Ancient Engineering · Bavaria, Germany & Upper Austria (shown: Ratgöbluckn erdstall, Perg, Austria)

The Erdstall Tunnels

Hundreds of claustrophobic medieval tunnel systems under Bavarian and Austrian farmland — dug with enormous effort, abandoned clean, and explained by nobody.

Mainstream: c. AD 950–1300 (radiocarbon-dated to the High Middle Ages)Alternative: c. 3000 BC or earlier (Heinrich Kusch's Neolithic claim) — inflated online into a '12,000-year-old tunnel network spanning Europe'48.25°, 14.63°

Engineering Lost Worlds

At a glance

The Erdstall Tunnels — Photo: Pfeifferfranz · CC BY-SA 3.0 AT

Scattered beneath farms, villages and churchyards across Bavaria and Austria — with cousins in France, where they are called souterrains — lie well over a thousand recorded erdstall tunnel systems, around 700 in Bavaria and 500 in Austria alone. A typical erdstall is a cramped network of low, narrow galleries, often barely a metre high, linked by 'Schlupf' squeeze-holes as tight as 40 centimetres that an adult can only pass by exhaling and wriggling through. Systems usually have a single concealed entrance, no second exit, and end blind. The tunnels were clearly dug deliberately and skilfully, with carved benches, lamp niches and neatly rounded vaults. Yet they are archaeologically almost sterile: excavations recover no tools, no storage residues, no burials, no rubbish — at most a few charcoal flecks and potsherds apparently swept in from above. Whatever people did in them, they did it without leaving anything behind, and no medieval document describes their construction or purpose. Local folklore filled the gap centuries ago: the tunnels are 'Schrazellöcher' (goblin holes), dug by dwarves, or penances imposed on sinners. Modern research has narrowed the when — but the why remains one of Central Europe's genuine archaeological open questions.

See it on the globe →

The mainstream view

What archaeology says

Systematic study is led by the Arbeitskreis für Erdstallforschung (Working Group for Erdstall Research), founded in 1973 in Roding, Bavaria, by Karl Schwarzfischer, which publishes the journal Der Erdstall. The dating question is now essentially settled: radiocarbon determinations on charcoal from tunnel floors and wall niches across multiple sites cluster consistently between the tenth and thirteenth centuries AD, with several samples between 950 and 1100. Erdstalls are High Medieval structures, contemporary with the villages above them.

Purpose is where mainstream researchers openly disagree. The refuge school, championed by Upper Austrian archivist Josef Weichenberger, sees them as short-term hiding places for villagers during the raids and feuds of the tenth to twelfth centuries — Hungarian incursions especially. Weichenberger famously tested the idea by spending 48 hours inside an erdstall; it is survivable, and the squeeze-holes would stop an armoured pursuer. Critics reply that a single-entrance tunnel is a death trap against smoke, that many erdstalls sit awkwardly far from dwellings, and that refuges elsewhere (like Kent's deneholes) show occupation debris that erdstalls lack.

The ritual school, associated with researchers in the Arbeitskreis such as Dieter Ahlborn, argues the deliberate narrowness, dead ends and sterility mark them as spiritual spaces — perhaps linked to folk beliefs about souls awaiting judgement, half-remembered pre-Christian ideas carried into the medieval countryside. It explains the cleanliness and the impractical design, but rests on no direct evidence at all. The storage hypothesis — cool larders — founders on damp, poor access and the total absence of food residues. The honest mainstream summary, offered by the Arbeitskreis itself: the tunnels are medieval, and their purpose is unknown.

Key evidence cited

Radiocarbon dates on charcoal from multiple erdstalls clustering between the tenth and thirteenth centuries AD
Iron-pick tool marks consistent with High Medieval mining technique
Distribution tracking medieval villages, churches and farmsteads rather than prehistoric sites
Lamp niches, benches and squeeze-hole design indicating purposeful medieval construction
Fifty years of systematic documentation by the Arbeitskreis für Erdstallforschung (founded 1973)

The alternative view

What the skeptics propose

The best-known alternative case comes from Heinrich Kusch, a speleologist and prehistorian formerly associated with the University of Graz, who with his wife Ingrid Kusch published Tore zur Unterwelt ('Gates to the Underworld', 2009). Kusch argues that many erdstall-type passages in Styria are far older than the Middle Ages — Neolithic, perhaps 5,000 years old or more — citing passages sealed beneath undisturbed layers he considers ancient, stone-dressing marks he attributes to pre-metal tools, and megalithic-looking entrance stones. In his reading, the medieval radiocarbon dates only show when tunnels were last cleaned or reused, not when they were first cut; the true builders belong to a forgotten prehistoric tradition of going underground.

The internet then did what the internet does: Kusch's regional claim was inflated into viral articles asserting a 12,000-year-old 'superhighway' of tunnels running from Scotland to Turkey, supposedly proving an Ice Age civilisation's escape network. Kusch himself never demonstrated any such connected network — erdstalls are short, isolated, blind-ended systems, physically incapable of linking anywhere to anywhere — and archaeological reviewers note that no tunnel of the type has ever produced a stratified prehistoric artefact.

The rebuttal to the core claim is straightforward but worth stating fairly: charcoal from construction niches (not later fills) dates to the tenth to thirteenth centuries; tool marks match iron picks; and erdstalls cluster around medieval settlement patterns, not prehistoric ones. Sceptics such as the researchers around the Arbeitskreis conclude Kusch has mistaken reused natural caves and undatable dressing marks for antiquity. Still, his challenge had one healthy effect: it pushed erdstall research to obtain the radiocarbon series that now anchors the field — a rare case where a fringe theory directly improved the mainstream evidence base.

Key evidence cited

Kusch's claimed sealed, stratigraphically old passages and pre-metal stone-working marks in Styria
Megalithic-appearing entrance stones at some Austrian sites
The argument that radiocarbon dates only fix later reuse and cleaning, not original excavation
The near-total artefact sterility, read as evidence the medieval dates reflect secondary use
Folklore of dwarf-dug 'Schrazellöcher', cited as folk memory of far older builders

Genuinely open questions

What were erdstalls actually for — no proposed purpose survives contact with all the evidence?
Why are the tunnels so archaeologically sterile when medieval refuges elsewhere are full of debris?
Why did construction stop, apparently abruptly, around the thirteenth century — and why does no medieval text mention them?

Worth knowing

Some erdstall squeeze-holes are so tight — down to about 40 centimetres — that researchers must exhale completely and corkscrew through; refuge-theory advocate Josef Weichenberger once spent 48 hours underground in one to prove hiding in an erdstall was at least survivable.