Bloody Mary

Bloody Mary is the ritual that has terrified sleepover guests and schoolyard dares across North America for generations. The rules are always whispered: go into a darkened bathroom, normally with the lights simply switched off rather than by candlelight, and chant her name — three times, thirteen times, forty-seven times, or until she answers; the count changes from telling to telling. More often than not it is done by a small group of girls, three to five of them, rather than alone. Then she comes. Some say her bloodied face appears in the glass; others that she reaches through to scratch the caller's face or drag her into the mirror's world. The names folklorists actually recorded in the field were Mary Worth and Mary Whales long before Bloody Mary, and the identification with Queen Mary I of England comes from print: it never turned up in their interviews. Alan Dundes read the game as a rite of passage rehearsing fears of blood and womanhood, an influential reading but a contested one. Bloody Mary needs no haunted house. She lives in every mirror.

Bloody Mary is one of the very few entries in this atlas that has been studied on the ground. Folklorists have sat in schools with the children who practise it, transcribed what they said, and deposited the texts in university archives where they can still be consulted. The record they built does not match the version that circulates online — starting with the name. One caveat before anything else. The foundational study — Janet Langlois, "'Mary Whales, I Believe in You': Myth and Ritual Subdued", Indiana Folklore 11, no. 1 (1978), pp. 5-33 — has never been digitised and we could not read it. Everything attributed to Langlois below comes from two independent scholarly readings that quote her with page numbers: Alan Dundes's 1998 article in Western Folklore, and Laura Winter's 2014 master's thesis in folklore at Memorial University of Newfoundland. Where the two differ, we say so. The name recorded in the field was not "Bloody Mary". Langlois did her fieldwork in the early 1970s at Holy Angels, an experimental Catholic elementary school for African-American children in northwest Indianapolis. She interviewed eighty students. Roughly twenty knew what they called "the Mary Whales thing", and about half of those had taken part in it; ten, by Winter's reading, were still doing it. The chant was not "Bloody Mary" but "Mary Whales, I believe in you". Twelve of her seventeen texts were collected in Indianapolis in 1973; the rest came from the Indiana University Folklore Archives, including one gathered in 1968 in Salem, Wisconsin, from several boys who did not know who "Mary Worth" was — only that she "comes leaping out of the mirror screaming and leashing out with her long, sharp fingernails and scratching the chanter's face to shreds". More than half the texts Langlois elicited herself were fused with a completely unrelated story: the Vanishing Hitchhiker. The first text in her article, from a twelve-year-old, described a girl who "stood on [the] corner when it rained and she had a long white dress on, and when someone stopped to give her a ride she would disappear in the back seat and just leave a wet spot with blood on the seat". The mirror ritual came attached to that legend, not to a queen. The names multiply, and none of them is a monarch. Winter lists what Langlois recorded: Mary Whales, Mary Lou, Mary Johnson, Mary Weathersby, Mary Worthington, Mary Worth. Add Mary Jane from the Knapps and Mary Wolf from Elizabeth Tucker's 1977 dissertation. Winter's summary of Langlois is blunt: "The origin of any of the names is not clear, nor is the origin of the accompanying narratives." The earliest printed description Dundes could point to is Mary and Herbert Knapp's children's-folklore anthology of 1976: a child goes into the bathroom alone at night, turns out the lights, stares into the mirror and repeats "Mary Worth" — "softly but distinctly, forty-seven times. She comes at you out of the mirror, with a knife in her hand and a wart on her nose." Jan Harold Brunvand, writing in 1986 about what he called "I Believe in Mary Worth", threw up his hands: "the precise origin of 'I Believe in Mary Worth' cannot be determined". Simon Bronner's American Children's Folklore (1988) catalogued them as "Mary Worth Rituals": "a girls' tradition common in elementary school" with the "atmosphere of the seance", the participants "huddled typically in a bathroom with the lights turned off", who "have to really 'believe' in her, or else she won't appear". In his notes Bronner remarks that "Bloody Mary" is "yet another name for variations of 'Mary Worth' rituals" — that is, in the scholarship of the 1980s the blood-name is the variant, not the original. The stories explaining who she was are as unstable as the names. Langlois collected a witch, sometimes one from the seventeenth-century Salem trials; a woman of African descent connected with a cult; a woman murdered by a jealous lover; and, in southern Wisconsin, a woman disfigured in an accident who tries to disfigure whoever summons her. Brunvand's later summary is the same shrug: "vague stories about some kind of tragedy suffered by the real Mary that disfigured her and made her determined to harm other young girls". Mary I of England appears nowhere in the field record. Winter found the identification in print, not in the mouths of children: The Daring Book for Girls (2007) offered Mary Tudor as the most likely historical explanation, and Winter notes flatly that the queen was "never mentioned by any of my informants". Dundes weighed the beheaded Mary of some versions against Mary, Queen of Scots and dropped it — "it is by no means certain that elementary school girls would necessarily be familiar with this historical figure" — mentioning Mary I only in a parenthesis, as the woman who had been "tarnished with the negative label". Elizabeth Báthory, endlessly repeated online, appears in neither Dundes's survey of the literature nor in Winter's 44,000-word thesis. It is not only small girls, and not only the United States. Dundes put the participants' ages at seven to twelve and noted that although a girl sometimes goes in alone, it is "much more often a matter of a small group of girls, three to five in number". But Tucker's 2005 Journal of American Folklore article worked from forty-two narratives collected over four years in a single residence hall at Binghamton University — university students, not children. The folklorist and play consultant Marc Armitage, running "play audits" in English schools, reported that in "around 65 per cent of the schools in the study group the story was well known amongst the child population", and that "very few adults within the schools knew of the existence of a toilet ghost character at their school". Winter's Catholic-school informants in St John's, Newfoundland, expected to see the Virgin Mary in the glass — an ecotype she found in no earlier publication — and her 2013 elementary-school informants had given Mary a brother, "Bloody Paul". On the Latin American thread we have to be honest: we found no academic documentation of a distinct Mexican or Chicano mirror-summoning ritual. What is documented is contact between traditions. Armitage explicitly compared his English toilet-ghost material with Langlois's findings and with La Llorona; one of Winter's Newfoundland students described Bloody Mary as a woman in white seen both in the mirror and walking by the river where she drowned. And Dundes's Californian texts came from Mexican-American, Chinese-American, Korean-American, Vietnamese-American and Irish-American girls — the ritual moves through multi-ethnic schoolyards rather than down a single national line.

There is nothing to visit. Bloody Mary has no house, no crossroads, no grave and no museum. She has no fixed geography at all: the point of the ritual is that it works in any dark bathroom, which is exactly why it spread. Any tour, attraction or address sold as "the real Bloody Mary" site is selling you something else. What does exist, and can be consulted, is the paper trail — the actual texts that folklorists collected from children, filed with dates, ages and places. Indiana University Folklore Archives (Bloomington, Indiana). This is where part of Langlois's material came from, including the 1968 Salem, Wisconsin text quoted above. The archive was founded in 1956 by Richard M. Dorson and grew to some 40,000 field collections gathered in Indiana, Michigan, Ohio and Kentucky before it closed in 1990; the surviving records and collections, dated 1807-2002 (bulk 1931-1982), amount to 70.4 cubic feet in 129 boxes and include subject files and index cards on beliefs, customs, games and legends. They are held by the IU University Archives, Herman B Wells Library E460, 1320 East Tenth Street, Bloomington, Indiana 47405-7000 ([email protected]). Consultation is in person and by prior arrangement: the finding aid states that select files are closed and that advance notice is required for access. Start with the finding aid on Archives Online and with the library's own research guide to the folklore collections, which also points to the Archives of Traditional Music and the Lilly Library; not all collections are processed, so if a box is not listed it is worth asking. UC Berkeley Folklore Archive (Berkeley, California). The archive Alan Dundes founded, and the destination of the seventy-five-plus Bloody Mary reports his undergraduates collected in 1996 — the ten printed in his article are a sample of what is filed there. It holds roughly a million items from more than two hundred countries and cultural groups, indexed by region and genre, and for each item it records the informant's name, age, occupation and ethnic background along with the circumstances of collection. It is at 110 Anthropology and Art Practice Building, University of California, Berkeley, CA 94720; access is "by appointment or by chance", arranged by writing to [email protected]. A digitisation programme is under way. USC Digital Folklore Archives — free, online, no appointment. If you want to read primary material tonight rather than fly to California, folklore.usc.edu is a publicly browsable database of items collected by USC students, each with informant details and a date. Its Bloody Mary entries include, for example, a version performed on 22 April 2011 by a thirteen-year-old student in Three Rivers, California, who learned it in the sixth grade from older classroom volunteers and reported that at her school it was a girls-only activity. This is the closest thing the legend has to a visitable site. And you can run the experiment yourself. Caputo stressed that his illusion "can be easily experienced and replicated as the details of the setting (in particular the room illumination) are not critical": a mirror, a weak lamp on the floor behind you, ten minutes of looking at your own eyes. Two practical notes. The published research uses adult volunteers who consented and were debriefed; the same paradigm has been used clinically to induce dissociation, so it is not a party trick to spring on a frightened child. And a fourth archive worth knowing about, for readers on the other side of the Atlantic: the Newfoundland texts Dundes cites, and the fieldwork behind Laura Winter's thesis, live in the Memorial University of Newfoundland Folklore and Language Archive. Winter's thesis, Dundes's article and the 2023 systematic review are all free to read online.