Ants at Work: How an Insect Society is Organized

  • Deborah Gordon
Free Press: 1999. 208 pp. $25
Mating time: ants display enormous variability in reproductive tactics across the 300 ant genera. Credit: CORBIS/GEORGE LEPP

It is certainly an odd sight in the early light of an Arizona desert morning to see a number of people digging themselves into the dry soil, crawling around with their noses close to the ground, hunting for something that is invisible from even a short distance. Still, the inhabitants of Paradise and Portal in southeasternmost Arizona are by now presumably quite used to it, because over the past decades the Chiricahua Mountains and adjacent San Simon Valley have become a favourite field site for researchers interested in the evolution, sociobiology and ecology of ants. One of them, Stanford professor Deborah Gordon, now writes about what she has discovered, in 17 years of observations, lab and field experiments and computer modelling, about Pogonomyrmex harvester ants.

Ants at Work is an entertaining mixture of personal travel report and popular science. And it is deliberately ‘Pogo-centric’, as the author herself says: it focuses almost exclusively on this genus, ignoring most other ants. It is also much more a private, ‘Gordon-centric’ oeuvre than a comprehensive and integrated treatise on ants, and three-quarters of the roughly 45 references in the appendix are by the author and her research team. Other research on Pogonomyrmex, including studies done just across the road from her own study site and a recently published book on harvester-ant biology, is left aside, as are other key publications, including several fundamental reviews on ant biology. These omissions might perhaps help a naive reader, whose knowledge does not go much beyond the film Antz, to keep on track and not go astray among the fascinatingly complex universe of social insects.

Those ant novices will be initiated straight to Pogonomyrmex, their foraging strategies, their conflicts with neighbouring colonies and how they make decisions, without the detour to other ants with alternative ways of life. And they will also learn a lot about how all this information is gathered and how researchers cope with the manifold problems that generally make field studies on animal behaviour quite tricky — such as temporal changes in activity patterns or motivation. Ants at Work vividly involves the reader in several ingenious experiments especially designed to overcome these and many other difficulties.

Readers more familiar with ants, however, will occasionally feel somewhat uneasy. Pogonomyrmex is only one of about 300 ant genera and it has a rather specialized ecology and behaviour. Generalizations, such as those made in the introduction that workers of all ants are sterile and new reproductives fly off to mate, unduly oversimplify the enormous variability of reproductive tactics in ants, with wingless sexuals mating in the nest or workers laying both fertilized and unfertilized eggs.

Furthermore, some aspects of ant biology presented to the reader as amazing news have for a long time and in much detail been known from other ants. It is common knowledge among social insect researchers, for example, that an ant colony operates without any central control, that workers can easily switch from one task to another, and that the meaning of signals used in communication is context-specific. That an ant queen is not in charge, claimed on the book jacket to be a revolution in our thinking on natural organization, is therefore trivial. On the other hand, the remark that no ant has power over another is somewhat puzzling, given the accumulating evidence for overt kin conflict and dominance hierarchies in some ant species.

One of the book's major themes is division of labour, a very acute problem that is being widely (and heatedly) discussed among social-insect researchers and has also elicited considerable interest in the artificial-intelligence community. Here, Gordon describes a series of experiments and observations which demonstrate that task allocation is highly flexible and that encounter rates between individuals may be important for task switching in Pogonomyrmex. She then uses this evidence to shatter the alternative hypothesis that an ant's task is innate, that is, that a worker does a single task throughout its life. However, to my knowledge, this latter viewpoint has never been held seriously, not even by George F. Oster and Edward O. Wilson, to whom it is attributed. This looks to me like fighting a straw man. I would have preferred an up-to-date overview on the state-of-art in this field or a thorough analysis of how Gordon's encounter-rate model compares with temporal polyethism (which is mentioned elsewhere in the book) or the many other models of decentralized control and self-organized division of labour. Even readers less familiar with social insects would not have been overtaxed by a balanced treatment of this subject.

To conclude, Ants at Work offers a lot of information on harvester ants and introduces several new, and some not so new, interpretations of ant behaviour, including the concept of self-organization, to a broader audience. This is probably the author's main aim, and it is well achieved. It is doubtful, however, whether the book will satisfy the much farther-reaching aims listed on the dust-jacket. I enjoyed reading the book, but it did not really revolutionize my understanding of ants.