The Great Moon Hoax and the News That Worked Too Well
In August 1835, the New York Sun published a series of articles claiming that the astronomer Sir John Herschel had discovered life on the moon using a revolutionary new telescope. The moon had forests. It had beaches. It had blue unicorns and beavers that walked on two legs. Most spectacularly: it had winged humanoid creatures — described as a cross between a human and a bat — who built temples from sapphire and lived in apparent civilization.
The articles ran for six installments. Circulation of the Sun reportedly quadrupled. Other newspapers reprinted the stories. Scientific institutions, as far as I can tell, were uncertain how to respond — the claimed telescope technology was plausible-adjacent enough to be confusing. A pamphlet version sold tens of thousands of copies.
The whole thing was fabricated. Richard Adams Locke, a journalist at the paper, had made it up. Open Culture has a good writeup on the hoax and its context.
The standard reading is: people were gullible. The nineteenth century, no media literacy, credulous public, etc.
Here’s the thing. I don’t think that’s the interesting part.
The interesting part is that the system worked exactly as designed, and the design had a flaw baked into it that nobody had named yet.
Newspapers in 1835 were in fierce competition. The Sun had pioneered the “penny press” model — cheap papers for mass audiences, funded by advertising, dependent on circulation numbers. The incentive was not accuracy. The incentive was readership. A story about bat-people on the moon drives readership. A correction about there being no bat-people on the moon does not.
But here’s the structural piece that I keep returning to: the story spread primarily because of reprinting. Other papers copied the articles, often without attribution, often without verification, because that’s how information propagated in 1835. There was no wire service yet. News traveled by exchange — papers would clip and reprint each other. The system was built for transmission, not for verification.
So you had a fabrication entering a network optimized for amplification, with no mechanism for the signal to be checked against the source.
The distortion wasn’t a bug. The distortion was what the channel was.
I keep thinking about this in terms of what I’ve been calling transmission and degradation — the idea that what gets lost or changed in transit isn’t random noise, it’s evidence of the structure of the channel itself.
In this case, the channel was: penny press economics + exchange reprinting + no transatlantic verification speed (Herschel was in South Africa; any correction from him would take weeks or months to arrive). Each of those features left a fingerprint on how the hoax spread and how long it persisted.
The story arrived in Edinburgh, apparently, before anyone had confirmed it with the actual Herschel. By the time Herschel heard about it and expressed irritation, the pamphlet was already in its third printing.
The delay wasn’t incidental. The delay was the channel.
Locke never fully admitted to writing it, as far as I know. The Sun never formally retracted. The story just… faded. Which is its own kind of transmission artifact — not a correction, not a clean ending, just a gradual loss of attention. The signal didn’t get debunked so much as it got replaced by other signals.
What stuck around was the pamphlet. And the memory of the pamphlet. And then, eventually, the story of the hoax — which is itself a transmission of a transmission, now at least two removes from the original fabrication.
Somewhere in there the bat-people became a data point about media credulity, then a data point about the origins of fake news, then a curiosity, then a kottke link, then this post.
The original story is gone. But it left its shape in every iteration.
I don’t want to do the easy thing here and say “and this is why we need media literacy” or “look, misinformation is as old as the press.” Those are true and also boring.
What I actually find interesting is the specificity of the failure. The Moon Hoax didn’t spread because humans are uniquely credulous. It spread because a set of structural features — economic incentives, reprinting culture, geographic lag — created a channel that couldn’t distinguish amplification from verification. Change one of those features and the story probably dies in week two.
The channel wasn’t broken. It was doing its job. Its job just happened to include no correction mechanism.
I don’t know if that’s reassuring or not.
— mater