The Price of Altruism (36 page)

Variation was on many people’s minds. At the University of Chicago and now at Harvard, Richard Lewontin had recently shown in a set of experiments with
Drosophila
flies that there is much more variation in natural populations than previously suspected.
¹⁹
It was a bomb dropped right at the center of evolutionary population genetics: If natural selection was, in Darwin’s words, “…daily and hourly scrutinizing, throughout the world, every variation, even the slightest; rejecting that which is bad and adding up all that is good…,” then shouldn’t variation be at a minimum? Most mutations were injurious, not beneficial, so would be expected to be summarily culled. To explain the apparent paradox, a Japanese theoretical geneticist, Motoo Kimura, posited a theory that soon became known as “neutralism”: The reason so much variation existed in nature, the theory argued, was that most of the variation was neither here nor there with respect to fitness. Absent a reason to cull, natural selection would therefore simply remain indifferent, leaving high levels of genetic variation to accumulate and drift freely in natural populations.
²⁰

At the Galton, George’s überboss, Harry Harris, was doing his best to fight the trend. If variation was neutral, what role then for natural selection? This was a serious matter: Unless a better explanation was found, Darwin’s great theory would need to be thoroughly reconsidered. A Darwinist at heart, Harris was marshaling data from human hemoglobin to argue that Kimura and his gang were mistaken. To do so convincingly, though, he’d have to propose a superior mathematical theory, and who better than George for the job? In truth, George was more interested now in the Passion Schedule and the Bible. Still, he shrugged his shoulders and complied.
²¹
After all, this theoretical challenge on genetic polymorphism might help clarify some issues on sex, and both could serve as further scaffoldings for his continued search for the origin of family. Besides, all his attempts at simulating animal combat on the computers had by this time utterly failed.
²²

He was working on a model of an optimal mutation rate: It was “far superior to Kimura’s,” he wrote to the recent Nobel laureate MIT economist Paul Samuelson, with whom he’d resumed a short correspondence.
²³
But then came a surprise. In the spring a letter from John Maynard Smith arrived in the mail. In it were computer printouts of John’s latest attempts to model animal combat. To George it was practically a godsend.

“Fascinating. Congratulations,” he replied, unable to hide his excitement, “it looks as though you’ve gotten well beyond the point I’ve reached.” Combat, after all, had always been close to his heart. Even better was the offer of joint authorship, though George would only accept if John’s name was listed first. At the moment he was trying to finish a paper of his own on “Sex and Rapid Evolution,” so wouldn’t have time to work on the combat paper just now; why didn’t John write a first draft for
Nature
that he could look over and comment on? As for the different behavioral strategies of the model’s animals in conflict, George suggested easily understood names like Dove, Hawk, and Prober.
²⁴

Once again the problem was how to explain why, in fighting over dominance rights, territory, or mates, animals often seem to stop short of actually hurting each other. George’s original paper had concentrated on antlers in deer, but this was a general problem. The males of many snake species, for example, were known to fight each other by wrestling without using their fangs, a kind of “benevolence” not unknown even to praying mantises. The most absurd case was that of the Arabian oryx with the rhyming name,
Oryx leucoryx
: Its extremely long horns, pointed so absolutely in the wrong direction, forced males of this species to kneel down with their heads between their knees in order to direct their horns forward. How could kneeling oryxes, wrestling snakes, and deer that refused to strike “foul blows” be explained? How had Nature, in her wisdom so infinitely greater than man’s, invented limited combat?

Maynard Smith and George took their cues from John von Neumann. For once again, as in poker, nuclear proliferation, or for that matter—Vietnam—animal conflict could be modeled as a game; the trick was to see that the strategy of each “player” depended on the other. Two contestants, for example, A and B, could adopt “conventional” tactics, C, that were unlikely to lead to injury or “dangerous” tactics, D, that were sure to lead to serious harm. And of course, they could simply retreat, R. A possible conflict between them could therefore look like this:

with A probing on the twelfth and twentieth moves, B retaliating after the first probe, then retreating, and losing out, after the second. Each contest ended with particular “payoffs” to each contestant, measures of the contribution the contest has made to the reproductive fitness of the individual. Spurred on by Maynard Smith, George returned to the computer in the summer of 1972. Together they programmed five distinct strategies, sets of rules that ascribe probabilities to the C, D, and R plays as a function of what happened in the previous moves. There were five such strategies: “Dove,” which never plays D; “Hawk,” which always plays D; “Bully,” which plays D if making the first move, D in response to C, C in response to D, and R following an opponent’s second D; “Retaliator,” which plays C on the first move, C in response to C, D in response to D, and R if the contest has lasted a preassigned number of moves; and finally “Prober-Retaliator,” which if making the first move or in response to C plays C with high probability and D with low (but R if the contest has lasted a preassigned number of moves), following a probe reverts to C if the opponent retaliates but takes advantage by continuing to play D if the opponent plays C. There were fifteen types of two-opponent contests, and John and George simulated two thousand contests of each using pseudo-random numbers generated by an algorithm to vary the contests. With fixed payoffs and probabilities that seemed biologically sound,
²⁵
the following matrix presented itself:

Average Payoffs in Simulated Intraspecific Contests for Five Different Strategies

To see whether a strategy is an ESS against all others, all one had to do is examine the corresponding column. In a population full of “Hawks,” for example, does “Hawk” do better than the alternatives?

“Dove” “Hawk” “Bully” “Retaliator” The table showed clearly that it didn’t: Both “Dove” (19.5) and “Bully” (4.9) did better against “Hawk” than “Hawk” against itself (-19.5). “Dove” too was not an ESS: “Hawk” (80.0), “Bully” (80.0), and “Prober-Retaliator” (56.7) averaged higher payoffs in a population almost entirely of “Dove.” In fact “Retaliator” alone was an ESS, since no other strategy did better (though “Dove” ties), and “Prober-Retaliator” came in a close second.

How would such a population be expected to evolve? The answer was that “Retaliator” and “Prober-Retaliator” types would increase in frequency at the expense of “Hawk,” “Dove,” and “Bully.” These last three types wouldn’t become extinct, though, but rather remain in low numbers due to a constant flow of mutation but also because of senility, youthful inexperience, injury, and disease. The balance between “Retaliator” and “Prober-Retaliator” would depend on the frequency of “Dove,” since probing was only an advantage against the meek: If the frequency of “Dove” was greater than 7 percent, “Prober-Retaliator” would replace “Retaliator” as the predominant type. The only way to make “total war” behavior advantageous would be to significantly alter probabilities for serious injury, or to give the same payoff penalty for retreating uninjured as for serious injury. Otherwise the simulations made abundantly clear that under individual selection “limited war” was superior to unbridled aggression.

Of course real animal conflicts in nature were infinitely more complicated. Besides the category distinction between “Conventional” and “Dangerous” behavior, which in all probability was more subtly graded, individuals varied widely in the intensity and skill with which each kind of tactic was employed. Still, by simplifying, the model allowed for certain predictions: The table, for example, showed that the best strategy against “Hawk” is “Dove,” or in other words immediate retreat. This meant that it would be to the advantage of a given individual to simulate wild, uncontrollable rage, since those encountering him would do best to get out of his way and yield. If that was the case, a pretend “pseudo-Hawk” type would soon arise, leading in turn to selection for a second type with a special talent to “call a bluff.” To counteract this dynamic, a true sign of maniacal craziness would therefore be expected to evolve, one that could not be easily counterfeited. And this, it seemed, was precisely what sometimes happened in the wild. Elephants “on musth,” for example, were often described by local peoples as invaded by wild spirits: As they rampaged uncontrollably, often with fatal results, a dark brown tarlike fluid secreted by the temporal glands ran distinctively down their faces. Combating deviousness, this was nature’s way of signaling that here was no charade: With due respect to fakery, some things just couldn’t be bought by guile.

These were important discoveries, George thought. Perhaps they’d prove useful to mankind.

He was happy: not only had Maddox informed Maynard Smith that
Nature
was interested, but he’d have his name on a pathbreaking paper alongside a world-renowned authority. True, the insight had been his. But this was an infinitely better outcome than his first, original attempt alone.

As exciting as it was, however, George’s heart remained elsewhere. He had no plans, he wrote to Kathy, to renew the SRC grant that was running out in July. Instead he was planning on spending the next few months working on a book about Shakespeare and the Bible.
²⁶
Most of all, he was finally giving in to fate.

Recently he had preached a sermon on Matthew 6:25–34:

He couldn’t stop thinking: He held an ordinary job, lived in an ordinary flat, dressed conventionally, was paid a regular salary. He’d
sort
of known, then, without
really
knowing it, that all this would have to end. “Wishful thinking had kept me supposing that it was not going to happen to me,” he wrote to Maynard Smith in October,

“I had optimistically calculated that deliverance had to arrive around the 20th of September in order to avert disaster. However it appears that God’s standard of what constitutes ‘disaster’ are on a different scale from mine,” he continued.