Letters to a Young Scientist (15 page)
If scientists know so little of raw biological diversity at the taxonomic level, we know even less of the life cycles, physiology, and niches of each species in turn. And for all but a very few localities on which biologists of diverse training have focused their energies, we are equally ignorant of how the idiosyncratic traits of individual species fit together to create ecosystems. Ponder these questions for a while: How do pond, mountaintop, desert, and rain forest ecosystems really work? What holds them together? Under what pressures do they sometimes disintegrate, and how and why? In fact, many are crumbling. Humanity’s long-term survival depends on acquiring answers to these and many other related questions about our home planet. Time is growing short. We need a larger scientific effort, and many more scientists in all disciplines. Now I’ll repeat what I’ve said when I began these letters: you are needed.
The female gypsy moth, located at the lower point of the active space, releases a pheromone cloud within which is a region of high concentration followed by the male. Drawing by Tom Prentiss (moths) and Dan Todd (active space of gyplure ©
Scientific American
). Modified from “Pheromones,” by Edward O. Wilson,
Scientific American
208(5): 100–114 (May 1968).
T
HE BEST WAY
I can explain the nature of scientific theories to you is not by abstract generalizations but by offering examples of the actual process of making theory. And because this part of science is the product of creative and idiosyncratic mental operations that are seldom put into words, I will stay as close to home as possible by using two such episodes in which I have been personally involved.
The first is the theory of chemical communication. The vast majority of plants, animals, and microorganisms communicate by chemicals, called pheromones, which are smelled or tasted. Among the few organisms that use sight and sound primarily are humans, birds, butterflies, and reef-dwelling fish. Working with the social behavior of ants in the 1950s, I became aware that these highly social insects use a variety of substances that are released from different parts of the body. The information they transmit is among the most complex and precise found in the animal kingdom.
As new information began to pour in, those of us conducting the early research saw that we needed a way to pull together the fragmented data and make sense of them. In short, we needed a general theory of chemical communication.
I was extremely fortunate during this early period to serve as the cosponsor of William H. Bossert, a brilliant mathematician working for a Ph.D. in theoretical biology. After completing his degree requirements in 1963, he was invited to join the Harvard faculty, and in a short time thereafter he received a tenured professorship in applied mathematics. While still a graduate student, he joined me in creating a theory of pheromone communication. The time was right for such an effort, and we were successful. On no other occasion in my scientific career has a project worked out as quickly and as well as did the collaboration with Bill Bossert.
To kick things off, I told him what I knew about the new subject. I laid out the basic properties of chemical communication as I had come to understand them. There was not a great deal of information to go on in this early period. From field and laboratory studies, I said, we knew that a wide variety of pheromones exist. It seemed logical that we should begin with a classification of the roles of all of those known, then try to make sense of each one in turn. The theory should deal not just with form and function of the pheromone molecules, which was the goal of most researchers, but also with their evolution. Put simply, we wanted to know what the pheromones are and how they work, of course, but also
why
they are one kind of molecule and not another.
Before giving you the theory, here are the specific “why” questions we meant for it to explain. Is the pheromone molecule used the best way possible, or is it one that was selected at random during evolution out of a limited array available for the job? What do the pheromone messages “look like” if you could see them spreading through space? Should the animal emit a lot of the pheromone or just a little in each message? How far and fast do the pheromone molecules travel through air or water, and why?
Here, then, in a nutshell, is the theory.
Each kind of pheromone message has been engineered by natural selection—that is, trial and error of mutations that occur over many generations resulting in the predominance of the best molecules, with the most efficient form of transmission allowed by the environment
. Suppose a population of ants is started by two ant colonies who compete with each other. The first colony makes one kind of molecule and dispenses it in a certain way, and the second colony makes another kind of molecule that is less efficient, or else is dispensed less efficiently, or both. The first colony will do better than the second, and as a consequence it will produce more daughter colonies. In the population of colonies as a whole, the descendants of the first colony will come to predominate. Evolution has occurred in the pheromone, or in the way it is used, or both.
Bossert and I agreed: “Let’s think about ants and other organisms using pheromones as engineers.” This thought took us quickly to ants recruiting other ants by laying a trail for them to follow. So, at the next picnic (or on your kitchen floor if the house is infested) drop a crumb of cake. It is logical to suppose that the ant scout that finds it needs to dribble out the trail pheromone at a slow rate in order to make the store of the substance she carries in her body last a long time. The piece of cake may be several ant-mile equivalents away. In this function, the ant is like an automobile engine designed for high mileage. In order to achieve such efficiency, the pheromone needs (in theory) to be a powerful odor for the ants following the trail. Only a few molecules should suffice. Also, the pheromone must be specific to the species using it, in order to provide privacy. It is bad for the colony if other ants from other species can pirate the trail, and even dangerous for the colony if a lizard or some other predator can follow the trail back to the nest. Finally, the trail substance should evaporate slowly. It needs to persist long enough for other members of the colony to track it to the end, and start laying trails of their own.
Then there are the alarm substances. When a worker ant or other social insect is attacked by an enemy, whether inside or outside the nest, it needs to be able to “shout” loud and clear, in order to get a fast response. The pheromone must therefore spread rapidly and continuously over a long distance. But it should also fade out quickly. Otherwise even small disturbances, if frequent, would result in constant pandemonium—like a fire alarm that cannot be turned off. At the same time, unlike the case for trail substances, there is no need for privacy. An enemy can gain little by approaching a location teeming with alert and aggressive worker ants.
Let me pause here to describe an easy way for you to smell an alarm pheromone yourself. Catch a honeybee from a flower in a handkerchief or other soft cloth. Squeeze the crumpled cloth gently. The bee will sting the cloth, and as it draws away it will leave the sting (which has reverse barbs) stuck in the cloth. When that happens, the immobile sting pulls out part of the bee’s internal organs. Let the bee move to the side, then crush the sting and the organs between two fingers. You will smell an odor that resembles the essence of banana. Its source is a mixture of acetates and alcohols in a tiny gland located along the shaft of the sting. These substances function as an alarm signal, and they are the reason other bees rush to the same site and add their own stings. Next, if the eviscerated bee hasn’t flown away, crush its head and smell that. The acrid odor you detect is from a second alarm substance, 2-heptanone, emitted by glands at the base of the mandibles. (Don’t feel bad about killing a worker bee. Each has an adult life span of only about a month, and it is only one of tens of thousands that make up a colony. The colony in turn is potentially immortal, since new mother queens replace the old ones at regular intervals.)
The next category of pheromones are the attractants, in particular the sex pheromones, by which females call to males for the purpose of mating. The phenomenon is widespread not only in social insects but also throughout the animal kingdom. Other attractants also include the scent of flowering plants, in which the flowers call to butterflies, bees, and other pollinators. The most dramatic substances of the kind are the sex attractants of female moths, which can draw males upwind for distances of a kilometer or more.
Finally, Bossert and I reasoned in our initial classification, there are the identification substances. An ant, upon smelling these substances, can tell whether another ant is from the same or a different colony. It can also identify a soldier, ordinary worker queen, egg, pupa, or larva, and if the latter, its age. Carrying a chemical badge of this kind with you at all times means wearing the pheromone like a second skin. An identity pheromone is a single substance or, more likely, a mix of substances. It needs to evaporate very slowly and be detectable only at a very close range. If you closely watch one ant or some other social insect approach another, say while running along a trail or entering a nest, you will see the two sweep each other’s body with their two antennae—a movement almost too fast for the eye to catch. They are checking body odor. If they detect the same odor, each will pass the other by. If the body odor is different, they will either fight or else flee from each other.
Reaching this point in the investigation, Bossert and I left the “adaptive engineering” method of evolutionary biology and passed into biophysics. We needed to envision the spread of the pheromone molecules from the body of the animal releasing them, and as precisely as possible. Obviously, as the pheromone cloud disperses, its density would decline—there would be fewer and fewer molecules in each cubic millimeter of space. Eventually there would be too few to smell or taste. Bossert then devised the crucial idea of “active space,” within which the molecules are dense enough to be detected by the receiving plant, animal, or organism. He constructed models (at last, a place for pure mathematics!) to predict the shape of the active space. We were now in a new phase in creating the theory of pheromone communication.
With the ant or any other broadcasting organism sitting on the ground in still air, the shape of the active space would be hemispherical—one half of a sphere cut in two—with the broadcaster at the center of the flat surface. When an organism releases the pheromone from a leaf or object off the ground and in an air current, the shape of the active space would be an ellipsoid (roughly, shaped like an American football), tapering to a point at each end. The broadcaster would be at one of the points, releasing the pheromone downwind. When a trail is laid on the ground in quantities sufficient for it to be detected over a long period of time, the space would become a very long semiellipsoid, in other words an ellipsoid cut in half lengthwise at ground level.
Next we turned our attention of the design to the molecule itself. Trail substances and identification odors should consist, we reasoned, of either relatively large molecules or mixes of large molecules. They should diffuse slowly. Alarm pheromone molecules should be chosen in evolution to be smaller in size. They should form a more limited active space, and dissipate quickly. The qualities of the active space depend on five variables that can be measured: the diffusion rate of the substance, the surrounding air temperature, the velocity of the air current, the rate at which the pheromone is released, and the degree of sensitivity of the organism receiving it. With these measurable quantities in place, the theory began to take shape in a form that could be taken into the field and laboratory, and used to study animals as they communicated.
Next, we left biophysics for a while and entered the realm of natural products chemistry to learn the nature of the pheromone molecules. It’s the same chemistry used widely in pharmaceutical and industrial research. It was our good luck that a recent major advance in molecular analysis put this part of the pheromone story within reach. By the late 1950s, the new technique of gas chromatography coupled with mass spectrometry made it possible to identify substances in quantities as little as a millionth of a gram, or less. Where previously chemists needed thousandths of a gram of pure substance to get the job done, now they needed only thousandths of a thousandth. The technique has allowed the detection of trace substances, including toxic pollutants, in the environment. Along with DNA sequencing (also requiring only a droplet of blood or the wipe of a wineglass), it also soon transformed forensic medicine. For us and other researchers it made possible the identification of pheromones carried in the body of a single insect. Ants commonly weigh between one and ten milligrams each. If a particular pheromone takes up only a thousandth or even a millionth of its body weight, it is still possible for researchers to make some progress in the characterization of the molecule. The chemists I worked with could obtain hundreds or thousands of ants. That was no great feat—it takes only a shovel and a bucket—and is one of the great advantages of working with ants. It became possible not only to isolate candidate pheromones but also to obtain enough of the material for bioassays—testing the material with live colonies to see if it evokes what theory suggests is the correct response.