöfði and Efri-Þórisstaðir (modern Þóroddsstaðir in Grímsnes) were most likely 54 independent farms in the possession of the see of Skálholt and both to the west of the river, Kaldárhöfði 55 being on the southern shore of the lake Þingvallavatn. Ás and Hamarsholt were both on the eastern side 56 of the river and Hamarsholt was the last settled farm as one headed towards the uninhabited interior. 57 Both Ás and Hamarsholt were dependent properties owned by a church (Tungufell and Hruni 58 respectively). 59 60 6 URL: http://mc.manuscriptcentral.com/rmed Page 7 of 32 Journal of Medieval History 1 2 3 Jón had written down what he was told immediately rather than at some later time by 4 which his memory of what he had been told had weakened. 5 6 These accounts certainly show how carefully Jón intended to record the stories of his 7 elderly informants about the plague of 1494-95 but they are far from convincing as 8 9 accurate accounts of mortality. Jón was in essence getting oral accounts from two 10 generations before his own, those who had been children during the plague but who 11 survived and were able to retell the stories to Jón in their old age. While we can 12 certainly conclude that mortality rates were remembered as being high, nothing much 13 else can be inferred. The first story almost certainly recounts what had become a 14 cliché in popular memory, namely that no more than ‘two or three’ people/children Fo 15 survived on each farm. The second story simply lacks any context: we do not know 16 what question prompted the informant to name four farms unaffected by the plague, 17 one of which was their own home. Gunnar Karlsson infers from the latter story that it 18 was ‘exceptional’ for farms to have escaped the 1494-95 plague entirely. He suggests rP 19 20 on the basis of this account, which relates to a district of about 120 farms, that ‘not 21 more than 10% of all farms in localities where the plague raged could have escaped 22 the epidemic.’ What seems more significant is that it was exceptional for a member of the clergy in the later sixteenth century to find anyone old enough to talk about the ee 23 24 plague and provide any kind of detail about it. 25 26 Most other annal entries giving numbers of deaths in the 1402-4 epidemic follow the rR 27 same kinds of patterns as those provided by Jón Egilsson, even if they are not quite so 28 29 obviously anecdotal. From New Annal the first one is perhaps the least certain but its 30 account that only the bishop and two lay servants remained at the see of Skálholt is cited as credible. However, in this case it is not even clear that the people who had left ev 31 32 had actually died or whether the occupants of the emptied or abandoned (aleyddi) 33 estate had simply left to avoid the plague.22 It does say that every priest remaining 34 there died but one might expect priests to stay longer than lay servants.23 ie 35 36 Other accounts for the 1402-4 outbreak in New Annal include stories where the 37 w numbers of living and dead appear to be of the kind which get retold in myths, 38 39 folktales or oral stories. For example, there are two competing stories for the same 40 year’s entry at the monastery of Þykkvibær. In one story, the abbot and six brothers died while another six survived. In the second story, only two brethren and one On 41 42 servant survive after Þykkvibær had been emptied of people a third time. It is not 43 clear which of these stories is correct but they do conform to patterns we see 44 elsewhere in these accounts. For the nunnery at Kirkjubær it is said that the abbess 45 and seven others died while six survived, making it look like a variation on the first ly 46 Þykkvibær story. Other statements in the sources that only six priests survived in the 47 diocese of Hólar and ‘barely 50’ in the southern see might also be the product of tales 48 49 passed down and exaggerated rather than on first-hand knowledge. If we assume that 50 the Icelandic church was operating efficiently, and had a full complement of priests, 51 then with only 56 survivors over 300 had died.24 There seems little reason to be 52 53 22 Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 25 54 23 Annálar 1400–1800, vol. 1 (Reykjavík, 1922-7), 10. 55 24 Cf Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 15–16, 23. An eighteenth- 56 century annal recalls a story for the later outbreak in which only a priest and a girl survive in a region 57 of eastern Iceland. They travel south to find seven people in one place and eleven people elsewhere. 58 These motifs are barely different to those in stories from earlier centuries. 59 60 7 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 8 of 32 1 2 3 certain how many priests typically operated at any one time or about the reality 4 behind these stories of priests’ deaths although Gunnar Karlsson suggests 430 in total. 5 The figure for the south has more of a ring of truth about it than yet another story of a 6 ‘lucky six or seven’, but we might perhaps expect numbers of active priests to decline 7 in an epidemic as (a) some were more regularly exposed to disease when blessing the 8 9 dying and (b) others chose to abandon their calling to avoid exposure to disease. 10 Either way, these are rather bland statements shaped heavily by oral tradition which 11 when retelling numbers is unlikely to have preserved meaningful ones on which to 12 base estimates of mortality. As Gunnar Karlsson has noted, a story of 795 people 13 being brought for burial in New Annal for 1403 is not credible.25 It would have been 14 hard to find 795 people anywhere near isolated Kirkjubær at any point in Iceland’s Fo 15 history so this figure looks like a complete exaggeration or scribal error. 16 17 Alongside the issue of numbers of dead, one might also question the claim that the 18 West Fjords escaped the 1494-95 epidemic; both Gottskálk’s Annal and the rP 19 20 seventeenth-century Skarðsárannáll record this. As we shall see, records for the West 21 Fjords for the 1440s suggest that there were almost no vacant farms in this region. If 22 we can make anything of these accounts of the distinctiveness of the West Fjords then it seems equally likely that the authors of annals, Gottskálk Jónsson (1524-90) and ee 23 24 Björn Jónsson (1574-1655) had confused the anecdotes they had been told about the 25 two outbreaks.26 This is a better reading of the evidence than to suggest that the West 26 Fjords was so disconnected from the rest of Iceland that it should have somehow rR 27 escaped the disease twice. 28 29 30 Farm abandonment ev 31 32 33 Beyond the annals, the other obvious texts that might relate to Iceland’s demography 34 in the fifteenth century are surveys of church properties and a handful of other ie 35 documents which record property transactions including one list of the properties of a 36 significant secular landowner. For both epidemics there are documents which relate 37 w largely to the north of Iceland and properties owned by the church. 38 39 40 Gunnar Karlsson dedicates considerable space to analysing these texts.27 His analysis of these documents seems sometimes to make a number of assumptions which our On 41 42 analysis questions: 43 1. that the Icelandic farming landscape was in excellent health in 1402 when the 44 first epidemic arrived with all known farms being occupied 45 2. that farms in Iceland were of relatively uniform quality, size and had the same ly 46 numbers of inhabitants 47 3. that all signs of under-utilisation recorded in later decades were due solely to 48 49 the 1402 and 1494-5 plague epidemics 50 51 If these assumptions are put to one side then a more nuanced picture of the farming 52 landscape appears in the fifteenth and early sixteenth century such has been attempted 53 54 55 25 Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 27. The figure might only be 675 56 but it still not credible. 57 26 See further, Jón Ólafur Ísberg, ‘Sóttir’, 188. 58 27 Ibid., ‘Plágurnar miklu’, 35–47; Gunnar Karlsson, ‘Plague without rats’, 271–73. 59 60 8 URL: http://mc.manuscriptcentral.com/rmed Page 9 of 32 Journal of Medieval History 1 2 3 for Norway recently.28 We will now look at the texts he considers as evidence for the 4 scale of the demographic impact in the same order as he does. 5 6 The first document is a record of a visitation of the bishop of Hólar, Iceland’s 7 northern diocese, to his churches, seemingly first in 1429 and then again in 1431-2.29 8 9 It is clearly a document(s) by which the bishop is trying to assess the wealth and 10 assets of his see, just as Icelandic bishops had done periodically before him. Gunnar 11 Karlsson makes the point that this kind of survey is inconsistent but the entries for 12 most churches within it are generally shorter and patchier than any similar survey 13 before it, rarely recording all the things one would otherwise expect for each church 14 (its properties, ownership of rights to driftage on the coast (reki), church contents, and Fo 15 the duties of priests).30 It only records the number of farms or number of empty farms 16 for some parishes. This is why Gunnar produces both a maximal level of deserted 17 farms of 35%, and minimal figure of 19%, dependent on whether or not one includes 18 those parishes where total numbers of farms are given but the number of vacant farms rP 19 20 is not. Certainly our preference would be to consider only those nine parishes where 21 vacant farms are actually recorded, i.e. those parishes which suggest that 35% of 22 farms appear to have been vacant in 1431-2. ee 23 24 The most important point is that this list contains only 17 churches, a fraction of all 25 the churches in the diocese. The previous comprehensive survey of churches in the 26 diocese in 1394 had recorded 81 churches. Almost all of those churches continued to rR 27 exist throughout the fifteenth century and beyond. Not all of them were necessarily 28 29 parish churches, but nonetheless only about half the diocese is covered by the 30 surviving parts of the 1429-32 survey. Indeed there are gaps in the manuscript where it might have been intended for further accounts to have been included. Thus the large ev 31 32 and relatively densely populated valley of Eyjafjörður is absent and so is the much 33 larger but sparsely populated region of Þingeyjarsýsla. The absence of these parts of 34 the survey fits with the idea that there was no longer a need to monitor every church’s ie 35 income because most, if not all, farms were occupied and paying tithe just as they had 36 been doing since 1394 and before. It could be that there were significant problems in 37 w certain parishes where church attendance and tithe income must have seemed low 38 39 but,on the other hand, the church may have recognised that for most parishes, the 40 many tens of them surveyed for which no empty farms were recorded, there was either no real need to carry out further surveys or else no need to transcribe the results On 41 42 of those surveys in 1431-2. It is impossible to say which of the latter possibilities was 43 the case but either situation is probably indicative of contemporaries’ perceptions that 44 life was going on as usual in most parishes. 45 ly 46 If we turn now to the detail of the surviving Hólar diocese, then we can see that there 47 are certain districts with high numbers of vacant farms and that there are some small 48 49 clusters of adjacent or near adjacent parishes. 50 51 Table 1 52 53 28 A. Dybdahl, ‘Climate and demographic crises in Norway in medieval and early modern times’, The 54 Holocene 22 (2012), 1159–67. 55 29 DI, vol. 4, 464–68; 510–14. Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 35–8. 56 30 Even the drawn-out and incomplete survey of Hólar’s churches carried out by Bishop Jón skalli in 57 the latter half of the fourteenth century gave a full account of each church according to conventional 58 practice. DI, vol. 3, 155–79. Thank you to Erika Sigurdson for pointing this out. 59 60 9 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 10 of 32 1 2 3 Deserted farms recorded in the visitations of the bishop of Hólar, 1431-231 4 5 Parish Total Occupied Vacant 6 no. 7 8 Farms 9 Urðir, Svarfaðardalur 17 7 10 10 Ufsir, Svarfaðardalur 7 4 3 11 Árskógur, Árskógarströnd 16 11 5 12 Þóroddsstaður, Reykjadalur 14 11 3 13 Helgastaðir, Reykjadalur 7 4 3 14 Auðkúla, Húnavatnssýsla 9 7 2 Fo 15 16 Breiðabólstaður, Vatnsdalur 14 8 6 17 Másstaðir, Vatnsdalur 5 4 1 18 Spákonufell, Skagaströnd 8 7 1 rP 19 Totals 97 63 34 20 21 22 The first three parishes listed here were close to one another, on the western, outer ee 23 part of outer Eyjafjörður while the list features two adjacent parishes in Reykjadalur 24 25 and two at the mouth of Vatnsdalur. The remaining two are outliers. What best 26 explains this high level of vacant farms and their particular distribution on the map? rR 27 Certainly there is no particular sign that these areas were especially prone to periodic 28 farm abandonment in the way that has often been identified for some parts of Iceland 29 which are typically further inland and at relatively high altitudes. Equally, however, 30 none of these districts stand out as being particularly wealthy or densely-populated. ev 31 The worst affected parish here, Urðir in upper Svarfaðardalur, was an area where in 32 later centuries smaller, dependent farms were periodically abandoned. There were 13 33 34 such unoccupied farms among 83 farms in the survey of the valley undertaken in 1712 (15% of households; for the survey see below), which, is higher than average but not ie 35 36 by much – 39 other farms in the sýsla of Eyjafjörður were abandoned in 1712 (11% of 37 all remaining farms) – but this is nowhere near what we see in the table. Considering w 38 this valley in isolation, then, it could be argued that fluctuations in Iceland’s climate 39 might have had an impact on farms’ viability and led to such vacancies. More likely, 40 however, what we are witnessing in Svarfaðardalur and nearby is the result of a On 41 combination of various influences on demography and decision-making by landlords 42 and tentants. It is hard to quantify the relative quality or merits of the Icelandic 43 44 landscape, but it is possible that Svarfaðardalur was more prone to settlement change 45 than many of the other areas listed.32 The 1402-03 epidemic could still have had ly 46 residual influence on the vacany of farms in these districts in the 1440s but it would 47 hardly have been the only one. 48 49 The lists of properties in the 1440s actually name the individual farms owned by the 50 monasteries and one secular owner who owned them. For these farms it has been 51 argued that they were either vacant, recently reoccupied or else vacant but utilised by 52 53 54 31 Total no. of farms includes the farm on which the church was located. Locations of parishes 55 32 One might also argue that the record of local historical traditions, Svarfdæla saga, reflects this: it is a 56 text far less well preserved and far less coherent than sagas recording other districts’ histories. See, for 57 example, R. Boyer, ‘Svarfdœla saga’ in Medieval Scandinavia. An Encyclopedia, ed. P. Pulsiano and 58 K. Wolf (New York: Garland Publishing Inc., 1993), 626–7. 59 60 10 URL: http://mc.manuscriptcentral.com/rmed Page 11 of 32 Journal of Medieval History 1 2 3 other farms. This allows us to consider the wealth of those individual farms. The best 4 and arguably only way to do that is to use a tax register which was compiled in 1702- 5 14. This document, the Jarðabók compiled by the Danish royal officials Árni 6 Magnússon and Páll Vídalín, gives a snapshot of Iceland. It is commonly used by 7 historians of Iceland to consider the economic history of the country for the period 8 9 before 1702-14. It is generally considered valid to do so because all the indications 10 are that settlement patterns, farming practices and technology remained largely 11 unchanged from the first time they were recorded in the Middle Ages right through to 12 the eve of the twentieth century. There was, for instance, a great deal of continuity in 13 the way farms’ values were measured – in an abstract unit called a hundred (hundrað, 14 abbreviated to ‘h’ below) – with the majority of farms having relatively stable values Fo 15 over the centuries. Rents were similarily expressed in a unit – the ‘ell‘- derived from a 16 measurment of length often applied to cloth. The vast majority of farms mentioned in 17 medieval written sources have survived into the modern era with most being 18 continuously occupied in those nine hundred years.33 rP 19 20 21 The Jarðabók of Árni and Páll followed several seventeenth-century royal surveys but 22 has a special significance because it records far more forms of data than the previous ones. It contains a vast array of subjective and numerical data about each farm, giving ee 23 24 an opinion on the quality of types of natural resources, as well as pasture, rents and 25 numbers of livestock. For the sake of simplicity the following analysis will use the 26 fuller 1702-14 figures for the two forms of rent which Icelandic farm tenants paid, the rR 27 land rent (landskuld), paid in cash, and what we can translate as cattle-hire 28 29 (leigukúgildi), a standard part of the rent for a tenant in Iceland, usually paid in kind, 30 which involved actually renting adult cows from the landowner. Changes in rents are often discussed in the 1702-14 register as well so that some entries comment on ev 31 32 whether a rent had gone up or down within living memory or, for later entries, how 33 the smallpox outbreak of 1707 had affected rents. 34 ie 35 Last, but not least, it is important to note that Jarðabók, no less than many other 36 sources, demonstrates the variability of the size and value of Icelandic farms. The 37 w largest farms were accorded a value of well over 100 h(undreds), and typically the 38 39 largest farm in a parish would be the farm with the parish church with a value of 60h 40 or more. The smallest farms were typically 6h or 10h. An average farm might have been rated at 20h or 24h and had a rent of 120 ells. We need to bear in mind this On 41 42 varied tenurial and economic landscape when we try to explain the impact of 43 epidemics in Iceland. This was not really possible for the possessions of the see of 44 Hólar but it is for the remaining survey documents. As we shall see, it was often the 45 case that farms that were vacant in the fifteenth or sixteenth century often had low ly 46 values in Jarðabók. There is a risk of circularity of argument here, that a low value for 47 a farm in Jarðabók was the result of earlier desertion, but farm values were generally 48 49 stable as Gunnar Karlsson has noted.34 50 51 33 52 Orri Vésteinsson, ‘A Divided Society: Peasants and Aristocracy in Medieval Iceland.’ Viking and 53 Medieval Scandinavia 3 (2007), 117–39 highlights the selectivity of some of those medieval sources, however. 54 34 In addition it should be noted that relative values between some regions and districts were stable 55 even immediately after the first outbreak. Two exchanges of land in Svarfaðardalur in 1409 and 1414 56 respectively show farms in that valley being exchanged for comparable farms elsewhere in the north. In 57 the first Tungufell, valued at 30h in 1712, is exchanged for a farm of the same value elsewhere in the 58 north (Hvoll in Húnavatnssýsla, 30h in 1705). In the 1414 transaction two farms totalling 76h (Sakka 59 60 11 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 12 of 32 1 2 3 4 The second document then, dates to 1446 and is a list of the confiscated property of 5 the outlawed landowner Guðmundur Arason, including his 180 farms, all in the West 6 Fjords. Of these 180, a mere three were definitely vacant, all in one small district 7 (Furufjörður, Reykjarfjörður and Skjaldarbjarnarvík) while two others in another 8 9 district are listed as if they are not occupied but are being utilised by another farm. 10 Four of these five properties were small, with only one being rated at over 12h in 11 1710 which, ironically, was the only one of the five which had been abandoned in the 12 smallpox epidemic.35 It could have been that Guðmundur was a very shrewd dealer on 13 the property market and did not own vacant farms but it seems more likely that there 14 were very few vacant farms in the West Fjords at all. In other words, any epidemic Fo 15 that had affected the West Fjords was undetectable by the 1440s. As suggested above 16 it might even have been the case that the first epidemic had had a limited impact on 17 the region. 18 rP 19 20 The third document is a survey of the monastery at Munkaþverá in Eyjafjörður dated 21 to 1446, mostly covering areas not covered by 1431-32 survey of Hólar’s property. 22 Gunnar Karlsson highlight the fact that Munkaþverá had 31 properties in its possession in 1446 of which six farms paid no rent and a further seven paid ee 23 24 ‘abnormally’ low rent (listed for all farms only in ells, without reference to the second 25 form of rent, cattle hire). His conclusion is that the document shows that all 13 of 26 these farms were vacant in 1446, supposing that those paying low rents were being rR 27 rented out as pasture to other farms. All in all this seems to be an unnecessarily 28 29 gloomy picture and not one which takes into account local geography, Jarðabók data 30 or a realistic assessment of what constituted rent on these farms. ev 31 32 Table 2 33 Farms with low rents in the list of property of the monastery of Munkaþverá in 34 1446.36 ie 35 36 Farm (location) Rent in ells Rent recorded Dýrleiki 37 w 38 1446 in Jarðabók (hundreds) 39 * denotes vacant in 1712 (ells) in Jarðabók 40 Belgsá, Fnjóskadalur 60 60 ells 12 On 41 *Helluvað, Mývatnssveit 60 60/>120 12 42 Brekka, Presthólahreppur 60 60/80/90/120 8 43 44 (60h) and Skáldalækur (16h)) were exchanged for three farms in Öxarfjörður worth a combined value 45 ly of 80 hundreds (Klifshagi (40h), Hafrafellstunga (20h) and Þverá (20h)). 46 35 DI, vol. 4, 687. Jarðabók, vol. 7, 83–84, 312–14, 321–22 records that in 1710 Furufjörður (24 47 hundreds, 120/160 ells) had been abandoned since the smallpox outbreak. Reykjarfjörður (12h, 70/60 48 ells) and Skjaldarbjarnarvík (6h; 90 ells) were occupied. Birnustaðir (6h; 20 ells) and Fjallaskagi (12h; 49 120 ells) were occupied, paying their rent in fish. Cf. Gunnar Karlsson and Helgi Skúli Kjartansson, 50 ‘Plágurnar miklu’, 43. 36 51 Jarðabók notes that Helluvað, Raufarhöfn and Skálpagerði had been abandoned since the smallpox 52 epidemic of 1707, Jarðabók, vol., 11. Helluvað’s rent had been 120 ells plus a small payment in 53 labour/goods. Brekka’s rent was 80 ells but had traditionally been 60 but also sometimes 120 or 90. Ormarslón’s rent had been reduced over the course of twenty years from 120 ells down to 60 and then, 54 after the smallpox outbreak, from 60 to 40. Jódísarstaðir and Höskuldsstaðir are identified here as the 55 farms of close to Munkaþverá itself as opposed to farms with those names in Reykjadalur. Of these, 56 Höskuldsstaðir was unusual in being nothing more than ruin within the boundaries of Jódísarstaðir in 57 1712. Þórðarstaðir’s rent was 90 ells but had been 110 twenty years before. 58 59 60 12 URL: http://mc.manuscriptcentral.com/rmed Page 13 of 32 Journal of Medieval History 1 2 3 Blikalón, Presthólahreppur 60 90/120? 15? 4 Rif, Presthólahreppur 60 60 5 5 *Raufarhöfn, Presthólahr. 60 60 5 6 7 Ormarslón, Þistilfjörður 90 40/60/80/120 8 8 Jódísarstaðir, Eyjafjörður 0 120/180 30 9 *Höskuldsstaðir, Eyjafjörður 0 0 0 10 *Skálpagerði, Eyjafjörður 0 60 10 11 Fífilsgerði, Eyjafjörður 0 60 10 12 Þórðarstaðir, Fnjóskadalur 0 90/110 20 13 Hof, Flateyjardalur 0 60 8 14 Fo 15 16 The farms belonging to Munkaþverá were actually a varied bunch. About half of the 17 31 properties lay in Eyjafjörður itself while the others were in districts to its east, in 18 the small valley of Fnjóskadalur and then dispersed across the north-east of Iceland at rP 19 locations where specific natural resources were to be found. One farm was in 20 Mývatnssveit, for example, where the salmon fishing was good (Helluvað); other 21 farms were in Presthólarhreppur on the coast and provided fishing, seal-hunting and 22 other resources; Fnjóskadalur itself was a more wooded valley than Eyjafjörður and ee 23 was also likely to have been home to charcoal production. What is important to note is 24 25 that it is mostly non-Eyjafjörður farms with specialist economic functions, and with 26 low values, which have the unusually low rents in 1446 (see Table 2). These farms rR 27 were not only small but also witnessed a lot of variability in their rents in the decades 28 up to Jarðabók’s compilation for this region (1712), even by the standards of its post- 29 1707 smallpox surveys. This would suggest that these farms were generally a little bit 30 unstable, more prone to having good years and bad years than other farms but it is ev 31 hardly clear that most of them were significantly worse off than in 1712. It is possible 32 that they had been vacant at some point between 1402 and 1446 but not at all certain 33 that this had been the case. 34 ie 35 36 For obvious reasons a stronger case can be made for the six farms paying no cash rent 37 in 1446 actually being empty at this point. On the other hand there is an equally valid w 38 case for most of these farms being occupied. It is notable again that these farms were 39 almost all small while almost all lay within a few kilometres of Munkaþverá in 40 Eyjafjörður itself. They could certainly still have been paying cattle hire payments to On 41 Munkaþverá if not landskuld and, given their proximity to the monastery and the 42 absence of any earlier statement on the nature of their tenancies, it is not out of the 43 question that they had always had unusual rent arrangements. 44 45 ly 46 The fourth document is the 1446 list of rents of the monastery at Reynistaður, Gunnar 47 Karlsson takes a maximalist approach to the figures, assuming 14 of the monastery’s 48 52 farms were abandoned. While there are unequivocally seven deserted farms among 49 that 52 – the last seven listed in Table 3 – the others might have been vacated at some 50 stage but were occupied in 1446. For these farms we have a ‘kúgildi’ value attached 51 to them that appears to be the farm’s cattle hire rent rather than value. The cattle hire 52 figure for the seven farms with (very) low cash rents are mostly consistent with those 53 found in Jarðabók. This would suggest that the farms were still reasonably productive 54 55 although, again, most farms in this table tend to be average-sized at best. In some 56 cases it would seem that a moderately-sized farm was getting some kind of rebate but 57 that is less clear for other farms. The two farms which Karlsson identifies as ‘let for 58 three years at a rapidly increasing rent from year to year’ probably had only recently 59 60 13 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 14 of 32 1 2 3 been reoccupied and where new tenants had been given a low introductory rent as an 4 incentive to move there. Four of the Reynistaður properties which Karlsson identifies 5 as abandoned or recently reoccupied had ceased to exist by c.1700 and two of those 6 are unidentifiable. These would seem to be cases where the farms had ceased to exist 7 beyond the memory of those alive when Jarðabók was compiled. It was not 8 9 impossible that here, as elsewhere, fifteenth-century farms were being let to very 10 small households because they were barely viable.37 11 12 Table 3 13 Farms with low rents in the list of property of the monastery of Reynistaður in 14 1446.38 Fo 15 16 Farm (location) Rent in Rent in Cattle Cattle hire Dýrleiki (h) 17 * vacant in 1709 or 1713 1446 Jarðabók hire in in Jarðabók in Jarðabók 18 (ells) (ells) 1446 rP 19 20 Hraun, Skefilsstaðahreppur 30 120 1 3 or 4 20 21 Malland, Skefilsstaðahr. 30 140 3 3 ? 22 *Selá, Skefilsstaðahr. 30 80 3 3 >7.5 *Brókarlækur, Skefilsstaðahr. 37.5 60 3 3 >10 ee 23 24 Hamar, Rípurhreppur 0? 60 or 120 4 3 or 4 20 25 Hryggir, Staðarhreppur 60 for 3 - 5 - - 26 yrs rR 27 Daufá, Lýtingsstaðahreppur 60 for 3 60 or 120 6 2, 3 or 4 20 28 yrs 29 Gnúpr? 20 - - - - Stapahóll? 20 - - - - 30 Instaland, Sauðárhreppur 30 30, 50 or 80 ? 3 12 ev 31 *Fossárteigur, Sauðárhr. 30 40 or 60 ? 2 10 32 Kárastaðir, Rípurhreppur 40 0, 60 or 140 ? 2, 3 or 5 24 33 *Lýsukot, Rípurhreppur 40 ? ? ? ? 34 *Ferjuhamar, Rípurhr. 40 ? ? ? ? ie 35 36 37 w 38 The general point to make about all these farms is that they were mostly small and 39 therefore atypical. Almost all of the other farms owned by Munkaþverá and 40 Reynistaður in 1446 were later recorded as having values of 20h or more.39 Indeed On 41 some of the remote or coastal properties may not have even been whole farms – 42 something suggested by the cartularies from 1525 discussed below – and are unlikely 43 ever to have supported many people as household sizes correlated roughly to a farm’s 44 45 ly 46 37 J. Hajnal, ‘The Icelandic Census of 1703 in perspective’, in Manntalið 1703 þrjú hundruð ára. 47 Tilefni afmælis, eds Eiríkur G. Guðmundsson and Ólöf Garðarsdóttir (Reykjavík: Hagstofa Íslands og 48 Þjóðskjálasafn Íslands), Table 2, p. 35, estimates that 22% of households in1703 as of 1–3 people. 38 49 Hamar appears in this list because no rent is listed for it. This may simply be an omission by the 50 scribe. Jarðabók again shows some of the local idiosyncracies of rental and ownership. Hryggir formed 51 an integrated part of the main estate of Reynistaður in 1713 and did not have its own rental value. 52 Gnúpr (nvp) and Stapahóll (stapol) are unidentifiable. Innstaland included an island, Instalandssker. 53 ‘Kallastader’ is identified as modern Kárastaðir. Kárastaðir had been abandoned for 12 years just before Jarðabók had been compiled. It is listed just before Ferjuhamar which is the name of a ruin at 54 Kárastaðir. ‘Lusabakke’ is most likely synonymous with the ‘Lijsekot’, a place-name remembered as 55 having been a farm at Helluland, next to Kárastaðir. Jarðabók, vol. 9, 60–1. 56 39 Jarðabók vol. 11. For Munkaþverá only the unusual Espihóll syðri (12h), probably originally an off- 57 shoot of its larger neighbour Espihóll nyðri, and Gásir (10h), essentially a seasonal port rather than a 58 farm, had values below 20h. 59 60 14 URL: http://mc.manuscriptcentral.com/rmed Page 15 of 32 Journal of Medieval History 1 2 3 value and its ability to feed its occupants. Such farms, or even parts of farms, were 4 almost certainly going to be the first farms to be vacated in a demographic crisis – or 5 at least exploited remotely rather than being occupied all year – and then to have been 6 last to be reoccupied permanently when the population expanded again. Thus a more 7 realistic reading of the Munkaþverá document would be that those 13 small, 8 9 ‘abandoned’ farms were occupied in 1446 but were neither populous nor successful. 10 Reynistaður’s properties do have lower rents than those of Munkaþverá, and seven are 11 clearly unoccupied. These low rents probably best reflect a general lowering of 12 property rents in response to a drop in the population. They represent a kind of 13 renegotiation between landlords and peasants as we see elsewhere in western Europe 14 in the fourteenth century. This is still a poor guide to actual population decreases, Fo 15 however. 16 17 The fifth document in Gunnar Karlsson’s list is the rental for the monastery of 18 Möðruvellir in Hörgárdalur, also within the wider valley system of Eyjafjörður, dated rP 19 20 to 1447. This one potentially provides the strongest case for a demographic crisis for 21 early fifteenth century Iceland. Thirty-nine of its 74 properties were uninhabited 22 leaving just 35 occupied. In this document there is absolute clarity that the occupied farms are paying both cash rent and cattle hire, many of them at levels higher than ee 23 24 those recorded in Jarðabók, and that these are average to large farms of 20h to 40h 25 with only one or two exceptions. In other words, this section of the monastery’s 26 property is in very good health. For the unoccupied farms, however, the picture is rR 27 generally as it is for the other monastic surveys – farms are generally of 20h or less 28 29 and some are hard to identify at all in other sources – although it is noteworthy that 30 two vacant farms were apparently larger ones, of 50h and 60h. Otherwise twelve of the 39 are accounted for by a single item, the whole of the remote island of Grímsey, ev 31 32 which was divided into twelve properties most of which were later valued at 10h-17h 33 each: these appear to have been small properties which depended heavily on fishing. 34 In 1713 the whole island’s livestock, for the now ten farms and one hjáleiga, ie 35 amounted to a meagre five cattle, 67 ewes and a few lambs.40 Even if Grímsey 36 contained twelve properties it seems unwise to regard them as normal or stable farms, 37 w for 1447 or at any time, however good the fishing might have been. Consideration 38 39 will be given to non-Black Death-related causes of abandonment below, but it is 40 worth noting here that in Grímsey’s case, that the island may have been harmed by a raid on the island in 1424 by the English (fishermen) active in Iceland at the time.41 On 41 42 43 Turning to the remaining 27 empty farms on the Möðruvellir list of 1447, 15 have a 44 monetary value listed for them which must be landskuld and presumably is a sign of 45 their resources being used in some way, but not of permanent occupation; three of ly 46 these are listed slightly differently and might actually have been occupied.42 The last 47 twelve were undoubtedly not productive. To sum up, this document does not 48 49 necessarily demonstrate devastation of the human landscape. If Grímsey were treated 50 51 52 40 53 Jarðabók vol. 10, 310-20. 41 Cf. Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 39. 54 42 The phrase ‘bykt firir’ with a value is also provided for three farms of the 15 vacant farms accorded a 55 landskuld (Vaglir, Hóll, Garðshorn) in the same way as it is used for both occupied and unoccupied 56 farms in the Reynistaður document. These farms were valued at 20h (Vaglir) and a combined 60h for 57 the other two which later formed part of the same farm. These look like they might have been 58 occupied. 59 60 15 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 16 of 32 1 2 3 as less significant, and those three other farms were occupied or used, then only about 4 half of Möðruvellir’s properties could be counted as empty. 5 6 For Möðruvellir in 1447 there was still a very high percentage of its farms which were 7 vacant but two features of these require particular emphasis. The first is that, as for 8 9 the other rentals above, the headline figure of the number of vacant farms probably 10 overstates what this represents in terms of population and overall productive capacity 11 of the land because on average they are smaller farms. The second point is important 12 and takes us back to the first document under discussion which showed a 13 concentration of empty farms in Svarfaðardalur in 1431-32. What we see in the 14 Möðruvellir document is that the greatest concentration of its vacant properties are Fo 15 also in Svarfaðardalur (at least seven of them), where it owns at most five occupied 16 farms in the valley. This contrasts with its holdings to the south of Svarfaðardalur 17 where it owned over 30 occupied farms (most valued at 30h or more) and only 10 or 18 so empty farms. In other words, the high levels of vacancies only occurred in rP 19 20 Svarfaðardalur. Our views of levels of occupancy are thus skewed by what could well 21 have been the high, anomalous figures for that valley.43 22 For the last document, a 1449 list of the properties of the diocese of Hólar, a ee 23 24 maximum of 23 of 187 properties were either empty or rented out for low rents.44 The 25 Hólar properties, both occupied and vacant, were distributed across northern Iceland 26 with a slight concentration of smaller, empty ones again in the north-east coastal rR 27 districts, just as with those owned by Munkaþverá.45 It is notable that Hólar’s large, 28 29 dispersed group of properties included no more than 12% of its number showing signs 30 of having been empty. ev 31 32 In sum it would seem that Gunnar Karlsson’s calculations for numbers of deserted 33 farms are too high. In terms of the population that the occupied farms might have 34 supported in the 1440s, it would seem most likely that this was higher too with the ie 35 obvious implication that the residual effects of plague might not have been sustained 36 to such a great extent forty years after the epidemic if, indeed, they remained 37 w detectable at all. 38 39 40 For the 1494-95 outbreak there is a similar range of surveys of church properties, this time all dated to 1525, and so still a significant time after the epidemic. These On 41 42 sometimes record individual vacant properties and sometimes do not and cover the 43 same landlords as before but with the addition of other church properties in the north 44 including the monasteries at Þingeyrar (to the west of Skagafjörður) and 45 Grenjaðarstaður (to the east of Eyjafjörður). The same general rules would seem to ly 46 apply for the 1525 documents as for those of the 1440s. Where there are no records of 47 individual farms it would seem safest to conclude that vacant farms were either rare or 48 49 within what was seen as being normal. Grenjaðarstaður in Reykjadalur is the one 50 51 43 52 Gunnar Karlsson’s statement that documents 1–6 cover 716 farms is not quite correct, ‘Plague 53 without rats’, 273. Some farms are counted twice because of Document 1’s overlap with many of the others. 54 44 23 is the number claimed by Gunnar Karlsson and Helgi Skúli Kjartansson, ‘Plágurnar miklu’, 39– 55 40, of which two might not be seen as independent. One of the farms listed here with a full rent is listed 56 as an empty farm for Möðruvellir in 1447. DI, vol. 5, 35–41. 57 45 Hólar owned few properties in Svarfaðardalur so it is not possible to compare occupancy rates for 58 Svarfaðardalur using texts 1, 5 and 6 as one might hope to. 59 60 16 URL: http://mc.manuscriptcentral.com/rmed Page 17 of 32 Journal of Medieval History 1 2 3 estate which requires no discussion because it had only 23 properties, the status of 4 which is not made clear in 1525, and for which we have no comparable data from the 5 1440s.46 6 7 The lists of properties for Munkaþverá (59 properties) and Möðruvellir (45 properties 8 9 plus part shares in the islands of Grímsey and Flatey) make no mention of vacancies. 10 Each monastery’s property portfolio had changed since the 1440s but with a large 11 core of their properties remaining the same. Both monasteries seem to have disposed 12 of some smaller properties or rights, including a mix of those occupied and empty 13 ones in the 1440s. Munkaþverá had also bought a significant number of new farms 14 and shares in farms while Möðruvellir had sold some of its smaller, unoccupied Fo 15 properties in Svarfaðardalur. Overall it would seem that these monasteries were more 16 likely to buy and sell small farms and the shares in farms of specific resources (ítök) 17 which had low rents. Monasteries’ buying and selling of these small farms would not 18 appear to be a clear indication of their permanent abandonment but more a sign of a rP 19 20 fairly active land market. 21 22 Reynistaður had had seven clearly vacant properties in 1446 among its 52 and also had seven properties out of 50 listed in 1525. Those earlier seven show limited ee 23 24 overlap with the later ones: three were now occupied, three had disappeared from 25 Reynistaður’s roster and only one remained empty. Three occupied farms that were 26 occupied in 1446 were now listed as empty and one of the two farms with a rR 27 temporarily reduced rent was now vacant. The farms which had been occupied in 28 29 1446 but were abandoned in 1525 were among the lower-valued among those 30 occupied ones. This all suggests that farms were being vacated and occupied fairly regularly. ev 31 32 33 For Hólar it appears the bishop owned whole districts in 1525 rather than individual 34 properties being listed with the result that none can be identified as empty. The ie 35 remainder of Hólar’s properties numbered 285 of which 21 were vacant. For the 36 districts where no vacancies were identified it would seem sensible to conclude again 37 w that there were no problems in the minds of those writing, something reinforced by 38 39 the fact that we have hardly any empty farms in those districts in the 1440s.47 The 7% 40 rate of vacancy for the diocese’s individual properties is therefore unlikely to be any higher for the whole of its possessions and, yet again, the empty ones are small to On 41 42 average in size. 43 44 Þingeyrar had 61 occupied properties and, listed separately, 35 vacant ones. The 45 latter, when they are identifiable, seem to have been mostly of 10h or less in Jarðabók ly 46 and lay in districts mostly to the west of Skagafjörður and where there were 47 concentrations of similar small properties. The occupied ones were almost all of 16h 48 49 or more. There seems to be no grounds on which to judge the occupational status of 50 the properties because no figures are given for rents. What is also noticeable, besides 51 the high rate of unoccupied properties generally, is the possibly continued 52 concentration of them around Breiðabólstaður in Vatnsdalur, close to Þingeyrar itself. 53 It is difficult to be certain of the boundaries of the 1440s parish of Breiðabólstaður, 54 55 46 Grenjaðarstaður was close to the parishes of Þóroddsstaður and Helgastaðir (see Table 1) but there is 56 still nothing meaningful to say about Reykjadalur in this context. 57 47 E.g. DI, vol. 5, 36. Hólar’s holdings described as ‘east of Héraðsvatn’ in Skagafjörður where there 58 were at least 10 properties providing full rents in 1449. 59 60 17 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 18 of 32 1 2 3 where six farms were empty, but if the parish extended westwards from 4 Breiðabólstaður then at least six vacant farms might have fallen within in it in 1525. 5 Regardless of whether the farms were quite the same ones in both lists, this does seem 6 to be a rare instance where seems to have been some kind of long-term but small- 7 scale abandonment of small properties. 8 9 10 In sum then there were relatively few empty farms in the north in 1525 in comparison 11 with the 1440s. The general pattern was still one where smaller properties were likely 12 to be vacant at any given time. By 1525 it looks as if the north generally seems to 13 have had a settlement pattern closer to what is generally perceived of as being normal, 14 whereby most farms were occupied. The fact that a large proportion of Þingeyrar’s Fo 15 properties were empty is not easy to explain based on the limited evidence. It may be 16 that there had been a disproportionately high local mortality in the 1490s or else 17 poorer management of the monastery’s properties since the epidemic(s), with fewer 18 potential tenants being enticed in relative to elsewhere. Perhaps Hólar and rP 19 20 Reynistaður had been able to encourage people to live on their farms. In this respect 21 the annals indicate that poorer people and their families migrated from the Vestfirðir 22 to the north to abandoned holdings in the north. 48 ee 23 24 ‘Plagues’ and other epidemics 25 26 Notwithstanding the view of the present authors that the 1402 and 1494-5 epidemics rR 27 were significant outbreaks of plague caused by Yersinia pestis, there were clearly 28 29 other causes of large-scale population decreases in medieval Iceland, including 30 outbreaks of disease. There can be little doubt that there were regular, small outbreaks of disease and famine which have not been recorded in the kinds of sketchy written ev 31 32 texts we have for the fifteenth and early sixteenth century. What information we do 33 have hints at occasional significant outbreaks of disease, although these get less 34 attention than they might in connection with the Black Death.49 If we look at annals ie 35 covering just the period c.1380 to c.1525, however, there were several recorded 36 epidemics, albeit rarely on the same scale.50 The entries for these other, hard-to- 37 w identify diseases – also given the rather unspecific name sótt (sickness) – are no less 38 39 credible than those for the epidemics of 1402 and 1494. These epidemics are 40 sometimes said to have affected the whole country and sometimes described as great or large (mikill) and as causing high mortality. They still differ from the ‘plagues’ of On 41 42 1402 and 1494, however, by not being accorded the countrywide impact and speed of 43 impact which strongly suggest that these two epidemics were widely recognised as 44 being somehow different from others.51 45 ly 46 47 48 48 Annálar 1400–1800, vol. 1, 74-5. This might have included not only people from around Þingeyrar 49 (Húnavatnssýsla) but possibly also those people from Vestfirðir whom Björn Jónsson claimed had 50 moved to the north after the 1494–95 epidemic. 49 51 Cf. Jón Ólafur Ísberg, ‘Sóttir’, 205, with reference to previous studies including the influential work 52 of Jón Steffensen, Menning og meinsemdir. Ritgerðasafn um mótunarsögu íslenzkrar þjóðar og baráttu 53 hennar við hungur og sóttir (Reykjavík, 1975). 50 Islandske Annaler indtil 1578. Udgivne ved G. Storm (Christiania, 1888), 213. 1347 actually seems 54 to have witnessed one of the largest deadly epidemics, a bólnasótt (small pox?) which lasted more than 55 a year and has a fuller write-up than many later epidemics. It is hard to compare the frequency of 56 epidemics over the long term in Iceland because the form of the textual evidence varies so much. 57 51 The Icelandic annals describe the outbreak of the Black Death in Norway as a deadly disease 58 (drepsótt). Islandske Annaler 213, 275, 354, 403. 59 60 18 URL: http://mc.manuscriptcentral.com/rmed Page 19 of 32 Journal of Medieval History 1 2 3 If we just tally up the instances of disease outbreaks then there are recorded outbreaks 4 in c.1380 with some annals using very similar wording to record a great epidemic in 5 Norway followed by one in Iceland. Disease in Norway is often identified before it 6 gets to Iceland.52 It might be, for example, that the entry for 1391 in one annal that an 7 Icelandic traveller dies three weeks after coming home from Norway that it is making 8 9 a connection between the ‘great epidemic’ (Sott micil) in Norway it has just 10 recorded.53 New Annal, probably written in the 1430s, records a pair of entries for 11 1420 and 1421 which are probably only credible as an account of a long term 12 outbreak of disease rather than anything more specific. It claims that in the first year 13 (in winter) everywhere was affected by a disease which killed people in their fifties, 14 and barely anyone older. The following year ‘many healthy people’ in their twenties Fo 15 died.54 For 1430 or 1431 the ‘great epidemic’ is recorded which seems to have 16 interrupted or caused the recording of estates by the bishop of Hólar.55 According to 17 Björn Jónsson’s Skarðsárannáll yet another ‘great plague’ caused ‘great mortality’ in 18 1472. For either 1479 or 1484 he records that a wealthy woman is identified as dying rP 19 20 as a result of sótt, presumably an indication of endemic disease or an allusion to 21 another outbreak. Finally another plague is recorded for 1511 or 1512.56 All these 22 events seem to have been recorded without the additional anecdotes attached to some accounts of the ‘plagues’ but any one of them could have had a significant role in ee 23 24 reducing population and emptying farms. In addition to these are a handful of 25 references to bad weather or bad farming conditions (as well as occasional references 26 to notably good weather) which are also unquantifiable in their impacts. These, like rR 27 the other epidemics, were not altogether unusual in Iceland but may well have had a 28 29 greater influence on settlement patterns than is allowed for.57 30 Climate variation and instability ev 31 32 33 The impact of climate change on the lives of farmers for the period c. 1400 to c. 1525 34 also needs to be taken into account. It has long been recognised that Europe and the ie 35 North Atlantic experienced a period of relatively warm temperatures period (the 36 Medieval Climate Anomaly) from c.900 to c.1250. This was followed by relatively 37 low temperatures, especially in the period c.1400 to c.1700.58 This must have w 38 39 decreased the viability of farms of marginal potential even if it is unclear how readily 40 people perceived such changes in the short term. Similarly recent research on past climates has highlighted significant changes in the predictability of the weather in the On 41 42 fifteenth and early sixteenth century; from about 1425 the weather in the North 43 Atlantic appears to have varied on a year-by-year basis in ways which people would 44 not have witnessed in preceeding decades.59 In the light of this it would seem that 45 ly 46 52 47 Islandske Annaler 281, 365, 413. 53 48 Islandske Annaler 367. 54 49 Annálar 1400–1800, vol. 1, 23. 55 50 Islandske Annaler, 370; Annálar 1400–1800, vol. 1, 53. 56 51 Annálar 1400–1800, vol. 1, 67, 71, 78. 57 52 Annálar 1400–1800, vol. 1, 24, 24–5, 83–84, 84; Islandske Annaler 295, 287, 416; Safn til sögu 53 Íslands, vol. 1, 44. Recorded for 1390, 1405, 1424, 1426, 1430, 1519, 1520. 58 M. Mann, Z. Zhang, S. Rutherford and others, ‘Global Signatures and Dynamical Origins of the 54 Little Ice Age and Medieval Climate Anomaly’, Science 326 (2009), 1256-60. 55 59 A. Dugmore, D. M. Borthwick, M. J. Church and others, ‘The Role of Climate in Settlement and 56 Landscape Change in the North Atlantic Islands: An Assessment of Cumulative. Deviations in High- 57 Resolution Proxy Climate Records’, Human Ecology 35 (2007), 169–78. Cf. J. Barrett, ‘What caused 58 the Viking Age?’, Antiquity 82 (2008), 673. 59 60 19 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 20 of 32 1 2 3 temporarily abandoned farms might not have been so readily risked by potential new 4 tenants who instead might have preferred to stay in established households. While 5 people may have been reluctant initially to leave their homes in crises, they may 6 equally have been resistant to going (back) to a poorer farm. This could explain some 7 part of the lower levels of farm occupancy we see. Furthermore recent research on 8 9 landscapes in southern Iceland does suggest continued management of the farming 10 landscape in the fifteenth and sixteenth centuries, lessening the impacts of climatic 11 deterioration. This implies a sufficiently large population capable of managing 12 pastures in much the same way as in previous centuries, regardless of where people 13 might have been living.60 14 Fo 15 16 Competing models of plague transmission in Iceland 17 18 Before we discuss how the plague reached Iceland and then how it was transmitted to rP 19 20 humans, it is first necessary to review the competing explanations already present in 21 the literature. Following the identification of Y. pestis as the aetiological agent 22 responsible for the third pandemic by Yersin in 1894, the role of insect vectors, such as the flea, in spreading the disease was first identified by Simond in 1898 and rapidly ee 23 24 became the most widely accepted epidemiological explanation of the phenomenon for 25 late medieval Europe just as for other times and places.61 26 rR 27 In simple terms, the rat-flea model requires that bacteria are spread from rat to rat by 28 29 flea bites. During the times of year when the conditions of warmth and humidity are at 30 an optimum, the fleas multiply and the disease becomes epizootic in the rodent population. The rapid spread of infection is thought to cause a ‘die-off’ of the rodent ev 31 32 population and, as the fleas leave the dead rats, they will then attack the less well- 33 favoured human hosts. Hence, the disease is incidental to the rodent-flea dynamic but 34 is no less unwelcome for being so. Xenopsylla cheopis was thought to be one of the ie 35 most efficient at transmitting plague because of its tendency to develop a ‘blocked’ 36 proventriculus, a valve that connects the oesophagus to the midgut, by forming a large 37 w mass of bacteria. When blocked fleas attempt to feed, blood containing large 38 39 quantities of bacteria from the blocking mass is refluxed into the bite site.62 40 It is important, however, not to be persuaded by an over-simplistic view of the On 41 42 biology of plague. There are numerous rodent species capable of sustaining enzootic 43 foci in various parts of the world.63 Black Rats are of particular interest simply 44 because they are often commensal, living in close contact with humans, and have a 45 habit of ‘hitch-hiking’ on human transport (hence the widely used names ‘roof-rat’ ly 46 and ‘ship-rat’). 47 48 60 49 R. Streeter, A. J. Dugmore and Orri Vésteinsson, ‘Plague and landscape resilience in premodern 50 Iceland’, Proceedings of the National Academy of Sciences of the United States of America 109 (2012) 51 3664–9. 61 52 A. Yersin, ‘La peste bubonique a Hong-Kong’, Annales de l’Institut Pasteur, 8 (1894), 662–7; P. L. 53 Simon, ‘La propagation de la peste’, Annales de l’Institut Pasteur 12 (1898), 625–87; G. Lamb, The Etiology and Epidemiology of Plague: A Summary of the Work of the Plague Commission (Calcutta: 54 Suprintendent of Government Printing, 1908). 55 62 A.W. Bacot and C. Martin, ‘Observations on the mechanism of the transmission of plague by fleas’, 56 Journal of Hygiene, Plague Supplement 3, 13 (1914), 423–39. 57 63 See, for example, R. Pollitzer, ‘Plague Studies: 6. Hosts of the infections’, Bulletin of the World 58 Health Organisation 6 (1952), 381-465 for thorough review. 59 60 20 URL: http://mc.manuscriptcentral.com/rmed Page 21 of 32 Journal of Medieval History 1 2 3 4 Although X. cheopis is perceived by some to be the most efficient flea vector, recent 5 studies suggest that this imay not actually be the case. This is partly because of a long 6 extrinsic incubation period before blockage formation, and subsequent transmission, 7 together with a high death rate amongst blocked fleas. Moreover, even though X. 8 9 cheopis is the most commonly implicated flea species, there are numerous other 10 species capable of driving and maintaining an epizootic. For instance, Nososphyllus 11 fasciatus, more often found in colder climates, can also develop a blocked 12 proventriculus and is also commonly found on rats and other rodents.64 13 14 Other disease vectors are possible. Spread between humans via the ‘human flea’ Fo 15 Pulex irritans and other ectoparasites such as the human body louse Pediculus 16 humanus cannot be ruled out entirely although these modes of transfer are not widely 17 accepted as being significant in the epidemiology of modern plague.65 However, 18 plentiful specimens of P. humanus and other species of louse have been found in rP 19 20 excavations of medieval farmhouses in Iceland.66 21 22 A major difficulty with the rat-flea model, in this context, is that there are no written records of rats being present in Iceland in the Middle Ages and there are no records ee 23 24 (as yet) of rat bones being found in archaeological deposits dating to the medieval 25 period. Moreover, it has been argued that the Black Rat could not have survived in 26 Iceland because it prefers warmer climates and, in any case, the environmental rR 27 conditions required by the ‘rat-flea’ X. cheopis to breed in sufficient numbers were 28 29 also absent. Karlsson is keen to stress that the annals report that the epidemics 30 continued throughout the winter when, in his view, fleas could not have been involved.67 ev 31 32 33 Based on what we know at present it is difficult to see how we can reconstruct the 34 fifteenth century ecology of commensal and sylvatic Icelandic rodents, together with ie 35 their fleas. As Gunnar Karlsson has rightly emphasised medieval written texts 36 produced in Iceland show a strong familiarity with mice but suggest that rats were 37 either unknown or exotic.68 Archaeological excavations up to 1996 had not yielded w 38 39 evidence of rat skeletons of medieval date in Iceland and, indeed, still have yet to be 40 found. Mice bones and evidence of rodent gnawing on other animal bones have been found on sites as early as the tenth-century (Hofstaðir in Mývatnssveit) and at other On 41 42 43 44 64 R. J. Eissen and others,,‘Early-phase transmission of Yersinia pestis by unblocked fleas as a 45 ly mechanism explaining rapidly spreading plague epizootics’, Proceedings of the National Academy of 46 Sciences of the United States of America 103 (2006), 15380–5; L. K. Little, ‘Plague historians in lab 47 coats’, Past and Present 213 (2011), 286 comments, ‘plague historians can henceforth safely retire 48 their exclusive fixation upon Rattus rattus and Xenopsylla cheopis.’; A. L. Burroughs, ‘Sylvatic plague 49 studies. The vector efficiency of nine species of fleas compared with Xenopsylla cheopis’, Journal of 50 Hygiene 45 (1947), 371–96. 65 51 M. Drancourt, L. Houhamdi and D. Raoult, ‘Yersinia pestis as a telluric, human ectoparasite-borne 52 organism’, The Lancet Infectious Diseases 6 (2006), 234–41; L. Houhamdi, H. Lepidi, M. Drancourt 53 and D. Raoult, ‘Experimental model to evaluate the human body louse as a vector of plague’, Journal of Infectious Diseases 194 (2006), 1589–96; Pollitzer, ‘Plague Studies’, 392. 54 66 Guðrún Sveinbjarnardottir, Egill Erlendsson, K. Vickers and others, ‘The palaeoecology of a high 55 status Icelandic farm’, Environmental Archaeology 12 (2007), 189. 56 67 Gunnar Karlsson, ‘Plague without rats’, 263, 267, 277-8. 57 68 Gunnar Karlsson, ‘Plague without rats’, 279–80. Jón Ólafur Ísberg, ‘Sóttir’, 195, could not get round 58 this problem either. 59 60 21 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 22 of 32 1 2 3 sites for before c.1700.69 Other excavations in recent decades have shown much later 4 evidence of the presence of rats in Iceland, through the evidence of gnawed animal 5 bones in both urban contexts (in Reykjavík at Tjarnargata 3c) and rural ones 6 (Bessastaðir near Reykjavík, Skútustaðir in Mývatnssveit, Möðruvellir in 7 Hörgárdalur).70 At present it may not be possible to distinguish between the evidence 8 9 of gnawing of rats and mice but the absence of rat bones might prove to be telling. 10 Equally it is still the case that relatively little attention has been paid to late medieval 11 archaeology in Iceland and that much Icelandic settlement archaeology was not done 12 sufficiently well to have necessarily found or recorded animal bones rigorously. That 13 said, one would have expected any large late medieval rat population to have left 14 skeletal traces, although it is possible that rats may have died in places where Fo 15 archaeologists are less likely to find them, i.e. within the confines of a dwelling or 16 outbuilding. This, however, does not undermine the argument of this article for 17 temporary appearances of rats in Iceland and for a more complex view of the plague’s 18 spread. rP 19 20 21 Pneumonic cases often occur in bubonic epidemics that are associated with epizootics 22 in commensal rodents. The infection is initially transferred from rodent to man by flea bites. If the infection spreads to the lungs the patient is said to have developed ee 23 24 secondary pneumonic plague. Plague may then spread via droplet infection to others 25 who are then said to have contracted primary pneumonic plague. 26 rR 27 The pneumonic model, though an attractive explanation of plague when rats and their 28 29 fleas are not obviously responsible, also has its problems since the aerosolisation of Y. 30 pestis appears to be an inefficient means of spreading plague. For instance, it was found that the distance that coughing patients projected bacteria in the plague wards ev 31 32 of Manchuria was limited to 1 metre and there is ample evidence that potential 33 contacts or even those who co-habit with sufferers frequently do not catch the 34 disease.71 Another problem is that the time from the onset of symptoms to death for ie 35 untreated patients is short (two to three days) and they are thought to be infectious 36 only when capable of projecting infected sputum whilst low contact frequency may 37 also be expected because patients rapidly become incapacitated.72 w 38 39 40 Nevertheless, it must be remembered that the most notorious modern pneumonic outbreaks which occurred in Manchuria and adjacent provinces of China in 1910/11, On 41 42 1917/18 and 1920/21 were responsible for 60,000, 16,000 and 9,300 deaths 43 44 45 69 ly T. H. McGovern, ‘The Archaeofauna’ in Hofstaðir. Excavations of a Viking Age Feasting Hall in 46 North-Eastern Iceland, ed. G. Lucas (Reykjavík: Fornleifastofnun Íslands, 2009), 172, 180, 221; M. T. 47 Hicks and R. Harrison, A Preliminary Report of the 2008 Midden Excavation at Skútustaðir, N. 48 Iceland. Unpublished NORSEC Report No. 45 (CUNY Northern Science & Education Center, 2008). 70 49 S. Perdikaris, C. Amundsen and T. H. McGovern, Report of Animal Bones from Tjarnargata 3C, 50 Reykjavík, Iceland. HERC/NORSEC Zooarchaeology Laboratory Report No. 1 (2002), 13; R. Harrison, 51 Möðruvellir in Hörgárdalur, N. Iceland: General Overview of the Archaeofauna Analyzed from the 52 2006-08 Midden Mound Excavations. HERC/NORSEC Zooarchaeology Laboratory Report No.59 53 (2011), 13, 35. 71 R.P. Strong and O. Teague, ‘Studies on pneumonic plague and plague immunization’, Philippine 54 Journal of Science 8 (1912), 129–268; H. Wang, Y. Cui, Z. Wang and others, ‘A Dog-Associated 55 Primary Pneumonic Plague in Qinghai Province, China’, Clinical Infectious Diseases 52 (2011), 185– 56 90; G. W. Gale, ‘An outbreak of pneumonic plague in the Kalahari’, South African Medical Journal 15 57 (1941), 369–73. 58 72 J.L. Kool, ‘Risk of person-to-person transmission of pneumonic plague’, 1166-72. 59 60 22 URL: http://mc.manuscriptcentral.com/rmed Page 23 of 32 Journal of Medieval History 1 2 3 respectively.73 These show that, under certain circumstances, this form of plague can 4 be responsible for large-scale disasters. Moreover, the Manchurian outbreaks took 5 place over winter in conditions of extreme cold, conditions under which fleas are 6 unable to breed or develop. 7 8 9 10 The transport of Y. pestis to Iceland: Imports, exports and fishing 11 12 If we are to explain the medieval epidemics of plague in Iceland in terms of the spread 13 of Y. pestis there needs to be a credible explanation of how it got to Iceland and then 14 how it spread amongst the human population. There is no need, as we shall see, to Fo 15 exclude the rat and its fleas as part of this phenomenon although, for the reasons 16 pointed out by Karlsson, it is hard to accept that the dominant phase of the epidemics 17 could be bubonic. 18 rP 19 20 It is well documented that modern plague moves along trade routes frequently 21 associated with ‘rat-favoured’ merchandise.74 The bacteria may be carried by rats 22 themselves or adult fleas which can survive months without feeding. The aetiological agent which caused the Icelandic epidemics was likely brought to Iceland through ee 23 24 such trading activities and, as mentioned above, the New Annal records that the first 25 epidemic arrived by ship in the harbour of Hvalfjörður on the south west coast. 26 rR 27 There was regularly trade between Iceland and Europe throughout this period with 28 29 Norwegian, English and Hanseatic merchants holding sway in succession.75 Before 30 the fourteenth century, Iceland’s exports were largely vaðmál (coarse cloth) but the emphasis changed to skreið (stockfish/dried fish) in the fourteenth century and this ev 31 32 trend continued with the arrival of English vessels from around 1412, hence one 33 modern scholar’s view that the English were the most likely carriers of the plague.76 34 The ships used by the English ranged from 16 tons to 80 tons but were relatively ie 35 small by European standards. The Hanse merchants, however, used ships which could 36 carry between 80 and 200 tons. Iceland imported a variety of goods. Of these, grain 37 w and cloth provide a suitable substrate for the inadvertent transport of rats and fleas: 38 39 such ships were certainly large enough for rats and fleas to have gone unnoticed. 40 Archaeologists have identified another important consequence of the stockfish trade On 41 42 which was the generation of large quantity of waste in seasonal fishing stations where 43 44 73 L.-T. Wu, ‘Plague in the Orient with special reference to the Manchurian outbreaks’ Journal of 45 ly Hygiene 21 (1922), 65. 46 74 G. Lamb, The Etiologyand Epidemiology of Plague: A Summary of the Work of the Plague 47 Commission (Calcutta): Suprintendent of Government Printing, 1908).; M. McCormick, ‘Rats, 48 communications, and plague: towards an ecological history’, Journal of Interdisciplinary History 24 49 (2003), 1–25; M. Sharif, ‘Spread of plague in the southern and central divisions of Bombay province 50 and plague endemic centres in the Indo-Pakistan subcontinent’, Bulletin of the World Health 51 Organisation 4 (1951), 75–109 75 52 M. Gardiner and N. Mehler, ‘English and Hanseatic trading and fishing sites in medieval Iceland: 53 report on initial fieldwork’, Germania 85 (2007), 385–427; B. Gelsinger, Icelandic Enterprise. Commerce and Economy in the Middle Ages (Columbia, SC: University of South Carolina Press, 54 1981). See also Björn Þorsteinsson and Guðrún Ása Grímsdóttir, ‘Hafskipahafnir á 14. og 15. öld’ in 55 Saga Íslands IV , ed. Sigurður Lindal (Reykjavík: Hið íslenska bókmenntafélag, 1989), 143–59; R. 56 Harrison, H. M. Roberts and W. P. Adderley, ‘Gásir in Eyjafjörður: International Exchange and Local 57 Economy in Medieval Iceland’, Journal of the North Atlantic 1 (2008), 99-119. 58 76 Jón Ólafur Ísberg, ‘Sóttir’, esp. 210. 59 60 23 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 24 of 32 1 2 3 large quantities of fish were processed ready to be exported or transported to other 4 centres where they would be consumed. The presence of a very high number of 5 cranial bones and vertebral elements combined with a comparative shortage of cod 6 caudal vertebrae characterise a fish processing site.77 Such material would have 7 provided an excellent source of food for rats. 8 9 10 Written evidence also includes references to islands or coastal areas where seasonal 11 fishing took place but this is now being more precisely charted by the physical 12 evidence of seasonal fishing booths, relatively small structures designed to 13 temporarily house a boat’s crew but not a farming household.78 This supported a 14 movement of workers from farms to fishing sites which, again, is another facet of Fo 15 behaviour which may have facilitated the movement of plague and will be discussed 16 below. 17 18 The establishment of enzootic foci on Iceland rP 19 20 21 Previous authors have tended to take at face value the annals’ account of the arrival of 22 plague at one particular point in Iceland and its subsequent spread throughout the country.79 It has also often been assumed that the plague was completely absent in ee 23 24 Iceland before the first recorded epidemic, and again between the first and second 25 epidemics. Neither of these assumptions need be true. If the plague travelled by ship 26 and was carried by either rats, or merchandise, there is the possibility that there were rR 27 multiple introductions at different landing sites where trade was carried out. This 28 29 means that although the plague gave the impression of a progressive spread across 30 Iceland, the phenomenon may, instead, have been a function of the movement of ships around the coast, as has been argued for the British Isles by Maddicott for both the ev 31 32 first and second pandemics.80 33 34 The literature of modern plague records many incidences of pneumonic plague ie 35 outbreaks throughout the world which are small and die out without developing into 36 widespread epidemics.81 Bearing in mind that pneumonic plague is outwardly 37 w indistinguishable from other kinds of bacterial pneumonia, save for the high mortality 38 39 40 On 41 42 43 77 Y. Krivogorskaya, S. Perdikaris and T. H. McGovern, ‘Cleaning up the farm: a later medieval 44 archaeofauna from Gjögur, a fishing farm of NW Iceland’, in Dynamics of northern societies 45 ly proceedings of the SILA/NABO Conference on Arctic and North Atlantic Archaeology, Copenhagen, 46 May 10th-14th, 2004, eds B. Grønnow, J. Arneborg and H.C. Gulløv (Copenhagen: Aarhus University 47 Press, 2005), 383–95Jón Ólafur Ísberg, ‘Sóttir’, 198 appears more sceptical about the possible role of 48 fishing stations as homes for rats. 78 49 R. Edvarðsson, S. Perdikaris, T. H. McGovern and others, ‘Coping with hard times in North-West 50 Iceland: Zooarchaeology, History, and Landscape Archaeology at Finnbogastaðir in the 18th century’, 51 Archaeologia Islandica 3 (2004), 20–48. 79 52 E.g. Gunnar Karlsson, ‘Plague without rats’, 268-71; Benedictow, Plague. 80 53 J. Maddicott, ‘Plague in Seventh-Century England’, in Plague and the End of Antiquity, ed. L. K. Little (Cambridge: Cambridge University Press, 2007), 171–214. 54 81 For example: H. Wang and others, ‘A Dog-Associated Primary Pneumonic Plague in Qinghai 55 Province, China’, Clinical Infectious Diseases. 52 (2011), 185-90; D. van Zwanenberg, ‘The last 56 epidemic of plague in England? Suffolk 1906–1918’, Medical History 14 (1970), 63–74; M. Baltazard, 57 M. Bahmanyar, C. Mofidi and B. Seydian, ‘Le foyer de Peste du Kurdistan 1. Recherches dans le 58 foyer’, Bulletin of the World Health Organisation 5 (1952), 441–72. 59 60 24 URL: http://mc.manuscriptcentral.com/rmed Page 25 of 32 Journal of Medieval History 1 2 3 rate,82 it is unlikely that small outbreaks would have been sufficiently noteworthy to 4 be recorded by annalists. Similarly, small outbreaks of bubonic plague might also 5 have gone unreported. 6 7 A necessary step in the development of a sustained plague epidemic, characterised 8 9 either by bubonic or pneumonic spreads is the establishment of a population of 10 animals which act as a reserve, or ‘focus’, of the disease. This leads us to consider the 11 possibility that rats (with fleas) could have established colonies on the mainland even 12 though it has been suggested that rats were not present in Iceland in the Middle Ages 13 and that Iceland is too cold for them to establish themselves.83 14 Fo 15 The presupposition that Iceland’s climate was too harsh for rats needs to be debunked 16 first of all. Despite its northerly latitude, Iceland’s climate is essentially maritime with 17 cool summers and mild winters since the Irminger Current, a branch of the North 18 Atlantic Drift encircles the south, west and north coasts .84 The inhabited coastal rP 19 20 regions tend to be milder than its uninhabited highland interior and it is potentially 21 significant that the fifteenth century was relatively mild.85 22 While there is no unambiguous written evidence for rats in Iceland, the fact that mice ee 23 24 were reported in Icelandic records is suggestive that other rodents might have been 25 there. It is even possible that the annalists used ‘mouse’ for different species of 26 rodent, even if there are arguments against this.86 It also should be noted that rats, rR 27 when present, are not always to be seen, even when associated with an epizootic. For 28 29 instance, Petrie et al describe searching for rats in mud-brick houses in Egypt where 30 plague-carrying fleas were found in abundance.87 These investigators had to dismantle the houses before a large numbers of dead rats could be found in burrows and shafts ev 31 32 which served as cesspits. 33 34 Black rats may be relatively scarce in contemporary Northern Europe, but the reasons ie 35 for this are unlikely to be solely temperature related since it has been found that this 36 species has established itself on a number of sub-Arctic islands.88 It seems that a mean 37 w winter temperature of 2°C may be a limiting factor for black rats living away from 38 39 human habitation although this may not be an effect of cold per se but on the ability 40 On 41 42 82 K.F Meyer ‘Pneumonic Plague’ Bacteriological Review 25 (1961), 249; K. L. Gage, D. T. Dennis, 43 K. A. Orloski and others, ‘Cases of cat-associated human plague in the Western US, 1977–1988’, 44 Clinical Infectious Diseases 30 (2000), 897 for examples of misdiagnosis. 45 83 ly Duncan and Scott, ‘What caused the Black Death?’; G. Twigg, The Black Death: a problem of 46 population-wide infection’, Local Population Studies 71 (2003) 40-52. 84 47 Markús Á. Einarsson, ‘Climate of Iceland’ in World Survey of Climatology: 15: Climates of the 48 Oceans, ed. H. van Loon (Elsevier: Amsterdam, 1984), 673–97. 85 49 G. Massé, S. J. Rowland, M.-A. Sicre and others, ‘Abrupt climate changes for Iceland during the last 50 millennium: evidence from high resolution sea ice reconstructions’, Earth and Planetary Science 51 Letters 269 (2008), 565–9. Based on measurements of a biomarker (IP25) found in stratified sediments 52 produced by sea ice algae but historical sources are unreliable or limited from 1430 to 1560. 86 53 Gunnar Karlsson, ‘Plague without rats’, 279-80 87 G. F. Petrie, R. E. Todd, RE and others, ‘A report on plague investigations in Egypt’, Tropical 54 Disease Bulletin 21 (1924), 875–6. 55 88 Z. Pucek, ‘A preliminary report on threatened rodents in Europe’ in Rodents: A World Survey of 56 Species of Conservation Concern, ed. W. Z. Lidicker. Occasional papers of the IUCN(SSC) 4 (1989), 57 26–32; Y. Frenot, S. L. Chown, J. Whinams, and others, ‘Biological invasions in the Antarctic: extent, 58 impacts and implications’, Biological Reviews 80 (2005), 45–72. 59 60 25 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 26 of 32 1 2 3 of the species to forage89 since Black rats have been found to infest cold stores at a 4 constant temperature of 4°C.90 The black rat is not niche dependent, it is an 5 opportunist and able to modify its behaviour to survive in a variety of environments. 6 There therefore seems no environmental reason for the black rat not to have made it to 7 Iceland and survived in the Middle Ages just as the attested mice which have a 8 9 smaller body mass. Interestingly, the flea X. cheopis, more frequently found in 10 warmer climates, also exhibits a surprising ability to survive in unexpected places. It 11 has been observed in Manchuria where the flea survived winters more severe than 12 those in Iceland through living close to heating systems.91 13 14 As noted above, the absence of archaeological rat bones in medieval Iceland is not Fo 15 definitive evidence of their absence. It had been thought that R. rattus was relatively 16 uncommon in northern Europe in the middle ages but they have now appeared in the 17 archeology in substantial numbers.92 The most recent presentation of the 18 archaeological evidence for presence rats in Norway demonstrates that they were rP 19 20 present in urban or proto-urban settlements in the period before 1350.93 These places, 21 particularly Bergen and Trondheim, were the primary points of contact with Iceland. 22 Given, the tendency of rats to stow away, and the regular contact between Iceland and Norway, and later England, it would be surprising if rats never arrived and ee 23 24 periodically established colonies in several locations in Iceland. 25 26 In the fourteenth and fifteenth centuries the North Atlantic economy was scaling-up, rR 27 as fishing became more industrialised, even before the documented arrival of English 28 29 fishing vessels in 1412. One result of this process was the creation of environments in 30 which rats might thrive. We have clear evidence that rats could survive in the major nodes of this exchange network, the emerging small towns of Norway. But at other ev 31 32 locations in Iceland the production of dried fish for export (skreið), produced waste 33 from the gutting and preparing of the fish which would have fed any rats brought to 34 Iceland in ships bringing grain or other foreign goods. While the sites of fish ie 35 processing were not necessarily the same as those which were directly visited by 36 foreign trading vessels, there were regular lines of communication between many 37 w coastal sites in Iceland such that from time to time rats could arrive and survive 38 39 sufficiently long to pass on infection. If rats or fleas, carrying Y. pestis subsequently 40 did turn up, the infection would spread to colonies already established close to human habitation thus facilitating the transfer of plague to humans. Plague bacteria and On 41 42 infected fleas are able to winter in nests and burrows, thus helping to maintain the 43 presence of the disease organism.94 44 45 ly 46 47 89 48 B. Studholme, ‘Ship rat (Rattus rattus) irruptions in South Island beech (Nothofagus) forest’, 49 Conservation Advisory Science Notes No 318 (Wellington: Department of Conservation, 2000), 1–9. 90 50 D. K. Kocher and V. R. Parshad, ‘Shorter tails: a thermal adaptation of rats in cold stores’, CSIRO 51 Rodent Research Newsletter 20 (2005), 6. 91 52 Ibid., 331-2. 92 53 P. L. Armitage, ‘Unwelcome companions: ancient rats reviewed’ Antiquity 68 (1994), 231–40. M. McCormick, ‘Rats, communications, and plague: towards an ecological history’, Journal of 54 Interdisciplinary History 24 (2003), 1–25 55 93 A. K. Hufthammer and L. Walløe, 'Rats cannot have been intermediate hosts for Yersinia pestis 56 during medieval plague epidemics in Northern Europe', Journal of Archaeological Science 40 (2013), 57 1754–5. 58 94 Pollitzer, Plague, WHO Monograph Series, No. 22 (Geneva: World Health Organisation, 1954), 319. 59 60 26 URL: http://mc.manuscriptcentral.com/rmed Page 27 of 32 Journal of Medieval History 1 2 3 The transfer of Y pestis from animal to man 4 5 As mentioned above, infection can be passed to humans via fleabites. However, 6 whilst there are many species e.g. the more northern N. fasciatus in addition to X. 7 cheopis identified as plague carriers, it seems most unlikely that any of them could 8 9 have been present in significant numbers outside animal burrows in winter months, at 10 a time when the annals tell us plague was still rampant. However, there is evidence 11 that both bubonic and primary pneumonic plague can be caught directly from plague- 12 affected animals without an insect vector. This possibility was noted by Benedictow 13 but he felt that it was of little significance since the prevailing view, until at least the 14 middle of the last century, was that it was not proven.95Nevertheless, he cited the Fo 15 review by Gage et al of the incidence of plague, caused by exposure to infected cats, 16 amongst American veterinary surgeons. The infection was passed on pneumonically 17 by exposure to infected droplets and bubonically by scratches. There was no evidence 18 of flea involvement.96 There are numerous other reports in the literature of outbreaks rP 19 20 of pneumonic plague which appear to have started when individuals handle or skin 21 infected wildcats, cats, dogs and rodents.97 Dogs and cats are now considered to be a 22 significant risk factor in rural areas of the USA where Y. pestis is enzootic; they have been well established as domestic animals in Iceland for many centuries.98 ee 23 24 25 The evidence summarised above shows that the transfer of plague bacteria from 26 animal to man can be facilitated by domestic animals acting as intermediate vectors. rR 27 In this scenario, rat fleas are not required to bite humans. Thus, even when rat fleas 28 29 are not plentiful, cats and dogs will seek out rodent pests and thus acquire plague 30 themselves which they convey to their human companions by the pneumonic route, the bubonic route or both. ev 31 32 33 The transfer of Y pestis from human to human 34 ie 35 The transfer of the bacterium via fleabite alone does not provide us with a satisfactory 36 explanation of a large and sustained epidemic of plague in Iceland since we would 37 w need to postulate the existence of large populations of rodents, a plethora of fleas and 38 39 a substantial die-off for which there is no evidence and, as discussed above, the 40 epidemics continued throughout the winter months. We can argue for a jump from animal to man from established enzootic foci but the case for multiple infections on a On 41 42 larger scale is more difficult to envisage. This means that we need to consider the 43 pneumonic route as a more probable explanation for the sustained Icelandic plague 44 epidemics. 45 ly 46 95 47 Benedictow, Plague, 214. 96 48 K. L. Gage, D. T. Dennis, K. A. Orloski and others, ‘Cases of cat-associated human plague in the 49 Western US, 1977–1988’, Clinical Infectious Diseases 30 (2000), 893–900. 97 50 M. L. Gupta and A. Sharma, ‘Pneumonic Plague, Northern India, 2002’, Emerging Infectious 51 Diseases 13 (2007), 664–6; Gage and others, ‘Cases of cat-associated human plague’, 894; Wang and 52 others, ‘A Dog-Associated Primary Pneumonic Plague’, 186; W. H. Kellog, ‘An epidemic of 53 pneumonic plague’, American Journal of Public Health 10 (1920), 600. 98 K. A. Orloski and S. L. Lathop, ‘Plague: a veterinary perspective’, Journal of the American 54 Veterinary Medical Association 222 (2003), 444–8; Stefán Aðalsteinsson, ‘Uppruni íslenska húsdýra’ 55 in Íslensk þjóðmenning. Uppruni og umhverfi, I, ed. Frosti Jóhannsson (Reykjavík: Bókaútgáfan 56 þjóðsaga, 1987), 41-44, 45; E. P. Jones, K. Skirnisson, T. H. McGovern and others, ‘Fellow travelers: a 57 concordance of colonization patterns between mice and men in the North Atlantic region’, BMC 58 Evolutionary Biology 12:35 (2012), doi: 10.1186/1471-2148-12-35. 59 60 27 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 28 of 32 1 2 3 4 It is well documented that, travellers infected with the bubonic form, but fit enough to 5 walk, are prone to develop secondary lung infection and then spread the disease in its 6 pneumonic form and outbreaks can also occur without the help of fleas as discussed 7 above.99 However, as previously discussed, the bacterium is not readily transferred 8 9 through air by aerosolized droplets. So, in order to understand how the pneumonic 10 form of the disease spreads more easily than expected, the role of human behaviour 11 needs to be considered. 12 13 It is widely reported that people who travel to visit the sick may catch the disease and 14 return to their own rural communities where they can infect others. The latency of two Fo 15 to three days between contact and the development of debilitating symptoms is 16 sufficient to allow this.100 Secondly, the bodies and clothing of the deceased retain the 17 potential to infect others since it has been frequently observed that individuals who 18 handle infected clothing, those responsible for preparing bodies for funeral, and those rP 19 20 attending funerals can catch the disease.101 By way of explanation live aerosolised Y. 21 pestis remain viable on metal, polyethylene or paper surfaces for up to five days 22 providing that the culture is maintained within controlled levels of temperature and relative humidity.102 It is reasonable to suppose that viability would be comparable on ee 23 24 articles of clothing or skin. Thus, it is not the sick (who are usually incapacitated) that 25 initiate contact as it might be with, say, influenza. Contact is made when healthy 26 individuals visit immobilised infected individuals, attend funerals or prepare the dead rR 27 for burial. 28 29 30 The Manchurian epidemics exhibit another set of circumstances which appear to facilitate transfer of the disease between humans. In conditions of intense cold and ev 31 32 overcrowding, the disease was first spread amongst Tarbagon fur trappers and then 33 amongst itinerant Chinese workers who were prepared to move long distances by rail 34 and road to get work.103 A similar epidemic was also noted more recently amongst ie 35 itinerant mine workers in the Congo although intense cold was clearly not a factor in 36 this case.104 37 w 38 39 40 99 Wang & Others: ‘A Dog-Associated Primary Pneumonic Plague in Qinghai Province, China’, On 41 Clinical Infectious Diseases. 52 (2011):185-90.; W. J. Simpson, A Treatise on Plague dealing with the 42 Historical, Epidemiological, Clinical, Therapeutic and Preventative aspects of the Disease 43 (Cambridge: Cambridge University Press, 1905); E. R. Brygoo and M. Gonon, ‘L’épidémie de peste 44 pulmonaire de Doany en octobre 1957’, Archives de l'Institut Pasteur de Madagascar 13 (1958), 865– 45 ly 936. 46 100 F. Thielmann and F. Cate, ‘A plague of plagues: the problem of plague diagnosis in medieval 47 England’, Journal of Interdisciplinary History 37 (2007), 371–93. 101 48 A. K. Chalmers, ‘Notes on cases of plague in Glasgow’, British Medical Journal Sept 29, 1900, 49 953–4 on clothing and funeral attendance; L-T. Wu, ‘The second pneumonic plague epidemic in 50 Manchuria’, Journal of Hygiene 21 (1923), 262–288 re clothing and preparing bodies; Simpson, Report 51 on plague in the gold coast in 1908, 11. 102 52 L. J. Rose, R. Donlan, S. N. Banerjee and M. J. Arduino, ‘Survival of Yersinia pestis on 53 Environmental Surfaces’, Applied and Environmental Microbiology 69 (2003), 2166–71. 103 A. S. Loukaskin, ‘The Tarbagan or the Transbaikalian Marmot as a carrier of plague’, Extrait des 54 Comptes Rendus du XIIe Congrès International de Zoologie, Lisbonne, 1935: 2097–111; M. Gamsa, 55 ‘The epidemic of pneumonic plague in Manchuria 1910–1911’ Past and Present 190 (2006), 147–84. 56 104 E. Betherat, K. M. Lamine, P. Formenty and others, ‘Epidémie de peste pulmonaire dans un camp 57 minier de la république du Congo: le réveil brutal d’un vieux fléau’, Médecine Tropicale 65 (2005), 58 511–4. 59 60 28 URL: http://mc.manuscriptcentral.com/rmed Page 29 of 32 Journal of Medieval History 1 2 3 In later medieval Iceland it was usual for many men to work spend part of their year 4 working on a farm and then migrate to fishing stations in the fishing season where 5 they lived in purpose built turf-walled dwellings. While such fishing was on a small 6 scale it would still have resulted in contact with any established rodent-based foci of 7 the disease as described above and a sufficiently crowded environment where the 8 9 disease, having made the jump from animal to man, could be passed on easily. 10 11 The migration of workers from farms to fishing sites where they came into contact 12 return of the sick and the deceased to farmsteads for care and burial, the movement of 13 people between farmsteads to visit the sick, the preparation of the dead for burial and 14 the attendance of funerals would all assist the spread of the disease. Travellers Fo 15 suffering from the bubonic form of the disease, later to develop secondary pneumonic 16 symptoms, would assist in the transfer of the disease inland. The point to emphasise 17 here is that the single route of the disease’s diffusion identified in the annals’ account 18 for 1402-4 is extremely simplistic and unlikely to reflect the reality. rP 19 20 21 As a further cautionary note, it must be said that conclusive proof that Y. pestis was 22 the disease agent in medieval Iceland cannot be based on incomplete reports of the disease symptoms and epidemiology.105 This could only be confirmed by finding ee 23 24 DNA, or antigens specific to this bacterium, in skeletal remains, and that is not easily 25 done.106 However, recent studies are beginning to uncover this evidence elsewhere in 26 Europe although there is likely some way to go before it is accepted as the consensus rR 27 view.107 As yet, no such studies have been carried out in Iceland. 28 29 30 Estimating mortality ev 31 32 Gunnar Karlsson’s estimates of mortality, of 50-60% for the first epidemic, 30-50% 33 for the second, rely on two blocks of evidence: firstly, the brief accounts given in the 34 written sources and secondly, the number of farms which were unoccupied, “some 40 ie 35 years after the first epidemic receded”. 36 37 w We have argued that the collection of anecdotes, stories and oral tradition do not 38 39 provide a sufficiently robust foundation on which to construct a statistical analysis. It 40 is also significant that of the nine items of evidence used by Gunnar in his analysis, seven appear to be connected to religious sites so recording the deaths of clergy and On 41 42 their servants. Benedictow has emphasised that there are strong reasons for not using 43 mortality figures from records concerning the clergy to judge mortality rates more 44 generally.108 If the annals were written by scribes associated with the church, the 45 ly 46 105 47 Jón Ólafur Ísberg, ‘Sóttir’, 180, on the sheer varieties of possible forms of transmission and 48 associated symptoms. 106 49 M. Drancourt, M. Signoli, L. V. Dang and others, ‘Yersinia pestis orientalis in remains of ancient 50 plague patients’, Emerging Infectious Diseases 13 (2007), 332–3; V. J. Schuenemanna, B. Kirsten, S. 51 DeWittec and others, ‘Targeted enrichment of ancient pathogens yielding the pPCP1 plasmid of 52 Yersinia pestis from victims of the Black Death’, Proceedings of the National Academy of Sciences of 53 the United States of America 108 (2011), 1–7. 107 Little, ‘Plague historians in lab coats’. 54 108 O. J. Benedictow, Plague in the late medieval Nordic countries. Epidemiological studies (Oslo: 55 Middelalderforlaget, 1992). Taking the English parish clergy as his example, he argued that although 56 priests might have had a significantly higher risk of exposure to the disease by carrying out their 57 pastoral duties, on balance they would have suffered a lower mortality. This is because the other 58 epidemiological factors such as better diet, better nursing care, better housing and relatively high 59 60 29 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 30 of 32 1 2 3 devastating effect that the plague might have had on the scribes’ immediate religious 4 community might inform their views to a greater extent than would a less intense 5 mortality elsewhere. 6 7 Gunnar’s use of farm ‘abandonment’ as an estimate of mortality is also questionable. 8 9 Here, we are invited to believe that the rate of unoccupancy, immediately following 10 the first epidemic, was such that a strong signal could still be detected some 40 years 11 after the event despite an assumed population growth of 1% per annum. Even if we 12 accept that all the surveys of the 1430s and 1440s used by Gunnar Karlsson record 13 genuinely empty farms then we see wildly different percentages of abandoned farms, 14 ranging anywhere from as few as 2% (3 of 180) in the West Fjords, to 35% for those Fo 15 of Hólar which suggests a rather patchy effect of the epidemic. Moreover, it can be 16 shown that no more than 20% of farms, mostly of below average size, were empty at 17 the time of the surveys. It is reasonable to believe that smaller farms would be more 18 susceptible to economic challenges than larger ones and that there would always be a rP 19 20 number of them unoccupied at any given time but that the number would vary 21 continuously due to environmental changes or other disease epidemics e.g. those 22 recorded in the New Annal for 1420, 1421 & 1430/31. ee 23 24 If we accept that the written sources are not credible, in the statistical sense, and that 25 farm vacancy rates some forty years after the event are not reliable indicators of the 26 mortality caused by original outbreak, then what evidence is there to provide an rR 27 estimate of mortality other than to accept that it was sufficiently high to excite the 28 29 interest of the annalists? Rather than relying on guesswork, it might be useful to 30 consider the impact of pneumonic plague in a modern epidemic where there is an acceptable record of mortality and population numbers. If, for instance, we consider ev 31 32 the famous Manchurian epidemic of 1911, we find that the town of Harbin suffered 33 the highest mortality which was 9,000 deaths.109 This town had an estimated 34 population of around 30,000 which suggests a mortality rate of no more than 30%.110 ie 35 The environment offered for the disease to spread in Manchurian towns was ideal 36 with overcrowding and a population largely ignorant of the need to take preventative 37 w measures. Nevertheless, outbreaks were mostly concentrated along the railway lines 38 39 which provided a means of rapid communication and there is no evidence of a wider 40 dissemination throughout the whole of Manchuria.111 In contrast, Iceland was entirely rural with its largely scattered population likely to have been less susceptible to the On 41 42 spread of disease. It therefore seems more likely than not that the mortality rate in 43 44 average age would serve to increase survivability of the disease once it had been contracted. This view 45 ly was based on a consideration of bubonic plague which is, indeed, survivable. If we accept the view that 46 the dominant phase of this disease in Iceland was the pneumonic, the opposite would be true since this 47 form of the disease is virtually unsurvivable (without modern antibiotics) and none of the 48 epidemiological factors considered by Benedictow apply except for higher risk of exposure. We would 49 therefore expect the clergy to have a higher mortality rate than the general populace. Indeed, a greater 50 opportunity for nursing care would increase the opportunities for passing on the disease. 109 51 L.-T. Wu, ‘Plague in the Orient with special reference to the Manchurian outbreaks’, Journal of 52 Hygiene 21 (1922), 66. 110 53 M. Gamsa, ‘The epidemic of pneumonic plague in Manchuria 1910–1911’, Past and Present 190 (2006), 154. Note that Gamsa indicates that the total deaths in Harbin were 1500. He attributes a 54 greater mortality of 7-8000 to the town of Fujiadian, adjoining Harbin, “amounting to about a third of 55 the Chinese town’s population”. 56 111 R Farrar, ‘Plague in Manchuria’ Proceeds of the Royal Society of Medicine, 5(Epidemiological 57 Section) (1912) 1-24. Farrar estimated an attack rate of only 2.25 per 1,000 population across 58 Manchuria as a whole. 59 60 30 URL: http://mc.manuscriptcentral.com/rmed Page 31 of 32 Journal of Medieval History 1 2 3 Iceland would have been less than that observed, locally, in Manchuria. It cannot be 4 argued that the disease in Iceland was more lethal than anywhere else since the death 5 rate (untreated) in modern times remains virtually 100%. 6 7 Given that one purpose of this article has been to question previous estimates of 8 9 mortality caused by the Black Death in Iceland, we would speculate that for the 1402- 10 4 outbreak no more than 25-30% of the population died as a result of it. There is every 11 reason for following Gunnar Karlsson’s rationale for seeing the 1494-5 outbreak as 12 having less impact – because the levels of vacant farms are lower in the sixteenth- 13 century estate surveys than in the fifteenth-century ones – and so 25% might be a 14 maximum figure for the 1494-5 event. Fo 15 16 17 18 Conclusions rP 19 20 21 Hopefully, this consideration of the scale and demographic impact of the Black Death 22 has illustrated the complexity of the issues involved. The Black Death outbreaks of 1402-4 and 1494-95 were no doubt rightly remembered as more deadly than ee 23 24 epidemics of other kinds. Mortality, however, was very likely less than the estimates 25 put forward by Gunnar Karlsson. 26 rR 27 The description of the plague in the available historical records are considered to be 28 29 insufficiently robust to permit a useful assessment of mortality and the correlation 30 between numbers of abandoned farms and the size of the population is uncertain. Even allowing for the population to have replenished itself, and assuming that it did, ev 31 32 after each Black Death epidemic and before the records, it seems unlikely that either 33 outbreak killed much more than about 25% of the population. This was still a huge 34 catastrophe but this has to be seen in the context of the continued impact of outbreaks ie 35 of disease, recorded and unrecorded, to which the Icelandic population was subjected. 36 37 w We argue that the plague bacteria were brought to Iceland by normal trading activities 38 39 and the vectors were rats, fleas or a combination of the two. On reaching Iceland 40 bacteria were transmitted to established colonies of commensal rats or sylvatic rodents via their mutual fleas. There may have been multiple introductions if ships On 41 42 visited several trading sites. 43 44 Colonies of rats, which took advantage of opportunities presented by the waste 45 generated by the stockfish industry, probably existed as small enzootic reservoirs of ly 46 the disease near trading sites and other human habitations. The disease would then 47 transfer to humans through fleabites or, pneumonically, by direct contact with rodents 48 49 or companion animals such as cats or dogs. The movement of the bacteria among the 50 human population would then be facilitated in pneumonic form by the migration of 51 workers to and from fishing and trading sites, the return of the deceased to the home 52 farmsteads, visiting the sick, preparing the deceased for burial and contact with the 53 deceased in funeral ceremonies. 54 55 The idea that the disease spread in a continuous wave throughout Iceland is unlikely 56 to be correct. The model presented here is one of multiple introductions to Iceland, the 57 establishment of small, multiple enzootic foci and intermittent transfer from animal to 58 59 60 31 URL: http://mc.manuscriptcentral.com/rmed  Journal of Medieval History Page 32 of 32 1 2 3 man over an extended period of time. A more sanguine approach to some of the 4 evidence for the impacts of the Icelandic epidemics – their being smaller and patchier 5 than has been suggested – supports this view. We do not have to postulate the 6 continuous presence of rats over centuries but accept that the environmental 7 conditions required to drive a significant epizootic and epidemic may have been at an 8 9 optimum on relatively few occasions. 10 11 12 13 14 Fo 15 16 17 18 rP 19 20 21 22 ee 23 24 25 26 rR 27 28 29 30 ev 31 32 33 34 ie 35 36 37 w 38 39 40 On 41 42 43 44 45 ly 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 32 URL: http://mc.manuscriptcentral.com/rmed