Theoretically, it would be possible that the distinctive morphology observed in LB1 is pathological and many different possibilities and arguments have been given over time. A very exhaustive and one of the first opponents of the new species theory, Jacob et al. (2006), have tried to make a case for the possible pathology of LB1 and many have followed after them (e.g., Martin et al., 2006b and Oxnard et al., 2010). One of their arguments is that LB1 is very asymmetrical, which may be an indication of growth anomalies. Aside from the fact that most crania, which have been buried, are asymmetrical due to taphonomy, it is actually quite normal for crania to be asymmetrical (Martin, 1914).

Another one of their arguments is that the neurocranium and the face of LB1 have archaic features which are all encountered routinely among Australomelanesians (Jacob et al., 2006). Their research (Jacob et al., 2006), for example, shows that 93.4% of the Rampassassa pygmies, which live on Flores close to Liang Bua cave, where H. floresiensis was found, have a neutral or negative chin. However, this was determined from photographs of living humans on which the bony structures cannot be seen, because they are covered with flesh. And, although each archaic feature may separately be encountered amongst modern populations, there is no record of one individual having all the reported archaic features.

They also note possible pathologies on the postcranial remains (Jacob et al., 2006). The right humerus shaft reportedly has very weakly marked muscle attachment sites and the deltoid tuberosity is nearly absent (Jacob et al., 2006). Morwood et al. (2005) have reported that the humeral torsion of 110° falls below the hominin range, and Jacob et al. (2006) conclude from this that the low humeral torsion of H. floresiensis is caused by a lack of muscle development. The tibial shafts are oval in cross section, suggesting to them (Jacob et al., 2006) a compromise between the need to support and move body mass and generally weak muscle development. CT scans of diaphyses have been reported to show thin cortical bone and very large marrow cavities (Jacob et al., 2006), used as further evidence that LB1 was abnormal. However, the arguments of Jacob et al. (2006) concerning the postcranial and cranial material from Liang Bua are based on several misconceptions about the postcranial skeleton, such as the degree of left/right asymmetry and the cortical bone thickness (Larson et al., 2007a).

A disease like microcephaly has been proposed as an explanation for the small brain size of LB1. Henneberg and Thorne (2004) have used measurements of a 4000-year-old Minoan microcephalic cranium and LB1 to make a statistical comparison of the two. They conclude that it cannot be rejected that LB1 and the Minoan microcephalic belonged to the same microcephalic population. Additionally, Weber et al. (2005) state that they find it possible that LB1 was a microcephalic, based on a comparison of the endocast of LB1 with the endocasts of 19 modern microcephalics. Martin et al., 2006a and Martin et al., 2006b compare the skull and endocast of H. floresiensis with microcephalics, and they also come to the conclusion that the size and shape of the cranium of H. floresiensis is very similar to one of these individuals. Therefore, they think it is likely that LB1 suffered from this disease as well. However, these assessments are mainly based on absolute dimensions, and ratios of dimensions, whereas the morphology of the endocast and the cranium are at least as important. For example, the maximum cranial breadth may be similar in size for LB1 and a microcephalic individual, but if they occur at a different position on the skull there is actually no similarity at all.

Laron Syndrome has also been proposed as a possible pathology for LB1 by Herschkovitz et al. (2007). The morphology of patients with Laron Syndrome has been compared with LB1 both quantitatively and qualitatively leading to the conclusion “that LB1 is but a local individual in a highly inbred, probably pygmy-like population (of Homo sapiens) in whom a mutation of the GH receptor had occurred”. However, Herschkovitz et al. (2007) write that “for comparative purposes we used photographs of “normal” sized skulls of LS individuals, not the microcephalic cases”. A comparison between LB1 and a microcephalic LS patient would have been more appropriate, since this is the affliction they propose. The current comparison casts doubt on whether both microcephaly and the other reported cranial features are actually encountered in the same individuals. In addition to this, they mention several other similarities between LB1 and patients with Laron Syndrome that do not match clinical literature. They write that patients with Laron Syndrome have abnormal limb proportions with disproportionately short legs (Herschkovitz et al., 2007). Rosenbloom et al. (1996) have, however, determined that adults with the syndrome have relatively short arms. Herschkovitz et al. (2007) also state that the maximum cranial breadth is in the mastoid region, whereas radiographs [Laron et al., 1996, p. 14, fig. 11] show that the maximum cranial breadth is at the parietals, which is comparable to healthy H. sapiens. And, there are also several features displayed by LB1, which are not mentioned by Herschkovitz et al. (2007), that also do not fit the diagnosis of Laron Syndrome. In LB1, the sutures are almost completely obliterated (Brown et al., 2004), whereas in Laron Syndrome patients, microcephaly is accompanied by delayed closure of the sutures (Kornreich et al., 2002 and Laron et al., 1996). A craniofacial feature of LB1 is that it has a sloping forehead (Brown et al., 2004), Laron Syndrome patients on the other hand have a very prominent forehead (Rosenbloom et al., 1996). Another feature of LB1 is that it has a relatively short first metatarsal and relatively large feet (Jungers et al., 2009b), whereas malformations of patients with Laron Syndrome include a short fourth metatarsal and relatively small hands and feet (Laron et al., 1996). The diagnosis of Laron Syndrome, thus, does not match the morphology of H. floresiensis.

Yet other researchers have proposed hypothyroid or myxoedematous cretinism as the disease from which LB1 suffered (Obendorf et al., 2008 and Oxnard et al., 2010). According to them, the cause for cretinism on Flores is the lack of iodine and they support this by stating that no marine food remains have been found in Liang Bua. Although this is true (van den Bergh et al., 2009), research on Pleistocene caves has shown that caves may only be used seasonally (e.g., de Lumley et al., 2004 and Marín Arroyo and González Morales, 2007) and a lack of marine fossils in Liang Bua does not necessarily mean a lack of marine resources in the diet of the occupying hominins; only further zooarchaeological research will be able to determine whether Liang Bua was used as a long duration habitat, a temporary seasonal habitat, or a hunting stopover. LB1 groups with ME cretins in the morphometric analysis of Obendorf et al. (2008). However, the three other fossil hominins included in their analysis also consistently group with the microcephalics even though they are not considered to be unhealthy. This clearly shows that their analysis is not suitable for distinguishing between healthy and unhealthy hominins. All their study indicates is that LB1 does not resemble a healthy H. sapiens. And there are several other problems with the cretinism hypothesis. The main argument for the hypothyroid cretinism hypothesis is that the pituitary fossa of LB1 is relatively large (Oxnard et al., 2010), which is not true. Although the area is damaged (Jungers et al., 2009c), the pituitary fossa is small if anything (Holloway in Wong, 2008). The authors (Oxnard et al., 2010) also cannot convincingly explain the small brain size of LB1. They make an educated guess that cretins derived from an endemic pygmy population might have brain sizes as small as that of LB1, but they fail to provide a single example. Oxnard et al. (2010), furthermore fail to take allometry into account in all their calculations. They use ratios, which eliminate size, but not allometry, which might influence the results greatly when the analyses deal with such a large range of sizes. One of the postcranial features which is supposed to support the cretin hypothesis is that the wrist bones of LB1 resemble those of cretins (Oxnard et al., 2010). However, their main argument that the trapezoid is bipartite is, again, false (Jungers et al., 2009c and Tocheri et al., 2007). Additionally Oxnard et al. (2010) write that Professor Alan Cooper has stated that “So far, however, only modern human DNA sequences have been found, consistent with handling of the specimens by multiple individuals prior to sampling”, they continue to write that it is likely that the modern human DNA sequences found, actually did belong to the fossil specimens and that this may be an indication that the fossils belong to pathological Homo sapiens instead of a new species. Another enquiry with Professor Alan Cooper (pers. comm.), however, resulted in the following statement: “The situation for any ancient specimen that has been handled by excavators, archaeologists, museum curators etc. is that it is expected to be contaminated with modern human DNA, which will generally be in relatively good condition (in comparison to the trace amounts of short, damaged fragments of ancient DNA). Hence, finding modern human DNA sequences in the hobbit material is totally expected, as no explicit steps were taken to limit contamination of the material. Furthermore, because we detected multiple modern human sequences in the samples, they are clearly due to contamination. This result says nothing about the hobbit being an extinct hominin, or otherwise.”

There are also several issues surrounding the (a)symmetry of the LB1 cranium. Most authors acknowledge some degree of asymmetry (Baab and McNulty, 2009, Brown et al., 2004, Eckhardt and Henneberg, 2010, Eckhardt et al., 2010, Jacob et al., 2006 and Kaifu et al., 2010), which is sometimes ascribed to antemortem developmental problems and sometimes to postmortem deformation. Falk et al. (2010) give an elaborate description of the conditions under which the cranium of LB1 was found, from which it is clear that postmortem deformation is most likely responsible for at least some of the asymmetries found in the cranium of LB1. Additionally, they point out severe weaknesses in the methodology for assessing the degree of asymmetry of Jacob et al. (2006).

So far, all proposed pathologies for LB1 have been refuted based on the skeletal evidence available (Argue et al., 2009, Brown and Maeda, 2009, Jungers et al., 2009c and Tocheri et al., 2007). It therefore seems highly unlikely that LB1 suffered from any currently known disease and one of the new species hypotheses should be accepted as being most likely.

The claim that the first H. erectus individuals to live on the island would have been a highly isolated inbred population and that they would have been adaptively constrained by a consequently narrow gene pool (Jacob et al., 2006), is not valid either. Stegodon migrated to Flores as well (van den Bergh, 1999 and van den Bergh et al., 2009); they too were a very small population in the beginning, but they adapted to their new environment. Many other animals that have migrated to islands started out as small populations, but they all adapted (van der Geer et al., 2010). There is no reason why hominins would not also be capable of adaptation. In fact, the small genetic variation expected in such a founder population increases the probability of evolution taking a different course than on the mainland.

A currently popular alternative is that H. floresiensis is a new species derived from some early australopith-like hominin. At present, there is no evidence for such a hominin in Asia, but its presence has been hypothesised by several authors on the basis of the Flores finds (Argue et al., 2009, Brown and Maeda, 2009 and Jungers et al., 2009c). Recently, an elaborate review of the current debate has lead to the conclusion that the most economical explanation for the morphology of H. floresiensis is that it is a descendent of an early australopith-like hominin (Aiello, 2010). H. floresiensis possesses both derived and primitive characters, and the presence of primitive characters mainly drives the hypothesis that H. floresiensis could not have been a descendant of H. erectus, but must have had an earlier, more primitive ancestor. The morphology of the wrist and foot bones also supports this view according to some (Jungers et al., 2009b and Tocheri et al., 2007).

Although there is not a complete H. floresiensis pelvic girdle available, the morphology of the parts which are available suggests great similarity with australopiths. This, together with other evidence, has contributed to the idea that H. floresiensis is a direct descendent of an australopith (or very early Homo) (Aiello, 2010). The pelvis, however, is not very suitable for phylogenetic investigations, because it has an important function in locomotion and childbirth, and as such is likely to be under high selection pressures.

Brown and Maeda (2009) believe that, considering the mandibular morphology, relative limb proportions, skeletal robusticity, wrist function and brain size, it is unlikely that H. floresiensis is an insular dwarf that is descended from H. erectus. LB1 has the limb proportions and endocranial volume of an African Pliocene Australopithecus (Morwood et al., 2005). And the two mandibles of H. floresiensis share symplesiomorphic characters with Australopithecus and early Homo (Brown and Maeda, 2009), which are at least uncommon in H. erectus. Reversals to a more ancestral state, however, are often encountered among island animals, as will be detailed below. The high robusticity of the lower limb bones of H. floresiensis is a functional adaptation, which is also explained below, and commonly seen in island animals. The wrist of H. floresiensis is more primitive than that of modern humans or Neanderthals as well. This eliminates the possibility that LB1 was a pathological modern human, but does not provide much insight into the ancestry of H. floresiensis, because no wrist bones of H. erectus have been found so far apart from one partial lunate. The morphology of the wrist of H. floresiensis can thus be consistent with both a H. erectus and an early hominin ancestry.

Cladistic analyses (Argue et al., 2009) suggest that H. floresiensis may have evolved in the Late Pliocene or Early Pleistocene, and was a descendent of an early species of Homo. Brown and Maeda (2009) argue that H. floresiensis arrived on Flores in the Middle Pleistocene and remained there relatively unaltered until its extinction. Insular artiodactyls exhibit sharp reversals to ancestral characteristics, which troubles cladistic analyses, as it groups islanders closer to less derived forms of the same taxon, rather than their own ancestors (van der Geer, 2005b). The same situation may be occurring for H. floresiensis, which is grouped with australopiths by cladistic analyses (Argue et al., 2009 and Argue et al., 2010), rather than H. erectus.

Whereas the morphology of LB1 may seem strange, or even impossible, when compared to other hominin fossils, it is not at all unexpected when compared to palaeontological finds from Flores and other islands (van Heteren and de Vos, 2007, Meijer et al., 2010 and Niven, 2007). The strangest case of endemism is arguably Hoplitomeryx. Its most striking characteristic is the presence of five horns, of which one is positioned between the eyes and the other four in pairs above the orbits. In spite of having horns, Hoplitomeryx is considered cervoid based on a cervid morphology of the molars, a distally closed metatarsal gully and the presence of a double lacrimal orifice on the rim of the orbit. Hoplitomeryx, thus, is a deer-like creature that developed horns instead of antlers. Compared to this animal, the ‘strange’ features displayed by LB1 seem relatively minor adaptations compared to a H. erectus base line. Appendix A gives a quantitative overview of changes undergone by other endemic animals. H. floresiensis is not any more different from H. erectus than the other endemics are from their ancestors or close relatives.

Initially, Brown et al. (2004) wrote that the most likely explanation for H. floresiensis “is long-term isolation, with subsequent endemic dwarfing, of an ancestral H. erectus population”. Although most researchers now accept an earlier ancestry, this is a very plausible idea. It has been shown that the environment on Flores during the Pleistocene was very suitable for insular dwarfs and giants to evolve and H. floresiensis may have been one (van den Bergh et al., 2009, Lucas, 2006 and Meijer et al., 2010).

It has been argued that it is improbable that all plesiomorphic traits of H. floresiensis were a consequence of island dwarfing, because some modern pygmies have greatly reduced body size without any evidence of evolutionary reversals (Jungers et al., 2009b). Modern pygmies, however, have large H. sapiens as ancestors, unlike H. floresiensis, and the different starting point would naturally result in a different end product. Furthermore, modern human pygmies belong to the same species as large H. sapiens, whereas H. floresiensis may have been a different species from H. erectus, although the possibility of a subspecies cannot be excluded at present. And because of seafaring, modern pygmies are much less genetically isolated than H. floresiensis was. Furthermore, island animals regularly show such reversals to a more primitive condition (van der Geer, 2005b and van der Geer et al., 2010). Phanourios, a pygmy hippopotamus, evolved a more lophodont dentition than its ancestor, which is a return to a more primitive condition (van der Geer et al., 2010). Hooijer (1954) describes Archidiskodon (=Elephas) celebensis as a dwarfed descendent of A. planifrons, but as being more primitive mostly based on the dentition. And the Japanese Hemimastodon (=Gomphotherium) annectens has molars which resemble those of the primitive genus Phiomia (van der Geer et al., 2010). Gymnesicolagus aff. gelaberti has a short diastema, which is a primitive character shared with continental ochotonids and insular leporids (Quintana and Agusti, 2007). Another good example is Hoplitomeryx. The orbital position of the horns of Hoplitomeryx is a return to a more primitive configuration (Leinders, 1983). The absence of ‘pogo-stick’ locomotion is characteristic of more primitive ruminants, but it is also found in Hoplitomeryx. This is characterised by a primitive astragalus, which can be found in both Hoplitomeryx and Myotragus (van der Geer, 1999, van der Geer, 2005b and van der Geer et al., 2010). And Hoplitomeryx has patellae which have returned to the primitive condition as well (van der Geer, 2005b). Primitive ruminants are very small compared to the derived forms. Island ruminants are also very small and have developed some features that are similar to the primitive ruminants. Analogously, australopiths are small compared to later members of the genus Homo, and H. floresiensis, similar in size to australopiths, has secondarily evolved primitive features as well. So, the presence of reversals to a more primitive condition should not be viewed as evidence against island dwarfing of H. floresiensis, but can be seen as supportive of this theory.

The external morphology of the LB1 cranium is significantly different from that of modern humans and similar to that of archaic Homo (Baab and McNulty, 2009), in particular H. erectus (Gordon et al., 2008 and Lyras et al., 2009). This similarity with H. erectus indicates that H. erectus is a very likely large-sized ancestor for H. floresiensis. Many of the features which are considered to be australopith-like can be explained as paedomorphic features relative to H. erectus. H. floresiensis has a high orbital, dental and brachial index, low humeral torsion, low tibial torsion and a high gonial angle, in addition to shortened lower limbs (van Heteren, 2008, van Heteren and de Vos, 2007 and van Heteren and Sankhyan, 2009). Of course, being adapted to the particular environmental challenges of Flores, H. floresiensis is not an exact copy of a H. erectus child. Neither are other island animals, which display paedomorphic features relative to their ancestors, exact copies of their ancestors’ young. In fact, descendents of the same ancestor evolved differently on different islands due to local circumstances or display radiation on the same island (van den Bergh, 1999, van der Geer, 2005a, van der Geer, 2008 and de Vos and van der Geer, 2002). The paedomorphic features found in H. floresiensis merely indicate the mode of dwarfing, but specific environmental adaptations have also occurred.

For example, contrary to the paedomorphic pattern described above, the head of H. floresiensis is relatively small, not relatively large, compared to the body. This is a specific energy-saving adaptation (Taylor and van Schaik, 2007) also reported for Myotragus (Köhler and Moyà-Solà, 2004). As the brain is a particularly expensive tissue (Aiello and Wheeler, 1995), any mutations reducing the cost of the brain without decreasing chances of survival, such as reorganisation in combination with reduced size (Falk et al., 2005), will be beneficial and carried into the next generation. The feet of H. floresiensis also do not follow the paedomorphic pattern, because they are relatively large, not relatively small. This is a specific locomotory island adaptation, which is explained further below.

H. floresiensis has been found to be megadont compared to H. sapiens and H. erectus (Brown et al., 2004 and Lucas, 2006). This is expected in the case of island dwarfing (e.g., elephants) (van den Bergh, 1999), as teeth tend to diminish in size more slowly than the rest of the skeleton, which has been shown in H. sapiens (Shea and Gomez, 1988).

That H. floresiensis is likely a descendant of H. erectus has led to the expectation that the earliest fossils on Flores, if they exist, should look more like H. erectus and be larger than the Liang Bua specimens, and in the case of gradual evolution, intermediate forms should be found; this would be proper evidence of the island dwarfing theory to be true (Lieberman, 2005). However, so far, no intermediate forms or large ancestors have been found on islands for any of the other island species apart from Myotragus (van der Geer et al., 2010). Myotragus, however, has been isolated on Mallorca for more than 5 million years. And the different Myotragus species are not examples of intermediate forms, each better adapted to the island than its ancestor. Rather, each species is adapted to the changing circumstances on Mallorca over time. And each species is a fully evolved island form in its own right. Insular dwarfing is a relatively rapid process and the odds of an individual from one of the first generations to be preserved in the fossil record are so small that, although not impossible, it is highly unlikely it will ever be uncovered. The conclusion of whether H. floresiensis is a result of insular dwarfing should, therefore, not depend on such evidence, but on palaeontological and palaeoanthropological comparisons with other insular taxa. Thus, as explained above, a very plausible explanation for the morphology of LB1 and associated specimens is that this population is derived from H. erectus and underwent insular dwarfing.