Colonel Percy Harrison Fawcett was the Royal Geographical Society’s kind of instrument: an artillery officer turned boundary surveyor who, across seven expeditions from 1906, mapped the disputed borderlands of Bolivia and Brazil through country that killed his porters and pack animals with metronomic regularity — while Fawcett himself, lean, driven, and eerily immune, acquired a reputation for surviving what no one else did. Conan Doyle borrowed his lectures for The Lost World.
Somewhere in the years of river tables and boundary stones, the surveyor became a believer. Sifting colonial documents — above all a 1753 Portuguese account of a ruined stone city in the interior — and his own observations of pottery and causeway-like earthworks, Fawcett concluded the Amazon had once held a complex civilisation, whose principal ruin he designated, with deliberate blankness, “Z.” The scientific establishment held the opposite as doctrine: the rainforest’s poor soils could never have supported more than scattered tribes; Z was a fantasy.
1925
Bankrolled by newspapers that syndicated his dispatches worldwide, Fawcett set out in 1925 with a party built for stealth rather than siege: himself, his 21-year-old son Jack, and Jack’s friend Raleigh Rimell. On 29 May 1925, from a spot he called Dead Horse Camp in the upper Xingu, he sent his wife Nina a last dispatch ending: “You need have no fear of any failure.” The three then walked east into unmapped country and out of history.
What followed became its own catastrophe. Fawcett had left instructions that no rescue be attempted — the country was too dangerous — and was ignored for decades: expeditions, freelancers, and fantasists went in by the dozen, and the commonly repeated (unverifiable, plausible) accounting is that around a hundred searchers died or vanished in the hunt. Bones produced in 1951 were debunked by dental records; “white god” rumours, captivity tales, and confession stories from Kalapalo oral tradition — one version has the party killed after breaching etiquette with a local group — have been offered and contested ever since. No verified trace of the three has ever been found.
Z, vindicated sideways
The best twist is archaeological. From the 1990s, researchers — Michael Heckenberger working with the Kuikuro of the Xingu foremost — mapped exactly what doctrine had forbidden: networks of large pre-Columbian settlements, plazas, causeways, fish weirs, and enriched terra preta soils, supporting populations in the tens of thousands precisely in Fawcett’s search grounds; lidar has since multiplied such finds across the basin. There was no marble city of Z — but there was a garden civilisation, depopulated by European disease before outsiders arrived to see it, its earthworks swallowed by forest. Fawcett was wrong in the particulars, dead because of the wrongness, and — in the argument that mattered, against the empty-Amazon consensus of his century — closer to right than the people who called him mad.