The most famous dream in the history of chemistry may never have happened. Friedrich August Kekulé told the story late in life, at a celebration held in Berlin on 11 March 1890 to mark the 25th anniversary of his benzene paper: a snake seizing its own tail, a fire, a reverie, and the sudden recognition that six carbon atoms could close into a ring. The story is charming, the ring is real, and the two things are not the same fact.
The molecule at the centre of the tale — benzene, C6H6 — had by then been puzzling chemists for four decades. Michael Faraday isolated it in 1825 from an oil-gas fluid produced by compression and supplied by the Portable Gas Company, reading his discovery paper to the Royal Society on 16 June of that year, as Chemistry World’s bicentenary account lays out in detail. He called it bi-carburet of hydrogen, a name that recorded what it was made of and said nothing about how it was put together. He did not know its shape. Nobody did.
A molecule that refused to behave
The problem with benzene was arithmetic. By the 1850s, chemists had settled on the idea that carbon formed four bonds. Hydrogen formed one. A saturated chain of six carbons should, by that logic, carry fourteen hydrogens — hexane, C6H14. Benzene had six. That is a staggering deficit of eight hydrogens for a molecule whose formula was, by then, well established.
Straight-chain drawings could not close the gap without piling double bond upon double bond upon triple bond, and the resulting sketches predicted reactivity that benzene refused to show. It did not decolourise bromine water the way an alkene should. It did not add hydrogen easily. It behaved like a saturated compound while wearing the formula of something wildly unsaturated. Chemists drew it as a chain and the chain lied.
The name itself came from Eilhard Mitscherlich in 1833, who distilled the compound from benzoic acid extracted from gum benzoin — a resin whose name traces back to Arabic alchemy. It was aromatic in the literal sense: it smelled of almonds, of the kernel at the heart of a peach stone. That almond note would later become the sensory signature of an entire class of chemistry.
The chain, the sausages, and the ring
Kekulé published his solution in 1865 in a paper titled On the constitution of aromatic substances. The date matters because the popular version of the story — dream, epiphany, ring — collapses nearly a decade of published argument into a single night of sleep. His first drawing was not even a hexagon. It looked, as Chemistry World’s Philip Ball notes in the same 200-year retrospective, more like a string of sausages: a chain of six carbons with the ends joined, alternating bonds, hydrogens hanging off the sides.
The clean hexagon that now decorates chemistry T-shirts came a year later, in 1866, first presented by Kekulé’s students. Adolph Claus produced the geometric stick form the same year. August Wilhelm Hofmann — the German chemist who had trained a generation of dye pioneers in London — made that representation popular. In 1872 Kekulé added the crucial refinement: two possible arrangements of alternating single and double bonds, interconverting so fast that experiment could not distinguish them.
He was almost right. It took Linus Pauling, drawing on Werner Heisenberg’s new quantum mechanics, to explain in the 1930s that benzene’s six carbon–carbon bonds are all identical, a resonance between the two Kekulé structures rather than a flip-flop between them. The electrons are smeared around the ring. The hexagon is real; the alternating double bonds are a bookkeeping fiction.
The dream that arrived 25 years late
Kekulé told the ouroboros story on 11 March 1890, at the Benzolfest convened by the German Chemical Society in the great hall of Berlin’s city hall to mark 25 years since his 1865 paper. Hofmann was there. Chemistry World’s bicentenary account also places Kaiser Wilhelm II among the guests, though historians of the Benzolfest have recorded the Kaiser and the dignitaries of the Reich as invited rather than present. Attendees were shown a “benzene tree”: coal and coal tar at the root and trunk, branches loaded with the colours the coal-tar industry had pulled out of them.
At that dinner, decades after the discovery, Kekulé described falling into a reverie by the fire, watching atoms dance, and seeing a snake bite its own tail. He also described an earlier reverie on the top deck of a London omnibus, in which the carbon chain first came to him; by his own account that one belongs to his London years, which puts it in 1855. Historians who read the fireside story literally place it in Ghent in the early 1860s, on the internal evidence of the speech alone. Nothing external fixes either. There are no contemporary letters, no lab notebook entries, no correspondence with colleagues from the 1850s or early 1860s that mentions either vision. The reveries appear only in retrospect, in a speech given by a famous chemist to a room full of admirers.
Why chemists failed for forty years
The reason the ring was hard to see is that nobody in the 1830s or 1840s was thinking about molecules as objects with a spatial arrangement. Chemistry was still largely a science of classification. Jöns Jacob Berzelius organised compounds by shared “radicals” — chunks of atoms that seemed to travel together through reactions. Structure, in the modern sense of atoms occupying specific positions relative to one another, was a fringe idea.
Kekulé’s larger contribution, more important than any one molecule, was to shift organic chemistry from classification to structure. Building on the English chemist Edward Frankland’s work on valency — without, as Ball notes in Chemistry World, always crediting him — Kekulé argued that carbon had four bonds and that those bonds could link carbon to carbon in chains of arbitrary length. Once you accepted the chain, the ring was a short step. But it required first believing that molecules had shapes at all.
The Dewar benzenes are a reminder that the hexagon was not the only skeleton on the table. In the late 1860s James Dewar, then at Cambridge, included the bicyclic structure that still carries his name in a list of possible C6H6 formulae — but he was never its advocate. As the American Chemical Society’s entry on Dewar benzene records, Dewar came down on the side of Kekulé’s structure. The molecule he lent his name to survived as a chemical curiosity rather than a rival: it was not made in the laboratory until 1962, nearly a century later, and it is strained enough to collapse back into ordinary benzene within a couple of days. Kekulé’s hexagon won on evidence. It predicted the right number of isomers for substituted benzenes.
What the ring unlocked
The commercial consequences were immediate and vast. Coal tar — the black, unappealing residue left behind when coal is baked to make gas — turned out to be dense with aromatic compounds. Charles Mansfield, working under Hofmann in London, first isolated benzene from coal tar in 1845 and by 1849 was producing it by industrial distillation. It became a solvent, in demand for the newly invented dry-cleaning trade, then a feedstock, then the raw material for an entire industry.
In 1856, nine years before Kekulé put the ring into print, an 18-year-old student of Hofmann’s named William Perkin was trying to synthesise quinine over the Easter holiday and produced instead a black residue that dyed silk brilliant purple. Perkin’s mauveine was the first commercially successful synthetic organic dye. It was made from aniline. Aniline was made from benzene, via nitration and reduction — a two-step aromatic route with no industrial precedent at the time, and exactly the sort of transformation Kekulé’s structural theory would soon make routine to describe.
In 1898, delivering the Kekulé Memorial Lecture in London, the British chemist Francis Japp said that three-quarters of modern organic chemistry was, directly or indirectly, the product of the benzene theory. The historian of chemistry Bill Brock, whom Ball quotes in the same piece, made the same point about direction of travel: after 1865, chemistry and the industry built on it were heading toward structure, with the hexagon as the sign over the door.
Intuition, evidence, and the myth chemists needed
The reason the ouroboros story survives despite thin documentary evidence is that it does useful cultural work. It makes chemistry heroic. It suggests that great structural insight arrives whole, from the unconscious, in a flash. It flatters the discipline’s self-image as a science of imagination as much as measurement.
The reality, as historians of science keep pointing out, is that Kekulé had spent years working on carbon chains, valency, and aromatic substitution before the hexagon appeared in print. The insight was earned through empirical grinding, not delivered by a serpent. Modern chemistry has its own version of the problem, and it is not flattering either. A Chemistry World essay on how chemists actually search a problem space describes a study of 69 scientists handed a simulated reaction with twelve variables to optimise. Most of them shrank the problem by fixing variables, then wandered the rest of it unsystematically, camping in high-reward regions and making longer and longer excursions as the returns fell away. The statistician Dennis Lendrem compares the pattern to the way animals forage across patchy resources. It is instinctive. It is also, measured against systematic experimental design, inefficient.
That question — how much of scientific discovery is pattern-recognition after long exposure, and how much is a genuine creative leap — has come back with new urgency in the age of machine learning. A Nature Research Custom Media feature produced for Mitsubishi Tanabe Pharma — partner content, not Nature editorial — describes researchers at the company using knowledge graphs and large language models, built on Elsevier’s SciBite semantic technology, to pull three-way relationships between drugs, diseases and genes out of the published literature and turn them into target hypotheses. Senior research scientist Kei Takedomi says in-house predictive models have cut the time chemists need to choose promising compounds by up to a third. The models find patterns humans would take years to notice. They do not dream.
Whether that kind of system could produce a Kekulé-style structural leap — a genuine reframing of what a molecule is, rather than an interpolation between known examples — remains open. What the benzene story suggests is that the hard part was never the drawing. It was the question underneath it: what if the chain has no ends?
The smell in the lab, still
Benzene itself is now known to be a human carcinogen, and undergraduate labs long ago swapped it for safer aromatic solvents. But the ring is everywhere. It is in aspirin, in paracetamol, in polystyrene coffee cups, in the TNT in demolition charges, in the pigments of nearly every synthetic dye. Tens of millions of tonnes of benzene are produced worldwide each year, most of it now from petroleum reforming rather than coal tar, and almost all of it destined for further reactions that begin with the hexagon and end in something useful or toxic or both.
The Long View has previously traced how helium was identified in the Sun before it was ever isolated on Earth — another 1860s puzzle where the evidence arrived before the substance could be handled. Benzene ran the opposite direction: the substance sat in Faraday’s crystallised distillate for four decades before anyone could say what shape it had.
The dream story will keep being told, because it is a good story and because chemistry teachers need something to say when a class of teenagers asks why a molecule is drawn as a hexagon. Somewhere in the middle of the tale, the actual work gets lost: the valency argument with Frankland, the sausage-string first draft, the students who cleaned it up into a hexagon, the isomer counting that settled the question, the resonance theory that arrived nearly seventy years later to explain why the double bonds were never really double.
The ring itself is unarguable. It sits in coal tar and in your bloodstream and on every organic chemistry textbook cover printed since 1900. Whether Kekulé really saw a snake by the fire in Ghent, or told a room full of Berlin chemists what they wanted to hear on a March evening in 1890, the hexagon he drew — or his students drew, or Claus drew — is the same hexagon still on the whiteboard this morning, in a first-year lab, being copied down by someone who has never heard the ouroboros story and does not need it to understand why the shape has to close.