The industrious ant

Sterile workers, farming insects, and an argument about where cooperation comes from

A A colony of ants presents an old difficulty in a particularly sharp form. Almost every individual in it is sterile, spends its life working for the reproduction of another animal, and leaves no descendants of its own. Any account of evolution that begins from the reproductive interest of the individual has to explain how such a thing could arise, and the explanation matters far beyond entomology, since the same problem appears wherever cooperation is costly to the cooperator.

B The most influential answer turns on relatedness. A worker that helps its mother produce sisters is propagating copies of its own genes at one remove, and in the ants, bees and wasps an unusual genetic system means that full sisters share more of their genes with each other than they would with their own offspring. On this reasoning, sterile helping is not self-sacrifice but an efficient reproductive strategy pursued by other means, and the fact that these groups have evolved this way of life repeatedly, while most animal groups never have, was for decades treated as decisive confirmation.

C The objections that have accumulated are serious and are not the objections of cranks. Termites are as thoroughly social as any ant and do not have the genetic system in question. Several groups that do have it have never evolved sterile castes at all. Detailed measurement of relatedness within colonies has produced values considerably lower than the argument requires, particularly where a queen has mated with several males or a colony contains several queens. A prominent group of theorists proposed in 2010 that the whole framework be replaced by an account resting on the advantages of defending a shared nest, and were answered by a letter of objection signed by well over a hundred researchers, which is not the usual way scientific disagreements are conducted.

D Whatever settles that argument, the practical consequences of colonial life are extraordinary, and they arise without any individual understanding them. A foraging ant that finds food returns to the nest depositing a chemical trail; other ants follow the trail and reinforce it if the food is still there; the trail evaporates if they do not. This arrangement locates the shortest route to a food source, abandons exhausted sources automatically, and requires no ant to know anything about routes, distances or the state of the colony's larder. The behaviour of the whole is a property of the interactions, not of the participants.

E The same logic, applied to other problems, produces the allocation of workers between tasks. An ant encountering many nest-mates engaged in one activity is less likely to take it up; an ant encountering few is more likely; and the colony consequently redistributes its labour force as circumstances change, without anyone counting anything. Elaborations on this principle produce results that look unmistakably like agriculture and animal husbandry. Several lineages cut vegetation not to eat it but to feed a fungus cultivated in underground chambers, which they weed, fertilise, protect with antibiotic-producing bacteria carried on their bodies, and eat; the fungus has been domesticated for millions of years and can no longer live independently. Others maintain herds of sap-sucking insects, defending them from predators and moving them to fresh plants in exchange for the sugary fluid they excrete. Neither behaviour was invented by anybody, and both are considerably older than the human versions. The comparison should not be pushed too far — no ant decides anything, and the arrangements are maintained by responses to chemical cues rather than by intention — but the functional similarity is close enough that the same words are used by researchers without embarrassment.

F The tendency to describe a colony as a single organism is therefore understandable and also treacherous. It captures something real: the colony reproduces as a unit, its members are differentiated like tissues, and selection plainly acts on colonies as wholes. But the metaphor conceals conflict that is demonstrably present. Workers in many species police one another's attempts to lay eggs, queens and workers differ in the ratio of sons to daughters each would prefer, and colonies containing multiple queens display exactly the tensions the metaphor denies. A body whose cells were at odds in this way would be diseased; a colony in which they are is normal. The useful formulation is that a colony is a society with unusually aligned interests rather than an organism with unusually independent parts, and the difference is not merely verbal: it predicts where conflict should appear, and the prediction has repeatedly been borne out.

G The strongest reason to keep studying these animals is neither the metaphor nor the dispute about its origins. It is that a colony solves problems no member can represent — allocating labour between tasks as conditions change, defending a territory, finding the best of several routes — using rules simple enough to be written on a postcard, and that the study of how it does so has already changed the way networks are routed and vehicles are scheduled. That a question about the origins of sterility should have produced a body of theory used to move data through the internet is a reasonable answer to anyone who asks what the study of insects is for.