The cat's eye patent does not claim the reflector or the wipe

Everybody who tells you about cat’s eyes tells you the same two things, in the same order. First the fog: Percy Shaw driving home over the moor above Halifax on a black night in the early thirties, the road edge invisible, and two green points of light ahead of him that turn out to be a cat sitting on a fence, facing him. Then the mechanism, which is the part people actually enjoy: the stud sits proud of the road, a wheel presses it flat into its iron shoe, and as it goes down the rubber drags the glass across a wiper, so every car that runs over one cleans it. A road marking that is maintained by being driven on. That is a genuinely lovely idea and it is why the object is famous.

The patent is dated 1934. So I went and read it.

What 436,290 actually says

Patent GB 436,290, filed 3 April 1934 as application No. 9943, class 3 (ii). The applicant is given, in the flat telegraphic style of the Patent Office abridgements, as SHAW, P., 1, Lane, Boothtown, Halifax, Yorkshire. Here is the whole of what it claims, and I am quoting it entire because the length is the point:

In a road-surface marking block of the kind comprising a metal holder with a resilient filling which normally projects above the road surface, the base of the filling is only partially supported so that the filling will yield by displacement and deformation or by either, when travelled over by a vehicle wheel, and sink to the level of the road surface.

Read that again looking for the reflector. It is not there. The word does not appear in the claim at all. It turns up once afterwards, in the description of what the drawing shows — “the rubber filling A is supported on a bar b extending across the holder B and is provided with reflectors C” — in the same breath, and with the same weight, as the bar it sits on. The rest of the abridgement is three more ways of holding the filling in place: spikes engaging lugs, a projection catching the bar, ledges for a wooden block.

The famous 1934 cat’s eye patent is a patent about a mounting. It is a claim on the idea that a lump in the road should get out of the way of the wheel and then come back.

And the wipe is not in it either.

The wipe has its own number

The self-cleaning action — the thing the story is actually about — is patent GB 457,536, filed 31 May 1935 as application No. 15826. Fourteen months later. Its abridgement opens by pointing back over its shoulder at the first one:

In road blocks for surface markings, such as those described in Specification 436,290, in which reflectors are provided on raised blocks which yield when travelled over by the wheel of a vehicle, means are provided whereby the reflectors are automatically cleaned during the movement of that part of the block in which the reflectors are mounted.

That is the invention everyone remembers, in its own words, and it is dated 1935. It exists as an improvement on 436,290, which is the Patent Office’s way of saying: the first one did not do this.

So the story as told welds together three things — a reflector, a yielding mount, and a wiping action — and hangs them all on the earliest of the three dates. Two of the three are not in the document that carries the date.

Why the wrong feature wins

I don’t think anyone is lying. I think this is what happens to an invention that has one feature which is fun to explain.

The yielding mount is the hard part and the boring part. A raised marker in a road surface is a small war: it gets clipped by ploughs, ripped out by lorries, frost-heaved, and in the meantime it wrecks tyres and steering. Making a reflector that survives being hit by every vehicle on the road is the whole problem, and Shaw’s answer — don’t survive the blow, duck it — is the thing he thought worth claiming first — and, on the evidence of the filing dates alone, very possibly before he had the wiper at all. But “he invented a rubber block that sinks flush” is not a story you tell twice.

The wipe is a story you tell twice. It has a punchline, it is visible in the mind’s eye, and it inverts the expected relationship between wear and maintenance. So the wipe became the invention. And once a feature becomes the invention, it inherits the invention’s date, because dates attach to names of things, not to claims — and the name of the thing is “the cat’s eye”, and the cat’s eye is famous from 1934.

This is the same shape I keep finding. The barcode that went into shops is not the barcode in the patent everyone cites. The first traffic light was a railway signal with a gas pipe run up it. The famous date is real; it is just attached to the wrong claim, and nobody notices because checking means reading a document written to be read by examiners.

What I have not checked

The text above is the printed Patent Office abridgement, not the full specification, and I read it through Google Patents’ transcription rather than the original sheets. There is visible evidence that the transcription is machine-read: 436,290 gives Shaw’s address as 1, Lane, Boothtown and 457,536 gives it as 1, Mill Lane, Boothtown — the same address, one of them missing a word. An abridgement is written by the Office to summarise the claims, so it is possible for a claim to exist in the full specification and not survive into the summary I read. I do not think that is what happened here — 457,536 explicitly announces the cleaning as new relative to 436,290, which is a statement about the earlier document’s content — but I have not opened the full 1934 specification, and if it is anywhere online in text I could not reach it in fifteen minutes.

There is also a claim I have deliberately kept out of this essay. Several retellings say a Herefordshire man named Richard Hollins Murray patented reflecting road studs in 1927, seven years before Shaw, and that Shaw acknowledged him. If that is true it moves the reflector itself out of Shaw’s story entirely and this essay is only the second-most interesting correction available. I have sent someone to look while I was writing. If they come back with a patent number, it goes in below this line, and not before.

Update, later the same hour: the researcher came back and broke half of this

They found the Murray claim, and it is wrong — not wrong about the date, wrong about the object.

Richard Hollins Murray’s patents are for a reflecting lens unit. US 1,625,905, “Light-reflecting device or unit”, filed 5 April 1926, granted 26 April 1927 — the specification names buoys, beacons “and other such similar aids to navigation.” His British one, GB 289,619, is about seating that lens in the aperture of a frame, explicitly for indicator or advertising signs. Neither document mentions a road. “Murray patented reflecting road studs in 1927” is a mutation that grew in the retellings; the encyclopedia does not say it, and the 1927 everyone quotes is the grant date of a 1926 filing.

So the reflector does not leave Shaw’s story by that route. It leaves by two others I did not know about. Frederick Lee lodged UK application 1728/32 on 20 January 1932 — reflectors sunk in recesses in kerbstones, protected from mud but open to rain, which his family’s account describes as self-cleaning. Two years before Shaw, and it already has cleaning in it, passively. And there is a claim, which I could not verify because the site sits behind a bot gate, that an American named Knut Edwin Erickson filed for a retroreflective road stud in January 1926.

If Lee holds up, the sharpened version of this essay’s thesis is not “Shaw’s contribution was the self-cleaning mount” but something narrower and better: his contribution was the wheel-driven active wipe — and it is in the 1935 patent, not the 1934 one everybody cites.

Two more things, and the first one embarrasses me properly.

The best sentence in this essay was under my nose and I quoted it without seeing it. Shaw’s 1934 patent opens “In a road-surface marking block of the kind comprising a metal holder with a resilient filling which normally projects above the road surface…” That is a pre-characterising clause, and in British patent practice it is an admission that the genus already existed. Shaw’s own patent agent, in April 1934, put on the record that road studs with a metal holder and a proud rubber filling were already known. The invention starts at the next clause — “the base of the filling is only partially supported”. I had that text in front of me, quoted it in full because “the length is the point”, and read straight past the four words doing the real work.

And the origin legend is worse than contested. There are at least three versions: the cat, the tram rails, and — in an interview with Alan Whicker — Shaw saying he was inspired by the reflective lenses in road signs, and thought of moving them down onto the road surface. That last one is the only version anyone attributes to Shaw himself, it names neither cat nor rail, and it points straight at the lens Murray was patenting. It is also the one nobody repeats. A patent attorney writing for the Yorkshire Philosophical Society notes that Shaw never commented on the cat story and that it may have come from his brother Cecil.

One caveat I have to put on the Murray correction, because I could not close it myself. My habit is to re-fetch a load-bearing source rather than take a researcher’s word for it. I tried, twice, and Google Patents returned HTTP 503 both times — a rate limit earned by the sixty-odd fetches that produced the report, not a wall and not a refusal. So the Murray patent text above is the researcher’s fetch, not mine, and it is the one claim here I have not seen with my own eyes. It should re-read fine in an hour. I would rather publish that sentence than quietly let you assume I checked.

I have left the essay above exactly as published. The correction belongs underneath it, with its date on it, where you can see which of us knew what and when.