Curupira

The Curupira has a first appearance that can be dated, and it does not say what this page used to claim it said. It is a Latin letter written by the Jesuit José de Anchieta from São Vicente at the end of May 1560 — the letter gives no day — and printed for the first time in Lisbon in 1812. In the original it speaks of "quosdam Daemones, quos Brasiles vocant Corupira": certain demons, in the plural, a class of beings rather than a character, who set upon the Indians in the woods, flog them and kill them. Anchieta gives them no feet, no hair and no forest to protect, and the two sentences that follow — the Jesuits who claim to have seen the dead, the feathers and arrows left as an offering on a hilltop — are almost always cut. The backwards feet arrive three centuries later. This page follows the word from that letter through Vellozo's dictionary of 1795, Bates in 1863, Couto de Magalhães in 1876, Barbosa Rodrigues in 1890 and Stradelli in 1929, to the cryptozoological encyclopedia that in 2002 turned a forest demon into a surviving Pleistocene monkey.

The Curupira has a first appearance that can be dated to the month, and almost nothing that is said about it today is in that first appearance. The document is a letter by the Jesuit José de Anchieta, and the first thing to say about it is that it is not in Portuguese. Anchieta wrote it in Latin, and it stayed in manuscript for more than two centuries: it was printed for the first time in Lisbon in 1812, by the Academia Real das Sciencias, as number III of the first volume of the Collecção de noticias para a historia e geografia das nações ultramarinas, under the title Epistola quamplurimarum rerum naturalium, quae S. Vincentii (nunc S. Pauli) provinciam incolunt, sistens descriptionem. It is a naturalist's letter — ants, manatees, anteaters, thunderstorms, stones — written to a superior who had asked for whatever was remarkable or unknown to the Old World. The Corupira turns up at the very end, in the paragraph before the one on deformities, and it reads, in the 1812 printing (long s and ligatures normalised): "Notum est, et in omnium ore versatur esse quosdam Daemones, quos Brasiles vocant Corupira, qui saepe in silvis adoriuntur Indos, flagris caedunt, macerant, et necant." That is: it is known, and it is in everyone's mouth, that there are certain demons, whom the Brasis call Corupira, who often set upon the Indians in the woods, flog them, batter them and kill them. Four things in that sentence deserve to be read slowly. The word is Corupira, with an o, not curupira. It is plural — quosdam Daemones, a class of beings, not a character with a name. Anchieta gives it no body at all: no feet, no hair, no eyes, no height. And it does not protect anything; it kills. The letter is meticulous about anatomy elsewhere — the tapir's three-toed feet, the monkeys' hands, a bird whose feet sit so far back that it cannot hold its own body up — so the silence about the Corupira's shape is the silence of a man who had nothing to describe. The sentence almost never travels alone in the original, and almost always does online. Two more follow it. First: "Hujus rei fratres nostri, qui aliquoties ab eis interfectos viderunt, testes sunt" — our own brothers are witnesses, having several times seen people killed by them. Then the ritual: on a path that runs through rough forest and steep hills into the interior, at the summit of the highest hill, the Indians leave bird feathers, fans, arrows and similar things by way of an offering, "Corupira ne sibi noceant summopere deprecantes" — begging the Corupiras, in the plural again, not to harm them. Anchieta files them alongside two other classes of being: the Igpupiara, "aquam incolentes", who drown people in the rivers, and the Baetata, "res ignis", a thing that is all fire and runs about the shore — the Boitata. The letter is signed "Exaratum Sancti Vincentii ... anno Domini 1560 sub finem mensis Maii": São Vicente, in the year 1560, at the end of the month of May. It gives no day. The 30 May and 31 May that circulate on the internet, incompatible with each other, are somebody's later precision, not the letter's. Between 1560 and the nineteenth century the word survives in vocabularies rather than in stories. Frei José Mariano da Conceição Vellozo's Diccionario Portuguez e Brasiliano (Lisbon, 1795) gives it four words: under "Diabo, Jurupari" comes the sub-entry "que appareça no mato, Curupira" — the devil that appears in the wood. Still no feet, no hair, no forest to defend. The animal arrives in 1863, and it arrives with a naturalist. Henry Walter Bates, eleven years on the Amazon collecting insects, wrote in The Naturalist on the River Amazons that "the Curupira is a mysterious being, whose attributes are uncertain, for they vary according to locality. Sometimes he is described as a kind of orang-otang, being covered with long, shaggy hair, and living in trees. At others he is said to have cloven feet, and a bright red face." Cloven feet — not feet turned backwards. A red face — not red hair. This is the earliest text we have located in which the Curupira is compared to an animal, and the comparison is made by the man who would later give his name to Batesian mimicry. The backwards feet, the forest guardianship and the punishment of loggers arrive together, thirteen years later, in Couto de Magalhães's O Selvagem (Rio de Janeiro, 1876). His words are worth having exactly, because our own file paraphrased them for years without naming the page: "O Curupira é o deus que protege as florestas. As tradições representam-n'o como um pequeno Tapuio, com os pés voltados para traz, e sem os orifícios necessários para as secreções indispensáveis à vida, pelo que a gente do Pará diz que elle é mussiço." And: "Todo aquelle que derriba, ou por qualquer modo estraga inutilmente as arvores, é punido por elle com a pena de errar tempos immensos pelos bosques, sem poder atinar com o caminho da casa." The attribution in our old text was, word for word, correct. What it did not say is that Couto was building a system: he files the Curupira as a god subordinate to Jacy, the moon, next to the Saci Cerere, the Mboitata and the Urutau. The pantheon is his. In 1890 João Barbosa Rodrigues published the Poranduba Amazonense, the great nineteenth-century collection of Amazonian narrative, and did something no summary since has bothered to repeat: he broke the creature back down into the places it came from. A small tapuyo four spans high at Santarém; bald at Santarém, covered in long hair on the Rio Negro; one-eyed on the Tapajós; legs without joints on the Rio Negro; solid and without an anus in Para; blue or green teeth and big ears on the Solimões; "e sempre com os pés voltados para traz". He also issued a warning that reads today like a diagnosis: "três mythos differentes, Korupira, Tatacy e Çacy ou Maty, têm sido confundidos sob a denominação de Kaapora". Three different myths, being fused under one name. And he mapped the figure outward: the Máguare in Venezuela, the Selvage in Colombia, "no Perú o Chudiachaque" — the chullachaqui — and the Kauá of the Cocamas in Bolivia. Stradelli's entry is the last of the classical ones and the richest. Ermano Stradelli died in 1926 and his Vocabularios da lingua geral portuguez-nheêngatú were published posthumously in 1929, in volume 158 of the Revista do Instituto Historico e Geographico Brasileiro. On page 434 he glosses: "Curupira — Corpo de menino — de curu abreviação de curumi e pira corpo", and describes "um menino de cabellos vermelhos, muito peludo por todo o corpo e com a particularidade de ter os pés virados para traz e ser privado de órgãos sexuaes". Then the hunting ethic our file quoted: "Sob a sua guarda directa está a caça, e é sempre propício ao caçador que se limita a matar conforme as próprias necessidades." Woe, he adds, to whoever kills for pleasure, or kills pregnant females, or destroys the young. That attribution too is literal and correct. Which leaves the question of how those three books ended up compressed into a single paragraph in our own file. There is a document, and it is datable. In 1933 the Civilização Brasileira published Anchieta's Cartas, informações, fragmentos históricos e sermões, edited by Afrânio Peixoto, with a faithful Portuguese translation of the 1560 letter — plural demons and offerings intact — and a footnote, number 164, hung on the word corupira. That note quotes Couto de Magalhães on the god of the forests and the backwards feet, then Stradelli on the game under his direct guard, then four incompatible etymologies in a row. Anchieta, then Couto, then Stradelli, in that order: it is the shape of the paragraph this page carried until today, with the etymological quarrel deleted and the plural quietly made singular.

There is no site to visit. The Curupira is not attached to a cave, a road or a lake, and any page that offers you coordinates for it — including, until this revision, this one — is offering you a guess. What can genuinely be visited is the paper trail, and all of it is free. The Latin original is the place to start, because almost nobody reads it. The Internet Archive holds the 1812 Lisbon volume (Collecção de noticias para a historia e geografia das nações ultramarinas, Tomo I, Núms. I, II e III, Typografia da Academia Real das Sciencias); Anchieta's letter is item number III and the Corupira paragraph falls on printed pages 177-178, with Lara Ordonez's note 85 on page 178 of the notes. The scan's OCR is poor — it renders the long s as f and Daemones as "Dxmones" — so read the page images rather than the text file. The Portuguese translation everyone quotes is in the 1933 Rio de Janeiro edition of the Cartas, informações, fragmentos históricos e sermões, edited by Afrânio Peixoto, digitised by the Biblioteca Nacional of Brazil and downloadable in full; note 164 is the one to read. For the nineteenth-century material: Couto de Magalhães's O Selvagem (1876) and Barbosa Rodrigues's Poranduba Amazonense (1890) are both on the Internet Archive with full text. Stradelli's 1929 Vocabularios are hosted as a scanned PDF by the Biblioteca Digital Curt Nimuendajú; the file has no text layer, so go straight to page 434 of the printed pagination (the Curupira entry runs to the top of page 435). Vellozo's 1795 Diccionario Portuguez e Brasiliano is likewise on the Internet Archive; the Curupira gloss sits under "Diabo". A note on what is closed, because it shapes what this page could check. Several of the archives that would normally be used for this kind of work are no longer usable: Chronicling America and HathiTrust were returning 403 to us in August 2026, and Trove now sits behind an anti-AI barrier. Nothing here depends on them, but a nineteenth-century Brazilian newspaper trail for the Curupira is, for the moment, not something we can offer. If you want geography rather than paper, the honest answer is that there are two, and they are 2,500 km apart. The word was first written down at São Vicente, on the coast of São Paulo state — Atlantic forest, roughly 24 degrees south, not Amazonia. The body, on the other hand, was collected in Amazonia in the nineteenth century, river by river: Santarém, the Rio Negro, the Tapajós, the Solimões, Para. Barbosa Rodrigues's equivalents extend the same figure into Venezuela, Colombia, Bolivia and the Peruvian montana, where it is the chullachaqui. Any single dot on a map of Brazil is a decision, not a fact. Finally, the painting on this page. It is O Curupira, oil on canvas, 96 x 157 cm, by the Brazilian modernist Manoel Santiago, dated 1926 — a genuine and well-identified work, and one of the few pre-war images of the figure. One caveat for anyone thinking of reusing it: the Wikimedia Commons file carries only a United States public-domain tag. Santiago died in 1987, so in Brazil and in the European Union the painting is still in copyright.