These “Hobbit”-like creatures first appeared at Liang Bua cave, on the Indonesian island of Flores, about 95,000 years ago. Previously, it was believed they had lived on Flores until quite recently, but new evidence published earlier this year suggests they were extinct by around 50,000 years ago.
Two hypotheses account for the evolutionary origin of Homo floresiensis.
The first is that Hobbits descend from Homo erectus, or “Java Man”, an archaic Asian hominin roughly similar in stature to us. A small population of Homo erectus, it is thought, got marooned on Flores and shrunk in body size.
The second hypothesis is that the ancestor of Homo floresiensis was an even more ancient hominin that was pint-sized to begin with. Candidates include Homo habilis or an Australopithecine, both known only from the fossil record of Africa.
Only months after the first Hobbit bones came to light, our friend and colleague Mike Morwood, co-discoverer of Homo floresiensis, set us the great challenge of unravelling the mystery of the Hobbit’s beginnings. And to do so, it was crucial to discover the identity of the first hominin colonisers of Flores.
We were already aware of the existence of ancient sites in the So’a Basin 70km east of Liang Bua. In the 1960s a Dutch priest named Theodor Verhoeven had discovered and excavated several sites with fossils of extinct megafauna and stone tools associated.
In the late 1990s, Mike’s work here with two of us (van den Bergh and Kurniawan) showed that hominin tool-makers were present by 840,000 years ago.
The So’a Basin at that time comprised a tropical savannah drained by numerous small stream channels. These grasslands teemed with pygmy Stegodon (an extinct Asian elephant), Komodo dragons and rats. Fossils from these species are preserved within rock strata that are exposed by erosion on the present land surface.
The fossil beds accumulated between 1.3 million to 500,000 years ago.
Our small-scale digs in the basin had already unearthed hundreds of fossils of Stegodon and other animals, as well as stone artefacts. In all likelihood ancestors of Homo floresiensis left these simple tools behind.
Frustratingly, however, we had never found a single bone or tooth from one of the tool-makers. We had to find these fossils.
Some observers thought that such elusive objects would only ever be found by chance, perhaps by local farmers, and probably not within our lifetimes.
The only way to accomplish our objective was to think big.
The big dig
In 2010, using funds from the Australian Research Council and the Geological Agency of Indonesia, we assembled an international team of researchers and recruited more than 120 workers from local villages.
At the site of Mata Menge we initiated one of the largest fossil digs ever undertaken in Southeast Asia since Eugene Dubois’s famous 1890s dig at Trinil in Java that uncovered the first known fossils of Homo erectus.
It took us five years of painstaking excavation in concrete-like sandstone, but on October 8, 2014 – only weeks before the project’s end – we finally found what we were looking for.
It was a young Indonesian woman, aspiring palaeontologist and PhD student Mika R Puspaningrum, who first identified it. A tiny hominin molar. More teeth followed, then a skull fragment and a piece of jaw emerged from the hard grey sandstone.
The Mata Menge fossils represent the remains of at least three or more hominin individuals, including an adult, and, astoundingly, two young children.
The sandstone containing these fossils was deposited at least 700,000 years ago, which is ten times older than the Homo floresiensis skeleton from Liang Bua.
The Mata Menge hominin is much smaller in size than Homo erectus from Java, but the teeth and jaw fragment do not resemble any pre-erectus hominin species. In fact, their closest affinity is with Homo floresiensis.
No one predicted the ancestor of the Hobbit would itself have looked like a Hobbit.
Although the Mata Menge hominins are remarkably Hobbit-like, the jaw fragment is from an adult that was 21% smaller in size than the tiniest Liang Bua Hobbit.
Homo floresiensis may actually be a bigger version of its ancestor!
Importantly, the lower molar from Mata Menge has five cusps instead of four (unlike the Liang Bua Hobbits, in which the fifth cusp is reduced), and most resembles those of Homo erectus in shape (but much smaller).
In sum, the Mata Menge fossils suggest that Homo floresiensis is indeed a kind of pygmy Homo erectus. It now appears that these castaways dwarfed in size soon after making landfall on Flores (or another nearby island, such as Sulawesi).
But Flores is full of surprises.
Until we find more complete hominin remains at Mata Menge, or even older fossil sites, we cannot be certain about the identity of the Hobbit’s ancestor and thus how this evolutionary saga began.
Our search for fossils – and funds – continues.
Iwan Kurniawan, curator at the Geology Museum in Bandung, Indonesia, was a contributor to this article.