The discovery that honeybees communicate with their hive-mates about food sources through patterned movement is one of the most widely quoted results in twentieth-century biology. When Karl von Frisch first proposed in the 1940s that the so-called waggle dance of returning foragers encoded both the direction and the distance of a nectar source, many of his colleagues were sceptical. Animals were not thought to be capable of such abstract communication. The subsequent sixty years of research have not only confirmed the essentials of von Frisch's account but have added a set of subtleties that have changed how biologists think about animal communication more generally.

The waggle dance itself is performed on the vertical surface of the comb inside the hive. A returning forager runs forward in a straight line while vibrating her abdomen from side to side, then loops back to her starting point and repeats the movement. The straight-line portion contains most of the information. Its direction relative to vertical corresponds to the direction of the food source relative to the sun. Its duration corresponds to the distance of the food source from the hive. A longer waggle indicates a food source further away. Other workers gather around the dancer, feel the vibrations through the comb, and use the information to fly out to the food.

The accuracy of the communication is impressive but not perfect. Bees translating a dance into a flight path routinely miss the exact location of the food source by several metres, and experiments that manipulated the angle of the dance by altering the hive's orientation produced correspondingly shifted flights. This scatter may not be an inefficiency; it may be a feature. Entomologist Dr. Hiro Nakamura has argued that slight variation in the receivers' interpretations causes the recruited bees to spread out across a small area around the indicated point, which is useful when the resource is a patchy field of flowers rather than a single bloom.

The dances are not used exclusively for flowers. When a colony is preparing to swarm - to split, with part of the workforce and the old queen leaving to found a new nest - scouts that have visited potential nest sites perform dances advertising those sites. The same grammar of direction and distance encodes the location of, for example, a hollow tree or a suitable cavity. Other scouts follow up on the dances, visit the sites, and, if they approve, add their own dances. Over a period of hours or days, the scouts converge on a single site, and the swarm moves.

Researchers in recent decades have drawn attention to the complexity of this decision-making. Each scout dances for as long as she judges her site worthwhile, and stops sooner if she visits a better one. Behavioural ecologist Dr. Aarti Kapoor describes the result as a form of 'distributed consensus': no individual bee compares the sites directly, yet the colony as a whole reliably chooses the best of several options. Simulations have shown that the algorithm implicit in the scouts' behaviour performs well even when some of the scouts are misinformed.

Honeybee dances also work in the dark. Inside the hive, there is no light, and the dancers rely on the pull of gravity as a reference for vertical. When researchers in Austria tilted a hive so that the dances were performed on a horizontal surface, the forages initially produced chaotic movements. After several hours they had reorganised their dances around the sun, which entered the hive through an opening in the experimental setup. This finding suggested that bees possess more than one reference system and can switch between them, although the switching is not instantaneous.

More recent work has looked at how dance information interacts with the colony's other sources of knowledge. Bees do not blindly follow dances; they use them alongside their own memories of previously successful food sites, their observations of other returning workers, and the chemical signals carried on flower petals. Dr. Kapoor calls the dance 'one voice in a larger chorus' and warns against interpretations that treat bee behaviour as a simple input-output system. What remains remarkable, and what still attracts new researchers to the field, is that such an elaborate pattern of communication exists at all in an animal whose brain, by mass, is a small fraction of a grain of rice.