In 2004 a team of australian and indonesian scientists who had been excavating a cave called Liang Bua on the Indonesian island of Flores announced that they had unearthed something extraordinary: a partial skeleton of an adult human female who would have stood just over a meter tall and who had a brain a third as large as our own. The specimen, known to scientists as LB1, quickly received a fanciful nickname—the hobbit, after writer J.R.R. Tolkien's fictional creatures. The researchers proposed that LB1 and the other fragmentary remains they recovered represent a previously unknown human species, Homo floresiensis. Their best guess was that H. floresiensis was a descendant of Homo erectus—the first species known to have colonized outside of Africa. The creature evolved its small size, they surmised, as a response to the limited resources available on its island home—a phenomenon that had previously been documented in other mammals, but never humans.

The finding jolted the paleoanthropological community. Not only was H. floresiensis being held up as the first example of a human following the so-called island rule, but it also seemed to reverse a trend toward ever larger brain size over the course of human evolution. Furthermore, the same deposits in which the small-bodied, small-brained individuals were found also yielded stone tools for hunting and butchering animals, as well as remainders of fires for cooking them—rather advanced behaviors for a creature with a brain the size of a chimpanzee's. And astonishingly, LB1 lived just 18,000 years ago—thousands of years after our other late-surviving relatives, the Neandertals and H. erectus, disappeared.

Skeptics were quick to dismiss LB1 as nothing more than a modern human with a disease that stunted her growth. And since the announcement of the discovery, they have proposed a number of possible conditions to explain the specimen's peculiar features, from cretinism to Laron syndrome, a genetic disease that causes insensitivity to growth hormone. Their arguments have failed to convince the hobbit proponents, however, who have countered each diagnosis with evidence to the contrary.

[break] A Perplexing Pastiche

Nevertheless, recent analyses are causing even the proponents to rethink important aspects of the original interpretation of the discovery. The findings are also forcing paleoanthropologists to reconsider established views of such watershed moments in human evolution as the initial migration out of Africa by hominins (the group that includes all the creatures in the human line since it branched away from chimpanzees).

Perhaps the most startling realization to emerge from the latest studies is how very primitive LB1's body is in many respects. (To date, excavators have recovered the bones of some 14 individuals from the site, but LB1 remains the most complete specimen by far.) From the outset, the specimen has invited comparisons to the 3.2-million-year-old Lucy—the best-known representative of a human ancestor called Australopithecus afarensis—because they were about the same height and had similarly small brains. But it turns out LB1 has much more than size in common with Lucy and other pre-erectus hominins. And a number of her features are downright apelike.

A particularly striking example of the bizarre morphology of the hobbits surfaced in May 2009, when researchers led by William L. Jungers of Stony Brook University published their analysis of LB1's foot. The foot has a few modern features—for instance, the big toe is aligned with the other toes, as opposed to splaying out to the side as it does in apes and australopithecines. But by and large, it is old-fashioned. Measuring around 20 centimeters in length, LB1's foot is 70 percent as long as her short thighbone, a ratio unheard of for a member of the human family. The foot of a modern human, in contrast, is on average 55 percent as long as the femur. The closest match to LB1 in this regard, aside from, perhaps, the large-footed hobbits of Tolkien's imagination, is a bonobo. Furthermore, LB1's big toe is short, her other toes are long and slightly curved, and her foot lacks a proper arch—all primitive traits.

“A foot like this one has never been seen before in the human fossil record,” Jungers said in a statement to the press. Characteristics of the pelvis, leg and foot make clear that the hobbits walked upright. But with their short legs and relatively long feet, they would have had to use a high-stepping gait to avoid dragging their toes on the ground. Thus, although they could probably sprint short distances—say, to avoid becoming dinner for one of the Komodo dragons that patrolled Flores—they would not have won any marathons.

If the foot were the only part of the hobbit to exhibit such primitive traits, scientists might have an easier time upholding the idea that H. floresiensis is a dwarfed descendant of H. erectus and just chalking the foot morphology up to an evolutionary reversal that occurred as a consequence of dwarfing. But the fact is that archaic features are found throughout the entire skeleton of LB1. A bone in the wrist called the trapezoid, which in our own species is shaped like a boot, is instead shaped like a pyramid, as it is in apes; the clavicle is short and quite curved, in contrast to the longer, straighter clavicle that occurs in hominins of modern body form; the pelvis is basin-shaped, as in australopithecines, rather than funnel-shaped, as in H. erectus and other later Homo species. The list goes on.

Indeed, from the neck down LB1 looks more like Lucy and the other australopithecines than Homo. But then there is the complicated matter of her skull. Although it encased a grapefruit-size brain measuring just 417 cubic centimeters—a volume within the range of chimpanzees and australopithecines—other cranial features, such as the narrow nose, mark LB1 as a member of our genus, Homo.

[break] Primitive Roots

Fossils that combine Homo-like skull characteristics with primitive traits in the trunk and limbs are not unprecedented. The earliest members of our genus, such as Homo habilis, also exhibit a hodgepodge of old and new. Thus, as details of the hobbits' postcranial skeletons have emerged, researchers have increasingly wondered whether the little Floresians might belong to a primitive Homo species, rather than having descended from H. erectus, which had modern body proportions.

An analysis conducted by Debbie Argue of the Australian National University in Canberra and her colleagues bolsters this view. To tackle the problem of how the hobbits are related to other members of the human family, the team employed cladistics—a method that looks at shared, novel traits to work out relationships among organisms—comparing anatomical characteristics of LB1 with those of other members of the human family, as well as apes.

In a 2009 paper in the Journal of Human Evolution, Argue and her collaborators reported that their results suggest two possible positions for the H. floresiensis branch of the hominin family tree. The first position is that H. floresiensis evolved after a hominin called Homo rudolfensis, which arose some 2.3 million years ago but before H. habilis, which appeared roughly two million years ago. The second position is that it emerged after H. habilis but still well before H. erectus, which arose around 1.8 million years ago. More important, Argue's team found no support for a close relationship between H. floresiensis and H. erectus, thereby dealing a blow to the theory that the hobbits were the product of island dwarfing of H. erectus. (The study also rejected the hypothesis that hobbits belong to our own species.)

If the hobbits are a very early species of Homo that predates H. erectus, that positioning on the family tree would go a long way toward accounting for LB1's tiny brain, because the earliest members of our genus had significantly less gray matter than the average H. erectus possessed. But Argue's findings do not solve the brain problem entirely. LB1 aside, the smallest known noggin in the genus Homo is an H. habilis specimen with an estimated cranial capacity of 509 cubic centimeters. LB1's brain was some 20 percent smaller than that.

Could island dwarfing still have played a role in determining the size of the hobbit's brain? When the discovery team first attributed LB1's wee brain to this phenomenon, critics complained that her brain was far smaller than it should be for a hominin of her body size, based on known scaling relationships. Mammals that undergo dwarfing typically exhibit only moderate reduction in brain size. But dwarfing of mammals on islands may present a special case. Eleanor Weston and Adrian Lister of the Natural History Museum in London found that in several species of fossil hippopotamus that became dwarfed on the African island nation of Madagascar, brain size shrank significantly more than predicted by standard scaling models. Based on their hippo model, the study authors contend, even an ancestor the size of H. erectus could conceivably attain the brain and body proportions of LB1 through island dwarfing.

The work on hippos has impressed researchers such as Harvard University's Daniel Lieberman. In a commentary accompanying Weston and Lister's report in Nature, Lieberman wrote that their findings “come to the rescue” in terms of explaining how H. floresiensis came by such a small brain.

Although some specialists favor the original interpretation of the hobbits, Mike Morwood of the University of Wollongong in Australia, who helps to coordinate the Liang Bua project, now thinks the ancestors of LB1 and the gang were early members of Homo who were already small—much smaller than even the tiniest known H. erectus individuals—when they arrived on Flores and then “maybe underwent a little insular dwarfing” once they got there.

Artifacts left behind by the hobbits support the claim that H. floresiensis is a very primitive hominin. Early reports on the initial discovery focused on the few stone tools found in the hobbit levels at Liang Bua that were surprisingly sophisticated for a such a small-brained creature—an observation that skeptics highlighted to support their contention that the hobbits were modern humans, not a new species. But subsequent analyses led by Mark W. Moore of the University of New England in Australia and Adam R. Brumm, now at the University of Wollongong, have revealed the hobbit toolkit to be overall quite basic and in line with the implements produced by other small-brained hominins. The advanced appearance of a handful of the hobbit tools at Liang Bua, Moore and Brumm concluded, was produced by chance, which is not unexpected considering that the hobbits manufactured thousands of implements.

To make their tools, the hobbits removed large flakes from rocks outside the cave and then struck smaller flakes off the large flakes inside the cave, employing the same simple stone-working techniques favored by humans at another site on Flores 50 kilometers east of Liang Bua called Mata Menge 880,000 years ago—long before modern humans showed up on the island. (The identity of the Mata Menge toolmakers is unknown because no human remains have turned up there yet, but they conceivably could be the ancestors of the diminutive residents of Liang Bua.) Furthermore, the Liang Bua and Mata Menge tools bear a striking resemblance to artifacts from Olduvai Gorge in Tanzania that date to between 1.2 million and 1.9 million years ago and were probably manufactured by H. habilis.

[break] Tiny Trailblazer

In some ways, the latest theory about the enigmatic Flores bones is even more revolutionary than the original claim. “The possibility that a very primitive member of the genus Homo left Africa, perhaps roughly two million years ago, and that a descendant population persisted until only several thousand years ago, is one of the more provocative hypotheses to have emerged in paleoanthropology during the past few years,” reflects David S. Strait of the University at Albany S.U.N.Y. Scientists have long believed that H. erectus was the first member of the human family to march out of the natal continent and colonize new lands because that is the hominin whose remains appear outside of Africa earliest in the fossil record. In explanation, it was proposed that humans needed to evolve large brains and long striding limbs and to invent sophisticated technology before they could finally leave their homeland.

Today the oldest unequivocal evidence of humans outside of Africa comes from the Republic of Georgia, where researchers have recovered H. erectus remains dating to 1.78 million years ago. The discovery of the Georgian remains dispelled that notion of a brawny trailblazer with a tricked-out toolkit because they were on the small side for H. erectus, and they made Oldowan tools, rather than the advanced, so-called Acheulean implements experts expected the first pioneers to make. Nevertheless, they were H. erectus.

But if proponents of the new view of hobbits are right, the first intercontinental migrations were undertaken hundreds of thousands of years earlier than that—and by a fundamentally different kind of human, one that arguably had more in common with primitive little Lucy than the colonizer paleoanthropologists had envisioned. This scenario implies that scientists might locate a long-lost chapter of human prehistory in the form of a two-million-year record of this pioneer stretching between Africa and Southeast Asia if they look in the right places.

This suggestion does not sit well with some researchers. “The further back we try to push the divergence of the Flores [hominin], the more difficult it becomes to explain why a [hominin] lineage that must have originated in Africa has left only one trace on the tiny island of Flores,” comments primate evolution expert Robert Martin of the Field Museum in Chicago. Martin is unconvinced that H. floresiensis is a legitimate new species. In his view, the possibility that LB1—the only hobbit whose brain size is known—was a modern human with an as yet unidentified disorder that gave rise to a small brain has not been ruled out. The question is whether such a condition can also explain the australopithecinelike body of LB1.

Many scientists are welcoming the shake-up. LB1 is “a hominin that no one would be saying anything about if we found it in Africa two million years ago,” asserts Matthew W. Tocheri of the Smithsonian Institution, who has analyzed the wrist bones of the hobbits. “The problem is that we're finding it in Indonesia in essentially modern times.” The good news, he adds, is that it suggests more such finds remain to be recovered.

“Given how little we know about the Asian hominin record, there is plenty of room for surprises,” observes Robin W. Dennell of the University of Sheffield in England. Dennell has postulated that even the australopithecines might have left Africa, because the grasslands they had colonized in Africa by three million years ago extended into Asia. “What we need, of course, are more discoveries—from Flores, neighboring islands such as Sulawesi, mainland Southeast Asia or anywhere else in Asia,” he says.

Morwood, for his part, is attempting to do just that. In addition to the work at Liang Bua and the Soa Basin of central Flores, he is helping to coordinate two projects on Sulawesi. He is eyeing Borneo, too. Searching the mainland for the ancestors of the Liang Bua hobbits will be difficult, however, because rocks of the right age are rarely exposed in this part of the world. But with stakes this high, such challenges are unlikely to prevent intrepid fossil hunters from trying. “If we don't find something in the next 15 years or so in that part of the world, I might start wondering whether we got this wrong,” Tocheri reflects. “The predictions are that we should find a whole bunch more.”