The signals inside a colony

The vibratory repertoire

The stop signal is the most interesting of these. It was long assumed to be a request for food. It is now understood as a negative feedback message: a bee that has met danger at a site will butt dancers advertising it, and the same signal is exchanged between scouts advocating competing nest sites. That cross-inhibition is what lets a swarm break a deadlock between two good cavities rather than splitting.

Trophallaxis

Bees pass liquid mouth to mouth, constantly. It distributes food, and it distributes information: pheromones, including queen substance, circulate through the colony this way, and so do the odours of what is currently being collected. Automated tracking of marked bees has mapped the resulting network, and it is dense enough that a substance introduced to one individual reaches a large share of the colony within hours.

It is the reason a colony behaves as a connected system. There is no broadcast channel in a beehive; there is a very large number of brief pairwise exchanges.

The chemical channel

• Queen mandibular pheromone: a multi-component blend that suppresses worker ovary development, inhibits the rearing of new queens, and holds a swarm together in flight. Its absence is detected within hours and triggers emergency queen rearing.
• Brood pheromone: released by larvae, and it shifts worker behaviour and foraging effort. The brood is an active participant in colony regulation rather than a passive load.
• Sting alarm pheromone: released when a bee stings, marking the target and recruiting defenders. Its principal component smells of banana, which is why beekeepers avoid bringing one to an apiary.
• Mandibular alarm pheromone: a second, shorter-range alarm compound released by biting.
• Nasonov pheromone: fanned from a gland at the tip of the abdomen at a nest entrance, at a water source and at a settling swarm. It means here, and bees adopt a characteristic stance to broadcast it.
• Footprint pheromone: deposited as bees walk, and used to mark recently visited flowers so foragers do not waste time revisiting an emptied one.

How this adds up to a decision

None of these signals is a command and none is addressed to the colony as a whole. Each shifts the probability that a particular bee does a particular thing. Colony-level outcomes — allocating foragers between patches, choosing a nest site, ventilating, defending — emerge from many individuals with different thresholds responding to local information.

That is worth stating plainly, because the shorthand of a hive mind hides exactly the mechanism that makes it interesting. Nobody is deciding. The decision is what the arithmetic of thousands of thresholds produces.