By Paul Rincon
BBC News science reporter
A US-British team of scientists has challenged the idea that the tiny skeleton from Indonesia dubbed the "Hobbit" is a new human species.
The Hobbit's (L) skull is very small
Writing in Science magazine, the team presents an alternative theory that the remains could be those of a modern human with a brain disorder.
Their arguments appear in a technical critique of previous research into the Hobbit brain also published in Science.
But the authors of that earlier paper have vigorously defended their work.
The skeletal remains were discovered by an Australian-Indonesian research team in the cave of Liang Bua on the island of Flores in 2003.
After carefully analysing the bones, the group declared them to be those of a human species previously unknown to science, and to which they gave the classification Homo floresiensis. (The specimen is also sometimes referred to as LB1 after the cave in which it was found).
The creature stood just 1m (3ft) tall and possessed a brain size of around 400 cubic centimetres (24 cubic inches) - about the same as a chimp's brain. Dating of the sediments around the remains indicated the Hobbit lived only 18,000 years ago.
LB1 caused a sensation when it was unveiled to the public through a publication in the academic journal Nature.
A subsequent study published in Science in April 2005 focussed on LB1's brain. A team led by Professor Dean Falk, of Florida State University in Tallahassee, compared a cast taken from the inside of the braincase with other similar casts from primitive and modern humans, including one individual with the condition microcephaly.
This disorder is characterised by a small brain and is sometimes associated with other defects.
Professor Falk's data supported the idea that LB1 was not a modern human but a creature new to science.
The brain volume of LB1 is similar to that of a chimp's
Now, biologist Robert Martin, of The Field Museum in Chicago, and colleagues have questioned this conclusion. They presented some of their arguments in a BBC documentary last year, but the Science paper represents the team's formal technical position.
"There is a fundamental problem of the tiny brain size combined with the sophisticated stone tools," Dr Martin told the BBC News website.
Some of the tools found with LB1 are of types previously only associated with modern humans (Homo sapiens).
Dr Martin also invokes a biological rule of scaling to argue that LB1 could not have been a dwarfed version of the older human species Homo erectus, as has been suggested.
H. erectus is known to have lived on nearby Java, and one theory proposed that a population of this species could have settled on Flores and evolved a small stature. This can happen in remote, isolated habitats, as organisms adapt to a scarcity of resources. The scaling rule is based on known instances of so-called insular dwarfing in mammals.
But these studies show that a reduction in body size is accompanied only by a comparatively modest reduction in brain size. Dr Martin and his colleagues argue that the brain of LB1 is far too small to be a dwarf hominid, or human-like species.
Dr Martin used the scaling law to get to a brain size of 400 cubic centimetres using H. erectus as the starting point. The scaling law predicts a creature only 2kg (4.4lbs) in weight.
However, Professor Falk questioned whether basing a calculation of dwarfing in a hominid on an example of dwarfing in an elephant - one of the models used by Dr Martin in his analysis - was appropriate.
According to one theory, LB1's ancestor was not H. erectus at all, but a smaller, ancestral hominid such as H. habilis; or Australopithecus, an even more ancient form. Some think this could explain the small brain of H. floresiensis without breaking the scaling law.
Chris Stringer, head of human origins at the Natural History Museum, UK, commented: "There are some interesting issues such as scaling of the brain and whether a human could have as small a brain normally as this creature seems to have.
"But if you look at the bigger picture, there are two jawbones and remains from the rest of the skeleton from several other individuals."
He told the BBC News website: "When we look at the rest of the material, including the post-cranial bones, we're finding this is a strange kind of human. It doesn't seem to be a modern, pathological individual. It seems to be a primitive human - one that's distinct from anything we've found so far."
Professor Stringer pointed to the robust character of the bones in general, the form of the shoulder blade and the thick, chinless jawbones as particularly indicative that researchers were dealing with a novel human species.
"Some of the material [at Liang Bua] is believed to go back to 70,000 years and the most recent material to 12,000 years. We're not talking about one individual at one point in time. This morphology is represented over a considerable period in time," he said.
Dr Martin also challenges Professor Falk's comparison of the adult LB1 with a specimen from a 10-year-old microcephalic. He contends the Indonesian example should have been matched against individuals with a mild form of microcephaly that permitted survival into adulthood.
The Field Museum researcher provides his own microcephalic specimens by way of comparison. But in their response, Dr Falk and colleagues described Dr Martin's comparison as "inadequate" and lacking "crucial details".