A fossil femur excavated by Eugène Dubois between 1891–1900 in the Lower/Middle Pleistocene bonebed of the Trinil site (Java, Indonesia) was recognised by us as that of a Hylobatidae. The specimen, Trinil 5703 of the Dubois Collection (Leiden, The Netherlands), has the same distinctive form of fossilization that is seen in many of the bonebed fossils from Trinil in the collection. Anatomical comparison of Trinil 5703 to a sample of carnivore and primate femora, supported by morphometric analyses, lead to the attribution of the fossil to gibbon. Trinil 5703 therefore provides the oldest insular record of this clade, one of the oldest known Hylobatidae fossils from Southeast Asia. Because living Hylobatidae only inhabit evergreen rain forests, the paleoenvironment within the river drainage in the greater Trinil area evidently included forests of this kind during the Lower/Middle Pleistocene as revealed here.
Citation: Ingicco T, de Vos J, Huffman OF (2014) The Oldest Gibbon Fossil (Hylobatidae) from Insular Southeast Asia: Evidence from Trinil, (East Java, Indonesia), Lower/Middle Pleistocene. PLoS ONE 9(6): e99531. https://doi.org/10.1371/journal.pone.0099531
Editor: Lorenzo Rook, University of Florence, Italy
Received: March 20, 2014; Accepted: May 5, 2014; Published: June 10, 2014
Copyright: © 2014 Ingicco et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. The raw coordinates are included within the Supporting Information files.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
The site of Trinil (East Java, Indonesia) is famous for being the locality that yielded the type specimen of Pithecanthropus erectus Dubois (, mostly cited as 1894), known as Homo erectus since Mayr . In this work, we report the discovery of a gibbon femur in the Trinil fossils assemblage, preserved in the Dubois Collection at the NCB Naturalis, Leiden (the Netherlands). The fossil was catalogued in the 1930s under the supervision of Dubois (Shipman 2001). The fossil was packed in a separate box labeled: “Semnopithecus - (Kleine soort) – l. femur” [Translated: Semnopithecus - (small species) – left femur]. Following examination of the specimen it became clear that it was attributable to a Hylobatidae, rather than to Semnopithecus.
Small apes are very scarce in the fossil record , making the reporting of even fragmentary new discoveries valuable. Moreover, as we show here, Trinil 5703 represents the oldest known presence of small apes in insular Southeast Asia [SEA], and provides further evidence of ever-wet forest habitat in an area of Homo erectus occupation. After describing the fossil femur anatomically, we compare it with SEA mammals of similar size in order to establish the taxonomic attribution of Trinil 5703.
Since its first description, two major amendments have been made to the Trinil fauna. First, Badoux  separated the fossils collected by von Koenigswald in the Punung area from the ones collected in Trinil, resulting in the creation of the distinct Punung fauna, considered to be Upper-Middle Pleistocene. Second, on the basis of excavation archives, de Vos and Sondaar  and de Vos et al.  reviewed the stratigraphic provenance of the fossils excavated at Trinil by Dubois in 1891–1900 and the Selenka expedition in 1906–1908, concluding that the bulk of the specimens came from a single bonebed named the Hauptknochenschicht. Following de Vos and Sondaar , the bonebed is referred to as Trinil H.K. in this paper. De Vos et al.  defined a biostratigraphic unit, the Trinil H.K. fauna, from the assemblage of vertebrate taxa, including Homo erectus, originating from this bonebed.
De Vos et al.  also reinterpreted the biostratigraphic position of the so-called Jetis fauna, which von Koenigswald  considered to be older than the Trinil fauna, but which according to de Vos et al.  is younger than their Trinil H.K. fauna. The Jetis fauna was further subsumed into a Kedung Brubus fauna, proposed and defined in the same paper. De Vos et al.  further defined Satir, Ci Saat, Ngandong, Punung and Wajak faunal association which have been highly influenced by eustatic fluctuations resulting in cyclic connections of Java to mainland Southeast Asia allowing migrations compared to endemic evolutions during insular conditions .
These new biostratigraphic zones (biozones) have been dated radiometrically – as follows: the Satir fauna at 1.5 myr, Ci Saat fauna at 1.2 myr, Trinil H.K. fauna at 1 myr, Kedung Brubus fauna at 0.8 myr, Ngandong fauna at 0.2–0.07 myr, Punung fauna at 0.12 myr, and Wajak fauna as Late Pleistocene to Holocene. The dates continue to be debated , , and especially so, the age of the Ngandong fauna –. No radiometric dating of Trinil bonebed has been reported, the age of 1 Myr for Trinil H.K. was mostly established on the basis of biostratigraphic correlation with assemblages collected at the Sangiran Dome, Java [e.g. 10]. Here, we use the coupled ESR/U-series and 40Ar/39Ar date of 0.8 myr that is obtained from the fossil-bearing beds at Ngebung 2, recently excavated by modern archaeological methods and referred to as Final Trinil H.K. fauna –. The biostratigraphical correlation of the Ngebung 2 fauna with Trinil H.K. is based on (a) the presence of Stegodon trigonocephalus and absence of Elephas hysudrindicus  which makes it older than the Kedung Brubus biozone where this latter species first appears ,  and (b) the presence of the cervid Axis lydekkeri ngebungensis that differs on subspecies level with Axis lydekkeri and has therefore been used to define this association as a Final Trinil H.K. fauna . The radiometric dates obtained at Ngebung 2 are consistent with the previous estimate age of ∼1 myr for the Trinil bonebed by von Koenigswald  and Leinders et al. . We therefore take 0.8 myr as the most reliable current age estimate for the Trinil H.K. and thus Trinil 5703.
Apart from the hominin remains, Trinil H.K. yielded the oldest undisputed primates in insular Southeast Asia , . These are the robust Colobinae Trachypithecus auratus robustus and the Cercopithecinae Macaca fascicularis (in Hooijer ). Von Koenigswald [30, p.63–64, 209] mentions the presence of the gibbon Hylobates and the siamang Symphalangus, and illustrates a molar of Hylobates moloch (formerly H. leuciscus) discovered in the Jetis biozone at Sangiran Dome. Hooijer  also mentions the presence of Hylobates in the Jetis and Trinil faunas (sensu von Koenigswald ). Later on, while describing the Punung fauna, Badoux  reallocated all the Hylobatidae formerly associated to the Trinil fauna (40 teeth) to this new biozone. Therefore, the oldest published Hylobatidae are from the Punung fauna of Java (128 Ka; , ). The two fossils collected by von Koenigswald  in Sangiran and assigned by him to the Jetis fauna, disappeared in the biostratigraphy of de Vos and Sondaar . Nevertheless, because they are reported coming from the Sangiran Dome, those two Hylobatidae should be older than the Punung fauna.
1.2. Stratigraphy of the trinil site
The Trinil H.K. bonebed lies near the base of the Kabuh Formation where it is in contact with the Pucangan Formation . The two formations dip gently southward in the vicinity of Trinil, as bedrock strata generally do along the southern edge of the Kendeng Hills (Figure1). The bonebed is exposed along the Solo River in the vicinity of Trinil due to river entrenchment of ∼10 m into the bedrock formations –. The Pucangan Formation exposed near the site is primarily tuffaceous conglomerate (sedimentary “breccia”) deposited by volcanic mudflows (lahars). The overlying Trinil H.K. is a volcaniclastic lens, deposited fluvially and filling a paleotopographic swale at the top of the Pucangan Formation; while the lens extends laterally for several hundred meters, it is only about a meter or two thick , –.
Lithologically, the Trinil H.K. is indurated gravelly sandstone. The stratum is very poorly sorted granulometrically and composed largely of fresh volcaniclastic materials. Besides carrying tens-of-thousands of well-preserved vertebrate fossils, including the remains of both terrestrial and aquatic animals, such as crocodile and fish , , , the Trinil H.K. also contained plant fossils, including broken tree trunks, and well-preserved fresh-water molluscs , , , .
Given the lithological and paleontological features of the Trinil H.K., it most likely accumulated from muddy flood waters that originated as a lahar on an active stratovolcano tens of kilometers to the south, descended the drainage to join a large river, passed through several biotopes, including tropical forest and grass-dominated areas, and inundated a swampy lowland , , , . The lowland was bordered a few kilometres north of Trinil by the marl bedrock of the Kendeng Hills, which contributed little clastic sediment to the bonebed relative to the more distant volcanic source terrain on the south  (see the paleogeographic map of Huffman and Zaim ).
Specimen 5703 is attributable to the Trinil H.K. bonebed, and thus this paleoenvironmental setting, for several reasons. First, Trinil 5703 has the same distinctive form of fossilization that is seen in many of the bonebed fossils of the Collection. The fossilization gives fossils like Trinil 5703 a brown, near glassy, appearance on surfaces where the cortical bone is freshly broken (Figure 2, 3c and 3d). Second, most of the Trinil fossils in the Dubois Collection originate from the Trinil H.K. The yellowish colour of the calcite crystals that have grown inside the femur medullar cavity further agrees with an origin of the fossil from the Trinil H.K. bonebed (Figure 3d).
A: Proximal view; B: Posterior view; C: Medial view; D: Lateral view; E: Anterior view.
A: radiograph in anterior view showing an opaque filling in the distal portion of the preserved medullar cavity; B: detail of the neck in posterior view, showing fluvial abrasion in weakly acid conditions; C: detail of the refitted break of the shaft that we attribute to breakage during excavation; D: detail of the distal break of the shaft, which is likely also excavation damage, and the calcitic filling inside the medullar cavity, as also seen on the radiograph.
Trinil 5703 is a partial left femur that preserves an incomplete proximal epiphysis and a portion of the shaft that is broken just below the nutrient foramen. Assuming that the specimen is from a fully-grown individual, as we establish below, Trinil 5703 therefore preserves 60% of its biomechanical length  (Figure2). The extremity of the lesser trochanter is missing, and the head and the greater trochanter are mostly lacking. Only a small portion of the head on the posterior aspect and part of the lateral aspect of the greater trochanter are preserved. The line of fusion of the head is preserved, however, and the fusion of the epiphysis and diaphysis are complete.
Considering that the proximal epiphysis fuses later than the distal epiphysis in apes, the length growth of the femur had ended. The developmental age is nevertheless uncertain, because epiphysation of the hip occurs early in the life of gibbons and other primates , , much before the eruption of the third molar, which is considered to be the boundary marker between youth and adulthood. Whether an adult or sub-adult by this criterium, the individual would have experienced no further growth along the epiphysed proximal end of its femur and had reached a practically full-grown stature.
Trinil 5703 exhibits damage attributable to both pre-burial events and excavation loss. Two fresh breaks occur on the shaft (Figure 3C and D). One broke the fossil in two. They were glued together decades ago and appear to refit perfectly. The edges along the break are sharp and not abraded, leading us to conclude that the fracture occurred during or after excavation. A chip of the shaft has been lost from along this break, and the gap reveals a brown, near glassy appearance of the fossilized cortical bone that from our own observations is typical of Trinil H.K. specimens in the Dubois Collection. The second fresh break occurs on the distal extremity of the femoral shaft, and is weakly abraded (Figure 3D), indicating the fracture was not strongly damaged during the fluvial transport it underwent. The distal break of the shaft reveals filling of the medullar cavity by well-crystalized, transluscent mineral that appears to be calcite (Figure 3D). The radiograph of the fossil shows that the calcite fills half of the preserved portion of the shaft (Figure 3A).
Other pre-burial taphonomic processes can be reconstructed from study of the surface of Trinil 5703. The cortical surface of the neck is damaged to the point that the trabeculae are prominently visible, except on the distal face. The femoral head and the greater and lesser trochanters are damaged to a similar degree (Figure 3B). The loss of bone surface of Trinil 5703 might well be due to wet and weakly acid conditions while the femur was at the ground surface after skeletonization but before transport because weathering rates are very fast in the tropics –.
Materials and Methods
The specimen we studied and describe in this paper is a a proximal femur labeled 5703 and curated in the Dubois Collection of the National Biodiversity Center Naturalis in Leiden, The Netherlands. Even though the fragmentary nature of Trinil 5703 precludes the acquisition of many classical anatomical measurements, some detailed morphological comparisons are possible, and although the fossil cannot be identified to a species level, its attribution to Hylobatidae is evident.
We first visually compared Trinil 5703 with femora of Southeast Asian carnivore and primate species of body sizes similar to gibbon, thus excluding very large taxa (Panthera and Pongo) and very small ones (Tarsius and Nycticebus) (Figure 4). The Trinil fauna contains three carnivores, Panthera tigris, Prionailurus bengalensis and Mececyon trinilensis, and two primates, Trachypithecus auratus robustus and Macaca fascicularis. Modern examples of Prionailurus bengalensis, Trachypithecus auratus robustus, and Macaca fascicularis were included in our sample. Mececyon is an extinct canid closely related to the extant Lycaon which is of similar shape with Cuon alpinus which as available to us. We therefore used the living species in our analyses.
3D models of the specimens obtained by surface scanning with a Nextengine Desktop 3D scanner.
We further compared our measurements on Trinil 5703 with those taken from the femora of extant Southeast Asian primates housed at the Muséum national d'Histoire naturelle (Paris, France) and the Naturalis Biodiversity Center (Leiden, The Netherlands), again excluding the very small and very large genera. This sample comprises 70 adults, 10 sub-adults and 6 juveniles when available in the collections for a total of 86 individuals from nine genera and 18 species (Table 1). For some of our analyses, we grouped the species into four natural clades, namely: small apes (Bunopithecus, Hylobates, Nomascus, Symphalangus), langurs (Presbytis, Trachypithecus) odd-nosed monkeys (Nasalis, Rhinopithecus) and macaques (Macaca fascicularis and Macaca nemestrina).
Eight linear measurements, taken with a digital calliper (Mitutoyo Digimatic Calliper with an accuracy of 0.01 mm), are used for metrical comparisons. The length of the neck was measured from the line of fusion of the head to the lesser trochanter (referred to as the LNW distance) and from the line of fusion of the head to the medial aspect of the greater trochanter (UNW). The other six measurements were the antero-posterior width of the neck (ANW), the minimum proximo-distal height of the neck (NH), the antero-posterior and medio-lateral shaft widths just below the lesser trochanter (ATW and MTW, respectively), and the maximum antero-posterior and corresponding medio-lateral shaft widths (ADW and MDW, respectively). We also measured the total length of the femur in our comparative sample in order to estimate the total length (TL) of Trinil 5703 by means of linear regression equations.
Our measurements of the specimens and the fossils (Table 2) are presented as bivariate plots. We estimate the original length of Trinil 5703 femur from our own calculated allometric regressions. This is because first, only  provided regressions for the femur of either small apes or monkeys, and second, when applied to our comparative material, none of those equations allowed us to retrieve or approximate the true femur lengths, so we developed our own equations. The proximo-distal femoral neck height (NH), because it is preserved in Trinil 5703, has been used to calculate allometric regressions of the femur length once log-transformed, as proposed by . Equations are presented for each three of the four natural clades established earlier, small apes, langurs and macaques (Table 3). Because Trinil 5703 is clearly out of the range both morphological and morphometrical of odd-nosed monkeys, we did not calculate any allometric regression. Femur length for Trinil 5703 was then estimated using each of the three equations from the logarithm of the neck length. We therefore obtained three different femur lengths for the fossil that we compared one by one with the respective femur length range of each primate clade.
Because the Javanese Pleistocene is characterized by numerous changes in mammal body size –, we applied the log-shape ratios method to our measurements, following methods described by , in order to remove size and describe shape only. We computed the log-shape ratios together, first in a principal component analysis (PCA), and second in a linear discriminant analysis (LDA) with cross validation on the four natural clades. The discriminant functions were calculated on the comparative material only, and the discriminant residuals of the fossil were calculated by post-multiplying the log-shape ratios of the fossil by the scores of the discriminant functions then obtained. By making Trinil 5703 a “hold-out” individual we obtained a post-classification for the fossil within one of the four pre-defined natural clades and percentages of Type I error of post-classification per clade (Table 4).
Hylobatidae are characterized by their very specialized locomotor behaviour, known as suspensory brachiation. Their anatomy reflects this specialization, especially in particularly slender and elongated limbs. Therefore, two characteristics may be regarded as important in the identification of femurs of small apes, which are the length of the femoral neck and the rectilinear shaft of the femur. Geometric morphometric methods have been successfully applied to long-bone curvature in primates by  so we applied it here. Because of the presence of the linea aspera on the posterior aspect of the femur that may complicate the record of the curvature, we decided to focus on the curvature of the anterior aspect. The curvature of the shaft was recorded directly perpendicular to the transverse plan of the femur in three dimensions from the freeware Landmarks (Institute for Data Analysis and Visualization, 2002) on 3D surface scans of the femurs acquired with a Nextengine surface scanner. We obtained a set of ten evenly-space horizontal landmarks on the mid-sagittal plane over the femoral shaft with the first and second points adjacent to the line just below the lesser trochanter and half of the shaft length respectively. This curvature was compared to our sample of primates. We first converted the three dimensional coordinate dataset into two dimensions through PCA on each individual one by one. The PCA scores for the two first dimensions of each individual have then been superimposed through baseline registration method as proposed by . A baseline was defined by the first and last landmarks of each curvature reset to the same fixed coordinate position, while the other landmarks followed this sum of rotation, translation and scaling. Newly obtained ‘x’ coordinates correspond to comparative landmarks along the femoral shaft, and the ‘y’ coordinates summarize the shape information. In this way, the ‘y’ coordinates are treated as variables at comparable positions (row number in the dataset) along the femoral shaft. Thus visual comparison of curvatures from one species to another one is straightforward. We also quantified the distances between Trinil femur shaft curvature and other species by calculating the mean distance between the Bookstein coordinates in baseline unit.
3.1 Identification of Trinil 5703 as a primate
Morphological comparison of Trinil 5703 to femora of Southeast Asia carnivores and primates of similar size indicates clearly that the fossil is a primate (Figure 4). Trinil 5703 is not a carnivore because (a) it is longer than Pardoxurus and Prionailurus; (b) the shaft and neck are narrower than those of Arctictis and Cuon; (c) the neck is much longer than in all the carnivores in Figure 4; and (d) the lesser trochanter is oriented more posteriorly than any of the carnivores. Trinil 5703 does have a salient gluteal tuberosity that is also seen in Paradoxurus, Prionailurus and Arctictis.
Trinil 5703 is generally similar to the primates shown in Figure 4: (a) the width of the shaft, both ATW over MTW and ADW over MDW, is within the range of the small apes, langurs and macaques (Table 2); (b) the posterior orientation of the lesser trochanter is within the range of the orientations seen in the small ape specimens; and (c) the estimated original length of Trinil 5703 is within the range of small apes and the small macaque M. fascicularis.
Equations have been calculated from linear regressions of the log-transformed TL over the log-transformed NH for each of the following clades: small apes, langurs and macaques (Figure 5). The equations are given in Table 3. They allowed us to estimate the femur length of Trinil 5703. The regression based on the macaque model produces the lowest estimate (133+/−13 mm), while the largest estimate (178+/−16 mm) is obtained by the small ape model. The regression based on the langur model gives a middle estimate of 171+/−39 mm. This latter estimate is largely out of the range of femur length for the langurs, Trinil 5703 being too small to be either attributed to Trachypithecus or Presbytis (Figure 5). The femur length estimate obtained by the macaque model clusters Trinil 5703 with the small size M. fascicularis. Nevertheless, the preserved portion of Trinil 5703, which corresponds to no more than half of its total length, as we established earlier in the preservation section, is nearly the size of the complete M. fascicularis femur (Figure 4). Therefore, the Trinil 5703 femur was unlikely to have had a length comparable to M. fascicularis, as predicted by the macaque model. The small ape regression model appears to provide the best estimate for the original length of Trinil 5703 femora (178+/−16 mm).
3.2 Identification of Trinil 5703 as Hylobatidae
Ellipses are 95% confidence interval for each natural clades: small apes, odd-nosed, macaques and langurs. Filled grey ellipses group the juvenile and sub-adult individuals of small apes and odd-nosed monkeys.
Although the femoral head of Trinil 5703 is mostly lacking, the small portion preserved on the posterior aspect likely suggests a globular shape comparable to what is observe in small apes. The head typically extends on the posterior aspect of the neck in Cercopithecidae such as macaques, langurs and odd-nosed monkeys (Figure3B). This characteristic is confirmed in Trinil 5703 with the small portion of the head preserved which clearly overhangs the neck (Figure 2B and 2E). The globular shape allows a high degree of abduction of the hind limb, which is essential in suspensory locomotion.
Trinil 5703 shows an elongated, thin and antero-posteriorly flattened femoral neck, similar to that found in all small apes (Figure7). The proportions of the neck in the fossil are similar to young and sub-adult small apes. Although the neck-shaft angle cannot be measured, the angle formed is obtuse. The anteversion of the neck cannot be estimated here without the distal portion of the femur. There is no tuberculum on the posterior aspect of the neck (Figure2B and 3B). Such a tuberculum is always present in catharrines and occurs in most hominoid taxa, but is usually absent in small apes , , .
The trochanteric fossa of Trinil 5703 femur is small and very rounded (Figure2B). Bacon  distinguishes three morphological types of primate trochanteric fossa. A trochanteric fossa located in a slot is characteristic of Strepsirhini, while a large and proximo-distally stretched oval fossa characterizes Cercopithecoidea, such as macaques, langurs and odd-nosed monkeys. A rounded fossa is only recorded in Hominoidea and Atelinae. Bacon  interprets the rounded shape of the trochanteric fossa of the kind that we observed in Trinil 5703 as the possibility for a greater mobility of the ligamentum teres during brachiating locomotion.
The lesser trochanter is on the posterior face of the shaft of Trinil 5703. The position of the lesser trochanter appears to vary within primate populations . This intraspecific variability also occurs in small apes. A mesial position of the lesser trochanter is reported for Hylobatidae , . In counterpart, Piganiol and Olivier  indicate that the lesser trochanter is oriented posteriorly in small apes. From our own observations, the position of the lesser trochanter seems to change during the life of the animal from posterior position among young and sub-adults and a mesial position in adults. In Figure2, the lesser trochanter of the adult Hylobates is oriented medially but of sub-adult Nomascus, posteriorly. Nevertheless, this age-related position of the lesser trochanter has to be confirmed by further studies with larger samples of young individuals, which is beyond the aim of this study. To summarize, the position of the lesser trochanter in Trinil 5703 is within the range of small apes, and may indicate a sub-adult developmental stage.
The shape of Trinil 5703 is typical for small apes. The femora in all small apes are straight in anterior view (this feature also is common in langurs). Moreover, the small-ape femoral shafts are straight in lateral view, as Dubois  recognized. Trinil 5703 has the straight anterior and lateral profile of small apes and langurs. We quantified the curvature in primate femurs and in Trinil 5703 through baseline registration (Figure 6). Trinil 5703 femur curvature is close to the genus Presbytis and Hylobates with a mean distance in baseline unit respectively of 0.010 and 0.014. The distance to the other primates is much greater, followed by Trachypithecus (0.037) and Nomascus (0.064), and finally the macaques M. fascicularis (0.073) and M. nemestrina (0.126). Moreover, there are features of Trinil 5703 that are seen in gibbons but not in other primates such as the shaft cross section morphology. While the cross section in female gibbons is circular, males have a spiral line, which resulting in an asymmetrical cross-section of the shaft. Trinil 5703 has a spiral line, although weakly developed.
Not all morphological details of Trinil 5703 fit those of modern small apes. In small apes of both sexes, the linea aspera is salient at the contact with the greater trochanter. This salience is not seen in Trinil 5703, neither at the contact of the greater trochanter nor more distally along the shaft. The salience of the linea aspera in Trinil 5703 is intermediate between those of small apes, macaques and langurs, falling within the 95% confidence intervals of the three groups. The weak development of the linea aspera in the femoral diaphysis of small apes is confirmed by shaft thickness (Figure 8). In keeping with these results, the index of platymeria for Trinil 5703 is 89.4 when measured below the lesser trochanter. A low platymeria index was found in gibbons by Manouvrier  (Table 2). Because Trinil 5703 is only partially preserved, a robustness index cannot be calculated for the specimen. Nevertheless, the shaft is thin and seems very elongated when considering its estimated length through both the macaque and small ape regression model.
Based upon the general morphology, shape and size of Trinil 5703, the specimen is a Hylobatidae. Nevertheless, some measurements on the fossil have values within the ranges of small apes, langurs and macaques. We employed log-shape ratios to clarify this issue. Trinil 5703 clusters with small apes using PC1 and PC2 in combination, which accounts for 64.57% of the total variance (Figure 9). A biplot of the PCA permits us to explain the clustering of the fossil by its very long and thin femoral neck (LNW and ANW variables), and by the discrete linea aspera on the shaft (ADW and ATW variables). In the LDA (Figure 10), the small apes are clearly separated from the three other natural clades — langurs, macaques and odd-nosed monkeys — by the first linear discriminant function (accounting for 88.82% of the total variance of the sample). The confusion matrix of the LDA (Table 4) shows a few misclassifications for the macaques clade, which plot with small apes. Indeed, only 14% (accounting for three individuals) of macaques are post-classified within small apes, while no langurs or odd-nosed monkeys have been misclassified as a small ape. These results increase the confidence that Trinil 5703 is a small ape.
The femoral neck-length and -height are the key parameters in establishing the taxonomic identity of Trinil 5703. These make an attribution to a primate the only reasonable one. These features, together with the general morphology of femur Trinil 5703 assessed by the metrical data, enable us to conclude that this specimen belongs to an adult or a subadult Hylobatidae.
Dubois' field crew collected Trinil 5703 while excavating the Trinil H.K. bonebed and catalogued the specimen as originating from the site. The color and state of preservation of Trinil 5703 are comparable to those exhibited by other fossils discovered in Trinil H.K., leaving little doubt that the specimen originated from the Homo erectus type stratum. Dubois [66, p.157] noticed that fossils from Trinil do not show prominent effects of fluvial transport. This is consistent with our own observations that Trinil 5703 is weakly abraded. From the geological evidence given above, as well as paleontological criteria , we conclude that the Trinil H.K. was deposited by a large river over a short period of time, with much of skeletal material in the bonebed apparently having come from deaths that had occurred shortly before deposition.
Although the femur Trinil 5703 had reached a practically full-grown stature, several features fit a subadult male: posteriorly positioned lesser trochanter, and weakly developed spiral line. Also the femur estimated original length is in the range of a present-day juvenile or sub-adult Hylobatidae, based upon metric comparisons of the fossil to modern material. Such a hypothesis could explain the presence of this isolated bone of a gibbon in Trinil assemblage, considering that high subadult mortality is not unusual for primates e.g. . Alternatively, the Javanese Pleistocene is characterized by numerous changes in mammal body size, including dwarfism in large species and gigantism in small species (for an overview of the history, see ). Indications for this are for instance Duboisia santeng , Mececyon trinilensis  and Stegodon trigonocephalus . Also, Hooijer  described a robust form of the small primate Trachypithecus auratus from the Trinil H.K. biozone. It is then possible that this gibbon evolved in isolation and therefore shows an unusual small size.
Gibbons are rare in the fossil record, and are found mostly as dentognathic remains . The oldest examples of Hylobatidae in the fossil record are of Pliocene age from Yunnan Province, China. Only two fossils are known from the Lower Pleistocene also from China. No Hylobatidae fossil has been discribed from SEA before the Late Pleistocene Punung fauna of Java , but the primate specimens from the Punung fauna are teeth, only one of which is identified as Hyolbates moloch . Ansyori  also identified an upper left second incisor of a Symphalangus sp. in the lower stratigraphical layers of Song Terus cave in the vicinity of Punung, and included the species in the Punung biozone. Trinil 5703, as a gibbon long-bone very likely attributable to the Lower/Middle Pleistocene Trinil H.K. fauna of Java, becomes the oldest known insular Hylobatidae fossil (on the insularity of the Trinil fauna, see ), one of the oldest representatives of the family in all SEA. Trinil 5703 further adds key skeletal material to the fossil record of a family that is principally known from dental specimens. The dating of Trinil 5703 supports the inference derived from molecular studies that gibbons first arrived in Sundaland near the Pliocene-Pleistocene boundary . Lower/Middle Pleistocene fossils of Hylobatidae, which would be contemporaneous with Trinil 5703, have been recorded in China, Thailand and Vietnam .
Our identification of Trinil 5703 also provides key information on the paleoenvironment in the paleo-drainage area of Trinil. The Trinil H.K. represents remains of a once living community . Although the predominance of herbivores in the assemblage indicates an open paleoenvironment, the lithological and paleontological features of the Trinil H.K. are better interpreted as representative of a broader set of paleoenvironments, including rain forest.
Paleoenvironmental studies have been conducted on the Javanese Lower/Middle Pleistocene deposits at Trinil and in the related lower Kabuh/Bapang sedimentary beds of Sangiran dome, 80 km to the west of Trinil. Study of the mammalian fauna, mainly from Trinil H.K., and the results of paleosol analyses at Sangiran indicate a savannah-like environment during this period –. Study of the ichthyofauna from Trinil H.K. gave similar results but also suggests the proximity of lakes and swamp forests. Based upon their study of the Trinil ichthyofauna, Joordens et al.  suggest that the grasslands were regularly inundated, and referred to the Trinil paleoenvironment as an “hydromorphic savanna”. Closed environment in the vicinity of Trinil during the Lower/Middle Pleistocene is also supported by the presence of the forest dweller bovid Duboisia santeng .
Palynological results from lower Kabuh/Bapang beds of Sangiran Dome testify to a mosaic of environments dominated by open vegetation (Poaceae) with ferns and rain forest-refuges [76, p. 131]. Palynological results from Trinil, although scarce, also testify to an open environment with the presence of a seasonal forest . Fossil imprints of leaves and seeds of Ficus and Indigofera were collected from the Trinil discovery sequence by the Selenka expedition . The fossils are considered to be typical of lowland rain forest .
The presence of a gibbon in Trinil H.K indicates that the closed forests were present near Trinil at the time of deposition. Small apes are mainly characterized by suspensory locomotion . Their anatomy shows profound specialization for arboreal living, such as elongated limbs, including the hindlimb. In Trinil 5703, the globular femoral head, thin- and elongated-femoral neck, small-, deep- and rounded-trochanteric fossa, thin femoral shaft, and associated small platymeria index indicate that the species represented had suspensory locomotion comparable to that of present-day small apes. The diet of this clade largely consists of fruits and leaves. Acquiring these foods constrain small apes to high forest canopies, where dense tree branching is absent, and their preferred food resources are available throughout the year , . Gibbons themselves are almost exclusively found in tropical lowland/upland evergreen rain forest. Mangrove, swamp and secondary forests are not appropriate for gibbons .
Evidence has been scarce in support of rain forest vegetation during Trinil H.K. , . The fossil Trinil 5703 thus points to the presence of evergreen rain forest during the Lower/Middle Pleistocene of eastern Java. The rain forest in which the Trinil 5703 individual evidently lived was part of the drainage basin of the river that carried the femur to the site. The habitat might have been in distant volcanic highlands to the south, possibly higher than 1600 meters, which is the present-day upper altitudinal limit for gibbons in Java .
The fossils that Eugène Dubois excavated from the Trinil site have drawn close scrutiny and often evoked controversy, because of their great importance to paleoanthropology. Some scholars have expressed concern about poor stratigraphic control during excavation, specifically that fossils from two geological formations might have been mixed [e.g., 85]. However, those who have looked closest at the provenience from archival and field evidence, including ourselves, conclude that the vast majority of the fossils Dubois collected at the site came from a single, one- to two-meter-thick bonebed, the Trinil H.K. , , , .
With our recognition of Trinil 5703 as the first Hylobatidae fossil from Trinil, we provide strong evidence that during the Lower/Middle Pleistocene the river basin upstream of the site contained a population of small-sized gibbons, who inhabited rain-forest, in addition to the Homo erectus and other mammals which have been known from the Trinil H.K. fauna for over a century.
Linear measurements used in this study.
We are grateful to Zora Gabsi for radiographs of the fossil. We are indebted to Antoine Balzeau for his comments and Martin Pickford for proofreading of an earlier version of this manuscript. We also would like to acknowledge the two editors (Lorenzo Rook and an anonymous one) and the three reviewers (Gert van den Bergh, Alexandra van der Geer and an anonymous one) for improving this manuscript.
Conceived and designed the experiments: TI JdV OFH. Analyzed the data: TI JdV OFH. Contributed to the writing of the manuscript: TI JdV OFH.
- 1. Dubois E (1894) Pithecanthropus erectus, eine menschenähnliche Uebergangsform aus Java Landesdrückerei, Batavia, 39p.
- 2. Mayr E (1944) On the terminology of vertical subspecies and species. National Research Council Bulletin 2: 11–16.
- 3. Jablonski NG, Chaplin G (2009) The fossil record of gibbons. In S Lappan, DJ Whittaker, The Gibbons Developments in Primatology, New York, 111–130.
- 4. Badoux DM (1959) Fossil mammals from two fissure deposits at Punung (Java) Drkkerij en uiteversmij v/h Kemink en zoon NV, Utrecht, 153p.
- 5. de Vos J, Sondaar PY (1982) The importance of the “Dubois Collection”. reconsidered. Mod Quat Res in South-east Asia 7: 35–63.
- 6. de Vos J, Sartono S, Hardja-Sasmita S, Sondaar PY (1982) The fauna from Trinil, type locality of Homo erectus: a reinterpretation. Geologie en Mijnbouw 61: 207–211.
- 7. von Koenigswald GHR (1934) Zur stratigraphie des javanischen Pleistocän. De ingenieur in Nederlandsch Indië 1(4): 185–201.
- 8. van den Bergh G, de Vos J, Sondaar P-Y (2001) The Late Quaternary palaeogeography of mammal evolution in the Indonesian Archipelago. Palaeogeogr Palaeoclim Palaeoeco 171: 385–408.
- 9. Watanabe N, Kadar D (1985) Quaternary geology of the hominid fossil bearing formations in Java; report of the Indonesian-Japan joint research project CTA-41, 1976–1979 Geological researcg and development centre (Special publication), Bandung, 379p.
- 10. Leinders JJM, Aziz F, Sondaar PY, de Vos J (1985) The age of the hominid-bearing deposits of Java: state of the art. Geologie en Mijnbouw 64: 167–173.
- 11. Westaway K, Morwood M, Roberts R, Rokus D, Zhao JX, et al. (2007) Age and biostratigraphic significance of the Punung rainforest fauna, East Java, Indonesia, and its implications for Pongo and Homo,. J Hum Evol 53(6): 709–717.
- 12. Storm P, Wood R, Stringer C, Bartsiokas A, de Vos J, et al. (2013) U-series and radiocarbon analyses of human and faunal remains from Wajak, Indonesia,. J Hum Evol 64: 356–365.
- 13. Sémah F, Saleki H, Falguères C, Féraud G, Djubiantono T (2000) Did Early Man reach Java during the Late Pliocene? Journal of Archaeological Science 27: 763–769.
- 14. Larick R, Ciochon RL, Zaim Y, Sudijono Suminto , et al. (2000) Lithostratigraphic context for Kln-199305-SNJ, a fossil Colobine maxilla from Jokotingkir, Sangiran dome. Int J Primatol 21(4): 731–759.
- 15. Bartstra GJ, Soegondho S, vand der Wijk A (1988) Ngandong man: age and artifacts. J Hum Evol 17: 325–337.
- 16. van der Plicht J, van der Wijk A, Bartstra G (1989) Uranium and thorium in fossil bones: activity ratios and dating. Applieds in Geochemistry 4(3): 339–344.
- 17. Swisher III CC, Rink WJ, Anton SC, Schwarcs HP, et al. (1996) Latest Homo erectus of Java: potential contemporaneity with Homo spiens in Southeast Asia. Science 274(5294): 1870–1874.
- 18. Grün R, Thorne A (1997) Dating the Ngandong humans. Science 276(5318): 1575–1576.
- 19. Yokoyama Y, Falguères C, Sémah F, Jacob T, Grün R (2008) Gamma-ray spectrometric dating of late Homo erectus skulls from Ngandong and Sambungmacan, Central Java, Indonesia. J Hum Evol 55(2): 274–277.
- 20. Indriati E, Swisher III CC, Lepre C, Quinn RL, Suriyanto RA, et al. (2011) The age of the 20 meter Solo river terrace, Java, Indonesia and the survival of Homo erectus in Asia, PloS ONE 6 (6).
- 21. Saleki H (1997) Apport d'une intercomparaison des methods nucléaires (230Th/234U, ESR et 40Ar/39Ar) à la datation des couches fossilifères pléistocènes dans le dôme de Sangiran (Java, Indonésie) Unpublished PhD thesis Muséum national d'Histoire naturelle, Paris.
- 22. Moigne AM, Due Awe R, Sémah F, Sémah AM (2004) The cervids from the Ngebung site (Kabuh series, Sangiran Dome, Central Java) and their biostratigraphical significance. Modern Quaternary Research in Southeast Asia 18: 31–44.
- 23. Sémah F, Falguères C, Sémah AM, Purnomo A, Djubiantono T, et al. (2011) Multi-proxy chronological approach of Homo erectus bearing sites in Indonesia: the Ngebung example 1st Congress of European Society for the Study of Human Evolution (ESHE), Leipzig.
- 24. Bouteaux A, Moigne A-M (2010) New taphonomical approaches: the Javanese Pleistocene open-air sites (Sangiran, central Java). Quat Int 223–224: 220–225.
- 25. van den Bergh G (1999) The Late Neogene elephantoid-bearing faunas of Indonesia. Scripta Geol 117: 1–388.
- 26. von Koenigswald GHR (1968) Observations of Pithecanthropus mandibles from Sangiran, Central Java Koninklijke Akademie van Wetenschappen, Amsterdam, Series B. 71: 1–9.
- 27. Jablonski NG, Tyler D (1999) Trachypithecus auratus sangiranensis, a new fossil monkey from Sangiran, Central Java, Indonesia. Int J Primatol 20(3): 349–352.
- 28. Larick R, Ciochon R L, Zaim Y (2001) Early Pleistocene 40Ar/39Ar ages for Bapang formation hominins, Central Jawa, Indonesia. Proc Natl Acad Sci USA 98(9): 4866–4871.
- 29. Hooijer DA (1962) Quaternary Langurs and macaques from the Malay Archipelago. Zoologische Verhandelingen 55: 1–64.
- 30. von Koenigswald GHR (1940) Neue Pithecanthropus Funde 1936–1938. Wetenschappelijke 28: 1–233.
- 31. Hooijer DA (1960) Quaternary gibbons from the Malay Archipelago. Zoologische Verhandelingen 46: 1–44.
- 32. Storm P, Aziz F, de Vos J, Kosasih D, Baskoro S, et al. (2005) Late Pleistocene Homo sapiens in a tropical rainforest fauna in East Java. J Hum Evol 49(4): 536–545.
- 33. Duyfjes J (1936) Zur geologic und stratigraphie des Kendenggebietes zwischen Trinil und Soerabaja (Java) De Ingenieur in Nederlandsch-Indië, IV Mijbouw & Geologic, De Mijningenieur Jaargang III. (8): 136–149.
- 34. Dubois E (1896) Pithecanthropus erectus, eine Stammform des Menschen Anat Anz XII (Band, Nr. 1): 1–14.
- 35. Selenka L, Blanckenhorn M (1911) Die Pithecanthropus-Schichten auf Java Geologische und paläontologische Ergebnisse der Trinil-Expedition (1907 und 1908) Verlag Von Wilhelm Engelmann, Leipzig.
- 36. Soeradi T, Shibasahi T, Kadar D, SudijonoIthara M, Kumai H, et al. (1985) Geology and stratigraphy of the Trinil area. In: Watanabe, N, Kadar, D (Eds), Quaternary Geology of the Hominid Fossil Bearing Formations in Java, No 4 Geological Research and Development Centre Special Publication, 49–53.
- 37. Carthaus E (1911) Zur Geologic von Java, insbesondere des Ausgrabungsgebietes. In: Selenka, ML, Blanckenhorn, M (Eds), Die Pithecanthropusschichten auf Java Geologische und Palaontologische Ergebnisse de Trinil-Expedition (1907 und 1908) Verlag von Wilhelm Engelmann, Leipzig, 1–33.
- 38. de Vos J, Aziz F (1989) The excavations by Dubois (1891–1900), Selenka (1906–1908), and the Geological Survey by the Indonesian-Japanese Team (1976–1977) at Trinil Java, Indonesia). J Anthrop Soc Nippon 97(3): 407–420.
- 39. Storm P (2012) A carnivorous niche for Java Man? A preliminary consideration of the abundance of fossils in Middle Pleistocene Java, CR Palevol 11(1–2): 191–202.
- 40. Dubois E (1891) Paleontologische Onderzoekingen op Java Verslage van het Mijnwezen 3e Kwartaal 1891.
- 41. van Benthem Jutting T (1937) Non marine mollusca from fossil horizons in Java, with special reference to the Trinil fauna. Zoologische Mededelingen 20: 83–180.
- 42. Joordens J, Wesselingh F, de Vos J, Vonhof H, Kroon D (2009) Relevance of acquatic environments for hominins: a case study from Trinil (Java, Indonesia). J Hum Evol 57(6): 656–671.
- 43. Dubois E (1892) Paleontologische Onderzoekingen op Java Verslage van het Mijnwezen 2e Kwartaal 1892.
- 44. Huffman OF, de Vos J, Balzeau A, Berkhout AW, Voight B (2010) Mass death and lahars in the taphonomy of the Ngandong Homo erectus bonebed, and volcanism in the hominin record of eastern Java Abstracts of the PaleoAnthropology Society 2010 Meetings PaleoAnthropology. 2010: A14.
- 45. Huffman OF (2001) Plio-Pleistocene environmental variety in eastern Java and early Homo erectus paleoecology - a geological perspective. In HT Simanjuntak, B Prasetyo and R Handini (Eds), Sangiran: Man, Culture, and Environment in Pleistocene Times, Proceedings of the International Colloquium on Sangiran Solo - Indonesia, 21st-24th September 1998, Jakarta, 231–256.
- 46. Huffman OF, Zaim Y (2003) Mojokerto Delta, East Jawa: Paleoenvironment of Homo modjokertensis - First Results Journal of Mineral Technology, vol.10(2) The Faculty of Earth Sciences and Mineral Technology, Institute Technology, Bandung.
- 47. Ruff CB (2002) Long bone articular and diaphyseal structure in Old World Monkeys and Apes I: Locomotor effects AJPA. 119: 305–342.
- 48. Todd TW (1930) Comparative youth: The physical aspect. Child Development 1(2): 79–89.
- 49. Schultz AH (1944) Age changes and variability in gibbons A morphological study on a population sample of a man-like ape. Am J Phys Anthrop 2(1): 1–129.
- 50. Behrensmeyer AK (1978) Taphonomic and ecological information from bone weathering. Paleobiology 4(2): 150–162.
- 51. Andrews P, Armour-Chelu M (1998) Taphonomic observations on a surface bone assemblage in a temperate environment. Bull Soc Géol France 169(3): 433–442.
- 52. Kölher M, Alba DM, Moyà-Solà S, MacLatchy L (2002) Taxonomic affinities of the Eppelsheim femur. Am J Phys Anthropol 119: 297–304.
- 53. Lyras GA, Van der Geer A, Rook L (2010) Body size of insular carnivores: evidence from the fossil record. J Biogeogr 37: 1007–1021.
- 54. Rozzi R, Winkler DE, de Vos J, Schiltz E, Palombo MR (2013) The enigmatic bovid Duboisia santeng (Dubois, 1891) from the Early-Middle Pleistocene of Java: a multiproxy approach to its palaeoecology. Palaeogeogr, Palaeoclim, Palaeoeco 377: 73–85.
- 55. Lomolino MV, van der Geer AAE, Lyras GA, Palombo MR, Sax D, et al. (2013) Of mice and mammoths: generality and antiquity of the island rule. J Biogeogr 40: 1427–1439.
- 56. Darroch JN, Mosimann JE (1985) Canonical and principal components of shape. Biometrika 72: 241–252.
- 57. Richmond BG, Whalen M (2001) Forelimb function, bone curvature and phylogeny of Syvapithecus, InL De Bonis, GD KoufosandP Andrews (eds), Phylogeny of the Neogene Hominoid Primates of Eurasia, Cambridge University Press, Cambridge, 338–360.
- 58. Bookstein FL (1991) Morphometric Tools for Landmark Data: Geometry and Biology. Cambridge: Cambridge University Press.
- 59. R Development Core Team (2009) R: A language and environment for statistical computing R Foundation for Statistical Computing, Vienna, Austria ISBN 3-900051-07-0, URL http://wwwR-projectorg.
- 60. Claude J (2008) Morphometrics with R. Springer, New York, 318p.
- 61. Bacon A-M (2001) La locomotion des primates du Miocène d'Afrique et d'Europe, analyse fonctionnelle des os longs du membre pelvien et systématique CNRS, Cahiers de paléoanthropologie, Paris, 136p.
- 62. Scherf H (2007) Locomotion-related femoral trabecular architectures in Primates (Paidopithex rhenanus, Pliopithecus vindobonensis) PhD Thesis, Universität Darmstadt, 203p.
- 63. Aiello L, Dean C (1990) An Introduction to Human Evolutionary Anatomy. Academic Press, London, 596p.
- 64. Piganiol G, Olivier G (1959) L'architecture du fémur des primates. Comptes rendus de l'Association des Anatomistes 104: 660–667.
- 65. Dubois E (1924) Over de voornaamste onderscheidende eigenschappen van het femur van. Pithecanthropus erectus Verslag van de gewone vergadering der Afdeeling Natuurkunde 35(3): 443–455.
- 66. Manouvrier L (1889) La platymérie Comptes rendus du congrès international d'anthropologie et d'archéologie préhistoriques, Xè session, Paris, 363–381.
- 67. Dubois E (1895) Sur le Pithecanthropus erect du pliocène de Java Résumé, Bulletin de la Société Belge de Géologie, d'Aléontologie et d'Hydrologie Procés-verbaux. 9: 151–160.
- 68. van der Geer A, Dermitzakis MD (2008) Dental eruption sequence in the Pliocene Panionin Paradolichopithecus arvernensis (Mammalia: Primates) from Greece Journal of Vertebrate Palaeontology. 28(4): 1238–1244.
- 69. van der Geer A, Lyras G, de Vos J, Dermitzakis M (2010) Evolution of island mammals, adaptation and extinction of placental mammals on islands, Wiley-Blackwell, Oxford, 500p.
- 70. Ansyori M (2010) Fauna from the oldest occupation in Song Terus cave, eastern Java, Indonesia Master thesis, Muséum national d'Histoire naturelle, Paris, 72p.
- 71. Chatterjee JJ (2006) Phylogeny and Biogeography of gibbons: A dispersal-vicariance analysis. Int J Primatol 27(3): 699–712.
- 72. de Vos J (1989) The environment of Homo erectus from Trinil HK- in Hominidae, Proceedings of the 2nd International Congress of Human Paleontology, Turin, September 28-October 3, 1987, Editorial Jaca Book: 225–229.
- 73. Bouteaux A (2005) Paléontologie, paléoécologie et taphonomie des mammifères du Pléistocène et du début de l'Holocène en Asie du Sud-est: un nouvel apport à la compréhension des comportements humains PhD Thesis, Muséum national d'Histoire naturelle, Paris, 371p.
- 74. Bettis E III, Milius A, Carpenter S, Larick R, Zaim Y, et al. (2009) Way out of Africa: Early Pleistocene Paleoenvironments inhabited by Homo erectus in Sangiran, Java. J Hum Evol 56: 11–24.
- 75. Brasseur B (2009) Dynamique et histoire des dépôts du Pléistocène inférieur et moyen ancien du dôme de Sangiran (Java Central, Indonésie) Caractérisation des surfaces d'occupation à Homo erectus PhD Thesis, Muséum national d'Histoire naturelle, Paris, 360p.
- 76. Sémah A-M (1986) Le milieu naturel lors du premier peuplement de Java, résultats de l'analyse pollinique PhD Thesis, Université de Provence, 2 vol., 188p.
- 77. Schuster J (1911) Monographie der fossilen Flora der Pithecanthropus-Schichten Abhandlungen, 70p.
- 78. Flenley JR (1979) The equatorial rain forest: A geological history. Butterworths, London, 162p.
- 79. Rumbaugh DM (1976) Suspensory behavior, locomotion, and other behaviors of captive gibbons, Gibbon and Siamang, a series of volumes on Lesser apes, vol.4 , Karger, Basel, 316p.
- 80. Kappeler M (1984) The gibbon of Java. In DJ Chivers, H Preuschoft, WY Brockelman, N Creel, The Lesser Apes, Edinburgh University Press, Edinburgh, 19–31.
- 81. Gupta AK, Chivers DJ (2004) Biomass and use of resources in south-east Asian primate communities. In JG, Fleagle, CH, Janson, Reed, KE, Primate Communities. Cambridge University Press, Cambridge, 38–54.
- 82. Sémah AM, Sémah F, Djubiantono T, Brasseur B (2010) Landscapes and Hominids' environments: changes between the Lower and the Early Middle Pleistocene in Java (Indonesia). Quaternary International 223–224: 451–454.
- 83. Sémah AM, Sémah F (2012) The rain forest in Java through the Quaternary and its relationships with humans (adaptation, exploitation and impact on the forest). Quaternary International 249: 120–128.
- 84. Nijman V (2004) Conservation of the javan gibbon Hylobates moloch: Population estimates, local extinctions, and conservation priorities. The Raffles Bulletin of Zoology 52: 271–280.
- 85. Day MH, Molleson TI (1973) The Trinil femora. In Symposia of the Society for the Study of Human Biology Day MH, (Ed), vol.2: 127–154.
- 86. Reuter HI, Nelson A, Jarvis A (2007) (2008) An evaluation of void filling interpolation methods for SRTM data. International Journal of Geographic Information Science Jarvis A, Reuter HI, Nelson A, Guevara E, editors. Hole-filled seamless SRTM data V4, International Centre for Tropical Agriculture (CIAT), available from http://srtmcsicgiarorg), 21(9): 983–1008.