The Rise of Mammals

From the top of Shifting Sands dune in the Serengeti Plain of Africa a million mammals are in motion. Wildebeests. Zebras. Gazelles. The plain is black with them. It is wildebeest calving season, and many of those giant bearded antelope have newborns trailing them. Others walk with the distended bellies of imminent birth. From a distance the movement seems a serene and constant march toward the southeast, where recent rains have made pastures greener. But a closer look reveals details of high drama.

A young Grant's gazelle suddenly dashes between the clusters of wildebeests, followed closely by its mother. A hyena races in pursuit. The mother slows and moves evasively to distract the hungry predator. But the inexperienced fawn makes a panicky turn. Within moments it falls victim to the jaws of the hyena. A few yards away, ears twitching, the mother stands helpless. Then, as if in frustration, she charges two jackals on the sidelines of the kill.

"She must be feeling emotion, but there's no way to prove it," says Patricia Moehlman, the wildlife biologist who has brought me to Shifting Sands, a 12-foot-high (3.5-meter-high) dune that is itself slowly migrating across the plain. Continues Moehlman, "She's a mother. Her brain may not work like ours, but I think there's pain. I think there's fear. And certainly stress. We feel connected to her because she's a fellow mammal."

Local Masai women regard the dune as a sacred fertility site. Moehlman calls it "a place of pilgrimage." Indeed, no place on Earth offers a more spectacular abundance of our fur-bearing, breast-feeding brethren, especially when the wildebeests are on the march. But the wildebeests are only part of the scene. Myriad mammal species graze, gallop, prowl, and wallow in this part of Africa.

In the nearby Ngorongoro Crater a mother hippopotamus nuzzles her pink newborn in a muddy pond, while a pair of lions leisurely copulate along the roadside. In a grove of acacia trees a group of giraffes, members of a family of mammals that until 20 million years ago were small forest dwellers, nibble at the top branches. A few miles away elephants—which scientists are just now realizing may come from one of the oldest of the modern mammalian lineages—lumber toward a midday bath in a rain-swollen stream. Quick-witted vervet monkeys dash down from the trees to steal food through the open door of a tourist van. Meanwhile, one of the few surviving black rhinoceroses in the area wanders stealthily through a stand of high grasses.

So many mammals—and such varied shapes and behaviors—throng this land that it's hard to believe any two could have descended from the same ancestor. Nonetheless, the amphibious hippo, with its lawnmower-like diet of up to a hundred pounds (45 kilograms) of grass a night, shares a common lineage with the three-inch-long (7.5-centimeter-long) naked mole rat—a subterranean, tuber-chomping hot dog with teeth, which lives like a termite in large colonies dominated by a queen.

Deep in their bones, all mammals are related. The earliest known mammals were the morganucodontids, tiny shrew-size creatures that lived in the shadows of the dinosaurs 210 million years ago. They were one of several different mammal lineages that emerged around that time. All living mammals today, including us, descend from the one line that survived. During the next 145 million years of evolution, the dominance of dinosaurs ensured that our distant mammalian ancestors remained no larger than a cat. But when a catastrophic asteroid or comet—maybe a few comets, as some scientists are now arguing—finished off the dinosaurs 65 million years ago, mammals got the most important evolutionary opportunity they would ever have. With dinosaurs gone, mammals could exploit the planet's resources themselves. Within a few million years of the impact the fossil record shows an explosion in mammalian diversity.

How did those little creatures transform into not only the hippo and the mole rat but also today's vast panorama of mammals with fur, hooves, and fangs, as well as others that swim hairless through deep oceans—or ride, like me, in a Land Rover across this grassland?

Only humans can ask that question, or hope to answer it. We are, in a sense, the ultimate mammals. To be sure, we share defining traits with the first mammals—traits that were evolving even as the morganucodontids scrambled for food among the dinosaurs: We are warm-blooded. We have specialized jaws, whose hinges came together early in our evolution to create the ear bones that let us hear better than other animals. We have complex teeth that let us grind and chew our food so that we get more nutrition out of it. We have hair. We are superb mothers whom evolution has supplied with physical adaptations—such as breasts and placental birth—that give mammalian young an important head start. We humans are among the most recent to evolve, and we use our big mammalian brains to reason and solve problems and struggle for goals beyond our basic needs. We ask about our past and wonder what it might tell us about the future.

From scratching around in the dirt to deciphering DNA—how did we get from there to here?

That question has never had an easy answer, but today new fossil discoveries and important new tools are illuminating our distant past more clearly than ever before. Less than half a century ago, trying to understand mammalian evolution was like exploring the universe with a primitive telescope. But now high-speed analysis of genetic evidence, painstaking reconstruction of past climates and continental movements, and dogged work with often minuscule bones are creating insights that are challenging some cherished assumptions.

In the late 1960s, evidence emerged that the world's landmasses were once assembled into one great continent, called Pangaea. Around 225 million years ago Pangaea began to split into a northern continent, called Laurasia, and its southern counterpart, Gondwana. Each continent carried its own cargo of animals. Based on the known fossil record, scientists believed that the ancestors of mammals alive today emerged in the north, and then migrated south, all the way to Antarctica and Australia, as land bridges episodically developed between the continents.

André Wyss, a paleontologist at the University of California, Santa Barbara, says that's known as the "Sherwin-Williams model of evolution," a reference to the paint company logo that shows paint dripping over a globe from north to south.

Recently paleontologists have dug deeper into the fossil record of southern continents. They are finding evidence of advanced mammals far older than any known in the north, perhaps turning the Sherwin-Williams world upside down.

On another front, geneticists comparing the genes of living mammals have found that certain groups thought to be very distant cousins—hippos and whales, say—are in fact next of kin. They have also found evidence that mammals began to diversify into today's 18 living orders much earlier than the fossil record shows. Fossils suggest that most modern groups appeared around 60 million years ago, after the dinosaurs were gone. Molecular data suggest they actually began diversifying about 100 million years ago.

"It's been a complete upheaval, says Mark Springer, an evolutionary geneticist at the University of California, Riverside. "We've come up with a very different family tree for mammals."

Many paleontologists angrily reject the DNA findings, arguing there must be something wrong with the molecular clocks the geneticists use to date their findings. The geneticists counter that paleontologists just haven't found the right fossils yet.

Scientists who trust the fossils and those who look to the genes agree on at least one thing: Mammals were starting to come into their own around the time of the morganucodontids. Their tiny jawbones—about an inch (2.5 centimeters) long—show just how different the mammalian form was from the giant reptile world.

Their jawbones were beginning to fuse into one piece. "This is very different from reptile jaws, which are made up of several bones," says paleontologist Rich Cifelli of the Sam Noble Oklahoma Museum of Natural History. "Modern mammals' bones migrated backward to become the small bones of the middle ear. That's why mammals hear so much better than reptiles."

The separation of the jaw and the ear bones allowed the skulls of later mammals to expand sideways and backward—enabling mammals to develop bigger brains. The teeth of the morganucodontids were another important innovation that later mammals would improve upon. The upper and lower molars of morganucodontid jawbones interlocked, letting them slice their food into pieces. That released more calories and nutrients.

"Reptiles don't cut up their food," says Cifelli. "They grab and gulp. But these little guys were so active they had to get every calorie they could out of what they ate. The more they could process their food in their mouths, the more energy it gave them."

Scientists believe mammary glands began as sweat glands at the bases of hairs. Both sweat glands and mammary glands produce water, salts, and proteins, all of which a newborn needs to survive.

The duck-billed platypus of Australia gives us a glimpse of how those primitive mammary glands worked. The platypus and the spiny anteater are the only surviving examples of a mammalian subgroup called monotremes.

"The platypus female doesn't have nipples," explains Peter Temple-Smith, a platypus specialist at the Melbourne Zoo. "Rather, there is a region where milk ducts come together and secrete milk onto hairs. The young then lick or suck the milk off the hairs."

Nipples, which concentrate milk ducts, probably emerged with the branch of mammals we know as marsupials—a group that includes kangaroos, koalas, and opossums. "The advantage of nipples is that they give the young something to hold on to," says Temple-Smith. "The marsupial mother can therefore continue to roam about and feed freely, carrying her baby wherever she goes in her pouch."

Back on the Serengeti, we see again how mammals emphasize maternal care. A newborn wildebeest stands between the legs of its mother, its skin still wet from birth. Suddenly the air fills with the cries of vultures. They descend and with their fierce beaks begin to tear into the placenta lying a few yards away. The mother wildebeest jerks her head. Their arrival has announced to every scavenger for miles around that there is fresh young meat here, and she urges her calf into as fast a gallop as its wobbly legs can manage.

"She's being a good mother," says Moehlman, the wildlife biologist. "If you aren't a good mom, your lineage dies out. That's what being a mammal is all about."

The bloody wildebeest placenta, which the scavenging birds fight over so aggressively, illustrates the physical investment that advanced mammalian mothers make in their young. Metabolically speaking, the placenta is very expensive for the mother to maintain. Yet it is invaluable. It not only nourishes the fetus in the womb; it also isolates the developing fetus from the mother's immune system. Otherwise, her immune cells would attack the fetus as a foreign object—after all, half its genes come from the father.

Reptiles and birds avoid immune system attack by surrounding the fetus in an eggshell and moving it out of the body. Monotremes such as the platypus still lay eggs. And marsupials solve the immune problem by delivering their embryos early.

Recent DNA studies suggest placental mammals began to diverge from marsupials as early as 175 million years ago. Thus far, the fossil record has not shown this, perhaps because paleontologists trying to date the split have only little teeth and jawbones to work with. The major differences between placentals and marsupials lie in the reproductive tract—which doesn't leave much fossil evidence. But the spectacularly complete new fossil of a protoplacental species found in China's Liaoning Province has given a concrete example that strengthens DNA researchers' claims that placentals began evolving much earlier than previously thought.

"This is the mother of all placental mammals," says Zhe-Xi Luo, a paleontologist at the Carnegie Museum of Natural History in Pittsburgh, proudly presenting a fossil of what resembles a pressed mouse with a long snout. It is so well preserved that some of its fur remains visible. "We call it Eomaia, which means 'dawn mother' in Greek.'

Luo and his colleagues estimate the fossil's age at 125 million years and have found anatomic markers that suggest that Eomaia, while not fully placental, was well on its way to becoming so. That placental development was so far along 125 million years ago makes it easier for paleontologists to accept the genetic evidence that says the first protoplacentals began to evolve 50 million years earlier.

Eomaia's mousy appearance makes it a pretty modest prize by today's mammal standards, but the little creature was the leading edge of a wave of mammalian evolution that had begun with the morganucodontids. Eomaia's placental progeny represented a huge leap, opening up evolutionary options that marsupials' pouch approach constrained. For instance, marsupials develop their forelimbs early in order to climb into the pouch. But placentals' extra time in the womb lets specializations such as the bat's wing and the seal's flipper evolve. The placenta also transports nutrients much more efficiently than milk ducts do. As a result, placental babies grow faster in utero and are more mature when they leave the womb.

For those reasons, most scientists regard the pouch strategy as archaic, and perhaps inferior, pointing out that placental mammals have dominated most of the world for the past 65 million years.

However, there are some dissenters. Marilyn Renfree, a marsupial specialist at the University of Melbourne, says that biologically speaking, "marsupials are every bit as good as other mammals—and in some ways superior." Marsupials have lower metabolic rates and can therefore survive in a broader range of conditions.

Mike Archer, director of the Australian Museum, also believes that the pouch has its advantages. "For marsupials there is such a thing as being a little bit pregnant," he says. After having two eggs fertilized, a kangaroo mother may have only one egg fully develop. Should food or water become scarce and the firstborn infant die, the embryo-in-reserve can implant itself after conditions improve. In an arid land such as Australia, these conditional pregnancies can be the best strategy.

But marsupials remain much less common than other mammals. Opossums and other marsupials exist in North and South America, but Australia is the only continent where marsupials—and monotremes—still rule.

Kangaroos, koalas, platypuses, and wombats: Why does Australia retain these supposedly antiquated mammals? According to the Sherwin-Williams model, marsupials, advanced mammals themselves 100 million years ago, migrated into Gondwana ahead of placentals.

They simply got on board the Antarctic-Australian landmass before it broke away from the rest of Gondwana. Placentals arrived too late—the Australian ship had already sailed.

That theory made a lot of sense until the late 1990s, when some revealing fossils began turning up in various parts of the old Gondwana—Patagonia, Madagascar, and Australia.

The new evidence, once again, came in the form of jawbones and teeth—a particular type known as tribosphenic molars. Such teeth work like a mortar and pestle, a further improvement on the slicing teeth of earlier mammals.

The ancestor of marsupials and placentals had tribosphenic teeth. Thus the discovery in the Southern Hemisphere of tribosphenic teeth as old as 167 million years, or 25 million years older than any found in the north, complicates the north-south model. Some explain the presence of these southern tribosphenic teeth by saying they must have developed independently in both hemispheres. Others say the innovation was too intricate to have evolved twice and that mammals must have evolved in the south, with subsequent generations moving north.

"It's good to remember that the evidence is still slim," says Oklahoma's Rich Cifelli. "I like to say that anyone who really stands up strongly for either theory is either nuts or thinks too highly of himself."

The tribosphenic controversy gets even deeper in Australia, where the husband-and-wife team of Tom Rich of the Museum of Victoria and Pat Vickers-Rich of Monash University have turned up three different mammals with tribosphenic teeth dating back 110 million years. The Riches say that these mammals weren't simply on the way to becoming placental, they were placental—something like hedgehogs, in fact.

Opponents of the Riches' theory argue that placentals—and certainly not the relatively advanced hedgehogs—were not supposed to be anywhere near Australia so long ago. Eomaia, that early forerunner of placentals, lived in Asia. If the Riches are right, we have to rethink how placentals traveled from Asia to the Southern Hemisphere. Rather than traveling down the Americas, Eomaia may have found an island-hopping shortcut to Australia. Or perhaps placentals were widespread much earlier than we think now, and there's just no record of them. They could even have originated in Gondwana and spread out from there. Placentals, suggest the Riches, might even have become extinct with the dinosaurs in Australia, making room for the marsupials to move in later.

Rich himself concedes, "Most radical ideas are wrong. It's wise to be wary of them—especially when they are your own."

Even more radical to many paleontologists has been the marriage of plate tectonics evidence and the placental family tree proposed by evolutionary geneticist Mark Springer and his colleagues. Springer is part of a new generation of researchers who examine the strands of an animal's DNA rather than scraping dirt from fossils at a dig. These molecular biologists read the sequences of genes in a living animal's DNA like an evolutionary history book. The scientists can then determine how closely these animals are related genetically and how long ago their ancestors diverged.

Troubling as it is to many paleontologists, Springer's reading of mammals' genetic history fits remarkably well with what geologists now know about the breaking up and subsequent motion of ancient continents. The oldest group of living placental mammals, according to Springer and his colleagues, arose in Africa just before the continent finished breaking away from the rest of Gondwana around 110 million years ago. Springer calls these animals afrotheres. They include elephants, aardvarks, manatees, and hyraxes. When Africa floated off, it carried these animals away to evolve on their own for tens of millions of years.

The fossil record for Africa from this period is almost blank. Nevertheless, Emmanuel Gheerbrant, a researcher for the National Center for Scientific Research in France, speculates that Africa "must have been a laboratory for some very peculiar animals."

One species Gheerbrant has discovered from this period in Africa is the oldest and most primitive known member of the elephant group, the proboscideans. The 55-million-year-old fossil of Phosphatherium escuilliei was discovered in Morocco. It was the size of a fox, and although it lacked a trunk, it had many dental and cranial features strikingly similar to modern elephants. Paleontologists had long thought elephants were one of the younger modern groups, evolving from ungulates that originated in Asia. But Gheerbrant's fossil, like the genetic evidence, suggests that proboscideans are in fact one of the oldest of the modern ungulate mammals.

One of the few fossil-rich regions in Africa—the Faiyûm Depression of Egypt—has not only these early elephants but also a strange assortment of hyraxes. Today hyraxes resemble guinea pigs. But 35 million years ago hyraxes took many forms. Some were the size of rhinoceroses; others had long legs like gazelles.

Most mammals on the African ark began to disappear around 20 million years ago, after Africa came into contact with the rest of the world again. But Africa wasn't the only ark. An ancient seaway split South America from Eurasia and North America for millions of years, and South America became home to what geneticist Springer calls xenarthrans, another group of placental mammals. South America's fossil record during its isolation is far better than Africa's, and includes such xenarthrans as sloths, armadillos, and anteaters.

Springer's data, in other words, indicate that the most recent common ancestor of placental mammals is Gondwanan. Contrary to more than a century of northern chauvinism, the northern continents have the youngest placental mammals. One group, the laurasiatheres, includes seals, cows, horses, whales, and hedgehogs. The other group, euarchontoglires, includes rodents, tree shrews, monkeys, and humans.

These genetic findings reveal more than simply which came first. They also redefine relationships among placental mammals. For one, anatomists have always assumed that bats were in the same superorder as tree shrews, flying lemurs, and primates. But genetic data place bats with pigs, cows, cats, horses, and whales.

The data further show that these superorders of living mammals started to diversify much earlier than the fossil record suggests. What gets fossilized is a record of an animal's shape. But geneticists contend that genes in an organism's mitochondria, the parts of a cell that are used to trace and date lineages, can be evolving rapidly without changing what would be left behind in the fossil record.

"An animal's shape may be heavily affected by its environment," says Úlfur Érnason, a geneticist at Sweden's University of Lund. "Crocodiles haven't changed much physically in 250 million years, yet they have a high rate of change in their mitochondrial DNA. Birds have a slow rate, yet they can evolve physically very rapidly."

However surprising the claims of geneticists seem at first, paleontologists and DNA researchers are finding that their theories can be complementary. Some stunning new fossils have confirmed a previously controversial DNA finding about whales. Most paleontologists long believed that whales and dolphins—or cetaceans—descended from an extinct line of carnivorous mammals that for unknown reasons became aquatic between 50 and 45 million years ago.

At the time of these fossils' discovery, molecular biologists were maintaining that new DNA work indicated the cetaceans were actually aligned closely with artiodactyls, an order that includes even-toed ungulates such as pigs, camels, deer, and hippopotamuses.

Paleontologists first dismissed this unlikely connection because nothing in the fossil record supported it. Then in September 2001 two teams of fossil hunters published finds that backed up the claims of the biologists. A group led by Hans Thewissen of Northeastern Ohio Universities College of Medicine found two species of the earliest known whales in 50-million-year-old deposits in Pakistan. Both had ear bones unique to whales, but the legs and anklebones of artiodactyls. "The first whales, it turns out, were fully terrestrial and good runners," Thewissen says.

Almost simultaneously, a group from the University of Michigan led by Philip Gingerich announced similar fossils from Pakistan that had the same dual traits. The evolutionary transition among major groups of mammals is rarely illustrated so clearly. And no other discoveries have linked fossils to DNA findings with such precision.

Until 65 million years ago dinosaurs dominated the land. The oceans swarmed with huge sharks and voracious marine reptiles. The dinosaurs and other large predators occupied the richest and most obvious evolutionary niches, keeping mammals at the margins.

Then an event occurred whose scale is still hard to comprehend. An object six miles (9.5 kilometers) across crashed near the present-day Yucatán Peninsula, punching out a crater 110 miles (177 kilometers) across. That impact may have been one of many over the next several hundred thousand years, each adding to the destruction. But the damage done by the Yucatán impact alone is impressive: Tsunamis 500 feet (150 meters) high battered North America. The temperature reached 500 degrees in parts of the world.

"Everything big bit it," says Kirk Johnson of the Denver Museum of Nature & Science. "The key to survival was to be small." Mammals fit that profile. They suddenly found themselves in a world without large carnivores. Restraints were off. Within 270,000 years they were diversifying and growing bigger.

Still, the majority of mammals didn't get much larger than a pig until the Eocene epoch, which began about 55 million years ago. Then a rapid increase in global temperature encouraged the spread of forests around the world—even near both Poles. This abundance of rich vegetation opened yet more ecological niches for mammals to exploit. Mammal diversity soared. One of the newcomers in the fossil record was our own order, the primates. The earliest primates belonged to the lemur branch. Today lemurs are confined to the island of Madagascar, where one species made it from Africa perhaps 50 million years ago, probably on rafts of storm-tossed debris.

A few million years later, more advanced primates appear in the fossil record of eastern Asia. These higher primates are anthropoids—monkeys, apes, and humans. Chris Beard, a specialist in primate origins at the Carnegie Museum of Natural History, has unearthed in China what may be the earliest known example, called Eosimias. These creatures evolved in the mid-Eocene as the world was cooling and concentrated in the midlatitudes where forests remained lush.

Beard says they "must have been frenetic little animals. Kind of caffeinated. They probably ate all the time. When you are that small, you have to. They probably lived in troops and maybe never left the tree they were born in." Despite its primitive anatomy, Eosimias had already adopted the monkeylike habit of walking along the tops of branches rather than leaping from tree to tree like earlier primates.

About 34 million years ago smarter, bigger, and more aggressive monkeys evolved. Fossils from the Faiyûm Depression, where Elwyn Simons of Duke University has led a dig since 1961, reveal how anthropoids were changing. Catopithecus, one of many anthropoids his team has uncovered, has a skull the size of a small monkey's, a relatively flat face, and a bony enclosure at the rear of its eye sockets. It is the first anthropoid to show the same arrangement of teeth humans have—two incisors, one canine, two premolars, and three molars—leading Simons to argue, "This is the first chapter of human history."

At the start of the long Miocene epoch—23.5 million to 5.3 million years ago—yet another major climate change occurred. The world was warming again and more seasonal climate patterns may have emerged. At higher latitudes, forests gradually gave way in many places to grassland meadows and savannas. Because grass is abrasive, some mammals developed new dentition. Horses, for instance, emerged as little leaf-eaters in the forests but later developed molars that are much better adapted to eating grass. Horses' crowns extend into the jawbones. As the crown gets ground down, new crown will emerge from the jaw to replace it.

Early in the Miocene, Africa's long isolation ended when it and Arabia came back into contact with Eurasia. That's when the ancestors of many mammals we think of as native to Africa arrived there. First came the ancestors of antelope, cats, giraffes, and rhinos. Later, around ten million years ago, North American mammals—camels, horses, and dogs—began to arrive. Almost every animal that roams the Serengeti today is a relative newcomer to the continent.

Africa gave back as well. Apes moved into Eurasia and flourished. Elephants and their relatives spread across the globe, reaching as far as the tip of Patagonia.

But geology and climate changed the world once again for mammals as the Miocene drew to a close. The Earth grew colder and drier still. Ice caps formed in the Arctic. The Sahara began taking over North Africa, and savannas spread across much of the continent.

The changing climate restricted the range of the primates to the equatorial zone. The surviving apes became larger and more specialized. Then, around seven million years ago, at least one offshoot of the African apes began walking on two legs.

As that bipedal ape evolved into what would become us, other mammals came and went. Most had to adapt to yet another global climate change about 2.5 million years ago, triggered in part by the formation of the Isthmus of Panama. Its formation blocked east-west ocean circulation and encouraged the Gulf Stream to grow stronger. As the Gulf Stream pumped more warm water closer to the North Pole, precipitation increased. Heavy snows became glaciers two miles (three kilometers) thick, which advanced and retreated in a series of more than 20 ice ages. Because big bodies retain heat better, many mammals, such as the woolly mammoth, grew larger. Even in the temperate zones of Australia, animals became immense. Australia was soon home to big meat-eating kangaroos, wombat-like creatures the size of trucks, and a marsupial lion twice as big as a leopard.

"It was a kick-ass big predator," says paleontologist Steve Wroe of the University of Sydney, as he admires a foot-long (0.3-meter-long) fossil skull of a marsupial lion from 40,000 years ago. "These big shearing teeth make it more highly adapted as a carnivore than any other known mammal." The teeth, explains Wroe, are "for butchery only. The animal would starve to death in a fruit and veggie shop."

Those big mammals, such as the marsupial lion and the killer kangaroo, disappeared between 100,000 and 20,000 years ago. Few controversies rage more fiercely in paleontology than why the megafauna vanished—not just in Australia but also in North America, where mammoths, horses, camels, and dozens of other large Ice Age mammals all vanished by about 11,000 years ago. Many scientists cite climate change. Others say it was humans, arguing that newly arrived Homo sapiens killed off the giants with their spears.

We humans may or may not have killed off the giant mammals of the Ice Age. But we are unquestionably threatening innumerable species today, as we expand relentlessly into ever more of their habitats. Signs of this encroachment appear all around the world. Manatees in Florida chopped up by boat propellers. Rhinos in the Ngorongoro Crater poached. Vast rain forests in Southeast Asia obliterated. All this done by the most intelligent of mammals. Evolution has given us this gift of intelligence, but are we too smart for our own good? If somehow we could rewind time to the dawn of anthropoids, what different path could we have gone down?

One answer lies some 5,000 miles (8,000 kilometers) from the Serengeti's vibrant mammalian spectacle, in the rain forests of Indonesia, Borneo, and the Philippines. There lives the tarsier, which the Carnegie Museum's Beard cites as an example of the primate road not taken. "Tarsiers are pretty weird," says Beard, "They can turn their heads 270 degrees. They're the primates' version of an owl. But they're the nearest living relatives of higher primates."

Tarsiers share a common ancestor with all anthropoids. We know this because like all higher primates, tarsiers lack a tapetum lucidum—the reflective layer in the eyes of nocturnal animals. The tapetum lucidum is critical to vision in low light levels and is what makes the eyes of night creatures glow when a flashlight shines on them.

Unlike most of their anthropoid relatives, tarsiers went back to a nocturnal lifestyle at some point and had to compensate by evolving enormous, spooky eyes. Shine a flashlight in a lemur's eyes at night, and they'll glow back at you. A tarsier's won't.

For humans, tarsiers represent what might have been. But because of us, today they're hard to find. Hunting, development, and the destruction of the rain forests have constricted the tarsiers' habitat.

In Borneo, where tarsiers are considered bad luck, few villagers worry about that. "If I see a tarsier, I go home," says a villager in Kampung Duras in Sarawak. Another local, Lemon Ales, agrees.

"They frighten people because of their big eyes. Also, they may leap on you and bite."

Other villagers regard the tarsier as totems, because the small agile creatures sometimes are seen in rice paddies holding on to the rice stems, as if guarding them.

Lemon and I head into the forest at twilight. The world looked like this in the Eocene, when primates were evolving. The forest is full of the same kinds of fruit-bearing trees that helped primates thrive in the vast forests then emerging around the world. The heavy air seems to press moisture into my skin, and my pores fight back with gushes of sweat. With flashlights we stumble on for several hours in the dark. But the tarsiers never show up. Or maybe they do: The locals warn that if the creatures don't move, you won't see them.

The Singapore zoo has tried to make sure its patrons won't be similarly disappointed. Its tarsiers are behind a glass wall in a simulated forest. "Only six zoos in the world have tarsiers," says C. S. Menon, an animal management officer. "Usually they stress out and die in captivity."

Eager, often pushy, visitors chatter in languages ranging from Dutch to Hindi to Japanese while waiting to board zoo trams that will take them to see what few can now see in the wild at night.

Many get off the tram, pause for a moment at the tarsier exhibit, and move on. That's because even here, under the best nighttime viewing conditions—the simulated light of a full moon—tarsiers are difficult to spot. I try to be patient.

Finally, like lightning, one flashes out of nowhere to grasp a cricket with both hands and land gracefully on a slender branch. It sits upright, munching and rotating its little head in an improbably wide arc. You can't look into the big eyes of our distant cousin without feeling awed by the distance we have traveled away from each other. Our intelligence may pose grave dangers to the world's wildlife and to ourselves, but it also lets us feel wonder. And concern. The jury's still out on where that will take us.

Then, faster than a moth's flutter, the tarsier is gone.