#1190 1926 · Ethyl Gasoline Corporation (Graham Edgar) · Petroleum / automotive
Edgar rated the unmeasurable gasoline by which of two reference fuels it matched
the problem
Engine 'knock' wrecked pistons and power, yet nothing could measure it, so no one could compare fuel quality
background
As engine compression rose in the 1910s and 20s, gasoline increasingly detonated inside cylinders — a violent 'flicker' that cost power, overheated engines, and wore out pistons. Yet knock resistance was impossible to put a number on: you could only fill a car and drive it to learn if a fuel knocked. Refiners produced fuel by countless non-standard processes, and the only way to know if a gasoline was good was to test it in a running engine.
The result was an industry that argued about quality it could not measure. Petroleum refiners, carmakers, and buyers traded vague claims — 'runs smooth', 'premium' — with no common unit to compare fuels across suppliers. The classic solution, cataloguing gasoline's chemical composition, had failed to give a useful handle on knock, because knock resistance is not read off composition.
what everyone would do
The obvious route was to catalogue gasoline's chemical composition — figure out which molecules cause knock and specify fuels by their ingredients. That had been the research program for years, but knock resistance stubbornly does not reduce to composition: two fuels of indistinguishable make-up can knock differently in different engines, so any number anchored to ingredients failed as soon as it was compared across refiners, and the argument over quality went nowhere.
what they saw
Edgar stopped measuring gasoline and rated its behaviour instead. He weighed fuel against a ladder of two pure reference fuels — iso-octane at 100, n-heptane at 0 — and the blend it matched was its number.
the move
Graham Edgar, a chemist at Ethyl Gasoline Corporation (a General Motors / Standard Oil venture), anchored a scale to two pure compounds: iso-octane, which never knocked in any engine, was set at 100; n-heptane, which always knocked, at 0. By blending the two in different proportions he built a ladder of reference fuels, and assigned every gasoline the number of the iso-octane/n-heptane blend whose knock behavior it matched — the octane number.
why it works
The number is stable because behaviour is more repeatable than composition: a fuel that knocks identically to a 90/10 iso-octane/n-heptane blend is interchangeable with any other 90-octane fuel, whatever its recipe, so the figure survives being compared across suppliers, engines, and countries. Because the two anchors are pure, easily sourced chemicals, any laboratory can rebuild the ladder, which made the number reproducible and therefore trustworthy enough to be a lawful unit of trade. Once refiners and carmakers both quoted the same figure, the quality argument moved out of vague words and into a single enforceable number — and that is the step that let the industry refine against it, price it, and regulate it, instead of arguing forever in adjectives.
the payoff
The octane number became the global standard for fuel quality and is still on every pump a century later.
where it breaks
It fails whenever a single scalar cannot capture the thing being rated: one octane number hides the fact that a fuel can knock like a 90 blend under one set of conditions and a 95 under another — which is why the industry later split into RON and MON, and why any single 'quality index' overclaims when the behaviour varies with context. It also depends on having two extreme, reproducible reference points; where no such pure anchors exist, or they are proprietary and unavailable, the ladder cannot be built and the number cannot be trusted. And it presumes the affected parties will accept the same number as authoritative, which a market with vested interests in a different measure often will not.
what came after
Edgar's two-reference-fuels ladder outlived its creator and its company. The octane number remains the universal currency of fuel quality — contested, but in the numbers rather than in incommunicable claims — and its method of anchoring a scale to reference extremes is the template used again for cetane (diesel), and countless other quality indices since.
references
- [1]Octane NumberEngineering and Technology History Wiki (IEEE), 2026ethw.org
- [2]The Octane ScalePolymer Science Learning Center, University of Southern Mississippi, 2020pslc.ws
- [3]September 19, 1887 — The developer of the octane rating system is bornAutomotive History, 2022automotivehistory.org