I began my career determined to study extreme weapons, so I set out to find the craziest, most bizarre animals that I could. I also wanted my research to take me someplace exotic. In my case, this meant the tropics, so I narrowed my search. My study animals needed to be easy to find in large numbers, to observe in the wild, and to rear in captivity. As fate would have it, the animals that best fit this bill were dung beetles. I resisted at first. After all, dung beetles lack the panache of elk or moose and, well, they eat dung.

In the process of weaving their histories together, it became clear that there was another species that belonged in the mix: humans. The more I sought common threads—themes uniting the stories of diverse animal species—the more apparent it became that these threads applied to our own weapons, too. In the end, my book about animal weapons evolved into a book about extreme weapons everywhere. I pored deeper into the literature surrounding our past, searching for the environments and circumstances in which our most elaborate weapons evolved. To my amazement, these circumstances truly were the same, and I realized I couldn’t tell one story without telling the other.

Unlike the rest of the body, which takes years to grow to adult size, antlers in even the largest bulls go from nothing to full size in just a few months. Antlers grow faster than any other bone in any animal, and this record speed racks up record energetic costs. Estimates from antlers of a related species, fallow deer, show that while males are growing antlers they more than double their daily energetic needs. In addition, growing antlers suck up so much calcium and phosphorus—minerals that make up the bone—that the males cannot possibly get enough from their food. Instead, they leach these vital minerals out of other bones and shunt them to the antlers, depleting the rest of their skeleton so severely that they experience a seasonal form of osteoporosis. Their bones get weak and brittle at precisely the time of the year—the rut—when they must hurl themselves against eight-hundred-pound rivals in incessant battles for access to females. By the end of the rut, the males will have fought so often and so hard that they've lost a quarter of their body weight, and they emerge from this season battered, starved, and brittle boned. If they cannot replenish their reserves in the few short weeks before winter, they'll starve.

Very specific factors must fall into place before weapon evolution launches into one of these races, and it turns out that the same special circumstances triggering arms races in animals also prompt humans to manufacture bigger and bigger weapons. … Analogous circumstances even bring about their collapse, as huge weapons come crashing down and the race dissolves. Ultimately, we'll see that our animal counterparts can teach us a surprising amount about ourselves.

Big weapons are ungainly, and most individuals fare poorly with them. For most weapons in most species, selection favors modest size and minimal cost—enough tooth to bite or grab prey, for example, but not enough to slow you down or impair your ability to maneuver. What this actually means is that selection on weapons is balanced: bigger weapons may be better for stabbing or biting, but they are also more expensive to produce and more difficult to carry. The result is a tenuous balance, much like a game of tug-of-war, which accounts for the multitude of unimpressive weapons adorning the vast majority of animal species.

Yet, every so often this balance is tipped. Here and there, sprinkled across the tree of life, are animal lineages where modesty of proportion was cast aside. The evolution of weapons in these species surged forward unencumbered. Gone were the shackles of balancing selection, and all that remained was selection favoring bigger and bigger weapon sizes. Here, individuals with the most grotesque or extreme armaments beat opponents with smaller weapons. In so doing, they secured opportunities to breed. Their progeny, as impressive in their weapons as their parents, quickly replaced earlier forms and advanced the population another notch in weapon size. As soon as the next new innovation arose—an even larger or more complex variant of the existing design—the process repeated itself. Again and again, individuals with the newest and biggest weapons won, replacing those with earlier and smaller forms, and this process pushed the population still further on the path to extreme. In short, the weapons of these species got caught up in an arms race.

The biggest animal weapons result from males competing with rival males over access to females. Part 2, "Triggering the Race," explores why this is true, and how competition leads to arms races. Competition drives rapid evolution of huge weapons, but only when two additional conditions are met. All three "ingredients" are necessary to trigger a race.

At some point in the past—possibly as recently as a few thousand years ago—mice spread into open areas along both the Gulf and Atlantic coasts, where they dug their burrows into sand dunes and grassy embankments. Beach mice now raced across a vastly different background than their inland ancestors, and in these new environs dark mice got plucked from the sand. By chance, some of the beach mice carried in their DNA new mutations to one or both of two genes involved in the production of dark pigments. Mice inheriting these mutations carried copies of the pigment-influencing genes that were just a little bit different from the copies carried by other mice (alternative versions of a gene are known as alleles), and as a result they developed with lighter fur. Mice bearing the new alleles survived better than mice inheriting the ancestral versions of the genes, and these survivors populated the beaches with their pups. Over time, mice with the new alleles increased in frequency, while those with the original alleles disappeared, and the result was an evolutionary shift from dark to white.

This may have solved logistical problems of production and distribution, but it also caused our troops to sometimes stand out when they were supposed to be blending in. After all, the solution with mice was two colors, not one, and the reality of diverse combat habitats is that no one pattern blends well in all places.

In the beginning (well, the first eighty thousand years of their one-hundred-thousand-year window), Nevada lake sticklebacks had almost no protective weapons (only one dorsal spine, rudimentary pelvic spines, and very few lateral plates). But then, eighty-four thousand years into the time sequence, this type of stickleback was replaced entirely by armored sticklebacks, meaning three long dorsal spines and full pelvic spines. Bell suspects that marine fish flooded into the lake around this time, because both forms co-occurred for about one hundred years before the early fish type disappeared. Remarkably, over the following thirteen thousand years, the defensive structures in this new fish regressed: in graded steps through time, the spines got shorter and shorter, until by the end of this period the new sticklebacks resembled the earlier form that they’d replaced. Lake-bound fish lost their weapons.

Of course, every story has exceptions. But with sticklebacks, the exception proves the rule. Dan Bolnick has been studying sticklebacks in Lake Washington, where fish have much bigger weapons than in other lakes. Bolnick found that this shift in armor happened very recently—fish samples collected prior to the 1960s had reduced armor typical of other lake sticklebacks. Efforts to stem pollution in this lake resulted in a dramatic improvement in the transparency of the water, and in these especially clear waters, introduced trout began to feed in earnest on sticklebacks. More predators translated almost immediately into bigger weapons.

In armor we see all of the processes that matter for the evolution of extreme weapons: individuals vary in the extent of their armament; these differences in weapon size affect the performance of their bearers (survival, growth, and reproduction in sticklebacks, and survival in soldiers); and, as a result, the sizes and shapes of these weapons evolve rapidly and dramatically over time. Weapons such as armor come at a price, and sometimes, when this cost is high enough, individuals with small weapons fare better than those with big ones. Indeed, most of the time, for most weapons, bigger is not better.

In hyenas, jaw-closing speed appears to have been traded for increased closing force. They have a tremendously powerful bite, and because of the shapes of their teeth, they use this bite for cracking bone rather than puncturing or slicing flesh.

Carnivore teeth aren't small because they couldn't or didn't evolve. They're small because individuals with unusually large teeth performed poorly when hunting their particular prey. Teeth and other major structures are almost always subject to trade-offs—a balance of opposing forces of selection. Bigger weapons may be better for killing prey, but they may also prevent an animal from catching prey in the first place.

Mammalian carnivores uncoupled the evolution of subsets of their teeth, so that each set evolved to function like a different tool. This way the mammalian jaw could carry three or four tools (for example, canines, molars, and premolars), and each could tackle a distinct task.

For example, theropod dinosaurs, notorious flesh-specialist predators (including Allosaurus, Carnotaurus, and the infamous Tyrannosaurus rex) had no obvious parallels to molars or premolars, no bladelike edges for shearing, no dome-shaped caps for crushing. Instead, all of their teeth were roughly similar in shape to canines. As a result, even though they did diverge in body size, allowing some partitioning of prey resources, theropod dinosaurs never diversified to fill the breadth of ecological roles seen in carnivorous mammals. In other words, there were no bone-cracking or saber-toothed theropods.

A survey of both living and extinct carnivores shows an astonishing frequency of natural tooth breakage, with one out of every four teeth chipped, cracked, or shattered.

Sit-and-wait, or ambush, predators take weapon evolution to an even greater extreme. Sabertooths were ambush predators who dropped from the branches of trees to plunge their daggers into the necks of unsuspecting prey. Like piranhas, ambush predators no longer chase after prey to hunt them down. In fact, most of them don't run or swim fast at all. Instead, they lurk motionless, often blending spectacularly with their backgrounds like a hunter in a blind, waiting for prey to come to them. When unlucky edibles happen by, these predators lunge from their hiding places, striking out with a snap of their jaws or a flick of their legs to snatch and incapacitate prey before they even recognize what is happening, much less have time to escape.

Most praying mantises are ambush predators, which explains their supersized forelegs with long curved spines. These insects get their name from a habit of holding their big forelimbs in front of their faces, a posture that resembles a person in prayer. In fact, these long raptorial limbs are spring-loaded with recoil and muscle, and their placement can be compared to the cocking of the hammer of a gun. These toothed limbs snap out from the body, grasping any prey that make the mistake of wandering into the "kill zone."

Mantis shrimp hurl their legs with incredible speed during these blows, which is nontrivial considering they are underwater. Sheila Patek and Roy Caldwell, studying the mechanics of the predatory strike of the peacock mantis shrimp, found that they accomplish this feat using a "click" mechanism in which a rigid latch locks the limb into a spring-loaded and cocked position. When this lock is released, the weapon snaps forward using elastic recoil from energy stored in the bending of the skeleton, exactly like releasing the recoil stored in a fully drawn archer's bow. Once released, their legs blaze through the water at strike speeds of up to sixty miles per hour which, when scaled down to the size of these "shrimp," means that the entire strike takes just two one-thousandths of a second.

The physics behind this last point is called the "law of the lever," and a good way to think about it is to consider a seesaw. If the pivot sits in the middle of the seesaw, then each end of the board moves the same total distance up or down, and each end moves through the air at the same speed (they travel the same distance in the same amount of time). But slide the pivot closer to one end of the board, and two things happen. The distance traveled by the two ends begins to differ, since the long end travels a more dramatic arc than the shorter end. And the speeds of the ends of the board diverge. Both ends of the seesaw complete their respective arcs in the same span of time (presuming that the board doesn't bend). But the long end travels farther than the short one, which means that the long end also travels faster than the short end of the board. The farther an object like a tooth sits from the hinge of a lever, the faster it moves when that lever rotates.

Carnivore jaws illustrate the other half of this principle. Cats and hyenas each sacrifice jaw speed in favor of jaw strength, with squat faces and canine teeth migrating relatively closer to the hinge. Here it helps to think of a nutcracker, or a pair of pliers: the closer an object is placed to the hinge, the more powerful the closing force that can be applied. Wolves, on the other hand, have longer jaws than hyenas or cats. Their canines bite with less force, but what is lost in force they gain in speed. Their canine teeth are situated farther from the hinge of the jaw and, as a result, move faster when the jaw closes.

Specific circumstances must fall into place before weapons launch into an arms race. Appreciating these “ingredients” reveals much about the function and diversity of nature’s most extravagant weapons, including why some species have them and most others do not.

In every way that matters for this book, jacanas are backward. Females are more aggressive than males, they are larger than males, they fight more viciously and frequently than males, and they have larger weapons. Usually it's the other way around. In flies, beetles, mastodons, crabs, and elk, males are armed, not females. Jacanas excepted, in every species with weapons confined to a single sex, males have those weapons. Why should just one sex have weapons? And why is it (almost) always the males?

Females of all animal species produce larger reproductive cells (called "gametes") than males. Eggs are bigger than sperm, and this difference in material investment is far more substantial than most of us appreciate. Humans are rather ordinary in this respect, but we're a good place to start. The female egg is the largest cell in the human body. It measures almost a fifth of a millimeter across—about the size of a period (.) on this page—and it's just visible to the naked eye. Sperm are the smallest cells in the body, and a hundred thousand could fit into the volume of a single egg.

Sexual selection differs from most forms of natural selection in ways tailor-made for pushing traits to their extreme. For one thing, sexual selection can be a lot stronger than natural selection. Whenever a small subset of males monopolizes access to large numbers of females, the disparity in reproductive success skyrockets. A few victorious males sire dozens or even hundreds of offspring, while the overwhelming majority of males sire none. When payoffs for success are high enough, weapons can evolve to really big sizes.

This is not how sexual selection works. In battles for access to reproduction, males compete with rival males. The environment that matters for performance is a social one—other males with whom a male does battle—rather than temperature or sea level or other physical features of the landscape. And this social environment evolves in tandem with the weapons. As antlers or horns become bigger, so, too, does the standard against which a male must contend. Like a sliding scale, each increase in weapon size resets the baseline of the population, selecting, in turn, for yet another increase.

Imagine a population of rhinoceros beetles where the average length of male horns is one-half inch. In this social environment, a few males stand out. A mutation increasing horn growth has given them three-fourths-inch horns. These males win battles the most frequently; they mate with the greatest proportion of female beetles; and they populate later generations with disproportionate numbers of their kind (including sons wielding three-fourths-inch horns). Over the next few dozen generations, the population shifts. It evolves in response to sexual selection so that now the average size of horns is three-fourths inches. … The benefits of a three-fourths-inch horn aren't so great anymore, because now everybody has horns this big. The evolutionary increase in horn length reset the standard for male competition. Into this new social milieu another mutation arises, this time leading to a one-inch horn. Males with these new alleles now have longer horns than their opponents, and they begin to win. So the new alleles sweep through the population as the largest horned males outcompete the earlier, three-fourths-inch rivals, until the population has evolved to this new weapon size. The scale has shifted once again. The population has ratcheted up to the new norm, and it stands poised and ready for the next mutation leading to still another increase in weapon size.

Most of the truly gargantuan armaments in the animal world owe their excesses to this form of competition. Reproduction is the "other half" of success, in an evolutionary sense (we've already looked at survival). But it's the half that really matters. When you strip the essence of life to its core, the only reason to survive is to have a shot at breeding, and, at the end of the day, it's how many offspring you produce that determines success or failure on the evolutionary stage. Individuals reproducing the most win, plain and simple. They contribute more copies of their genome to subsequent generations of the population than other individuals do. Their alleles persist, while others gradually disappear.

In a sense, males compete with rival males regardless of whether sexual selection proceeds through female choice or male competition and, in many respects, the process—the intensity, consistency, and social nature of selection—is the same regardless. Why, then, do some species embark on a trajectory of overt competition leading to the evolution of weapons, while others end up dancing or singing with displays? Here is where the ingredients for arms races come into play, for these are the pieces that must fall into place if sexual selection is to trigger evolution of extreme weapons. The first ingredient is competition, the essence of all sexual selection. The second is economic defensibility.

Imagine a food resource spread out uniformly across a landscape—grass, if you're a grazer, extending as far as the eye can see. As a male, where would you stand guard? Even if it were absolutely necessary for females to visit places with grass to feed, and even if they were willing to mate with you if you happened to be there when they fed, where would that location be? For food resources broadly distributed in space, there is no obvious location that is better than any other. It would be impossible for a male to anticipate where females were likely to visit since they could find their food anywhere. A male could still invest in the production of weapons, and he could use his weapons to deflect rivals from a patch of turf. But why should he bother to guard one particular area if all of the surrounding areas are just as good? And why should he pay the price of producing a weapon and fighting, if other males without weapons or territories do just as well as he does? In the parlance of economists, such behavior would not be cost-effective.

If, instead, those same food resources were sparse, and especially if they were clumped into rare but concentrated patches, then the male would face a very different set of payoffs for guarding a territory. He would still pay a price for producing the weapon, and for expending time and energy fighting to keep rival males out of his territory. But now these territories would matter, and the benefits he could glean from guarding them might be significant. Females would be much more likely to visit him, since the resources they needed were few and far between, and one of the only places where they could access them was inside his territory.

Females are not the only ones to approach males as they roll their balls away. Rival males constantly challenge one another over ball ownership, and vigorous battles are commonplace. But these fights occur out in the open on the exposed surface of the soil. Furthermore, the objects of these fights—balls of dung—are themselves mobile and malleable. Balls are pushed, pulled, even torn in half during fights as the males tussle around and around, clinging to and scrambling over the rolling balls. (These battles are great fun to watch, incidentally. At the Barro Colorado Field Research Station we'd paint numbers on the backs of rival males and place them on a dung pile centered in the bull's-eye of a horizontal dartboard, betting on the winner and cheering them on as they fought to roll their marble-sized balls out of the ring.) Despite their pugnacity, not one of the thousands of ball-rolling dung beetle species has horns.

A second strategy adopted by many dung beetles is tunneling. Females of these species fly into dung and immediately begin to excavate burrows into the soil below. Once they've dug sufficiently deeply—a foot to a yard, depending on the species—they begin pulling pieces of dung down into the tunnels to stash them away from the other dung-feeding insects above. Females may make fifty or more trips to bury sufficient dung to provision just a single egg, and they'll repeat the process for a string of successive eggs. While females are working on this arduous task, male beetles fight among themselves for tunnel ownership. A victorious male will guard the entrance to a tunnel—not so much to keep other species away from the food as to keep rival males of the same species away from the female. While in residence, a male will mate repeatedly with the female, but he will often get kicked out of the tunnel by an intruding, larger male. Males of tunneling species often have horns.

We've identified two of the three critical ingredients for arms races: intense competition, generally arising among males as they battle for access to females, and ecological situations that cause resources to be localized and economically defensible. There's one final ingredient, and it involves the details of the fights themselves—the way that males face each other in battle. Males must face each other one on one, rather than all together in a scramble. For bigger weapons to perform better than smaller ones, the battles must be matched and "symmetrical," with comparably armed contestants challenging each other face-to-face. … Oddly, this last ingredient has been almost entirely overlooked by biologists. To appreciate its significance, we must turn instead to attrition models of military forces, and to the century-old insights of an eclectic automotive and aeronautical engineer.

Lanchester designed his square law for modern warfare and, without question, the vast majority of interest has focused on that category of equations. But it's his linear law that is most relevant to the evolution of extreme weapons. By contrasting ancient and modern warfare, Lanchester helped define the circumstances in which large weapons will, and will not, be cost-effective. When opponents can concentrate their fire—gang up on an opponent simultaneously—investing in large weapons is probably a mistake. On the other hand, when soldiers duel each other at close range, face-to-face and one-on-one, then the better fighter is likely to win. Because fighting ability often depends on the size of a weapon, duels can lead to situations where bigger and bigger weapons prevail.

Burrows are probably the most widespread ecological situation leading to the evolution of extreme weapons. They are localized and readily defendable. Males can guard the entrances to tunnels where females reside, and in so doing, block rival males from approaching the females. But tunnels also physically restrict access in a way that aligns the interactions of opponents. A rival male dung beetle has to enter the tunnel before he can challenge the guarding male. Ten males couldn't attack at once even if they wanted to, because there isn't space for more than one rival to enter at a time. The restricted confines of tunnels align battles so that they necessarily occur as a series of successive duels. Ball-rolling species, on the other hand, face no such restricted access. Males can challenge from all angles at once, and very often battles among ball rollers entail chaotic scrambles between three or four males. In dung beetles, species that fight one-on-one often have elaborate horns; species that fight in chaotic scrambles do not.

This is why we often find not just a single species with huge weapons but entire clades packed with species after species all armed to the teeth. Inherited characteristics like asymmetrical parenting tip the balance in favor of arms races for all descendant members of the clade. All that is needed is for the remaining two pieces to fall into place. If many of these species also share another one of the ingredients, say an inclination for using habitats such as burrows that result in defensible resources or choke points, the balance is tipped still further. The result can be explosions of animal diversity, as species after species within these clades launch onto trajectories of rapid weapon evolution.

What these broad patterns reveal is a simple and surprisingly universal rule: after the final ingredient for arms races falls into place, entire clades of descendant species can all experience rapid evolution of extreme male weapons. Mastodons and flies could not be more different from each other. They lived at different times, in different habitats, and fed on different foods. One was more than 120 million times the size of the other. One had enlarged teeth, and the other had chitin protrusions from the forehead. Yet, the same three ingredients triggered evolution of extreme weapons in both cases. So it is with wasps, beetles, crabs, earwigs, elephants, and antelope. Despite extraordinary differences among these species, arms races are arms races, and the circumstances leading to big weapons are the same.

Resource allocation trade-offs shape the development of all animals, but most of the time the effects are trivial. When animals begin to invest unusually heavily into particular structures, however, the effects of trade-offs become more pronounced. Weapons caught up in arms races get very big very fast, and in these species resources channeled into growth of the weapons can drastically impair bodily functions. In insects, this sometimes means reduced growth of other body parts.

Not only are the claws bulky and heavy structures to build, they are energetically costly structures to maintain. Crab claws are not mere hollow threats; they are packed with powerful muscles that can crush the skeletal shells of rival males. Muscle tissue is incredibly energy demanding, because muscle cells are loaded with dense concentrations of mitochondria, the microscopic organelles responsible for converting stored nutrients and oxygen into usable energy. Mitochondria are often called the “cellular power plants,” and inside muscle cells they provide energy needed to contract the muscle and close the claw. … Because of their many mitochondria, muscle cells are expensive to maintain even when they are resting, and males with big claws have the most muscle. Male fiddlers burn energy like crazy to keep their muscle cells alive. Resting metabolic rates of males with big claws are almost 20 percent higher than those of females (who lack enlarged claws) simply because of the costly muscles inside the claw.

Fallow deer (Dama dama) slide in neck and neck with caribou as the living species with the most extreme antler sizes. Fallow deer are native to Eurasia, and archaeological excavations in Israel suggest that they were an important source of meat for people as far back as the Paleolithic period (nineteen thousand to three thousand years ago). This species of deer was carried across central Europe by the Romans and introduced into the United Kingdom by at least the first century CE.

In both moose and caribou, the calcium and phosphorus demands were so high that animals had to “borrow” these minerals from the other bones in their bodies in order to build their antlers. They could not get enough from their diet, so they pulled calcium and phosphorus from their own skeletons, reallocating them to the antlers. This is truly a form of deficit spending for these animals because it is not sustainable. The depleted skeletal reserves have to be replenished through feeding after the rut, and failure to do so is generally catastrophic. … All told, antlers in these animals turned out to be every bit as costly to a male as reproduction was to a female: the cost of building and using antlers was energetically and nutritionally equivalent to the cost of producing and nursing two fawns to weaning. Antler growth dramatically reduces bone mass overall, rendering males more fragile, more brittle, and much more prone to bone breaking. In essence, antler growth induces a seasonal form of osteoporosis exactly when animals are engaged in the most physically demanding and dangerous activities of their lifetimes.

The time when Irish elk disappear coincides precisely with a period of rapid climate change called "the Younger Dryas." This would have lowered the quality of available food and made it even more difficult for males to replenish the calcium and phosphorus cost of their weapons.

From fossil skeletons it's possible to determine the body sizes and proportions of these giant deer, and Moen, Pastor, and Cohen fit these values into their model to estimate how much the animals must have paid for the growth of their incredible weapons. Not surprisingly, Irish elk antlers appear to have been impressively costly, half again as much as the antlers of moose or caribou, and demanding almost two and a half times the basal metabolic energy requirements each day to grow. The calcium and phosphorus demands were severe, and seasonal osteoporosis was likely especially dire in this species.

In the end, only the largest, fittest, best-armed males prevail in the competition for reproduction. For the fallow deer in Phoenix Park, one male in ten managed to mate at all, and the vast majority of copulations (73 percent) went to just 3 percent of the bucks. Such extremes in reproductive success—90 percent failure rates and extraordinary success by just a very few individuals—lead to intense sexual selection, and much of this is directed toward bulk, stamina, and big weapons.

Not surprisingly, nutrition had a pronounced impact on beetle growth. Genitalia in poorly fed males were 7 percent shorter than in well-fed males. Wings and legs each were about 20 percent smaller. Horns, however, differed by almost 60 percent, meaning horn growth was three times as sensitive to nutrition as wings and legs, and almost nine times as sensitive as genitalia.

The best way to tell if you can beat opponents in battle is always to fight them. Charge with everything you've got, hold nothing back, and, at the end of the brawl, you'll have your answer. The problem with fighting is it's dangerous. Sometimes the risk is simply a by-product of being distracted. Crabs are well protected from one another, since their exoskeletons are like armor, but fighting crabs are distracted crabs and they make easy targets for gulls and grackles. Other times distraction is deadly because of where the fight takes place. Bighorn sheep and ibex smack heads on narrow ledges of steep cliffs, and a single misstep can be disastrous. Even a broken leg is fatal, so males pay constant attention to their opponents and their footing as they fight.

For every contest settled in a full-blown brawl, hundreds are resolved peaceably. Fiddler crabs have the largest weapons relative to their size of any living animal but, because claws act as deterrents, they almost never have to use them in fights.

Caribou are similarly cautious. One study followed more than 11,600 male-male contests over two years; only six escalated into outright battle—less than one-twentieth of 1 percent.

The most dangerous form of sneak attack, at least against our modern forces, may also be the least appreciated. Cyberattacks don't sound very scary, and it's tough to imagine how they could threaten our security beyond the occasional hassle of a usurped credit card password or identity theft. But hacking may prove to be this country's greatest danger, capable of crippling the entirety of our military forces.

The fact that these crucial weapons were compromised is scary enough, but the truly terrifying part of these events was the realization that the Chinese were not simply pirating information. It now appears their plan was to inject code that would, when activated, give them complete control of our systems.

Armed with crossbows, however, ordinary farmers could shoot down the best-trained, best-armed knights of the day. Suddenly, sitting astride a horse was a problem, rather than a tactical advantage, because mounted knights made easy targets. Bolts fired up at them could slip beneath armor plates—into the armpits, for example—and direct hits penetrated regardless. Horses could be toppled, too, bringing the mass of muscle and metal crashing to the ground so that a knight lay underfoot, as helpless as an overturned turtle.

No arms race lasts forever. As weapons get bigger they also get dramatically more expensive. Eventually, populations reach a new balance where the now-higher costs neutralize the reproductive benefits. Bigger stops being better, and the arms race stalls. Populations stabilize, hovering at the new weapon size. How big the weapons are at this stalling point depends on where the balance is finally reached; animal weapons under strong sexual selection, for example, may attain astonishing proportions before costs catch up and place the process in check.

However, when sneaks start doing well, they may erode the payoffs to fighting males substantially. Together with traditional costs, reproductive success lost to cheats can put the brakes on continued weapon evolution, helping define the point where populations begin to stabilize. In fact, if cheaters start doing too well, they may erode payoffs to weapons so drastically that the direction of selection reverses; big weapons become a liability. Instead of stalling, these races collapse.

When a few species of stalk-eyed fly stopped roosting on hanging rootlets in Malaysian streams, two of the three arms race ingredients vanished; females stopped gathering in harems on hanging threads (they were no longer economically defensible), and fights stopped being duels. With two of the race-stimulating ingredients gone, the flies' arms races collapsed. They lost their extreme weapons.

The first three sections of this book focus on animal weapons, plucking snippets from military history as needed to make a point or to illustrate a parallel. Just how far do these parallels extend? The final chapters delve more deeply and completely into humanity's greatest arms races, revealing startling similarities with animals, as well as important differences.

Although the information relevant to rifle construction is transmitted through documents and computers, rather than DNA, the result is faithful copying of the design from rifle to rifle. Rifles coming off of an AK-47 assembly line are all AK-47s and not M16s or Stg 44s. Yet, through accident or design, engineers are constantly adjusting rifle design, probing possibilities, and testing variations on the theme. Most of these experiments fail, but every now and then new design features work, and are rapidly incorporated into newer models. Most important of all, the realities of markets and battlefields act like agents of selection, culling assault rifles that are too expensive to produce, that jam or misfire, or that are more cumbersome or awkward than other available alternatives. The conditions of modern warfare shape the evolution of assault rifles in much the same way that natural agents of selection, such as battles between rival males, shape the evolution of elk antlers.

Artillery and fortifications are manufactured structures, yet they evolve just like the weapons of animals do. Advances in the effectiveness of artillery select for new and better designs of fortresses, and vice versa, in a back-and-forth cycle that can spiral into an arms race. Because fortresses are fixed in place, rivals in these races have defined roles as attackers and defenders, much like animal predators and prey. The final examples pit attacker against attacker, in matched contests more reminiscent of battles between beetles or elk.

After centuries of stasis, a single change in technology forever altered the behavior of ancient Mediterranean warships. Around 700 BCE, a simple pole of cast bronze, mounted at the waterline of the bow, turned what had been vessels of transport into weapons.1 Galleys, powered primarily by people pulling oars, began to charge into the sides of other galleys, attempting to breach their hulls and sink them. Ships fought rival ships at close range, and one-on-one, fulfilling the final ingredient of an arms race; faster ships prevailed. Speed required lots of oars, and more oars meant bigger ships. … Shipbuilders added oars, added rowers to each oar, and even added entire tiers of oars. In just a few centuries slender ships 10 feet wide and 90 feet long, powered by fifty men, jumped to double-hulled megaships 420 feet long, powered by four thousand men.3 The biggest ships were magnificent, but they'd gotten so big their weight offset the advantage of extra oars. They were too large to be fast, and too awkward to actually close in on and ram their rivals; the pendulum of ship evolution had swung too far and these ships were nautically worthless.

A mere ten years after Orville and Wilbur Wright soared over sand dunes in Kitty Hawk, North Carolina, aircraft started shooting down other aircraft in battle. … The first attempts to mount machine guns onto airplanes fared badly; bullets whizzed straight through the spinning propeller, splintering wood when they hit the blades. The French tried a crude fix, mounting steel wedges to the inside of each blade to deflect the bullets, but it was the Germans who finally solved the problem, when they mechanically coupled propellers with the firing mechanism of the machine gun, synchronizing bullets so they shot in between the blades of the propeller.14 Within a matter of weeks the French had copied, and then improved upon, this design, and for the remainder of the war both sides flew planes with forward-facing machine guns.

Now pilots could challenge other pilots directly, and dogfighting was born. Aerial duels pushed planes to their performance limits, and pilots quickly learned the capabilities of both their own and their opponents' machines, often attempting to exploit subtle differences in relative speed, climb rate, or turning radius.15 Bigger wasn't better, but faster speeds and greater maneuverability were, and planes became locked into an arms race just the same. Pilots could occasionally work around limitations of their machines, using clever tactics, skill, or trickery, but the advantage clearly went to the pilot with the superior plane, and each side raced to develop better and better machines.

Constant flight speeds made bombers predictable targets—“sitting ducks”—not unlike fixed townships and cities, and in many ways the evolution of these planes paralleled the evolution of walled fortifications and castles.26 Bombers did not attack rival bombers, like fighters attacked fighters. Instead, their survival depended upon defense, thwarting the advances of enemy fighters. In these confrontations, exposed planes were dead planes, so rotatable, protruding turrets with machine guns were added to the top and bottom, and guns were added to the nose, tail, and flanks. As with castles, the idea was to leave no flank undefended, and bombers soon carried guns and crew sufficient to provide covering fire to all sides of the craft. Even the names of these planes reflected the defensive logic of their design, such as the B-17 “flying fortress,” or the B-29 “superfortress.”

When states compete with states, the “individuals” that confront each other are rival governments, and the relevant weapons their respective military forces. New states are born from time to time, and other states disappear, but this isn't the evolutionary turnover that matters here, because arms races between states happen much faster than this, beginning and ending during the political lifetimes of the involved states. Rather, it's the military establishments within each state that grow or recede. Circumstances conspire to spark an arms race when rivals face each other in such a way that bigger militaries suddenly become much better than smaller ones. The state with more or better weapons gains an advantage that prompts the other side to catch up, launching them both into increasingly extravagant back-and-forth cycles of military spending. … For political arms races, the best parallel with animals is not the gradual turnover in populations we've focused on thus far, but rather a confrontation between two rival males—two crabs, for example, that face off in the sand. Who will back down? An arms race between states erupts in the same way a contest between crabs escalates; neither yields, pushes give way to grabs and jabs, grappling to smashing, and then pounding, and finally unrestricted war. When it comes to warfare, states behave just like crabs on a beach.

Like crabs, states have mandatory expenses that they must pay first, before they can shunt resources into weapons. Crabs are made from millions of cells, which need to be fed and protected. If the cells die, the crab dies, and most of its mandatory expenditures revolve around keeping these cells alive. States are made up of people, and necessary expenses protect and nurture these people—things such as education and welfare, police forces and highways. Only when funds are left over after mandatory expenses are accounted for can states invest in militaries, weapons, and other luxury items.

Submarines all but guaranteed a retaliatory strike capability, since they were always moving and invisible, and therefore impossible to target. Bombers flew rotating shifts, so that some were airborne all the time, and by the late 1970s the United States was developing "stealth" bombers that would be just as invisible as submarines. Railcars, hidden silos, submarines, and bombers all combined to form a swirling maze of missile platforms, making them very difficult to hit.

Weapons of this arms race could now exterminate civilization on a planetary scale; in principle, they could do this many thousands of times over.

The "dead hand," as it was called, was a satellite-based network of sensors that, once activated, could automatically fire Soviet ICBMs to previously targeted coordinates without a human having to push a button.

But deterrence has its limits, and we may be facing those limits now. … In crabs, beetles, flies, and caribou—indeed, in all animals with extreme weapons—deterrence works for very good reasons, and it works only when specific conditions are met. It boils down to choosing battles wisely. It never pays to shy from a fight you might win, but it often pays to walk away from the ones you’re likely to lose. The trick lies in predicting the outcome beforehand. In order to do this, potential combatants must have a reliable method for evaluating each other’s fighting ability.

For animal weapons to be honest signals, they must be expensive—exorbitantly expensive. So expensive, in fact, that only the top-condition males can afford them. Costs keep signals honest. If anybody could afford big weapons, then all males would have them, and differences in weapon size would be meaningless. Only when most males cannot possibly afford them will big weapons provide reliable signals of fighting strength. Then, and only then, will it pay for males with small weapons to walk away.

Early in the Cold War weapons of mass destruction met this requirement. They were prohibitively expensive, and only the richest two superpowers had nukes. But as the race progressed, warheads got cheaper. The cost of conventional weapons—submarines, fighters, and carriers, for example—soared, but the nuclear warheads themselves got smaller and cheaper. Pretty soon, England and France were testing nuclear warheads, then China and South Africa. By the 1970s India had successfully tested nuclear warheads, too, and by the 1990s so had Pakistan. Now, Israel and North Korea have them as well. The most important precondition for deterrence is disappearing.

Biological weapons weren’t very expensive to begin with, and in recent years their price has plummeted. Today, it’s possible to assemble perfect copies of the world’s most dangerous diseases—pathogens such as the 1918 strain of avian influenza, responsible for roughly one hundred million deaths—in a simple basement laboratory for a few thousand dollars. If anybody can make these weapons, then anybody can use them, and this throws the essential logic of deterrence out the window.

We are racing toward a world where lots of states possess weapons of mass destruction, regardless of the size and relative strength of their conventional fighting forces. Nuclear and biological weapons break the rules—they cheat—by providing states with few resources a means to bring down wealthier rivals. If history is any lesson here, then weapons of mass destruction are likely to erode the cost-effectiveness of expensive conventional forces. Just as longbows and muskets foretold the end of medieval armor, and exploding artillery the end for sailing warships and castles, so, too, may we be nearing the point where low-cost nuclear and biological weapons spell the end for expensive conventional military forces.

7. Hopi Hoekstra and her colleagues have revealed how mutated versions of two genes arose and spread in the beach populations, causing animals in these areas to develop with white fur. … At some point in the past, mutations altered the sequences of these two genes, and these changes caused beach mice to develop with lighter fur. … Both of these mutations lightened the color of mouse fur, and together they yielded very white mice.

Another reason parental care is generally provided by females, rather than males, has to do with the certainty of parentage. For many animal species females retain the eggs inside their bodies until after they are fertilized. A female who cares for these eggs can be certain that they are hers, and not those of another female, so energy and time invested are well spent. Males have no such assurance, for precisely the same reason: if fertilization happens inside females then males run the risk that sperm from a rival male actually fertilizes the offspring. Expending resources for the care of unrelated offspring is not cost-effective. Consequently, because females have invested the most already, and because they generally have higher certainty of genetic parentage than their mates, selection favors the evolution of maternal care of offspring more often than it does paternal care.

The concept described here is called the “operational sex ratio” (OSR). Whereas the sex ratio is simply a head count of the number of males and females in a population (and, in all but a very few exceptional species, this ratio hovers near 1:1), the operational sex ratio accounts for the fact that not all individuals are actually available for reproducing at any point in time. It is defined as the ratio of reproductively available males to reproductively available females. The OSR can skew in the direction of females, as it does in jacanas, but it is typically skewed toward an excess of available males. The extent of skew is a good metric for the intensity of sexual selection likely to be acting in the population.

To an extraordinary degree, powerful lords sequestered young women inside the walls of their castles, plucking them from their homesteads to work as maids and attendants. There is abundant evidence that these lords mated prolifically with harems of these women, often siring dozens of illegitimate children. In some cases, they prevented these women from marrying other men; in others, they sequestered them as virgins, and then married them off after they had had a child with the castle lord.

The Grants showed that year-to-year fluctuations in rainfall led to dramatic shifts in the types and amounts of seeds available to the birds, and this resulted in selection favoring deep beaks in some years but thin beaks in others. Although natural selection was directional and strong for most years of this period, the pattern of selection oscillated, so that the net effect was stasis. Despite multiple bouts of rapid change, birds at the end of the sample period had roughly the same bill shapes as those at the beginning.

My colleagues and I used information from the DNA sequences of approximately fifty species of the dung beetle genus Onthophagus to arrange taxa into a nested series of groups based on their relatedness. The resulting tree, called a phylogeny, describes the history of these animals and can be used to trace the evolution of particular traits such as horns. This study showed that horns were gained and lost repeatedly in the history of these beetles.

Keith Philips and I used a phylogenetic tree depicting the branching relationships among dung beetle species to test for an association between evolutionary gains or losses of male horns and the tunneling versus rolling behavior of each species. We found that tunneling behavior strongly predicted the evolution of horns, and when species switched from tunneling to ball-rolling behavior, they subsequently lost their male horns.

The French marched into battle using traditional close-range tactics of knights in armor, but the English combined the strength of their knights with a new type of weapon, the longbow, wielded by thousands of archers. The English were able to concentrate arrow fire in a way that the French could not, and, despite being drastically outnumbered at the outset of the battle, the English won the day. For reasons we come back to in later chapters, this battle and others like it (for example, the Battle of Crécy) marked significant turning points in the nature of battle, spelling the beginning of the end for knights resplendent in suits of expensive armor.

Trade-offs among developing structures are widespread in insects, and they constitute a dramatic cost of growing elaborate weapons. But this particular cost—stunted growth of other structures—applies primarily to insects such as beetles, flies, ants, and bees. It does not apply to any animals that I am aware of outside of insects. The reason almost certainly has to do with the way these particular insects develop. Specifically, it has to do with when, during development, all of the various adult structures are formed. Exaggerated weapons of sexual selection always grow at the end of development, around the time when males reach sexual maturity. Deer, elk, and moose all begin antler growth only after males are young adults. Elephants and boars grow their tusks after they are adults. Even shrimp and crabs begin the process of enlargement of their fighting claws around the time that their gonads mature. In all of these animals, bodies have already grown and organs, tissues, and appendages are already at or near their adult proportions, long before the weapons begin to grow. Weapons cannot stunt the growth of other traits in these animals because all of the other structures are produced first. In contrast, beetles, bees, flies, and ants all undergo metamorphosis as they develop, and this means that they grow their adult body parts at the same time as their weapons. Simultaneous growth exposes these structures to the insidious effects of resource limitation and allocation trade-offs.

Masatoshi Matsumasa and Minoru Murai were able to measure the energetic costs of fiddler crab claws in action by tracking changes in blood glucose (a sugar used to power activity) and blood lactate (a chemical by-product of metabolism and an indication of energy burned) as animals performed various behaviors. By measuring baseline lactate levels in resting animals, and comparing this to the elevated levels they detected during fights, Matsumasa and Murai showed that the energetic cost of waving a claw was substantial.

A number of theoretical models also conclude that small, poor-quality males pay a steeper price for big ornaments or weapons and that, as a result, it’s not cost-effective for them to invest in full-sized structures.

Here I build this story around costs; specifically, the observation that costs are steeper for individuals with fewer resources. This idea is an integral assumption of most models of animal signaling, and it surely applies most of the time. But there are exceptions. As I was writing this book, a doctoral student in my lab, Erin McCullough, was systematically unraveling this notion for the rhinoceros beetle we study. Her work shook the field, because everyone—including myself—assumed that the giant pitchfork horns in these beetles were costly. How could they not be? These weapons are two-thirds the length of the beetle, and they splay forward in front of the animal’s face like a massive pitchfork. Yet, these horns turned out to be remarkably inexpensive to produce, and virtually cost-free to fly around with.

15. Surprisingly few studies have actually measured the strength and nature of selection on extreme weapons—it's incredibly labor-intensive to do. But those that have often find that as weapons get bigger and bigger the success of males rises up to a point, beyond which success begins to drop. Males with the very largest weapons tend to do worse than males with slightly smaller weapons. If males with the biggest weapons of all fared the best, selection would be "open-ended" and directional. The fact that the very biggest males did slightly worse shows us that selection in these populations is stabilizing, and it suggests these populations may be at or close to their balance point.

I’m not the first to make such comparisons. Fun early accounts of similarities in the design of human and animal tools, including weapons, is provided in the treatise by the Reverend J. G. Wood, Nature’s Teachings: Human Invention Anticipated by Nature (London: William Glaisher, High Holborn, 1903). More rigorous and contemporary contrasts between weapons of animals and humans are made by Robert O’Connell in his superb books Of Arms and Men: A History of War, Weapons and Aggression (Oxford: Oxford University Press, 1989), and Soul of the Sword: An Illustrated History of Weaponry and Warfare from Prehistory to the Present (New York: The Free Press, 2002).

But the details are a bit more complex, because alleles influencing the antlers of offspring come from both parents, not just the bulls. Cows and bulls each carry the full complement of the elk genome. Genes important for antler growth may be silenced in cows (since cows do not grow antlers), but they will still be included in her eggs and passed to offspring. Thus, the antlers of sons will reflect the combination of alleles inherited from both parents.

The randomness of mutation has confused many people, since it implies that evolution also is random. (If evolution is random, how can exquisite adaptations be explained?) The trick lies in recognizing the difference between the source of variation—where the raw material necessary for evolution comes from—and what happens to this variation once it is there. Natural selection is anything but random. It’s no accident that weapons performing poorly are culled, while those performing well are retained and expanded. Given sufficient time and enough variation to work with, natural selection will push the evolution of weapons in directions that are anything but random. Consequently, new mutations infusing genetic variations into biological systems is random, but the evolution that subsequently unfolds in those populations very often is not.

Perhaps the most telling evidence that fighter aircraft evolution is unfolding at the level of the plane rather than the pilot is the fact that the newest planes don't even have pilots. Pilots were instrumental in creating the conditions that started this race, but the aircraft arms race appears to be proceeding full tilt without them.

The earliest human paintings—some more than 30,000 years old—portray animals with extreme weapons. Wooly rhinoceros from Chauvet Cave (Above) and Irish elk from Lascaux Cave (Below).

Elk pay a steep price for their weapons. Bulls double their daily energetic needs and leach essential minerals away from other bones in order to grow antlers.

Many animals, such as this katydid, rely on camouflage to avoid being eaten by predators. Leaf-mimic katydids even rock back and forth when they walk, resembling a leaf fluttering in the wind.

Special Forces snipers also rely on camouflage, choosing from among a selection of specialized suits the one most likely to match the particular background of each mission. Exquisite background matching is not cost effective for most other situations, and the majority of soldiers wear uniforms that provide only moderate camouflage—a universal color pattern designed to work reasonably well against a range of different backgrounds.

Weapons in ambush predators experience very different selection from weapons in other predators. Instead of running, swimming, or flying fast to catch prey, ambush predators snatch them with a quick strike from claspers. As a result, many have large weapons.

Males in many animals, such as these flies, fight over restricted, economically defensible resources visited by females. In these situations, big weapons are often beneficial. Above: Stalk-eyed flies defend hanging rootlets on which females roost. Below: Antlered flies defend egg-laying holes in the bark of fallen trees.

Vehicles also often fight in one-on-one duels and, as in animals, this can trigger an arms race. Above: The addition of a battering ram precipitated explosive growth in ship size in oared galleys of the ancient Mediterranean, settling in the end on the “five.” Below: Side-mounted cannons had the same effect on sailing galleons.

Fallow deer and moose pay exorbitantly for their weapons, and for the stamina and energy needed to win fights during the rut, suffering gashes, infections, and depleted energy reserves. Above: Three-quarters of fallow deer bucks die without ever succeeding in defending territories, and 90 percent fail to mate even once in their lifetime. Below: Bull moose double their daily energetic demands while growing antlers, and a third die from injuries sustained during combat.

Battling for access to females is dangerous, resulting frequently in serious wounds and, occasionally, death. Above: Fighting male oryx. Below: Male antelope died after their horns locked together in combat.

Because fighting is dangerous, males assess each other beforehand, comparing weapons and pushing. Claws function as deterrents, and fights only escalate when males are evenly matched.

Termite soldiers have even bigger heads than army ant soldiers, in part because, unlike ants, they do not need to run long distances. Instead, they position themselves at the entrances to tunnels, biting anything that attempts to enter. … By restricting access to a small number of tunnels, termites remove the numerical advantage of the army ants, forcing their soldiers to confront termite soldiers in duels the termites are likely to win.

Fortification styles evolved in tandem with ever more effective artillery. Early structures relied on square-sided, protruding towers to provide flanking fire along the walls, but corners proved vulnerable to flying boulders. Below: Later castles had round towers to deflect artillery. Even when struck directly, these towers were less likely to shatter.

Cannons destroyed even the most magnificent castles, shattering towers, breaching walls, and collapsing the arms race, until a new style of fortress emerged. Star forts sit low to the ground and rely on extensive earthworks to absorb the impact of cannons, and angled, pointy walls to deflect cannonballs fired from any direction.

Medieval knights fought as individuals in duels with rival knights. Above: The most common battles were tournaments—structured spectacles with strict rules—that tended to favor the more heavily armed and better trained contestor. Below: Even in full battle, knights confronted other knights in hand-to-hand combat that often unfolded as duels.