hide announcement WIN your very own meteorite! Subscribe, gift or renew a subscription to Cosmos and automatically go into the draw – Shop now!

The struggle to understand the Hobbit


The discovery of Homo floresiensis in 2003 threw up many questions about the history of our own species. More than a decade later, they remain unanswered. Debbie Argue, biological anthropologist at the Australian National University, explains why.


An artist’s impression of H. floresiensis.
An artist’s impression of H. floresiensis.
Katrina Kenny

Way back in 2003 at a conference on archaeology I met a couple of students who told me that something amazing was coming out of Indonesia. Then they said they couldn’t say anything more.

Interesting? Tantalising? Yes, indeed. I waited months for some kind of evolution bombshell to be unveiled in the journal Nature – which was surely where such a dramatic find would be revealed. Months went by. Still nothing. Did I read too much into what the breathless students had said?

But on 28 October 2004 there it was in black-and-white. Amid a torrent of publicity: a new species of hominin that was named Homo floresiensis.

This incredible discovery that would come to rattle the status quo on human evolution comprised a series of bones excavated in Liang Bua cave on the island of Flores in Indonesia. The breakthrough was made by a team of Indonesian and Australian researchers led by the late Mike Morwood, then at the University of New England in New South Wales, and Thomas Sutikna at the Indonesian Centre for Archaeology in Jakarta. The first author on the paper was Peter Brown, also at the University of New England. As I read the publication, my reaction was probably much the same as that of everyone else: “Whaaat? How could this be?”, but also “How wonderful! What a mystery!”

Liang Bua cave.
Liang Bua cave.
University of Wollongong

In one fell swoop this publication challenged much of what we thought we knew about human evolution. Could there really have been a population of tiny beings one metre tall living at the same time as Homo sapiens? Such a scenario could hardly have been imagined before this discovery was made.

The Flores excavation was intended to deliver insights into the origins of the first Australians, but these bones were not Homo sapiens. Peter Brown, Mike Morwood and colleagues compared their characteristics to the bones of Homo erectus, H. ergaster, H. georgicus, H. sapiens, and Australopithecus africanus. They found that Homo floresiensis – affectionately known as ‘The Hobbit’ – had a mix of archaic and modern characteristics that had never been seen together in one species. Yet the bones were thought to be only 12,000 years old – placing them well within the arrival time of modern humans in the region. Since 2016, we’ve known that they are actually between 60,000 and 90,000 years old.

At the time of the initial discovery I was studying one-to-two million-year-old skulls from Africa for my PhD, so of course my interest was piqued. I was astonished at the similarity with the older African skulls. Despite the report suggesting H. floresiensis lived as recently as 12,000 years ago, and half a world away from Africa on an Indonesian island, I had to add this new find into my thesis. It was just too good to ignore.

The Hobbits were small. In fact, a partial skeleton (dubbed Liang Bua 1, or LB1) excavated at 5.9 metres depth is so tiny that at first it was thought to be the remains of a child. But analysis of the mandible revealed that all the molars had erupted, indicating a mature adult.

The brain cavity of 426 cubic centimetres is small, even for such a short being (our cavities average between 1300 and 1500 cubic centimetres). The forehead slopes back, the skull is low, yet analyses of brain imprints show the species possessed an expanded frontal cortex. This implies they were capable of sophisticated actions such as planning and learning from mistakes, and was able to pass information from generation to generation.

The jaws lack a chin, and instead have some ape-like bone structures internally, below the lower incisors. Wrist bones are also ape-like. H. floresiensis had relatively short legs, which resulted in its arms extending much lower than ours, and its shoulders would have been shrugged and hunched forward. H. floresiensis would have walked upright but with a somewhat odd gait because its feet were quite long compared to its legs. It had to lift those feet up higher than we do just to get ground clearance.

Skulls of H. floresiensis (left) and H. sapiens (right).
Skulls of H. floresiensis (left) and H. sapiens (right).
Professor Peter Brown, University of New England

None of the bones discovered were the result of deliberate burial. In fact, the excavation revealed that LB1 had, after death, sunk into mud in a shallow pool of water where she was slowly covered by silt.

Peter Brown and colleagues originally proposed two competing hypotheses about the origins of H. floresiensis. One is that the species is an early hominin similar to the earliest identified in the Homo genus. The fossils for these species are known only from Africa and are between one and two million years old. This implies that the ancestors of H. floresiensis could have got to Flores Island in the vicinity of a million years ago and survived there until at least 60,000 years ago.

Their alternative hypothesis was that H. floresiensis is a dwarfed descendant of Homo erectus, which is the only known non-sapiens hominin to once have existed in Indonesia. Its remains have been found on the island of Java. According to this view, the dwarfing of H. erectus was an evolutionary response to being isolated on an island with a limited food supply. Just as the Asian elephant evolved into the dwarfed Flores stegodon after becoming marooned on the island, H. erectus could have met a similar fate.

Some fresh additions to the origin debate came in 2016, when Gerrit van den Bergh at the University of Wollongong and colleagues announced the discovery of a small partial jaw and some teeth from excavations in a different part of the island, the Soa Basin, 74  kilometres east-southeast of Liang Bua cave.

The fossils were dated to around 700,000 years old, 600,000 years earlier than H. floresiensis in Liang Bua. They appear to solve a long-standing mystery. Flores has long been known for stone tools dating back to a million or more years ago. The makers of these tools may at last have been found. The adult jaw is fragmentary and smaller than H. floresiensis, but the discoverers consider it to be “H. floresiensis- like” although this remains a provisional interpretation until new skeletal materials are found. Nevertheless, they conclude that the remains derive from H. erectus.

There is also a third hypotheses for the origin of the Hobbit. Within a month or two of the announcement of H. floresiensis in 2004, Maciej Henneberg at the University of Adelaide and Alan Thorne (now deceased) at Australian National University proposed that LB1, at least, represented a modern human with microcephaly, a disorder characterised by a marked reduction of brain growth.

The incidence of hereditary microcephaly, however, is low in modern populations (for instance one in 30,000 in Japan and one in 2,000,000 in Scotland). Examples have been found in only five archaeological excavations so far. Laron Syndrome – sometimes dubbed “cretinism” – and Down Syndrome have also subsequently been evoked as explanations for the strange diminutive morphology of H. floresiensis. But any explanation must account for more than just the tiny skull: there is also the archaic head shape, facial features and shoulder configuration, the ape-like structure of the lower jaw and wrist, and the overly-long arms and feet.

And even if one was so lucky as to find an individual with one of these rare syndromes in excavated skeletal material, the rest of the bones in the population should be those of normal modern humans. But all the bones of H. floresiensis represent statures of one metre or less.

It is now 14 years since the discovery. We still do not have a resolution to the fundamental question: where does H. floresiensis fit on the human evolutionary tree?

The most recent study, by José Diniz-Filho of the Federal University of Goiás, in Brazil, and Pasquale Raia of University of Naples Federico II, favours the island dwarfing solution for H. floresiensis.

They employed a computer model to analyse the relative body weight and brain volumes of several specimens of H. floresiensis and H. erectus. Their H. erectus sample included 1.8 million-year-old fossils from Dmanisi, in the Republic of Georgia, and the Zhoukoutian remains, from China. Their finding was that H. floresiensis could have derived from a population of isolated H. erectus. They note, however, that even if H. floresiensis descended from an early species of Homo, that does not necessarily mean that the dwarfing occurred after they arrived on the island. The ancestor may also have been tiny to begin with.

With my colleagues, Colin Groves, William Jungers and Mike Lee, I also set out to test the two hypotheses for H. floresiensis. We knew we needed to do something different from preceding studies if we were to resolve the issue. Most of these studies focused primarily on the skull and jaw. What had not been done before was to also compare characteristics of the arms, legs and shoulders.

Debbie Argue holds a reconstructed H. floresiensis skull.
Debbie Argue holds a reconstructed H. floresiensis skull.
Stuart Hay / ANU

Because H. floresiensis displayed a number of characteristics seen in a two- to four- million-year-old group of hominins known as Australopithecines, we also included these. The species we used in our comparative analyses are: Australopithecus afarensis, A. africanus, A. sediba, H. habilis, H. georgicus, H. naledi, H. ergaster, H. floresiensis, H. erectus (from Java, in Indonesia) and H. sapiens.

We performed analyses to flesh out the relationships between hominin species. On the family tree generated by our research H. floresiensis formed a branch with 1.75 million-year-old H. habilis. On the other hand, H. erectus lay on a different part of the tree and formed a branch with the 1.5 million-year-old species from Africa, H. ergaster. We performed statistical tests on these results. In particular, we wanted to be absolutely sure H. floresiensis was not closely related to H. erectus. But we found no statistical support for such a relationship.

When two species branch in a phylogenetic tree, that suggests they shared a common ancestor that was not shared by any other species in the analysis. We therefore hypothesise that H. floresiensis shared a common ancestor with H. habilis. We do not know who that ancestor was or when it lived, but it would have to be older than the oldest H. habilis specimen known, so older than 1.75 million years. The implication is that the H. floresiensis ancestor evolved before that time in Africa, dispersed from that continent, and arrived on Flores earlier than 700,000 years ago, judging by the age of the jaw and teeth found at Soa Basin. This represents a hitherto unknown movement of very early hominins out of Africa.

It is difficult to see how this disagreement will be resolved. Maybe we need to place the question of the origins of H. floresiensis in a holding basket until we know more about the species. So far, we only know it from one cave and, probably, one other site on Flores.

There are still puzzles surrounding H. floresiensis. How did this species arrive on Flores? It is an island that is considered to have been separated from other landmasses since at least the Early Pliocene, 5.3 million years ago.

When did H. floresiensis arrive, and when did it become extinct? Some would even ask: is it extinct? The Hobbit could have arrived any time after its lineage emerged in Africa, at least 1.75 million years ago. As to its extinction, we simply do not know when this occurred. So far, the youngest date we have for members of the species is 60,000 years ago, but other finds could bring the date forward. Excavations continue in Liang Bua cave, and there are hundreds of other caves on Flores to investigate.

Could the Hobbit still exist in the wild mountain forests of Flores? When H. floresiensis was announced, the media picked up on the local folklore that small human-like creatures roam the forests. Descriptions of sightings are well recorded and quite detailed. The similarity to H. floresiensis is intriguing. But most researchers would say ‘show me the bones!’

Dr Debbie Argue is a Visiting Fellow at the School of Archaeology and Anthropology at the Australian National University in Canberra.
Latest Stories
MoreMore Articles