If bees are that smart and still disappearing we should probably start checking the skies for Vogons

Right now, the second most read story on the Guardian is that

Bees can solve complex mathematical problems which keep computers busy for days, research has shown. The insects learn to fly the shortest route between flowers discovered in random order, effectively solving the "travelling salesman problem", said scientists at Royal Holloway, University of London. ... Computers solve the problem by comparing the length of all possible routes and choosing the one that is shortest. Bees manage to reach the same solution using a brain the size of a grass seed.

No, they don’t.

The travelling salesman problem is NP complete, meaning that it belongs to a class of problems for which no efficient solution has been found. Mathematicians haven’t yet proven that no such solution exists, but crucially nor have bees. NP problems are considered difficult because they get impractical very quickly when they get big. A brute-force solution to Travelling Salesman requires you to check N! routes, where N is the number of places you need to visit. If there were 22 flowers then a computer could spend the age of the universe doing that, but the research being reported here involved only four flower patches, and a computer could do that in the blink of an eye. What will take a long time to find the fastest route between four flowers is a bee.

The bees in this research took about two hours to settle on the best route around the four patches, putting them a long, long way behind even the very worst computers. Which is fair enough, because as the Guardian was at pains to point out, they have brains the size of grass seeds. A bee can no more calculate the fastest route between a large number of flowers than it can play Minesweeper. This should be obvious, and indeed is when you think about it for (0.3 to 0.4) × blink of an eye. What a bee can do is to come up with a pretty good stab at the fastest route between flowers. That’s still impressive. It’s one of many deeply impressive things that bees do. But it’s not better than a computer.

As well as being interesting in and of itself, bee behaviour may be able to inform the next generation of smarter algorithms, which solve large-scale NP problems in manageable times, at the cost of potentially missing the absolute best solution. For example, you can get to within 0.02% of the optimal solution much faster if you use a “bee colony optimisation” algorithm. This name is not entirely a coincidence: bees’ behaviour has been used as a model for route optimisation for some time. That’s not to slag off this research, but the claim that “bees can solve complex mathematical problems” is only true the same sense that a goalkeeper can solve differential equations in his head.

To me, and to the editors of Natural News, the opening line of the Guardian piece strongly implies that bees can solve the problem better than a computer, which is simply false. But that line, along with most of the story, is lifted wholesale from the journal’s ‘forthcoming articles’ page. Presumably there was a press release, since the same research was reported even earlier by the BBC.

This is interesting stuff that should be in newspapers and the public eye, but scientific journals shouldn’t get it there by misrepresenting it.