What we have corrected
A public record of our own errors: what an entry claimed, what is true, and the source that settles it.
Corrections, newest first
473 corrections published: 4 coordinates · 310 facts · 158 sources · 1 pattern
-
It said: The entry placed the attraction near the Gare Saint-Lazare and gave 2013 as its opening year.
It is: It was at 18 rue de Paradis, in the tenth arrondissement, more than two kilometres away and on a different metro line. And it opened in May 2011: the press covered it on the 5th and the 7th, and the first visitors went in on Friday the 6th. The entry was two years and two kilometres out.
Source: Le Parisien, 5 and 7 May 2011.
-
It said: The entry said Curioni named the species Lariosaurus balsami after Larius, the Latin name of the lake.
It is: Larius explains the genus, not the species. The epithet balsami honours Giuseppe Balsamo Crivelli, who described the specimen in 1839 and declined to name it. Two people, two halves of the name, and the entry gave both to one of them.
Source: Curioni’s and Balsamo Crivelli’s original descriptions.
-
It said: The entry had the man paddling at night by firefly light, and called the girl a drowned bride.
It is: In the poem the entry is built on, it is she who paddles the white canoe. And she is not drowned: the poem buries her — “they made her a grave”. Both details were reversed from the only source the entry had.
Source: Thomas Moore, “The Lake of the Dismal Swamp”.
-
It said: The entry described the kongamato as local tradition about dangerous water spirits.
It is: The magistrate who recorded it in 1923 states explicitly that the Kaonde did not consider it a supernatural being like the mulombe, but a dangerous animal. Turning an animal into a spirit is the opposite of what the source says, and it is the kind of change that makes a cryptid sound older than it is.
Source: Frank H. Melland, In Witch-Bound Africa (1923), ch. XIX.
-
It said: The entry dated the accident to 5 September 1951, and placed the boiler in the lobby with the boy playing by the sweet counter.
It is: It was Thursday 2 August 1951. The heater was under the stairs, and the boy was coming down from the balcony. The local press covered it at the time, photograph included, so every one of those details was checkable.
Source: Albuquerque Journal and Albuquerque Tribune, August 1951.
-
It said: The entry closed with a longtime Kaysville neighbour, in a local history interview, saying kids had dared each other to go up there after dark for as long as he could remember.
It is: The line exists nowhere. What makes this one worse is that real interviews with real neighbours do exist and are findable: there was no need to invent one, and inventing one puts a made-up sentence next to their real ones.
Source: The local oral history interviews that do exist.
-
It said: The entry treated the tower as having been the MGM Grand continuously.
It is: It was the MGM Grand from 1973 to 1986, and then Bally’s. Names on the Strip change hands, and an entry that skips the changes reads a single history into a building that had several.
Source: Property records of the tower’s successive operators.
-
It said: The entry gave a verse — “on the ninth stane they buried him then” — as a rhyme collected from local tradition.
It is: That verse does not exist. And the boiling of Lord Soulis is not folk tradition either: it comes from a literary ballad written by John Leyden and published by Walter Scott in the Minstrelsy. A known author is not oral tradition, and the difference is the whole point of the entry.
Source: Leyden, “Lord Soulis”, in Scott’s Minstrelsy of the Scottish Border.
-
It said: A hotel employee was quoted in a local magazine: “she’s supposed to be one of the more elegant ghosts in Washington”.
It is: The quotation appears in no locatable source. Its only echo is the advertising headline of a ghost tour company — which is where the phrasing came from, not from anyone at the hotel.
Source: Search of the locatable sources; the tour company’s own advertising.
-
It said: The entry described the burials as Congregationalist families.
It is: The transcription made for the local historical society says otherwise. It is a small thing to get wrong and an easy thing to check, which is exactly why it is worth writing down.
Source: Katharine A. Prichard’s transcription for the Mattatuck Historical Society.
-
It said: The Great Salt Lake was filed under the category “forest”, which the site printed on the page and on its listing card.
It is: It is a lake. The catalogue had no category for bodies of water at all — twenty types and not one of them wet — so one was added in the same change.
Source: The site’s own category list.
-
It said: The entry called the house a National Historic Landmark and described it as a municipal museum.
It is: It is not. It has been a New York City landmark since 1966 and on the National Register since 1975, which are different designations. The distinction matters because the entry was using the wrong one as a badge of importance.
Source: New York City Landmarks Preservation Commission; National Register of Historic Places.
-
Government House (Saskatchewan)
It said: The entry called the female figure by the main staircase “the housekeeper”.
It is: The only published sighting describes a woman in Victorian clothing seen through a window. The housekeeper is a label the entry added, and it is the kind of addition that turns an account into a character.
Source: Global News (2015).
-
It said: The entry said the Summer School of the Arts arrived after the sanatorium closed in 1972.
It is: The school was opened on 29 July 1967 by the provincial premier — five years before the sanatorium closed. For those five years the two things shared the site, which is a more interesting fact than the one the entry invented.
Source: Provincial records of the school’s opening.
-
It said: The entry opened by placing the house near where Grace Sherwood was tried by water in 1706.
It is: The house was built 124 years after that. Opening an entry with an event that predates the building by more than a century is not false in the details, but it is misleading in the frame: it lends the house a history that is not its own.
Source: The building’s own construction date.
-
It said: The entry gave three phenomena — children’s laughter in the great hall, footsteps in the minstrels’ gallery, a child in Tudor dress by the dais — as recorded by the heritage body that runs the palace.
It is: None of the three can be attributed to any source. The only traceable ghost account from Eltham is a newspaper piece of 27 June 2009, drawn from the body’s first inventory of incidents, and it describes something else entirely: a former employee who died a week after retiring.
Source: The Daily Telegraph, 27 June 2009.
-
It said: The entry placed the creature in a set of Chilean localities.
It is: Those localities return zero folkloric matches in the surveys of 1904-1913, and one of them returns zero appearances of any kind — while control searches in the same sweep return dozens or hundreds for the places that are genuinely attached to the being. The real localities are the ones given in chapter XI of the standard collection.
Source: Vicuña Cifuentes, chapter XI, pp. 38-39.
-
It said: The map pin sat at 25.0N, 137.0E — a round number copied from an encyclopedia, in open ocean.
It is: Everything this entry describes — the Myojin-sho eruption, the loss of the Kaiyo Maru No. 5 with 31 people aboard on 24 September 1952, the Bayonnaise Rocks — happens around 31°55'N, 140°01'E, in the Izu volcanic arc. The pin was about 1,000 km from its own subject.
Source: The events named in the entry itself.
-
It said: The entry made the ghost a nurse in period uniform walking the upper corridors.
It is: In all five published versions — across three newspapers between 2012 and 2025 — the ghost is the mother whose baby was taken from her, and the nurse is the living person who took it. The entry had swapped the victim for the official.
Source: Otago Daily Times (2012); Critic (2012, 2019, 2021, 2025).
-
It said: The entry said the legend drew scientific attention in 1893, when a huge carcass washed up at Do Son and French naturalists who examined it concluded it was an oarfish or a decomposed basking shark.
It is: The year is 1883, not 1893. And the rest does not hold either: the examination and the conclusion attributed to those naturalists are not in the record.
Source: Heuvelmans’s account of the stranding.
-
It said: The entry said salt was scattered to stop the flying head from rejoining its body.
It is: In the Chilean sources salt is what the chonchón comes to collect — “come back tomorrow for salt” — not what stops it. The remedies that are actually documented are turning the body face down, hiding or destroying the ointment, the cross of Solomon, the Magnificat and the Twelve Words Redoubled.
Source: Mapuche informants collected by Guevara (1908); El Obrero de Ovalle (1904).
-
It said: The whole description hung on a fabricated quotation: a nineteenth-century sentry who left his post after feeling a cold hand on his shoulder in the empty guardhouse.
It is: Neither the sentence nor the sentry appears in any corpus consulted, while positive controls on the castle’s documented ghosts return results in the same search. The entry was rebuilt around what is documented, starting with four soldiers on the morning of Saturday 23 April 1729.
Source: Full-text search with positive controls; castle records.
-
It said: An anonymous night watchman was quoted hearing “someone walking in shoes that don’t belong to anyone in the building”.
It is: Two full sweeps of the Internet Archive book corpus returned zero matches for the sentence and its variants — with positive controls in the same corpus returning thousands of results for the building itself. The search works; the quotation does not exist.
Source: Full-text search with positive controls.
-
It said: The entry called the keeper’s son adopted, had him die in a sawmill accident, and said the keeper was never found — contradicting itself one sentence later with a body recovered months afterwards. The Spanish version turned “24 miles” into 24 kilometres.
It is: The keeper is named in the US Lighthouse Service rolls, and the station log, the Marquette Mining Journal of 27 November 1902 and the federal case files tell a different sequence. The pin was also 200 m off the position the Coast Guard publishes.
Source: US Coast Guard Light List vol. VII (2025), light no. 14705; NARA RG 26; Mining Journal, 27 Nov 1902.
-
It said: The entry for Bermuda was not about Bermuda: it was about the Bermuda Triangle, and it contradicted our own entry on the Triangle.
It is: It also quoted islanders saying “Bermuda didn’t ask for the Triangle” — zero matches, against 132 and 891 for two control phrases in the same corpus — and gave fourteen men lost on Flight 19 by adding the crew of the rescue plane to the five Avengers. The entry has been rewritten to be about the islands.
Source: Naval History and Heritage Command on Flight 19.
-
It said: The entry named a patient, Jimmy Kennedy, as the source of one of the asylum’s ghost stories.
It is: His name and the asylum appear together zero times in the book corpus, with the search repeated and the same zero returned — while control searches in the same corpus on the same day returned some 22 books for the hospital’s own name. In the university research site devoted to the hospital, 37 articles deep, “Kennedy” appears once, and it is about a television programme.
Source: Full-text search with positive controls; Charles Sturt University research site.
-
It said: The entry placed the barmanou especially among the communities of Balochistan, and its map pin agreed.
It is: The fieldwork this all comes from was done in the Chitral district, in the Hindu Kush of northern Pakistan — more than a thousand kilometres away. Balochistan appears in no source. The coordinates pointed at Quetta, which is at least consistent with the error: the pin and the text were wrong together.
Source: Jordi Magraner’s field report (1992).
-
It said: The entry said Raymond Morris had restored the castle from 1971.
It is: David Maxwell bought the castle in 1971 and had the tower partly restored; he sold it to Raymond Morris in 1985.
Source: Historic Environment Scotland, listed building record LB16664
-
It said: The entry said the Green Lady was a young woman of the Sibbald or Leslie household, and that some accounts placed her after Culloden, in 1746.
It is: Every ghost book consulted identifies her as a Lundie, and no source connects Balgonie's ghost to Culloden. The one Jacobite-era episode attached to the castle is a raid by Rob Roy MacGregor in January 1716.
Source: Richard Jones, Haunted Castles of Britain and Ireland (2003); Martin Coventry, The Castles of Scotland (1997)
-
It said: The entry gave 1998 as the closing year, described J Ward as a wing of the same building, and cited unnamed “staff records” for a hanging.
It is: The hospital closed in 1993 and was dismantled in April 1994. J Ward is the former Ararat gaol, a different building two kilometres away, proclaimed an asylum ward in December 1886 and closed in January 1991. No record of the hanging has turned up anywhere. Two quotations were removed as fabricated.
Source: Public Record Office Victoria, VA 2841; Victorian Heritage Register.
-
It said: The entry followed, paragraph by paragraph, the structure and the gaps of an encyclopedia article, and closed with an unattributed sceptical formula about what “modern scholars” think.
It is: It was rebuilt against the primary source: the Florentine Codex, book XI, chapter IV, folios 70v-73v, read in Nahuatl, in Sahagún’s Spanish and in two scholarly translations.
Source: Florentine Codex, book XI; Anderson and Dibble; García Garagarza.
-
It said: The entry credited the naturalist Ivan Sanderson with a 1925 sighting of his own on Mount Salak, in Java.
It is: Sanderson was born in 1911: he was fourteen years old in 1925, and he was never in Java. His giant bat encounter happened in 1932, in the Assumbo mountains of Cameroon, and it was a different cryptid altogether. Two creatures had been welded into one. The Javanese witness on record is Ernst Bartels, who heard the cry in 1927 and published in 1962 and 1966.
Source: Sanderson, Animal Treasure (1937) and his 1972 book, both read in full; Bartels in Tong-Tong (1962) and Fate (1966).
-
It said: A shepherd’s account “collected in the 1970s” was quoted: it walked upright like a man but stooped forward, and when it whistled the dogs shook until dawn.
It is: It appears in no source. The detail about the dogs is the kind that makes a testimony feel real, which is why its absence matters.
Source: Full-text searches for the quotation.
-
It said: The entry said folklorists read the Tarasque as a memory of the Rhône’s floods or of pagan river cults.
It is: “Folklorists”, unnamed and collective, is an attribution to nobody. The standard study of the Tarasque exists, has an author and a date, and says something more specific than that.
Source: Louis Dumont, La Tarasque (Gallimard).
-
It said: A local rhyme was attributed to the Östersund festivities: “Storsjön is deep and Storsjön is wide, and something older than the town lives inside”.
It is: Zero results for its fragments in full-text search. The monster has a genuine nineteenth-century paper trail and a protection order; it did not need a rhyme.
Source: Full-text search of the Internet Archive book corpus.
-
It said: A lumberjack was quoted on catching one and finding only tears in the sack.
It is: The squonk has exactly one primary source: a short entry in a 1910 book of lumberwoods creatures, digitised and read here in full and compared sentence by sentence with our own. The quotation is not in it. When a creature has one source, checking is not hard.
Source: William T. Cox, Fearsome Creatures of the Lumberwoods (1910).
-
It said: An anonymous rancher was quoted, before investigators in the 1990s: “it looked at me like it knew exactly what I was thinking”.
It is: Full-text search returns two results, both of them unrelated novels. The line reads like testimony because it is written like testimony.
Source: Full-text search of the Internet Archive book corpus.
-
It said: A rancher’s account “kept in local press archives” was quoted, describing a wolf’s head, a hyena’s sloping back and jaws that could snap a bone.
It is: Phrase searches in the newspaper archive return nothing. The animal itself is real enough as a specimen; the account was not.
Source: Chronicling America (Library of Congress).
-
It said: A trapper interviewed by folklorists in the 1930s was quoted in Cajun French: “mon père disait que le rougarou marche la nuit du vendredi…”
It is: It appears in no web search and in no book of the full-text corpus. Writing an invented quotation in the community’s own language is not a flourish: it is what makes it convincing.
Source: Full-text search of the Internet Archive book corpus.
-
It said: A line was given as collected traditional narrative: “she left her body sleeping, and returned only in pieces before dawn”.
It is: It is not in the digitised primary source for the passage, read in full. The being is well attested; the sentence dressed as fieldwork was not.
Source: W. W. Skeat, Malay Magic (1900).
-
It said: The entry treated the Ozark Howler as long-standing regional folklore.
It is: The exact phrase returns zero results in the Library of Congress newspaper archive for 1800-1963, and zero for its variant. Full-text book search returns about fifteen hits, all in popular titles published after 2008. Whatever it is, it is recent.
Source: Chronicling America; Internet Archive full-text search.
-
It said: A Tohoku fisherman’s nineteenth-century warning was quoted: “if she asks you to hold her child, do not — the bundle grows heavier than stone and will not let you go”.
It is: Four different search strings, including the fisherman himself, return zero results each. Neither the man nor the line exists. The motif does — but it belongs to a different regional legend, which is where it has now been sent back.
Source: Full-text search of the Internet Archive book corpus.
-
It said: An “old saying” was quoted: “wash na yer feet in the sea by nicht, or the Nuckelavee will tak’ ye outright”.
It is: It is not in the collector who recorded the Nuckelavee, and it appears in no search. Written in dialect, it reads older than anything actually attested — which is what dialect spelling is doing there.
Source: Walter Traill Dennison’s Orkney collections.
-
It said: The entry quoted, inside testimony supposedly collected by British colonial officers, the line “look upon it and you will never sleep soundly again”, and reported an expedition finding.
It is: The quotation does not exist, and the expedition finding never happened. The region the entry named was not even a real administrative division.
Source: Full-text search of the Internet Archive book corpus.
-
It said: A muleteer’s warning “recorded in the nineteenth century” was quoted: “don’t cross where the water turns in circles, the fox’s tail waits down there”.
It is: It is in neither of the two standard Chilean sources for this being, nor in any search: the only hits are open wikis. The whirlpool belief is genuine; the muleteer was not.
Source: Guevara (1908); Lenz, Diccionario etimolójico (1905-1910).
-
It said: The entry quoted Martin Sheffield’s 1922 letter describing a “long-necked, swan-headed creature”.
It is: That wording is not from any 1922 source: it is the phrasing of a modern popular article about the case, put back into the letter it was summarising. The quotation marks turned a summary into testimony.
Source: Sheffield’s letter to Clemente Onelli (1922).
-
It said: A Falmouth saying of autumn 1976 was quoted: “when Morgawr comes to the bay, best keep the nets ashore that day”.
It is: It does not exist, searched whole and in fragments. The 1976 flap is real and well documented; the proverb attached to it was not.
Source: Searches of the 1976 press coverage.
-
It said: The entry quoted Doris Harrison, a real named witness, telling reporters the creature was “big, black and hairy, and it smelled just terrible, like something dead”.
It is: That wording appears in no source. She was fifteen at the time and what she did tell the St. Louis Post-Dispatch in 1972 — reprinted by the same paper in 2022 — is different, and specific. The entry had also demoted the chief of police to a municipal constable and turned a thirteen-centimetre mark into three-toed prints.
Source: St. Louis Post-Dispatch (1972, reprinted 2022); Associated Press (2012).
-
It said: The entry had Hichens reporting that witnesses described the fur as “grey, like that of a donkey”.
It is: In his actual text the donkey is a comparison of size, not colour: “a gigantic cat, striped like a tabby, but as large as a donkey”. A striped cat had been turned into a grey one by moving one word.
Source: Hichens, Discovery, December 1937, pp. 369-373.
-
It said: The entry began the minhocão’s written history in 1877, and quoted a nineteenth-century settler on ground opening “as though plowed by a monstrous plow”, with the stream running blood-coloured for two days.
It is: The quotation is in none of the primary sources. What the original actually says is that the river ran turbid for fourteen days. And the written history starts thirty years earlier: a note read to the French academy of sciences on 28 December 1846, reprinted in English in July 1847 — describing a different animal altogether, aquatic, black and short, that seizes horses and cattle by the belly.
Source: Fritz Müller (1877) and his letter to Darwin, 25 March 1877; Saint-Hilaire (1846).
Log updated on October 06, 2026.