On a rocky hillside near Çıralı in southern Turkey, flames emerge through openings in the ground. Yanartaş, often associated with the ancient Chimaera, belongs among historical natural mysteries because the phenomenon is real while its explanations have changed. Ancient writers described an enduring fire; a nineteenth-century naval surveyor recorded both a burning vent and local beliefs about it; modern investigators have measured the gas and debated how much formed without biological organic matter. The strongest evidence concerns combustible gas escaping through rock, rather than a supernatural fire.
The ancient report and its limits
In Natural History, Book 2, section 236, Pliny the Elder describes Mount Chimaera in the territory of Phaselis as burning by day and night. Citing Ctesias, he adds the claim that water increased its fire whereas earth or dung extinguished it. This is an ancient literary report, containing attributed information, rather than a surviving controlled experiment. It establishes a tradition about remarkable Lycian flames; it does not supply a continuous observation log linking every ancient vent to a particular modern opening.
The identification becomes explicit in Francis Beaufort’s travel account. His confidence that he had found Pliny’s locality is itself part of the history of interpretation. A historical place-name, a story about an unceasing flame and the modern gas field should not be treated as three independently proven descriptions of one unchanged burner. “Eternal” is a memorable name, but it is not a measured age for an individual flame.
What Beaufort actually saw
Beaufort’s Karamania, published in 1817, recounts his Admiralty survey of the coast in 1811–1812. On printed pages 44–47 he describes noticing a small steady light in the hills from his ship during the preceding night. Local inhabitants told him about the yanar and offered horses and guides. He therefore approached a phenomenon already familiar to residents; the account is not evidence that he discovered an unknown fire.
His party rode through partly cultivated ground and followed a wooded rocky glen. At the corner of a ruined building, he found a wall undermined by an opening approximately three feet across, shaped like an oven mouth. The flame gave intense heat but little visible smoke. He detached small pieces of caked soot, yet described the walls as scarcely discoloured. Plants grew nearby and a little stream ran down the hill. His description combines a dramatic fire with ordinary vegetation and water within the surrounding landscape.
He also noticed crumbly serpentine rock and loose limestone, and reported no volcanic products in the neighbourhood. Although his terminology included “volcanic flame” and “crater”, the observations did not amount to evidence of a volcanic eruption. Lower on the hillside he saw another opening that appeared to him to have carried a similar flame previously. That inference deserves to remain separate from his direct observation of the burning opening.
The guide told Beaufort that only one flame had existed within living memory, that its appearance had remained constant, and that it produced neither earthquake noises nor ejected stones. He said water could not quench it and shepherds used it to cook. Then came a different kind of assertion: the fire would not roast stolen meat. Beaufort recorded that belief, but supplied no test of it. It is evidence for an attributed local story about the flame, not evidence that the gas distinguished honest food from stolen food.
Nearby ruins included what Beaufort identified as a Christian church, with layers of plaster and painted inscriptions. These remains show why the locality attracted antiquarian attention alongside curiosity about the flame. The useful distinction in his account is between the light he saw, the opening he inspected, the geological materials he described and the statements his guide supplied. Combining them into a single miraculous eyewitness narrative would erase the account’s most informative details.
Measuring the fuel: two generations of gas studies
Hakan Hoşgörmez, Giuseppe Etiope and Mehmet Namık Yalçın’s 2008 study in Geofluids reported molecular and isotope analyses of methane, other hydrocarbons, hydrogen, carbon dioxide, helium and nitrogen. Their abstract describes a mixture of gas derived from heated organic material in sedimentary rocks and abiogenic gas associated with low-temperature alteration of the Tekirova ophiolite. It explicitly rejects mantle or magma degassing as the methane’s origin. The authors estimated the abiogenic contribution at about half the total gas. That is a published interpretation of analytical evidence, not simply an inference from seeing flames.
The distinction matters because “natural gas” does not identify one formation process. Thermogenic gas forms through heating of organic material. Abiotic gas forms through reactions not directly involving such organic material. Serpentinization is the alteration of certain iron- and magnesium-rich rocks by water, a process capable of generating hydrogen. Etiope and Barbara Sherwood Lollar’s 2013 review explains why recognising abiotic methane requires careful geochemical evidence, especially where different gas sources mix. An unusual flame colour or long history cannot perform that analysis.
In their 2011 Earth and Planetary Science Letters study, Etiope, Martin Schoell and Hoşgörmez reported direct methane and carbon-dioxide flux measurements at the gas field and nearby ophiolite. Their abstract describes vents across an approximately 5000-square-metre outcrop and estimates annual methane release of at least 150–190 tonnes. Combining molecular and isotope measurements with modelling of thermogenic gas formation, they suggested a dominant abiotic component of roughly 80–90 percent mixed with thermogenic gas.
That proportion is a source interpretation, distinct from the percentage of methane in the escaping gas. It also differs from the 2008 estimate. The changing figures should not be presented as one timeless established measurement. The 2011 authors identified another unresolved issue: sustaining the emission required effective continuing synthesis or a pressurised accumulation of gas. Their interpretation therefore explained important observations while leaving questions about generation and storage. For this case file, the 2011 research is represented through its publisher and author-repository abstract; the repository’s full PDF was not accessible.
A small water seep changes the investigation
D’Arcy Meyer-Dombard and colleagues visited in February 2012 and published their original microbiological investigation in 2015. Besides ignited gas vents, they found a small fluid seep not described in the earlier reports they cited. Water emerged beside a burning vent and divided into two shallow channels downstream. They sampled fluids, sediments and biofilms, measured chemistry and examined microorganisms through microscopy, culturing and genetic tests.
Water at the source had a measured pH of 11.95, falling to 9.4 at the sampled downstream location. Those values describe strongly alkaline water, not the temperature of the flame. The researchers interpreted the chemistry as consistent with serpentinization at depth and found evidence of microbial processes in the surface ecosystem. Their culture and gene results describe biological potential under sampled conditions; they do not establish that microbes produced all the methane feeding the fire. They considered the water seep probably intermittent. An enduring gas field can therefore contain a changing small-scale environment.
New measurements and a mechanism still being tested
Murat Şentürk’s 2026 Minerals paper reports four newly sampled outlets and compares their gas compositions and isotope patterns with published datasets. Its results section gives methane concentrations of 85.5–91.6 percent and hydrogen of 2.6–5.3 percent. Methane’s carbon-isotope composition differed markedly from that of ethane and propane, while the latter two were much closer to one another. Outlet-to-outlet differences were described without claiming that they alone proved separate origins.
The paper’s abstract treats the results as compatible with a multistage history, including thermogenic heavier hydrocarbons and possible modification involving hydrogen, fractures and ophiolitic rocks. Catalytic conversion of carbon dioxide is framed as a hypothesis requiring tests, rather than an established reaction operating throughout the natural field. The abstract and indexed results were accessible for this case; the complete publisher PDF was not. This qualification keeps a recent interpretation from being presented as a final solution.
What remains established, and what remains open
The historical and scientific records answer different questions. Beaufort supplies a detailed nineteenth-century observation and an attributed cooking belief. Pliny supplies an older written tradition. Gas analyses identify a combustible mixture and support explanations involving geological processes; field microbiology reveals life in a chemically unusual seep. None of these sources establishes the stolen-meat claim, an uninterrupted lifetime for one flame or an exact reconstruction of every reaction underground. Yanartaş remains compelling because repeated observations can coexist with a changing scientific explanation.
The directory pin uses the documented camera location of the accompanying 2021 photograph, rounded to four decimals, as a public hillside reference. It does not identify a laboratory sampling point or guarantee that a particular fissure remains ignited. Related reading: Costești’s trovants and living-stone stories, Other Mysteries and the Cases directory.