Here, there be dragons

When fossils of a diminutive, recently extinct race of humans were discovered in a cave on Flores, one of the most easterly islands of the Indonesian archipelago, researchers can be forgiven for assuming they were down-sized by the peculiar selection pressures that act upon insular species.

On Flores, natural selection has morphed several familiar species to unfamiliar sizes. There be dragons: three metre, 200kg monitor lizards, also known as the Komodo dragon, plus the fossilised remains of an extinct giant rodent, Spelaeomys, and extinct Stegodon pigmy elephants.

The tiny skulls of the Flores fossils ignited a heated, sometimes ad hominin, debate about whether they represented a new species of Homo, or were merely microcephalic mutants of a Floresian form of H. erectus, or even a microcephalic H. sapiens.

But a new cladistic analysis by an Australian National University PhD student suggests the diminutive, gracile humans who lived on Flores as recently as 12,000 years ago, now popularly known as 'hobbits', might not have been dwarfed by insular life.

Debbie Argue, of the Australian National University's School of Archaeology and Anthropology, has produced strong evidence that Homo floresiensis was tiny because its African ancestors were tiny.

If Argue is right, H. floresiensis descends from the first hominins to leave Africa, and this might have happened some 2.25 million years ago, around the time when the first, primitive Homo species was emerging in Africa.

If so, the hobbits' forebears could have colonised the Indonesian archipelago up to half a million years before the first large hominin species, Homo erectus - Java Man - crossed the deepwater gaps separating Java and Lombok, and Sumbawa and Flores, by means that did not involve swimming.

---PB--- Comparisons

Argue, who is supervised by renowned ANU human evolutionist Professor Colin Groves, described the results of her morphometric and morphological analysis at the ANU's recent Archaeological Science Conference in Canberra.

She compared anatomical structures of the type specimen of H. floresiensis, LB1, with several modern humans, and many ancient hominins such as H. erectus, H. ergaster, H. habilis, and the Dmanisi specimens.

Cranially, it is most similar to H. ergaster or H. habilis - Argue says this indicates only that H. habilis and H. floresiensis shared an ancestor. But H. floresiensis had long arms in proportion to its legs, and is close to the primitive arm-to-leg ratio of the gracile australopithecine, Australopithecus garhi.

When A. garhi was discovered in 1996, it was hailed as a missing link between Australopithecus and Homo, but is now regarded as the most advanced of the australopithecines.

A. garhi, which lived about 2.2 million years ago, had a cranial capacity of 450cc, less than a third that of modern humans and 33cc larger than that of LB1, the type species of H. floresiensis, which has been set at 417cc according to a study last year by Dean Falk and colleagues. Given the pronounced variation in modern human brain size, A. garhi and H. floresiensis are in the same headspace.

But australopithecines never made it out of Africa. Or did they?

"Floresiensis seems to have evolved around the time of A. garhi, given its primitive arm-leg ratio, whereas H. habilis was moving towards the modern human ratio around the same time," Argue says.

"If we're right, it means some hominin must have moved out of Africa about two million years ago, which is half a million years earlier than the Dmanisi hominin, which is supposedly the earliest out of Africa.

"Not long after H. floresiensis was discovered, someone drew attention to the skull's resemblance to one of the sub-adult skulls from Dmanisi, but morphologically, the Dmanisi crania are very similar and they are quite different from H. floresiensis.

"The 'hobbit' had a relatively small head in relation to its body, and it stood about one metre tall. Australopithecines started at just under one metre, and range up to 1.2 metres, so height is not really an issue.

"We don't have to invoke the island law - the hobbit has some very ancient australopithecine characteristics, which suggest its ancestors were tiny.

"The question is: how did it get to Flores, and when? It's likely to have left Africa between two million and 2.25 million years ago, but that doesn't mean it got to Indonesia immediately - it could have evolved along the way.

"But its presence on Flores means it must have been present on islands back up the Indonesian archipelago. If we knew where to look.

"There is a new Australian-Indonesian program to excavate other targeted areas and it could have reached Flores via Sulawesi, as some others have suggested."

---PB--- Skull and mandibles

Argue confined her cladistic analysis to the skull and mandibles of H. floresiensis and the comparative sample. She compared 40 character states related to the face, basal area of the skull, the dimensions of the back of the skull, and 30 characters of the mandibles of the various species. She is confident of her conclusions.

"If someone had found these skulls in Africa, nobody would have argued they are microcephalic," she says. A lone microcephalic individual from Flores, or even two, would cast doubt upon the claim that H. floresiensis is a new species of Homo.

But the cave on Flores has yielded partial remains of at least two other adults and one child of similar or even smaller proportions.

Argue says stratigraphic and other evidence indicates that they occupied the cave for more than 80,000 years, from about 94,000 years ago until around 12,000 years ago - around the time a major volcanic eruption devastated the island, possibly extinguishing the tiny and extraordinary Floresians.