Front view of Homo floresiensis skull Click to enlarge image
This skull (cast) is part of a fairly complete unfossilised skeleton of Homo floresiensis known as LB1. It is the type specimen of the species and dates to between 100,000 and 60,000 years old. The other skeletal remains of LB1 include leg bones, parts of the pelvis, hands and feet, and some other fragments. It is assumed to belong to a female aged about 30 years old. She stood about 1 metre tall, had a brain about 380-420cc and weighed about 25 kilograms. Image: Carl Bento
© Australian Museum

Conflicting interpretations and debates surround the remains of these tiny humans from Indonesia. H. floresiensis are not our ancestors but their unusual features and recent survival indicates our human family tree is more complex than once thought.

Background of discovery


The human remains date from about 100,000 to 60,000 years old, but archaeological evidence (mostly associated stone tools) suggests H. floresiensis lived at Liang Bua from at least 190,000 to 50,000 years ago (recent dates published in Nature, March 2016). These dates make it one of the latest-surviving humans along with Neanderthals, Denisovans and our own species H. sapiens.

Their disappearance coincides with that of other local fauna such as the Stegodon, the giant marabou stork and various vulture species. The loss was originally attributed to a volcanic eruption that occurred on Flores approximately 12,000 years ago but the recently published dates nullify this suggestion. Instead, it is now considered possible that the arrival of modern humans played a role. Although there is no evidence of modern humans in Liang Bua cave until 11,000 years ago, our species was moving through the region about 50,000 years ago.

Important discoveries

A joint Australian-Indonesian team, looking for evidence of the early migration of Homo sapiens from Asia to Australia, stumbled on the remains of a small human in the cave of Liang Bua, Flores, in 2003. The discoverers (Peter Brown, Michael Morwood and colleagues) argued that a variety of primitive and derived features identified the remains as that of a new species. Descriptions of some of the remains and the new species designation were published in October 2004.

The remains include a largely complete skeleton with skull (LB1) and parts of at least eleven other individuals. These remains come from different levels and range in date from 100,000 to 60,000 years old. An arm bone, from a deeper level and dating to about 74,000 years old, is provisionally assigned to H. floresiensis. A more accurate designation is difficult to make as LB1 lacks an arm bone to make comparisons with.

Stone tools have been recovered from a number of levels and range in dates from 190,000 to 50,000 years old.

As the remains are relatively young and unfossilised, researchers hoped to find mitochondrial DNA. Initial efforts were unsuccessful, but the research continues.

Excavations from 2007 to 2014 used new dating techniques to understand the complex cave stratigraphy. The publication of the revised dates (in Nature, March 2016) led to a reassessment of the causes of the species' extinction.

Key specimens

  • LB1 – type specimen discovered in September 2003. It is unfossilised. The remains consist of a fairly complete skull and partial skeleton including leg bones, parts of the pelvis, hands and feet, and some other fragments. It is assumed to belong to a female aged about 30 years old. She stood about 1 metre tall, had a brain volume of about 380-420cc and weighed about 25 kilograms. The body was not deliberately buried but covered soon after death by fine sediments, when still partially fleshed.
  • LB6 – a partial skeleton belonging to a shorter individual than LB1. It has a more V-shaped jaw and is assumed to be a child, possibly only 5 years old.

In 2016, scientists announced they had discovered the lower jaw and teeth from at least one adult and possibly two children of what may be an early form of H. floresiensis. These fossils were found at Mata Menge, about 70kms east of Liang Bua cave on Flores and date to 700,000 years old.

Skeleton of LB1, <i>Homo floresiensis</i>
Skeleton of LB1, Homo floresiensis Image: University of Wollongong
© University of Wollongong

What the name means

The genus name Homo is the Latin word for ‘human’ or ‘man’. The species name floresiensis recognises the island of Flores in Indonesia where the remains were found.

They are commonly referred to as the ‘hobbits’, after the Lord of the Rings characters, in reference to their small size and large feet.


All remains come from the cave of Liang Bua on the island of Flores in Indonesia. Flores lies towards the eastern end of the Indonesian island chain.

Flores has always been separated from mainland Asia - even at low sea levels the water-crossing was at least 24 kilometres. It is known that other animals reached Flores by swimming or floating on debris. How or when H. floresiensis reached the island is unknown.

Relationships with other species


When first discovered, it was suggested that H. floresiensis was possibly descended from Javanese H. erectus. However, more detailed analysis of skeletal remains has uncovered traits more archaic than Asian H. erectus and more similar to australopithecines, H. habilis or the hominins from Dmanisi in Georgia (classified as Homo ergaster or Homo georgicus). Most scientists that accept H. floresiensis as a legitimate species now think its ancestor may have come from an early African dispersal by a primitive Homo species similar in appearance to H. habilis or the Dmanisi hominins. This means that it shared a common ancestor with Asian H. erectus but was not descended from it. Cladistic analysis supports the lack of a close relationship with H. erectus.

The recently announced (2016) discovery of a jawbone and some teeth from Mata Menge on Flores help fill in the time gap between H. floresiensis and its earlier ancestor. Stone tools that may have been made by H. erectus (or a similar species) were discovered on Flores. These date to 840,000 years ago, so indicate that a hominin species was probably living on the island at that time.

Whatever the origins of the ancestral population, it is accepted that the population underwent long term isolation on the island and some insular dwarfing (although they were probably small to start with) which resulted in an endemic 'dwarf' species H. floresiensis. This is a common phenomenon seen in other mammals in similar environments.


Modern humans arrived in Indonesia between 55,000 and 35,000 years ago, and may have interacted with H. floresiensis, although there is no evidence of this at Liang Bua.

Interestingly, local legends exist in Flores of the Ebu Gogo – small, hairy, cave dwellers similar in size to H. floresiensis. It is suggested that perhaps the hobbits survived longer in other parts of Flores to become the source of these stories.

A new species or a small Homo sapiens?

Doubts that the remains should be classified as a new species are voiced by a number of scientists, some vehemently. They claim that the remains come from a modern human with some sort of physical disorder. The alternate suggestions include:

  • the remains are from a very small human that suffered from some type of disease that causes microcephaly, a developmental disorder of the brain that causes it to be much smaller than normal
  • the remains are from a human with Laron syndrome, a disorder that results in pituitary dwarfism (published in 2007 by a team from Tel Aviv University, Israel)
  • the remains are those of dwarfed Homo sapiens similar to the small-bodied humans that inhabited the Micronesian island of Palau between 1400 and 3000 years ago. These people shared some features with the H. floresiensis specimens, but not all. Detailed analysis of the Palau specimens is unlikely to settle arguments over the status of H. floresiensis but they do suggest that some of its unusual features could be due to environment rather than ancestry.

Many of those rejecting the new species status focus only on the remains of LB1, and ignore the other remains that show many of the same characteristic features. In contrast, a number of recent analyses of the skull, face, foot and wrist have confirmed the many unusual primitive features of H. floresiensis remains and stated that they are more similar to australopithecines. For instance:

  • a cladistic analysis done in 2009 supported H. floresiensis as a separate species (Journal of Human Evolution Online as of 4 August 2009)
  • a study using 3D-morphometrics showed that the skull of LB1 differs significantly from all H. sapiens skulls, including those of small-bodied individuals and microcephalics, and is more similar to the skull of Asian Homo erectus (Baab & McNulty, Journal of Human Evolution, 2008)
  • a team of experts created detailed maps of imprints left on LB1’s braincase and concluded that the remains belonged to a new species. Comparisons of different parts of the brain showed it was nothing like a microcephalic's and is also different from modern humans. The endocasts also revealed that parts of the frontal lobe and other features were consistent with higher cognitive processes. (Falk et al, PNAS, January 29, 2007,)
  • H. floresiensis wrist is almost indistinguishable from an African ape or early hominin wrist and is nothing like those of modern humans or Neandertals. The distinctive shapes of wrist bones form during pregnancy and as most pathologies and growth disorders affect the skeleton well after that, this demonstrates that the remains come from a new Homo species. (Matt Tocheri et al, Science 21 September 2007)
  • studies on brain-size reduction in dwarf hippos from Madagascar revealed that brains shrank to volumes well below predicted sizes. This refutes a key argument used by sceptics who claim the brains are too small for dwarfing alone to be the cause (hence they cite microcephaly or similar disorders), (Weston & Lister, Nature, 7 May 2009)

Reconstruction of <i>Homo floresiensis</i> based on skeleton LB1by Atelier Daynes, Paris.
Reconstruction of Homo floresiensis based on skeleton LB1by Atelier Daynes Paris. Image: Eunostos
© n/a

Key physical features

Body size and shape

  • very small stature of about 1 metre (the remains of 5 individuals were used for this average), which is less than the average for short populations of humans such as pygmies (who average about 1.4-1.5 m tall)
  • the wide pelvis and hunched shoulders give it a different body shape compared with H. sapiens


  • combination of features not seen in other hominin species
  • has features that are not found in H. sapiens, particularly in the temporal and frontal lobes
  • small brains averaging 380cc, about the size of a chimpanzee. This suggests there was active selection for a small brain size, possibly related to the reduced energy requirements of small brains.
  • enlarged Broadmann area 10, an area of the brain that appears to help with complex cognitive activities


  • cranial shape is long and low and closer to that of H. erectus than H. sapiens
  • receding and small forehead
  • thick bones within the range of H. erectus and H. sapiens
  • flat face
  • brow ridges over each eye that do not form a continuous brow ridge as in Indonesian H. erectus
  • narrow nose

Jaws and teeth

  • lacks the bony point on the chin found in modern humans
  • relatively large jaw and teeth that resemble H. erectus but with more primitive features
  • premolar roots different from H. sapiens
  • small post-canine and canine teeth
  • parabolic or V-shaped dental arcade typical of Homo
  • bony shelf at the front of the lower jaw which is a primitive feature not seen in H. erectus

Limbs and pelvis

  • bones and joints of the arm, shoulder and the lower limbs suggest that H. floresiensis was more similar to early humans than modern humans
  • characteristic bipedal foot that includes a big toe aligned with other toes and a locking mechanism on the middle of the foot to help stiffen the arch after heel lift occurs
  • several primitive features include a relatively long foot for its body size (70% as long as the thigh bone, compared with 55% for modern humans), a flat arch lacking the spring-like mechanism used to store and release energy during running, and a short big toe. These features are similar to ancient hominins such as H. habilis (OH8) and australopithecines and suggest the gait was different from and less efficient than modern humans.
  • unusual low twist in the upper arm bone
  • wide leg bones compared to the length
  • relatively short and curved clavicle
  • shape of the shoulder blade resulted in the shoulder being moved forwards slightly as if hunched
  • wrist bones differ significantly from the those of modern humans and are more similar to African apes or australopithecines. They lack features that evolved with the ancestors of modern humans at least about 800,000 years ago. In particular, the trapezoid bone is pyramidal in form, whereas modern humans have a boot-shaped trapezoid.
  • primitive flared ilium blades in the pelvis, similar to australopithecines, and females have wider pelvises than H. sapiens females
  • relatively long arms

Foot of LB1 compared to a modern human foot.
Foot of LB1 compared to a modern human foot. Image: University of Wollongong
© University of Wollongong



Stone tools were found in a number of different layers dating from 190,000 to 50,000 years ago. Tools include simple flakes, points, perforators, blades and microblades which were possibly hafted as barbs. Some were found with the remains of LB1, but most came from the same location as the remains of the extinct pygmy elephant Stegodon. This suggests that H. floresiensis was hunting these small elephants. Stone tools produced by heavier percussion were also recovered from layers not associated with H. floresiensis occupation. These tools date to about 102,000 years ago. The makers are unidentified.

There has been some speculation that the stone tools associated with H. floresiensis were actually made by H. sapiens. The basis for this is purely the belief that humans with such small brains couldn’t make such sophisticated stone tools – there is no other evidence in support of this. However, those studying the tools claim they are not as sophisticated as they appear and regard them as ‘simple’.

Analysis of the residues and polish on some of the tools revealed they were used for working wood and fibrous materials, perhaps to make spear shafts or items such as traps. Cut marks on the Stegodon bones also suggest some of the tools were used to process meat.

Precursors to this tool kit may come from earlier sites on Flores. Tools excavated from Mata Menge (about 50km from Liang Bua) in 2004-5 are at least 700,000 years old, and those from the Soa Basin date to about 880,000 years old. Tool kits from both sites show some similarities and technological continuity with those found in Liang Bua cave. The identity of the makers is unknown, but they could possibly be ancestral to H. floresiensis.


There is evidence of the use of fire in Liang Bua cave. The remains of numerous juvenile Stegodon have charred bones, possibly indicating that H. floresiensis was able to control fire for cooking.


There are no traces of pigments, ornaments or deliberate burials in the layers associated with H. floresiensis – all of which characterise the modern human levels from the upper parts of the cave.

Environment and diet

Flores is a heavily forested tropical island with mountain peaks reaching over 2000 metres. The environment during H. floresiensis time would have been similar. The nature of their environment and the limited food sources typical of such islands provides strong clues to the evolution of H. floresiensis. When a small population becomes separated, changes can occur very quickly. This particular environment favours reduced energy requirements with dwarfing a response to this. Several dwarf species, including Stegodon, have been recovered on Flores and other small islands.

This species shared the island with pygmy elephants Stegodon, giant rats and large lizards like Komodo dragons. Evidence of cut marks on the Stegodon bones from Liang Bua cave show that H. floresiensis was at least hunting and eating this animal.

Further reading:

Brown, P.; et al. (27 October 2004). ‘A new small-bodied hominin from the Late Pleistocene of Flores, Indonesia’. Nature. 431 (7012): 1055–1061

Morwood, M. J.; et al. (13 October 2005). ‘Further evidence for small-bodied hominins from the Late Pleistocene of Flores, Indonesia’. Nature. 437 (7061): 1012–1017.

Callaway, E. (8 June 2016). 'Hobbit' relatives found after ten-year hunt’. Nature. 534 (7606): 164–165.