Face to face with ‘the Hobbit’ of Indonesia

By Susan Hayes 15 June 2015
Reading Time: 9 Minutes Print this page
The fossil of a totally unexpected tiny species of human was discovered in Indonesia in 2003. We look at what we know about this discovery and what facial reconstruction may reveal

Every skull has a quirk that doesn’t quite fit the pattern. And every human skull was once a living person – someone with particular likes and dislikes, who had family and friends, who ate, who slept and who dreamed. Each and every skull is the only one of its kind, and all skulls demand respect for the lives they once lived. That being said, the skull of the ‘Hobbit’, Homo floresiensis, took being unique to a whole a new level. From the moment she was excavated, this ancient and totally unexpected new species of human turned what we knew about human evolution on its head.

My specialty is facial anthropology and forensic facial approximation, which, to the lay person, could be described as facial reconstruction. Many researchers in my field prefer the term facial approximation, because the results of estimating the ‘average’ appearance of soft tissues from fossils are, of necessity, always going to be an approximation. No-one is average, nor is their skull.

My job – with the help of research funding from the Australian Geographic Society – was to take the only skull we now have from the 11 or so known individuals of H. floresiensis and use computer software to attempt to reconstruct her facial appearance. More on this and the results later – but let’s start at the beginning of this story with the remarkable and surprising discovery made in 2003.

Flores: a special place in human history 

WHAT NOT MANY people realise is that the home of H. floresiensis, the Indonesian island of Flores, occupies a special place in the history of the planet. Ever since it began emerging above the sea some 10 million years ago, Flores – located about 500km east of Bali – has been an island. And the fossil record of vertebrate animals that it holds shows very little variation – a consequence of never being connected to a larger landmass.

The older inhabitants, those that lived on Flores more than 1 million years ago, include a 1m-tall pygmy elephant (Stegodon sondaari), a giant tortoise and the Komodo dragon. More recent inhabitants can be found within stratigraphic layers dated to 900,000 and 700,000 years ago. By this time, the pygmy elephant had been replaced by a 1.8m-tall dwarf species (Stegodon florensis), and a giant rat (Hooijeromys nusatenggara) had appeared.

To palaeontologists, fossil evidence of a limited number of endemic vertebrate animals, including dwarf and giant versions of mainland animals, is a clear indication that the bizarre phenomenon known as the ‘Island Rule’ is at play. Being small has distinct survival advantages on an island: smaller mammals need less food, which is often a restricted resource anyway. In such an environment, therefore, smaller individuals are likely to be physically fitter, live longer, and produce more surviving offspring.

Consequently, over long periods of time, large vertebrate animals become smaller in order to adapt and survive. In an interesting twist, some smaller vertebrates, such as the rat, then become bigger to fill the empty niche left by the shrinking large vertebrates (islands do not have many medium-sized mammals).

Flores is also unique because it is the only island, so far, to have yielded fossil remains and artefacts proving the presence of early human relatives. These ‘hominins’ include extinct humans such as Homo erectus and Homo neanderthalensis, as well as Homo sapiens. H. floresiensis was discovered on 7 September 2003, deep within an excavation against the eastern wall of a huge cave known as Liang Bua. A gaping limestone formation, Liang Bua has been an archaeological excavation site since the 1950s. The team who discovered her was searching for evidence about the migration of our own species (early H. sapiens) to Australia, via Indonesia, many tens of thousands of years ago.

Starting in 2001, Professor Raden Panji Soejono of the Indonesian National Centre for Archaeological Research, and legendary Australian archaeologist Professor Mike Morwood (see page 95), led a series of Indonesian-Australian excavations in Liang Bua, employing a new technique enabling them to excavate safely to depths of more than 10m. The focus of these excavations was a search for fossil evidence of the first early humans to arrive on Flores. Although stone tools clearly show hominin occupation from about 1 million years ago, the bones of early toolmakers of our own species were proving elusive.

No-one was expecting to find a new hominin species of only 1m (3.5 feet) tall, but this is exactly what happened. Embedded within a matrix of silty clay 6m down, and displaying disproportionately large feet for such a small frame, ‘Hobbit’ was announced to the world in 2004 as Liang Bua 1 (LB1), and she has held international scientific attention ever since.

Although her bones were demineralised and soft (often described as having the consistency of wet blotting paper), LB1’s skull and skeleton were very well preserved. One estimate later dated her to 18,000 years old, which indicates her species overlapped with our own in time. But they may or may not have lived with them on Flores itself, as modern humans did not arrive until an estimated 11,000 years ago.

If this dating holds true, it could mean that H. floresiensis is the human species to have most recently coexisted with our own on the planet – Neanderthals, for example, were extinct by perhaps 30,000 years ago.

Puzzlingly, LB1 exhibits a mixture of skeletal features both primitive and modern, complicating our understanding of how she fits into the hominin family tree. This mix of features is also present in the other 10 H. floresiensis fossils, although these later finds – which are up to 38,000 years old – are not as complete.

Many international scientists are trying to solve the puzzle of where H. floresiensis fits within human evolution. According to some researchers, the species is a dwarf version of Homo erectus. Fossils of H. erectus have been excavated from the much larger Indonesian island of Java (the famous ‘Java man’, for example), and the evidence shows they lived there between 1.5 million and 150,000 years ago.

According to this hypothesis, a small population of H. erectus became stranded on Flores. Big tsunamis occur in the Indonesian region every 100 years or so, and it is possible a tsunami washed a small group out to sea and carried them to the island’s shores (during the 2004 Banda Aceh tsunami, for example, some people were flushed many kilometres offshore and were found alive, days later, clinging to rafts formed from plant debris).

Finding themselves completely isolated on the small island of Flores, this group of H. erectus could have then become subject to the Island Rule – across thousands of generations and 1 million years they gradually shrank to the size of H. floresiensis – similar to what happened on islands with other large-bodied mammals (such as the dwarf elephants) stranded there. In other words, a selective pressure towards smaller body size would have been continuously acting on a H. erectus population on Flores.

Moreover, because the fossil evidence shows that the island did not have large mammalian carnivores such as hyenas, tigers or wolves, the advantage of being bigger to defend oneself from large predators was not necessary. Nor would this group of hominins have needed a large body size to travel long distances in search of food – the island is only 60km at the furthest from north to south, and 300km from east to west. Being big, therefore, was not necessary for the survival of either H. floresiensis or the pygmy elephants, whereas being smaller meant living well on limited resources.

But nobody, at present, can rigorously test this hypothesis. Although skeletal body remains of H. floresiensis (LB1 is the only specimen found with its skull) are comparatively plentiful, this is not the case for H. erectus. So far no H. erectus fossils have been excavated from Flores, and, despite many of their fossil crania (mostly just skull-caps) having been found on Java, few other skeletal remains have been unearthed.

So a thorough comparison between the skeletons of both species – which would help determine the ancestry of Hobbit – is not yet possible. The search, however, continues.

Facial reconstruction of the Hobbit

MY ROLE IN the H. floresiensis story began when I joined the Centre for Archaeological Science (CAS) at the University of Wollongong in 2012. Professor Richard Roberts, CAS director and Australian Laureate Fellow, introduced me to Professor Mike Morwood, and I started work with Mike and Thomas Sutikna, an archaeologist from Indonesia, on the facial approximation of LB1. Thomas and Mike had worked together for many years, and Thomas (together with Wahyu Saptomo, Rokus Due Awe, Jatmiko and Sri Wasisto from Arkenas) was instrumental in the discovery of H. floresiensis.

Working with Hobbit was certainly challenging, but Mike and Thomas were great collaborators, and I knew they were the two people with the most experience regarding this strange little hominin. Our experiment aimed to discover what would happen if we applied what we knew about forensically approximating the faces of modern humans to an archaic, extinct hominin.

I had never worked with such an archaic human fossil before, so, although there is always extensive background research before starting to approximate a face, this project also included a steep learning curve about the relevant palaeoanthropology (the research that has been undertaken with H. floresiensis has been extensive since 2004). To make the approximation, we used CT scans of the skull, 3D imaging and computer modelling.

We found that Hobbit herself (she’s the only individual of H. floresiensis for which we have a skull) looked more like us than many people might have expected. And, although by today’s standards of human evolution she’s not exactly what you’d call pretty, she is certainly distinctive. Hobbit doesn’t have those hyper-feminine, modern-human features such as big eyes, nor is there much of a forehead, and she doesn’t have a chin at all. But our results show that she looks a lot less like a chimpanzee than some of the illustrations created by many palaeo-artists.

At the Australian Archaeological Association conference in late 2012, the response to our results was, for me, astounding. Mike and Thomas were well used to all the attention, but I was not. Within hours I was receiving emails from all over the world, and our Hobbit face was all over the internet.

My worry was that we had not yet submitted our research for publication, and, in the sciences, international peer-review is essential. Mike’s response to the paper we submitted in early 2013 was that it was too focused on the technical process and not easily understood by the wider community. As ever, he was right.

Sadly, Mike died in July 2013 with projects yet to do. This was shortly after our research appeared in the international Journal of Archaeological Science, so he never knew how we were received by the media – which was that we thought Hobbit had a very modern face. To be fair, our results do show a more modern facial appearance than some Hobbit faces produced by palaeo-artists. This is probably because our approach was based on evidence, and because we applied knowledge of faces derived from large numbers of modern humans. That was unavoidable, because we know very little about the soft tissues of ancient hominins.

The results of reconstructing a face

Our research paper concluded with a comparison of our results with nine other facial reconstructions of Hobbit’s skull, produced by international palaeo-artists. Most of these faces are on display in natural history museums around the world. Because every artist was working from the same skull, we thought it would be interesting to see how our face compared with the rest.

We used statistical methods to analyse the common facial landmarks (eyes, noses, mouths, etc.), and the results were surprising. All nine facial reconstructions are very, very different. The eyes, mouths and noses are different widths and lengths – even though Hobbit’s skull clearly shows where these features sit, each of the palaeo-artists have arranged them in different places.

This could be due to how, or if, the damaged parts of the bones of the H. floresiensis skull were ‘reconstructed’ by each artist. Every fossil is distorted by being in the earth, and, because Hobbit’s bones were like butter within their silty clay encasement, further inadvertent damage occurred both during and after excavation. The first thing we had to do, therefore, was try to ‘repair’ the bones – some were twisted out of shape, which is common with fossils, some surface areas were missing and some bones were broken. Naturally, our ‘repairs’ were undertaken on the CT scans of Hobbit’s skull, not the actual bones.

Another issue with some reconstructions is that many, if not most, of the methods used don’t yield scientifically accurate results. The scientific methods for approximating facial appearance are relatively new, and are published in academic journals, not popular science books, so they are not easily -available.

This is unfortunate because these visually compelling displays of evolution attract public interest to archaeological science, and, by association, support less exciting areas of research. I’m now thinking about producing a guide for palaeo-artists to inform them on what’s new, what still works and what methods to avoid.

My next project is with the only South-East Asian Homo erectus fossil to still have his facial bones (a specimen known as Sangiran 17). Preparation for this project was completed with AG Society sponsorship, and later this year I hope to experiment with my Indonesian colleagues from the Geology Museum in Bandung, Java, in producing a 3D model of his face (currently my results are two-dimensional computer models).

Understanding the Hobbit

NOW THAT WE have a more accurate facial approximation of Hobbit, it helps us to understand this species as being people that in many ways were like us, despite their diminutive size. But what happened to these people and did they overlap with the first modern humans to arrive on Flores?

Evidence of the presence of Hobbits in the form of stone tools and butchered pygmy elephant bones is found in layers of earth at the Liang Bua cave dated from 95,000 to 13,000 years ago. While the 11 fossils themselves date from 38,000 to as early as 18,000 years ago, if this dating is correct.

Current thinking from some experts is that a volcano – ash from which is found at Liang Bua – on Flores sent these people to extinction about 12,000 years ago. To further complicate the matter, there are legends in Flores of small, hairy cave-dwellers known as ‘Ebu Gogo’, which some people see as roughly fitting the description of Hobbits.

But many of the world’s cultures have legends about ‘little people’, and, being human, Hobbit was probably not very hairy – or so the evidence suggests. The conclusions we can draw from these legends, therefore, are not at all clear cut.

It could be that other skeletons of H. floresiensis were discovered many years ago, and that they now form part of the cultural memory of Indonesians there today. In much the same way that now-extinct animals such as the marsupial lion and other Australian megafauna probably relate to some of the legendary creatures in Aboriginal lore across our continent.

Future research will surely unveil more fascinating details about H. floresiensis – an unexpected species that has captured the world’s imagination.