Next I enlisted a friend, the Harvard biochemist John Law, in an attempt to identify the structure of the trail substance molecule. We were joined by a gifted undergraduate student, Christopher Walsh, who in later years was to become a leading molecular biologist and president of the Dana Farber Cancer Institute. We were a capable team, but we encountered a technical snag: we learned that each ant carries less than a billionth of a gram of the critical substance in its Dufour’s gland at any one time. The problem, however, was not insoluble. The late 1950s and early 1960s were the dawn of coupled gas chromatography and mass spectrometry, which allows the identification of organic substances down to millionths of a gram. That meant we needed tens of thousands or hundreds of thousands of ants, each with its vanishing trace of pheromone, to produce the minimum amount required for analysis.
How to gather such huge quantities? From my field experience, I knew of a relatively easy way. When fire ant nests are flooded in nature by rising stream waters, the workers float to the surface in tightly packed masses. Their bodies form a living raft within which the queen and brood are safely tucked. The colony floats downstream until it reaches solid ground, and there the workers proceed to excavate a new nest. After I had explained this phenomenon to Law and Walsh, we traveled to Jacksonville, Florida, one of the southern cities closest to Boston where fire ants are abundant. We took a rental car to the farm country west of the city, where we found two-foot-high fire ant mounds dotting pastures and grassy strips along the roads. There were as many as fifty to an acre, and within each mound lived 100,000 or more ants. Pulling the car over to the verge of the interstate highway, we shoveled entire nests into the water of a slow-moving stream passing through one of the culverts. The soil settled to the bottom, and large portions of each colony rose to the surface. We scooped up seething masses of ants in kitchen strainers and plopped them into bottles of solvent. Law and Walsh soon learned the source of the ants’ common name: the sting of a worker feels like heat from a match brought too close to the skin. And every ant in the nest tries to sting you, ten times or more in succession if you don’t squash it first. We took scores of stings on our hands, arms, and ankles, each of which produced an itching red welt. A day or two later, many of these sites erupted into white-tipped pustules. I suspect that my distinguished colleagues resolved then and there to stay with laboratory biology. Having paid the price, we returned home with enough material to proceed with the analysis of the trail pheromone.
Even with enough raw material, however, the structure of the molecule proved elusive. As Law and Walsh closed in on the active part of the spectrographic array, the peak most likely to be the pheromone diminished to levels too low to analyze further. Was the substance unstable during the separation procedure? Possibly, but now we were running out of extract. In the end the two chemists deduced that the material is a farnesene, a terpenoid with fifteen carbon atoms arranged in a basic structure previously found most commonly among the natural products of plants. They fell short of determining the exact structure, in which the location of every double bond is specified. That difficult feat was accomplished twenty years later by Robert Vander Meer and a team of researchers at the U.S. Department of Agriculture Laboratory in Gainesville, Florida. They discovered that the fire ant trail pheromone is actually a mixture of farnesenes, one of which is Z, E-α-farnesene, augmented by at least two other similar compounds. One gallon of the mixture would be sufficient, in theory at least, to summon forth the inhabitants of 10 million colonies.
For several years following my identification of the glandular source of the trail substance I pursued my goal of deciphering as much of the ants’ chemical language as possible. As I looked more closely at the fire ant trail, I stumbled on a second phenomenon of social behavior, mass communication. The amount of food or the size of an enemy force cannot, I noticed, be transmitted by signals from a single scout. Such information can be conveyed only by groups of workers signaling to other groups. By laying trails on top of one another during a short interval of time, multiple workers, say a group of ten, can signal the existence of a larger target than one identifiable by only a single worker. A hundred workers acting together can raise the range of the smell volume still more. When the food site becomes crowded or the enemy subdued, fewer workers in the group lay trails, so that excess pheromone evaporates and the signal diminishes, and a smaller number of nestmates thus respond.
The information contained in the combined action of masses of individuals coming and going to a target is surprisingly precise. Later writers pointed to a parallel action in masses of brain cells, and the similarity that exists between the brain, the organ of thought, and the insect colony, the superorganism. The first to make the abstract comparison, I believe, was Douglas Hofstadter in Gödel, Escher, Bach: An Eternal Golden Braid, an ingenious disquisition on the nature of organization and creativity. The question then arose and has since been asked many times: Does the resemblance mean that an ant colony can somehow “think”? I believe not. There are too few ants, and those are too loosely organized to form a brain.
I moved on to pheromones that attract and alarm ants. The simplest such substance I found, almost certainly the most elementary pheromone ever discovered, was carbon dioxide. Fire ants use it to hunt subterranean prey and to locate one another in the soil. The most bizarre pheromone, if the generic term can even be used in this case, is the signal of the dead—the means by which a corpse “announces” its new status to nestmates. When an ant dies, and if it has not been crushed or torn apart, it simply crumples up and lies still. Although its posture and inactivity are abnormal, nestmates continue to walk by it as though nothing has happened. Two or three days pass before recognition dawns, and then it is through the smell of decomposition. Responding to the odor, a nestmate picks the corpse up, carries it out of the nest, and dumps it on a nearby refuse pile.
I thought: maybe with the right chemicals I could create an artificial corpse. It should be possible to transfer the odor from one object to another. When I soaked bits of paper with an extract of well-seasoned corpses, the ants carried them to the refuse piles. Thinking back to the basic idea of the chemical releaser I asked, will any decomposition substance trigger the removal instinct, or will the ants respond to just one or two? I found out that a quick answer was possible, because biochemists had already identified a large roster of compounds found in rotting insects. Don’t ask me why such research had been conducted. The scientific literature is filled with such information, and however arcane it often proves useful in unexpected ways. Such was the case in my own (also arcane) study. With two newly recruited assistants I gathered an array of the putrid substances and offered them to my ants on bits of paper, one by one. They included skatole, a component of feces; trimethylamine, one of the essences of rotting fish; and several of the more pungent fatty acids that contribute to rancid human body odor. For weeks my laboratory smelled like the combined essences of sewer, garbage dump, and locker room. In contrast to the responses of my human nose and brain, however, the ants’ responses to the chemicals were consistently narrow. They removed only the paper scraps treated with oleic acid or its ester.
The experiments proved that the ants are neither aesthetic nor meticulously clean in any human sense. They are programmed to react to narrow cues that reliably identify a decaying body. By removing the source they unconsciously safeguard colony hygiene. To test this conclusion about the simplicity of ant behavior I asked, finally, what would happen if a corpse came to life? To find out, I daubed oleic acid on live workers. Their nestmates promptly picked them up, even though they were struggling to get free, and carried them to the refuse pile. There the “living dead” cleaned themselves for a few minutes, rubbing their legs against their body and washing the legs and antennae with their mouthparts, before venturing back to the nest. Some were hauled out again, and a few then yet again, until they became clean enough to be certifiably alive.
A new sensory world was opening to biologists. We
came fully to appreciate the simple fact that most kinds of organisms communicate by taste and smell, not by sight and sound. Animals, plants, and microorganisms employ among their millions of species an astonishing diversity of devices for transmitting the chemicals. The pheromones are usually sparse enough in the bodies of the organisms to make detection difficult for human beings. Animals are unfailingly ingenious in the methods by which they manufacture and deploy these substances. In the late 1950s I was one of no more than a dozen researchers who studied them in ants and other social insects. It was a bonanza that lay before us. We discovered new forms of chemical messages everywhere we looked, and with minimal effort.
In 1961 I invited William Bossert, a Harvard graduate student in applied mathematics, to join me in a project to synthesize all existing knowledge about chemical communication within a single evolutionary framework. Bill possessed in consummate degree the mathematical skills I so conspicuously lacked. At that time he was also pioneering the use of computers in the modeling of evolutionary change. One day he took me into the computer room in Harvard’s Aiken Computation Laboratory, pointed to the spinning tape disks and futuristic control panels, and instructed me that here was housed the future of theoretical biology. Now was the time, he urged, to come aboard and master the powerful new technology. He failed to recruit this naturalist, however. I was just too overwhelmed by the alien culture, easing my way about like an eighteenth-century Pacific islander invited to inspect the armory and rigging of H.M.S. Endeavour. In the years thereafter, as hardware with the computing capacity of the Aiken room shrank to the size of a suitcase, Bossert continued his efforts, but I was never motivated to join him. I had no desire to struggle for years in a field in which I could never hope to become more than mediocre.
Instead, on this occasion, I gave Bossert everything I knew or could find about the chemistry and function of the known pheromones and let him devise the models of their dispersal and detection. He incorporated evaporation and diffusion rates of the known or likely candidate molecules, with estimates of the numbers disseminated and the densities required for animals to recognize them. Together we conceived a series of different forms of expanding gases and theorized on active spaces—the zones within which the molecular densities are high enough to trigger a response. Active spaces are hemispheric in shape when the pheromone is released from one spot in still air, and half-ellipsoidal when the material is released into a steady wind or streaked along the ground into still air. We factored in the size of the molecule as it affects the rates of evaporation and diffusion. We showed that the potential variety of signals dramatically increases—it goes up exponentially—as the size of the pheromone molecule is enlarged within a homologous series. We observed that the substance can either evoke an immediate response or else change the physiology of the animal and its propensity to respond over relatively long periods. When the theory was finally stitched together and all the evidence weighed, we concluded that animals have selected chemicals during their evolution that are well suited to particular meanings. For example, the molecules used as alarm pheromones are smaller in size and have higher response concentrations than those used for sexual attraction, allowing their active spaces to flash on and off more quickly. As a rule, the pheromones chosen are among the ones conceivably most effective in transmitting a particular message.*
Even though this theoretical study of the most general properties of pheromones was progressing well, I stayed close to ants and pushed my laboratory research. In time I estimated that the workers and queens of each colony use somewhere between ten and twenty kinds of pheromones to regulate their social organization. The number undoubtedly varies according to species. But this spread, ten to twenty, is only an educated guess, and remains no more than a guess as I write, thirty years later. The reason is that beyond a few of the most obvious classes, such as the trail and alarm pheromones, the bioassays and chemical analyses turned out to be increasingly difficult. I soon realized that to stay ahead in the field I would have to devote all my time to it and acquire advanced technical training in histology and chemistry. In the late 1960s, ten years after I performed my first crude experiments, the field of pheromone studies was being flooded by a small army of gifted researchers prepared to make this commitment. So I pulled out, an outclassed elder at thirty-five, returning to experiments on chemical communication only when I saw the possibility of a quick result with low technology.
Now we have come to 1969. For some it is the easily remembered year when the Pharaoh’s ants began to steal culture media from the molecular biologists, for others as the year student radicals stalked the campus and Harvard Square with raised fists and revolutionary slogans stamped on their T-shirts. For me it marked a significant change of another kind. My interest in pheromones and island biogeography, my two passions of the previous ten years, had begun to wane. But in September a young scientist with whom I had been corresponding, Bert Hölldobler, knocked at the door of my office in the Biological Laboratories. A lecturer in zoology from the University of Frankfurt, he had come to spend a year as a visiting scholar under my sponsorship. I was about to enter the most sustained and productive collaboration of my research career, built upon a close friendship and a common lifelong commitment to the study of ants.
Although we made no such lofty analysis at the time, we met as representatives of two national cultures in behavioral biology, whose melding would soon lead to a better understanding of ant colonies and other complex societies. One of the contributing disciplines was ethology, European in origin and Hölldobler’s forte, the study of whole patterns of behavior under natural conditions. Though the product of many minds across two generations, in 1969 ethology had become associated in popular tradition with the leadership of Lorenz, Tinbergen, and von Frisch and was well on its way to cosmopolitan status. The other foundation discipline, the one in which I had been more intensively educated, was population biology. Mostly of American and British origin, it was radically different from ethology in its approach to behavior. It addressed entire ensembles of individuals, how they grow, how they spread over the landscape, and, inevitably, how they retreat and vanish. Modern population biology, now also cosmopolitan, attempts to span wide stretches of space and time, and consequently it relies as much on the disciplined imagination invested in mathematical models as it does on studies of live organisms. Its techniques are closely allied to those of demography, in the sense that it combines the births, deaths, and migrations of individual members to construct a statistical picture of the whole society.
The key to the second, higher-level approach is the perception that an insect colony is a population. Some colonies, like the queen and 20 million worker force of the African driver ant, have more inhabitants than entire countries. Like human populations, the only way to understand such ensembles fully is to trace the lives and deaths of their separate members. Both the population-level and the individual-level bodies of information, however, require ethology to create a complete science. This discipline alone addresses in concrete terms the heart of social organization, from communication and nest construction to caste and the division of labor. The final element in the mix is evolution. The behavioral descriptions and population analysis are the historical products of natural selection. Put together, population biology, ethology, and evolutionary theory form the content of the new discipline of sociobiology, which I was to define in 1975 as the systematic study of the biological basis of social behavior and of the organization of complex societies.
Bert Hölldobler and I were edging toward sociobiology. We were, however, first and foremost entomologists, committed to the study of insects. At the time we met he was thirty-three, seven years my junior, but had independently arrived at the conviction that ants are worthy of scientific study no matter how it is done, by ethology, sociobiology, or any other biological discipline. We nevertheless also foresaw, in our early occasional conversations, that ethology and population biology are complementary approaches to the s
tudy of social behavior, and potent in combination.
It could easily have ended there, as a declaration of common interests. At the end of his second year Hölldobler returned to Frankfurt to resume what he foresaw as a lifetime academic career in Germany. At just this time, however, John Dunlop, dean of Harvard’s School of Arts and Sciences, decided to increase faculty representation in behavioral biology. He authorized the appointment of three new professors and placed me in charge of the search committee. In time, after sifting through many letters and evaluations from consultants, we identified this same Dr. Hölldobler as the most promising young scientist in the world working on the behavior of invertebrate animals. He was accordingly invited to come to Harvard as a full professor. He accepted, returning to Cambridge in 1972.
Thereafter we shared the fourth floor of the newly constructed laboratory wing of the Museum of Comparative Zoology. Our contact was close, and we collaborated with increasing frequency in projects in teaching and research. But it was not to be a lifetime arrangement. Sixteen years later, in 1989, Bert returned to Germany, this time to the University of Würzburg in Bavaria, where he had been asked to create a special department devoted to social insects in the newly founded Theodor Boveri Institute of Biological Science. By that time he had come to be greatly appreciated in his native country. Germany, like most other European countries, had a growing interest but weak representation in ecology and related subjects. Bert’s hybrid experience in behavior and population biology uniquely qualified him for national leadership—and continues to do so as I write. In 1991 he received the Leibniz Prize, Germany’s highest award in science.
Nearly two decades of residence in America had turned Bert Hölldobler into a lover of the Arizona mountains, where he spent summers with his family, and of country-western music. Underneath the new American, however, remained the Bavarian—practical, solid, warm-natured and humorous, flexible, altogether the antithesis of the stereotypical Prussian, a difference he was quick to point out whenever the German national character became the subject of conversation. The bluegrass songs of Doc Watson, he once noted in passing, reminded him of Bavarian folk music. Bert above all was rooted to the earth, a naturalist, perhaps, by hereditary predisposition. Fluent English came slowly during his stay in America, and he never lost a marked accent. But it was an asset at Harvard University, where students assumed, correctly, that they were receiving German science and philosophy straight from the source. They consistently gave his courses the highest ratings.
Naturalist 25th Anniversary Edition Page 27