Sunrise on the Serengeti, and life on the savanna is in full swing. Zebras and wildebeests graze the dewy grass; elephants and giraffes munch on acacia leaves; and lions and hyenas survey the scene, looking for their next meal. To visit this place is, in some ways, to see the world as it looked to our ancestors millions of years ago, long before humans began to wreak havoc on the planet—or so the conventional wisdom goes. Indeed, much of eastern Africa is often thought of as a pristine ecosystem, largely unchanged by our kind in the more than two million years since our genus, Homo, arose.

But new research paints a rather different picture of this supposedly unaltered place. In my studies of the fossil record of African carnivores, I have found that lions, hyenas and other large-bodied carnivores that roam eastern Africa today represent only a small fraction of the diversity this group once had. Intriguingly, the decline of these carnivores began around the same time that early Homo started eating more meat, thus entering into competition with the carnivores. The timing of events hints that early humans are to blame for the extinction of these beasts—starting more than two million years ago, long before Homo sapiens came on the scene.

The rise of this new meat eater—and the loss of the big carnivores—would have triggered large-scale changes farther down the food chain, affecting the prey animals and even the plants those creatures ate. Thus, if my hypothesis is correct, our forebears began radically transforming ecosystems far earlier than previously thought, at a time when ancestral population sizes were quite small. Homo, it seems, has been a force of nature from the outset.

Vanished Carnivores
Fossil carnivores—which is to say, members of the Carnivora order of mammals—have captivated me ever since I first read about them in the books of Finnish paleontologist Björn Kurtén as a teenager. Back then, I just thought they were cool, and I knew that they played an essential role as regulators of herbivore populations, which would explode without these predators to keep them in check. Only after I began studying carnivore fossils professionally, however, did I come to appreciate how their relationship with humans has evolved over millions of years.

For two decades I have studied thousands of carnivore fossils from eastern and southern Africa, trying to get a handle on how the modern carnivore community evolved over the past seven million years. I have conducted much of this research in collaboration with Margaret E. Lewis of Richard Stockton College, who is an expert on carnivore bones from the neck backward, whereas I specialize in their teeth and skulls. Our work has yielded a much higher-resolution view than was previously available of how many kinds of carnivores there were in Africa at different times during this interval, which also spans the entire known history of human evolution. As we amassed more and more data, we gained a much clearer picture of the species that thrived and failed over time, and we began to realize that the decline of the large carnivores (those weighing 21.5 kilograms or more) coincided with a shift among human ancestors from a mostly vegetarian diet to one that relied more heavily on animal foods. To our great surprise, it looked as though our early ancestors might have been at fault for the loss of these species.

Snapshots of a few key fossil sites provide a sense of the transformation the African carnivore community underwent. The carnivores from the early part of this seven-million-year interval were nothing like the ones found today. Fossils dating to between 7.5 million and five million years ago from the site of Lothagam on the western shore of Lake Turkana in northern Kenya reveal sabertooth cats, strange long-legged hyenas, giant bear dogs (neither bears nor dogs but members of an extinct family of carnivores, the Amphicyonidae), and a leopard-size member of the mustelid family to which badgers belong. Smaller carnivores related to today's civets and mongooses also prowled there.

By four million years ago a familiar face had joined the carnivore cast. At the nearby site of Kanapoi, sabertooths and other now extinct lineages were still present, but the most common carnivore there was a hyena species ancestral to the brown hyena found in southern Africa today. Fast-forward another few hundred thousand years, and the carnivore community is even more recognizable. The 4.4-million- to 3.6-million-year-old site of Laetoli in the Serengeti of Tanzania, famed for its fossilized trail of footprints belonging to hominins (members of the human family), has remains of modern-looking cats along with the sabertooths. Early spotted hyenas, several dog species, a giant civet and a variety of smaller carnivores lived there, too. At Hadar in Ethiopia, the final resting place of the 3.2-million-year-old Lucy skeleton, sabertooths, hyenas and dogs abound, along with giant otters that have no modern counterpart.

These and other sites in the time range of four million to 2.5 million years ago all tell the same story. Each has a slightly different mix of carnivore species, depending on the environmental setting, but all have the same general kinds of carnivores. For example, all the sites have hyenas, but they differ in the species of hyena that lived there. And more important, none indicates that these animals were any worse off as a result of the presence of hominins.

After peaking around 3.5 million years ago, the number of large carnivore species declined gradually over the next million and a half years or so, mostly because the rate at which new species originated slowed down while the extinction rate held steady. Still, on the whole, the big carnivores reigned supreme during this time; our small, slow, defenseless ancestors were merely food. But the tide was about to turn.

The record after two million years ago shows unmistakable changes in the composition of carnivore communities. With extinction rates increasing and origination rates remaining low, the number of large species began to nosedive, particularly after 1.5 million years ago. Not only did individual species die out, but entire groups of species, such as the sabertooth cats, disappeared. As these beasts of yore declined, modern species—including the lions, leopards and jackals that inhabit Africa today—came to account for an ever increasing proportion of carnivore communities. By around 300,000 years ago the archaic carnivore lineages had all been winnowed out in eastern Africa and the modern carnivore community was in place.

A Wolf in Sheep's Clothing
The general pattern Lewis and I observed in our data fit with our intuitive understanding of the evolutionary history of African carnivores in that it confirmed that there were more kinds of large carnivores in the past than there are today. What we had not anticipated was the steep downturn after 1.5 million years ago. It was this timing that hinted our Homo ancestors might be at fault.

For the first few million years of human evolution, hominins were relatively small-brained, chimpanzee-size creatures that subsisted primarily on plant foods. But by 1.5 million years ago a new kind of hominin was on the scene—one that was bigger, smarter and armed with stone tools. This was Homo erectus (sometimes called Homo ergaster), the first member of the human family to really look like us—and the first to start eating much in the way of meat. Perhaps, Lewis and I reasoned, competition with this human predator, which was incorporating increasing amounts of animal protein from large herbivores into its diet, could explain the carnivore decline.

That explanation seemed promising, but the timing of the events nagged at me. If competition with H. erectus was to blame, then the steep decline in eastern Africa's large carnivore species should have started well before 1.5 million years ago because H. erectus had emerged by nearly 1.9 million years ago. Species numbers are a blunt instrument at best for tracking the progress of an entire order of mammals over time because a reduction in numbers of one of its group can be masked by an increase in another. If two sabertooth species go extinct but are replaced by lions and leopards, the numbers will remain the same, but the community will have undergone a major change because lions and leopards can take a broader range of prey than sabertooths could.

It occurred to me that I could get a better sense of what had befallen the large carnivores if I understood not just how many species there were at any given time but how diverse their ecological roles were. Carnivores vary a lot in how they make a living. The cats, for example, are highly adapted to eating meat and thus qualify as hypercarnivores. But other carnivores are omnivorous—dogs, for example, will eat a wide variety of food in addition to meat. Still others, such as raccoons, are hypocarnivores, eating very little meat and subsisting mainly on fruits and vegetables.

I decided to build on work of my former postdoctoral student Gina D. Wesley-Hunt, now at Montgomery College, who had investigated the evolution of North American carnivores over the past 60 million years. As part of her study, Wesley-Hunt identified a set of traits related to the function of the jaws and teeth of carnivores. By studying these traits, she could quantify just how different species in a single carnivore community were from one another in terms of the kinds of foods they ate and hence their ecological roles. Using the fossil-coding scheme she developed to identify the function of the jaws and teeth (that is, the eating preferences they had evolved for), I coded 78 carnivore species—29 large and 49 small—from the African fossil record of the past 3.5 million years. I then analyzed the data, looking at how the number of different kinds of carnivores with different ecological roles living within the same community changed over time.

To visualize the diversity of form and inferred eating preferences in these fossil carnivores, I plugged the data from the coding scheme into a statistical analysis, thereby creating a two-dimensional plot that I call the morphospace. This morphospace represents the diversity of form (and inferred function) that exists within a group of related organisms, in this case the carnivores that lived in Africa over the past 3.5 million years. Plotting separate morphospaces for carnivores from distinct time intervals and comparing them offers a sense of how carnivore anatomy and eating habits shifted over time.

The results proved startling. As Lewis and I reported in March in PLOS ONE, it turns out that large carnivores that live in eastern Africa today occupy only a small fraction (less than 1.5 percent) of the morphospace of the carnivores in the 3.5-million- to three-million-year interval, when species diversity was at its highest. The group has lost nearly 99 percent of its so-called functional richness, which is to say today's carnivores fill far fewer kinds of ecological roles than their predecessors did. Moreover, the measured decrease in this functional richness began in the interval between two million and 1.5 million years ago, which means that the process must have started before that time—bringing the onset of this major decline in line with the origin of H. erectus. Although our work focused on carnivores from eastern Africa, modern large carnivores are basically the same across the continent. Thus, it is likely that the loss of functional richness we found in this region is representative of what happened to all of Africa's large carnivores.

Human activity is not the only possible cause of this loss of Africa's carnivores. Climate change has been implicated in many faunal changes in Africa over the course of the past few million years, and at first glance comparisons of climate and species numbers imply it is a front-runner in this case as well. Studies of modern carnivore species, however, suggest that the influence of climate on the functional richness of modern carnivore communities is slight. In general, carnivores are insensitive to climate and related environmental change, unlike mammalian herbivores, which are dependent on the distribution of plant food, which in turn is largely determined by climate. Furthermore, if climate change was the culprit, then the smaller carnivores should have declined, too—but they did not. Both the species richness and functional richness of the small carnivores have held up over most of the past 3.5 million years and may even have increased.

Nevertheless, to determine whether human activity was responsible for the decline of these carnivores, it would help to know how important meat was to early Homo. Archaeologists have long debated this question. Some think meat and hunting were critically important; others hold that meat was a marginal component of the diet at best, with the hominins merely scavenging a few scraps that carnivores rejected. But they generally agree that Homo did begin to obtain more protein from animals, perhaps including fish and shellfish, between two million and 1.5 million years ago in the Early Pleistocene period.

Anthropologist Henry Bunn of the University of Wisconsin–Madison envisions the transition to a meatier diet unfolding in three steps, the timings of which dovetail nicely with the idea that competition with early hominins drove many big carnivores to extinction. First, hominins occasionally butchered bones using primitive stone tools or naturally flaked blades. At this stage, which Bunn puts at around 2.6 million to 2.5 million years ago, based on the available archaeological evidence, they had only a slight ability to obtain meat. The second stage involved more habitual butchery, along with the skills to break bones to get at the marrow inside and to transport meat-rich parts of carcasses to a home base or similar. Bunn estimates that hominins reached this stage around 2.3 million to 1.9 million years ago and that by this point they could obtain meat on a regular basis through scavenging and possibly making their own kills. In the third stage, hominins butchered animal remains extensively and had access to intact carcasses because they were better at appropriating carnivore kills or possibly because they were routinely hunting the animals themselves. Bunn dates this last stage to between 1.8 million and 1.6 million years ago.

Thus, although they lacked the lethal teeth and claws and the sheer physical strength of the sabertooth cats and other large carnivores, hominins were able to level the playing field through their rapidly evolving intelligence and social cooperation—there is strength in numbers even if brawn is lacking. And in lean times, hominins would have had a distinct advantage over carnivores, especially hypercarnivores such as the sabertooths, because, being omnivorous, they had a much larger array of foods they could fall back on when their top choices were unavailable. During the worst times of the year, then, the hominin competitive edge would have been the greatest. (That the remaining large carnivores are all hypercarnivores reflects the fact that there were many more kinds of hypercarnivores to start with than omnivores or hypocarnivores.)

Food for Thought
Like any nascent hypothesis, this one comes with a series of problems that need resolution. The most significant of these issues concerns the timing of the events described here, both in terms of when the carnivores began going downhill and when humans started to pose a competitive threat to them. We need a clearer picture of what happened and when to draw firm conclusions about cause and effect. In addition, scientists do not know whether hominins were sufficiently numerous and competitive to cause such massive change to the carnivore community.

Pinpointing when the carnivore decline began requires either the discovery of additional fossils from the 2.5-million- to two-million-year time interval or more refined techniques for analyzing the fossils we already have. I am currently working on developing such techniques. What I can definitely say at this point is that the onset of change in the carnivores had occurred by 1.8 million years ago and that the most refined analysis at present suggests that it occurred shortly before two million years ago. Whether this can be accurately matched up with events in hominin evolution, however, is not yet clear. Although Bunn's timetable is fully compatible with the scenario I have presented, it has not gone unchallenged. Other scholars suspect the first two stages occurred considerably later than he proposes.

Resolution of the issue of hominin numbers and competitive ability may never come. These aspects of early hominins are currently mostly a matter of opinion. Undoubtedly, population density was low, but how low is unknown. It may be possible to generate a series of simulations of both factors to see whether the hypothesis is viable given reasonable values for either or both. But hard evidence of how many hominins were around and how successful they were in getting hold of prey that would have otherwise ended up in a sabertooth's belly may always elude science's grasp. The absence of these data does not demonstrate that my hypothesis is false, however.

I hope that researchers skeptical of my hypothesis will come up with some ingenious ways of testing it. To that end, another aspect of this idea bears mention. Attempts to explain ecosystem change typically provide a bottom-up perspective, looking at how climate factors affect plants and how changes in those organisms affect the rest of the food chain up to the top predators. My hypothesis about eastern Africa's large carnivores provides a top-down view, considering how change in the top predators could affect the primary producers at the bottom of the food chain, such as grasses and trees.

The reintroduction of wolves to Yellowstone National Park and their effect on the herbivores living there, and by extension on the vegetation of the park, provide a stunning example of the impact of change among top predators. As the wolves became more plentiful, the elk they preyed on diminished in numbers. This in turn led to less feeding pressure on the plants and to lusher vegetation in those places where the herbivores were previously particularly common [see “Lessons from the Wolf,” by Jim Robbins; Scientific American, June 2004].

The entry of early Homo into the carnivore niche in Africa could have triggered an even more dramatic cascade of ecological disruption than the one that occurred at Yellowstone. Whereas wolves were once a natural part of the Yellowstone ecosystem, meaning that the other species there retained at least some adaptations to their presence, early Homo had no such precedent. One would expect the introduction of such a new predator to have even greater consequences for the ecosystem than the reintroduction of one that had been there originally. Perhaps, then, the smoking gun in the case of the disappearing carnivores will turn up not among the remains of our hominin ancestors or the large carnivores themselves but among the remnants of herbivores and plants whose world was upended when Homo developed a taste for meat.