Discovering Ant Language

In 1953, while I was a graduate student at Harvard University, I heard a lecture by Konrad Lorenz on ethology. The experience illustrates the principle that new fields are impelled by one to several great ideas expressible in a few words. The one offered by Lorenz that captured my imagination was the concept of the sign stimulus.

By | November 17, 1986

In 1953, while I was a graduate student at Harvard University, I heard a lecture by Konrad Lorenz on ethology. The experience illustrates the principle that new fields are impelled by one to several great ideas expressible in a few words. The one offered by Lorenz that captured my imagination was the concept of the sign stimulus.

Animal behavior, Lorenz said, is organized into modules of fixed-action patterns, complex sequences of sensory and motor actions that accomplish something for the organism, such as capturing a prey or courting a mate.

A sign stimulus is a particular feature in another organism or in the inanimate environment that triggers a fixed-action pattern. Thus when a male stickleback fish on his territory sees a rival male nearby and rushes out to challenge the intruder, he is not responding to the whole image of a male stick-leback as human beings see it but to the red belly of the rival. The experimenter can get the same response (in other words, the fixed-action pattern) merely by showing the male a red spot painted on a piece of paper.

The sign stimulus was and remains a highly heuristic construction. It says, in effect, that you can dissect out large pieces of behavior by presenting artificial stimuli, such as spots of different color on pieces of paper, until the right one summons the fixed-action pattern. It is somewhat like finding the right access code for a computer program by trial and error. This is a crude way of putting the matter in light of our current understanding of animal behavior, but it contains the kernel of truth that helped launch the new field of ethology.

That afternoon, listening to Lorenz, it occurred to me that the social behavior of ants might be broken open by a similar approach.

Let me explain at this point (and, again, I am oversimplifying) that there are two tribes ofbiologists. One is devoted to solving some fundamental problem or other by looking around for the right organism with which to accomplish the trick. The other, to which I belong, is devoted to glorifying their favorite organism-in my case, the ant-by finding important problems that it is ideally suited to solve.

Lorenz's examples were mostly from fishes and birds, and the sign stimuli were all sight and sound. I thought to myselfthat the social behavior of ants could be understood by defining chemical sign stimuli. The lower the organism in anatomical and behavioral complexity, the more it should depend on chemosensory cues.

But I didn't rush into the laboratory to test this idea. For one thing, ethology was not yet fashionable. In fact, it was opposed by powerful factions in behaviorist and comparative psychology, which insisted that al most all behavior is based on learning and is therefore shaped by the environment. In other words, there are few if any fixed-action  patterns, and the concept of the sign stimulus is mostly illusory. For another thing, I
was preparing for a lengthy trip to the South Pacific and Australia to study the classification and process of species formation in ants, research that led in part to the theory of island biogeography-but that is another story.

In 1958, back at Harvard as an associate professor, I decided to look into what I was calling "chemical releasers" (they were dubbed "pheromones" in 1959 by Adolf Butenandt and Peter Karison) and to do it by studying the trail system of the fire ant (Solenopsis invicta), a favorite species of mine.

At the time, entomologists were well aware that ants follow odor trails, but they regarded the chemicals as signposts for orientation. I decided to look for the releasers that induced the trail-following behavior. Could the key substance come from a particular organ in the posterior part of the abdomen?

As a first step, I dissected out the rectal sac and two principal glands of the poison apparatus and washed them individually in insect Ringer's solution-not an easy procedure, since these organs are barely visible to the naked eye. I then crushed each in turn on the tip of an applicator stick and smeared it in an artificial trail across the glass plate being used by the ants as a foraging arena.

I expected that at best some of the ants would follow the line after they were stimulated by the presence of food. But when I tested Dufour's gland, an insignificant finger-shaped organ located at the base of the sting, an astonishing thing happened.

Worker ants poured out of the nest by the dozens, ran the length of the artificial trail, and milled around in confusion at its end. The same response occurred when I used an extract of the gland's contents in ether or ethanol. I could write my initials with ants frantically following behind in tight columns.

Dufour's gland, I saw at once, contains a chemical substance that not only guides the ants but also attracts them out of the nest. The contents of one gland are enough to activate a large group of foragers. Stretched out in a line, the pheromone is not just the guidepost, it is the entire message.

Soon afterward I pinpointed alarm substances in the mandibular glands of harvester ants. In the same year, Martin Lindauer and his coworkers in Germany, working independently, demonstrated alarm substances in the mandibular glands of the major caste of leaf cutter ants. Work on the pheromones of ants expanded rapidly, paralleled closely by rapid advances in the chemical communication of honeybees by Charles G. Butler and others.

Within 10 years the number of known signal categories in the social insects, some still very poorly understood, had grown to more than 10. Today it is generally agreed that more than 90 percent of signals are chemical in nature.

More often than not the substances work in complex mixtures, creating changeable grades of nuances, and their effects are sometimes enhanced or damped by learning. By studying them experimentally in this fashion, we have made considerable progress toward understanding the matrix of colonial life in the insects.

In the peculiarly human way that science unfolds, the "crude" concept of the sign stimulus developed by Lorenz and other early ethologists proved exactly right.

Wilson is Frank B. Baird Jr. Professor of Science and curator in entomology, Museum of Comparative Zoology, Harvard University, Cambridge, MA 02135.

Popular Now

  1. Genetic Test Solves Royal Mystery
    Notebook Genetic Test Solves Royal Mystery

    Genetic analyses lay to rest conspiracy theories about death of Belgian King Albert I, who lost his life in a rock climbing accident more than 80 years ago.

  2. Investigation Finds Pathologist Guilty of Systemic Misconduct
  3. Bacteria and Humans Have Been Swapping DNA for Millennia
  4. Curious George
    The Scientist Curious George

    George Church has consistently positioned himself at genomics’ leading edge.