Researchers raised 5 colonies of young bees with no experienced dancer to copy, and the waggle dance those bees invented carried a distance error they never corrected in 38 days of life

A honey bee comes back from a good flower patch and does something no other insect does.

She runs a short straight line across the comb, shaking her body side to side, then loops back and runs it again.

The angle of that run tells the hive which way to fly, and how long it lasts tells them how far to go.

Which everybody has known since the middle of the last century.

What nobody tested was whether a bee has to be taught it.

What the dance actually says

Karl von Frisch worked out the code of the waggle dance and won a Nobel Prize for it in 1973.

The trick is that the comb hangs vertically in the dark, so the bee cannot point at anything directly.

Instead she uses straight up as the sun, and the angle of her run off vertical is the angle the others must fly relative to it.

Then the duration of the waggle run carries the distance, with a longer run meaning a farther trip out from the hive.

So a follower crowds in, touches the dancer with her antennae, reads both numbers off a body she cannot see, and flies out to a point she has never visited.

It is the only symbolic language anybody has found in an animal without a backbone.

The assumption everyone carried

And because the whole system is so precise, it was read for decades as pure instinct.

A behavior this exact looked like something written into the animal rather than picked up on the job.

Even the dialects fit that story. Different races and species of honey bee encode distance differently, which von Frisch and Martin Lindauer had documented by the 1950s.

But a dialect can be genetic, and everyone read it that way, as a difference in wiring.

What nobody had done was take the teacher away and see what a bee produced without one.

Because in a normal hive that is impossible. A young bee is surrounded by older dancers from the day she emerges.

Ten colonies in Kunming

So a team in southwest China built hives where that could not happen.

Working in Kunming, researchers from the Chinese Academy of Sciences and the University of California set up ten colonies, five of them experimental and five as controls.

In the five experimental hives, every bee was exactly the same age.

Which meant no young bee could watch an older one dance, because there were no older ones.

They all reached the age of their own first dances, around nine days old, having never seen the behavior performed.

Then the researchers filmed what those bees did on the comb when they came back from a real flower patch.

What stuck and what did not

The naive bees danced anyway, which already answers half the question. Nobody had to show them the moves at all.

But the dances were a mess. Their waggle angles scattered far more widely than the controls, so the direction they gave was unreliable.

And their waggle runs ran too long, which told the hive the food was farther than it was, as the paper in Science reported.

Then came the split in the results that makes this whole study worth reading twice.

By about twenty days old, having watched other bees and practiced, they had fixed the direction. Their angles matched the control colonies.

The distance error never corrected. It stayed with them through the rest of their roughly 38 day lives, as the university described it.

They had not simply made a mistake. They had built a new dialect and kept it, the way a habit outlives its cause at a solar site or an offshore project.

What that makes the dance

So the honest summary is not that von Frisch was wrong.

His code holds exactly as he wrote it, and the Nobel stands. What falls is the assumption that the whole thing is hardwired and finished at birth.

Part of it is learned, by watching, in a window that closes before the bee gives her first performance.

Which is a description of social learning with a critical period, and the lead author reaches straight for the human comparison.

Language in a child works the same way. Early exposure sets something that later effort cannot fully undo.

One caveat is worth keeping. This is ten colonies in one place, and nobody has yet shown how much of the variation between wild populations works this way.

But if a dialect can be created by accident in a single generation and then held for life, then something in that hive is being passed on rather than inherited.