By Emilia Wellesley · Published May 5, 2026 · Updated May 6, 2026
On the morning of 29 May 1925, Colonel Percy Harrison Fawcett walked out of a place called Dead Horse Camp in the Brazilian Mato Grosso and into the green wall of the Amazon. With him were his eldest son Jack, twenty-one and freshly inoculated for an expedition his father had been planning since the close of the Great War, and Jack’s friend Raleigh Rimell, also twenty-one, already nursing an infected foot. Fawcett’s last dispatch ran in the British and American press through the summer; after that, the rumour file fattened and the men did not. A century on, the Royal Geographical Society in Kensington still keeps the field notebook he sent home before vanishing, and the question of what happened to him still has at least three plausible answers, none of them archived to the standard a historian would prefer.
Direct answer. Colonel Percy Fawcett, his son Jack, and Raleigh Rimell vanished in May 1925 while searching the Mato Grosso for a lost city Fawcett called “Z.” The Royal Geographical Society holds his pre-departure papers; later expeditions, two confessions, and Kalapalo oral testimony recorded by Orlando Villas Bôas in 1951 narrowed the search but never closed it. Killed, starved, or assimilated remain the three live hypotheses.
The Man Who Walked Into the Mato Grosso
Fawcett (1867–presumed 1925) trained as an artillery officer, surveyed the Bolivia–Brazil frontier for the Royal Geographical Society between 1906 and 1909, and returned to the Amazon basin five more times before his disappearance. The RGS gold medal he received in 1916 was earned the hard way: by walking transects through territory that had killed two of his rivals. He read maps the way a textual scholar reads a manuscript, treating cartographic blanks as sites of latent evidence rather than empty zones [1]. He had also read, by 1920, a manuscript catalogued as Manuscript 512 in the National Library of Rio de Janeiro, an anonymous bandeirante account from 1753 describing a ruined stone city found deep in the Bahian sertão. He believed it.
By 1924 he was sixty-five. His Bolivian survey work had earned him a reputation for surviving conditions that broke other men, and his correspondence with North London relatives (now held with his archive at the Royal Geographical Society) makes clear he saw the 1925 expedition as a final attempt rather than a routine survey. He named his target “Z” in letters to his backers because, he wrote to his wife Nina, naming it would commit him to finding it. The letter is dated 14 March 1925, and it reads less like a survey plan than a vow.
The 1925 Expedition: A Small Party by Design
Fawcett’s earlier guides and patrons had assumed he would go in force. He did not. The party that set out from Cuiabá in April 1925 was three people: Fawcett, Jack, and Rimell. Two Brazilian camaradas accompanied them as far as Bakairí Post, plus a small string of pack animals and two dogs. Fawcett’s reasoning was operational rather than romantic: a small party could be fed off the land, would alarm indigenous communities less, and could move faster through terrain that had stalled the larger Rondon expeditions of the previous decade [2]. The choice was also financial. The expedition was funded by a syndicate of newspapers, the North American Newspaper Alliance chief among them, and Fawcett had agreed to file a series of dispatches in lieu of cash advance.
The last of those dispatches, dated 29 May 1925 and posted from Dead Horse Camp at the headwaters of the Upper Xingu, was carried out of the bush by one of the camaradas. It described the party as well, the country as harder than expected, and Fawcett’s intent to push north-east into territory he had not previously surveyed. He instructed Nina that if no further word came, no rescue should be sent: “we shall be in country no white man has penetrated, and may be detained for a long time.” The letter is reproduced in his son Brian’s edited posthumous compilation Lost Trails, Lost Cities (1953), and in original at the RGS archive (Fawcett Papers, classmark RGS/PHF/1) [3].
What the RGS Archive Actually Holds
The Fawcett papers at the Royal Geographical Society include his 1906–1909 survey notebooks, a draft of the unfinished Exploration Fawcett manuscript, correspondence with Sir John Scott Keltie and Hugh Robert Mill, the original Manuscript 512 transcription Fawcett carried, and his final letter from Dead Horse Camp. They do not include any post-29 May 1925 communication. The historian of exploration David Grann, who consulted the archive while researching The Lost City of Z (2009), has noted that the gap in the archive is itself evidence: a man who wrote daily for two decades stops writing on 29 May, and the silence is sharper than any positive report [4].
The Recovery Missions: Eighteen Expeditions, One Hundred Dead
By the centenary of the disappearance, the literature counts at least thirteen named recovery expeditions and several hundred amateur attempts. The earliest official mission, led by George Miller Dyott in 1928 under RGS sponsorship, returned with a Kalapalo report that three white men had passed through their village and left east five years earlier; one of the men was tall, blue-eyed, and limped. Dyott concluded that Fawcett had been killed by the Kalapalo. He was wrong on the conclusion and partially right on the trace evidence; later research suggests the killing claim was confused with a separate incident.
In 1932 a Swiss trapper named Stefan Rattin emerged from the bush claiming to have met an old Englishman held captive by an Aricuna chief. The story collapsed under questioning. In 1951 Orlando Villas Bôas, the eldest of three brothers central to twentieth-century Brazilian indigenist work, announced that the Kalapalo had confessed to killing Fawcett’s party in 1925 and had produced bones he claimed were the colonel’s. London anthropologists examined the remains; the femoral measurements did not match Fawcett’s known height. The bones were returned to the Kalapalo and reburied in the late 1990s [5]. Estimates of the cumulative death toll across all recovery attempts run to roughly one hundred. The number is rough because several Brazilian indigenist accounts list deaths the foreign press never recorded.
- 1928, Dyott (RGS-sponsored): reached Bakairí Post; collected first Kalapalo testimony; concluded killing.
- 1934–35, Churchward: claimed contact via psychics; produced no usable evidence.
- 1951, Villas Bôas brothers: recovered alleged remains; bones later disqualified by femoral length.
- 1996, Brazilian–Anglo party: detained by Kalapalo; rescued by Brazilian air force; no new findings.
- 2005, Grann field trip: documentary follow-through; Kalapalo elders restated the 1925 visit but disputed the killing.
Kalapalo Testimony and the Limits of Translation
The most evidentiarily serious witness in this case is the Kalapalo people themselves, and their testimony has been collected, retranslated, and contested across four generations. The 1925 visit is not in dispute. Kalapalo elders interviewed in the 1950s, 1980s, and 2000s consistently describe three white men passing through the village in the dry season of that year, accepting food, and leaving east toward the country of the Suyá. The eastward bearing matters because the Suyá in 1925 were hostile to outsiders. Whether the Kalapalo killed the party themselves, watched them leave alive, or watched them leave alive and die elsewhere is the live disagreement, and the disagreement is partly an artefact of translation.
The anthropologist Ellen Basso, who recorded extensive Kalapalo narrative discourse in the 1980s, observed that Kalapalo testimony about external visitors is structured as a sequence of moral judgements about the visitor’s behaviour rather than as a Western-style chain of physical events; an interlocutor who failed to read the moral register would mistake one for the other [6]. Read in the moral register, the 1951 “confession” reads as a description of a party that had transgressed (camp etiquette, smoke fires, possibly a refusal to share goods) rather than a narrative of murder. The bones, in any case, did not fit. What the Kalapalo say with consistency, across decades and interviewers, is: “they were here; they went on; we do not know what happened beyond the river.” The historian who wishes to convert that into a positive account of death has to add evidence the village does not supply.
Three Live Hypotheses
The serviceable hypotheses are three, and a careful reader holds all three open. Killed by indigenous parties: pressed by Dyott in 1928 and partially by Villas Bôas in 1951; weakened by the femoral mismatch and by Basso’s reading of the testimony; not impossible if the party crossed into Suyá territory. Died of starvation, disease, or jaguar attack: Rimell’s foot infection was already serious at Dead Horse Camp; the Mato Grosso in late dry season is unforgiving; mortality among small parties of this period was high without external violence. Assimilated: a small but persistent thread, fed by the 1932 Rattin claim and by intermittent reports through the 1940s of a “white chief” living among unidentified communities. The assimilation hypothesis has no positive evidence that survives critical examination, but it has not been disproved, and Fawcett’s own correspondence shows a man drawn to that fate as much as repelled by it [7].
Why the Question Stays Open
In a case like this, the absence of a body is not by itself decisive; the Mato Grosso swallows bodies routinely. What is decisive, or would be, is a contemporary written record from someone who saw the party after Dead Horse Camp. None has surfaced. Manuscript 512, the bandeirante account that drew Fawcett east in the first place, has itself been re-examined in the last two decades and is now considered a probable composite, embellished by a copyist, and not a reliable map [8]. The city of “Z” Fawcett went looking for is increasingly understood, after Michael Heckenberger’s 2003 Xingu archaeological surveys, to have existed in a different form: not a single monumental capital but a network of fortified earthwork settlements supporting populations in the tens of thousands, abandoned by the early seventeenth century. Fawcett’s intuition about pre-Columbian Amazonian complexity has been substantially vindicated, even as his specific city remains undocumented.
What a Historian Will and Will Not Say
A century after Dead Horse Camp, the responsible historical position is the cautious one: Percy Fawcett, his son Jack, and Raleigh Rimell entered the Mato Grosso in late May 1925 and did not come out. They were last reliably observed by the Kalapalo within a few weeks of their last dispatch. Beyond that, the record holds three reconstructions that fit the surviving evidence and discriminate between them only weakly. The romance of the story, which Fawcett himself helped craft and which David Grann’s reportage has refurbished for a modern readership, is real but should not be confused with proof. What survived was the letter, the Kalapalo memory, and the negative space where a man who wrote daily fell silent. The discipline a case like this asks for is the discipline of reading that silence carefully and not filling it in.
