# How Lighthouses Learned to Blink Their Names

My Weird Prompts — an AI-generated podcast. Daniel Rosehill sends in a prompt; AI hosts Corn and Herman discuss it. Verify important claims independently.

- Episode: 6653
- Published: 2026-11-17
- Duration: 18:25
- Audio: https://episodes.myweirdprompts.com/audio/lighthouse-flashing-patterns-navigation-20162.m4a
- Tags: history-of-technology, physics
- Web page: https://myweirdprompts.com/episode/lighthouse-flashing-patterns-navigation-20162/
- JSON: https://myweirdprompts.com/api/episodes/lighthouse-flashing-patterns-navigation-20162.json

## Summary

Every lighthouse flashes in its own distinct pattern, turning a simple beam into a readable name at sea. When coloured glass proved too dim—costing up to three-quarters of a lamp's intensity—engineers like Robert Stevenson and John Hopkinson turned to mechanical shutters and angled prisms to create timed darkness for free. From counting flashes with a stopwatch using Bowditch's rules to reading the Light List and IALA's modern grammar, this is the story of how mariners turned blinking lights into known positions.

## The prompt

> Every lighthouse along a coast flashes in its own pattern: one long flash every ten seconds, groups of three, a colour sector. How did that system come about, and how did a ship's navigator actually use those signatures to know exactly which light they were looking at before GPS?

## Show notes

Every lighthouse flashes a signature, but the blink was a latecomer: before it, every coastal light looked identical from seaward, from Tynemouth's coal fires to St Michael's Mount's oil bowl. The first fix was architectural—Wintertonness's 1618 double tower—but a system costing two towers per letter collapses once a coast has forty lights. Colour was the instinctive answer and a costly one: red glass delivers only about 40 percent of a lamp's intensity, green 25 percent, so the flash won on economics, since a shutter buys a new signature for free. John Hopkinson's 1874 prism panels multiplied flashes per rotation, and by the 1890s a navigator timed first flash to first flash, read the Light List's fixed grammar, and knew the light's identity—position came from chart and bearing, never brightness.

## Sources

- [)%20Lighthouse%20construction%20and%20illumination,%20Stevenson%20(1881).pdf](https://www.pharology.eu/resources/books/1) — Thomas Stevenson, *Lighthouse Construction and Illumination* (1881), retrieved 2026-11-17.
- [Trinity House, Portland Bill Lighthouse, retrieved 2026-11-17.](https://www.trinityhouse.co.uk/lighthouses-and-lightvessels/portland-bill-lighthouse)
- [United States Coast Guard, CG-193 (1949), via Distant Reader, retrieved 2026-11-17.](https://distantreader.org/stacks/trust/ufl1/ufl1_ark+=13960=t08w4h347.txt)
- [Wikipedia, Light characteristic, last edited 2026-09-23, retrieved 2026-11-17.](https://en.wikipedia.org/wiki/Light_characteristic)
- [Thomas Tag, Chronology of Lighthouse Events, United States Lighthouse Society, retrieved 2026-11-17.](https://uslhs.org/node/1606)
- [Encyclopaedia Britannica 1911, vol. 16, Lighthouse [Optical Apparatus], via Wikisource, retrieved 2026-11-17. [primary for the 1874](https://en.wikisource.org/wiki/Page:EB1911_-_Volume_16.djvu/656) — 1893 technical detail]
- [Recent Improvements in Lighthouse Illumination, *Nature*, 1895-11-01, retrieved 2026-11-17.](https://doi.org/10.1038/053056a0)
- [DNB 1901 supplement, Hopkinson, John (1849](https://en.wikisource.org/wiki/Dictionary_of_National_Biography%2C_1901_supplement/Hopkinson%2C_John_%281849-1898%29) — 1898), retrieved 2026-11-17.
- [Aids to Navigation, US Naval Institute *Proceedings*, Vol. 7/3/17, July 1881, retrieved 2026-11-17.](https://www.usni.org/magazines/proceedings/1881/july/aids-navigation)
- [CaptainsGround, Reading a Light's Characteristic: Light List, Period and Range, citing Bowditch Ch. 5 §§502](https://captainsground.com/lessons/light-list-and-light-characteristics) — 503, retrieved 2026-11-17.
- [ShipShape.wiki, Light descriptions and light ranges, retrieved 2026-11-17.](https://shipshape.wiki/articles/light-descriptions-light-ranges/)
- [The Fresnel Post, Why Some Beams Are Red or Green: Sector and Leading Lights, retrieved 2026-11-17.](https://crystalfern.xyz/articles/colored-sector-lights/)
- [Lighthouse Accommodation, Needles Lighthouse, Isle of Wight, retrieved 2026-11-17.](https://lighthouseaccommodation.co.uk/listings/needles-lighthouse/)
- [Cornish Rock Tors, Fact File: St Anthony Lighthouse, 2020-11-17, retrieved 2026-11-17.](https://www.cornishrocktors.com/fact-file-st-anthony-lighthouse/)
- [Victorian Web, Robert Stevenson (1772](https://www.victorianweb.org/technology/engineers/robertstevenson.html) — 1850), retrieved 2026-11-17.
- [Correspondence. Improvements in Dioptric Apparatus for Lighthouses, Institution of Civil Engineers, 1899, retrieved 2026-11-17.](https://doi.org/10.1680/imotp.1899.18975)

## Transcript

Corn: A dark coast, two in the morning, and a light blinking out at sea. That blink isn't decoration — it's the light telling you its name. Daniel wants to know where that system came from, and, better, how a navigator with no GPS actually read it. His phrasing is that every light "flashes in its own pattern" — one long flash every ten seconds, groups of three, a colour sector — and the question is how a blinking light became a known position.

Herman: And the answer starts somewhere unexpected, which is that the blink was a latecomer. Before any of that, every light on a coast looked exactly the same from seaward. A fire was a fire.

Corn: Coal fires at Tynemouth in the fifteen-forties, tallow candles at Eddystone, a stone bowl of oil and wick at St Michael's Mount going back to around fourteen hundred. All of it, from a ship, is one thing: a glow.

Herman: And the first fix wasn't optics at all — it was architecture. Wintertonness, sixteen eighteen, the first English double-tower light built specifically so it would look different from everything else. Two lights in line is unmistakable. It also costs you two complete towers, two keepers, two of everything.

Corn: A signalling system where the alphabet costs a building per letter.

Herman: It works beautifully — for a coast with four lights. Then the coast has forty and you're out of towers. There's a line from an eighteen eighty-one piece in the US Naval Institute Proceedings that I love: "On a coast so thickly lighted as our own, the question of distinguishing a light beyond mistake is a most important one, the means being fewer than may be supposed."


Herman: It's a polite Victorian way of saying we're running out of options. So the arc from there is: colour first, which turned out to be expensive. The flash second, which turned out to be free. And then the part nobody talks about — the man with the watch.

Corn: Start with the colour. Because that's the instinctive answer, right? Make one light red, one green.

Herman: Robert Stevenson thought so too — he began trials of coloured glass in eighteen-oh-nine, red glass chimneys, green glass sheets, the lot. And before eighteen twenty-two you did get Scottish lights distinguished partly by showing red and white alternately. But colour has a physics bill, and it's brutal. A red beam is only about forty percent the intensity of the same lamp burning white. Green is twenty-five percent.

Corn: So red costs you more than half your light and green costs three quarters. On a coastline, where the entire point is being seen from as far away as possible.

Herman: And this was an era when every candle-power was fought for. You didn't spend it on pigment. If you insisted on a coloured sector, the textbooks of the time say it should invariably be reinforced with extra mirrors or condensing prisms just to claw back to parity with white. You're buying back the light you threw away on colour.

Corn: Which is why colour survived as a specialist tool — marking particular dangers, bearings — rather than as the naming system. The flash is where it gets good.

Herman: Eighteen eleven. Stevenson's contribution there is beautifully mechanical: a vertically opening screen that intermittently obscures the lamp. That's it. You don't need a new light source to make a new signature, you need a shutter.

Corn: The light was never the hard part. The darkness was.

Herman: That's actually the whole discipline in one line. By eighteen twenty-five he's introducing proper flashing on the Scottish coast — five seconds of light, five seconds of dark. His son Thomas then spends the eighteen-forties improving the revolving apparatus with reflecting prisms above and below the lenses, so the whole height of the optic is doing work instead of just the middle band.

Corn: The family trade. Three generations of Stevensons, all making the dark more efficient.

Herman: Thomas laid out the logic in three stages in his eighteen eighty-one book, and it's almost a little history of civilisation in itself. First you mark the danger — any light will do. Second, you make the light seen from far off — now optics and burners matter. And only third, "when lights are multiplied on the same line of coast," do you need distinctions in character, so one is constantly in view and another waxes to a flash and wanes to extinction. Identity is a third-generation problem.

Corn: Which is a pattern we've hit before — nobody standardises until duplication forces it. Then the combinatorics. That's Hopkinson.

Herman: John Hopkinson, engineer at Chance Brothers in Birmingham, eighteen seventy-four. What he does is divide the lenses of a revolving light with panels of reflecting prisms set at an angle, so a single rotation of the apparatus gives you two, three, even four flashes instead of one. Same lamp, same rotation, same fuel — a completely different signature from a rearrangement of glass.

Corn: No new light. No new colour. Just geometry.

Herman: And the timing template is still recognisable today: half-second flashes, roughly two seconds between them, a long eclipse of ten seconds, whole period about fifteen seconds. The Casquets got the triple-flashing apparatus in eighteen seventy-seven. Pendeen in Cornwall got a quadruple. And Britannica says it flat out: "With the introduction of group-flashing characteristics the necessity for using colour as a means of distinction disappeared."

Corn: The flash beat the colour on economics. Colour rented space from the lamp; the flash got the signature for free.

Herman: And the failures are just as instructive. Wigham's gas system at Galley Head in eighteen seventy-eight raised and lowered the flame to make groups — of six or seven flashes. And Britannica's dry objection is that the uncertainty in the number of flashes in each group is found to be an objection.

Corn: A signature you can't count is not a signature. It's weather.

Herman: Then the Americans tried something gloriously literal. Captain Mahan's system had the light flash its own catalogue number. Minot's Ledge, off Boston, flashed one-four-three — its number in the register — repeating every thirty seconds. Exhibited at the Chicago Exhibition in eighteen ninety-three.

Corn: A lighthouse reciting its own serial number into the dark.

Herman: Every thirty seconds, one flash, four, three. And it lost, for the reason you'd expect: recognising a rhythm takes one cycle, counting digits takes longer, and you're doing it in a swell, at night, with the deck moving. Plus the mariner now has to memorise a national numbering scheme instead of reading a chart.

Corn: And a rhythm survives fatigue in a way arithmetic doesn't. A man who's been awake for ten hours can feel da-dum-dum. He cannot reliably count to four.

Herman: That's the constraint that shapes everything after — a character has to be countable under bad conditions. So by the eighteen-nineties you have this combinatorial signature space: flash or occult, quick or very quick, groups of one to four, periods from a few seconds to half a minute, colour held in reserve for bearings.

Corn: So the signature is not just a light, it is a sentence the coast speaks to you, if you have the book to read it.

Hilbert: Hilbert's Mid-Show Haiku. The Light List. Every coast keeps its own blink. A watch, a chart, and the dark.

Corn: And the book is the Light List.

Herman: Which is the question Daniel's really asking, I think — fine, the lights have names. What did a navigator actually do at two in the morning? The procedure is beautifully disciplined. You sight a light, and before you trust it, you time it. The rule from Bowditch: start the watch on the first flash of the group and stop it on the first flash of the next group. The full period, first flash to first flash. Not the gap.

Corn: Because the gap is where you fool yourself.

Herman: There's a classic error the training material warns about. You see a light marked in the list as two flashes then one, six-second period, red. You start your watch, count two flashes, the single flash comes, you stop. Three seconds. You conclude it's a different light, or that the list is wrong. You timed a fragment inside the cycle.

Corn: Three seconds of confidence in a six-second truth.

Herman: And the number in brackets trips people up too — in that notation the parentheses count flashes per group. Says nothing about the rate. The period is always stated separately, in seconds, after. It's a fixed grammar: characteristic, colour, period, height, nominal range, in that order. A navigator fluent in it reads a string like a sentence.

Corn: So the vocabulary — walk me through what a man had to hold in his head.

Herman: Fixed, F, steady light. Flashing, Fl, light shorter than dark. Occulting, Oc, the opposite — light longer than dark, a light that blinks darkness at you instead. Isophase, equal. Then quick, sixty to eighty flashes a minute, very quick, eighty to a hundred and sixty. There's even one that flashes Morse A — dot-dash — on safe-water marks, because if you're going to invent a language you may as well use one that already exists.

Corn: And the practical trap?

Herman: At distance, the dark bits register as the events. An occulting light — mostly on, briefly off — gets reported as flashing, because what your eye catches is the going-out. The vocabulary is precise and the eyeball isn't.

Corn: Here's the assumption I want killed, because I'd have made it. Bright light near, dim light far.

Herman: No. Brightness cannot give you distance — fog, haze, a dirty lens, the burner's condition that week, all of it changes what you see. What the light tells you is which light it is. Where you are comes from the chart and a bearing: the chart gives you the light's nominal range and position, the compass gives you a line, and the fix is the intersection of knowing and measuring.

Corn: The light says who. You work out where.

Herman: And the sectors are where the chart stops being optional. A sector light shows different colours in different arcs of bearing — and the thing everyone gets wrong is reading red as the edge of the channel.

Corn: It's not a wall.

Herman: It's a bearing across danger. Portland Bill's red sector covers the Shambles sandbank — while you hold the red, you're looking across that bank. The Needles is the great early case: a red light replaced the fixed white in March eighteen forty, explicitly to avoid confusion with St Catherine's up the coast, and a green sector came by eighteen eighty-four. That station now runs white sectors, red sectors, a green one — one tower speaking in dialects depending on where you stand.

Corn: Two boats can look at the same tower, see different colours, and both be right.

Herman: Which is why colour without the chart is incomplete information — the change from white to red is tempting to read as a universal command, and it isn't. It means something only at that light, on that bearing.

Corn: The modern end of all this?

Herman: IALA, the international association, has codified the whole grammar into numeric constraints — and every one of those numbers is a scar from the failures. Single flashing light: four to thirty flashes a minute, light duration under two seconds, and darkness at least three times the flash. Group flashing, period no more than twenty seconds. The buoyage system doesn't allow a fixed light at all — stated reason, lack of differentiation. A modern characteristic reads like a full sentence: group occulting, three, white, ten-second period, fifteen metres, ten-mile range.

Corn: Three times the flash — that's Wigham's ghost. Six-or-seven is banned by regulation now.

Herman: It's the countability rule written into law. And the maintenance regime is remarkable — the Light List updated against a weekly Notice to Mariners. A distributed database with a weekly patch, running on paper, worldwide, for over a century.

Hilbert: The flash beat the colour, no question.

Hilbert: Though the man you're all skipping is whoever decided the dark must be three times the light. Hopkinson got the prisms, that one got nothing. Herman, you said fifteen-second period — half a second on, two between, ten off. Corn took nine seconds to ask his next question just now, I was watching the clock. I've been timing lights for years, so it stands out.

Herman: Timing lights.

Hilbert: From car parks. Coastal viewpoints, after dark, stopwatch on the phone. You'd be surprised how many charted lights are not running to time. There's one I won't name — I've got the timing diagram on the back of an envelope here, pencil, to scale — where the eclipse inside the group runs too close to the eclipse between groups. That's a design fault. Anybody with a watch and patience can see it. I keep a notebook of discrepancies. Fourteen entries. Most of them, if I'm honest, are me — cold hands, phone latency. But not all of them.

Corn: You're filing bug reports against lighthouses.

Hilbert: Somebody has to. The keepers used to log their own revolution times every watch, you know. When that stopped, the lights stopped being checked and nobody noticed. I dropped the notebook in a puddle at Sennen last month. Dried it on the dashboard. Page twelve is gone.

Corn: Page twelve had the good one, I assume.

Hilbert: It had the only one I was sure of. Corn, that cough of yours, we're cutting it. Where was I. The envelope, the envelope's fine, it was in my coat. The darts team thinks I'm mad, Clive especially, he waited in the car with the heater on while I did forty minutes on a headland in February. But he read the chart for me, so he's invested now.

Corn: Clive from the darts team, lighthouse verification unit.

Herman: The thing I like about the envelope diagram is that it's exactly what Thomas Stevenson would have drawn. That's the actual working method of this whole field — a light's character is a shape in time, and the only way to argue about it is to draw the shape.

Corn: One thing still unresolved, before we wrap. Nobody has actually settled where the red sector came from.

Herman: No, and I looked — you get station dates, the Needles in eighteen forty, St Anthony early in the last century, but no inventor, no documented first instance. The best-attested driver is the boring one: avoiding confusion with the light next door. Not danger-marking at all.

Corn: The safety feature may have started as a naming collision fix. That's unresolved and I'd send Daniel down that hole happily. The larger thought for me is the patch discipline. A weekly correction cycle, a rule that no two neighbours may speak alike, held together by consensus across every maritime nation for a century — no single authority ever owned it, and it never fell over.

Herman: If our software ran like the Light List, we'd have nothing to complain about.

Corn: True. If today's question blinked in your direction, the reviews are where you flag it — takes a second, and it keeps this coast lit. Producer credit to Hilbert Flumingtop, who has an envelope we will never see. This has been My Weird Prompts, find us at my weird prompts dot com. See you tomorrow, Herman.

Herman: See you tomorrow.
