One of the myths of ancient Greece tells of Caenis, “loveliest of the maids of Thessaly,” who, while walking alone on an isolated shore, was spied by Poseidon—god of the sea, elder brother of the king of the gods, and sometime rapist. Mad with lust, the god attacked her on the spot. Afterwards, he took pity, and asked what he might give in reparation. Manhood, was her answer. She wished to be transformed into a man—not just any kind of man, but one extravagantly male, a warrior and “invulnerable.” Then she would never again be subjected to such a humiliation. Poseidon agreed. The metamorphosis was completed. Caenis became Caeneus.

Time passed. Caeneus fathered a child. With his sharp and expertly wielded sword he killed many. But the swords and spears of his adversaries could not penetrate his body. The metaphor here is not hard to fathom. Eventually Caeneus became so full of himself that he scorned the gods. He erected his spear in the marketplace and made the people worship it and sacrifice to it. He insisted, on pain of death, that they worship no other gods. The symbolism is again lucid.

Extreme arrogance, of which this is a fair example, was called by the Greeks hubris. It was almost exclusively a male trait. Sooner or later it would attract the attention and then the retribution of the gods—especially toward those humans insufficiently deferential to the immortals. The gods craved submission. When news of Caeneus’s effrontery finally reached Zeus, whose desk was doubtless piled high with such casefiles, he ordered the centaurs—chimeras, half-man, half-horse—to execute his merciless judgment. Dutifully they attacked Caeneus, taunting him: “Do you not remember at what price you gained this false appearance of a man? … Leave wars to men.” But the centaurs lost six of their number to Caeneus’s swift sword. Their lances bounced off him “like a hailstone from a roof.” Disgraced at being “conquered by an enemy but half-man”—a hollow complaint, coming from a centaur—they resolved to smother him with timber, destroying vast stands of trees “to crush his stubborn life with forests for our missiles.” He had no special powers concerning breathing, and after a struggle they managed to subdue and then to suffocate him. When the time came to bury the body, they were amazed to find that Caeneus had reverted back to Caenis; the invincible warrior had become, once again, the vulnerable young woman.³

Perhaps poor Caenis had overdosed on the stuff that Poseidon used to effect the metamorphosis. There is a proper amount of whatever it is that makes one male, the ancient Greeks recognized, and too much or too little can get you into trouble.

——

The testicles of a sparrow are about a millimeter long and weigh about a milligram. (That’s one of the reasons you never hear that someone’s hung like a sparrow.) With testes intact, the scrappy birds enter into their mainly linear hierarchy, chase away other birds who invade their territory, and, if they’re high-ranking, make successful overtures to fertile females. But reach under those feathers, remove those two tiny organs, and, after the bird has recovered, all of these traits are lost, or nearly so. Aggressive birds become submissive, territorial birds become complacent about intruders, passionate birds lose interest in sex. Now inject a certain steroid molecule into the sparrow and it regains its plucky enthusiasm for sex, aggression, dominance, and territoriality.

Shortly after castration, male Japanese quails stop strutting, crowing, and copulating. They also fail to elicit the interest of female Japanese quails. Treat them with that same steroid and they’re back to strutting, crowing, and copulating, and the females find them irresistible once more. Castrate a young male fiddler crab and he will never develop his distinctive asymmetrical giant claw.

Humans have understood some of this for thousands of years. Captured warriors were castrated so they’d make no trouble. We still describe an ineffective leader as a “political eunuch.” Chieftains and emperors castrated men so they could guard the harems without succumbing to temptation (or at least—the compromise sometimes reached—without impregnating any of the residents); and so their loyalties to the leader would not be adulterated by family ties or other distracting affections and obligations. It is remarkable that almost exactly the same molecule should produce such fundamental changes in behavior in sparrows, quail, crabs, and humans.

The steroid molecule that works these transformations like some wizard’s potion is testosterone. Along with other, similar molecules, it’s called an androgen. It’s manufactured (from, of all things, cholesterol) mainly in the testicles,⁴ enters the bloodstream, and induces a complex set of behaviors that we recognize as characteristically male. Here too, the connection is acknowledged in the language, as in the expression “He’s got balls”—meaning he’s shown exemplary courage and independence, he’s not a coward or a sycophant.

In newly formed groups of male monkeys, the higher the rank in a forming dominance hierarchy, the more testosterone is found to be circulating in the blood. But when the hierarchy settles down to symbolic encounters, and betas are routinely submitting to alphas, the correlation vanishes. The more testosterone an animal has, the farther away he’s willing to roam to challenge and dominate potential rivals.⁶ With high testosterone levels there’s a cross-species tendency for dominance within the group to be extended to dominance over a piece of territory. The boss and the landlord become one.

In the brains of many animals are specific receptor sites to which the testosterone molecule and other sex hormones chemically bind, and which are in charge of hormone-induced behavior. There may be separate brain centers responsible for strutting, crowing, bullying, fighting, copulating, defending territory, and fitting into the dominance hierarchy; but each center has a button pushed by testosterone. The behavior is actuated once the testosterone migrates from the testicles through the blood to the brain. In the individual brain cells, the presence of testosterone activates otherwise untranscribed and ignored segments of the ACGT sequence, synthesizing a set of key enzymes. As with many hormones, testosterone is at the nexus of an array of positive and negative feedback loops that maintain the concentration of the molecule circulating in the blood.

Male animals don’t just endure, but seem to delight in, testosterone-mediated scuffles, intimidation, and combat. Mice will learn to run a complex maze when the only reward or reinforcement is the opportunity to have a tussle with another male. There are abundant similar examples in our species. Activities that are central to leaving many offspring tend to be entered into with enthusiasm. Sex itself is the most obvious example. Aggression is in the same category.

Even among animals with very short gestation periods, such as mice, the delay between conception and birth is too long for the animal to associate cause and effect. To leave it to mice to figure out the connection between copulation and the creation of the next generation is to condemn their genes to extinction. Instead there must be an absolutely overwhelming need for sex and—as a means of reinforcement—a delight in partaking of it. This is just the DNA creatively demonstrating its control in the most overt and clear-cut way.

A deal has been struck: The animal will forgo food, will conform to extreme postural indignities, will put its very life at risk so its strands of DNA can join up with the strands from some other animal of the same species. In exchange, there will be a few moments of sexual ecstasy, one of the currencies in which the DNA pays off the animal that carries it around and nurtures it. There are many other examples of DNA-mediated delight in activities tending toward adaptive fitness—including parental love for children, joy in exploration and discovery, courage, camaraderie, and altruism, as well as the standard array of testosterone-driven traits making bosses and landlords.

Hormones similar to testosterone play a central role in the development, of sexual organs and sexual behavior all the way down to the aquatic fungi. Steroids must have evolved very early to be so widely distributed today, perhaps going a fair way back to the invention of sex around a billion years ago.

This trans-species use of the same molecule for roughly the same sexual purpose has some bizarre consequences. For example, the chief sex pheromone in the pig is 5-alpha-androstenol—chemically similar to testosterone. It’s mixed in with the boar’s saliva (as testosterone is present in the spit of men). When a sow in heat smells this steroid on a slavering boar, she promptly adopts the come-hither mating posture. Oddly, truffles, the French culinary delight, produce exactly the same steroid and in a higher concentration than in boar spit. This seems to be why pigs are used by gastronomes to find and unearth truffles. (How strange it must seem to the sows, always falling in love with little black pieces of fungus, only to have them cruelly snatched away by humans.) Since truffles are fungi, in which steroids play key sexual roles, perhaps tormenting sows is just an accidental side-effect—or perhaps it serves the function of inciting pigs to dig so the spores are spread more widely and the Earth is covered with truffles.

Now in light of all this, what are we to make of the fact that 5-alpha-androstenol is also copiously produced in the underarm perspiration of men?⁷ Long ago—before institutionalized hygiene, before the present perfumed and deodorized age—might it have played a part in human and prehuman courtship and mating behavior? (The noses of women, we cannot help noting, are often at the same level as the armpits of men.*) Might this have something to do with the willingness of the rich to spend exorbitant sums on tiny pieces of a nearly tasteless cork-like substance?

A genetically male embryo deprived of testosterone and other androgens will emerge with what look very much like female genitals. Conversely, the genitalia of a genetically female embryo subjected to high levels of testosterone and other androgens will be masculinized: If smaller amounts of the steroid are present, perhaps she’s born with only a somewhat bigger clitoris; if larger amounts, her clitoris becomes a penis, and her labia majora fold over to become a scrotum. She may develop a normal-looking male penis and scrotum, although the scrotum will have no testicles within. (She’ll also have nonfunctioning ovaries.) Such girls as they grow up are found to prefer guns and cars to dolls and kitchen supplies, boy to girl playmates, and enjoy rough-housing and the outdoors; they may also find women sexually more attractive than men.⁸ (There’s no evidence for the converse—for example, that most tomboys have excessive amounts of androgens.)

The difference between male and female, not genetically but on so fundamental a matter as which set of external genitalia you are to have, depends on how much male steroid you encountered in the first few weeks after conception. Leave that bit of developing embryonic tissue alone and it will become a female. Suffuse it with a little testosterone-like hormone and it will become a male, † The tissue is spring-loaded to respond to the androgen (the word literally means “male maker”), which serves as a means of internal communication. There are buttons on the developing embryo that only androgens can push. Once they are pushed, substantial machinery, whose existence you might otherwise never have guessed, takes over and works mythic transformations.

Across widely different animal species, another class of sex hormones, the estrogens, curbs aggressiveness in females, and yet another, progesterone, increases the feminine inclination to protect and care for the young. (The words signify, respectively, something like estrus-maker and gestation-promoter.) Mother rats, as all mammals, are attentive to their offspring: They build and defend nests, nurse the pups, lick them clean, retrieve them when they wander away, and teach them. None of this behavior is evident in virgin females, though, who studiously ignore newborn pups, or even make some efforts to avoid them. But prolonged treatment with the female hormones progesterone and estradiol—bringing the hormone levels of virgins up to those typical of late pregnancy—results in the emergence of marked maternal behavior. Rats with high levels of estrogen are also less anxious and fearful and less likely to engage in conflict.¹⁰

These female hormones are produced mainly in the ovaries. But when we see a calm, competent, and loving mother, most of us are not driven to exclaim “Man, has she got ovaries!” The reason doubtless has something to do with the ready accessibility of testicles for accidental or experimental removal, dangling as they do in vulnerable external sacs*—quite differently situated than the ovaries, which are locked away for safekeeping within the vault of the body. But clearly ovaries must equally be counted as among the family jewels.

The female hormones control the estrus cycle—which culminates when the females are ovulating and, usually, broadcasting olfactory and visual cues that they’re available for mating. In many species this doesn’t happen often and doesn’t last long; cows, for example, are interested in sex for about six hours every three weeks. Cows don’t date much. “For most species,” writes Mary Midgley,¹¹ “a brief mating season and a simple instinctive pattern makes of it a seasonal disturbance with a definite routine, comparable to Christmas shopping.” In a wide variety of mammals, from guinea pigs to small monkeys, mating outside of estrus is not only discouraged by the female, it’s also made physically impossible by an organic chastity belt: The vagina is sealed by a membrane or plug grown specially for the purpose, or—even more decisive—it’s fused shut.

In contrast, among most humans and some apes, sex is not only possible but is equally probable at virtually any phase of the cycle. Some humans monitor the cycle (by measuring small changes in body temperature) and then avoid sex around the time of ovulation. This Church-condoned contraceptive technique is the mirror image of the practice of most animals—who garishly advertise ovulation and avoid sex at all other times. It is a reminder of how far from our ancestors our culture has taken us, and what fundamental changes in us are possible.

For many animals the ovulation cycle is a few weeks in length. Not many species have periods almost exactly equal to the lunar cycle (the time from new moon to new moon). Whether this peculiarity of humans is more than a coincidence—and if so, why it should be—is unknown.

Mammals suckle their young, but only the females are appropriately endowed.* It’s one of the few cases where the definition of a major classification category in biology, or taxon, is determined by the characteristics of only one of the sexes. Giving milk is also hormonally mediated. Mother’s milk is essential for the young, who are born helpless, unable to digest the adult diet. This is another reason that females spend more time with, and therefore have a greater investment in, the young. The males are generally more interested in other things—dominance, aggression, territoriality, many sex partners.

The connection between steroids and aggression applies with surprising regularity across the animal kingdom. Remove the principal source of sex hormones and aggression declines, not just among the mammals and birds, but in lizards and even fish. Treat castrated males with testosterone and the aggression returns. Give estrogen to intact animals and aggression diminishes, again across all these species. The repeated use of these same steroids for the same functions, turning aggression on and off, for so many different animals, is a testament both to their effectiveness and to their antiquity.

Aggression is adaptive, but only in controlled amounts. The repertoire of aggressive behavior is on call, awaiting only to be disinhibited. The steroids, their production titrated by the social environment and the biological clocks, do the disinhibiting. This being the case, why is it that males are so often more aggressive than females? If the females can generate a little less estrogen and a little more testosterone, can’t they become as aggressive as males? Something like gender equality in aggression occurs in wolves, tree squirrels, laboratory mice and rats, short-tailed shrews, ring-tailed lemurs, and gibbons. In the southern flying squirrel, males are not territorial but females are, and most quarrels between the sexes are initiated by the females—and won by them.¹³ The clear fact that males are more aggressive than females among us humans (where blood plasma testosterone is about ten times greater in men than in women) by no means commits the rest of the animal kingdom, or even the rest of the primates, to the same arrangement.

As anyone knows who has seen a pet tomcat drag himself home after an absence of a day or two—with an eye closed, an ear torn, his fur matted and bloody—testosterone exacts a price. What happens if you take a male animal—let’s say, someone less combative than tomcats out for a night on the town—and equip him with an implant that keeps his testosterone blood levels high? When this is done to sparrows, hardy territorialists, there seems to be no significant increase in the sparrow murder rate. But when male cowbirds are implanted, their numbers markedly decrease;¹⁴ many birds are now observed with unusually serious injuries, clearly obtained in combat with their fellows. Unlike sparrows, cowbirds establish dominance hierarchies but do not have core territorial refuges into which they can flee. Bluff can escalate into serious fighting if you’re simultaneously charged up with testosterone and have no tradition of sanctuary. Another steroid deficit: Male birds with artificially high testosterone levels are less inclined to feed their hatchlings.¹⁵ Macho males tend to neglect their family responsibilities.

Sex hormones are now manufactured by pharmaceutical companies, and widely used—legally and illegally. We can learn something about their role in Nature by asking why people use them. Anabolic steroids are molecules very like, but usually not identical to, testosterone. They’re taken mainly by: (1) bodybuilders and athletes (who widely believe that certain feats of strength can be accomplished only by young men on steroids); (2) young men who wish to macho up, usually to attract women or other men; and (3) those who wish to disinhibit their meanness (nightclub bouncers, hit men in organized crime, prison guards, and so on).¹⁶ The enhanced musculature does not come about through steroids alone; it also requires vigorous and systematic exercise. One of the side effects is facial and back acne. Anabolic steroids don’t seem to grow hair. Large doses lead to dysfunction and atrophy of the testicles—perhaps the body’s response to excessive testosterone titers; too much testosterone is socially sufficiently dangerous that a mechanism may have evolved so that tendencies toward excessive production aren’t passed on to future generations.

Estrogen is taken by women, usually post-menopause or post-hysterectomy, to preserve sexual interest and lubrication, to slow loss of bone calcium, and to achieve a more youthful complexion. Bodybuilding and transsexual women may take anabolic steroids because they strikingly redistribute weight—from thighs to chest and biceps, for example. Transsexual men taking estrogen redistribute weight the other way, grow breasts, and feminize the nipples and areolae; there’s also a general mellowing of temperament. Bearing in mind these consequences of taking sex hormones as an adult, and the much more profound influence they have on the embryo—actually determining which sexual organs will be present—it seems likely that far subtler changes in hormone levels might influence not just dominance, territoriality, aggression, care for the young, gentleness, anxiety level, and talent for conflict resolution, but also sexual appetite and preference.

——

Bulls, stallions, and roosters are made into steers, geldings, and capons because humans find their machismo inconvenient—the very same male spirit that the castrators likely admire in themselves. One or two skilled motions of the blade—or a deft bite by a reindeer-herding Lapp woman—and the testosterone levels are down to manageable proportions for the rest of the animal’s life. Humans want their domestic animals to be submissive, easily controlled. Intact males are an awkward necessity; we want just enough of them to father a new generation of captives.

Something similar although less direct happens within the dominance hierarchy. From pit vipers to primates, the loser in ritual combat often experiences a steep decline in testosterone and related sex hormones, making him less likely to challenge the leadership at a later time, and therefore less likely to be injured. On a molecular level, he’s learned his lesson. With fewer circulating steroids, he’s now less ardent in his pursuit of females—at least when high-ranking males are around. This also is to the liking of the alphas. Again, decreases in testosterone levels following defeat are usually much more marked than any increases following victory.

Back to the testicles of sparrows: In a breeding area each little piece of territory has a male sparrow who will defend it against all comers.* Suppose a meddling ornithologist captures one of these territorial males and removes him from the territory. What happens? Other males from adjacent areas—many of them not previously able to defend a territory—move in. Of course they have to threaten and intimidate before they’ll be taken seriously. So the general level of sparrow anxiety rises, both among the newcomers and among unreplaced sparrows in adjacent territories. Political tensions become high. If now we monitor the bloodstreams of the sparrows in the course of their disputes (which from our point of view, of course, seem petty, but to them it’s Quemoy and Matsu), we find that everyone’s testosterone level has risen—the newly introduced males who are trying to establish their territories, and the males of neighboring territories who are now required to do more in the way of defending than has been their recent practice. Something similar is true for many animals.

Those who have more testosterone, by and large, become more aggressive. Those who need more testosterone, by and large, generate it. Testosterone seems to play a vital role as both the cause and the effect of aggression, territoriality, dominance, and the rest of the “boys-will-be-boys” constellation of male behavioral traits. This seems to be true for widely differing species, including monkeys, apes, and humans.

In springtime, stimulated by the increasing day length, the testosterone level in male perching birds and songbirds (such as jays, warblers, and sparrows) goes up; they develop plumage, unveil a scrappy temperament, and begin singing. Males with larger repertoires breed earlier and produce more chicks. The repertoires of the most attractive males range up to dozens of distinguishable songs. Musical variety is the means by which more testosterone is converted into more birds.

When eggs are being laid, the male testosterone level remains high; they’re protecting their mates. Once the females begin incubating the eggs and are uninterested in sexual advances, male testosterone levels fall. Suppose that the females are now given estrogen implants so they remain sexually alluring and receptive, despite their new maternal duties. Then the testosterone levels in the males remain high. As long as the female is sexually available, the male is inclined to be nearby and protective.¹⁷

These experiments suggest that an important selective advantage may accrue if a species breaks out of the estrus constraint. Continuous female sexual receptivity keeps the male around for all sorts of useful services. This is just what seems to have happened—maybe through a small adjustment in the DNA code for the internal estrogen clock—in our species.

Testosterone-induced behavior must be subject to limits and constraints. If it were carried to counterproductive lengths, natural selection would quickly readjust the concentration of steroids in the blood. Testosterone poisoning to the point of maladaptation must be very rare. In nectar-eating birds, bats, and insects it’s possible to compare the energy expended in male steroid-driven defense against poachers with the energy that could be extracted from the flowers being guarded.* In fact, territoriality typically turns on only when the energy benefit exceeds the energy cost, only when there are so few delectable flowers to suck that it pays for you to expend the effort to chase away the competition. Nectar-eaters are not rigid territorialists. They won’t fight all comers to protect a wasteland of stones. They make a cost-benefit analysis. Even in a rich garden of nectar-bearing flowers, often no territorial behavior is seen in the morning—because plentiful nectar has been accumulating at night when the birds were asleep. In the morning, there’s enough to go around. Toward noon, when birds from far and wide have been feeding and the resource begins to get scarce, territoriality turns on.¹⁸ Wings outstretched, beaks lunging, the locals drive away the intruders. Maybe they feel they’ve been nice guys long enough, but now they’ve had it up to here with these foreigners. Fundamentally, though, it’s an economic, not a patriotic decision; practical, not ideological.

——

Many animals may do it, but at least among rats and mice it’s well-demonstrated: Fear is accompanied by a characteristic odor, a fear pheromone, easily recognized by others.¹⁹ Often, as soon as they sense you’re afraid, your friends and relatives run away—useful for them, but not very helpful for you. It may even encourage the rival or predator who has prompted your fear in the first place.

In the heads of goslings and ducklings and chicks at the moment they peck their way out of the egg is, a classic experiment suggests, a rough knowledge of what a hawk looks like. No one has to teach it to them. Hatchlings know. They also know fear. Scientists make a very simple silhouette—cut out of cardboard, say: There are two projections which could be wings. They flank a body which is longer and rounded at one end and shorter and stumpy at the other. If the silhouette moves with the long projection first, it looks like a flying goose, wings spread, long neck preceding. Move the silhouette overhead, neck first, over the hatchlings and they go about their business. Who’s scared of a goose? Now move the same silhouette stumpy end first—so it looks like a hawk with wings outstretched and long tail trailing—and there’s a flurry of peeps and trepidation. If this experiment has been properly interpreted,²⁰ somehow, inside the sperm and the egg that made that chick, encoded in the ACGT sequence of their nucleic acids, there’s a picture of a hawk.

Perhaps this inborn fear of raptors is akin to the fear of “monsters” that almost all babies manifest around the time they become toddlers. Many predators who are circumspect when a human adult is around would happily attack a toddler. Hyenas, wolves, and large cats are only a few of the predators that stalked early humans and their immediate ancestors. When the child begins to amble off on its own, it helps for it to know—in its marrow—that there are monsters out there. With such knowledge, it’s much more likely to come running home to the grown-ups at the slightest sign of danger. Any mild predisposition in this direction will be resoundingly amplified by selection.*

In grown-up chickens there’s a set of more organized and systematic responses, including specific auditory alarm calls that alert every chicken within hailing distance of the ominous news: A hawk is overhead. The cry announcing an aerial predator is distinctly different from that announcing a ground predator—a fox, say, or a raccoon. Since the bird sounding the alarm is also giving away its presence and location to the hawk, we might be tempted to consider it courageous, its behavior evolved through group selection. An individual selectionist might argue—how convincingly is another matter—that the cry works to stir other chickens into motion, whose scurrying might distract the hawk and save the bird that sounded the alarm.

Experiments by the biologist Peter Marler and his colleagues²¹ show that, at least among cockerels, a propensity to make alarm calls depends very much on whether there’s a companion nearby. With no other bird present, the cockerel may freeze or gaze up into the sky when seeing something like a hawk, but he doesn’t cry out in alarm. He’s more likely to sound the alarm if there’s another bird within earshot; and, significantly, he’s much more likely to cry out if his companion is another chicken—any chicken—rather than, say, a bobwhite. He’s indifferent to plumage, though; chickens with very different color patterns are worthy of being warned. All that counts is that the companion be another domestic fowl. Maybe this is just sloppy kin selection, but it certainly edges toward species solidarity.

So is this heroism? Does the cockerel understand the danger he subjects himself to, and then, despite his fear, bravely cry out? Or is it more likely that squawking when there’s a companion nearby but not when you’re alone is a program in the DNA, and nothing more? See a hawk, see another chicken, cry out, and no agonizing moral struggle. When one of the combatants in a cockfight continues, although bleeding and blinded, to fight to the death, is he displaying “invincible courage” (as an English admirer of cockfighting has described it), or is this just a combat algorithm gotten out of hand, escaping the inhibition subroutines? Indeed, in humans does the hero have a lucid grasp of the danger, or is he or she merely following one of our preprogrammed subroutines? Most heroes report that they just did what came naturally, without much conscious thought.

The two sexes are not equally likely to produce alarm calls. In another study by Peter Marler and his colleagues,²² cockerels cried out in alarm every time a hawk silhouette was presented; but hens made such calls only 13% of the time.* Castrated cockerels are much less likely to sound the alarm—except when they have testosterone implants, in which case the call rate goes back up. So testosterone plays a role not just in dominance hierarchies, sex, territoriality, and aggression, but also in providing early warning of predators, whether we hold the bearer to be hero or automaton.

——

Preadolescent female mice have a molecule in theirurine that induces testosterone production in males who get a whiff. In turn, the males’ urine now contains pheromones which, when sniffed by the immature female, quicken her sexual development. She matures early if there are males around, and late if there aren’t—a positive feedback loop that saves unnecessary effort. (As you might expect, female mice who can’t detect odors never come into heat.) What’s more, normal pregnant females who sniff the urine from males of a different strain of mice spontaneously abort their pregnancies; they resorb the embryos back into their bodies and quickly come into heat.²³ This is convenient for the alien males. If the resident males don’t like it, it’s up to them to stop strangers from coming around with their abortion-inducing aromas.

In mice, as for many other animals, testosterone begins to be manufactured in earnest at puberty, and that’s when serious aggression against other mice begins. In adult males, the more testosterone, the quicker will be the attack when a strange male appears at the territorial frontiers. Again, castrate the males and their aggressiveness declines. Again, deliver testosterone to the castrates and their aggressiveness increases. Male mice are given to “marking” their environment with tiny dribbles of urine—a practice they pursue with redoubled effort when other mice are around (or when they come upon some unfamiliar object, maybe a hairbrush). Because of embryo resorption, if the males are to leave progeny at all, they must be the chief urinators in their territory. Maybe marking is like nametags on luggage, “no trespassing” signs on private property, or heroic portraits of the national leader in public places. The doughty little mouse is singing “This land is my land” and “She belongs to me.” Even when he’s not physically present he wants passersby to take careful note of his proprietorship. As you might suspect, castrate the mouse and urinary marking declines strikingly; resupply testosterone and his compulsion to mark is rekindled.

Normal female mice are infrequent urinators. They are not inveterate markers. But what happens if anatomically normal female infants are jolted with testosterone? Then they begin marking often. (If a similar experiment is done in dogs, adult females who were given testosterone before birth adopt the urination posture of the males; they lift one leg and trickle the urine down the other—one more indignity visited at the hands of the scientists.) When female rats with ovaries surgically removed are supplied with testosterone, they become aggressive, alternating a masculine propensity for confrontation with distinctly feminine sexual behavior. But one thing about giving testosterone to normal females early in their lives: When they grow up, the males find them much less attractive.

While testosterone in the blood is intimately connected with the expression of aggression in male animals, it is by no means the whole story. There are, for example, molecules in the brain that repress aggression. Hereditary strains of rats that are unusually violent turn out to have less of these inhibitory brain chemicals than more peace-loving strains. Aggressive rats are calmed when there are more of these chemicals in their brains; peaceful rats are agitated when there is less of these chemicals. If you’re a rat, busy watching violence in other rats—mice-killing, say—your level of inhibiting brain chemicals drops.²⁴ You’re now more likely to be violent yourself, and not just toward mice. Your repressed aggressive tendencies have been disinhibited. And everybody else’s. Hostility can then rapidly spread through your group, expressed differently by different individuals. Perhaps that’s what happened with Calhoun’s rats, so confined that aggression and despair spread in waves, reflected and amplified from multiple foci through the community. Violence is contagious.

In experiments performed by Heidi Swanson and Richard Schuster,²⁵ rats were given a complex cooperative task to learn, having to run together over specific floor panels in a particular sequence. If they succeeded, they were rewarded with sugar water; if they didn’t, they found themselves racing around the experimental chamber for the fun of it. Nobody taught them what to do, or at least not directly. It was trial and error. The experiment was tried on pairs of males, pairs of females, pairs of castrated males, and pairs of castrated males with testosterone implants. Some of the rats had previously lived alone.

Here’s how it turned out: Females, as well as male castrates, learned fairly quickly. Normal males and castrates with administered testosterone learned much more slowly. Males who had previously lived alone did still worse. Some pairs of previously solitary male rats—pairs with intact testicles as well as pairs of testosterone-jolted castrates—never learned at all.

For the solitary males this is just what you might expect: Because you live alone you have little experience in cooperating, so probably you’re not going to do very well on a demanding test of cooperation. But then, why should females who’ve been living alone be able to figure it out? The answer seems to be that if you’re a solitary male, a loner, and you have to perform a complex task in coordination with someone else, testosterone makes you stupid. Every pair of males who ordinarily lived alone and couldn’t figure out how to pass the test was engaged in violent combat. Communal living, by contrast, tended to calm them down.

Swanson and Schuster conclude that the learning deficits were not so much due to aggression per se, as to aggression in the context of the dominance hierarchy. Those who tended to be the winners in ritualized (or real) combat—almost always it was the same individuals—would strut and saunter with hair erect, threatening, feinting, and occasionally attacking. The subordinates would crouch, close their eyes, and either freeze for long periods or hide. But tendencies to strut or crouch or hide are not well suited for the gymnastic cooperation needed to get that sugar water.

Cooperation has strong democratic overtones. Extreme dominance/submission hierarchies do not. The two are strongly incompatible. In these experiments, females intimidated others and fought as did the males, but today’s winner was often yesterday’s loser, and vice versa—unlike the males. Cowering and freezing were less common, and the female style of aggression didn’t impede social performance as much as her male counterpart’s.

The unfolding richness and complexity of testosterone-induced sexual behavior—dominance, territoriality and all the rest—is one means by which males compete to leave more offspring. It’s not the only possibility. We’ve already mentioned selection at the level of competition among sperm cells, as well as those species in which the male leaves a vaginal plug when he’s done to frustrate those who come after him. Male dragonflies attempt to undo the competition retroactively: Projecting from the male’s penis is a whip-like prong that attaches itself to the mass of sperm previously deposited in the female. When he withdraws, he takes his rivals’ semen with him. How much more direct the dragonflies are than the birds and mammals—our males violent, consumed with jealousy, spitting out threats and accusations, longing for exclusive sexual access to at least one female. The dragonfly male is spared much of this; he merely rewrites his mate’s sexual history.

We’ve concentrated on aggression, dominance, and testosterone because they seem to be of central importance in understanding human behavior and social systems. But there are many other behavior-eliciting hormones fundamental for human well-being, including estrogen and progesterone in females. The fact that complex behavioral patterns can be triggered by a tiny concentration of molecules coursing through the bloodstream, and that different animals of the same species generate different amounts of these hormones, is something worth thinking about when it’s time to judge such matters as free will, individual responsibility, and law and order.

Had Poseidon more carefully measured out whatever it was he gave to Caenis, the matter would not have come to Zeus’ attention. Had Poseidon’s own testosterone titer been lower, or had there been enforceable penalties against gods raping humans, Caenis might have lived a happy and blameless life. As it was, Caeneus was afflicted by hubris, surely; but only because of the rape and its aftermath. He was guilty of disrespect for the gods, but the gods had shown disrespect for her. There is not a hint that the piety of Thessaly would have been troubled had Poseidon left Caenis alone. She had been minding her own business, walking along the beach.