The dancing plague of 1518

On Wednesday 14 or Thursday 15 July 1518, in Strasbourg, then a free city of the Holy Roman Empire on the Rhine, a woman began to dance in the street. She kept it up for days. The city council had her taken to a shrine of St Vitus, the saint who was believed both to send this sickness and to cure it, at the Hohlenstein, a hollow rock in the hills above Saverne, a small town to the north-west. There, a chronicle says, she became quiet. While she was still on the road, other people began to dance near the city stables, and within four days there were about thirty-four of them, women and men. The council forbade drums and pipes, lodged the dancers in two guild halls under guard, and on Friday 23 July ordered three wagons to carry them to the shrine in three groups, with priests engaged to sing mass over them. That morning three councillors questioned the dancers' families. A schoolmaster said that his son had watched the dancers and begun to dance himself; a wheelwright said that his wife, who had had fits before, had run after a band of poor people passing with a bagpipe. Two days later the council wrote to its bishop that the physicians called the trouble a natural illness of the heat and the stars, and that it was coming to an end.

In early August the sickness "arose again", in the council's words. On 2 August the council banned all public dancing until Michaelmas, 29 September; its decree speaks of up to fifty people dancing day and night at once. On 11 August it made every household and brotherhood responsible for its own sick. On 31 August it voted a hundredweight candle for the saint. By then, a Strasbourg family's chronicle records, more than two hundred people had danced in the space of four weeks.

The cause was not diagnosed in 1518 and is still disputed. The physicians of the city called it a natural illness. The chroniclers who wrote the story up over the following century saw a curse of St Vitus, a punishment from God, and a good deal of fraud. In 1952 E. Louis Backman, a Swedish physician who wrote a history of religious dance, proposed that the dancers had been poisoned by ergot, a fungus that grows on rye and whose poison causes convulsions and gangrene; that explanation became the popular one. In 2008 John Waller, a historian of medicine, argued in a book on the outbreak that the dancers were in a trance brought on by hunger, fear and belief in the saint's curse, and that the trance spread from person to person by sight. Gregor Rohmann, a German historian of the Middle Ages, and Robert Bartholomew, a sociologist who studies outbreaks of mass hysteria, read the dancing as a religious rite rather than an illness.

The facts are disputed as well as the cause. The popular account has four hundred dancers, fifteen deaths a day, a first dancer named Frau Troffea, and a council that built a stage and hired musicians so that the afflicted could dance the sickness out. The texts that survive give fifty, a hundred, two hundred, several hundred and four hundred dancers. The council's own minutes burned with the city library in 1870, when Strasbourg was besieged in the Franco-Prussian War, and are known only from extracts made in the eighteenth century. The modern retellings rest on a few chronicles written sixty to a hundred years after the event and on one reading of one lost manuscript.

We read every text that describes the dancing, in the order in which they were written, from the council's own papers of July and August 1518 to the chronicles of the following century, in the printed editions and, where they exist online, the page images. We asked of each text when it was written and what it says about numbers and deaths, and we traced each element of the modern story, the four hundred, the daily deaths, the stage, the name Troffea, to the text in which it first appears. We then tested the four explanations against that record and against three kinds of evidence the earlier writers did not have: reconstructions of the weather of the growing seasons of 1515 to 1518 from tree rings and written weather records, the grain prices of the same years, and the modern medical case series of outbreaks of dancing, fainting and convulsions that spread through schools and villages by suggestion.

The texts written in 1518 describe poor people set off by watching others dance or by the sound of a bagpipe, a verdict of natural illness from the physicians, up to fifty afflicted at once, and no death. Every element of the familiar story enters later, at a date that can be fixed: the name Troffea within fifteen years of the event, the deaths and the stage about sixty years after it, the four hundred in 1879, the daily death toll in the present century. Ergot fails on eight counts, the weather of the crop year among them. What remains is a behavioural epidemic among the poor of a city at the end of two years of dearth. We put the odds of that at about 85 percent, which makes a behavioural epidemic the probable cause: the evidence clearly favours it but does not prove it. The alternatives, and the assumptions that would move the figure, are set out in the panel below. The percentages record our judgment of the evidence; no measurement produced them.

What follows. Section 1 lists every text that describes the dancing, by date of writing, and shows how they reach us. Section 2 quotes the texts written in 1518 and the first texts to mention deaths, in the original and in translation. Section 3 traces how two hundred dancers became four hundred and fifteen new cases a day became fifteen deaths. Section 4 dates the council's measures. Section 5 tests the four explanations, ergot, a psychogenic epidemic, a religious rite, and other disease or fraud, against the record and against the reconstructed weather and grain prices of 1514 to 1519. Section 6 dates each element of the modern story. Section 7 explains the probabilities. Sections 8 to 10 say what is new here, what would overturn the verdict and where this report fell short, and section 11 lists the sources.

Strasbourg's own papers on the dancing exist in two forms. A few originals survive in the city and departmental archives. The rest is known through extracts that the city archivist Jacob Wencker made in the eighteenth century from the council protocols of 1518, which then burned with the municipal library in the siege of 1870. Those extracts were printed by the Strasbourg priest and historian Léon Dacheux in 1892 and 1899 under the misleading title "Annales de Sébastien Brant"; Brant, the author of The Ship of Fools, was the city's chancellor in 1518, but the compilation is Wencker's. Everything else is chronicle, and the chronicles were written between a few years and 180 years after the event.

The table lists every text that describes the dancing, in the order they were written, with what each says about numbers and deaths. Every text written within a generation of 1518 gives at most a few hundred dancers, and none of them mentions a death.

The passages below are quoted from the editions named, checked against the page images, with our translations. They are arranged in the order they were written. The Imlin chronicle comes first because its hand belongs to a Strasbourg burgher writing between 1516 and 1543, and because it is the only early text that gives day-precise dates.

Read in the order they were written, the texts give numbers that grow with the distance from the event, and deaths appear only when that distance has reached sixty years. The council's own extract says "uff fünfzig personen" were dancing day and night. The Imlin chronicle, the one private record from within a generation that gives figures, counts 34 within four days and "more than 200" within four weeks. The Augsburg chronicler Wilhelm Rem, writing in 1518, heard that "in one day a good fifteen people" were seized. No text written before about 1580 mentions a death.

The figure of 400 that runs through Midelfort, Waller and every popular account is a reading of one word in one manuscript. Rodolphe Reuss edited the Imlin chronicle from the original in 1873 and again in 1875 and printed "mehr dan 200" both times. Georg Witkowski, a psychiatrist writing in 1879, transcribed the same passage with "mehr denn 400", and Alfred Martin took Witkowski's text into his 1914 article, from which Midelfort and Waller took theirs. The manuscript burned in 1870 with the rest of the Strasbourg municipal library, so the two readings cannot be arbitrated on the page. Two things favour Reuss. He was editing the manuscript as a text, with the stated aim of printing it "nach der Originalhandschrift"; Witkowski was quoting it inside a psychiatric essay. And the independent line of the same tradition, the Duntzenheim chronicle as summarised by Jung in 1830 and the anonymous Kleine Strassburger Chronik of the early seventeenth century, both have "more than two hundred". We therefore treat 200 as the chronicle's figure and 400 as a nineteenth-century variant.

Even 200 is not a count of people dancing at once. Imlin's phrase is "in 4 wuchen wurden ir mehr dan 200 dantzet": over four weeks more than 200 had danced. The council's fifty and Specklin's "over 100 at one time" are figures for a single moment. The three numbers are compatible with an outbreak that peaked at several dozen and touched a couple of hundred people over six weeks, most of them briefly.

Half the modern literature lists a Strasbourg dancing plague of 1418 alongside the one of 1518. There was none. The council's Instruction of 23 July is dated "Veneris post Magdalene etc. XVIII", with the century left to be understood. Johann Schilter, printing it in his 1698 edition of the medieval Strasbourg chronicle of Jakob Twinger von Königshofen, read the abbreviation as 1418 and headed the whole section "Vom Veitz-Tantz An. 1418". Hecker copied the year in 1832; B. G. Babington's English translation of Hecker carried it into the medical literature; Boersch, writing in 1836, first printed 1518 and then recanted; the medical historian Heinrich Häser inserted "14.." into the German text; and in 1902 August Bernoulli, editing the Basel chronicle, found a notice that his manuscript dated to 1518 and moved it to 1418 because Schilter's council orders said so. Alfred Martin exposed the error in 1914 from the original, whose hand, magistrates' names and correspondence with the council extract all give 1518. Rem's dated Augsburg notice and the bishop's letter of 25 July 1518 put it beyond doubt. The consequence for the Basel notice is the reverse of what its editor intended. Restored to 1518, it becomes a witness written within twenty years of the event, in another city, using the same words as the anonymous verses Schilter printed. The phrase "many hundreds" is therefore early, and because it appears in the same words in two texts it is a stock expression rather than a tally.

The only contemporary text with the number fifteen in it is Rem's: "daß ain tag wol bei 15 menschen ankam", that on one day it came upon a good fifteen people. It is a rate of new cases on a bad day, reported in Augsburg from hearsay, and it says nothing about anyone dying. No text of the sixteenth or seventeenth century gives any daily figure for deaths, and the earliest text that mentions deaths at all, Specklin's, was written about sixty years after the event by a man born in 1536. Section 6 follows the modern texts in which this became "fifteen deaths a day", and says how much of that descent is traced and how much inferred.

The council's measures can be dated from the surviving papers and the extract, once the feast-day dates are converted with the Strasbourg calendar (St Margaret 15 July, St Adelphus 29 August, according to Grotefend's handbook of medieval dating). The weekdays check: 22 July 1518 was a Thursday, 23 July a Friday, 11 August a Wednesday, 31 August a Tuesday. The sequence is not the one in the modern retelling. The council never prescribed more dancing. Its first recorded decision restricts music; its commission of 23 July organises transport to the shrine and a diet; its mandate of 2 August bans dancing outright.

Three details in the 23 July minutes matter for what follows. Anthony Silinger, the wheelwright whose wife Apolonia was dancing, "had come in from the harvest": the grain harvest was under way in the last week of July. He thought at first that his wife was having one of her old attacks "von Sanct Martzolff", which Martin glosses as epilepsy, and she only got up and followed the dancers when "the poor people came past with a bagpipe". And the schoolmaster Burgawer's son Bernhart "had watched and this morning began to dance", which is the same mechanism Gebwiler describes when he says the disease attacked many "just from looking on". The documents of 1518 therefore describe cases that began by watching or by hearing music, among people whom the council calls "the poor people who are now dancing". Those are the features of a behavioural epidemic among the destitute; a poison in bread does not spread by sight or sound.

The stage on the horse-market with hired musicians and paid dancers, the detail on which the modern story of a council that made things worse depends, is Specklin's, written about 1580. The 1518 papers show the opposite policy: guild halls with guards, families and brotherhoods made responsible for their own, a cooling diet without wine, quiet music only, and a ban. The council did pay musicians in some form, since the letter of 25 July says the physicians allowed string music and the notices later forbid all music; but nothing in 1518 speaks of platforms built to keep the dancers going.

Four kinds of explanation have been offered since 1518. The physicians of 1518 called it a natural illness of heat and the stars. Chroniclers from Specklin onward saw a curse, a punishment and a good deal of fraud. Backman in 1952 proposed poisoning by ergot, and that reading became the popular one. Waller in 2008 argued for a trance state induced by hunger and fear, spread by belief in the curse of St Vitus; Rohmann in 2013 and Bartholomew in 2000 read the dancing as a religious rite rather than an illness. Each is tested here against the record as sections 1 to 4 reconstruct it, and against three bodies of evidence the earlier writers did not have: the climate of the growing seasons of 1515 to 1518, the grain prices of the same years, and the modern case series on outbreaks of this kind.

Ergot

Ergot is the hard black body, the sclerotium, that the fungus Claviceps purpurea forms in place of a grain when it infects rye at flowering; it is milled into the flour with the grain. Its alkaloids cause two diseases. The gangrenous form, St Anthony's fire, produces burning pain, dry gangrene and loss of limbs. The convulsive form produces formication (a feeling of insects crawling under the skin), cramps of the hands and feet, painful spasms and convulsions, and in the great German epidemics of 1581 to 1597 a case fatality of ten to twenty percent (George Barger, whose 1931 monograph is still the standard history of ergot, p. 39). The hypothesis that the dancers of 1518 were poisoned by ergot fails on eight counts. The argument most often made against it, that the poison would have decayed in storage, is not among them, for the reason given at the end of this section.

The season is wrong. Ergotism follows the new harvest, because the sclerotia go into the flour with the freshly threshed grain; the German epidemics in Barger's collection begin in late summer and autumn, after the new grain is milled. The dancing began on 14 or 15 July, when the 1518 rye was still standing. Anthony Silinger had "come in from the harvest" on 23 July, so the cutting had just begun; the bread of the first week of the outbreak was baked from the crop of 1517, or from older stock in the granary.

The weather of the crop year is wrong. Ergot needs a wet spring for the sclerotia to germinate and a cool, wet flowering in late May and June so that the florets stay open (Miedaner and Geiger 2015, section 3; Barger 1931, pp. 98–99). The Old World Drought Atlas, a reconstruction of summer soil moisture across Europe from tree rings, gives the summer of 1517 at Strasbourg an index of −6.8, the third driest summer of the sixteenth century, with every cell of the Upper Rhine below −3 and drought across ninety-six percent of France (Cook and others 2015; the values are our extraction from the atlas, kept with this report). The spring of 1517 was the tenth driest of the century in the published reconstruction of past rainfall, and May and June 1517 were 2.1 and 1.2 degrees above the 1961 to 1990 mean in the reconstruction of past temperatures from written weather records. The price record shows that the harvest of 1517 was small, as a drought harvest is; but a drought does not produce ergot.

The regional form is wrong. Barger found that the recorded epidemics west of a line running through Lorraine were gangrenous and those east of it convulsive, and that "west of the border-line typical convulsive ergotism was never recorded" (Barger 1931, p. 28). Strasbourg lies on that line, and the Alsatian ergotism that the hospitals of the Order of St Anthony treated, and that the Isenheim altarpiece, painted for one of those hospitals, depicts, was St Anthony's fire. Nothing in 1518 mentions burning limbs or gangrene, and nothing mentions the formication and cramps of the convulsive form either.

The course in the individual is wrong. Convulsive ergotism keeps its victims in bed with spasms for weeks. The dancers of 1518 were on their feet in the streets for days, and one of them followed a bagpipe. Neither Barger's German and Russian case descriptions nor the 1951 outbreak of mass poisoning at Pont-Saint-Esprit in southern France, which is the case usually cited for florid behaviour and which the European Food Safety Authority now attributes to organic mercury rather than ergot (EFSA 2012, p. 95), contains sustained dancing.

The spread is wrong. A poison spreads through a household's bread. The record of 1518 shows a boy who "watched and this morning began to dance", a woman set off by "the poor people who came past with a bagpipe", and Gebwiler's statement that the disease seized many "just from looking on". The commission's minutes record no common household, bakery or meal among the sick.

The fatality is wrong, so far as the record goes. No text written before about 1580 records a death, and the council's own words for the first group sent to Saverne are "fast erledigt", almost all relieved. A case fatality of ten to twenty percent among fifty to two hundred people would be five to forty deaths. The council, which recorded its payments for the wagons, the priests and the candle, records none. The silence is in the council's own papers, which are the sources most likely to have recorded a death. The burial registers that could settle the question were not searched.

The physicians did not see it. Strasbourg's physicians examined the sick for some days and told the council in writing that this was a natural illness due to a conjunction of the stars and the heat of the season, to be treated with a cooling diet without wine. They and the St Anthony hospitals of the region knew ergotism; they did not name it.

The end is wrong. The outbreak peaked in the last ten days of July, relapsed in early August, and was over by the end of that month, when the council voted its thank-offering. Grain from a poisoned store would have been eaten, and would have produced cases, for months.

The one argument in the popular literature against ergot that does not hold is chemistry. It is often said that any ergot in the 1517 crop would have lost its potency by July 1518. Barger reports that whole sclerotia kept dry lose their alkaloids only slowly, a fourteen-year-old sample still meeting the pharmacopoeia (Barger 1931, pp. 211–212), and modern storage studies agree. We do not use the argument. The rye harvest of 1515, which had classic ergot weather, could in principle have reached the poor through the granary distributions of 1517 to 1518. Nothing in the record describes such stock, the dearth of 1516 to 1517 would have drawn old grain down first, and the eight counts above stand whatever the chemistry. Ergot keeps a few percent because one poisoned person inside a larger behavioural outbreak cannot be excluded. The outbreak as a whole does not look like a poisoning.

A behavioural epidemic: the psychogenic reading

Waller's argument has two parts. The first is that the people of Strasbourg in 1518 were hungry, frightened and primed by belief in the curse of St Vitus. The second is that under such conditions people enter dissociative trance, and that trance spreads by sight. The climate and price record bears out the first part in substance and corrects it in detail. Three harvests and vintages failed in a row, 1515 from cold and wet, 1516 from heat, 1517 from drought (Johann Wencker's chronicle, Dacheux 1892 no. 3008). The rye price at Strasbourg reached 160 pfennig a Viertel in 1516, the highest in Hanauer's price series since 1437, and the winter of 1516 to 1517 was 2.4 degrees below the century's mean in the European reconstruction, more than two standard deviations. The hospital was "so full of starved and sick" that it needed thirty Viertel of grain a week; the city sold subsidised rye at seven shillings against a market of seventeen or eighteen and fed 450 destitute strangers through the winter of 1517 to 1518 (Jacob Wencker's extracts, Dacheux 1892 no. 3237d; Otto Winckelmann's history of the city's poor relief, 1922, pp. 43 and 73). The council's papers of July 1518 speak throughout of "die armen leute", the poor people who were dancing. The prices of 1518 were back to normal after the harvest, so the dancing began in the last month of the pre-harvest gap of a two-year dearth. Waller's premise, a city in dearth, is therefore borne out. Several of the specifics he adds do not: the English sweating sickness, an epidemic fever of the period, did not reach Strasbourg in 1517 (Specklin's notice is a misdating of 1529, when the city's printers issued their regimens against it), the "smallpox hospital" is the Blatterhaus, the house for syphilis, the leper-house accounts of 1518 to 1519 show eight inmates and two deaths rather than a surge, and the Bundschuh "rebellions" of 1517 (the Bundschuh were the peasant conspiracies of the Upper Rhine, named after the peasant's laced shoe) were a conspiracy scare whose substance recent work doubts (Sheik 2021/22).

The second part is a claim about mechanism, and the record supports it directly. The commission minutes of 23 July are the only eyewitness testimony of the outbreak, and every case in them is a case of contagion by sight and sound: a schoolmaster's son who watched and then began, a wheelwright's wife with a history of fits who followed the bagpipe, a beggar's daughter, a labourer's wife. Gebwiler, who taught at the cathedral school that summer, wrote that the disease seized people "just from looking on". The modern name for an outbreak with those features is mass psychogenic illness of the motor type, which the psychiatrist Simon Wessely distinguished in 1987 from the anxiety type: it arises against chronic strain, comes on and goes off gradually, produces abnormal movement and trance imagery, and lasts weeks to months. The 1518 outbreak lasted about six weeks with a relapse. Its size, up to fifty at once in the council's mandate and more than two hundred over four weeks in the Imlin chronicle, is an ordinary size in the modern series (nineteen at Le Roy, New York, in 2011 to 2012; 194 cases in the Nepal school study; over 1,600 in Afghan schools between 2009 and 2012). And across more than nine thousand modern cases of this kind the recorded death toll is zero, which is also what the record of 1518 shows. The deaths and the hundreds, which were the two objections to the psychogenic reading, are later additions to the story (sections 3 and 6); without them the reading fits the record.

The reading has one weakness. The texts of 1518 describe the group dancing "day and night" but give no individual durations; the only one reported is the first woman's, six days in the Imlin chronicle and four in the later Kleine Strassburger Chronik, both at second hand. Whether anyone danced for days without pause, as the modern retelling insists, the record of 1518 does not say. The physicians' regimen, guards to stop the sick from hurting themselves, and a hospital that kept dancers for days all suggest exhausted people in fits rather than people dancing continuously for days.

A rite rather than an illness

Martin argued in 1914 that the dance at the shrine was itself the cure, Bartholomew in 2000 that the medieval dancers were pilgrims and sectaries rather than patients, and Rohmann in 2013 that the dancing mania was an illness constructed by religion, in which a divinely sent frenzy is discharged by a rite at the saint's altar. The evidence for this reading is real and older than 1518: a curse "sant Vitstantz" fined in the town law of Rottweil, in Swabia, in 1485; a poor man dancing at the Wasserkirche, a church in Zurich, "on St Vitus's day" in 1452; sick dancers from the priory of Saint-Jean-de-Bassel given a shilling at Obernai in 1463; and the annual dancing pilgrimages from the Breisgau, across the Rhine, to St John at Biesheim and St Vitus at Wasenweiler that the physician Johann Schenck von Grafenberg described later in the century. Specklin says the disease struck those "denen man St. Veitstantz fluchte", on whom the dance had been wished. The belief existed before 1518, and it determined both the form the behaviour took and the remedy sought for it.

What the 1518 papers do not show is dancers who chose the rite. The pilgrimage was organised, paid for and policed by the council: wagons from the hospital and the cathedral works, priests engaged in advance, one pfennig per sick person taken from the alms, named individuals sent in three groups. The physicians called it an illness, the episcopal vicar accepted their diagnosis and added public prayers, and the council's first measures were secular. The red shoes with chrism crosses, the most ritual detail in the whole story, appear only in Specklin, sixty years later. Wencker in 1637 says that some ran to the shrine themselves and "fell down dancing before the image", which is the one sign of voluntary pilgrimage, and it was written 119 years after the event. The ritual reading explains why the dancing looked as it did and why St Vitus was the cure. It does not explain why the poor people of Strasbourg began in the third week of July 1518, and Rohmann himself says of that question, "we do not yet know what actually happened".

Other organic causes, and fraud

Apolonia Silinger had a history of attacks "von Sanct Martzolff", which Martin glosses as epilepsy, before she joined the dancers. Individual epilepsy, Sydenham's chorea (the jerking disorder that can follow a streptococcal infection) or fever will have been present among dozens of destitute people in July, and the physicians' own diagnosis named heat. None of these spreads by watching, and none produces a six-week outbreak that ends with a pilgrimage. An encephalitis would have killed. Fraud was the chroniclers' favourite: the chronicler of Limburg on the Lahn called the dancers of 1374 "duisserie unde ketzerie", devilry and heresy, Specklin says "much roguery went with it", and Paracelsus made the first dancer a wife feigning illness to annoy her husband. Beggars will have joined a crowd that the city was feeding, and the council's instruction that each sick person's pfennig be paid from the alms shows that the dancers were on relief. Imitation and opportunism belong inside any behavioural outbreak of this kind; they are not a rival explanation of it.

The popular account of the dancing plague has a named first dancer, four hundred victims, fifteen deaths a day, and a council that built a stage and hired musicians to keep the sick dancing. None of those four elements is in the record of 1518. The Strasbourg museums' exhibition of 2018, with Élisabeth Clementz and Georges Bischoff, had already challenged the abundance of contemporary testimony behind the familiar story and treated Frau Troffea as a later construction rather than a documented participant. This section builds on their correction and extends it to the numbers and the daily toll, dating each element's entry into the story where the texts allow and saying where they do not.

Frau Troffea is Paracelsus's. He treats St Vitus's dance in two writings, and in the book on invisible diseases, written about 1531 to 1532 and first printed in 1565, he tells its origin as a fable. He names no city and no year.

Paracelsus's Troffea is a moral tale about imagination and belief, and it is the tale of a fraud. His earlier book on the diseases that rob people of reason, of about 1525 to 1526, had already sorted St Vitus's dance into three kinds, one from the imagination, one from lust and one from the body, with cures that include burning a wax image of the sufferer and confinement in the dark on bread and water; it too names no city. The identification of his unnamed wife with the unnamed woman of Strasbourg is at least as old as Backman's 1952 book, Midelfort supplied the date 14 July by counting back from the Imlin chronicle, and Waller's book fused the two into a person with a husband and a street. No Strasbourg source of any century names the first dancer. Lynneth Miller's 2017 statement in the journal Dance Research that four of six chronicles call her Lady Troffea, now repeated in Wikipedia, is untrue of every chronicle we read. The modern retelling uses Paracelsus's name for the first dancer and leaves out his statement that she was feigning illness.

Fourteen eighteen is Schilter's. In 1698 he printed the council's Instruction and six verses from an unidentified manuscript chronicle under the heading "St. Veits Tantz An. 1418", reading the "etc. XVIII" of the Instruction's date as 1418. Hecker took the date from Schilter in 1832, Babington carried it into English in 1844, and it is still in Bartholomew's 2000 essay. Martin disposed of it in 1914; there was no dancing at Strasbourg in 1418.

Four hundred is a reading of the Imlin chronicle. The manuscript, as Reuss printed it in 1875, says that "in four weeks more than 200 had danced". Witkowski in 1879 read the same passage as 400, Martin printed Witkowski's figure in 1914, Midelfort followed Martin in 1999, and Waller, citing Backman, Midelfort and Martin, wrote "as many as 400" in his 2008 article and "about 400 men, women and children" in 2009. His book knew the other reading, "Duntzenheim says only two hundred were involved", and preferred 400. A press interview about the book put 400 into the first version of the Wikipedia article on 8 August 2008. The council's own figure, for one moment in late July, was up to fifty; the chronicle's two hundred is a count over four weeks. Section 3 gives the whole chain.

Dozens of deaths, fifteen a day is the most recent element, and the one whose descent we can only partly trace. No text before Specklin, about 1580, mentions a death; Specklin says "many danced themselves to death", Kleinlawel in 1625 "many stayed dead of it", Wencker in 1637 "several died of it". Those three late statements are the whole basis for Waller's "the chronicles agree, dozens were dying from exhaustion". The daily figure has one candidate source in the sixteenth century, the Augsburg merchant Wilhelm Rem: "on one day a good fifteen people were struck by it". Rem's verb is ankam, came upon; he is counting new cases on the worst day, and he mentions no death. Martin quoted the sentence in 1914 to prove that the outbreak belonged to 1518 and not 1418, and used it for nothing else. The earliest text we found in which fifteen cases have become fifteen deaths is Waller's essay in the medical journal The Lancet of February 2009: "One chronicle states that it claimed, for a brief period at least, about 15 lives a day". The essay names no chronicle, and its further-reading list gives only Backman and Midelfort; Midelfort has no fifteen. Waller's book has "one chronicle suggests that (at least for a time) fifteen were dying each day". We could not see the endnote to that sentence. The snippets of the book's endnotes that a library full-text search returns show Waller citing Rem's Chronica newer geschichten twice, so he knew the chronicle directly, but which sentences those notes serve the snippets do not show. Backman's book, which translates the chronicles into English, was not available to us, and a search of its text for the figure found nothing. So the descent from Rem's count of the sick to Waller's count of the dead is an inference from the wording and from the absence of any other source with the figure; it is not a chain of citation traced page by page. If Backman's translation of Rem already reads "died", the conversion is his; if Waller's note cites Martin, it is Waller's. What is established is narrower and still matters: the only sixteenth-century text with fifteen and a day counts the sick, and says nothing of the dead. Wikipedia added the death rate on 8 April 2016, citing the Lancet; Miller repeated it in 2017 without a source, and the online magazine The Public Domain Review in 2018 reasoned from it to a toll "into the hundreds".

The stage and the musicians are Specklin's. Writing around 1580 he says the council set the dancers on a platform at the horse market and in the guild halls, "and hired people to dance with them and drums and pipes". The papers of 1518 show the council forbidding drums on 22 July, sending the sick to the shrine on 23 July, prescribing string music only on the physicians' advice on 25 July, and banning dancing altogether on 2 August. Whether Specklin preserves a memory of the first days that the extracts leave out cannot be settled. The papers of 1518 show that the council's policy was to contain the dancing rather than to encourage it. Two smaller points belong here. Waller writes that the physicians declared the disease natural "having excluded astrological and supernatural causes"; the magistrate's letter of 25 July says they attributed it to a conjunction of the stars and the heat. And the first version of the Wikipedia article, written from a press interview in August 2008, turned Waller's "dozens were dying" into "most of them died", which stood for two years.

The scholarly literature never held these four elements together. Martin's 1914 source study, Midelfort's 1999 chapter and the Strasbourg historians of the 2018 exhibition, Clementz and Bischoff, all give a small outbreak, no contemporary death, an invented Troffea and a council that sent the sick away. The popular account descends from Hecker through Backman to Waller and Wikipedia; this report corrects that account.

The probabilities at the head of this report are a judgment on the record as reconstructed here. They are given for the primary cause of the outbreak; imitation, begging and individual illness sat inside whichever cause was primary. The rows of the table beside them are the concessions a critic might ask for. Accept Specklin's and Kleinlawel's deaths, written sixty and a hundred years after the event, and the organic causes gain, because a behavioural epidemic does not kill. Read Rem's fifteen as deaths, as Waller does, and the same happens. Allow ergot a route through granary stock from the wet harvest of 1515, and give the cold June of 1518 full weight, and ergot rises to about twelve. Read the physicians' "natural illness of heat and the stars" as a real epidemic disease, and the other organic causes rise to fifteen. Take Wencker's "ran there themselves" and Specklin's red shoes as evidence that the dancers chose the rite, and weight moves from illness to rite without leaving the behavioural share. Grant all of these at once and the behavioural reading still leads at about 55, because the contagion by sight, the small size, the timing before the harvest and the drought of 1517 survive every assumption in the table.

The account that the texts written in 1518 support is this. On Wednesday 14 or Thursday 15 July 1518 one unnamed woman began to dance in the street. Within four days about thirty-four people were dancing; on the worst day about fifteen new cases appeared; the council's mandate of late July speaks of up to fifty afflicted at once, and the Imlin family chronicle, written within the generation, counted more than two hundred over four weeks. The first is a figure for one moment, the second the only contemporary cumulative count, and nothing in the record contradicts either. They were poor: a schoolmaster's son, a wheelwright's wife with a history of fits, a beggar's daughter, a labourer's wife, a journeyman's son. The council restricted drums on 22 July, sent the sick to St Vitus at the Hohlenstein in three wagon-loads on 23 July, told the bishop on 25 July that the physicians called it a natural illness and that it was coming to an end, banned public dancing until Michaelmas on 2 August, ordered the brotherhoods to keep their sick indoors when the sickness "arose again" in early August, made families and brotherhoods responsible for their own on 11 August, and voted a hundredweight candle for the saint on 31 August. Nobody is recorded as having died in any of these texts.

Nothing here overturns what historians already think about the outbreak. Alfred Martin's source study of 1914 dated the outbreak to 1518, printed the council's papers and disposed of Schilter's 1418. Erik Midelfort in 1999 gave the sober English account: one woman about 14 July, thirty-four within four days, the sick sent to Saverne, deaths only in Kleinlawel. The Strasbourg historians who mounted the 2018 exhibition, Élisabeth Clementz and Georges Bischoff, reduced the outbreak to "quelque cinquante personnes", found no death in any contemporary document, exposed Troffea as Paracelsus's invention and dismissed ergot. John Waller's premise, a city in dearth and fear, is confirmed by the climate and price record, and his mechanism, contagion of a trance-like state by sight, is the one the commission minutes describe. This report agrees with all of them and builds on them.

What it adds, so far as the literature we could reach shows, is listed below. Almost all of it is a new reading of evidence that was already in print. The climate and price datasets are the one body of evidence not brought to the question before, and they are new only in their application here. Each item carries a tag saying which kind it is.

• The daily death toll has one candidate source, and it counts the sick. new analysis of a printed text The only sixteenth-century sentence with fifteen and a day is Wilhelm Rem's "daß ain tag wol bei 15 menschen ankam", on one day it came upon a good fifteen people. Martin quoted it in 1914 only to fix the year. Waller's Lancet essay of 2009 and his book give fifteen dying a day, and Wikipedia took the figure from the Lancet in 2016. That the death rate descends from Rem is our inference from the wording and from the absence of any other source with the figure. Fragments of Waller's endnotes show that he cites Rem's chronicle, but not which sentence the citation serves; Backman's translation was not available to us. The chain is therefore supported by the wording of the texts; we have not traced it through citations. Clementz said in 2016 that no contemporary author mentions a death; we have not seen anyone identify the sentence behind the fifteen.
• Four hundred is a reading of 1879. new analysis of printed editions Reuss's edition of the Imlin chronicle from the manuscript (1875) reads "mehr dan 200"; the independent Kleine Chronik reads "über 200"; the 400 enters with Witkowski in 1879 and reaches Waller through Martin and Midelfort. Whether the manuscript can settle it we could not check.
• An early witness belongs to 1518 rather than 1418. new analysis of a printed edition An anonymous Basel chronicle written before 1537 records "several hundred people, women and men, at Strasbourg" dancing "day and night", and names the plague "sant Vietz tantz". Its manuscript dates the entry 1518; the editor Bernoulli moved it to 1418 in 1902 on the strength of Schilter's misdated council orders, and it has been invisible since. It is the earliest narrative written outside Strasbourg and the earliest source for "hundreds", and its wording matches the verses Schilter printed.
• The weather of the crop year rules out ergot on its own terms. evidence not previously brought to the question The bread of July 1518 came from the harvest of 1517. The Old World Drought Atlas makes the summer of 1517 the third driest of the century at Strasbourg and the spring of 1517 among the driest; May and June were warm. That is the opposite of the wet, cool flowering the fungus needs. The one favourable sign in the window, the cold June of 1518, belongs to a crop that had not been cut when the dancing began. We know of no previous test of the ergot hypothesis for 1518 against reconstructed weather. The extraction from the atlas was run a second time from scratch and gives the same value; the reconstruction itself is the published work of the climate scientists cited.
• The outbreak ended in August rather than October. a calendar correction Martin read the council's "Dienstag post Adolphi" as late September and concluded the epidemic ran at least to 29 September; St Adelphus falls on 29 August, and the candle was voted on Tuesday 31 August, as Clementz's chronology also has it. The magistrate told the bishop on 25 July that the sickness was "coming to an end". The outbreak lasted about six weeks, with one relapse.
• The council contained the dancing; it did not stage it. new reading of documents Martin printed Drums restricted on 22 July, wagons to the shrine on 23 July, a cooling diet and string music only on the physicians' advice on 25 July, a general ban on 2 August, families and brotherhoods made responsible on 11 August. The platform on the horse market with hired dancers is Specklin's, written about 1580, and it is the detail on which the modern story of a council that made things worse depends.
• The pilgrimage was imposed, which weakens the ritual reading. new reading of the Instruction of 23 July The Instruction of 23 July shows the council hiring wagons and priests and paying each sick person's pfennig from the alms. Dancers who chose the rite appear only in Wencker (1637).
• Several of Waller's supporting details do not hold. new analysis of published work The English sweat did not reach Strasbourg in 1517; Specklin's notice is a misdating of 1529. The "smallpox hospital" is the Blatterhaus, the house for syphilis. The leper house had eight inmates and two deaths in 1518 to 1519. The Bundschuh of 1517 was a conspiracy scare rather than a series of rebellions.
• Two arguments against ergot made in the literature do not hold, and this report does not use them. published literature applied The claim that ergot alkaloids would have decayed by July is unsound (Barger 1931, pp. 211–212), and Pont-Saint-Esprit, the modern case usually cited for ergot-induced mass behaviour, is now attributed to organic mercury (EFSA 2012, p. 95). The case against ergot rests on season, weather, form, course, spread, fatality, diagnosis and end; chemistry plays no part in it.
• The earlier outbreaks on the Upper Rhine are listed and dated. new analysis of printed editions Strasbourg has no record of dancers in 1374; Königshofen's chronicle has only the floods. The fifteenth-century Upper-Rhine records are Zurich 1418 or 1428, Zurich 1452 on St Vitus's day, and the cluster of 1463 at Metz, Obernai and Eberhardsklausen under St John. The curse "sant Vitstantz" is fined in the Rottweil town law of 1485. "Strasbourg 1418" is Schilter's error.

• A document of 1518 that records deaths among the dancers: a hospital or parish register, or the full narrative that the council's own register says existed "in my chronicle", Sebastian Brant's lost account. Deaths would cut the behavioural share to about 70 percent and reopen the organic causes.
• The German original of the magistrate's letter of 25 July 1518 (ADBR 12 J 1495), if it describes burning limbs, gangrene, or cramps of the hands and feet. We have it only in a 2018 French translation.
• A chronicle notice of a wet May and June 1517 at Strasbourg. The tree-ring drought value for 1517 is regional and robust, but the seasonal rainfall reconstructions for the sixteenth century are weak, and one wet fortnight at flowering would not show in them.
• An ergotism epidemic recorded in Alsace or Baden in 1518, for instance in the accounts of the St Anthony hospital at Isenheim. None is known to Barger or to the later literature.
• Evidence that the sick went to St Vitus of their own motion before the council organised the transport of 23 July. That would move weight from illness to rite without touching the verdict against ergot.
• The Imlin manuscript reading 400. It would restore Witkowski without changing the council's fifty or the absence of deaths.

• Hieronymus Gebwiler's contemporary account (ed. Stenzel 1926, pp. 74–75) was not read; it is known here through Midelfort and the 2018 exhibition dossier. The council register's reference to a full account "in my chronicle" points to a lost narrative by Brant.
• The letter of 25 July 1518 was read only in the French translation published for the 2018 exhibition. The mandate of 2 August 1518 (AVES 1 MR 3, f. 72r) is not transcribed anywhere we could reach.
• Witkowski's 1879 paper (Allgemeine Zeitschrift für Psychiatrie 35) could not be found online, so the 200 or 400 variant rests on Reuss's edition against Martin's report of Witkowski.
• Backman's 1952 book was not obtained and is used through Bartholomew, Midelfort, Waller and about sixty snippets from a library full-text search. Waller's articles of 2008 and 2009 were read in full, but his book only in about a hundred and twenty such snippets, so the endnote behind "fifteen were dying each day" was not seen; the derivation from Rem rests on the wording and on the absence of any other source with the figure. Midelfort's chapter was read in the text of a scan on archive.org whose provenance we could not confirm.
• Rohmann's 2013 monograph was not read; his argument is taken from his 2020 article, from reviews and from an interview. Clementz's transcriptions of the 1518 documents in the 2018 catalogue (pp. 136–152) were not obtained; the originals are read here in Martin's 1914 edition.
• Hanauer's price volume of 1878 could not be opened; the prices come from two later transcriptions that agree on the ranking of years but not on the 1516 rye figure.
• The gridded climate reconstructions have low skill before 1659. The drought atlas value for 1517 is safe because it is extreme and regional; the seasonal rainfall figures are not. Wencker's one dated weather notice can belong to 1517 or to 1518, and the report uses both readings.
• Three words in the council's papers remain unexplained: "pfrymen" in the Instruction, "termen" in the decision of 22 July, and "brucken", which Martin reads as "beucken", kettledrums.
• No Strasbourg parish or hospital register was searched. The absence of deaths is an absence in the council extracts, the letter, Rem, Imlin and Gebwiler as reported, which is the whole contemporary layer, but it is still an argument from silence, and silence is not proof that nobody died.