DNA deliberations

Other features of the Flores fossils are indicative of ME cretinism, according to Obendorf, but this is another area that Brown disputes.

In normal humans the adult mandible has an angle close to 90 degrees, but in children it is shallower - as it is in adult cretins, whose skeletons exhibit juvenile states.

The double-rooted pre-molar of LB1 could be explained by the fact that ME cretins sometimes retain some deciduous teeth into their 30s. Obendorf has images taken from a 3-D scan of the jaw shown in a BBC TV documentary, which shows that the roots and the crown of this tooth shows the characteristic features of the deciduous or 'milk' molar that would have first appeared in that position.

Brown says the fossil only has adult premolars. He also disputes their reference to an unfused area on the frontal bone, which he says is excavation damage.

The aberrant trapezoid bone of the wrist, identified as a primitive feature shared with australopithecine hominins, is also characteristic of ME cretins. Obendorf says delayed ossification in cretins may result in carpal bones having two or more separate pieces that may remain unfused even in adults.

In the original Science paper describing LB1's skeleton, the trapezoid looked small, but Obendorf says one of the authors told him the image had been scaled up by 25 per cent to facilitate comparison with the other bones of the hand, and so it was even smaller than expected, and consistent with it not being a complete bone.

Obendorf says a DNA sample, which is most likely to be obtained from a tooth, should resolve the issue of whether the "hobbits" are a different human species, or pathologically deformed ME human cretins.

He is surprised that DNA has not yet been recovered. "I heard rumours a year or two ago that they had recovered DNA, but got a result that suggested it was contaminated by modern DNA.

"If we're right, the DNA should be not very different from the DNA of the indigenous, modern Floresians, some of whom worked at the site."