It was a bit like finding a penguin in the Arctic, and a yeti, and a talking dog, all rolled up into one.

The remains found in September 2003 at Liang Bua, on the Indonesian island of Flores, broadly resembled early hominins, members of the ‘tribe’ that includes modern humans – but of a kind that had lived in Africa and gone extinct there more than two million years before. On the other hand, they also bore a resemblance to the local equivalent of the yeti, described in folktales but unknown to science. But the most extraordinary thing about them was that a hominin with a chimp-sized brain was surrounded by stone tools of a kind that did not appear elsewhere until the advent of modern humans.

“I would have been less surprised if someone had uncovered an alien,” remarked Peter Brown, the principal author of the article which announced the find, last October in the journal Nature. He’s still not quite sure how to make sense of it. Half a year on, his peers now have several stories to decide between.

His team’s report described a partial skeleton that included the skull, lower jaw, pelvis and right leg. Its height was given as 105cm, it was judged to be adult, and its brain size is around 400cm3, about a third that of a typical modern human brain. It was dated to around 18,000 years ago; though it was not fossilised.

The authors assigned Liang Bua 1 (LB1) to a new species, Homo floresiensis. To much of the world, however, it was what members of the investigating team had dubbed it: the ‘hobbit’.     

For a brief moment, while palaeoanthropologists paused to take in what they were being asked to absorb, these labels were unchallenged. Then the palaeoanthropologists found their feet and resumed the ways of their tribe, re-igniting old feuds and starting new ones. A faction emerged to claim that LB1 was not extraordinary but merely odd: a small modern human with a disorder that had produced an abnormally small brain.
 
Among this group is Teuku Jacob, the doyen of Indonesian palaeoanthropology, who took possession of the specimen and was bitterly criticised by the investigators for doing so. They have since protested that when they got the material back, they found that it had been seriously damaged.
 
The controversy about the care of the specimen has drawn some of the attention away from the scientific questions that the discovery has raised. These are only increasing as the discoveries and studies proceed. A number of other finds have been made in the Liang Bua cave: a second lower jawbone, a number of bone fragments representing other individuals, and LB1’s arms. These have yet to be described in formal publications, but a paper has appeared in the journal Science discussing a ‘virtual endocast’ made using CT scans to produce 3D images of the inside of LB1’s skull.
 
At this stage there are three main interpretations of the Liang Bua bones. One is the story told in Nature, that members of Homo erectus, a species much larger in brain and body, underwent the phenomenon known from other mammalian species as ‘island dwarfing’ to produce a new species, Homo floresiensis. A second is that H. floresiensis and H. erectus both evolved from a common small-brained, diminutive and unidentified ancestor. The third is that there is no such species as H. floresiensis, only the skeleton of somebody with a condition known as secondary microcephaly.

It isn’t just the formal names that are significant. Attempts have been made to establish ‘ebu gogo’, the name of the creature from the Florinese folk tales, as an alternative to ‘hobbit’. It hasn’t caught on. But although it is more exotic, it is less fanciful; and it might turn out to be closer to the truth. If LB1 is not simply a lone pathological specimen, the two nicknames express the fundamental difference between the remaining possibilities. Ebu gogo are creatures lacking human mental abilities. Hobbits are like real people, only smaller.

The basis for the latter assumption is the association of the bones with a relatively sophisticated range of stone artefacts, including blades, pointed tools and small blades that may have been fitted to hafts as barbs. This appears to short-circuit the familiar story of hominin evolution, in which crude stone tools appear around the same time as the genus Homo, and ’toolkits’ containing differently specialised artefacts do not emerge until around the time that Homo sapiens does. There is little evidence that the smaller australopithecines, hominins comparable in brain size to ‘Florrie’, ever made even the simplest tools. Is association enough to justify the claim that hominins with chimp-sized brains made the kind of tools found at Liang Bua?

Peter Brown, who is an associate professor at the University of New England in New South Wales, considers that it is. “If you go by the evidence from the caves so far,” he say, “with stone tools going down to 95k [thousand years ago], and up until around 12k, similar types of stone tools, using a similar tradition and similar material, fairly uniform, and the only skeletal material we have in association with them so far is Homo floresiensis, then you go with them being the toolmakers.”

How far you go with such a claim is another matter. It doesn’t satisfy the principle that extraordinary claims require extraordinary evidence. Indeed, it is based on the absence of evidence. "It only needs the find of one modern human fossil - even a single tooth - to invalidate the assumption that the only tool-maker in the cave before 12,000 years ago was Homo floresiensis," says Chris Stringer, Head of Human Origins at the Natural History Museum in London.

Stringer tells the cautionary tale of how Louis Leakey, perhaps the most famous palaeoanthropologist of all, “doggedly and increasingly desperately searched the deposits at Olduvai Gorge for 25 years, for the remains of the earliest tool-maker. When his wife Mary found the skull of ‘Zinjanthropus’ in 1959, Louis quickly declared that the search was over, since it was found close to stone tools. Two years later, the discovery of Homo habilis fossils in the same levels led to a complete rethink, with the announcement that Homo habilis was actually the first tool-maker.”

Had the Liang Bua tools likewise been found in East Africa, they would have been classified as Middle Stone Age artefacts and therefore the work of modern humans. Dates of 95,000 years for the oldest examples would not have raised any eyebrows. But as the archaeologists began to find the tools at Liang Bua, before the bombshell of the hominin remains, they thought they were looking at evidence that modern humans had been in Australasia for 40,000 years or so longer than had been hitherto supposed. Modern humans would have reached Flores during the dispersal that took them to Australia, where they are believed to have arrived by around 55,000 years ago. Could they have been in the region getting on for twice as long as that?

"I can't think of any good theoretical reason why not," says Rob Foley, Director of the Leverhulme Centre for Human Evolutionary Studies at Cambridge University. And whenever it was that they arrived, argues Chris Stringer, they surely wouldn’t have shared Flores with Homo floresiensis for tens of thousands of years without entering the shelter of the Liang Bua cave: “Even the large-brained and powerful Neanderthals gave way to modern humans in Europe, so what would have kept Homo sapiens out of Liang Bua - surely not ‘hobbits’?”

The very presence of the small hominins on Flores has been taken as a sign of their intelligence. Getting there involved crossing water, a feat achieved by only a limited number of mammals from the Asian side of Wallace’s Line, the sea barrier between the Asian and Australian continental masses. Elephants managed it, being large herd animals that are good at swimming. So did rats, which are thought to have drifted across on uprooted trees rafts of vegetation. The Australian zoologist John Calaby once suggested that the first humans to reach Australia might also have been “unwilling castaways carried off in floods”, and that “perhaps even one young pregnant female” might have been sufficient to populate the continent. But his colleagues have generally felt that for hominins to get beyond Asia, they needed watercraft, language and organization.

They must have had these over 800,000 years ago, argued Peter Brown’s colleague Mike Morwood, in a 1998 paper reporting dates of such antiquity for stone tools found on Flores. Speaking last December, he repeated the suggestion. A month later, however, he co-wrote a newspaper article citing new evidence that had become available: not from scientific research, but from the Boxing Day tsunami. Calaby’s observation, hitherto used ironically, looked different in the light of the news story about a pregnant woman who had survived after five days at sea, clinging to an uprooted tree.

Accidental colonisations such as these would have involved very few individuals, which might make the dwarfing hypothesis more plausible. ‘Florrie’ has a very small brain, even relative to her body size. In this respect she is comparable to australopithecines and early Homo, rather than to erectus or sapiens. Over the course of hominin evolution, the story is clear: brains triple in size. Yet they are costly organs; for women giving birth to large-headed infants, and to maintain thereafter. An adult human’s brain accounts for 2 per cent of the body’s weight, but demands 20 per cent of the body’s energy intake. If the benefits had not exceeded the costs, large brains would not have evolved. The dwarfed-erectus hypothesis implies that on Flores, hominins found a way of avoiding the price that the rest of their genus was forced to pay.

Isolation might, however, have eased the pressure. According to one influential school of thought, large brains evolved in response to the pressure of competition within complex social systems. Such competition might be reduced in a population which started small, and remained so. Robin Dunbar, of the Evolutionary Psychology and Behavioural Ecology Research Group at the University of Liverpool, believes that in Homo, the main reason for large group size is to regulate access over wide areas to resources such as waterholes. “You could plausibly argue that in the kind of habitat you have in the Sunda shelf islands [which include Flores], that is much less of an issue,” he says.

For the critics who argue that Florrie is a pathological specimen, however, such discussion is premature at best. Maciej Henneberg, Head of the Department of Anatomical Sciences at the University of Adelaide in South Australia, insists that “a pathology should be excluded before one creates a new species”. He remains unpersuaded by the virtual-endocast paper, whose principal author was Dean Falk of Florida State University. It compared LB1’s braincase with fossil hominin and modern specimens, but the only microcephalic one was of a different kind to that which he and his colleagues have suggested LB1 might have been. That, in his opinion, “is like comparing a patient with tuberculosis to a patient with a broken leg.” There are, he observes, some 150 types of microcephaly described in the literature.

His remark illustrates the scale of the challenge posed by the Liang Bua finds. Writing in Nature, Rob Foley and his co-author Marta Lahr described LB1 as “the most extreme hominin ever discovered”. It has provoked correspondingly extreme reactions. “One thing is for sure,” observes Robert Kruszynski, a curator who looks after the anthropological collections at the Natural History Museum. “The LB finds are testing anthropological knowledge to its limits and perhaps beyond them.”

Nor is DNA likely to resolve the issues. Both the investigators and the critics have taken steps to analyse material, but tropical conditions are considered poor for the preservation of DNA, and samples are likely to be contaminated by DNA from the people who have handled and breathed on the specimens.

In the endocast study, LB1 most closely resembled a “classic” erectus specimen. But Dean Falk and her fellow authors were clear that LB1’s brain was not “a miniaturized version” of an erectus or sapiens brain. They were inclined  to interpret its similarities with erectus as a sign not that the Flores hominins were descended from erectus, but that they and erectus were both descended from an unknown hominin with a small body and brain.

Having seen LB1’s long arms and other new information from Flores, Peter Brown is likewise leaning towards the idea that “there was some more australopithecine-like ancestor involved”. The trouble with the Liang Bua finds is that, in Brown’s words, they are like “a grab-bag of hominid spare parts”. Their scope for interpretive rearrangement is far from exhausted. In his initial letter to Nature, Brown proposed putting LB1 in a genus of its own. We shouldn’t assume that she will remain Homo floresiensis for ever. And in time we may even be able to decide whether she was ebu gogo, a dwarf or a hobbit.