American Scientist

Ants at Work: How an Insect Society is Organized. Deborah M. Gordon. 208 pp. The Free Press. 1999. $25.

Because the basic unit of scientific communication, the primary research paper, is short and specific, books remain important in science. The writer of a scientific book can synthesize a body of reported knowledge into a broader, more meaningful presentation. In Ants at Work, Deborah Gordon assembles the bits and pieces of knowledge generated over 15 years of studying the red harvester ant, Pogonomyrmex barbatus, to build a synopsis of her scientific work.

This is not a scientific monograph written for the specialized audience of biologists interested in social insects. Rather, it caters to a general audience. Its language is nontechnical so that any literate person can read and understand the book. Indeed, it is a selection of both the Book-of-the-Month and the Quality Paperback book clubs. Because it is a book about science more than one of science, the key issue for this reviewer was how well it succeeded in presenting its scientific subject to a general readership in an interesting and accurate way.

Gordon, like many biologists who study social insects, is fascinated by ants because they present a mysterious juxtaposition of two levels of biological organization: individual and colony. Somehow the actions of the lower-level units combine to produce the achievements of the higher-level unit. Exactly how this works is, of course, a puzzle common to many branches of biology, but with social insects both the lower-level and the higher-level units are macroscopic, hence the puzzle is particularly striking. In introducing her book, Gordon does a fine job of conveying this central issue in the biology of social insects: "Zoom in, you see ants—zoom out, you see a colony. Ants and colonies are both there in front of you, all the time."

This book also provides the reader with a vivid picture of how a colony of harvester ants develops over several years from a solitary, fertilized queen digging a tunnel to safety underground into a mature colony containing the queen plus some 10,000 workers. The author has a knack for describing colonies as functioning units. For example, in describing the daily foraging maps of colonies, and how these maps reflect interactions between colonies, she uses the lovely imagery of "slow, silent dialogues between colonies." And in describing the competition between colonies for food, she nicely likens them to trees in a forest, for just "as the branches of neighboring trees jostle for light, so the foragers of neighboring colonies divide up the seeds on the ground."

Gordon also succeeds admirably in depicting the logistical challenges she has surmounted in her fieldwork. She candidly recounts, for example, her need as a graduate student to rely on Rent-a-Wreck cars to reach her field site, her difficulty with sines and cosines as she and her undergraduate assistants tried to map colonies by triangulation, and how Boy Scouts needing a merit-badge project helped her construct the 200 wire-screen traps needed for one study. Early on in the book the author characterizes her fieldwork as "waking up at 4:30 to follow ants around and slowly roast," and by the end of the book one feels keenly how hard she has worked.

Gordon has shown great diligence in presenting her work interestingly. I wish the same could be said for the accuracy of its presentation. The problem of accuracy arises at the outset—the dust jacket blurb. Here the reader is informed that "Ants at Work overturns all standard ideas of insect society hierarchy. Gordon shows that an ant colony operates without any central control" and that the author's "surprising and deceptively simple message that the queen is not in charge represents a fundamental shift in modern biology." This is utter bull sauce. I know of no biologist of ants or other social insects who has ever held the view that an ant colony operates with central control, with the queen in charge. So claiming that decentralized control in ant colonies is a fundamental advance in biology is plainly inaccurate.

Such inaccuracy arises not only on the dust jacket but also in the heart of the book. One example is the way the author repeatedly describes the individual ant as bumbling or not very smart. "It can't make complicated assessments. It probably can't remember anything for very long. Its behavior is based on what it perceives in its immediate environment." Such assertions fly in the face of numerous studies that demonstrate that ants and other social insects make complex assessments of their social contexts in deciding how to behave, have long-term memories capable of storing information for weeks or months, and can perform sophisticated feats of path integration, which involve responding to perceptions other than those of the immediate environment. Clearly, the author does not always accurately represent her ants.Ants at Work provides a clear statement of the enduring mystery of how ant colonies work, a vivid depiction of the natural history of one species of harvester ant and an engaging account of how one biologist has pursued her scientific work. It is a truly interesting read. Regrettably, though, it is not a thoroughly accurate one.

A second example of serious inaccuracy is the author's discussion of the concept of caste in social insects. She erects a straw-man definition of caste that she says characterizes earlier work on this topic, most notably the 1978 book by George F. Oster and Edward O. Wilson, Caste and Ecology in the Social Insects. According to Gordon, previous investigators thought each individual worker was specialized on one particular task for its entire life, and they defined a caste as a group of workers permanently specialized on a particular task. Gordon then discusses her work that shows that individual ants often switch tasks (for example, from nest building to food collecting), a phenomenon that, in her view, debunks the concept of caste. The only concept that is debunked here, however, is Gordon's self-created definition of caste. Oster and Wilson fully recognized that social insect workers switch tasks routinely and can even undergo wholesale changes in behavior during their lives. Hence they were careful to define caste as any set of colony members that specialize on particular tasks for prolonged periods of time. Clearly, the author does not always accurately represent the work of other biologists studying ants.