By the early 1960s population biology was gaining substantial independent strength, and my confidence in its canonical relation to sociobiology rose. In late July of 1964, when I met with the “Marlboro Circle” in Vermont—Egbert Leigh, Richard Levins, Richard Lewontin, and MacArthur were the others—I represented the idea of sociobiology as a possible derivative of population biology. Societies are populations, I argued, and amenable to the same modes of analysis.
I saw that the quickest way to make the point was to use population biology in a solid account of caste systems and division of labor in the social insects. I was already well prepared for the task. In 1953 I had traced the evolution of caste in ants in a more descriptive manner, using measurements of scores of species from around the world. I showed how the divergence in anatomy among queens, soldiers, and ordinary (“minor”) workers is the consequence of changes in allometry, the differential growth of different organs. Allometry, just by increasing or diminishing one dimension of the body relative to another, can produce larger or smaller heads, full-blown or shriveled ovaries, and other divergent products in any part of the final adult form. The idea was not new. It had been advanced earlier by Julian Huxley in his 1932 book, Problems of Relative Growth; he in turn had been inspired by D’Arcy Thompson’s analysis of the evolution of morphological gradients, published in the 1917 classic On Growth and Form. I took the ants from there, following in plausible sequence the evolution of castes from one basic type in small steps all the way to multiple forms differing among themselves radically. Then I gave the subject a new twist. To allometry I added demography, the relative numbers of individuals of different castes within each colony. When allometry and demography are joined closely, the probable evolution of caste becomes much clearer. The anatomy of a particular caste member obviously determines the efficiency of its labor role; a soldier, for example, functions best with large, sharp mandibles and powerful muscles to close them. But the number of soldiers, I pointed out, is also crucial. If there are too few fighting specialists, the colony will be overwhelmed by enemies. If there are too many, on the other hand, the colony cannot gather enough food to care for the young. It follows that colonies must regulate the birth and death rates of the various caste members created by allometry. In later studies I came to call the phenomenon “adaptive demography.” I interpreted it as a population-level trait of an advanced society.
Julian Huxley was intrigued by my employment of allometry and demography. When he visited Harvard in 1954, he asked to see me. My faculty advisers were impressed by the request, and I was thrilled to meet the great evolutionary scholar and humanist. Our common interest, we agreed, was a classic topic of general biology. The problem of ant castes had attracted the attention of Charles Darwin, who saw it as a threat to the theory of natural selection. Although Darwin had construed the idea of relative growth intuitively, Huxley and I knew that the ideas and data of our own studies had produced the first full and quantitative evolutionary explanation.
In 1968 I refined the idea of adaptive demography and developed several new principles of caste evolution with the aid of models in linear programming. In 1977 I was joined in a further, year-long study by George Oster, an exceptionally gifted and resourceful applied mathematician from the University of California, Berkeley. This time we explored the theory of caste evolution throughout the social insects. We were able to add other concepts from population biology to my earlier formulation. Oster led the modeling effort. His range of analytic techniques was awesome, affirming his generally held reputation as the mathematically most competent of all theoretical biologists. He often played with novel approaches, and he then had to lead me through the steps before we were able to continue the conversation. My role was the same as the one I had adopted with Bill Bossert in the synthesis of chemical communication fifteen years previously. At the beginning of each new avenue of exploration, I fed in all that I knew about caste and division of labor, information that often consisted of no more than doubtfully related fragments, along with the best intuitive conclusions I could draw. Oster then built formal models with what we could see—or guess—of the empirical relationships and trends, extending our reach in space and time. I responded with new evidence and guesses, he reasoned and modeled again, I responded, he modeled, I responded.* During breaks we gossiped and explored our other common interests. A magician of professional grade, he once dazzled me with sleights of hand I could not fathom even when he repeated them a foot or two from my concentrated gaze. I found my incapacity deeply disturbing. I was a proud scientific materialist, but I had to ask, How much else seemingly real in the world is an illusion? I learned a principle that others have established, often from painful experience: never trust a scientist to evaluate “evidence” of telekinesis and other feats of the paranormal; go instead to an honest magician.
Through the 1960s I searched for other ideas to add to the sociobiology armamentarium. One that I fashioned from population biology was the evolutionary origin of aggression. In his early writings, and again in his famous book On Aggression in 1966, Konrad Lorenz postulated aggression to be a widespread instinct that cannot be suppressed. It wells up in organisms and, like a crowded liquid, seeks release in one form or another. In human beings, Lorenz suggested, it is better released in organized sports than in war. In 1968, in the first of two Man and Beast symposia sponsored by the Smithsonian Institution, I showed that a more precise explanation consistent with the growing body of evidence from field studies is the role of aggressive behavior as a specialized density-dependent response.* As populations increase in density, those of many species are constrained by a growing resistance from one or more factors. Among these density-dependent responses are the rise in per capita mortality from predation and disease, the loss of fertility, a greater propensity to emigrate, and—aggression. Whether aggressive behavior originates at all during evolution depends on whether other density-dependent factors reliably intervene to control population growth. Even then the form it takes can vary, emerging as territorial defense, dominance hierarchies, or all-out physical attack and even cannibalism, depending on the circumstances in which population limits are attained. Thus aggression is a specialized response that evolves in some species and not others. Its occurrence can in principle be predicted from a knowledge of the environment and natural history of the species.
The elements of sociobiological theory came from many sources. But when the most important idea of all came along, I at first resisted it with all my ability. In 1964 William Hamilton published his seminal theory of kin selection in the Journal of Theoretical Biology, in a two-part article titled “The Genetical Evolution of Social Behaviour.” In the decades since, a sizable research industry has been built upon this single paper. Some of Hamilton’s reasoning and conclusions have been challenged, then defended by enthusiasts, only to be challenged, and defended, again. The core of the theory has stood up well. Its essence, like that of all great ideas, is simple, of the kind that evokes the response, “Obviously that is true (but why didn’t I think of it?).” Conventional Darwinism envisages natural selection as an event occurring directly between generations, from parent to offspring. Different lineages carry different genes, most of which prescribe traits that affect survival and reproduction. How an organism grows in body form, how it searches for food, how it avoids predators: these are among the traits affected by genes. The genes therefore determine survival and reproduction. Because by definition lineages that survive and reproduce better create more offspring in each generation, their hereditary material comes to predominate in the population over many generations. The increase of one set of genes at the expense of another is (again by definition) evolution by natural selection. The history of life has been guided by the appearance of new genes and the rearrangement of chromosomes bearing the genes through random mutations. These ensembles are winnowed by natural selection, which is the increase or decrease of particular combinations of genes and chromosomes through the differential
survival and reproduction of the organisms carrying them.
In one important respect this traditional process of natural selection can be called just one type of kin selection. Parents and offspring are, after all, close kin. But Hamilton observed that brothers, sisters, uncles, aunts, cousins, and so forth are also kin; and he thought about what this truism means for evolution. The other kin share genes by common descent no less than parents and offspring. So if there is any interaction among them that is influenced by genes, say, a hereditary tendency toward altruism, or cooperation, or sibling rivalry, the interaction will result in a change in survival and reproduction and should equally well cause evolution by natural selection. Perhaps the ancillary forms of kin selection drive most forms of social evolution.
What made Hamilton’s idea immediately attractive was that it helped to resolve the classic problem in evolutionary theory of how self-sacrifice can become a genetically fixed trait. It might seem on first thought—without considering kin selection—that selfishness must reign complete in the living world, and that cooperation can never appear except to enhance selfish ends. But no, if an altruistic act helps relatives, it increases the survival of genes that are identical with those of the altruist, just as the case in parents and offspring. The genes are identical because the altruist and its relative share a common ancestor. True, the corporeal self may die because of a selfless action, but the shared genes, including those that prescribe altruism, are actually benefited. The body may die, but the genes will flourish. In the enduring phrase of Richard Dawkins, social behavior rides upon the “selfish gene.”
Hamilton had traveled that high road of science once described by the great biochemist Albert Szent-Gyorgyi, “to see what everyone has seen and think what no one has thought.” But I am reasonably sure that had Hamilton expressed kin selection in merely abstract terms, the response to his formulation would have been tepid. Other biologists upon reading it would have said, “Yes, of course, and Darwin had a somewhat similar idea, did he not?” And, “Correct me if I’m wrong, but haven’t notions of this kind been discussed off and on for a long time?” Yet Hamilton did succeed dramatically (although few learned about the theory until I highlighted it in the 1970s). He did so because he went on to tell us something new about the real world in concrete, measurable terms. He provided the tools for real, empirical advances in sociobiology. As Hamilton told me later, he was able to pull off his feat for three loosely related reasons. First, he was “bothered” by the problem of altruism; was the Darwinian explanation completely sound, or was it not? Second, he had a working knowledge of social insects, to which the altruism problem eminently applied. And third, he was intrigued by the mathematics of kinship, into which—impelled by the first two concerns—he had been guided through reading the work of the geneticist Sewall Wright. The closer the kinship, of course, the larger the fraction of genes shared as a result of common descent. Wright had devised an ingenious way of expressing the exact fraction shared, by a measure he called the coefficient of relationship. Working problems with it is an interesting mental game not unlike calculating the odds in gambling. What, for example, is the fraction of genes shared with a second cousin, or a half-sister’s full niece? This number, the degree of kinship, Hamilton saw to be crucial in the evolution of altruism. Even this tributary idea is intuitively straightforward. You may be willing to risk your life for a brother, for example, but the most you are likely to give a third cousin is a piece of advice.
With these points in mind, Hamilton now joined the natural history of wasps and other social insects with the calculus of kin selection. At this point he was aware of two more important pieces of relevant information affected by kinship, this time from entomology. One is that most social insects, including the ants, bees, and wasps, are members of the insect order Hymenoptera. The only exceptions are the termites, composing the order Isoptera. The other important fact is that the Hymenoptera have an unusual sex-determining mechanism called haplodiploidy, in which fertilized eggs, with two sets of chromosomes, produce females, and unfertilized eggs, with only one set of chromosomes, produce males. Turning to the coefficient of relationship (or the “concept of relatedness,” as he later named it), Hamilton saw that because of haplodiploidy sisters are more closely related to each other—have more genes in common—than are mothers and daughters. At the same time, they are much less related to their brothers. From the occurrence of haplodiploidy alone, he concluded that all of the following should be true if social behavior has evolved in the insects by natural selection.
The Hymenoptera should have given rise to many more groups of social species than other orders, very few of which are also haplodiploid.
The worker caste of these species should always be female.
In contrast, the males should be drones, contributing little or no labor to the colony and receiving little attention from their sisters.
All these inferences are in fact true, and they admit of no easy explanation except kin selection biased by haplodiploidy.
I first read Hamilton’s article during a train trip from Boston to Miami in the spring of 1965. This mode of travel was habitual for me during these years, the result of a promise to Renee that I would avoid trips by air as much as possible until our daughter, Catherine, reached high school age. I found an advantage in the restriction. It gave me, in the case of the Miami run, eighteen hours in a private roomette, trapped by my pledge like a Cistercian monk with little to do but read, think, and write. It was on such journeys that I composed a large part of The Theory of Island Biogeography. On this day in 1965 I picked Hamilton’s paper out of my briefcase somewhere north of New Haven and riffled through it impatiently. I was anxious to get the gist of the argument and move on to something else, something more familiar and congenial. The prose was convoluted and the full-dress mathematical treatment difficult, but I understood his main point about haplodiploidy and colonial life quickly enough. My first response was negative. Impossible, I thought; this can’t be right. Too simple. He must not know much about social insects. But the idea kept gnawing at me early that afternoon, as I changed over to the Silver Meteor in New York’s Pennsylvania Station. As we departed southward across the New Jersey marshes, I went through the article again, more carefully this time, looking for the fatal flaw I believed must be there. At intervals I closed my eyes and tried to conceive of alternative, more convincing explanations of the prevalence of hymenopteran social life and the all-female worker force. Surely I knew enough to come up with something. I had done this kind of critique before and succeeded. But nothing presented itself now. By dinnertime, as the train rumbled on into Virginia, I was growing frustrated and angry. Hamilton, whoever he was, could not have cut the Gordian knot. Anyway, there was no Gordian knot in the first place, was there? I had thought there was probably just a lot of accidental evolution and wonderful natural history. And because I modestly thought of myself as the world authority on social insects, I also thought it unlikely that anyone else could explain their origin, certainly not in one clean stroke. The next morning, as we rolled on past Way cross and Jacksonville, I thrashed about some more. By the time we reached Miami, in the early afternoon, I gave up. I was a convert, and put myself in Hamilton’s hands. I had undergone what historians of science call a paradigm shift.
That fall I attended a meeting of the Royal Entomological Society of London (crossing on the Queen Mary) to give an invited lecture on the social behavior of insects. The day before my session I looked up Bill Hamilton. Still a graduate student, he was in some respects the typical British academic of the 1950s—thin, shock-haired, soft-voiced, and a bit unworldly in his throttled-down discursive speech. I found that he lacked the terminal digits of one hand, lost during the Second World War, when as a child he tried to make a bomb in the basement laboratory of his father, an engineer with experience in rock-blasting who invented bombs for the British Home Guard—for use in case of a German invasion. As we walked about the streets of London, rambling o
n about many subjects of common interest, he told me he had experienced trouble getting approval for his Ph. D. thesis on kin selection. I thought I understood why. His sponsors had not yet suffered through their paradigm shift.
The next day I devoted a third of my hour-long presentation to Hamilton’s formulation. I expected opposition, and, having run through the gamut of protests and responses in my own mind, I had a very good idea of what the objections would be. I was not disappointed. Several of the leading figures of British entomology were in the audience, including J. S. Kennedy, O. W. Richards, and Vincent Wigglesworth. As soon as I finished, they launched into some of the arguments I knew so well. It was a pleasure to answer them with simple prepared explanations. When once or twice I felt uncertain I threw the question to young Hamilton, who was seated in the audience. Together we carried the day.
The time was approaching to write a synthesis of knowledge about the social insects. I dreamed of spinning crystal-clear summaries of their classification, anatomy, life cycles, behavior, and social organization. I would celebrate their existence in a single well-illustrated volume. A work of this magnitude had not been attempted in thirty-five years, the last being Franz Maidl’s rather opaque Die Lebensgewohnheiten und Instinkte der staatenbildenden Insekten, and was badly needed. The literature was scattered through hundreds of journals and books, in a dozen languages, and it varied enormously in quality. The study of social insects had been balkanized for a hundred years: experts on ants seldom spoke to those on termites, honeybee researchers lived in a world of their own, and students of halictine bees and social wasps cultivated their subjects to one side as minor arcane specialties. I wanted to create a showcase for sociobiology using insects and, in so doing, demonstrate the organizing power of population biology. That much, I believe, my book accomplished. The Insect Societies, published in 1971, conveyed my vision of the social insects and, in the final paragraph, I looked to the future:
Naturalist 25th Anniversary Edition Page 29