Giulia Tofana is one of the best-known names in the history of poison. Popular books and websites describe her as a Sicilian woman who sold a clear, tasteless liquid called Aqua Tofana (also spelled Acqua Tofana or Acqua Toffana) to wives who wanted to be rid of their husbands, and who is said to have confessed to hundreds of deaths before being executed. The poison's name later reached Mozart's deathbed and the pages of 18th- and 19th-century travel writers and encyclopedias.
The documents tell a much smaller and more uncertain story. Giulia was never charged, tried or punished. She died in Rome in 1651, apparently in her bed. Everything the archives say about her comes from an investigation that began eight years later, in 1659, into a network of women who were making and selling poison in Rome. In that investigation witnesses and prisoners named her as the woman who had brought the recipe from Sicily and taught it to her stepdaughter, Gironima Spana. On that evidence she was the maker and seller of a poison later used by others. No court found, and no surviving document shows, that she killed anyone herself.
Her story comes in three layers: what the archival record shows, what later writers claimed, and what modern historians conclude.
On the evidence, Giulia Tofana was not a serial killer. No killing by her own hand is established, she was never charged, and the deaths investigated in 1659 happened after she died. Her fame rests on her name, attached in hundreds of popular retellings to the claim that she was one of the deadliest women in history, and the distance between that claim and what the record actually supports is the heart of her story.
The surviving 17th-century record has three parts. The first is the criminal file of the 1659 investigation, kept in the State Archives of Rome (Archivio di Stato di Roma). It was found and first described by the Italian historian Alessandro Ademollo in a short book of 1881, and it was studied in full by the American historian Craig A. Monson, whose book The Black Widows of the Eternal City (University of Michigan Press, 2020) is built on a notary's transcript of about 1,450 pages. The Italian legal historian Simona Feci has since worked through more than 700 documents on the investigations for her book L'acquetta di Giulia (Viella, 2024). The second part is the account written in the 1660s by Cardinal Sforza Pallavicino in his life of Pope Alexander VII. The third is a group of chronicles, diaries and street ballads from the time.
The modern research summarized by National Geographic in July 2026 places Giulia's birth in Coriglione (the old form of Corleone), inland from Palermo, rather than in Palermo itself, as the older tradition had it. Her death record, made in Rome in 1651, gives her age as about 70, which puts her birth around 1581. Her married name, according to Monson and Feci, was Giulia Mangiardi.
She was married twice. After the death of her first husband she married again, and her second husband had children of his own. One of those stepchildren was Gironima Spana, who would become the central figure of the 1659 case. An 1881 account of the trial file records Spana's own statement that her parents had died in Sicily and that she was Giulia's stepdaughter, not her daughter.
In Sicily and later in Rome, Giulia worked in the borderland between medicine, beauty and folk magic. National Geographic, drawing on Monson, Feci and the historian Mike Dash, describes her as a wise woman, healer and marriage broker who offered fortune-telling, cosmetics, skin creams and love potions, and who arranged marriages and property deals for well-off clients. Monson notes that marriage brokers of the period were often suspected of witchcraft, and that she had that reputation. Feci and others think she probably inherited the poison recipe and sold it only now and then.
The family moved from Sicily to Rome in 1624, with Giulia, her second husband and his daughters traveling together. The reason is not certain. National Geographic reports that the move followed the sudden death of a man, and that his maid claimed Giulia might have been involved.
The trial file of 1659, as summarized in 1881 by the Sicilian scholar Salvatore Salomone-Marino from Ademollo's book, gives more detail. It says Giulia had to leave Palermo with her stepdaughter after it was discovered that a rich Genoese gentleman, Ippolito Larcari, had died of a poison she had prepared and that a man named Spadafora had given him. This is the only death in any source that ties Giulia to a specific victim. It rests on testimony given in Rome some 35 years after the event, about a death in another city, and no charge or trial in Sicily is known. It is an allegation in the record, not a finding.
Salomone-Marino, writing in 1881, believed the flight from Palermo happened around 1633. The modern archival work dates the move to 1624, and another member of the later network, Maria Spinola, was said in the same file to have been living in Rome since 1627. The 1633 date was Salomone-Marino's belief; 1624 rests on the archival work.
In Rome, according to the trial file as summarized in 1881, Giulia went on working as a marriage broker and was said to "understand physiognomy," that is, to read character from faces. She formed connections with people of high rank. One of the women hanged in 1659, Giovanna De Grandis, a Roman laundress who removed stains for a living, had been married off by Giulia to her third husband, a Sicilian named Franchi.
The same summary says Giulia had a working relationship with a cleric known as Father Girolamo of Sant'Agnese, who lived at San Lorenzo fuori le Mura and supplied the arsenic that apothecaries would not have sold to her. National Geographic, drawing on the modern research, likewise describes "a renegade priest" who supplied the arsenic and two go-betweens who sold the poison. The 1881 account also says Maria Spinola, a Sicilian nicknamed Grifola who had come to Rome in 1627, appears to have been in Giulia's confidence, although she became active in the trade only under Spana.
Giulia's second husband, according to the modern research, had once been prosperous but lost the family's money. National Geographic says Spana's family fortune was squandered by her stepfather. Washington University's summary of Monson's book adds that Spana was orphaned, left Sicily as a girl of nine, saw her stepfather waste her inheritance, and was married at 14 to a husband who spent freely and was unfaithful.
Witnesses in 1659 described how the liquid was made, and their words were printed in 1881. One of the main accused women said it was made with arsenic and lead, boiled in a new, tightly sealed pot until the water dropped by a finger's width. The water left behind was clear. Taken in wine or soup, she said, it caused vomiting and then fever, and five or six drops a day would cause death in 15 to 20 days without changing the taste of the food. A maidservant described watching her mistress grind antimony and add filed lead and a piece of crystalline arsenic before sealing the pot with dough and boiling it for about an hour. The court tested the seized liquid, and a version made up by its own experts from these instructions, on dogs.
The flasks were sold under several names. The 1659 poison was called Acquetta (little water), Acqua di Perugia, Acqua di Palermo and, on some bottles, Manna of Saint Nicholas, after a liquid said to flow from the saint's tomb at Bari. Modern writers stress that the mixture was ordinary arsenic with other metals, not a mysterious or undetectable substance. A 2015 article in Revista Clínica Española gives the same three ingredients, arsenic, lead oxide and antimony, as the usual description.
Giulia died in Rome in 1651. Her death record, according to National Geographic, gives her age as 70 and points to natural causes. Salomone-Marino, reading the trial file in 1881, wrote that she seems to have died around 1651, probably in her bed, and he dismissed stories of her death in a convent or under torture in prison as unconfirmed. No one appears to have suspected her of anything at the time. Her trade, the testimony says, passed to Gironima Spana, together with the help of Father Girolamo.
Spana, by her own account a widow of a Florentine (his name is given as "Carrozzi" in the 1881 summary of the file), lived in a large furnished house on the Via della Lungara, near the Tiber. She was known in Rome as an astrologer and seer. A report written at the time, quoted in 1881, says she was believed to find lost things and foretell events, that ladies of the nobility sought her out, and that she kept a servant, a manservant and two sons at their studies. Cardinal Pallavicino, in his life of Alexander VII, called her "a certain Girolama Siciliana," clever and attractive, who had drawn in many important people with her fortune-telling and had more than once been called before the Holy Office of the Inquisition and talked her way free.
Monson's book describes the buyers as ordinary Roman women in unhappy or dangerous marriages: two daughters of an aging linen seller; the wife of a jealous dyer; a woman who said she had come across the poison only while looking into her own son's death; a 20-year-old mother of two on her second marriage; the wife of a violent butcher; and, among the better-off, a duchess and a wealthy heiress. Feci has described them as "very normal, ordinary women," the wives of butchers, linen sellers and dyers. In a review of Monson's book, the historian Trevor Dean of the University of Roehampton notes that the suspicious deaths in the file fall between 1655 and 1659, beginning with the husbands of two sisters and their aunt in 1655. Their symptoms were fever, severe stomach pain, vomiting and great thirst, and none of the doctors who treated them seems to have suspected poison. Rome had just come through the plague of 1656, and Mike Dash has suggested that some deaths blamed on poison may have been disease.
Pallavicino's account says the authorities first learned of the poison through a woman who had used it on her husband. Months later she confessed to a priest, was told she might be obliged to reveal the danger to the authorities, and offered to tell everything in return for immunity, which she received.
To obtain the poison itself as evidence, the court set a trap. The wife of one of the criminal court's clerks was dressed as a noblewoman and placed in a borrowed palace furnished for the purpose, with court officers dressed as footmen. A go-between told one of the sellers, an old woman named Giovanna, that a great lady wanted help with an unbearable husband and would pay well. Giovanna came, was convinced by the scene, produced the flask and explained how to use it. When the "lady" asked whether her husband could be saved if he reformed during his illness, Giovanna told her the antidote. Officers then came out from behind the bed curtains and arrested her with the flask in her hand. The sealed liquid was given to two animals; the one left untreated died, and the one given the antidote recovered.
National Geographic, on the modern research, identifies the seller caught in the trap as Giovanna De Grandis, who offered to obtain poison for an invented client in a sting organized by the city's lieutenant governor. Dean's review names that official as Lieutenant Governor Stefano Bracchi. After her arrest and interrogation, De Grandis confessed, and her statements led to the arrest of further sellers and buyers. Pallavicino calls Giovanna a Sicilian, while the 1881 summary of the trial file calls De Grandis a Roman; the two sources disagree on that point.
According to Pallavicino, a woman who lived with Spana as a servant, though she was really her sister, showed the court where the flasks were hidden. The maidservant's testimony about the recipe, quoted above, appears to come from the same household. National Geographic reports that Spana's longtime housemaid testified against her.
Pallavicino writes that five women were finally held as the principal makers and sellers. Four of them, after denying it for many days, confessed when faced with the weight of the evidence. Spana alone denied everything to the end. The judges nevertheless held her guilty, and she was condemned to the ordinary penalty with the others. Dean's review says Spana met Bracchi repeatedly and denied or turned aside every question and every piece of testimony against her. National Geographic, likewise, says she denied involvement to the end.
To obtain information about the other sellers, Pallavicino says, the court promised pardons to many of the women who had bought and used the poison. Two noblewomen who had allegedly obtained it, Dean notes, were not pursued by the investigator. The Folger Shakespeare Library's account of the case says noblewomen implicated in it were pardoned by the pope. Rome's elite, in other words, was largely kept out of the courtroom, and the burden of punishment fell on poor and working women.
On Saturday, July 5, 1659, five women were taken from the Tor di Nona prison through the streets of Rome on carts and hanged together in the Campo de' Fiori. They were Gironima Spana, Giovanna De Grandis, Maria Spinola, Graziosa Farina and Laura Crispolti (the name is also spelled Crispoldi). The five names, the date and the place are recorded in a ballad printed soon after, which Salomone-Marino found matched the trial file in every particular he could check, and the five women and the date appear in all modern accounts. In that ballad Spinola went up the ladder first, then De Grandis, then Farina, then Spana, with Crispolti last.
Pallavicino explains why the gallows was put in the Campo de' Fiori rather than at the usual place by the Ponte Sant'Angelo: the square was busier, and the authorities wanted the largest possible crowd. Despite the extreme heat of the season and the hour, he says, houses emptied, and a single shop window overlooking the square rented that day for four gold doubloons, about as much as a year's lodging. He also records that Spana had hidden a small knife in her sleeve, meaning to take her own life before the executioner could, and that she gave it up to a member of the confraternity that comforted the condemned. On the ride to the gallows she at times cried out that she was innocent, that the witnesses had lied and that the judge was unjust, but she died, he writes, as a Christian.
Monson's modern reconstruction of the scene draws on the records of the confraternity that attended the condemned, according to Dean's review. A Roman gossip, quoted by Monson, remarked right after the hangings that the recent plague had given the women cover, and that their poison might therefore have killed "five hundred husbands." That remark, a guess made in the street on the day of the executions, is one of the earliest sources of the large numbers that later attached to Giulia's name.
The investigation did not stop with the five hangings. Dean's review describes a follow-up inquiry that produced another confession and another death sentence, which was commuted to exile, and a group of women who were incriminated but never convicted. Monson's publisher states that only six women were hanged in all, and National Geographic reports that only one of the buyers went to the gallows, two received papal immunity for testifying, and the others were banished, with three of the exiles later having their sentences remitted. These summaries do not fully agree on the number of executions after July 1659; the point is noted below as unresolved.
Pallavicino records that the executions were followed by a strict edict. Making, keeping or selling poison without a written license from the magistrates was forbidden on pain of the severest penalties, and each license had to state the substance, the quantity and the names of the people involved. Physicians were ordered to report any suspicion of poisoning in their patients, and the symptoms of the poison and its antidote were published. Washington University's summary of Monson's book adds that the popes kept the transcript locked away for centuries so that the recipe would not spread.
Giulia herself appears in none of this as a defendant. She had been dead for eight years. Her name entered the file because the women on trial, and their servants, described where the recipe had come from.
After 1659 the story passed to the writers, and what they added is claim, not record, each piece with its own author and date. None of it is supported by the 1659 record.
The earliest detailed account, Pallavicino's, never names the poison and never mentions Giulia. He calls the ringleader only "Girolama Siciliana." The Folger Shakespeare Library notes that the women hanged in 1659 never called their product Aqua Tofana. In the ballad printed after the executions, the liquid is "the water of Palermo," which "the vile Tofania used to make," a reference to a Palermo poisoner rather than to Giulia. The name "Aqua Tofana" spread through later chroniclers. The Folger points to the German historian Johann Wilhelm von Archenholz, whose England und Italien (1785) named the poison and described it as a weapon an old husband could not guard against.
A separate set of Sicilian records concerns poisoners in Palermo a generation earlier. The diary of the Palermo notary Baldassare Zamparrone, quoted by Salomone-Marino in 1881, records the execution in February 1633 of Francesca La Sarda as the maker of a poison of which a single drop killed within three days, and on July 12, 1633, the execution of Teofania di Adamo (also written Thofania d'Adamo) for giving poisoned water to several people who died of it. Later Sicilian writers credited Teofania with inventing the poison, and in Sicilian speech "la gnura Tufania" became a byword for a woman suspected of poisoning. Mike Dash's chapter in the academic volume Toxicology in the Middle Ages and Renaissance (2017) gives La Sarda's execution as February 1632; the 1881 transcription gives 1633.
The idea that Giulia was Teofania di Adamo's daughter, and that "Tofana" was a family name taken from her mother, can be traced to Salomone-Marino's article of 1881. He reasoned that Giulia held the secret of the poison, that she left Palermo (by his dating) around the time Teofania was executed, and that in Sicily children were often given a parent's unusual first name as a surname. He then proposed that Giulia was Teofania's daughter and heir. He also wrote, plainly, that he had no documents to give a sure answer and that the matter could not be settled without the Palermo trial file.
Modern research has not found that link. National Geographic reports that Feci and other historians see no evidence that the two women were related, and Monson treats both the surname "Tofana" and the mother story as later inventions. The name the archives give her is Mangiardi.
The number most often attached to Giulia in modern retellings is 600, usually as a claim that she confessed to poisoning 600 men between about 1633 and 1651, or that her product killed that many. The documents do not support it. The 1659 record concerns a few dozen buyers and a smaller number of deaths, all after 1655. The large figures seem to have grown in stages: the Roman gossip of 1659 who guessed "five hundred husbands"; travel writers of the early 18th century who wrote of a Naples poisoner who had killed "many hundred people"; and 19th- and 20th-century compilers who joined these to Giulia's name. National Geographic reports that historians dismiss the claim of hundreds as exaggerated and that no one knows how many men died.
The most persistent legend gives Giulia a different end. In this version she kept selling poison for decades, took refuge in a convent, was seized, confessed under torture and was executed, in some tellings in 1709, in others years later, with her body thrown over the convent wall. National Geographic summarizes the legend as ending with her either retiring to a convent or being captured and executed, depending on the version.
That story belongs to a different woman, or women, in Naples in the early 1700s. The German traveler Johann Georg Keysler, writing from Naples in letters later printed in English (London, 1758), reported that "Tophana, the noted female poisoner, from whom the Agua Tophana took its name," was still alive in prison there, that visitors went to see her out of curiosity, that she was a small old woman who had belonged to a religious sisterhood, and that her life had been spared for that reason, "tho' she sent many hundred people out of the world." Keysler's Naples prisoner was alive around 1730, nearly 80 years after Giulia's death in Rome, and was not executed by his account. Whoever she was, she cannot have been the woman who died in Rome in 1651. Salomone-Marino, in 1881, already set aside the stories of Giulia dying in a convent or under torture in prison as unconfirmed. No known document connects Giulia to any event in 1709.
The poison's name gained its widest audience through Mozart. The first biography of the composer, by Franz Xaver Niemetschek (Prague, 1798), written with help from his widow Constanze, describes a drive the couple took in the Prater in Vienna after Mozart returned from Prague in the autumn of 1791. Already ill and working on the Requiem, he spoke of death and told her he was writing the Requiem for himself, and that he felt certain someone had given him poison. Niemetschek does not name the poison.
The name appears in a later account. In 1829 the English musician Vincent Novello and his wife Mary visited Constanze in Salzburg. According to their travel diaries, published in 1955, she told them that about six months before his death Mozart had become convinced that someone had given him "acqua toffana," and that she had considered it an absurd idea and tried to talk him out of it. Medical writers have since rejected poisoning as the cause of Mozart's death; a 2015 review in Revista Clínica Española called the theory clearly untenable, and favored kidney disease. Mozart's remark tells us only that, by the late 18th century, "Acqua Tofana" was a name a frightened man in Vienna could reach for. It says nothing about Giulia.
The 19th century turned the story into a genre. Ademollo's 1881 book, I misteri dell'Acqua Tofana, was the first to use the Roman trial file, and he suggested that the recipe had passed from Naples to Sicily and then to Rome through Giulia, giving the poison her name. Salomone-Marino disagreed in the same year, placing the origin in Palermo with Teofania di Adamo and proposing the mother-daughter link discussed above. Popular histories, encyclopedias and sensational novels repeated and expanded both ideas. The Folger Library notes that over the centuries the story was steadily embellished, and Monson's publisher describes earlier accounts as sensational and often colored by ideas of witchcraft.
Three bodies of modern work matter most.
Craig A. Monson, Paul Tietjens Professor Emeritus of Music at Washington University in St. Louis and the author of earlier archival studies of convent life in early modern Italy, published The Black Widows of the Eternal City with the University of Michigan Press in September 2020. The book is the first full-length study of the 1659 investigation and is based on the newly recovered notary's transcript of about 1,450 pages, along with birth, marriage and death records and decisions of the governor's court. Monson argues that Spana, the figure at the center of the legend, may never have poisoned anyone herself, while dozens of women who did use the poison escaped the gallows. He identifies the woman known as Giulia Tofana as Giulia Mangiardi, places her move to Rome in 1624, and treats her surname and her supposed descent from Teofania di Adamo as later inventions. Reviewing the book, Trevor Dean praised its use of the archive and its picture of the buyers as women in unhappy or abusive marriages, while cautioning that Monson writes for general readers, recasts indirect testimony as direct speech, and sometimes imputes thoughts to his characters that the records do not contain.
Simona Feci, a legal historian at the University of Naples L'Orientale, published L'acquetta di Giulia: Mogli avvelenatrici e mariti violenti nella Roma del Seicento (Giulia's little water: poisoning wives and violent husbands in 17th-century Rome) with Viella in December 2024. Her publisher describes the case as one of the very few all-female criminal plots in history and the most sensational crime story of its decade. National Geographic quotes her calling Giulia "a mask" rather than a real person, by which she means that the figure in the legend has little to do with the woman in the archive.
Mike Dash, a historian whose chapter on Aqua Tofana appears in Toxicology in the Middle Ages and Renaissance, edited by Philip Wexler (Academic Press, 2017), written before Monson's book appeared, reached a similar view from the printed sources. He traced the poison to Palermo around 1630, treated Giulia as a shadowy figure who probably died unsuspected around 1651, and called the claimed link to Teofania di Adamo unproven. National Geographic quotes him describing a "criminal magical underworld" of apothecaries, astrologers and dubious clerics, in which the women "probably didn't start out as poisoners at all."
Taken together, the modern work agrees on these points. Giulia Mangiardi was a real Sicilian woman who lived in Rome from 1624 and died there in 1651. Testimony given in 1659 named her as the source of the recipe and as someone who sold it. She was never accused in her lifetime. Her stepdaughter and four other women were hanged eight years after her death. The surname, the Palermo parentage, the death toll and the execution story were added afterward.
The usual test for a serial killer, based on the FBI's definition, requires two or more victims killed in separate events. Giulia Tofana does not meet it on the evidence, for four reasons. No killing by her own hand is established in any document. The only death tied to her by name, that of Ippolito Larcari in Palermo, rests on secondhand testimony given decades later and was never charged. She was never arrested, charged or tried, and the investigation that named her began eight years after she died. The deaths examined in 1659 were attributed to other women who bought or sold the poison. Her name lives on because it is so widely attached to claims of mass killing, not because the record bears them out.
Her name. Archives: Giulia Mangiardi (Monson; Feci). Tradition: Giulia Tofana. The 1881 summary of the trial file calls her "Giulia Tofana di Palermo," but it is not clear whether that wording is the file's own or the 19th-century writer's.
Her birthplace. Coriglione (modern research as reported in 2026) versus Palermo (older tradition and the 1881 summary).
The move to Rome. 1624 (modern archival work) versus about 1633 (Salomone-Marino's 1881 inference).
The Larcari death. An allegation in the 1659 file, about a death in Palermo years earlier. Not tested in any court.
Relation to Teofania di Adamo. Proposed in 1881 without documents; rejected or unproven in modern scholarship.
Number of victims. No figure for Giulia is supported. "600" has no traceable documentary basis; "five hundred" was a Roman gossip's guess in 1659; "many hundred" comes from a traveler's description of a different woman in Naples around 1730.
Her death. Natural death in Rome in 1651 (death record, as reported by National Geographic) versus legends of capture, torture and execution in 1659, 1709 or later. The legends are not supported.
Executions after July 5, 1659. Five women were hanged that day. Monson's publisher says six were hanged in all, and a widely repeated list names a sixth woman hanged in March 1660, but Dean's review of the same book describes a later death sentence commuted to exile. The total is left as five on July 5, 1659, with one possible later execution.
Spana's confession. Pallavicino and National Geographic say she never confessed. Some popular summaries say she confessed in June 1659; no strong source for that was found.
The poison's makeup. Testimony describes arsenic, lead and antimony; later writers added belladonna and other ingredients without documentary support.
About 1581: Giulia (later called Mangiardi, and in legend Tofana) is born in Sicily, at Coriglione according to modern research.
1624: Giulia, her second husband and his children, including Gironima Spana, move from Sicily to Rome, reportedly after the sudden death of a man in Palermo.
1627: Maria Spinola, a Sicilian later hanged in the case, is recorded as living in Rome.
February 1633 (or 1632): Francesca La Sarda is executed in Palermo as a maker of poison.
July 12, 1633: Teofania di Adamo is executed in Palermo for poisoning. No documented link to Giulia.
1624 to 1651: Giulia lives in Rome; testimony of 1659 says she made and sold the poison during these years (start date not established).
1651: Giulia dies in Rome, recorded as about 70, apparently of natural causes and unsuspected. The trade passes to Spana.
1655 to 1659: A series of husbands die in Rome with similar symptoms.
1656: Plague in Rome.
Early 1659: A sting operation leads to the arrest of Giovanna De Grandis; the investigation under Lieutenant Governor Stefano Bracchi follows.
1659: Spana and other sellers are arrested; buyers testify, many under promises of pardon.
July 5, 1659: Gironima Spana, Giovanna De Grandis, Maria Spinola, Graziosa Farina and Laura Crispolti are hanged in the Campo de' Fiori.
1659 to 1660: Follow-up proceedings; other women are exiled, pardoned or released. A strict edict regulates the sale of poisons in Rome.
1660s: Cardinal Sforza Pallavicino writes his account of the case in his life of Alexander VII, naming neither Giulia nor the poison.
About 1730: Johann Georg Keysler describes an old woman called "Tophana" held in prison in Naples.
1785: Johann Wilhelm von Archenholz names Aqua Tofana in England und Italien.
1791: In Vienna, Mozart tells his wife he believes he has been poisoned; in 1829 she tells visitors he named "acqua toffana."
1798: Niemetschek's biography of Mozart records the poisoning remark without naming the poison.
1881: Alessandro Ademollo publishes the first study of the Roman trial file; Salvatore Salomone-Marino proposes that Giulia was Teofania di Adamo's daughter.
September 2020: Craig A. Monson publishes The Black Widows of the Eternal City.
December 2024: Simona Feci publishes L'acquetta di Giulia.
July 2026: National Geographic publishes a review of the legend and the modern research.